From 4b9bc370cb85492356c37c33e702d0340aecafac Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 11:57:07 -0600 Subject: [PATCH 01/16] document expose_functions method * rename internal function `expose_stan_functions` (method name still `expose_functions`). Not having identical names makes it easier to document the exposed one in the way we've been doing for other methods. * add doc page for `expose_functions` method --- R/model.R | 52 ++++++++++++++++++++++++++++++++++++++++++++++------ R/utils.R | 2 +- 2 files changed, 47 insertions(+), 7 deletions(-) diff --git a/R/model.R b/R/model.R index e95c13759..2ddaaff79 100644 --- a/R/model.R +++ b/R/model.R @@ -188,6 +188,7 @@ cmdstan_model <- function(stan_file = NULL, exe_file = NULL, compile = TRUE, ... #' [`$exe_file()`][model-method-compile] | Return the file path to the compiled executable. | #' [`$hpp_file()`][model-method-compile] | Return the file path to the `.hpp` file containing the generated C++ code. | #' [`$save_hpp_file()`][model-method-compile] | Save the `.hpp` file containing the generated C++ code. | +#' [`$expose_functions()`][model-method-expose_functions] | Expose Stan functions for use in R. | #' #' ## Model fitting #' @@ -325,10 +326,6 @@ CmdStanModel <- R6::R6Class( "- ", new_hpp_loc) private$hpp_file_ <- new_hpp_loc invisible(private$hpp_file_) - }, - expose_functions = function(global = FALSE, verbose = FALSE) { - expose_functions(self$functions, global, verbose) - invisible(NULL) } ) ) @@ -394,7 +391,10 @@ CmdStanModel <- R6::R6Class( #' (`log_prob()`, `grad_log_prob()`, `constrain_pars()`, `unconstrain_pars()`) #' @param compile_hessian_method (logical) Should the (experimental) `hessian()` method be #' be compiled with the model methods? -#' @param compile_standalone (logical) Should functions in the Stan model be compiled for used in R? +#' @param compile_standalone (logical) Should functions in the Stan model be +#' compiled for use in R? If `TRUE` the functions will be available via the +#' `functions` field in the compiled model object. +#' #' @param threads Deprecated and will be removed in a future release. Please #' turn on threading via `cpp_options = list(stan_threads = TRUE)` instead. #' @@ -584,7 +584,7 @@ compile <- function(quiet = TRUE, self$functions$hpp_code <- get_standalone_hpp(temp_stan_file, stancflags_standalone) self$functions$external <- !is.null(user_header) if (compile_standalone) { - expose_functions(self$functions, !quiet) + expose_stan_functions(self$functions, !quiet) } stancflags_val <- paste0("STANCFLAGS += ", stancflags_val, paste0(" ", stancflags_combined, collapse = " ")) withr::with_path( @@ -1749,6 +1749,46 @@ diagnose <- function(data = NULL, } CmdStanModel$set("public", name = "diagnose", value = diagnose) +#' Expose Stan functions to R +#' +#' @name model-method-expose_functions +#' @aliases expose_functions +#' @family CmdStanModel methods +#' +#' @description The `$expose_functions()` method of a [`CmdStanModel`] object +#' will compile the functions in the Stan program's `functions` block and +#' expose them for use in \R. This can also be specified via the +#' `compile_standalone` argument to the [`$compile()`][model-method-compile] +#' method. +#' @param global (logical) Should the functions be added to the Global +#' Environment? The default is `FALSE`, in which case the functions are +#' available via the `functions` field of the [CmdStanModel] object. +#' @param verbose (logical) Should detailed information about generated code be +#' printed to the console? Defaults to `FALSE`. +#' @template seealso-docs +#' @examples +#' \dontrun{ +#' stan_file <- write_stan_file( +#' " +#' functions { +#' real a_plus_b(real a, real b) { +#' return a + b; +#' } +#' } +#' " +#' ) +#' mod <- cmdstan_model(stan_file) +#' mod$expose_functions() +#' mod$functions$a_plus_b(1, 2) +#' } +#' +#' +expose_functions = function(global = FALSE, verbose = FALSE) { + expose_stan_functions(self$functions, global, verbose) + invisible(NULL) +} +CmdStanModel$set("public", name = "expose_functions", value = expose_functions) + # internal ---------------------------------------------------------------- diff --git a/R/utils.R b/R/utils.R index 3eef2590d..937d48021 100644 --- a/R/utils.R +++ b/R/utils.R @@ -907,7 +907,7 @@ compile_functions <- function(env, verbose = FALSE, global = FALSE) { invisible(NULL) } -expose_functions <- function(function_env, global = FALSE, verbose = FALSE) { +expose_stan_functions <- function(function_env, global = FALSE, verbose = FALSE) { if (os_is_wsl()) { stop("Standalone functions are not currently available with ", "WSL CmdStan and will not be compiled", From 711b5799a8335dd232091043575905c730b0b87a Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 14:04:54 -0600 Subject: [PATCH 02/16] add to doc for new methods --- DESCRIPTION | 4 +-- R/fit.R | 98 ++++++++++++++++++++++++++++++++++++++++++++++++----- R/model.R | 28 ++++++++++++--- 3 files changed, 115 insertions(+), 15 deletions(-) diff --git a/DESCRIPTION b/DESCRIPTION index 46a5814db..b656d0d98 100644 --- a/DESCRIPTION +++ b/DESCRIPTION @@ -1,7 +1,7 @@ Package: cmdstanr Title: R Interface to 'CmdStan' -Version: 0.5.3 -Date: 2022-04-24 +Version: 0.6.0 +Date: 2023-07-25 Authors@R: c(person(given = "Jonah", family = "Gabry", role = c("aut", "cre"), email = "jsg2201@columbia.edu"), diff --git a/R/fit.R b/R/fit.R index 0bffa1cc5..a8afd2376 100644 --- a/R/fit.R +++ b/R/fit.R @@ -82,7 +82,7 @@ CmdStanFit <- R6::R6Class( invisible(self) }, expose_functions = function(global = FALSE, verbose = FALSE) { - expose_functions(self$functions, global, verbose) + expose_stan_functions(self$functions, global, verbose) invisible(NULL) } ), @@ -301,12 +301,18 @@ init <- function() { CmdStanFit$set("public", name = "init", value = init) #' Compile additional methods for accessing the model log-probability function -#' and parameter constraining and unconstraining. This requires the `Rcpp` package. +#' and parameter constraining and unconstraining. #' #' @name fit-method-init_model_methods #' @aliases init_model_methods -#' @description The `$init_model_methods()` compiles and initializes the -#' `log_prob`, `grad_log_prob`, `constrain_variables`, and `unconstrain_variables` functions. +#' +#' @description The `$init_model_methods()` method compiles and initializes the +#' `log_prob`, `grad_log_prob`, `constrain_variables`, `unconstrain_variables` +#' and `unconstrain_draws` functions. These are then available as methods of +#' the fitted model object. This requires the `Rcpp` package. +#' +#' Note: there may be many compiler warnings emitted during compilation but +#' these can be ignored so long as they are warnings and not errors. #' #' @param seed (integer) The random seed to use when initializing the model. #' @param verbose (boolean) Whether to show verbose logging during compilation. @@ -317,6 +323,9 @@ CmdStanFit$set("public", name = "init", value = init) #' fit_mcmc <- cmdstanr_example("logistic", method = "sample") #' fit_mcmc$init_model_methods() #' } +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] #' init_model_methods <- function(seed = 0, verbose = FALSE, hessian = FALSE) { if (os_is_wsl()) { @@ -358,9 +367,13 @@ CmdStanFit$set("public", name = "init_model_methods", value = init_model_methods #' \dontrun{ #' fit_mcmc <- cmdstanr_example("logistic", method = "sample") #' fit_mcmc$init_model_methods() -#' fit_mcmc$log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +#' fit_mcmc$log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' log_prob <- function(unconstrained_variables, jacobian_adjustment = TRUE) { if (is.null(private$model_methods_env_$model_ptr)) { stop("The method has not been compiled, please call `init_model_methods()` first", @@ -392,9 +405,13 @@ CmdStanFit$set("public", name = "log_prob", value = log_prob) #' \dontrun{ #' fit_mcmc <- cmdstanr_example("logistic", method = "sample") #' fit_mcmc$init_model_methods() -#' fit_mcmc$grad_log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +#' fit_mcmc$grad_log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' grad_log_prob <- function(unconstrained_variables, jacobian_adjustment = TRUE) { if (is.null(private$model_methods_env_$model_ptr)) { stop("The method has not been compiled, please call `init_model_methods()` first", @@ -425,10 +442,14 @@ CmdStanFit$set("public", name = "grad_log_prob", value = grad_log_prob) #' @examples #' \dontrun{ #' fit_mcmc <- cmdstanr_example("logistic", method = "sample") -#' fit_mcmc$init_model_methods() -#' fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +#' fit_mcmc$init_model_methods(hessian = TRUE) +#' fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' hessian <- function(unconstrained_variables, jacobian_adjustment = TRUE) { if (is.null(private$model_methods_env_$model_ptr)) { stop("The method has not been compiled, please call `init_model_methods()` first", @@ -460,6 +481,10 @@ CmdStanFit$set("public", name = "hessian", value = hessian) #' fit_mcmc$unconstrain_variables(list(alpha = 0.5, beta = c(0.7, 1.1, 0.2))) #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' unconstrain_variables <- function(variables) { if (is.null(private$model_methods_env_$model_ptr)) { stop("The method has not been compiled, please call `init_model_methods()` first", @@ -521,6 +546,10 @@ CmdStanFit$set("public", name = "unconstrain_variables", value = unconstrain_var #' unconstrained_draws <- fit_mcmc$unconstrain_draws(draws = fit_mcmc$draws()) #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' unconstrain_draws <- function(files = NULL, draws = NULL) { if (!is.null(files) || !is.null(draws)) { if (!is.null(files) && !is.null(draws)) { @@ -565,6 +594,7 @@ unconstrain_draws <- function(files = NULL, draws = NULL) { self$unconstrain_variables(variables = par_list) }) }) + unconstrained } CmdStanFit$set("public", name = "unconstrain_draws", value = unconstrain_draws) @@ -587,6 +617,10 @@ CmdStanFit$set("public", name = "unconstrain_draws", value = unconstrain_draws) #' fit_mcmc$variable_skeleton() #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' variable_skeleton <- function(transformed_parameters = TRUE, generated_quantities = TRUE) { if (is.null(private$model_methods_env_$model_ptr)) { stop("The method has not been compiled, please call `init_model_methods()` first", @@ -617,9 +651,13 @@ CmdStanFit$set("public", name = "variable_skeleton", value = variable_skeleton) #' \dontrun{ #' fit_mcmc <- cmdstanr_example("logistic", method = "sample") #' fit_mcmc$init_model_methods() -#' fit_mcmc$constrain_variables(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +#' fit_mcmc$constrain_variables(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) #' } #' +#' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], +#' [unconstrain_variables()], [unconstrain_draws()], [variable_skeleton()], +#' [hessian()] +#' constrain_variables <- function(unconstrained_variables, transformed_parameters = TRUE, generated_quantities = TRUE) { if (is.null(private$model_methods_env_$model_ptr)) { @@ -1233,6 +1271,20 @@ CmdStanFit$set("public", name = "code", value = code) #' [`$time()`][fit-method-time] | Report total and chain-specific run times. | #' [`$return_codes()`][fit-method-return_codes] | Return the return codes from the CmdStan runs. | #' +#' ## Expose Stan functions and additional methods to R +#' +#' |**Method**|**Description**| +#' |:----------|:---------------| +#' [`$expose_functions()`][fit-method-expose_functions] | Expose Stan functions for use in R. | +#' [`$init_model_methods()`][fit-method-init_model_methods] | Expose methods for log-probability, gradients, parameter constraining and unconstraining. | +#' [`$log_prob()`][fit-method-log_prob] | Calculate log-prob. | +#' [`$grad_log_prob()`][fit-method-grad_log_prob] | Calculate log-prob and gradient. | +#' [`$hessian()`][fit-method-hessian] | Calculate log-prob, gradient, and hessian. | +#' [`$constrain_variables()`][fit-method-constrain_variables] | Transform a set of unconstrained parameter values to the constrained scale. | +#' [`$unconstrain_variables()`][fit-method-unconstrain_variables] | Transform a set of parameter values to the unconstrained scale. | +#' [`$unconstrain_draws()`][fit-method-unconstrain_draws] | Transform all parameter draws to the unconstrained scale. | +#' [`$variable_skeleton()`][fit-method-variable_skeleton] | Helper function to re-structure a vector of constrained parameter values. | +#' CmdStanMCMC <- R6::R6Class( classname = "CmdStanMCMC", inherit = CmdStanFit, @@ -1718,6 +1770,20 @@ CmdStanMCMC$set("public", name = "num_chains", value = num_chains) #' [`$output()`][fit-method-output] | Pretty print the output that was printed to the console. | #' [`$return_codes()`][fit-method-return_codes] | Return the return codes from the CmdStan runs. | #' +#' ## Expose Stan functions and additional methods to R +#' +#' |**Method**|**Description**| +#' |:----------|:---------------| +#' [`$expose_functions()`][fit-method-expose_functions] | Expose Stan functions for use in R. | +#' [`$init_model_methods()`][fit-method-init_model_methods] | Expose methods for log-probability, gradients, parameter constraining and unconstraining. | +#' [`$log_prob()`][fit-method-log_prob] | Calculate log-prob. | +#' [`$grad_log_prob()`][fit-method-grad_log_prob] | Calculate log-prob and gradient. | +#' [`$hessian()`][fit-method-hessian] | Calculate log-prob, gradient, and hessian. | +#' [`$constrain_variables()`][fit-method-constrain_variables] | Transform a set of unconstrained parameter values to the constrained scale. | +#' [`$unconstrain_variables()`][fit-method-unconstrain_variables] | Transform a set of parameter values to the unconstrained scale. | +#' [`$unconstrain_draws()`][fit-method-unconstrain_draws] | Transform all parameter draws to the unconstrained scale. | +#' [`$variable_skeleton()`][fit-method-variable_skeleton] | Helper function to re-structure a vector of constrained parameter values. | +#' CmdStanMLE <- R6::R6Class( classname = "CmdStanMLE", inherit = CmdStanFit, @@ -1821,6 +1887,20 @@ CmdStanMLE$set("public", name = "mle", value = mle) #' [`$output()`][fit-method-output] | Pretty print the output that was printed to the console. | #' [`$return_codes()`][fit-method-return_codes] | Return the return codes from the CmdStan runs. | #' +#' ## Expose Stan functions and additional methods to R +#' +#' |**Method**|**Description**| +#' |:----------|:---------------| +#' [`$expose_functions()`][fit-method-expose_functions] | Expose Stan functions for use in R. | +#' [`$init_model_methods()`][fit-method-init_model_methods] | Expose methods for log-probability, gradients, parameter constraining and unconstraining. | +#' [`$log_prob()`][fit-method-log_prob] | Calculate log-prob. | +#' [`$grad_log_prob()`][fit-method-grad_log_prob] | Calculate log-prob and gradient. | +#' [`$hessian()`][fit-method-hessian] | Calculate log-prob, gradient, and hessian. | +#' [`$constrain_variables()`][fit-method-constrain_variables] | Transform a set of unconstrained parameter values to the constrained scale. | +#' [`$unconstrain_variables()`][fit-method-unconstrain_variables] | Transform a set of parameter values to the unconstrained scale. | +#' [`$unconstrain_draws()`][fit-method-unconstrain_draws] | Transform all parameter draws to the unconstrained scale. | +#' [`$variable_skeleton()`][fit-method-variable_skeleton] | Helper function to re-structure a vector of constrained parameter values. | +#' CmdStanVB <- R6::R6Class( classname = "CmdStanVB", inherit = CmdStanFit, diff --git a/R/model.R b/R/model.R index 2ddaaff79..f32f7e424 100644 --- a/R/model.R +++ b/R/model.R @@ -388,12 +388,15 @@ CmdStanModel <- R6::R6Class( #' not modified since last compiled. The default is `FALSE`. Can also be set #' via a global `cmdstanr_force_recompile` option. #' @param compile_model_methods (logical) Compile additional model methods -#' (`log_prob()`, `grad_log_prob()`, `constrain_pars()`, `unconstrain_pars()`) +#' (`log_prob()`, `grad_log_prob()`, `constrain_variables()`, +#' `unconstrain_variables()`). #' @param compile_hessian_method (logical) Should the (experimental) `hessian()` method be #' be compiled with the model methods? #' @param compile_standalone (logical) Should functions in the Stan model be #' compiled for use in R? If `TRUE` the functions will be available via the -#' `functions` field in the compiled model object. +#' `functions` field in the compiled model object. This can also be done after +#' compilation using the +#' [`$expose_functions()`][model-method-expose_functions] method. #' #' @param threads Deprecated and will be removed in a future release. Please #' turn on threading via `cpp_options = list(stan_threads = TRUE)` instead. @@ -1752,7 +1755,7 @@ CmdStanModel$set("public", name = "diagnose", value = diagnose) #' Expose Stan functions to R #' #' @name model-method-expose_functions -#' @aliases expose_functions +#' @aliases expose_functions fit-method-expose_functions #' @family CmdStanModel methods #' #' @description The `$expose_functions()` method of a [`CmdStanModel`] object @@ -1760,9 +1763,16 @@ CmdStanModel$set("public", name = "diagnose", value = diagnose) #' expose them for use in \R. This can also be specified via the #' `compile_standalone` argument to the [`$compile()`][model-method-compile] #' method. +#' +#' This method is also available for fitted model objects ([`CmdStanMCMC`], [`CmdStanVB`], etc.). +#' See **Examples**. +#' +#' Note: there may be many compiler warnings emitted during compilation but +#' these can be ignored so long as they are warnings and not errors. +#' #' @param global (logical) Should the functions be added to the Global #' Environment? The default is `FALSE`, in which case the functions are -#' available via the `functions` field of the [CmdStanModel] object. +#' available via the `functions` field of the R6 object. #' @param verbose (logical) Should detailed information about generated code be #' printed to the console? Defaults to `FALSE`. #' @template seealso-docs @@ -1775,11 +1785,21 @@ CmdStanModel$set("public", name = "diagnose", value = diagnose) #' return a + b; #' } #' } +#' parameters { +#' real x; +#' } +#' model { +#' x ~ std_normal(); +#' } #' " #' ) #' mod <- cmdstan_model(stan_file) #' mod$expose_functions() #' mod$functions$a_plus_b(1, 2) +#' +#' fit <- mod$sample(refresh = 0) +#' fit$expose_functions() # already compiled because of above but this would compile them otherwise +#' fit$functions$a_plus_b(1, 2) #' } #' #' From 88097c644a9871af161ec55379425245cdfc34be Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 14:08:19 -0600 Subject: [PATCH 03/16] generate rd files for new doc --- man/CmdStanMCMC.Rd | 15 +++++ man/CmdStanMLE.Rd | 15 +++++ man/CmdStanModel.Rd | 1 + man/CmdStanVB.Rd | 15 +++++ man/fit-method-constrain_variables.Rd | 7 ++- man/fit-method-grad_log_prob.Rd | 7 ++- man/fit-method-hessian.Rd | 9 ++- man/fit-method-init_model_methods.Rd | 17 +++-- man/fit-method-log_prob.Rd | 7 ++- man/fit-method-unconstrain_draws.Rd | 5 ++ man/fit-method-unconstrain_variables.Rd | 5 ++ man/fit-method-variable_skeleton.Rd | 5 ++ man/model-method-check_syntax.Rd | 1 + man/model-method-compile.Rd | 10 ++- man/model-method-diagnose.Rd | 1 + man/model-method-expose_functions.Rd | 83 +++++++++++++++++++++++++ man/model-method-format.Rd | 1 + man/model-method-generate-quantities.Rd | 1 + man/model-method-optimize.Rd | 1 + man/model-method-sample.Rd | 1 + man/model-method-sample_mpi.Rd | 1 + man/model-method-variables.Rd | 1 + man/model-method-variational.Rd | 1 + 23 files changed, 199 insertions(+), 11 deletions(-) create mode 100644 man/model-method-expose_functions.Rd diff --git a/man/CmdStanMCMC.Rd b/man/CmdStanMCMC.Rd index 98b7f4374..218ee49c3 100644 --- a/man/CmdStanMCMC.Rd +++ b/man/CmdStanMCMC.Rd @@ -55,6 +55,21 @@ methods, all of which have their own (linked) documentation pages. \code{\link[=fit-method-return_codes]{$return_codes()}} \tab Return the return codes from the CmdStan runs. \cr } +} + +\subsection{Expose Stan functions and additional methods to R}{\tabular{ll}{ + \strong{Method} \tab \strong{Description} \cr + \code{\link[=fit-method-expose_functions]{$expose_functions()}} \tab Expose Stan functions for use in R. \cr + \code{\link[=fit-method-init_model_methods]{$init_model_methods()}} \tab Expose methods for log-probability, gradients, parameter constraining and unconstraining. \cr + \code{\link[=fit-method-log_prob]{$log_prob()}} \tab Calculate log-prob. \cr + \code{\link[=fit-method-grad_log_prob]{$grad_log_prob()}} \tab Calculate log-prob and gradient. \cr + \code{\link[=fit-method-hessian]{$hessian()}} \tab Calculate log-prob, gradient, and hessian. \cr + \code{\link[=fit-method-constrain_variables]{$constrain_variables()}} \tab Transform a set of unconstrained parameter values to the constrained scale. \cr + \code{\link[=fit-method-unconstrain_variables]{$unconstrain_variables()}} \tab Transform a set of parameter values to the unconstrained scale. \cr + \code{\link[=fit-method-unconstrain_draws]{$unconstrain_draws()}} \tab Transform all parameter draws to the unconstrained scale. \cr + \code{\link[=fit-method-variable_skeleton]{$variable_skeleton()}} \tab Helper function to re-structure a vector of constrained parameter values. \cr +} + } } diff --git a/man/CmdStanMLE.Rd b/man/CmdStanMLE.Rd index c5044acc7..01acae4d9 100644 --- a/man/CmdStanMLE.Rd +++ b/man/CmdStanMLE.Rd @@ -45,6 +45,21 @@ all of which have their own (linked) documentation pages. \code{\link[=fit-method-return_codes]{$return_codes()}} \tab Return the return codes from the CmdStan runs. \cr } +} + +\subsection{Expose Stan functions and additional methods to R}{\tabular{ll}{ + \strong{Method} \tab \strong{Description} \cr + \code{\link[=fit-method-expose_functions]{$expose_functions()}} \tab Expose Stan functions for use in R. \cr + \code{\link[=fit-method-init_model_methods]{$init_model_methods()}} \tab Expose methods for log-probability, gradients, parameter constraining and unconstraining. \cr + \code{\link[=fit-method-log_prob]{$log_prob()}} \tab Calculate log-prob. \cr + \code{\link[=fit-method-grad_log_prob]{$grad_log_prob()}} \tab Calculate log-prob and gradient. \cr + \code{\link[=fit-method-hessian]{$hessian()}} \tab Calculate log-prob, gradient, and hessian. \cr + \code{\link[=fit-method-constrain_variables]{$constrain_variables()}} \tab Transform a set of unconstrained parameter values to the constrained scale. \cr + \code{\link[=fit-method-unconstrain_variables]{$unconstrain_variables()}} \tab Transform a set of parameter values to the unconstrained scale. \cr + \code{\link[=fit-method-unconstrain_draws]{$unconstrain_draws()}} \tab Transform all parameter draws to the unconstrained scale. \cr + \code{\link[=fit-method-variable_skeleton]{$variable_skeleton()}} \tab Helper function to re-structure a vector of constrained parameter values. \cr +} + } } diff --git a/man/CmdStanModel.Rd b/man/CmdStanModel.Rd index fbe191f82..94ec10d61 100644 --- a/man/CmdStanModel.Rd +++ b/man/CmdStanModel.Rd @@ -29,6 +29,7 @@ methods, many of which have their own (linked) documentation pages: \code{\link[=model-method-compile]{$exe_file()}} \tab Return the file path to the compiled executable. \cr \code{\link[=model-method-compile]{$hpp_file()}} \tab Return the file path to the \code{.hpp} file containing the generated C++ code. \cr \code{\link[=model-method-compile]{$save_hpp_file()}} \tab Save the \code{.hpp} file containing the generated C++ code. \cr + \code{\link[=model-method-expose_functions]{$expose_functions()}} \tab Expose Stan functions for use in R. \cr } } diff --git a/man/CmdStanVB.Rd b/man/CmdStanVB.Rd index b11a361f4..4b4d53ada 100644 --- a/man/CmdStanVB.Rd +++ b/man/CmdStanVB.Rd @@ -48,6 +48,21 @@ all of which have their own (linked) documentation pages. \code{\link[=fit-method-return_codes]{$return_codes()}} \tab Return the return codes from the CmdStan runs. \cr } +} + +\subsection{Expose Stan functions and additional methods to R}{\tabular{ll}{ + \strong{Method} \tab \strong{Description} \cr + \code{\link[=fit-method-expose_functions]{$expose_functions()}} \tab Expose Stan functions for use in R. \cr + \code{\link[=fit-method-init_model_methods]{$init_model_methods()}} \tab Expose methods for log-probability, gradients, parameter constraining and unconstraining. \cr + \code{\link[=fit-method-log_prob]{$log_prob()}} \tab Calculate log-prob. \cr + \code{\link[=fit-method-grad_log_prob]{$grad_log_prob()}} \tab Calculate log-prob and gradient. \cr + \code{\link[=fit-method-hessian]{$hessian()}} \tab Calculate log-prob, gradient, and hessian. \cr + \code{\link[=fit-method-constrain_variables]{$constrain_variables()}} \tab Transform a set of unconstrained parameter values to the constrained scale. \cr + \code{\link[=fit-method-unconstrain_variables]{$unconstrain_variables()}} \tab Transform a set of parameter values to the unconstrained scale. \cr + \code{\link[=fit-method-unconstrain_draws]{$unconstrain_draws()}} \tab Transform all parameter draws to the unconstrained scale. \cr + \code{\link[=fit-method-variable_skeleton]{$variable_skeleton()}} \tab Helper function to re-structure a vector of constrained parameter values. \cr +} + } } diff --git a/man/fit-method-constrain_variables.Rd b/man/fit-method-constrain_variables.Rd index 31cdbe9db..cf3e8c2ee 100644 --- a/man/fit-method-constrain_variables.Rd +++ b/man/fit-method-constrain_variables.Rd @@ -28,7 +28,12 @@ the constrained scale \dontrun{ fit_mcmc <- cmdstanr_example("logistic", method = "sample") fit_mcmc$init_model_methods() -fit_mcmc$constrain_variables(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +fit_mcmc$constrain_variables(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/fit-method-grad_log_prob.Rd b/man/fit-method-grad_log_prob.Rd index e0a58d487..f7c8819f8 100644 --- a/man/fit-method-grad_log_prob.Rd +++ b/man/fit-method-grad_log_prob.Rd @@ -23,7 +23,12 @@ Stan model's \code{log_prob} function and its derivative \dontrun{ fit_mcmc <- cmdstanr_example("logistic", method = "sample") fit_mcmc$init_model_methods() -fit_mcmc$grad_log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +fit_mcmc$grad_log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/fit-method-hessian.Rd b/man/fit-method-hessian.Rd index 3674e926b..0fb4bb0f9 100644 --- a/man/fit-method-hessian.Rd +++ b/man/fit-method-hessian.Rd @@ -22,8 +22,13 @@ Stan model's \code{log_prob}, its derivative, and its hessian \examples{ \dontrun{ fit_mcmc <- cmdstanr_example("logistic", method = "sample") -fit_mcmc$init_model_methods() -fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +fit_mcmc$init_model_methods(hessian = TRUE) +fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/fit-method-init_model_methods.Rd b/man/fit-method-init_model_methods.Rd index 6561b0cde..165e804a6 100644 --- a/man/fit-method-init_model_methods.Rd +++ b/man/fit-method-init_model_methods.Rd @@ -4,7 +4,7 @@ \alias{fit-method-init_model_methods} \alias{init_model_methods} \title{Compile additional methods for accessing the model log-probability function -and parameter constraining and unconstraining. This requires the \code{Rcpp} package.} +and parameter constraining and unconstraining.} \usage{ init_model_methods(seed = 0, verbose = FALSE, hessian = FALSE) } @@ -16,13 +16,22 @@ init_model_methods(seed = 0, verbose = FALSE, hessian = FALSE) \item{hessian}{(boolean) Whether to expose the (experimental) hessian method.} } \description{ -The \verb{$init_model_methods()} compiles and initializes the -\code{log_prob}, \code{grad_log_prob}, \code{constrain_variables}, and \code{unconstrain_variables} functions. +The \verb{$init_model_methods()} method compiles and initializes the +\code{log_prob}, \code{grad_log_prob}, \code{constrain_variables}, \code{unconstrain_variables} +and \code{unconstrain_draws} functions. These are then available as methods of +the fitted model object. This requires the \code{Rcpp} package. + +Note: there may be many compiler warnings emitted during compilation but +these can be ignored so long as they are warnings and not errors. } \examples{ \dontrun{ fit_mcmc <- cmdstanr_example("logistic", method = "sample") fit_mcmc$init_model_methods() } - +} +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} } diff --git a/man/fit-method-log_prob.Rd b/man/fit-method-log_prob.Rd index cbdbb93a0..c35e48431 100644 --- a/man/fit-method-log_prob.Rd +++ b/man/fit-method-log_prob.Rd @@ -20,7 +20,12 @@ The \verb{$log_prob()} method provides access to the Stan model's \code{log_prob \dontrun{ fit_mcmc <- cmdstanr_example("logistic", method = "sample") fit_mcmc$init_model_methods() -fit_mcmc$log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2, 1.1)) +fit_mcmc$log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/fit-method-unconstrain_draws.Rd b/man/fit-method-unconstrain_draws.Rd index 947e47fe2..6764ecb53 100644 --- a/man/fit-method-unconstrain_draws.Rd +++ b/man/fit-method-unconstrain_draws.Rd @@ -36,3 +36,8 @@ unconstrained_draws <- fit_mcmc$unconstrain_draws(draws = fit_mcmc$draws()) } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/fit-method-unconstrain_variables.Rd b/man/fit-method-unconstrain_variables.Rd index 472c1488f..04bfc78e4 100644 --- a/man/fit-method-unconstrain_variables.Rd +++ b/man/fit-method-unconstrain_variables.Rd @@ -23,3 +23,8 @@ fit_mcmc$unconstrain_variables(list(alpha = 0.5, beta = c(0.7, 1.1, 0.2))) } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/fit-method-variable_skeleton.Rd b/man/fit-method-variable_skeleton.Rd index 859dab332..2116d0ad2 100644 --- a/man/fit-method-variable_skeleton.Rd +++ b/man/fit-method-variable_skeleton.Rd @@ -26,3 +26,8 @@ fit_mcmc$variable_skeleton() } } +\seealso{ +\code{\link[=log_prob]{log_prob()}}, \code{\link[=grad_log_prob]{grad_log_prob()}}, \code{\link[=constrain_variables]{constrain_variables()}}, +\code{\link[=unconstrain_variables]{unconstrain_variables()}}, \code{\link[=unconstrain_draws]{unconstrain_draws()}}, \code{\link[=variable_skeleton]{variable_skeleton()}}, +\code{\link[=hessian]{hessian()}} +} diff --git a/man/model-method-check_syntax.Rd b/man/model-method-check_syntax.Rd index 885061116..9193a684d 100644 --- a/man/model-method-check_syntax.Rd +++ b/man/model-method-check_syntax.Rd @@ -80,6 +80,7 @@ The Stan and CmdStan documentation: Other CmdStanModel methods: \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, diff --git a/man/model-method-compile.Rd b/man/model-method-compile.Rd index 9ff75f473..b4d7b35f8 100644 --- a/man/model-method-compile.Rd +++ b/man/model-method-compile.Rd @@ -60,12 +60,17 @@ not modified since last compiled. The default is \code{FALSE}. Can also be set via a global \code{cmdstanr_force_recompile} option.} \item{compile_model_methods}{(logical) Compile additional model methods -(\code{log_prob()}, \code{grad_log_prob()}, \code{constrain_pars()}, \code{unconstrain_pars()})} +(\code{log_prob()}, \code{grad_log_prob()}, \code{constrain_variables()}, +\code{unconstrain_variables()}).} \item{compile_hessian_method}{(logical) Should the (experimental) \code{hessian()} method be be compiled with the model methods?} -\item{compile_standalone}{(logical) Should functions in the Stan model be compiled for used in R?} +\item{compile_standalone}{(logical) Should functions in the Stan model be +compiled for use in R? If \code{TRUE} the functions will be available via the +\code{functions} field in the compiled model object. This can also be done after +compilation using the +\code{\link[=model-method-expose_functions]{$expose_functions()}} method.} \item{threads}{Deprecated and will be removed in a future release. Please turn on threading via \code{cpp_options = list(stan_threads = TRUE)} instead.} @@ -142,6 +147,7 @@ The Stan and CmdStan documentation: Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, diff --git a/man/model-method-diagnose.Rd b/man/model-method-diagnose.Rd index 371b71619..7f9cde7d4 100644 --- a/man/model-method-diagnose.Rd +++ b/man/model-method-diagnose.Rd @@ -122,6 +122,7 @@ The Stan and CmdStan documentation: Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, diff --git a/man/model-method-expose_functions.Rd b/man/model-method-expose_functions.Rd new file mode 100644 index 000000000..c01aa3c1a --- /dev/null +++ b/man/model-method-expose_functions.Rd @@ -0,0 +1,83 @@ +% Generated by roxygen2: do not edit by hand +% Please edit documentation in R/model.R +\name{model-method-expose_functions} +\alias{model-method-expose_functions} +\alias{expose_functions} +\alias{fit-method-expose_functions} +\title{Expose Stan functions to R} +\usage{ +expose_functions(global = FALSE, verbose = FALSE) +} +\arguments{ +\item{global}{(logical) Should the functions be added to the Global +Environment? The default is \code{FALSE}, in which case the functions are +available via the \code{functions} field of the R6 object.} + +\item{verbose}{(logical) Should detailed information about generated code be +printed to the console? Defaults to \code{FALSE}.} +} +\description{ +The \verb{$expose_functions()} method of a \code{\link{CmdStanModel}} object +will compile the functions in the Stan program's \code{functions} block and +expose them for use in \R. This can also be specified via the +\code{compile_standalone} argument to the \code{\link[=model-method-compile]{$compile()}} +method. + +This method is also available for fitted model objects (\code{\link{CmdStanMCMC}}, \code{\link{CmdStanVB}}, etc.). +See \strong{Examples}. + +Note: there may be many compiler warnings emitted during compilation but +these can be ignored so long as they are warnings and not errors. +} +\examples{ +\dontrun{ +stan_file <- write_stan_file( + " + functions { + real a_plus_b(real a, real b) { + return a + b; + } + } + parameters { + real x; + } + model { + x ~ std_normal(); + } + " +) +mod <- cmdstan_model(stan_file) +mod$expose_functions() +mod$functions$a_plus_b(1, 2) + +fit <- mod$sample(refresh = 0) +fit$expose_functions() # already compiled because of above but this would compile them otherwise +fit$functions$a_plus_b(1, 2) +} + + +} +\seealso{ +The CmdStanR website +(\href{https://mc-stan.org/cmdstanr/}{mc-stan.org/cmdstanr}) for online +documentation and tutorials. + +The Stan and CmdStan documentation: +\itemize{ +\item Stan documentation: \href{https://mc-stan.org/users/documentation/}{mc-stan.org/users/documentation} +\item CmdStan User’s Guide: \href{https://mc-stan.org/docs/cmdstan-guide/}{mc-stan.org/docs/cmdstan-guide} +} + +Other CmdStanModel methods: +\code{\link{model-method-check_syntax}}, +\code{\link{model-method-compile}}, +\code{\link{model-method-diagnose}}, +\code{\link{model-method-format}}, +\code{\link{model-method-generate-quantities}}, +\code{\link{model-method-optimize}}, +\code{\link{model-method-sample_mpi}}, +\code{\link{model-method-sample}}, +\code{\link{model-method-variables}}, +\code{\link{model-method-variational}} +} +\concept{CmdStanModel methods} diff --git a/man/model-method-format.Rd b/man/model-method-format.Rd index 777e7da01..1d7ed11bd 100644 --- a/man/model-method-format.Rd +++ b/man/model-method-format.Rd @@ -79,6 +79,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, \code{\link{model-method-sample_mpi}}, diff --git a/man/model-method-generate-quantities.Rd b/man/model-method-generate-quantities.Rd index da75ac5f3..5a80bb736 100644 --- a/man/model-method-generate-quantities.Rd +++ b/man/model-method-generate-quantities.Rd @@ -172,6 +172,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-optimize}}, \code{\link{model-method-sample_mpi}}, diff --git a/man/model-method-optimize.Rd b/man/model-method-optimize.Rd index a8735e7fe..42f81d07a 100644 --- a/man/model-method-optimize.Rd +++ b/man/model-method-optimize.Rd @@ -300,6 +300,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-sample_mpi}}, diff --git a/man/model-method-sample.Rd b/man/model-method-sample.Rd index 2c6b693db..e0955abb4 100644 --- a/man/model-method-sample.Rd +++ b/man/model-method-sample.Rd @@ -412,6 +412,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, diff --git a/man/model-method-sample_mpi.Rd b/man/model-method-sample_mpi.Rd index a99b96952..89981273f 100644 --- a/man/model-method-sample_mpi.Rd +++ b/man/model-method-sample_mpi.Rd @@ -311,6 +311,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, diff --git a/man/model-method-variables.Rd b/man/model-method-variables.Rd index f2fb296f5..aa609ddda 100644 --- a/man/model-method-variables.Rd +++ b/man/model-method-variables.Rd @@ -40,6 +40,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, diff --git a/man/model-method-variational.Rd b/man/model-method-variational.Rd index a64a5a232..21c92fe24 100644 --- a/man/model-method-variational.Rd +++ b/man/model-method-variational.Rd @@ -304,6 +304,7 @@ Other CmdStanModel methods: \code{\link{model-method-check_syntax}}, \code{\link{model-method-compile}}, \code{\link{model-method-diagnose}}, +\code{\link{model-method-expose_functions}}, \code{\link{model-method-format}}, \code{\link{model-method-generate-quantities}}, \code{\link{model-method-optimize}}, From a1e3790041b82ca883d0e07d2ab114c1a88a7af6 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 14:23:20 -0600 Subject: [PATCH 04/16] update NEWS --- NEWS.md | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) diff --git a/NEWS.md b/NEWS.md index 495c4d477..aaf69021e 100644 --- a/NEWS.md +++ b/NEWS.md @@ -1,3 +1,37 @@ +# cmdstanr 0.6.0 + +### Major new features + +* New `expose_functions()` method to expose Stan functions to R by @andrjohns in #702. See `?expose_functions`. +* New methods for accessing log_prob, grad_log_prob, hessian, un/constrain variables by @andrjohns in #701. See `?init_model_methods`. + +## Other changes + +* mod$variables works w includes in precompile state (fix #680) by @MKyhos in #682 +* Update broken link for Stan OpenCL support page by @erictleung in #686 +* Add newline to check syntax output by @rok-cesnovar in #689 +* Allow exposing functions without sampling by @andrjohns in #705 +* Expose skeleton by @andrjohns in #706 +* WSL - Run cmdstan and models under WSL filesystem by @andrjohns in #696 +* Bugfix - Deep copy method/function environments by @andrjohns in #709 +* Add option for including jacobian adjustments in hessian method by @andrjohns in #710 +* WSL Optimisations and Bugfixes for CI by @andrjohns in #711 +* add stancflags from make/local by @rok-cesnovar in #690 +* Update co-authors by @andrjohns in #715 +* Update model methods parameter naming and extract skeleton function by @andrjohns in #724 +* Add method for unconstraining all parameter draws by @andrjohns in #729 +* Improve efficiency of variable matching by @sbfnk in #736 +* Add verbosity to download output and errors by @andrjohns in #745 +* Update handling of show_messages, add show_exceptions by @andrjohns in #746 +* Rtools43 support by @andrjohns in #755 +* Add stanc M1 make patch, suppress boost warnings by @andrjohns in #756 +* more examples of summary method by @gravesti in #751 +* Fix model$format and model$check_syntax for compiled models with include-paths by @adrian-lison in #775 +* Generalise RTools config/support by @andrjohns in #777 +* New posterior vignette by @gravesti in #719 +* Add moment-matching support to $loo() method by @andrjohns in #778 +* replace \ with function by @jsocolar in #789 + # cmdstanr 0.5.3 ### New features From ceabf488b0e38570400cc65f7449bb06e7c429f5 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 14:23:31 -0600 Subject: [PATCH 05/16] add new vignette to _pkgdown.yml --- _pkgdown.yml | 1 + vignettes/posterior.Rmd | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/_pkgdown.yml b/_pkgdown.yml index 2f99caefa..70db6a30b 100644 --- a/_pkgdown.yml +++ b/_pkgdown.yml @@ -70,6 +70,7 @@ articles: and using CmdStanR in R Markdown documents. contents: - cmdstanr-internals + - posterior - r-markdown - deprecations - profiling diff --git a/vignettes/posterior.Rmd b/vignettes/posterior.Rmd index 3d246a6f1..fca84be29 100644 --- a/vignettes/posterior.Rmd +++ b/vignettes/posterior.Rmd @@ -81,4 +81,4 @@ fit$summary(variables = NULL, "Strictly Positive" = strict_pos) # fit$print(variables = NULL, "Strictly Positive" = strict_pos) ``` -For more information, see [posterior::summarise_draws()], which is is called by `$summary()`. +For more information, see [posterior::summarise_draws()], which is called by `$summary()`. From 18dfd3cd204952c3366698c1f16260863e900ab2 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 14:27:39 -0600 Subject: [PATCH 06/16] minor doc fixes --- R/fit.R | 35 +++++++++++++------------ man/fit-method-constrain_variables.Rd | 9 ++++--- man/fit-method-grad_log_prob.Rd | 8 +++--- man/fit-method-hessian.Rd | 14 +++++----- man/fit-method-unconstrain_variables.Rd | 4 +-- 5 files changed, 36 insertions(+), 34 deletions(-) diff --git a/R/fit.R b/R/fit.R index a8afd2376..f9dd4d8d2 100644 --- a/R/fit.R +++ b/R/fit.R @@ -396,10 +396,10 @@ CmdStanFit$set("public", name = "log_prob", value = log_prob) #' @description The `$grad_log_prob()` method provides access to the #' Stan model's `log_prob` function and its derivative #' -#' @param unconstrained_variables (numeric) A vector of unconstrained parameters to be passed -#' to `grad_log_prob` -#' @param jacobian_adjustment (bool) Whether to include the log-density adjustments from -#' un/constraining variables +#' @param unconstrained_variables (numeric) A vector of unconstrained parameters +#' to be passed to `grad_log_prob`. +#' @param jacobian_adjustment (bool) Whether to include the log-density +#' adjustments from un/constraining variables. #' #' @examples #' \dontrun{ @@ -434,16 +434,16 @@ CmdStanFit$set("public", name = "grad_log_prob", value = grad_log_prob) #' @description The `$hessian()` method provides access to the #' Stan model's `log_prob`, its derivative, and its hessian #' -#' @param unconstrained_variables (numeric) A vector of unconstrained parameters to be passed -#' to `hessian` -#' @param jacobian_adjustment (bool) Whether to include the log-density adjustments from -#' un/constraining variables +#' @param unconstrained_variables (numeric) A vector of unconstrained parameters +#' to be passed to `hessian`. +#' @param jacobian_adjustment (bool) Whether to include the log-density +#' adjustments from un/constraining variables. #' #' @examples #' \dontrun{ -#' fit_mcmc <- cmdstanr_example("logistic", method = "sample") -#' fit_mcmc$init_model_methods(hessian = TRUE) -#' fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) +#' # fit_mcmc <- cmdstanr_example("logistic", method = "sample") +#' # fit_mcmc$init_model_methods(hessian = TRUE) +#' # fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) #' } #' #' @seealso [log_prob()], [grad_log_prob()], [constrain_variables()], @@ -471,8 +471,8 @@ CmdStanFit$set("public", name = "hessian", value = hessian) #' @description The `$unconstrain_variables()` method transforms input parameters to #' the unconstrained scale #' -#' @param variables (list) A list of parameter values to transform, in the same format as -#' provided to the `init` argument of the `$sample()` method +#' @param variables (list) A list of parameter values to transform, in the same +#' format as provided to the `init` argument of the `$sample()` method. #' #' @examples #' \dontrun{ @@ -641,11 +641,12 @@ CmdStanFit$set("public", name = "variable_skeleton", value = variable_skeleton) #' @description The `$constrain_variables()` method transforms input parameters to #' the constrained scale #' -#' @param unconstrained_variables (numeric) A vector of unconstrained parameters to constrain -#' @param transformed_parameters (boolean) Whether to return transformed parameters -#' implied by newly-constrained parameters (defaults to TRUE) +#' @param unconstrained_variables (numeric) A vector of unconstrained parameters +#' to constrain. +#' @param transformed_parameters (boolean) Whether to return transformed +#' parameters implied by newly-constrained parameters (defaults to TRUE). #' @param generated_quantities (boolean) Whether to return generated quantities -#' implied by newly-constrained parameters (defaults to TRUE) +#' implied by newly-constrained parameters (defaults to TRUE). #' #' @examples #' \dontrun{ diff --git a/man/fit-method-constrain_variables.Rd b/man/fit-method-constrain_variables.Rd index cf3e8c2ee..78c5b8098 100644 --- a/man/fit-method-constrain_variables.Rd +++ b/man/fit-method-constrain_variables.Rd @@ -12,13 +12,14 @@ constrain_variables( ) } \arguments{ -\item{unconstrained_variables}{(numeric) A vector of unconstrained parameters to constrain} +\item{unconstrained_variables}{(numeric) A vector of unconstrained parameters +to constrain.} -\item{transformed_parameters}{(boolean) Whether to return transformed parameters -implied by newly-constrained parameters (defaults to TRUE)} +\item{transformed_parameters}{(boolean) Whether to return transformed +parameters implied by newly-constrained parameters (defaults to TRUE).} \item{generated_quantities}{(boolean) Whether to return generated quantities -implied by newly-constrained parameters (defaults to TRUE)} +implied by newly-constrained parameters (defaults to TRUE).} } \description{ The \verb{$constrain_variables()} method transforms input parameters to diff --git a/man/fit-method-grad_log_prob.Rd b/man/fit-method-grad_log_prob.Rd index f7c8819f8..ef42d6c3d 100644 --- a/man/fit-method-grad_log_prob.Rd +++ b/man/fit-method-grad_log_prob.Rd @@ -9,11 +9,11 @@ given vector of unconstrained parameters} grad_log_prob(unconstrained_variables, jacobian_adjustment = TRUE) } \arguments{ -\item{unconstrained_variables}{(numeric) A vector of unconstrained parameters to be passed -to \code{grad_log_prob}} +\item{unconstrained_variables}{(numeric) A vector of unconstrained parameters +to be passed to \code{grad_log_prob}.} -\item{jacobian_adjustment}{(bool) Whether to include the log-density adjustments from -un/constraining variables} +\item{jacobian_adjustment}{(bool) Whether to include the log-density +adjustments from un/constraining variables.} } \description{ The \verb{$grad_log_prob()} method provides access to the diff --git a/man/fit-method-hessian.Rd b/man/fit-method-hessian.Rd index 0fb4bb0f9..fe384c9d9 100644 --- a/man/fit-method-hessian.Rd +++ b/man/fit-method-hessian.Rd @@ -9,11 +9,11 @@ for a given vector of unconstrained parameters} hessian(unconstrained_variables, jacobian_adjustment = TRUE) } \arguments{ -\item{unconstrained_variables}{(numeric) A vector of unconstrained parameters to be passed -to \code{hessian}} +\item{unconstrained_variables}{(numeric) A vector of unconstrained parameters +to be passed to \code{hessian}.} -\item{jacobian_adjustment}{(bool) Whether to include the log-density adjustments from -un/constraining variables} +\item{jacobian_adjustment}{(bool) Whether to include the log-density +adjustments from un/constraining variables.} } \description{ The \verb{$hessian()} method provides access to the @@ -21,9 +21,9 @@ Stan model's \code{log_prob}, its derivative, and its hessian } \examples{ \dontrun{ -fit_mcmc <- cmdstanr_example("logistic", method = "sample") -fit_mcmc$init_model_methods(hessian = TRUE) -fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) +# fit_mcmc <- cmdstanr_example("logistic", method = "sample") +# fit_mcmc$init_model_methods(hessian = TRUE) +# fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2)) } } diff --git a/man/fit-method-unconstrain_variables.Rd b/man/fit-method-unconstrain_variables.Rd index 04bfc78e4..69c6e308e 100644 --- a/man/fit-method-unconstrain_variables.Rd +++ b/man/fit-method-unconstrain_variables.Rd @@ -8,8 +8,8 @@ unconstrain_variables(variables) } \arguments{ -\item{variables}{(list) A list of parameter values to transform, in the same format as -provided to the \code{init} argument of the \verb{$sample()} method} +\item{variables}{(list) A list of parameter values to transform, in the same +format as provided to the \code{init} argument of the \verb{$sample()} method.} } \description{ The \verb{$unconstrain_variables()} method transforms input parameters to From f7d72611d21aa18581deb73c0b37fb7fff21b05d Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 14:32:31 -0600 Subject: [PATCH 07/16] avoid pkgdown error about missing topics --- R/csv.R | 2 +- R/example.R | 2 +- R/utils.R | 2 +- _pkgdown.yml | 2 ++ man/read_sample_csv.Rd | 1 + man/stan_threads.Rd | 1 + man/write_stan_tempfile.Rd | 1 + 7 files changed, 8 insertions(+), 3 deletions(-) diff --git a/R/csv.R b/R/csv.R index 9ab528e1f..5d3824428 100644 --- a/R/csv.R +++ b/R/csv.R @@ -422,7 +422,7 @@ read_cmdstan_csv <- function(files, #' Read CmdStan CSV files from sampling into \R #' #' Deprecated. Use [read_cmdstan_csv()] instead. -#' +#' @keywords internal #' @export #' @param files,variables,sampler_diagnostics Deprecated. Use #' [read_cmdstan_csv()] instead. diff --git a/R/example.R b/R/example.R index 388c555ef..755f33703 100644 --- a/R/example.R +++ b/R/example.R @@ -188,7 +188,7 @@ write_stan_file <- function(code, #' Write Stan code to a temporary file #' #' This function is deprecated. Please use [write_stan_file()] instead. -#' +#' @keywords internal #' @export #' @inheritParams write_stan_file write_stan_tempfile <- function(code, dir = tempdir()) { diff --git a/R/utils.R b/R/utils.R index 937d48021..7e732c0a5 100644 --- a/R/utils.R +++ b/R/utils.R @@ -235,7 +235,7 @@ generate_file_names <- #' Set or get the number of threads used to execute Stan models #' #' DEPRECATED. Please use the `threads_per_chain` argument when fitting the model. -#' +#' @keywords internal #' @name stan_threads NULL diff --git a/_pkgdown.yml b/_pkgdown.yml index 70db6a30b..1dea923a9 100644 --- a/_pkgdown.yml +++ b/_pkgdown.yml @@ -111,6 +111,8 @@ reference: - write_stan_json - write_stan_file - draws_to_csv + - as.mcmc.list + - as_draws.CmdStanMCMC - title: "Using CmdStanR with knitr and R Markdown" contents: - register_knitr_engine diff --git a/man/read_sample_csv.Rd b/man/read_sample_csv.Rd index 76bb5463c..4c7661ae7 100644 --- a/man/read_sample_csv.Rd +++ b/man/read_sample_csv.Rd @@ -13,3 +13,4 @@ read_sample_csv(files, variables = NULL, sampler_diagnostics = NULL) \description{ Deprecated. Use \code{\link[=read_cmdstan_csv]{read_cmdstan_csv()}} instead. } +\keyword{internal} diff --git a/man/stan_threads.Rd b/man/stan_threads.Rd index e79e882ff..e8cf8470d 100644 --- a/man/stan_threads.Rd +++ b/man/stan_threads.Rd @@ -19,3 +19,4 @@ The value of the environment variable \code{STAN_NUM_THREADS}. \description{ DEPRECATED. Please use the \code{threads_per_chain} argument when fitting the model. } +\keyword{internal} diff --git a/man/write_stan_tempfile.Rd b/man/write_stan_tempfile.Rd index be3073ffa..e663302fe 100644 --- a/man/write_stan_tempfile.Rd +++ b/man/write_stan_tempfile.Rd @@ -20,3 +20,4 @@ is used.} \description{ This function is deprecated. Please use \code{\link[=write_stan_file]{write_stan_file()}} instead. } +\keyword{internal} From 87159463f4463a0d0eb2cb1085a0765fc8053d4c Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 15:17:38 -0600 Subject: [PATCH 08/16] Update _pkgdown.yml --- _pkgdown.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/_pkgdown.yml b/_pkgdown.yml index 1dea923a9..ca5375be8 100644 --- a/_pkgdown.yml +++ b/_pkgdown.yml @@ -111,7 +111,7 @@ reference: - write_stan_json - write_stan_file - draws_to_csv - - as.mcmc.list + - as_mcmc.list - as_draws.CmdStanMCMC - title: "Using CmdStanR with knitr and R Markdown" contents: From 3303838e3651ad479494eb6fc923e5a7f294f6b7 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 15:19:35 -0600 Subject: [PATCH 09/16] minor vignette fixes * now need to use underscore in file name for including the knitr settings (to avoid it being treated as its own vignette) * fix array syntax in a few cases --- .../opencl-files/bernoulli_logit_glm.stan | 2 +- vignettes/articles-online-only/opencl.Rmd | 2 +- .../children/{settings-knitr.Rmd => _settings-knitr.Rmd} | 0 vignettes/children/comparison-with-rstan.md | 8 ++------ vignettes/cmdstanr-internals.Rmd | 6 +++--- vignettes/cmdstanr.Rmd | 2 +- vignettes/deprecations.Rmd | 2 +- vignettes/posterior.Rmd | 4 ++-- vignettes/profiling.Rmd | 2 +- 9 files changed, 12 insertions(+), 16 deletions(-) rename vignettes/children/{settings-knitr.Rmd => _settings-knitr.Rmd} (100%) diff --git a/vignettes/articles-online-only/opencl-files/bernoulli_logit_glm.stan b/vignettes/articles-online-only/opencl-files/bernoulli_logit_glm.stan index 133521db4..3d5514364 100644 --- a/vignettes/articles-online-only/opencl-files/bernoulli_logit_glm.stan +++ b/vignettes/articles-online-only/opencl-files/bernoulli_logit_glm.stan @@ -2,7 +2,7 @@ data { int k; int n; matrix[n, k] X; - int y[n]; + array[n] int y; } parameters { vector[k] beta; diff --git a/vignettes/articles-online-only/opencl.Rmd b/vignettes/articles-online-only/opencl.Rmd index bdf35bf77..bc8599efd 100644 --- a/vignettes/articles-online-only/opencl.Rmd +++ b/vignettes/articles-online-only/opencl.Rmd @@ -13,7 +13,7 @@ vignette: > %\VignetteEncoding{UTF-8} --- -```{r child="../children/settings-knitr.Rmd"} +```{r child="../children/_settings-knitr.Rmd"} ``` ## Introduction diff --git a/vignettes/children/settings-knitr.Rmd b/vignettes/children/_settings-knitr.Rmd similarity index 100% rename from vignettes/children/settings-knitr.Rmd rename to vignettes/children/_settings-knitr.Rmd diff --git a/vignettes/children/comparison-with-rstan.md b/vignettes/children/comparison-with-rstan.md index 65233450b..314266c82 100644 --- a/vignettes/children/comparison-with-rstan.md +++ b/vignettes/children/comparison-with-rstan.md @@ -9,15 +9,11 @@ results to output files. ### Advantages of RStan -* Advanced features. We are working on making these available outside of RStan -but currently they are only available to R users via RStan: - - `rstan::log_prob()` - - `rstan::grad_log_prob()` - - `rstan::expose_stan_functions()` - * Allows other developers to distribute R packages with _pre-compiled_ Stan programs (like **rstanarm**) on CRAN. +* Avoids use of R6 classes, which may result in more familiar syntax for many R users. + ### Advantages of CmdStanR * Compatible with latest versions of Stan. Keeping up with Stan releases is diff --git a/vignettes/cmdstanr-internals.Rmd b/vignettes/cmdstanr-internals.Rmd index 93c656f31..164b80085 100644 --- a/vignettes/cmdstanr-internals.Rmd +++ b/vignettes/cmdstanr-internals.Rmd @@ -13,7 +13,7 @@ vignette: > %\VignetteEncoding{UTF-8} --- -```{r child="children/settings-knitr.Rmd"} +```{r child="children/_settings-knitr.Rmd"} ``` ## Introduction @@ -127,7 +127,7 @@ is missing a lower bound and a prior for a parameter. stan_file_pedantic <- write_stan_file(" data { int N; - int y[N]; + array[N] int y; } parameters { // should have but omitting to demonstrate pedantic mode @@ -461,7 +461,7 @@ the save file. rm(fit); gc() fit <- readRDS(temp_rds_file) -fit$summary() +fit$print() ``` ## Developing using CmdStanR diff --git a/vignettes/cmdstanr.Rmd b/vignettes/cmdstanr.Rmd index 9350ea541..31f35bce3 100644 --- a/vignettes/cmdstanr.Rmd +++ b/vignettes/cmdstanr.Rmd @@ -13,7 +13,7 @@ vignette: > %\VignetteEncoding{UTF-8} --- -```{r child="children/settings-knitr.Rmd"} +```{r child="children/_settings-knitr.Rmd"} ``` ## Introduction diff --git a/vignettes/deprecations.Rmd b/vignettes/deprecations.Rmd index 6fb1b6e8a..920aebed6 100644 --- a/vignettes/deprecations.Rmd +++ b/vignettes/deprecations.Rmd @@ -13,7 +13,7 @@ vignette: > %\VignetteEncoding{UTF-8} --- -```{r child="children/settings-knitr.Rmd"} +```{r child="children/_settings-knitr.Rmd"} ``` ## Introduction diff --git a/vignettes/posterior.Rmd b/vignettes/posterior.Rmd index fca84be29..e1f9e30e8 100644 --- a/vignettes/posterior.Rmd +++ b/vignettes/posterior.Rmd @@ -12,7 +12,7 @@ vignette: > %\VignetteEncoding{UTF-8} --- -```{r child="children/settings-knitr.Rmd"} +```{r child="children/_settings-knitr.Rmd"} ``` ## Summary @@ -81,4 +81,4 @@ fit$summary(variables = NULL, "Strictly Positive" = strict_pos) # fit$print(variables = NULL, "Strictly Positive" = strict_pos) ``` -For more information, see [posterior::summarise_draws()], which is called by `$summary()`. +For more information, see `posterior::summarise_draws()`, which is called by `$summary()`. diff --git a/vignettes/profiling.Rmd b/vignettes/profiling.Rmd index 59323bb1a..1f9eb0dc5 100644 --- a/vignettes/profiling.Rmd +++ b/vignettes/profiling.Rmd @@ -13,7 +13,7 @@ vignette: > %\VignetteEncoding{UTF-8} --- -```{r child="children/settings-knitr.Rmd"} +```{r child="children/_settings-knitr.Rmd"} ``` ## Introduction From 1dffad3791a319b4d4ddd6c25dd9efc20451055f Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 15:38:14 -0600 Subject: [PATCH 10/16] Update NEWS.md --- NEWS.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/NEWS.md b/NEWS.md index aaf69021e..863988233 100644 --- a/NEWS.md +++ b/NEWS.md @@ -5,7 +5,7 @@ * New `expose_functions()` method to expose Stan functions to R by @andrjohns in #702. See `?expose_functions`. * New methods for accessing log_prob, grad_log_prob, hessian, un/constrain variables by @andrjohns in #701. See `?init_model_methods`. -## Other changes +### Other changes * mod$variables works w includes in precompile state (fix #680) by @MKyhos in #682 * Update broken link for Stan OpenCL support page by @erictleung in #686 From 0af482832c6e1f6898e78bb4cffa4107a9bc90c0 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 19:08:16 -0600 Subject: [PATCH 11/16] temporary work around pkgdown formatting issue with posterior draws_summary the hack of using print.data.frame in chunks with echo=FALSE is used because the pillar formatting of posterior draws_summary objects isn't playing nicely with pkgdown::build_articles(). --- vignettes/children/comparison-with-rstan.md | 2 - vignettes/cmdstanr.Rmd | 52 +++++++++++--- vignettes/posterior.Rmd | 75 ++++++++++++++++++--- 3 files changed, 107 insertions(+), 22 deletions(-) diff --git a/vignettes/children/comparison-with-rstan.md b/vignettes/children/comparison-with-rstan.md index 314266c82..ed1628f1b 100644 --- a/vignettes/children/comparison-with-rstan.md +++ b/vignettes/children/comparison-with-rstan.md @@ -22,8 +22,6 @@ package and new CRAN releases of both **rstan** and **StanHeaders**. With CmdStanR the latest improvements in Stan will be available from R immediately after updating CmdStan using `cmdstanr::install_cmdstan()`. -* Fewer installation issues (e.g., no need to mess with Makevars files). - * Running Stan via external processes results in fewer unexpected crashes, especially in RStudio. diff --git a/vignettes/cmdstanr.Rmd b/vignettes/cmdstanr.Rmd index 31f35bce3..dcd24d15a 100644 --- a/vignettes/cmdstanr.Rmd +++ b/vignettes/cmdstanr.Rmd @@ -18,11 +18,11 @@ vignette: > ## Introduction -CmdStanR is a lightweight interface to [Stan](https://mc-stan.org/) for R users -(see [CmdStanPy](https://github.com/stan-dev/cmdstanpy) for Python) that -provides an alternative to the traditional [RStan](https://mc-stan.org/rstan/) -interface. See the [*Comparison with RStan*](#comparison-with-rstan) section -later in this vignette for more details on how the two interfaces differ. +CmdStanR (Command Stan R) is a lightweight interface to +[Stan](https://mc-stan.org/) for R users that provides an alternative to the +traditional [RStan](https://mc-stan.org/rstan/) interface. See the [*Comparison +with RStan*](#comparison-with-rstan) section later in this vignette for more +details on how the two interfaces differ. **CmdStanR is not on CRAN yet**, but the beta release can be installed by running the following command in R. @@ -38,7 +38,6 @@ later in examples. ```{r library, message=FALSE} library(cmdstanr) -check_cmdstan_toolchain(fix = TRUE, quiet = TRUE) library(posterior) library(bayesplot) color_scheme_set("brightblue") @@ -188,7 +187,7 @@ first argument specifies the variables to summarize and any arguments after that are passed on to `posterior::summarise_draws()` to specify which summaries to compute, whether to use multiple cores, etc. -```{r summary} +```{r summary, eval=FALSE} fit$summary() fit$summary(variables = c("theta", "lp__"), "mean", "sd") @@ -203,6 +202,24 @@ fit$summary( ) ``` +```{r, echo=FALSE} +# NOTE: the hack of using print.data.frame in chunks with echo=FALSE +# is used because the pillar formatting of posterior draws_summary objects +# isn't playing nicely with pkgdown::build_articles(). +options(digits = 2) +print.data.frame(fit$summary()) + +print.data.frame(fit$summary(variables = c("theta", "lp__"), "mean", "sd")) + +print.data.frame(fit$summary("theta", pr_lt_half = ~ mean(. <= 0.5))) + +print.data.frame(fit$summary( + variables = NULL, + posterior::default_summary_measures(), + extra_quantiles = ~posterior::quantile2(., probs = c(.0275, .975)) +)) +``` + #### CmdStan's stansummary utility CmdStan itself provides a `stansummary` utility that can be called using the @@ -334,11 +351,20 @@ the `$sample()` method demonstrated above. We can find the (penalized) maximum likelihood estimate (MLE) using [`$optimize()`](https://mc-stan.org/cmdstanr/reference/model-method-optimize.html). -```{r optimize} +```{r optimize, eval=FALSE} fit_mle <- mod$optimize(data = data_list, seed = 123) fit_mle$summary() # includes lp__ (log prob calculated by Stan program) fit_mle$mle("theta") ``` +```{r, echo=FALSE} +# NOTE: the hack of using print.data.frame in chunks with echo=FALSE +# is used because the pillar formatting of posterior draws_summary objects +# isn't playing nicely with pkgdown::build_articles(). +options(digits = 2) +fit_mle <- mod$optimize(data = data_list, seed = 123) +print.data.frame(fit_mle$summary()) # includes lp__ (log prob calculated by Stan program) +fit_mle$mle("theta") +``` Here's a plot comparing the penalized MLE to the posterior distribution of `theta`. @@ -354,10 +380,18 @@ We can run Stan's experimental variational Bayes algorithm (ADVI) using the [`$variational()`](https://mc-stan.org/cmdstanr/reference/model-method-variational.html) method. -```{r variational} +```{r variational, eval=FALSE} fit_vb <- mod$variational(data = data_list, seed = 123, output_samples = 4000) fit_vb$summary("theta") ``` +```{r, echo=FALSE} +# NOTE: the hack of using print.data.frame in chunks with echo=FALSE +# is used because the pillar formatting of posterior draws_summary objects +# isn't playing nicely with pkgdown::build_articles(). +options(digits = 2) +fit_vb <- mod$variational(data = data_list, seed = 123, output_samples = 4000) +print.data.frame(fit_vb$summary("theta")) +``` The `$draws()` method can be used to access the approximate posterior draws. Let's extract the draws, make the same plot we made after MCMC, and compare the diff --git a/vignettes/posterior.Rmd b/vignettes/posterior.Rmd index e1f9e30e8..f9d339dc4 100644 --- a/vignettes/posterior.Rmd +++ b/vignettes/posterior.Rmd @@ -15,14 +15,28 @@ vignette: > ```{r child="children/_settings-knitr.Rmd"} ``` + +```{r, include=FALSE} +options(digits=2) +``` + ## Summary We can easily customise the summary statistics reported by `$summary()` and `$print()`. -```{r} +```{r eval=FALSE} fit <- cmdstanr::cmdstanr_example("schools", method = "sample") fit$summary() ``` +```{r echo=FALSE} +fit <- cmdstanr::cmdstanr_example("schools", method = "sample") +print.data.frame(fit$summary()) +``` By default all variables are summaries with the follow functions: ```{r} @@ -30,23 +44,33 @@ posterior::default_summary_measures() ``` To change the variables summarised, we use the variables argument -```{r} +```{r eval=FALSE} fit$summary(variables = c("mu", "tau")) ``` +```{r echo=FALSE} +print.data.frame(fit$summary(variables = c("mu", "tau"))) +``` We can additionally change which functions are used -```{r} +```{r eval=FALSE} fit$summary(variables = c("mu", "tau"), mean, sd) ``` +```{r echo=FALSE} +print.data.frame(fit$summary(variables = c("mu", "tau"), mean, sd)) +``` To summarise all variables with non-default functions, it is necessary to set explicitly set the variables argument, either to `NULL` or the full vector of variable names. -```{r} +```{r eval=FALSE} fit$metadata()$model_params fit$summary(variables = NULL, "mean", "median") ``` +```{r echo=FALSE} +fit$metadata()$model_params +print.data.frame(fit$summary(variables = NULL, "mean", "median")) +``` Summary functions can be specified by character string, function, or using a formula (or anything else supported by [rlang::as_function]). If these arguments are named, those names will be used in the tibble output. If the summary results are named they will take precedence. -```{r} +```{r eval=FALSE} my_sd <- function(x) c(My_SD = sd(x)) fit$summary( c("mu", "tau"), @@ -57,28 +81,57 @@ fit$summary( Minimum = function(x) min(x) ) ``` +```{r echo=FALSE} +my_sd <- function(x) c(My_SD = sd(x)) +print.data.frame(fit$summary( + c("mu", "tau"), + MEAN = mean, + "median", + my_sd, + ~quantile(.x, probs = c(0.1, 0.9)), + Minimum = function(x) min(x) +)) +``` + Arguments to all summary functions can also be specified with `.args`. -```{r} +```{r eval=FALSE} fit$summary(c("mu", "tau"), quantile, .args = list(probs = c(0.025, .05, .95, .975))) ``` +```{r echo=FALSE} +print.data.frame(fit$summary(c("mu", "tau"), quantile, .args = list(probs = c(0.025, .05, .95, .975)))) +``` The summary functions are applied to the array of sample values, with dimension `iter_sampling`x`chains`. -```{r} +```{r eval=FALSE} fit$summary(variables = NULL, dim, colMeans) ``` +```{r echo=FALSE} +print.data.frame(fit$summary(variables = NULL, dim, colMeans)) +``` -For this reason users may have unexpected results if they use [stats::var()] directly, as it will return a covariance matrix. An alternative is the [distributional::variance] function. -```{r} -fit$summary(c("mu", "tau"), distributional::variance, ~var(as.vector(.x))) + +For this reason users may have unexpected results if they use `stats::var()` directly, as it will return a covariance matrix. An alternative is the `distributional::variance()` function, +which can also be accessed via `posterior::variance()`. +```{r eval=FALSE} +fit$summary(c("mu", "tau"), posterior::variance, ~var(as.vector(.x))) ``` +```{r echo=FALSE} +print.data.frame(fit$summary(c("mu", "tau"), posterior::variance, ~var(as.vector(.x)))) +``` + Summary functions need not be numeric, but these won't work with `$print()`. -```{r} +```{r eval=FALSE} strict_pos <- function(x) if (all(x > 0)) "yes" else "no" fit$summary(variables = NULL, "Strictly Positive" = strict_pos) # fit$print(variables = NULL, "Strictly Positive" = strict_pos) ``` +```{r echo=FALSE} +strict_pos <- function(x) if (all(x > 0)) "yes" else "no" +print.data.frame(fit$summary(variables = NULL, "Strictly Positive" = strict_pos)) +# fit$print(variables = NULL, "Strictly Positive" = strict_pos) +``` For more information, see `posterior::summarise_draws()`, which is called by `$summary()`. From 907d0e83deaa5af49cff40e0c0f98232ae1e8e40 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 19:09:49 -0600 Subject: [PATCH 12/16] Update posterior.Rmd --- vignettes/posterior.Rmd | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/vignettes/posterior.Rmd b/vignettes/posterior.Rmd index f9d339dc4..465391d8b 100644 --- a/vignettes/posterior.Rmd +++ b/vignettes/posterior.Rmd @@ -19,7 +19,7 @@ vignette: > NOTE: the hack below of using print.data.frame in chunks with echo=FALSE is used because the pillar formatting of posterior draws_summary objects isn't playing nicely with pkgdown::build_articles(). When that is fixed -using options(digits=2) won't be necessary anymore. +using options(digits=2) also won't be necessary anymore. --> ```{r, include=FALSE} options(digits=2) From 71f85cc619b19a02efb41b10fe9a2fd0e3129590 Mon Sep 17 00:00:00 2001 From: jgabry Date: Tue, 25 Jul 2023 19:12:14 -0600 Subject: [PATCH 13/16] website for v0.6.0 --- docs/404.html | 160 +- docs/LICENSE-text.html | 139 +- docs/LICENSE.html | 148 +- .../articles/articles-online-only/opencl.html | 227 +-- docs/articles/cmdstanr-internals.html | 896 ++++++----- docs/articles/cmdstanr.html | 705 +++++---- .../cmdstanr_files/figure-html/plot-mle-1.png | Bin 37984 -> 41641 bytes .../figure-html/plot-variational-1-1.png | Bin 38005 -> 41428 bytes .../figure-html/plot-variational-2-1.png | Bin 37488 -> 41075 bytes .../cmdstanr_files/figure-html/plots-1.png | Bin 37827 -> 41578 bytes docs/articles/deprecations.html | 199 ++- docs/articles/index.html | 173 +-- docs/articles/posterior.html | 309 ++++ .../header-attrs-2.18/header-attrs.js | 12 + docs/articles/profiling.html | 325 ++-- docs/articles/r-markdown.html | 260 ++-- docs/authors.html | 230 ++- docs/index.html | 153 +- docs/news/index.html | 527 +++---- docs/pkgdown.css | 83 +- docs/pkgdown.js | 4 +- docs/pkgdown.yml | 12 +- docs/pull_request_template.html | 175 +-- docs/reference/CmdStanDiagnose.html | 218 +-- docs/reference/CmdStanGQ.html | 436 +++--- docs/reference/CmdStanMCMC.html | 242 +-- docs/reference/CmdStanMLE.html | 233 +-- docs/reference/CmdStanModel-1.png | Bin 12494 -> 15795 bytes docs/reference/CmdStanModel-2.png | Bin 12652 -> 16009 bytes docs/reference/CmdStanModel.html | 906 ++++++------ docs/reference/CmdStanVB.html | 237 +-- docs/reference/Rplot002.png | Bin 4617 -> 6854 bytes docs/reference/as_draws.CmdStanMCMC.html | 435 +++--- docs/reference/as_mcmc.list.html | 184 +-- .../cmdstan_default_install_path.html | 174 +-- docs/reference/cmdstan_default_path.html | 179 +-- docs/reference/cmdstan_model-1.png | Bin 12494 -> 15795 bytes docs/reference/cmdstan_model-2.png | Bin 12652 -> 16009 bytes docs/reference/cmdstan_model.html | 909 ++++++------ docs/reference/cmdstanr-package-1.png | Bin 12494 -> 15795 bytes docs/reference/cmdstanr-package-2.png | Bin 12652 -> 16009 bytes docs/reference/cmdstanr-package.html | 911 ++++++------ docs/reference/cmdstanr_example.html | 571 ++++--- docs/reference/draws_to_csv.html | 252 ++-- docs/reference/eng_cmdstan.html | 179 +-- .../reference/fit-method-cmdstan_summary.html | 485 +++--- docs/reference/fit-method-code.html | 263 ++-- .../fit-method-constrain_variables.html | 179 +++ .../fit-method-diagnostic_summary.html | 264 ++-- docs/reference/fit-method-draws-1.png | Bin 61243 -> 72272 bytes docs/reference/fit-method-draws-2.png | Bin 614623 -> 671367 bytes docs/reference/fit-method-draws.html | 534 +++---- docs/reference/fit-method-grad_log_prob.html | 173 +++ docs/reference/fit-method-gradients.html | 192 +-- docs/reference/fit-method-hessian.html | 172 +++ docs/reference/fit-method-init.html | 234 ++- .../fit-method-init_model_methods.html | 180 +++ docs/reference/fit-method-inv_metric.html | 332 ++--- docs/reference/fit-method-log_prob.html | 167 +++ docs/reference/fit-method-loo.html | 257 ++-- docs/reference/fit-method-lp-1.png | Bin 82170 -> 86194 bytes docs/reference/fit-method-lp.html | 212 +-- docs/reference/fit-method-metadata.html | 422 +++--- docs/reference/fit-method-mle.html | 212 +-- docs/reference/fit-method-num_chains.html | 178 +-- docs/reference/fit-method-output.html | 539 +++---- docs/reference/fit-method-profiles.html | 338 ++--- docs/reference/fit-method-return_codes.html | 200 +-- .../fit-method-sampler_diagnostics.html | 288 ++-- docs/reference/fit-method-save_object.html | 230 +-- .../fit-method-save_output_files.html | 318 ++-- docs/reference/fit-method-summary.html | 380 +++-- docs/reference/fit-method-time.html | 223 +-- .../fit-method-unconstrain_draws.html | 185 +++ .../fit-method-unconstrain_variables.html | 165 +++ .../fit-method-variable_skeleton.html | 171 +++ docs/reference/index.html | 477 ++---- docs/reference/install_cmdstan.html | 365 ++--- docs/reference/model-method-check_syntax.html | 335 ++--- docs/reference/model-method-compile.html | 424 +++--- docs/reference/model-method-diagnose.html | 325 ++-- .../model-method-expose_functions.html | 231 +++ docs/reference/model-method-format.html | 325 ++-- .../model-method-generate-quantities.html | 510 +++---- docs/reference/model-method-optimize-1.png | Bin 12494 -> 15795 bytes docs/reference/model-method-optimize-2.png | Bin 12652 -> 16009 bytes docs/reference/model-method-optimize.html | 1131 +++++++------- docs/reference/model-method-sample-1.png | Bin 12494 -> 15795 bytes docs/reference/model-method-sample-2.png | Bin 12652 -> 16009 bytes docs/reference/model-method-sample.html | 1310 ++++++++--------- docs/reference/model-method-sample_mpi.html | 651 ++++---- docs/reference/model-method-variables.html | 283 ++-- docs/reference/model-method-variational-1.png | Bin 12494 -> 15795 bytes docs/reference/model-method-variational-2.png | Bin 12652 -> 16009 bytes docs/reference/model-method-variational.html | 1151 +++++++-------- docs/reference/read_cmdstan_csv.html | 566 +++---- docs/reference/read_sample_csv.html | 163 +- docs/reference/register_knitr_engine.html | 188 +-- docs/reference/set_cmdstan_path.html | 210 +-- docs/reference/stan_threads.html | 170 +-- docs/reference/write_stan_file.html | 318 ++-- docs/reference/write_stan_json.html | 282 ++-- docs/reference/write_stan_tempfile.html | 175 +-- 103 files changed, 12938 insertions(+), 14948 deletions(-) create mode 100644 docs/articles/posterior.html create mode 100644 docs/articles/posterior_files/header-attrs-2.18/header-attrs.js create mode 100644 docs/reference/fit-method-constrain_variables.html create mode 100644 docs/reference/fit-method-grad_log_prob.html create mode 100644 docs/reference/fit-method-hessian.html create mode 100644 docs/reference/fit-method-init_model_methods.html create mode 100644 docs/reference/fit-method-log_prob.html create mode 100644 docs/reference/fit-method-unconstrain_draws.html create mode 100644 docs/reference/fit-method-unconstrain_variables.html create mode 100644 docs/reference/fit-method-variable_skeleton.html create mode 100644 docs/reference/model-method-expose_functions.html diff --git a/docs/404.html b/docs/404.html index 0b36386a3..ced1d14e2 100644 --- a/docs/404.html +++ b/docs/404.html @@ -1,74 +1,34 @@ - - - - + + + + - Page not found (404) • cmdstanr - - - + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + + + + + + + + + + - - - - -
-
- + +
+ + + - - -
+

Page not found (404)

@@ -177,31 +137,31 @@

Page not found (404)

+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

+

Site built with pkgdown 2.0.7.

-
+
+ + - - diff --git a/docs/LICENSE-text.html b/docs/LICENSE-text.html index 697b1bfa8..5ed5129d6 100644 --- a/docs/LICENSE-text.html +++ b/docs/LICENSE-text.html @@ -1,74 +1,12 @@ - - - - - - - -License • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -License • cmdstanr - - + + - - -
-
- -
- -
+

License

@@ -179,31 +108,27 @@

License

+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/LICENSE.html b/docs/LICENSE.html index 21ed26a88..2dcbb7c1a 100644 --- a/docs/LICENSE.html +++ b/docs/LICENSE.html @@ -1,74 +1,12 @@ - - - - - - - -BSD 3-Clause License • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -BSD 3-Clause License • cmdstanr - - + + - - -
-
- -
- -
+

BSD 3-Clause License

@@ -174,45 +103,38 @@

BSD 3-Clause License

-

Copyright (c) 2019, Stan Developers and their Assignees
-All rights reserved.

+

Copyright (c) 2019, Stan Developers and their Assignees All rights reserved.

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

-
    -

  1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

    2. +

    1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

    2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

    3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

      4. -
    -

    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

    +

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/articles/articles-online-only/opencl.html b/docs/articles/articles-online-only/opencl.html index 98da606fa..274f8275e 100644 --- a/docs/articles/articles-online-only/opencl.html +++ b/docs/articles/articles-online-only/opencl.html @@ -26,6 +26,8 @@ + +
+

How does CmdStanR work?

-

Jonah Gabry and Rok Češnovar

+

Jonah Gabry and +Rok Češnovar

- Source: vignettes/cmdstanr-internals.Rmd + Source: vignettes/cmdstanr-internals.Rmd
-
-

-Introduction

-

This vignette is intended to be read after the Getting started with CmdStanR vignette. Please read that first for important background. In this document we provide additional details about compiling models, passing in data, and how CmdStan output is saved and read back into R.

-

We will only use the $sample() method in examples, but all model fitting methods work in a similar way under the hood.

+
+

Introduction +

+

This vignette is intended to be read after the Getting +started with CmdStanR vignette. Please read that first for +important background. In this document we provide additional details +about compiling models, passing in data, and how CmdStan output is saved +and read back into R.

+

We will only use the $sample() method in examples, but +all model fitting methods work in a similar way under the hood.

-library(cmdstanr)
-check_cmdstan_toolchain(fix = TRUE, quiet = TRUE)
+library(cmdstanr) +check_cmdstan_toolchain(fix = TRUE, quiet = TRUE)
-
-

-Compilation

-
-

-Immediate compilation

-

The cmdstan_model() function creates a new CmdStanModel object. The CmdStanModel object stores the path to a Stan program as well as the path to a compiled executable.

+
+

Compilation +

+
+

Immediate compilation +

+

The cmdstan_model() function creates a new +CmdStanModel object. The CmdStanModel object +stores the path to a Stan program as well as the path to a compiled +executable.

-stan_file <- file.path(cmdstan_path(), "examples", "bernoulli", "bernoulli.stan")
-mod <- cmdstan_model(stan_file)
-mod$print()
+stan_file <- file.path(cmdstan_path(), "examples", "bernoulli", "bernoulli.stan") +mod <- cmdstan_model(stan_file) +mod$print()
data {
   int<lower=0> N;
-  array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
+  array[N] int<lower=0,upper=1> y;
 }
 parameters {
   real<lower=0,upper=1> theta;
@@ -171,209 +182,256 @@ 

   y ~ bernoulli(theta);
 }
-mod$stan_file()
-
[1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1/examples/bernoulli/bernoulli.stan"
+mod$stan_file()
+
[1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2/examples/bernoulli/bernoulli.stan"
-mod$exe_file()
-
[1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1/examples/bernoulli/bernoulli"
-

Subsequently, if you create a CmdStanModel object from the same Stan file then compilation will be skipped (assuming the file hasn’t changed).

+mod$exe_file()
+
[1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2/examples/bernoulli/bernoulli"
+

Subsequently, if you create a CmdStanModel object from +the same Stan file then compilation will be skipped (assuming the file +hasn’t changed).

-mod <- cmdstan_model(stan_file)
-

Internally, cmdstan_model() first creates the CmdStanModel object from just the Stan file and then calls its $compile() method. Optional arguments to the $compile() method can be passed via ....

+mod <- cmdstan_model(stan_file)
+

Internally, cmdstan_model() first creates the +CmdStanModel object from just the Stan file and then calls +its $compile() +method. Optional arguments to the $compile() method can be +passed via ....

-mod <- cmdstan_model(
-  stan_file, 
-  force_recompile = TRUE, 
-  include_paths = "paths/to/directories/with/included/files", 
-  cpp_options = list(stan_threads = TRUE, STANC2 = TRUE)
-)
+mod <- cmdstan_model( + stan_file, + force_recompile = TRUE, + include_paths = "paths/to/directories/with/included/files", + cpp_options = list(stan_threads = TRUE, STANC2 = TRUE) +)
-
-

-Delayed compilation

-

It is also possible to delay compilation when creating the CmdStanModel object by specifying compile=FALSE and then later calling the $compile() method directly.

+
+

Delayed compilation +

+

It is also possible to delay compilation when creating the +CmdStanModel object by specifying +compile=FALSE and then later calling the +$compile() method directly.

-unlink(mod$exe_file())
-mod <- cmdstan_model(stan_file, compile = FALSE)
-mod$exe_file() # not yet created
-
character(0)
+unlink(mod$exe_file()) +mod <- cmdstan_model(stan_file, compile = FALSE) +mod$exe_file() # not yet created
+
character(0)
-mod$compile()
-mod$exe_file()
-
[1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1/examples/bernoulli/bernoulli"
+mod$compile() +mod$exe_file()
+
[1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2/examples/bernoulli/bernoulli"
-
-

-Pedantic check

-

If you are using CmdStan version 2.24 or later and CmdStanR version 0.2.1 or later, you can run a pedantic check for your model. CmdStanR will always check that your Stan program does not contain any invalid syntax but with pedantic mode enabled the check will also warn you about other potential issues in your model, for example:

+
+

Pedantic check +

+

If you are using CmdStan version 2.24 or later and CmdStanR version +0.2.1 or later, you can run a pedantic check for your model. CmdStanR +will always check that your Stan program does not contain any invalid +syntax but with pedantic mode enabled the check will also warn you about +other potential issues in your model, for example:

    -
  • Distribution usages issues: distribution arguments do not match the distribution specification, or some specific distribution is used in an inadvisable way.
    • -
  • Unused parameter: a parameter is defined but does not contribute to target.
    • -
  • Large or small constant in a distribution: very large or very small constants are used as distribution arguments.
    • -
  • Control flow depends on a parameter: branching control flow (like if/else) depends on a parameter value.
    • -
  • Parameter has multiple twiddles: a parameter is on the left-hand side of multiple twiddles (i.e., multiple ~ symbols).
    • -
  • Parameter has zero or multiple priors: a parameter has zero or more than one prior distribution.
    • -
  • Variable is used before assignment: a variable is used before being assigned a value.
    • -
  • Strict or nonsensical parameter bounds: a parameter is given questionable bounds.
    • +
  • Distribution usages issues: distribution arguments do not match the +distribution specification, or some specific distribution is used in an +inadvisable way.
    • +
  • Unused parameter: a parameter is defined but does not contribute to +target.
    • +
  • Large or small constant in a distribution: very large or very small +constants are used as distribution arguments.
    • +
  • Control flow depends on a parameter: branching control flow (like +if/else) depends on a parameter value.
    • +
  • Parameter has multiple twiddles: a parameter is on the left-hand +side of multiple twiddles (i.e., multiple ~ symbols).
    • +
  • Parameter has zero or multiple priors: a parameter has zero or more +than one prior distribution.
    • +
  • Variable is used before assignment: a variable is used before being +assigned a value.
    • +
  • Strict or nonsensical parameter bounds: a parameter is given +questionable bounds.
-

For the latest information on the checks performed in pedantic mode see the Pedantic mode chapter in the Stan Reference Manual.

-

Pedantic mode is available when compiling the model or when using the separate $check_syntax() method of a CmdStanModel object. Internally this corresponds to setting the stanc (Stan transpiler) option warn-pedantic. Here we demonstrate pedantic mode with a Stan program that is syntactically correct but is missing a lower bound and a prior for a parameter.

+

For the latest information on the checks performed in pedantic mode +see the Pedantic +mode chapter in the Stan Reference Manual.

+

Pedantic mode is available when compiling the model or when using the +separate $check_syntax() method of a +CmdStanModel object. Internally this corresponds to setting +the stanc (Stan transpiler) option +warn-pedantic. Here we demonstrate pedantic mode with a +Stan program that is syntactically correct but is missing a lower bound +and a prior for a parameter.

-stan_file_pedantic <- write_stan_file("
-data {
-  int N;
-  int y[N];
-}
-parameters {
-  // should have <lower=0> but omitting to demonstrate pedantic mode
-  real lambda;
-}
-model {
-  y ~ poisson(lambda);
-}
-")
-

To turn on pedantic mode at compile time you can set pedantic=TRUE in the call to cmdstan_model() (or when calling the $compile() method directly if using the delayed compilation approach described above).

-
mod_pedantic <- cmdstan_model(stan_file_pedantic, pedantic = TRUE)
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56281c2c0.stan', line 4, column 2: Declaration
-    of arrays by placing brackets after a variable name is deprecated and
-    will be removed in Stan 2.32.0. Instead use the array keyword before the
-    type. This can be changed automatically using the auto-format flag to
-    stanc
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56281c2c0.stan', line 11, column 14: A
-    poisson distribution is given parameter lambda as a rate parameter
-    (argument 1), but lambda was not constrained to be strictly positive.
-Warning: The parameter lambda has no priors.
-

To turn on pedantic mode separately from compilation use the pedantic argument to the $check_syntax() method.

-
mod_pedantic$check_syntax(pedantic = TRUE) 
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 4, column 2: Declaration
-    of arrays by placing brackets after a variable name is deprecated and
-    will be removed in Stan 2.32.0. Instead use the array keyword before the
-    type. This can be changed automatically using the auto-format flag to
-    stanc
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 11, column 14: A
-    poisson distribution is given parameter lambda as a rate parameter
-    (argument 1), but lambda was not constrained to be strictly positive.
-Warning: The parameter lambda has no priors.
-Stan program is syntactically correct
-

Using pedantic=TRUE via the $check_syntax() method also has the advantage that it can be used even if the model hasn’t been compiled yet. This can be helpful because the pedantic and syntax checks themselves are much faster than compilation.

-
file.remove(mod_pedantic$exe_file()) # delete compiled executable
-[1] TRUE
-rm(mod_pedantic)
-
-mod_pedantic <- cmdstan_model(stan_file_pedantic, compile = FALSE)
-mod_pedantic$check_syntax(pedantic = TRUE)
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 4, column 2: Declaration
-    of arrays by placing brackets after a variable name is deprecated and
-    will be removed in Stan 2.32.0. Instead use the array keyword before the
-    type. This can be changed automatically using the auto-format flag to
-    stanc
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 11, column 14: A
-    poisson distribution is given parameter lambda as a rate parameter
-    (argument 1), but lambda was not constrained to be strictly positive.
-Warning: The parameter lambda has no priors.
-Stan program is syntactically correct
+stan_file_pedantic <- write_stan_file(" +data { + int N; + array[N] int y; +} +parameters { + // should have <lower=0> but omitting to demonstrate pedantic mode + real lambda; +} +model { + y ~ poisson(lambda); +} +")
+

To turn on pedantic mode at compile time you can set +pedantic=TRUE in the call to cmdstan_model() +(or when calling the $compile() method directly if using +the delayed compilation approach described above).

+
mod_pedantic <- cmdstan_model(stan_file_pedantic, pedantic = TRUE)
+Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac4361020c.stan', line 11, column 14: A
+    poisson distribution is given parameter lambda as a rate parameter
+    (argument 1), but lambda was not constrained to be strictly positive.
+Warning: The parameter lambda has no priors. This means either no prior is
+    provided, or the prior(s) depend on data variables. In the later case,
+    this may be a false positive.
+

To turn on pedantic mode separately from compilation use the +pedantic argument to the $check_syntax() +method.

+
mod_pedantic$check_syntax(pedantic = TRUE)
+Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model_febb1e69c7387a0e64cf13583e078104.stan', line 11, column 14: A
+    poisson distribution is given parameter lambda as a rate parameter
+    (argument 1), but lambda was not constrained to be strictly positive.
+Warning: The parameter lambda has no priors. This means either no prior is
+    provided, or the prior(s) depend on data variables. In the later case,
+    this may be a false positive.
+Stan program is syntactically correct
+

Using pedantic=TRUE via the $check_syntax() +method also has the advantage that it can be used even if the model +hasn’t been compiled yet. This can be helpful because the pedantic and +syntax checks themselves are much faster than compilation.

+
file.remove(mod_pedantic$exe_file()) # delete compiled executable
+[1] TRUE
+rm(mod_pedantic)
+
+mod_pedantic <- cmdstan_model(stan_file_pedantic, compile = FALSE)
+mod_pedantic$check_syntax(pedantic = TRUE)
+Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model_febb1e69c7387a0e64cf13583e078104.stan', line 11, column 14: A
+    poisson distribution is given parameter lambda as a rate parameter
+    (argument 1), but lambda was not constrained to be strictly positive.
+Warning: The parameter lambda has no priors. This means either no prior is
+    provided, or the prior(s) depend on data variables. In the later case,
+    this may be a false positive.
+Stan program is syntactically correct
-
-

-Stan model variables

-

If using CmdStan 2.27 or newer, you can obtain the names, types and dimensions of the data, parameters, transformed parameters and generated quantities variables of a Stan model using the $variables() method of the CmdStanModel object.

+
+

Stan model variables +

+

If using CmdStan 2.27 or newer, you can obtain the names, types and +dimensions of the data, parameters, transformed parameters and generated +quantities variables of a Stan model using the $variables() +method of the CmdStanModel object.

-stan_file_variables <- write_stan_file("
-data {
-  int<lower=1> J;
-  vector<lower=0>[J] sigma;
-  vector[J] y;
-}
-parameters {
-  real mu;
-  real<lower=0> tau;
-  vector[J] theta_raw;
-}
-transformed parameters {
-  vector[J] theta = mu + tau * theta_raw;
-}
-model {
-  target += normal_lpdf(tau | 0, 10);
-  target += normal_lpdf(mu | 0, 10);
-  target += normal_lpdf(theta_raw | 0, 1);
-  target += normal_lpdf(y | theta, sigma);
-}
-")
-mod_v <- cmdstan_model(stan_file_variables)
-variables <- mod_v$variables()
-

The $variables() method returns a list with data, parameters, transformed_parameters and generated_quantities elements, each corresponding to variables in their respective block of the program. Transformed data variables are not listed as they are not used in the model’s input or output.

+stan_file_variables <- write_stan_file(" +data { + int<lower=1> J; + vector<lower=0>[J] sigma; + vector[J] y; +} +parameters { + real mu; + real<lower=0> tau; + vector[J] theta_raw; +} +transformed parameters { + vector[J] theta = mu + tau * theta_raw; +} +model { + target += normal_lpdf(tau | 0, 10); + target += normal_lpdf(mu | 0, 10); + target += normal_lpdf(theta_raw | 0, 1); + target += normal_lpdf(y | theta, sigma); +} +") +mod_v <- cmdstan_model(stan_file_variables) +variables <- mod_v$variables()
+

The $variables() method returns a list with +data, parameters, +transformed_parameters and +generated_quantities elements, each corresponding to +variables in their respective block of the program. Transformed data +variables are not listed as they are not used in the model’s input or +output.

-names(variables)
+names(variables)
[1] "parameters"             "included_files"         "data"                  
 [4] "transformed_parameters" "generated_quantities"  
-names(variables$data)
+names(variables$data)
[1] "J"     "sigma" "y"    
-names(variables$parameters)
+names(variables$parameters)
[1] "mu"        "tau"       "theta_raw"
-names(variables$transformed_parameters)
+names(variables$transformed_parameters)
[1] "theta"
-names(variables$generated_quantities)
-
character(0)
-

Each variable is represented as a list containing the type information (currently limited to real or int) and the number of dimensions.

+names(variables$generated_quantities) +
character(0)
+

Each variable is represented as a list containing the type +information (currently limited to real or int) +and the number of dimensions.

-variables$data$J
+variables$data$J 
$type
 [1] "int"
 
 $dimensions
 [1] 0
-variables$data$sigma
+variables$data$sigma 
$type
 [1] "real"
 
 $dimensions
 [1] 1
-variables$parameters$tau
+variables$parameters$tau 
$type
 [1] "real"
 
 $dimensions
 [1] 0
-variables$transformed_parameters$theta
+variables$transformed_parameters$theta 
$type
 [1] "real"
 
 $dimensions
 [1] 1
-
-

-Executable location

-

By default, the executable is created in the same directory as the file containing the Stan program. You can also specify a different location with the dir argument.

+
+

Executable location +

+

By default, the executable is created in the same directory as the +file containing the Stan program. You can also specify a different +location with the dir argument.

-mod <- cmdstan_model(stan_file, dir = "path/to/directory/for/executable")
+mod <- cmdstan_model(stan_file, dir = "path/to/directory/for/executable")
-
-

-Processing data

-

There are three data formats that CmdStanR allows when fitting a model:

+
+

Processing data +

+

There are three data formats that CmdStanR allows when fitting a +model:

  • named list of R objects
  • JSON file
  • R dump file
-
-

-Named list of R objects

-

Like the RStan interface, CmdStanR accepts a named list of R objects where the names correspond to variables declared in the data block of the Stan program. In the Bernoulli model the data is N, the number of data points, and y an integer array of observations.

+
+

Named list of R objects +

+

Like the RStan interface, CmdStanR accepts a named list of R objects +where the names correspond to variables declared in the data block of +the Stan program. In the Bernoulli model the data is N, the +number of data points, and y an integer array of +observations.

-mod$print()
+mod$print()
data {
   int<lower=0> N;
-  array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
+  array[N] int<lower=0,upper=1> y;
 }
 parameters {
   real<lower=0,upper=1> theta;
@@ -383,92 +441,113 @@ 

   y ~ bernoulli(theta);
 }
-# data block has 'N' and 'y'
-data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-fit <- mod$sample(data = data_list)
-

Because CmdStan doesn’t accept lists of R objects, CmdStanR will first write the data to a temporary JSON file using write_stan_json(). This happens internally, but it is also possible to call write_stan_json() directly.

+# data block has 'N' and 'y' +data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1)) +fit <- mod$sample(data = data_list)
+

Because CmdStan doesn’t accept lists of R objects, CmdStanR will +first write the data to a temporary JSON file using +write_stan_json(). This happens internally, but it is also +possible to call write_stan_json() directly.

-data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-json_file <- tempfile(fileext = ".json")
-write_stan_json(data_list, json_file)
-cat(readLines(json_file), sep = "\n")
+data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1)) +json_file <- tempfile(fileext = ".json") +write_stan_json(data_list, json_file) +cat(readLines(json_file), sep = "\n")
{
   "N": 10,
   "y": [0, 1, 0, 0, 0, 0, 0, 0, 0, 1]
 }
-
-

-JSON file

-

If you already have your data in a JSON file you can just pass that file directly to CmdStanR instead of using a list of R objects. For example, we could pass in the JSON file we created above using write_stan_json():

+
+

JSON file +

+

If you already have your data in a JSON file you can just pass that +file directly to CmdStanR instead of using a list of R objects. For +example, we could pass in the JSON file we created above using +write_stan_json():

-fit <- mod$sample(data = json_file)
+fit <- mod$sample(data = json_file)
-
-

-R dump file

-

Finally, it is also possible to use the R dump file format. This is not recommended because CmdStan can process JSON faster than R dump, but CmdStanR allows it because CmdStan will accept files created by rstan::stan_rdump():

+
+

R dump file +

+

Finally, it is also possible to use the R dump file format. This is +not recommended because CmdStan can process JSON faster than R +dump, but CmdStanR allows it because CmdStan will accept files created +by rstan::stan_rdump():

-rdump_file <- tempfile(fileext = ".data.R")
-rstan::stan_rdump(names(data_list), file = rdump_file, envir = list2env(data_list))
-cat(readLines(rdump_file), sep = "\n")
-fit <- mod$sample(data = rdump_file)
+rdump_file <- tempfile(fileext = ".data.R") +rstan::stan_rdump(names(data_list), file = rdump_file, envir = list2env(data_list)) +cat(readLines(rdump_file), sep = "\n") +fit <- mod$sample(data = rdump_file)
-
-

-Writing CmdStan output to CSV

-
-

-Default temporary files

+
+

Writing CmdStan output to CSV +

+
+

Default temporary files +

-data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-fit <- mod$sample(data = data_list)
-

When fitting a model, the default behavior is to write the output from CmdStan to CSV files in a temporary directory.

+data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1)) +fit <- mod$sample(data = data_list)
+

When fitting a model, the default behavior is to write the output +from CmdStan to CSV files in a temporary directory.

-fit$output_files()
-
[1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-202203181228-1-9912f0.csv"
-[2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-202203181228-2-9912f0.csv"
-[3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-202203181228-3-9912f0.csv"
-[4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-202203181228-4-9912f0.csv"
-

These files will be lost if you end your R session or if you remove the fit object and force (or wait for) garbage collection.

+fit$output_files()
+
[1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-202307251455-1-48e885.csv"
+[2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-202307251455-2-48e885.csv"
+[3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-202307251455-3-48e885.csv"
+[4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-202307251455-4-48e885.csv"
+

These files will be lost if you end your R session or if you remove +the fit object and force (or wait for) garbage +collection.

-files <- fit$output_files()
-file.exists(files)
+files <- fit$output_files() +file.exists(files)
[1] TRUE TRUE TRUE TRUE
-rm(fit)
-gc()
-
          used (Mb) gc trigger (Mb) limit (Mb) max used  (Mb)
-Ncells 1135293 60.7    2208007  118         NA  2208007 118.0
-Vcells 2043878 15.6    8388608   64      32768  3361903  25.7
+rm(fit) +gc()
+
          used (Mb) gc trigger  (Mb) limit (Mb) max used  (Mb)
+Ncells 1260687 67.4    2431888 129.9         NA  2431888 129.9
+Vcells 2216221 17.0    8388608  64.0      32768  3504936  26.8
-file.exists(files)
+file.exists(files) 
[1] FALSE FALSE FALSE FALSE
-
-

-Non-temporary files

-

To save these files to a non-temporary location there are two options. You can either specify the output_dir argument to mod$sample() or use fit$save_output_files() after fitting the model.

+
+

Non-temporary files +

+

To save these files to a non-temporary location there are two +options. You can either specify the output_dir argument to +mod$sample() or use fit$save_output_files() +after fitting the model.

-# see ?save_output_files for info on optional arguments
-fit$save_output_files(dir = "path/to/directory")
+# see ?save_output_files for info on optional arguments +fit$save_output_files(dir = "path/to/directory")
-fit <- mod$sample(
-  data = data_list, 
-  output_dir = "path/to/directory"
-)
+fit <- mod$sample( + data = data_list, + output_dir = "path/to/directory" +)
-
-

-Reading CmdStan output into R

-
-

-Lazy CSV reading

-

With the exception of some diagnostic information, the CSV files are not read into R until their contents are requested by calling a method that requires them (e.g., fit$draws(), fit$summary(), etc.). If we examine the structure of the fit object, notice how the Private slot draws_ is NULL, indicating that the CSV files haven’t yet been read into R.

+
+

Reading CmdStan output into R +

+
+

Lazy CSV reading +

+

With the exception of some diagnostic information, the CSV files are +not read into R until their contents are requested by calling a method +that requires them (e.g., fit$draws(), +fit$summary(), etc.). If we examine the structure of the +fit object, notice how the Private slot +draws_ is NULL, indicating that the CSV files +haven’t yet been read into R.

-str(fit)
+str(fit)
Classes 'CmdStanMCMC', 'CmdStanFit', 'R6' <CmdStanMCMC>
   Inherits from: <CmdStanFit>
   Public:
@@ -476,14 +555,21 @@ 

     cmdstan_diagnose: function () 
     cmdstan_summary: function (flags = NULL) 
     code: function () 
+    constrain_variables: function (unconstrained_variables, transformed_parameters = TRUE, 
     data_file: function () 
     diagnostic_summary: function (diagnostics = c("divergences", "treedepth", "ebfmi"), 
     draws: function (variables = NULL, inc_warmup = FALSE, format = getOption("cmdstanr_draws_format", 
+    expose_functions: function (global = FALSE, verbose = FALSE) 
+    functions: environment
+    grad_log_prob: function (unconstrained_variables, jacobian_adjustment = TRUE) 
+    hessian: function (unconstrained_variables, jacobian_adjustment = TRUE) 
     init: function () 
+    init_model_methods: function (seed = 0, verbose = FALSE, hessian = FALSE) 
     initialize: function (runset) 
     inv_metric: function (matrix = TRUE) 
     latent_dynamics_files: function (include_failed = FALSE) 
-    loo: function (variables = "log_lik", r_eff = TRUE, ...) 
+    log_prob: function (unconstrained_variables, jacobian_adjustment = TRUE) 
+    loo: function (variables = "log_lik", r_eff = TRUE, moment_match = FALSE, 
     lp: function () 
     metadata: function () 
     num_chains: function () 
@@ -503,20 +589,26 @@ 

     save_profile_files: function (dir = ".", basename = NULL, timestamp = TRUE, random = TRUE) 
     summary: function (variables = NULL, ...) 
     time: function () 
+    unconstrain_draws: function (files = NULL, draws = NULL) 
+    unconstrain_variables: function (variables) 
+    variable_skeleton: function (transformed_parameters = TRUE, generated_quantities = TRUE) 
   Private:
     draws_: NULL
     init_: NULL
     inv_metric_: list
     metadata_: list
+    model_methods_env_: environment
     profiles_: NULL
     read_csv_: function (variables = NULL, sampler_diagnostics = NULL, format = getOption("cmdstanr_draws_format", 
-    sampler_diagnostics_: 1 2 2 1 2 1 1 2 1 1 1 1 2 2 1 2 1 1 2 1 2 1 1 1 1 2 1 1  ...
+    sampler_diagnostics_: 1 1 2 1 2 1 1 1 2 1 1 1 1 2 2 1 1 1 1 2 1 1 1 1 2 2 2 2  ...
     warmup_draws_: NULL
     warmup_sampler_diagnostics_: NULL
-

After we call a method that requires the draws then if we reexamine the structure of the object we will see that the draws_ slot in Private is no longer empty.

+

After we call a method that requires the draws then if we reexamine +the structure of the object we will see that the draws_ +slot in Private is no longer empty.

-draws <- fit$draws() # force CSVs to be read into R
-str(fit)
+draws <- fit$draws() # force CSVs to be read into R +str(fit)
Classes 'CmdStanMCMC', 'CmdStanFit', 'R6' <CmdStanMCMC>
   Inherits from: <CmdStanFit>
   Public:
@@ -524,14 +616,21 @@ 

     cmdstan_diagnose: function () 
     cmdstan_summary: function (flags = NULL) 
     code: function () 
+    constrain_variables: function (unconstrained_variables, transformed_parameters = TRUE, 
     data_file: function () 
     diagnostic_summary: function (diagnostics = c("divergences", "treedepth", "ebfmi"), 
     draws: function (variables = NULL, inc_warmup = FALSE, format = getOption("cmdstanr_draws_format", 
+    expose_functions: function (global = FALSE, verbose = FALSE) 
+    functions: environment
+    grad_log_prob: function (unconstrained_variables, jacobian_adjustment = TRUE) 
+    hessian: function (unconstrained_variables, jacobian_adjustment = TRUE) 
     init: function () 
+    init_model_methods: function (seed = 0, verbose = FALSE, hessian = FALSE) 
     initialize: function (runset) 
     inv_metric: function (matrix = TRUE) 
     latent_dynamics_files: function (include_failed = FALSE) 
-    loo: function (variables = "log_lik", r_eff = TRUE, ...) 
+    log_prob: function (unconstrained_variables, jacobian_adjustment = TRUE) 
+    loo: function (variables = "log_lik", r_eff = TRUE, moment_match = FALSE, 
     lp: function () 
     metadata: function () 
     num_chains: function () 
@@ -551,33 +650,43 @@ 

     save_profile_files: function (dir = ".", basename = NULL, timestamp = TRUE, random = TRUE) 
     summary: function (variables = NULL, ...) 
     time: function () 
+    unconstrain_draws: function (files = NULL, draws = NULL) 
+    unconstrain_variables: function (variables) 
+    variable_skeleton: function (transformed_parameters = TRUE, generated_quantities = TRUE) 
   Private:
-    draws_: -7.82818 -7.71698 -7.1161 -7.09583 -7.51485 -8.04188 -7. ...
+    draws_: -7.16701 -7.08773 -6.93018 -6.84097 -6.84713 -6.90778 -7 ...
     init_: NULL
     inv_metric_: list
     metadata_: list
+    model_methods_env_: environment
     profiles_: NULL
     read_csv_: function (variables = NULL, sampler_diagnostics = NULL, format = getOption("cmdstanr_draws_format", 
-    sampler_diagnostics_: 1 2 2 1 2 1 1 2 1 1 1 1 2 2 1 2 1 1 2 1 2 1 1 1 1 2 1 1  ...
+    sampler_diagnostics_: 1 1 2 1 2 1 1 1 2 1 1 1 1 2 2 1 1 1 1 2 1 1 1 1 2 2 2 2  ...
     warmup_draws_: NULL
     warmup_sampler_diagnostics_: NULL
-

For models with many parameters, transformed parameters, or generated quantities, if only some are requested (e.g., by specifying the variables argument to fit$draws()) then CmdStanR will only read in the requested variables (unless they have already been read in).

+

For models with many parameters, transformed parameters, or generated +quantities, if only some are requested (e.g., by specifying the +variables argument to fit$draws()) then +CmdStanR will only read in the requested variables (unless they have +already been read in).

-
-

-read_cmdstan_csv()

-

Internally, the read_cmdstan_csv() function is used to read the CmdStan CSV files into R. This function is exposed to users, so you can also call it directly.

+
+

read_cmdstan_csv() +

+

Internally, the read_cmdstan_csv() function is used to +read the CmdStan CSV files into R. This function is exposed to users, so +you can also call it directly.

-# see ?read_cmdstan_csv for info on optional arguments controlling 
-# what information is read in
-csv_contents <- read_cmdstan_csv(fit$output_files())
-str(csv_contents)
+# see ?read_cmdstan_csv for info on optional arguments controlling +# what information is read in +csv_contents <- read_cmdstan_csv(fit$output_files()) +str(csv_contents)
List of 8
  $ metadata                       :List of 40
   ..$ stan_version_major  : num 2
-  ..$ stan_version_minor  : num 29
-  ..$ stan_version_patch  : num 1
-  ..$ start_datetime      : chr "2022-03-18 18:28:03 UTC"
+  ..$ stan_version_minor  : num 32
+  ..$ stan_version_patch  : num 2
+  ..$ start_datetime      : chr "2023-07-25 20:55:39 UTC"
   ..$ method              : chr "sample"
   ..$ save_warmup         : num 0
   ..$ thin                : num 1
@@ -594,14 +703,14 @@ 

   ..$ num_chains          : num 1
   ..$ id                  : num [1:4] 1 2 3 4
   ..$ init                : num [1:4] 2 2 2 2
-  ..$ seed                : num 1.13e+09
+  ..$ seed                : num 1.01e+09
   ..$ refresh             : num 100
   ..$ sig_figs            : num -1
-  ..$ profile_file        : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-profile-202203181228-1-6c529e.csv"
-  ..$ stanc_version       : chr "stanc3 v2.29.1"
+  ..$ profile_file        : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-profile-202307251455-1-07c6cb.csv"
+  ..$ stanc_version       : chr "stanc3 v2.32.2"
   ..$ sampler_diagnostics : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
   ..$ variables           : chr [1:2] "lp__" "theta"
-  ..$ step_size_adaptation: num [1:4] 1.062 0.9 0.966 0.973
+  ..$ step_size_adaptation: num [1:4] 0.938 0.853 0.941 0.842
   ..$ model_name          : chr "bernoulli_model"
   ..$ adapt_engaged       : num 1
   ..$ adapt_delta         : num 0.8
@@ -612,9 +721,9 @@ 

   ..$ threads_per_chain   : num 1
   ..$ time                :'data.frame':    4 obs. of  4 variables:
   .. ..$ chain_id: num [1:4] 1 2 3 4
-  .. ..$ warmup  : num [1:4] 0.004 0.005 0.005 0.004
-  .. ..$ sampling: num [1:4] 0.012 0.013 0.018 0.014
-  .. ..$ total   : num [1:4] 0.016 0.018 0.023 0.018
+  .. ..$ warmup  : num [1:4] 0.004 0.004 0.004 0.004
+  .. ..$ sampling: num [1:4] 0.012 0.011 0.011 0.012
+  .. ..$ total   : num [1:4] 0.016 0.015 0.015 0.016
   ..$ stan_variable_sizes :List of 2
   .. ..$ lp__ : num 1
   .. ..$ theta: num 1
@@ -624,143 +733,178 @@ 

   ..$ total : int NA
   ..$ chains:'data.frame':  4 obs. of  4 variables:
   .. ..$ chain_id: num [1:4] 1 2 3 4
-  .. ..$ warmup  : num [1:4] 0.004 0.005 0.005 0.004
-  .. ..$ sampling: num [1:4] 0.012 0.013 0.018 0.014
-  .. ..$ total   : num [1:4] 0.016 0.018 0.023 0.018
+  .. ..$ warmup  : num [1:4] 0.004 0.004 0.004 0.004
+  .. ..$ sampling: num [1:4] 0.012 0.011 0.011 0.012
+  .. ..$ total   : num [1:4] 0.016 0.015 0.015 0.016
  $ inv_metric                     :List of 4
-  ..$ 1: num 0.532
-  ..$ 2: num 0.497
-  ..$ 3: num 0.514
-  ..$ 4: num 0.539
+  ..$ 1: num 0.491
+  ..$ 2: num 0.494
+  ..$ 3: num 0.449
+  ..$ 4: num 0.511
  $ step_size                      :List of 4
-  ..$ 1: num 1.06
-  ..$ 2: num 0.9
-  ..$ 3: num 0.966
-  ..$ 4: num 0.973
+  ..$ 1: num 0.938
+  ..$ 2: num 0.853
+  ..$ 3: num 0.941
+  ..$ 4: num 0.842
  $ warmup_draws                   : NULL
- $ post_warmup_draws              : 'draws_array' num [1:1000, 1:4, 1:2] -7.83 -7.72 -7.12 -7.1 -7.51 ...
+ $ post_warmup_draws              : 'draws_array' num [1:1000, 1:4, 1:2] -7.17 -7.09 -6.93 -6.84 -6.85 ...
   ..- attr(*, "dimnames")=List of 3
   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
   .. ..$ variable : chr [1:2] "lp__" "theta"
  $ warmup_sampler_diagnostics     : NULL
- $ post_warmup_sampler_diagnostics: 'draws_array' num [1:1000, 1:4, 1:6] 1 1 1 1 0.888 ...
+ $ post_warmup_sampler_diagnostics: 'draws_array' num [1:1000, 1:4, 1:6] 0.86 1 0.94 1 0.999 ...
   ..- attr(*, "dimnames")=List of 3
   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
   .. ..$ variable : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
-
-

-as_cmdstan_fit()

-

If you need to manually create fitted model objects from CmdStan CSV files use as_cmdstan_fit().

+
+

as_cmdstan_fit() +

+

If you need to manually create fitted model objects from CmdStan CSV +files use as_cmdstan_fit().

-fit2 <- as_cmdstan_fit(fit$output_files())
-

This is pointless in our case since we have the original fit object, but this function can be used to create fitted model objects (CmdStanMCMC, CmdStanMLE, etc.) from any CmdStan CSV files.

+fit2 <- as_cmdstan_fit(fit$output_files())
+

This is pointless in our case since we have the original +fit object, but this function can be used to create fitted +model objects (CmdStanMCMC, CmdStanMLE, etc.) +from any CmdStan CSV files.

-
-

-Saving and accessing advanced algorithm info (latent dynamics)

-

If save_latent_dynamics is set to TRUE when running the $sample() method then additional CSV files are created (one per chain) that provide access to quantities used under the hood by Stan’s implementation of dynamic Hamiltonian Monte Carlo.

-

CmdStanR does not yet provide a special method for processing these files but they can be read into R using R’s standard CSV reading functions.

+
+

Saving and accessing advanced algorithm info (latent dynamics) +

+

If save_latent_dynamics is set to TRUE when +running the $sample() method then additional CSV files are +created (one per chain) that provide access to quantities used under the +hood by Stan’s implementation of dynamic Hamiltonian Monte Carlo.

+

CmdStanR does not yet provide a special method for processing these +files but they can be read into R using R’s standard CSV reading +functions.

-fit <- mod$sample(data = data_list, save_latent_dynamics = TRUE)
+fit <- mod$sample(data = data_list, save_latent_dynamics = TRUE)
-fit$latent_dynamics_files()
-
[1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-diagnostic-202203181228-1-263c4e.csv"
-[2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-diagnostic-202203181228-2-263c4e.csv"
-[3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-diagnostic-202203181228-3-263c4e.csv"
-[4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-diagnostic-202203181228-4-263c4e.csv"
+fit$latent_dynamics_files()
+
[1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-diagnostic-202307251455-1-156ae3.csv"
+[2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-diagnostic-202307251455-2-156ae3.csv"
+[3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-diagnostic-202307251455-3-156ae3.csv"
+[4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-diagnostic-202307251455-4-156ae3.csv"
-# read one of the files in
-x <- utils::read.csv(fit$latent_dynamics_files()[1], comment.char = "#")
-head(x)
+# read one of the files in +x <- utils::read.csv(fit$latent_dynamics_files()[1], comment.char = "#") +head(x)
      lp__ accept_stat__ stepsize__ treedepth__ n_leapfrog__ divergent__
-1 -7.29215      0.999807   0.944718           2            3           0
-2 -6.81105      0.988812   0.944718           1            3           0
-3 -8.51410      0.694033   0.944718           1            3           0
-4 -7.89903      1.000000   0.944718           1            1           0
-5 -8.72776      0.925309   0.944718           1            1           0
-6 -6.87320      1.000000   0.944718           2            3           0
-  energy__    theta    p_theta   g_theta
-1  7.38144 -1.84247 -0.6346060 -1.358890
-2  7.27073 -1.34029  1.4398900 -0.510444
-3  9.26904 -2.53206 -1.8452700 -2.116310
-4  8.47008 -2.22096  1.6048800 -1.825390
-5  8.72848 -2.63117  0.0571749 -2.194090
-6  8.50747 -1.44232 -2.7149800 -0.705762
-

The column lp__ is also provided via fit$draws(), and the columns accept_stat__, stepsize__, treedepth__, n_leapfrog__, divergent__, and energy__ are also provided by fit$sampler_diagnostics(), but there are several columns unique to the latent dynamics file.

+1 -6.77840 0.964447 0.816376 2 3 0 +2 -6.74817 0.998632 0.816376 1 3 0 +3 -6.74817 0.828194 0.816376 1 3 0 +4 -6.75537 0.956585 0.816376 2 3 0 +5 -7.01137 0.961708 0.816376 2 3 0 +6 -6.77984 0.967943 0.816376 1 3 0 + energy__ theta p_theta g_theta +1 7.09195 -1.265290 -1.074960 -0.3592230 +2 6.78971 -1.087340 -0.391278 0.0254425 +3 7.76283 -1.087340 -1.933750 0.0254425 +4 7.08469 -1.018340 -1.101660 0.1842200 +5 7.09451 -0.631842 -0.553509 1.1651200 +6 7.23052 -1.269250 -1.288770 -0.3673810 +

The column lp__ is also provided via +fit$draws(), and the columns accept_stat__, +stepsize__, treedepth__, +n_leapfrog__, divergent__, and +energy__ are also provided by +fit$sampler_diagnostics(), but there are several columns +unique to the latent dynamics file.

-head(x[, c("theta", "p_theta", "g_theta")])
-
     theta    p_theta   g_theta
-1 -1.84247 -0.6346060 -1.358890
-2 -1.34029  1.4398900 -0.510444
-3 -2.53206 -1.8452700 -2.116310
-4 -2.22096  1.6048800 -1.825390
-5 -2.63117  0.0571749 -2.194090
-6 -1.44232 -2.7149800 -0.705762
-

Our model has a single parameter theta and the three columns above correspond to theta in the unconstrained space (theta on the constrained space is accessed via fit$draws()), the auxiliary momentum p_theta, and the gradient g_theta. In general, each of these three columns will exist for every parameter in the model.

+head(x[, c("theta", "p_theta", "g_theta")]) +
      theta   p_theta    g_theta
+1 -1.265290 -1.074960 -0.3592230
+2 -1.087340 -0.391278  0.0254425
+3 -1.087340 -1.933750  0.0254425
+4 -1.018340 -1.101660  0.1842200
+5 -0.631842 -0.553509  1.1651200
+6 -1.269250 -1.288770 -0.3673810
+

Our model has a single parameter theta and the three +columns above correspond to theta in the +unconstrained space (theta on the constrained +space is accessed via fit$draws()), the auxiliary momentum +p_theta, and the gradient g_theta. In general, +each of these three columns will exist for every parameter in +the model.

-
-

-Saving fitted model objects

-

As described above, the contents of the CSV files are only read into R when they are needed. This means that in order to save a fitted model object containing all of the posterior draws and sampler diagnostics you should either make sure to call fit$draws() and fit$sampler_diagnostics() before saving the object fit, or use the special $save_object() method provided by CmdStanR, which will ensure that everything has been read into R before saving the object using saveRDS().

+
+

Saving fitted model objects +

+

As described above, the contents of the CSV files are only read into +R when they are needed. This means that in order to save a fitted model +object containing all of the posterior draws and sampler +diagnostics you should either make sure to call fit$draws() +and fit$sampler_diagnostics() before saving the object +fit, or use the special $save_object() method +provided by CmdStanR, which will ensure that everything has been read +into R before saving the object using saveRDS().

-temp_rds_file <- tempfile(fileext = ".RDS") # temporary file just for demonstration
-fit$save_object(file = temp_rds_file)
-

We can check that this worked by removing fit and loading it back in from the save file.

+temp_rds_file <- tempfile(fileext = ".RDS") # temporary file just for demonstration +fit$save_object(file = temp_rds_file)
+

We can check that this worked by removing fit and +loading it back in from the save file.

-rm(fit); gc()
-
          used (Mb) gc trigger (Mb) limit (Mb) max used  (Mb)
-Ncells 1159655 62.0    2208007  118         NA  2208007 118.0
-Vcells 2197871 16.8    8388608   64      32768  3895933  29.8
+rm(fit); gc()
+
          used (Mb) gc trigger  (Mb) limit (Mb) max used  (Mb)
+Ncells 1286546 68.8    2431888 129.9         NA  2431888 129.9
+Vcells 2365428 18.1    8388608  64.0      32768  4130519  31.6
-fit <- readRDS(temp_rds_file)
-fit$summary()
-
# A tibble: 2 × 10
+fit <- readRDS(temp_rds_file)
+fit$summary()
+

[38;5;246m# A tibble: 2 × 10
[39m
   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-  <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-1 lp__     -7.25  -6.97  0.732 0.309 -8.71   -6.75   1.00    2009.    1714.
-2 theta     0.250  0.235 0.118 0.118  0.0821  0.463  1.00    1445.    1640.
+ 
[3m
[38;5;246m<chr>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m 
[3m
[38;5;246m<num>
[39m
[23m +
[38;5;250m1
[39m lp__ -
[31m7
[39m
[31m.
[39m
[31m28
[39m -
[31m7
[39m
[31m.
[39m
[31m00
[39m 0.751 0.339 -
[31m8
[39m
[31m.
[39m
[31m81
[39m -
[31m6
[39m
[31m.
[39m
[31m75
[39m 1.00 
[4m1
[24m833. 
[4m2
[24m545. +
[38;5;250m2
[39m theta 0.253 0.239 0.121 0.125 0.080
[4m5
[24m 0.470 1.00 
[4m1
[24m594. 
[4m1
[24m821. -
-

-Developing using CmdStanR

-

CmdStanR can of course be used for developing other packages that require compiling and running Stan models as well as using new or custom Stan features available through CmdStan.

-
-

-Troubleshooting and debugging

-

When developing or testing new features it might be useful to have more information on how CmdStan is called internally and to see more information printed when compiling or running models. This can be enabled for an entire R session by setting the option "cmdstanr_verbose" to TRUE.

+
+

Developing using CmdStanR +

+

CmdStanR can of course be used for developing other packages that +require compiling and running Stan models as well as using new or custom +Stan features available through CmdStan.

+
+

Troubleshooting and debugging +

+

When developing or testing new features it might be useful to have +more information on how CmdStan is called internally and to see more +information printed when compiling or running models. This can be +enabled for an entire R session by setting the option +"cmdstanr_verbose" to TRUE.

-options("cmdstanr_verbose"=TRUE)
-
-mod <- cmdstan_model(stan_file, force_recompile = TRUE)
+options("cmdstanr_verbose"=TRUE) + +mod <- cmdstan_model(stan_file, force_recompile = TRUE)
Running make \
-  /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10 \
+  /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565 \
   "STANCFLAGS +=  --name='bernoulli_model'"
 
 --- Translating Stan model to C++ code ---
-bin/stanc --name='bernoulli_model' --o=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10.hpp /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10.stan
+bin/stanc --name='bernoulli_model' --o=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565.hpp /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565.stan
 
 --- Compiling, linking C++ code ---
-clang++ -std=c++1y -Wno-unknown-warning-option -Wno-tautological-compare -Wno-sign-compare -D_REENTRANT -Wno-ignored-attributes      -I stan/lib/stan_math/lib/tbb_2020.3/include    -O3 -I src -I stan/src -I lib/rapidjson_1.1.0/ -I lib/CLI11-1.9.1/ -I stan/lib/stan_math/ -I stan/lib/stan_math/lib/eigen_3.3.9 -I stan/lib/stan_math/lib/boost_1.75.0 -I stan/lib/stan_math/lib/sundials_6.0.0/include -I stan/lib/stan_math/lib/sundials_6.0.0/src/sundials    -DBOOST_DISABLE_ASSERTS          -c -include-pch stan/src/stan/model/model_header.hpp.gch -x c++ -o /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10.o /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10.hpp
-clang++ -std=c++1y -Wno-unknown-warning-option -Wno-tautological-compare -Wno-sign-compare -D_REENTRANT -Wno-ignored-attributes      -I stan/lib/stan_math/lib/tbb_2020.3/include    -O3 -I src -I stan/src -I lib/rapidjson_1.1.0/ -I lib/CLI11-1.9.1/ -I stan/lib/stan_math/ -I stan/lib/stan_math/lib/eigen_3.3.9 -I stan/lib/stan_math/lib/boost_1.75.0 -I stan/lib/stan_math/lib/sundials_6.0.0/include -I stan/lib/stan_math/lib/sundials_6.0.0/src/sundials    -DBOOST_DISABLE_ASSERTS                -Wl,-L,"/Users/jgabry/.cmdstan/cmdstan-2.29.1/stan/lib/stan_math/lib/tbb" -Wl,-rpath,"/Users/jgabry/.cmdstan/cmdstan-2.29.1/stan/lib/stan_math/lib/tbb"      /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10.o src/cmdstan/main.o        -Wl,-L,"/Users/jgabry/.cmdstan/cmdstan-2.29.1/stan/lib/stan_math/lib/tbb" -Wl,-rpath,"/Users/jgabry/.cmdstan/cmdstan-2.29.1/stan/lib/stan_math/lib/tbb"   stan/lib/stan_math/lib/sundials_6.0.0/lib/libsundials_nvecserial.a stan/lib/stan_math/lib/sundials_6.0.0/lib/libsundials_cvodes.a stan/lib/stan_math/lib/sundials_6.0.0/lib/libsundials_idas.a stan/lib/stan_math/lib/sundials_6.0.0/lib/libsundials_kinsol.a  stan/lib/stan_math/lib/tbb/libtbb.dylib stan/lib/stan_math/lib/tbb/libtbbmalloc.dylib stan/lib/stan_math/lib/tbb/libtbbmalloc_proxy.dylib -o /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10
-rm -f /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/model-dfb56119ce10.o
+clang++ -std=c++1y -Wno-unknown-warning-option -Wno-tautological-compare -Wno-sign-compare -D_REENTRANT -Wno-ignored-attributes -I stan/lib/stan_math/lib/tbb_2020.3/include -O3 -I src -I stan/src -I stan/lib/rapidjson_1.1.0/ -I lib/CLI11-1.9.1/ -I stan/lib/stan_math/ -I stan/lib/stan_math/lib/eigen_3.4.0 -I stan/lib/stan_math/lib/boost_1.78.0 -I stan/lib/stan_math/lib/sundials_6.1.1/include -I stan/lib/stan_math/lib/sundials_6.1.1/src/sundials -DBOOST_DISABLE_ASSERTS -c -include-pch stan/src/stan/model/model_header.hpp.gch -x c++ -o /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565.o /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565.hpp +clang++ -std=c++1y -Wno-unknown-warning-option -Wno-tautological-compare -Wno-sign-compare -D_REENTRANT -Wno-ignored-attributes -I stan/lib/stan_math/lib/tbb_2020.3/include -O3 -I src -I stan/src -I stan/lib/rapidjson_1.1.0/ -I lib/CLI11-1.9.1/ -I stan/lib/stan_math/ -I stan/lib/stan_math/lib/eigen_3.4.0 -I stan/lib/stan_math/lib/boost_1.78.0 -I stan/lib/stan_math/lib/sundials_6.1.1/include -I stan/lib/stan_math/lib/sundials_6.1.1/src/sundials -DBOOST_DISABLE_ASSERTS -Wl,-L,"/Users/jgabry/.cmdstan/cmdstan-2.32.2/stan/lib/stan_math/lib/tbb" -Wl,-rpath,"/Users/jgabry/.cmdstan/cmdstan-2.32.2/stan/lib/stan_math/lib/tbb" /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565.o src/cmdstan/main.o -Wl,-L,"/Users/jgabry/.cmdstan/cmdstan-2.32.2/stan/lib/stan_math/lib/tbb" -Wl,-rpath,"/Users/jgabry/.cmdstan/cmdstan-2.32.2/stan/lib/stan_math/lib/tbb" stan/lib/stan_math/lib/sundials_6.1.1/lib/libsundials_nvecserial.a stan/lib/stan_math/lib/sundials_6.1.1/lib/libsundials_cvodes.a stan/lib/stan_math/lib/sundials_6.1.1/lib/libsundials_idas.a stan/lib/stan_math/lib/sundials_6.1.1/lib/libsundials_kinsol.a stan/lib/stan_math/lib/tbb/libtbb.dylib stan/lib/stan_math/lib/tbb/libtbbmalloc.dylib stan/lib/stan_math/lib/tbb/libtbbmalloc_proxy.dylib -o /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565 +rm -f /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/model-17ac47a020565.o 
-fit <- mod$sample(
-  data = data_list,
-  chains = 1,
-  iter_warmup = 100,
-  iter_sampling = 100
-)
+fit <- mod$sample( + data = data_list, + chains = 1, + iter_warmup = 100, + iter_sampling = 100 +)
Running MCMC with 1 chain...
 
-Running ./bernoulli 'id=1' random 'seed=911816340' data \
-  'file=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/standata-dfb528f42a87.json' \
+Running ./bernoulli 'id=1' random 'seed=1523483778' data \
+  'file=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/standata-17ac41edf7bd5.json' \
   output \
-  'file=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-202203181228-1-036a9c.csv' \
-  'profile_file=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-profile-202203181228-1-7b073f.csv' \
+  'file=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-202307251455-1-7b6962.csv' \
+  'profile_file=/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-profile-202307251455-1-281032.csv' \
   'method=sample' 'num_samples=100' 'num_warmup=100' 'save_warmup=0' \
   'algorithm=hmc' 'engine=nuts' adapt 'engaged=1'
 Chain 1 method = sample (Default) 
@@ -790,16 +934,16 @@ 

 Chain 1     num_chains = 1 (Default) 
 Chain 1 id = 1 (Default) 
 Chain 1 data 
-Chain 1   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/standata-dfb528f42a87.json 
+Chain 1   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/standata-17ac41edf7bd5.json 
 Chain 1 init = 2 (Default) 
 Chain 1 random 
-Chain 1   seed = 911816340 
+Chain 1   seed = 1523483778 
 Chain 1 output 
-Chain 1   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-202203181228-1-036a9c.csv 
+Chain 1   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-202307251455-1-7b6962.csv 
 Chain 1   diagnostic_file =  (Default) 
 Chain 1   refresh = 100 (Default) 
 Chain 1   sig_figs = -1 (Default) 
-Chain 1   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMNhFrl/bernoulli-profile-202203181228-1-7b073f.csv 
+Chain 1   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpOiRpoh/bernoulli-profile-202307251455-1-281032.csv 
 Chain 1 num_threads = 1 (Default) 
 Chain 1 Gradient evaluation took 6e-06 seconds 
 Chain 1 1000 transitions using 10 leapfrog steps per transition would take 0.06 seconds. 
@@ -834,11 +978,13 @@ 

 
 
       

-  
Developed by Jonah Gabry, Rok Češnovar.

+  

+
Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

 

 
 

-  
Site built with pkgdown 1.6.1.

+  

+
Site built with pkgdown 2.0.7.

 

 
       

@@ -847,5 +993,7 @@ 

   
 
 
+  
+
   
 
diff --git a/docs/articles/cmdstanr.html b/docs/articles/cmdstanr.html
index 5c8946717..ac4bb8ffa 100644
--- a/docs/articles/cmdstanr.html
+++ b/docs/articles/cmdstanr.html
@@ -26,6 +26,8 @@
 
 
 
+    
+
     

       

+      

   

     

       
Getting started with CmdStanR

-                        
Jonah Gabry and Rok Češnovar

+                        
Jonah Gabry and
+Rok Češnovar

             
       
-      Source: vignettes/cmdstanr.Rmd
+      Source: vignettes/cmdstanr.Rmd
       
 
     

 
     
     
-

-

-Introduction

-
CmdStanR is a lightweight interface to Stan for R users (see CmdStanPy for Python) that provides an alternative to the traditional RStan interface. See the Comparison with RStan section later in this vignette for more details on how the two interfaces differ.

-
CmdStanR is not on CRAN yet, but the beta release can be installed by running the following command in R.

+

+
Introduction
+

+
CmdStanR (Command Stan R) is a lightweight interface to Stan for R users that provides an
+alternative to the traditional RStan interface. See the Comparison with RStan section
+later in this vignette for more details on how the two interfaces
+differ.

+
CmdStanR is not on CRAN yet, but the beta release
+can be installed by running the following command in R.

 
-# we recommend running this is a fresh R session or restarting your current session
-install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))

-
CmdStanR (the cmdstanr R package) can now be loaded like any other R package. We’ll also load the bayesplot and posterior packages to use later in examples.

+# we recommend running this is a fresh R session or restarting your current session
+install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
+

CmdStanR (the cmdstanr R package) can now be loaded +like any other R package. We’ll also load the bayesplot +and posterior packages to use later in examples.

-
-
-

-Installing CmdStan

-

CmdStanR requires a working installation of CmdStan, the shell interface to Stan. If you don’t have CmdStan installed then CmdStanR can install it for you, assuming you have a suitable C++ toolchain. The requirements are described in the CmdStan Guide:

+library(cmdstanr) +library(posterior) +library(bayesplot) +color_scheme_set("brightblue")
+ +
+

Installing CmdStan +

+

CmdStanR requires a working installation of CmdStan, +the shell interface to Stan. If you don’t have CmdStan installed then +CmdStanR can install it for you, assuming you have a suitable C++ +toolchain. The requirements are described in the CmdStan Guide:

-

To double check that your toolchain is set up properly you can call the check_cmdstan_toolchain() function:

+

To double check that your toolchain is set up properly you can call +the check_cmdstan_toolchain() function:

+check_cmdstan_toolchain()
The C++ toolchain required for CmdStan is setup properly!
-

If your toolchain is configured correctly then CmdStan can be installed by calling the install_cmdstan() function:

+

If your toolchain is configured correctly then CmdStan can be +installed by calling the install_cmdstan() +function:

-install_cmdstan(cores = 2)
-

Before CmdStanR can be used it needs to know where the CmdStan installation is located. When the package is loaded it tries to help automate this to avoid having to manually set the path every session:

+install_cmdstan(cores = 2) +

Before CmdStanR can be used it needs to know where the CmdStan +installation is located. When the package is loaded it tries to help +automate this to avoid having to manually set the path every +session:

    -

  1. If the environment variable "CMDSTAN" exists at load time then its value will be automatically set as the default path to CmdStan for the R session. This is useful if your CmdStan installation is not located in the default directory that would have been used by install_cmdstan() (see #2).

    2. -

  2. If no environment variable is found when loaded but any directory in the form ".cmdstan/cmdstan-[version]", for example ".cmdstan/cmdstan-2.23.0", exists in the user’s home directory (Sys.getenv("HOME"), not the current working directory) then the path to the CmdStan with the largest version number will be set as the path to CmdStan for the R session. This is the same as the default directory that install_cmdstan() uses to install the latest version of CmdStan, so if that’s how you installed CmdStan you shouldn’t need to manually set the path to CmdStan when loading CmdStanR.

    3. +

  3. If the environment variable "CMDSTAN" exists at load +time then its value will be automatically set as the default path to +CmdStan for the R session. This is useful if your CmdStan installation +is not located in the default directory that would have been used by +install_cmdstan() (see #2).

    4. +

  4. If no environment variable is found when loaded but any directory +in the form ".cmdstan/cmdstan-[version]", for example +".cmdstan/cmdstan-2.23.0", exists in the user’s home +directory (Sys.getenv("HOME"), not the current +working directory) then the path to the CmdStan with the largest version +number will be set as the path to CmdStan for the R session. This is the +same as the default directory that install_cmdstan() uses +to install the latest version of CmdStan, so if that’s how you installed +CmdStan you shouldn’t need to manually set the path to CmdStan when +loading CmdStanR.

-

If neither of these applies (or you want to subsequently change the path) you can use the set_cmdstan_path() function:

+

If neither of these applies (or you want to subsequently change the +path) you can use the set_cmdstan_path() function:

-set_cmdstan_path(PATH_TO_CMDSTAN)
-

To check the path to the CmdStan installation and the CmdStan version number you can use cmdstan_path() and cmdstan_version():

+set_cmdstan_path(PATH_TO_CMDSTAN) +

To check the path to the CmdStan installation and the CmdStan version +number you can use cmdstan_path() and +cmdstan_version():

-
[1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1"
+cmdstan_path() +
[1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2"
-
[1] "2.29.1"
- -
-

-Compiling a model

-

The cmdstan_model() function creates a new CmdStanModel object from a file containing a Stan program. Under the hood, CmdStan is called to translate a Stan program to C++ and create a compiled executable. Here we’ll use the example Stan program that comes with the CmdStan installation:

+cmdstan_version()
+
[1] "2.32.2"
+ +
+

Compiling a model +

+

The cmdstan_model() function creates a new CmdStanModel +object from a file containing a Stan program. Under the hood, CmdStan is +called to translate a Stan program to C++ and create a compiled +executable. Here we’ll use the example Stan program that comes with the +CmdStan installation:

-file <- file.path(cmdstan_path(), "examples", "bernoulli", "bernoulli.stan")
-mod <- cmdstan_model(file)
-

The object mod is an R6 reference object of class CmdStanModel and behaves similarly to R’s reference class objects and those in object oriented programming languages. Methods are accessed using the $ operator. This design choice allows for CmdStanR and CmdStanPy to provide a similar user experience and share many implementation details.

-

The Stan program can be printed using the $print() method:

+file <- file.path(cmdstan_path(), "examples", "bernoulli", "bernoulli.stan") +mod <- cmdstan_model(file)
+

The object mod is an R6 reference object of class CmdStanModel +and behaves similarly to R’s reference class objects and those in object +oriented programming languages. Methods are accessed using the +$ operator. This design choice allows for CmdStanR and CmdStanPy to provide a +similar user experience and share many implementation details.

+

The Stan program can be printed using the $print() +method:

-mod$print()
+mod$print() 
data {
   int<lower=0> N;
-  array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
+  array[N] int<lower=0,upper=1> y;
 }
 parameters {
   real<lower=0,upper=1> theta;
@@ -207,26 +249,31 @@ 

   theta ~ beta(1,1);  // uniform prior on interval 0,1
   y ~ bernoulli(theta);
 }
-

The path to the compiled executable is returned by the $exe_file() method:

+

The path to the compiled executable is returned by the +$exe_file() method:

-mod$exe_file()
-
[1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1/examples/bernoulli/bernoulli"
- -
-

-Running MCMC

-

The $sample() method for CmdStanModel objects runs Stan’s default MCMC algorithm. The data argument accepts a named list of R objects (like for RStan) or a path to a data file compatible with CmdStan (JSON or R dump).

+mod$exe_file()
+
[1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2/examples/bernoulli/bernoulli"
+ +
+

Running MCMC +

+

The $sample() +method for CmdStanModel +objects runs Stan’s default MCMC algorithm. The data +argument accepts a named list of R objects (like for RStan) or a path to +a data file compatible with CmdStan (JSON or R dump).

-# names correspond to the data block in the Stan program
-data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-
-fit <- mod$sample(
-  data = data_list, 
-  seed = 123, 
-  chains = 4, 
-  parallel_chains = 4,
-  refresh = 500 # print update every 500 iters
-)
+# names correspond to the data block in the Stan program +data_list <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1)) + +fit <- mod$sample( + data = data_list, + seed = 123, + chains = 4, + parallel_chains = 4, + refresh = 500 # print update every 500 iters +)
Running MCMC with 4 parallel chains...
 
 Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
@@ -260,79 +307,99 @@ 

 
 All 4 chains finished successfully.
 Mean chain execution time: 0.0 seconds.
-Total execution time: 0.3 seconds.
-

There are many more arguments that can be passed to the $sample() method. For details follow this link to its separate documentation page:

+Total execution time: 0.4 seconds. +

There are many more arguments that can be passed to the +$sample() method. For details follow this link to its +separate documentation page:

-

The $sample() method creates R6 CmdStanMCMC objects, which have many associated methods. Below we will demonstrate some of the most important methods. For a full list, follow this link to the CmdStanMCMC documentation:

+

The $sample() method creates R6 CmdStanMCMC objects, +which have many associated methods. Below we will demonstrate some of +the most important methods. For a full list, follow this link to the +CmdStanMCMC documentation:

-
-

-Posterior summary statistics

-
-

-Summaries from the posterior package

-

The $summary() method calls summarise_draws() from the posterior package. The first argument specifies the variables to summarize and any arguments after that are passed on to posterior::summarise_draws() to specify which summaries to compute, whether to use multiple cores, etc.

+
+

Posterior summary statistics +

+
+

Summaries from the posterior package +

+

The $summary() +method calls summarise_draws() from the +posterior package. The first argument specifies the +variables to summarize and any arguments after that are passed on to +posterior::summarise_draws() to specify which summaries to +compute, whether to use multiple cores, etc.

-fit$summary()
-
# A tibble: 2 × 10
-  variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-  <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-1 lp__     -7.27  -7.00  0.709 0.344 -8.70   -6.75   1.00    1852.    2114.
-2 theta     0.247  0.232 0.119 0.123  0.0804  0.466  1.00    1611.    1678.
-
-fit$summary(variables = c("theta", "lp__"), "mean", "sd")
-
# A tibble: 2 × 3
-  variable   mean    sd
-  <chr>     <dbl> <dbl>
-1 theta     0.247 0.119
-2 lp__     -7.27  0.709
-
-# use a formula to summarize arbitrary functions, e.g. Pr(theta <= 0.5)
-fit$summary("theta", pr_lt_half = ~ mean(. <= 0.5))
-
# A tibble: 1 × 2
-  variable pr_lt_half
-  <chr>         <dbl>
-1 theta         0.969
-
-
-

-CmdStan’s stansummary utility

-

CmdStan itself provides a stansummary utility that can be called using the $cmdstan_summary() method. This method will print summaries but won’t return anything.

-
-
-
-

-Posterior draws

-
-

-Extracting draws

-

The $draws() method can be used to extract the posterior draws in formats provided by the posterior package. Here we demonstrate only the draws_array and draws_df formats, but the posterior package supports other useful formats as well.

-
-# default is a 3-D draws_array object from the posterior package
-# iterations x chains x variables
-draws_arr <- fit$draws() # or format="array"
-str(draws_arr)
+fit$summary() +fit$summary(variables = c("theta", "lp__"), "mean", "sd") + +# use a formula to summarize arbitrary functions, e.g. Pr(theta <= 0.5) +fit$summary("theta", pr_lt_half = ~ mean(. <= 0.5)) + +# summarise all variables with default and additional summary measures +fit$summary( + variables = NULL, + posterior::default_summary_measures(), + extra_quantiles = ~posterior::quantile2(., probs = c(.0275, .975)) +)
+
  variable  mean median   sd  mad    q5   q95 rhat ess_bulk ess_tail
+1     lp__ -7.27  -7.00 0.71 0.34 -8.70 -6.75    1     1852     2114
+2    theta  0.25   0.23 0.12 0.12  0.08  0.47    1     1611     1678
+
  variable  mean   sd
+1    theta  0.25 0.12
+2     lp__ -7.27 0.71
+
  variable pr_lt_half
+1    theta       0.97
+
  variable  mean median   sd  mad    q5   q95  q2.75 q97.5
+1     lp__ -7.27  -7.00 0.71 0.34 -8.70 -6.75 -9.165 -6.75
+2    theta  0.25   0.23 0.12 0.12  0.08  0.47  0.065  0.52
+
+
+

CmdStan’s stansummary utility +

+

CmdStan itself provides a stansummary utility that can +be called using the $cmdstan_summary() method. This method +will print summaries but won’t return anything.

+
+
+
+

Posterior draws +

+
+

Extracting draws +

+

The $draws() +method can be used to extract the posterior draws in formats provided by +the posterior +package. Here we demonstrate only the draws_array and +draws_df formats, but the posterior +package supports other useful formats as well.

+
+# default is a 3-D draws_array object from the posterior package
+# iterations x chains x variables
+draws_arr <- fit$draws() # or format="array"
+str(draws_arr)
 'draws_array' num [1:1000, 1:4, 1:2] -6.78 -6.9 -7.05 -6.85 -6.75 ...
  - attr(*, "dimnames")=List of 3
   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
   ..$ chain    : chr [1:4] "1" "2" "3" "4"
   ..$ variable : chr [1:2] "lp__" "theta"
-
-# draws x variables data frame
-draws_df <- fit$draws(format = "df")
-str(draws_df)
+
+# draws x variables data frame
+draws_df <- fit$draws(format = "df")
+str(draws_df)
draws_df [4,000 × 5] (S3: draws_df/draws/tbl_df/tbl/data.frame)
  $ lp__      : num [1:4000] -6.78 -6.9 -7.05 -6.85 -6.75 ...
  $ theta     : num [1:4000] 0.284 0.186 0.162 0.196 0.252 ...
  $ .chain    : int [1:4000] 1 1 1 1 1 1 1 1 1 1 ...
  $ .iteration: int [1:4000] 1 2 3 4 5 6 7 8 9 10 ...
  $ .draw     : int [1:4000] 1 2 3 4 5 6 7 8 9 10 ...
-
-print(draws_df)
+
+print(draws_df)
# A draws_df: 1000 iterations, 4 chains, and 2 variables
    lp__ theta
 1  -6.8  0.28
@@ -347,41 +414,52 @@ 

 10 -7.5  0.42
 # ... with 3990 more draws
 # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
-

To convert an existing draws object to a different format use the posterior::as_draws_*() functions.

-
-# this should be identical to draws_df created via draws(format = "df")
-draws_df_2 <- as_draws_df(draws_arr)
-identical(draws_df, draws_df_2)
+

To convert an existing draws object to a different format use the +posterior::as_draws_*() functions.

+
+# this should be identical to draws_df created via draws(format = "df")
+draws_df_2 <- as_draws_df(draws_arr)
+identical(draws_df, draws_df_2)
[1] TRUE
-

In general, converting to a different draws format in this way will be slower than just setting the appropriate format initially in the call to the $draws() method, but in most cases the speed difference will be minor.

-
-
-

-Plotting draws

-

Plotting posterior distributions is as easy as passing the object returned by the $draws() method directly to plotting functions in our bayesplot package.

-
-mcmc_hist(fit$draws("theta"))
+

In general, converting to a different draws format in this way will +be slower than just setting the appropriate format initially in the call +to the $draws() method, but in most cases the speed +difference will be minor.

+
+
+

Plotting draws +

+

Plotting posterior distributions is as easy as passing the object +returned by the $draws() method directly to plotting +functions in our bayesplot +package.

+
+mcmc_hist(fit$draws("theta"))

-
-

-Sampler diagnostics

-
-

-Extracting diagnostic values for each iteration and chain

-

The $sampler_diagnostics() method extracts the values of the sampler parameters (treedepth__, divergent__, etc.) in formats supported by the posterior package. The default is as a 3-D array (iteration x chain x variable).

-
-# this is a draws_array object from the posterior package
-str(fit$sampler_diagnostics())
+
+

Sampler diagnostics +

+
+

Extracting diagnostic values for each iteration and chain +

+

The $sampler_diagnostics() +method extracts the values of the sampler parameters +(treedepth__, divergent__, etc.) in formats +supported by the posterior package. The default is as a +3-D array (iteration x chain x variable).

+
+# this is a draws_array object from the posterior package
+str(fit$sampler_diagnostics())
 'draws_array' num [1:1000, 1:4, 1:6] 1 2 2 2 2 1 1 1 1 2 ...
  - attr(*, "dimnames")=List of 3
   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
   ..$ chain    : chr [1:4] "1" "2" "3" "4"
   ..$ variable : chr [1:6] "treedepth__" "divergent__" "energy__" "accept_stat__" ...
-
-# this is a draws_df object from the posterior package
-str(fit$sampler_diagnostics(format = "df"))
+
+# this is a draws_df object from the posterior package
+str(fit$sampler_diagnostics(format = "df"))
draws_df [4,000 × 9] (S3: draws_df/draws/tbl_df/tbl/data.frame)
  $ treedepth__  : num [1:4000] 1 2 2 2 2 1 1 1 1 2 ...
  $ divergent__  : num [1:4000] 0 0 0 0 0 0 0 0 0 0 ...
@@ -393,12 +471,14 @@ 

  $ .iteration   : int [1:4000] 1 2 3 4 5 6 7 8 9 10 ...
  $ .draw        : int [1:4000] 1 2 3 4 5 6 7 8 9 10 ...
-
-

-Sampler diagnostic warnings and summaries

-

The $diagnostic_summary() method will display any sampler diagnostic warnings and return a summary of diagnostics for each chain.

-
-fit$diagnostic_summary()
+
+

Sampler diagnostic warnings and summaries +

+

The $diagnostic_summary() method will display any +sampler diagnostic warnings and return a summary of diagnostics for each +chain.

+
+fit$diagnostic_summary()
$num_divergent
 [1] 0 0 0 0
 
@@ -406,93 +486,113 @@ 

 [1] 0 0 0 0
 
 $ebfmi
-[1] 1.017555 1.250490 1.078559 1.237357
-

We see the number of divergences for each of the four chains, the number of times the maximum treedepth was hit for each chain, and the E-BFMI for each chain.

-

In this case there were no warnings, so in order to demonstrate the warning messages we’ll use one of the CmdStanR example models that suffers from divergences.

-
-fit_with_warning <- cmdstanr_example("schools")
-
Warning: 76 of 4000 (2.0%) transitions ended with a divergence.
+[1] 1.0 1.3 1.1 1.2
+

We see the number of divergences for each of the four chains, the +number of times the maximum treedepth was hit for each chain, and the +E-BFMI for each chain.

+

In this case there were no warnings, so in order to demonstrate the +warning messages we’ll use one of the CmdStanR example models that +suffers from divergences.

+
+fit_with_warning <- cmdstanr_example("schools")
+
Warning: 185 of 4000 (5.0%) transitions ended with a divergence.
 See https://mc-stan.org/misc/warnings for details.
-

After fitting there is a warning about divergences. We can also regenerate this warning message later using fit$diagnostic_summary().

-
-diagnostics <- fit_with_warning$diagnostic_summary()
-
Warning: 76 of 4000 (2.0%) transitions ended with a divergence.
+
After fitting there is a warning about divergences. We can also
+regenerate this warning message later using
+fit$diagnostic_summary().

+
+diagnostics <- fit_with_warning$diagnostic_summary()

+
Warning: 185 of 4000 (5.0%) transitions ended with a divergence.
 See https://mc-stan.org/misc/warnings for details.

-
-print(diagnostics)

+
+print(diagnostics)

 
$num_divergent
-[1] 22 19 33  2
+[1]   8 120  17  40
 
 $num_max_treedepth
 [1] 0 0 0 0
 
 $ebfmi
-[1] 0.3826859 0.3953390 0.2349030 0.2268031

-
-# number of divergences reported in warning is the sum of the per chain values
-sum(diagnostics$num_divergent) 

-
[1] 76

-
-
-

-CmdStan’s diagnose utility

-

CmdStan itself provides a diagnose utility that can be called using the $cmdstan_diagnose() method. This method will print warnings but won’t return anything.

-
-
-
-

-Create a stanfit object

-

If you have RStan installed then it is also possible to create a stanfit object from the csv output files written by CmdStan. This can be done by using rstan::read_stan_csv() in combination with the $output_files() method of the CmdStanMCMC object. This is only needed if you want to fit a model with CmdStanR but already have a lot of post-processing code that assumes a stanfit object. Otherwise we recommend using the post-processing functionality provided by CmdStanR itself.

+[1] 0.29 0.24 0.26 0.30 +
+# number of divergences reported in warning is the sum of the per chain values
+sum(diagnostics$num_divergent)
+
[1] 185
+
+
+

CmdStan’s diagnose utility +

+

CmdStan itself provides a diagnose utility that can be +called using the $cmdstan_diagnose() method. This method +will print warnings but won’t return anything.

+
+
+
+

Create a stanfit object +

+

If you have RStan installed then it is also possible to create a +stanfit object from the csv output files written by +CmdStan. This can be done by using rstan::read_stan_csv() +in combination with the $output_files() method of the +CmdStanMCMC object. This is only needed if you want to fit +a model with CmdStanR but already have a lot of post-processing code +that assumes a stanfit object. Otherwise we recommend using +the post-processing functionality provided by CmdStanR itself.

+
+stanfit <- rstan::read_stan_csv(fit$output_files())
+
+
+
+

Running optimization and variational inference +

+

CmdStanR also supports running Stan’s optimization algorithms and its +algorithms for variational approximation of full Bayesian inference. +These are run via the $optimize() and +$variational() methods, which are called in a similar way +to the $sample() method demonstrated above.

+
+

Optimization +

+

We can find the (penalized) maximum likelihood estimate (MLE) using +$optimize().

-stanfit <- rstan::read_stan_csv(fit$output_files())
-
-
-
-

-Running optimization and variational inference

-

CmdStanR also supports running Stan’s optimization algorithms and its algorithms for variational approximation of full Bayesian inference. These are run via the $optimize() and $variational() methods, which are called in a similar way to the $sample() method demonstrated above.

-
-

-Optimization

-

We can find the (penalized) maximum likelihood estimate (MLE) using $optimize().

-
-fit_mle <- mod$optimize(data = data_list, seed = 123)
+fit_mle <- mod$optimize(data = data_list, seed = 123) +fit_mle$summary() # includes lp__ (log prob calculated by Stan program) +fit_mle$mle("theta")
Initial log joint probability = -9.51104 
     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
 Optimization terminated normally:  
   Convergence detected: relative gradient magnitude is below tolerance 
-Finished in  0.1 seconds.
-
-fit_mle$summary() # includes lp__ (log prob calculated by Stan program)
-
# A tibble: 2 × 2
-  variable estimate
-  <chr>       <dbl>
-1 lp__        -5.00
-2 theta        0.2
-
-fit_mle$mle("theta")
-
theta 
-  0.2
-

Here’s a plot comparing the penalized MLE to the posterior distribution of theta.

-
-mcmc_hist(fit$draws("theta")) + 
-  vline_at(fit_mle$mle("theta"), size = 1.5)
+Finished in 0.2 seconds. +
  variable estimate
+1     lp__     -5.0
+2    theta      0.2
+
theta 
+  0.2
+

Here’s a plot comparing the penalized MLE to the posterior +distribution of theta.

+
+mcmc_hist(fit$draws("theta")) +
+  vline_at(fit_mle$mle("theta"), size = 1.5)

-
-

-Variational Bayes

-

We can run Stan’s experimental variational Bayes algorithm (ADVI) using the $variational() method.

-
-fit_vb <- mod$variational(data = data_list, seed = 123, output_samples = 4000)
+
+

Variational Bayes +

+

We can run Stan’s experimental variational Bayes algorithm (ADVI) +using the $variational() +method.

+
+fit_vb <- mod$variational(data = data_list, seed = 123, output_samples = 4000)
+fit_vb$summary("theta")
------------------------------------------------------------ 
 EXPERIMENTAL ALGORITHM: 
   This procedure has not been thoroughly tested and may be unstable 
   or buggy. The interface is subject to change. 
 ------------------------------------------------------------ 
-Gradient evaluation took 4e-06 seconds 
-1000 transitions using 10 leapfrog steps per transition would take 0.04 seconds. 
+Gradient evaluation took 6e-06 seconds 
+1000 transitions using 10 leapfrog steps per transition would take 0.06 seconds. 
 Adjust your expectations accordingly! 
 Begin eta adaptation. 
 Iteration:   1 / 250 [  0%]  (Adaptation) 
@@ -509,92 +609,115 @@ 

 Drawing a sample of size 4000 from the approximate posterior...  
 COMPLETED. 
 Finished in  0.1 seconds.
-
-fit_vb$summary("theta")
-
# A tibble: 1 × 7
-  variable  mean median    sd   mad    q5   q95
-  <chr>    <dbl>  <dbl> <dbl> <dbl> <dbl> <dbl>
-1 theta    0.267  0.250 0.117 0.117 0.105 0.487
-

The $draws() method can be used to access the approximate posterior draws. Let’s extract the draws, make the same plot we made after MCMC, and compare the two. In this trivial example the distributions look quite similar, although the variational approximation slightly underestimates the posterior standard deviation.

-
-mcmc_hist(fit$draws("theta"), binwidth = 0.025)
+
  variable mean median   sd  mad  q5  q95
+1    theta 0.27   0.25 0.12 0.12 0.1 0.49
+

The $draws() method can be used to access the +approximate posterior draws. Let’s extract the draws, make the same plot +we made after MCMC, and compare the two. In this trivial example the +distributions look quite similar, although the variational approximation +slightly underestimates the posterior standard deviation.

+
+mcmc_hist(fit$draws("theta"), binwidth = 0.025)
Posterior from MCMC

Posterior from MCMC

-
-mcmc_hist(fit_vb$draws("theta"), binwidth = 0.025)
+
+mcmc_hist(fit_vb$draws("theta"), binwidth = 0.025)
Posterior from variational

Posterior from variational

-

For more details on the $optimize() and $variational() methods, follow these links to their documentation pages.

+

For more details on the $optimize() and +$variational() methods, follow these links to their +documentation pages.

-
-

-Saving fitted model objects

-

In order to save a fitted model object to disk and ensure that all of the contents are available when reading the object back into R, we recommend using the $save_object() method provided by CmdStanR. The reason for this is discussed in detail in the vignette How does CmdStanR work?, so here we just demonstrate how to use the method.

-
-fit$save_object(file = "fit.RDS")
-
-# can be read back in using readRDS
-fit2 <- readRDS("fit.RDS")
-
-
-

-Comparison with RStan

-
-

-Different ways of interfacing with Stan’s C++

-

The RStan interface (rstan package) is an in-memory interface to Stan and relies on R packages like Rcpp and inline to call C++ code from R. On the other hand, the CmdStanR interface does not directly call any C++ code from R, instead relying on the CmdStan interface behind the scenes for compilation, running algorithms, and writing results to output files.

-
-
-

-Advantages of RStan

-
    -
  • Advanced features. We are working on making these available outside of RStan but currently they are only available to R users via RStan: +
    +

    Saving fitted model objects +

    +

    In order to save a fitted model object to disk and ensure that all of +the contents are available when reading the object back into R, we +recommend using the $save_object() +method provided by CmdStanR. The reason for this is discussed in detail +in the vignette How +does CmdStanR work?, so here we just demonstrate how to use the +method.

    +
    +fit$save_object(file = "fit.RDS")
+
+# can be read back in using readRDS
+fit2 <- readRDS("fit.RDS")
    +
    +
    +

    Comparison with RStan +

    +
    +

    Different ways of interfacing with Stan’s C++ +

    +

    The RStan interface (rstan package) is +an in-memory interface to Stan and relies on R packages like +Rcpp and inline to call C++ code from +R. On the other hand, the CmdStanR interface does not directly call any +C++ code from R, instead relying on the CmdStan interface behind the +scenes for compilation, running algorithms, and writing results to +output files.

    +
    +
    +

    Advantages of RStan +

    -
    • -
  • Allows other developers to distribute R packages with pre-compiled Stan programs (like rstanarm) on CRAN.
    • +

  • Allows other developers to distribute R packages with +pre-compiled Stan programs (like rstanarm) on +CRAN.

    • +

  • Avoids use of R6 classes, which may result in more familiar +syntax for many R users.

-
-

-Advantages of CmdStanR

+
+

Advantages of CmdStanR +

    -

  • Compatible with latest versions of Stan. Keeping up with Stan releases is complicated for RStan, often requiring non-trivial changes to the rstan package and new CRAN releases of both rstan and StanHeaders. With CmdStanR the latest improvements in Stan will be available from R immediately after updating CmdStan using cmdstanr::install_cmdstan().

    • -

  • Fewer installation issues (e.g., no need to mess with Makevars files).

    • -

  • Running Stan via external processes results in fewer unexpected crashes, especially in RStudio.

    • +

  • Compatible with latest versions of Stan. Keeping up with Stan +releases is complicated for RStan, often requiring non-trivial changes +to the rstan package and new CRAN releases of both +rstan and StanHeaders. With CmdStanR +the latest improvements in Stan will be available from R immediately +after updating CmdStan using +cmdstanr::install_cmdstan().

    • +

  • Running Stan via external processes results in fewer unexpected +crashes, especially in RStudio.

  • Less memory overhead.

    • -

  • More permissive license. RStan uses the GPL-3 license while the license for CmdStanR is BSD-3, which is a bit more permissive and is the same license used for CmdStan and the Stan C++ source code.

    • +

  • More permissive license. RStan uses the GPL-3 license while the +license for CmdStanR is BSD-3, which is a bit more permissive and is the +same license used for CmdStan and the Stan C++ source code.

-
-

-Additional resources

-

There are additional vignettes available that discuss other aspects of using CmdStanR. These can be found online at the CmdStanR website:

+
+

Additional resources +

+

There are additional vignettes available that discuss other aspects +of using CmdStanR. These can be found online at the CmdStanR +website:

To ask a question please post on the Stan forums:

-

To report a bug, suggest a feature (including additions to these vignettes), or to start contributing to CmdStanR development (new contributors welcome!) please open an issue on GitHub:

+

To report a bug, suggest a feature (including additions to these +vignettes), or to start contributing to CmdStanR development (new +contributors welcome!) please open an issue on GitHub:

@@ -610,11 +733,13 @@

-

Developed by Jonah Gabry, Rok Češnovar.

+

+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

+

Site built with pkgdown 2.0.7.

@@ -623,5 +748,7 @@

+ + diff --git a/docs/articles/cmdstanr_files/figure-html/plot-mle-1.png b/docs/articles/cmdstanr_files/figure-html/plot-mle-1.png index 1998cf9a6d4336390c48fd73053000863dd500de..ea9c074429a91760ed53dbd3b60bbe74b9aee4fd 100644 GIT binary patch literal 41641 zcmeFabyQSe_&16p3Wy4dh)An|NUD?)gNcZMii9-M-CfQX5S3I~QV~#6dT68(3F#g} zLSkqJ7#Qvzloam$M#kOx+Y)%co$J@Y!H|}vDn7@(fvzNIeYQwY%BLSn+kdho14rge zIMGyUUGSSLFO~1GT#WO;ly<(v8$Q1vhl^)6)#=60^(m3*T=ktQ?)Z1&=Nt2#(N39n zy7fvsoh7`}I2=wP4#F1#FPR4Pxe-T5{p>P-nark>Rz|$0k*z~8eI7?KH_z4d$h2Uo znv`80<%s#`IG!Ix%&|?uTEqxvYm2Y^-^!7ndoh=>}exBn9B7hBYM z>MvqCmtIzT%4l;}%)DHL51rP8I3QWIbgsA7UyIg!eS-z*J}`4aI##e$=5u)9%~fUA ztW)Lm0xy(8kD6Cj^J_6b04v~QI64459&>?E4v@T zL)=!^R;TyPztvFbx`UyNd+uH7h};&lU*;FI&g!z7S9UaonbxfpD-XARQ6A=zJBq<` zJm@smDV@bN#`#_Ph(nH%e5vrPfVjF~G^>=pYFgBXr(Dv@5SnTnhivOye67RsD?d?H z(6@S3sJAo>P6ItZ1^a5vB1|Jp5YH~o$<-e%zDo|WftS%}sLi)mmK$YFrk|&~Mb=VYJF#tjqFTKb?* z?gkuYw2v8>yn0aFaKPof2$cMst3j@<%fx()rK6}oM>4G3Y+9fH&@+2wdrrYq&|+g} za)f)qqw(n`(_g98$NN2W!qPL|jlM3a`Z@6INb;3Sr;iR6MP3}s8Of#m5pz5vP*p;v7kW)C(YaNv<^2iNGW3Cq-4ssTXoYRHOc ziizP|0@wIB2i$aUaKY69@c%6MkAnk=glyb7)D?Mfd*^lk*n=TPrEedVl6adY$-}SK zY2p0$(|G9nwTM{~t`$?>t7jy{%UGV* z$KCs7+Ir{bMfrHQ>Gx_Gy#BnX`|@8s{^`h;m>7d5D^4|q9*p= zZ^RI*id~c3d~)Q5P;bw{4=48&1+9=!FY?Uhv%4+?@bSCsgr5AjPt*8aR0sEIH4^Ut z%#P^g{XL|=-UlpKZMRl7`kteuHJ&{1h+rS_(;8bH-KSL;HPD*??WG(0G>wb&+DG#L zgGxPf?)SuIFL0do{8m*})pR~FA>n}omCVM3H)o@88lMCXwVC-yNkgNIWbLY`uCCUw zvO?Wg_NIZd#m2_U)XoyX5+D>56dHDR=;p#QB;q3p%mV^l=_9H}6|Sv*mS0<$t*chy z2wJvH66(!7bo%t^Z_ao`N*U9Gfyx1CX=$B?RTCs5cU1KB;`I#;3zC{wZ_fXS5>AGq z4#37CC}cNtH~R`|Vc{DFKQ8$5fBv>#XP*^A|wsV$jC@d zOPjq!IDX|uY#xM&MlhizFV^q#-FS|@rJR0&c<2O@^?isx)}HzaR@|Un_onet%Bx6| zr9rnwsD?Pd3Dt=mVu3A@5X0*b;$pY4I^`i=tXazpJ7>J6<%G&ET0M-mAl+#u9E1m8 zXcoWj!`LCXkjW~t`4$NnhE(|CWq6mHpc~0|4ejCUqV7k(-IW32G)VRg4+#y+6YmvG z<{Ffx-|5Hi59Mj=j?8O&?d=Y!z**At??>$I5yWyrz0|e>@6PXR^UuRD7U142Aq|YX zhWF=Tm(#>>Hzn_(y}^Oj{dtiP2IlRLz| zF$^JX;Ku)^yiIifw7g*(f4ty-SHp&>jvzUOu`Z*7D~PsDSqVXO7R)<3Eelg*^*Ha2 z!!T~(MFz<3Cg89miYk-Z?^k-(4%7DSS9R80~t3wYQvG{fM-p*K01Ll9yp7! zmEKHy@65F{^x~2nQ=O)jnc-jjWC~ye=3Ox!78L3Oix~NQ^Hu{Os4F&igX_>|Z&PhE z!*wJmIt5YUv^c^-whO5KfXV^LBmmDsiSnh*8#^yljsUPcCCheyZ{Wb~VWc&7itv~v z*o;s+EYAFJ9c+RtfcQ|foP+4A{m;opLt|#rjKe*|ltKjeuyY|lbq<<-V47EC zxZ}dq1cVGF%Es1AHHM5((6)vWLCB!WmQXyt%wg3i%@JHS1ba5|Q$!!;j+7^zGa{osPE-A@%h1(!vVdg(oxDud0k;sw^@T*nEXhPIjoF+bT+ja`13@PXzJ6 zkbyDC!0(;X=#Yw%6X>)jf7#e?>7E4vg(uHGYH!sDIYCszAIDHIm6*M^sxxp{@_RpA zy}BG!ltp0f6xW?^QMWdB?RlZtx7Hqd4##=Z(=vs^1!JbQXkOSfZilFweP|!Q61pGr z3Qa!?@3z#+fh(KR^a%kTjx(;KXone>xk@NOrQP&iVOd_w1LlS@ND0 zrRHIERS3%2*?DHXJxx$+EMFS+j5{|sS7uZYkfYG>aHVT?jJT;s%#%gFRAg?{V((XHnE{9SW|_h9aeXCW zE*UjQ_v{BHN5%ix{OspO9;qEE57UC#p^*8u$c5F_)rS0QCxY-`NSnk-M4^+0AKTcq z>(_Hk!zJ+5Nk^2NOtrLzn;no9-ZWZNw-6*SHzFEov#E!ry6H#uk*g?hn1DAg+% z69cVr^8XEcqBSD1iY*I~M#>Hn#YIrMd8)7A+BAfK^X!(nS9b%CPfKZE-x*AfFY ztknEsRl{}?2$(lWb)0a+&iaHo4mCXyZ6C$gRmvgF0=+Et+k`Xy~M<6(!P-8HgxHKC;@0hLJCxpNf4E_ipFi zjV|g#p9{lzvP=(>(E2>;{hI%{Mf#alfz!ipGEndCDCA$vj0?mIE{kRcFck@iyYKLX zlG5mB+2|{Z-r~DEfCYeOS&b0mxVFTL{Yy8O3cdeh{5lylo6Dy)NCp*fsE`T0voVeh zbfgvP)dVUJ$<97~0D{0@>C<8!R9=k@GpAk~kx}wu4$)@N@b-sH#8V`if}3g z#N2tcLy8gmM7;;t+QGq{FKsIAU+rLn9sMt<9dO(Rnt%o1ii_99t|FlFm!OL@6n(r^h2kDPay+`7t!ArLEJ$N`8LL1=iGH@dJumaRL4WYeHGFJJ{g=?)@sx1I}Kg7rOP z5+x`&Sipb3Ebi%UI{;Qs&>s_nQnQBOULw#>@;6;mitZN(SoN*5Wr;bp+2Y29kkkS8 zcXIwRCc7K@ZGGBzKyBw+OwB&2pA@ieDI3zg85B$!3_4N2BlPciWA8i!KBx4r;|+Ty z{wL~9QT)F6XIwi3^3R<%FzZiX{0AUA1@aF-cKFKy$UgwtA&~#?0*H!i#`38$Jm(wa z$4k;@Yir5K$Pm||wsl5a<3}Ngb6>s~WBA}Y7y&kx<^1j}pXHMWK+bSCYL=mnVbW>F zoc>0|Fsos*S2er7_kENAy1%JgCVOS6Y^jH&^5RsITD!xfM@Z8w8kfy$8Yu`@Lp$8B ztz(!<>iWo6cRAd^ciADLNX*n1)ohAcIpQJ$Z*6J0j|!MD2a6T@tYd*ic!vT&i6=I~ zartYt1zB59PD`1>O8`QuhrA?e5QLq>A)c$w%1na@%s~7w5{Xn2DAB?l zx)+H)fXYEh3~B2}angQCNqLmR!w8wZWIS9OWj)Sy-^V zCMd`z5ze12{Y$GZr|?00c8i80cHD0col4s|(Bpm1wrJ*VKO_RNZeH?@H6YBmwRIvLz*Qt9c_r%ZNs zb{bAjGoIbAiYW2Ty_fUBf{wpPxS)famzP&k=r9T7Dd=^$t-amG|Bz*Hq@>b%IYUqI zL9dd*P1lVIVC$)on^wCrB4fz|Q0bm0weveNtF3&sh#piK>5tOvED&sBBP;|YY(|%c zZTF_3LcMQ6QQ+hYqKzT!daOtca9kdLGl9JcGOe+QP_H9FiQBH%HZ<59gH58pWm9jC zXXgd{0iT9*WK6B|_U-e_-Y+Us`|P7=eFi>K&zHoEEd@XugwF{VlDW#~Od z$VPzQ#jd*Ovrp??PgJZMnR7PBUNn9PWHv^#70Sn|3(?LsSI!*$R$jhj!>*JNIBm_3 zj!80ccbYV!4p3g63mHl_u_2?|%uGR@;x1V0DNp?M$GWwpd*{aO=V+WVp5$c3$JWG0 z99|yh>(qjvgwXXzwoS+aq!AVhQX6%tC#15W{Fi&bnKp-vL@!oHU9o9$V2?hvqbF<7 z(;}(P#=NF7i<2;fT7h;dv7g|$b%#+>(f34TK^Gj2dK)Rc(=EbmO-Tz#TvJ% zqn%H2GJ&F=>BI|ACffOj^%^&<0ND1wDQ~lae_GyUqW=tg7i|9X<=sf~e<*yFFvN0m z!ipeuM0kVBYpr})4SmKYvM)rVJ?^>>E?N)E3D{$1%k&g$UH9N6-*cZAIrQ-87nnob zVx4Cq)~qK2(8!x}oyfibf-e4Ti1Ww1U`zmtCI8Fjg*`o@Z{(Lsyz(mIxKeA};J-EI)%JV6NaKw-xos)6o&eWrJb`|{vN=jxv#`f3K^L@cU>p_U=KB-~zk*qPCHuG!d8(uXdqYwf758RtLdGDb zP94(Ohx?d<L=8;Uy^Yg>}+`T+!t@utmn~ zLN9uNlC|;Tm3T1)0+7ZVY6myU*lnKbc;<0pry-W#sp?w^D!!fI-2)|dWeV7XS4wPV^t zy0R*~lsU=`Qy3_x?YNtt(y$9CE4AAhc71(2E#?cQgD05OcC zuFuij9Hd;8vB;(19(%)l@XyrJ@^rRAK0S$bL2Ev5@-SR1lY0V-%0+m%$B!%dY1N4f z8Z#*~Oa(&Xd&(=2c|en)KEJ1Y zVPV0)vlD?}-`ZN*jhaBBtpfj)0K&q;bZguQx^fTBFD}A6A{|aCDJeNJ8$&?`(72Ta=j$NMj)`@D()qr@~m#^{y0h@`G>22S5Sli0n#B|BO*F#q5_CUB_p*7r6&nljxy9(8Nnn5Oqh1=bua;e z&BXFO$%a|NeeQEg1<4eoiJ9MxH_Od-Kk2 zaD>^Mp2A1j^dz-!ypzRZ%zfRcw{WFOcd0QpWzrT&rReB^W-ZCK+2|U?g~vxqF`pUf zKgVnioelRc0vB1h=G#1jhGz4&+HHFp|9dP*_yke9m;T%ox(M@+t}^?8S2%2AEB)<< z$wls(ZRX|kfOgSapo%j?bVF8C;;^v9$4LExSy9zmeqNTry27K4UUxSJ4W@PPV(%SE z5;n2w-BbJ>#fZUfSZX82>-&ho1>nDnXycW=LtZJ^$Xn6*%4=J98#oC@P5T~%yuOC? zZHnDqO`Habo3{LLe!D!j=6dfUt+520lXV|P!Ve52FKh(6F0seFPJ*iIzbS7CllGsM zcbO>cpJ8u6$e$YjKVRO3B>&)R8-T1E$v?#26>k2Yf)q^EFF;yGrn;p?+Rw75y+Vb< zdTa$`uUDwO5!VsqxQ3UEbIcp{HkL~3KNR&Fr;H5VT4y1c!;tHCkR^wqk#nQo(GZf@?Rt|4$5lic&}`uR`q z6K$Q7n78K^l$|| z!cHP0B2m=r7ZR8a@d)*UrJ<;H>}G{uzkcz1XdQ+`qYphs1&Sa8e?re>W@W`6JHcQ) zS+I1s(D#SHhoPFH)xqjO%)lp0>&&N%y^-kI3oyHoP9CoHNd_rJqG6L|?n{JSx%ES) zy**Xm2SOW-y10h!MDi8n7UkRZ2LI0ck+9cBhmqav{$S%lFL4kdj7`4~zsq5=HoxH= zESF5I#zB81e3$zFaGn1^N!gj@CdD2^1VkMluuL0n=ajQ|VFpGE2o|k-{rbVs-TsD) z{}U|J4);VLH`(NlFh=_l5Uk2`>H8ck&IUDwY*g$H!nxH2J~#(Pl!|Y#b!QuOdoE-M zQ$;RV{N8_#?KDyhRx&V~y*B>qi~S^CM9huF9(Ec(#9{`S&T{kCPIs6NV0De!*({s9 z49sv~+U7iC)QzR_F@CGDLdRd;`p%R$J2t!fMC&sgZsXeLf8kFk0zy+%fy;hFhAm3+ zcd`$Tga{8-OiFLtEMW)=B89;rdRuD~_;g=5Y>F1cYKRl<^+?#2#33l7d>8huHG)d^ z-pOmTll8dkYT{4j+58s3u0TjkOAJn*Q>zfjZH{oq1#AVVY=w{Wz~&TpT;PHeCU~;{ zo!+6T%`RvQfy&m7i*3$y=fwday#oP}hjxXcKd)D^gDPF&Es}jyh66b9b9!yq-c~w8 zh~Kq2{mt=gT>P8zE`9ye@=nVHD|-J7dndyGkNPr9^*MFAP-IS#(aM71Qk5^I8(Lc- ztmN2UzW?wBtuW-Nsxf`%Noy0%i)dj?WK?2eNWkAfP^1rHLfX8gUKY!jw5Q*eJ6L(9 zVIk0Ktb^07G;kmGy|Qg$ae#(~^k?IkzT_!<)0`?=o%$5>ykn;=2b!+#CC}?}Ob=O*zLlD<3<}GXpgJuj9=?^iR~A zqWFFB&$xC9!4@FdLq2oY8}?ljcHYb3`S(JKYm_s&F(21 zgf6U>8L%F|PyV1MHS`7bV}A{)@Mc6%vTg7nqN{ctQD2jXa zv*=qIai@%!q|0nhbL@^SO{Fh3aPKzB)S{a<#g833MYjmX$5WuT?5Eerxz%(iM1Cx~ ztNS$qt>s;7m*^IKbX{cLJQbPt$eYV#?duu8_6J`Yv2v=~)e4Kn#D!`FBVa&qL@1PvvNo4{wGz`lpGT40Gj zAu;jKH!0wifjJ)aYuV&HydmKwxXA7UA=4DbR>PG{1x39R2>((exotUG%*it!SAuop zP7F^Ki;9Xy+nJr?NQQJDWC_Fn9wI~r*ZO{W*nsHK@QGV@yFCgvg|y>>)*qxx%{iR= zMEysHiut3zYja+zsHk`prZnFyK-M1OG-_pP+JCLt`{(}L;2xqX=x?*Xxfb#dV(YVgM)+2WrGYdy!LMuBGCoi4r@rI zx)i?5u$d!yOZjniBE`5!H#fJzrP*PJ+7`P+!xO*10%5n1@aHQK)3~+b&|pq_WA7~T ztxreIF-`rf@)fN9EcK|jC*LmG&TtHTO(M0hP#v@>c@{C z{tl_BDslF23gZF6Te_YyUZG{!A`O9}?KHbM{-+{g1 zXA_IJBjuLU*MpM{`OyZyKZmtL8|xB`15N+ycq1kpfcz8nPBV)9XIwi3^3R=iW@@(sa4}JAGgL!jtF{SdXkHeW@Rq{;L1&0aD=&7dYxLCWK_}sX$o}5XW@bCgu zOQM<6umD$Vu9L;6!XHc$>c+j$HFh`t*%n4^H^@rg=Pq{Pw(|1(#0EpiNu))T)a4ya zZnNs}Q$(m(Kq5aeaeI?nvEw1n0RmLckGP<{x#r9T+P?7Eo|gT-G!z1c)~zkIPf)Wt z>0?|cX{5)z_t2Wego^VG(=TnYo9%`N<0rTWAO4(`vHq3Ezr)%n18tuYj~wY<6%cR; zEgZ;nwyCpYpKp>`koXuUQzX9LTc)%UK&CdM1_m4X!KOIZFJL)fXsw>oWqZrihY`q> ze?BvFABNtI+c7GgwzKOprHe=b;qcmFcbfVnOJ~9HQQX< zkrsSDDI0j9OTIYE5h^H^V!#r#nyJXzS8$ou`%k4k_!Q8AdKSTsy@nJ=j!w2s$#!@p z$deT>*QVd0r{3tq0LV>m;M3{tw#ccV5>j)P`ewv|@A9XL|D$&ZAKLh+tq{YZE>14r zZiao%yxqfENQdoh{(VFx0z~z|G9+>zJs5x~>x`Y4-YNeLa{`@T=(^f&@ZUZbln&Gs zn#U@|Vn`>ekapMgJjK^IN%`|dq>{0m?}1g`t+s;1>I293iyZ5TF8Y*#Dy zfOn7_*=ONUNTnAvelTnH!FK1_X1YH3_u!1DxVNOf?QdWYb`h{NbHV#amFg==nrKOm4RM_Lpn*d{5 zkEle0TS?;DY)rC*pj@db22Lla^jEGoSSqI(pNh%YB>;)=LDmRuRzgwzxACwSl;bce>_M&`>I^4Pp z87TLRx9|`C#6CFbsb2ihXInSw!YV&2AF1}2KH?(c#ZZNo6}svje_tF!dWO$ZzcEv+ zWQLiT@wh}N@8v=qxq?TqrVCxUIXT5S7GvCJ^`mG!CY8$89uCq^Tn?;|tDdN><)tWj zeLI6wifo&;{~{e3S5RII$%xlHP##<~t+%>7PlpjmAl5z&L&oD+d-90Z=?b5^ICyvf zPNm#=-||q7w^W9{i0(-wvrN)acQW=QjZR%98DU}JAbuOuJNblUY!6l`4v`*c)z%vP zws?F#=xlS$+3l+TD6i5!GuW5W8qms8DdNlEx+pS9$}H`oQ612nphS8?`=e0k$HjoG zyy$Z-L1RzzZ7s?!^WmmmB8I-HnsjPdi?| zW;fj=CzDM>&L3W>SdA)4rlEM6){>r?`68hB6ufpm>ZaE4%XiK4aU-2hxg7djst!7> zR=;Jd@)B6O8-9nz4nLu_gD5H{SMVFz=)F|VU{9&-XwB7B6rH#wb^f;z)W-tvi@3cr z^uh87J(wOC(nz6Sjc{pr8XeEe{1hV<=kXg=J^`YUf1e|EhsOzFaPJ-`QY~&X&&o}c zEEeb%xNsdiuuA9WAu&Vm12c9&(w32r0;E>5dPcU7fAWI}HT>`G-O~AAE46xQ6f=B# zgu$xwp%tvQmWO@1d5uTVLx0kCR_JvWgV*R$#C6k0ug{-4yKZhHg}~a@5<_=UsVj6- zrSVk3FRvcHrty9Gd+DWz>o#2<2%c8HNxf_16FEr`2e(U$AN#pnO#^&U8TF&m@O6WV zBm$vA@Hzjk>@Ssp!_geSL{hLyfT5GVD{bN1Np|dL=IDK9P`T(Mh=DD0-03`fMY@e5 zFi7djv24EwR=jI`e_Df?1KW*_qPe{fNg)Rer_LJh$UfiOY@h@D$jg!V&^EB%HcoZG zQ3M94F53V0{;g5{*2_INAiC?zH%@JPWm~-VF#(eDu!@}8o1E<#;egZpsBta2nTQ4e;;X-Ic6|sDpzwis!cln;P4Bm1ynZfCT^V0#4mE3M= z7e#J)`HuN8tXqz1U+bi{ptMfw8-H|db$e-_raxwXy%(EqetH7`-qTV%d6AmT5)-t zcZ}#Z4giA6U;qlW*A>imdcDv)TscCkz)pSBcjrd01^CC`dRg6Ds&03>Uoqv9OG7Na zt>K#}ERH-RI=~~A=oWW<=ldIxfYu7436K3)P2-p%pfJ@rhqw?^%H`yW(!GDWVE2^Q zVTtw#0LY&fU!IAH2IWTz(Eq^SCnJj^1;sJxgxX;;Bq>SHmCvR3bG&y2KUrhbM8w5A7u=V|&~+UH;wKgjQgvr!rA zroAI>u15-8Y&rDmebTKBw%f%n6b@qNY!|vid@HPM*b(*)7q%Qq9{u|7cCmM8>wrvc zkPZg>;P5+aIn?9voxi7yz2l<=Wcp9t+w)+o0sOyJcfEAlyC-{Dnot}rX1>Lw@9QRv@8ri8798x4@@*J=N88vy3%Ew`qt3GVRyrB;u7VDchZS^KEQv_X~U{imRj2?Ji&OPq0@4wWmz><<$56Y&*(6cCA;}kUmiJ8j+ssAzMC-8-L z%@l=>T68TMMxj?Nt!1>^7k~MaY4<9PeJS>szL|&pu-rFb^VMAXf1%7fAr*(=F;94J z3qjw1@)YfnuRQat6_fL4WAuuh9G^KVCcig@;HEM@v}PCIIzPKhe}|b>UZ7PD3H(JT zeICP@NN%1HxDx#}|3u4cDZ1~9i_5hwAul*s1{-e$vJk0$eQO@_+6xC(4%~*Pw#WZ3 zIa!cu>m7R^m7f%N#(Yq(HUH@yJGvG)x9vMoy{}Keh=hzFOfBL7Q*3{44e=4c zDq2km(R}iB;&;s}e1xObEgx-Ez1k*?791YHnQ`xf0n-y8{*n~PDkLNGeWXIJypLey znM#)#a;p#iYCrXJn(nBP5v>B0E&s!mj@up4Te7X>#a5WYlvV4y*Tvf%XL6=6iqbLX zgBU!{aS@tz`YRH}fvH$hK(oWymk2UB@&$TBe_y zKYg^jaB$5nE?d`LAf5Blh$F7t@+vQ3a(4rqID{0fCiYrU&ErS$8UCEZeq(_+CIqw|NDjS!?hPaCAaJWTh26etV zWo5jCI7#)nDs=&q>V0*3U@R*t|NH~+4O0|}z?F-zfTWw`PmXcf&V2J{z8Y+t{Ajks z;Ob4@A-s$@F#!NWmB~eMu}1jrj;rtE4sfpy*WN4^7!2-JgfKbPeX%?ywf)5V_UAvvJ|{=cHzj^RE}9f-xKgmHM<*zd6pt1Q7hIPk9Md6f zX}WLsG-q_4{@`6wLlU0_@5R!(P!`950@8sadj8G5rw>?6gu~4%Ul)(P&)~Rht$$eB zRZeNq>GWFRF?RbVJ@p(TibvyFi*hfMUtM-x(=u>fGtyz3jxwKKlVc${3-{`C>NdZ{ ziy>*QGma=^>Au<(NUA?jv3B>g;nJH03cRw`q7S6gh+_D|GZk`@m<2~llKwVgD1xH^ zmfU?#7K*9rsOzP7HDO_Mx|+i=R8HX@on7kP%4VZRNuj(L9I!x?zeuitkdX{#cY~wYG1?GV-dF6uvN%ggL9Jm&M>g=85RJtlOtC$7{iW9wRF3 zdq$3Ws6|_+EeHT_J4ATeW!?lfaVpT-xe2M4$*utZ@&$aq_n8`#m$EkQ^2wK|M`x~T zK9H4{z7_t(U+;$d19sUYN}FO6&yEly1EZ3+5U=tO@L!yV4xbG zm^teB&H)lk`jVh>k;iz2s;e!m-SCt(4{VXIY88c0HZf1D;jW$+Kwsh*S68C->YTQq zE~Lo4iU_mN6Vla$w4P|9N!)o54w#^ z0ujJsQ0m%WA!(h){=5TdJ-a|+*q&|EK9m9RFlQz%bUWwWdUOW6HWnD(BEI*~KE;Ef zDcVTw!<}uwx?G*8gx)^Cx!5EH!d_C|rG2<_9)7wx>3yZMd~KF&@|i5(!v%6byvr&8 zKeWV-2%CYpn4PAUhEWfSOc=Cp6~hbdl5>|4Tm~t{eQ6Y z`j+pWV?CL*wY3~p$Sw}8+R!h$&2bzW-~7HzWOlv)Yf)Cvj3KmVo@EV16m}8r=zF24 ziaMLwAy+abYdRL?0)G9BW@**YrqAyaS!JMtZeOw#U*o_uPioICDjF_znbo-hggZq9 zS~D!4Hl^$YQoKlVucVEtVhns;{JO`P%jfj-=UY@8K8k3Z=m3kRLd%Eba*oiisvc)6 z=F)3%p)kt3ys}sUJuVu+^7)f~)2By$KDUk&tSx%4o$0qs6g{7;IF|VdOPFsl)mKi> zg5P94E~9861s_YweQi`9kd+I*k+rBfp#?>^PbFF9$FI;~2SNLgSo`NAn!7f@>Z$(1KV=D)1deA;Uvw1uQFY(cX@8QBzHl60W>`(RrmV=;;a? z8ZZ8tN@6E68PJ%IB`ih)j>Ybv2i|EUJknUxGoCIZYn=<6ysKoVW-xyScxwY2c7&tx z$;Vq2trTT*Ee#?bpeGBRcvt5dOTg}0G=D2LXjeK5ov)Y-w{J~WR$h_s1${|S2mD#@ zj=Weul7bG9HzYt{7P8!p+6)s!W5hyEvlXt)yc;s$wiLkbc2S9hh=mz@r^#Y|MPQh| z2VIUDAwdv$rk$`tIG{N6+rCUWt)ZOy3eZT=!}IXG?;3%Mr@7nJePOl?b$UNDd}*;u z>$J`gPfDoM>X@>dk;9n$tB#<{YQ@TH^U7{On@NjIpWqk1hFSHH_oE)So7D-p7cN#r zFk_RA85S_jTppBOQZ_>=&*N%$7rX2i2H-v$7(3sPO8;xz=##To+-MxFv8!meR$Lmo ztq@ey4s$;mN>n2p7MR2X=5noWK6A!_ojSs4QM_u0Kr6v9%VX~2R8&-b;OBS|y#s4R z&=q47%V#TeB%%sRL-ES1x+ZoUYn(G3&aXPLb*6f4*j>*E@Y(p|Cc#0?@iJrt3!oe0 zmZw?<9Ci-fITn2#ddbHwzmmMf7-s&McPL2helE6JL>mb4JX>g2;@JO+vABvcg0i0l zHFK^{DM>X12P{uFoHVem3}gjQ~}2ZrUH@9#>bbZUwn?h%*DC`!=8Eu5yZ>7&BV3745v>E zWNzoIas;nm{T(42T*X=4PGa~vKqXk-5in#gUxoe|j_8xETF^m1$=C5G{JhF0!7_2# zE7nOzRIYCc_@(1XUjyuRj(br;jSOAjhl#_x&gg)+8+XsceDbBDRX3O4;YnVz%WO;> zK@F99L=1*?U{9D()S;WrfDs;kpRrqq+c)=^US#s|`+KkdtrSXza#Qh!HFgvCbt8 zixGoE7kiyw*(hY?ubdaDI;cT`M`LfCO6D|TZfVJYZZ-sJZ!RCjE=hb8iF`CLoM$t8 zbEzf+#PJHjr!2ALyd+ts`UURNJbHve&_EO>fiVuFINTMbA{`;(CTtz@NG z_X_WM=+aMLkqNtr44KsK&buobc~3wrk`FzRHs2y2czhmhcRsm$k~R$N=EvH#=?bNw zz4gn&yoyr0mDfCaqwxufx!k46k6&3_#G*hH76p(!CYe8Z3YTj7`xZr_)9M6spCuq2 z3mm{ie@aEZQM1a+DhKQ*qQ1h8ab8MJUfx-PWj^{R$c7uv)%#uTig3a#eF6Ne6b^kg z3s*8)dQhen;lOR<$VCBAoy+8UAV5wX5`cx|T1}MZ^*72w*u#vA2~>YpzHyl02f#$C zcd&56z-Vmt8g`{t=qh&LNaWIdvXcOyiE!&hLgmGqZTe|7IT6p+D*#6TY~(82PnUfG zVGoH_iY)8f&O zK$QnywVGj-bQ+f68G*&hn>-*UQ5$`E7pj6KWMV8n0n_Mm3ZtW_9y_B)*9%@P_n->0 zUNv)JqqsAePTT7Dt92K?R@foS+Q(A9zvGIBUy!A&Ry>IflBFM*E4_<-D#*#Jl z8<9*}aCQ*OQoPencXlkJ3W{9_NZtvB(~16a#xG7 z7FU{;UcA1=0dFFnF5_F%J@-uxuY9=UUhIJCCUaaFpm7`U;$!P_L{C)Bt|M^`L@dHM zSyQYVB<3HJr4!}+^<~l?VQ%StN;G+4hJ!=3EOPtClVnA)&{}MV9k{7i9Lqn}$`md) zyBN*D>K!;3e{kP2E}pysiH3>$NA=R6U70gD0)bUH9G68iN~+>;HWIW2>F6; z#Bjwd7@%&yIU6=IGx#h_&qfC-9f??58NT%(C-B>^*a&J%5H1^%Cb8ZajrB%1YTJt- zd4MwI7-bIz^!Z7OS832_V~@hckpi1j3^(?!E-y8m`yuQ$X+O+Y#?>jRms26aYVxZb z8w{6zHZ0GNm6VYwhkk#NoY{^29p0jc_?#NW4uIjl>K<*5_nj!4d7fwsR^HhU>TD&FtarkX^s<8>XrR;-ZaGsft{ETXBy-ocFZQ{T%-~dlo zwDoir#6nF-e6I^rn%q7uTZRR>9761%^e_Q}7UttA6)zMm+oO>N7C>D1obEzV<;W2k z#|ijaACVL9Tn43a=*VM7C$IK2t%}%7Sfhy+l1%Co)chWV%a@b}c;JstzUZn~uAXJ? zWQ@7L2O=`HKH3{FPOk`_9&5G5EGKl9diIyaim!Lb6i@Xm)mmvjl&!>0-&^Q>`DCYI zF6+wV8pbH)hA)HI5hEvbX9+f-Qd7v@{}xDXzcO9nE0I<{YzSCww3|IPcp2iJy=)n1 zoa=N}2R)I+F3>cFf{#yv;1HI8!i=ZYE!ZaE0jvUSBjy1-aIJx|6(#pYJ4~Yt-UAy>MBFs^|6HCiOtNp`j;uW>J))KOzb9A zK^VWqJYi~^kJh8r?1fzsvAk{o4v(GC{zk7615)!8hBkNhf;L8Eu`6N8Yf|R7vz9$J z!|J*3Jnt&yzT*u8WKWom2sab}Tax&5WJ6a~|xBeoP0Dmyh(! ze^}VN3*x$^Lq8A{PGx?_BV2*i9>e~rOiwgqkGZx zKs1Q!6?aoaC*PL)GKhzs;RZzk&u~PiGl)h;m#<-Mody4+>WYV`PZ_gwIw?5IFTk?g=4L>_Dm_Mg&LKyC?G27Vd{ z3CE+hRxxGw0s!-A?qM=QEj8bM{=a-wPj0K30G2G>1rYZ3%87Y_NgIi06k0h55P_;3 zQJ&UTV+~?~{c6+4+nt#_^A=S47qDk3IOQVW3~1#|H-N_iiLr6L518ulw-W@fu&MIg zB`ia4jpScXRw=O9<*?Z2#+V{7T}smTPA^Gh!LDNwgjBNr*hgxwB23SYTmfepC$Qv5 zm$hC_Hg-W-S5fEeC$*F*flfF%b~o^s|H4P@-N5y`~%_AZ91`G0SO3dtt_We z$%so(`gY7O2`4#)Mij^e38%-BF{8Kq2Y`L8bP>gariJKalLcw-5&oJdBPtp{S2JlF1ZG&+S|G{8j} zGZR?V$J>d8)8FKKog|?^zC`})Lq4aTK&`2_XpOn@C^JL*K>i<0Wu+;^~l{fqoMuQ*+{=>xubh53w(?Srb+s#VzIq-aWxmQ%64>N zb=X(@`sZg=|7lLpOkj=G2$Dfvh(4QZ=N^TgMv8vx0?AQ;-!Cuq2pm|~ZxVk&LSVj0 zy+JgzD78_X&lNB~=y+;18(q>}Vl$3GRF`|iO;41g=wBl z0g|RAfq0%Nyn5jLJ?tN4K|Dsqgc!<&=3|oAmXqnkwLNBU3?!fynZvbpc?v;03|7YEgJsu!S^5L>;Tja7H0wrDf`8e zSqWMcD+B?hJ*7QLuVD-E9-FNN=`UQ(E?zye#cY|4kAG#kK^+_T?-fO&2laEf{D80;Q|hdi(@`mQmC&-)!z z$`EQ~XYw8EX$C(^3k0#VC2lu^ZtY9ZhQJ&a+8Gn%v)FX1u(cFWu=gLwrWDu! zg0&mkLa?)CIBPOvvtAqdKxKLES6`c&H!EMm>`)02l#<$Z%6>*?>~8fXdjp z_MV1@x0%FaXC-<%m1PwCH&xu^zUnT=va{hvxH7MPcbY+{B zRK7W10+8-32wA!HJv~ak%Xo4k`l!%rX2sZ_0#F`nO(r5?cz_Sc{NB5D&}fc@EteO z9{`mk0t7(k;$u8r4jG@Lir_qNQ{ir@^pj#8=$&v~&%g+>?7E5SudZfjk9B@ynptjRb zS^aSfTaNi{Z}lD!R@dN*iZyK<23y!t?^|?>NZ`~0sF#IPcl5SV+l`-~c3kXxJpjsa zsuLXB>-%1pU z!F|A5^~8!`^D9w+xgl1DcuHcStG^uSJJrTP#l)!ikuVkMbv{0x4!@u#C=xzU99S?H z?a0`v=32NqVN{}w8B~6Ol|j@z(a_g*D!z(Xb7-zwu&$|~apGSuhYVUCCUG9#TF1E; z1r|ni#z`-4dU2DO%Gdu6m3_3zEAk+}S`6|IKLx!CDyyS_xrnDiP5MixFsMRkyEjF* z7CaDoCNOddFn)>!P#hN}_!L%S?H4a%NQ4ALlhUaeF!wOgF(Bb`57II-cNxeqhzBY3l2Ekq5N@y4HmWGy`Z4QfDJvP zVg{R5@#%iA7a#IsZ6=!lQ_@vxZ-az6KVxsbx60Oiu6>})GH&9q04OsDrDdCHS;(t^ z(H>A@DhA|Lz9VswnxvU;B3<&%+vOlNF9O6W(#pMR{IxHl4O>A{R5VBCXFUR2M$5p$@d@Ji-7~b1={U_J9Q5g?U zf5Chg$w_QZ@Ed%@XjMdl8Afi4Z6c_4LJ_vOT?;ZMYTt>=sl4Y`^JZ(~@vkr>V(X43 z3F#!QciAIOP`7FnB6G%U#WE{~sce3}DJ!+IzZLnU|L78LW1cj9&y!64_8#yl)I8PE za!|YNkT=S=c$wB!Et+RDM#!G`(6N$x!#Z(b-@*PgtpY->4X+Z4d;>9A3c^yBI%4tb zzdmx<&W-%cgTFvut>gC{599_FScGecVb14tigIrnnxul|QFSLZ3ckraGhR?r^py-? zm6x_ZPkW7{($hMvab^aa`0E$0E)2Tac6*IwdQ)CY?Vko!wTx$ot)IJqC1A&qk5{(& zCCvJVFe1k(=T*7W(fRoT`d}f9G=&Xd^D`)NW!MR$RhHuIv@)Pt&6<_Ko0$@Z94oGf z*3^~iKc+BHBdgZG3d+^TfNm`d;|+Cm2s#b26w8(d(-xVjzigc=rHz6ZkbBCvZ7K!i zjK)Bl3w&NdAg8**>Ti9p5lP>posl=5g>W3qbug^>xBwu+cTl&VtH1T5i0`t!p)zAB z`6$TQ7P)*HRu3nEa7BKPD|}lo8DM4X^Tb)$rW>l?Pyq|TV^%>XHbFC~U>R$syWbM! zf(E6$Vo=~n`D#PK$6?TsexZ-L&fNmg&H@U%`!zJ~A{e;1odRkVo0Wj^*jWQMe<`W4 zxmha3XdC3TfFPRIpu z)dgN(`!Dn@lJTaX0`&sBCPDE;Cg=Sq5F&MdD~JYp2f!+Yk^)(1Jny z3;E2GvCibz=}aR8Yc<+>dW2&wiOiX5U`vDQ+{h1Ach<~%cX9#br%r+L%p_^bI0z(k z)J}DFJtWt!U%x~>RzT7zskAKjEOtdp^UXPJ9@W|M!I(h>VkbZccp0qT9$ zK}$F-H@=qb0k}hE6lB|`>NX#&?=7JKDl}!Y^j*xn>#NG^8U^1qAV-STzk0VhfqH%j zNd48Umw(0j9$uaHcKW}@1FF(C-JjoUe9l4{G*dGcG540%WKREg<>KEtplT3U z6mEF*VP^U~$z#3J;e~6q&#iF20qmWI*t|(wEh7bS&s0-*RaUa<2Jmjf9R|GhSGfU?7jbos3s}F1QYga#-0vhf)-coJ)cK`Ev z)lNIlZ%90R=peg%OyS(#w|{`0qB{>B0;f@P3eLcbCq|}7u?1N ztv^j(>eE}p4RE-~QXi{?=wbs}+xCO^+$^yaBA8 z4*&mi{(nm=x42sWQI~3)ZU6uMZl5tjUmrB8e|Vz1Tx8vfq}{W=m6w#XSVZsJ69XLj zf4FAzxmyXj-zxlS9xI*8>Nh;Tq}o3fxFW|^whX$t4OnM6q=6^DH(%t2H}@M3D6l3u z{Q!=Azt}jTg>pED}@ythyBiut5^&g;FuC2;QY(AVqn^XL3JEdS3yrvA^z zlXX9jotbHTSg~D3XdZA5=+oho-+*In!ft&slAtv!RZ+8^UU25uUB@c__-+I+CJm=m zFkf@+77GoL%%8LC^EvBwiQqX~<<3_Fv%J=|n)})7_a3_FE-!lySSdZ(DgXC{ySB|F zcF^nuv^+eq0+@YJR#IMZdjOhYoviNvZsVpI5&u8W|9|Ik@>jbb*S7E5y2Hlj`^zHR<6`~Dv_n_n*$17}C~WeVCGemHLb&u~t~qfW(R;aZoM`|GQ{@cxao`2WN% z;L4S2)d7Ft>9JX;!+0}j9RJaoxAU}KkHB5Ji`Td7x190@beMZ3j-0gQkW&A9y?T!m85Eq~{Y5@!6)dz~SkO~8McmNqMe1RRl zy(cYjEbn1i1>EGlp6T>TWIZr10sXcBbb`vg6d8COi6jR~C85A=Y2l4eLtxnt%@r3w z2V~?M^}#!a@ZbV72s<8FoSI7_<#!bQ4x+#oLui5}jx$s^z5rM1+?(BlYe(J!E}-AE zHq60shzd&;us(n9dkohyE;!j=nh7X~Br15H)D>fk87G?I=MCjflY>?f>dqueo|1 UpUDQC1!iFIboFyt=akR{0OJvwx&QzG literal 37984 zcmeIb1yodR*e(nR0-^$?(QLk;5?prpzul7k|lG)Os=NJt~y z>Cg<#ki)-61p)D#wf^s%^Tpy_>s@QR+56e|bH{aG&;7(aQ&pBbcI3f`cOx z5TK!FMyqjBGCVJn^L*W-=D4_*$1y^%PgXa-KkgjGm}FXaa|K&{y8OIyvM175T+4#$?v>$) zX|Iu9M^rD#^IR^S*S*YmTxU|akxhe{6~m>aqE27H4adL@+?;{0I9{>f(hQXpbxut) zmiM+T<)AlBCsI`9aG&-(X-^WVIyY#7gE>`}q4xm)^4d4zHltO)^#cxoWS#rDx3Q`+^EaD-_|8=uf|rYlffC?T2sT(wN)#l$dWrlx{NYmjW2 z9?R0MM=C*?eBhIPhjN!fI2>LN_>jRhesnuMP31AqEF;@A$ulROWx~-ze|qwx1)&iLj~Q?kBCuFhY?tIAQ=*0cBxB91tF-HbRoOiGW_KghwI=8k5f zPg64Icv>BxWp=S*?&)&wFy(zbFpAsMT(qs>QtW--%UmeGp*eERc*#<&4^bgkHVn9P zPL?xEcn*EZwb|&4f^D??5VD$ZM62VN(|OY~7TiwXI-4UM>NiUaM%vO0MufGGVhFts zK_VxB?O;Tp-fYdG7 z*N0g>BYlQpbsl_-KF9nItRA{;5td%%{k*9&DpyWxZ61zvI z`}_*kCTE)+#xkmp5Bi%&e##0P_x)P=W60=8>IDwUqsZcDwu#T92-@#)$FoArkT_`` z7eq>IZ(ug4y4+$-9N~vW)>*~S+S!*#`;!heteI6`d|*0MFpZNW8Rxom)0MTNjnXpvY5TuFZ+02b)Iq0rsrJCqd%_6_*Z;!q9gbq<`zf#>^GvC2KoETd zCtUyYhlxI~1>H-zc}_G@D!a^!PjvnSSuU$~wi*CLd_KiX@m03U5XY5a+IrEk)28rM zq{GOoBC#=bPyfAoAmixQse0_d@fXtm3amQ>o=-elYs15V_g)dW)WPO{Agf3(yn!d-xQRFEDEQFRH;8|G<`AAU@la{O=eB?i5=P z;G+`tlE3T@{w?_pm2shK5uH&odlc~-G7?Xb<)*r19cKC)dw@I8OF-L>79Nz_c(v;y zNmIJNu(RdNU&{K`*zifR-1-kwclIQi-;$+i1AIf_ibi&)>N36feIx@f!-Ggb!i9LLkYzS~7b$!EAtMDM^ZVoAg9KVe&ow#GFW z=w0%10wM^C#!nHK!gMWY)Lx~cHFoUvPpM2EGYHxI5nKU&pN!s^S-fe}W0T0* zv%5KH7?35yK3h;QN;4|=kA1ks0QF$dE4P+q-s_mfYj}8|cx+e%;`8WnAr*%p$U=+@ zdk$LkN(;xX$N)aUcs+WyoyAPv1Q0gu1auw|8zI)U#3yW<)dNTU6{}eQ)5*Hm!u(vR zA53 zpGdVUmK#C1#LeT zrgI8C7oW_MikhtyP#W}bLNp77y2oM-z`KZ`v>0F_7nipbKP{5R0o^3&_L<{sGL$R& z?22$zEp&`I3H6}rYCONw2%P}b=VX!oGB>n^n07Mwo2p(?m5~ z4a=Jwk$1ih{W1>J4(dlGpXbD%Z)NtSVB;ocqv)!_9{49FIJ24VO)PA5Z?CWX((vd_ zfMOv#1(o(F`nLxr6BBjobUpR@%Nu@bIa&eWl_8c*g1f}rcHU1cjCQuxOsi-2i4DL- zc>v_rOHpdtWw7l>8^Qp~Wd)4-?C%Mm0ptedT~q>MVjsz+Edg`1UVO*!Kdsa4P90fs zK%mV$2XJbdin=-^rH~KPa`km0X3jndLqz(wI$W;4%j7t_qqo@QS&JgfenqBb(X6d`P*1Rz~e&BEiBT} zZqQ7zSltz;d233j$C_B<*}ZO2OseF3=m0{9)x`bA7%58zYyUA%^o6q^|!Wv-N6G1@8_1- zMzPJ}K9%1Nl06qt1kJx1#RYsC=Nj|=q4?lIanG?mZEcwriUWix&bt%;(7bd%>hM1r zl?IxXz)>UKYH5!N80|m@<~%SbHjECU^B_6_lsEv}1F-#>Gyj*l(+bamZ7b=HYWL5~ zd|j%K;xx!x?_RF+nKnm*e1PGatrvK?!bPn4^vTwK5~L&03a2{NV4E^Nbd6gj&qjn42xj~9AR7)y8%?v!&!J~dMY8tj%A z+$qnFM&jk3*Da{g#m*%;hYo4{vIe{KQA2wXp}xD&UAlj@Soc0i?z=0&u+;Dr691G z?pd3;U6ra2_VQdQRk0a+WW2Jdwdu4R@>;u*WQQ&A&jDQaMVQ4_m+zBNlTJ(AV?!F9 zjx&hNr#otr6ReqaV+SdOssj$GQZ_8}Tg6Z$feR?EbfaVr9MSw`L1P)YXvqrp4%%uF zYg0oBKy{0gjpu*8BYgoX&?HqU#E`4<3{fo6araEfwn3zgaHRx=X6FKwcYJe)!1h2( zA_}i|c-s#0$e0e0w}W@Llyjh~1K&7^qhCI75DWh$VxgF2c2btLA-@xAa~i2oct}f_N%Q$#{r2wE~OLyA-)Sp{!$Um zt)!cw8;@@G7HWLR&Df`9QD6WEN;}VeLgYvn+%@0YM<<5>Wo>&M^!Kw_1nUjWoh9S{ zsJk5j=!&lp&CBf!w(G(In5DKPN!&Z*^_$MH18jx8)$00I=H8L#067x48SPTJlO%WF z-7@2Qpa=1{Gtd5mMP&mD$xd7SyF1uB&#y@LM;X<6@0l0){V=o>$gu|<-cq~tfrSrT z`XGk4k>mihcER-lN^b$>Pn0-FOIxF_p8*qnkcW5bv<=_*L|jb3y3g_+^=;;O=A;O1RlV&o zWXEA0vGmxfx9eXQi1E-AuRNb`XMSlgz;nUID%x@=b?K?@~^Jlht4JAC0yV(_KRX!)IrB6LA-arzv`aWH}usKHJPfZC2iUF!z z8`BN=TMSnc-+QjP!9BiKxLq_F^T$Jf0PDX-dW*)kWXdGg{d~LXzDIUN`bG4h+>gmj~@fKqdu=^&6=9p_Marp z7lUo&6!PbL`i8g@Z`%nv3}n07)a#G_h#E+h>)|@Rq0@y_iR+`=u)tWeAzMCF)(iF} z9R5RGBCZtYz%CJlh3ucbaiBp*mXDx=e3u3Oe5(#WMV|v%l9wG`xPn@UodQTn2fCvw!aGl*CX7U;r&i zjf{W#we!KW_YaD}H^BA37S8T{2HFCU3$0A~M>|{mLl%l(0&O!dN?^Ao?0pruqlu;G zwH(L(K`#`5+4GQQwAFsE0=vlq0KtS31^+<;x!8fWZ}w&WvBORg4?4Y@$zhFiVClac z?ZCbN4&47}Ok2J7%r&KWqtjp2R&K!6h3s;aLZ6L7*IdnSRoE|B->3mQahS5@gc#HG zc#d9OxpFAd3|NQtx2;|Qt2MN;RfoiLi;*f4%OgCCp|-lbCbS{$pGS7B=l}ik%ne{r z1;183k|Jxk%7LC+j+qLYZw>&A>b`Exs9Bc@}@{WU) zbC7cWgDK}=f7iiwy`Q=MV1L)a(CsG&{9g{;+VPuJckja0CUea>8tY%+dAH?u72lsj zpl$~NOi}_N_@{J=i$_ebv?N8lI*xS6%%2&6>tK3BUKhDemCcE9qqnJKG3%*OBqxL9 zn6BhMra_5nIDjJpOp!~wPVH!>aGjf5P|}XGaWa!fkxcy!_6R zfgp1lLcL0JXp?T{Ps;dS70N{Z%I!|48 z*6{!MLbM8?6}U%}(0-l5ClbUS>R@%Yb9+R&OSMo*zyp7Gdbw{2h2NQyA$fASP*_-i z?8LJ#L`Fi=1TyDMhz3=yTjH}{wOUwO!&ACT9xb310wPhSG+yy*B~K?rpRE&I*F0tqaM@qU6$aErW`bS(`&7F zI%0iVa_vtvZi~kj`rxrgCC^j4e{?$x?*ytP-{##vLhB{*_1uujheZFd*GG270sYzu z7K{BY@<00Hd>XZBGP`|E<79im$Up)sO9n|G*m&sS$A!22!I3U)G8AnZ%=-4fYppTl zc0Oxn-C6Q~U}h&Y2~pVjc+~PFpmtg_`UoCph-pVyOckAZ%nOs89z7Jb%^BUkG@^p8o@= zJpq3~B3RCy`x81lx&b|Meeu6nyE>BO#)Xw7u!Db!zH8OyZtBn1}(UOd!1iR3(S*L&5`A+sd z`@A*RzKuIBwkLn*b3RdAV)XY8%iUhVagOiO3ll$i%<0)@Prv`>4gQ+ZKWwRF{bHPV z3B0W+#)f3TL33`7@1L)6pvM6RrL&vd-#>BjdH@FnamLc`68ul3jWRM0^x{_V9DIC_ zg8uhvSEpO**n61Wt!svx3%b!19i7H~4X!t2O6r!Tt9c4Wz!accTYv~v1}qloG`WX< zoFZ>WObQp9p4P*wM%(=B2kXl)sYA6&}S7i8%Lb>1hGc za$~dpr)K7VM1eK_`R=abcEug|w^M2EBUs&n{O$nZ;5S4+8VR0ebb{-^33iVasCNqX1oBb^_J7`>5ZW@-iy+x zNR|w%rUzl8UFJ;XKJh+O^dYIQ!M$(3@o;+h4+0P>-2#v_x*>9oOXud? zWa%SPla46xor|Z5(He3v1~i5+)5`txy1g%UrN8q@wiN|!y2{08yK$228-ZqWO$L+Z z4@+c3A9z}~M52(`c52+A{IN0qQW3()?@u_CIb0A7n4=z;g7vELtrz6AzSqz7;B(d~iOw)w!$i~rW%qIfKs zdEeu6_-LVR7hk7DR(ahQ30O>hl`FQqgy0EVZo@==4<$OPC0zo}?bY5Ve)#O5F>x8M z{8%!Bj^@Ah5^ne=Nfi&Xo6bp`Kc)n{W!8LVAG*)y3H=~*=ORD-Igdsw|A5J=mlf`e`LaxFo}TF= zKF6)nfY+dxaEU=ue@*m5&k)xL(gW>eu2rVP`?;Ov@&nKxyO|4xZ<` z2Efl*BTL^ssLja71w|@kT8+KDo0XmBvV^M5AViCMPJWl;nWX(HZqbI~s+PLUSL2c; z^FZ1eNF9le$=CiWOaP}fi^T7b)SF{ndUe- zwHB^Hm}rWYe$hXQ2dMxygON@b_SYluIx*V5U~YAHQDqyl!KY@9q6-~FcO8xO6QHH_ z2$oP%4Vk7XYgy%l$CF;!z5V`e(9$tcV5&!$YT}xUMTa%%%^r9GIz=if7ZYu*S*{KE zO%iWW-QSf)Cr&;(pc}%G7x-U9zD$ML)6Z&HbZ+)#SIA9L-C|xItDICPPkij5!P8DQ z2(x3r%V7DFCALPA<4kEH7c8n3yf&9>6kktu>q*Q5&X{dx+* z)$>3ViOIr=0*dhJRGldz=I%uB#X$kz~xb0J9ILU-@dM+v;fMlp7U&`6#`@I!}I30tUeA7LAy`_;ywWUN94Dis% zkN3iLgKb0Oe}s89-D2;03d)c#G9|aHgq+M+NH5D{c=7_n7;#_hIUHW99SeT~t4V1Tl(cCfLj-xCzS}|pHvO@vcB8JTY0s92;BabWuvns+u&t#y?Sh$ zfbs1K%l4pT47``LtaXIXJIiIPLUU7P+UK34|4qQbV5?J|y}@An`!|*bT@VU6;B#W! zM%q!e_xk+1S_^ePFCW1zBNl}>pcPDL-9K1lXnr%&bv?EV_MVyp=HbAmNfRdAyeHyAk<1{U2&4j zIX}Qk?_#LL#qc_pGS7Q`&4rs$$+pMh8OF;j`XaR;moVyctOc|JHJ@h@&QAV!Su5V8 zNtHk5{%YlM*1Ak&=25^iD~(dH4p(2{bH}3-ZVBZ)WP|8Q0O{i8w9M0+lVwR9~kl7GpD=4K{|4a*={ zz7%o3{K-_h=k!Ar!N-S&&{odtQIUfD!Bc+e1DsdQ=F-B5npg!{GGop(btLmI_1fQN zrCInQ(BW&Pw)DPP(LYbC;*AbUia|J|qO3jZV)-Kk(+}HYVIy{R3~dkxBMvdgCE5Th z%Nm7^Sl8Jz*gMgD%Oob`RUi1&wTN!E_tGa9s@8@o*IMx4Xo;>yh+s}L{={VLM=R_{n@z3hj$VYw<@n(tHW8jl}hCmf(3i6W?> zDJcGSk9Qbh4{_POwXx_kc$-Wmw4(tWCWx;R7)C)4VcFHI(j1wfu(Ix7g*ET9c8Y>8 zefP5Z1bFe_WzYvJQ0~ZvQr~U#oM&v(yhfs?$wX^%__66$9ot%$X`aP+lIkipY*Fl} zQvWxmpc?wc1i9L5(ArdvDZ*vyjCR-01PtvZ2oxE7YPGStm`Uh0BGy+)?^+xm?V;s_ zJ>S@yHO;8ZcQSVugjTe+S+rF=f@W-$x!y#HHi{}#)buFSq~^wnc|afddjJE%`gTmU|^ z%0Xtl7$!16(Rjo7l0>zT#A@|V4KSa2n(nUg_~M%Q(rx9wS3(W%HE*1L3}n4Lf~Di4 z`5CmaX^qT>1;(T;_Mg-qMRm0{jRWU2Dcy%>kc+tcoTtsV-1te(`k zC?GITA>8BCBz5%tSypVB2wF%K;na9zq~e+6E9F)*LH%s-+EUR}^|Q#IyLk_vBXfJn z5FlPj)mS;Sv6MKtu|!DN_&gXWP(_#G1v3XKnQdpg-51dpfYkXa?-t(v#VF_n5@4QY zW&G)1KT`1%?eRI7fVz9A0<-G=nU^gSsHujg2Dbm5$gp?*vp=r>>e;8qP-PVr1IbI5 zRxE*d<-Pm#3YPLg9~NUa);1k2EG%5Q183J(OL+XSic`G;4Gj%VCU2e0rJF7&x>-W@ zZLdnCEUB)ZkY|}x|wbjd^ z{vUfTFIKn#><*xnY60-?sGw5IuL?2ncv&~ye0@)k`}c=e-xzpxn1q1OXDi9^V@K%H zB*3FR+*d+-7MkNF7gzYbHr4zm4E*@l%cjOVfFmBE_NUmIa=}X-rm5j!VXGa~IdGbx zrKX{<_cXG+&*nbZ7f1q>5MZa}C#Sg8GC6T9klhe zX1%r&-ep^~@leMm5Bc^ss#@@)%i4+jV8DkBALGTGkz!&GJ~27{tj&AGX%sWOv4$?1 ztKv+}e|wb<0eAqP3-Am=>~cM`O?G)M=G*mdPoWbiQ>Tx%*!lMI8_|kGF8*y5GVeBO z2OyXIJF1Oql!=!}WqOBSa;D@Op+)qlz|7LopYe=??cfSgo@WnFF52f|X0y9&*2au( zf}7%$EGfCQ7s}c}vPVZZ(~TT@v zAGzwidDa$G=~=)BG3zS?(+J!4u4k28#JV}P4py`CP&wVrlS9k%>Jr50l~0qb(uNWE zRf0AZbW2p_)L=FLHclfiTU{T6XHw7O!daAn7ZYI8i9asKiZ&>v1%W4xL&_d(jIm7^ zH+UJMQ9V?Y!BbxJJI53K3suxT*rPVu**! zf|*r$q3PFiQw@qR*7hCBiXl9Qr=@x_OsYZJdgZQM#ac!l@}J(1mE=xrbdd7$O!lc| zA=1p~fLRIX9Ff1py7+iPj2EXqs~6wP2`YW}u~=R|;gwBVP-@>R%ypW+S$k(I594*h z{(Lr4%1?E^n~eGI(rmC32FC1sAJY*OE}tC*$`_^l?bjNW7qJr-#Ue*wce?}ct@$FW zLpB}#TSn^j9@$QHW=*hmA+5`2MaO`xi!Brh;F67Ibrp$oQB9bMmDtdzC}oI{wOCpB z?4h-x+=cEm0;L!W(B&_p*dG94K0!UO97>0nw5^-*6!^Ix5PB`$P z%()tCdtoSfV2||&VXO8u!Sa!!XZ9OX?{p)y{H+5#J0gH)hB#Ec?ZsYB`7w21K=#M# zl-kagj~CbvRV$?C4TsSUx$7;2<3Qbsnk5J5L7bS|x*ts(Lni86ofO@9Y`fTBPHw`c z{e_3>1F0U!AtwrI`yu1W=_vg0U%bLToT)Nh+$Id3?PuVZP@|0V~yO zZXlT%PU6c7xqu$(=h*ZhjC3JR;;;Kp%@!6ei=-+0bjV?SfQ;>HE?wmknqX<(VTL-f zFG5lyEHxZ(gXoWVSt_{|RBsF4+u3KAIaG*g58C6RWg^#K2>89`(Ob=YGKA2l)j;SG zCY%GpQl5kFlE{%u_^0X4)F5i*rDZ z>As7DN~`rCN5%BIgih{qL+p&E#z|`A@@WPGa0M9MgVyT#TsHAF;K^qLPA(mwoLPLE z*xzB&hJX&PD#0R|wfdqR{q}mV=VuMNE-d8K;W1O^F!dFU4!TN3(vQ7SfxPB@gH^*g z_VaRcKm2p0SBCC2&KG}{&#mAm94}jjxVvMFd8yoQO)Z^>!PcGGVv5-+>X(Pl9~k4vhWD7P$gUdaG{{{p{U$q;4?;+!z4ehuc!p7!jA66*I^6j@{42n1CN>I%PT z?S3O4;^ilXBE(ddd~pEVYw8%Y#jNtS#i;zL&?kJIjA!^mT$2q(~Qa*HV?Uc+H5q6uxF~SI5&Vf#P3}s`nuj9-4S1giA^sQ z7}hcFVjU?o{SYoEMX6$q)jbn=!Xh1t|2hNAj&z_$k3;HqQqH}~mvU5XBRcdNi1{)T z?AaexJgTo|1;9szdzWpmNAYh4Enh*^hxBNAVl@DhzEh5(4p#}}UwJn=B4(spXra2{ ztPe1yTY84DAs2TBx+wReS``sO52fpaacHfxB(-&IBcZ<5x-m%U_Du{LV@ z*m&O9!#3beYpLEOtL(}wn~I-k&@=1l>RMw@Dc;tilS3Q6p5a*YS8=FL$W@g-$WI|I zOcd)LA}|DbmW~y_Q1B?ni9uczNX3-F006RZ*^>>>xeDp9Y-V9WG z`j9YFFM_12*L675lpgkpck*X`x=GY5IzV5tsc|b%>;z(C^!tmlA)^(9(!_jn;K!bo z!3^vgOcN@$#bq}vXYt!Sna@pT+m&P6=LgV^*FIIpIk59MM=u-RGX{y*gPi@V83N{> zRr40D8($qw#8!(`3860nx}fN9=L|hBjCl{b&(8so+j#6bLv2olYtg;2##kxXFBtO2 z3pP3W(SX0q;Ib`u6sP2BQATd>-IeIa9<(48pU*y%XEYE#_(s?tx#VC`ott&!D6AP9 zdM1c5o>{LA7w*&K*B`(_G9w-nJ)(~x(qghZZ$5;u@e+2xB7KYia3{I0KK7Tgh0GG7 zaYH97))myB1c$dKV+SR-@UpzTJ&sm^R}!t~-G&hlYZs~9O~JsWWt-Q}2#I#<=4@Ja zL*P#z zyih~mS#hV`RP~<-8JQTyS_)B|P(JnJjJ0hLqM4Me!efG&P_vV(ltDw7!MRzjtbp~D z78f>ltgz@f-B+|(bncF0vH0OR(UsZUl2^d)^8@xk9=FliTmxWRh zRNoQjrg$R;_1igE_tPfvx9C<*nDttvk?CH>AG1&QYG@+%U2B?aP1XUGG28)_4Nj_) z4ARcCjD^?VjRgGU>&b*gCX{r8NwMT~9*vTsq?AYgh#VMf+ci_BTdfXA}f#vru0me~0eC@N+p- zp@U!-`CO2|DF%u8ESux1*h#;b6Ie)uc0QdmRr3#Y*O+*hI^IWF619C za7#hu4CdwxSCkd@{KoTZhUbS|G71(N6{h`Yv~7ITL07edW|#fnfG=>@3ncSQ78Ryh zVF^?$MS@3(L>B@y4KEeGUBv$*MAmJtJb;GEwj;yD0Ku7V-V`S|I&x@5!hYSfvt{ZN z_|}w*ZzVR4Ze5P%8Zup%>;JJFR4i2i!NV}VoSGgh=Q**E>id~tG%iIH>p8*2Y=Fp} zx-{M&YQ)Ms)(MAFs>BDp3_Oj-DCX(k11s6_Ex7o6?GeQ**|#yv-K&1a7@8+wAFe2` z)iJq>@{WqiQcp=?=`^p-zz!<2FdrFel%-PAIeYwWR?+^xgg;|(^E@D z!Y}7ZDO-0_5moP5ffE8j$D``yGBzH)fUQK@Bn;-g9I5Detuppo&~WL~<*3i0^)FHQJDssktnAi^mi3hbKA1fc!*!;{b1Sz}+n{E?Qi=)fySIVc(pMgIXWx*`3K5?a6Zli94ps`H#UF1U5E6xzv zo~&8znf!$Ll;uOyaW2l*z@|j$9so03+IloM0jK37f~o~}eP5y(lvsGhu0k=5Xre<& zHKhcpAxB1*rrH5ew9QF06~m^XbE<1%-4qMMwc$mShZe#c9cm!w=qYA`&GX!Ri#}qm zOhS~8E_oPZ)Z26&YXgTx%NAYjm-7G-AVOARq0vk^U7!&K)^%*d7N1xeAY@%UPM3?47jy0!tSM}ywL zHM^-sB+MkHXjBnp_qnkU_8RZd87a~;@qR-~=JAo)7hXm+rTXdC5<4FEJEt}O>BEN)^<^GQ zW9<+xao^!bF?@|IfZp9ty%}A6@UboE%i10O+eH#pP49ToGApvMCgU0X6Z0yi6~SIi z7^BtX+fkq}kGl5Zgj~VEJ{sPWDhiNTYiC2LE^B8mELT<4f_eqQBg?1SLkgv*{SBH@_MDb`R ztK6MaT{fMWruD*I*j^?f$IsV-fSeJ;Rvnl+*ghD=k3 zT=XOsvGX5a$@pAIuil}-njel8;rc3^>{3{KDZ}-)Ye6YJ1s>?uZ%vaE2GbFH3#5M* z9{x9|{ruOdS2@KpcH`iXrr)@B`4(W1t!c6UW*5|dhA#UmlV4oZhpAS2Z<0#%Bza^01jH`wvQzK#Xu!Plp9 zE~H07+1%bqz!h>qxCuvqXx{I6wj*56Uh&FR22Yvu4CtTC__HF>&>!FPZ45yBze;Ny=>tl;7$HDy60|NtWZA_UXOr zuAx)u+~Btsyy~lSEp6A_v|+(9wB*QEHn^X|Y_~&=dl)@kmH{oYL+xgSoJ<~nm5^FW z;2f#i{db?jS35@3-@V9@b-M|JmDUGO_4y-HWmn{3q_TNpe@)u_?t=w~o$P|87ECsT znYl7No~q1yaCY;e)>&}ag0&jTw$mWV!*!*HF)_+1(`iK1p)CFLf40 zrN*pZhn-TRl)C{+iofRRRVS!Yxoq&PQG79`zIyO|h)iXd0`Zb&)245LTcnPW+toyQ zY~I}m+@uTe8+%6G97Jl1+D*`56exqhAxR1o(JG%9S}JCEqu%t`?1h!cJ~xk|5G^wS zmHO8zLH;)uH^r;2!8Ds}%gWay%nOC9#*#VetaZ%bevpXkkZWdCD&w}F{iK4%C>*E4 z{_)`e0?IH?8|B>iA9Lt;;+c;$+}Gu!9xKn&h&C%B8Z&#<3pS~knbXCT!i`O5yW7fF zUtT2SdS72Xq&r4_!Hs4th|pXIG1l1rqD`rCMcKJ4@+&rP{lNi>8qm(&eLhE=f;}Y{5>VOi5&fWRmp-d*kUv3s+N*Qr z+u1NPXTx7jkMbF`81;SN&d%y=wuZyW2{(0Na&jxxnylk#N+kt{0`;~v)`H#Tu(U`a zv5ijs_@9*|c?BO54i+1gfMxtBVjp`*MxG=g*=YQ(J{MdgFr8hG7!n`zQx{W8G&Ys5 zqFQ@C&a1HbX3Gcwi%$xu@qODCSLb)gi@9n=EyeF;AzRC@*;i}}hFrjcjEcbv6Wc$ATnx-szVOtS&LpI} zx&H7wlX1NX;ndV-)s?k;kt=R)ur`?+pe6`gRj68H=A_hV@cV_b%YQ2=mTs8<5b592*d1o`mz>XYA zmk?c_Z5-z~pZ1Q-ay8@oLfOYYPAeMZ;Ly8Sn4y$Jl@rXscWy*Yaa#uba!p2wimk01 zSiy?l;uZ8xfk#gthg2&m#Zxcnp?a29e9omL#=O7h5ytcMT!*B>!1fOm10`4h<3|!x zr?=aT#2OAHS}}m?>UDa^V5vQAA|!z^Df|#FnHmu6DOwel7YybZW(d`}_-CmYUn`Kl zjKM3;+vom8Kav)%969cTHfa+MtD8$qFDx|I5l$SDiRIS2H5pfL8#w=dBx|I}XP&@u zOM^hsoCnYbN(y&%d-kkBIr`}ID`8=R9P!q^-cQzrEgEVAsqZ}P3VqXy!X=9`#>8_j zChKgcS4O^Uyq^5G^g6V3+XuK%b3!GQ3MAf{gBublxXv}Y%qd&s+P+BIdnByt+d ztr|6%s7W@zyO|1B4hp_u(5)Sz%#-rzwtkyYICj2f3-GbK?W(Iks&C&SA#4BcR5KLb z+LfGQ*7@RPcaLK2Vpx4}9X~)blvu-k7kOpf6kwV+R`XYxjL%9KY;>sgl9_;D!ynot z z0AUTdp3<|qvA7`~D-Hha*w1Z3Ew#q$SFDN{t0-dLhnolbI>^F6G99OxTh=1krsAF~ zte`~h^m<7`Jghib8mY=w72b7IAnVxN3IusqB`&t=Y`i^t@gKwpm>n1koG(!GtsD?D z+Hh98Ik*AfFBvoQ2S4@I#)d(D^GiJ)J?5ekE*BZ*F~!5mg09^MBm4= z^~IbF{#u5Fv0+xOK-ABzt+TOz*G&p}-Na--PZ8VVbj51@CT!UF=G=-~^6@uaOWkiE zQu6cm$*oO8#*Iv-+7fsy7a0T>8%LZBJ*f+1BY0W0+vy=t#jQhYUE8D|TB_oFEhFfI>sYLg9SF~M=L`?AW@70+kmPEk&vWr5*9 z^C&t7n#PdmxO>isi@eXkZsFWWGfrXO26s(xGpOl zGigMC9cKS_y>Z5dv7uP!QMlyR9pIeFr^QTz(WWvth>@^jj6Upw5__SO@H(+t^5#%vp&~+)p@hsTqD-e02~p;mkae}DJ(U8}X;uXotbeumF8>}T)eEq_gNFX>@Y0s?})QdchB zARr*>ARyQoMZ63A{@I)wL0!x+&cn%E1NFUW6@z7E=Wl;>tJ~{V$Hn3Mg8c={jw5{~MQ!7g zjGwxi7qZY;qtRq}$sLb6AG9P2emUN2NPwX(Pu03hcwyx;N%PG`k}@IlaK>KB(hLHX zgOTJ@skI@`PCwGQ%y}~YDWRt>b|4>N8vA@CN8aJoYDy=wRgV}zGsk}o}jyzoh!u| z%|xFfYr_7h%16ccY{kr@#q2?vTQIi?E+dok&2{JA-g^3b?;` z^mt%D@1bKci~+2b3IDq-Tkc9&26h=t7F4p7bR)h;55l+Hi3-!Kssm= ze7z&;qIwIL@JWLQA|<6I2GVlF)DJyL^mv@_2qaMX&}!xFsz|XiNN`KMOZ(n#%rY^p zsYgsvB4`(@Yi!@-nVr`x`#hPX#10*QV3FtSIX)kI$6Z>a4T<<OK_D{2(6xZj=Vc@XWfc%)!Dm7eFc*QiK4NN3pX5D z%Wk4zf{ph}KdQ(bu1#PVo!4-*WbPtpaV^lsuB|JG@K?2*K43@b57AM?e6JhW-8vQAhO7-`6{%pqqTbulxuI-~>{a zE+{(g80(ZaH)waJof?pQftnDt91@JUB*h}da>#T4#gJnsJSnJ-8$K=z-$5xZafO2I z(S<9=nd_fY?jk;R3by~1KO^kEOF8GstG3xWaaK_i^%hy}cUG=n9-qi=A4txf&P2Cm z&=&=`BF)gr8HJPRMw?))R+si+Im?ee1Uq1a#1xG1yMJ6b`WMgKryu`_^Coz_*kuPI zj}?aJ?O`~)yKBd5W?WweFW)WhwEJ_w-z3~-WVD_n_(q9mV0ib$xxa}0_3n8Jifo6w z#w56B1ip(t`g@c=XG}*-+#v){C&V)!ifosCf4=gIK*T;m!U;M?^?QFX@b^f>9U{B& zumQds21DT3kxNUjY^J~RVS&W9}kM37uUzh#Z3t4G>0K4EeWq4XlZRFX>V`u85`rwcb<s=qRj&6}1-+w;kjC-3zqC{A9EcB<~mzvD4GC@Ux`%Kethjp8L`OIsUh zUS1x9pHD^K)L3hpAKL9=G#}#r{rf5T`A2WmUbS;5ir9U-)z2Z@;k^kLZdH7z%oyU3 zts=C;hYAP~{sD$s!!Wv{dA!unXgeKk&!Q!T9~eUQrtHV8TXh}L3skM#M^PLSmH*>n z?CO(@DGiVGz^mub@qNJ_JxRrD+P3tUwy6_8DRGBccy1&Kkax5SV&E;s!zlN?SCB3! zhqT!$3W=B%(YC(r3JQ+^dU<{**$kq7_`0qemuJ*sYQ!aK8pfMoEiz|1M14BP6kc)Z{>o^g7HFQlT>hX>k$j1wV#BU9`3vEq~P7!_386h zaeNpI^$oD>n=zLR?z{XXfH|R=+7NU8eu6RdZn4PWqzAa~Mzer9$<{Y<;5M{NR212g zF&VtL?>3xZv__h=IWyxJ_AUX!7%;>*orVSX-B2(mepwCVZ;Rc|7z$ZO-5h*Ee7Z>uK`11zNx>&eb0Zb#xBf_41ySD`oexaer4i?er+|LN_E-0>C zy0O@YY)Wlf{1%arII1H22eiu<0%%=5H0lq)Qvv3ZcZ`!umhS3uP2Ljo$KP zG6D^QImi(L_*P^|5y(mXYAbiocvu7{QB955^R?>s7gG)%D~z_}fvA&5LdgZTyC47rE{KV)PtERDez+Vf zq!i`Ej$SUOVj76o`K+o&(!;7>^szv{1gYa`y|!=7H#qa?9}M=729x~w>ZCU`Nl{wk zq4P@I6wXaobwmNp*7&CWSikPz0kUE4;^#uVgJ z0A2ryVpiqK7Wge?^nvW`)v4mi)dAksFcd0hi2aFM3knmGHYMQRd-KEPmWXX;9S_1V z6@JU=^;jE`N200iCO9cqc*t{*siNIaB$qOKNKlY}Ypl(v4_N%ERN?$Oiyv|EzRbU& zdgN2r+xB_6nu`a0a<%GBxAV7~8%Q+ojB8HUGOt(eF z(R+3j@nt+cH-aY(rU5pgGeux7heN5s2!}n5;1CW)N zn3z9lXlHkZ4;7QU|K#=VW&W?#U9M5i3Z+ADNi>Tho#uw0HS>n-b2xulTbHVuwk{kt z7Tcw!Ge>*H*bEjy#;4P7w>+6@clRDm{=C&gBKufs*ep>f#ZBHtic4R`5(>qp562dU zG;Wq$F*i5&9Q-b&U0OzGF}osnvonCCBh2ct!oHL+D|xG36S7HOUS1ytZ)(BF zsExI=1sJNy4%N=e#T=A%N3~%WXEWq>pI#nwL57X0l69Op9CAa5!nkn1kgyDzVOGfa zLnEb-kkI5#G|Thpi~{P2z%!J_u#k4d%JlRH#?vu}*fS!EySlnoiVH{#Lx*yYASj74 zGc#jh=vYal;nnc;$Yi3>=F_s>n1%87H)-ETo$c1J>@c*ZSBxc$aAdz=RQ~masMVDM z=fh2Ge}NgC17XnB(-m*Z3M><-bfAa%zUs^BwmJNhH4cCKibotE>m4<-2QGnnvGP7e zlc<&i?-I9P*TbnX!x=+%D9YGo5B?d&woQ#3k$?XEms*@S-c2{*z31W~LQ~4cilUVh z`GPf*s1=qX@xL5Ca6QGEI}j_HhdqEN=R_d=luO9F63f)Pa<`56^UE~HfoT)x$&yeQ z6Vh9Gr4$DFl^Y)@nN?Y8NVu_`MDPN_9k`h_=XM_g#1Yt$z!W#5(NBmnv4VWcY*x|Z z1An;}8g_rw<>$0t4DJ5<)X(|zYQVci+sm)xxm|Y^xSq#Vl{g58EJHW*Zn0bDlNda= zqhA2;LLD5q4eK`Tt=JrJ$(&Xa&uusdKwS=QNZy<8Y>Efb3Sh03+ZT7qgaDxoZYyhE z!(9Z}0ZRjnOw&60N1V7Zkt7giBxQ%=DSQy#0-=IYo*BV4O^pdBBW6x)8Gf6r=%AcuvugjH zYVoFkXJ_@?$g&=_+x&fsWFGI?(wbSL4I_pw{9{4ZQ=Rkf@lRuOj4mFzN^DlTo4faq zTrm2pt35ue$C7+zHFlMSBRbr%87!m8eH zBYhM0bRz-IMn3Jsa~qIeTtF10fCos6g#peAAF9A}8%_prmW(qFACNi$Aay$JyG6j8 zgaMC418&o(Uc(2Zvj9kE4UTTB#BX6C7r4!p5{?f@2LX@{RSMz>(g@W-;H#l~4o@|f z_$$a#YXfb<@!>33@XGrCQs9V2nxXSA2oRoxnA+z1Q_FlcMIn24-A_rQD``CGSz4M= zljrWxx)eZjVymcsiY>(&Xj6$~Gyx1vwOEZUq{1vFOq5RjFjZ!8K*y5jNN%&)pBh^% z1KyFL*-g^Vl#+hn;_S0}3qGfz{3BZw;g@P7z5|7LE2}8hNHRA5aH{{hwgt9Q#i=nt z9f##O?0;mpV>oYY>?n?#&6$}$=Up$-+ zob~ds9~vYfUvxDzJjW^5WqRuH-qOZK%ueOy9`dz$k41~cXXD?<#VVt=0!_3$VDhW` zjWj=*?2jx)@6#{5WaeIcs*C)b_S2x52P)^b3V{A3#cW>A?x4Zn@QpZ61mb4iqy6*EwdTT{F&-+ zo%>G^xdH;>$Ogd!;yAP%Jdp#I3+*Dpb8%}FBO93PjXwE3TPL(Brhaxt>sou;5j;xq z1}KG;kNBxAG(PpH%txK`k6|uI)K(c!&|4kU`kr1Zc(pY==(vZA7*XS_Wt?V8O~_yQ~}EKJYHXh@%-Z(2~;1GC!a7DaJV zw!4a_Xiv4AEm*>5508opZqh!zR<Sat zH3A5iSz)wh(GykYq3ky@&S$Rg$F2VcCxQN38(+nv|8qe9e^g((j8jD!{a*q7FMY*^ z2iLxXP#{i#iu$h%`KJ~=1p2?nD;7`EbOi)&kzF#1cSO=WZRGg3R2+z(0fZP_bX2Q=d%8?oqvA%uiSv~UvTnQ zj`uG(`F|)lK{U=?y|IyRz^}nnV&obzv)DPoJt`GZ`wV7UM`i?9Kk*cNd82N)r7+BOJs^srInS}x8YAkq|B%XFp>_f?q!#I^Yinyg+>J{ z>@_F+GmGbES8u(Tc^P(Clt)(A#N>yYbhX*bgoK3eZ3(ZV3@Gyk$&9}sR>qCQ6($pT zma}T@qQJtU26op0vm>MHRDew}Of{&yFmNto6mv`XL1XJ=;aI;MaCnV-zsbJ^c9n6-HK6RB>|&%@M% zm96&O?lC0W$~K^ga;N}197*SV@Gp1#6J&-jFovKR_?wmTv$Ka^4Xjvv(-a+CK76&B zRJI^F=C+u$--v{+u%$cluE&{CiF)?+tWHPc?nLV@`_(pZZQ4#{XIU({-Wb9M9FpjqA##0{e4Ggpv$z1IW>-?o6o z@7kK0l5gLxYgX<~EG#M#vzN_}?J6xT-JLl6ijbbg+PzRft$044K6d!q`uN@V;j=C} zI{A(WCObrr^}$GCr_+t)qs_=bcl}z$4-*u|Y@kv=8)~*186NK0*Vh;1cjr7pm=G*; zmrP7dD1N}yMD|+8@j?ylXscxl+ltxQ*^*r8xQIBpC<%^F?fU<^gEvE!|FQ*K)h9D` zOZ)f!R(?Nl`L_YVqt!v)pApl%v0zl3QQUueSkJ=IjmHlw<`yg&5x-ep2aV;@UmE^3 zhi?XL|C+-$d9j}5{M+WSS&@cl@T?iEfd;Y_ zfh;E6(enYY62fEsP90CLIR)27(JX7A$pOs>iQv zuxZ-`5m?J0I~pIfl@D7YHVj$Ky-*y&*>v?a^h z@-`d+=z@90Gvj9GunuUDDaxVSRJpCvjD`cc5aNxLxY_wn%v+fS$@@Q5%!o!s=eCv> zVjb7%%j^TL$`#+ge{TrDo}v7G7QlztE|6J68<2+;bS(te!pH2aTz*?ZE`_nS9K+G0 zw?8y=@PnPsgP0I8M*qCxdGvb?4UJbz)=ctG;A2HXrZx>mp|ED1c7+Wj=Xbvt$W zrD=Vq@UM7tYPV7u?No+Hp-mzk$V*WLF0`O+TYrA3+up&!p>$+~H958`%L*mxs6|GR zjVvxNujzkve?_!oV{H6{Khvngq1iggRsjU{h}2zI}k3hez6s zC&c|G=0JZOq0`6lvpo00Xgl6_h4KswSjiU14o^tvUWdgzV>qPsyx`}GI03?vFj=Cr zFt%_lwBY?rZ*T9KqyV+X1GqY@Ir!Q1nu)sIb=fSK0BY5-AXRu!y#yN!`qv{(02BFT zzNndaFKEp_ujSXw9cej_8!F(8_rPi{ES?#Uxr+jGcRSO33b(mm0)1VQ@Qf32y@ zAEedbc|5Dk!~lkf-v!}8xMkp~=BqU;aFb1O0qh;#(HMaHt~|5{s4!%7Gh5w;-^W2O z8&kA;z_veb!0-rIzit+=qr=loKnNH@s3?sGTHU~$mTx8NI z3EHTy_M6^$o4?pksLYgTA_Z$D*7N~8Vba(f;_GZ;?dAA;_WjNNjlQOFBFy~cvaJ0=JF6HADcS;R&^h<;71d@EIu+!xte_k1&E_&quKRFIE97uG8HjU5}3SigW< z0hV^Xv%ursnSpngj!f>_{?c#BMhgIGSv8!+yE7<9^Y{(BP$6nh}LfXNnI zZ-2D?rQZ(mDzr!SpP0AE^q-i2>&9;@`De|WLHy5%{lBti-HiUhxvj>G;3mQ}tCUWy zI=-!9cF`H7bzc=;U5rWSLT4QN1yRwv0Q&4zXxf?*MG&vnV7o{?7L7(D5;NXi#+0L6 zL`XX1Xfu9?B0FH00oFdDM-Kj$7Et9!$5MoI1c$EPq8-l9etJw#_BZN)fCz>Li!58a z(TnNU_g!HZ_p!3v{u1RTM)=C*;?C>8x_;%~z;8`fp0wQ+iU+M>0A7gpTqoONL0j}E z#Sv=NHHwD)reHfBp9BXzX%|yuY{`SR!qgLxVe+`|V*iuVla)@|V9mA*|3nZhw|+bMbgS-cKJjnT+l>xL<^NYnC%jk8%*{Q# z>;8rHr-%9)Z&1BjGfv zYHCg%(M#HG=(zf`6EB!}kM0vQ`!u_DhH^LQ0aYCxx$_)c{`qM^(8gn)Xs~!5!B6Az zYJCs4{-oKOCDE=ZY(eMvF6B*4_kUUZX<4GBvwFw{S+*F!83%R`gC=Z#VookD*}{k8 zO?C0~9JEv9(o3Q{@{^9CTHoW@&_fCJxby9Ky9Ix@!doG*Zdao?Tin)t?My|#IzJ0PzvtoIWzrCW*#&Kax%}Yr2G02s7@7}HF z_V^M7icUlyd%Cjs_cJUolXAnjTdh@l|ADb z<>&ld6JN_e$tF-Q`)|sVNUICXcw_4 zy$R=uN!QxG{_4pHhu_C+j*|g zq&NPAe?nK0;OdXCu6I|MY0n4$3by1~!OYHI$lJRGu?P>drS$lO@r)j&3F1HPf2_l0 zJU8>0=J@aP9N^Fy?=>e~^LA__a731{%dnz>ZE}>Mv!%5j=j6Gck zsu{Nx%Nitg6}ekQ-Ob%5lzj5cKO1&1cO)Epzvm|#qv>Jvq3Yy@=q+*RraBNHf@uA% zt(I7nFW~0-J1@X>k>{QFafg-a+VGNsfiP7fJVK!dLYbGpLyr67#!i53K~^`Z@f2E} ziGezaw~G9p{#H%_iyr_C)F|{;`s)~vpJE*W%+mh9b>fKqTsq>^i8;jqN~I+URrO1) zvQIlNZV}9eCLv4#7gWSg^%<_#QpJ=Qrm@`8yfBZ8-=A~u_7tdj5WN%IczN4};^ksu zbUZKn3DT!y#Cl;SKT;igUQpyV#d6&GhRY-Y>2@(h7Hm+}ZPM!$p zJ3Avgo&%nMoL%;Ad3@_<5$uo_?6|&vyA5n847{Y%Y~z3axKO+RjFj}z-r*ZMd3-*)m}vz;lF`QM+^g?nB>nc$ z(b4TJ`07M>gSqYu1J(UtKPe?D(#(qp(6+ov-B;m9e&nqAyD8H+f&1|K6CL`y4+plG zsNKvjuwORuE(*2rH4as|&TsaXbE^g^%IO+ZeC|Ny1GUS3vp+}dORwJpI1S;=)2C&Z zVmQ;i&o@aF?|z?QlO*|Gc9v(REyGNF)uhRAtmOkU$wG2+a$Rd{Yhq6me{xNYVEoP7 zT9ZGvbPo|zTnQ`&yPGe*n~BQT|MawywXF{oJ8h*%hEP86NUGv`%vm>)6TXXDg|1E< z%1RN${Y1T=7ZyK$|MG&PG}nIgUBeaK5>FC-_H&-GSF;6Xk9aOp7xeBQ@9gYcqIIuU zXSqy-R6D~$o8;;8I8^4{7Ts@9(;bLc(rd{bxwe-=a_Q5vahoZ1nXjtE>G5!z5|O5F zqXkMD>IJXi;kH679d0kx+ndF?s3urY$rkNw+YcJt@Ia@JGxkJ2d#Joohu$JGF!Z}5 zc)bjpLc1yZAMZD)FbXu@PF=MrPf5v5FqX76nBveX75^;qgg;UB8;x>OVzpzgUe3)p zUoEjkMlGpz@ITE02K%r}MQf|HDBk(OC%)-a&LBp+o+qk!7-}YeUQ}2!oOx;AgT3aK zv`obDvzsXW9o#|`W_6vj5qG_N2W)`wsWlI){vUzHyMW?58V@MqjHIek*ncYLhDi|bQPo{r_KCYQrNEJsZc@`}v>ZA$?#F2v>PWc$ZYu@vTbZ+MYFxP0 zDEbpjBD?vZ>hvauA<`I=V5N4Yb9cfclZR<}$n?H_zg^C`UJW&-1LA2ik>>xy4=T|b zn0 zD=XnSd-q))(1gy)PCS{B`gI^`Y7KyxTg+lpRih_>4%@Rjh~u^Z;%VqsJgcGvyjC{S!?UiUo=t?V?Bqn zCkR0Ul@&^MMe=5n+#g|jL!GcJI85;;OMbelL9Qu(pnNPnj| z{`-NM8N7L=6u601&@cuUbUT-QGBiKLa(#_NMSfwIJCrACE@Yy`u20Jc9!LfD6*;eR zcX>NtA!sO7>oY!_mg62Hj#}(#P;44ob4etn&37@`coC$#iAAsmJQy8+T-Vb@=^*tw z0Bxa}FR*y{*)j7oNg>Ba#DY|49;BEXVu#MH*HuDQKw>^|tX<@W&uqRcKKC zoooDIwTMA#!0WEe{fBV!m-x+<;-eF6rkt<-^mj=dbs2Z`J8u;=?B~V|lqo9Dr3fJR zd1Sh&{F-Y#R6*H$+3+uvcz80!a_ti=QbYK7Sri|4yX;7iwVZybcL)I@p5#p$I$3?5 z`me4iV{XKPX6CKA&x4QO7a1!j?|U_cxZp8#9%Ky&Wg0MrfhMItTb3B@Vfwa}3nmKR zsTatQ{ge~80;E)8yn}%95<423%BoW1sbi{vI%ci-YO#SL$QN~}%2BY!uPt$&9fghEJ?gp0Tx1EU!LUo|tE zc|4?r0zcgBsq|`#YHbmZKR>|c#Ek4^Jfz7&CDxL7dEChUITW~!CUwnU9iD%(E}8|9 z#;>P@=g?6DSrFTpaU~bxCQY6SkQSQ4j_1JBR48K2pi$AmP1-weK-xcX{~?n9WO4Hz zH=|bmTiS>b(nM6shkTWwy<~m`-?cMJBz+Fkd@9sd?|6N=kwYwKwWGI3Z|z|wuZ5<= zL7~(+?dtptOE=*MJAlgFV`3)QPq`7a+V;W?uUz6f8b+DiKl6Hv0J*L%VAH=prIp@@(r< zg|S~gIcBW{Qk=i>d_W`jxMy>a`BU&a5pODlC34LTiILpbRKC7WLW1N-C6VW8-rwFr zk-Lt@$W5v*;E9mz2ih|g^L*vnt20;3I?C-+u0^W?qRS6A;l`c2#f%^#-Ayud`Y)#Z zR62Q10D6gfe`Kc{*t6#byu{7lntO9ee>-TFL(35OUNE{ok{BYj^?`kEhPS=BdI9NM?{a@)|i-T=falEm>Y$F`sgI31yiz));1}@7Co01}$Qlubo3axe z`gA!)@=p(`1#2jevvAlUgBNoTy$HjIVcwXiWjgoqiR#<=Sr}V|dME8-SzM0y8S<$u zX!431W~y4us2PYjqK~@ywRohB)5v}R!=UwI#lv28f}@4%`dK9G? zMT;GK%ebG^`5BxJljU z4L7#3BG)j#zC4x6EFI)IF`>EeE#^r_pObA|)TQJep&aG9Y2CW_lbH9ObLS_+(4metdr` zyLMbXMfU=RyaDDuHfGk!`(R~JHhMp-)Qh}UOyOAJGjFRdE<*L|jD>|!E)zKc@v;%; z`JXcA00ioJ`u%zP@yT6n*JGtuAf4v`A zKS77ZARE8)ff9^TiuwS(-Pcok6CE9avMALqbRqhIG*3Z#%8&I@>l>NqganSU4;A#x z*L56w9`r2s`SG+Gtw;0UnXIvEXfHqaxftdhDA(C)J#?qT=FO>bD-R*fJFUQn*z`~K z`H4lEYk+zJJdNay@x$s{F;wJ#IQic=+5*bxg_R|Ee&U7VVESFpmFORuc=q%Wo&_})`73zkG1>V2K+m_`(K{=I&2LmoVndoFDFK)jMR>3aQyy8O7GLtLzS89$M0SV zhOZ7yev@S2B5?+KqBqlD$-H6?5f67JR8H1BG*}&`LOFF!X#9YLAH7y;?JkHrZzOel zN*?qK1nXRbS(=%Vjh^23t?VsLozuPY0?7)>P{GS^fvA>!|o~u`_$lc9x zJfpc}b!k={{J=?zZ53DPLc0}%gzr(cC+#nuw|;gBCcFt51s6-*FvcEYB{&y#4~I(f zo_6yz*_Wbj&IeB&r@|&?ws1(<6H^$oASBS^B3zLkn57iQ>E6`|k%bm@|!BQ+-bKc1wT2un<7I2TAfEQY95FrAAkI9kkt30>l_`x^%wNKmm-<`gio(lMy*OcLTTm>MY@!LKim*J z3Y4#Cl~HE0m5LG^o2_^unwe?>wfIS(#7Cg(<h^Rx5g>m+zEWQ(`66co7!U(_vTJX+Gx+ zc57w>2PZD`{Xmd{oCyC>hYp(uvPDp29qk2=Sze&mdv0D6a)vehPMgUwUt1pS6CGf= z2BZVg`Os~(v|A6zc=CBk#!4A5)`em`*J3NwUg#GNhngiz)dslK*0m@1I>%4#w95 zLj2i*mLkP4(|&;lE~_kxs}xyT?{D_*R1ll?mg^TQ=IZ`L=t?HW{#JgSZr+pYN)7Y9 zZ&SEf&SM#5&f`$HuE_H2b?*vA_W63(mmUZb6kGWb7Zn*b5S_w<0f1OV{>W)a2o*Ah znMZm=B(6I$enNHb9{16tEEe*!ZMVWKu2M3He7u8lL(2~v^nET9G%S`6lIU_dX_FTO z;V5_AXAHkPQ+!I9*U&sVdVG-}wj|ZGF`gHkQ(O{@f)qpkMBV4Yf|E}J8j&7rqf$k5 zf&A+W+3#ecue-T~ciDVL>>e~hJbMSY>o47tiz8mwt6eSuhKc^#p12plzi~k(tfUX%RCFc=DQvOGhI9PefNf7;S~f6&;s!y zi!e)9pPcLP4UekzRaQ364|~?z3Z9b!%i-}c=uS>qu&1HEhDdogD4ixCGC;nF_V9z73eOCD2mP(DGN zsOfXfFK5~h1!(=tSCtPtOTszo8sH!KdZ5X4;v zzy{};XFha7XIFnjVN4jVp9kedL>pOp12I}f$Blpg6Y`RmXD!rIJDaZxPc98b)kaZb z5eL5ldmv}gZib+d3uZE^W0l?W(DP0eSJv44j)Y1Q$h`coBW{hfTjdM{P3581xKWhT z;!Q=XRYhNo6W&2y=OGKw|FsaKNGX_K?aY&y=DyQbUHPkudLR7!K9pPWma`q#zbIPM zLyt<(fOM4%5p=kx(01TU*-784M^D|xq+YAUthue7ZdeOwC?21uXV^)w)gCh22YEkKU*uz4KTa z2o4lvm3c{#o0nIbRXCqy{auRnL6~JO8&)P(5Qdp=)>#MYMG(8vUK5@!vxZ=zSq*Xp z{6y+!d$cRVb3pp+Z30|(s&?hoX)(_ELI-A>^{z$}sw>ji!+~d`u&YBOqnsY8G3h7N?R&ZTLn9oT2lgT4g6Tv#}PK$on zMZXmHwPn|qQ$zK3yF3S~g1yMNv%!yb=b`AXG=_tNJb?P|I!l=U`uGtC+v(xCk1Pc$ zNNtgKqCMx(=Q>R%qiVy?WS8rnl_e2`&gBQgYyD=Ej?~ zT`U|&-p3W~)YwNrK*ua~>4M_SM>-J4_uSpNuSHniWu3m_b^2xn$%Uuj?k8OKI3o)d zNR7g6Bwn7gyMYLvSQo`+<)bwga(46MW z>j&(K5FYN#s*My8a>cHoY9;wWgM;<*k2v(BIu^(z6(QIy5bU9k^YN|%dB%~#g$(4t zkU7uSAN0FR%h(u@c@w$)M`TBvlE{L!?x>wUT=@OG5VR;!K$5Ol;s(Gzi)8({7yT3xtKOBT*dUM0y!giai$E}JdAR57cCu)+iy3S)Sn2zuUuvqw>A9<&&tFjv~r9ly*DZD*vD(MZ?HNj(JAn4*xN+^N+ z9W_E?e!Lv*)5Ne`=XKGeeow`Y)8)0}jti|uY2Rf=r0&A$2-QvZp+PGRvw;nL=i=R` zwo%3`tb@uy+*+aL`_A@seRg5}xbDKpNcr@!&HyGnJyx%L#-hb>{A|0q zf?)r8?${5!OBQ-H=)14bM2nS6*$>aG&O7Y}g=i4RB%{p2p%@B53S?NYC~Fn&Wj5Vg zt~b)~eoZ}pc_g9ooQ7qNA!M%GbnU%|&+H9txB12%Ft}dHL2QKgtf+I{jf*!SPqhZY zdz6raC8fryMh^tCshy}VtwvE1;gpgRkpRSh$()#(;ZoBXR;oCEHtpW$r*^9K+LMc| zT4Nd@LuL#Xgnr3#C#CUwHp5uCDEh}k6Cl6~UcRiFY9OVWVX9=?m7F&bny8Wrg&~x2 zc92-kB5wWA>0{W7o#DaW2r&1PTJQk;3O1K(e{waSkeE+1T5WmZp*S&=@^L}>)!Uogva@BElx?+}Mdi0gcc!o;U$p><0 z`lJF`WJxtFGL)i*f`wW@>hYw~9q3H@Jpc_zo5!hwhF_uZ!E+8^!DrTPmzqN%;kxUN z{L8xgV;V)HJg})Y)C|L8W1JzDgBDgU8V66^?2D4y8Il~G%{dD6`13s|eiP)E2~TsY z4&`?mlw@$@J9@kB!dgIWl$hALAz3SDg`e?(IJNP6PnR;mxjOMcPZ!tdP84PwwP11^ z1QaY4Ao86waO33w5vOjiJ=8b;{WwUG=0SL{w`g^Mhbo!LSMa#l3;1eG$vWoVr!u~T zqGvD{sq@k(TI_w&lilYl&5nM+<9h7iju7kh2p<|$6f;9k@iIUdXP8T34pqz@@O zYg7|%-D%&<++L96+IIv+Jgt#v<#3LscpSX6aob+(q=Mikwcph7+zmz0#$zjXYN z+4)8YtJM(dbK^UIG*|6v=qm=vVo9|CDtyc*rW2HbU#u5r2oTjoz-Vhix~;q&h`gg&}?o9F5CSo(y`b3 z7{*uH)qkYNi-Eoivye43H17-o(piO_oH~WV9>SRW*46yYp5vMA+(n)39^Bd8_tqV@ z*WHK7XOch$!I%Kc7}i-CZTFZ@(0TC6og3m>kc^an@5baqm}t6$Rzau2$9(Y z-GAB4<{qT8x?t+690(al;{oHGa~(|ATo->pLB{IIGS*1PHzhNI}C6U$O>bGTs59;lwE~`H*yz5b`vtn2~E3YxIpit~jsBXCreNSS{7+JVHVh!qU zgK^GmV?3+AQMW&Nb-4i#N)AxWNY6pZ9SpDZ{_aJ9&~hz)gpJhr!n8x3S0;<4w0L!;s)G;E0 zB5uT9h_uABoO}qZ{?{6pCP1+&jxQ*xfR*mt$g8;5We`t6cZE@Uj}o3=42tIfi#}R% zZ*iCX;1ZzDbf#@8e&fX8?esSS(vZn<-*J!yMdUG{=y%A%u@O&+Yy^;|(yU;DYgSO$ z4oYB8&*qx~CjZ4WAP###TEggiKAcRU`1}qmCrcdB-g0`~R^~_N0@51i74g(s4HQ9j zE@m&X4L51>T7a|>XLdZNpqyq0E8X87s#xGA?Oie;&0gdMoVwa|Id73R;sVK7 z7RvoqUY=0S0`R}2Adr;G-1^&H{#rYjI?~z~OgH+o-2Ez0%IhocjM3b0dBjx7yF+?uZh4Z4ldP9_6a=~zLIwkslJKCxNs%-*Ee8|<{F z;w>q#Y_QzuA*nNmNe6MVNo8hbWbkV|;`d=4oky}ILxJDVkxsJ#rdHlkF~P&sOu*EP z%P(*?LhXQsfOfiCgR}&wGiYOmn-_{2#o`JHT=Bxc?;>Q#%v(~9eGL=f$zPqRT3YG> zM?2Hs`ucMJCqD)q>R_ns((^W0n-pXg>IX%fZ$r6hi=G$VMUe3;ePJ*^_MvNbsFwT* z4c8ZJA}H2D86s4YURnfsw`j|2txZB1`GGFJH(^I?ROS`MyDM3v_Ap=-m4y*^+SoK5I_)d(=x`wOXRboXeAF`cN383`{KXEv(uWU9_X%-^-~KMu z1J$11&GjUns&sEuQhA_ZRlqMi^O>pq@=&w>X>8}DyMrJ|-)D;igkBj-@{giAJLxR* zr$4!B7A;jG<`tgwYhQm3+73xXS~e-1zNpJp;Z&>!4RW|ktD1AZ(hHM@jcJ3>9IDi z_vJa;WtS?Xw)@gR51G#5#YgBvpDC#(QA;%r7muo;+D|tu(|%d~A?I;ebT*{6Uvidq zVQ)X$GH;v)Gg^e<_zF@T#yFv$KgA9y4LqK350{jZ9m%F%wh|W3yg+)&x%N`+yqa0j z;sMHnUi!Z0SQ2%L3Txxb-eA00BM*wMwAStuY^5h#E$4CHj~T1Qa0)LE$G8QFJ--j- z>x*Ex(Z_+~(~(f{F9_8^>c{sm?`qzVABC|5FvzeEN@W}R4fY?FKcO88?3&(XH&nA1 zDi7HrK~?D-lyX%rypakl0Y=tkAe20Jv`5>O_u7ToN;&t_V+H80W7;bX3UUv?%7l>X zY#3^f&9&KH=4<=yfQLUY$*$&SbD)BT{L@lf?fKEfsm!bbE@;(C4=TlDRwl_o{_@4o zlm|0odh`N)U^|nfFDIMByZIszpOzwa(;s|1N!@pT;JHEx6t`r8KxZ|7+-Bd3QIy+R z?2cmoW=*Q5L^TCUNj`6dZq-eTNryITa%9Bj{##wa46i)IU;K9<8FpMl1FVWVf0=w|8PEnf?Vu^QXH4F|oF>g~r zM_UMrjYEO0%&bKprlB&knB?;Jx6*;7I?LaKXHGE-IVbLk#X#@^!kXOy0lf!Kp*!s4 zUh?W)OhpcT0d-djU!dnxP?^oE*XxK)yXypr_r`BDB50tlVKTng>>IT6DR+Yf#yucE zvqV-d98^2{%pKIif(Q=Rz7(dbTQH#C|#Gq(; zQPh_Y*+j{}t`Tm|aO&nI4C1{Tb9-L_l=o3c=Fsv?_ayNYF5aBxhEj}Qh<1YqAlfpE z;-&>jke2<>zb3Xe=PoWd-=I+1@s>R~gJv@MIyU9hJ^8aWN36eDSXcl|ymRtg zk>df6${k0xOrD+h_*)yY&8P&fKz!77>f=sdCYFxA`!ZzUb0)RSqQ)7H=eqCid@Qh{ z=pbfC?Yfcy|q;>KSXCT#2R(;30Da~kh4NaB;6UK8X3xo>3r2eU_hv5 zW+gG2z*U>b1;NRcM76XMz|j5LP6ICjR)C9mWzfF9j`)*t>0psGu7Id~qd~t;9s#me zrE!dOF!x@_S4cTc>%9}#G7YizlbUmxED#X0#e#x+4Qs3ZYd2uC;(OcwM|;;E4P_dK zQ4Yolu@g>3PE=H^)FL}(w#%7NN+p$b+wCHfF_w|knAV_0TMU(PE4#*2a!D!56tlI_ zQj#vj8MneP#;v(A`}^8_qvGtJ`_Il_-g)@ALeg=lA>G(J7;}6dsZCC3k>%UX9 zC*2M5SPvY=TBCGv-*#V69apd^864bcQ{!IUT6Pn;#+(&{fEN?%5s}ji;ela$z4x*H zRJoU|s427_N4eg|SCG0V&VYXfjp@B;M(vpZ8e%0Sf_3;4h4QvfDCsMVLlPQ5geo`h z4eLtvhz+Jno6U;(&}owa>Whk2?Ao=}AA^oc!Q!IfI)}KW*-@+5PmpGj8r_Iy&9n9(K2CMMTkg6HJZ^&93!NqE&eHZNQ18aERPU8Z= zlWlw(cyLKzkfpV-3X_<@l}E2vgP+Rmzpf|(C#0)xHc#np(sP*Ho*+hp6~BW#4#ih4 zH|H0J0EnkOeq}^VlRv|cKZ{LYTl@m_}+#l%WdUHG$4!`oY}UA|hV$lmto$R-lzFX=FLyOt7_J%Mq3kc{={r zT!GXR3DpNdGPfNW^Lqephoquq%6Z*vDA?Tk328`I0p1iAuWhk%MP_CuFXv}1Ob9~^ zF82b<@^_c2E|yPBFgao-=}(h5M$0gG%-RSti4fJbXy2pc&fS|!A0AgTitvp;C2sGU zWh^4t0h>smq+a#dE(dC)L#)-e^6(0Cw*wHfWioMDdiY!MxvwGjsHe5L#Q*7~*6`*j zZ1ILl(0sB>K<<=kZJG&&8#6dd``agh<{=y2Uv|IZoIGSLT);YFz9OXNoSnt-MYRa7 zBIK^KQ4sakf%jXAD~ydtR%9=m1iY0FQp&{$N=YRE7X4{1G4AMaH@!9QfW2(6dN}m{ z!wBhK3pO>;%rSZtYzHgzRgQLHFT5yotn=&wa9Oq>1SQ+Iy$9-tgzmA3^Lcx0v9Wm5 zDI)ov;$&qS17pC;YKYZCLs>%D>}#9}36h=Wt(kCYl$Cv={KbzqTE*SA9D6)&bN$%? zy3|oF8y27Kz@_v9p73C)`ig)X=}N#NErH1eiI|C}%0xo?*F&6hD?R76uXm@4s*Ph2 zYl8<$l~Z!RaD9R(iPi?P43DwHRl~ZniVX2lx84Xn=!2>IPk{L)5}Igm&%T1->)46T z@L<(crq=Iep;zt(zb-w*v*n1l#GmS+-04{8CmQ7K51y;{Dg<)@FXL&8+502w>b(c5 zocpMctslnNNhUf|whq{2ozIY0aHJK8?@syy>(QAaN5tNUKF_xF7+nrd2XBD5nIEHF ze=G!6rOqyw)f7um7144L<5ZU4*Yk`OcYbx2T}bI$f;%?G9aVJAVJAMI zlr=Z|g3ZBL0H8TNyHcC*b@z=@1VT@xdPnKuc#ToWyjYBE@mspSG5tzk5oPBSKu}Pb z`J{NVN*=@fi_$=}+5i4?-GQ!s{OPOElvS`u8GLlt&dEJtvhN>*%WbaExQxUP#`WYt zp??i(j0)_Qn3+$SuYnw?b1B6$ly<1^L`wga({`XRf%1kA7X*8_J9!nNvrVH&DA6#G ze7}6=qvn?ZntQuGKzY{ZU84b6QQhr2eO&-jl4;sf`{*UQH-6Rz3*>wVwnd0t(Y>-zRx&biNhKKJLo&pG!w<9l99dF!Tqo9O81 zwyLU}(V?Sb?4YAtAG2`-xbsb5T_hbH!?x?EPoK9rt$f8DTijI0mn?No@!&k{Ub>(h{s@NygFc;I_mov-)Weo!p3>^g)3Y~P=J{$;o` z>N>AP@64O)0t!1%vVM5m9neiR{J?kUM#{(&;n(lazh>~^|6Mn&{OTJ!GcI2ut1uwX zrnstuI1q`4VY^&LyEZ@16EY0$;-MGXmnPg_|2?Z>psY;h<_BoqR%o5D;Ip73LA>ks z_kKjRjo`V;yP79+r>IuAwDZdA9(CStc_&@@-**c-%C1kDMt2xal0GvxU;E8mCTknX z^?gTaHl6;SXqKOu=&%<@ADNsL;!pOYl(j!%!g5buW?*vp+5SxK_}f&0|EkV@kxU8K z;;t|DA$jMV?X)Rt_tEyx5`*r;8U3INf84n+rwz@9UTz4Yg^6^Qb(I>rQ+bPNiwtKW^?!ftfrxK{#KkJe|;;r?#i?+brty5gdQ0_jTMlF;Y*8p8Rk- zSx&Js5AA-B{-hQo#oB^yLEBUYHQO7@K<)LP6rQ^nQ`qD(c=NE;z8k_fKeshRIisl` zO|Z@BCRnMf$Eq)Gf)+R!mMM?RYndy=3+vfl&5C<%aQxylJ!cgi zH2-$tMym~mGksHE6kpnWnR>~|>)qScr_fXD65|}>gt_zCs9VI@#M#7&(-S2wT`|R+c$E9ETdarFl96hpoOMgi$ zKd#_M;eoFS+pk4k@ZP^FOB+uh?WTw*pz&+y4kUCued(*16I9m%>%@{Uz+ zF`n+7Lc;Ayo!fk6*D&eU34v8#l}<;*#kPhz5zp*7fW4F z<+;82lUY+0AP8xTORAQdnsi6N^+vjN53kcPfU9-jZy)%hqoa?dpS#1@5xaiwdVLJ- z;jxfc{&aLobgE}g>fT;A+^PQ7qSb)uXLgFczMCr{W%`2NhP$y`Dm&jWGHKa5e%C$g zQRk!L`1+ZSymRD9$GSndG&4T~!^z0~|J{9dfboFfrUzOwWR+9xgUZFDzupw$8d9S? zQ}70^&0Uz>CInaCu%R1*07ZFoeNhZ5i2r@|4&6F>hK;OTN^2KuO89>T2}6tt)A|ub z?oOUV5M?cx8;Mx^!itl5745?z8)t}D{@zRt&y=dRRO+k6(lP$(K8(! zHf5Y`AnWDsNFUIY# zW;9GowbG;=g&Qdz>E1LIIJh@_i3tVoJfA0;G$KN)YR~2F7b7TKwE9Fy{y~mlXWV4n z3xnt-uWuaYiFL$0sgh!&)erZ8r=Ww!ewz{SqT~T4&Y&@}TQIT_y3|6s5K6BVtwPva z`XcgI6y$p9s7(n#M@$7@3vQai^w@WZP851N7!yX`SE7_08%9jDQgucTY~ZM~URFiPPf$e1kC7@jDJ?=tRS=eb|idPUbA~ z-0)EP?y`1o1m2#7^Tj$^_|^Tk0 zNUWnaN9?ba<2bGdpe5(FC9D`z(82qF)P3vr#Vix<@7vcvys}pD0#~i6da1Q@Jt#Tp zC$8}~Bt65H=WZkQwYNFVfpUE+@JFd71UcS>tH=4ClRUD-Ko9RepU1Zi{bRPY*!Ys3 zf%^JH9S#8qu{hZ6xUA<92T9rX4@>JZ<&AZWf!{v-2lk`vExSki#~@SQ!qfF2T6ga4 zimCHgNFy#8hQzV~^58pp*=H8s?{|deJSB~z-^&!&Fv#6BJY` z@Jojz^RmMsoA^7lnfu$9)Upuml=v|09A)|T%X6vd^DlG<0&|0$B1*1o$y`6V*4-!h z+??T5JpZ#sfWUK5T!g{HCffTJj(#nCr)X`7daEA2UtRi$+q^g9f2Vjcj)wFYvjPrC2Vl%lu{4!UZ(6HVu7D!l$^3J z6w~bgSX?0&lS7p{Lp>p^eTA225nYz>(Nms8dLQ~*Vdo|}$|Zr2nCk26xSz|FR!lUkRh8M*Bgse|z7?cu9W<%z#A*qqd5HYJosA=5Q}=g`#P7n6dyga$G|G6I0F!d>9C$$qJ{qJsRdzY8Q3VQqn>xkzzxRT$^6 z(qbtOvx0%u#xIZlR9zxX@3whr@Cval&~aphYV2quJ9+*#HLifPVx zYmK`DGUXw?NV_&XrDrCq+3~-{Dsl#mbLlfz`%|l_7~JIct3W=H0fXsMe|7iX5xMd3 zz{SN2GtAUA>`ty-nRnnlK-HyZ?XNCw#GE7>S*;I&==9Z)#>!9k0OC-%(zkmBa{>JT z-wi;pR`Gn_t*Lr3EP;WnfomQcvlnaTv_L$wL!UJQnU7m%^|8j)OATg?K$hxbjX>52 zWG&|UJ4BgE{wJNpwg>GkALj?wRw9n$O{1TEY?`Goo$@@>>8*kI zBEs|p`?3|s4`7t}fnI#uUw_3FSCaFN(${<}=sKzAlb@`u6$SCeWq-s1Sm9iN$hc1K z`Ll2c7HFQ71Xblrf(!;7g@PA0mi3FK|5zbl$w-17GPh>Tpu8VnCwdzH) zC*L>yewE~uCBAp5+nfus4ud%nzD-qc=Ucbrf>j^XBDFtYu9o?!nAr=Z|G*U6>vX?W z=3AV=+}~Qo3o1LjLt5on%7umLqsaV{5!pg+)NPO$;(trqSD~Xwk3z2SK3@F7+B|Nd zvw!F)^>kUh3;u78<)u^SpoHmy{cctKMlNtXZ``$w%rdw^N^HKP;k5vPQH(bZ&e;N7 z25#45T+2t=%MBn&JwqXX-&&LhBpQN&27e=`~zPVq*XJpLxHZ<#PZ z9sk3St2L%)PXAb@? zaZ3ZAxdN2dDxSm1A0uB=^#aAxK(Yp|IX?JfGHd3vz|k~HStF3S;NUO8{x3@)(w}7u zzAX4tKViqb))A^KNCo4-fWYi$sY9Sc(9zEJ!(WHb%{wkGF|ZtpJ}bAd4x{p)#$U|f12oI4^Zy5JulaXa>h=|mUpw2DSY^!4tK_rj;5Ng}8}tJAc)xMIK(bhxMl z+k$k9J3~ncIMtH#erHx*#UTQxN<1ndKIaOT@*V@&slv47lgdI17Fq{VhqOWB)LO;! zDqBA1?Ly8^o z3A(#HNzQFU;y+paG)sj)1oWa0#mQBRB!Pm&9a3ezreTTNjnLsGN5oN5h1< zD|^wcMWe*J5Rl7+kh43zvilSSa_Qf7TV-3v1lYKV7u0TFRNkU-R*4#rYpuzPdS$Jb z7b$~=$A5tRsBh1&xw8P4plS`|i!5u_+5Z31t{v0A6{jtXTYyWEfu|tsDdw^hf`c0r zkN)hrO004-jzL2E)D`CGGV%pT6!YUcxQOIR=;$NH)a!MPD~kq+NnV2EGnR8{WM{yp z7-g*r+eL(VZ|7}u66Q^G!A|G*Y!9cukm7BrHw#VX-`EP~b+1qLHuSYv%c} zX%m=cNATc{7eXeVO}5V_aP_*#WBfkZ^Y6(3C+@Z6wIIS0eVqqO zMhKJioWIsswx?r)X{S}tf*dZesYG5`D`D|gf|9U} za#J@2v=yCrm4%%@KOKp&qGF`j^DBA&ng{&hZj}k_J`NB4`5g{R1137SB@n{${kD~3 zBKQk1lg}M*r~ePMW~H(Qn~NRJ;Y?S4x(94DQP63-wRq&P5EHFcycC;2)!WWpZb}88@nu@;M{iQ

f3QBM&a=BE=;{ASl&xU`a4}h`^fHJ4sFOM;Y++33FLZA?HVWlOXst9y7z!_g*ZQQm2nxvB*@Re3l5ML zncd=6m52jyo1OPsS$2kv!-7CV`qfh_4+?1qzU&;Wi2c1|QG*u?+5)nYkciNgcGAx_ zw*m|a{7?SkSw%PL2r!^95yvd7x-C%+!?W$@^EiFRs3V=5-<#{X^Zj%wjDXcCjkE6C zUdHtJ@yUC4?)-ja*>FL^_p7zWcfT}#3EzEYj}X#fzTtibs%5IL!$RNshD)m(eHAZ1 zTz(<0H02QgUEvG5Vn@c#TffP8#W_{$EtOlKt>0XnJ35kVh}9+6UXB=Vvx;$`Y7vm1 z^b@sOS5y3ML9QO;C2_(kidSM;$2`TsFF8NxTZZ*dn^f{dFKR$n(M;WGrH{o-k2M!7 zdr`ldX3*~#eX1Br}y};U4|Zj z?=IF0i+u7racR5e?4eEpz-iJ*V*AR5s!?1a&||G0V-cHH7y z5(~jqw_kekFh9+(KCR2XurN8ns%=-6^=vgtY=DF!9Cwv1^_ymb=>xIJ1(>fRLtRoR zL%peCYyIlie+t!P{TY{}nYZcMBBq=(AW#atX|qaRE@j8iihi77 z`ReODfC)C4oI1El>AV|N@}zo>JX+<@Jzdl4Z`ZF9=ZO=2wQZHsSJw5sq^azHKFB-`i)&mIms#Hh zrL(2o-LvT4A|=HV0Oh8l?#axna_NO8a8?c<^;lO}zob?8-WYTXA>=&MDy1{ir-ZZK zTV6MOVB>I$yMvJ6MO_EE$-80tQhb>d?$&~085bO#+ z&)5+f)>er4l*=k&%Q%LX6y)Pjveam_hsJm^bToTsNTV%TPjakUbPz4 z>$&1aPFr6JnAKveHt1f3SGtR)hiN?-z?*){RWIm`r(wbv{{iNyoGZdkKrB-^#u@9O z+s69E^NGfb^a+q`vHQ8BiKS!Wvf}=B82s1H=Z>6cT(bUr4=%Y#uLdos+A?}`$=&%* z5sLVwN^PHt5*P!RCKc*Vr8tE zh5;)K+1oL=%40e3xMI9uXxGYyB9sMUf}=jNa|7CCw(OS(9{+#X=_%z|h9jem$+l+BsYjyuBhw-u|VE+^xOJ>CfzvwvWahNyxr6JwXY@SRB|40`NNMg zq385d-zCRMZb?Y(ztJDuo^r#|GQV~}ZbckgKLf1dlJ*Jq;@xMqNvQ*wn1OQ+d9F|H zdB%x8?_Q|~_JrTz)I7K^^})(9x#>$F1>fKe-2BlBi;=;waf@aNx5a~VDOZ=Z*7O8p zf(=SQGCRkk5-8`3k9PZUO>Mp0>N!CvCg3XZ=+yHz42_V#_#z;j(bLO>@^zeGNT) zD=n~;d))o)jHkWlf4gVn$_4-!+`iyt^1?G%FMZ8?_@lOObVf&mrkpx^0P9;OieHz9 zfAS0*j+o@j*_jp39@rO9d?mA(^lpENz?U5}4K(}>DKzkdEcOF{a zjsSu4J4Ti5t@Sav_5qF#&AZq@#!2Z;5Ji+WG)k^)#2S78^<`IX99$B}%)bl23iyi! zqTjwErn$t#0k^WbB)QHn94;k+2Q&iGFtS_T&wolQDFZQvc6EHyh2_ovqrE)|KsGVg z>{Ys(4PaD3Jmj+SGQ9m2-ZT`*w$|_k`2Y2YYd!wIBh=GxT=Phd`_~HBv$r6}yLb~5 zlJ)J2Oa~a(D}G4d7nyXCcYtZc9+psW)iSwF{2gTLtIn+uJ=f}L4Oi{sdgON2Hcrcc zDfM9{ECVJcDY^ZD`|GRO^72OVq^4XgsNJOzo7_%dJ|jrj>*YRRqn|F9Dy9?i@@<91 ztH4_t%(v0U^yG_>1y-J=>|^i`1@~^>E-j4|Eb29b&0LX2sCUv#!GUgX)5A7iICkO(40s+-;NR1bVuS#Bj$ReFfJ%#GVj#m6*JsZBGg z1{4oF=zqJVjKy4FBGs~LR{A-^+%Jskr;J}G-84O0Z)`MPABz4e?$C~j`}Ra@xI6@k z7!+%k8L*HO9kmOW$?z-|!sy{bbEgR}yQ_Y;1Si$t8)Xq8*E*Y~jQgp$BI97JNtN^r zDyq3^tmtB;L3t11Th7!F0i|5!7SH6F(C434NEC*QI&uo(V5#nJs$qv~+Rl6XG`;&g z;GeMNRH2MPH!(%h-odbDv$>&}!#FO!=*x9nA0cx_zZPdfta&P0*!r7piIGA=US&0W{jtR8_$5gnL*E9(yWUbARm zN;f9b*hoIn_m%L@dJvsD+uxCv^BKsMpw^xuL}Yqf>f3)d(TmELjW%jL-uQUvp{8%D zzw~H;;XBSqXSd@ZE69KPY0}H)s?%dKC!B926o^yNQ>o<)P1;R&LRp>? zrOaUUD%=KTU(Aq42CM>4kKrehXJm`5wm*GOPCaY*TarGR1A7HkPjGV-zRGxT0yV zqxWQ>K!YE6#HrPfI-(ywDa2po=VS=XB#AtzUTq4!TZv(t%yS>oiJa?hd2ZKWCc9y; zvc8oORr#U8=7ra@l3&AclIFW zaNcz1{>q^bEF#*6Qj&!Yb98e%zJ>E$J@!1h1NonyN+lf69_636$YzAGcRxU(kNm zbsNUwMh61?bzC==pRqp|%BkdeZk=-z%f>3;hl_yIp}WAg3VADi?6g@LqAwJ% zMG4Gw&^-#Dt7dr>J4TIHne9_DrK3;TP*?s&@>lra;NbVEr>cx~4xn{@PfDt*t1BcD zFqfwT`a8#+1|wjgSkrF|qxslqu%FI|WD}3sTaO5TPr-=-$5^h(g|Ej*9Y5~+$rg3% zsUdaZ;eyYbAC>&?e!+X%^v2PGq>z6*DL};Wn<`)GiRmA4+K)=bArOd-Xnx;IZS4YI z3LBT$CAglCT^cbp%B{;?TO&|WN_Jy&xRP?YRw*R$aq4-!VUVk zea$w9hI6GOBJ*X;DB`NMUWb^c>Zel7XQy0;`GBDEX2f5r?@=-*kCwLkBF01AV!Esp@aoxaM{F8*;~WX zX7vf0CTVtAvny=!4CjugO>MM@(e%aj?(K~+mQ>5bpEjc@%w^ND>l)~TlUS=6m`(YZ zj|58iB-L3#PFU(%I88`hxUuTpSJ967`@KunbEHf9;FNWqRnd83bL`q{5q$0Hazx$n zoQ|izUk;1<22^l5m2G!8Q75hJ+O_^^`!5;?Z8Q)(1P;bwz~OW0 z%PWf1*H1*I^NS3qq>f^k3z1F=MRhuxWAt2V3>!a_&g*`$>jx@5$9{>-=o|U65@r;< zl?KW>=HNw8u9B%SoY!wqHHWjvzOw*@mZrRGHASFsX@xM4v@3N*2Qz8s7_0BxgS$r( zNQ`t=iLR&M4n`Oi-U`x|#zu}Mc%Uf<^rL@ezqzi9HPT7n1}3wW)EUl6=LA)VMxovffv(rA84N1@HP7;&APv|kSTt|7GL-nXwlH*0BXD3!?VuBL2DB=ivD3jFjTGh0 zVez$8jVMDK0*9Vq&MZ8S1BK(VStkuU@|=1jBnB{|{v5ZjQoLg+Z_~*`ZL{+Ro{ZsQ z34E=ToHG;ewRHDZm>40Swi!r)p-+i3v!ClndBuCv9sBwQgwUQ=Y8=dtotPj%)?#+LsC2Au1d2Hb18{qMuu~YJf(F;Aaiq zmJ^_HErr3H6)?3X*l{z8x}1k8;t4J9VuD+2NSr*Jg@)O7d7|F1*#wKKKf|)f;%MJq z(t>Ab=3$?eonD`h4XD|Y(@a=zhzOBt7yW?^^@zFAFF91^|hyYhe7dbcL%tOhg#0joM$oPKA;N> zYHY-v)0l&?HH`8;%M=_(0`K3wr~RJij>CyeS>w`SW5a349it~KwTt~46N9kQA_K4e zhKio}UN8b8yw@}hG_yfI@TeN7iO8L@2}VinmvJ3;>kF^VJTc3+yY}?wA*CJ{tJ^fE zL9Lo{y4u8%2Qf$_;>WJ>02L<9PQ-N2l6p1EzyF>)0ibA((+Y zf$_1@B%W57N@ad4wdufA>)~$BKv!BMw?RZKhk0ttThif<@xcKF---Og5Fq9Z~_ zdK8k!Ya2P+X&T)|Z@1TQJa-_BD4R)2NzR9iHTIe!Xveefd7#8rP8fRx-_q(GHFjSv zmt0f&BT6LBNb?8PxJ>5L*%pf*t#txog>4o%{Etcu_L>l*wN}`hfDgU z_Dd;_nQ5XR)#JH+oivx3;Pk7R1t@#Vk_=SK!n%%NM3M{rRZE4V}s(v`KXx zhI1zpupN92Q4(q1Ld=s`i*^P7CtsLS`hR^8MMsTDoEYfu53Sa&cDOy$EJQhHAVvG? zJiVtxmCucgLnG$Ti{m)dn%tV2v6r z?F!hMUxb!XR%Z!6=bs897&FO%8ZvJE%*a96TQdz?r8*qlA~HVFEZBl|PitPZ)g;FY zaZ<4WXHnlv^{(74P*>WIRrwW`d@i3T;eb@sX^7>{G)Y14m-Pnc4RA zheo@2=hj9Y&~m?7T?BbXe3-V0&n?xEHe^}H>^CCOA#rD0A6wJtt=->7(@+hvxWaaW zA4QWi;UXCeD$01DredQ#M>)$V!$tYt)A1q39yQV?Oi;i9C9yQWj~$(iQRjs5V!eAm zFmZZ$l>a0aOnh#uJq?*sjUpWIb}OGzmzibVP;=lv20zJRks@wM_#U+9Sfl5~flHs6 z(jvCW4yR9dJ#5k!0u^^FF_ugAAYHJv-s-0;{k1oAl7y;&MpIEQ)EUCRU*3|xf_I6 z$h#0)wv+HK>+db$oE@GXFGam~ej67>{{GIA8+fsWW|qE6Y!!qTO478&v2rCEC)vL* zO4ACR!t(TVp>@3{szXR8lxU}DO$gHXkja#{UmF68Go6TV)Gn4+41OVN8tj}4oTp*I zb-=YHd7F;iu&s$2u^PaCVj6S%emp;m+}Zm$jkbqw&+8Huz|q&bLL!_+@QEgSv1pZI zeRNZ?LPc?o%autxoDdD0cS=5wi8h{PH@fL*e~);%yg#7Q_46gDhb8AwZ}@==DdX{N znP-7S2hvNV`8rZY+HA0FZPuLX@m$VwY24q`mVNyfd+nfXkYJw-T`Cl!7)SdAGLg$otYP zr>S1^qnNM=>$DEU%dJ2G(sZ+YDgLE7+`h7e*qo*C5Mf+AoiuLHzm;H+P4*7$n_;oR zK}DYp6+xNa(mefT{JZ`1P+DSg$)UP#FM1?hYL2LXi@n6V&P_pc9Qaf}!+l~{I{kY^3rpH(& z54i7K^AhNVO5GhCqO}Pq!FEs0z1CDnau9;6^H#HwTXCv>RBg^-qo(*`DNC|}0B3uq zMU{hnSCPAo6*axK{W>NR04;!D>|KWIEAiq>SVNtuFHBj~K-6}s6bNx6oht<%R=aL( zLYSah;)Gd!X zdG)I@N3(r877RTDQy&tUFt6`!SEJ6Jux2Zd^%6TeKTZw8m zsmQZ_e6l(Q+Nh^*H<$)7K`T+llcsP2ELz0IQWnLT`vh{Sj05=qZADn@I(_LrL-u|3 z0k)WQ5P)l!k+$z0_x|;W`&pg~@}+|=RFrwj1~QdOEOgf7;B#qwV=@gB)n zlZF}AOOYUQo9Kv#xp%ptUY&`Y>PCa`FTVJ-B563QN}?+!s=&gdu|LFwLN+GY(#QspLVh89nZq2w%*bnJ>_+Ei0WcRl^ciP zNZ&>ka2A;7wT6nRv~WcuvgCPr6SVy2bSP{nDKSkuliUjzai|SBN^aHJ=DY=TH}6@~ z{0f4qQ=>L`-|y6{u{*n;c9ub?qTF~ZT&}(xMW|4)Wd;v}zEI`t8G0EE07J0_$-O@p2-D4%);UmDl%L>swJicm@4%CEgsL77~@N zQDC!$97+4~t6k!)Vw%_MHeAWc=v;kL4$XD$VBBybr&cP#%7pl$D8}Hi6GfpTtO9*N z-@C6xr_E_r+>@sHx49V(_{0WQGIZ&{$ zGD;lMfpHPuW!1_YiFu|VZ|C)&4?#F;5*yAkROFKFK$}$Hynv(5v^I+k+PGbeTDZ~mWl5zL={&9~x&AHdDuron?ZYh=s8dT1W+&s{JV24yRg zfWU4@hFTpv#J@^vA{{(ob6{Z$TFc$667@zSvAQ}~+rob3#5GXcH0$lNLo+s&Q36qC zJgr1}x|m{*tfJjmk#rfs7|p@M?NwJk7K@lKY&bq9BzodTr4+N7uRK`TKSG0^0sl(7q!-3(5!$56UIq zrSoQ>OEVj2!>n0|l`wz=IPIUGn&*E!2To>)4it9iaKPPm2rX0apW4+ZJKrrYz@q5V zqIrLol#Rm!gSI}g3GLCi7V~g0$!`92W{}Pp^OOC~=zKp8@8%uSVlyIz53@>>2arTF6Y2ugM+Y}iF^A@r&)2I5QJ*Ijs(yDUCe3GApRr({4 zLj6txe2E8Mj->{AYq*n!oh_e)kA}3NRo$J{e8Bmf{B-3q>eN$94?Xjo^^(RwQZ)v?jR#LQQ)nM6y$ zK!*UNLO+uc3Fqoig{(-NkPR{?-xrXDsL4*fv5hcF%k@;viF*0XX5X0@^+M%@BEk5w zvC=?ARriZ(ss)O|{ZkONZ>TZ?LV)kh?0U|>b(?B1R(4{wLxuZAhCRMbSGFOwqMX^? zrWyo$ki0<@ZMDQ=XJt~qfT2yh3<()x$KzgHck{}Tz%thpqX-C0uD?6bN2l^2j%PdO zdRDdiQA!A;Ao1jcyWPHG>3HC=P4UOcA-2N}`G>0b2+8KNzoI#ZmpSqQE-geUe}8en z0g-X-$=2f0dn7BD4rR!#_#)Z<-+{HdqaLFEA%h*n&J!5?ts~XqqUc~ze?I{NRmkE< zeDg#^YYV75H!-p)<@<0>{A50*=M0hcYT9iz?<%92s89uIPh*KWu*@Oy5Ayez!Q**j z3fV?^@@L3_e6?nJ9H?)qr$UHeB3ZbwGOJ6ym>Qz~UR z6uDCgF3-k#PPR%m-@I_^=1_3ojF&6+M(M4Eno?OW{iyK}lsTlbaPKrLzv}mRjs6QF z+A2h>`!<+KGxDJYIW)~Ks_~&!&rEC3l%?^Gj&V&qzRmqclXr$q`;`>(M7~^GLi90F zw_Nm;N~G)!KEs&)Z=4GYOCWs=WWQEbB}3+Iem(0CK4J&%7^b#dl+bpeeUjMbY~^HT z^~#3gX^Goj*%fsRr&6oGy5KvcVv-ZS%GfLThND+$~ zHjY-q3HL}JCnhGr6uf#t!#tNcu;Wenw?_&nv30*bj;LEQY$9R59Z-GRG ztcRWGHxOuk6UeD{d|pP@(2Zc_%{Q-FX8x!6(iz|<566T!wn4WgQA>WjYIXUg;K_u_ zY*63~JP$k6ze}Lse%dWQ5}dwhL46e%09mY2km?N`7R}f;PrgDIOq0G~d7{&>R|)q# z$@WvxK}_jR8BpUe7v$l7fBx=QSt1LppD_4Kd^E zSZOi_VYk2kyN~nOOH)N*3;jxOEn?P^uXpjIosKOukt-clat%M(g}T2UGnSpk`v4l^ ziGotTmbso)Q_Tq({3P&lq)=dHe%a0RB~URRaR@iN3N*{8Ui4(%7_3dN(v(qiV~O_7 zR8h-pIf})rq!#nejnV zu8g@NI8424%xjN<_hM{}R0G?SxZe4DY&?$I z`VK6EmpLbgbp6?8>~icz+l+T~DK=%>r8N>-FTh%E+Eo^M=w7yssp>VaWXWwx48RG> z2TM)1ST5?Te~ne0wdI>@wNGIa_Og7Qjk2oql!R+D&OY(Ic+^;VTKPo_a)1Crbuj1+ zURRNKb3buNTg%;|GAD5KdInSkI~onk^gL}x-fp0fm}9)V%4m9aEqq?9^A~{byhD&d zR-w&Rkb-H_?{`%9=A01MgZ0qn+xmpg6rKyn6kOi8r(SiA&Qy3>LDU|qK)Xrg*yliJ ziJGediHSugENs-0rV;}46XGX`!N;0Tcw12nTQ6$AK~zt>K#(EM5Vb_(ux2vHbAmDF zG4rfd_F%RhIX^0miPq8Hh6yF;zX>cO;S)N-bY3{p3PkOx^0KOI zx2j?}iqf9jZ8ooizZc~i-Fk!C@+YpCi1>bIZTsJ$ zV5UsGLb046LHUQ5I8HlFKV4&vsg+C*Fbt492Bjyo7ya6%(An;{2PWPGzC16&Jd95~ zcKyh&pu!$)4S_0*O1?~gc;G1KshR{Nt-F7`zrx`U8GVD2?2i!m^sZT#;~<118xlPT zGpf8o$gEmm-1#ckF8&R6*6u9u^gOk$T`Y`BmxidziEANhUI|Eb2jEl|jR0~-6t+u^ Q4*XL+t92&rl}k<7#PPyMQ-20 zz`*Iiz&I3vdl>xXClh8U1_m~v-mP0Q2DgN7nHZXw%b99uY71){Yn$t7-VqkSz~J)o zl2_F|ElAc_~#~kZeSzf(iesdm^s=uVDZ4wUo z*446@jo8!~PmvMEeAemq>jPKS**rRD6GQQ2 zIMMW{+PAN-JX06Ez?kT{QP%$K$S~ca67~^`>GoGV9P!CaXR8G(uO$Ub(KQsftU6~2 zb{dqmyGVPdUvebtK9sW<$ZQ_a>p>hLKhRhVs>JWFk*jjzzsroi`doA;_dMLh?)1ANH6`=4aJjin`azXP+(Q43R&=a-3q-(YxMK#~t(N#mx&-FZ*YYE5z}YDJFym3aqQM zW|MuR=XF;LJ!)BT$!cw3#ac1!EpDGLHw6KuCq0`&=8QmvvtlMgG_iV?rL%WmM}=g& zUlnFVQ3U0nj315PqzYfzi6E~^=oeSJi4&I|klyxz5a~Tueh?LhPdTFdZX`zc-BMG} zM0W|(@&-osUHy%U@savV0#ebfFpn_neo~>@bJ1MgT-}VuB0^Zd(GJdTFqInh5r0T8P_`pN zKQ>LDrPepfnC{!?e&jPmSVCl_L^oy2+DQeodLW(;|79zIro%+5!~Ut;QV zs{3W3?8Ho?`RJ!=!ah&Eu#C*$u@|M41N{&2Q?D?S9~~@?WE{^O$va&iOPCp`JBX3y zaD}VHM08`FqTMb=+alsH*BZkOgi;naQO}1%b*sA7*PiI~7r-$-@Wt9L-mzsUd)SS| z)#Ot8MNyivHi>3@LD|-rz6;~m;{vtK)pdKvxkHWJ7&(!fhc@YL3BGDJR|A0H(!MA9 zNKz7m8T`h@!1U0=zy`lC!9Pmy4+8@$5^MV}I31CPwtpXrKwe}IeCLgU0mTr#ee&jD-kEyo-BADngs3$@5Vrb9QwZA(1=UGZ(;vCK=x_GEZ1lqNs+J4Ky z5g)_F?Vy5YV51oiaZcW`gOB{_AnY(U_S7+m($fP1+P)mx5W=Wvi&AlUOSqh5dU|n^(04{ z5S@6w{Q|R!iVFLu072Yn=CqL8-7(r4>qwR_cYQlm)x{y(iV(`8S2v z78NaZsrKsy=8iP4n@a9To%|%WHOnD!TDP%oU>t(UuVwN(o;7uje3wMJ&kkVMaImM0 z^xIz|pHVt-k0v`;RWn3Ytw?3CDdXedh1&=0E#n+;IlQ)ONDHVdEM`07ns!Ckx!YrE z=|>WA7Y)DVi57+fKG#M*EiBhgo=t<=%I5C?|F8b2{{poowUHC&WV0IcqZZak3^33@ z%_cI`ZD+88-YR!EZQJYqM(V_YW3ZJcGVTBKcK3CaLC3Ok)=_(8WFT-zwO8_TXxh#K z9WydpJ-$!c+r5Oc0jU?9PKrg-wh`!9d*346c31W_0#n1s?Qq~$kVMmVJLp)}*Gb_6 zn)_ob1!TlI2=foYXHmDEg*1nk?cS)BVb26Hg%9525JcU!5dm(;KSl3V7ErWM!GzeB zkxaYbI4xSAlwv!=zqEo-H5;}hZf@ZZGqXBTGS2y9gKgCh2c7Y^!NCNuc}fd> zZW{OjJzL={A_RW5r#K`})$;S;^$w@zrQX6h&JoU3{L*5wH2=^gaCF2r@6T~#d(UKEz{AFbV;HoDRnI@t`_=OCiF=M zBmA|_QpeZ)FY?l;4RM`JWa!kCj*d?6(o#-DRqoko0$4#o!6izslKOhtc;mT?*muZ+ zBoey|r`Zu?kK`O9mKNKi$i!rCm2Tm`)7nyLVyapJetPSN-ZK}3A&xKx>2Q@I%T^Eo zLPJAckt5N=qN1YGFfd>NfyNa9?+$D(bi#9O8dv|gKFe9$oo`2)l9Dn)%qWFK{t5&F z5oHu0(Zd%eRXRARogu_30l|T_Lt}FLdlth>Za1ch^ekHEj##|PWa-h-%$Il(<8KrtJjKF*&D@{V{C2mr!*O{~X^&^Pxdi zInc4VwQFcxq7cx!)hnWoDEsvU8V5QiwQMBycb}1eaXb!CRoiz3)tH`Oz@z}GhRwEz z?w8~Cn^w*Oc~+lWK*R0A0Nmdg%;w%6=Kew`L zJ3YEMzM&K8Rf&=-7#QSus0^*#;!%HAr9K? zN#cJ&4XhS>)`@aRCwU3AzhdKBGW499>TGR)ivgo1@N}nL`*}0`*v0^nK+uob52{Un zz560?Vb!tr5DZus{?Ws>p87yigxA2_qYQ|>U--imvH zggA%oj+rN#%Z-4%wP0IZM-2tU_W(IpkM28(=5jn>Z>e}gmVZg>_x))i(4In(ub0tW zmXU+@sF8jjJt&z!FMII%NwMSW7n*1;N5%r=J|EP88up(QpM}-GcX{=P(RSiVoac1d zTxyqHV%XG6$`hdcwcN7NsAv0$+voxPfov2u4>U7cE8>QHvrJI&V3TAi|H`A&iP@oS z&QpQ4I8)q2GL!qdB_j@o^0K;e>yMGVu-*lNqU1*_ek&92c^Pi(r`52#*jF4+M&L(W z--;?w9$2y$aiJ5X$@)DOX&u_NC7q87*%$XqTtu;$MpDw|8Z#OK?+NVq)G_gdeVy8g zhCV4zig`O*=$B?$>HB-0?h?+%2Frl3TiD@03CMacoZ}?3aqms z%&Q7@0EN(j04h{^hk5^KxBC&z0x^~TpE>9U{<6tp^w3BQSkXF_cWJAZk5L?(_Zt7B ztWr6ygQMwz0TCX-$^~Bb1AomU;#KL4TV3vb`K0}ZfqX237UY+=TB)&d8v%1Ys=-kT zmnn~<4OMS8ZaML{5bYs88lg>xtXPR5(*;<0%Cv??kKuZiVWu1>au_(taQ=FM63KM# zu`R?q*vT#Zt$_3vj(BW^Ce>rb135rIq|vRdo5hX?{ou1&N!?r(TJq-4b4repU{g0z zbyHvW(|N0qYPio}_T|3Nn^>Fk{3be`lM!}Kw5GN8PhafiP|!ReFYAKI-;no$$A$!q@&daug=OHCIu zzEM&HRlb!Kd*;D-7bqF*&6|+T{zT*PDA_COPO|pv?_LrfcA>GCZvnUr?MG&M6H`;y z3a2{ZRyS_k@JcngiN%fEv3@;%OjTK#ARF<?b;<#566EqA{s&;J-V9H=aYMu$m54fmCtJ)X4AM zKjl7#ss3GjVnMUn*a%A!-?O}5rrjz$=D9|Lcy9&P-{aKQ9l`(vUKWL5=+p zIR()YU2}50sLkXY$Q-1#h^w~EW}Ef@h8|cxj<22Q+p)~gAQ7EyzRL@PACFginPgw^ zbJTKdVWhAjam5*$+P@LT!*}xuv>K~v-A~V-H>qVT`Qp^(?cPk?!&upfm&W2&nn<|b04S8nZm6$Do&#Pj+R)5`T3xqEiF1UbGB2UN>JNOLO|Q)! z4bbTVFw)H+Kfl)}b{c~D0!l$Isho-$28ku1J=9*!C8)2;_yTSgtv8MaB;NypMAa(= zp@vW~Kt}@EjSA4v5q2ODb|nHd3A-p5lW%OYB5Htrb{=BtEo7j9`YNL+AaOr47SK?c zdLWR0O5O#Ae@Z@JMv#A=ydMDndGdaF;QoW_{V5FeA6y?GSpQ=34g=Zx@h>Ltmj~p( zB3{Mu(T{c*45MIAwsSr?m>_J;zLD!>U!lHCs?&;ix3|qIux=&rvX1H>y5)wq=F;7D zY!jinVY86+O%CRuk*zwp%f%}0LpFsqxW|+J6dXnF^yu5mHPb1tsZ!0jRPY?& z#sQA&_NPHU)9(-}X&m0^R3Lv+Yw0%P#eg-TSN}bqz0B`5;8m|5u!{f2$^jNrCo{QK zc!An6=W%r&uZ{}8_;Ke2F4xZFc;(yo*Ou5_yPZ#I>%H9V5}3l=7kOcPxC2WQ(^b4B zenYF5q9EkHf zpx*NiuuTvAT)w@}>_NC3u+kv!)ZW+cJPNAQ|IyN&sKd>=TFTPl>uX?+rl ztEzr%noUkjxK2(^-mRXng0Tb{-Bi*a8PI1-Ef7^L5)n3!QD;8%Fv36m(u^~LK*x^b$sH%)D!=Dbfr z@g!%RM783zt#Rwh7CV-af4tFIwH^iLv%;S|DKEo2dK0e_x5C5sl$*eH64Ro1x+V84 z$@8ZCU{wp4L_LuRYlmUl0#3?}BY4w;GG~xwOJU>csA)!uXj>gm1T#j;GvZhD_5})@PQfRQZ6!vOLJG#wg-48`PVFGaK_+w7MdmPS z>lpy0xLOCN#Jz>FUAKYV243_0sK6=I37AhdFmw2FLip|zd*4C=CKnm5;A*2bO=uS| z>-^z3d(?I>avM~sDj@!7CKv6%l8t|Z-h%=Fp8jciH*Ne=ZTM&2e|gtGpZ>cN^$*DQ zA;}hh_=nzmW7(E`{wIelq34b7j$&a`5-;g&YF}OZ=G<=hAZ99ntKhU{_UAAZ+<*tH zU*EvsQKa_?E`AxMu@u3kAjcxkiz3p$0NkD8@*~NB^Op*1IuKcN6Dwas4wBeE283K$ z4Wq{YpysR%pB!inD(0>IoDlpMKY(Dv@{}>R1q?OE3iJ!{tzmGPMmUYjnlCgljj*lN z&rJtzF5M4H;tJe-9wq_i&E679wf^ZRjE_J?(j+%T;z+Oh1|yswcf-L%WM1k*8sA%$1s$wgqBpi^l@Z>kBuSP-uExcLUjAl-}3{?Ef(=)H29uyg} z;;;Kl2~EZSpa}OxW!O_NCY)o+NxSOVi#^2BpyFTKjAl=Uj1Q>zM`{?LsrdH+f%La3 zy*;3UKkvr}6@LnL!{ogw-EO7?R(-DIQcx-X=(vi87&UUT8e&IHZ&R3W6ov(9%{rW5m$0XIe1Tv*u z7gfhwRs(9g#E;RI4BguZQYa_=szmfRfyWu>R5orQ@)P*i=mlg!s+$vVZA~KMv8nN#@&f1io;$2iNFsEt z99OnFn+tY?pkv)mhIL4%#6^_V=i`$Iyvd{RYWNMwsbGm7Cv$U>u0qFlEtBT6KB2`M z8yg<{glik?t1_KKA%&;Uz*X2rGpB;BGpT<|`B`!OjM>sqNU6t#`s#7&)2AOV3J-X{ zdL^(hBoN6C8_P_zv9>Px`c=?&eZ@*|ay$?kzY*Jw+|uC(77Ck`YF$~L2qMDuhY;W3 zWd35>N)7gP6iy^$)`FSE%t{p)YRAF-4w9{5$|y6`R}k6&mkgifn6{wzJ-&LVyv*#ohsCuy-~h$D{y>%#d%b*resAwj{eu~ zI(R5K-~$)FMT>0f64rrzk%9BXX3eG&QIm2_VD}moe-U9v9Cvo4_4YlOSjm8hDr;on zZPO}f0v3x4vBDKc+VV=a@EG;6iFwJL-b8}+lm#h4bGrntk=Qnav_#Q?BgLm8q3P

So zo=Q=uqbV_1W=k#C(uVq~j3*dNkgS3K{{C!N1|Nd;o0?-vA*ipu0Chx)tNLDOh?@im z;}WU{Xo%ZoFs6q>$+Ejj*xSQrU}>>;f`Jk0tBgW`oqvz#K*P>2gSdCuDtT99JE}we z_zHCW@>HhKuKKn|i_{kctSjb?P02)kHS#1F)8(jUM$~F71Iww)8#8|w#`h@+dH+8p zA26f;P{~48S1PNwEMX zY*m2#2%5`KUSN;BY&xiY3m-u2xy8L)}L=j%<>Km?HiX|Y;u z^>)?c;5#0?1DhCVU%Kq+?OxLY$}cYXAE5k$-A1bEKUKIVIrU^u4_G_1F)B4dj$Eot);Ca$sj@L9U-{=ND7jx;Mpag>hDy1&-VP zQi<;=l+=P~>JQ*NO?V5LjopQfU7j&oQcHvtg0-C$9>(-*%85@_1?{By-WIR-^_{(I zauGY1gp21xxb|T7h1}U}w%_9MNh$8_U7a?*WROujmaPouOPZOOxDXlGn|Y6|A9QOpzOBD!H`%kJ*c;McUWdT+zR_1h%v; zUob5Klb#F=43)}^KF|iNru4R?K)Kx9+*yt~_LPNB#3#MWEnp4{5oz*L16wJmwpLjYQTc30JKw4PeP4$P+EXy0*8|&QM>9hu$#nt+`tO;RYowYsa6r0 zDT*4Q%y5Lv0<-smJvaXMM-hn|qahkxg}RHMuftOVX*%xDF4@wDhL+jce&PsscL^Jf zbI;%1O9AnPhB{=tevx@$kH1ags&rPyZ4U8wn`ks7kMkyG53xDV0AkC(Hf|EE{_DlA z3=4^K=0bhh`1)_59s>mWKo7IlR6V%r`R+qiWJ{&Bw@x0L8((kABkrOeQp|p_?Ht&% zV7tQbjT>)v1*$9fK${qZxNfgADlL<{3_Ma2b`*XsyHb-4&AZb=0n%|Z=%C*DfB|y= z@dGitbZ*bkcRPtod`}9%Y^Y!QE@j?}lfa^YZXN6%`=W^F$QCaE5Ld6MyuY`}T?_}R z1yQtshYyW*u>(K^VH+N}fM#r_sZWa6)ujC=7;aHuIm3?D_TN)%4n#8;8j9u!1?(xHHyKQy9wzzsxWE>(m2 z_y$ZY1n^h*;EX65(fsdLyJx_SANy>sLcC{UEw;-U;oMx3x)G&gcVbg8(|I$s@R;*r zAX_2V*pBPM8iCdmByaDX*I_?%u*KvIz`rxB z!+Q+LC#<-Q%atLo_09Ub9uqY;(^Z-ycGi*lf7AuN4~*l}P_pu_`o7|MrbKiUroYu` zm+zU{@o*pSC;Y*|wIA-H$*J%F2PUO{^>e@RZM!3c2K08sND8d}87>}h3Q@Y6T7mtE^L~{&<&U9} z*vM0e9xu6kkr%SJw^tIV4qLD_H6=EJgWZ;irK6UV$+85bOMIA|MR3wtyX`^1Qm_-t z#i-e=VrXo-TMZY@4qe?4=?!~|P2M@>Sj#eUII_tNtEtvqP@RfE+WHbVCL5DqhX~oh zli&V83O68_7NNnxsOc>P&{GFZ^+-(k53dJ z7Uc;EXqF}Tf~0_fljQGB^jkIpeFao|IxhCnwp8|aPz2;N;s5&v-Lsi(!yw};KZyCn zL|Pg#KS4!TQH*<(r1dxs)~69X?6U$Q?{zP^N_gVCLT+p0DQi6^z`99f?ukizSoU1} zv)-Ji9)i#c$m)|z7Y^apluYpQb~cPE%XlY-Po*S|wsj@J!x|b)3J@FJli`XBh~(9C z9U&>=?0eP;?N!51Mo4e?H?;;dI?}z4iHC?CCpm^X_kpqFi{!+8Y{@`lAHh=4l7i9P z+b%npvF``O3yWk*Z-T=Jjcvw~`k0#JUaW_pF`%d?t`3wAk|C)5B61Y_KN~viqh!n3 z{Ue8kLY+6qiFLZKvwDipeY%~xO|7r6`Sb-EzJ*g^ZvO>2VhF=@TifDY zaZ69TezqeCuLSzHyBvdau0x!Pd*4Y?$4&%Yrgt?vbIhpZynNv!RorStqn|Msw^a~* z2U^1WrZcvFkIQEG7B896N}P@;eAb_1jfa?D2qFK2w;>mM-X(+>4~wA|BE$z*B!_`a zVV`2QBNI4BHkKgh&ZCA&6>gE98*>bD`)Equ$s7rb+VYaZ4Si#0*&ye>~ zE2xht;+BjR?0NsKAdcj*ug&?znd9d=3|@V}LNUZsmHDGX&4Kd+$cT*bbtLr6B^t;Mebu-YdVaNCM_Fdfb^>I8*rem2SMF|XT)`8`X{l9`L@8Hr=m~J#|9itO{bAF$=b?IT5<7(d4^|Om<2yNx)>+^C% zO2;?sYdI2L-TY(kur>nF>8^Xn_I6s=43<@wN85?*xNbB`2p#`zQ?OiMmo0Ah>G$2L z3R_$x?z(<8Y^Ere-8H<+GD|h-&;2XNrn^cHjj#{o638h)q! zbr4LpGnCk=O8bVe6Za|tfL^m!>>aZZ%!wonaDN-deuqd7yNyK%&sZ^@quF`027Jx^-+2r6rv2wxphpsexJwebQ9_V zAd%{S+01Sn8|*h zU8tm1?00&KwvufQIBlWCTF#*gmLZ z1Oys+8u**Je<&K4z!pQcj$w}GvsS>U`xLQ@6*8{?lyHmHpj&`Q4_X(;;NHpBwj<-87OIW|wFp;oqXr^q zq$_A058{Cu4W2-`fm;449fjBZ$4bAHRW_4*hMLn@C5X%HVdt2f>;lg9>(>+2<4N`@ zBa#`bAy$$oOwtc0UQ7hP@=0;fWh!r+uq)bPBXD>v)KUN<^A_sh;4t+RkMz825KCsX zT%?|H=FE}Aq$HNG7fNBIC;O;Hx6T*eG2r0akiy{rQxo5sI?v-c=O~2-%+heM`yv$% zDL1s&OZu6r7dbhOjE{?_MhAFw6)qj|i(P*GlWw08$w1Qdo0#zsSVT!bXo-(Jctyis z7=PKI`2!x+HABWW9TzAIZdte6(vAAg&NKE74$8Cx?q9j|entm3b+WhC2bsc0M-%Qx z;-?fq)lJsS^n+;<+KzlUyw?q)MIJOIzIoTb(29SjQD5>$E>uTnSO^c%l%yzr?NLGh zSI%hYnx(KrPO5VD#WH@L`}zzr)vl}GNVS_T64Imd$L}EL(8)5uL30M!0vSN+xUAeA zVeR_?Ne3SBc2Wt3hIH`)(lwrtd~xqEcG05Z18}!Tc?u8qNpio&4;X zP3$b_+zgxxtAkCu-c`%}GM7Yh4VVIw4##wnA2|I>B%yc(B#|iw;(FuiCX24GHtLMc zyYC|2eFLZx=?}b^iMD(nPCtPj#?A^kBS6{ERJ+ON8f3WHUCuC2__O0g_eKc6tBs-B z^hX!=BT-wYnqG>2;=1Z{!TJE9G{h@=-DITq9h5|Vw64@2-%0@PLI6L&43OLZHVoJ| zTnS2=lJ9z62@ZEN33+lBKf`2K$dS0Cdknxx^N5b@f#&ZLc%C}>jz5LYk=F2}NRBcR z*dnd4c@m?nR+OUAK+-MH7G)pv-`m*!$QNMFf7{a!498uW?;M0$s3XqGYWD{?6+!}E zuAE}Rc20)5IV-%4YG^Y3`U#^Z*|CE;l#*CaR=1b5ReoMUrf>;w6L0(s zI4ATLxS6_YY?qY%b`>SH2Yt}O0$w-2)x{4)@ABdX^N5h_xC+=5+sjJPgA#?+<9!dM z6i^%N-|_9pblBkED%)EJB!|<$>gsiGkH8Ej)U*} zjdwmqK!b*e)RBF0Y=_cBN`Zi8^;7{(vZVO^uJ;6C~40R*#Bg0pA4Ct{z=+d&GOAvowxH%_+g^@ zpA>=FL?2rY5+cYCtb&5nY^RyA%e|m3c$dTfdsXPW>ZD6ca>*ABe! zkDJMy00T_C#O#XN?UYoDZ^@y29Og+dk0_#kfLPgzeMaDGLVn*0j!WoA)tO?Ub>_NR ztOCbWQkD5Rz25f`aeEI)O*$s&YJe2>=_R6*A|UL;HOt^ebuCaeE~S=B4~&L zf^`L8SumA_iTjej(yjz&EM*esI#hAhGhyGk2nhKsVdLY!z!BC={7iGJZrHpClcup^{DEW1w_{EQl5>b#Wj{DzrGLQkzkb>yHtVhp0g8$v zbRDM|QK&YoN>J|B(j4n>bUKbu<#?>+LJN(oR|2T&NDFf6p;%K0o5!DDDCk;b*sMF= zrx+Y4J&|{(g!jjpd;MABIV~sefx!7UuIFE@@rHalQE&xp)qz05Wj?-F~#tgHm_S5c@g0{?Y-;@4b9_ zeL;0y)Fo|GHFXVhj~tuS+V%C-e>vw_HCUo(p>O_#XQQ+_ z_NeYhSBL8>qAaNNr;7z^O&1su(NI_*N_9&!6ePuXl#i!*g*svy#Tb$OgsNf(51Cxn z{N{-l78YiOmwSrkRQu)W`(y{s?*212ea5=xKD%F(KZjlz2fI3eBb!e07g`e?isu;6 zt&aG#^n=8{x#)9vZ01@zU7uV?!^zSWnB9k2aTR#&5&!PDPvOhg;xdWwpF6DnT(y%Hs&+iYe`8sDy?8TDFI16jvZ#7yGjfe&7znMGcg`jwv{SuiH~l!uZBo4`yNPD3)F?OkUx0?_v$_qU@`ct59gObs zkDKd@O>V}WvN!T?-jq(CZe!fsl+_W|3F5YR#9efyrWG5OwE@#6O-WjzH)u&Q8g*SY zA`svjqLSBLtf-!Qrro2pAK4Y_w&-2l+!qM+y!V5%nK&@-B~=w&8Z- zF&m^Rs%|(R4cB4J;@UX+H6q6BP|+c91I5pr{H&n`1!?M~qIh1n&4?7Lr6SA8;+9{q zENQXP%{Z3B8~k2!{Sv2bZ%YX`NSwX5A#L$k^0>Nk*{SETvGX>q22r!Cc?vrAMt50S zXR8<5;vXj|g^gy)CTp>C>yh_#GUqO7l(=3p>np2kXdn$>Q&+S6ei)`e{0#oO-(j#s@KwA_sXYr` z1yxgOw$6`W`ShybNBMG(`#;n2TR8tHk-r4{hW~-a_NnGt>76xw@&1Mvl+|&IM{#*J z*XA_7zBnOz4tS2n;zVm@)(NcSv&5tmZqBw=(zEK$=IDy9+Za$wp;8P3}_c$(H8j$=f6 z?t^nMiyCf*4%55@y8(oVjH%E)3atoisagVcf-8@LW7cnq#`Qrt0C7`L~U zhH}I5;DtB7hj6Bf^n38HY1o$E=#tH-MIK}J^Q=$Fefb7~B$V(6%vKD9k>tkqhX*s8 z-#`KjM&=5h3SZC|{r|^LXDkq1;9ybxJ@QCs&sPfdqvC-VB^N!ZI4@45l@cQD0Ek8R z46K~SnP+}+@jPbHAgc*ph)JQ{C!?(c;gxQIf~34gZ;6}Wm%{Z)6CzbrRl!w5plJvz zkh|U}(?mmk>2?-{a-tm(m0q*(? z$i~^#Kmb10+oY&|>?JuXerIQ=M)QaJ8X-nmc>UkJlI;4)+0;&}+4j0^dHX`MvPWyW zj>kvM&TsmJsmS^FqqKDut6zwJ;MMXXRPNP^+9N<$4w6`@uDE$m? z(Zp#kr1v>PIIfwgR4)nX)ZEYy0^!qUx1YhOE+9t|2N)f1W)s*{03*Y{pqzCfP{KQ^ zT3E@t!>rtN<|~CFYo%F}yrSE9O5Vi}pdcTuZ&3~ovM4fb(~EO^dsQcZrv(o=$~3*L zel@lg78Z8!j@1j(NJ{ycN;7^y_W2Q(S0O`({djV3wfctnGttf00iP{~-K*J(^7Y}e zLW$MLPX`TvbYhOA5J`QDg|N&rdid8H_}y>_$Om41t*13(UpCI0BnLt!-GFayuyxpU zJ5~~Oaus~n4pb$-24KyrDp9N4+i5;(I)=@W>9f|)XwhEjjREPBtajrNDnXh!!koF9 z8y0f71JZ4p-n`~)v6^{E5kO>AVM3imVuEHrI9gnnjC#+t->-5~7Id|@Q>7Fvwh{ZWE)TkrO5v8DwvcdYut9=Eyf2h)>PSrR#xG58 zH&SVy%Vn}tZZBOO-nk<;KiyT2e2s8zH_l@Xb(s1rDnqj|?O}m z=Y|6tkfv4%ocQoy(4l)(&DYSIZcx+LIX>ZG-yu2v#gE>DL*BI%G^EqilD;j|$X#Gf z$8{UmA4E5o^Utc8e!-!lZx_z}w)M@ZlLb&ZIG!0)DwO(sQQ2~<^!3r`4hECC)jmX4 zR#pcj6j6vMsGL&Ym<_DVMZ#mF5~zhWmsEi>GofNQ#)Zp61DKY|CIwcq4;@TJ@+Lls zR}@8h(be|*Ukb@BJ3hBe&xf7b=yI>W)MvIjn8Tptz2M8)kS$X~I*Jwl$%Ik$^)W@i zO}J*s*R8D*3&(W_cxk189aZ|OM1q4-31=FT=~$g-D4buPb{cpsEU=~tL&EU_NKR5c zJ|REy|sJ3L_+!tXKXrNuW4{PRgKZnELVuc-PoEbk|zV* z>eb<2y-!lbjnGHf zeHj6>A2A*n2KcG;5e0hF;e&8u`qH`&Vez{~NZ;g;Ad+CS7>j#!tUil0>>+=nw z{rz(>eIpAwJixO*>wW86r_Ew>T49nBKSergMu=G2CohmA@{4MvLawoXPBC?T{3*SPWkxf`rqfFC&X|VBnSc+!*<3vVd-XcRNq_7G zOG*U}JIKtYTWg)H=f5y*oh-ymg0!B&Ky|6M>)KWRbpTrmXS_^Y?dFEus7&%NBlP09 zur@MV)CdrrHtI{w*@Y)w-I&TTU7gHkXt@vgBD4S~aGXAw8mT~$$my?^^cL^IiAU#U zhvEW6t?$~=pI1lXS#%=^4em%V{=;X0VO;GX>a)0B!@s$v#SFVz!$n870kfHxW!~yp z71G__rA-By77h4{gKyNfIJ+IAx;79@-R*I7 z#c<7!0)}VLmw`;^RGgLvE3YOXvDvg3^N8zYc6Z^#y)dxz9afKfvX_sNw*}06K%%hL z*Vje0fPT#bODqUlstXE3=hW;+W2O_N9hu9b9h|r)Ga5=jYzyd~-`Xsjc&=__^~>7w zTxoXE`lJdw5+%4sqx}-e-aZ7whDP3j33ZX;lTMkn1?8N!YLx9|QC}7_u320Qoif9a z3Vk#iY>*<#cdV(Uh2mo2QlA$4;~29W5df;*fH`MH<{@OEDB}qfndBj^Rj-(n|JZv(f|YhDRNdN|J3tk2{>Jw8wiH>twolR_Q0%Z z%3k|ZsM%^f!wUyR0N34pq0 zZdT>QulQC*Wc0_XxftW7-_j6JTs$o}e-Y{S%UDCEY~59F0f6bAf{FT_r~mlP?q`4L zGj}Rp-%zf!Pmh;pwbFy}5Fy67Olzw}n-0eQnjavAJ%uChUzX%`uTMquZGpFpBa%q~ zU1m+sMQrj=)Ont|T5_X%xn{b(d&9Z{V@`B=?TavtBfDL!aK80;a&`^P?3T`jXzSt~ z1Po_;q`8c;JFV&Ip3AA!yuHfs7QWu96TxcdMT#_d|+a;HD~A=b=kMK48X`t^jOE8Gc@Cwlzj88bp5B3?0(Z8DZibKzbZ;pI!4H`!b?5%wGc*edc04h*((CH8}g zRg?TyU^gFq=RqADLINqQ$li)DjyRNN{LhHWE+XFjV#}WfrKKQE68~DqDT^@Ww=1bQlsha)yEpd?(*20P>sawp7`gQzk9tWnESMZY&M? zf%sc^03iKeAX6!bkHaO|$GYwBI{a(_%=Ig`#eZyNP6 z_N^)MND0_d@n`*RqK8d`9i&x;?;E!IErpet!{G10|4h#7YpMghfRn;#?e2tvve5}- zcJ_Wm(|3I=DV5MIqBxAxM6I6t8hA=Z99i&-2{^&Hb&knT7GfG#dyp=t4Q7BzU;-A)dl@}>n% zJC&MA6GSH`vrgpnztSjLUnBs6wRjeh?*-??k6=5& zf0-(RNc(p{_cTon%&?yFc_s%+=7Wq3vF z)lglQ`YwPAr>1DFy6C7?1t0L^Mv3CnQyfw9$n~0^Xo=Mu)y%%0=#7thg>Qzv+C%?} z`8g5C`eJs`8S8eV>@OZHZLjF=Y^7{-eY+J>y2_1f_K9`KV+}y)7Sx8R|#s_mjvHAGO>|qhRubh+O!aZV8AMgvgdbzN| zD@x_}E0nONjwCmAzXuVbSOiJp$`JH3>y>l0{1Fc7B*SmJomMA>^;0x~5yvM0KbN#S zf#w_E$!8yky;T_1Guyuu0;#$!2bTnyuR&Fk>7V}x4)XiL?;pCzd9;5|?~@5q>HjZF z?OSt(rQ7Z#Uqw(cs39QNC;G)t<BPr!@`+RS3tx`V9Gx4kl`gzy`0LXAbEbA><25D%af_Y%aU119qLb3G`)Og_YcQQ~ zyYr{b%8vKeZm!jq0{9O}w&O67PAXTocyIz~n_DIbru~?-l@jxC#J4MMGK|v|bl$Fj z`?6Tt*icLvzg$Ix_PSCfvx3vW)Kj$z)zD7=E zR}2|oJ(7dCbzbCdpLexl{O^-kxBIqp$AA4ab+D3=?BGu6Gb01)?Cfj{ z2u2AgM1_XXI05>ik(8W;53>D~7uCpm$#maKHg`pJM-ISi`3!-)H2@GDJ9FlYJd?8= z!E-Y#tca)^G@4M;ZZZ6Ho@rP0Ui0D?VL4?_Kaj#8#@E--~7wfR)p2hnjZ=SrPPZ+!CwKL5CY>drm z=4G!Y&Fa1Tm5aX?k>%4~V5&#Ka^RDf{ev3$r*isBbol1pe)60ASRAo2AJ>a4!Isi_ z^IUnj9K~&mq|HSuNkxfVt<>X>=j(+_K@blLG1g%>GG*QINZ2DXVBV)31Vgqx6kGp% zn&s$Em7g~h(jI(yYFaNyL2oq)D*DJ^DwP!{9YM`E4I?_&Zl*nrizsn1!w>cb#R3gL zmvCp=z#I;9hgFMe@3{J={WFVTii>71g8ZcUKAS4m+ZzgNwD*1CQeOo`n!IZv#l}t5 zrjz4>T0Ym>5N8n{NEnoj?>h8)DE8yd26Z4;bIX2?^4bJ-9z2k_t22IJ$r@kk_a;Iz zlV5;1Sj=UWb5BrQR0&!=RQyI!?NoEJvSQ;aU6Q!sk30TET;U7cF~DAKD;jRu#R)vX za?X6NZ(e@c$J1}tEmAFmhu~sZcB4=`S;N)uNjt3L+ygsOg5<9bLAr$#pdxak1F)*) zAMfFpWefUW-5{q=@a7&0nr=#U;t2eH@%plt7tQ*J@MrFyfuMeKdIMB)A`Ij{Y1UQ6 zmKC+AIk>IDi};zCq>ftvTJOOm7kh5?a4K^0tQSa@09ns6XWTHOGY+ZcrTw^j#t9~- zT$NTNsoRb@*WH6vU(CA@3euZf=i4=_`<~4MBhHS0eOa#nRQTRIE_-ot+V?)2=3-j~ zYTz{d#H(ggc3gOXqEa=D7S^c=|Q!e zGLP)S=@J6BybrsuE>Zv_IF=`avEx8qWX~GPLoI)b(uV+w$Kd*$l;QmFSMH7or@_V$OtZy zCdn!XvMMWNoZ;L6`E8zVw=n-oeOi-(@`*_*Jj z*IZAUcn`oFSzuSTbn0c?cz+1VkN*fg`tAcmvG(NFoxx)V3-o~OgfZ{Tt<7J1`jCo) zz5TVAlPx?n4U^4QL@L=l%_kn_n}U+tg2(7THy(wpXvAq=Tyt(?`er0v_WtP$I z`$Y19PTL%0R(FCFCG#c6mHEV`B2Xom2FU{X8wwit>uoUsc!;3eMZ@7>S3ANP@h_TKmGTQ0{>Jli|4+8y;*2d)OnAlWUN=*KG>=w{0ZuwLCnuO0v(r-U8PSu>-9ABxpL)cuFl`GSOgWvy%A`C|2n;(>@E3?o$jV0t83$*Vt!AZ` zLumE}sJtA_?y@)d1tw1tn+f-9eb*HNE1^O@DrRhzaFE)Jq{l&8*!!IRwMnyc%esO0 zq%s3Qw$1k#?$F&HpR;yI-UdWP;v#EMMe9J6cw{1r2vn*%M|RQj0&+WG0_X42I&$nj z`Y=%FxaN80#x*dezcf6iK@IRc z=K4xY5upc-lkF$f8f4jo8l1%^5X#T%^mSAfKhByd&hyLg2bIrPc&3XF3Fg%TSiA&E zt?wO6jmnOcq}==Xek9E!ncVExjHYt6{{NdV3jF=gyZwFxw^m_78nBRDvh({E_z`^xAxy$*J{r5~zP`HZ zdb{@qs1EOb2AtVB4KiZqnq|NVKs#Ws)zbC(v7J|egU{DN<)+u8`#GD>ewzMt-|>qt z_4j@Wy14J5**su@2I`#Wltdh=RTi>_7Yi7x)f#=pJNQ*}V;+U*>F@iI^l;s-_5QO= zG;4qZHF~SQw@K%5%>Q{N{Sa_)vg@08-Vv4`_y3%|-vaD>t}iJ6ab>w(=$6;L~+4{8WRggjTB9wYP(IHd9cIM$#7 znlZH$6MkC=>J5q~CG9Kt`GDPCqR-+{$Dvr)*He$&zj;&gQO3_hPd}a3=La^opScxp zK3@DtUhmP0Z#SOJD3Q;K+>~-MVE0>i`K2%+3N*f;x^?Y!q#og7kq*v>z{;@xM9`y- zcc$6bL{y(9KeF_@J1PFr6#c@GD$CP`$7O`q?f>`d$(7aX_gMiKZs_0Ci>U+-avuVA z4rJ!l|ND99?#W-b?|s{506ZY%!x`iAEen^WSw5NITvUAU__w|-Awu;z-Zc!bfeS*y zLQ-l9Uin$S6-k=?XvNoxd%zjfFq7xMfo%%OKC4$N3f}E}u6DidiL$)rCl9^|q;_Kv zu>1(VRX7VgFbfLx0NF;Mfo5lbHRXvzNlAap_y4{-;ds{G&u{bVtMjU698|vh`K#Uc zJHT)8$4SIN%@)Lz?OJ5 ze-?P$0+hKAtYCR`;tg=}(A2!BdPP&)Y`)YL#jE$`e?Du@zyH_O^-exM`=8ss=LAk; z9S#!|zFEjM^ZvhY+jp)^WjlY~?oR8}WA(t%;My<; z$LV=2ZNSq$dElb3GI!++OaDffPHu- zo#pUK4wQlsY8oE`=RU5v@9R3RbNQUhb>4SDS{lmitovE% z=;+v0RW9n#(J^+@(QSq>ZvjTW32us`qhr`+qokx|r=+apWbfpn>wMF~QrXhc(#6I? zNBIIB-MQz_^-Zh~>hG41EzK7?+8ppbB_(VdMLeRwUgzt>uAdake1~q~aQlMup`XWk z;%$zf8JK@-BPhT7{EiRr`ojCDMj!YO+h$;&io7Y%dczPPFxxVx{QBEomeJBP`Q_my zcF@Ld;!qsUAJgkP-pdwRB5V}ddyHOqf3^s^^+$gFP;ITW+XufEcE1)8q2QM%ULN0c zV4xb-g~jpK^&)192~?}mY%S$Y|MWcS$UN6@+GWK!2 zKR9X%=?u6Mwol~2V_uy8$Lx}@K>9OEE%F~G4Bw0)1C#3nGFVppU6$aH#!H_iGG1x$ zwL$|(prT7XcD2X=&5(RCzulHa~sp>F)*eBmjxI97Lmow%(-+qvob zew+}=^(OFf=c?PB_^0u?fwe?E{-4>Bp*?r#{l-PCY@`t_QptCooEIK{I)LAKEmgkO zAT1&40@+NYcz4|~xyPn)Tirg1xzA6}-7|=fVg3;&wU=%Gvty#y3V&Vr1TrY(f4jr1 z$PE_Q5aZ=xl=8fUG~SFzRK%R0loh2@Sj-}h6h zqN-s5vj>la4~p$OoWvW(|5^9;v)f;#uWS>qk38afd$RQFcdt!+IuOaH6;FS9oUfx; z-Hz@nfPcIKk=)S%?||DiyhJ#fO8Yq)JT1R`4N~6jI_!4TYQL?B+o!JY@h)&`wHXGH zYle|D+)82c-wYZ>UMm(KJ==7R)vwISs8)GWPUE(GvWUL@&HSV{H^i^a(epIY`IUN> zGh1yjnja*5RtU5EObFvS4jSkkl<#kLg}6ZOqV9h(={xCS=3;i2#}%S%``rt7)^0o} z={@U^O_X*w;X@~JB9JMze(-1{MRI^~}KRk^MlaGmQfds)U? zzGl8><##JoI|uATJ+0Ip`}N&rIInTrk7qXV1%~&mfzh&` zEpt&EO!81B-U1Z7Yi7Dlfi`puz||(;Z$I!yM@OGXzcj+wow#}FdNYLfv?V(>ydkQhoHY@f+};y&If_gs2B~^Dm{OBH<&(xnI-55Bjd5Nf-mkf zAG5!BNr~s!`Lk~m)gE5F%9^OD1ZLoTb2un;z(j^e<3~sNOmkyTW`6IeKtuC%DbA-^ z4)#K#(+*BXx%$IJ!mO;U>L$+Lr`tr&z`TQ3apPix3I8RU5Xm+ci7IP|k4n{rc=rt= zQcx(bzAh)0zpx)1xMQ|yN{Ynf51rjXnBDys8lBz1;Yph2bbt9dcM~|DeaxJ3;pYig zdIN=G>fpHB{1BY{^Z~W6jyN0m3*8gTs4RJUtvy&DOLTVPEVcy5Yf}v#i)G)AhT@A;-Ix`QMB5maJJBO?jHZT_6io;^vH)R zIbeNzPRzew4c&P5H`;CRYPG@rr|JecKR5i~MVckWa(MR7QJiKhWOeSPk!O3}3&}t~ zthEk0@B_e_9=zTc1Kc0DqXSnrotEQ|-vP&O{WmOB?P1tGZlv;c8=wu>wfo)=+jVPd z8u(USQHBxQxiGC!KS^uslxE=vuPFBW3)sL9E8bVs25*;>s~dR)$tj9|VveR4vB3 zcEzR1zjJDpAT}0W#>fp;uk{EU%C%%Ezm0FhoK}Mi1M~k=0BPp)EW~|__ zv)st$3ladTuBBnetR;ZvX2%CnK*(ecu4+T=2c0EzzR5XR(kAjJq+Vk5y&Ci8!@k&m z&GK8)E-bUG1Y@u0$8_N5EOGI{zrJGZPEKbT*fu-Zu0d_!{;Kx?y4HsgjOiH+naA!~ z$j+%e)fZ9yH+|NkeMSuhAV#{sAo#=k)p8fU?)#He+DMD3Vt+4%lkSgQFVedc*gEqi ztG~?b1Na7#0p0pLWxmKP(9O3y{x7K=mT~tu)@&2q`ef}moah{)<+f#gzf!PQtEBk< zc*A{}P#_VetsAP<*BnoHEG>1wtYn5rj>VC~Z^4iupvqs#J&3334glj<1D5m6O^-;A zK6fwjs(tS{6g=VAaL-s?`$r;|sOB!AZo7|n4A(q!oj1|h?EcW6g+XT(Fn{*5yE?gd zK?7N*xaY5sm0o%6S-Zt^kMCELxkeHerF@dtdAjpdvxca%QkW(g7Hcs-2zGiNep}dL ztwg|W4D&7D+pG_%{aZN}#TXALQvA0fp|xu*{>H$Zc^n|}dHaVon_X6j7F9q)1Rcsb zmTq6Obd@cjA^*QEY?To|W)6tdmkw+tUkn!yTQx=glc4vt@srN+rscf_I zN@{rCQ*2#^WMv7-vy;famPt&zv`ToMbuDI%ilSh0`RMpPi@9I*3S@l_a~KbWp!0cf ztNJPp2iLI8Duf0~5-u*-SId>YAR!A`DRTPEulB4{mDD6#-Pou#=oBoN=&X=ZX|+H) zcnf`qY5I*4O%$tge{UbT;K9_+@TqOq7i_7$e$u>x3J8w7l2i4 zJo_68H+Z#5XE)Jr=*B8A+t7^--Ppj`MZd9O3mdkuVGA3!@H=c;9k_193>z`Se?zVT z&M%5JiuK2m%ZwUzgN|knIUUm=f{hs0`dq|5XZGB%0`q@~@LPx+t$XR$r`} zWVXH;NTCjM-76Q=*{7JxT40J z%VYgr2?m8Bs!Z9k4J{tk)GydUT zg?nRyk#bZ}C3H;QV@--!b$XLoe`V7bljUS^?S(l|ZGqe^SvRW$INb=8Y8tz{FRgKf zD>K{y7`88vZ578?SEzE;fHZ$&=@oX-n6ts~b+lze&DJW(g1K!N*D5c&g!&th@+TB; zK*|QB{5}TR0Md1V)GA+SFL1Pap>M=&%&c4PbmcYQn5ZXkC*MC52TViVJ7A;Yz$!<{ zleoHBEH|<%Y<2RpxPam|AkqsysCM359K2oOOr2h36v=a%34K{YY<1awAskY41@4{P zTQ!H=>p2v3N`sYB@EBprV*x|?%P)ndtKz1P;QQ5nGB>&&E@7S^*vJV5v@K`*`BOs( zIVUJX&wq6uFzf8eJXaR8s@ccW6F3p9F(Rrh?Fzh^wY$_6ce53KocL*~TtU7ssfMQz zCQI>_*6Zg!IjXpH<0kr2AkZtti?J^eUqxbQ)~NUVfhSRJf*U7?jgrU{&*=m9zgS_# zAH8DjJ3I%!aT!321_s^XzE*Rreve3Hbs8%XcBpz{Ux-fAWoz!0h1bYS>~VFE(VwWl zI6@DQxaptZzlEx-9=H|SmbS4t!n{MPgo#)B!C!tC=!R}yL&gdPyo!qb-1mPn|ApQ} zXTr8M2aqc(faw6C)4kzaCD_%J->>eGSy^T+rf+O?ky9HIW!|x&x{J24VWYf?8(_2u z>l>(hpYDI0d+8JI+p4_ES5C--$rLci`81`kYnuFpiK3c@(#V)6d=bS`^1rt{KDO)~ zof&|zW@n4sauf#C=`a`^w&cJ7u5OA)Um>NM%+@-_N=V=IYPrv?&m4&30M5_^ zb zAAk9}_~EGSE4s2`aNGES-mADme$L&V^LQ>EqBHaZXsV72MRZ)ZWlb?L?>^5fLn*oZ zzx0{y2dwgyz}5@P0J7%dCD32`vh|G2dXzITzul=-k{VRayF%UStU>Dr(BQVQ@UoIC zZm@dUH!fUkDA%%XEL{8#H79WX*>J9&}KDKu9!vQJK(KOYv;P-D?_XB7;FsqmW_p%1@RxxYJFu- zi=}S?{98{KcJZGpZ>!R}Q_EFVnp)ptXolk;uK4+Bay`zBbN&lfvR=}RO6no71v}pU zZAgkjKsb?6B64+EYZF7NupZ;sGURiiREv3K^cNneti0S8YFmR$H0?cU8XH>9R~1JL zhvpqxRTN%TbToIT^#Vx0XIjNVXEpG!PY7_1aV(+gVk$ce?z&Jbz|XT!DUNdrFc^+M+Z2U@2@qm? zJg0~DdQc4=9XBuKk|zZcongGunk)STz)R_`8OAG&N+4r(5OsJ3DPyqSU!n38d$zM^ zOv3BMC>Z<{D8-AuLA$OFik%t2t;4b4gx&Lx6FsO^ec+sL3J>CzEAHEbfZ||gaq3D& z&8yf~O%GR)snZQHn=ML0O+a~qe|2pSmVdly2|Jl~3CxOfFy;T^UA?hYvGzm@kah;9 z;Dgo=V-@@Mvdu=GhIQ1zh;;O#^Joe}p$e-Eb62R#FHh3Ex%ig&T3}*0!6MKtl&o*v zPX0Wbla_muBq=V7PP`g5C)(pwpswXEfs^Fg372`A@Kd_WZH>_-)?k2&%)4Ktu=Mnv{V z257i$T_hw|8`$&ZfWkE|^9PGKW9BIg@Zg@=@8paJuufcEIAguU?+<%9n*)!4Kgb2I z4itM9T7X7=y)pvf0>v>B;)~Ne9|Z2b+by?JfC3(JU>M&s&f}lbTZU;h_$kyqioSDe z!3><$fRA%cN6#g!epPxE!(st8I32(8^xy1RWOdakpm1LQo$!|;P{%Vs{my6`{zY@v zykEU)&EIq10dmgRpl?UkN_8nM1qgpcIyEh~e5APa!=f5W@Bvz>7P03qnqvVJuR_vr zo$Jq?3%n;ly7Q*sFPd{u7tl&IKlop&S6={)m)wi>X_5DXKUQlWJcKH5-&08C-cYk2C>SNv~ z&TPxT`S=CebK>2Oy%WsztlAhl<$bm6`&e!j9cNHVWC>;76c(z;%O3w$2CDakaBl=f zBq$WlXN-^$1TQH`Cm4YuU!2jKVFvXEtwzbMI*QePp>th-+41|Gq0T(9uXq22Ut4v| z(Yyx`|JpGZBRK8z$fp|7IQcuf=cJ8z5TKkcKhm*X9`GW##Nz6S@|#!)bHZ&4AJbZv zmBmppad(iIQ2GdJyO9rd?vqGX7Uig4rSsriR?Wg)=CuqJI(KTfsClf1ia?kDIY!xj z`M7~M2FsJuF2pX<$A)C2xAD#Hnkv4h*Va@1m|4eQL5@|?Y|r|z)zKmUOstWs`A>#x z==jqa2J9xqd&XJ@Yaaxj`>l!Zw5xuIKRR0742H~(Hm;EVl_+JvSD0*@`P>! zQr6)9A{1{x%6~bem|MCk?UhKAvf>Cbcg`JD8gf@t>t1U(lpBRlZAN$^S3t7h& zfLSjqRqYy2I;G*gd=j%5cSZ{5-Zas7kiGReb^BOte0lmvGp+|QpYk*XxBt5Gh_z># zHxcU;E}=} zRF}ekLMr-|x;~J)fN$=e_~Or{dQZiD%(24rV%|MwJ4VlN# zbBs4KxJj?&A#aNhxTbxk^0{e&*bR9(&1&f0f)U`w+$t^B*_Q9G5AKOy=hIWPBusl3 z5d0PHY&Bayj^VljgmC|jkAq(-bqf7S`!jG=CHc&GH3L!k{x`#um&M->r_YUGof15; z?`F(k0Haz&0mA)N?E?R$0M>C;Q{Vj#YpUsCZl^_P$7ie8AL~|jpPGuL16Q)Bln|a9 zDy@ksfr70rV|!LwdRxY2(K$Jv3|j;=3E?CM1k+WC3`e_}j^&vAi+72q=mUAK=_@;M zL7hFuBAj`k+N~D!3|j>IlysTWavc73`|r4r>i|IU`+Zly>lh{A@T+)%pm5{ql_2^z zcx-TO5xUj_$p2$)0l&1^^_h#&Wj;p@q9AnxdM85ZRYLZ0JS6B}=Eijs+~YGoI$!HH z{d3pJSCsF}fja8AEK@fUkGij8ghh)7J;P%&07IGUt#9Q?pX`ygU@mWWk2-!7SzhGA zyB6fQu7T5rV-nQ$M)S{pD}7>XS$Ok${Pa^g@Gn{h`{TXw`#0$l+^1>}#iiPz%3KrT z@BKUd;yadmTIxv>THvy6lUpPm_ye;_$@V&E#x;XMc9JqrZ`*4nkI;vt7WSZRYaaz4 zWBa5n9McR`7tg9y1*S;YS3O#3A6TQ5P=FZR^9;lFbJxV10DiL{aPD8Hga}xo(HuK~=9-GHx&UmJjipm_4TBbP^)Py` zbaJ2Dn6|!t*m#mekDHx7R*;1dG&1$jb?)6XHIyR%x~yc*^H{$>^OKi%X^f}p<2Yw;q95<92HBtSJ|@hb zI^rOH?cE7_rITV0oU-@3X!rc37J8Xgr9{%DgYhr5Hg`6a5`;*){5nelJGOocNP#cV zq~+gISlp3TF(BA1BR#+Mem=M(`c);*kB)R{yN@{uEzL zk_ZqbIo-vsRoaEAt^cX2^dm5px#Z2wlI@a;3vRucYnz(;@pZJ6o1hESIOfu_MAoO7 z%V?8kNeH%u_1V9yn#T7DW_vGZef8kJ0hn2NR{A~MK5k{XeU5++>*wZyuE#?2`}D16 zaIfLMHNSGAd(T#73ug?Qkkb8RP!hkemij*i83Cc~J?ir3V!((xXUCaI?zzvLZ3k{? zHK=He8eEec%S=}exs6eN*UC4me30*LM#g&F18jo}*oWkucZ};b;{mE7!L6;Y?09{o zGeE`(Bd70PAN-*ZRSgtJ5--9T3v)JF|A4GLCq{&Tjg4& z#6p<2U%oCt=(vAB1XCrMadl zBva`Y(f-ZPNz&Z&T(i;=TN*jfO^7}6d4oQEa@5zXzVSx}dQh8Zvsr!N{F&?5j`4vs zBS!BZ{Fp~!5he}Zos~9~c)+@6cc`;js5Ko^w~?Qlt^0JvrUGNL@Lg@4$$7(Rwmu2= z6&aOJNP4Yq)48^JU3%D@@lU>kl!jX%{VbzX+VS4zTT?PJCfPc)=iBNnCNVn$`Lzjc zQ`92)=e`bK29NlvB_wo&aotmNL z_nwHI#9?HJ9&Sn^FK6C|%^xrf2Pvh#Jl;yRovBZ%?9vZ+t-XV5{VA)d}ZW*sA8 zEva|x?2h}H#!w*ICz@bMc9XcDyIpZ56jT`0UspSUrAR>M+wtzs7b>@DG;H22ql?5unOT7BqI)3$&bG&Y3_c3fWt2DpQot zD4&YQDTTK^o}y-y9uF}N{?QlJ>W3Hz$k^k}7YgE25_hlENjaK!d7LTyHAmaN^=(Oj zo+^%(fxhJ`;_sfMt8Kz%;e4hE13vpI%KqYvPtj+19~#b;>+gL+2otWzExfhxk=FtK zT+7Bt-{|4n?g$eP{o-El(OE`;BeYeoYne|DeQA5lejq)qqGq`6fRpir=R(h4)hx?F z_wU+mzv_8Fytz4kSi83Ip7MH8m7P0z#nkBe-<~NZcOjOSi}#XR=olgh;xgE0~ZAiK>NQsNIJEjt;&n&Nog% zUTVotegIGhY|1h=tp72~J5aozQsE`Lq99`4DR%wqK+tp1 zJn!&b*6?+)8F(v34=8u1R-E`tJv#*mvm`Cwe*9Oqy%e&_aRL#OU{1qdgd_+A5$gMF z?*1iW(geJQPeaCE3h{r9eir?1i!R`>67cCuAHG___=S+=)e*q2M!eX=zj*h*eg(E| zZmDT!uvpKCHtN&Y)zw8yPoYbu#8Pd(XuSMro+2jUB+8`?%1bImzmKL}Bl| zt<)`CFVtTmI;ufqeSHQL;V=ANKSlhHx|vHK9Hfmc@n32%SQ0j1L)Lpc@+=!)4Ft=K zAMU(2Ut3$-C6rzA!vaeZ0q#yj&=CTLyRGerYmC`lB07qCsR*liie^iSJPcdhMWTfY z_~m~E!5gOg2(Iz7zS%7B*wJ%H+9HYa;G1Q8;{0YhW3Oc!6Lm}X$nqf`@bfp|?++&n zf{e^+Z2<#O^q^00#x;M|rfs02suYvp^KEc&kVJp?UW2&mJ($R-g5_DVZXGccGb^?V z6Got=@A`%bQ(rtJ4No3&S~yr2)-;dKW(4ryT+occu{TkJ(zz!Q;4pKLf&b;wsRp5( z?!pR{cpxeTCj$ix(Jx!0u5E9-yAdBF|4*4oHT@+^l zD9zO(wB~?5x%b~dpJO3BDi{N-z6#Iqc-SzMe@yfff&A2!{W^r`!ORo_ZH^z9tYl(t4y)o{<`7M2E)%^QZgKaF=6+94RRpJ%=yW##bkvNtK;#za3) z@bmaheRVQ*+?|@>J#sT`R4Fa>D3Yo<;(w20gny7_t_EM*aHDf3@|&rVjGaN2;WXxk z+1H{k3*Yws&MWrL?!yHW%^w)c(rX7*ibUNq$pB+`b7jFdOZUXxt=p$OL;`t*k+G#) zB5sqC%RO)u|Iah%=Q;T-JKkvmfF^%c98e&0VMYn3-uYJyGw(Qz9j&Q@uk5mkb@t#P z_mhz3nJ`|_Rz|N8vJFAxB$ldt@NTVLYcJtgss9JdD3h7iQZS`n7V7V#I)gF!WnhK8 zR9Y%8icTon)7f6&>VAC4w5`XSg z&JIuQY)Nd+-dNmO62JyU^C0>fXAzY=a@H{AjM3pi5;dd2tg@*63t;79!hjd4In6N# zd=L?K5wl3VBqFb;f^14v_2zP9Y8nZOIhgwqPLJyNp_s*bi!KT~|L^~d(%^Lyf2 zz0@dYr=~r6y)b6?93ggcSl{;!C=4=kL0PI!Gi>n^iOxb28`CT%5qh8(ez+@d&;j{} zEDq32>${K`1LDCT$)G}nFpn+_&l=V=VjBWT5ruGKhYn2-i)otv1e|+dqAO5B=690E6HdS-7}7f_vls1?@b7G!qRxC zh+@1p1HHsfQ~oLy=0TR~U0pJ-V!M$@1m;F%z8|XY`2k5QvJe(IVyb~kXh}cAGv;k$ z*SX-OBl*SD0F%%?sd5(HOd#@`di1z^h>D(G_|U&%F?~Sgp10(^1qAW!NaL`Fq#9R~ z60t_BAuF(MK1jVC*DzmZPpWFNI`2N|-7u5B;Boc<`5~hE9S6X^XAnh4A4$#&|Fnv*vUw=tLHDS_F+ZH| zB{ZTn^|{v*vbG>q1g$DH^tyAV{aRe{k_qXx#oU)O0hsLZ{A`NYwVKkE3 zvy2FJ##gytnx#h)2;(rAN92OS@X!ZT?0Jg~EU1qj{SX*#4mw|S2wxUGf1#+N9Uq0i zO>%1{Cc9`1@uTwmoRLg z?{)3=Mu^AE?JfW-8g+lw=w_Erw;PF$sq&ow@_EGo#w~s#DnAATzXckkR0zZLplKhJ z4!Z%fZB?UjXEX0Dl*WWY}*7n}Cair4<0u_|(&QZ0L#PyUDl%2|kQ+f*Z zGHJ>NL_%C0g}k_4IcXFI6^p0P!l1AjjT^L19dIS<)-l6Sfe@#VvmmUSAITeopTQy< zYvgeb4sh|?)33zs{1X>}#h+PiCTXVwV1xIoezFXH4$UbhNaxr~W%~Igpp`hrrEtMs z!6qZwFNuk72`|H4TgjP#Tkom({v>XpfzFj}>#X6#%#{6HcKP_+qf#I1g-Y9$A&naa zbrZ7{^!~ zhP~^#SG!ZW0O#jMdSXcvjUgZ!MD++QIQfTl3|Wk`D~*xwAa!?h`#4)BJnG)f(_1Fk zj0YY6dbd$FP}JLiLl73~Ayp~tuAq>^HQVt<$eEr7sBR*F-#MQxm5TB4BUY7Dd&*MW zI)|0f6(9$hqWQvhCFG(PuR6jA#F+v;O>)$zuJTxzi=T7*cH-6U8L}m@Y&5%E237)< zdCEk&An7g~{T7#5Lp#EYevgsa@x-ajiRu5rzn%xtTY=kHwE9U*wci(DYFpE-lIVdp00%*y9sRt z4s*ucu?0;xValaGoC5Ui-BsY+(_J;4mYodni^Ydh&Z66n6*%r1mm2R2`G9C zhhwc$#tF&IX22Ehq~GaqYV#YIHr0y=E|8zH>sVw7FYKB;m2Gxh$*Fyzal6=TS@ z-yY%^+txutEpMMr%c!21Ku5Gye-v>XTO(>~y$6b^EL`%Z6Ur%`|g<1@NQsx%24IOmH0ab!o6k(@ghm zNL=g~sEK_Bcf6Al>Da9PfRZdV122JxnJr;(q2Kp<1nnk=L%!;fD9L}HaLMhhPnyw6eLvVL8?H@ zD+xr3(ADT{=|V^;jOEs}4|WXbV9w;@)QOiqv^QI1(sH?~Z?yR1x?xEC!gq_^*?@Q; zvQN`z<_o^oIW1ffx&W|-!9`YfKle0>L^Jy1_h@zhq$TJ$5Og+FkS^8<8wy7m^fDruX<7lrsFge9}bxx zEVhzqrgbw;0WFl!sIn2-sc$s%ph=mNhCmKneN@lU85K@Qf|YO@a!dqbY^EU%l~41r zy|OnbZw^v_wNt!?Qj2g`p-c!(x#{mUz-dF^n5bma)tKJ*5#e!4G!eS8O_2=6JT%8> z&930{wD>BN$<`DC5%-sSk86Z;JE?vna}rAj-|l-#`=6?iLH0am-uS%&2@cP&?v)bY z$8zwt*n@-a04PTDD%YPWI$A?s9q2$L%L-z8V(`fktu9LC(vt*?rUdAW<_q6rjcgic zaHidnn-PQaa;0AWI$kcxanXM0%{c!z!lL)Cn8-}d#e9j+Nxfk&7rYSlb~4TZSu}K- z$?=P*A0!;+JU7vlpvk|_t!{Hj-XmZi3?sX_1`Bx=gyLTKP*7K86`~P{2yFK*&zP^MuLiF497{|o?Kp~1YFYKpx2`>R-Ny1$$^=Y8{w+N zuQAyQK2sN3fL@JN53ER_1q87-{aV9RW7fF&1%VJ%Pg_ny4H^34au00!Sg-GD7d@0QmoWE)d` z&XWbb#!0gsn&ANyh=l>^EuWOV^zz|kiv@T17hnUIzg5Lx3nP$X>V~%)~4!Zqz$* zKWhN`I_Ojg;h!hV??(VrbBu3JLM(snwJMZFR+j8#*6`%;sDWsCYi+d8aGK-tRF8o; zLUNwx{=_mV+n6J2C+l^B9YU3JZuN52@EsTw&|A76_?{7?ul2}zR<>oPvU8q0a~E{Z zQdfC?d$TuY)Y9=&5gwfWG^Qair&1WLD#)K|uR`>0P)MzCxFeN+Nu7V_Cni}|BzbD` zp9s+u<5(B>22`?Ia+7%5zDbMLPfl-TPbMGzW%PMQ)QYH}BOi}%j&MqeUg7={&uAm& zYK$71#vZ9^9yhLpRk+T5wYOHFnptfV0IAQtiom&#IVsG@bP@(598<$GL020pBXrn1h)IojOFE#C?~95qQffezGdPtCj{sF@jc(Kn;{wCLwqGqvHkk$=1#PC#Xo+e}q+;SywD ziK5_A!A7GyGM|3G8L*_bOZv+t3dW!H7Lqa^;=!J-Xnsh7p)0m0k7;^&Gz91x>Uxbp zD@;?S2R$;ND4RoK^%WIHl@+H`4pM_7f+9Lo5O*wqwNc?NUG_}P_C~f%SXi6@>|%6= z!m(cbZcv6M4q}5?=M0->J9XF1(>`OGANr`C1j(S8Fz-*~nAseA-8etz6)aNTP~O{k zmztmx)-kD$%*%(4!*BadALJK4Hsqd`QE?%q?J^O4zEc9>cH>?_u*hd3F*rVB+d@(K zH+LUaU635*+xbvq*${U81S_g?h+3Y=aq+s+P>+!!h6VefqNWFAiVCmsgg;`tbFDYl zOVt@JNp`Pq*3GB59zqToVvi5uZDVKNdU=)kRT}*8EzWvhlV|*6p}@Te0bEys^3MB% z*A~u?7KXm_wy5L1itacz*TKIOB{?8dk&#AEV-493O`^WcHALY1--3|uor>D<i|{f?AHS5LPq~UpbODKM0%>)!)_oxICm4s95|NBi(hv zy{X;QtJueGE>(E`o}DGjBwqUSqh6}K5zijvcQ2{_y}HesMEfhmM7@P`wFNGKFVjx6 zZbdULpJ?SWfThFWSFLA4b-NVIrzXZq=R={>Il|&Y`oj%ux~4C55>h}04KadflNlZr zDXA4g6wSwIwE6$TxKLuJ@j3S*IQcjvs-rtIwUECX3AqQE?%OYMQeueHW-?hYN7c)@ z0rfV59|;S)MLR%PMDP|`6}XPQ^!_pkURE*pGcs!od>*;|8u=cw==@3fBhRB4NH&5b zU5ZtMjLu@S?7{^5g|vDHtXF1J>;uz3_3I*m#f3qQ&kyNO%?T15Ehylr(UFKSd(qab zI!BS)wPu&!{8J1EHW4d3cZ(J*E)W#0nTQ;m^mP~ZK(Qa=_f^U`>^tOL3b-w{JsV{nT zEu?7^CBnnC7msoEdJRs`1+&CcdZd>DH3mqujTw`Oi5@iut`F+QlNCvmqPFB_ae4Sn8f0H4;e#O0w#-rS^tIOThp@WQf63uRRok2CJvh#Q(|7AMf`hhAni%_-~p%UKp3 z^&6cS$k*!4p%g@zsdm2nCM6;<^z9yIiZA{^S36HEI?G|y^1W=b(<(tO=?5Se@Tons zqCr^|H5E`F!Gy^vInU) z5KMjXsSih=_*8G#FxwhjJQ=W-iV>$y8842S*@v4C`Ndm7{9% zf|gN39MS_$V+;olMtR*H}2s0aoTP^z6VI$__jy2O^28P^6EuRNd1 z;exR-SqyeE&iUF*A;-^nOkukRTnEFH=TMp27sWBj1OKKmT+<`Sqej$fK z_6HnKJ&?KMi(>H5rV#a2Ah{AS(egN!iqpRe_ydtqSKrK9A%BPjjZNxac~eR*$ukE2 zy`vVuvxMjUvJi~NL<~-zv68uY2YKc+)$t0(D?cj=Q6FK@+%&@)Z7+Y@P^}4F8}^oZ ztHB;yROA4soY85}3x@G07v_kgVjV=){z?JsyycHqS8R?wZp60<8tEFgKDr^0bJgb1 z8N@uB)R6WWe!1h5k2ul5|UXFvQ8=zq9SDrA$!a0R5Z*GvI{AD?>$5IUgsDY8OL_a zbAH!RR`GrQ_`W``=leXr{_uL;?sMPQ^}gPFUDy3a_O>_?{waJcEG#04o7eARVIAth z!a5j+dl=klVmT0ug@r?`ckP<2{x$Jy#sn^Ji~j6pwG3!QIA}U85-NRKp%NAA_{3UR4vlA$m-g2G=6>ScH#V zB`yBa=hKH)DrP?YBzDr3{ZpFkC!81OmKs*XKQ>X~jb~p<&GpIB&#&oO9uA&z9D$i5 zU<6)SY)XDGMr^iI3G7471F4n6Wo1_^iX9q=92(eJJ>H*x&vbxlu%xhSa*DnJ*1njD zMC*(v$ci6$-Rs;D$5(x3KpP8nygWta8P3(!Z+PvGmhj3141(zg$V$_(6i$SaOsCX; zcz5x&+6}gIF>a``?$<{~7#0pk zsKwV9-o_%<=#~#Ur0^Fumc@eA6^QF==@-<3@hu(i1``gG&|>usoM%t6MKaJPN$Z_| zUE{8xd%1G<^-|U_#bay-1iOyjmG%bysK;-wvLRdtXOG^G7A{lx8tN~)uEw5utb$R< zMKzev;v1Lc+QN#JLdXZ)0dIavf>UmcocGgbM87#Hq%%j7s%2O>c@Fe^lkWDZJ`h$I zc4|o0ht6kGS-9+YfO{3B_~p}y<6P?G*S+XT^z4-zFJW0m zVP(oaL46w^ZR8UDjAoc_h$)%Um1oE!bbP8CYuSMJ*o2|wf&{08_lqU!Gd@4KPM(gS z_hzoTr{$*Ac;yx`Po>`(bFI1T#+IiC819Dgy5_l#T3@X|>1YM?rPY_;3gad1sPCxP zuYTWdqiVY6&r`N0+FLv?r;J@G26bjyN848t=ZJ z$urLV{XV`!j|5qh))lB?yq3e)nGYf+vMWv99CNUfvgho)~`Z( z(=j;>!hQ(%*O`Zzg_**ieN*Y>R6Mxe3bc$Oakr@u9pau*el?#kic4+l7m*Y5JIrkV zd2s5|!P`cIFBv36PM&c#%(Z|DF7lNa zOQ6;%yPrmBn};3dTVobLDrWGL^u--)SkQn%w6iz`JK%R!ba+Z_2ewYO|}3=st)?($K`&fR?dJ}8nzh?Xe1C$mA2O&g%{^~w(;UMbvgcrQs@Zwof{*mj9zXtm2sb@vK z{C_NR>q%Qu(k$e&FF1d_^lN>jSprA4p51)`A;HE*(vrUa6MvgeVOx*vgCo+>tB^j} z!B1rSpcvAJpx$p(Aza)ZJBZx#eTK&E2|2jWsNqKrKIO%q$b?+uYpD+P?f8av!>|u#l3TzOYjN2&vQL zTqKJ_D`+E2K|uj?o0}b~nH?y1Ys=0Nh=i7mZy`Y5FsD1H>DKGaOVh-ehsy^ zFi20OT%uFs@Ix=AMs@O?l zp{feI@!Q)gPp#xSEjt&jYu5#HzV=vhw?uQSNuTY z?P97TsY7{sV^YgB!GDbSo2gyS(3{XzGu7@R>6ea^X4U(`i+2yYIh6W2Aa)-OnyB|? zbFO;_psW9;-pw=r5Wa^w`p0IwnFM3E{!8J4HcELS`Y@u$bppy+@bDaZ$vRWCE;Wr? zZFtzRv#YbyZ21MVkD6-#r9V-=(_TXFE_w{tOreTzk>NY7Gd}dayj$2I=MTKO7GtUB z2c(`K2Vs1BHE-T+_O;4XKHJbd+an#@Gk@6g6nxyCRHNH)02waZbmt3V^iDrGkwKH=f^L>OKd*;~JO-G%gpni&Q%?H=>@!%$tIoOG!g zN%nRpB9D<~MJoq=-o4mOFz5rH6$y2T&0pIM&!#`YatH^((*9ZRPrUwm{WTN4iD$;q z^<9d1bL3}`Tfp}JpO)w%x2->0-Wm@vk6W;+@h#=BJxS|?Z^g1TY%hLs{N ziioIPzI^$jNBGD|{J8jK-k#h(HAU4>)N*jH>otBcGH7f>87oJV*TPEZ3QqlzcDc>521<`1i(!DW*< zh^(z065#W`rge<8L#rn+G?BGoU449}$}+u&!v})2wnO%ds(-=mCC66AC6o?d6~E^x z>5|ck8lJHzL$S&K$xpwK4s`)IC*p*U?#&AgVUPfS-=AXD(~6_L`bugATBFNmg~H6! zr=f_#C)RYm<7iu=J`#!Ld?cQfYGFuu>(slc+>mi$Y4wZFi*2XF9`lvisR;_tWNE$b zHnjfi3Xp+K8>W17$?_aFdrgm%;@D5omQd8|*sR*ghxs_; z^^gz~SI>>quLj2>6?=m^ClCnLudDMvo6T%(ZHKO?k$yYhkk4G))^?wtpP!{Sds4KI z30o%2(l}?eQyMXlkxAMC@8$$9PhPoDuQou?!SYumbOp3iPJa%5*`1}oiJPFR1#77!3fRLga6aw@c-q-FcxGCe8XY@DbTS1u6 zs(XAN0rUgEvKw~C{rw*1j;u7KPxU`Fu^2m%{{4+5dCB+9Pvp(fts8vx%@(}3mM(&h znMm-QR$nQleYl6$#a(%uDRj2w`mW zCTK0uLzwl*^ex9`U;OaH=E@vQ!$Wx7aNOv;C&jIYgs4oU)-6*T3tn0e(OqxjoWa9E zzY`*RuL|BbL1K$6^}m=MXy`WV$+z!eFf#T{G3PXLQqdXI8$b<)RUqzI??z>~%d=zA zE-`_-?`}@f5d2Ihi#OBkAUVC)oxc9tEU`7o0&tOYL*r^?e+>mZ(Lk{Nz?>)K^pJ37 zsPoTpD^B)BxwyL@62vKYV+!&NQUjj#%QUva7Q(e(epIkFRaxxQhy1?0KAAcj8x~eD zVy%2*6OL`|+-C(|>(0)mJb(?yUmM}{s}WgK=>92)8cr-^L-!>X+-%Qu`CB0m0E04! z5YxeF!ecvYLp3~;Wg_4%1I_qT3Y%SuzwVv60-}P3LwRrAJ_Wnur?;8)Q@` z#g46sy6$vss5$j6vFDLuL)=~14@d@)Km~z`&Bk3WEol9ee@eg2UeP)Gw~wDgAtufR^OAq`J3&*0b+HqlWJO6jt7Yw)2c2HzawN z1{>!to*Ph7Qqwfkt#{7fuO-`Nj9=mvNO6;tmnWaDl$>5xLR~Z6_lUszx5x*J$imU& zfV(><|MjLk#-)>2E#D;$TjCXd9C$yhL6?~~*>j82HZJ4VM3&DZ~H7X-gV5F8(Qy-KOB5%G_PM&9bB}tZ9>ov7BFQ`)y~qhyN#ui}*>=kI)%OrFC^Dpm)=sA#Z-9D= z_Ju%Zp=8xUCl1U%$fBLUGS%>IA)L`9sOLxSlWfN<^|?ix+4G7*6V+Utj@L@9HTxEQjf z#S?4Ua39~_ga*o0Zz96ScdkB6r0{nArQF)4rn^s9KWNH?3bI5;M_(GRg5u4yGzg_W z*4K}vrla$ZmI+ghou3=&MN4%Sv>Eok29DdiLpchqE2(hTp8luCGM`EO9h23nV-*ss z{ndmyds7|hgzcABcnep?V$H0qD4cL<$g*6DhKC<_c68X7*UZc>gwcvPt_|ylk$&Sy zN=?1z=kH$wgQ<4H!yGlOPjg{G5f2MVXa%z_R;B9HS2s2`z8|;0N*@bBj!}1acPFM6 zagbY7v2LU@YG2{?#7ckR`t4u#1!Aj=BK`=wjad6CtGlbS>GpO)NwdzXX*fx3p_P|u%5noX;#Oh%feRx-`9-{|q6Xp~`w*e~Yf07*|yjV0gRp7CZxiraG^ z(&wnekNhK(x~0hcLDt2e%YLJlwRLm*!iW{DEk1bl=q9vV18-bFB|(TTc|sUdo7xNh zOyHoy@7JyGwj*1Q7BK?UFcOCC?V)2qyYbzO?8LWW&$5I1G|y?Gmd5WaN24_&zSq{i zUs#wMS1nLx^{M)S2+FBAimpzi_7g!aXV$6>E*kA5B}j#0Bi)pmFa3hFA+G3W=klls z$~M`8PJxq3p^$LC6`4_ogMco6JibE$#z5hAJS%1sMPJS*aXDxd4MoMg8jV_tjkxI0 z=D->DzgIe)P}Fsc#Z%Z0xuV;?)=%eeHgl1$(YAS4vU!DK`tw%@ z*O+|>Ee*o{Am!YT`vgkHU`F_-vAyX8@c}DmYB$o|v8o*;8C3#lyUZ`P+hpJIBYi_8 zDA8A?Q|*&Vr(|)vQN&h0Ay+ZEqMp({!#f477hQ$*gJ(0 z82x{;yk$RrUHmUqUlB~vm${)u*qp>8)gmjI$j43nx*Bp~u2F841Vc+UtBOKvg=%cd z<5ou3|2Dqrqd?<-H14qfi9;+iQeB7=iM@o#C1}*?Ep@}A4*dgzz_=F?0aeV4V{VRZ z&61R^T5Uf~>3GFPzIx`TN(>+HcmK2v&;T$&fL-HpXghXpvw|O}TSU)~XhL(F-~T;0 zgqd_LJJoLwI5!|S0|wT7Lc;&gIeyJbstoC)wZGc6w+92I25?Z@H8K7TBIXVb$Z#G} z^IF_3Lz@uNzXECZ%212lRhlg@iUCTBDI1`?-fjTNehBqAVRGBLG0-ie4);ixxx#FL+1spKcQagdL0>) ziJFoUD`>D-zioY_)ymp>s_z_eqj!;3RFvMY>U%t{L2s@Ct#Dp@b=GhkeZ#u!Drld? zfB*h0mlD+7-Mz^ni}XZH(fTU#)s?{PQxqrM#c2Jx^U$8|w5u#bbmOq8zYC_nLOyz+ zaOqte4*wy*zp3>Dgaae6FvQ+r3jJ$HK)Z>^l`H3a#eYa5SEP&!GhiTe=NAxg?a6?7 z7#!dWvn!c3R_9?mo^34QbzO)0u(ta(_og zFI9C1hkP~;4nKuNRcG`=KB0^Q2&r(pLihOic-G2pIGhutCN|p|kdug2YyU%5AHpS7 zp_7x7PcSKfT=IpbrL3-ShhwUWinh!qB7%bRl;U(*Px6l1Saz&el0#RiNW(DeFQUh} z?IEN3>eh=B6{OZ@8iYEBCM7AQa!60SEgBme8^4o7-;^!n8y~2tq0t{wlwV`}Nu;=G z7`I1G{o;oun8Up^^~}VRS#}a1)N9&^_rW1(HX7*(O4tX70l;W=h^Fn+Y{U}bmPpJ1Cq#9nZUl-Uikh?Ct?t%U)bK{(SLOpfB zHtGb8|CYR&N%`@Q%DeanLjRAkH+kc?c>eE{<)U*BM7p-qWnARYQlGo18X12bvk4jg z(XOVfwF+c(ZSD>o#bpi>h2Iszw=@;6iNV{$%L*>t!qWC!DKiR7Fc&q z;{x(*bi=CBn{8CMj7fQX4$|{|o`N#_!X*V@=9` zSe@^VMBNWyB_Wnv?o`m|YQ)_!2Og$Vqa;f2!)ElfJL((&#$j;sy7&&{?;xO;EHDlx zdW!oP2OD44K9t1J$b0AYV^daG=s50TZoE^%8&?2mST2*huNfeUt5lBD1o z?(my-|BZu7 z@tihNb&Vx-%HM1rY&aOSj=i>x|4eewp-@+wziq`{k_LncYplV^Vs(9+e4A!f^G98NjU5~D6xY85Jx~!6iwhC_G0s4 zY(UY2Y*z{P9GTur4N%nS71JM@k9yhq-;%d5_K(Vcu;@R=-r|kl7ypg2TyzdXqDU{R(#MNlHd?M~ zrjA~(gI8KkNE~c{Wr{ z3)%E3s&uW^ZSB8P89_gTt~KfS?xek51$*lx!Cpbh`zm|uWZhe!+o>goNOEtrn+}Hl zH}Mw1|Dk#d6~8Y2G1qQ{{O=Seb-CE2oSoVwi!?y0PJ~LvNr*;Gd9}ANG*sG{cRef1 zEgG3!x6WaN{#vL?t*?>boBdPVujbpmn3T9%VZR0_f&hoEnN?%wl<4Zp*VAtpd&@Dc<*MW^Y6_qxb$3#0f#$k4xGwCC8gXd6~tP`+}spuwG6c2e~@CbhPfVu zNvVVK_G4?49c5;`Lm8QwkOh4-ttm~<{kA!v1Dl!Tuw&~0! zA&TM;F*S#My=~-RY@MeBMOb#{`k&p?Yj)v%>=4hE!w&?pepdle4P5BuM#l=@U;qXM8vr%|GArK_W9BEUONQNQXqKYAzqz*(>VC(mX0Z#xj;-^4p*=s$Y)#y?d5L-l{` zV3+?j%!Mw@U&Zo?J5PSt-u8-A9iiU07jatGp@;Cso`>Up(pNn8&hm!(LtlA3GP&mm za3~zSj;I`yD|^;=h9mdWbCIh9&wprWxo|~lm=V*Ut{k%bI`&hpS6&5qKY!$@OAfLv z-n?)wbie|s5M8$waeoPX_R_9ZS6*t%wcK^#ewMx;b~8Udejjg-=KvlurZHxpLljR* zKuYr8*(xb*+xVTfI{XPX8A3in{=%lA-I;#`yNbL4XbcM(8QXc@o&F-E73XYEtXS-$ zeG&j?0cc`lhK5^9-T5*WGzRZxmV*2J^1T<r8;+z965xaQr(=DgK?w6u~$1Us#1O9Xi`{snhWJPEdC8FNO_KQg-9JWKuWj7*%Kw}vt&)2VZ zv*q^W^!=e#M&#Ukb8jS|l8~X$8QPr)+k6+7^jjogZl+x6hQ0aE>oVZpw1w9;o9N&D z7`+c*&;Mr^9?ocZ;0-escKf=e)Kclah?JuS9MzQmmvxx#L-&$pFKMt?gGHs<2uGtk z=2^(>cRZfguy(KbH)U?R0<!5_7;kJBKTRO`7%sKm|okF1w|K0~Azo0Z}ZYl15y3a-v(wl&z$np0nzd_We0a;#T z&fVIZ=JZ~;H?x?JUEZ6f&x-nhv>Y)WdB5k%L1=Rn@Z&jyZ64aK(2BuoNnmxXd#j{p zkTZZ?Wjl7gvx9SAP*`R|%2H4|e$v&;W>f$GI7}rU~-UsLW&3w1;gKV$`gb1*h{_Clxoh+s+2YnZR zTIxdps;yp)x+G_DzLG`s7rN+a0P=p09=WrV>B2wa1e*JFS9(e~`B1XgGWImnea8L8 zbx>d2rQ6E4u85x4Il(s1c^P6P9T_&dH9#0eC4s)b{9T0sb?jP5ykG)pe+-5tm2zOn$tULip+f`6|Cirv8E_tuaYjRo=HU z;s-3#)6<>5-i!8Y$<=NA)Z!#7OJ?G1bwB=Xn5<{4tYOSfV++YSNg z7qxG_hR>^Jn`u8Nuzy-UCVv#Fh+ii0mdYqgLjw0gd+OubKyOVeTrGKeM+Uv7NQq^w z)!4*>z}CmF1(=hnn1bYLi;oTcPjxrbY##zPlmsAyT0bh)SJgsJzm_QbBa&lB%pQ7Fl zlcMDd%C4}gO?54Pu6`9G_Dh(<)qwChz6suGBM)G^b8vC+W;;j?F@I9<<#FVj9vgUj z|LeuyuSkJcE*1yxKI7b-nafrS*QX*nw>KmN=m*5|S|#KAN}z7YM(5LrvGsZHeDz-UHb3z(OTnNha8}=&c997r&*Y+<`KM;pWG=Z{68-A!`Zn5JR53HQnwD?&p|D<$veGp?)!H#h9~9cVZ42 zLttT5VEQI;t-g#aCw<)gAPX!|@be{Qla7I7;^(+GNX%sdkOo!a_t}jsEz!Oig|3T7 zbqLM6vOK<*otNEsx*-hU&`C7(58Zx}Jb+vH<*+a3nOrLs+?xB33GZiZkS3fC!_ChG z|8U+@n7B+0a2f6iR|zq7Gxnq|){${y4CFks*KB?6Dgk_M7h$&-wWq1u=0r#ZVBSkG zXvrV&{T>Z!fw4Uibl3N>JtqP4*4rGx|KrsEO{_MZ2DaS8k@V0`TfU8REoI<+4bq(3 zIO%^EAvT`B`3j)!%F5MadoRt)6ky8DI(&MEx!EzuzlCp7VH@F?JO3#D|3g++yhvO4 zr<_0u@#FNJ*hjTvgG!!eN=Z2ogetXA4=3sp->bBE$sTPHDGU>RM%VFFyTf$Q^0$?i z)doU${$T8%eVh-}!q*pq7seKk_OHCEjB^sbX2TvYimGgEge=~Q*=>0<>q8q8g(F%v?cV2nAB zT2DhYOu zXL%#GPlzC!0F+twBY9yF>%@She&6$Ik=U)A4<4~+;d908^;sS06_D_fQPLT{h#;?w zV*j!*_%KhZ#KGaX_-=Xr1(mMW0RmHAvnhRny;ZH86hQhT-}Cz%I4?#4l50O;?N%E$ zYukGKHWXh10PyAK>m(B9dpb?>dm z!yf^PACHjRXL}WQ3IL4(S&*?$uB;qnHdw{Jp4n?=aCWd308I?xV7QM?c#FxVy4}68 z2iO>9{DyJudRDu3`;l7@AID@$liE4kIAN9xOIH1fnk?E zhcJM=$UlPr#Zi#|TY{&_#WC%PbFcHzMEJ$if?S#G@)wBC2AIQA&rW2`9olE7K#}+?pjJWj&0ro|Ek&(>8h2SePKGFDN7%jx zGM$+AW#upKP(?4kTW?F+rvU8tUq=m~) zcBGNzZ<`Zl44 zmOfGcVUwTlvyBDVR4GL5ba3IzPJ>6L-&zEx)V3^67tkm~F$7$3JHvyAJjx*UG9x2H zjyy9huD%tdlO9c6V`tB{J@zO6VnHnkjt%ovS~O%``02qT)}F?N3?pRf5%(n2`2NLW zHNnt{y`ZiqlO`d@SUKkf`&$H!u_Bf5lAp&XhRduhNsfoJz0>{gw12r1Q3Ag6QDbxC z%Iy&DuNUdGZwH{o?Aj9SGzH?7oD*frWo}+NIyLS^lbw8aawelt_(7#t`%nf!o1O!w z#jtR>J7dKMt8UTYnPdh^Yl;E)SgvUtyYJGatfhQqzR=H=qV^9@R^nSd zwBbDV$lY_pw8EqL{mSZh2 znfdQtIDt84+BbRRM$2RG$5(5Wo4aaPI}{f52|+-=s$u_uHo8JYrl56%bDXcd{Q45(@XA8{P2UF2=|De+ler!{Kxx0 zP$I7ALig_ut@;)@##Ha7GM7|ePrWhCM%OEvp$nt3%MzQ)(d~~u_(ClS6-EkTPNCC> zg>7krW-Sjy{G1IoTw8&U_OOdIISz)Om>PILQWk7H&$T~dv%l4;oSGf$K~|5=e2v8zVaP4#_QZa4dl@kCy* zq6@x(e+LQ~Q>HIvae26ZbI&BdIGuTYEtk1M@FLRgsj||Q$@2QXv4w`f9{AB%Gqj>v z;fO~~5uD7q@P+sg%I`J0NfqBoQmcQeFK(TXW8UHWhhl@2@o&3Ab`IWc;V-sCc+xf6w~ znlXM|1bIsXu8ibZii(OQ!$HQXt3}r#;iA|V4xB_E_9%QomCO1vCNPko3@(AlohtC= zy9~1#bScnL!*dDwPEcklP*6V@kt{IR`!M}*cR!Z>O?HLZ_=_98)4IK#KPC zMA9|H?SuFjogaC>@7Y0OT+YHZGctIiI2}#F+7Ku6>PBUY-T|9Z&E7O?;4Q0n(EzXX z;MCx+cNEkYCp!Curk=`3JcFZA$ZKm3%D}!sB5)(^FBZ$XX6l8IWG?)bi1SZxqua zW%}MLsN&5H*GgjR>*?Wj8y7mH7;jYRnwXQUCcC)IhrfFlWpq7>#Z)1FjVN0WmbB25 z+qH)~hM`X>le)wQjR>pLf7z}L{D08_2(v|~818e>p9Tm8=q^8@+$a9B#rSF?u-40a zkA4DEHvbP!uKRVQ!qQj?e0M8=*H|?`z}o0RzO7}zWkbb|C$aG*-`+9ghAGahE2`!W zxZgPq7VL5-ylrp|Tw^&eHvM&ouDhA6F zsr4?-1eC)hlX-J>6*G^hy8#=WBfDAJK~p&%9cp{t4OFfSlFKPHfDXHHvg$qzZ!fGxX@p8kc1Gul88PvU==0Fo~ZIO;|u1% z!crF5nwhyMe^q=1Wz!p3b9Dl9bPW`lx!KwLXpKo7;kggfi#lO$)nF6?vS_*2Y8FOGj1e`yz7 zK5L9#G}Zuu^fYWqLO0kwtA-=@CB{2LuNI;SkT5&ri(e{-v^lbWcqbJfrXp;+^i+jC zF$8C$w0-gzfOZ*7b6MG>_|TiKEYrafk+o8hm+_gfD`@sFS!auBtb2|QXnvf?=r&qb zTN|uD?hGb(X9IM|jP0lO)XOp;)GWuqLROIl$b6H@^!DLSP@62_9|=x{ zwWH~5M?bDjlr3Fj^2^Hmu23-her-_}b+1OBgwZGwqE47Yy2JA|L{LkE{V&OH zW2K)LW8YmtC(Mr zI8MK)D!^3fBc;QQq6gWPGT7FJc?%`-WP}!47Q82Cs`xI=47hUMuFIXw==P>M&v>M! z=gUxf4Pom~z@InY10JYXTOU)9ND*>uW^ugP-F)}LjLX+B)O<9Cucsej7aTBpb5}P( z9840&3~+9juWoi^Kr7I^UdTFM%T8wG3eag#vEyHcI%t@)%Z4=Y858Oq2i`CbD=?g+ zcJ-8&mR^3tK4nuW;9csp(j4Y(h-n1583E~K>M_onsrXTGGsw8$sl=m5JMoFM=Gz~+ zY$`vvno2-F`YNj0oinw=?67;n4radOIQUg|pmt#;_nlRq>AjC}%GnpchX`aQjboB$ zL>S<5nJ1g%v3343o^`jc!KANoEa?cVd}3%>T+XazlHN(b9l!fXcKL@xGP!-YB#h772Ss^@?ys;Ph7`N)+W|JC8O2Y@&EZZd;vBE2(q zI<<(n2ldGT&Z!GCwCM66o&|-rcqPh#LeV#tm}To;hbnr&d-Deks*D}Rm6l6qf^O*q zn~GIE=v}v8#XKMf9_XEcokI3ntBKN0g|w*Y85t25B)5clkKmEsbcLwYnK&V`M@qmq zsmTYSXX^%S^UE(C4|_WN;m5*|LCb8{P}xuAUmWT;s1tlA6^&XQF)px}NUn5gj1WiG z%(dK;c^^lF;oEF@J0GUSXyF`&mboLm1vBoL3i?`qy%TCqW6+%!;rP}83`Y754~U2( z(0wIc>H2f)Tvihu)7DN%>y?+_bxTHB|4Th;KnF(p3=mzu+@%|#RhZ~*152$J4iaHR zO`nwdI+T8dJh2_4RoSs0Qi&vH4+p+9>PpiSf~oWH7h^2_*aLd@LF7Wa8XZ_(Xr9Fy z=>ofq!CV-J4^U5F($Hxo(`j?E)ru;6nZut3=Y_Ww%}{&t+}Z| z^MP$4uenCdf3`CN6v&`ry%=t^)vURX<}13jz9ansnjU!VD%L6}*e%n;Yv7~drTIu0mE){W*vkk{;VV5>-G)`; znD!|8T8wq!)RFE5UV3EW{Y2Fe1OL=JYuEI%5>;*oW13$fT{tNh9V;U1FnaK%F|Z~5 zgLG({U*3;9lkE7608C~u*t|}{kC|^Wtyy&DdZ1YEp%-tkuxN56u3wcKCz%Ap6hUyF z@LN%;2{83==8S%t24nDn9*f~?p4lG|tjF1a5(~C$d)SXrdyv2p>J0%&;)TjJ#R4_dl8d>jl51zFsPIFb&;}Q+L_d;-;>dCorme^K$j0j z6P%IuZYLKvA{?;K#k>3i=sKjTD74ac@o@PLyDFR3qe z);s!H!wtZ38*i9HJG>mi8v&5#6oAUGHZ2k2pdzn1Do%a0U@m0pN4o)M+)rBO^s*?b z#GJ91A3>5dG&KDs1e{WjOZ~fzMt~Qs&WAI0v0oTCxMfDQfjm`KPE^jhT9tE$1(cHg z+7XEkm;y2+@KBSWqcFxGrG~8=VEi=E6aH*AQ+Dtm5W;~6BJf+Dta+qfsu!5bTYKq` zG4Yq`1uwaQPMtfsS93v*NY)P$Ux7HGrLV6q5eK{+J=DQwK0l<~!0g3J8f)cQxwYOk z;NVCtVh*ODH^dh5e5~i1^4(@2B#t18?~N*7Tes#+_6zB(_Svrd^cqt@R;{b_@q;74 z2Y82mqOY+i9c)VObLADStuDobatz*M2rxkDqbLe2En4n~JA)^a#|$k2$#Gj;G0==E z*iUAm-+9`fozJ8;2_C&`@8rR6K3wLHNjKh%f~X|)3mf%h;Rqz8QW@tpb~8s+y~%cG zl=M5B`wI|>X-tUbbrs6(5!SBeq+PFkug2H%~bgDqxcYRf+O21fhh0b#G0 zQZmlxS?ecp`o^mZ9L|1rV(4cnHwYI)&#)4mwP%IyZ; zLtq;rjTsD6YcD%T+d8<_pp1aaT4g&`FE}TaIAV(l)Y1Wuy|kO*tTlB=1Rk#o(dCYC zwQ&X2r;PkMKzCJH@l&PeS@D}*@dpp6}yI@z9<2NN za-6*6MW%+FE+rX}aR9<_c$RDgkTRSJ+cG=yJNQDcWE+khP|GXjD!@;D0J@gUV8j9w zUKmwUqwh~1Z4S7^>oPRWEo|mRSEA!WEMwJTfEEaB3pqPOGkD!W@9s@!h+@T*occy= zdh#gHYeJe8Ze&rNsXG1>$c%e==ZnkUklP8=V%VcaIh;c$t8&3o)9eQjm2U?t6a5i6 z;|UIrD1{l_nWaMo`3-=V;SdV%_5rvtzQ7O4DNr`tNR2f&Nw0E7P2lN>FlyONBa3dI$te|dq zEsrKRNXw%%1_CGb+v2L|Z08%rOAA&JN;Wy|mD%PaO!bCY{Vr7*za2FWa>UVPkO{a+ zhL~}&4g+z=Y8ucj-997MK%Z#YaRDMOst+uj0?UiMfeD2$YRf2l$nWJ~-wTpDP99P; z{T)o48g32kr!5?Gk3C{LHwHd|wBcY%OslPSwLkyx4(}D1@YDnKtV#s7r9Rw z3l?F@DK8=4ySqGJe#Z(bP8IW79=u}I>J(ia-=G_PXQ&sZUlx)GoR2~~qH4_0^bOG% z^@_1&Zzg9d=|~h*Eag++%@r7*o3zmVl;oOUv%3EQ#8;zA@U*NRbhzhAx>~$Nl-?KP zwWXQUFdR$ZTrZmp5H9i-F4vw;xOZQ9Kv9D^%-$@;(Jr_b^;O0pX31A;bwmXH)KMCa z5y3L2Dckm3?~%8>F{YCPD~3Q1hN~Wm;c+C^S`%KKNG%1LI3U`cw@!HO`=x8j^$x8u z@&{MiTUzAI&|bEa43n78NXcNgo+vzW(F%D?5~&G-L!v^|LZVNpX~rtZ275bfu*!yu zgmM2~zgaAaTi~k^;^1j^_?j4?wOqUPMU4JU1n`QiQWc_?Dn(pAI-=~?W-9&J9?B-J zN2mZoj7o<$4O%;CS{s{Mh`<~qJ^&qE<$baD-2^#(LlU4^?%>%fpW9bhKFpIbCm8^a z(Fk)`#1lYdX}*pm}EBfSeX56RgzI} z&NLjo4l?(>0#Wht?Dlg(@Clwy*!mi3QUB>A>d~Aa(9TV4vaU083Au<{m)>+>^eY}l z;yye(dKUb67esf-b`kEEtLs(KEA54}11~H1=(y`_J<+!3P4R*`otA5agoZGF^j${T zO#LiM-$5)#;21eGc<`MmJ2}xSxrW07J2s^wvih(?+g~L|L?{wu80>CB?J%1-p0IKC z0B2Mv0tnHwqJ~2*c5Ae@Wlu1PB#gg1Sm91^p?@QY@%MehABfP7jnBYWTWRta&HI_o zSe!H7!3?2y$F|Iml#=N)b%1@iq%+)Z;HmFY&QmhLoKt!SgbMw-Vfq*uRm85Se^M2L zQOT{xxf~VR6)=_TFD@8~c+aXZY<{_32&pWrkoxm;kc(kvZ^o%wIm}00W0)FDK>wV? zzWam}!x8?zdZVz`Utci(tFHc^&g=ZQCYAn8j?t@o&>ZTgk?T#cUDgMMQz2x}B@hrW zvu6)_2$X>Q{4d3|`8IdLvy%w-ZlvN?qk-&M6A){;M0i|@8NFE_L?SdVYcpMZ#K#G2 zrxpqeNRBvY=u{ENhzaX{!%%u2>}81S;n>P=0viwP^{Y~10k%B!`Vps(bE$WHq;(wLTkQ`Hj&)2?0bVr$9o=Gs$*gY@n1VwRd#VIKZRrSI;9x zMm5N}N*y}BsBATT0AJFTg8jTi1}_g!6nj}rMf2!n;`xN?t-9r@yPyx`bgoLo)}2kz zH)aiQ0twpp6!QCg$c`Jc@#vaWuGl^)>8L|J-04l9#jH`mDJ`^ zOQYfgV$^`%nOY=1`eS5O%fdJ4|oG;Lj;4EXMZ%!Si<2rYK34camvqxqt>3O*oM zn%=G&a+CrU0X5?ix z3#|FxI88Su{<=mYBxjZfObEzga`0yubf!$#sG%+f11kd#3KZf0syo9ARj`4*74TDOW?yrnGt)wdX?iD^3C*n!m^B(Z2^Q-mr|x$t!kYR9Gi zkEyAt^`FZkQ5!{_rE*SUzD`=!we=|{IuvROEi4F3yhEFaFEeNEM9MTKdeCO>9^|33 zn@CQ=WJCPBPu?O_-R?O9mgR9)bh1L(R!5X94-}K(_JPe{1BaauxO$Fe@gSqe5e)Jy zRj40TKBcQLWPz55kwsX)>YA4F^WjTj%TgV{oQT>_f`n8FaOO@XfbtS)J2xkX>6-vx zXe?Qcy%#AVta{(=}E6Yx3qE@ru`2lJcH(qSn(E)Qvdc$3lex=$94 z2S{%GmhtdVP{JAi%A1bK4T} zCykwg)EFE~WY6e0Q}z&1&Irp}0D&f_uvX9}x}HIXF`d`PodR#K0KpmQ zZNC1OKYT%DbwmXBCNZHY18gEKD-W4JMbvGODn0X2HFGd$GD6&w<0%S6A}pPm+uzrq z|EpHlBG-3~Zl&Rw-wi!OLn0-^ba{z5jB4)m6cAbeDI@9O001GuxXV?^1y(tU1a$s^ zU*+ELnz-|7y1zjJr<^a?wBqE!gq*KmSq$jyZWmTZ3>1L}sKrp0Zfv8{}6 zIv0T)D{mj)qY@$5w=~%SBB@+|n~2td6i|MkoLbDK@#(r3eOuZu_D%b1^k=CLO;h@- z0;lU_J@qpxonEl?8u#UZ(p|CKAkP+|aqhB$8faI-QH+b&E2ApHb3_G-ac=9%QxYUEV|rn zaIXoG1!6b};TE#-uX6lAC6{3G0t+#$J+hYmDpq(ia1bQ~IZ2ZQNUu0}s6%hJHv^v)FA+ zH!x!jl4u`+)%}r;5J(m(k1AxedRw2q-G4?;fbHj7CCd3`t3x;ZG|r=x{8cvO5juM= z9z*nUy^|_gjR0)MV~yw@$lmoi`b^FMb7t(tP48cgVL8JSMIV zDl-Had)*XX2>Jo~fkBgAJd$EM_BGn&PK20qmZR+&rN1?cYL_;>rG0?R^M~)d$hp9(UK#E($K&n5SI%xh`He& zp*De3S*~;~FC;qU!y_#EH{Jwew*J&Xg+lMQ-5LE zFJRFSQM6|J+~xD@tmXiV{@1p;5BJG+BCN&OF9O_J{IP8baMbGQC*XLfh`iLdw}p?7 zwdU9VzO8!w?^FGH$$h}uy3I8E+7uc6d!3vg{A(TwKLid7>)OeH#%QO$5r04B@;_h& zXIL^{P(J=eM%Yr6u1kz{DFTYp$5I9h-X=A#k|Nj^N|03V+*Z-fC9anPEH4!*K$N`!< zl@`;Dnqng^{PyDw;CO{_QqsO1H7}gyx!PsRB0esh^tR~M{cG369_9ReR0JG}YrE`k zU%Pkf&BvA>o4YMm+a8md{LXeNE4SE}k{gCF=K$Avih`CuoN?1HK!`bubZ|cW{eJ)c z<57=#Kz-6_Do>LiS?b*l(kMJR-9o!o;VGzMy&hMcTl5Olpei_HcwFW8QEA}bLoVPz zMM4E|hNO#M{dRQT&QuB8s*)MNsgM_$nbVs2tTgtE)%0v%ALREtS?43$nr?A@y_H7y zEJ}f%;4oZfk!?Hg`;W){)7Sh2j+gCj2ld|P04J+YX4`+;*gv&G#UE+!HE;!_@g3t^ zg@@n=Bdq8-a5X$$cA4MYTgj=08~<(J|99>2=c{Tz_~usy*HzAVXMC6YI?%ZE&&J1h ztKXhIYu2H5`#%p0fc0R@ze#U5t}D8=|EO5>lk@WSbv24Te;!|2>g~K|8kz5|D#q16LsRmkJ1aPADSES&w> zczgA#72BT%R9S7mbR0Av^#M3gWqIVt7vRMAw+q0_&t9u0DOu|&e!jl{pXvTBKHtCI ztgcSn`4%)y;XJ3d9ynR{P14`z-2c=2|9H=^EKb`2Tu|`&c|F8 zXZ(KqQ{wsg`Sbmg@`0n!ivMOCCcDi5?w*~X7n=9`vD6#GpK|+I4bsoe5fFX_oB}Am z0c@BBJvCDi(TB%~LO>g^rz9ES2igt+t2dDudpUv5GM7A#$TRTF17bKx0?YH%1WO!e zfN*>PE?d7hyXP^YsDXPA#8|)sT*_j!VGfQ1NLZ?X3uWH>9s>;=qgV@dAz{aR?D($s z7}pA34QF5^*4#@1>)K z@G0PmE`8?HE0Ih@_TvI`V9?x4k-@$ffsrW{xEF7FqDK literal 37827 zcmeIb2T)Vp*Efm?DxwroL69Oyks|OYf|L*yMWm|~MF{`bLQlnz1Lplx5`<2AmF4EA@chS&nMef)J-1&NBOB4+a9h3E$GdebBRL?lrIyhf)yl!TpYGH5T zY;AT?^)wBQ+@nW&##RUQ_JSe{v&9bA-Dyrpc*;bU3C*#+*m$>dlx&f0*Cifen{(>% zr^)VUYoTLp0sPWaHR6ik4`s~nwIbazDN8x~bxer0EvC_I*3 z98zFYTHCcS9EF38^tfPq7#|mi8$Ih0q!s5(lNf3m$gUo)sE~Jk2W{95ZIBQP3KI@ghg=#1ju__eS@jW5{_%sH6&FbkBv+vQwvPzluhMOI`Hl`!lzEeSgunLerxPhJ zv;>+;?-1d6XS;1G(08-i>V4~S^-Di$on6U7u_&}=m=d5d zE-ty}ymJavVW#I(a&Xi&DOt9$Tr9CU!jh%;j;Gr@#hNOt%2KX0Bx|)feVP+nh zz5!8*3~zz}Pll$os9)_P1FI9&(|+>>!~1X2La`E-*79u)C*yBEJSC0|7@TI+PXtvM zB*lcECP5_f_f`oi-8YHa<@!;|ZE0@N-5@$*$H3E*?2Md&f|B~VQ>Q<|4RVEFvq17( z;lTr4A2rb-HCvDUZ%BTRv(ghN*#+~mN^xsoxala+EUhM~ z_VmvDLH>|$QU?yj@jVs(bjdu>yixuFlT7t9ei!qZ!p3H=Edm#jvH>LlqaLTK$d=~M z`*P|lFCb-E+UwivZEC{W>`mmM_67mP=k$@qtuAA(hb=j8Nw|LOY>sxWr@V)Zv}Hg> zWUuZbZ-;G#PoVYlWhRb)*53&&axkh;ol(*<2gOV1*Nk6buLI}q*L$hkH&pRogvN;Cteq%>aAuk+;JOhdR*qt zVe4?+F68N}?GlP2=J%A|e|T@Mu06W%{#}OalHTsJDQu6}jf=NcXE>OrK;PbGPjZ>G zf1BOfuYCF3vu(n@34^#}Tea*5?+U0XAK<@dSL}TkGne2FQ&;LtM^w2?*(%cC@P+su z2kljU?|A1xZ$O#ug%bgGJ4754xQ+Oxlj6CfHduZByOHC+5k~ef_YjZD6K`dn}MbMM6(yX-> z00=n?L$#Zlnlz%o^$waX53FhEfU7OQe@@^(4GnE9?XNrZU9nq#U2jEFe>@)kGMI)2 zOrxf9>aypS$?oFd#1Xyi-$!JR75B{9MYnf$tLWeG7on#=0)Bb^%$a=*oQFkOPJ_>| zLJrr6G>aTMyj$e>&jcHkquZ$ReYBgx|InZfS2cO-nw(k)T-CG|5g z_2=UzEnp)`BQ8OcU)r*drFsc~Z_{j{rQ5;6w|TM2g#S-8A(6d3njCLz;^maI_`O&Z zm9sQITnxiL_?r@z8}C?}nwBPdlv3UsmsCCYi>e}8=$Z6!;H`fd#&j#KD%bOOJC=X$ zYo-lHc2VlE8A{IMv^FJ~=Y61U=79wGt7&3EztY{w{9d=)y?K$cj^E!yQ53Nk$geSdPT33sm6rCS<|7LW# z^mN!4Sv1&K#I_pH#(3V->8p1?IZ*00ZXlV%? z@u!zpx{faD9i1!SJ`*qEP2t~``9t{;&u^D2ol9cbSP?GoXDKk~DeGX^Py~EJe2&4o zqAGt(xja)Cu*k4{2Y2cs*R5uGDnEuDU1p&>*RS&z5nzig2%pX6{^Z)GbXV2q|F6n? zuu}BM4?Eb8sD{BQ4O5l)DSxS@V#ik&)?w?dHc*_;oWo0j3JT?SX&>;~pz@b$hq&1C z(kyJc8=F6d89v*tYhPUX@#)zQ?7MikEKAnNX`zR;rDrgJLQG|l4aswa{bPTAA>!=4DJJs4dnbm@{hsH z08FXwd|$OT8VunmqNDJ!jDou4|*&{XTmVsC3EnLk!EwRf79oNOW?kqO9pm; zmVcuIm4?0$a74z{cHO+pzn1+}JuQ*obM)8=%=i3!U!!_zG}jWcE4`$(GqO$ibc^Wk zn#^}t%F~wZ#l-`_o3<)_WX4qE5{Jp_X{8`{c+qb$2;Gj^JuC%;r+VRF8H92tvUm|p zoU}EADBGu(Eq&=`zByrpTF)M;y$(`M4DBDZL*t0!kgrb(cMd!{b-`r>-7BF)!P`P7 z!q7X@NEqo1w`b5e_#n6*uN+|6SQBYBKyTX>oXT_xf`rr>aFY_*8F97d_Cp7;8 z?^Xc3H=cOK`-h~KDSdtsfcK%!<9~toI6&~KXg}B=8u3HA|FMhiv#Lpx0oIK0*;@b` z#LyD)3n}%J6g*F@8+-xQOJe2u9TixBxEtW458bRYE#2_m0f_f&SSBGi5M0-Udn0SM$9 zso4i22f3dA{o(JLm%2CBh8-+B02|uscVR{Me!D;vbFHzRbpvk!QUDrVhK)6a4?xAg z<$Zj>kIa0cf3E2V7wNF}>*iUb=5#<7w?##8WBs~)yS#h%Z{+_X4gvhAbo&k6#`ea- zQc(PtU-(#nAGfoF{f!B<9L3!Kmll8?O8+qTH?Tc*lmAV2N(?Q;{jj(_b(#b z(X}beO<``?w9HM*WZ8sSx=ok`T-Sf;GlOVlLVx)$4GBl*X)Pkb*cMW<)Jzt}E!7k# z%b&FB^%+YkhI8hhiC!C-!~&&)($+G~pS9GkQ+>n5w1hjxM0a#||DA11@k8mNFdyp) zkL|Xa>T4aH$Pu8BQRsf=XOyX|-RJnPWPwMPgB_~GXI-9y$OMQt!<>=*@p^d;t+sDg zZsQrk2*e)KVk36OOL3iRF3~pqx4z#SQV&D?#Xs&61_CL2LC!x?wPp9aKI8Z!(2$A( zqUVAzE5D5gzX?Ro^#w9Ye~F$Q0hiye|LV}MuwxzNEYEd;=y|YpR^yLZmji|mK=k~; z$p8MvQGw|>AY}+NYC5)l%wKZ-A6>NS2e;QNYeq2d^_*G6^!DWAxeSm!aKct)&)23$-R z6(b*h3HZgQ3C3%3sKZ$1wBG9NZW+G&VLtss%!@lvs%zMczyM)xeztnZ&k(GVu1*5} zlndkt+2eum$^3Bq=n7jA@2REy{vocvl=7Vc8ArumSgY;7gaDNC4S`a==fb(al=1=P zQnF0ivohDZi@iW8pWURaapRPc576rRE6M9~(-k7_0V(0Va0l0o74IXUSuW2+*IM=} zZ@|}qIDT`vKV`Tn-A!fQFrhzE7SLmMSS9O2k8YxO3{6dq*+b zB(=UKp4b3#U0GQQJg)17g3p;Upy3v6o$SY2F0b`mzXF8VL?A{=OzWgOP0L=F?@qa= zlc{B4e=r~4IT8I(Z2ed0WxO_}1w1*O`d5C0(RHF(a3JRUVkGcQMfSDAALxzfq52|+ z>vJ0r2zavCyTH^HFLtl5imo_?Qot#c-fQ`bFWLk6u{|~?4zKe;D;0`az!?$t>mz@y za@Ai@?_2=9sruWF$1781iswd#&_9;5Je9~HAT-Rpd1h@*c6D^H1K^D|7yOIu zn-X3R#+xd=mLG^s8(xQAo1pazH>byj&J-Ks9eiZ=EHN9Tt<~9EGC$XiTACaouW30zxB=z9i2|RM&WQ5cw8F*s+U>>o zZ?{vSrs@xh@xis*cXTQ1d~#R~5l1EY9B~el%WYK49x#OApg6>Hip)4DOA3*_ipEvr zF0E--O$Gwg6x+_I6#(#%aD**fnil3Qo?wPxwjxQ_ZK`Drd`h;jspNg<+0uJ_PKgh8 zVKp9I2dG*CeaE`APVK&-vs z0eoBk@nZQw+G2cC_h>XzZ{O(c#npPK4iGt;_)rplh3rug<#FLx733AOtE0zHQQJz4 zWVPllO0F$!g4qfbeho~6IL7MPcf*Fg4*4lgi}bk71Y#oUsGN&rUuET+qkWMH`Y4*2 zH7j2eRDe~0kB`?c-(9!xmCP@h5#TH2_QJ;b2p>Q2*;`8V^v3LK0n(8c$?~NY=B>M6 z0rD+Xj(raABRd;n+H(t;ipcfz= zwq^L29;{Hie1Y%-%82jZ?AUlgm)U`i;7#%Vl*6WY*J#+L^{zFUP3v95jZJV~aTvhG zCOEHh=NsERdX*~GOIPEP z^6!B6?KgORY8Bu-_7`|L6Wf@X&`1g8qB+nKIxTnPnUJ!yZye|99wNt`KojxsUMrKo z#3^(=VnCSJD>g8)>ez+Y3Kp$wm`x5AkJ&hgBRk1)nLf?jEB1r%0h;D}Q)1gzV=BWd zrVIZ_U`)TGii?U{Uax6=RYp{a%JsjXc+cJ_s2g{IePN`Ac*E=A!@P`KF}T=+6%m*;Mbf_WA=C z{+l+9h?C*PcF@&b?W+mLqdJ+;kmxxVT)&rg*9V)@&~?!BLq!qC00crF%2VCV=h)8E zh3OpGHmbt77G-|XXsZFBP3Aq$5iI>t_#{@@JvG>ODy$Vx%#FboZuElhUI1ZK8>cfP zSq(w8rApHc^MgyHEv@!DyQ(3fOdDjK;=l*8&F$hF!yB$iYHCe>rE`j{upOqS*1wvR`v z`LmM?p>a@44__hXnU&BBK>+%$dQ}J5*SPlcj0aKNTX4X3H^GnOTSc&XZ*|m62{eyn zLDXU^sLz&;{#mKxVgsTn+m}_V(UeqC`MsGqk9J^dnFGPowY@VH*uajjP3%LtnSEoL zJNMIBz7YUwN7*(Zs}(;(<37iQqZg4yMc!Pzo#iM=h{~`zLzbjcduLoYX!e5v>304z zirjZdeQM{w4^~3M5(%JCC{Mp}Ai23Og6h9EH@R$fUjqp7Uv^(_?pOk<^P4-CHg_y- z?pWH~OStJX0eA7=@R@xgI%n4JxaBz2_bAWCPPtZCA7g};!1eqJHqq2_oMpJsY5)n2 zjV+(!!tH|2ITyFac#Vt*P^?f6aMDs+CvQ+!+{&B0dZy3nedph}EO`DU_;XC41$`iC z6e_Jdz=TA{gDUZT@(egEDR2ot9c)DW#3kVI=2z*nd<>`nCg#svSUc!0LC1byo9%Zf zO2Z%C5xH)?eDRas1h((0(PR1v*VjJ1n%{F!_tFgL&;DgUL=zBsCmKcE-=JaujA5z* zlo|Yt8ri7P{py&%st6juRtu!P1*gzz>2pmX3c(62L%{y5_C~J)+wY@+4Mas@*3UPn zoB-3E18OJ11^bjX-iuN03e>d!rxxva+_;*PN2{;q5;3te&G3YfY13?8!>Lty@aafzVIVv-X6avZZ;cKA zXHRaL9CJP3#X3|O7s;?8e9+a_FdyNWOK2^08#`koq%c-DBb=71UNZ3WlUEI^;{%_Q z*sFqTSg|?`;yy#chtlipmp#_%sm^vpi&(a8H)~5!PD@Re5iZ$1@YT;grDQ?Aev{OR zL!4_tA@E-L)A^FQnxPi#_-BpWLAIYC$8?mo1cxT|$guw9beU!yK%)2AH9!7EqDp{7 z$Bo%oH^xiwZa|%EJR9%+WlLoXRh>2`|I-})AK8vb8ozg#f31Xp&5dQ?6gTj$U8XPE%T1258>y0`EMNmIq%TU%S)Kkg>qLnZnfG%OWJw9 zfFKtCmv6N!qlftVd$v{|UC*f%OKV66C_!Bu-&lQV5InSduu6`TYy0`uGH!G7bpl^i zkI_R?EoMmL-#`PbNr%_0r~v2G)4fhIf(9CAAy4B==H_xv&>FhR<4V*|B^KpF7Y>k@^O5>WTNJyXy z)5O74aU=E-&`}$UVqu}Il>04pH<0-7s#y5EU5^Bq%M}oJTbmD$59EKuz*NlIA0)(c{t9z9;x3vDIX8J zlJ|C?!+$$(O;)#0w(UQAur^F7AS{&mX2-1S+~8g+P`(h+nF`MldOGoYF8-MD?UiVY z2dOn}$tO`0F@<-1eA#*yb6Wy4Ia<=X_h8R5*8)h$5l;op_`JN0bu>1U7VDVa7vKH8 zs#UTgJmZ?qy9W9Dd+WQ6iYQ+0+&1lh9F<1R@({y}=L-y`IJIx8^S>R^+m(7r<0&Tk zvh8oczYN5tT>!3%`Bo^ZWkvPt7Jo0s5WM-pzmOAAF+U=0s0T!)3 zo!|KWGtJ|Opswsd?q7}riTwmPBh9j+#b2NOeoIdUfXJ73^X=bwyUDdRMT9j5`=6{W z&pEHCZ+O;^>2cb08wrSV7-`mu)2Kv$0NaPgD#8bl?6|->71JERfa2y_IoGh^XK00Z z$#_sYEY;+n2&n7mZVC6TsgpBK7Z?bY?rvpI0JsYdyn1GXDD0f1RPNzgA)#}Jj5&mJ%~iqYu@u0KPuuMq7w31<}2 zq7~Q+;LD-Bgg4t4xlx3eJyxnrvcCx$9$m zw0*}ws3Y%|ETKQRPWM6}vY+D|VS3`U29+(B9Y!rvy z$mZNS#EtZBcN_5$%F~DB_pn|lXd^EA$NXCJio%&z&>lHda=qJ+i#wxm*6*3TP*=Fb zm#jD*X;Vw}H!;Fp!r|tW#Jj`no!{_s-{*1yvT^?ar4M)Ar4`Q-(L7OT^2+8d1Ps!ELm+T;d`@YGb;;u?_DK{b9}f7hqBKh zm*(Dag4=z|G~s(0cxJcFmwCSWL??uY;m7Yw^hRop!^C)I3W>V%zFdTtz||gd^7n6 zCH+)2G;V4~S$*jVgjO!8h8hVt!>g+yU)s)<jJ?k&#=mE!J z^S1uTPVFG%y?sw?wUKX%iZ2&?aR&FM3YhIuI8=2PDlK&Fr1sobk7~K^cT)XrH7lSt z;=TIbU?o~V%4oWurDRN0as{pf4*Yiw9v{Nps>HRMFQsW#*H4=72JM!Uy-(kjJ5`J~ zUAWyX;MbcPH?YXdg|p%D%e-zxc+t|%HU*LyFy24-9&cEd$f@fV#1E@)t-^Iou^sE^ z;ovI!TBz@qcYln{D5Qfmb!_q(hmS+LmN_?n#3y|V^{m9#O47q7tY-GP6qVi@9$vGy zMGj`{!^fzWm#SeMd)wg7H$)ZePQ>cFYkn3Z#d}V1<~iL^bJtjk>1Ib+wi#^8Qs4Dvb=q_0nsbcy;H!;%U5=Uk_P05(#ruTADI$A^R)yU>b8*X4`lVM09PR|{zs z21-Ah4M4v%N()W;N%!C1dAIr4-TCg$gHBqNhWI%@5>ZLaz`|0j7$y4h_{pI8jyKKC z=CQO-8%DQb&sa-zxI9gvoLEy)!t*Q3NFg*k!0JO zH;1boN~Wq%6MDVNS(u1H-~3>(vM~R=>%&(qw@Q79cnsXwUS+v336!3m2ro2#4Dpq2 z@h70ug9vEe@gFZ9*EP-ZkNztBTcrC7$fa%pWg*=Fo8UhXe)$Q`BS7x=u{QbA#@g|p z{5B#^gINNsDvUDhk0bo(y>2Pu*0_bk=lShbe6$(-H~AXY=F3^RK1~ zSNNLkYR|8np3?=fY$dpT7&o|rVM07pPnGPDrN;IX+?C5;_POb5IExkT{!=_=t?kpvcw>ni$I*t`Z#J)9)RIj-pm9)C zC*Fw|c&x5nN=9bxn6$mJWEIA;LPkwC(1c^CY0w)w8%cv$+!5HqjkSz+Yt0M^2|@bR zx||CY6R>tZ_0}PHK~35Zbuy2eYbf0wKqeQ5vW4LXH@tKcvO0}J`=Y74umpEf3BD*K z*xt*XzPldNNJ<5))2GEE+}Ir&EbmT5iAgG@KBwvh?bH07NV-o77H%Ex(#HHS(!daK zSQUmNJvc~(X!M@a#gT1L^*nmdByUj3mwVwR-eYM_MMoDe+y+DGtAzSN;ff14cC~X2 zI+IAL`8|u5Dqk-?lQ3$PxRUivJrmc=(LjM#2XRYw<=LUPKJd1#y+zO&bf+GZTC-oo zW+{ruT)nT6GFN!281)us5^n1V}FEt$%KdNG;qm`;y57p#d}&8>U` z4LrcBSSDJFM@Zpl*RSKAyXS2=;?0SWYMb|r0xw~geW$J6y7XMUZGAkvqeVE<42xXH zkd=#1-=7EoAc{jBkyV9V$PkE_?c_jPhG1q|a;cfs;9;FGC!d+&FE%&cYI$VA`Kl&D zZBgG}KH%}rGAGKBj*p8WZZZU(eGeV1dgKsL>wc}sRoavCDa^oRAzCY3nuqawSbNMK z32S*ttT(*wLjWT&|E)Xw{?s_Ci{u6-trrNA8^|AU8`v&T2Vl?CPfg6uGrt#OoY)T}7E;fE#deowG(|^Br%|p4qS@z?EqMV( zR_Vs>9CSa7Hb}izIL&MJ7G#O#^tBnOi&k!wa5xs!riNr-d4mKGvSEc|JO{2%Md2Z} zl3!+LT^CcV+m7@E;(B?j<45jGSc#@|)6-Az0YbfeKDJ`&n9@y*<)B+%n?ND?o&DiV z+}Cyjx{@@-dpB`kn*`A40FwYV`T+ zwVt#~hdU^TdJi~MA=xBpg6rD0M})Uv`J|JnmM=;*q$=abV|8=fP6kAu5f^`XitM;= zz7Fo&`>bWXj0W`5Lvj>|BT9}{%~e;JF_;{{O#}pQop5^&rK$iULNpNl#1;-QRaV{} zoVh45b}4?Lh)VakXu~{ECS)B~XJofD5+{_z3G%QzIy8UcJ(+F3Fg>JvNv;f6nbnC= z08>B9=;%`A>UoFKq+7@Q4EFRNBEP>2REwi_V%oxmY9et|Qe-FQX(vaj%491F$hU$8 zM-$8y0DTkJ!52V>BP>A@J!eqzsBM|PSH79`grr9T{Ee-neWX=T-Tm-Eb!7*o?JGxl zKEWmP;oN}cV*Q+pMyA#k%zm>B;6n^TerM=OSXH$_FAP^oAfH9qPc1s0d5LY?S-Y$` z`m`27-O;gwY{|ia+NrSPc&OY{S>)K|>JwDzvcdFKn2!X>LcSjDL{<6WCgKb|mF#BY z*hU7z`Isn9b`f}#hBuEPgn=M;6JE+e(N?R6AN+QDojjHaz;yyp5(6_Hkg?D-ku?!Z z`x5?zs=M~f6udQfYdH0?om@^L^iu9?5$%9NXn6r+sFJ-yN7hjM6nWv>d$)F0Y3M%? z?Rs}*;$${2*F@!A-!gO*V4Gp=&y#)cK43BJ>Tte1^_`;uC!R(yEXyx&=q_^OG9wvP z+Lb=5U4sp7oXUeTLrI0!zK00+erlv<>;*HA!&Y^S6!xlur5u~I!kZ%n#*8GO1P@lG zwFpKxl4~D2RXq~t4ZcfcuHYcQFa$(niQG#JuU_0LjS|X(-opAkTlAVWn`(YW75+nT z9L_QL)FSYGb_Z_!(!#qNr9df0DWL$=>wE=M#~w}slYN>aS zu@I%>W;8`bd_!8j1I1>xphH#PX5x?z!A=X)*XP>2sDpbT{;9^yg5*1UDWkpS2PP6F zjf-Z-IX*NI9J*L%W-F79bifQlvu{uqoVrkFX~|QW)L0ulM9&%hlcuH0+1o1K`6+C7 z7Pia{C~g((0$5}Ow*@{^gN^CN|AAFadapTKPm- z;vl?oz#+kn>W7j=zeBDJC7<)$;Z#{pA`SFsmZcZ;*qwZe+3qW?`Zh86pE;$l7W>{U z$v`C@@o;DJ^`)+i>mSnTdYM%Ur}1b9xaQd-_s#;pJz)q{a##Il_<_KESyeB)CDz&D zCvMlM6}~q~;D|x@Zw5@B|AG%UQp7o!&*!mU@tAC%ie$wZ-?MX&$K*6Io$qAFUjce; zz=Cvu_?Y`C;-b~u{v_D!==TAq%%$$Uxdr{&TOMTvlT_ig(2FX;Vw|cDa1AKaJ1|xl z5wP9Ffk+_gc$s|XAvH`FU=gvK3>sNTd*rlWrVH|-##+8~&=%?+FL}YI(^ZWj6@<%nvY37LmGA8M|LBVP0%hqSB>Om14l&<|EGg z*fDqzi`6B>8e8Q0=oH$S$1$uIRPY>>8`E{n- zNr=(Q6Nv3R^YJfT>(l@ZZvay@+={4h_Up&|*Q(XCQ3KE6w^AHWql@&YnsXtxu^!N5 zFJ1C2G%RKz&D4i8ges4Z*ycKZVuL=nB0}m~zglSU25lY09EDzh@BdfvWk0ddr{R8xU z-QoY8;i7u~V~FG+iQ&s=^rFaEI4I47?jxpHEo16tB6IHRvx>tYh$U? znc7Fvc>4S+Gjd~jQ!{^n!YX%&Ebu*e3VB=E z&zx9H%nYC-rNm*k6O&S?tA24FG9Er_efhD>fCeZGEklX z-=VMJ$Y+dl2pH1-oQgU;jM8oQXc#z>80Bnk8wHKh_)Mt$B4xmH`O`M4hMlK^rdr*U zZj!y?1YbMK0RowkUOxuW0o+lm(<`p7$-B zK8#l0s}|9Vc_11i=B3qTx6lk5Pd-D%Z-&d)C+%M(kpY7>eGd3#v94%w zlYZ9gbg1k>uWwJ%#j(8l#g!+-slk~)?FUZf5w3iJj13y{LSnfLKB) zWPtBy8>S$s+B(IV&$EF9WMGlC_smc@gj}jMNBwdP4x_8$?~f@TJP?d~*nu{?9pnFQ zsebTeae9rU%Hcld;9?%Q5I&k0?Ox6OFgrW}%Lid?*>0WW%d=O9&8gct(7C+(j`mr=2O7Zs=rt_)$G zz`bl{eIP)X%B@CdQ%xIgKlwoLNs8BEb93{^pc2t_QdyYi?99u50&}SG!YO*8ZA!7& zIkWFD_g$?hwxO}caGy9|ftn-Io{|9z!h?fqC|0;M320|h_@ZQKYjgP;Pc_b;{b6Dv z>Z`lFqcj@h3PS?`+He_Bc~IHs5js;6is;dnsE{_UhkQdjP9+*{KEQqDprbl*`carG1;Ah!w*VnT`31dPa&-#cw>fB+h|k;-JSAY9pr z5xPCBL&~o{t5|;tgZmf$wFEhOrf*JnWI`7k{8YGe2I!jMy`8+#sMI!>SLIoT3;7~m zSxR37y8=#A)1sd`9Qfq6q$_7!Kx)bgO#w$UgzM(x&nl;|t?n%6(@_V?%U4ZKoC2Wr zN6C5D6;o&wrJiUzcpo&OSK7%3P4k9Sto|*Q*u99Pi&tBy@yL(mBF1-Ws95m^XXVwi z5b#&hYg)EpT*8R_)Fk&OLGjcU_2b^VS0=f9xlfiqZA9;Q-t~JjK~SD&@iQw#1!ms9 zOXPs~8n1Np%AN1)$Qo&Kf=h-7|{8!7}evvZcNZWIy~`o5L5Uig)G zlHs)vyDSe96*aU+RvL*pT!8gTQ?@j$9DRzof(Z2_JBO+UJ*m!EEIGQXE$~TP+6px+ z?UH@!mk5^(e@)23aS74%suvlM^)i*BTJ237w8*Q@o4uEOM`Ntk0KzpirmN-WY#3_v z$`@2a?@}wxW#z5W3{wjoQBiZ^VI8_JVsN&Qzoz$Wg#TB&tDiy2M>rSeM6~^%KArc< zB1L8{wTM9cr(cSAoRwZ#X9wDK6v>cb(bw7`WfC<| zJ7Rzv&L3Lu_Z1iVbPgAkC3!)<9xGBj{48Lk^l)A@M|@Kb2UKgsn0@Gj?nQCta|3FX0M+AlHH0BmoH*p(vD$gEbGEt+1;hBp%&L6Cc@p7-5|FM z1(LxR;9P>rkfJ!!vRr zYk!)D1bqcboI@H^s7VGr%@5x*8J$2mz%?T;TRoki6}Dl^7NGez8D1AR&U|xx0Z*#O z<_dRFB!Y;xp<^Pv{?4wUz85@76RT^#dB8?6B$&y|T16!zkkUUwEtDn^HeGp5)IP2m zmxP5Fkwn*IxqQ2q(4%B0eiG|TBgG`t*io`#f_7c5&0ks$cJDW|nl4o8k9k_#8!-Kr zM}KTe%xjPQmvo%2EapPJw>GeiCN#7HD`WR4y3z&0JyBb9Fyk>OJXzBNb#^Y-~%?nu6ZvrSE+rvq-RM8-B6cK`H0i4vyzqjFs5P09 z+LQnzE`64pw~SL1szG|l>jvAcZ?}Jw50rmO(~+Yq47)Gf;4tm=lC_I>`+hcTihPua zVAsspyrJP}co67_@+ZX^7_mgGE~HE@w=L&{d%|nqOy(r=!`a|fL$+$gCt`+wlW=oy&LV%92wY(KT57SO)NQn+Usm+o2r^e z6I`ywTK8-Ar1HL@yCOoXVETytv7^J1VC$)0w-EF@#g8wPln=}E=xY?Phg0H8Bx)LU$wC*3vPnj|Ya~;p1%3~mJX*-K0?N(jP4y17&RWEc1?Zs++VzJA z2IHneO_c>}Ttgv&(~0Ze*5^~IZd}!;+vV?Q@AP1x-w!ZBYaiJqa^Q`zB>chy{xq>y zug!g4)IH+w0IBz{{8?gz6SbdaBqbfUqAR6T9uwrNUt-Zc?*H=I81b6RD3LA7l8~VE z;1FM-nxeuGC}@~B+UUe;Bd1g=pvPv=1-F$BnAitk4S2qjn-64Mr=&HIV{^NQB&2rLGiP635(bmf#`uq;Bf#Gt2he^-qmB|6V; zCN<4}C!BTY%#23|1f=@32nRj z^}gXeph4$M^rKg98Qv}r$}@{;$m_9l(?bV^e#2VcN^gV3xCTXt;H(rnF298zsWFOk zz3_FUg?At0*ur?d|8$i|@v9u|JNT2i-5`=D@&5b8kQ-j=*Oep^if1)fTfA!iM#TAdvc?{~^CtQoJA3?F9ANtPT5}PlL9Up^wkc%;XIEwghN9^Jt-s+f~uO@_&{jc@f=T$t6CoJgcn-= z(U@xA*y=5b)n$r^9@w=qvd#k+m}uTy5hLsnlsw3UdClQ|Zb2io{#>Epg~bnN7R}1F z&wVn}n1SjZeJ)0*Tzaxr(bqs1UxO4vRu%s6Q~!YTg}sz(FHuLHK(#_ct&X;;*N#1r z41V<0TK>cd*vSh-=_3nEOIq7M+xeJ$M)Agz3bZE1rliN79E;}J!WPN=eA}3dyp-UW zgWuPrZeMg>JL=?x7Z*v}*Fr3jbmocbXD-DB!Hm$HhCBH63TO1>CQ+KeaG zSF)uBgax0q4n8E<3E5|HC3Il6$nXg0^vJj&Tr2ZmYN_pK(4omPIyd!IXjYr|*Bzi9 ztkTEo)Q}zb`mXZ@Q@G@XFrZPRh>P4VV!aTlo96Fq7mAlW{YVq4$v8H%DCgp>s8tj~ zUKHsKdiq&jbF5K%Ojltn0*j7wO<^xy14*kBtP6O4I;fuk-{w!NvRM(*=EbIxtRyJ^ f8y79C^OUH=GD?@qc>-vFe`;s7RMJkH{` + +

+
-

Handling deprecated Stan features with the canonicalizer in CmdStanR

-

Rok Češnovar and Jonah Gabry

+

Handling deprecated Stan features with the +canonicalizer in CmdStanR

+

Rok Češnovar +and Jonah Gabry

- Source: vignettes/deprecations.Rmd + Source: vignettes/deprecations.Rmd
-
-

-Introduction

-

This vignette demonstrates how to handle cases where your Stan program contains deprecated features resulting in deprecation warnings. In most cases, the Stan-to-C++ compiler can be used to automatically update your code to a non-deprecated feature that replaces the deprecated one. This vignette showcases how that automatic conversion can be done using CmdStanR.

-

The automatic conversion of deprecated features to non-deprecated features is done using the so-called “canonicalizer”, which is part of the Stan-to-C++ compiler. We recommend using CmdStan 2.29.2 or later when using the canonicalizer and this vignette. The minimum CmdStanR version to run the code in the vignette is 0.5.0.

+
+

Introduction +

+

This vignette demonstrates how to handle cases where your Stan +program contains deprecated features resulting in deprecation warnings. +In most cases, the Stan-to-C++ compiler can be used to automatically +update your code to a non-deprecated feature that replaces the +deprecated one. This vignette showcases how that automatic conversion +can be done using CmdStanR.

+

The automatic conversion of deprecated features to non-deprecated +features is done using the so-called “canonicalizer”, which is part of +the Stan-to-C++ compiler. We recommend using CmdStan 2.29.2 or later +when using the canonicalizer and this vignette. The minimum CmdStanR +version to run the code in the vignette is 0.5.0.

-library(cmdstanr)
-check_cmdstan_toolchain(fix = TRUE, quiet = TRUE)
+library(cmdstanr) +check_cmdstan_toolchain(fix = TRUE, quiet = TRUE)
-
-

-Deprecation warnings

-

The following logistic regression model uses several deprecated language features, resulting in several warnings during compilation.

+
+

Deprecation warnings +

+

The following logistic regression model uses several deprecated +language features, resulting in several warnings during compilation.

-stan_file <- write_stan_file("
-data {
-  int<lower=1> k;
-  int<lower=0> n;
-  matrix[n, k] X;
-  int y[n];
-}
-parameters {
-  vector[k] beta;
-  real alpha;
-}
-model {
-  # priors
-  target += std_normal_log(beta);
-  alpha ~ std_normal();
-  
-  y ~ bernoulli_logit(X * beta + alpha);
-}
-")
-mod <- cmdstan_model(stan_file)
-
Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmplsuXnS/model-840d5c7122c.stan', line 6, column 2: Declaration
+stan_file <- write_stan_file("
+data {
+  int<lower=1> k;
+  int<lower=0> n;
+  matrix[n, k] X;
+  int y[n];
+}
+parameters {
+  vector[k] beta;
+  real alpha;
+}
+model {
+  # priors
+  target += std_normal_log(beta);
+  alpha ~ std_normal();
+
+  y ~ bernoulli_logit(X * beta + alpha);
+}
+")
+mod <- cmdstan_model(stan_file)
+
Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMBUSHs/model-17e6a34f96f68.stan', line 6, column 2: Declaration
     of arrays by placing brackets after a variable name is deprecated and
-    will be removed in Stan 2.32.0. Instead use the array keyword before the
+    will be removed in Stan 2.33.0. Instead use the array keyword before the
     type. This can be changed automatically using the auto-format flag to
     stanc
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmplsuXnS/model-840d5c7122c.stan', line 13, column 2: Comments
+Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMBUSHs/model-17e6a34f96f68.stan', line 13, column 2: Comments
     beginning with # are deprecated and this syntax will be removed in Stan
-    2.32.0. Use // to begin line comments; this can be done automatically
+    2.33.0. Use // to begin line comments; this can be done automatically
     using the auto-format flag to stanc
-Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmplsuXnS/model-840d5c7122c.stan', line 14, column 12: std_normal_log
-    is deprecated and will be removed in Stan 2.32.0. Use std_normal_lpdf
+Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpMBUSHs/model-17e6a34f96f68.stan', line 14, column 12: std_normal_log
+    is deprecated and will be removed in Stan 2.33.0. Use std_normal_lpdf
     instead. This can be automatically changed using the canonicalize flag
     for stanc

The first warning is about using the deprecated array syntax

int y[n];
-

which should be replaced with the new syntax using the array keyword:

+

which should be replaced with the new syntax using the +array keyword:

array[n] int y;
-

The second warning is about using the deprecated commenting symbol #, which should be replaced by //.

-

The last warning is about the use of the deprecated _log suffix for probability density and mass functions. In this case the _log suffix should be replaced with _lpdf. For probability mass functions the suffix _lpmf is used.

-

We can go and fix these issues manually or use the canonicalizer as outlined in the next section.

+

The second warning is about using the deprecated commenting symbol +#, which should be replaced by //.

+

The last warning is about the use of the deprecated _log +suffix for probability density and mass functions. In this case the +_log suffix should be replaced with _lpdf. For +probability mass functions the suffix _lpmf is used.

+

We can go and fix these issues manually or use the canonicalizer as +outlined in the next section.

-
-

-Using the canonicalizer

-

The canonicalizer is available through the canonicalize argument of the $format() method of the CmdStanModel class. The arguments accepts TRUE and FALSE values, in which case all or none of the features of the canonicalizer are used. It can also accept a list of character vectors that determine which features of the canonicalizer to use.

-

The canonincalizer in CmdStan 2.29.2 supports four features: parentheses, braces, includes and deprecations. The parentheses and braces features clean up the use of parentheses and braces, while includes will replace #include statements with the code from the included files. See the canonicalizer section of the Stan User’s Guide for more details.

-

In this vignette we will be using the deprecations feature that replaces deprecated Stan model features with non-deprecated ones if possible.

+
+

Using the canonicalizer +

+

The canonicalizer is available through the canonicalize +argument of the $format() method of the +CmdStanModel class. The arguments accepts TRUE +and FALSE values, in which case all or none of the features +of the canonicalizer are used. It can also accept a list of character +vectors that determine which features of the canonicalizer to use.

+

The canonincalizer in CmdStan 2.29.2 supports four features: +parentheses, braces, includes and +deprecations. The parentheses and +braces features clean up the use of parentheses and braces, +while includes will replace #include +statements with the code from the included files. See the canonicalizer +section of the Stan User’s Guide for more details.

+

In this vignette we will be using the deprecations +feature that replaces deprecated Stan model features with non-deprecated +ones if possible.

-mod$format(canonicalize = list("deprecations"))
+mod$format(canonicalize = list("deprecations"))
data {
   int<lower=1> k;
   int<lower=0> n;
@@ -219,15 +252,25 @@ 

   
   y ~ bernoulli_logit(X * beta + alpha);
 }
-

By default, the format function will print the resulting model code. We can see that all three issues were resolved. y is now defined using the new array keyword, the comment uses // and the std_normal_log() is replaced with std_normal_lpdf().

-

You can also use the $format() method to write the updated version of the model directly to the Stan model file. That can be enabled by setting overwrite_file = TRUE. The previous version of the file will automatically be backed up to a file with the .stan.bak suffix. If that is not desired or you are using a version system and making a backup is redundant, you can disable it by setting backup = FALSE.

+

By default, the format function will print the resulting model code. +We can see that all three issues were resolved. y is now +defined using the new array keyword, the comment uses // +and the std_normal_log() is replaced with +std_normal_lpdf().

+

You can also use the $format() method to write the +updated version of the model directly to the Stan model file. That can +be enabled by setting overwrite_file = TRUE. The previous +version of the file will automatically be backed up to a file with the +.stan.bak suffix. If that is not desired or you are using a +version system and making a backup is redundant, you can disable it by +setting backup = FALSE.

-mod$format(
-    canonicalize = list("deprecations"),
-    overwrite_file = TRUE,
-    backup = FALSE
-)
-mod$print()
+mod$format( + canonicalize = list("deprecations"), + overwrite_file = TRUE, + backup = FALSE +) +mod$print()
data {
   int<lower=1> k;
   int<lower=0> n;
@@ -259,11 +302,13 @@ 

 
 
       

-  
Developed by Jonah Gabry, Rok Češnovar.

+  

+
Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

 

 
 

-  
Site built with pkgdown 1.6.1.

+  

+
Site built with pkgdown 2.0.7.

 

 
       

@@ -272,5 +317,7 @@ 

   
 
 
+  
+
   
 
diff --git a/docs/articles/index.html b/docs/articles/index.html
index 4daf22f4d..dfb4bb983 100644
--- a/docs/articles/index.html
+++ b/docs/articles/index.html
@@ -1,74 +1,12 @@
-
 
-
-  
-  
-
-
-
-Articles • cmdstanr
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Articles • cmdstanr
-
-
-
-  
+
+    
 
-  
     

-      

-      
 
       
 
-      

-
-

+      

   

     

       
Articles

@@ -174,50 +103,46 @@ 
Articles

 
     

       
Getting started

-      
Installing CmdStan, fitting models, and accessing results.

+      
Installing CmdStan, fitting models, and accessing results.

 
-      

-        
Getting started with CmdStanR

-        

-      

-    

+      
Getting started with CmdStanR

+        

+      

     

       
More details

-      
More information about compilation, passing in data, how CmdStan ouput is written to CSV and read back into R, profiling Stan programs, running Stan on GPUs, and using CmdStanR in R Markdown documents.

-
-      

-        
How does CmdStanR work?

-        

-        
R Markdown CmdStan Engine

-        

-        
Handling deprecated Stan features with the canonicalizer in CmdStanR

-        

-        
Profiling Stan programs with CmdStanR

-        

-        
Running Stan on the GPU with OpenCL

-        

-      

-    

+      
More information about compilation, passing in data, how CmdStan ouput is written to CSV and read back into R, profiling Stan programs, running Stan on GPUs, and using CmdStanR in R Markdown documents.

+
+      
How does CmdStanR work?

+        

+        
Working with Posteriors

+        

+        
R Markdown CmdStan Engine

+        

+        
Handling deprecated Stan features with the canonicalizer in CmdStanR

+        

+        
Profiling Stan programs with CmdStanR

+        

+        
Running Stan on the GPU with OpenCL

+        

+      

   

 

 
 
-      

-      

-  
Developed by Jonah Gabry, Rok Češnovar.

+      

+  
Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

 

 
 

-  
Site built with pkgdown 1.6.1.

+  
Site built with pkgdown 2.0.7.

 

 
-      

-   

+      

 
   
 
 
-  
-
+  
 
+  
 
diff --git a/docs/articles/posterior.html b/docs/articles/posterior.html
new file mode 100644
index 000000000..0e0439f75
--- /dev/null
+++ b/docs/articles/posterior.html
@@ -0,0 +1,309 @@
+
+
+
+
+
+
+
+Working with Posteriors • cmdstanr
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+    
+
+    

+      

+
+
+      
+
+      

+  

+    

+      
Working with Posteriors

+            
+      
+      Source: vignettes/posterior.Rmd
+      
+
+    

+
+    
+    
+
+

+
Summary
+

+
We can easily customise the summary statistics reported by
+$summary() and $print().

+
+fit <- cmdstanr::cmdstanr_example("schools", method = "sample")
+fit$summary()

+
Warning: 130 of 4000 (3.0%) transitions ended with a divergence.
+See https://mc-stan.org/misc/warnings for details.

+
   variable  mean median  sd mad     q5 q95 rhat ess_bulk ess_tail
+1      lp__ -58.9  -59.2 5.0 5.1 -66.97 -50    1      224       84
+2        mu   6.6    6.7 4.3 4.3  -0.55  14    1      394      115
+3       tau   5.8    5.0 3.7 3.4   1.33  13    1      223       92
+4  theta[1]   9.7    9.0 7.3 6.3  -0.99  23    1     1066     2034
+5  theta[2]   7.0    6.8 5.9 5.6  -2.54  16    1      900     2321
+6  theta[3]   5.5    5.8 6.9 6.1  -6.25  16    1      841     2201
+7  theta[4]   6.8    6.9 6.2 6.0  -3.13  17    1      781     2193
+8  theta[5]   4.6    5.0 6.0 5.8  -5.63  14    1      513      940
+9  theta[6]   5.5    5.7 6.3 5.6  -5.61  15    1      782     1784
+10 theta[7]   9.5    9.1 6.2 5.8   0.14  20    1      882     2164
+11 theta[8]   7.1    7.0 7.3 6.3  -4.80  18    1      976     2151

+
By default all variables are summaries with the follow functions:

+
+
[1] "mean"      "median"    "sd"        "mad"       "quantile2"

+
To change the variables summarised, we use the variables argument

+
+fit$summary(variables = c("mu", "tau"))

+
  variable mean median  sd mad    q5 q95 rhat ess_bulk ess_tail
+1       mu  6.6    6.7 4.3 4.3 -0.55  14    1      394      115
+2      tau  5.8    5.0 3.7 3.4  1.33  13    1      223       92

+
We can additionally change which functions are used

+
+fit$summary(variables = c("mu", "tau"), mean, sd)

+
  variable mean  sd
+1       mu  6.6 4.3
+2      tau  5.8 3.7

+
To summarise all variables with non-default functions, it is
+necessary to set explicitly set the variables argument, either to
+NULL or the full vector of variable names.

+
+fit$metadata()$model_params
+fit$summary(variables = NULL, "mean", "median")

+
 [1] "lp__"     "mu"       "tau"      "theta[1]" "theta[2]" "theta[3]"
+ [7] "theta[4]" "theta[5]" "theta[6]" "theta[7]" "theta[8]"

+
   variable  mean median
+1      lp__ -58.9  -59.2
+2        mu   6.6    6.7
+3       tau   5.8    5.0
+4  theta[1]   9.7    9.0
+5  theta[2]   7.0    6.8
+6  theta[3]   5.5    5.8
+7  theta[4]   6.8    6.9
+8  theta[5]   4.6    5.0
+9  theta[6]   5.5    5.7
+10 theta[7]   9.5    9.1
+11 theta[8]   7.1    7.0

+
Summary functions can be specified by character string, function, or
+using a formula (or anything else supported by [rlang::as_function]). If
+these arguments are named, those names will be used in the tibble
+output. If the summary results are named they will take precedence.

+
+my_sd <- function(x) c(My_SD = sd(x))
+fit$summary(
+  c("mu", "tau"), 
+  MEAN = mean, 
+  "median",
+  my_sd,
+  ~quantile(.x, probs = c(0.1, 0.9)),
+  Minimum = function(x) min(x)
+)        

+
  variable MEAN median My_SD  10% 90% Minimum
+1       mu  6.6    6.7   4.3 0.98  12   -11.7
+2      tau  5.8    5.0   3.7 1.81  11     0.9

+
Arguments to all summary functions can also be specified with
+.args.

+
+fit$summary(c("mu", "tau"), quantile, .args = list(probs = c(0.025, .05, .95, .975)))

+
  variable 2.5%    5% 95% 97.5%
+1       mu -2.0 -0.55  14    15
+2      tau  1.1  1.33  13    15

+
The summary functions are applied to the array of sample values, with
+dimension iter_samplingxchains.

+
+fit$summary(variables = NULL, dim, colMeans)

+
   variable dim.1 dim.2     1     2     3     4
+1      lp__  1000     4 -58.8 -58.4 -59.0 -59.4
+2        mu  1000     4   6.8   6.7   6.6   6.1
+3       tau  1000     4   5.7   5.6   5.7   6.1
+4  theta[1]  1000     4   9.9   9.5   9.8   9.5
+5  theta[2]  1000     4   7.4   7.2   7.0   6.3
+6  theta[3]  1000     4   5.8   5.7   5.6   4.8
+7  theta[4]  1000     4   6.9   6.7   7.0   6.7
+8  theta[5]  1000     4   4.9   4.8   4.6   4.1
+9  theta[6]  1000     4   5.7   5.8   5.6   4.8
+10 theta[7]  1000     4   9.6   9.8   9.4   9.2
+11 theta[8]  1000     4   7.0   7.3   7.0   7.0

+
For this reason users may have unexpected results if they use
+stats::var() directly, as it will return a covariance
+matrix. An alternative is the distributional::variance()
+function, which can also be accessed via
+posterior::variance().

+
+fit$summary(c("mu", "tau"), posterior::variance, ~var(as.vector(.x)))

+
  variable posterior::variance ~var(as.vector(.x))
+1       mu                  19                  19
+2      tau                  14                  14

+
Summary functions need not be numeric, but these won’t work with
+$print().

+
+strict_pos <- function(x) if (all(x > 0)) "yes" else "no"
+fit$summary(variables = NULL, "Strictly Positive" = strict_pos)
+# fit$print(variables = NULL, "Strictly Positive" = strict_pos)

+
   variable Strictly Positive
+1      lp__                no
+2        mu                no
+3       tau               yes
+4  theta[1]                no
+5  theta[2]                no
+6  theta[3]                no
+7  theta[4]                no
+8  theta[5]                no
+9  theta[6]                no
+10 theta[7]                no
+11 theta[8]                no

+
For more information, see posterior::summarise_draws(),
+which is called by $summary().

+

+  

+
+  
+
+

+
+
+
+      

+  

+
Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+

+
+

+  

+
Site built with pkgdown 2.0.7.

+

+
+      

+

+
+  
+
+
+  
+
+  
+
diff --git a/docs/articles/posterior_files/header-attrs-2.18/header-attrs.js b/docs/articles/posterior_files/header-attrs-2.18/header-attrs.js
new file mode 100644
index 000000000..dd57d92e0
--- /dev/null
+++ b/docs/articles/posterior_files/header-attrs-2.18/header-attrs.js
@@ -0,0 +1,12 @@
+// Pandoc 2.9 adds attributes on both header and div. We remove the former (to
+// be compatible with the behavior of Pandoc < 2.8).
+document.addEventListener('DOMContentLoaded', function(e) {
+  var hs = document.querySelectorAll("div.section[class*='level'] > :first-child");
+  var i, h, a;
+  for (i = 0; i < hs.length; i++) {
+    h = hs[i];
+    if (!/^h[1-6]$/i.test(h.tagName)) continue;  // it should be a header h1-h6
+    a = h.attributes;
+    while (a.length > 0) h.removeAttribute(a[0].name);
+  }
+});
diff --git a/docs/articles/profiling.html b/docs/articles/profiling.html
index 58cce66d3..0bb97c2aa 100644
--- a/docs/articles/profiling.html
+++ b/docs/articles/profiling.html
@@ -26,6 +26,8 @@
 
 
 
+    
+
     

       

+      

   

     

       
Profiling Stan programs with CmdStanR

-                        
Rok Češnovar, Jonah Gabry and Ben Bales

+                        
Rok Češnovar,
+Jonah Gabry and Ben Bales

             
       
-      Source: vignettes/profiling.Rmd
+      Source: vignettes/profiling.Rmd
       
 
     

 
     
     
-

-

-Introduction

-
This vignette demonstrates how to use the new profiling functionality introduced in CmdStan 2.26.0.

-
Profiling identifies which parts of a Stan program are taking the longest time to run and is therefore a useful guide when working on optimizing the performance of a model.

-
However, be aware that the statistical assumptions that go into a model are the most important factors in overall model performance. It is often not possible to make up for model problems with just brute force computation. For ideas on how to address performance of your model from a statistical perspective, see Gelman (2020).

+

+
Introduction
+

+
This vignette demonstrates how to use the new profiling functionality
+introduced in CmdStan 2.26.0.

+
Profiling identifies which parts of a Stan program are taking the
+longest time to run and is therefore a useful guide when working on
+optimizing the performance of a model.

+
However, be aware that the statistical assumptions that go into a
+model are the most important factors in overall model performance. It is
+often not possible to make up for model problems with just brute force
+computation. For ideas on how to address performance of your model from
+a statistical perspective, see Gelman (2020).

 
-library(cmdstanr)
-check_cmdstan_toolchain(fix = TRUE, quiet = TRUE)

+library(cmdstanr)
+check_cmdstan_toolchain(fix = TRUE, quiet = TRUE)
-
-

-Adding profiling statements to a Stan program

-

Consider a simple logistic regression with parameters alpha and beta, covariates X, and outcome y.

+
+

Adding profiling statements to a Stan program +

+

Consider a simple logistic regression with parameters +alpha and beta, covariates X, and +outcome y.

data {
   int<lower=1> k;
   int<lower=0> n;
   matrix[n, k] X;
-  int y[n];
+  array[n] int y;
 }
 parameters {
   vector[k] beta;
@@ -169,7 +181,9 @@ 

 
   y ~ bernoulli_logit(X * beta + alpha);
 }
-

A simple question is how much time do the prior calculations take compared against the likelihood? To answer this we surround the prior and likelihood calculations with profile statements.

+

A simple question is how much time do the prior calculations take +compared against the likelihood? To answer this we surround the prior +and likelihood calculations with profile statements.

profile("priors") {
   target += std_normal_lpdf(beta);
   target += std_normal_lpdf(alpha);
@@ -177,132 +191,163 @@ 

 profile("likelihood") {
   target += bernoulli_logit_lpmf(y | X * beta + alpha);
 }
-

In general we recommend using a separate .stan file, but for convenience in this vignette we’ll write the Stan program as a string and use write_stan_file() to write it to a temporary file.

+

In general we recommend using a separate .stan file, but +for convenience in this vignette we’ll write the Stan program as a +string and use write_stan_file() to write it to a temporary +file.

-profiling_bernoulli_logit <- write_stan_file('
-data {
-  int<lower=1> k;
-  int<lower=0> n;
-  matrix[n, k] X;
-  int y[n];
-}
-parameters {
-  vector[k] beta;
-  real alpha;
-}
-model {
-  profile("priors") {
-    target += std_normal_lpdf(beta);
-    target += std_normal_lpdf(alpha);
-  }
-  profile("likelihood") {
-    target += bernoulli_logit_lpmf(y | X * beta + alpha);
-  }
-}
-')
-

We can then run the model as usual and Stan will collect the profiling information for any sections with profile statements.

+profiling_bernoulli_logit <- write_stan_file(' +data { + int<lower=1> k; + int<lower=0> n; + matrix[n, k] X; + array[n] int y; +} +parameters { + vector[k] beta; + real alpha; +} +model { + profile("priors") { + target += std_normal_lpdf(beta); + target += std_normal_lpdf(alpha); + } + profile("likelihood") { + target += bernoulli_logit_lpmf(y | X * beta + alpha); + } +} +')
+

We can then run the model as usual and Stan will collect the +profiling information for any sections with profile +statements.

-# Compile the model
-model <- cmdstan_model(profiling_bernoulli_logit)
-
-# Generate some fake data
-n <- 1000
-k <- 20
-X <- matrix(rnorm(n * k), ncol = k)
-
-y <- 3 * X[,1] - 2 * X[,2] + 1
-p <- runif(n)
-y <- ifelse(p < (1 / (1 + exp(-y))), 1, 0)
-stan_data <- list(k = ncol(X), n = nrow(X), y = y, X = X)
-
-# Run one chain of the model
-fit <- model$sample(data = stan_data, chains = 1)
+# Compile the model +model <- cmdstan_model(profiling_bernoulli_logit) + +# Generate some fake data +n <- 1000 +k <- 20 +X <- matrix(rnorm(n * k), ncol = k) + +y <- 3 * X[,1] - 2 * X[,2] + 1 +p <- runif(n) +y <- ifelse(p < (1 / (1 + exp(-y))), 1, 0) +stan_data <- list(k = ncol(X), n = nrow(X), y = y, X = X) + +# Run one chain of the model +fit <- model$sample(data = stan_data, chains = 1)
-
-

-Accessing the profiling information from R

-

The raw profiling information can then be accessed with the $profiles() method, which returns a list containing one data frame per chain (profiles across multiple chains are not automatically aggregated). Details on the column names are available in the CmdStan documentation.

+
+

Accessing the profiling information from R +

+

The raw profiling information can then be accessed with the +$profiles() method, which returns a list containing one +data frame per chain (profiles across multiple chains are not +automatically aggregated). Details on the column names are available in +the CmdStan +documentation.

-fit$profiles()
+fit$profiles()
[[1]]
         name   thread_id total_time forward_time reverse_time chain_stack
-1 likelihood 0x107702e00 0.79556800   0.64104100   0.15452700       52380
-2     priors 0x107702e00 0.00536866   0.00314377   0.00222489       34920
+1 likelihood 0x102122e00 0.71089600   0.58314100    0.1277560       51969
+2     priors 0x102122e00 0.00482875   0.00293865    0.0018901       34646
   no_chain_stack autodiff_calls no_autodiff_calls
-1       34920000          17460                 1
-2              0          17460                 1
-

The total_time column is the total time spent inside a given profile statement. It is clear that the vast majority of time is spent in the likelihood function.

+1 34646000 17323 1 +2 0 17323 1 +

The total_time column is the total time spent inside a +given profile statement. It is clear that the vast majority of time is +spent in the likelihood function.

-
-

-Comparing to a faster version of the model

-

Stan’s specialized glm functions can be used to make models like this faster. In this case the likelihood can be replaced with

+
+

Comparing to a faster version of the model +

+

Stan’s specialized glm functions can be used to make models like this +faster. In this case the likelihood can be replaced with

target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);
-

We’ll keep the same profile() statements so that the profiling information for the new model is collected automatically just like for the previous one.

+

We’ll keep the same profile() statements so that the +profiling information for the new model is collected automatically just +like for the previous one.

-profiling_bernoulli_logit_glm <- write_stan_file('
-data {
-  int<lower=1> k;
-  int<lower=0> n;
-  matrix[n, k] X;
-  int y[n];
-}
-parameters {
-  vector[k] beta;
-  real alpha;
-}
-model {
-  profile("priors") {
-    target += std_normal_lpdf(beta);
-    target += std_normal_lpdf(alpha);
-  }
-  profile("likelihood") {
-    target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);
-  }
-}
-')
+profiling_bernoulli_logit_glm <- write_stan_file(' +data { + int<lower=1> k; + int<lower=0> n; + matrix[n, k] X; + array[n] int y; +} +parameters { + vector[k] beta; + real alpha; +} +model { + profile("priors") { + target += std_normal_lpdf(beta); + target += std_normal_lpdf(alpha); + } + profile("likelihood") { + target += bernoulli_logit_glm_lpmf(y | X, alpha, beta); + } +} +')
-model_glm <- cmdstan_model(profiling_bernoulli_logit_glm)
-fit_glm <- model_glm$sample(data = stan_data, chains = 1)
+model_glm <- cmdstan_model(profiling_bernoulli_logit_glm) +fit_glm <- model_glm$sample(data = stan_data, chains = 1)
-fit_glm$profiles()
+fit_glm$profiles()
[[1]]
         name   thread_id total_time forward_time reverse_time chain_stack
-1 likelihood 0x10550ce00 0.44504100   0.44354000   0.00150062       17161
-2     priors 0x10550ce00 0.00457269   0.00296803   0.00160466       34322
+1 likelihood 0x1066bde00 0.45516500   0.45357200   0.00159287       17695
+2     priors 0x1066bde00 0.00399743   0.00242302   0.00157441       35390
   no_chain_stack autodiff_calls no_autodiff_calls
-1              0          17161                 1
-2              0          17161                 1
-

We can see from the total_time column that this is much faster than the previous model.

+1 0 17695 1 +2 0 17695 1 +

We can see from the total_time column that this is much +faster than the previous model.

-
-

-Per-gradient timings, and memory usage

-

The other columns of the profiling output are documented in the CmdStan documentation.

-

The timing numbers are broken down by forward pass and reverse pass, and the chain_stack and no_chain_stack columns contain information about how many autodiff variables were saved in the process of performing a calculation.

-

These numbers are all totals – times are the total times over the whole calculation, and chain_stack counts are similarly the total counts of autodiff variables used over the whole calculation. It is often convenient to have per-gradient calculations (which will be more stable across runs with different seeds). To compute these, use the autodiff_calls column.

+
+

Per-gradient timings, and memory usage +

+

The other columns of the profiling output are documented in the CmdStan +documentation.

+

The timing numbers are broken down by forward pass and reverse pass, +and the chain_stack and no_chain_stack columns +contain information about how many autodiff variables were saved in the +process of performing a calculation.

+

These numbers are all totals – times are the total times over the +whole calculation, and chain_stack counts are similarly the +total counts of autodiff variables used over the whole calculation. It +is often convenient to have per-gradient calculations (which will be +more stable across runs with different seeds). To compute these, use the +autodiff_calls column.

-profile_chain_1 <- fit$profiles()[[1]]
-per_gradient_timing <- profile_chain_1$total_time/profile_chain_1$autodiff_calls
-print(per_gradient_timing) # two elements for the two profile statements in the model
-
[1] 4.556518e-05 3.074834e-07
+profile_chain_1 <- fit$profiles()[[1]] +per_gradient_timing <- profile_chain_1$total_time/profile_chain_1$autodiff_calls +print(per_gradient_timing) # two elements for the two profile statements in the model
+
[1] 4.103770e-05 2.787479e-07
-
-

-Accessing and saving the profile files

-

After sampling (or optimization or variational inference) finishes, CmdStan stores the profiling data in CSV files in a temporary location. The paths of the profiling CSV files can be retrieved using $profile_files().

+
+

Accessing and saving the profile files +

+

After sampling (or optimization or variational inference) finishes, +CmdStan stores the profiling data in CSV files in a temporary location. +The paths of the profiling CSV files can be retrieved using +$profile_files().

-fit$profile_files()
-
[1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpsMj0TM/model_20fabfb2aee52e6a18d30460e35ae184-profile-202203181228-1-107daa.csv"
-

These can be saved to a more permanent location with the $save_profile_files() method.

+fit$profile_files()
+
[1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmppMVUoY/model_6580008f67848265f3dfd0e7ae3b0600-profile-202307251456-1-5100ca.csv"
+

These can be saved to a more permanent location with the +$save_profile_files() method.

-# see ?save_profile_files for info on optional arguments
-fit$save_profile_files(dir = "path/to/directory")
+# see ?save_profile_files for info on optional arguments +fit$save_profile_files(dir = "path/to/directory")
-
-

-References

-

Gelman, Andrew, Aki Vehtari, Daniel Simpson, Charles C. Margossian, Bob Carpenter, Yuling Yao, Lauren Kennedy, Jonah Gabry, Paul-Christian Bürkner, and Martin Modrák. 2020. “Bayesian Workflow.” https://arxiv.org/abs/2011.01808.

+
+

References +

+

Gelman, Andrew, Aki Vehtari, Daniel Simpson, Charles C. Margossian, +Bob Carpenter, Yuling Yao, Lauren Kennedy, Jonah Gabry, Paul-Christian +Bürkner, and Martin Modrák. 2020. “Bayesian Workflow.” https://arxiv.org/abs/2011.01808.

@@ -317,11 +362,13 @@

-

Developed by Jonah Gabry, Rok Češnovar.

+

+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

+

Site built with pkgdown 2.0.7.

@@ -330,5 +377,7 @@

+ + diff --git a/docs/articles/r-markdown.html b/docs/articles/r-markdown.html index fe246b59b..eeb298b19 100644 --- a/docs/articles/r-markdown.html +++ b/docs/articles/r-markdown.html @@ -26,6 +26,8 @@ + +
+

R Markdown CmdStan Engine

-

Mikhail Popov

+

Mikhail +Popov

- Source: vignettes/r-markdown.Rmd + Source: vignettes/r-markdown.Rmd
-

R Markdown supports a variety of languages through the use of knitr language engines. One such engine is the stan engine, which allows users to write Stan programs directly in their R Markdown documents by setting the language of the chunk to stan.

-

Behind the scenes, the engine relies on RStan to compile the model code into an in-memory stanmodel, which is assigned to a variable with the name given by the output.var chunk option. For example:

-
```{stan, output.var="model"}
-// Stan model code
-```
-
-```{r}
-rstan::sampling(model)
-```
-

CmdStanR provides a replacement engine, which can be registered as follows:

+

R Markdown supports a variety of languages through the use of knitr +language engines. One such engine is the stan engine, which +allows users to write Stan programs directly in their R Markdown +documents by setting the language of the chunk to stan.

+

Behind the scenes, the engine relies on RStan to compile the model +code into an in-memory stanmodel, which is assigned to a +variable with the name given by the output.var chunk +option. For example:

+
```{stan, output.var="model"}
+// Stan model code
+```
+
+```{r}
+rstan::sampling(model)
+```
+

CmdStanR provides a replacement engine, which can be registered as +follows:

-

By default, this overrides knitr’s built-in stan engine so that all stan chunks are processed with CmdStanR, not RStan. Of course, this also means that the variable specified by output.var will no longer be a stanmodel object, but instead a CmdStanModel object, so the code above would look like this:

-
```{stan, output.var="model"}
-// Stan model code
-```
-
-```{r}
-model$sample()
-```
-
-

-Example

-
// This stan chunk results in a CmdStanModel object called "ex1"
-parameters {
-  array[2] real y;
-}
-model {
-  y[1] ~ normal(0, 1);
-  y[2] ~ double_exponential(0, 2);
-}
+library(cmdstanr) +check_cmdstan_toolchain(fix = TRUE, quiet = TRUE) + +register_knitr_engine()
+

By default, this overrides knitr’s built-in stan engine +so that all stan chunks are processed with CmdStanR, not +RStan. Of course, this also means that the variable specified by +output.var will no longer be a stanmodel +object, but instead a CmdStanModel object, so the code +above would look like this:

+
```{stan, output.var="model"}
+// Stan model code
+```
+
+```{r}
+model$sample()
+```
+
+

Example +

+
// This stan chunk results in a CmdStanModel object called "ex1"
+parameters {
+  array[2] real y;
+}
+model {
+  y[1] ~ normal(0, 1);
+  y[2] ~ double_exponential(0, 2);
+}
-ex1$print()
-#> // This stan chunk results in a CmdStanModel object called "ex1"
-#> parameters {
-#>   array[2] real y;
-#> }
-#> model {
-#>   y[1] ~ normal(0, 1);
-#>   y[2] ~ double_exponential(0, 2);
-#> }
+ex1$print() +#> // This stan chunk results in a CmdStanModel object called "ex1" +#> parameters { +#> array[2] real y; +#> } +#> model { +#> y[1] ~ normal(0, 1); +#> y[2] ~ double_exponential(0, 2); +#> }
-fit <- ex1$sample(
-  refresh = 0,
-  seed = 42L
-)
-#> Running MCMC with 4 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> Chain 3 finished in 0.0 seconds.
-#> Chain 4 finished in 0.0 seconds.
-#> 
-#> All 4 chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.6 seconds.
-
-print(fit)
-#>  variable  mean median   sd  mad    q5   q95 rhat ess_bulk ess_tail
-#>      lp__ -1.50  -1.17 1.24 0.96 -3.94 -0.18 1.00     1304     1536
-#>      y[1] -0.01  -0.01 0.99 0.99 -1.67  1.60 1.00     1993     2262
-#>      y[2] -0.07  -0.04 2.90 2.05 -4.79  4.54 1.00     2050     1420
+fit <- ex1$sample( + refresh = 0, + seed = 42L +) +#> Running MCMC with 4 sequential chains... +#> +#> Chain 1 finished in 0.0 seconds. +#> Chain 2 finished in 0.0 seconds. +#> Chain 3 finished in 0.0 seconds. +#> Chain 4 finished in 0.0 seconds. +#> +#> All 4 chains finished successfully. +#> Mean chain execution time: 0.0 seconds. +#> Total execution time: 0.7 seconds. + +print(fit) +#> variable mean median sd mad q5 q95 rhat ess_bulk ess_tail +#> lp__ -1.50 -1.17 1.24 0.96 -3.94 -0.18 1.00 1304 1536 +#> y[1] -0.01 -0.01 0.99 0.99 -1.67 1.60 1.00 1993 2262 +#> y[2] -0.07 -0.04 2.90 2.05 -4.79 4.54 1.00 2050 1420
-
-

-Caching chunks

-

Use cache=TRUE chunk option to avoid re-compiling the Stan model code every time the R Markdown is knit/rendered.

-

You can find the Stan model file and the compiled executable in the document’s cache directory.

+
+

Caching chunks +

+

Use cache=TRUE chunk option to avoid re-compiling the +Stan model code every time the R Markdown is knit/rendered.

+

You can find the Stan model file and the compiled executable in the +document’s cache directory.

-
-

-Using RStan and CmdStanR engines side-by-side

-

While the default behavior is to override the built-in stan engine because the assumption is that the user is probably not using both RStan and CmdStanR in the same document or project, the option to use both exists. When registering CmdStanR’s knitr engine, set override = FALSE to register the engine as a cmdstan engine:

+
+

Using RStan and CmdStanR engines side-by-side +

+

While the default behavior is to override the built-in +stan engine because the assumption is that the user is +probably not using both RStan and CmdStanR in the same document or +project, the option to use both exists. When registering CmdStanR’s +knitr engine, set override = FALSE to register the engine +as a cmdstan engine:

-register_knitr_engine(override = FALSE)
-

This will cause stan chunks to be processed by knitr’s built-in, RStan-based engine and only use CmdStanR’s knitr engine for cmdstan chunks:

-
```{stan, output.var="model_obj1"}
-// Results in a stanmodel object from RStan
-```
-
-```{r}
-rstan::sampling(model_obj1)
-```
-
-```{cmdstan, output.var="model_obj2"}
-// Results in a CmdStanModel object from CmdStanR
-```
-
-```{r}
-model_obj2$sample()
-```
+register_knitr_engine(override = FALSE)
+

This will cause stan chunks to be processed by knitr’s +built-in, RStan-based engine and only use CmdStanR’s knitr engine for +cmdstan chunks:

+
```{stan, output.var="model_obj1"}
+// Results in a stanmodel object from RStan
+```
+
+```{r}
+rstan::sampling(model_obj1)
+```
+
+```{cmdstan, output.var="model_obj2"}
+// Results in a CmdStanModel object from CmdStanR
+```
+
+```{r}
+model_obj2$sample()
+```
-
-

-Running interactively

-

When running chunks interactively in RStudio (e.g. when using R Notebooks), it has been observed that the built-in, RStan-based engine is used for stan chunks even when CmdStanR’s engine has been registered in the session as the engine for stan. As a workaround, when running chunks interactively, it is recommended to use the override = FALSE option and change stan chunks to be cmdstan chunks.

-

Do not worry: if the template you use supports syntax highlighting for the Stan language, that syntax highlighting will be applied to cmdstan chunks when the document is knit/rendered.

+
+

Running interactively +

+

When running chunks interactively in RStudio (e.g. when using R +Notebooks), it has been observed that the built-in, RStan-based +engine is used for stan chunks even when CmdStanR’s engine +has been registered in the session as the engine for stan. +As a workaround, when running chunks interactively, it is +recommended to use the override = FALSE option and change +stan chunks to be cmdstan chunks.

+

Do not worry: if the template you use supports syntax highlighting +for the Stan language, that syntax highlighting will be applied to +cmdstan chunks when the document is knit/rendered.

@@ -253,11 +285,13 @@

-

Developed by Jonah Gabry, Rok Češnovar.

+

+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

+

Site built with pkgdown 2.0.7.

@@ -266,5 +300,7 @@

+ + diff --git a/docs/authors.html b/docs/authors.html index c44f2bc06..be12ad2f9 100644 --- a/docs/authors.html +++ b/docs/authors.html @@ -1,74 +1,12 @@ - - - - - - - -Authors • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Authors and Citation • cmdstanr - - + + - - - -
-
-
- -
+
-
-

Authors

+
+
+

Authors

+
+ + +
  • +

    Jonah Gabry. Author, maintainer. +

    +
    • +
  • +

    Rok Češnovar. Author. +

    +
    • +
  • +

    Andrew Johnson. Author. +

    +
    • +
  • +

    Ben Bales. Contributor. +

    +
    • +
  • +

    Mitzi Morris. Contributor. +

    +
    • +
  • +

    Mikhail Popov. Contributor. +

    +
    • +
  • +

    Mike Lawrence. Contributor. +

    +
    • +
  • +

    William Michael Landau. Contributor. +

    +
    • +
  • +

    Jacob Socolar. Contributor. +

    +
    • +
+
+
+

Citation

+ Source: DESCRIPTION +
-
    -
  • -

    Jonah Gabry. Author, maintainer. -

    -
    • -
  • -

    Rok Češnovar. Author. -

    -
    • -
  • -

    Ben Bales. Contributor. -

    -
    • -
  • -

    Mitzi Morris. Contributor. -

    -
    • -
  • -

    Mikhail Popov. Contributor. -

    -
    • -
  • -

    Mike Lawrence. Contributor. -

    -
    • -
  • -

    William Michael Landau. Contributor. -

    -
    • -
  • -

    Jacob Socolar. Contributor. -

    -
    • -
+ +

Gabry J, Češnovar R, Johnson A (2023). +cmdstanr: R Interface to 'CmdStan'. +https://mc-stan.org/cmdstanr/, https://discourse.mc-stan.org. +

+ 
@Manual{,
+  title = {cmdstanr: R Interface to 'CmdStan'},
+  author = {Jonah Gabry and Rok Češnovar and Andrew Johnson},
+  year = {2023},
+  note = {https://mc-stan.org/cmdstanr/, https://discourse.mc-stan.org},
+}
@@ -213,22 +165,20 @@

Authors

-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/index.html b/docs/index.html index 079d4d868..c14b85fd5 100644 --- a/docs/index.html +++ b/docs/index.html @@ -19,11 +19,11 @@ + The CmdStanR interface is an alternative to RStan that calls the command + line interface for compilation and running algorithms instead of interfacing + with C++ via Rcpp. This has many benefits including always being compatible + with the latest version of Stan, fewer installation errors, fewer unexpected + crashes in RStudio, and a more permissive license."> + +
-
-

-CmdStanR +
+

CmdStanR

-
-

-Overview

-

CmdStanR is a lightweight interface to Stan for R users (see CmdStanPy for Python).

+
+

Overview +

+

CmdStanR is a lightweight interface to Stan for R users (see CmdStanPy for Python).

If you are new to CmdStanR we recommend starting with these vignettes:

-
-

-Goals

+
+

Goals +

  • A clean interface to Stan services so that CmdStanR can keep up with Stan releases.

  • R code that doesn’t interface directly with C++, only calls compiled executables.

  • Modularity: CmdStanR runs Stan’s algorithms and lets downstream modules do the analysis.

    • -

  • Flexible BSD-3 license.

    • +

  • Flexible BSD-3 license.

-
-

-Installation

-
-

-Installing the R package

+
+

Installation +

+
+

Installing the R package +

You can install the latest beta release of the cmdstanr R package with

-# we recommend running this is a fresh R session or restarting your current session
-install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
+# we recommend running this is a fresh R session or restarting your current session +install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))

This does not install the vignettes, which take a long time to build, but they are always available online at https://mc-stan.org/cmdstanr/articles/.

To instead install the latest development version of the package from GitHub use

-# install.packages("remotes")
-remotes::install_github("stan-dev/cmdstanr")
+# install.packages("remotes") +remotes::install_github("stan-dev/cmdstanr")
-
-

-Installing CmdStan

+
+

Installing CmdStan +

If you don’t already have CmdStan installed then, in addition to installing the R package, it is also necessary to install CmdStan using CmdStanR’s install_cmdstan() function. A suitable C++ toolchain is also required. Instructions are provided in the Getting started with CmdStanR vignette.

-
-

-Contributing

+
+

Contributing +

There is a lot of work still to be done and contributions are very welcome! If you are interested in contributing please comment on an open issue or open a new one if none are applicable.

-
-

-License

+
+

License +

CmdStanR, like CmdStan and the core Stan C++ code, is licensed under the following licenses:

@@ -199,50 +200,54 @@

-

Developed by Jonah Gabry, Rok Češnovar.

+

+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

+

Site built with pkgdown 2.0.7.

@@ -251,5 +256,7 @@

Dev status

+ + diff --git a/docs/news/index.html b/docs/news/index.html index 6eb05a3af..2e82f3765 100644 --- a/docs/news/index.html +++ b/docs/news/index.html @@ -1,74 +1,12 @@ - - - - - - - -Changelog • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Changelog • cmdstanr - - + + - - -
-
- -
- -
+

Changelog

- Source: NEWS.md + Source: NEWS.md
-
-

-cmdstanr 0.5.2

-
    -

  • Refactored toolchain installation and checks for R 4.x on Windows and added support for Rtools42. (#645)

    • -

  • Expanded the use of CMDSTAN environment variable to point to CmdStan installation or directory containing CmdStan installations. (#643)

    • -

  • New vignette on how to handle deprecations using the $format() method. (#644)

    • -
+
+

cmdstanr 0.6.0

+
+

Major new features

+
+
+

Other changes

+
-
-

-cmdstanr 0.5.1

-
    -
  • Temporarily disable format="draws_rvars" in the $draws() method due to a bug. Until this is fixed users can make use of posterior::as_draws_rvars() to convert draws from CmdStanR to the draws_rvars format. (#640)
    • -
+
+

cmdstanr 0.5.3

+
+

New features

+
  • On Windows, users can now install and use CmdStan with WSL (Windows Subsystem for Linux). Set wsl=TRUE in install_cmdstan() to install CmdStan for use with WSL. This can offer significant speedups compared to native Windows execution. (#677, @andrjohns)
    • +
+
+

Bug fixes

+

  • In cmdstan_default_path() we now ignore directories inside .cmdstan that don’t start with "cmdstan-". (#651)

    • +

  • Fixed Windows issue related to not locating grep.exe or when it is located in a path with spaces. (@weshinsley, #661, #663)

    • +

  • Fixed a bug with diagnostic checks when ebfmi is NaN.

    • +

  • Fixed a bug that caused issues when using ~ or . in paths supplied to the cmdstanr_write_stan_file_dir global option.

    • +

  • Fixed a bug that caused the time() method fail when some of the chains failed to finish succesfully.

    • +
-
-

-cmdstanr 0.5.0

-
-

-Bug fixes

-
    -

  • Fixed bug that caused stdour/stderr not being read at the end of optimization. (#522)

    • -

  • Fixed issue with handling NA as the reported external process status. (#544, @myshkin)

    • +
    +

    cmdstanr 0.5.2

    +

    • Refactored toolchain installation and checks for R 4.x on Windows and added support for Rtools42. (#645)

      • +

    • Expanded the use of CMDSTAN environment variable to point to CmdStan installation or directory containing CmdStan installations. (#643)

      • +

    • New vignette on how to handle deprecations using the $format() method. (#644)

      • +
    +
    +

    cmdstanr 0.5.1

    +
    • Temporarily disable format="draws_rvars" in the $draws() method due to a bug. Until this is fixed users can make use of posterior::as_draws_rvars() to convert draws from CmdStanR to the draws_rvars format. (#640)
      • +
    +
    +

    cmdstanr 0.5.0

    +
    +

    Bug fixes

    +

    • Fixed bug that caused stdour/stderr not being read at the end of optimization. (#522)

      • +

    • Fixed issue with handling NA as the reported external process status. (#544, @myshkin)

    • Fixed issue with handling models with no parameters and CmdStan 2.27+.

      • -
    -
    -
    -

    -New features

    -
      -

    • Default directory changed to .cmdstan instead of .cmdstanr so that CmdStanPy and CmdStanR can use the same CmdStan installations. Using .cmdstanr will continue to be supported until version 1.0 but install_cmdstan() will now default to .cmdstan and CmdStanR will first look for .cmdstan before falling back on .cmdstanr. (#454)

      • -

    • New method diagnose() for CmdstanModel objects exposes CmdStan’s diagnose method for comparing Stan’s gradient computations to gradients computed via finite differences. (#485)

      • -

    • New method $variables() for CmdstanModel objects that returns a list of variables in the Stan model, their types and number of dimensions. Does not require the model to be compiled. (#519)

      • -

    • New method $format() for auto-formatting and canonicalizing the Stan models. (#625)

      • -

    • Added the option to create CmdStanModel from the executable only with the exe_file argument. (#564)

      • +
    +
    +

    New features

    +

    • Default directory changed to .cmdstan instead of .cmdstanr so that CmdStanPy and CmdStanR can use the same CmdStan installations. Using .cmdstanr will continue to be supported until version 1.0 but install_cmdstan() will now default to .cmdstan and CmdStanR will first look for .cmdstan before falling back on .cmdstanr. (#454)

      • +

    • New method diagnose() for CmdstanModel objects exposes CmdStan’s diagnose method for comparing Stan’s gradient computations to gradients computed via finite differences. (#485)

      • +

    • New method $variables() for CmdstanModel objects that returns a list of variables in the Stan model, their types and number of dimensions. Does not require the model to be compiled. (#519)

      • +

    • New method $format() for auto-formatting and canonicalizing the Stan models. (#625)

      • +

    • Added the option to create CmdStanModel from the executable only with the exe_file argument. (#564)

    • Added a convenience argument user_header to $compile() and cmdstan_model() that simplifies the use of an external .hpp file to compile with the model.

      • -

    • Added the cmdstanr_force_recompile global option that is used for forcing recompilation of Stan models. (#580)

      • -

    • New method $code() for all fitted model objects that returns the Stan code associated with the fitted model. (#575)

      • -

    • New method $diagnostic_summary() for CmdStanMCMC objects that summarizes the sampler diagnostics (divergences, treedepth, ebfmi) and can regenerate the related warning messages. (#205)

      • -

    • New diagnostics argument for the $sample() method to specify which diagnostics are checked after sampling. Replaces validate_csv argument. (#205)

      • -

    • Added E-BFMI checks that run automatically post sampling. (#500, @jsocolar)

      • -

    • New methods for posterior::as_draws() for CmdStanR fitted model objects. These are just wrappers around the $draws() method provided for convenience. (#532)

      • -

    • write_stan_file() now choose file names deterministically based on the code so that models do not get unnecessarily recompiled when calling the function multiple times with the same code. (#495, @martinmodrak)

      • -

    • The dir argument for write_stan_file() can now be set with a global option. (#537)

      • -

    • write_stan_json() now handles data of class "table". Tables are converted to vector, matrix, or array depending on the dimensions of the table. (#528)

      • -

    • Improved processing of named lists supplied to the data argument to JSON data files: checking whether the list includes all required elements/Stan variables; improved differentiating arrays/vectors of length 1 and scalars when generating JSON data files; generating floating point numbers with decimal points to fix issue with parsing large numbers. (#538)

      • -

    • install_cmdstan() now automatically installs the Linux ARM CmdStan when Linux distributions running on ARM CPUs are detected. (#531)

      • -

    • New function as_mcmc.list() for converting CmdStanMCMC objects to mcmc.list objects from the coda package. (#584, @MatsuuraKentaro)

      • -
    -
    -
    -
    -

    -cmdstanr 0.4.0

    -
    -

    -Bug fixes

    -
      -

    • Fixed issue with retrieving draws with models with spaces in their names. (#453)

      • -

    • Fixed bug with spaces in path to the temporary folder on Windows. (#460)

      • -

    • Fixed issue with not reporting model executable name clashing with folder name. (#461)

      • -
    -
    -
    -

    -New features

    -
      -

    • New function as_cmdstan_fit() that creates CmdStanMCMC/MLE/VB objects directly from CmdStan CSV files. (#412)

      • -

    • read_cmdstan_csv() now also returns chain run times for MCMC sampling CSV files. (#414)

      • -

    • Faster CSV reading for multiple chains. (#419)

      • -

    • New $profiles() method for fitted model objects accesses profiling information from R if profiling used in the Stan program. Support for profiling Stan programs requires CmdStan >= 2.26. (#434)

      • -

    • New vignette on profiling Stan programs. (#435)

      • -

    • New vignette on running Stan on the GPU with OpenCL. OpenCL device ids can now also be specified at runtime. (#439)

      • -

    • New check for invalid parameter names when supplying init values. (#452, @mike-lawrence)

      • -

    • Suppressing compilation messages when not in interactive mode. (#462, @wlandau)

      • -

    • New error_on_NA argument for cmdstan_version() to optionally return NULL (instead of erroring) if the CmdStan path is not found (#467, @wlandau).

      • -

    • Global option cmdstanr_max_rows can be set as an alternative to specifying max_rows argument to the $print() method. (#470)

      • -

    • New output_basename argument for the model fitting methods. Can be used in conjunction with output_dir to get completely predictable output CSV file paths. (#471)

      • -

    • New format argument for $draws(), $sampler_diagnostics(), read_cmdstan_csv(), and as_cmdstan_fit(). This controls the format of the draws returned or stored in the object. Changing the format can improve speed and memory usage for large models. (#482)

      • -
    -
    -
    -
    -

    -cmdstanr 0.3.0

    -
    -

    -Bug fixes

    -
      -

    • Fixed reading inverse mass matrix with values written in scientific format in the CSV. (#394)

      • -

    • Fixed error caused by an empty data list. Previously if a model didn’t require data then data had to either be NULL or be a non-empty list, but now list() is allowed. (#403)

      • -
    -
    -
    -

    -New features

    -
      -

    • Added $sample_mpi() for MCMC sampling with MPI. (#350)

      • -

    • Added informative messages on compile errors caused by precompiled headers (PCH). (#384)

      • -

    • Added the cmdstanr_verbose option for verbose mode. Intended for troubleshooting, debugging and development. See end of How does CmdStanR work? vignette for details. (#392)

      • -

    • New $loo() method for CmdStanMCMC objects. Requires computing pointwise log-likelihood in Stan program. (#366)

      • -

    • The fitted_params argument to the $generate_quantities() method now also accepts CmdStanVB, posterior::draws_array, and posterior::draws_matrix objects. (#390)

      • -

    • The $optimize() method now supports all of CmdStan’s tolerance-related arguments for (L)BFGS. (#398)

      • -

    • The documentation for the R6 methods now uses @param, which allows package developers to import the CmdStanR documentation using roxygen2’s @inheritParams. (#408)

      • -
    -
    +

  • Added the cmdstanr_force_recompile global option that is used for forcing recompilation of Stan models. (#580)

    • +

  • New method $code() for all fitted model objects that returns the Stan code associated with the fitted model. (#575)

    • +

  • New method $diagnostic_summary() for CmdStanMCMC objects that summarizes the sampler diagnostics (divergences, treedepth, ebfmi) and can regenerate the related warning messages. (#205)

    • +

  • New diagnostics argument for the $sample() method to specify which diagnostics are checked after sampling. Replaces validate_csv argument. (#205)

    • +

  • Added E-BFMI checks that run automatically post sampling. (#500, @jsocolar)

    • +

  • New methods for posterior::as_draws() for CmdStanR fitted model objects. These are just wrappers around the $draws() method provided for convenience. (#532)

    • +

  • write_stan_file() now choose file names deterministically based on the code so that models do not get unnecessarily recompiled when calling the function multiple times with the same code. (#495, @martinmodrak)

    • +

  • The dir argument for write_stan_file() can now be set with a global option. (#537)

    • +

  • write_stan_json() now handles data of class "table". Tables are converted to vector, matrix, or array depending on the dimensions of the table. (#528)

    • +

  • Improved processing of named lists supplied to the data argument to JSON data files: checking whether the list includes all required elements/Stan variables; improved differentiating arrays/vectors of length 1 and scalars when generating JSON data files; generating floating point numbers with decimal points to fix issue with parsing large numbers. (#538)

    • +

  • install_cmdstan() now automatically installs the Linux ARM CmdStan when Linux distributions running on ARM CPUs are detected. (#531)

    • +

  • New function as_mcmc.list() for converting CmdStanMCMC objects to mcmc.list objects from the coda package. (#584, @MatsuuraKentaro)

    • +
-
-

-cmdstanr 0.2.2

-
-

-Bug fixes

-
    -

  • Fixed bug with reading Stan CSV when grep used coloring by default (#364,#371)

    • -

  • Depend on posterior v0.1.3 to avoid a potential error in $summary(). (#383)

    • -
+
+

cmdstanr 0.4.0

+
+

Bug fixes

+

  • Fixed issue with retrieving draws with models with spaces in their names. (#453)

    • +

  • Fixed bug with spaces in path to the temporary folder on Windows. (#460)

    • +

  • Fixed issue with not reporting model executable name clashing with folder name. (#461)

    • +
+
+

New features

+

  • New function as_cmdstan_fit() that creates CmdStanMCMC/MLE/VB objects directly from CmdStan CSV files. (#412)

    • +

  • read_cmdstan_csv() now also returns chain run times for MCMC sampling CSV files. (#414)

    • +

  • Faster CSV reading for multiple chains. (#419)

    • +

  • New $profiles() method for fitted model objects accesses profiling information from R if profiling used in the Stan program. Support for profiling Stan programs requires CmdStan >= 2.26. (#434)

    • +

  • New vignette on profiling Stan programs. (#435)

    • +

  • New vignette on running Stan on the GPU with OpenCL. OpenCL device ids can now also be specified at runtime. (#439)

    • +

  • New check for invalid parameter names when supplying init values. (#452, @mike-lawrence)

    • +

  • Suppressing compilation messages when not in interactive mode. (#462, @wlandau)

    • +

  • New error_on_NA argument for cmdstan_version() to optionally return NULL (instead of erroring) if the CmdStan path is not found (#467, @wlandau).

    • +

  • Global option cmdstanr_max_rows can be set as an alternative to specifying max_rows argument to the $print() method. (#470)

    • +

  • New output_basename argument for the model fitting methods. Can be used in conjunction with output_dir to get completely predictable output CSV file paths. (#471)

    • +

  • New format argument for $draws(), $sampler_diagnostics(), read_cmdstan_csv(), and as_cmdstan_fit(). This controls the format of the draws returned or stored in the object. Changing the format can improve speed and memory usage for large models. (#482)

    • +
-
-

-New features

-
    -

  • Added support for native execution on the macOS with the M1 ARM-based CPU. (#375)

    • -

  • Added threading support via threads argument for $optimize() and $variational() (was already available via threads_per_chain for $sample()). (#369)

    • -
+
+

cmdstanr 0.3.0

+
+

Bug fixes

+

  • Fixed reading inverse mass matrix with values written in scientific format in the CSV. (#394)

    • +

  • Fixed error caused by an empty data list. Previously if a model didn’t require data then data had to either be NULL or be a non-empty list, but now list() is allowed. (#403)

    • +
+
+

New features

+

  • Added $sample_mpi() for MCMC sampling with MPI. (#350)

    • +

  • Added informative messages on compile errors caused by precompiled headers (PCH). (#384)

    • +

  • Added the cmdstanr_verbose option for verbose mode. Intended for troubleshooting, debugging and development. See end of How does CmdStanR work? vignette for details. (#392)

    • +

  • New $loo() method for CmdStanMCMC objects. Requires computing pointwise log-likelihood in Stan program. (#366)

    • +

  • The fitted_params argument to the $generate_quantities() method now also accepts CmdStanVB, posterior::draws_array, and posterior::draws_matrix objects. (#390)

    • +

  • The $optimize() method now supports all of CmdStan’s tolerance-related arguments for (L)BFGS. (#398)

    • +

  • The documentation for the R6 methods now uses @param, which allows package developers to import the CmdStanR documentation using roxygen2’s @inheritParams. (#408)

    • +
+
+

cmdstanr 0.2.2

+
+

Bug fixes

+

  • Fixed bug with reading Stan CSV when grep used coloring by default (#364,#371)

    • +

  • Depend on posterior v0.1.3 to avoid a potential error in $summary(). (#383)

    • +
+
+

New features

+

  • Added support for native execution on the macOS with the M1 ARM-based CPU. (#375)

    • +

  • Added threading support via threads argument for $optimize() and $variational() (was already available via threads_per_chain for $sample()). (#369)

    • +
-
-

-cmdstanr 0.2.1

-
-

-Bug fixes

- +
+

cmdstanr 0.2.1

+
+

Bug fixes

+
+
+

New features

+
  • +compile() and check_syntax() methods gain argument pedantic for turning on pedantic mode, which warns about issues with the model beyond syntax errors. (#361)
    • +
-
-

-New features

-
    -
  • -compile() and check_syntax() methods gain argument pedantic for turning on pedantic mode, which warns about issues with the model beyond syntax errors. (#361)
    • -
-
-
-
-

-cmdstanr 0.2.0

-
-

-Bug fixes

-
    -

  • Fix potential indexing error if using read_cmdstan_csv() with CSV files created by CmdStan without CmdStanR. (#291, #292, @johnlees)

    • -

  • Fix error when returning draws or sampler diagnostics for a fit with only warmup and no samples. (#288, #293)

    • -

  • Fix trailing slashes issue for dir in cmdstan_model() and output_dir in fitting methods. (#281, #294)

    • -

  • Fix dimensions error when processing a list of matrices passed in as data. (#296, #302)

    • -

  • Fix reporting of time after using fixed_param method. (#303, #307)

    • -

  • With refresh = 0, no output other than error messages is printed with $optimize() and $variational(). (#324)

    • -

  • Fix issue where names of generated files could clash. (#326, #328)

    • -

  • Fix missing include_paths in $syntax_check(). (#335, @mike-lawrence)

    • -
-
-
-

-New features

-
    -

  • CSV reading is now faster by using data.table::fread(). (#318)

    • -

  • install_cmdstan() gains argument version for specifying which version of CmdStan to install. (#300, #308)

    • -

  • New function check_cmdstan_toolchain() that checks if the appropriate toolchains are available. (#289)

    • -

  • $sample() method for CmdStanModel objects gains argument chain_ids for specifying custom chain IDs. (#319)

    • -

  • Added support for the sig_figs argument in CmdStan versions 2.25 and above. (#327)

    • -

  • Added checks if the user has the necessary permissions in the RTools and temporary folders. (#343)

    • -
-
-
-
-

-cmdstanr 0.1.3

-
    -
  • New $check_syntax() method for CmdStanModel objects. (#276, #277)
    • -
-
-
-

-cmdstanr 0.1.2

-
    -

  • User is notified by message at load time if a new release of CmdStan is available. (#265, #273)

    • -

  • write_stan_file() replaces write_stan_tempfile(), which is now deprecated. With the addition of the dir argument, the file written is not necessarily temporary. (#267, #272)

    • -
-
-
-

-cmdstanr 0.1.1

- -
-
-

-cmdstanr 0.1.0

-
    -
  • Beta release
    • -
+
+

cmdstanr 0.2.0

+
+

Bug fixes

+

  • Fix potential indexing error if using read_cmdstan_csv() with CSV files created by CmdStan without CmdStanR. (#291, #292, @johnlees)

    • +

  • Fix error when returning draws or sampler diagnostics for a fit with only warmup and no samples. (#288, #293)

    • +

  • Fix trailing slashes issue for dir in cmdstan_model() and output_dir in fitting methods. (#281, #294)

    • +

  • Fix dimensions error when processing a list of matrices passed in as data. (#296, #302)

    • +

  • Fix reporting of time after using fixed_param method. (#303, #307)

    • +

  • With refresh = 0, no output other than error messages is printed with $optimize() and $variational(). (#324)

    • +

  • Fix issue where names of generated files could clash. (#326, #328)

    • +

  • Fix missing include_paths in $syntax_check(). (#335, @mike-lawrence)

    • +
+
+

New features

+

  • CSV reading is now faster by using data.table::fread(). (#318)

    • +

  • install_cmdstan() gains argument version for specifying which version of CmdStan to install. (#300, #308)

    • +

  • New function check_cmdstan_toolchain() that checks if the appropriate toolchains are available. (#289)

    • +

  • $sample() method for CmdStanModel objects gains argument chain_ids for specifying custom chain IDs. (#319)

    • +

  • Added support for the sig_figs argument in CmdStan versions 2.25 and above. (#327)

    • +

  • Added checks if the user has the necessary permissions in the RTools and temporary folders. (#343)

    • +
+
+

cmdstanr 0.1.3

+
  • New $check_syntax() method for CmdStanModel objects. (#276, #277)
    • +
+
+

cmdstanr 0.1.2

+

  • User is notified by message at load time if a new release of CmdStan is available. (#265, #273)

    • +

  • write_stan_file() replaces write_stan_tempfile(), which is now deprecated. With the addition of the dir argument, the file written is not necessarily temporary. (#267, #272)

    • +
+
+

cmdstanr 0.1.1

+
+
+

cmdstanr 0.1.0

+
  • Beta release
    • +
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/pkgdown.css b/docs/pkgdown.css index 1273238dd..80ea5b838 100644 --- a/docs/pkgdown.css +++ b/docs/pkgdown.css @@ -56,8 +56,10 @@ img.icon { float: right; } -img { +/* Ensure in-page images don't run outside their container */ +.contents img { max-width: 100%; + height: auto; } /* Fix bug in bootstrap (only seen in firefox) */ @@ -78,11 +80,10 @@ dd { /* Section anchors ---------------------------------*/ a.anchor { - margin-left: -30px; - display:inline-block; - width: 30px; - height: 30px; - visibility: hidden; + display: none; + margin-left: 5px; + width: 20px; + height: 20px; background-image: url(./link.svg); background-repeat: no-repeat; @@ -90,17 +91,15 @@ a.anchor { background-position: center center; } -.hasAnchor:hover a.anchor { - visibility: visible; -} - -@media (max-width: 767px) { - .hasAnchor:hover a.anchor { - visibility: hidden; - } +h1:hover .anchor, +h2:hover .anchor, +h3:hover .anchor, +h4:hover .anchor, +h5:hover .anchor, +h6:hover .anchor { + display: inline-block; } - /* Fixes for fixed navbar --------------------------*/ .contents h1, .contents h2, .contents h3, .contents h4 { @@ -264,31 +263,26 @@ table { /* Syntax highlighting ---------------------------------------------------- */ -pre { - word-wrap: normal; - word-break: normal; - border: 1px solid #eee; -} - -pre, code { +pre, code, pre code { background-color: #f8f8f8; color: #333; } +pre, pre code { + white-space: pre-wrap; + word-break: break-all; + overflow-wrap: break-word; +} -pre code { - overflow: auto; - word-wrap: normal; - white-space: pre; +pre { + border: 1px solid #eee; } -pre .img { +pre .img, pre .r-plt { margin: 5px 0; } -pre .img img { +pre .img img, pre .r-plt img { background-color: #fff; - display: block; - height: auto; } code a, pre a { @@ -305,9 +299,8 @@ a.sourceLine:hover { .kw {color: #264D66;} /* keyword */ .co {color: #888888;} /* comment */ -.message { color: black; font-weight: bolder;} -.error { color: orange; font-weight: bolder;} -.warning { color: #6A0366; font-weight: bolder;} +.error {font-weight: bolder;} +.warning {font-weight: bolder;} /* Clipboard --------------------------*/ @@ -365,3 +358,27 @@ mark { content: ""; } } + +/* Section anchors --------------------------------- + Added in pandoc 2.11: https://github.com/jgm/pandoc-templates/commit/9904bf71 +*/ + +div.csl-bib-body { } +div.csl-entry { + clear: both; +} +.hanging-indent div.csl-entry { + margin-left:2em; + text-indent:-2em; +} +div.csl-left-margin { + min-width:2em; + float:left; +} +div.csl-right-inline { + margin-left:2em; + padding-left:1em; +} +div.csl-indent { + margin-left: 2em; +} diff --git a/docs/pkgdown.js b/docs/pkgdown.js index 7e7048fae..6f0eee40b 100644 --- a/docs/pkgdown.js +++ b/docs/pkgdown.js @@ -80,7 +80,7 @@ $(document).ready(function() { var copyButton = ""; - $(".examples, div.sourceCode").addClass("hasCopyButton"); + $("div.sourceCode").addClass("hasCopyButton"); // Insert copy buttons: $(copyButton).prependTo(".hasCopyButton"); @@ -91,7 +91,7 @@ // Initialize clipboard: var clipboardBtnCopies = new ClipboardJS('[data-clipboard-copy]', { text: function(trigger) { - return trigger.parentNode.textContent; + return trigger.parentNode.textContent.replace(/\n#>[^\n]*/g, ""); } }); diff --git a/docs/pkgdown.yml b/docs/pkgdown.yml index f3150bdec..39643d5e4 100644 --- a/docs/pkgdown.yml +++ b/docs/pkgdown.yml @@ -1,14 +1,16 @@ -pandoc: 2.11.4 -pkgdown: 1.6.1 +pandoc: 3.1.1 +pkgdown: 2.0.7 pkgdown_sha: ~ articles: - articles-online-only/opencl: opencl.html - children/settings-knitr: settings-knitr.html + opencl: articles-online-only/opencl.html + settings-knitr: children/settings-knitr.html cmdstanr-internals: cmdstanr-internals.html cmdstanr: cmdstanr.html + deprecations: deprecations.html + posterior: posterior.html profiling: profiling.html r-markdown: r-markdown.html -last_built: 2022-03-18T18:23Z +last_built: 2023-07-25T20:34Z urls: reference: https://mc-stan.org/cmdstanr/reference article: https://mc-stan.org/cmdstanr/articles diff --git a/docs/pull_request_template.html b/docs/pull_request_template.html index f6580eab7..f82c3fd8e 100644 --- a/docs/pull_request_template.html +++ b/docs/pull_request_template.html @@ -1,74 +1,12 @@ - - - - - - - -NA • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -NA • cmdstanr - + + - - - -
-
- -
- -
+

NA

-
-

-Submission Checklist

-
    +
    +

    Submission Checklist

    +
    • +Run unit tests
    • - -Run unit tests
      • -
    • - -Declare copyright holder and agree to license (see below)
      • -
    -
    -
    -

    -Summary

    +Declare copyright holder and agree to license (see below) +
+
+

Summary

Please describe the purpose of the pull request.

-
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/CmdStanDiagnose.html b/docs/reference/CmdStanDiagnose.html index e313ab968..4046c2686 100644 --- a/docs/reference/CmdStanDiagnose.html +++ b/docs/reference/CmdStanDiagnose.html @@ -1,76 +1,13 @@ - - - - - - - -CmdStanDiagnose objects — CmdStanDiagnose • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanDiagnose objects — CmdStanDiagnose • cmdstanr + + - - - - -
-
- -
- -
+

CmdStanDiagnose objects

- Source: R/fit.R + Source: R/fit.R

A CmdStanDiagnose object is the object returned by the -$diagnose() method of a CmdStanModel object.

+$diagnose() method of a CmdStanModel object.

- -

Methods

- +
+

Methods

CmdStanDiagnose objects have the following associated -methods:

- - - - - - - -
MethodDescription
$gradients()Return gradients from diagnostic mode.
$lp()Return the total log probability density (target).
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
- - -

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +methods:

MethodDescription
$gradients()Return gradients from diagnostic mode.
$lp()Return the total log probability density (target).
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other fitted model objects: -CmdStanGQ, -CmdStanMCMC, -CmdStanMLE, -CmdStanVB

- -

Examples

- 
# \dontrun{
-test <- cmdstanr_example("logistic", method = "diagnose")
-
-# retrieve the gradients
-test$gradients()
-
#>   param_idx      value     model finite_diff        error
-#> 1         0 -1.3398900  29.94130    29.94130  4.70503e-08
-#> 2         1  1.8968100 -36.55540   -36.55540  2.82879e-08
-#> 3         2 -0.0268544  -7.27184    -7.27184  3.28137e-08
-#> 4         3 -1.4508300  25.98480    25.98480 -1.55207e-08
# }
+
The Stan and CmdStan documentation:
Other fitted model objects: 
+CmdStanGQ,
+CmdStanMCMC,
+CmdStanMLE,
+CmdStanVB

+
-
+
+

Examples

+ 
# \dontrun{
+test <- cmdstanr_example("logistic", method = "diagnose")
+
+# retrieve the gradients
+test$gradients()
+#>   param_idx     value    model finite_diff        error
+#> 1         0 -0.185026 11.68660    11.68660 -2.89634e-09
+#> 2         1 -1.579220 12.33760    12.33760  2.86842e-08
+#> 3         2 -0.391481  4.28029     4.28029 -2.55717e-08
+#> 4         3 -0.580364 18.08200    18.08200 -1.15854e-08
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/CmdStanGQ.html b/docs/reference/CmdStanGQ.html index 5f560e4ca..24ddc1db4 100644 --- a/docs/reference/CmdStanGQ.html +++ b/docs/reference/CmdStanGQ.html @@ -1,77 +1,14 @@ - - - - - - - -CmdStanGQ objects — CmdStanGQ • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanGQ objects — CmdStanGQ • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

CmdStanGQ objects

- Source: R/fit.R + Source: R/fit.R

A CmdStanGQ object is the fitted model object returned by the -$generate_quantities() method of a -CmdStanModel object.

+$generate_quantities() method of a +CmdStanModel object.

- -

Methods

- +
+

Methods

CmdStanGQ objects have the following associated methods, -all of which have their own (linked) documentation pages.

Extract contents of generated quantities object

- - - - - -
MethodDescription
$draws()Return the generated quantities as a draws_array.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$code()Return Stan code as a character vector.
- - - - -

Summarize inferences

- - - -
MethodDescription
$summary()Run posterior::summarise_draws().
- - - - -

Save fitted model object and temporary files

- - - - - -
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
- +all of which have their own (linked) documentation pages.

+

Extract contents of generated quantities object

+
MethodDescription
$draws()Return the generated quantities as a draws_array.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$code()Return Stan code as a character vector.
+
+

Summarize inferences

+
MethodDescription
$summary()Run posterior::summarise_draws().
+
+

Save fitted model object and temporary files

+
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
-

Report run times, console output, return codes

- - - - - -
MethodDescription
$time()Report the total run time.
$output()Return the stdout and stderr of all chains or pretty print the output for a single chain.
$return_codes()Return the return codes from the CmdStan runs.
+
+

Report run times, console output, return codes

+
MethodDescription
$time()Report the total run time.
$output()Return the stdout and stderr of all chains or pretty print the output for a single chain.
$return_codes()Return the return codes from the CmdStan runs.
- - - -

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other fitted model objects: -CmdStanDiagnose, -CmdStanMCMC, -CmdStanMLE, -CmdStanVB

- -

Examples

- 
# \dontrun{
-# first fit a model using MCMC
-mcmc_program <- write_stan_file(
-  "data {
-    int<lower=0> N;
-    int<lower=0,upper=1> y[N];
-  }
-  parameters {
-    real<lower=0,upper=1> theta;
-  }
-  model {
-    y ~ bernoulli(theta);
-  }"
-)
-mod_mcmc <- cmdstan_model(mcmc_program)
-
#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model-d654db71ed5.stan', line 3, column 4: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc

-data <- list(N = 10, y = c(1,1,0,0,0,1,0,1,0,0))
-fit_mcmc <- mod_mcmc$sample(data = data, seed = 123, refresh = 0)
-
#> Running MCMC with 4 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> Chain 3 finished in 0.0 seconds.
-#> Chain 4 finished in 0.0 seconds.
-#> 
-#> All 4 chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.7 seconds.
-#> 

-# stan program for standalone generated quantities
-# (could keep model block, but not necessary so removing it)
-gq_program <- write_stan_file(
-  "data {
-    int<lower=0> N;
-    int<lower=0,upper=1> y[N];
-  }
-  parameters {
-    real<lower=0,upper=1> theta;
-  }
-  generated quantities {
-    int y_rep[N] = bernoulli_rng(rep_vector(theta, N));
-  }"
-)
-
-mod_gq <- cmdstan_model(gq_program)
-
#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model-d654260c210e.stan', line 3, column 4: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc
-#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model-d654260c210e.stan', line 9, column 4: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc
fit_gq <- mod_gq$generate_quantities(fit_mcmc, data = data, seed = 123)
-
#> Running standalone generated quantities after 4 MCMC chains, 1 chain at a time ...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> Chain 3 finished in 0.0 seconds.
-#> Chain 4 finished in 0.0 seconds.
-#> 
-#> All 4 chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.5 seconds.
str(fit_gq$draws())
-
#>  'draws_array' int [1:1000, 1:4, 1:10] 0 0 0 0 1 0 1 0 0 0 ...
-#>  - attr(*, "dimnames")=List of 3
-#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   ..$ variable : chr [1:10] "y_rep[1]" "y_rep[2]" "y_rep[3]" "y_rep[4]" ...

-library(posterior)
-
#> Warning: package ‘posterior’ was built under R version 4.1.2
#> This is posterior version 1.2.1
#> 
-#> Attaching package: ‘posterior’
#> The following objects are masked from ‘package:stats’:
-#> 
-#>     mad, sd, var
as_draws_df(fit_gq$draws())
-
#> # A draws_df: 1000 iterations, 4 chains, and 10 variables
-#>    y_rep[1] y_rep[2] y_rep[3] y_rep[4] y_rep[5] y_rep[6] y_rep[7] y_rep[8]
-#> 1         0        0        0        0        0        0        0        0
-#> 2         0        1        1        1        1        1        1        1
-#> 3         0        0        0        0        0        1        1        1
-#> 4         0        0        0        0        0        1        0        1
-#> 5         1        1        1        1        1        0        1        0
-#> 6         0        0        0        0        1        0        1        1
-#> 7         1        0        0        0        1        1        1        0
-#> 8         0        1        0        0        0        1        0        1
-#> 9         0        1        0        1        1        1        1        1
-#> 10        0        1        0        1        0        0        0        0
-#> # ... with 3990 more draws, and 2 more variables
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
# }
+
The Stan and CmdStan documentation:
Other fitted model objects: 
+CmdStanDiagnose,
+CmdStanMCMC,
+CmdStanMLE,
+CmdStanVB

+
-
+
+

Examples

+ 
# \dontrun{
+# first fit a model using MCMC
+mcmc_program <- write_stan_file(
+  "data {
+    int<lower=0> N;
+    int<lower=0,upper=1> y[N];
+  }
+  parameters {
+    real<lower=0,upper=1> theta;
+  }
+  model {
+    y ~ bernoulli(theta);
+  }"
+)
+mod_mcmc <- cmdstan_model(mcmc_program)
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model-16218796affe.stan', line 3, column 4: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+
+data <- list(N = 10, y = c(1,1,0,0,0,1,0,1,0,0))
+fit_mcmc <- mod_mcmc$sample(data = data, seed = 123, refresh = 0)
+#> Running MCMC with 4 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> Chain 3 finished in 0.0 seconds.
+#> Chain 4 finished in 0.0 seconds.
+#> 
+#> All 4 chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.6 seconds.
+#> 
+
+# stan program for standalone generated quantities
+# (could keep model block, but not necessary so removing it)
+gq_program <- write_stan_file(
+  "data {
+    int<lower=0> N;
+    int<lower=0,upper=1> y[N];
+  }
+  parameters {
+    real<lower=0,upper=1> theta;
+  }
+  generated quantities {
+    int y_rep[N] = bernoulli_rng(rep_vector(theta, N));
+  }"
+)
+
+mod_gq <- cmdstan_model(gq_program)
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model-162182f764088.stan', line 3, column 4: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model-162182f764088.stan', line 9, column 4: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+fit_gq <- mod_gq$generate_quantities(fit_mcmc, data = data, seed = 123)
+#> Running standalone generated quantities after 4 MCMC chains, 1 chain at a time ...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> Chain 3 finished in 0.0 seconds.
+#> Chain 4 finished in 0.0 seconds.
+#> 
+#> All 4 chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.5 seconds.
+str(fit_gq$draws())
+#>  'draws_array' int [1:1000, 1:4, 1:10] 0 0 0 0 1 0 1 0 0 0 ...
+#>  - attr(*, "dimnames")=List of 3
+#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   ..$ variable : chr [1:10] "y_rep[1]" "y_rep[2]" "y_rep[3]" "y_rep[4]" ...
+
+library(posterior)
+#> This is posterior version 1.4.1
+#> 
+#> Attaching package: ‘posterior’
+#> The following objects are masked from ‘package:stats’:
+#> 
+#>     mad, sd, var
+#> The following objects are masked from ‘package:base’:
+#> 
+#>     %in%, match
+as_draws_df(fit_gq$draws())
+#> # A draws_df: 1000 iterations, 4 chains, and 10 variables
+#>    y_rep[1] y_rep[2] y_rep[3] y_rep[4] y_rep[5] y_rep[6] y_rep[7] y_rep[8]
+#> 1         0        0        0        0        0        0        0        0
+#> 2         0        1        1        1        1        1        1        1
+#> 3         0        0        0        0        0        1        1        1
+#> 4         0        0        0        0        0        1        0        1
+#> 5         1        1        1        1        1        0        1        0
+#> 6         0        0        0        0        1        0        1        1
+#> 7         1        0        0        0        1        1        1        0
+#> 8         0        1        0        0        0        1        0        1
+#> 9         0        1        0        1        1        1        1        1
+#> 10        0        1        0        1        0        0        0        0
+#> # ... with 3990 more draws, and 2 more variables
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/CmdStanMCMC.html b/docs/reference/CmdStanMCMC.html index 6ca6e51ab..e7ecaec5d 100644 --- a/docs/reference/CmdStanMCMC.html +++ b/docs/reference/CmdStanMCMC.html @@ -1,78 +1,15 @@ - - - - - - - -CmdStanMCMC objects — CmdStanMCMC • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanMCMC objects — CmdStanMCMC • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

CmdStanMCMC objects

- Source: R/fit.R + Source: R/fit.R

A CmdStanMCMC object is the fitted model object returned by -the $sample() method of a CmdStanModel object. -Like CmdStanModel objects, CmdStanMCMC objects are R6 +the $sample() method of a CmdStanModel object. +Like CmdStanModel objects, CmdStanMCMC objects are R6 objects.

- -

Methods

- +
+

Methods

CmdStanMCMC objects have the following associated -methods, all of which have their own (linked) documentation pages.

Extract contents of fitted model object

- - - - - - - - - - -
MethodDescription
$draws()Return posterior draws using formats from the posterior package.
$sampler_diagnostics()Return sampler diagnostics as a draws_array.
$lp()Return the total log probability density (target).
$inv_metric()Return the inverse metric for each chain.
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$num_chains()Return the number of MCMC chains.
$code()Return Stan code as a character vector.
- - - - -

Summarize inferences and diagnostics

- - - - - - - -
MethodDescription
$summary()Run posterior::summarise_draws().
$diagnostic_summary()Get summaries of sampler diagnostics and warning messages.
$cmdstan_summary()Run and print CmdStan's bin/stansummary.
$cmdstan_diagnose()Run and print CmdStan's bin/diagnose.
$loo()Run loo::loo.array() for approximate LOO-CV
- - +methods, all of which have their own (linked) documentation pages.

+

Extract contents of fitted model object

+
MethodDescription
$draws()Return posterior draws using formats from the posterior package.
$sampler_diagnostics()Return sampler diagnostics as a draws_array.
$lp()Return the total log probability density (target).
$inv_metric()Return the inverse metric for each chain.
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$num_chains()Return the number of MCMC chains.
$code()Return Stan code as a character vector.
+
+

Summarize inferences and diagnostics

+
MethodDescription
$print()Run posterior::summarise_draws().
$summary()Run posterior::summarise_draws().
$diagnostic_summary()Get summaries of sampler diagnostics and warning messages.
$cmdstan_summary()Run and print CmdStan's bin/stansummary.
$cmdstan_diagnose()Run and print CmdStan's bin/diagnose.
$loo()Run loo::loo.array() for approximate LOO-CV
-

Save fitted model object and temporary files

- - - - - - -
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
$save_latent_dynamics_files()Save diagnostic CSV files to a specified location.
+
+

Save fitted model object and temporary files

+
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
$save_latent_dynamics_files()Save diagnostic CSV files to a specified location.
+
+

Report run times, console output, return codes

+
MethodDescription
$output()Return the stdout and stderr of all chains or pretty print the output for a single chain.
$time()Report total and chain-specific run times.
$return_codes()Return the return codes from the CmdStan runs.
+
+

Expose Stan functions and additional methods to R

+
MethodDescription
$expose_functions()Expose Stan functions for use in R.
$init_model_methods()Expose methods for log-probability, gradients, parameter constraining and unconstraining.
$log_prob()Calculate log-prob.
$grad_log_prob()Calculate log-prob and gradient.
$hessian()Calculate log-prob, gradient, and hessian.
$constrain_variables()Transform a set of unconstrained parameter values to the constrained scale.
$unconstrain_variables()Transform a set of parameter values to the unconstrained scale.
$unconstrain_draws()Transform all parameter draws to the unconstrained scale.
$variable_skeleton()Helper function to re-structure a vector of constrained parameter values.
- -

Report run times, console output, return codes

- - - - - -
MethodDescription
$output()Return the stdout and stderr of all chains or pretty print the output for a single chain.
$time()Report total and chain-specific run times.
$return_codes()Return the return codes from the CmdStan runs.
- - - - -

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other fitted model objects: -CmdStanDiagnose, -CmdStanGQ, -CmdStanMLE, -CmdStanVB

+

The Stan and CmdStan documentation:

Other fitted model objects: +CmdStanDiagnose, +CmdStanGQ, +CmdStanMLE, +CmdStanVB

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/CmdStanMLE.html b/docs/reference/CmdStanMLE.html index fd63166fb..1eb913bf5 100644 --- a/docs/reference/CmdStanMLE.html +++ b/docs/reference/CmdStanMLE.html @@ -1,76 +1,13 @@ - - - - - - - -CmdStanMLE objects — CmdStanMLE • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanMLE objects — CmdStanMLE • cmdstanr - - + + - - -
-
- -
- -
+

CmdStanMLE objects

- Source: R/fit.R + Source: R/fit.R

A CmdStanMLE object is the fitted model object returned by the -$optimize() method of a CmdStanModel object.

+$optimize() method of a CmdStanModel object.

- -

Methods

- +
+

Methods

CmdStanMLE objects have the following associated methods, -all of which have their own (linked) documentation pages.

Extract contents of fitted model object

- - - - - - - - -
MethodDescription
draws()Return the point estimate as a 1-row draws_matrix.
$mle()Return the point estimate as a numeric vector.
$lp()Return the total log probability density (target).
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$code()Return Stan code as a character vector.
- - - - -

Summarize inferences

- - - -
MethodDescription
$summary()Run posterior::summarise_draws().
- - +all of which have their own (linked) documentation pages.

+

Extract contents of fitted model object

+
MethodDescription
draws()Return the point estimate as a 1-row draws_matrix.
$mle()Return the point estimate as a numeric vector.
$lp()Return the total log probability density (target).
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$code()Return Stan code as a character vector.
+
+

Summarize inferences

+
MethodDescription
$summary()Run posterior::summarise_draws().
-

Save fitted model object and temporary files

- - - - - -
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
+
+

Save fitted model object and temporary files

+
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
+
+

Report run times, console output, return codes

+
MethodDescription
$time()Report the total run time.
$output()Pretty print the output that was printed to the console.
$return_codes()Return the return codes from the CmdStan runs.
+
+

Expose Stan functions and additional methods to R

+
MethodDescription
$expose_functions()Expose Stan functions for use in R.
$init_model_methods()Expose methods for log-probability, gradients, parameter constraining and unconstraining.
$log_prob()Calculate log-prob.
$grad_log_prob()Calculate log-prob and gradient.
$hessian()Calculate log-prob, gradient, and hessian.
$constrain_variables()Transform a set of unconstrained parameter values to the constrained scale.
$unconstrain_variables()Transform a set of parameter values to the unconstrained scale.
$unconstrain_draws()Transform all parameter draws to the unconstrained scale.
$variable_skeleton()Helper function to re-structure a vector of constrained parameter values.
- -

Report run times, console output, return codes

- - - - - -
MethodDescription
$time()Report the total run time.
$output()Pretty print the output that was printed to the console.
$return_codes()Return the return codes from the CmdStan runs.
- - - - -

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other fitted model objects: -CmdStanDiagnose, -CmdStanGQ, -CmdStanMCMC, -CmdStanVB

+

The Stan and CmdStan documentation:

Other fitted model objects: +CmdStanDiagnose, +CmdStanGQ, +CmdStanMCMC, +CmdStanVB

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/CmdStanModel-1.png b/docs/reference/CmdStanModel-1.png index d7b8b97f8c19824bd7f4936ec52513cc8fae745f..27fd88217456b4a7d0da11a09599ea685749e06d 100644 GIT binary patch literal 15795 zcmeHu2UJtp+IGN#j)ie-BOr{@Y!pF2YOo9#5KyYrpfn{ykPaaPE90P|AR=8wL;|73 z&;kTSsRDx#Is~Ky2oPElNJ#ki!I^c}x7NRIzw5jATQ`fW6HZRq=Pl3sJn!DQf5Fsn z+jnB$!CKDfQHN+A?_Ee23`BFc8_N~X-M8I^%$xx$Wtq&!(VF}v-MIJmm5mR%TrqH zBa!fLZ0GfZTs=H8Pj46R)9)=h4&OU?izRd|@y(4K*+Lg!J@u?$U7d-cUhnYx-mvN!BbtBUKM; zhncpex%fmTzR9YMma?HP9}LFfDD?Zz$9sEw)zVrBqQ4Z7$@b~3Z_9GU7sVtLd%L>c z7I{ItL43s4zsBRYZ{Hqv^3)v?iDZXF7VB*)C@7elJn84C*b3Pn_Pysdk6Gc_!2J>$ zERv`<`x;H!Sku@m6I^g)(%{TWBrY?Qy>|oGV-(OIw zKc0&x*x)!xl%dZFOTKAC;4+n^!+)E?2YAS`)nW^~wQ^}0DcSg)@o*}#O_LJ6{q zmg`SS`y8hzVDOtc@~r9GH}Z((&B2x|4VpHA+QAVAcfY(d_@xwZ8=VhYFUI1at$+?PG(Tf{q1lM%p|dV4!o?dZ!mQ{f#?iC?)nJtwHH8 zQrO_|ovPj*g84alDI3C8&iVl^U&?xasgp7orO?J?i-<^96-jo}doeGLDPb$0*Uy2* zpZ{XWY4!)`Q$*)qK$Sm0$+8!mS;d;(|8SB4mk>nH1zFQs#58RNW-GV~2DQZ8Dik5L zz(6|Ytu8fDD{C*+;Z`$5QK%l4E*N`Nh0Lt5{)<>!i)d?nWqun+#zy;OZs_WZN!J0( zI0*3aw*nZvBW6SOuY%Xh=YEEU>F-?#VgKI@g?|kC??lo+PG5Uw%q~jA92tK%Sv%*!G0ai>eORRSV z^HeVTp=jei2=_OY!)9&e(gD`sjMTBBU{BRgCG}!p`wv%uKgRtJKK4I&`kK`?m*q~9 z&1zY392>6Ws1Xe}rm78|F!<_vr-%*7&Dr6@y=~lwXfvMVD5*n&3KDMc)DvOYp2bo_ zV6#QtPt?lg^h$ZiDg){__d1`Ne(R|S_0Yd6^6(_`21hS~0_l2z~eh4W*6{)E3>+l2p1y4JmF{FK1&GODcEhDF)Y z@+Y@%-%d>a%#$2eZ7RbPyR}(+)w3l<8F%{ObTY;&QppsBX4=xymGz-&DfST0DC?oo zm>7vmb$~d4N^F*vG#zqU^W@35%VEz^;w9Gu0!Y?2HU!brIRrxgl1w|G zY28z`rg!%}Xw1|Ja?}8pZyQ@!{=myYiVNR!R!^c{SO2VWiWnbRA84tzpCTy5im6#q zw_&iKl)DViwFabcuAD7h8Sqq3+FX9Nq|0nSiV4@S1;z5Or=b7MwZEIa|1X)&iv}Fb zvEas@~!R=$z~g4FOAmo(Yd znHg`IbQS4jhx%KLBf7?K6e?nDTYf%s@0FZT*^wDHTl>?=K7DFPc5n!smw ze@?tBr6Gwcl-oBCqz5?|K34JLVT1IGYE8>bGoW<8LQ0oO0oeIFw)YefQc_ZYAO(0{ zU0t0jf3P*I(%$rz2)CW33qw=8Cv2yu_b=*hT=(w(qiRq<`YQ< zXFHwS%#&OPi-1AdAo`)z)?#%WKHXloAtvPUV#?kJkcbHsLSe(;qFPZ{u7x zrCU*Ww-G5#NtK4BU!sUZ&0q`~P(R5BwaOy5ajALN*T$uD61s13;0bNGHan9>5ZktR zwma6AD#d@xJPZ)c8**}ALfj2%@Qa(p6Wml_g_+cU5j6FH^o0$w%5ptdOR`qyD&nBQ zYjA1+9|Puvs*ZCF<`Cbuvbb?d2>K^I6t?GUui z&O%@+1O&Q&mCdOiONjub7eAOwfXn@-)$6|zZ~xBRgj!i~adCf;-8{qe))yXbC5dnF zWtEx0^GR)zhhy{Ym~Ax^&56Ga!5mDy!yXLpDk0G&tgupcJB z2s5v%+jqP{PtPB2aqQd{YJJc6k`V0KONMu)0V!AZ=gxqEs2!v8visbL301~-qWJ9v zY=Iy502E64P@8Ux2Gxq%3BStdp2&yeP9|89hBe@j%7;DsdC<*AKx@0fpT`YMU><0A zCgNKN9No1huV_M5=$oM@y#q)3NE*CeuTl+^vG>hZ$P)5vdsGG*TI~Xk2OKL z^iQ9lSrb_>+p_p*3cmu8=T~0;j=(JqGDGz(>TZTr%|q3xiWC;+D_A;8A@wtV+OqV8 zhO@)bl~MuIgjId{c7Q|jd!8zjtyP0cY0j>kOULic#F`1i@*SX};3M58L`nLwi3dva z;Iz*_VeUe4`2_p?-q{{yI-f1jx$y;BYLAa)n3 zV^Jn;?N|K=ipaUy%VD5N;-H?UF-in=eda@$US9X-ykE98ec~NwemZv?bUm&$<8YpJ z73&nq<4ATl4}I-s5E_k!`bb4yC@^1w$D5j(PR`Ee)9;r|O|%u^0|8lbo*nrvLIcB| z0*pKb*eR{Bpx~lcZ-+n-PD`&kT?C4ON`j>o_5J&sqVm7w7Zq7E7>wKA{$W>KT!yJP z#>$fAHbdV2Tj$ySi$;G!BS;T}_P^V5jiDETIxbTO&ffh=yWXS&a@jvUL;p%z<;jzbC}sftPUBs+U&w`KjpQSTD-EF`MQjpg zu3L~2IFrKp+0hW02d|QKV5i>+l&4BAx4ssddCaV}VoEK3EBag0s(3IpD=?MATxqra z*4NhG_}`w}Up?jh=Tq_c$5t#@6krIg1^8PCQq?M&DY!k7%i!IF2QfLa2R1)m@`^aP zX@l3PhXq^ZEk{TCOD+V8GLsqB8EO8O*x@Yh@S_C3Dz za$nAO$U87|apm=tT|ewtScoy6VS96UwKqH-JCw+fi+QCO9d~cv(ta$BK}>6@WG;lv zU}$I%HZNYR2eA)c8)UyB>!CwJ0OJA0JVnwVgKfkF{@KWT z<*AU<34I0V~6V81Zzw9`CEQ|hri5xAL##X=Ae+5mzN7IzH#(fE@>c;76`_k z^}3=$w{3j>5yJ8a-;-6zX0zKpNpjZfcB&={nPm25*99^?*jClT1KmXX;2*O3{r&xS zz3B)YpESB+WgBt$RM7US*N5w*Que93yNWpuw91I=_rE|ij=-v<_~sqVwq)a;{C00c zd6EX&+p_m7wQqw{CsC#+ff_vaD*XAv-?_RQ2x!|iRCrc;p;v7k|<*N6AowEYkKmsL%v1lu@ zF`{NCMEoyYO((vT)Ra%??rt|{PbQeu=MH_zrtN^SZ$jRVIQ){RFyTFv-4|o!1G@^t z0onv!-x)y@uHfto!fQJsZvF60MeKSCo%LtqZ}VCs^S_sXa`9hj3r}<))bi0sb!l2Q z3@a67TB8DGF`aD}8}BdNGxrQU08`Lk^P&u*awRr3Z0N~=d;IYO#c2hywaxoI1NXs9 z)^ZdUBcpv3;bHGPb)YHNHB37)CLk#$yjxNKhc!gWPx3ts0_5EbJFmbEFv6#UD9Hb? z4b%_JdME(fFJ{XsO&y}&WbJQKbaP-77psxi!`M&O;zaI!)rG=IoHh}>lN=kI0+w== z^j~iJ7SvB;O8e_?*)}P-BU(293r7U_ocfH^fj#gc0_S=j}aoskeK}=vb22)TC4_(2%WpUcGdrw zM0(41UrY4-yzs&TUGhDn*m<)1^`ja5jP)6LdD2wRfYZ_1HwG?*_Mo=qT%XLA4O#hA z-`mN{eme4SA@U2=9jCgyR-*JTZ`~Tbf}yTcO|pESvzfN=Rf9zbZ;kwCSIg(cDwB|Vlb`1NfEcQ zw#svK+Tt*&6P|&GF#(^Ghx+DHo0=4FBYgTN_m#46Wl4@TcUI_{NxtWbH6-P}q4EU} zl1${H+^UBG@#sF6fOv2W{VE?n3q{vI>_GEyBsQ8qa_X=G!A9jm^E7Ef@= zMCU6useSYbaDBzG&b{H`HbUDU_ayq=yTbRvQ%uh&Ci7uUe$8XrdA&A6kBUV) z_0X{E5cW~q5B#Vr7i1k%xkpzn_DLt_0=PrNxWIczOvFhmb1LuM%<+T=<&`4#VK`B% zSi7h+mj_N~rgCI5_DDXtWXwRGu|y%%2#zkR#+~15ix@baA>>fOJ$+(|d^;A4z3(kd zRn<|nHyU!&t*uPv8;#H%iyU?GQ{;qM;cVojuJr{)*=U@MW=Upu-1%IijF?A{_(|~e%R=aNiG(EctP^|mJ3UF3 z*{qN)S9#8bFY^>l)XtOCRgMsf@6zoN2C0n+DNVUon1PltsgLSY=&u25Nqf{i)Km9M z{lK0yFY+cH8K*okA)zUcwmL(7vXf{XiY+ER-1?-KU!w3qFmU|&=c{vKk$7ojzk+}} zlPyEE8-Lyxi-gM=w(OV;2)`hjzSl8DL2uIEP&T3aE04UXcBJZpxnlI^-jh4hv0gD- zMa`Obhbl-*yLh}?^0U3eukbrGaTPTkqMS{7Vy4x;kX*O}JIxkeJf(d^8b|K610LEc zud5LL=y*4e?4X#r_+@tXKAY53;@&59oBEpgDM~`ZdwSu8>+70FMIB}C61v`o(cLd6 z>%>I~eIl~;r~2zrcELA37-U&z6ZXz}_q@v*XzDNl%Miyt@3U3$ad!R*i&aQ%_+Y$U zZ)tFY&Z}p5hJrtOswIVf_lCRAKZ;0&`816B zdXntaz5BYRe4S6bHzy~6rKWU|T{Yma+FS7^N$R9_@%PH9;I+tRYrzlwlA!AGhD^p# zBzwOyzb!mePQCYF&}94dD}x4@nNqs%Xq=54VvmB|(=Cc6FY0%{|M(zW=51W2{>_V^ zqr99=;bzO!?=X~j5{;d5G0EK|XMi+rb-3|qcXy$iaT%GsP}+jc?H;YVGmbi(h{~%_ zH`8TV#*E;za~r0*+pe?Zze6DHj{&;Ot9WBu?N05cX-dnT$uO^cBU6xwnM?ON3y*oqMdM{1(u$D4M=-j0f;BQaCgEFQrHOAaBUao5mkm!3Jc z6v5uagmR+c{0TF^iH!J}f(n`CaYJ}wteIX2g|Xq#D}%HsHj6cSYl~uCRh53T9&aic zCs-CSSJXd1CFc<=kCaXoHWmg>_vj9U58d56_CU3{iy129WHPjF0APn`*x}8i96Svo5Eve0V?eT+B)>ZpBbqW6O(-6I_SIu_SbJ4+#aGpk=fD zf&^;@gQ@rXZQL+ZfdtM)6G9E^j*#XcA7(3<;zH?{W?B9jr}xDPsMo8lrJp%<(N>yt2k6_-!DnJd>gCgt-$5=jSd2D<^ZjF71uMZ#bz z{cPj*4mltRH21fU!rfZJCu|U(oY`I)6)YYx1yH! zAh1WNeD=d@(gx+Y6R&V9eRRFz&RzH1I|nl^WNmMA6AG zpQwu3aOkbUA}RgUYeWlaIvq@t476>MNrsWbBUj2r<9 zo>ufczEnla!O_~I_Vwu=%$9aqZ`Wgt9NM)?&ptx1ULfH zhEwumi&FSp)LNxEe$7khyzIkCDHO@hI_J_8aqp4?i1`6dr3MK0-r>n=n1y2nuP2Q^ z<*^U6KR_K>0AF(4imkG#GWDI6Y*N^;IPF|a-;LV+WAqJg%7J~zxSs3fBxzI4(WiZ( z3<6Y91W&grAGj+iNo{R?vSG^}nSD6nMMKFD8H6wR;fI->JSjR750u_@^E}cUJ3LA3VvG96WfFm1mwEkrKqBR>C6Y&Kh;w9QK_Oo} zh)Rd*f7paEa&3qrlEHw5VKwJI?)2kMWO}r|`Wds;)sQa6V})UGJ;KkoPf|N)Dm~}2 zn|N29&N-K5(L)qq7;9!Fdpte5`Ju+Z^W8_SU4e<0ms@bDAbv(($C@^PimeK!ZtkIX zDp#1?-|?;6+4kgq@fwHtmp7;)o)rT2?;jZWhDcNe-tY|6gu#CM(v#!}?!QuIl<{Q> z{5;~t_YXij)cFf`Of?O1qtilSClG))cE2$Qhjwf<<6EKsy8md#-mcG7@<_b9!K`#E}w?HPxPe z2l9Xx(%r(OtP1*YtNi5^Iy@~Btjz|i;_`y0?rjsFZw{(dMRnvr2UYTB`mmh-HRY_XMt!D7cI zv-3NnaaKW67j`wiQLruZbFwwm!Q3AWz8UiFP$W;{`o#f24a4ql-<#M_6&x*@s!Ek27 zsY`&3V2<_i$rK*>de6IS*Dn_VN(u3_b1`A8Y-H1PG0iE`Llg#cLI7o5^;&#?liiUx zb|}S@lnAAQ)N-96M53M=WdPkU_9{xTArihaB(rY|CMXExnMZGtm*!M{^*s2g6qGC1 z&}cJI>mZN=V5FLv8tNG{0@&U;l3hPtEU#=5FC%?m5fo?DpqV~!(Rang^tb7c8Ylt= zW$eyQQ!p$5(a-DJfa7@Q!I1PDwOH*IkaKkh*dn-D5k$;>?Lb#>TNgNz=x$=2>yRK* zI#(mE>QeK-MyJR@kK%n%YmFO?e9mwNmb6cC7$6nUy&P;SvM6%9RM}yP-Mu++`lhK&YOvOT9qatWCjK zG(dL9u_ekG$167dy8Ecuymt8PZK(%$$vCzL?itW^fzjy?gVMBs!^GZ0RycncjpOyk z%7ARYP#stt^0i6%E<30ibD`pL53uT~YeB#?3(he;^FUUwQOy(jf;eEVq z+qRAVc1F|$08r)1s1noSAyg+~MjoQWTmThhFO>`pU5)~)gn#pat{~U{GEy?p)%6ZT zpy;EX8dSk>On~FUV@Bp1m83Y~5K#iM_&JDY)#j5fX`VXZNn>QPZStB22L~m4k(FT~ zTpo{u5(iT56=d`CWkHn-DK0dnOZh(mbsCUl{xUsXnzk~g5i6rX3-i-k84{(_;j1A) z145ende>vhQWT{Q1a+(oa=F1q1tcsMKL;h5O5uxMAoJdrgK$H{ynp{*hdZ8bTjqaF z4@fDHR+@+yz@(eQz->?5fdUc1DeJ6IEbjAN<6<%hk{toXlTYzCpCL0Vp$v>du3QNp z)(DpdwG8}i*>kK86xw*-(qtU37}p3OlU-U91aw6yxMl2v^&f$pH3Ndf#>QpP22Kqlhy)q9aAu03s+TBE8I5Lsbxwjs;PWUIL*6b=6S>G%)l~R73&+ zDFGn_a2Uae1T`dpkOU3I5FipwkPwplo;dfP-#&Zq&W^kH?%v;dFt5*f&-=dL`uUb~ zt{-tYr1kekP$B|R>xBaR5PZ!sG zyY$Y;fz7#*dk$-_(NNH;x%Fzv^+l~`G)^7evOi`2zU2Mi9X+~v+qQ4F-r^Wvjj?ZL zF`rY1;#NO7@xzVyvka0!B9+hTDK&BKmNdPLA0HYoTO_CJ3;t}jMy0YzRrqA}#Dr3O zM})*pzIlGJPZ{Td*>U>Xlm_{Y3fuvng8YBxxh+CyI6vjID+XQJS)ES^iQdQi3ZBwl z;(a2br^TsfwK8v>fHJ>XJu!;HOYaF+gUO_Ii}Wu16sU|BHF#8Ii$pklN%;nNN6YmW zmG}NSO2#`*5xsX!;YcpT05Z(y=44w27$E4!ubSm!@e*%%Oggei4<4<{(ELf(ezmju;_-;a+5|Yx|kM2UQFK?P>*fM9``G@z$ ze^M!LiQ#x~j?;FPZsg5}V90R|!c}04>_5Fzg>%HGth(hDShrI}Jz@CP#m>;s zNQnwwDk>$STBJpIDPFkZKc5eaYMm%cCb-%~?CKvtOKs@Wc5Jq4C!Ws=mP^(_3v60V zb`kyxh^L7U&;z6)`5lj5xK4#Mc+b4-s20Kkj4|Hnwit~6y+D7{#{*y((D2Klm>>zg zc*2RNTFSlJG~b@p#IIukF+5BK`ddNl>q8}m_oAP0MJPNTA7~Da@ZAA4yk_#5H@()e z9@xNg^<%2K9#!Dz6Fy+?B_6u?Xn)CIGwQV!1hq8^{N;N8yY>GC%tzmd^NFU1;pMn? z1}c^ULfftb=CWTz2}dg@x(F!v&N^V8n(<@dbmB^&iuqY}B5c0afz-dzO#dL6|J>z5 z!gA=I@rmO>AubC;4V~4(Ozmr3`1WEUjJ3n*z>3Z9(T@;S@H?mKy?0Of)^j(431g*2 zdTpH#*{%uGHoy{DkG!JBAO5!F?W!;1?f+akd|{~iUq$>eHX7KkZ4cR;wLy$72GIC2 zN~jk??kD%bz^ShesD1b{wMm>3{GjzbnR8Niy)e*4m?n3J4_g2#AEU4A!>{+^eWyO< z>mpQV4FbTnyaeF;hzn`Z#lt`m26|`D8aN#P?RfsbN=Uyq82w))Ud&hwB?+VW4+A#f z3tA0R2@Q5`DK1@dut++z6&T9+*$VvViKw7fA@)4qY7GeFNuEGbp_S3~_+al~@~PK? zPM76f3>{!6YAFVPCD8v%$`%M`^MEtP#IicBssMNq!i!GZ$NL8Mi2RiomDToPsuCmU zbT$!=sZiV5Z)4u9iWk|qSkV?cs;Rh;V@u@#3Q;e&t@Iy$6-*b|E#@y&%0oIt?XpW> z2X6(He)I1nv@ety{!1Fnza#JRPy@v~wY(P_C%!Kd-k=G+=y^W8CWaZviLF3pmDIiK z_*9Tfg7_3~3vF8E7}KZTTLH5E8lFn2p}Be$h7Efu!U|RD+4%eF2YQ1Va4htr*uZ8} zwPW(wcHvui-`ZWhQzuDqp53nt@9aPJERkL+a<9sco=@a~SQc6jD}(wx(I+aZRe0k( z{1C50BT(*fU#fih%}C|hHsr83wX=tQM9&iw)jSxj2%@kSm~VToTA5ntMdx`9Y6DiD zuK)trQb6*I3M&Ovu|aJC68wI1@TB!m3sd;yL4qqFG2$)O5T^7G7g)bh`1jUl|I*7r zaj8GrQO@C~y0kLUo!2|WcAu-`3p+1nnar0|PR#DS&`_{zD%2%vPaQB>pr8BS#^eyE z2zWh9I{Han1FK5RkCSTzdRUhK`1cjJ|I)(y|HS`K6}A5y$xnDF0ehFS-_*dFUu=}5 zmf_)0OLS>qp+U9-8Z1fnpuYmO-Bela(td(Seh^Z}+U@rMPd^e>hi|Ri4nVkQ3;2P~ z-e=>*b39O2$H|H&6y1~|-}V1sPk}332y#=?O)KGv0%SS{|!};z7MASud!7 z+G63HDk;s+M9oT1Bs5E|oXXTb?ZPyxz#mvhZ~>)Z+iSQdm2^nZSnfjdAh^cYjUEwJ z@nIMUmHvb2|97rT|3O99XLB_UQ#2Y9Oh@?z*I0&YeI@^WllvZ6tvAS#r3MAg9!PPmf@W&*3K zn(77t>_54K2g$?mu19A;6#qNAnqS6);jH5IqnX#Ac!msT0xJLQ6;P{z%a^)@r5F`P zlD?TfDQ|c}ZJaeRM*X8qe2@2eP>-QPTz6cP*JF@x3wXBv#Id06PwU1i1Ih4uXe_8B z7IuCpDILy6!1aamq5qTF#vfEV+zE%Eya4}le1~}ALqkVZJ-}O&W5O*}%NF{Qz539L zq4-yDPMdUawW~|=0eT6JVSAt1;$wKz#Z}~?mm^B|qc2x-uY!{H${)MBbLyA{su5(F zR+bN5vvOxU)S2JKDD-=7QLlr~f`LvPdUEH-!5Im^4G8OPV z1pU<)1br0`Dt($op)?3V106~mZGY)Ma$#F>Z}rJ@OzZ4dyHC8U zjNNblcfwM-P?6_$Yptl_*pF8O1 zw3mYV3T{bif9_?t+k8%zUpRgJxp$oD1X~6K)n#0gzcMH)S@#)x{`wXHOQUB2O|OJWFzO4gJr89g1j z!Nr>8Qp0wyt{Ul{m9&RP&`gZYtJ6lEIwY(pQY)b(RStGVmaIoJqSTUJ<6-@2U`eRXu*nR>2@)@~OpqU6m6 z!dZrq)GS=O@4QO%?#)QcW&6@$yWAXyg62l@mtG4@U*^CtDIjtSz=&DmmmJ{5z*5op3KRQZ?yB5aarJy?8UDY}HGl%O&f1V** zKev%N&FP$$@^Lq1t*9Dw?VrP_F(cF!ocwCk2rB%hj68U2=BF?jscb=(8T9ngE8B@| zhcWW{?$IFQWdTpv-P!(wJnd(T`LFC}t7*T+gEc(nGYP%gQvA zGe%6b=YB+Q>?%*B-*dK++8~3g46V+Y_H16BCCH*t?&5Mozz(-^C0A<12l>jm)ZK^e zcj{>LJfO&Q4NuPQ{tKPB+)lD6|bR=dbax+1ooZS2owWkwa=%_ohZ^#}*;j!KPi- z8h!W9YEqN|*@v|U+|x@-#+>c_H{N4h*oiZ5V8oXpl#FoxMc7R?f+_b16yPW}(xhhx zPFEyr>2#VJJzp|%CnLU+tE&|nr9J0B*x2)|IX&lfmk2w=ip=1YY2?}JDVL6<-mg{Z ztbIl4qjjBVRc#$gZm{Tlt{5wE{_1pmnk9@YNztax{_m+;oY;f#yMD;}viI#MPgZnB zU)>{fN{?3T>? z_ABNlqhN_-VU{g>dH(}gu*0LixiVvS#5fKm4W_vzU9YR0H%$?LtEi+E`{5^_(_U1Q z@%+GabB)slYl{f=7X}1Z5qgbpZ|F#;$g`~ZQu}d*>^*&+5#27xXrIu>OAAv^yC>c5 ze|WOJ3+uV=gRcHnQ+v4h21xwzzYTLRe08*iWI0KTxND$vJVX4_=^5|%@#8JMk?WI? zb#---?B3p9UI4)-#B^obz3nByzD}#+g7a~>go6a(4&>SNdl@yYG;VVcyCtDP$-u}u@20@GZg9%d zkYW@-aP&EG;!(uHd4<>h86PH7Rn01M3JRoLcOJ1ML(eO5ZZw2VLh)a%IkmW=j24-6 zG8czixPtJR7NQGTO;J2Q0|SG;UU#Faii)8Dt6M>5bd!*~CL6<|qebHZ1omM;GIA#} z97zxob#!#tLyZyRwpVfLMT!SwRm5+VTVqTq7ZICWj~~~ZKBJVBz6&244OZsV*48%I z+uKj^h3E!qPmip3XlO`nbf9;Mh- zwExEcWHiaqRzPv*l{W+pMh`AlGJjzf`R;`f$jV0W5dp}`DJ--yo-ADI>zN{k<{j(k3nGtDF!WjF;F=H<;luI7pOY?(DyNfTMsM+)$I ztl)koqqCS4J@tBsCuBu4qxw`m&O!Nx2Y2;ElbEgEHUl(v$hC0gbHD6g*G!|)^bTX) zc|pjDYMwg!U9qeWBd&985&WE$UseriK6yCa5j0p&xWH@@j&Uvso=9JHW0Gn+04vJM z$}%tQQDU5El1s4Cwl-6YD`1l2X0IX+S$;Po1tJJprK8hUmi^PE1_n9-o8i<#WQfx+ z5>A&JIrDU8W@e>{^JxXU8z*~G;5}6NlUJRd`7Ml=I_l_KuS}VM^(4KZJ(@|hG*k%P zyHR8wH_JrHj$uE{`?SOz-lgC$037dXaYnccw%46+N;C zK(ty}Sz*}d&PC^vk3J^?V;e-1S(6fpq^Htqi!~|5w;;g_Gk1g_ycFCpFGc9t?oJ;$ zU6O&Kc6T$Q#n%-asQZBh*bL+&a1P@y!uERD0dYP5d<3O2tTR=$PqWeY?Ab9z-p1uc zL02vMre6~S0R~)jdZwqS{U_^k#fSoCt-qrd#a8UY^0sN574xcOB{?}cr61Ac*4rah z40kKqRJ1B>(N?2x{NHPu^!qrd`sbw}oor%9R;4c0*QP!ldKB1KS=Ol*H^~NzHUT4B zAE6sszj4Ad?%j*xhD@i#(}{`F;p_EFjcr#26)Ze+ZK>ftH6}8AJb__~0K2#8g$Z+A zKGL;-4H1H$Kri{b@IA+?Rh6H?duKznCLlL>#!eRjz+%Tw9qWiqNn>?|{U~wJg#n=FFU~>SG>h|H@ z+J4;zmjVfx+>t9pM!BPbUK=v!Cxe)@rY><*AfXV4V-8`&%!GN-!aKVZAlQQzMu3%; zij43#zx?tG#+-Q*3NpIXSf75T{rj}7_puiq5FMT&)uB$qOR=JMWKDB(^Vy3YHjP!S z&9m$7F3QJV-xf1IHkNp6WG!+hX{KXSLq#>kV7?1gK!GKY%AT z0?rF38$D!ZR#u$W>-D~i6oU$Id3hv4hM?G;m**{hl~sWH*yKfuw%3?zL=+Yj7Rn@R z!WXYPmb+Tt&I~mjAEZ6X){311Nx|~;GhtyfrRkDH05`!v%RZLTg%X4y$>}P~h}E8jTMbO@Sv*zM)F%`8I-y_G?O3vkZ?A zvTT(!rp@gfmO9)GG`QRpM~z5tJ_RT&e)8lA2`kyJJ$mDZE0dR%Ec=n+!B~3$csSwB zrTW~fhCMLQ0x)j5bw+a`&|2+-aroUEv)ZVOwcYw@=M`e^00QD>r@e9u3)x2l$tIM- zND{L>IXStoikOH29v?8e^S=t-f74uO{Wdt!NK&xDYH7tfV3oCEkIFvR7TIwg*hA^P zQjYX2Ncq7Py4?nAbsa8!7`^9B!41TK!SK8x#g*!tN}N3ZNhn{}!RW%OSW?9p>=<>} z;^RoLv8`&2i=oaW8K-LkJc7jHPv;A_nc4L3=iHsFoB^7_QnIpwg&fNvO)CQQKr0YWU@V7jkT7n|W0ju<})lxLmp)j9?2yqm|x$OwD z1&rtL__G+SId@NOx2Rzt>x?aCG6&H*B$FNrfBP2xY!52>IW>bg3t(~UQvPLRq824Y zYao~Y_M@O;ecd4&R(If}OoUMHTAK`jLhT(N{Az>OR5Z|qD6lEt=A_bX<*6NNxT|eTl_E=uJgy{S{ zTA!jN?&4QeOKlfMS1z9fH8bx%kd&{=M~84yLSQ^U`H)5HSauYDQqRiDx7kreD7mjN zxawIwc0{WdZOs39cZ7r)V`%q(ptDc*I&(uSjOM1Je%eCY$A*=IG-AE=R%LIpcC1WIpXKG zBO{$|YYcihhZ~oH#umrd!vFp1nwZ`XU46T#Gc{yQia$1fFVZM2iv-N?Ae;$rSbo>| z@@(bZ0$F!@f8zE)A3fBOdYyF1N!c7Ptrj#dZr%hIniQsgW7BPB;RtDAod5K?o}(xe zpy!J!3)OPSxVuzYKEVa_7o+)scnIl+4EBg7YEHK|!?s9h7 z0~PBZDk0{wVkM@JTGJnwOkP3m1U~3vIcB)mPng?SXculwNkB@CDUJna>$evCJaTto zmsRG-cd^aCGV@2Qv|_hWLNo_LFZIdEf-U}k@^YZbmM1Q+;nolR_ync{dB9fZ diff --git a/docs/reference/CmdStanModel-2.png b/docs/reference/CmdStanModel-2.png index 093449f41694c87ed4f6a948f6425a8e02c8a550..4ed9fe5c527f0931750caf802c560ea60a2989a0 100644 GIT binary patch literal 16009 zcmeHu2~<<(y7oqjv<}c#1(B&%Dk=ghqRhc#72`k@1!Sg`83bfbfCQzLDgrVn%3M(p z0z~F91d4!w5CN0O7?dGEAYn=Zge3Rdc>1q@{r8?*?>#;D{%5U|71j>fdw=`e@B2Q_ z`@HYBbLXO|$<|G~HbD@y6?x`+GYFChfgtf8Ki>$x8F={X9N299@tn!`U<+O=Cu?zF z^M&sjD}M-5jDr8f;>PElA?O9V09x@qQKFma%k$;w-$$B!$Vzr?d4%x@R@JMEQQoBW_zaiZc)Cj-A5R>dnhN0X7ndoK>DKSaUqq``PlG4*r zh_xRIX-ulWBO#~PJv}4C!kp6Fe9F$wPTT%7K84L=&X2ZomZo~GHIfS%4fe#UjrsXC z_4P%Ii;FGMjv{-UfTM>%5SUCR^|B*jW8?`#%+mAx{D&RSTA|pU(bhQ6;;V0)o0|*k zMN3Od)SEPuqp$&>pZeSML&V;Zc(TM*>Ogjme9Pz$7jns@`g*0ee!9YUndn1(RNC;> zS+pEDdJwfiv%zrw99=ID$0xZKx`w@4iWC{KEQQ=qPIEPR92h^M7TBJ3r#Z(`H9H@% z_?(lFnoh{C7e(fe=g=^401j@3N9kdkUyxe%;r@&Z{t|p?zhCrf7ubHfZ~ji#{*BY0 zzt`}oqqAwgRlO%08s)ZZR9@M^eMTpvJG~F>eCzwp(@@;LT9((uU{ra26Wb9*2*(gJ z_>f5laQG|W@Ze;be&&!+cfpi>V{s)c`M9*??95C7=d;+?A+`?^&O3z8mSdJ*a5lYD zq!}yJwmF$4wms!+ar^ODD#hy<7(GQU<}~L`vysWP2F(~xnWnZ5xsMR*>gsxR&Ky;c z@-)h%t*oH0A4{eB>wMcWUQnU!Z8++W=FV6V>1=Cqgk=$!eP7>WFuR^;&#e6XaX%|d zOUoG9MxkNITvkz0kw>*{MaG6u&WBykr_oPq>gv2ez_P(4E%4@c4i5Bbes1qnVH~ly zp|;Pg)nN3- zPu9PSQiLJxQC`EjL2Qrj98w;$(a?h4cC7&IkQ?>ceJh0Z@g$ze_}EYxte z8$G2l0U!EreR*khE>5KgFpx$v%wg$iX>a|i3pkh|%%&N05*@eF#p!gp9Nqk6{@Zt? zJ}-;Z&Iu~bIC$mxK>EXoala|k&Thb{n&O1b`doJ+%d%bC#n!=rHa@<)m09$M>xD~t&^uli1pkJ2V0#F?F>_y!)zJZc_Fvgy3nH1 zNQ6T9h@`TGuA;Btyah?c|ACnJJHfTpMjARZr%BQ!(6NjcqV_c*KXSUF&@!xsQ;~0J zj^J@hQ>3JVXKI+`8Ss2EJAR;TiV&YyuGcDLULd$gu2oL-o}xIt!8GUE6yPh&IYeHR zw$4cAZr!DeQs4wl1>bx~=U|+xxr3f3@q-H;Tg$mcW}^@VDGGCY-W;cU5#e!-GR3i? z?eHfk^YlMPcKw}}`p2C9{0CTc7JA{QeFO{Ht4-0Cf)t*o^580dF&MX6Bb;}r3wc)g z1q6+43lWRRrd-Ek*@6>>X%^|R4xB=seV;lp)SOexuJ#fKM(om(fJP?MFXRtDsKaW2*oinfuJvic_3aA|1dU)Z6*CkPKix{5R+v2KHGly=X;1{W-(b4duN*fmD`FsH zB+ZG`YudW_UfatM zf|lC~Yw{6#Y#M^!;?CTE9J3hq1_N|t15tHugK<`K&GZu{jUa3+<0Q$0+k|!dAP&g* zpHjBsv8zbtj^q5MTyvb?Fa}fr4T6bJG~gd8|Nr8x{K<94?^nNFYh+BI4RCgke8a&Q zG7u0Usda}`UtDHpmu2h&hM2;hII+1s>RqDrjl-p-rFO#x3T2YgN@W=hYHt@8r4+tN z?re1iHAbQ;3P?k;vtBwiK<_#bnUBZ}?M$)_Il`r7wAfQ7mg;p3-{%#CXWPNl#xfOW zQz+&@_pMX7}uDVa#jn zqjN4ofn*V^k)NS%Lx5x8)!07; z-@WxsK0eImdFl*K?H(D~vN1Alt>28q)>eq%KcO+Ji#}w9FZZ{uv>%q=vat&n8`}zn zur$=mXw0G2Dt%&WCu4D)+5qyY+dPis)ilwi8*cjZjFeJ~raPm55TyDbybQ-LhfkH~ zlSH!Cgpn`L?@dfOzZaIVrF{i1rB<+vbvv3LB`wWQkq8?9y|(5gY=RnY+IZy)N8Bl- zRSxx>cSWB?VQoHXg#x2a8wgqhAZRV|=7DyV9iSyJ1yr_DoM*KyHK_J_jdQ7-g*iX& zE;Qm@VIHVgwCca^x7uWLlBg-4+5BXqhA7sIL??7lqKD{S$I9&q@__wP&+i4FJWKn^ z)(dg8uemubFc_oy?Ngw=DS<_J-dns1VE-U3t#&Bp3c4mlsR0i3tzfar_q9# ztxY5K9{jJr4l4mopy}IPUDuB8ikWVVlr-n8#V9ct3|KCMN&#*Y{hz*2V@E@+J%QO* z>bl+#z^+vRezvo<9h~N)fSwKa5^dN9$JHly$bRR!e<~RMN!`JJ-TR+bEjLj?N{djQ zsxVJx+B4-c-H*JZz8W|MK`C(78;Mx!C1g@?LPR1}%cxh@v}Tn;S_H>IkO>NuAQ9pV zs>uR9IdkG+A5ci5lFiw+7`_J7u=`8+?1p=i<|tEZLX}lz;li*T6wx%|91KVKANGS` zg=!{P2+rlm`{33z-U13UD6-Z-2lPHsSQwU^F(A+U8cwpO6q-kxQ9Vg(|*TIS?{Uj`amqBDhDSDa7cy2=*j0zs>`d+s;&>-g#Jg6R%vI3=*_E|)KC{0`zys-f~X5~*rN zeL475s!gFKI2u1r42N&*&-07j@bc6Y+>vH2eEH@)iHqEJ<|412i4156ZyQ*6C6TdY0wZh29)N)B9Pi*)DN{ zzBZ2%q^}N2poN1n(KU4jgPw1_wepi1Lec1z1!lL~QhWpdJw5W;4L?6Ka0NiWHmIA# zU7!j4MV*qk>H%B~!S!u>G^l#((I#!@rp2PmAFH>WUBOvOp2xt8w&nswl+sEv6-= zUKFrik)D8aTN35rk%dJ0pG&H)EKJmv`8Jt3`Fz#?8{ z{u66zk9Qa-ETK|kHL!TPJYznOYG+PArvhX7sLGt5@34XLTL%Fz4nW~N(Vij;=nURB z^5}~?g|tQK;cEW4qEJBkmH z0?PLW*2#Mb>HbR2!Cb+VO==846c z<69#@UEbT<3&bn$ZR7l~RyeZ$5Vmlb!-fPH`jibVSYf%I$sB3nnmNenAEUD> z_?Yr%!-5hwddwrUe z)Rop8TeTl;4ip|r&)X4w^vS+6M|RxVu{ZKB^WXk7V)N!pMfLBEH#n8Hu`-7IS<2_Q zPn?doO{9f)C59PkJsz6TW$?mK8U7h`|9UhRUDI=UjGaeK%)?&dKK^wqG>)?`M9gi! z1mHsGa=t}E_JnNukACDWsspzZ_PD4nVxmp8s$)nRdSZ~|w~UXi_I)1@hm4W?;dT%N zojv|yP=4xzb@Fj|IL}0eS8`tDe2bw;}%&@w|{I{efR8>Z{~kxi2U00NOFduzCa+D zwf#&o3w`aulRZ$~Uq4VG5{XtXvT`T)y*XxRXlV7q@kIAw#};!uWr@Ri zW`V!JoR2wut-r(Dvr+%pv4_RQ#X%ak#4i%pyWmw6E*` z4VC{D>;8uUtL53WZD?iYw(w{Q9CeOFvjC7)A!wh31pCAU%fw@)Yft&>!X4>Oew@?Y zV3h@evJ|%0N&d?6@EiQ$RPLj}-ji)jHgJH0S2|huZ=vJg-EvJdy!Zu4M!IOByA%COyqmR+ZEAxDSb!SlCa{O>0E z{|9*fYOa9Wzjlv3e(xdV%SyxMa0xzMI=z&;_*-AhlWb_a>AFQ!kki+Y-?h~ZpY6W1 z*IfCW68%yeA?V7EI_T0mGmyK@k+8j*$Aa#ukcKY>Y>Mc5ukgtnF?322_S`$0PFK?K zLT~-xqFeTxgO~HJB#6+jeeDfg3nuq9h<%(&Qp?=rGR@`~=NaD2>tWu#d*A6#KUia1 zO3Nrp=rpV|SVWI5>xGjks(b%lUXyBqi$9qs{?6gkF+;HnfTt57-tD%hPuF_CUQ3dP zU1vaG(D#ngAAP5O?8rk~QMOgJ(4_Qk?2%`*wMvW;Z^HF%tk{g?dKxf})=mZX8hazf z3eFsiHz*EA?QD^-6o8y?6|7cXN4N}o4FY4aEMxR!)8{Q zm70#pPO!G@M(#a+G{s=~`?y;?XdP0J&KXX?cbfWjliL&JcOtP1%qN#;thCzR3)-S5 zWB}3;IE;a+WQLScl_Gr6h>vLck8`>PiNASvsQSyQj)U8zZnk~ShQa7U{1fz*tHdSa4V>n?#5uLr3= ze%5?zXnLeH8J-7pXvfDr|49<|6EwH!mBZ2VKJfY0HJy1adL^@yHD z+iizn?;F~B4`-gMvT@_F2h+4H0@8Xe&Q-kI9o}|nG1~NSb*6!?@HA|Pkr6YdX)K}) z7&z2rUtd}I^u;4;6qPl28j?AOxmO5-*7_E3I} z`f>(yfmNU?uDF_VzJ*TDR;;gQ)>4BdT@|oFnSLfje}}}jWq#<>CK}Iek9*LvQA-i| zOMNEt8I1Fzy(^i!cXd9E@9L@=?E~LOwbDAoWi!kp2NOG;M`Jm}KQa#v*52IfB)Yvid$nG#gh3&8_} z;N@;uQ7GY)-F-&>Hv8&;NrJjKNMhF6y4Wk$q9uDxEf@ZoBe@mu>xG3Uzt`3^NV$SX0=Q^)zg~_nF3nu@ zSsB#L>hsF!-X}bSKT^O>&qkswdxjeIjGQICB5P`Da?->%GDrP`Eu#qR z_r*>RIYjzi{_A(hmR+NyQ*Z8ahA7(g**&qkS=*VRDpQ(`t}U2oYOsDM^PIg-?c{cT z5gzYJvKPgi&eqyIeDmhbEz;7Jn2<>~Aggr8?w?#(>P5K*1e8tc9Lio@T|MI3m8Q}* zGU6D-8np8L_K_8kOHoP~w~Ffn598ydghR*OBJKFb$RkIOCe031%BFqpGWH=+&-^~TbngreQ1uh(KSf=%QLh6YfD=${#47VQNvplo=r$YPdv7& z@Um*W<2dnHT<5A_xfRjnYP9%RrA>Z#?smb|;wK~DA6h&>wtwGg(9k$h+^W>!oNV_{ z1>LM>AJ41j5hYY_J+37A8(Uh^$_Vb%g;H;?zLM{FbQ~_h$4DheJ$qC=+c&K*y4W|e$xBXP;Vg}(t}hT zuUqUom!J^%_EMpeQiN2{OmQn1PBkOs$3}PP5mFP=9lduV!m@thfkdvD2Tj&+iWT>nR&Lf&r3!TUc3a|G{*Li&4EwQ{X?Ptxw#pb1ILE~F0<QGV1LM;~oBE z@g@~ha?l~W*}-5@qcrw@T=>j2tKuI|bs6#oWxm^(Wo~UK3D(RCY$`P>hk|Ud+|tZv z1cD}fi^Hqb$_XA$#2c8~GBR?>{a9*e9_s~qaUyK;q;flp74+K|s$&Zsny6r&J?Evm z;^qjh=Xel57a?%hs9*jt6#GC$SvVI}>eQAXXy0tiEc2&_J34+NYwLRg#F;<0&y}&Z zNT$=Ppt6g2Mg*XQa~jnusmHGO&(6x)hy2i`GkVXR@pI%*l7tUX`1@pnFPcjH#@hichKY~w@JMwa~8e3vmq1tv{I8R;SyvXupHahpBHuR>)GYLx@>$GpxYqy_c6*Al~N zQESR;Ba*%Cn~XinJ?R2IU4-?nnaZ&ZdH=I6=j~-4VjrYV1Jfv~Wt}}Dl$QquB)L{F z=uhrU(e~WiBEcPcD&bSw_Zo#=NK|QKuXV9!xJGciC5-(VrIg~g`=j?bL3}`3rU?o+ zUVe3Xm|%gQJ$D}J&$8pQ$Oen&HEL%{?nHZTc!TagEV)hq+pFH3PR#f!l+T?-CV<8r zC_-nsMzzB;afaM?Nt@l*?*(5u0OA>k7@w^i^gEX@R!#_?t$+Dh*4F-F_ORPg*fUF9 z?nKWy%?r=W+%H^(tn(H98;ZVNme3420Rl5A)w^atcoYleR9Vj{%(KX?W2r(~PapU)f2D~H#HkMcGzdDtRS|eKEou_)9p_jT%{F}BOhtavo{!tEc^RQji zptQ8KtCN#crEEQVZN4>81x7%Q?HRS!4OxQz;vw&2sa8ZxU=fkquhRC|HRJIPHH#VU z*&1kjF1S7!CtSRD@1B!jm7}}(S_?N?aXLF*Wo03mlt#dE@W_NzpZdNPwAiM%(_}HS zU@Nh$sPabp{@k(K^4_mb)7<5~=e2A)yVG7A_L-RFS4S?|=F|AS`?; zQ#)NiHYAVrdlEXoP))nfEUMLDhcWaY)Iy5S75OaSP|yP*&AZYFVycgq<)CH{j>5&f zD}JvJx2mj&ay^%?*G@!&hZxxY_Kt%|BTs9hSE75}Rep3^bv_-hADWr-`QR0HZ;C1_+t+33a22>_^LwvxCL% z(x!zjeI?Gw1Rv~TDjyb72P^lade z3==RN$e{6)Fuw;wE3`?L_o+J!T29+e9eTk&(8&;bzx6g)dUJARE(%dMF!1o8RYJ68 z{qk{PF0s$cG|hoIjVE-)n<(kd7DM|i)Fxm=)d^;OP$dlfyoObY3Qowt4L@0eX5uWc zQt#5El(9O>t8b$G{+gG3?0g}v4L~9cO>h@&b8~ZxEqDw3N-Ak3b&bUqdT-7YQ?T7j zeATa^8QRyeutLTe9Ul(4#;Rsyz=>*H6NffM2$l}o*M(I6JX5sjDhLWTNC(p~1=C^J z-wZ}EECYn6BD%WU$zo8_IdHLSvATO{cENTQ1pD+rCUxABv}!J+VfzTC<&I=c^6n^w z9xmfE$r3}Qk9{m-SHIuXK2g>DaohAiEr+tQX9laa%B|JG8#r1`5A=)z%2=Qe2Q~QM zK_kxu=o~ZO&p<0zi*-43^$KvLJzzH((0O%!1?o4r%yh8@a-V3=p4=1IaUVQ1{vS4U zv?kyL1Wq?vh>uf2W?-iC-g=aS_h2ZSSQ~ymyZaT0({)e1IVs(lrt1$A>e?bg~o^y*Mk$!TZv)OS*DZQrp>h!gVj)&@Kkl!E0$R8KBkbQ_8>L7+$xY8hLbRW3Y8IW6V z*K#s$C3ywtVqsk-J(wO|&93J!0p0|3r)Ng62my3K*K8#{!?UP*FUpW}Wmn8;(Dq#HBV&&tOKyy|ge_$& zhYyctge|nC`Tq7rU)2o|YZH%Pi-IPsoVXBQUfpl8!N%rf#aJ+oP0az;o1wyjn-az0{&u3zD)^rVnkn%z%1|5W*COZowA*dQzXete$d+@H57aDO8gEI!gR)p%=fF(XI!J5h zO!cOos3t8VnBB1fWBqcwR8|cR&^os-aes=sq{1zIR%B+pzhg=tp}DVu&X_W z^cxDKIdfuV3=6F>E}qFP}ZrAl+KBlyx_qbCJF0QQd6r+ zA1b_Zz^W#|9hP^JN+FIUKOoAz1}fC!5W)~8Fk!FZI$gXaAwwr&Cj|AM0uC{5ot~W? zVzkcad2kygmCqMiq%PEh2dR-jGCmBKP9D1{sq{(}vC*{N=i;9&N%C`szw)Faa`AG zZoelF?FF|bVa!YIGf@r?IOaY0XuKoY6$tO^RxLxpbi+$5;Z&y39Tqo3O4rv0F9QNG z)?e=mOD4hD1(^V%v+{b!G7Tr7sUUGUTv7e%G!pgQJOFd0Iv^euc>GOoWB1sSjrcPv zqYO$okOZnk`sAdW6k=@-aK_pINibmK95(rDRDud2q*<+pLMb0;jfsh=h8J(H=Yj}v zY*EB6zel#j`KGfZy7n&S94GEKQMmDa!WfW(=}XvT2P_&!4ycsUZnVhpljf^NW6veP z!V}2#T1DJ=`-48<^Rf&urx|s)Gp5Cg@d+D!$=fWeaUT5ku){R(4N%}fZ+apVVAak> zlS3Eq4uS(8VQf?3)h3BpZY;2^K7w2T>~t6utwXu@;A$yhesc|xX~%8{As4SACnwh? zE^Tlh*pjg=bNTxTa%QH=3PH50$ruH(t`0Nm>#A#UhR5Kd1Qf!VD(**jsVj=c5>Z`> z!GrODp~8SJ?2NWwdU6F%0q=vhRfjAc!UOXu9!9)svNNDfe%d^TfD4BUs_*19e~Hui z@C&F}U~z0tz?gg2&Q=VDjK*w}Mf0anxD=1_6j*wwhi8CQHFv!vkO;)z!b(oGLS_JP znt^(r4cC?MQT|jb*Ks>G#&V)FwbEgfexp@!1@_S(nvo9hY*pff90uhR11=drP1Os) zR~#39FjKgcdl^qD!EmP2{Mz@cV}ii)2fQi)7D&`{-QX?s2`hvDX^Q)^hY%okc!*B! TDKcAs_5^v#^!uW}INtg<4VB|G literal 12652 zcmeHt2Ut^CyY5z0MnM_J89_mr5=2y_NRh5%!GM6m4818$X+c`(MEw;V1ayZYCCM-t zG7cgN0!9eIf=sAlfPfG{2~8w{P(lbKXYF{}bDs16^WSsl{`a1H+|Q#cthT@PeeYM- zzJAKqa^r^W8z2bUh(2-D9)i}!K#XLNrc`+1u1C+!>b$$x2{|T$ zf!W8mad$ZVxwc$Ra*`m)Jj^H{iig3r35jE3(i$mQf8u|*)yyNy>=IN)t!SBmnmw}Y zCf!`6ki=mBYNBpr5hI6m)IY3HGLw$rWGl9tLAdYq+8xX#r!oeOjFWF0F*%Dn!RR8i z#uMkmc`p~6l!Wq9q75hQD;wEThSJIySOtb#kx~CVn4*>L`LLIp#2-!T*bDlHVm^;q zm0qZ>x{c;tff?K!111ms`k1jDdEp_Ai;775b7sAil2cYgYmvWqt4*&)+RQJI=@4YK zz7}a9@)}R1!P&3YLfTKuKd}zM!6(@K!&Rm;a1MfcjFa`sh{ieX8N>@-TZK6+V!k+f zP98|)B~bby6Vut4Y{Tf;_+|~4lSQoHD!Oeq!%JOi0&I5!%{Pcy>fl_DC~^v3$inP0 z6j+vlrQWE3f2i^!C&6uX^1LmXTfV#8HohaJ*W|=Pcaf7yDAppD^8?HoBnnewr}H*e zMQ#Q3C8%Fi95T0IRA3kIaskJ+dV?{^bC=>xZsV1)uw^@{Ib_8GHlpgp$$-iIvS4ye zn19ytkmd2Ro)1A7wRr6Dvd0#xw3`L&ZM1hQr7Xt0Tago(TTcs_taJcBz_QE#6YQa5 zkt_t=nCq+*itx)X#&-yO%*Ip)W0K+`<~9Kw#D7)`)n7L#n~h**(zgMBb8o9qh)TAI z1FIgG0Z)KKsV3_JWI5H+`$P<g?^c9W@h9SA=RN~2M-~N8snDu7)){`w-P|IM;m^XT848Xgkd4s> z@0>nWn%%b}1~VlGeAOYXe+-E4c8#xs>M!Sx`Q9=Wb|{XLcGCRt^O?;>hJY-U06E6_O6kzb6iuo72!Cff&u+s#?M8D&%Yc0DnD^$X&mvojrjm{4UWy6{x;C7S+Jl0 zfkO#m0_F|+!ho(JOYdz3!elxb%xoCdei5R!&y1?uaOAD_hd ze_o9Ln<)6_<%nk2Kc8Q;o^3c#J)2*IMZ_02x5o3}Xxl(j!~D0&x;6~GIKo4GNXWb} z#R*~|A@kQ4)_rt5ntzZ@yQb}gTJP5#OvZEzdpk@v9k7cvogM!QDp)gHaFUVXk$6U& z)!~M8&sUBWCo2J}p!7@ia;o(SH%Tp-T4@6pc3)7$DKPZZCXGn8{0U@oyRZg@prR2> zv>dV|o{X-VtH--x;WBWn)j0gSCLxC}?AU7{TB?Sz(Fnn~qVJ}kZFbh=ii3+rvbn2= zVq8!5iSllvm>`;g!d_G@P9gCN1f@uGg-`iE!i!I}kH22`3Dl1R97=E=HA0uiy_1{0 z^3=h=Rh^PuWC=o^BcP@<^x}50ua!Bq;&+pxYLIH32Kgy;@H7aNK=KOWr%6fvhjt_j zd-Xjs`>0Xn$^C>KAcCq6-6ax92cFqS(U1GWQ?b^c}Kz~5H} z`{+S`ZWbp9F6fLXlAajQjiOuM7HfDV%~w+}H$b6o`T^o5>#=@V30plv95{YcG@Fa^ z+HSTQ8ajh$8%~!0d=OtS@cPK$!Wd_-TVjX~RNsNH3nyD%aUwb!Nw?0MfdHdE#mAbC zK@tjR$n$?fVXhL?Mv>y>j(d1!Gm4>@Jk&m>wTT$i0;t%7R`S->;N)pfi=yCc3%iqa#j3n z%ZAHWxOtYHD3lt*chC0Mhf=0*YoGrOemn7K{h7Rz1lz(#YWRak8n}T!RB!wK{KaB} z+x!h0J%?`GTK??Wv#Rdb+?DwZTKQjkN<7Z<*b!y(!j^uc0D&nNt12)IpvQBKKyh&Q zT4`ikf6e}5P58#IGsQ?zq`rv#y+=WI9>HzUsVvx_H8(5IZBt$8@{dd-$a$M|k$lfX zVZ(V1VW~fUUs-0Y&h<-*-bZ&Mj?_EN{h=B)%!8>BCt7}5yafuY`VxWsgX1ro-pN2q z4gc)vk7r(zGRFI;#z+7A|87TT3=_%tao^a7k1Pz{)jDtdGI!sj@NuqTmX-lQ##>2x zaGYgl9!otok!jR+r}Ab7N=Pk^faGPQ?3%t-stju?>JfgaHmS>3Az+Nx)=!_3ay{d= zlJ%~X$C^reZPX#9{h{T^eJZ5=isELMcGv5Tga_&*VY>8JksF2*3K3%m9dGj41viES z8Xv=82T1)EEM(~Ic0^`>^Y#BiU|{WpbZW+~#;x!b+wzsGOLfH+`vSfZ=3wb$)1Jsp zL;W8`yfohPefcT;%)?sp#6DW&`?KrK9Qk#q4W3G8HXtg7Gj}u0Hl5@bJIqBdo(pe$ zTkHU>4E?WDaP(E@j!R+(sVE8p(s%OOG84jk1_#V^a-!kh;)pq29b%aZPmgdmR+*)aRVPoV%#ozX z9WtH&E+7B!n~<0>McCx(QvW-=f!9s<@WTN4H9_avlZ#b9vS{q?#z}bQ8_gJ1BF6Y> zQz=GGnWYbqFazJHN53p6`X+HbqIJ*%mH()Ze@#K5`iW5!#@a1tICmr6N!ga!H*=+U zOIY2&K4Alwa^D551^S%^1B>qBso^oY=Ba2DZ# z8j}oX0`Y+-Y#=`7%tMYQgJ(b>=j2)bWVM#(Vn1ED>Vi51ZI*ruu<7?`;&1;==%9b*MiV+j2B~E{3Twbmw6>Up4i4AajZH`*Rx# z4&R-juF!5sSbeUGYl#zXSAt~-V&TuWu1)^&t2bgKLWJtHOOK#XkRXny7FgsJsibV=qi z8}9cE;)WXAX({Yz&md>Xy6JmijIFiwg}2HVGX?zM!*yb zCt||){=`9R4Ye)$v7v!tFIbe<#%z8lbSJ>i)r7~HE^&K$Z{;yiUgwOhw8*}0Hbgkj z4k5v{qVN&_gX?SqoIsy2HOq2$gE5_Xm%+{>Plcv6k91jyi3;R53=xJ><~Uk4JBzwu3&t;M}L03;|j1Vfj8-Xvz zRsI;X5~eYtm9fKQq1q;3mG@Mpz4({zrBoju9};rRKYGS~DQIn)iWXT~YhX$%ZPkoi zt%P=^j*}jug(=ZMulWTF!PHHMB7`~HPPVWk&`3%-2RtroFp^^i? z`z%D)th_*Zl~kHXMMp2(@1m^z*=rrU_d`E5qH%F?(T?R>wuhW>VY0iFS3C1kIYZzx z7-k+=8rg97?p?1Fe6lTNHaDcN%EQ3Oh+5SB<;RVS%k{TtXvdSb) zhL&0{>}y>*UnwUi=QUUv#7dRDrHYF{2vDSr9~(w*Lb+5wco62OPJpzXFRMr>#l`mv z^75WU1qTNwQ<&6PfB#WRF@+-7gFO@F7^&SATAi%N$btJ09yl=j#Gx8f-RFf4@2Cpt zBR+x^eLtQ)b~)J!w(D~Bq9uwJ+!R#J`;Fq}#`VdW7PkCOR0O)JrKQDG?^@YfT$b{j zPJ9hJ$qx29Q680)BwjDONo8^2y@SjHS3phQ)IuLztIiOEqtfhoGAk9o9Eb@jHQa@! zFk^WNTEAiJ%BNlx9+8|Ym9jl{S<-6Xl+oVSmNfcO8DpBnv@z6dX=@Xc0KW=Qa_C*w z8d@KCPEb8niK8Q$XL(6#?VDD}nTaK>>klPw3|vNdMF0IPRTMR{e{QgREia1YfjzG= z@&?toz#EjLslGW}GC$!vqadj4SA;Hiwd=&U6V0}eGy5nM3eP;TSG*!pP#2znBaI2_ zWyS}g5?#P+b`2IY9%;;by!5`rPHPi%uIJ~M5`X|aIg`c?!S2BpO0qR$8XFr)IjG8R zH+&~EM-^x}rl!I!c4V&?)vt8nBdP}K5=GAdvqKL)-{UoI%l4K1`OIZge^P$FNI=eH zaoIK16-#%ZIp)_dC1&vpCs9T->1hgWfFGr*E*DiF1wVBQqU}OkX<+wNE^i?%zDU?w zmVbM=F@xt^lDJCSc}1l{z-9LtF_Yh0Yve>xNJeAz&|JgM>AgS{?b-IBuJ73@kRImq zVae0WvdMcGFS#ka$s;Gd*AF+Bp0bJ!hdW*A5m&poC&Q0v_=#MnL^<6{h&LrWq=wn_9NQB#@1H+ zhRT(}i6;b$^mqaHq3dLdwFbSPat)F{L~oSP;ssT3?PJl~DRO3f8Fe(zi&#*QkU(Jy zWzk~K()sM@nkgRSO!jLZvf-b?8b@D^>Q+N*j9k=5^;{+}zxhY&S*R@%gA_2|tHkD}GTISW4G3F%cnSTT(4Vpm$ja z%*FzWF8R%g_P*VGsUu^jsbD-T$&H~OgW=g4lFn47!gsm?HDd|chd&^EOx1}G4dv3v znTC&hSJzLEk1bm0>(399Gp~bP-IkO#<5rCeW+XBSrqjP~)CXvyM=kMJM1sN7Rzeaw ze0gyuhZN;Kw0W=5Wx)$VE9s3-EV^7D0mJ>bb&%4%i|M^J(Nl3DrSlG|qMkOFUc}^)31(*tj>e+Jv0h3nb%2@bQNn%gP!hKyzarvqOR8p4m6S z0c%;Vk|0>I)3xkiO6?(*)%~aO@6tz|v?>z<0|Sp>7u=*4Q;Lx7oZDvLB<1!^$0{l+ zB!2qaR*lnF-H?=`wpg3mDp~8E4)S0y)FJub0Y)tHP_whM)>|KwGeu9B@$KL43Cr`7 z8vvx#;1hHien4N@Wb1q3e%K|hR7drH)u{nw88RXQ{Qb=zbn-oFqRqJXF0Fp$lxQxN zWC7-tcYepRo==&rhQ9Px#HHTn=_3R3)yl*`A2%2lI z&*-(*AfR$dB;LO2{ZLSG*|218ZmtyA#gfW8Az3SN;tFMdA3&%xOtRwxx^nv*bCuhI z{K$K`Zz?A%$(fRyoJu?%uZo7LR4TXyDzv3U%mb4{k6bdS!Sa$?`uqF6P`&O}v4;ip z!()%MpuCQ|52sA~x|9QmxfM%O_rfQ{Zj9LHhk4#EU8nuXGn>=1EnDK4XL1=!4F(q+ zyOECqUfoew&kZp?s$kbP?GHYA0vRp@<;~tfk*q=9c~eb3!T@~M^cN81;C8uZae~l0 z`&MGUP0jiK;LPK&+$paD&L~O&C8VR1cBLBnn&2twLPK<2F6SvqvX{|5l49Lo(OPAe z%9^0*Gj4H*+iP-FFwzV%!&Rcw?Byls44STj6CR6ZCNjKu4-?V+HdJzE!|N*K;E~uG zx>fj$k$?XC`xROYGCR?3m#~XFW|2LT8h{)fcPsLjdo-T>WWnwjo*h30outmF+P9VB zR^j}LxI5Nx-@_FXe1fTk`y*SFQ9qm8y*%Ap)$G?7(==NbetkrM&GG1)-3yNWBR^XQ zg03F+cxs#vFIS^5w&4$IS_@HSCVqqx+dh9wxLRG9rbZl&*=pU7qV8f0S~LeMs0n?s zORZ|zz*8h1&X@uc=t6Qj`O9q^Sbfzm#_GbFmeZNpv>FpyX`|RYwrP3Zta~CW%h7xq;<2q>G8@|m{0W5ha)R(=gLwIZh9BspE=vcE3^UPi-ZyJB6o%8mHWd;y@P zed<1_=hSOEw{EMYwGedaQD2FS9I99-sv08z|0tpy;2fuq6lT1YQSe31z|YPnapN ziuojf{3}^00O!21PxLPv8LW&^K1i??F$W%z-AO=G1;k@jW04d^6m|NMUn$Bda*Qol z4dE`HmO{CVs4G#ZVV0tkXm%pWFbW$vCOHqwprbjQ3)5tmeHRC_@qW`*6@(|t8F27; zo5g;HEUdyyO0d~#JQ6T zcksyOZWE45>_CbL6!Z^_yT(bmJm57;6Yp6 jhb<9ELGSrG diff --git a/docs/reference/CmdStanModel.html b/docs/reference/CmdStanModel.html index 0e7c77d84..d8b10625e 100644 --- a/docs/reference/CmdStanModel.html +++ b/docs/reference/CmdStanModel.html @@ -1,78 +1,15 @@ - - - - - - - -CmdStanModel objects — CmdStanModel • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanModel objects — CmdStanModel • cmdstanr - - - - - - - - - - - + + - - - -
-
- -
- -
+

CmdStanModel objects

- Source: R/model.R + Source: R/model.R
-

A CmdStanModel object is an R6 object created -by the cmdstan_model() function. The object stores the path to a Stan +

A CmdStanModel object is an R6 object created +by the cmdstan_model() function. The object stores the path to a Stan program and compiled executable (once created), and provides methods for fitting the model using Stan's algorithms.

- -

Methods

- +
+

Methods

CmdStanModel objects have the following associated -methods, many of which have their own (linked) documentation pages:

Stan code

- - - - - - - -
MethodDescription
$stan_file()Return the file path to the Stan program.
$code()Return Stan program as a character vector.
$print()Print readable version of Stan program.
$check_syntax()Check Stan syntax without having to compile.
$format()Format and canonicalize the Stan model code.
- - - - -

Compilation

- - - - - - -
MethodDescription
$compile()Compile Stan program.
$exe_file()Return the file path to the compiled executable.
$hpp_file()Return the file path to the .hpp file containing the generated C++ code.
$save_hpp_file()Save the .hpp file containing the generated C++ code.
- +methods, many of which have their own (linked) documentation pages:

+

Stan code

+
MethodDescription
$stan_file()Return the file path to the Stan program.
$code()Return Stan program as a character vector.
$print()Print readable version of Stan program.
$check_syntax()Check Stan syntax without having to compile.
$format()Format and canonicalize the Stan model code.
+
+

Compilation

+
MethodDescription
$compile()Compile Stan program.
$exe_file()Return the file path to the compiled executable.
$hpp_file()Return the file path to the .hpp file containing the generated C++ code.
$save_hpp_file()Save the .hpp file containing the generated C++ code.
$expose_functions()Expose Stan functions for use in R.
+
+

Model fitting

+
MethodDescription
$sample()Run CmdStan's "sample" method, return CmdStanMCMC object.
$sample_mpi()Run CmdStan's "sample" method with MPI, return CmdStanMCMC object.
$optimize()Run CmdStan's "optimize" method, return CmdStanMLE object.
$variational()Run CmdStan's "variational" method, return CmdStanVB object.
$generate_quantities()Run CmdStan's "generate quantities" method, return CmdStanGQ object.
-

Model fitting

- - - - - - - -
MethodDescription
$sample()Run CmdStan's "sample" method, return CmdStanMCMC object.
$sample_mpi()Run CmdStan's "sample" method with MPI, return CmdStanMCMC object.
$optimize()Run CmdStan's "optimize" method, return CmdStanMLE object.
$variational()Run CmdStan's "variational" method, return CmdStanVB object.
$generate_quantities()Run CmdStan's "generate quantities" method, return CmdStanGQ object.
- - - - -

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

-
- -

Examples

- 
# \dontrun{
-library(cmdstanr)
-library(posterior)
-library(bayesplot)
-
#> This is bayesplot version 1.8.1.9000
#> - Online documentation and vignettes at mc-stan.org/bayesplot
#> - bayesplot theme set to bayesplot::theme_default()
#>    * Does _not_ affect other ggplot2 plots
#>    * See ?bayesplot_theme_set for details on theme setting
#> 
-#> Attaching package: ‘bayesplot’
#> The following object is masked from ‘package:posterior’:
-#> 
-#>     rhat
color_scheme_set("brightblue")
-
-# Set path to CmdStan
-# (Note: if you installed CmdStan via install_cmdstan() with default settings
-# then setting the path is unnecessary but the default below should still work.
-# Otherwise use the `path` argument to specify the location of your
-# CmdStan installation.)
-set_cmdstan_path(path = NULL)
-
#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.29.1

-# Create a CmdStanModel object from a Stan program,
-# here using the example model that comes with CmdStan
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-mod <- cmdstan_model(file)
-mod$print()
-
#> data {
-#>   int<lower=0> N;
-#>   array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
-#>   y ~ bernoulli(theta);
-#> }

-# Data as a named list (like RStan)
-stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-
-# Run MCMC using the 'sample' method
-fit_mcmc <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  parallel_chains = 2
-)
-
#> Running MCMC with 2 parallel chains...
-#> 
-#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.2 seconds.
-#> 

-# Use 'posterior' package for summaries
-fit_mcmc$summary()
-
#> # A tibble: 2 × 10
-#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
-#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.

-# Get posterior draws
-draws <- fit_mcmc$draws()
-print(draws)
-
#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 -6.8 -6.8
-#>         2 -6.9 -6.8
-#>         3 -7.0 -7.0
-#>         4 -6.9 -7.1
-#>         5 -6.7 -7.0
-#> 
-#> , , variable = theta
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 0.28 0.21
-#>         2 0.19 0.20
-#>         3 0.16 0.17
-#>         4 0.20 0.36
-#>         5 0.25 0.34
-#> 
-#> # ... with 995 more iterations

-# Convert to data frame using posterior::as_draws_df
-as_draws_df(draws)
-
#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
-#>    lp__ theta
-#> 1  -6.8  0.28
-#> 2  -6.9  0.19
-#> 3  -7.0  0.16
-#> 4  -6.9  0.20
-#> 5  -6.7  0.25
-#> 6  -7.1  0.36
-#> 7  -9.0  0.55
-#> 8  -7.2  0.15
-#> 9  -6.8  0.23
-#> 10 -7.5  0.42
-#> # ... with 1990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# Plot posterior using bayesplot (ggplot2)
-mcmc_hist(fit_mcmc$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-# Call CmdStan's diagnose and stansummary utilities
-fit_mcmc$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181224-1-14c234.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181224-2-14c234.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit_mcmc$cmdstan_summary()
-
#> Inference for Stan model: bernoulli_model
-#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
-#> 
-#> Warmup took (0.0090, 0.0080) seconds, 0.017 seconds total
-#> Sampling took (0.032, 0.026) seconds, 0.058 seconds total
-#> 
-#>                 Mean     MCSE  StdDev     5%   50%   95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8      781    13466      1.0
-#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0  2.3e+02  4.0e+03  1.0e+00
-#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0  1.0e+00  1.7e+01  2.0e+13
-#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0  1.9e+03  3.2e+04  1.0e+00
-#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0  8.9e+01  1.5e+03  1.0e+00
-#> divergent__     0.00      nan    0.00   0.00  0.00  0.00      nan      nan      nan
-#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6  7.7e+02  1.3e+04  1.0e+00
-#> 
-#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47      796    13730      1.0
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).

-# For models fit using MCMC, if you like working with RStan's stanfit objects
-# then you can create one with rstan::read_stan_csv()
-
-# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
-
-
-# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
-# and also demonstrate specifying data as a path to a file instead of a list
-my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
-fit_optim <- mod$optimize(data = my_data_file, seed = 123)
-
#> Initial log joint probability = -9.51104 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.2 seconds.

-fit_optim$summary()
-
#> # A tibble: 2 × 2
-#>   variable estimate
-#>   <chr>       <dbl>
-#> 1 lp__        -5.00
-#> 2 theta        0.2 

-
-# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
-fit_vb <- mod$variational(data = stan_data, seed = 123)
-
#> ------------------------------------------------------------ 
-#> EXPERIMENTAL ALGORITHM: 
-#>   This procedure has not been thoroughly tested and may be unstable 
-#>   or buggy. The interface is subject to change. 
-#> ------------------------------------------------------------ 
-#> Gradient evaluation took 7e-06 seconds 
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.07 seconds. 
-#> Adjust your expectations accordingly! 
-#> Begin eta adaptation. 
-#> Iteration:   1 / 250 [  0%]  (Adaptation) 
-#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
-#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
-#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
-#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
-#> Success! Found best value [eta = 1] earlier than expected. 
-#> Begin stochastic gradient ascent. 
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
-#>    100           -6.262             1.000            1.000 
-#>    200           -6.263             0.500            1.000 
-#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
-#> Drawing a sample of size 1000 from the approximate posterior...  
-#> COMPLETED. 
-#> Finished in  0.2 seconds.

-fit_vb$summary()
-
#> # A tibble: 3 × 7
-#>   variable      mean median    sd   mad     q5      q95
-#>   <chr>        <dbl>  <dbl> <dbl> <dbl>  <dbl>    <dbl>
-#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
-#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
-#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  

-# Plot approximate posterior using bayesplot
-mcmc_hist(fit_vb$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-
-# Specifying initial values as a function
-fit_mcmc_w_init_fun <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function() list(theta = runif(1))
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2 <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function(chain_id) {
-    # silly but demonstrates optional use of chain_id
-    list(theta = 1 / (chain_id + 1))
-  }
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.5
-#> 
-#> 
-#> [[2]]
-#> [[2]]$theta
-#> [1] 0.3333333
-#> 
-#> 

-# Specifying initial values as a list of lists
-fit_mcmc_w_init_list <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = list(
-    list(theta = 0.75), # chain 1
-    list(theta = 0.25)  # chain 2
-  )
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_optim_w_init_list <- mod$optimize(
-  data = stan_data,
-  seed = 123,
-  init = list(
-    list(theta = 0.75)
-  )
-)
-
#> Initial log joint probability = -11.6657 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.
fit_optim_w_init_list$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.75
-#> 
-#> 
# }
+
The Stan and CmdStan documentation:

+
-
+
+

Examples

+ 
# \dontrun{
+library(cmdstanr)
+library(posterior)
+library(bayesplot)
+#> This is bayesplot version 1.10.0
+#> - Online documentation and vignettes at mc-stan.org/bayesplot
+#> - bayesplot theme set to bayesplot::theme_default()
+#>    * Does _not_ affect other ggplot2 plots
+#>    * See ?bayesplot_theme_set for details on theme setting
+#> 
+#> Attaching package: ‘bayesplot’
+#> The following object is masked from ‘package:posterior’:
+#> 
+#>     rhat
+color_scheme_set("brightblue")
+
+# Set path to CmdStan
+# (Note: if you installed CmdStan via install_cmdstan() with default settings
+# then setting the path is unnecessary but the default below should still work.
+# Otherwise use the `path` argument to specify the location of your
+# CmdStan installation.)
+set_cmdstan_path(path = NULL)
+#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.32.2
+
+# Create a CmdStanModel object from a Stan program,
+# here using the example model that comes with CmdStan
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+mod <- cmdstan_model(file)
+mod$print()
+#> data {
+#>   int<lower=0> N;
+#>   array[N] int<lower=0,upper=1> y;
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
+#>   y ~ bernoulli(theta);
+#> }
+
+# Data as a named list (like RStan)
+stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
+# Run MCMC using the 'sample' method
+fit_mcmc <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  parallel_chains = 2
+)
+#> Running MCMC with 2 parallel chains...
+#> 
+#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.2 seconds.
+#> 
+
+# Use 'posterior' package for summaries
+fit_mcmc$summary()
+#> # A tibble: 2 × 10
+#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
+#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.
+
+# Get posterior draws
+draws <- fit_mcmc$draws()
+print(draws)
+#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 -6.8 -6.8
+#>         2 -6.9 -6.8
+#>         3 -7.0 -7.0
+#>         4 -6.9 -7.1
+#>         5 -6.7 -7.0
+#> 
+#> , , variable = theta
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 0.28 0.21
+#>         2 0.19 0.20
+#>         3 0.16 0.17
+#>         4 0.20 0.36
+#>         5 0.25 0.34
+#> 
+#> # ... with 995 more iterations
+
+# Convert to data frame using posterior::as_draws_df
+as_draws_df(draws)
+#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
+#>    lp__ theta
+#> 1  -6.8  0.28
+#> 2  -6.9  0.19
+#> 3  -7.0  0.16
+#> 4  -6.9  0.20
+#> 5  -6.7  0.25
+#> 6  -7.1  0.36
+#> 7  -9.0  0.55
+#> 8  -7.2  0.15
+#> 9  -6.8  0.23
+#> 10 -7.5  0.42
+#> # ... with 1990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# Plot posterior using bayesplot (ggplot2)
+mcmc_hist(fit_mcmc$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+# Call CmdStan's diagnose and stansummary utilities
+fit_mcmc$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251434-1-21173a.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251434-2-21173a.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit_mcmc$cmdstan_summary()
+#> Inference for Stan model: bernoulli_model
+#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
+#> 
+#> Warmup took (0.0030, 0.0040) seconds, 0.0070 seconds total
+#> Sampling took (0.011, 0.011) seconds, 0.022 seconds total
+#> 
+#>                 Mean     MCSE  StdDev     5%   50%   95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8    781    35502      1.0
+#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0    235    10662  1.0e+00
+#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0    1.0       46  2.0e+13
+#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0   1874    85179  1.0e+00
+#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0     89     4050  1.0e+00
+#> divergent__     0.00      nan    0.00   0.00  0.00  0.00    nan      nan      nan
+#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6    775    35215  1.0e+00
+#> 
+#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47    796    36197      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+
+# For models fit using MCMC, if you like working with RStan's stanfit objects
+# then you can create one with rstan::read_stan_csv()
+
+# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
+
+# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
+# and also demonstrate specifying data as a path to a file instead of a list
+my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
+fit_optim <- mod$optimize(data = my_data_file, seed = 123)
+#> Initial log joint probability = -9.51104 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+
+fit_optim$summary()
+#> # A tibble: 2 × 2
+#>   variable estimate
+#>   <chr>       <num>
+#> 1 lp__        -5.00
+#> 2 theta        0.2 
+
+
+# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
+fit_vb <- mod$variational(data = stan_data, seed = 123)
+#> ------------------------------------------------------------ 
+#> EXPERIMENTAL ALGORITHM: 
+#>   This procedure has not been thoroughly tested and may be unstable 
+#>   or buggy. The interface is subject to change. 
+#> ------------------------------------------------------------ 
+#> Gradient evaluation took 7e-06 seconds 
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.07 seconds. 
+#> Adjust your expectations accordingly! 
+#> Begin eta adaptation. 
+#> Iteration:   1 / 250 [  0%]  (Adaptation) 
+#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
+#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
+#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
+#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
+#> Success! Found best value [eta = 1] earlier than expected. 
+#> Begin stochastic gradient ascent. 
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
+#>    100           -6.262             1.000            1.000 
+#>    200           -6.263             0.500            1.000 
+#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
+#> Drawing a sample of size 1000 from the approximate posterior...  
+#> COMPLETED. 
+#> Finished in  0.1 seconds.
+
+fit_vb$summary()
+#> # A tibble: 3 × 7
+#>   variable      mean median    sd   mad     q5      q95
+#>   <chr>        <num>  <num> <num> <num>  <num>    <num>
+#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
+#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
+#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  
+
+# Plot approximate posterior using bayesplot
+mcmc_hist(fit_vb$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+
+# Specifying initial values as a function
+fit_mcmc_w_init_fun <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function() list(theta = runif(1))
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2 <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function(chain_id) {
+    # silly but demonstrates optional use of chain_id
+    list(theta = 1 / (chain_id + 1))
+  }
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.5
+#> 
+#> 
+#> [[2]]
+#> [[2]]$theta
+#> [1] 0.3333333
+#> 
+#> 
+
+# Specifying initial values as a list of lists
+fit_mcmc_w_init_list <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = list(
+    list(theta = 0.75), # chain 1
+    list(theta = 0.25)  # chain 2
+  )
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_optim_w_init_list <- mod$optimize(
+  data = stan_data,
+  seed = 123,
+  init = list(
+    list(theta = 0.75)
+  )
+)
+#> Initial log joint probability = -11.6657 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+fit_optim_w_init_list$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.75
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/CmdStanVB.html b/docs/reference/CmdStanVB.html index 5eb677f7f..51c738a30 100644 --- a/docs/reference/CmdStanVB.html +++ b/docs/reference/CmdStanVB.html @@ -1,77 +1,14 @@ - - - - - - - -CmdStanVB objects — CmdStanVB • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanVB objects — CmdStanVB • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

CmdStanVB objects

- Source: R/fit.R + Source: R/fit.R

A CmdStanVB object is the fitted model object returned by the -$variational() method of a -CmdStanModel object.

+$variational() method of a +CmdStanModel object.

- -

Methods

- +
+

Methods

CmdStanVB objects have the following associated methods, -all of which have their own (linked) documentation pages.

Extract contents of fitted model object

- - - - - - - - -
MethodDescription
$draws()Return approximate posterior draws as a draws_matrix.
$lp()Return the total log probability density (target) computed in the model block of the Stan program.
$lp_approx()Return the log density of the variational approximation to the posterior.
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$code()Return Stan code as a character vector.
- - - - -

Summarize inferences

- - - - -
MethodDescription
$summary()Run posterior::summarise_draws().
$cmdstan_summary()Run and print CmdStan's bin/stansummary.
- - +all of which have their own (linked) documentation pages.

+

Extract contents of fitted model object

+
MethodDescription
$draws()Return approximate posterior draws as a draws_matrix.
$lp()Return the total log probability density (target) computed in the model block of the Stan program.
$lp_approx()Return the log density of the variational approximation to the posterior.
$init()Return user-specified initial values.
$metadata()Return a list of metadata gathered from the CmdStan CSV files.
$code()Return Stan code as a character vector.
+
+

Summarize inferences

+
MethodDescription
$summary()Run posterior::summarise_draws().
$cmdstan_summary()Run and print CmdStan's bin/stansummary.
-

Save fitted model object and temporary files

- - - - - - -
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
$save_latent_dynamics_files()Save diagnostic CSV files to a specified location.
+
+

Save fitted model object and temporary files

+
MethodDescription
$save_object()Save fitted model object to a file.
$save_output_files()Save output CSV files to a specified location.
$save_data_file()Save JSON data file to a specified location.
$save_latent_dynamics_files()Save diagnostic CSV files to a specified location.
+
+

Report run times, console output, return codes

+
MethodDescription
$time()Report the total run time.
$output()Pretty print the output that was printed to the console.
$return_codes()Return the return codes from the CmdStan runs.
+
+

Expose Stan functions and additional methods to R

+
MethodDescription
$expose_functions()Expose Stan functions for use in R.
$init_model_methods()Expose methods for log-probability, gradients, parameter constraining and unconstraining.
$log_prob()Calculate log-prob.
$grad_log_prob()Calculate log-prob and gradient.
$hessian()Calculate log-prob, gradient, and hessian.
$constrain_variables()Transform a set of unconstrained parameter values to the constrained scale.
$unconstrain_variables()Transform a set of parameter values to the unconstrained scale.
$unconstrain_draws()Transform all parameter draws to the unconstrained scale.
$variable_skeleton()Helper function to re-structure a vector of constrained parameter values.
- -

Report run times, console output, return codes

- - - - - -
MethodDescription
$time()Report the total run time.
$output()Pretty print the output that was printed to the console.
$return_codes()Return the return codes from the CmdStan runs.
- - - - -

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other fitted model objects: -CmdStanDiagnose, -CmdStanGQ, -CmdStanMCMC, -CmdStanMLE

+

The Stan and CmdStan documentation:

Other fitted model objects: +CmdStanDiagnose, +CmdStanGQ, +CmdStanMCMC, +CmdStanMLE

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/Rplot002.png b/docs/reference/Rplot002.png index 9e2206dc923b3c26ec67b5c726a540871e830198..522adfde35bfcddf573348b37c266fa69dbf5187 100644 GIT binary patch literal 6854 zcmdU!c~n#9y2gW`Xa$jiGDD?R5RpMp#*nH3914_TL^RB0k}(2dlu)fQ%XI<-0*WnY z6c9nC0Eq^IGKygcgak+ggakrBl0c@rY3n&X_xASm{BiDGH*4*+_u6YG-~PVud7t<9 zzRCF3*-`#8_0K>cko?IL(ClWb7<-xYQz-f!$X-6n<20m-Yu_)lQE$W0v zGzhf)^NmC59W_M=0;va{gdRN;U%WsKe-s{vCUYhd$pTwyn(h7DCrV5{e{!qt_{E`e zE&n%JlRuVk-3wKfJ;(%pQUD?%a@)Nq>Pt7J4q?p6$?`z?VeXLKkx*jZYSOIrmY08ohy{k&a=go2-V5* z5WkR+5N4po5A)48Fc`ns*uzN%kGTsJ6~dHBUVeUU#jT4eDR#hBg`I7_C(@(XI`-l` zqxRV|A03YoQ!_KA&Rm^-VEi(DKO~Y{eEGdln2ZJ0Hi+KvcpV(^j9tc~C)S1K8B>px zB3r_CqWI-=v^fx5PO9V*#A-OB8AH^Bbu4eU7fuz`2b==Ay*zmi^dDcRN`3JINc!td zQd?4GHpzju{eBI;!2kNd%j@}NhTVq(gM#AxjId`t@?7%*l;fGnd7h24f@gyxye;Sx z;>9Qyg7N@^dBl`2wNCPKRj!GgXbUdf2Z$4c$;4ouSKPc9u4A=OnQqw%j3%w{e-yF;DH15=S`gawH}Uv_=6-H zHB{tX^60`r1)plgNw81CPDFfD=DpqSjpwxLum#Y_Gwr8Ei)2Qg!0AVjw55zZNbQdz z=wDOTJZl?1Vo=9kI1i1bP4rgg)ja>so@72N62SzHrfz6S94J&qwU3d6TO^_+1L6X9 zu>h;9TUcPlZlrxd_r|5{UlhxBQ*4g<(6^ZsJCdL=PjqMclqOe|?b}}$dHX0OZ}~0M zEoFL3xb#zMs1*vwVx_|C;~=O$zf~@2Ok!=Wp#U+{SuN;HykX=BKhgk>-ju%6ryPIG zqAW>_({H}tLF4vp_Q~o0GYgqyr{3F|CDIkI#PM>$i}{_Q#%4cwo&;`ItlJ9{=Y-x; zjU_7iMUBb68vA6TZnju}{%0Thr>pq#`h+YzFQBKlccC4QtbA_RFRayA?b6LG3qD-G z(3t+_oY!^0l1WQ<_Gjt>cII!c9^kSU<~hQ>8K7qG`i&@2KT~&pef+7{BQ1?B@#g9s zU0q!r{b5&zVzmN19I8jUOS%#&pY9$ouN~O;ofD}}U1cBwE|_+@aTEvH)qk(4sp*`j zXUYYCvpz`(!_MP|L*P-Lsgs0m&%)TVdt1yT~nOeFtI8uPoYw73t+ zu75sMSbnci&|DJYEsMRZ({!7d;LPgt$AF%7Zu>J8Z%8g|{?s{Ew%CO3XdFY-ZXg@` zz29UGQnF8tyII3XCNAw>n`n$Z3yxLg@L7L0W&sEhB2Mk8ARkx#HF|t z7a5Ss$36hS=-HD}msq`rAyp!aJo|HH`tqcDAW-JAIBNUWx9eAM5w zK8wXFkAPU96stosIc@l<0O<=xRnv{zJZR1uhZS>0iFPTqhb*|{2^3YIh8%jItmf#E zSA&v>gUf~6p@4@rmn^bhcu*XjlMd)QLyViLiyjc{+vx8Jm4TL;L_lG~Jz>XSuw8({ za1#@Q3xZZ>Xf{sTt}+8@8g5gC4osW1+K=Un_32CBwRc6esQ5Z54KU>yI-YGUPG*$E zSkKO`u2b0#4_S|YDnz4Sd#udBU`QS>SJu3xj-FJVwEX*`1#r4cAi>f96jJFw_a9pOL}dWj7wRvNJ| zC=^QS3BNeLQqgtpL&R(kIkYgH-xp2g$4#_-&#JM2A4-?-U7W*3Kc$B6#WVSF{dxy2E2-%Qva z$kg(JQ>I^NG?2QBz6u|4qRa-vV2THw?;CX=+P!;snSN$LX=(02aC85w#TaWhY%bhw z1LRCsHyG^^kjIxFWOfj z*oZMZ$25ZVn5K+IjzdjjqbiWeYwGHh2~#oN8i~HXz885nLjl%xt}y#$fC4l+XL@?N zHpLDCLD48Se>AVZgrol^Nj$361Hp@N*5etFc>cSZP7(5P_5ofb;m1@bKmk!`gR&}2 zP3K{!TKQV3Bi*0&-^XCbQC^zj0fHY3(K^8Q2l<$l>*7(uRuLyl30>0(Mq$#Q?1&)z z2r5W)Q=B~N*8WY9@ZkdW-AMybm_n)4Z|LoY*s4yp7P+=Rco?gx3N5yVbp+qmJAmxk z=auv(swhbI_VLsoYJ^ljHOaKGf`f)v~5+Xn4u-lCx+bs?2nR8z?ALp1Xyt@8_#{B(CFcQ$A<$6kvQMf zG=i_8v2kQC>z3U06-86A(oYL<4dw$-L2`oz#r$CB!qW2j3Eb|}T_hyB-h0gXBDOcX z(ACh$$N`nRIUb2Ta`x=`mRA$U0drkjTRR7b_xw^9l~ERm^Yil_5H=D!+bY_yx_#>+ z@o*pM9sMG_xe@U9(22MaO(G8|JPmMh|q2lsHp+d0rGY@rn#%+)y%Gr z9mZ7)%VRl|xoM|c^~}JE&KEQXN?$GYozcZG3L+M~Cg?r_)GlTE<;!h2xT~@&40dOF zlpXH9HW7M6ZYjz#yQ8C{@}B3dsr#Af+cKg^?asvVkk%F?1DTj``=^EG(K$qA!$AFt zChDXL7y!(dKFhrP@WEIuv~!RFThI-(4Qgv^Yg$^8hx=z|XZw?q((eC-dV6E*ugDwQ zHx(uGMyUJ_4yQsW^i!Mu5uWjR>YRzIlz3T!&dOg0?4>ozfL?gI-uD>j?K|k^MvQLY zdHCK|L$3ImP^!StPWrd1e&75~b9@As^bjWG@VEeo$9Yq8QESb2!wVF#_dwENH{C>s zQW7o(vJl3VyEoZhlTr0~qP*bD3qwTgPfGnH{Xgvz{DWYV)ug|EUU6fP6RF+wJqsag z?IawcW`RQ#n`T9A=uxvkZLd#e|mSUoT%=}wW02fM6L#(RE?>ydK2T6bPs1$dm z1B*)-B{R;-Iwi06`BO~3B<{h3lrRRmfi?`(t#kQ=@B(?p@~0|q5}4>6wIhl+D_9P= z%Ak;szBX{AC)j$M3I2{%%Q+$r^>LESa9-6Fcd~V|q|L(nTW6+$*?!iG?3Bnn*LSFt zuri|8KEX`XR`paTwQODSe3ag9>mZ%YwtYLzSSW$_ zujb1!-i=Z(zbXJpKidy1B_{}Cw{vK7wSR(G343;nlW>l$m22T;M{BSUeGJx;tQsp2 zSp`3<124=goaR47wOG>dTKn*zurYB2^ze+$=(4VnEqEft?|}_X;?szo;FtWZseuQQ zQ2o&T3_ceCN5=wn>rg0_w1v z+-t7tc{0EF)ghiu?fC#yVL*>-G|#B+LVnF5TD~6uO#rFR6w(5QBH*sO6`iud^#)n0 zL_T9)w+y0=s-E1d=|E58zvyDxI&e} zOI+-;$82}Qptufy*E@=>dAS{I&6dU)Ac&)*qk}|YmYOA!dMDD;q;~8Gv*X>n@0f2` zn}I%-N8SLN_gnk!I>Gm8RaKzN-XRt%47eh0aY2LG_#dPJK#G5f?ymcl%%@o)KVdKi z0F=?&Gn9aKA&@R(u+(VoO#q^1p;mrZW;bEInuH_zIWM6wG>>Bosfd8 ze|X55Hbz;sEnPDX8+~5@FBQJmV-$u~W$;>N!J}@zDvcs8g|Sj9Np9i3sO~NK>`lr# zsg#&z_wfOrXwe3=`!eWX8G=qcTda9F(lT?H!)?G_u9$gc6t~iA(fM<}S0Pv)nPIco z=syoBq1RoNRUq0-to5-%zA2V>d9?{tU$PPAlTSgA)j>DThljZi)p3ilCL;o^BvNAT z+HDs*wOrff6Pl0|Ti?L%$-z2N{3Txs;SM?YLj3w%1*2#}7^rE33`+|-;?XC=*MqE` z1c-omfwr8-r>o%NhjWI1_fv|B?D$R_dUUB^d{DBDXb$||TQ-kDDiz0QT3!0e7s>i5 zr~cZ6aK?hes_iE(G)!nKfIoiwN=~oy9ta}lGY}iI8Xt!P+fs|fFhw_7Ls~n9m?u@4 zMG?*Wo2d~!Jw0UYa=e2eHBCEFpC($2)9bx~t{bRdx~e1mqN-)NE1p%;N%v$Yy1z5j zBisX(S|W72mOrAZJUPK;IRD7=Y*Um@$Vip}2=#nsE&19z*m2Y&`H~QxqRXiK+ zzFaHgpd_DO2g)&$xvAz3t%KCm?Fi7ywHA*W)=Bo~O5j_8A9ZG}{rwA6Laa-jNM zbHlx;JxR&ljI;pm%;a5&$j~c^r}qR1Z)k(n*+~-uIuNBCF%Ku{xqGc`17hp zdMV5B*Abl41Sgu8T;`5i#(enJRw7*C;H4GX6;OOaJa=@X`_wVg#(hys)CM%K9Y6U4 zf2|*-4H1m?)5QXIu8McE=^9N6X!RlQmFf32KPLkU+2!tiITj#{23le$5DkWHtCf)* zmw8kb0gs&vz%ZZR3DHj5p=BM)dU;ir;a0kSW<0td{L8?=Id10jD_`3IE3mlSZ46mq z1NPNE$23MwAMnlq{_be?G6ozh`TdtVR-xmJT!^-gcqxtH6q)GOG9@3Gej%bMz;Rao zsY|M_%6T{(&f_&ukr$F@%t|xv-n|Pn?>R3OQcT(;B6jIvt_qHCKaS|sLQujRIsmn( zC_seIwlfGuy7*_^KvcyKJ@I<6XwQmloVw1(%i8U%PNFjDL)9s<<52{8&TvE9>eHGy zT0)SmaEX9enta*FZWGW;k`^-nktNBpw>{GyxdQMF?Nbf)U;n;Xl7OZvMd^)4vr9q1 O=cK(ewCb48mHz|?xeaCj literal 4617 zcmd^Cc~nzppT1$U3P`JUMMxbWSP`MJN`z=#Xk=0AieZVeDg==w2myjR1+5BWQVkF` z$02Hw5I_PUWKqhJfe`sXAS4i!K!7AbfFy=36Xs~Y`D0GcnfW^XXYM)gec$^#?|biW zf1Z2yq`$Z6y6x)#05FCBaO6h-01dB>FN}1O^d9CT-DGsx$NPwG)?J#zbgXV#8}~y% z0st6qSRJ6h4Q4C>z*jHeM-H92Q79iwzZ0Gv(5ev8k|WR&)lj1Icj5{-9(m_S$4#M| z1RIXqy+QwIuw|*0u)FuyZ#Ed9(W+f=&b{Xq z=}vqX7e8uEWpc!n=gM_Aw3{^@LrRe3)8;g6TsXvRE=&@}#PDcpp}~LZsJms}oh;EB z0x>wdwbDg=+Id_nq|q3VZFhQSd{SBv-%@^LU;ZcAY8^PSh&)uwQ+NqzZ+;)z3yA*0~<<^~Uqpl-`Q|Bd^YQ41cR= zStCF@3L7fj1SiNLZGB{H>0pDyI>)01H;$WRaqN8tB{>x_GHHb=UaE#9EXO7|6bt(M z^z&T(5B(W|S~U@g@aXAm)YJwD+Qx;xv&pb`&P2ur#~_2P!LO5y1Kh9grJ<`9WyX=N zSTJ(o3&+NvY58H(foj@Di7R#sa-rVnQ>wTda?j9}hvhDPEQmv5k$Ek?-)IpXm!`|~>dfxAGBvCe&@k>}v1HI6*u$1)^1m{Ya$dX>7q1!c zS3kc9_P=VCc`X;HvAU0wm7E|KPrD|^|BK0yt^HYrxBlAn?KtDWZk|e~p!%}jApplK zdX3zdrg+5T2A*e4wT2Ej?*DUQ^^cV$+gGXAJ*o2^HFXU@bpREYu4;Hdry&NQEV~<; z{twH2n7b8;ugJv_y#|<1q{6sgHu4xO&8#(Aj|v85`MK_}shj4721)!YAUqrHXE8Mr zLW_8^s4E$+$grOosQ=f7>d$RvNKz`Ow1rU;C8#~6roG>lGD|3km~*=|cr^18;i9=s z-JYZpOhoux^3B1vZ=?Q8kBw#%6oV`gHDP(uzTdPri{okb3e`tDMLeLmSNydh^Z^jk z5-8euwtss~j!Xa9zdE7+_g>PW6Cf&9o{I1G9&kH54NA0PQ<_}hQ}vJ%xkvKVB)xXv zyl>{EY^fSoE}`?*I`+hye!`f3muIEe89Yi5i<}ZJkSg^6x-I+JIlQ?FU9u4gKrZze z)QfynvI6ZzfC}yV3{|$_H~m$eHtGYEClcofIZc+baB^f*+m|3C50BacVRcXXR6j@D ztl;ecWvwQ;%|DK55x;KH!RINW^PHxW2pcGtdtj8Dw9$#f+(~QIt=)PiF?WOU!li;V zW?v!^N-H=ZYA=E_ogX%r-U0LxV-3yM@6e+J;*{L<{O|v;>-`0Iyqdz9@cPxJeLG+H zE{UAa`TklS{}S;LxkettoK=QgDVBP)PFb_rXBw0Xjd8J_`O_ay$BqX6ZQABMH;ITa zx4D9^wDqflqwPROq+TX%YG)AG#)%=G)Wjy-R)&SYMP&_gT|Q9^_P0?;1Y!U$>=Lu4 zK87I}>tc3;jHtDb4pl!Z=}^(Rrdc}RbiD`Y3xGuKsm|O36ZL8(f8V`;gg9p1iJ$%$Z zW}xe(Iy9t;P^86?aMDkI$Q%A!x;AtlRC-bXdDe@=SdOnyO&E|AndeYT+_MeH%*&Wa zyt(>QFS;cCz1Qm&l*6^C`0gj@Dyi^8^uDDJ4sN&MvfmzUL}pq{$}Yw?neHDzPmv+jrWyrOfu`jBfu#3x~gUX78=>Iv%?awU{+^dw+sVpNsWHmIc8xLFjy{ zpN^2WHgr~);6jTVYp>|+B;?C&`}QQ3Kp9jGT^8|s0q8I??x>1q3!Be2$MlhSD>3S1 zRH_tQCr2`*J^gnht0k)y)*~ z1kg+pd~+n7r*ykfRl{vz4jspahFmMB{hg|qvX2D05Ynbw>9CaOo|Maxx>JzJj;rw_ zMnkm|{T!|+9J(s{mHV;mhro`{30x}vdfuKT7)q0;6;Gsw{DOAQPY3cW=bSiFmE<>+ zSt4fhc>#pCKE8Wy8kKV*s`lP51~=*-mo=A(2Xd0#1AX;_3l7Igs( zaWbzzlq|@(c93~?fe}`J)0W6p;@@3T(|*cVC~;)BLOzsGxFbL+-`CNW<;A!9~kG=HuDu+82NeJmJ47ThvI*aE?GV} z=c}$92;$%=Ijl;}q@ok+3^;o3H6LDHR96;yVpQqn<>j!NO~zFWdwY91A|N1t*8I2Y zHamBhOv)4EgOw8MTu4>PGA>|&(H2B@kW^<^`Bb^>(sddma#aVi-gCSCDLDk}+H2%K7E)Ysoo>+H#0pWRevt>NC<#Q&ZEwR}t< zOvegTKXpZkmCL86er$b17cEcmUn6JuT~YIr515oKhV>``cz?XOjdfsYrpXqUCTA{A z5m`9icjwEk<_9p35Oocn{Tgc9`L_R(~l~CQ+tPnJl(Y+p) zsf?kglFlONBXJ4-fFdWkab&EZcGt{oiK(c2S}jT%dEhowEcRRVE~2W(UolJAnLft8 zsb2|xa4aOMV2t|IJGYqrW>wq8Wv{qlwnerXc;&;B!`Z3cxt4MZ&(!%%Au*k2ZoC_L z-}cZd*_JaqJ8S0bmoTP#;`LT=Hs5#``UEB5JvjE14o;hSUAfo60eqtRV!ec#++{Va z3%;2;l>y5uDq6mbE5S^Rw70ioRaZ+wJ%T1LxNCW8RVqUp7#tj&E~{!~rp}L_iabi{ zx3e&FurS`Tm5P$O=fx8i_!J5y&Yc4#5C|br0##SGW$F)^pwIvJ0Gf#5Y3C9qRGo*_ z!7CvkAeeI*$AW1WHTSKAN$Y{}@m$=iZK5{h&JloR&q;s}VP3P;!-nHINRS{#`W zcCF=*SCRO=bq%7%3y}FP&Q*9bFt{Bj;t{xVTp`z;e9d>ZNrKhBwuULHX~r(_3_Fh6 zFEJYEf3LF9Jey;Q@Zt$$FZHwprmC`*Zi5Y2Ks5IV+|gSawROr)4$-!5-`nQ)X+zT( zu)I8NsZ8Lx*P6DIGq)uzR@6SGdk4`PR-&*%B8l}q^qrujf%_w=FVuaWII-J4CRuP| zw)cDmb&BB>Ig*Q3f+^ vYSOxr`w@PTMYvuH=DEN5-D95@C#SW5eCer>|B_;}DuW;OKSKT+^( diff --git a/docs/reference/as_draws.CmdStanMCMC.html b/docs/reference/as_draws.CmdStanMCMC.html index 864386a62..6be705dc4 100644 --- a/docs/reference/as_draws.CmdStanMCMC.html +++ b/docs/reference/as_draws.CmdStanMCMC.html @@ -1,77 +1,14 @@ - - - - - - - -Create a draws object from a CmdStanR fitted model object — as_draws.CmdStanMCMC • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Create a draws object from a CmdStanR fitted model object — as_draws.CmdStanMCMC • cmdstanr - - - - - - - - - - - - - - + + -
-
- -
- -
+

Create a draws object from a CmdStanR fitted model object

- Source: R/fit.R + Source: R/fit.R

Create a draws object supported by the posterior package. These -methods are just wrappers around CmdStanR's $draws() +methods are just wrappers around CmdStanR's $draws() method provided for convenience.

- 
# S3 method for CmdStanMCMC
-as_draws(x, ...)
-
-# S3 method for CmdStanMLE
-as_draws(x, ...)
-
-# S3 method for CmdStanVB
-as_draws(x, ...)
+    

+    
# S3 method for CmdStanMCMC
+as_draws(x, ...)
+
+# S3 method for CmdStanMLE
+as_draws(x, ...)
+
+# S3 method for CmdStanVB
+as_draws(x, ...)
+
+# S3 method for CmdStanGQ
+as_draws(x, ...)

+    

 
-# S3 method for CmdStanGQ
-as_draws(x, ...)
+
+

Arguments

+
x
+

A CmdStanR fitted model object.

-

Arguments

- - - - - - - - - - -
x

A CmdStanR fitted model object.

...

Optional arguments passed to the $draws() -method (e.g., variables, inc_warmup, etc.).

-

Details

+
...
+

Optional arguments passed to the $draws() +method (e.g., variables, inc_warmup, etc.).

+
+
+

Details

To subset iterations, chains, or draws, use the -posterior::subset_draws() method after creating the draws object.

- -

Examples

- 
# \dontrun{
-fit <- cmdstanr_example()
-as_draws(fit)
-
#> # A draws_array: 1000 iterations, 4 chains, and 105 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration   1   2   3   4
-#>         1 -67 -65 -65 -67
-#>         2 -65 -66 -65 -64
-#>         3 -67 -68 -65 -67
-#>         4 -65 -67 -66 -68
-#>         5 -64 -67 -68 -66
-#> 
-#> , , variable = alpha
-#> 
-#>          chain
-#> iteration    1     2    3    4
-#>         1 0.46 0.450 0.50 0.49
-#>         2 0.21 0.098 0.30 0.34
-#>         3 0.31 0.712 0.40 0.26
-#>         4 0.45 0.242 0.39 0.56
-#>         5 0.46 0.848 0.28 0.24
-#> 
-#> , , variable = beta[1]
-#> 
-#>          chain
-#> iteration     1     2     3     4
-#>         1 -0.43 -0.55 -0.38 -1.19
-#>         2 -0.64 -0.87 -0.91 -0.76
-#>         3 -1.16 -0.35 -0.31 -0.16
-#>         4 -0.62 -1.15 -0.89 -1.25
-#>         5 -0.56 -0.37 -0.69 -1.10
-#> 
-#> , , variable = beta[2]
-#> 
-#>          chain
-#> iteration     1      2      3      4
-#>         1 -0.33 -0.575 -0.167 -0.291
-#>         2 -0.51  0.026 -0.341 -0.179
-#>         3 -0.36 -0.126 -0.401 -0.464
-#>         4 -0.49 -0.319 -0.023 -0.047
-#>         5 -0.39 -0.397  0.022 -0.209
-#> 
-#> # ... with 995 more iterations, and 101 more variables

-# posterior's as_draws_*() methods will also work
-posterior::as_draws_rvars(fit)
-
#> # A draws_rvars: 1000 iterations, 4 chains, and 4 variables
-#> $lp__: rvar<1000,4>[1] mean ± sd:
-#> [1] -66 ± 1.5 
-#> 
-#> $alpha: rvar<1000,4>[1] mean ± sd:
-#> [1] 0.38 ± 0.22 
-#> 
-#> $beta: rvar<1000,4>[3] mean ± sd:
-#> [1] -0.66 ± 0.25  -0.28 ± 0.23   0.69 ± 0.27 
-#> 
-#> $log_lik: rvar<1000,4>[100] mean ± sd:
-#>   [1] -0.517 ± 0.101  -0.401 ± 0.148  -0.501 ± 0.220  -0.446 ± 0.151 
-#>   [5] -1.187 ± 0.287  -0.586 ± 0.189  -0.640 ± 0.125  -0.276 ± 0.135 
-#>   [9] -0.690 ± 0.166  -0.744 ± 0.235  -0.277 ± 0.125  -0.491 ± 0.234 
-#>  [13] -0.660 ± 0.207  -0.364 ± 0.175  -0.279 ± 0.110  -0.274 ± 0.089 
-#>  [17] -1.597 ± 0.300  -0.478 ± 0.109  -0.232 ± 0.078  -0.114 ± 0.081 
-#>  [21] -0.211 ± 0.090  -0.565 ± 0.147  -0.328 ± 0.140  -0.136 ± 0.068 
-#>  [25] -0.454 ± 0.122  -1.519 ± 0.347  -0.304 ± 0.122  -0.445 ± 0.086 
-#>  [29] -0.723 ± 0.226  -0.693 ± 0.191  -0.483 ± 0.161  -0.423 ± 0.110 
-#>  [33] -0.409 ± 0.128  -0.064 ± 0.052  -0.587 ± 0.182  -0.324 ± 0.133 
-#>  [37] -0.701 ± 0.227  -0.310 ± 0.149  -0.180 ± 0.112  -0.680 ± 0.128 
-#>  [41] -1.138 ± 0.260  -0.931 ± 0.198  -0.414 ± 0.265  -1.176 ± 0.192 
-#>  [45] -0.361 ± 0.119  -0.581 ± 0.130  -0.302 ± 0.128  -0.324 ± 0.085 
-#>  [49] -0.319 ± 0.082  -1.286 ± 0.332  -0.289 ± 0.096  -0.833 ± 0.144 
-#>  [53] -0.399 ± 0.130  -0.373 ± 0.143  -0.384 ± 0.136  -0.319 ± 0.193 
-#>  [57] -0.658 ± 0.119  -0.950 ± 0.352  -1.362 ± 0.345  -0.977 ± 0.162 
-#>  [61] -0.542 ± 0.100  -0.868 ± 0.309  -0.116 ± 0.075  -0.904 ± 0.248 
-#>  [65] -2.034 ± 0.615  -0.507 ± 0.138  -0.276 ± 0.084  -1.065 ± 0.239 
-#>  [69] -0.435 ± 0.086  -0.638 ± 0.236  -0.613 ± 0.206  -0.464 ± 0.171 
-#>  [73] -1.490 ± 0.370  -0.948 ± 0.197  -1.133 ± 0.386  -0.375 ± 0.141 
-#>  [77] -0.880 ± 0.140  -0.491 ± 0.175  -0.766 ± 0.191  -0.540 ± 0.195 
-#>  [81] -0.163 ± 0.101  -0.223 ± 0.139  -0.343 ± 0.084  -0.275 ± 0.094 
-#>  [85] -0.131 ± 0.077  -1.130 ± 0.323  -0.823 ± 0.129  -0.782 ± 0.241 
-#>  [89] -1.283 ± 0.323  -0.260 ± 0.135  -0.386 ± 0.130  -1.504 ± 0.354 
-#>  [93] -0.735 ± 0.217  -0.318 ± 0.090  -0.386 ± 0.112  -1.576 ± 0.295 
-#>  [97] -0.431 ± 0.102  -1.052 ± 0.371  -0.695 ± 0.142  -0.392 ± 0.099 
-#> 
posterior::as_draws_list(fit)
-
#> # A draws_list: 1000 iterations, 4 chains, and 105 variables
-#> 
-#> [chain = 1]
-#> $lp__
-#>  [1] -67 -65 -67 -65 -64 -67 -65 -66 -66 -65
-#> 
-#> $alpha
-#>  [1] 0.46 0.21 0.31 0.45 0.46 0.27 0.34 0.56 0.28 0.52
-#> 
-#> $`beta[1]`
-#>  [1] -0.43 -0.64 -1.16 -0.62 -0.56 -1.13 -0.66 -0.88 -0.42 -0.91
-#> 
-#> $`beta[2]`
-#>  [1] -0.327 -0.506 -0.364 -0.489 -0.387  0.114 -0.647 -0.539  0.094 -0.208
-#> 
-#> 
-#> [chain = 2]
-#> $lp__
-#>  [1] -65 -66 -68 -67 -67 -66 -67 -68 -67 -65
-#> 
-#> $alpha
-#>  [1] 0.450 0.098 0.712 0.242 0.848 0.382 0.178 0.062 0.682 0.147
-#> 
-#> $`beta[1]`
-#>  [1] -0.55 -0.87 -0.35 -1.15 -0.37 -0.23 -1.14 -0.71 -0.93 -0.65
-#> 
-#> $`beta[2]`
-#>  [1] -0.575  0.026 -0.126 -0.319 -0.397 -0.109 -0.369 -0.415 -0.210 -0.334
-#> 
-#> # ... with 990 more iterations, and 2 more chains, and 101 more variables
# }
+posterior::subset_draws() method after creating the draws object.

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example()
+as_draws(fit)
+#> # A draws_array: 1000 iterations, 4 chains, and 105 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration   1   2   3   4
+#>         1 -66 -65 -65 -64
+#>         2 -66 -65 -64 -66
+#>         3 -65 -66 -65 -66
+#>         4 -65 -65 -66 -65
+#>         5 -67 -66 -66 -67
+#> 
+#> , , variable = alpha
+#> 
+#>          chain
+#> iteration    1    2    3     4
+#>         1 0.37 0.21 0.66 0.314
+#>         2 0.34 0.35 0.28 0.748
+#>         3 0.58 0.25 0.41 0.023
+#>         4 0.20 0.37 0.45 0.297
+#>         5 0.24 0.37 0.72 0.328
+#> 
+#> , , variable = beta[1]
+#> 
+#>          chain
+#> iteration     1     2     3     4
+#>         1 -0.54 -0.69 -0.71 -0.55
+#>         2 -0.98 -0.91 -0.67 -0.70
+#>         3 -0.77 -0.30 -0.82 -0.63
+#>         4 -0.41 -0.66 -1.07 -0.75
+#>         5 -0.26 -0.78 -0.72 -0.45
+#> 
+#> , , variable = beta[2]
+#> 
+#>          chain
+#> iteration      1      2      3      4
+#>         1 -0.738 -0.427 -0.183 -0.256
+#>         2 -0.545 -0.076 -0.220 -0.433
+#>         3 -0.168 -0.494 -0.400 -0.201
+#>         4 -0.281 -0.029 -0.388  0.072
+#>         5  0.054 -0.460 -0.027  0.122
+#> 
+#> # ... with 995 more iterations, and 101 more variables
+
+# posterior's as_draws_*() methods will also work
+posterior::as_draws_rvars(fit)
+#> # A draws_rvars: 1000 iterations, 4 chains, and 4 variables
+#> $lp__: rvar<1000,4>[1] mean ± sd:
+#> [1] -66 ± 1.4 
+#> 
+#> $alpha: rvar<1000,4>[1] mean ± sd:
+#> [1] 0.37 ± 0.22 
+#> 
+#> $beta: rvar<1000,4>[3] mean ± sd:
+#> [1] -0.67 ± 0.25  -0.27 ± 0.22   0.67 ± 0.27 
+#> 
+#> $log_lik: rvar<1000,4>[100] mean ± sd:
+#>   [1] -0.517 ± 0.101  -0.404 ± 0.151  -0.497 ± 0.215  -0.451 ± 0.152 
+#>   [5] -1.177 ± 0.282  -0.595 ± 0.194  -0.639 ± 0.124  -0.281 ± 0.134 
+#>   [9] -0.696 ± 0.170  -0.736 ± 0.229  -0.284 ± 0.127  -0.503 ± 0.249 
+#>  [13] -0.654 ± 0.209  -0.361 ± 0.173  -0.281 ± 0.106  -0.278 ± 0.087 
+#>  [17] -1.585 ± 0.286  -0.482 ± 0.109  -0.234 ± 0.076  -0.114 ± 0.078 
+#>  [21] -0.214 ± 0.088  -0.572 ± 0.150  -0.333 ± 0.139  -0.137 ± 0.067 
+#>  [25] -0.455 ± 0.121  -1.516 ± 0.346  -0.310 ± 0.123  -0.447 ± 0.085 
+#>  [29] -0.723 ± 0.222  -0.697 ± 0.195  -0.492 ± 0.165  -0.426 ± 0.108 
+#>  [33] -0.410 ± 0.128  -0.064 ± 0.050  -0.584 ± 0.182  -0.327 ± 0.132 
+#>  [37] -0.697 ± 0.220  -0.313 ± 0.145  -0.180 ± 0.109  -0.682 ± 0.131 
+#>  [41] -1.124 ± 0.254  -0.931 ± 0.201  -0.413 ± 0.269  -1.172 ± 0.186 
+#>  [45] -0.360 ± 0.119  -0.582 ± 0.132  -0.307 ± 0.130  -0.326 ± 0.084 
+#>  [49] -0.321 ± 0.080  -1.289 ± 0.340  -0.290 ± 0.095  -0.834 ± 0.146 
+#>  [53] -0.405 ± 0.132  -0.371 ± 0.144  -0.387 ± 0.137  -0.322 ± 0.189 
+#>  [57] -0.657 ± 0.118  -0.947 ± 0.345  -1.358 ± 0.345  -0.977 ± 0.161 
+#>  [61] -0.542 ± 0.101  -0.881 ± 0.324  -0.118 ± 0.073  -0.899 ± 0.249 
+#>  [65] -1.999 ± 0.597  -0.509 ± 0.135  -0.278 ± 0.082  -1.056 ± 0.235 
+#>  [69] -0.437 ± 0.085  -0.641 ± 0.239  -0.608 ± 0.208  -0.462 ± 0.171 
+#>  [73] -1.479 ± 0.361  -0.947 ± 0.192  -1.146 ± 0.401  -0.373 ± 0.139 
+#>  [77] -0.875 ± 0.138  -0.489 ± 0.178  -0.762 ± 0.186  -0.544 ± 0.202 
+#>  [81] -0.164 ± 0.099  -0.225 ± 0.139  -0.345 ± 0.082  -0.277 ± 0.091 
+#>  [85] -0.131 ± 0.075  -1.126 ± 0.328  -0.822 ± 0.130  -0.773 ± 0.243 
+#>  [89] -1.276 ± 0.323  -0.261 ± 0.132  -0.386 ± 0.129  -1.490 ± 0.344 
+#>  [93] -0.740 ± 0.216  -0.319 ± 0.088  -0.391 ± 0.113  -1.566 ± 0.280 
+#>  [97] -0.433 ± 0.103  -1.050 ± 0.364  -0.693 ± 0.143  -0.392 ± 0.099 
+#> 
+posterior::as_draws_list(fit)
+#> # A draws_list: 1000 iterations, 4 chains, and 105 variables
+#> 
+#> [chain = 1]
+#> $lp__
+#>  [1] -66 -66 -65 -65 -67 -65 -64 -64 -66 -64
+#> 
+#> $alpha
+#>  [1] 0.37 0.34 0.58 0.20 0.24 0.48 0.36 0.35 0.21 0.35
+#> 
+#> $`beta[1]`
+#>  [1] -0.54 -0.98 -0.77 -0.41 -0.26 -0.80 -0.62 -0.74 -0.55 -0.51
+#> 
+#> $`beta[2]`
+#>  [1] -0.738 -0.545 -0.168 -0.281  0.054 -0.315 -0.204 -0.103  0.067 -0.446
+#> 
+#> 
+#> [chain = 2]
+#> $lp__
+#>  [1] -65 -65 -66 -65 -66 -68 -66 -68 -65 -65
+#> 
+#> $alpha
+#>  [1]  0.211  0.348  0.245  0.371  0.366  0.050  0.135 -0.055  0.417  0.247
+#> 
+#> $`beta[1]`
+#>  [1] -0.69 -0.91 -0.30 -0.66 -0.78 -0.47 -0.86 -0.85 -0.32 -0.72
+#> 
+#> $`beta[2]`
+#>  [1] -0.427 -0.076 -0.494 -0.029 -0.460  0.199 -0.272  0.096 -0.201 -0.249
+#> 
+#> # ... with 990 more iterations, and 2 more chains, and 101 more variables
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/as_mcmc.list.html b/docs/reference/as_mcmc.list.html index 363137cd0..62098ec93 100644 --- a/docs/reference/as_mcmc.list.html +++ b/docs/reference/as_mcmc.list.html @@ -1,81 +1,18 @@ - - - - - - - -Convert CmdStanMCMC to mcmc.list — as_mcmc.list • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Convert CmdStanMCMC to mcmc.list — as_mcmc.list • cmdstanr - - - - - - - - - - - + + - - - -
-
- -
- -
+

Convert CmdStanMCMC to mcmc.list

- Source: R/utils.R + Source: R/utils.R
@@ -191,53 +119,53 @@

Convert CmdStanMCMC to mcmc.list

compatibility.

- 
as_mcmc.list(x)
- -

Arguments

- - - - - - -
x

A CmdStanMCMC object.

+
+ 
as_mcmc.list(x)
+
-

Value

+
+

Arguments

+
x
+

A CmdStanMCMC object.

-

An mcmc.list object compatible with the coda package.

+
+
+

Value

+ -

Examples

- 
# \dontrun{
-fit <- cmdstanr_example()
-x <- as_mcmc.list(fit)
-# }
+
An mcmc.list object compatible with the coda package.

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example()
+x <- as_mcmc.list(fit)
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/cmdstan_default_install_path.html b/docs/reference/cmdstan_default_install_path.html index 252412389..45d262399 100644 --- a/docs/reference/cmdstan_default_install_path.html +++ b/docs/reference/cmdstan_default_install_path.html @@ -1,75 +1,12 @@ - - - - - - - -cmdstan_default_install_path — cmdstan_default_install_path • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -cmdstan_default_install_path — cmdstan_default_install_path • cmdstanr - + + - - - -
-
- -
- -
+

cmdstan_default_install_path

- Source: R/path.R + Source: R/path.R
-

Path to where install_cmdstan() with default settings installs CmdStan.

+

Path to where install_cmdstan() with default settings installs CmdStan.

+
+ +
+ 
cmdstan_default_install_path(old = FALSE, wsl = FALSE)
- 
cmdstan_default_install_path(old = FALSE)
+
+

Arguments

+
old
+

Should the old default path (.cmdstanr) be used instead of the new +one (.cmdstan)? Defaults to FALSE and may be removed in a future release.

-

Arguments

- - - - - - -
old

Should the old default path (.cmdstanr) be used instead of the new -one (.cmdstan)? Defaults to FALSE and may be removed in a future release.

-

Value

+
wsl
+

Return the directory for WSL installations?

-

The installation path.

+
+
+

Value

+ + +

The installation path.

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/cmdstan_default_path.html b/docs/reference/cmdstan_default_path.html index ef1e8bda0..624f12f6f 100644 --- a/docs/reference/cmdstan_default_path.html +++ b/docs/reference/cmdstan_default_path.html @@ -1,76 +1,13 @@ - - - - - - - -cmdstan_default_path — cmdstan_default_path • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -cmdstan_default_path — cmdstan_default_path • cmdstanr - - + + - - -
-
- -
- -
+

cmdstan_default_path

- Source: R/path.R + Source: R/path.R
@@ -181,47 +109,56 @@

cmdstan_default_path

version.

- 
cmdstan_default_path(old = FALSE)
+
+ 
cmdstan_default_path(old = FALSE, dir = NULL)
+
+ +
+

Arguments

+
old
+

See cmdstan_default_install_path().

+ -

Arguments

- - - - - - -
old

See cmdstan_default_install_path().

+
dir
+

Path to a custom install folder with CmdStan installations.

-

Value

+
+
+

Value

+ -

Path to the CmdStan installation with the most recent release +

Path to the CmdStan installation with the most recent release version, or NULL if no installation found.

+
+
+

Details

+

For Windows systems with WSL CmdStan installs, if there are side-by-side WSL +and native installs with the same version then the WSL is preferred. +Otherwise, the most recent release is chosen, regardless of whether it is +native or WSL.

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/cmdstan_model-1.png b/docs/reference/cmdstan_model-1.png index d7b8b97f8c19824bd7f4936ec52513cc8fae745f..27fd88217456b4a7d0da11a09599ea685749e06d 100644 GIT binary patch literal 15795 zcmeHu2UJtp+IGN#j)ie-BOr{@Y!pF2YOo9#5KyYrpfn{ykPaaPE90P|AR=8wL;|73 z&;kTSsRDx#Is~Ky2oPElNJ#ki!I^c}x7NRIzw5jATQ`fW6HZRq=Pl3sJn!DQf5Fsn z+jnB$!CKDfQHN+A?_Ee23`BFc8_N~X-M8I^%$xx$Wtq&!(VF}v-MIJmm5mR%TrqH zBa!fLZ0GfZTs=H8Pj46R)9)=h4&OU?izRd|@y(4K*+Lg!J@u?$U7d-cUhnYx-mvN!BbtBUKM; zhncpex%fmTzR9YMma?HP9}LFfDD?Zz$9sEw)zVrBqQ4Z7$@b~3Z_9GU7sVtLd%L>c z7I{ItL43s4zsBRYZ{Hqv^3)v?iDZXF7VB*)C@7elJn84C*b3Pn_Pysdk6Gc_!2J>$ zERv`<`x;H!Sku@m6I^g)(%{TWBrY?Qy>|oGV-(OIw zKc0&x*x)!xl%dZFOTKAC;4+n^!+)E?2YAS`)nW^~wQ^}0DcSg)@o*}#O_LJ6{q zmg`SS`y8hzVDOtc@~r9GH}Z((&B2x|4VpHA+QAVAcfY(d_@xwZ8=VhYFUI1at$+?PG(Tf{q1lM%p|dV4!o?dZ!mQ{f#?iC?)nJtwHH8 zQrO_|ovPj*g84alDI3C8&iVl^U&?xasgp7orO?J?i-<^96-jo}doeGLDPb$0*Uy2* zpZ{XWY4!)`Q$*)qK$Sm0$+8!mS;d;(|8SB4mk>nH1zFQs#58RNW-GV~2DQZ8Dik5L zz(6|Ytu8fDD{C*+;Z`$5QK%l4E*N`Nh0Lt5{)<>!i)d?nWqun+#zy;OZs_WZN!J0( zI0*3aw*nZvBW6SOuY%Xh=YEEU>F-?#VgKI@g?|kC??lo+PG5Uw%q~jA92tK%Sv%*!G0ai>eORRSV z^HeVTp=jei2=_OY!)9&e(gD`sjMTBBU{BRgCG}!p`wv%uKgRtJKK4I&`kK`?m*q~9 z&1zY392>6Ws1Xe}rm78|F!<_vr-%*7&Dr6@y=~lwXfvMVD5*n&3KDMc)DvOYp2bo_ zV6#QtPt?lg^h$ZiDg){__d1`Ne(R|S_0Yd6^6(_`21hS~0_l2z~eh4W*6{)E3>+l2p1y4JmF{FK1&GODcEhDF)Y z@+Y@%-%d>a%#$2eZ7RbPyR}(+)w3l<8F%{ObTY;&QppsBX4=xymGz-&DfST0DC?oo zm>7vmb$~d4N^F*vG#zqU^W@35%VEz^;w9Gu0!Y?2HU!brIRrxgl1w|G zY28z`rg!%}Xw1|Ja?}8pZyQ@!{=myYiVNR!R!^c{SO2VWiWnbRA84tzpCTy5im6#q zw_&iKl)DViwFabcuAD7h8Sqq3+FX9Nq|0nSiV4@S1;z5Or=b7MwZEIa|1X)&iv}Fb zvEas@~!R=$z~g4FOAmo(Yd znHg`IbQS4jhx%KLBf7?K6e?nDTYf%s@0FZT*^wDHTl>?=K7DFPc5n!smw ze@?tBr6Gwcl-oBCqz5?|K34JLVT1IGYE8>bGoW<8LQ0oO0oeIFw)YefQc_ZYAO(0{ zU0t0jf3P*I(%$rz2)CW33qw=8Cv2yu_b=*hT=(w(qiRq<`YQ< zXFHwS%#&OPi-1AdAo`)z)?#%WKHXloAtvPUV#?kJkcbHsLSe(;qFPZ{u7x zrCU*Ww-G5#NtK4BU!sUZ&0q`~P(R5BwaOy5ajALN*T$uD61s13;0bNGHan9>5ZktR zwma6AD#d@xJPZ)c8**}ALfj2%@Qa(p6Wml_g_+cU5j6FH^o0$w%5ptdOR`qyD&nBQ zYjA1+9|Puvs*ZCF<`Cbuvbb?d2>K^I6t?GUui z&O%@+1O&Q&mCdOiONjub7eAOwfXn@-)$6|zZ~xBRgj!i~adCf;-8{qe))yXbC5dnF zWtEx0^GR)zhhy{Ym~Ax^&56Ga!5mDy!yXLpDk0G&tgupcJB z2s5v%+jqP{PtPB2aqQd{YJJc6k`V0KONMu)0V!AZ=gxqEs2!v8visbL301~-qWJ9v zY=Iy502E64P@8Ux2Gxq%3BStdp2&yeP9|89hBe@j%7;DsdC<*AKx@0fpT`YMU><0A zCgNKN9No1huV_M5=$oM@y#q)3NE*CeuTl+^vG>hZ$P)5vdsGG*TI~Xk2OKL z^iQ9lSrb_>+p_p*3cmu8=T~0;j=(JqGDGz(>TZTr%|q3xiWC;+D_A;8A@wtV+OqV8 zhO@)bl~MuIgjId{c7Q|jd!8zjtyP0cY0j>kOULic#F`1i@*SX};3M58L`nLwi3dva z;Iz*_VeUe4`2_p?-q{{yI-f1jx$y;BYLAa)n3 zV^Jn;?N|K=ipaUy%VD5N;-H?UF-in=eda@$US9X-ykE98ec~NwemZv?bUm&$<8YpJ z73&nq<4ATl4}I-s5E_k!`bb4yC@^1w$D5j(PR`Ee)9;r|O|%u^0|8lbo*nrvLIcB| z0*pKb*eR{Bpx~lcZ-+n-PD`&kT?C4ON`j>o_5J&sqVm7w7Zq7E7>wKA{$W>KT!yJP z#>$fAHbdV2Tj$ySi$;G!BS;T}_P^V5jiDETIxbTO&ffh=yWXS&a@jvUL;p%z<;jzbC}sftPUBs+U&w`KjpQSTD-EF`MQjpg zu3L~2IFrKp+0hW02d|QKV5i>+l&4BAx4ssddCaV}VoEK3EBag0s(3IpD=?MATxqra z*4NhG_}`w}Up?jh=Tq_c$5t#@6krIg1^8PCQq?M&DY!k7%i!IF2QfLa2R1)m@`^aP zX@l3PhXq^ZEk{TCOD+V8GLsqB8EO8O*x@Yh@S_C3Dz za$nAO$U87|apm=tT|ewtScoy6VS96UwKqH-JCw+fi+QCO9d~cv(ta$BK}>6@WG;lv zU}$I%HZNYR2eA)c8)UyB>!CwJ0OJA0JVnwVgKfkF{@KWT z<*AU<34I0V~6V81Zzw9`CEQ|hri5xAL##X=Ae+5mzN7IzH#(fE@>c;76`_k z^}3=$w{3j>5yJ8a-;-6zX0zKpNpjZfcB&={nPm25*99^?*jClT1KmXX;2*O3{r&xS zz3B)YpESB+WgBt$RM7US*N5w*Que93yNWpuw91I=_rE|ij=-v<_~sqVwq)a;{C00c zd6EX&+p_m7wQqw{CsC#+ff_vaD*XAv-?_RQ2x!|iRCrc;p;v7k|<*N6AowEYkKmsL%v1lu@ zF`{NCMEoyYO((vT)Ra%??rt|{PbQeu=MH_zrtN^SZ$jRVIQ){RFyTFv-4|o!1G@^t z0onv!-x)y@uHfto!fQJsZvF60MeKSCo%LtqZ}VCs^S_sXa`9hj3r}<))bi0sb!l2Q z3@a67TB8DGF`aD}8}BdNGxrQU08`Lk^P&u*awRr3Z0N~=d;IYO#c2hywaxoI1NXs9 z)^ZdUBcpv3;bHGPb)YHNHB37)CLk#$yjxNKhc!gWPx3ts0_5EbJFmbEFv6#UD9Hb? z4b%_JdME(fFJ{XsO&y}&WbJQKbaP-77psxi!`M&O;zaI!)rG=IoHh}>lN=kI0+w== z^j~iJ7SvB;O8e_?*)}P-BU(293r7U_ocfH^fj#gc0_S=j}aoskeK}=vb22)TC4_(2%WpUcGdrw zM0(41UrY4-yzs&TUGhDn*m<)1^`ja5jP)6LdD2wRfYZ_1HwG?*_Mo=qT%XLA4O#hA z-`mN{eme4SA@U2=9jCgyR-*JTZ`~Tbf}yTcO|pESvzfN=Rf9zbZ;kwCSIg(cDwB|Vlb`1NfEcQ zw#svK+Tt*&6P|&GF#(^Ghx+DHo0=4FBYgTN_m#46Wl4@TcUI_{NxtWbH6-P}q4EU} zl1${H+^UBG@#sF6fOv2W{VE?n3q{vI>_GEyBsQ8qa_X=G!A9jm^E7Ef@= zMCU6useSYbaDBzG&b{H`HbUDU_ayq=yTbRvQ%uh&Ci7uUe$8XrdA&A6kBUV) z_0X{E5cW~q5B#Vr7i1k%xkpzn_DLt_0=PrNxWIczOvFhmb1LuM%<+T=<&`4#VK`B% zSi7h+mj_N~rgCI5_DDXtWXwRGu|y%%2#zkR#+~15ix@baA>>fOJ$+(|d^;A4z3(kd zRn<|nHyU!&t*uPv8;#H%iyU?GQ{;qM;cVojuJr{)*=U@MW=Upu-1%IijF?A{_(|~e%R=aNiG(EctP^|mJ3UF3 z*{qN)S9#8bFY^>l)XtOCRgMsf@6zoN2C0n+DNVUon1PltsgLSY=&u25Nqf{i)Km9M z{lK0yFY+cH8K*okA)zUcwmL(7vXf{XiY+ER-1?-KU!w3qFmU|&=c{vKk$7ojzk+}} zlPyEE8-Lyxi-gM=w(OV;2)`hjzSl8DL2uIEP&T3aE04UXcBJZpxnlI^-jh4hv0gD- zMa`Obhbl-*yLh}?^0U3eukbrGaTPTkqMS{7Vy4x;kX*O}JIxkeJf(d^8b|K610LEc zud5LL=y*4e?4X#r_+@tXKAY53;@&59oBEpgDM~`ZdwSu8>+70FMIB}C61v`o(cLd6 z>%>I~eIl~;r~2zrcELA37-U&z6ZXz}_q@v*XzDNl%Miyt@3U3$ad!R*i&aQ%_+Y$U zZ)tFY&Z}p5hJrtOswIVf_lCRAKZ;0&`816B zdXntaz5BYRe4S6bHzy~6rKWU|T{Yma+FS7^N$R9_@%PH9;I+tRYrzlwlA!AGhD^p# zBzwOyzb!mePQCYF&}94dD}x4@nNqs%Xq=54VvmB|(=Cc6FY0%{|M(zW=51W2{>_V^ zqr99=;bzO!?=X~j5{;d5G0EK|XMi+rb-3|qcXy$iaT%GsP}+jc?H;YVGmbi(h{~%_ zH`8TV#*E;za~r0*+pe?Zze6DHj{&;Ot9WBu?N05cX-dnT$uO^cBU6xwnM?ON3y*oqMdM{1(u$D4M=-j0f;BQaCgEFQrHOAaBUao5mkm!3Jc z6v5uagmR+c{0TF^iH!J}f(n`CaYJ}wteIX2g|Xq#D}%HsHj6cSYl~uCRh53T9&aic zCs-CSSJXd1CFc<=kCaXoHWmg>_vj9U58d56_CU3{iy129WHPjF0APn`*x}8i96Svo5Eve0V?eT+B)>ZpBbqW6O(-6I_SIu_SbJ4+#aGpk=fD zf&^;@gQ@rXZQL+ZfdtM)6G9E^j*#XcA7(3<;zH?{W?B9jr}xDPsMo8lrJp%<(N>yt2k6_-!DnJd>gCgt-$5=jSd2D<^ZjF71uMZ#bz z{cPj*4mltRH21fU!rfZJCu|U(oY`I)6)YYx1yH! zAh1WNeD=d@(gx+Y6R&V9eRRFz&RzH1I|nl^WNmMA6AG zpQwu3aOkbUA}RgUYeWlaIvq@t476>MNrsWbBUj2r<9 zo>ufczEnla!O_~I_Vwu=%$9aqZ`Wgt9NM)?&ptx1ULfH zhEwumi&FSp)LNxEe$7khyzIkCDHO@hI_J_8aqp4?i1`6dr3MK0-r>n=n1y2nuP2Q^ z<*^U6KR_K>0AF(4imkG#GWDI6Y*N^;IPF|a-;LV+WAqJg%7J~zxSs3fBxzI4(WiZ( z3<6Y91W&grAGj+iNo{R?vSG^}nSD6nMMKFD8H6wR;fI->JSjR750u_@^E}cUJ3LA3VvG96WfFm1mwEkrKqBR>C6Y&Kh;w9QK_Oo} zh)Rd*f7paEa&3qrlEHw5VKwJI?)2kMWO}r|`Wds;)sQa6V})UGJ;KkoPf|N)Dm~}2 zn|N29&N-K5(L)qq7;9!Fdpte5`Ju+Z^W8_SU4e<0ms@bDAbv(($C@^PimeK!ZtkIX zDp#1?-|?;6+4kgq@fwHtmp7;)o)rT2?;jZWhDcNe-tY|6gu#CM(v#!}?!QuIl<{Q> z{5;~t_YXij)cFf`Of?O1qtilSClG))cE2$Qhjwf<<6EKsy8md#-mcG7@<_b9!K`#E}w?HPxPe z2l9Xx(%r(OtP1*YtNi5^Iy@~Btjz|i;_`y0?rjsFZw{(dMRnvr2UYTB`mmh-HRY_XMt!D7cI zv-3NnaaKW67j`wiQLruZbFwwm!Q3AWz8UiFP$W;{`o#f24a4ql-<#M_6&x*@s!Ek27 zsY`&3V2<_i$rK*>de6IS*Dn_VN(u3_b1`A8Y-H1PG0iE`Llg#cLI7o5^;&#?liiUx zb|}S@lnAAQ)N-96M53M=WdPkU_9{xTArihaB(rY|CMXExnMZGtm*!M{^*s2g6qGC1 z&}cJI>mZN=V5FLv8tNG{0@&U;l3hPtEU#=5FC%?m5fo?DpqV~!(Rang^tb7c8Ylt= zW$eyQQ!p$5(a-DJfa7@Q!I1PDwOH*IkaKkh*dn-D5k$;>?Lb#>TNgNz=x$=2>yRK* zI#(mE>QeK-MyJR@kK%n%YmFO?e9mwNmb6cC7$6nUy&P;SvM6%9RM}yP-Mu++`lhK&YOvOT9qatWCjK zG(dL9u_ekG$167dy8Ecuymt8PZK(%$$vCzL?itW^fzjy?gVMBs!^GZ0RycncjpOyk z%7ARYP#stt^0i6%E<30ibD`pL53uT~YeB#?3(he;^FUUwQOy(jf;eEVq z+qRAVc1F|$08r)1s1noSAyg+~MjoQWTmThhFO>`pU5)~)gn#pat{~U{GEy?p)%6ZT zpy;EX8dSk>On~FUV@Bp1m83Y~5K#iM_&JDY)#j5fX`VXZNn>QPZStB22L~m4k(FT~ zTpo{u5(iT56=d`CWkHn-DK0dnOZh(mbsCUl{xUsXnzk~g5i6rX3-i-k84{(_;j1A) z145ende>vhQWT{Q1a+(oa=F1q1tcsMKL;h5O5uxMAoJdrgK$H{ynp{*hdZ8bTjqaF z4@fDHR+@+yz@(eQz->?5fdUc1DeJ6IEbjAN<6<%hk{toXlTYzCpCL0Vp$v>du3QNp z)(DpdwG8}i*>kK86xw*-(qtU37}p3OlU-U91aw6yxMl2v^&f$pH3Ndf#>QpP22Kqlhy)q9aAu03s+TBE8I5Lsbxwjs;PWUIL*6b=6S>G%)l~R73&+ zDFGn_a2Uae1T`dpkOU3I5FipwkPwplo;dfP-#&Zq&W^kH?%v;dFt5*f&-=dL`uUb~ zt{-tYr1kekP$B|R>xBaR5PZ!sG zyY$Y;fz7#*dk$-_(NNH;x%Fzv^+l~`G)^7evOi`2zU2Mi9X+~v+qQ4F-r^Wvjj?ZL zF`rY1;#NO7@xzVyvka0!B9+hTDK&BKmNdPLA0HYoTO_CJ3;t}jMy0YzRrqA}#Dr3O zM})*pzIlGJPZ{Td*>U>Xlm_{Y3fuvng8YBxxh+CyI6vjID+XQJS)ES^iQdQi3ZBwl z;(a2br^TsfwK8v>fHJ>XJu!;HOYaF+gUO_Ii}Wu16sU|BHF#8Ii$pklN%;nNN6YmW zmG}NSO2#`*5xsX!;YcpT05Z(y=44w27$E4!ubSm!@e*%%Oggei4<4<{(ELf(ezmju;_-;a+5|Yx|kM2UQFK?P>*fM9``G@z$ ze^M!LiQ#x~j?;FPZsg5}V90R|!c}04>_5Fzg>%HGth(hDShrI}Jz@CP#m>;s zNQnwwDk>$STBJpIDPFkZKc5eaYMm%cCb-%~?CKvtOKs@Wc5Jq4C!Ws=mP^(_3v60V zb`kyxh^L7U&;z6)`5lj5xK4#Mc+b4-s20Kkj4|Hnwit~6y+D7{#{*y((D2Klm>>zg zc*2RNTFSlJG~b@p#IIukF+5BK`ddNl>q8}m_oAP0MJPNTA7~Da@ZAA4yk_#5H@()e z9@xNg^<%2K9#!Dz6Fy+?B_6u?Xn)CIGwQV!1hq8^{N;N8yY>GC%tzmd^NFU1;pMn? z1}c^ULfftb=CWTz2}dg@x(F!v&N^V8n(<@dbmB^&iuqY}B5c0afz-dzO#dL6|J>z5 z!gA=I@rmO>AubC;4V~4(Ozmr3`1WEUjJ3n*z>3Z9(T@;S@H?mKy?0Of)^j(431g*2 zdTpH#*{%uGHoy{DkG!JBAO5!F?W!;1?f+akd|{~iUq$>eHX7KkZ4cR;wLy$72GIC2 zN~jk??kD%bz^ShesD1b{wMm>3{GjzbnR8Niy)e*4m?n3J4_g2#AEU4A!>{+^eWyO< z>mpQV4FbTnyaeF;hzn`Z#lt`m26|`D8aN#P?RfsbN=Uyq82w))Ud&hwB?+VW4+A#f z3tA0R2@Q5`DK1@dut++z6&T9+*$VvViKw7fA@)4qY7GeFNuEGbp_S3~_+al~@~PK? zPM76f3>{!6YAFVPCD8v%$`%M`^MEtP#IicBssMNq!i!GZ$NL8Mi2RiomDToPsuCmU zbT$!=sZiV5Z)4u9iWk|qSkV?cs;Rh;V@u@#3Q;e&t@Iy$6-*b|E#@y&%0oIt?XpW> z2X6(He)I1nv@ety{!1Fnza#JRPy@v~wY(P_C%!Kd-k=G+=y^W8CWaZviLF3pmDIiK z_*9Tfg7_3~3vF8E7}KZTTLH5E8lFn2p}Be$h7Efu!U|RD+4%eF2YQ1Va4htr*uZ8} zwPW(wcHvui-`ZWhQzuDqp53nt@9aPJERkL+a<9sco=@a~SQc6jD}(wx(I+aZRe0k( z{1C50BT(*fU#fih%}C|hHsr83wX=tQM9&iw)jSxj2%@kSm~VToTA5ntMdx`9Y6DiD zuK)trQb6*I3M&Ovu|aJC68wI1@TB!m3sd;yL4qqFG2$)O5T^7G7g)bh`1jUl|I*7r zaj8GrQO@C~y0kLUo!2|WcAu-`3p+1nnar0|PR#DS&`_{zD%2%vPaQB>pr8BS#^eyE z2zWh9I{Han1FK5RkCSTzdRUhK`1cjJ|I)(y|HS`K6}A5y$xnDF0ehFS-_*dFUu=}5 zmf_)0OLS>qp+U9-8Z1fnpuYmO-Bela(td(Seh^Z}+U@rMPd^e>hi|Ri4nVkQ3;2P~ z-e=>*b39O2$H|H&6y1~|-}V1sPk}332y#=?O)KGv0%SS{|!};z7MASud!7 z+G63HDk;s+M9oT1Bs5E|oXXTb?ZPyxz#mvhZ~>)Z+iSQdm2^nZSnfjdAh^cYjUEwJ z@nIMUmHvb2|97rT|3O99XLB_UQ#2Y9Oh@?z*I0&YeI@^WllvZ6tvAS#r3MAg9!PPmf@W&*3K zn(77t>_54K2g$?mu19A;6#qNAnqS6);jH5IqnX#Ac!msT0xJLQ6;P{z%a^)@r5F`P zlD?TfDQ|c}ZJaeRM*X8qe2@2eP>-QPTz6cP*JF@x3wXBv#Id06PwU1i1Ih4uXe_8B z7IuCpDILy6!1aamq5qTF#vfEV+zE%Eya4}le1~}ALqkVZJ-}O&W5O*}%NF{Qz539L zq4-yDPMdUawW~|=0eT6JVSAt1;$wKz#Z}~?mm^B|qc2x-uY!{H${)MBbLyA{su5(F zR+bN5vvOxU)S2JKDD-=7QLlr~f`LvPdUEH-!5Im^4G8OPV z1pU<)1br0`Dt($op)?3V106~mZGY)Ma$#F>Z}rJ@OzZ4dyHC8U zjNNblcfwM-P?6_$Yptl_*pF8O1 zw3mYV3T{bif9_?t+k8%zUpRgJxp$oD1X~6K)n#0gzcMH)S@#)x{`wXHOQUB2O|OJWFzO4gJr89g1j z!Nr>8Qp0wyt{Ul{m9&RP&`gZYtJ6lEIwY(pQY)b(RStGVmaIoJqSTUJ<6-@2U`eRXu*nR>2@)@~OpqU6m6 z!dZrq)GS=O@4QO%?#)QcW&6@$yWAXyg62l@mtG4@U*^CtDIjtSz=&DmmmJ{5z*5op3KRQZ?yB5aarJy?8UDY}HGl%O&f1V** zKev%N&FP$$@^Lq1t*9Dw?VrP_F(cF!ocwCk2rB%hj68U2=BF?jscb=(8T9ngE8B@| zhcWW{?$IFQWdTpv-P!(wJnd(T`LFC}t7*T+gEc(nGYP%gQvA zGe%6b=YB+Q>?%*B-*dK++8~3g46V+Y_H16BCCH*t?&5Mozz(-^C0A<12l>jm)ZK^e zcj{>LJfO&Q4NuPQ{tKPB+)lD6|bR=dbax+1ooZS2owWkwa=%_ohZ^#}*;j!KPi- z8h!W9YEqN|*@v|U+|x@-#+>c_H{N4h*oiZ5V8oXpl#FoxMc7R?f+_b16yPW}(xhhx zPFEyr>2#VJJzp|%CnLU+tE&|nr9J0B*x2)|IX&lfmk2w=ip=1YY2?}JDVL6<-mg{Z ztbIl4qjjBVRc#$gZm{Tlt{5wE{_1pmnk9@YNztax{_m+;oY;f#yMD;}viI#MPgZnB zU)>{fN{?3T>? z_ABNlqhN_-VU{g>dH(}gu*0LixiVvS#5fKm4W_vzU9YR0H%$?LtEi+E`{5^_(_U1Q z@%+GabB)slYl{f=7X}1Z5qgbpZ|F#;$g`~ZQu}d*>^*&+5#27xXrIu>OAAv^yC>c5 ze|WOJ3+uV=gRcHnQ+v4h21xwzzYTLRe08*iWI0KTxND$vJVX4_=^5|%@#8JMk?WI? zb#---?B3p9UI4)-#B^obz3nByzD}#+g7a~>go6a(4&>SNdl@yYG;VVcyCtDP$-u}u@20@GZg9%d zkYW@-aP&EG;!(uHd4<>h86PH7Rn01M3JRoLcOJ1ML(eO5ZZw2VLh)a%IkmW=j24-6 zG8czixPtJR7NQGTO;J2Q0|SG;UU#Faii)8Dt6M>5bd!*~CL6<|qebHZ1omM;GIA#} z97zxob#!#tLyZyRwpVfLMT!SwRm5+VTVqTq7ZICWj~~~ZKBJVBz6&244OZsV*48%I z+uKj^h3E!qPmip3XlO`nbf9;Mh- zwExEcWHiaqRzPv*l{W+pMh`AlGJjzf`R;`f$jV0W5dp}`DJ--yo-ADI>zN{k<{j(k3nGtDF!WjF;F=H<;luI7pOY?(DyNfTMsM+)$I ztl)koqqCS4J@tBsCuBu4qxw`m&O!Nx2Y2;ElbEgEHUl(v$hC0gbHD6g*G!|)^bTX) zc|pjDYMwg!U9qeWBd&985&WE$UseriK6yCa5j0p&xWH@@j&Uvso=9JHW0Gn+04vJM z$}%tQQDU5El1s4Cwl-6YD`1l2X0IX+S$;Po1tJJprK8hUmi^PE1_n9-o8i<#WQfx+ z5>A&JIrDU8W@e>{^JxXU8z*~G;5}6NlUJRd`7Ml=I_l_KuS}VM^(4KZJ(@|hG*k%P zyHR8wH_JrHj$uE{`?SOz-lgC$037dXaYnccw%46+N;C zK(ty}Sz*}d&PC^vk3J^?V;e-1S(6fpq^Htqi!~|5w;;g_Gk1g_ycFCpFGc9t?oJ;$ zU6O&Kc6T$Q#n%-asQZBh*bL+&a1P@y!uERD0dYP5d<3O2tTR=$PqWeY?Ab9z-p1uc zL02vMre6~S0R~)jdZwqS{U_^k#fSoCt-qrd#a8UY^0sN574xcOB{?}cr61Ac*4rah z40kKqRJ1B>(N?2x{NHPu^!qrd`sbw}oor%9R;4c0*QP!ldKB1KS=Ol*H^~NzHUT4B zAE6sszj4Ad?%j*xhD@i#(}{`F;p_EFjcr#26)Ze+ZK>ftH6}8AJb__~0K2#8g$Z+A zKGL;-4H1H$Kri{b@IA+?Rh6H?duKznCLlL>#!eRjz+%Tw9qWiqNn>?|{U~wJg#n=FFU~>SG>h|H@ z+J4;zmjVfx+>t9pM!BPbUK=v!Cxe)@rY><*AfXV4V-8`&%!GN-!aKVZAlQQzMu3%; zij43#zx?tG#+-Q*3NpIXSf75T{rj}7_puiq5FMT&)uB$qOR=JMWKDB(^Vy3YHjP!S z&9m$7F3QJV-xf1IHkNp6WG!+hX{KXSLq#>kV7?1gK!GKY%AT z0?rF38$D!ZR#u$W>-D~i6oU$Id3hv4hM?G;m**{hl~sWH*yKfuw%3?zL=+Yj7Rn@R z!WXYPmb+Tt&I~mjAEZ6X){311Nx|~;GhtyfrRkDH05`!v%RZLTg%X4y$>}P~h}E8jTMbO@Sv*zM)F%`8I-y_G?O3vkZ?A zvTT(!rp@gfmO9)GG`QRpM~z5tJ_RT&e)8lA2`kyJJ$mDZE0dR%Ec=n+!B~3$csSwB zrTW~fhCMLQ0x)j5bw+a`&|2+-aroUEv)ZVOwcYw@=M`e^00QD>r@e9u3)x2l$tIM- zND{L>IXStoikOH29v?8e^S=t-f74uO{Wdt!NK&xDYH7tfV3oCEkIFvR7TIwg*hA^P zQjYX2Ncq7Py4?nAbsa8!7`^9B!41TK!SK8x#g*!tN}N3ZNhn{}!RW%OSW?9p>=<>} z;^RoLv8`&2i=oaW8K-LkJc7jHPv;A_nc4L3=iHsFoB^7_QnIpwg&fNvO)CQQKr0YWU@V7jkT7n|W0ju<})lxLmp)j9?2yqm|x$OwD z1&rtL__G+SId@NOx2Rzt>x?aCG6&H*B$FNrfBP2xY!52>IW>bg3t(~UQvPLRq824Y zYao~Y_M@O;ecd4&R(If}OoUMHTAK`jLhT(N{Az>OR5Z|qD6lEt=A_bX<*6NNxT|eTl_E=uJgy{S{ zTA!jN?&4QeOKlfMS1z9fH8bx%kd&{=M~84yLSQ^U`H)5HSauYDQqRiDx7kreD7mjN zxawIwc0{WdZOs39cZ7r)V`%q(ptDc*I&(uSjOM1Je%eCY$A*=IG-AE=R%LIpcC1WIpXKG zBO{$|YYcihhZ~oH#umrd!vFp1nwZ`XU46T#Gc{yQia$1fFVZM2iv-N?Ae;$rSbo>| z@@(bZ0$F!@f8zE)A3fBOdYyF1N!c7Ptrj#dZr%hIniQsgW7BPB;RtDAod5K?o}(xe zpy!J!3)OPSxVuzYKEVa_7o+)scnIl+4EBg7YEHK|!?s9h7 z0~PBZDk0{wVkM@JTGJnwOkP3m1U~3vIcB)mPng?SXculwNkB@CDUJna>$evCJaTto zmsRG-cd^aCGV@2Qv|_hWLNo_LFZIdEf-U}k@^YZbmM1Q+;nolR_ync{dB9fZ diff --git a/docs/reference/cmdstan_model-2.png b/docs/reference/cmdstan_model-2.png index 093449f41694c87ed4f6a948f6425a8e02c8a550..4ed9fe5c527f0931750caf802c560ea60a2989a0 100644 GIT binary patch literal 16009 zcmeHu2~<<(y7oqjv<}c#1(B&%Dk=ghqRhc#72`k@1!Sg`83bfbfCQzLDgrVn%3M(p z0z~F91d4!w5CN0O7?dGEAYn=Zge3Rdc>1q@{r8?*?>#;D{%5U|71j>fdw=`e@B2Q_ z`@HYBbLXO|$<|G~HbD@y6?x`+GYFChfgtf8Ki>$x8F={X9N299@tn!`U<+O=Cu?zF z^M&sjD}M-5jDr8f;>PElA?O9V09x@qQKFma%k$;w-$$B!$Vzr?d4%x@R@JMEQQoBW_zaiZc)Cj-A5R>dnhN0X7ndoK>DKSaUqq``PlG4*r zh_xRIX-ulWBO#~PJv}4C!kp6Fe9F$wPTT%7K84L=&X2ZomZo~GHIfS%4fe#UjrsXC z_4P%Ii;FGMjv{-UfTM>%5SUCR^|B*jW8?`#%+mAx{D&RSTA|pU(bhQ6;;V0)o0|*k zMN3Od)SEPuqp$&>pZeSML&V;Zc(TM*>Ogjme9Pz$7jns@`g*0ee!9YUndn1(RNC;> zS+pEDdJwfiv%zrw99=ID$0xZKx`w@4iWC{KEQQ=qPIEPR92h^M7TBJ3r#Z(`H9H@% z_?(lFnoh{C7e(fe=g=^401j@3N9kdkUyxe%;r@&Z{t|p?zhCrf7ubHfZ~ji#{*BY0 zzt`}oqqAwgRlO%08s)ZZR9@M^eMTpvJG~F>eCzwp(@@;LT9((uU{ra26Wb9*2*(gJ z_>f5laQG|W@Ze;be&&!+cfpi>V{s)c`M9*??95C7=d;+?A+`?^&O3z8mSdJ*a5lYD zq!}yJwmF$4wms!+ar^ODD#hy<7(GQU<}~L`vysWP2F(~xnWnZ5xsMR*>gsxR&Ky;c z@-)h%t*oH0A4{eB>wMcWUQnU!Z8++W=FV6V>1=Cqgk=$!eP7>WFuR^;&#e6XaX%|d zOUoG9MxkNITvkz0kw>*{MaG6u&WBykr_oPq>gv2ez_P(4E%4@c4i5Bbes1qnVH~ly zp|;Pg)nN3- zPu9PSQiLJxQC`EjL2Qrj98w;$(a?h4cC7&IkQ?>ceJh0Z@g$ze_}EYxte z8$G2l0U!EreR*khE>5KgFpx$v%wg$iX>a|i3pkh|%%&N05*@eF#p!gp9Nqk6{@Zt? zJ}-;Z&Iu~bIC$mxK>EXoala|k&Thb{n&O1b`doJ+%d%bC#n!=rHa@<)m09$M>xD~t&^uli1pkJ2V0#F?F>_y!)zJZc_Fvgy3nH1 zNQ6T9h@`TGuA;Btyah?c|ACnJJHfTpMjARZr%BQ!(6NjcqV_c*KXSUF&@!xsQ;~0J zj^J@hQ>3JVXKI+`8Ss2EJAR;TiV&YyuGcDLULd$gu2oL-o}xIt!8GUE6yPh&IYeHR zw$4cAZr!DeQs4wl1>bx~=U|+xxr3f3@q-H;Tg$mcW}^@VDGGCY-W;cU5#e!-GR3i? z?eHfk^YlMPcKw}}`p2C9{0CTc7JA{QeFO{Ht4-0Cf)t*o^580dF&MX6Bb;}r3wc)g z1q6+43lWRRrd-Ek*@6>>X%^|R4xB=seV;lp)SOexuJ#fKM(om(fJP?MFXRtDsKaW2*oinfuJvic_3aA|1dU)Z6*CkPKix{5R+v2KHGly=X;1{W-(b4duN*fmD`FsH zB+ZG`YudW_UfatM zf|lC~Yw{6#Y#M^!;?CTE9J3hq1_N|t15tHugK<`K&GZu{jUa3+<0Q$0+k|!dAP&g* zpHjBsv8zbtj^q5MTyvb?Fa}fr4T6bJG~gd8|Nr8x{K<94?^nNFYh+BI4RCgke8a&Q zG7u0Usda}`UtDHpmu2h&hM2;hII+1s>RqDrjl-p-rFO#x3T2YgN@W=hYHt@8r4+tN z?re1iHAbQ;3P?k;vtBwiK<_#bnUBZ}?M$)_Il`r7wAfQ7mg;p3-{%#CXWPNl#xfOW zQz+&@_pMX7}uDVa#jn zqjN4ofn*V^k)NS%Lx5x8)!07; z-@WxsK0eImdFl*K?H(D~vN1Alt>28q)>eq%KcO+Ji#}w9FZZ{uv>%q=vat&n8`}zn zur$=mXw0G2Dt%&WCu4D)+5qyY+dPis)ilwi8*cjZjFeJ~raPm55TyDbybQ-LhfkH~ zlSH!Cgpn`L?@dfOzZaIVrF{i1rB<+vbvv3LB`wWQkq8?9y|(5gY=RnY+IZy)N8Bl- zRSxx>cSWB?VQoHXg#x2a8wgqhAZRV|=7DyV9iSyJ1yr_DoM*KyHK_J_jdQ7-g*iX& zE;Qm@VIHVgwCca^x7uWLlBg-4+5BXqhA7sIL??7lqKD{S$I9&q@__wP&+i4FJWKn^ z)(dg8uemubFc_oy?Ngw=DS<_J-dns1VE-U3t#&Bp3c4mlsR0i3tzfar_q9# ztxY5K9{jJr4l4mopy}IPUDuB8ikWVVlr-n8#V9ct3|KCMN&#*Y{hz*2V@E@+J%QO* z>bl+#z^+vRezvo<9h~N)fSwKa5^dN9$JHly$bRR!e<~RMN!`JJ-TR+bEjLj?N{djQ zsxVJx+B4-c-H*JZz8W|MK`C(78;Mx!C1g@?LPR1}%cxh@v}Tn;S_H>IkO>NuAQ9pV zs>uR9IdkG+A5ci5lFiw+7`_J7u=`8+?1p=i<|tEZLX}lz;li*T6wx%|91KVKANGS` zg=!{P2+rlm`{33z-U13UD6-Z-2lPHsSQwU^F(A+U8cwpO6q-kxQ9Vg(|*TIS?{Uj`amqBDhDSDa7cy2=*j0zs>`d+s;&>-g#Jg6R%vI3=*_E|)KC{0`zys-f~X5~*rN zeL475s!gFKI2u1r42N&*&-07j@bc6Y+>vH2eEH@)iHqEJ<|412i4156ZyQ*6C6TdY0wZh29)N)B9Pi*)DN{ zzBZ2%q^}N2poN1n(KU4jgPw1_wepi1Lec1z1!lL~QhWpdJw5W;4L?6Ka0NiWHmIA# zU7!j4MV*qk>H%B~!S!u>G^l#((I#!@rp2PmAFH>WUBOvOp2xt8w&nswl+sEv6-= zUKFrik)D8aTN35rk%dJ0pG&H)EKJmv`8Jt3`Fz#?8{ z{u66zk9Qa-ETK|kHL!TPJYznOYG+PArvhX7sLGt5@34XLTL%Fz4nW~N(Vij;=nURB z^5}~?g|tQK;cEW4qEJBkmH z0?PLW*2#Mb>HbR2!Cb+VO==846c z<69#@UEbT<3&bn$ZR7l~RyeZ$5Vmlb!-fPH`jibVSYf%I$sB3nnmNenAEUD> z_?Yr%!-5hwddwrUe z)Rop8TeTl;4ip|r&)X4w^vS+6M|RxVu{ZKB^WXk7V)N!pMfLBEH#n8Hu`-7IS<2_Q zPn?doO{9f)C59PkJsz6TW$?mK8U7h`|9UhRUDI=UjGaeK%)?&dKK^wqG>)?`M9gi! z1mHsGa=t}E_JnNukACDWsspzZ_PD4nVxmp8s$)nRdSZ~|w~UXi_I)1@hm4W?;dT%N zojv|yP=4xzb@Fj|IL}0eS8`tDe2bw;}%&@w|{I{efR8>Z{~kxi2U00NOFduzCa+D zwf#&o3w`aulRZ$~Uq4VG5{XtXvT`T)y*XxRXlV7q@kIAw#};!uWr@Ri zW`V!JoR2wut-r(Dvr+%pv4_RQ#X%ak#4i%pyWmw6E*` z4VC{D>;8uUtL53WZD?iYw(w{Q9CeOFvjC7)A!wh31pCAU%fw@)Yft&>!X4>Oew@?Y zV3h@evJ|%0N&d?6@EiQ$RPLj}-ji)jHgJH0S2|huZ=vJg-EvJdy!Zu4M!IOByA%COyqmR+ZEAxDSb!SlCa{O>0E z{|9*fYOa9Wzjlv3e(xdV%SyxMa0xzMI=z&;_*-AhlWb_a>AFQ!kki+Y-?h~ZpY6W1 z*IfCW68%yeA?V7EI_T0mGmyK@k+8j*$Aa#ukcKY>Y>Mc5ukgtnF?322_S`$0PFK?K zLT~-xqFeTxgO~HJB#6+jeeDfg3nuq9h<%(&Qp?=rGR@`~=NaD2>tWu#d*A6#KUia1 zO3Nrp=rpV|SVWI5>xGjks(b%lUXyBqi$9qs{?6gkF+;HnfTt57-tD%hPuF_CUQ3dP zU1vaG(D#ngAAP5O?8rk~QMOgJ(4_Qk?2%`*wMvW;Z^HF%tk{g?dKxf})=mZX8hazf z3eFsiHz*EA?QD^-6o8y?6|7cXN4N}o4FY4aEMxR!)8{Q zm70#pPO!G@M(#a+G{s=~`?y;?XdP0J&KXX?cbfWjliL&JcOtP1%qN#;thCzR3)-S5 zWB}3;IE;a+WQLScl_Gr6h>vLck8`>PiNASvsQSyQj)U8zZnk~ShQa7U{1fz*tHdSa4V>n?#5uLr3= ze%5?zXnLeH8J-7pXvfDr|49<|6EwH!mBZ2VKJfY0HJy1adL^@yHD z+iizn?;F~B4`-gMvT@_F2h+4H0@8Xe&Q-kI9o}|nG1~NSb*6!?@HA|Pkr6YdX)K}) z7&z2rUtd}I^u;4;6qPl28j?AOxmO5-*7_E3I} z`f>(yfmNU?uDF_VzJ*TDR;;gQ)>4BdT@|oFnSLfje}}}jWq#<>CK}Iek9*LvQA-i| zOMNEt8I1Fzy(^i!cXd9E@9L@=?E~LOwbDAoWi!kp2NOG;M`Jm}KQa#v*52IfB)Yvid$nG#gh3&8_} z;N@;uQ7GY)-F-&>Hv8&;NrJjKNMhF6y4Wk$q9uDxEf@ZoBe@mu>xG3Uzt`3^NV$SX0=Q^)zg~_nF3nu@ zSsB#L>hsF!-X}bSKT^O>&qkswdxjeIjGQICB5P`Da?->%GDrP`Eu#qR z_r*>RIYjzi{_A(hmR+NyQ*Z8ahA7(g**&qkS=*VRDpQ(`t}U2oYOsDM^PIg-?c{cT z5gzYJvKPgi&eqyIeDmhbEz;7Jn2<>~Aggr8?w?#(>P5K*1e8tc9Lio@T|MI3m8Q}* zGU6D-8np8L_K_8kOHoP~w~Ffn598ydghR*OBJKFb$RkIOCe031%BFqpGWH=+&-^~TbngreQ1uh(KSf=%QLh6YfD=${#47VQNvplo=r$YPdv7& z@Um*W<2dnHT<5A_xfRjnYP9%RrA>Z#?smb|;wK~DA6h&>wtwGg(9k$h+^W>!oNV_{ z1>LM>AJ41j5hYY_J+37A8(Uh^$_Vb%g;H;?zLM{FbQ~_h$4DheJ$qC=+c&K*y4W|e$xBXP;Vg}(t}hT zuUqUom!J^%_EMpeQiN2{OmQn1PBkOs$3}PP5mFP=9lduV!m@thfkdvD2Tj&+iWT>nR&Lf&r3!TUc3a|G{*Li&4EwQ{X?Ptxw#pb1ILE~F0<QGV1LM;~oBE z@g@~ha?l~W*}-5@qcrw@T=>j2tKuI|bs6#oWxm^(Wo~UK3D(RCY$`P>hk|Ud+|tZv z1cD}fi^Hqb$_XA$#2c8~GBR?>{a9*e9_s~qaUyK;q;flp74+K|s$&Zsny6r&J?Evm z;^qjh=Xel57a?%hs9*jt6#GC$SvVI}>eQAXXy0tiEc2&_J34+NYwLRg#F;<0&y}&Z zNT$=Ppt6g2Mg*XQa~jnusmHGO&(6x)hy2i`GkVXR@pI%*l7tUX`1@pnFPcjH#@hichKY~w@JMwa~8e3vmq1tv{I8R;SyvXupHahpBHuR>)GYLx@>$GpxYqy_c6*Al~N zQESR;Ba*%Cn~XinJ?R2IU4-?nnaZ&ZdH=I6=j~-4VjrYV1Jfv~Wt}}Dl$QquB)L{F z=uhrU(e~WiBEcPcD&bSw_Zo#=NK|QKuXV9!xJGciC5-(VrIg~g`=j?bL3}`3rU?o+ zUVe3Xm|%gQJ$D}J&$8pQ$Oen&HEL%{?nHZTc!TagEV)hq+pFH3PR#f!l+T?-CV<8r zC_-nsMzzB;afaM?Nt@l*?*(5u0OA>k7@w^i^gEX@R!#_?t$+Dh*4F-F_ORPg*fUF9 z?nKWy%?r=W+%H^(tn(H98;ZVNme3420Rl5A)w^atcoYleR9Vj{%(KX?W2r(~PapU)f2D~H#HkMcGzdDtRS|eKEou_)9p_jT%{F}BOhtavo{!tEc^RQji zptQ8KtCN#crEEQVZN4>81x7%Q?HRS!4OxQz;vw&2sa8ZxU=fkquhRC|HRJIPHH#VU z*&1kjF1S7!CtSRD@1B!jm7}}(S_?N?aXLF*Wo03mlt#dE@W_NzpZdNPwAiM%(_}HS zU@Nh$sPabp{@k(K^4_mb)7<5~=e2A)yVG7A_L-RFS4S?|=F|AS`?; zQ#)NiHYAVrdlEXoP))nfEUMLDhcWaY)Iy5S75OaSP|yP*&AZYFVycgq<)CH{j>5&f zD}JvJx2mj&ay^%?*G@!&hZxxY_Kt%|BTs9hSE75}Rep3^bv_-hADWr-`QR0HZ;C1_+t+33a22>_^LwvxCL% z(x!zjeI?Gw1Rv~TDjyb72P^lade z3==RN$e{6)Fuw;wE3`?L_o+J!T29+e9eTk&(8&;bzx6g)dUJARE(%dMF!1o8RYJ68 z{qk{PF0s$cG|hoIjVE-)n<(kd7DM|i)Fxm=)d^;OP$dlfyoObY3Qowt4L@0eX5uWc zQt#5El(9O>t8b$G{+gG3?0g}v4L~9cO>h@&b8~ZxEqDw3N-Ak3b&bUqdT-7YQ?T7j zeATa^8QRyeutLTe9Ul(4#;Rsyz=>*H6NffM2$l}o*M(I6JX5sjDhLWTNC(p~1=C^J z-wZ}EECYn6BD%WU$zo8_IdHLSvATO{cENTQ1pD+rCUxABv}!J+VfzTC<&I=c^6n^w z9xmfE$r3}Qk9{m-SHIuXK2g>DaohAiEr+tQX9laa%B|JG8#r1`5A=)z%2=Qe2Q~QM zK_kxu=o~ZO&p<0zi*-43^$KvLJzzH((0O%!1?o4r%yh8@a-V3=p4=1IaUVQ1{vS4U zv?kyL1Wq?vh>uf2W?-iC-g=aS_h2ZSSQ~ymyZaT0({)e1IVs(lrt1$A>e?bg~o^y*Mk$!TZv)OS*DZQrp>h!gVj)&@Kkl!E0$R8KBkbQ_8>L7+$xY8hLbRW3Y8IW6V z*K#s$C3ywtVqsk-J(wO|&93J!0p0|3r)Ng62my3K*K8#{!?UP*FUpW}Wmn8;(Dq#HBV&&tOKyy|ge_$& zhYyctge|nC`Tq7rU)2o|YZH%Pi-IPsoVXBQUfpl8!N%rf#aJ+oP0az;o1wyjn-az0{&u3zD)^rVnkn%z%1|5W*COZowA*dQzXete$d+@H57aDO8gEI!gR)p%=fF(XI!J5h zO!cOos3t8VnBB1fWBqcwR8|cR&^os-aes=sq{1zIR%B+pzhg=tp}DVu&X_W z^cxDKIdfuV3=6F>E}qFP}ZrAl+KBlyx_qbCJF0QQd6r+ zA1b_Zz^W#|9hP^JN+FIUKOoAz1}fC!5W)~8Fk!FZI$gXaAwwr&Cj|AM0uC{5ot~W? zVzkcad2kygmCqMiq%PEh2dR-jGCmBKP9D1{sq{(}vC*{N=i;9&N%C`szw)Faa`AG zZoelF?FF|bVa!YIGf@r?IOaY0XuKoY6$tO^RxLxpbi+$5;Z&y39Tqo3O4rv0F9QNG z)?e=mOD4hD1(^V%v+{b!G7Tr7sUUGUTv7e%G!pgQJOFd0Iv^euc>GOoWB1sSjrcPv zqYO$okOZnk`sAdW6k=@-aK_pINibmK95(rDRDud2q*<+pLMb0;jfsh=h8J(H=Yj}v zY*EB6zel#j`KGfZy7n&S94GEKQMmDa!WfW(=}XvT2P_&!4ycsUZnVhpljf^NW6veP z!V}2#T1DJ=`-48<^Rf&urx|s)Gp5Cg@d+D!$=fWeaUT5ku){R(4N%}fZ+apVVAak> zlS3Eq4uS(8VQf?3)h3BpZY;2^K7w2T>~t6utwXu@;A$yhesc|xX~%8{As4SACnwh? zE^Tlh*pjg=bNTxTa%QH=3PH50$ruH(t`0Nm>#A#UhR5Kd1Qf!VD(**jsVj=c5>Z`> z!GrODp~8SJ?2NWwdU6F%0q=vhRfjAc!UOXu9!9)svNNDfe%d^TfD4BUs_*19e~Hui z@C&F}U~z0tz?gg2&Q=VDjK*w}Mf0anxD=1_6j*wwhi8CQHFv!vkO;)z!b(oGLS_JP znt^(r4cC?MQT|jb*Ks>G#&V)FwbEgfexp@!1@_S(nvo9hY*pff90uhR11=drP1Os) zR~#39FjKgcdl^qD!EmP2{Mz@cV}ii)2fQi)7D&`{-QX?s2`hvDX^Q)^hY%okc!*B! TDKcAs_5^v#^!uW}INtg<4VB|G literal 12652 zcmeHt2Ut^CyY5z0MnM_J89_mr5=2y_NRh5%!GM6m4818$X+c`(MEw;V1ayZYCCM-t zG7cgN0!9eIf=sAlfPfG{2~8w{P(lbKXYF{}bDs16^WSsl{`a1H+|Q#cthT@PeeYM- zzJAKqa^r^W8z2bUh(2-D9)i}!K#XLNrc`+1u1C+!>b$$x2{|T$ zf!W8mad$ZVxwc$Ra*`m)Jj^H{iig3r35jE3(i$mQf8u|*)yyNy>=IN)t!SBmnmw}Y zCf!`6ki=mBYNBpr5hI6m)IY3HGLw$rWGl9tLAdYq+8xX#r!oeOjFWF0F*%Dn!RR8i z#uMkmc`p~6l!Wq9q75hQD;wEThSJIySOtb#kx~CVn4*>L`LLIp#2-!T*bDlHVm^;q zm0qZ>x{c;tff?K!111ms`k1jDdEp_Ai;775b7sAil2cYgYmvWqt4*&)+RQJI=@4YK zz7}a9@)}R1!P&3YLfTKuKd}zM!6(@K!&Rm;a1MfcjFa`sh{ieX8N>@-TZK6+V!k+f zP98|)B~bby6Vut4Y{Tf;_+|~4lSQoHD!Oeq!%JOi0&I5!%{Pcy>fl_DC~^v3$inP0 z6j+vlrQWE3f2i^!C&6uX^1LmXTfV#8HohaJ*W|=Pcaf7yDAppD^8?HoBnnewr}H*e zMQ#Q3C8%Fi95T0IRA3kIaskJ+dV?{^bC=>xZsV1)uw^@{Ib_8GHlpgp$$-iIvS4ye zn19ytkmd2Ro)1A7wRr6Dvd0#xw3`L&ZM1hQr7Xt0Tago(TTcs_taJcBz_QE#6YQa5 zkt_t=nCq+*itx)X#&-yO%*Ip)W0K+`<~9Kw#D7)`)n7L#n~h**(zgMBb8o9qh)TAI z1FIgG0Z)KKsV3_JWI5H+`$P<g?^c9W@h9SA=RN~2M-~N8snDu7)){`w-P|IM;m^XT848Xgkd4s> z@0>nWn%%b}1~VlGeAOYXe+-E4c8#xs>M!Sx`Q9=Wb|{XLcGCRt^O?;>hJY-U06E6_O6kzb6iuo72!Cff&u+s#?M8D&%Yc0DnD^$X&mvojrjm{4UWy6{x;C7S+Jl0 zfkO#m0_F|+!ho(JOYdz3!elxb%xoCdei5R!&y1?uaOAD_hd ze_o9Ln<)6_<%nk2Kc8Q;o^3c#J)2*IMZ_02x5o3}Xxl(j!~D0&x;6~GIKo4GNXWb} z#R*~|A@kQ4)_rt5ntzZ@yQb}gTJP5#OvZEzdpk@v9k7cvogM!QDp)gHaFUVXk$6U& z)!~M8&sUBWCo2J}p!7@ia;o(SH%Tp-T4@6pc3)7$DKPZZCXGn8{0U@oyRZg@prR2> zv>dV|o{X-VtH--x;WBWn)j0gSCLxC}?AU7{TB?Sz(Fnn~qVJ}kZFbh=ii3+rvbn2= zVq8!5iSllvm>`;g!d_G@P9gCN1f@uGg-`iE!i!I}kH22`3Dl1R97=E=HA0uiy_1{0 z^3=h=Rh^PuWC=o^BcP@<^x}50ua!Bq;&+pxYLIH32Kgy;@H7aNK=KOWr%6fvhjt_j zd-Xjs`>0Xn$^C>KAcCq6-6ax92cFqS(U1GWQ?b^c}Kz~5H} z`{+S`ZWbp9F6fLXlAajQjiOuM7HfDV%~w+}H$b6o`T^o5>#=@V30plv95{YcG@Fa^ z+HSTQ8ajh$8%~!0d=OtS@cPK$!Wd_-TVjX~RNsNH3nyD%aUwb!Nw?0MfdHdE#mAbC zK@tjR$n$?fVXhL?Mv>y>j(d1!Gm4>@Jk&m>wTT$i0;t%7R`S->;N)pfi=yCc3%iqa#j3n z%ZAHWxOtYHD3lt*chC0Mhf=0*YoGrOemn7K{h7Rz1lz(#YWRak8n}T!RB!wK{KaB} z+x!h0J%?`GTK??Wv#Rdb+?DwZTKQjkN<7Z<*b!y(!j^uc0D&nNt12)IpvQBKKyh&Q zT4`ikf6e}5P58#IGsQ?zq`rv#y+=WI9>HzUsVvx_H8(5IZBt$8@{dd-$a$M|k$lfX zVZ(V1VW~fUUs-0Y&h<-*-bZ&Mj?_EN{h=B)%!8>BCt7}5yafuY`VxWsgX1ro-pN2q z4gc)vk7r(zGRFI;#z+7A|87TT3=_%tao^a7k1Pz{)jDtdGI!sj@NuqTmX-lQ##>2x zaGYgl9!otok!jR+r}Ab7N=Pk^faGPQ?3%t-stju?>JfgaHmS>3Az+Nx)=!_3ay{d= zlJ%~X$C^reZPX#9{h{T^eJZ5=isELMcGv5Tga_&*VY>8JksF2*3K3%m9dGj41viES z8Xv=82T1)EEM(~Ic0^`>^Y#BiU|{WpbZW+~#;x!b+wzsGOLfH+`vSfZ=3wb$)1Jsp zL;W8`yfohPefcT;%)?sp#6DW&`?KrK9Qk#q4W3G8HXtg7Gj}u0Hl5@bJIqBdo(pe$ zTkHU>4E?WDaP(E@j!R+(sVE8p(s%OOG84jk1_#V^a-!kh;)pq29b%aZPmgdmR+*)aRVPoV%#ozX z9WtH&E+7B!n~<0>McCx(QvW-=f!9s<@WTN4H9_avlZ#b9vS{q?#z}bQ8_gJ1BF6Y> zQz=GGnWYbqFazJHN53p6`X+HbqIJ*%mH()Ze@#K5`iW5!#@a1tICmr6N!ga!H*=+U zOIY2&K4Alwa^D551^S%^1B>qBso^oY=Ba2DZ# z8j}oX0`Y+-Y#=`7%tMYQgJ(b>=j2)bWVM#(Vn1ED>Vi51ZI*ruu<7?`;&1;==%9b*MiV+j2B~E{3Twbmw6>Up4i4AajZH`*Rx# z4&R-juF!5sSbeUGYl#zXSAt~-V&TuWu1)^&t2bgKLWJtHOOK#XkRXny7FgsJsibV=qi z8}9cE;)WXAX({Yz&md>Xy6JmijIFiwg}2HVGX?zM!*yb zCt||){=`9R4Ye)$v7v!tFIbe<#%z8lbSJ>i)r7~HE^&K$Z{;yiUgwOhw8*}0Hbgkj z4k5v{qVN&_gX?SqoIsy2HOq2$gE5_Xm%+{>Plcv6k91jyi3;R53=xJ><~Uk4JBzwu3&t;M}L03;|j1Vfj8-Xvz zRsI;X5~eYtm9fKQq1q;3mG@Mpz4({zrBoju9};rRKYGS~DQIn)iWXT~YhX$%ZPkoi zt%P=^j*}jug(=ZMulWTF!PHHMB7`~HPPVWk&`3%-2RtroFp^^i? z`z%D)th_*Zl~kHXMMp2(@1m^z*=rrU_d`E5qH%F?(T?R>wuhW>VY0iFS3C1kIYZzx z7-k+=8rg97?p?1Fe6lTNHaDcN%EQ3Oh+5SB<;RVS%k{TtXvdSb) zhL&0{>}y>*UnwUi=QUUv#7dRDrHYF{2vDSr9~(w*Lb+5wco62OPJpzXFRMr>#l`mv z^75WU1qTNwQ<&6PfB#WRF@+-7gFO@F7^&SATAi%N$btJ09yl=j#Gx8f-RFf4@2Cpt zBR+x^eLtQ)b~)J!w(D~Bq9uwJ+!R#J`;Fq}#`VdW7PkCOR0O)JrKQDG?^@YfT$b{j zPJ9hJ$qx29Q680)BwjDONo8^2y@SjHS3phQ)IuLztIiOEqtfhoGAk9o9Eb@jHQa@! zFk^WNTEAiJ%BNlx9+8|Ym9jl{S<-6Xl+oVSmNfcO8DpBnv@z6dX=@Xc0KW=Qa_C*w z8d@KCPEb8niK8Q$XL(6#?VDD}nTaK>>klPw3|vNdMF0IPRTMR{e{QgREia1YfjzG= z@&?toz#EjLslGW}GC$!vqadj4SA;Hiwd=&U6V0}eGy5nM3eP;TSG*!pP#2znBaI2_ zWyS}g5?#P+b`2IY9%;;by!5`rPHPi%uIJ~M5`X|aIg`c?!S2BpO0qR$8XFr)IjG8R zH+&~EM-^x}rl!I!c4V&?)vt8nBdP}K5=GAdvqKL)-{UoI%l4K1`OIZge^P$FNI=eH zaoIK16-#%ZIp)_dC1&vpCs9T->1hgWfFGr*E*DiF1wVBQqU}OkX<+wNE^i?%zDU?w zmVbM=F@xt^lDJCSc}1l{z-9LtF_Yh0Yve>xNJeAz&|JgM>AgS{?b-IBuJ73@kRImq zVae0WvdMcGFS#ka$s;Gd*AF+Bp0bJ!hdW*A5m&poC&Q0v_=#MnL^<6{h&LrWq=wn_9NQB#@1H+ zhRT(}i6;b$^mqaHq3dLdwFbSPat)F{L~oSP;ssT3?PJl~DRO3f8Fe(zi&#*QkU(Jy zWzk~K()sM@nkgRSO!jLZvf-b?8b@D^>Q+N*j9k=5^;{+}zxhY&S*R@%gA_2|tHkD}GTISW4G3F%cnSTT(4Vpm$ja z%*FzWF8R%g_P*VGsUu^jsbD-T$&H~OgW=g4lFn47!gsm?HDd|chd&^EOx1}G4dv3v znTC&hSJzLEk1bm0>(399Gp~bP-IkO#<5rCeW+XBSrqjP~)CXvyM=kMJM1sN7Rzeaw ze0gyuhZN;Kw0W=5Wx)$VE9s3-EV^7D0mJ>bb&%4%i|M^J(Nl3DrSlG|qMkOFUc}^)31(*tj>e+Jv0h3nb%2@bQNn%gP!hKyzarvqOR8p4m6S z0c%;Vk|0>I)3xkiO6?(*)%~aO@6tz|v?>z<0|Sp>7u=*4Q;Lx7oZDvLB<1!^$0{l+ zB!2qaR*lnF-H?=`wpg3mDp~8E4)S0y)FJub0Y)tHP_whM)>|KwGeu9B@$KL43Cr`7 z8vvx#;1hHien4N@Wb1q3e%K|hR7drH)u{nw88RXQ{Qb=zbn-oFqRqJXF0Fp$lxQxN zWC7-tcYepRo==&rhQ9Px#HHTn=_3R3)yl*`A2%2lI z&*-(*AfR$dB;LO2{ZLSG*|218ZmtyA#gfW8Az3SN;tFMdA3&%xOtRwxx^nv*bCuhI z{K$K`Zz?A%$(fRyoJu?%uZo7LR4TXyDzv3U%mb4{k6bdS!Sa$?`uqF6P`&O}v4;ip z!()%MpuCQ|52sA~x|9QmxfM%O_rfQ{Zj9LHhk4#EU8nuXGn>=1EnDK4XL1=!4F(q+ zyOECqUfoew&kZp?s$kbP?GHYA0vRp@<;~tfk*q=9c~eb3!T@~M^cN81;C8uZae~l0 z`&MGUP0jiK;LPK&+$paD&L~O&C8VR1cBLBnn&2twLPK<2F6SvqvX{|5l49Lo(OPAe z%9^0*Gj4H*+iP-FFwzV%!&Rcw?Byls44STj6CR6ZCNjKu4-?V+HdJzE!|N*K;E~uG zx>fj$k$?XC`xROYGCR?3m#~XFW|2LT8h{)fcPsLjdo-T>WWnwjo*h30outmF+P9VB zR^j}LxI5Nx-@_FXe1fTk`y*SFQ9qm8y*%Ap)$G?7(==NbetkrM&GG1)-3yNWBR^XQ zg03F+cxs#vFIS^5w&4$IS_@HSCVqqx+dh9wxLRG9rbZl&*=pU7qV8f0S~LeMs0n?s zORZ|zz*8h1&X@uc=t6Qj`O9q^Sbfzm#_GbFmeZNpv>FpyX`|RYwrP3Zta~CW%h7xq;<2q>G8@|m{0W5ha)R(=gLwIZh9BspE=vcE3^UPi-ZyJB6o%8mHWd;y@P zed<1_=hSOEw{EMYwGedaQD2FS9I99-sv08z|0tpy;2fuq6lT1YQSe31z|YPnapN ziuojf{3}^00O!21PxLPv8LW&^K1i??F$W%z-AO=G1;k@jW04d^6m|NMUn$Bda*Qol z4dE`HmO{CVs4G#ZVV0tkXm%pWFbW$vCOHqwprbjQ3)5tmeHRC_@qW`*6@(|t8F27; zo5g;HEUdyyO0d~#JQ6T zcksyOZWE45>_CbL6!Z^_yT(bmJm57;6Yp6 jhb<9ELGSrG diff --git a/docs/reference/cmdstan_model.html b/docs/reference/cmdstan_model.html index 61303d512..e466aec34 100644 --- a/docs/reference/cmdstan_model.html +++ b/docs/reference/cmdstan_model.html @@ -1,81 +1,18 @@ - - - - - - - -Create a new CmdStanModel object — cmdstan_model • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Create a new CmdStanModel object — cmdstan_model • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Create a new CmdStanModel object

- Source: R/model.R + Source: R/model.R
-

-Create a new CmdStanModel object from a file containing a Stan program -or from an existing Stan executable. The CmdStanModel object stores the +

+Create a new CmdStanModel object from a file containing a Stan program +or from an existing Stan executable. The CmdStanModel object stores the path to a Stan program and compiled executable (once created), and provides methods for fitting the model using Stan's algorithms.

See the compile and ... arguments for control over whether and how compilation happens.

- 
cmdstan_model(stan_file = NULL, exe_file = NULL, compile = TRUE, ...)
+
+ 
cmdstan_model(stan_file = NULL, exe_file = NULL, compile = TRUE, ...)
+
-

Arguments

- - - - - - - - - - - - - - - - - - -
stan_file

(string) The path to a .stan file containing a Stan -program. The helper function write_stan_file() is provided for cases when +

+

Arguments

+
stan_file
+

(string) The path to a .stan file containing a Stan +program. The helper function write_stan_file() is provided for cases when it is more convenient to specify the Stan program as a string. If -stan_file is not specified then exe_file must be specified.

exe_file

(string) The path to an existing Stan model executable. Can -be provided instead of or in addition to stan_file (if stan_file is -omitted some CmdStanModel methods like $code() and $print() will not -work). This argument can only be used with CmdStan 2.27+.

compile

(logical) Do compilation? The default is TRUE. If FALSE -compilation can be done later via the $compile() -method.

...

Optionally, additional arguments to pass to the -$compile() method if compile=TRUE. These -options include specifying the directory for saving the executable, turning -on pedantic mode, specifying include paths, configuring C++ options, and -more. See $compile() for details.

- -

Value

- -

A CmdStanModel object.

-

See also

- -

install_cmdstan(), $compile(), -$check_syntax()

-

The CmdStanR website -(mc-stan.org/cmdstanr) for online -documentation and tutorials.

-

The Stan and CmdStan documentation:

-
+stan_file is not specified then exe_file must be specified.

-

Examples

- 
# \dontrun{
-library(cmdstanr)
-library(posterior)
-library(bayesplot)
-color_scheme_set("brightblue")
 
-# Set path to CmdStan
-# (Note: if you installed CmdStan via install_cmdstan() with default settings
-# then setting the path is unnecessary but the default below should still work.
-# Otherwise use the `path` argument to specify the location of your
-# CmdStan installation.)
-set_cmdstan_path(path = NULL)
-
#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.29.1

-# Create a CmdStanModel object from a Stan program,
-# here using the example model that comes with CmdStan
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-mod <- cmdstan_model(file)
-mod$print()
-
#> data {
-#>   int<lower=0> N;
-#>   array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
-#>   y ~ bernoulli(theta);
-#> }

-# Data as a named list (like RStan)
-stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
exe_file

+
(string) The path to an existing Stan model executable. Can
+be provided instead of or in addition to stan_file (if stan_file is
+omitted some CmdStanModel methods like $code() and $print() will not
+work). This argument can only be used with CmdStan 2.27+.

 
-# Run MCMC using the 'sample' method
-fit_mcmc <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  parallel_chains = 2
-)
-
#> Running MCMC with 2 parallel chains...
-#> 
-#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.2 seconds.
-#> 

-# Use 'posterior' package for summaries
-fit_mcmc$summary()
-
#> # A tibble: 2 × 10
-#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
-#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.

-# Get posterior draws
-draws <- fit_mcmc$draws()
-print(draws)
-
#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 -6.8 -6.8
-#>         2 -6.9 -6.8
-#>         3 -7.0 -7.0
-#>         4 -6.9 -7.1
-#>         5 -6.7 -7.0
-#> 
-#> , , variable = theta
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 0.28 0.21
-#>         2 0.19 0.20
-#>         3 0.16 0.17
-#>         4 0.20 0.36
-#>         5 0.25 0.34
-#> 
-#> # ... with 995 more iterations

-# Convert to data frame using posterior::as_draws_df
-as_draws_df(draws)
-
#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
-#>    lp__ theta
-#> 1  -6.8  0.28
-#> 2  -6.9  0.19
-#> 3  -7.0  0.16
-#> 4  -6.9  0.20
-#> 5  -6.7  0.25
-#> 6  -7.1  0.36
-#> 7  -9.0  0.55
-#> 8  -7.2  0.15
-#> 9  -6.8  0.23
-#> 10 -7.5  0.42
-#> # ... with 1990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# Plot posterior using bayesplot (ggplot2)
-mcmc_hist(fit_mcmc$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-# Call CmdStan's diagnose and stansummary utilities
-fit_mcmc$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181224-1-1bc5e7.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181224-2-1bc5e7.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit_mcmc$cmdstan_summary()
-
#> Inference for Stan model: bernoulli_model
-#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
-#> 
-#> Warmup took (0.0050, 0.0050) seconds, 0.010 seconds total
-#> Sampling took (0.018, 0.016) seconds, 0.034 seconds total
-#> 
-#>                 Mean     MCSE  StdDev     5%   50%   95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8      781    22972      1.0
-#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0  2.3e+02  6.9e+03  1.0e+00
-#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0  1.0e+00  3.0e+01  2.0e+13
-#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0  1.9e+03  5.5e+04  1.0e+00
-#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0  8.9e+01  2.6e+03  1.0e+00
-#> divergent__     0.00      nan    0.00   0.00  0.00  0.00      nan      nan      nan
-#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6  7.7e+02  2.3e+04  1.0e+00
-#> 
-#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47      796    23422      1.0
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).

-# For models fit using MCMC, if you like working with RStan's stanfit objects
-# then you can create one with rstan::read_stan_csv()
 
-# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
compile

+
(logical) Do compilation? The default is TRUE. If FALSE
+compilation can be done later via the $compile()
+method.

 
 
-# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
-# and also demonstrate specifying data as a path to a file instead of a list
-my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
-fit_optim <- mod$optimize(data = my_data_file, seed = 123)
-
#> Initial log joint probability = -9.51104 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.

-fit_optim$summary()
-
#> # A tibble: 2 × 2
-#>   variable estimate
-#>   <chr>       <dbl>
-#> 1 lp__        -5.00
-#> 2 theta        0.2 

+
...

+
Optionally, additional arguments to pass to the
+$compile() method if compile=TRUE. These
+options include specifying the directory for saving the executable, turning
+on pedantic mode, specifying include paths, configuring C++ options, and
+more. See $compile() for details.

 
-# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
-fit_vb <- mod$variational(data = stan_data, seed = 123)
-
#> ------------------------------------------------------------ 
-#> EXPERIMENTAL ALGORITHM: 
-#>   This procedure has not been thoroughly tested and may be unstable 
-#>   or buggy. The interface is subject to change. 
-#> ------------------------------------------------------------ 
-#> Gradient evaluation took 6e-06 seconds 
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.06 seconds. 
-#> Adjust your expectations accordingly! 
-#> Begin eta adaptation. 
-#> Iteration:   1 / 250 [  0%]  (Adaptation) 
-#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
-#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
-#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
-#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
-#> Success! Found best value [eta = 1] earlier than expected. 
-#> Begin stochastic gradient ascent. 
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
-#>    100           -6.262             1.000            1.000 
-#>    200           -6.263             0.500            1.000 
-#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
-#> Drawing a sample of size 1000 from the approximate posterior...  
-#> COMPLETED. 
-#> Finished in  0.1 seconds.

-fit_vb$summary()
-
#> # A tibble: 3 × 7
-#>   variable      mean median    sd   mad     q5      q95
-#>   <chr>        <dbl>  <dbl> <dbl> <dbl>  <dbl>    <dbl>
-#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
-#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
-#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  

-# Plot approximate posterior using bayesplot
-mcmc_hist(fit_vb$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

+

+    

+    
Value

+    
 
-# Specifying initial values as a function
-fit_mcmc_w_init_fun <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function() list(theta = runif(1))
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2 <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function(chain_id) {
-    # silly but demonstrates optional use of chain_id
-    list(theta = 1 / (chain_id + 1))
-  }
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.5
-#> 
-#> 
-#> [[2]]
-#> [[2]]$theta
-#> [1] 0.3333333
-#> 
-#> 

-# Specifying initial values as a list of lists
-fit_mcmc_w_init_list <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = list(
-    list(theta = 0.75), # chain 1
-    list(theta = 0.25)  # chain 2
-  )
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_optim_w_init_list <- mod$optimize(
-  data = stan_data,
-  seed = 123,
-  init = list(
-    list(theta = 0.75)
-  )
-)
-
#> Initial log joint probability = -11.6657 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.
fit_optim_w_init_list$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.75
-#> 
-#> 
# }
+
A CmdStanModel object.

+    

+    

+    
See also

+    
install_cmdstan(), $compile(),
+$check_syntax()

+
The CmdStanR website
+(mc-stan.org/cmdstanr) for online
+documentation and tutorials.

+
The Stan and CmdStan documentation:

+    

 
-
+
+

Examples

+ 
# \dontrun{
+library(cmdstanr)
+library(posterior)
+library(bayesplot)
+color_scheme_set("brightblue")
+
+# Set path to CmdStan
+# (Note: if you installed CmdStan via install_cmdstan() with default settings
+# then setting the path is unnecessary but the default below should still work.
+# Otherwise use the `path` argument to specify the location of your
+# CmdStan installation.)
+set_cmdstan_path(path = NULL)
+#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.32.2
+
+# Create a CmdStanModel object from a Stan program,
+# here using the example model that comes with CmdStan
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+mod <- cmdstan_model(file)
+mod$print()
+#> data {
+#>   int<lower=0> N;
+#>   array[N] int<lower=0,upper=1> y;
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
+#>   y ~ bernoulli(theta);
+#> }
+
+# Data as a named list (like RStan)
+stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
+# Run MCMC using the 'sample' method
+fit_mcmc <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  parallel_chains = 2
+)
+#> Running MCMC with 2 parallel chains...
+#> 
+#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.2 seconds.
+#> 
+
+# Use 'posterior' package for summaries
+fit_mcmc$summary()
+#> # A tibble: 2 × 10
+#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
+#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.
+
+# Get posterior draws
+draws <- fit_mcmc$draws()
+print(draws)
+#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 -6.8 -6.8
+#>         2 -6.9 -6.8
+#>         3 -7.0 -7.0
+#>         4 -6.9 -7.1
+#>         5 -6.7 -7.0
+#> 
+#> , , variable = theta
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 0.28 0.21
+#>         2 0.19 0.20
+#>         3 0.16 0.17
+#>         4 0.20 0.36
+#>         5 0.25 0.34
+#> 
+#> # ... with 995 more iterations
+
+# Convert to data frame using posterior::as_draws_df
+as_draws_df(draws)
+#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
+#>    lp__ theta
+#> 1  -6.8  0.28
+#> 2  -6.9  0.19
+#> 3  -7.0  0.16
+#> 4  -6.9  0.20
+#> 5  -6.7  0.25
+#> 6  -7.1  0.36
+#> 7  -9.0  0.55
+#> 8  -7.2  0.15
+#> 9  -6.8  0.23
+#> 10 -7.5  0.42
+#> # ... with 1990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# Plot posterior using bayesplot (ggplot2)
+mcmc_hist(fit_mcmc$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+# Call CmdStan's diagnose and stansummary utilities
+fit_mcmc$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251434-1-79f79a.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251434-2-79f79a.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit_mcmc$cmdstan_summary()
+#> Inference for Stan model: bernoulli_model
+#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
+#> 
+#> Warmup took (0.0040, 0.0030) seconds, 0.0070 seconds total
+#> Sampling took (0.011, 0.011) seconds, 0.022 seconds total
+#> 
+#>                 Mean     MCSE  StdDev     5%   50%   95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8    781    35502      1.0
+#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0    235    10662  1.0e+00
+#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0    1.0       46  2.0e+13
+#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0   1874    85179  1.0e+00
+#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0     89     4050  1.0e+00
+#> divergent__     0.00      nan    0.00   0.00  0.00  0.00    nan      nan      nan
+#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6    775    35215  1.0e+00
+#> 
+#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47    796    36197      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+
+# For models fit using MCMC, if you like working with RStan's stanfit objects
+# then you can create one with rstan::read_stan_csv()
+
+# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
+
+# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
+# and also demonstrate specifying data as a path to a file instead of a list
+my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
+fit_optim <- mod$optimize(data = my_data_file, seed = 123)
+#> Initial log joint probability = -9.51104 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+
+fit_optim$summary()
+#> # A tibble: 2 × 2
+#>   variable estimate
+#>   <chr>       <num>
+#> 1 lp__        -5.00
+#> 2 theta        0.2 
+
+
+# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
+fit_vb <- mod$variational(data = stan_data, seed = 123)
+#> ------------------------------------------------------------ 
+#> EXPERIMENTAL ALGORITHM: 
+#>   This procedure has not been thoroughly tested and may be unstable 
+#>   or buggy. The interface is subject to change. 
+#> ------------------------------------------------------------ 
+#> Gradient evaluation took 6e-06 seconds 
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.06 seconds. 
+#> Adjust your expectations accordingly! 
+#> Begin eta adaptation. 
+#> Iteration:   1 / 250 [  0%]  (Adaptation) 
+#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
+#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
+#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
+#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
+#> Success! Found best value [eta = 1] earlier than expected. 
+#> Begin stochastic gradient ascent. 
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
+#>    100           -6.262             1.000            1.000 
+#>    200           -6.263             0.500            1.000 
+#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
+#> Drawing a sample of size 1000 from the approximate posterior...  
+#> COMPLETED. 
+#> Finished in  0.1 seconds.
+
+fit_vb$summary()
+#> # A tibble: 3 × 7
+#>   variable      mean median    sd   mad     q5      q95
+#>   <chr>        <num>  <num> <num> <num>  <num>    <num>
+#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
+#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
+#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  
+
+# Plot approximate posterior using bayesplot
+mcmc_hist(fit_vb$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+
+# Specifying initial values as a function
+fit_mcmc_w_init_fun <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function() list(theta = runif(1))
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2 <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function(chain_id) {
+    # silly but demonstrates optional use of chain_id
+    list(theta = 1 / (chain_id + 1))
+  }
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.5
+#> 
+#> 
+#> [[2]]
+#> [[2]]$theta
+#> [1] 0.3333333
+#> 
+#> 
+
+# Specifying initial values as a list of lists
+fit_mcmc_w_init_list <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = list(
+    list(theta = 0.75), # chain 1
+    list(theta = 0.25)  # chain 2
+  )
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_optim_w_init_list <- mod$optimize(
+  data = stan_data,
+  seed = 123,
+  init = list(
+    list(theta = 0.75)
+  )
+)
+#> Initial log joint probability = -11.6657 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+fit_optim_w_init_list$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.75
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/cmdstanr-package-1.png b/docs/reference/cmdstanr-package-1.png index d7b8b97f8c19824bd7f4936ec52513cc8fae745f..27fd88217456b4a7d0da11a09599ea685749e06d 100644 GIT binary patch literal 15795 zcmeHu2UJtp+IGN#j)ie-BOr{@Y!pF2YOo9#5KyYrpfn{ykPaaPE90P|AR=8wL;|73 z&;kTSsRDx#Is~Ky2oPElNJ#ki!I^c}x7NRIzw5jATQ`fW6HZRq=Pl3sJn!DQf5Fsn z+jnB$!CKDfQHN+A?_Ee23`BFc8_N~X-M8I^%$xx$Wtq&!(VF}v-MIJmm5mR%TrqH zBa!fLZ0GfZTs=H8Pj46R)9)=h4&OU?izRd|@y(4K*+Lg!J@u?$U7d-cUhnYx-mvN!BbtBUKM; zhncpex%fmTzR9YMma?HP9}LFfDD?Zz$9sEw)zVrBqQ4Z7$@b~3Z_9GU7sVtLd%L>c z7I{ItL43s4zsBRYZ{Hqv^3)v?iDZXF7VB*)C@7elJn84C*b3Pn_Pysdk6Gc_!2J>$ zERv`<`x;H!Sku@m6I^g)(%{TWBrY?Qy>|oGV-(OIw zKc0&x*x)!xl%dZFOTKAC;4+n^!+)E?2YAS`)nW^~wQ^}0DcSg)@o*}#O_LJ6{q zmg`SS`y8hzVDOtc@~r9GH}Z((&B2x|4VpHA+QAVAcfY(d_@xwZ8=VhYFUI1at$+?PG(Tf{q1lM%p|dV4!o?dZ!mQ{f#?iC?)nJtwHH8 zQrO_|ovPj*g84alDI3C8&iVl^U&?xasgp7orO?J?i-<^96-jo}doeGLDPb$0*Uy2* zpZ{XWY4!)`Q$*)qK$Sm0$+8!mS;d;(|8SB4mk>nH1zFQs#58RNW-GV~2DQZ8Dik5L zz(6|Ytu8fDD{C*+;Z`$5QK%l4E*N`Nh0Lt5{)<>!i)d?nWqun+#zy;OZs_WZN!J0( zI0*3aw*nZvBW6SOuY%Xh=YEEU>F-?#VgKI@g?|kC??lo+PG5Uw%q~jA92tK%Sv%*!G0ai>eORRSV z^HeVTp=jei2=_OY!)9&e(gD`sjMTBBU{BRgCG}!p`wv%uKgRtJKK4I&`kK`?m*q~9 z&1zY392>6Ws1Xe}rm78|F!<_vr-%*7&Dr6@y=~lwXfvMVD5*n&3KDMc)DvOYp2bo_ zV6#QtPt?lg^h$ZiDg){__d1`Ne(R|S_0Yd6^6(_`21hS~0_l2z~eh4W*6{)E3>+l2p1y4JmF{FK1&GODcEhDF)Y z@+Y@%-%d>a%#$2eZ7RbPyR}(+)w3l<8F%{ObTY;&QppsBX4=xymGz-&DfST0DC?oo zm>7vmb$~d4N^F*vG#zqU^W@35%VEz^;w9Gu0!Y?2HU!brIRrxgl1w|G zY28z`rg!%}Xw1|Ja?}8pZyQ@!{=myYiVNR!R!^c{SO2VWiWnbRA84tzpCTy5im6#q zw_&iKl)DViwFabcuAD7h8Sqq3+FX9Nq|0nSiV4@S1;z5Or=b7MwZEIa|1X)&iv}Fb zvEas@~!R=$z~g4FOAmo(Yd znHg`IbQS4jhx%KLBf7?K6e?nDTYf%s@0FZT*^wDHTl>?=K7DFPc5n!smw ze@?tBr6Gwcl-oBCqz5?|K34JLVT1IGYE8>bGoW<8LQ0oO0oeIFw)YefQc_ZYAO(0{ zU0t0jf3P*I(%$rz2)CW33qw=8Cv2yu_b=*hT=(w(qiRq<`YQ< zXFHwS%#&OPi-1AdAo`)z)?#%WKHXloAtvPUV#?kJkcbHsLSe(;qFPZ{u7x zrCU*Ww-G5#NtK4BU!sUZ&0q`~P(R5BwaOy5ajALN*T$uD61s13;0bNGHan9>5ZktR zwma6AD#d@xJPZ)c8**}ALfj2%@Qa(p6Wml_g_+cU5j6FH^o0$w%5ptdOR`qyD&nBQ zYjA1+9|Puvs*ZCF<`Cbuvbb?d2>K^I6t?GUui z&O%@+1O&Q&mCdOiONjub7eAOwfXn@-)$6|zZ~xBRgj!i~adCf;-8{qe))yXbC5dnF zWtEx0^GR)zhhy{Ym~Ax^&56Ga!5mDy!yXLpDk0G&tgupcJB z2s5v%+jqP{PtPB2aqQd{YJJc6k`V0KONMu)0V!AZ=gxqEs2!v8visbL301~-qWJ9v zY=Iy502E64P@8Ux2Gxq%3BStdp2&yeP9|89hBe@j%7;DsdC<*AKx@0fpT`YMU><0A zCgNKN9No1huV_M5=$oM@y#q)3NE*CeuTl+^vG>hZ$P)5vdsGG*TI~Xk2OKL z^iQ9lSrb_>+p_p*3cmu8=T~0;j=(JqGDGz(>TZTr%|q3xiWC;+D_A;8A@wtV+OqV8 zhO@)bl~MuIgjId{c7Q|jd!8zjtyP0cY0j>kOULic#F`1i@*SX};3M58L`nLwi3dva z;Iz*_VeUe4`2_p?-q{{yI-f1jx$y;BYLAa)n3 zV^Jn;?N|K=ipaUy%VD5N;-H?UF-in=eda@$US9X-ykE98ec~NwemZv?bUm&$<8YpJ z73&nq<4ATl4}I-s5E_k!`bb4yC@^1w$D5j(PR`Ee)9;r|O|%u^0|8lbo*nrvLIcB| z0*pKb*eR{Bpx~lcZ-+n-PD`&kT?C4ON`j>o_5J&sqVm7w7Zq7E7>wKA{$W>KT!yJP z#>$fAHbdV2Tj$ySi$;G!BS;T}_P^V5jiDETIxbTO&ffh=yWXS&a@jvUL;p%z<;jzbC}sftPUBs+U&w`KjpQSTD-EF`MQjpg zu3L~2IFrKp+0hW02d|QKV5i>+l&4BAx4ssddCaV}VoEK3EBag0s(3IpD=?MATxqra z*4NhG_}`w}Up?jh=Tq_c$5t#@6krIg1^8PCQq?M&DY!k7%i!IF2QfLa2R1)m@`^aP zX@l3PhXq^ZEk{TCOD+V8GLsqB8EO8O*x@Yh@S_C3Dz za$nAO$U87|apm=tT|ewtScoy6VS96UwKqH-JCw+fi+QCO9d~cv(ta$BK}>6@WG;lv zU}$I%HZNYR2eA)c8)UyB>!CwJ0OJA0JVnwVgKfkF{@KWT z<*AU<34I0V~6V81Zzw9`CEQ|hri5xAL##X=Ae+5mzN7IzH#(fE@>c;76`_k z^}3=$w{3j>5yJ8a-;-6zX0zKpNpjZfcB&={nPm25*99^?*jClT1KmXX;2*O3{r&xS zz3B)YpESB+WgBt$RM7US*N5w*Que93yNWpuw91I=_rE|ij=-v<_~sqVwq)a;{C00c zd6EX&+p_m7wQqw{CsC#+ff_vaD*XAv-?_RQ2x!|iRCrc;p;v7k|<*N6AowEYkKmsL%v1lu@ zF`{NCMEoyYO((vT)Ra%??rt|{PbQeu=MH_zrtN^SZ$jRVIQ){RFyTFv-4|o!1G@^t z0onv!-x)y@uHfto!fQJsZvF60MeKSCo%LtqZ}VCs^S_sXa`9hj3r}<))bi0sb!l2Q z3@a67TB8DGF`aD}8}BdNGxrQU08`Lk^P&u*awRr3Z0N~=d;IYO#c2hywaxoI1NXs9 z)^ZdUBcpv3;bHGPb)YHNHB37)CLk#$yjxNKhc!gWPx3ts0_5EbJFmbEFv6#UD9Hb? z4b%_JdME(fFJ{XsO&y}&WbJQKbaP-77psxi!`M&O;zaI!)rG=IoHh}>lN=kI0+w== z^j~iJ7SvB;O8e_?*)}P-BU(293r7U_ocfH^fj#gc0_S=j}aoskeK}=vb22)TC4_(2%WpUcGdrw zM0(41UrY4-yzs&TUGhDn*m<)1^`ja5jP)6LdD2wRfYZ_1HwG?*_Mo=qT%XLA4O#hA z-`mN{eme4SA@U2=9jCgyR-*JTZ`~Tbf}yTcO|pESvzfN=Rf9zbZ;kwCSIg(cDwB|Vlb`1NfEcQ zw#svK+Tt*&6P|&GF#(^Ghx+DHo0=4FBYgTN_m#46Wl4@TcUI_{NxtWbH6-P}q4EU} zl1${H+^UBG@#sF6fOv2W{VE?n3q{vI>_GEyBsQ8qa_X=G!A9jm^E7Ef@= zMCU6useSYbaDBzG&b{H`HbUDU_ayq=yTbRvQ%uh&Ci7uUe$8XrdA&A6kBUV) z_0X{E5cW~q5B#Vr7i1k%xkpzn_DLt_0=PrNxWIczOvFhmb1LuM%<+T=<&`4#VK`B% zSi7h+mj_N~rgCI5_DDXtWXwRGu|y%%2#zkR#+~15ix@baA>>fOJ$+(|d^;A4z3(kd zRn<|nHyU!&t*uPv8;#H%iyU?GQ{;qM;cVojuJr{)*=U@MW=Upu-1%IijF?A{_(|~e%R=aNiG(EctP^|mJ3UF3 z*{qN)S9#8bFY^>l)XtOCRgMsf@6zoN2C0n+DNVUon1PltsgLSY=&u25Nqf{i)Km9M z{lK0yFY+cH8K*okA)zUcwmL(7vXf{XiY+ER-1?-KU!w3qFmU|&=c{vKk$7ojzk+}} zlPyEE8-Lyxi-gM=w(OV;2)`hjzSl8DL2uIEP&T3aE04UXcBJZpxnlI^-jh4hv0gD- zMa`Obhbl-*yLh}?^0U3eukbrGaTPTkqMS{7Vy4x;kX*O}JIxkeJf(d^8b|K610LEc zud5LL=y*4e?4X#r_+@tXKAY53;@&59oBEpgDM~`ZdwSu8>+70FMIB}C61v`o(cLd6 z>%>I~eIl~;r~2zrcELA37-U&z6ZXz}_q@v*XzDNl%Miyt@3U3$ad!R*i&aQ%_+Y$U zZ)tFY&Z}p5hJrtOswIVf_lCRAKZ;0&`816B zdXntaz5BYRe4S6bHzy~6rKWU|T{Yma+FS7^N$R9_@%PH9;I+tRYrzlwlA!AGhD^p# zBzwOyzb!mePQCYF&}94dD}x4@nNqs%Xq=54VvmB|(=Cc6FY0%{|M(zW=51W2{>_V^ zqr99=;bzO!?=X~j5{;d5G0EK|XMi+rb-3|qcXy$iaT%GsP}+jc?H;YVGmbi(h{~%_ zH`8TV#*E;za~r0*+pe?Zze6DHj{&;Ot9WBu?N05cX-dnT$uO^cBU6xwnM?ON3y*oqMdM{1(u$D4M=-j0f;BQaCgEFQrHOAaBUao5mkm!3Jc z6v5uagmR+c{0TF^iH!J}f(n`CaYJ}wteIX2g|Xq#D}%HsHj6cSYl~uCRh53T9&aic zCs-CSSJXd1CFc<=kCaXoHWmg>_vj9U58d56_CU3{iy129WHPjF0APn`*x}8i96Svo5Eve0V?eT+B)>ZpBbqW6O(-6I_SIu_SbJ4+#aGpk=fD zf&^;@gQ@rXZQL+ZfdtM)6G9E^j*#XcA7(3<;zH?{W?B9jr}xDPsMo8lrJp%<(N>yt2k6_-!DnJd>gCgt-$5=jSd2D<^ZjF71uMZ#bz z{cPj*4mltRH21fU!rfZJCu|U(oY`I)6)YYx1yH! zAh1WNeD=d@(gx+Y6R&V9eRRFz&RzH1I|nl^WNmMA6AG zpQwu3aOkbUA}RgUYeWlaIvq@t476>MNrsWbBUj2r<9 zo>ufczEnla!O_~I_Vwu=%$9aqZ`Wgt9NM)?&ptx1ULfH zhEwumi&FSp)LNxEe$7khyzIkCDHO@hI_J_8aqp4?i1`6dr3MK0-r>n=n1y2nuP2Q^ z<*^U6KR_K>0AF(4imkG#GWDI6Y*N^;IPF|a-;LV+WAqJg%7J~zxSs3fBxzI4(WiZ( z3<6Y91W&grAGj+iNo{R?vSG^}nSD6nMMKFD8H6wR;fI->JSjR750u_@^E}cUJ3LA3VvG96WfFm1mwEkrKqBR>C6Y&Kh;w9QK_Oo} zh)Rd*f7paEa&3qrlEHw5VKwJI?)2kMWO}r|`Wds;)sQa6V})UGJ;KkoPf|N)Dm~}2 zn|N29&N-K5(L)qq7;9!Fdpte5`Ju+Z^W8_SU4e<0ms@bDAbv(($C@^PimeK!ZtkIX zDp#1?-|?;6+4kgq@fwHtmp7;)o)rT2?;jZWhDcNe-tY|6gu#CM(v#!}?!QuIl<{Q> z{5;~t_YXij)cFf`Of?O1qtilSClG))cE2$Qhjwf<<6EKsy8md#-mcG7@<_b9!K`#E}w?HPxPe z2l9Xx(%r(OtP1*YtNi5^Iy@~Btjz|i;_`y0?rjsFZw{(dMRnvr2UYTB`mmh-HRY_XMt!D7cI zv-3NnaaKW67j`wiQLruZbFwwm!Q3AWz8UiFP$W;{`o#f24a4ql-<#M_6&x*@s!Ek27 zsY`&3V2<_i$rK*>de6IS*Dn_VN(u3_b1`A8Y-H1PG0iE`Llg#cLI7o5^;&#?liiUx zb|}S@lnAAQ)N-96M53M=WdPkU_9{xTArihaB(rY|CMXExnMZGtm*!M{^*s2g6qGC1 z&}cJI>mZN=V5FLv8tNG{0@&U;l3hPtEU#=5FC%?m5fo?DpqV~!(Rang^tb7c8Ylt= zW$eyQQ!p$5(a-DJfa7@Q!I1PDwOH*IkaKkh*dn-D5k$;>?Lb#>TNgNz=x$=2>yRK* zI#(mE>QeK-MyJR@kK%n%YmFO?e9mwNmb6cC7$6nUy&P;SvM6%9RM}yP-Mu++`lhK&YOvOT9qatWCjK zG(dL9u_ekG$167dy8Ecuymt8PZK(%$$vCzL?itW^fzjy?gVMBs!^GZ0RycncjpOyk z%7ARYP#stt^0i6%E<30ibD`pL53uT~YeB#?3(he;^FUUwQOy(jf;eEVq z+qRAVc1F|$08r)1s1noSAyg+~MjoQWTmThhFO>`pU5)~)gn#pat{~U{GEy?p)%6ZT zpy;EX8dSk>On~FUV@Bp1m83Y~5K#iM_&JDY)#j5fX`VXZNn>QPZStB22L~m4k(FT~ zTpo{u5(iT56=d`CWkHn-DK0dnOZh(mbsCUl{xUsXnzk~g5i6rX3-i-k84{(_;j1A) z145ende>vhQWT{Q1a+(oa=F1q1tcsMKL;h5O5uxMAoJdrgK$H{ynp{*hdZ8bTjqaF z4@fDHR+@+yz@(eQz->?5fdUc1DeJ6IEbjAN<6<%hk{toXlTYzCpCL0Vp$v>du3QNp z)(DpdwG8}i*>kK86xw*-(qtU37}p3OlU-U91aw6yxMl2v^&f$pH3Ndf#>QpP22Kqlhy)q9aAu03s+TBE8I5Lsbxwjs;PWUIL*6b=6S>G%)l~R73&+ zDFGn_a2Uae1T`dpkOU3I5FipwkPwplo;dfP-#&Zq&W^kH?%v;dFt5*f&-=dL`uUb~ zt{-tYr1kekP$B|R>xBaR5PZ!sG zyY$Y;fz7#*dk$-_(NNH;x%Fzv^+l~`G)^7evOi`2zU2Mi9X+~v+qQ4F-r^Wvjj?ZL zF`rY1;#NO7@xzVyvka0!B9+hTDK&BKmNdPLA0HYoTO_CJ3;t}jMy0YzRrqA}#Dr3O zM})*pzIlGJPZ{Td*>U>Xlm_{Y3fuvng8YBxxh+CyI6vjID+XQJS)ES^iQdQi3ZBwl z;(a2br^TsfwK8v>fHJ>XJu!;HOYaF+gUO_Ii}Wu16sU|BHF#8Ii$pklN%;nNN6YmW zmG}NSO2#`*5xsX!;YcpT05Z(y=44w27$E4!ubSm!@e*%%Oggei4<4<{(ELf(ezmju;_-;a+5|Yx|kM2UQFK?P>*fM9``G@z$ ze^M!LiQ#x~j?;FPZsg5}V90R|!c}04>_5Fzg>%HGth(hDShrI}Jz@CP#m>;s zNQnwwDk>$STBJpIDPFkZKc5eaYMm%cCb-%~?CKvtOKs@Wc5Jq4C!Ws=mP^(_3v60V zb`kyxh^L7U&;z6)`5lj5xK4#Mc+b4-s20Kkj4|Hnwit~6y+D7{#{*y((D2Klm>>zg zc*2RNTFSlJG~b@p#IIukF+5BK`ddNl>q8}m_oAP0MJPNTA7~Da@ZAA4yk_#5H@()e z9@xNg^<%2K9#!Dz6Fy+?B_6u?Xn)CIGwQV!1hq8^{N;N8yY>GC%tzmd^NFU1;pMn? z1}c^ULfftb=CWTz2}dg@x(F!v&N^V8n(<@dbmB^&iuqY}B5c0afz-dzO#dL6|J>z5 z!gA=I@rmO>AubC;4V~4(Ozmr3`1WEUjJ3n*z>3Z9(T@;S@H?mKy?0Of)^j(431g*2 zdTpH#*{%uGHoy{DkG!JBAO5!F?W!;1?f+akd|{~iUq$>eHX7KkZ4cR;wLy$72GIC2 zN~jk??kD%bz^ShesD1b{wMm>3{GjzbnR8Niy)e*4m?n3J4_g2#AEU4A!>{+^eWyO< z>mpQV4FbTnyaeF;hzn`Z#lt`m26|`D8aN#P?RfsbN=Uyq82w))Ud&hwB?+VW4+A#f z3tA0R2@Q5`DK1@dut++z6&T9+*$VvViKw7fA@)4qY7GeFNuEGbp_S3~_+al~@~PK? zPM76f3>{!6YAFVPCD8v%$`%M`^MEtP#IicBssMNq!i!GZ$NL8Mi2RiomDToPsuCmU zbT$!=sZiV5Z)4u9iWk|qSkV?cs;Rh;V@u@#3Q;e&t@Iy$6-*b|E#@y&%0oIt?XpW> z2X6(He)I1nv@ety{!1Fnza#JRPy@v~wY(P_C%!Kd-k=G+=y^W8CWaZviLF3pmDIiK z_*9Tfg7_3~3vF8E7}KZTTLH5E8lFn2p}Be$h7Efu!U|RD+4%eF2YQ1Va4htr*uZ8} zwPW(wcHvui-`ZWhQzuDqp53nt@9aPJERkL+a<9sco=@a~SQc6jD}(wx(I+aZRe0k( z{1C50BT(*fU#fih%}C|hHsr83wX=tQM9&iw)jSxj2%@kSm~VToTA5ntMdx`9Y6DiD zuK)trQb6*I3M&Ovu|aJC68wI1@TB!m3sd;yL4qqFG2$)O5T^7G7g)bh`1jUl|I*7r zaj8GrQO@C~y0kLUo!2|WcAu-`3p+1nnar0|PR#DS&`_{zD%2%vPaQB>pr8BS#^eyE z2zWh9I{Han1FK5RkCSTzdRUhK`1cjJ|I)(y|HS`K6}A5y$xnDF0ehFS-_*dFUu=}5 zmf_)0OLS>qp+U9-8Z1fnpuYmO-Bela(td(Seh^Z}+U@rMPd^e>hi|Ri4nVkQ3;2P~ z-e=>*b39O2$H|H&6y1~|-}V1sPk}332y#=?O)KGv0%SS{|!};z7MASud!7 z+G63HDk;s+M9oT1Bs5E|oXXTb?ZPyxz#mvhZ~>)Z+iSQdm2^nZSnfjdAh^cYjUEwJ z@nIMUmHvb2|97rT|3O99XLB_UQ#2Y9Oh@?z*I0&YeI@^WllvZ6tvAS#r3MAg9!PPmf@W&*3K zn(77t>_54K2g$?mu19A;6#qNAnqS6);jH5IqnX#Ac!msT0xJLQ6;P{z%a^)@r5F`P zlD?TfDQ|c}ZJaeRM*X8qe2@2eP>-QPTz6cP*JF@x3wXBv#Id06PwU1i1Ih4uXe_8B z7IuCpDILy6!1aamq5qTF#vfEV+zE%Eya4}le1~}ALqkVZJ-}O&W5O*}%NF{Qz539L zq4-yDPMdUawW~|=0eT6JVSAt1;$wKz#Z}~?mm^B|qc2x-uY!{H${)MBbLyA{su5(F zR+bN5vvOxU)S2JKDD-=7QLlr~f`LvPdUEH-!5Im^4G8OPV z1pU<)1br0`Dt($op)?3V106~mZGY)Ma$#F>Z}rJ@OzZ4dyHC8U zjNNblcfwM-P?6_$Yptl_*pF8O1 zw3mYV3T{bif9_?t+k8%zUpRgJxp$oD1X~6K)n#0gzcMH)S@#)x{`wXHOQUB2O|OJWFzO4gJr89g1j z!Nr>8Qp0wyt{Ul{m9&RP&`gZYtJ6lEIwY(pQY)b(RStGVmaIoJqSTUJ<6-@2U`eRXu*nR>2@)@~OpqU6m6 z!dZrq)GS=O@4QO%?#)QcW&6@$yWAXyg62l@mtG4@U*^CtDIjtSz=&DmmmJ{5z*5op3KRQZ?yB5aarJy?8UDY}HGl%O&f1V** zKev%N&FP$$@^Lq1t*9Dw?VrP_F(cF!ocwCk2rB%hj68U2=BF?jscb=(8T9ngE8B@| zhcWW{?$IFQWdTpv-P!(wJnd(T`LFC}t7*T+gEc(nGYP%gQvA zGe%6b=YB+Q>?%*B-*dK++8~3g46V+Y_H16BCCH*t?&5Mozz(-^C0A<12l>jm)ZK^e zcj{>LJfO&Q4NuPQ{tKPB+)lD6|bR=dbax+1ooZS2owWkwa=%_ohZ^#}*;j!KPi- z8h!W9YEqN|*@v|U+|x@-#+>c_H{N4h*oiZ5V8oXpl#FoxMc7R?f+_b16yPW}(xhhx zPFEyr>2#VJJzp|%CnLU+tE&|nr9J0B*x2)|IX&lfmk2w=ip=1YY2?}JDVL6<-mg{Z ztbIl4qjjBVRc#$gZm{Tlt{5wE{_1pmnk9@YNztax{_m+;oY;f#yMD;}viI#MPgZnB zU)>{fN{?3T>? z_ABNlqhN_-VU{g>dH(}gu*0LixiVvS#5fKm4W_vzU9YR0H%$?LtEi+E`{5^_(_U1Q z@%+GabB)slYl{f=7X}1Z5qgbpZ|F#;$g`~ZQu}d*>^*&+5#27xXrIu>OAAv^yC>c5 ze|WOJ3+uV=gRcHnQ+v4h21xwzzYTLRe08*iWI0KTxND$vJVX4_=^5|%@#8JMk?WI? zb#---?B3p9UI4)-#B^obz3nByzD}#+g7a~>go6a(4&>SNdl@yYG;VVcyCtDP$-u}u@20@GZg9%d zkYW@-aP&EG;!(uHd4<>h86PH7Rn01M3JRoLcOJ1ML(eO5ZZw2VLh)a%IkmW=j24-6 zG8czixPtJR7NQGTO;J2Q0|SG;UU#Faii)8Dt6M>5bd!*~CL6<|qebHZ1omM;GIA#} z97zxob#!#tLyZyRwpVfLMT!SwRm5+VTVqTq7ZICWj~~~ZKBJVBz6&244OZsV*48%I z+uKj^h3E!qPmip3XlO`nbf9;Mh- zwExEcWHiaqRzPv*l{W+pMh`AlGJjzf`R;`f$jV0W5dp}`DJ--yo-ADI>zN{k<{j(k3nGtDF!WjF;F=H<;luI7pOY?(DyNfTMsM+)$I ztl)koqqCS4J@tBsCuBu4qxw`m&O!Nx2Y2;ElbEgEHUl(v$hC0gbHD6g*G!|)^bTX) zc|pjDYMwg!U9qeWBd&985&WE$UseriK6yCa5j0p&xWH@@j&Uvso=9JHW0Gn+04vJM z$}%tQQDU5El1s4Cwl-6YD`1l2X0IX+S$;Po1tJJprK8hUmi^PE1_n9-o8i<#WQfx+ z5>A&JIrDU8W@e>{^JxXU8z*~G;5}6NlUJRd`7Ml=I_l_KuS}VM^(4KZJ(@|hG*k%P zyHR8wH_JrHj$uE{`?SOz-lgC$037dXaYnccw%46+N;C zK(ty}Sz*}d&PC^vk3J^?V;e-1S(6fpq^Htqi!~|5w;;g_Gk1g_ycFCpFGc9t?oJ;$ zU6O&Kc6T$Q#n%-asQZBh*bL+&a1P@y!uERD0dYP5d<3O2tTR=$PqWeY?Ab9z-p1uc zL02vMre6~S0R~)jdZwqS{U_^k#fSoCt-qrd#a8UY^0sN574xcOB{?}cr61Ac*4rah z40kKqRJ1B>(N?2x{NHPu^!qrd`sbw}oor%9R;4c0*QP!ldKB1KS=Ol*H^~NzHUT4B zAE6sszj4Ad?%j*xhD@i#(}{`F;p_EFjcr#26)Ze+ZK>ftH6}8AJb__~0K2#8g$Z+A zKGL;-4H1H$Kri{b@IA+?Rh6H?duKznCLlL>#!eRjz+%Tw9qWiqNn>?|{U~wJg#n=FFU~>SG>h|H@ z+J4;zmjVfx+>t9pM!BPbUK=v!Cxe)@rY><*AfXV4V-8`&%!GN-!aKVZAlQQzMu3%; zij43#zx?tG#+-Q*3NpIXSf75T{rj}7_puiq5FMT&)uB$qOR=JMWKDB(^Vy3YHjP!S z&9m$7F3QJV-xf1IHkNp6WG!+hX{KXSLq#>kV7?1gK!GKY%AT z0?rF38$D!ZR#u$W>-D~i6oU$Id3hv4hM?G;m**{hl~sWH*yKfuw%3?zL=+Yj7Rn@R z!WXYPmb+Tt&I~mjAEZ6X){311Nx|~;GhtyfrRkDH05`!v%RZLTg%X4y$>}P~h}E8jTMbO@Sv*zM)F%`8I-y_G?O3vkZ?A zvTT(!rp@gfmO9)GG`QRpM~z5tJ_RT&e)8lA2`kyJJ$mDZE0dR%Ec=n+!B~3$csSwB zrTW~fhCMLQ0x)j5bw+a`&|2+-aroUEv)ZVOwcYw@=M`e^00QD>r@e9u3)x2l$tIM- zND{L>IXStoikOH29v?8e^S=t-f74uO{Wdt!NK&xDYH7tfV3oCEkIFvR7TIwg*hA^P zQjYX2Ncq7Py4?nAbsa8!7`^9B!41TK!SK8x#g*!tN}N3ZNhn{}!RW%OSW?9p>=<>} z;^RoLv8`&2i=oaW8K-LkJc7jHPv;A_nc4L3=iHsFoB^7_QnIpwg&fNvO)CQQKr0YWU@V7jkT7n|W0ju<})lxLmp)j9?2yqm|x$OwD z1&rtL__G+SId@NOx2Rzt>x?aCG6&H*B$FNrfBP2xY!52>IW>bg3t(~UQvPLRq824Y zYao~Y_M@O;ecd4&R(If}OoUMHTAK`jLhT(N{Az>OR5Z|qD6lEt=A_bX<*6NNxT|eTl_E=uJgy{S{ zTA!jN?&4QeOKlfMS1z9fH8bx%kd&{=M~84yLSQ^U`H)5HSauYDQqRiDx7kreD7mjN zxawIwc0{WdZOs39cZ7r)V`%q(ptDc*I&(uSjOM1Je%eCY$A*=IG-AE=R%LIpcC1WIpXKG zBO{$|YYcihhZ~oH#umrd!vFp1nwZ`XU46T#Gc{yQia$1fFVZM2iv-N?Ae;$rSbo>| z@@(bZ0$F!@f8zE)A3fBOdYyF1N!c7Ptrj#dZr%hIniQsgW7BPB;RtDAod5K?o}(xe zpy!J!3)OPSxVuzYKEVa_7o+)scnIl+4EBg7YEHK|!?s9h7 z0~PBZDk0{wVkM@JTGJnwOkP3m1U~3vIcB)mPng?SXculwNkB@CDUJna>$evCJaTto zmsRG-cd^aCGV@2Qv|_hWLNo_LFZIdEf-U}k@^YZbmM1Q+;nolR_ync{dB9fZ diff --git a/docs/reference/cmdstanr-package-2.png b/docs/reference/cmdstanr-package-2.png index 093449f41694c87ed4f6a948f6425a8e02c8a550..4ed9fe5c527f0931750caf802c560ea60a2989a0 100644 GIT binary patch literal 16009 zcmeHu2~<<(y7oqjv<}c#1(B&%Dk=ghqRhc#72`k@1!Sg`83bfbfCQzLDgrVn%3M(p z0z~F91d4!w5CN0O7?dGEAYn=Zge3Rdc>1q@{r8?*?>#;D{%5U|71j>fdw=`e@B2Q_ z`@HYBbLXO|$<|G~HbD@y6?x`+GYFChfgtf8Ki>$x8F={X9N299@tn!`U<+O=Cu?zF z^M&sjD}M-5jDr8f;>PElA?O9V09x@qQKFma%k$;w-$$B!$Vzr?d4%x@R@JMEQQoBW_zaiZc)Cj-A5R>dnhN0X7ndoK>DKSaUqq``PlG4*r zh_xRIX-ulWBO#~PJv}4C!kp6Fe9F$wPTT%7K84L=&X2ZomZo~GHIfS%4fe#UjrsXC z_4P%Ii;FGMjv{-UfTM>%5SUCR^|B*jW8?`#%+mAx{D&RSTA|pU(bhQ6;;V0)o0|*k zMN3Od)SEPuqp$&>pZeSML&V;Zc(TM*>Ogjme9Pz$7jns@`g*0ee!9YUndn1(RNC;> zS+pEDdJwfiv%zrw99=ID$0xZKx`w@4iWC{KEQQ=qPIEPR92h^M7TBJ3r#Z(`H9H@% z_?(lFnoh{C7e(fe=g=^401j@3N9kdkUyxe%;r@&Z{t|p?zhCrf7ubHfZ~ji#{*BY0 zzt`}oqqAwgRlO%08s)ZZR9@M^eMTpvJG~F>eCzwp(@@;LT9((uU{ra26Wb9*2*(gJ z_>f5laQG|W@Ze;be&&!+cfpi>V{s)c`M9*??95C7=d;+?A+`?^&O3z8mSdJ*a5lYD zq!}yJwmF$4wms!+ar^ODD#hy<7(GQU<}~L`vysWP2F(~xnWnZ5xsMR*>gsxR&Ky;c z@-)h%t*oH0A4{eB>wMcWUQnU!Z8++W=FV6V>1=Cqgk=$!eP7>WFuR^;&#e6XaX%|d zOUoG9MxkNITvkz0kw>*{MaG6u&WBykr_oPq>gv2ez_P(4E%4@c4i5Bbes1qnVH~ly zp|;Pg)nN3- zPu9PSQiLJxQC`EjL2Qrj98w;$(a?h4cC7&IkQ?>ceJh0Z@g$ze_}EYxte z8$G2l0U!EreR*khE>5KgFpx$v%wg$iX>a|i3pkh|%%&N05*@eF#p!gp9Nqk6{@Zt? zJ}-;Z&Iu~bIC$mxK>EXoala|k&Thb{n&O1b`doJ+%d%bC#n!=rHa@<)m09$M>xD~t&^uli1pkJ2V0#F?F>_y!)zJZc_Fvgy3nH1 zNQ6T9h@`TGuA;Btyah?c|ACnJJHfTpMjARZr%BQ!(6NjcqV_c*KXSUF&@!xsQ;~0J zj^J@hQ>3JVXKI+`8Ss2EJAR;TiV&YyuGcDLULd$gu2oL-o}xIt!8GUE6yPh&IYeHR zw$4cAZr!DeQs4wl1>bx~=U|+xxr3f3@q-H;Tg$mcW}^@VDGGCY-W;cU5#e!-GR3i? z?eHfk^YlMPcKw}}`p2C9{0CTc7JA{QeFO{Ht4-0Cf)t*o^580dF&MX6Bb;}r3wc)g z1q6+43lWRRrd-Ek*@6>>X%^|R4xB=seV;lp)SOexuJ#fKM(om(fJP?MFXRtDsKaW2*oinfuJvic_3aA|1dU)Z6*CkPKix{5R+v2KHGly=X;1{W-(b4duN*fmD`FsH zB+ZG`YudW_UfatM zf|lC~Yw{6#Y#M^!;?CTE9J3hq1_N|t15tHugK<`K&GZu{jUa3+<0Q$0+k|!dAP&g* zpHjBsv8zbtj^q5MTyvb?Fa}fr4T6bJG~gd8|Nr8x{K<94?^nNFYh+BI4RCgke8a&Q zG7u0Usda}`UtDHpmu2h&hM2;hII+1s>RqDrjl-p-rFO#x3T2YgN@W=hYHt@8r4+tN z?re1iHAbQ;3P?k;vtBwiK<_#bnUBZ}?M$)_Il`r7wAfQ7mg;p3-{%#CXWPNl#xfOW zQz+&@_pMX7}uDVa#jn zqjN4ofn*V^k)NS%Lx5x8)!07; z-@WxsK0eImdFl*K?H(D~vN1Alt>28q)>eq%KcO+Ji#}w9FZZ{uv>%q=vat&n8`}zn zur$=mXw0G2Dt%&WCu4D)+5qyY+dPis)ilwi8*cjZjFeJ~raPm55TyDbybQ-LhfkH~ zlSH!Cgpn`L?@dfOzZaIVrF{i1rB<+vbvv3LB`wWQkq8?9y|(5gY=RnY+IZy)N8Bl- zRSxx>cSWB?VQoHXg#x2a8wgqhAZRV|=7DyV9iSyJ1yr_DoM*KyHK_J_jdQ7-g*iX& zE;Qm@VIHVgwCca^x7uWLlBg-4+5BXqhA7sIL??7lqKD{S$I9&q@__wP&+i4FJWKn^ z)(dg8uemubFc_oy?Ngw=DS<_J-dns1VE-U3t#&Bp3c4mlsR0i3tzfar_q9# ztxY5K9{jJr4l4mopy}IPUDuB8ikWVVlr-n8#V9ct3|KCMN&#*Y{hz*2V@E@+J%QO* z>bl+#z^+vRezvo<9h~N)fSwKa5^dN9$JHly$bRR!e<~RMN!`JJ-TR+bEjLj?N{djQ zsxVJx+B4-c-H*JZz8W|MK`C(78;Mx!C1g@?LPR1}%cxh@v}Tn;S_H>IkO>NuAQ9pV zs>uR9IdkG+A5ci5lFiw+7`_J7u=`8+?1p=i<|tEZLX}lz;li*T6wx%|91KVKANGS` zg=!{P2+rlm`{33z-U13UD6-Z-2lPHsSQwU^F(A+U8cwpO6q-kxQ9Vg(|*TIS?{Uj`amqBDhDSDa7cy2=*j0zs>`d+s;&>-g#Jg6R%vI3=*_E|)KC{0`zys-f~X5~*rN zeL475s!gFKI2u1r42N&*&-07j@bc6Y+>vH2eEH@)iHqEJ<|412i4156ZyQ*6C6TdY0wZh29)N)B9Pi*)DN{ zzBZ2%q^}N2poN1n(KU4jgPw1_wepi1Lec1z1!lL~QhWpdJw5W;4L?6Ka0NiWHmIA# zU7!j4MV*qk>H%B~!S!u>G^l#((I#!@rp2PmAFH>WUBOvOp2xt8w&nswl+sEv6-= zUKFrik)D8aTN35rk%dJ0pG&H)EKJmv`8Jt3`Fz#?8{ z{u66zk9Qa-ETK|kHL!TPJYznOYG+PArvhX7sLGt5@34XLTL%Fz4nW~N(Vij;=nURB z^5}~?g|tQK;cEW4qEJBkmH z0?PLW*2#Mb>HbR2!Cb+VO==846c z<69#@UEbT<3&bn$ZR7l~RyeZ$5Vmlb!-fPH`jibVSYf%I$sB3nnmNenAEUD> z_?Yr%!-5hwddwrUe z)Rop8TeTl;4ip|r&)X4w^vS+6M|RxVu{ZKB^WXk7V)N!pMfLBEH#n8Hu`-7IS<2_Q zPn?doO{9f)C59PkJsz6TW$?mK8U7h`|9UhRUDI=UjGaeK%)?&dKK^wqG>)?`M9gi! z1mHsGa=t}E_JnNukACDWsspzZ_PD4nVxmp8s$)nRdSZ~|w~UXi_I)1@hm4W?;dT%N zojv|yP=4xzb@Fj|IL}0eS8`tDe2bw;}%&@w|{I{efR8>Z{~kxi2U00NOFduzCa+D zwf#&o3w`aulRZ$~Uq4VG5{XtXvT`T)y*XxRXlV7q@kIAw#};!uWr@Ri zW`V!JoR2wut-r(Dvr+%pv4_RQ#X%ak#4i%pyWmw6E*` z4VC{D>;8uUtL53WZD?iYw(w{Q9CeOFvjC7)A!wh31pCAU%fw@)Yft&>!X4>Oew@?Y zV3h@evJ|%0N&d?6@EiQ$RPLj}-ji)jHgJH0S2|huZ=vJg-EvJdy!Zu4M!IOByA%COyqmR+ZEAxDSb!SlCa{O>0E z{|9*fYOa9Wzjlv3e(xdV%SyxMa0xzMI=z&;_*-AhlWb_a>AFQ!kki+Y-?h~ZpY6W1 z*IfCW68%yeA?V7EI_T0mGmyK@k+8j*$Aa#ukcKY>Y>Mc5ukgtnF?322_S`$0PFK?K zLT~-xqFeTxgO~HJB#6+jeeDfg3nuq9h<%(&Qp?=rGR@`~=NaD2>tWu#d*A6#KUia1 zO3Nrp=rpV|SVWI5>xGjks(b%lUXyBqi$9qs{?6gkF+;HnfTt57-tD%hPuF_CUQ3dP zU1vaG(D#ngAAP5O?8rk~QMOgJ(4_Qk?2%`*wMvW;Z^HF%tk{g?dKxf})=mZX8hazf z3eFsiHz*EA?QD^-6o8y?6|7cXN4N}o4FY4aEMxR!)8{Q zm70#pPO!G@M(#a+G{s=~`?y;?XdP0J&KXX?cbfWjliL&JcOtP1%qN#;thCzR3)-S5 zWB}3;IE;a+WQLScl_Gr6h>vLck8`>PiNASvsQSyQj)U8zZnk~ShQa7U{1fz*tHdSa4V>n?#5uLr3= ze%5?zXnLeH8J-7pXvfDr|49<|6EwH!mBZ2VKJfY0HJy1adL^@yHD z+iizn?;F~B4`-gMvT@_F2h+4H0@8Xe&Q-kI9o}|nG1~NSb*6!?@HA|Pkr6YdX)K}) z7&z2rUtd}I^u;4;6qPl28j?AOxmO5-*7_E3I} z`f>(yfmNU?uDF_VzJ*TDR;;gQ)>4BdT@|oFnSLfje}}}jWq#<>CK}Iek9*LvQA-i| zOMNEt8I1Fzy(^i!cXd9E@9L@=?E~LOwbDAoWi!kp2NOG;M`Jm}KQa#v*52IfB)Yvid$nG#gh3&8_} z;N@;uQ7GY)-F-&>Hv8&;NrJjKNMhF6y4Wk$q9uDxEf@ZoBe@mu>xG3Uzt`3^NV$SX0=Q^)zg~_nF3nu@ zSsB#L>hsF!-X}bSKT^O>&qkswdxjeIjGQICB5P`Da?->%GDrP`Eu#qR z_r*>RIYjzi{_A(hmR+NyQ*Z8ahA7(g**&qkS=*VRDpQ(`t}U2oYOsDM^PIg-?c{cT z5gzYJvKPgi&eqyIeDmhbEz;7Jn2<>~Aggr8?w?#(>P5K*1e8tc9Lio@T|MI3m8Q}* zGU6D-8np8L_K_8kOHoP~w~Ffn598ydghR*OBJKFb$RkIOCe031%BFqpGWH=+&-^~TbngreQ1uh(KSf=%QLh6YfD=${#47VQNvplo=r$YPdv7& z@Um*W<2dnHT<5A_xfRjnYP9%RrA>Z#?smb|;wK~DA6h&>wtwGg(9k$h+^W>!oNV_{ z1>LM>AJ41j5hYY_J+37A8(Uh^$_Vb%g;H;?zLM{FbQ~_h$4DheJ$qC=+c&K*y4W|e$xBXP;Vg}(t}hT zuUqUom!J^%_EMpeQiN2{OmQn1PBkOs$3}PP5mFP=9lduV!m@thfkdvD2Tj&+iWT>nR&Lf&r3!TUc3a|G{*Li&4EwQ{X?Ptxw#pb1ILE~F0<QGV1LM;~oBE z@g@~ha?l~W*}-5@qcrw@T=>j2tKuI|bs6#oWxm^(Wo~UK3D(RCY$`P>hk|Ud+|tZv z1cD}fi^Hqb$_XA$#2c8~GBR?>{a9*e9_s~qaUyK;q;flp74+K|s$&Zsny6r&J?Evm z;^qjh=Xel57a?%hs9*jt6#GC$SvVI}>eQAXXy0tiEc2&_J34+NYwLRg#F;<0&y}&Z zNT$=Ppt6g2Mg*XQa~jnusmHGO&(6x)hy2i`GkVXR@pI%*l7tUX`1@pnFPcjH#@hichKY~w@JMwa~8e3vmq1tv{I8R;SyvXupHahpBHuR>)GYLx@>$GpxYqy_c6*Al~N zQESR;Ba*%Cn~XinJ?R2IU4-?nnaZ&ZdH=I6=j~-4VjrYV1Jfv~Wt}}Dl$QquB)L{F z=uhrU(e~WiBEcPcD&bSw_Zo#=NK|QKuXV9!xJGciC5-(VrIg~g`=j?bL3}`3rU?o+ zUVe3Xm|%gQJ$D}J&$8pQ$Oen&HEL%{?nHZTc!TagEV)hq+pFH3PR#f!l+T?-CV<8r zC_-nsMzzB;afaM?Nt@l*?*(5u0OA>k7@w^i^gEX@R!#_?t$+Dh*4F-F_ORPg*fUF9 z?nKWy%?r=W+%H^(tn(H98;ZVNme3420Rl5A)w^atcoYleR9Vj{%(KX?W2r(~PapU)f2D~H#HkMcGzdDtRS|eKEou_)9p_jT%{F}BOhtavo{!tEc^RQji zptQ8KtCN#crEEQVZN4>81x7%Q?HRS!4OxQz;vw&2sa8ZxU=fkquhRC|HRJIPHH#VU z*&1kjF1S7!CtSRD@1B!jm7}}(S_?N?aXLF*Wo03mlt#dE@W_NzpZdNPwAiM%(_}HS zU@Nh$sPabp{@k(K^4_mb)7<5~=e2A)yVG7A_L-RFS4S?|=F|AS`?; zQ#)NiHYAVrdlEXoP))nfEUMLDhcWaY)Iy5S75OaSP|yP*&AZYFVycgq<)CH{j>5&f zD}JvJx2mj&ay^%?*G@!&hZxxY_Kt%|BTs9hSE75}Rep3^bv_-hADWr-`QR0HZ;C1_+t+33a22>_^LwvxCL% z(x!zjeI?Gw1Rv~TDjyb72P^lade z3==RN$e{6)Fuw;wE3`?L_o+J!T29+e9eTk&(8&;bzx6g)dUJARE(%dMF!1o8RYJ68 z{qk{PF0s$cG|hoIjVE-)n<(kd7DM|i)Fxm=)d^;OP$dlfyoObY3Qowt4L@0eX5uWc zQt#5El(9O>t8b$G{+gG3?0g}v4L~9cO>h@&b8~ZxEqDw3N-Ak3b&bUqdT-7YQ?T7j zeATa^8QRyeutLTe9Ul(4#;Rsyz=>*H6NffM2$l}o*M(I6JX5sjDhLWTNC(p~1=C^J z-wZ}EECYn6BD%WU$zo8_IdHLSvATO{cENTQ1pD+rCUxABv}!J+VfzTC<&I=c^6n^w z9xmfE$r3}Qk9{m-SHIuXK2g>DaohAiEr+tQX9laa%B|JG8#r1`5A=)z%2=Qe2Q~QM zK_kxu=o~ZO&p<0zi*-43^$KvLJzzH((0O%!1?o4r%yh8@a-V3=p4=1IaUVQ1{vS4U zv?kyL1Wq?vh>uf2W?-iC-g=aS_h2ZSSQ~ymyZaT0({)e1IVs(lrt1$A>e?bg~o^y*Mk$!TZv)OS*DZQrp>h!gVj)&@Kkl!E0$R8KBkbQ_8>L7+$xY8hLbRW3Y8IW6V z*K#s$C3ywtVqsk-J(wO|&93J!0p0|3r)Ng62my3K*K8#{!?UP*FUpW}Wmn8;(Dq#HBV&&tOKyy|ge_$& zhYyctge|nC`Tq7rU)2o|YZH%Pi-IPsoVXBQUfpl8!N%rf#aJ+oP0az;o1wyjn-az0{&u3zD)^rVnkn%z%1|5W*COZowA*dQzXete$d+@H57aDO8gEI!gR)p%=fF(XI!J5h zO!cOos3t8VnBB1fWBqcwR8|cR&^os-aes=sq{1zIR%B+pzhg=tp}DVu&X_W z^cxDKIdfuV3=6F>E}qFP}ZrAl+KBlyx_qbCJF0QQd6r+ zA1b_Zz^W#|9hP^JN+FIUKOoAz1}fC!5W)~8Fk!FZI$gXaAwwr&Cj|AM0uC{5ot~W? zVzkcad2kygmCqMiq%PEh2dR-jGCmBKP9D1{sq{(}vC*{N=i;9&N%C`szw)Faa`AG zZoelF?FF|bVa!YIGf@r?IOaY0XuKoY6$tO^RxLxpbi+$5;Z&y39Tqo3O4rv0F9QNG z)?e=mOD4hD1(^V%v+{b!G7Tr7sUUGUTv7e%G!pgQJOFd0Iv^euc>GOoWB1sSjrcPv zqYO$okOZnk`sAdW6k=@-aK_pINibmK95(rDRDud2q*<+pLMb0;jfsh=h8J(H=Yj}v zY*EB6zel#j`KGfZy7n&S94GEKQMmDa!WfW(=}XvT2P_&!4ycsUZnVhpljf^NW6veP z!V}2#T1DJ=`-48<^Rf&urx|s)Gp5Cg@d+D!$=fWeaUT5ku){R(4N%}fZ+apVVAak> zlS3Eq4uS(8VQf?3)h3BpZY;2^K7w2T>~t6utwXu@;A$yhesc|xX~%8{As4SACnwh? zE^Tlh*pjg=bNTxTa%QH=3PH50$ruH(t`0Nm>#A#UhR5Kd1Qf!VD(**jsVj=c5>Z`> z!GrODp~8SJ?2NWwdU6F%0q=vhRfjAc!UOXu9!9)svNNDfe%d^TfD4BUs_*19e~Hui z@C&F}U~z0tz?gg2&Q=VDjK*w}Mf0anxD=1_6j*wwhi8CQHFv!vkO;)z!b(oGLS_JP znt^(r4cC?MQT|jb*Ks>G#&V)FwbEgfexp@!1@_S(nvo9hY*pff90uhR11=drP1Os) zR~#39FjKgcdl^qD!EmP2{Mz@cV}ii)2fQi)7D&`{-QX?s2`hvDX^Q)^hY%okc!*B! TDKcAs_5^v#^!uW}INtg<4VB|G literal 12652 zcmeHt2Ut^CyY5z0MnM_J89_mr5=2y_NRh5%!GM6m4818$X+c`(MEw;V1ayZYCCM-t zG7cgN0!9eIf=sAlfPfG{2~8w{P(lbKXYF{}bDs16^WSsl{`a1H+|Q#cthT@PeeYM- zzJAKqa^r^W8z2bUh(2-D9)i}!K#XLNrc`+1u1C+!>b$$x2{|T$ zf!W8mad$ZVxwc$Ra*`m)Jj^H{iig3r35jE3(i$mQf8u|*)yyNy>=IN)t!SBmnmw}Y zCf!`6ki=mBYNBpr5hI6m)IY3HGLw$rWGl9tLAdYq+8xX#r!oeOjFWF0F*%Dn!RR8i z#uMkmc`p~6l!Wq9q75hQD;wEThSJIySOtb#kx~CVn4*>L`LLIp#2-!T*bDlHVm^;q zm0qZ>x{c;tff?K!111ms`k1jDdEp_Ai;775b7sAil2cYgYmvWqt4*&)+RQJI=@4YK zz7}a9@)}R1!P&3YLfTKuKd}zM!6(@K!&Rm;a1MfcjFa`sh{ieX8N>@-TZK6+V!k+f zP98|)B~bby6Vut4Y{Tf;_+|~4lSQoHD!Oeq!%JOi0&I5!%{Pcy>fl_DC~^v3$inP0 z6j+vlrQWE3f2i^!C&6uX^1LmXTfV#8HohaJ*W|=Pcaf7yDAppD^8?HoBnnewr}H*e zMQ#Q3C8%Fi95T0IRA3kIaskJ+dV?{^bC=>xZsV1)uw^@{Ib_8GHlpgp$$-iIvS4ye zn19ytkmd2Ro)1A7wRr6Dvd0#xw3`L&ZM1hQr7Xt0Tago(TTcs_taJcBz_QE#6YQa5 zkt_t=nCq+*itx)X#&-yO%*Ip)W0K+`<~9Kw#D7)`)n7L#n~h**(zgMBb8o9qh)TAI z1FIgG0Z)KKsV3_JWI5H+`$P<g?^c9W@h9SA=RN~2M-~N8snDu7)){`w-P|IM;m^XT848Xgkd4s> z@0>nWn%%b}1~VlGeAOYXe+-E4c8#xs>M!Sx`Q9=Wb|{XLcGCRt^O?;>hJY-U06E6_O6kzb6iuo72!Cff&u+s#?M8D&%Yc0DnD^$X&mvojrjm{4UWy6{x;C7S+Jl0 zfkO#m0_F|+!ho(JOYdz3!elxb%xoCdei5R!&y1?uaOAD_hd ze_o9Ln<)6_<%nk2Kc8Q;o^3c#J)2*IMZ_02x5o3}Xxl(j!~D0&x;6~GIKo4GNXWb} z#R*~|A@kQ4)_rt5ntzZ@yQb}gTJP5#OvZEzdpk@v9k7cvogM!QDp)gHaFUVXk$6U& z)!~M8&sUBWCo2J}p!7@ia;o(SH%Tp-T4@6pc3)7$DKPZZCXGn8{0U@oyRZg@prR2> zv>dV|o{X-VtH--x;WBWn)j0gSCLxC}?AU7{TB?Sz(Fnn~qVJ}kZFbh=ii3+rvbn2= zVq8!5iSllvm>`;g!d_G@P9gCN1f@uGg-`iE!i!I}kH22`3Dl1R97=E=HA0uiy_1{0 z^3=h=Rh^PuWC=o^BcP@<^x}50ua!Bq;&+pxYLIH32Kgy;@H7aNK=KOWr%6fvhjt_j zd-Xjs`>0Xn$^C>KAcCq6-6ax92cFqS(U1GWQ?b^c}Kz~5H} z`{+S`ZWbp9F6fLXlAajQjiOuM7HfDV%~w+}H$b6o`T^o5>#=@V30plv95{YcG@Fa^ z+HSTQ8ajh$8%~!0d=OtS@cPK$!Wd_-TVjX~RNsNH3nyD%aUwb!Nw?0MfdHdE#mAbC zK@tjR$n$?fVXhL?Mv>y>j(d1!Gm4>@Jk&m>wTT$i0;t%7R`S->;N)pfi=yCc3%iqa#j3n z%ZAHWxOtYHD3lt*chC0Mhf=0*YoGrOemn7K{h7Rz1lz(#YWRak8n}T!RB!wK{KaB} z+x!h0J%?`GTK??Wv#Rdb+?DwZTKQjkN<7Z<*b!y(!j^uc0D&nNt12)IpvQBKKyh&Q zT4`ikf6e}5P58#IGsQ?zq`rv#y+=WI9>HzUsVvx_H8(5IZBt$8@{dd-$a$M|k$lfX zVZ(V1VW~fUUs-0Y&h<-*-bZ&Mj?_EN{h=B)%!8>BCt7}5yafuY`VxWsgX1ro-pN2q z4gc)vk7r(zGRFI;#z+7A|87TT3=_%tao^a7k1Pz{)jDtdGI!sj@NuqTmX-lQ##>2x zaGYgl9!otok!jR+r}Ab7N=Pk^faGPQ?3%t-stju?>JfgaHmS>3Az+Nx)=!_3ay{d= zlJ%~X$C^reZPX#9{h{T^eJZ5=isELMcGv5Tga_&*VY>8JksF2*3K3%m9dGj41viES z8Xv=82T1)EEM(~Ic0^`>^Y#BiU|{WpbZW+~#;x!b+wzsGOLfH+`vSfZ=3wb$)1Jsp zL;W8`yfohPefcT;%)?sp#6DW&`?KrK9Qk#q4W3G8HXtg7Gj}u0Hl5@bJIqBdo(pe$ zTkHU>4E?WDaP(E@j!R+(sVE8p(s%OOG84jk1_#V^a-!kh;)pq29b%aZPmgdmR+*)aRVPoV%#ozX z9WtH&E+7B!n~<0>McCx(QvW-=f!9s<@WTN4H9_avlZ#b9vS{q?#z}bQ8_gJ1BF6Y> zQz=GGnWYbqFazJHN53p6`X+HbqIJ*%mH()Ze@#K5`iW5!#@a1tICmr6N!ga!H*=+U zOIY2&K4Alwa^D551^S%^1B>qBso^oY=Ba2DZ# z8j}oX0`Y+-Y#=`7%tMYQgJ(b>=j2)bWVM#(Vn1ED>Vi51ZI*ruu<7?`;&1;==%9b*MiV+j2B~E{3Twbmw6>Up4i4AajZH`*Rx# z4&R-juF!5sSbeUGYl#zXSAt~-V&TuWu1)^&t2bgKLWJtHOOK#XkRXny7FgsJsibV=qi z8}9cE;)WXAX({Yz&md>Xy6JmijIFiwg}2HVGX?zM!*yb zCt||){=`9R4Ye)$v7v!tFIbe<#%z8lbSJ>i)r7~HE^&K$Z{;yiUgwOhw8*}0Hbgkj z4k5v{qVN&_gX?SqoIsy2HOq2$gE5_Xm%+{>Plcv6k91jyi3;R53=xJ><~Uk4JBzwu3&t;M}L03;|j1Vfj8-Xvz zRsI;X5~eYtm9fKQq1q;3mG@Mpz4({zrBoju9};rRKYGS~DQIn)iWXT~YhX$%ZPkoi zt%P=^j*}jug(=ZMulWTF!PHHMB7`~HPPVWk&`3%-2RtroFp^^i? z`z%D)th_*Zl~kHXMMp2(@1m^z*=rrU_d`E5qH%F?(T?R>wuhW>VY0iFS3C1kIYZzx z7-k+=8rg97?p?1Fe6lTNHaDcN%EQ3Oh+5SB<;RVS%k{TtXvdSb) zhL&0{>}y>*UnwUi=QUUv#7dRDrHYF{2vDSr9~(w*Lb+5wco62OPJpzXFRMr>#l`mv z^75WU1qTNwQ<&6PfB#WRF@+-7gFO@F7^&SATAi%N$btJ09yl=j#Gx8f-RFf4@2Cpt zBR+x^eLtQ)b~)J!w(D~Bq9uwJ+!R#J`;Fq}#`VdW7PkCOR0O)JrKQDG?^@YfT$b{j zPJ9hJ$qx29Q680)BwjDONo8^2y@SjHS3phQ)IuLztIiOEqtfhoGAk9o9Eb@jHQa@! zFk^WNTEAiJ%BNlx9+8|Ym9jl{S<-6Xl+oVSmNfcO8DpBnv@z6dX=@Xc0KW=Qa_C*w z8d@KCPEb8niK8Q$XL(6#?VDD}nTaK>>klPw3|vNdMF0IPRTMR{e{QgREia1YfjzG= z@&?toz#EjLslGW}GC$!vqadj4SA;Hiwd=&U6V0}eGy5nM3eP;TSG*!pP#2znBaI2_ zWyS}g5?#P+b`2IY9%;;by!5`rPHPi%uIJ~M5`X|aIg`c?!S2BpO0qR$8XFr)IjG8R zH+&~EM-^x}rl!I!c4V&?)vt8nBdP}K5=GAdvqKL)-{UoI%l4K1`OIZge^P$FNI=eH zaoIK16-#%ZIp)_dC1&vpCs9T->1hgWfFGr*E*DiF1wVBQqU}OkX<+wNE^i?%zDU?w zmVbM=F@xt^lDJCSc}1l{z-9LtF_Yh0Yve>xNJeAz&|JgM>AgS{?b-IBuJ73@kRImq zVae0WvdMcGFS#ka$s;Gd*AF+Bp0bJ!hdW*A5m&poC&Q0v_=#MnL^<6{h&LrWq=wn_9NQB#@1H+ zhRT(}i6;b$^mqaHq3dLdwFbSPat)F{L~oSP;ssT3?PJl~DRO3f8Fe(zi&#*QkU(Jy zWzk~K()sM@nkgRSO!jLZvf-b?8b@D^>Q+N*j9k=5^;{+}zxhY&S*R@%gA_2|tHkD}GTISW4G3F%cnSTT(4Vpm$ja z%*FzWF8R%g_P*VGsUu^jsbD-T$&H~OgW=g4lFn47!gsm?HDd|chd&^EOx1}G4dv3v znTC&hSJzLEk1bm0>(399Gp~bP-IkO#<5rCeW+XBSrqjP~)CXvyM=kMJM1sN7Rzeaw ze0gyuhZN;Kw0W=5Wx)$VE9s3-EV^7D0mJ>bb&%4%i|M^J(Nl3DrSlG|qMkOFUc}^)31(*tj>e+Jv0h3nb%2@bQNn%gP!hKyzarvqOR8p4m6S z0c%;Vk|0>I)3xkiO6?(*)%~aO@6tz|v?>z<0|Sp>7u=*4Q;Lx7oZDvLB<1!^$0{l+ zB!2qaR*lnF-H?=`wpg3mDp~8E4)S0y)FJub0Y)tHP_whM)>|KwGeu9B@$KL43Cr`7 z8vvx#;1hHien4N@Wb1q3e%K|hR7drH)u{nw88RXQ{Qb=zbn-oFqRqJXF0Fp$lxQxN zWC7-tcYepRo==&rhQ9Px#HHTn=_3R3)yl*`A2%2lI z&*-(*AfR$dB;LO2{ZLSG*|218ZmtyA#gfW8Az3SN;tFMdA3&%xOtRwxx^nv*bCuhI z{K$K`Zz?A%$(fRyoJu?%uZo7LR4TXyDzv3U%mb4{k6bdS!Sa$?`uqF6P`&O}v4;ip z!()%MpuCQ|52sA~x|9QmxfM%O_rfQ{Zj9LHhk4#EU8nuXGn>=1EnDK4XL1=!4F(q+ zyOECqUfoew&kZp?s$kbP?GHYA0vRp@<;~tfk*q=9c~eb3!T@~M^cN81;C8uZae~l0 z`&MGUP0jiK;LPK&+$paD&L~O&C8VR1cBLBnn&2twLPK<2F6SvqvX{|5l49Lo(OPAe z%9^0*Gj4H*+iP-FFwzV%!&Rcw?Byls44STj6CR6ZCNjKu4-?V+HdJzE!|N*K;E~uG zx>fj$k$?XC`xROYGCR?3m#~XFW|2LT8h{)fcPsLjdo-T>WWnwjo*h30outmF+P9VB zR^j}LxI5Nx-@_FXe1fTk`y*SFQ9qm8y*%Ap)$G?7(==NbetkrM&GG1)-3yNWBR^XQ zg03F+cxs#vFIS^5w&4$IS_@HSCVqqx+dh9wxLRG9rbZl&*=pU7qV8f0S~LeMs0n?s zORZ|zz*8h1&X@uc=t6Qj`O9q^Sbfzm#_GbFmeZNpv>FpyX`|RYwrP3Zta~CW%h7xq;<2q>G8@|m{0W5ha)R(=gLwIZh9BspE=vcE3^UPi-ZyJB6o%8mHWd;y@P zed<1_=hSOEw{EMYwGedaQD2FS9I99-sv08z|0tpy;2fuq6lT1YQSe31z|YPnapN ziuojf{3}^00O!21PxLPv8LW&^K1i??F$W%z-AO=G1;k@jW04d^6m|NMUn$Bda*Qol z4dE`HmO{CVs4G#ZVV0tkXm%pWFbW$vCOHqwprbjQ3)5tmeHRC_@qW`*6@(|t8F27; zo5g;HEUdyyO0d~#JQ6T zcksyOZWE45>_CbL6!Z^_yT(bmJm57;6Yp6 jhb<9ELGSrG diff --git a/docs/reference/cmdstanr-package.html b/docs/reference/cmdstanr-package.html index 5d51c60f8..1208c80d2 100644 --- a/docs/reference/cmdstanr-package.html +++ b/docs/reference/cmdstanr-package.html @@ -1,77 +1,14 @@ - - - - - - - -CmdStanR: the R interface to CmdStan — cmdstanr-package • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStanR: the R interface to CmdStan — cmdstanr-package • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

CmdStanR: the R interface to CmdStan

- Source: R/cmdstanr-package.R + Source: R/cmdstanr-package.R
-

https://mc-stan.org/about/logo/ - Stan Development Team

+

Stan Development Team

CmdStanR: the R interface to CmdStan.

- -

Details

- +
+

Details

CmdStanR (cmdstanr package) is an interface to Stan -(mc-stan.org) for R users. It provides the +(mc-stan.org) for R users. It provides the necessary objects and functions to compile a Stan program and run Stan's algorithms from R via CmdStan, the shell interface to Stan -(mc-stan.org/users/interfaces/cmdstan).

-

Different ways of interfacing with Stan’s C++

+(mc-stan.org/users/interfaces/cmdstan).

+
+

Different ways of interfacing with Stan’s C++

-

The RStan interface (rstan package) is +

The RStan interface (rstan package) is an in-memory interface to Stan and relies on R packages like Rcpp and inline to call C++ code from R. On the other hand, the CmdStanR interface does not directly call any C++ code from R, instead relying on the CmdStan interface behind the scenes for compilation, running algorithms, and writing results to output files.

+
-

Advantages of RStan

+
+

Advantages of RStan

-
    -

  • Advanced features. We are working on making these available outside -of RStan but currently they are only available to R users via RStan:

      -

    • rstan::log_prob()

      • -

    • rstan::grad_log_prob()

      • -

    • rstan::expose_stan_functions()

      • + - - -

      Advantages of CmdStanR

      - -
        -

      • Compatible with latest versions of Stan. Keeping up with Stan -releases is complicated for RStan, often requiring non-trivial -changes to the rstan package and new CRAN releases of both -rstan and StanHeaders. With CmdStanR the latest improvements -in Stan will be available from R immediately after updating CmdStan -using cmdstanr::install_cmdstan().

        • -

      • Fewer installation issues (e.g., no need to mess with Makevars -files).

        • +
+ +
+

Advantages of CmdStanR

+ +

  • Compatible with latest versions of Stan. Keeping up with Stan releases +is complicated for RStan, often requiring non-trivial changes to the +rstan package and new CRAN releases of both rstan and +StanHeaders. With CmdStanR the latest improvements in Stan will be +available from R immediately after updating CmdStan using +cmdstanr::install_cmdstan().

    • +

  • Fewer installation issues (e.g., no need to mess with Makevars files).

  • Running Stan via external processes results in fewer unexpected crashes, especially in RStudio.

  • Less memory overhead.

  • More permissive license. RStan uses the GPL-3 license while the license for CmdStanR is BSD-3, which is a bit more permissive and is the same license used for CmdStan and the Stan C++ source code.

    • -
- - -

Getting started

+
+
+
+

Getting started

CmdStanR requires a working version of CmdStan. If -you already have CmdStan installed see cmdstan_model() to get started, -otherwise see install_cmdstan() to install CmdStan. The vignette -Getting started with CmdStanR +you already have CmdStan installed see cmdstan_model() to get started, +otherwise see install_cmdstan() to install CmdStan. The vignette +Getting started with CmdStanR demonstrates the basic functionality of the package.

-

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

-
- -

Examples

- 
# \dontrun{
-library(cmdstanr)
-library(posterior)
-library(bayesplot)
-color_scheme_set("brightblue")
-
-# Set path to CmdStan
-# (Note: if you installed CmdStan via install_cmdstan() with default settings
-# then setting the path is unnecessary but the default below should still work.
-# Otherwise use the `path` argument to specify the location of your
-# CmdStan installation.)
-set_cmdstan_path(path = NULL)
-
#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.29.1

-# Create a CmdStanModel object from a Stan program,
-# here using the example model that comes with CmdStan
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-mod <- cmdstan_model(file)
-mod$print()
-
#> data {
-#>   int<lower=0> N;
-#>   array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
-#>   y ~ bernoulli(theta);
-#> }

-# Data as a named list (like RStan)
-stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-
-# Run MCMC using the 'sample' method
-fit_mcmc <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  parallel_chains = 2
-)
-
#> Running MCMC with 2 parallel chains...
-#> 
-#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.2 seconds.
-#> 

-# Use 'posterior' package for summaries
-fit_mcmc$summary()
-
#> # A tibble: 2 × 10
-#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
-#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.

-# Get posterior draws
-draws <- fit_mcmc$draws()
-print(draws)
-
#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 -6.8 -6.8
-#>         2 -6.9 -6.8
-#>         3 -7.0 -7.0
-#>         4 -6.9 -7.1
-#>         5 -6.7 -7.0
-#> 
-#> , , variable = theta
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 0.28 0.21
-#>         2 0.19 0.20
-#>         3 0.16 0.17
-#>         4 0.20 0.36
-#>         5 0.25 0.34
-#> 
-#> # ... with 995 more iterations

-# Convert to data frame using posterior::as_draws_df
-as_draws_df(draws)
-
#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
-#>    lp__ theta
-#> 1  -6.8  0.28
-#> 2  -6.9  0.19
-#> 3  -7.0  0.16
-#> 4  -6.9  0.20
-#> 5  -6.7  0.25
-#> 6  -7.1  0.36
-#> 7  -9.0  0.55
-#> 8  -7.2  0.15
-#> 9  -6.8  0.23
-#> 10 -7.5  0.42
-#> # ... with 1990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# Plot posterior using bayesplot (ggplot2)
-mcmc_hist(fit_mcmc$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-# Call CmdStan's diagnose and stansummary utilities
-fit_mcmc$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181224-1-8f6e7b.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181224-2-8f6e7b.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit_mcmc$cmdstan_summary()
-
#> Inference for Stan model: bernoulli_model
-#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
-#> 
-#> Warmup took (0.0050, 0.0040) seconds, 0.0090 seconds total
-#> Sampling took (0.015, 0.019) seconds, 0.034 seconds total
-#> 
-#>                 Mean     MCSE  StdDev     5%   50%   95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8      781    22972      1.0
-#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0  2.3e+02  6.9e+03  1.0e+00
-#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0  1.0e+00  3.0e+01  2.0e+13
-#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0  1.9e+03  5.5e+04  1.0e+00
-#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0  8.9e+01  2.6e+03  1.0e+00
-#> divergent__     0.00      nan    0.00   0.00  0.00  0.00      nan      nan      nan
-#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6  7.7e+02  2.3e+04  1.0e+00
-#> 
-#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47      796    23422      1.0
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).

-# For models fit using MCMC, if you like working with RStan's stanfit objects
-# then you can create one with rstan::read_stan_csv()
-
-# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
-
-
-# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
-# and also demonstrate specifying data as a path to a file instead of a list
-my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
-fit_optim <- mod$optimize(data = my_data_file, seed = 123)
-
#> Initial log joint probability = -9.51104 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.

-fit_optim$summary()
-
#> # A tibble: 2 × 2
-#>   variable estimate
-#>   <chr>       <dbl>
-#> 1 lp__        -5.00
-#> 2 theta        0.2 

-
-# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
-fit_vb <- mod$variational(data = stan_data, seed = 123)
-
#> ------------------------------------------------------------ 
-#> EXPERIMENTAL ALGORITHM: 
-#>   This procedure has not been thoroughly tested and may be unstable 
-#>   or buggy. The interface is subject to change. 
-#> ------------------------------------------------------------ 
-#> Gradient evaluation took 5e-06 seconds 
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.05 seconds. 
-#> Adjust your expectations accordingly! 
-#> Begin eta adaptation. 
-#> Iteration:   1 / 250 [  0%]  (Adaptation) 
-#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
-#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
-#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
-#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
-#> Success! Found best value [eta = 1] earlier than expected. 
-#> Begin stochastic gradient ascent. 
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
-#>    100           -6.262             1.000            1.000 
-#>    200           -6.263             0.500            1.000 
-#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
-#> Drawing a sample of size 1000 from the approximate posterior...  
-#> COMPLETED. 
-#> Finished in  0.1 seconds.

-fit_vb$summary()
-
#> # A tibble: 3 × 7
-#>   variable      mean median    sd   mad     q5      q95
-#>   <chr>        <dbl>  <dbl> <dbl> <dbl>  <dbl>    <dbl>
-#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
-#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
-#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  

-# Plot approximate posterior using bayesplot
-mcmc_hist(fit_vb$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-
-# Specifying initial values as a function
-fit_mcmc_w_init_fun <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function() list(theta = runif(1))
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2 <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function(chain_id) {
-    # silly but demonstrates optional use of chain_id
-    list(theta = 1 / (chain_id + 1))
-  }
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.5
-#> 
-#> 
-#> [[2]]
-#> [[2]]$theta
-#> [1] 0.3333333
-#> 
-#> 

-# Specifying initial values as a list of lists
-fit_mcmc_w_init_list <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = list(
-    list(theta = 0.75), # chain 1
-    list(theta = 0.25)  # chain 2
-  )
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_optim_w_init_list <- mod$optimize(
-  data = stan_data,
-  seed = 123,
-  init = list(
-    list(theta = 0.75)
-  )
-)
-
#> Initial log joint probability = -11.6657 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.
fit_optim_w_init_list$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.75
-#> 
-#> 
# }
+
The Stan and CmdStan documentation:

+
-
+
+

Examples

+ 
# \dontrun{
+library(cmdstanr)
+library(posterior)
+library(bayesplot)
+color_scheme_set("brightblue")
+
+# Set path to CmdStan
+# (Note: if you installed CmdStan via install_cmdstan() with default settings
+# then setting the path is unnecessary but the default below should still work.
+# Otherwise use the `path` argument to specify the location of your
+# CmdStan installation.)
+set_cmdstan_path(path = NULL)
+#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.32.2
+
+# Create a CmdStanModel object from a Stan program,
+# here using the example model that comes with CmdStan
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+mod <- cmdstan_model(file)
+mod$print()
+#> data {
+#>   int<lower=0> N;
+#>   array[N] int<lower=0,upper=1> y;
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
+#>   y ~ bernoulli(theta);
+#> }
+
+# Data as a named list (like RStan)
+stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
+# Run MCMC using the 'sample' method
+fit_mcmc <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  parallel_chains = 2
+)
+#> Running MCMC with 2 parallel chains...
+#> 
+#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.2 seconds.
+#> 
+
+# Use 'posterior' package for summaries
+fit_mcmc$summary()
+#> # A tibble: 2 × 10
+#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
+#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.
+
+# Get posterior draws
+draws <- fit_mcmc$draws()
+print(draws)
+#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 -6.8 -6.8
+#>         2 -6.9 -6.8
+#>         3 -7.0 -7.0
+#>         4 -6.9 -7.1
+#>         5 -6.7 -7.0
+#> 
+#> , , variable = theta
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 0.28 0.21
+#>         2 0.19 0.20
+#>         3 0.16 0.17
+#>         4 0.20 0.36
+#>         5 0.25 0.34
+#> 
+#> # ... with 995 more iterations
+
+# Convert to data frame using posterior::as_draws_df
+as_draws_df(draws)
+#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
+#>    lp__ theta
+#> 1  -6.8  0.28
+#> 2  -6.9  0.19
+#> 3  -7.0  0.16
+#> 4  -6.9  0.20
+#> 5  -6.7  0.25
+#> 6  -7.1  0.36
+#> 7  -9.0  0.55
+#> 8  -7.2  0.15
+#> 9  -6.8  0.23
+#> 10 -7.5  0.42
+#> # ... with 1990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# Plot posterior using bayesplot (ggplot2)
+mcmc_hist(fit_mcmc$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+# Call CmdStan's diagnose and stansummary utilities
+fit_mcmc$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251435-1-2078cc.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251435-2-2078cc.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit_mcmc$cmdstan_summary()
+#> Inference for Stan model: bernoulli_model
+#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
+#> 
+#> Warmup took (0.0040, 0.0030) seconds, 0.0070 seconds total
+#> Sampling took (0.011, 0.011) seconds, 0.022 seconds total
+#> 
+#>                 Mean     MCSE  StdDev     5%   50%   95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8    781    35502      1.0
+#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0    235    10662  1.0e+00
+#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0    1.0       46  2.0e+13
+#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0   1874    85179  1.0e+00
+#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0     89     4050  1.0e+00
+#> divergent__     0.00      nan    0.00   0.00  0.00  0.00    nan      nan      nan
+#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6    775    35215  1.0e+00
+#> 
+#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47    796    36197      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+
+# For models fit using MCMC, if you like working with RStan's stanfit objects
+# then you can create one with rstan::read_stan_csv()
+
+# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
+
+# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
+# and also demonstrate specifying data as a path to a file instead of a list
+my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
+fit_optim <- mod$optimize(data = my_data_file, seed = 123)
+#> Initial log joint probability = -9.51104 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+
+fit_optim$summary()
+#> # A tibble: 2 × 2
+#>   variable estimate
+#>   <chr>       <num>
+#> 1 lp__        -5.00
+#> 2 theta        0.2 
+
+
+# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
+fit_vb <- mod$variational(data = stan_data, seed = 123)
+#> ------------------------------------------------------------ 
+#> EXPERIMENTAL ALGORITHM: 
+#>   This procedure has not been thoroughly tested and may be unstable 
+#>   or buggy. The interface is subject to change. 
+#> ------------------------------------------------------------ 
+#> Gradient evaluation took 6e-06 seconds 
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.06 seconds. 
+#> Adjust your expectations accordingly! 
+#> Begin eta adaptation. 
+#> Iteration:   1 / 250 [  0%]  (Adaptation) 
+#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
+#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
+#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
+#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
+#> Success! Found best value [eta = 1] earlier than expected. 
+#> Begin stochastic gradient ascent. 
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
+#>    100           -6.262             1.000            1.000 
+#>    200           -6.263             0.500            1.000 
+#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
+#> Drawing a sample of size 1000 from the approximate posterior...  
+#> COMPLETED. 
+#> Finished in  0.1 seconds.
+
+fit_vb$summary()
+#> # A tibble: 3 × 7
+#>   variable      mean median    sd   mad     q5      q95
+#>   <chr>        <num>  <num> <num> <num>  <num>    <num>
+#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
+#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
+#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  
+
+# Plot approximate posterior using bayesplot
+mcmc_hist(fit_vb$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+
+# Specifying initial values as a function
+fit_mcmc_w_init_fun <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function() list(theta = runif(1))
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2 <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function(chain_id) {
+    # silly but demonstrates optional use of chain_id
+    list(theta = 1 / (chain_id + 1))
+  }
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.5
+#> 
+#> 
+#> [[2]]
+#> [[2]]$theta
+#> [1] 0.3333333
+#> 
+#> 
+
+# Specifying initial values as a list of lists
+fit_mcmc_w_init_list <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = list(
+    list(theta = 0.75), # chain 1
+    list(theta = 0.25)  # chain 2
+  )
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.4 seconds.
+#> 
+fit_optim_w_init_list <- mod$optimize(
+  data = stan_data,
+  seed = 123,
+  init = list(
+    list(theta = 0.75)
+  )
+)
+#> Initial log joint probability = -11.6657 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+fit_optim_w_init_list$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.75
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/cmdstanr_example.html b/docs/reference/cmdstanr_example.html index f337bb94a..996f52635 100644 --- a/docs/reference/cmdstanr_example.html +++ b/docs/reference/cmdstanr_example.html @@ -1,75 +1,12 @@ - - - - - - - -Fit models for use in examples — cmdstanr_example • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Fit models for use in examples — cmdstanr_example • cmdstanr - - + + - - -
-
- -
- -
+

Fit models for use in examples

- Source: R/example.R + Source: R/example.R
@@ -179,249 +107,260 @@

Fit models for use in examples

Fit models for use in examples

- 
cmdstanr_example(
-  example = c("logistic", "schools", "schools_ncp"),
-  method = c("sample", "optimize", "variational", "diagnose"),
-  ...,
-  quiet = TRUE
-)
-
-print_example_program(example = c("logistic", "schools", "schools_ncp"))
+
+ 
cmdstanr_example(
+  example = c("logistic", "schools", "schools_ncp"),
+  method = c("sample", "optimize", "variational", "diagnose"),
+  ...,
+  quiet = TRUE
+)
+
+print_example_program(example = c("logistic", "schools", "schools_ncp"))
+
-

Arguments

- - - - - - - - - - - - - - - - - - -
example

(string) The name of the example. The currently available -examples are

    -

  • "logistic": logistic regression with intercept and 3 predictors.

    • +
    +

    Arguments

    +
    example
    +

    (string) The name of the example. The currently available +examples are

    • "logistic": logistic regression with intercept and 3 predictors.

    • "schools": the so-called "eight schools" model, a hierarchical meta-analysis. Fitting this model will result in warnings about divergences.

    • "schools_ncp": non-centered parameterization of the "eight schools" model that fixes the problem with divergences.

      • -
    +

To print the Stan code for a given example use +print_example_program(example).

+ + +
method
+

(string) Which fitting method should be used? The default is +the "sample" method (MCMC).

-

To print the Stan code for a given example use -print_example_program(example).

method

(string) Which fitting method should be used? The default is -the "sample" method (MCMC).

...

Arguments passed to the chosen method. See the help pages for -the individual methods for details.

quiet

(logical) If TRUE (the default) then fitting the model is -wrapped in utils::capture.output().

-

Value

+
...
+

Arguments passed to the chosen method. See the help pages for +the individual methods for details.

-

The fitted model object returned by the selected method.

-

Examples

- 
# \dontrun{
-print_example_program("logistic")
-
#> data {
-#>   int<lower=0> N;
-#>   int<lower=0> K;
-#>   array[N] int<lower=0, upper=1> y;
-#>   matrix[N, K] X;
-#> }
-#> parameters {
-#>   real alpha;
-#>   vector[K] beta;
-#> }
-#> model {
-#>   target += normal_lpdf(alpha | 0, 1);
-#>   target += normal_lpdf(beta | 0, 1);
-#>   target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);
-#> }
-#> generated quantities {
-#>   vector[N] log_lik;
-#>   for (n in 1 : N) {
-#>     log_lik[n] = bernoulli_logit_lpmf(y[n] | alpha + X[n] * beta);
-#>   }
-#> }
fit_logistic_mcmc <- cmdstanr_example("logistic", chains = 2)
-fit_logistic_mcmc$summary()
-
#> # A tibble: 105 × 10
-#>    variable      mean  median     sd    mad       q5      q95  rhat ess_bulk
-#>    <chr>        <dbl>   <dbl>  <dbl>  <dbl>    <dbl>    <dbl> <dbl>    <dbl>
-#>  1 lp__       -66.0   -65.7   1.45   1.27   -68.8    -64.3    1.00     1146.
-#>  2 alpha        0.384   0.382 0.217  0.207    0.0306   0.749  1.00     2115.
-#>  3 beta[1]     -0.662  -0.658 0.253  0.249   -1.08    -0.256  0.999    2154.
-#>  4 beta[2]     -0.273  -0.272 0.225  0.221   -0.651    0.0966 1.00     1882.
-#>  5 beta[3]      0.679   0.676 0.272  0.269    0.241    1.14   1.00     2000.
-#>  6 log_lik[1]  -0.514  -0.509 0.0959 0.0951  -0.682   -0.369  1.00     2063.
-#>  7 log_lik[2]  -0.407  -0.386 0.151  0.150   -0.695   -0.193  1.00     2242.
-#>  8 log_lik[3]  -0.494  -0.464 0.212  0.194   -0.888   -0.209  1.00     2079.
-#>  9 log_lik[4]  -0.453  -0.434 0.154  0.152   -0.732   -0.236  1.00     1963.
-#> 10 log_lik[5]  -1.19   -1.17  0.276  0.276   -1.66    -0.768  1.00     2212.
-#> # … with 95 more rows, and 1 more variable: ess_tail <dbl>

-fit_logistic_optim <- cmdstanr_example("logistic", method = "optimize")
-fit_logistic_optim$summary()
-
#> # A tibble: 105 × 2
-#>    variable   estimate
-#>    <chr>         <dbl>
-#>  1 lp__        -63.9  
-#>  2 alpha         0.365
-#>  3 beta[1]      -0.632
-#>  4 beta[2]      -0.259
-#>  5 beta[3]       0.648
-#>  6 log_lik[1]   -0.515
-#>  7 log_lik[2]   -0.394
-#>  8 log_lik[3]   -0.469
-#>  9 log_lik[4]   -0.442
-#> 10 log_lik[5]   -1.14 
-#> # … with 95 more rows

-fit_logistic_vb <- cmdstanr_example("logistic", method = "variational")
-fit_logistic_vb$summary()
-
#> # A tibble: 106 × 7
-#>    variable       mean  median    sd   mad       q5      q95
-#>    <chr>         <dbl>   <dbl> <dbl> <dbl>    <dbl>    <dbl>
-#>  1 lp__        -65.9   -65.6   1.42  1.24  -68.7    -64.2   
-#>  2 lp_approx__  -1.99   -1.68  1.40  1.24   -4.76    -0.361 
-#>  3 alpha         0.402   0.398 0.234 0.243   0.0390   0.785 
-#>  4 beta[1]      -0.598  -0.605 0.233 0.249  -0.971   -0.230 
-#>  5 beta[2]      -0.252  -0.256 0.199 0.192  -0.585    0.0849
-#>  6 beta[3]       0.779   0.773 0.211 0.215   0.450    1.13  
-#>  7 log_lik[1]   -0.520  -0.514 0.104 0.101  -0.711   -0.364 
-#>  8 log_lik[2]   -0.370  -0.354 0.122 0.115  -0.604   -0.202 
-#>  9 log_lik[3]   -0.470  -0.439 0.198 0.184  -0.841   -0.206 
-#> 10 log_lik[4]   -0.440  -0.430 0.132 0.123  -0.699   -0.251 
-#> # … with 96 more rows

-print_example_program("schools")
-
#> data {
-#>   int<lower=1> J;
-#>   vector<lower=0>[J] sigma;
-#>   vector[J] y;
-#> }
-#> parameters {
-#>   real mu;
-#>   real<lower=0> tau;
-#>   vector[J] theta;
-#> }
-#> model {
-#>   target += normal_lpdf(tau | 0, 10);
-#>   target += normal_lpdf(mu | 0, 10);
-#>   target += normal_lpdf(theta | mu, tau);
-#>   target += normal_lpdf(y | theta, sigma);
-#> }
fit_schools_mcmc <- cmdstanr_example("schools")
-
#> Warning: 497 of 4000 (12.0%) transitions ended with a divergence.
-#> See https://mc-stan.org/misc/warnings for details.
#> Warning: 1 of 4 chains had an E-BFMI less than 0.2.
-#> See https://mc-stan.org/misc/warnings for details.
fit_schools_mcmc$summary()
-
#> # A tibble: 11 × 10
-#>    variable   mean median    sd   mad        q5   q95  rhat ess_bulk ess_tail
-#>    <chr>     <dbl>  <dbl> <dbl> <dbl>     <dbl> <dbl> <dbl>    <dbl>    <dbl>
-#>  1 lp__     -55.9  -57.1   7.17  7.48 -66.2     -43.1  1.18     15.5    NA   
-#>  2 mu         6.44   6.31  4.01  3.68  -0.00706  13.7  1.15    592.    632.  
-#>  3 tau        4.63   3.66  3.77  3.63   0.760    11.9  1.18     15.5     8.46
-#>  4 theta[1]   8.81   7.10  6.76  4.99  -0.300    21.0  1.06    399.   1279.  
-#>  5 theta[2]   6.86   6.89  5.35  4.29  -1.81     16.0  1.15    969.   1298.  
-#>  6 theta[3]   5.67   6.13  6.26  4.75  -4.95     16.1  1.15    901.   1189.  
-#>  7 theta[4]   6.38   5.90  5.49  4.55  -2.34     16.0  1.15    750.   1636.  
-#>  8 theta[5]   4.78   5.75  5.41  4.27  -4.90     13.5  1.12    673.    863.  
-#>  9 theta[6]   5.42   5.79  5.57  4.58  -4.12     14.6  1.15    765.    878.  
-#> 10 theta[7]   8.53   7.16  5.81  4.96   0.719    19.3  1.05    244.   1167.  
-#> 11 theta[8]   6.87   5.78  6.41  4.86  -3.12     18.3  1.15    999.   1605.  

-print_example_program("schools_ncp")
-
#> data {
-#>   int<lower=1> J;
-#>   vector<lower=0>[J] sigma;
-#>   vector[J] y;
-#> }
-#> parameters {
-#>   real mu;
-#>   real<lower=0> tau;
-#>   vector[J] theta_raw;
-#> }
-#> transformed parameters {
-#>   vector[J] theta = mu + tau * theta_raw;
-#> }
-#> model {
-#>   target += normal_lpdf(tau | 0, 10);
-#>   target += normal_lpdf(mu | 0, 10);
-#>   target += normal_lpdf(theta_raw | 0, 1);
-#>   target += normal_lpdf(y | theta, sigma);
-#> }
fit_schools_ncp_mcmc <- cmdstanr_example("schools_ncp")
-
#> Warning: 1 of 4000 (0.0%) transitions ended with a divergence.
-#> See https://mc-stan.org/misc/warnings for details.
fit_schools_ncp_mcmc$summary()
-
#> # A tibble: 19 × 10
-#>    variable     mean   median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>    <chr>       <dbl>    <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#>  1 lp__     -46.9    -46.6    2.45  2.39  -51.3   -43.4  1.00     1434.    2266.
-#>  2 mu         6.39     6.33   4.36  4.25   -0.764  13.5  1.00     2726.    2142.
-#>  3 tau        4.77     3.95   3.75  3.45    0.374  11.7  1.00     1988.    1457.
-#>  4 theta_r…   0.376    0.379  0.949 0.970  -1.17    1.93 1.00     4065.    2603.
-#>  5 theta_r…   0.0728   0.0765 0.897 0.895  -1.42    1.55 1.00     3969.    2465.
-#>  6 theta_r…  -0.135   -0.129  0.965 0.981  -1.70    1.45 1.00     4040.    2528.
-#>  7 theta_r…   0.0355   0.0394 0.913 0.905  -1.46    1.54 1.00     4126.    2685.
-#>  8 theta_r…  -0.272   -0.269  0.917 0.888  -1.76    1.25 1.00     4273.    3098.
-#>  9 theta_r…  -0.167   -0.168  0.922 0.910  -1.69    1.32 1.00     3993.    2775.
-#> 10 theta_r…   0.386    0.410  0.946 0.910  -1.22    1.90 1.00     4227.    2942.
-#> 11 theta_r…   0.0653   0.0824 0.938 0.952  -1.48    1.58 1.00     4858.    2868.
-#> 12 theta[1]   8.83     8.22   6.62  5.73   -0.909  20.7  1.00     3668.    3136.
-#> 13 theta[2]   6.81     6.70   5.51  5.09   -2.08   15.6  1.00     4198.    3392.
-#> 14 theta[3]   5.49     5.74   6.46  5.69   -5.57   15.5  1.00     3698.    2838.
-#> 15 theta[4]   6.68     6.59   5.85  5.35   -2.69   16.1  1.00     4391.    3338.
-#> 16 theta[5]   4.76     5.04   5.74  5.33   -5.40   13.5  1.00     4192.    3211.
-#> 17 theta[6]   5.40     5.60   5.92  5.32   -4.58   14.4  1.00     4953.    3365.
-#> 18 theta[7]   8.85     8.39   6.04  5.70   -0.292  19.4  0.999    3832.    3162.
-#> 19 theta[8]   6.92     6.72   6.45  5.79   -3.11   17.5  1.00     4113.    3205.

-# optimization fails for hierarchical model
-cmdstanr_example("schools", "optimize", quiet = FALSE)
-
#> Initial log joint probability = -373.952 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>       99       114.115      0.330131   2.59089e+09      0.3881           1      169    
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>      179       242.979      0.047195   2.42406e+16       1e-12       0.001      373  LS failed, Hessian reset  
#> Optimization terminated with error: 
#>   Line search failed to achieve a sufficient decrease, no more progress can be made
#> Finished in  0.1 seconds.
#>  variable estimate
-#>  lp__       242.98
-#>  mu           1.05
-#>  tau          0.00
-#>  theta[1]     1.05
-#>  theta[2]     1.05
-#>  theta[3]     1.05
-#>  theta[4]     1.05
-#>  theta[5]     1.05
-#>  theta[6]     1.05
-#>  theta[7]     1.05
-#> 
-#>  # showing 10 of 11 rows (change via 'max_rows' argument or 'cmdstanr_max_rows' option)
# }
+
quiet

+
(logical) If TRUE (the default) then fitting the model is
+wrapped in utils::capture.output().

 
-
+
+
+

Value

+ + +

The fitted model object returned by the selected method.

+
+ +
+

Examples

+ 
# \dontrun{
+print_example_program("logistic")
+#> data {
+#>   int<lower=0> N;
+#>   int<lower=0> K;
+#>   array[N] int<lower=0, upper=1> y;
+#>   matrix[N, K] X;
+#> }
+#> parameters {
+#>   real alpha;
+#>   vector[K] beta;
+#> }
+#> model {
+#>   target += normal_lpdf(alpha | 0, 1);
+#>   target += normal_lpdf(beta | 0, 1);
+#>   target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);
+#> }
+#> generated quantities {
+#>   vector[N] log_lik;
+#>   for (n in 1 : N) {
+#>     log_lik[n] = bernoulli_logit_lpmf(y[n] | alpha + X[n] * beta);
+#>   }
+#> }
+fit_logistic_mcmc <- cmdstanr_example("logistic", chains = 2)
+fit_logistic_mcmc$summary()
+#> # A tibble: 105 × 10
+#>    variable      mean  median    sd    mad       q5      q95  rhat ess_bulk
+#>    <chr>        <num>   <num> <num>  <num>    <num>    <num> <num>    <num>
+#>  1 lp__       -65.9   -65.6   1.40  1.25   -68.6    -64.3    1.00     1066.
+#>  2 alpha        0.381   0.385 0.216 0.223    0.0256   0.732  1.00     1952.
+#>  3 beta[1]     -0.667  -0.657 0.250 0.249   -1.09    -0.266  1.00     1902.
+#>  4 beta[2]     -0.275  -0.271 0.223 0.222   -0.655    0.0809 1.00     2026.
+#>  5 beta[3]      0.691   0.680 0.272 0.261    0.257    1.14   1.00     2036.
+#>  6 log_lik[1]  -0.516  -0.508 0.100 0.0986  -0.696   -0.364  1.00     2121.
+#>  7 log_lik[2]  -0.399  -0.376 0.146 0.142   -0.665   -0.195  1.00     2027.
+#>  8 log_lik[3]  -0.496  -0.463 0.218 0.204   -0.895   -0.205  0.999    2110.
+#>  9 log_lik[4]  -0.448  -0.433 0.154 0.149   -0.731   -0.227  1.00     1978.
+#> 10 log_lik[5]  -1.19   -1.17  0.276 0.277   -1.65    -0.776  1.00     2176.
+#> # ℹ 95 more rows
+#> # ℹ 1 more variable: ess_tail <num>
+
+fit_logistic_optim <- cmdstanr_example("logistic", method = "optimize")
+fit_logistic_optim$summary()
+#> # A tibble: 105 × 2
+#>    variable   estimate
+#>    <chr>         <num>
+#>  1 lp__        -63.9  
+#>  2 alpha         0.364
+#>  3 beta[1]      -0.632
+#>  4 beta[2]      -0.259
+#>  5 beta[3]       0.649
+#>  6 log_lik[1]   -0.515
+#>  7 log_lik[2]   -0.394
+#>  8 log_lik[3]   -0.469
+#>  9 log_lik[4]   -0.442
+#> 10 log_lik[5]   -1.14 
+#> # ℹ 95 more rows
+
+fit_logistic_vb <- cmdstanr_example("logistic", method = "variational")
+fit_logistic_vb$summary()
+#> # A tibble: 106 × 7
+#>    variable       mean  median     sd    mad       q5     q95
+#>    <chr>         <num>   <num>  <num>  <num>    <num>   <num>
+#>  1 lp__        -66.7   -66.3   1.91   1.73   -70.5    -64.4  
+#>  2 lp_approx__  -2.03   -1.75  1.37   1.23    -4.66    -0.332
+#>  3 alpha         0.519   0.518 0.223  0.224    0.153    0.895
+#>  4 beta[1]      -0.690  -0.687 0.234  0.244   -1.06    -0.298
+#>  5 beta[2]      -0.296  -0.297 0.262  0.251   -0.720    0.144
+#>  6 beta[3]       0.546   0.548 0.309  0.310    0.0501   1.06 
+#>  7 log_lik[1]   -0.454  -0.448 0.0931 0.0926  -0.622   -0.311
+#>  8 log_lik[2]   -0.530  -0.498 0.215  0.215   -0.951   -0.244
+#>  9 log_lik[3]   -0.488  -0.446 0.238  0.228   -0.906   -0.178
+#> 10 log_lik[4]   -0.535  -0.516 0.191  0.180   -0.875   -0.267
+#> # ℹ 96 more rows
+
+print_example_program("schools")
+#> data {
+#>   int<lower=1> J;
+#>   vector<lower=0>[J] sigma;
+#>   vector[J] y;
+#> }
+#> parameters {
+#>   real mu;
+#>   real<lower=0> tau;
+#>   vector[J] theta;
+#> }
+#> model {
+#>   target += normal_lpdf(tau | 0, 10);
+#>   target += normal_lpdf(mu | 0, 10);
+#>   target += normal_lpdf(theta | mu, tau);
+#>   target += normal_lpdf(y | theta, sigma);
+#> }
+fit_schools_mcmc <- cmdstanr_example("schools")
+#> Warning: 391 of 4000 (10.0%) transitions ended with a divergence.
+#> See https://mc-stan.org/misc/warnings for details.
+fit_schools_mcmc$summary()
+#> # A tibble: 11 × 10
+#>    variable   mean median    sd   mad       q5   q95  rhat ess_bulk ess_tail
+#>    <chr>     <num>  <num> <num> <num>    <num> <num> <num>    <num>    <num>
+#>  1 lp__     -55.6  -56.0   6.40  7.14 -65.7    -45.9  1.14     19.4   333.  
+#>  2 mu         6.95   7.68  4.20  3.72  -0.0865  14.5  1.07    246.     40.2 
+#>  3 tau        4.24   3.19  3.55  3.11   0.592   11.2  1.18     15.7     9.70
+#>  4 theta[1]   9.05   7.63  6.17  5.01  -0.247   20.1  1.11    575.   1129.  
+#>  5 theta[2]   7.29   7.96  5.17  4.40  -1.33    15.6  1.06    383.    660.  
+#>  6 theta[3]   6.04   7.16  6.37  5.11  -4.96    15.8  1.05    345.    591.  
+#>  7 theta[4]   7.16   7.85  5.55  4.64  -2.09    16.3  1.04    287.    330.  
+#>  8 theta[5]   5.44   6.31  5.68  4.65  -4.51    14.3  1.02    185.     57.9 
+#>  9 theta[6]   6.24   7.03  5.68  4.70  -3.84    15.5  1.02    162.     83.5 
+#> 10 theta[7]   8.94   7.77  5.69  4.69   0.261   18.8  1.15    489.   1084.  
+#> 11 theta[8]   7.45   8.18  6.27  5.20  -2.53    16.9  1.06    527.   1455.  
+
+print_example_program("schools_ncp")
+#> data {
+#>   int<lower=1> J;
+#>   vector<lower=0>[J] sigma;
+#>   vector[J] y;
+#> }
+#> parameters {
+#>   real mu;
+#>   real<lower=0> tau;
+#>   vector[J] theta_raw;
+#> }
+#> transformed parameters {
+#>   vector[J] theta = mu + tau * theta_raw;
+#> }
+#> model {
+#>   target += normal_lpdf(tau | 0, 10);
+#>   target += normal_lpdf(mu | 0, 10);
+#>   target += normal_lpdf(theta_raw | 0, 1);
+#>   target += normal_lpdf(y | theta, sigma);
+#> }
+fit_schools_ncp_mcmc <- cmdstanr_example("schools_ncp")
+fit_schools_ncp_mcmc$summary()
+#> # A tibble: 19 × 10
+#>    variable         mean   median    sd   mad       q5    q95  rhat ess_bulk
+#>    <chr>           <num>    <num> <num> <num>    <num>  <num> <num>    <num>
+#>  1 lp__         -46.9    -46.7    2.45  2.37  -51.5    -43.4  1.00     1390.
+#>  2 mu             6.45     6.53   4.16  4.02   -0.486   13.3  1.00     3299.
+#>  3 tau            4.63     3.83   3.59  3.39    0.324   11.6  1.00     1997.
+#>  4 theta_raw[1]   0.343    0.373  0.968 0.963  -1.26     1.90 0.999    3956.
+#>  5 theta_raw[2]   0.0288   0.0326 0.919 0.900  -1.49     1.54 1.00     4784.
+#>  6 theta_raw[3]  -0.150   -0.150  0.930 0.871  -1.70     1.39 1.00     4594.
+#>  7 theta_raw[4]   0.0241   0.0138 0.929 0.917  -1.51     1.53 1.00     3987.
+#>  8 theta_raw[5]  -0.254   -0.266  0.923 0.936  -1.77     1.29 1.00     4286.
+#>  9 theta_raw[6]  -0.138   -0.151  0.936 0.911  -1.69     1.44 1.00     4732.
+#> 10 theta_raw[7]   0.354    0.386  0.935 0.917  -1.22     1.88 1.00     4376.
+#> 11 theta_raw[8]   0.0685   0.0563 0.960 0.955  -1.47     1.65 1.00     4469.
+#> 12 theta[1]       8.82     8.19   6.70  5.73   -0.854   21.4  1.00     3878.
+#> 13 theta[2]       6.71     6.68   5.47  4.95   -2.08    15.8  1.00     4371.
+#> 14 theta[3]       5.58     5.85   6.23  5.28   -4.90    15.2  1.00     4089.
+#> 15 theta[4]       6.54     6.57   5.71  5.16   -2.62    16.0  1.00     4307.
+#> 16 theta[5]       4.90     5.24   5.62  5.24   -5.14    13.5  1.00     3910.
+#> 17 theta[6]       5.65     5.84   5.75  5.24   -4.27    14.7  1.00     4570.
+#> 18 theta[7]       8.69     8.23   5.89  5.49    0.0330  19.2  1.00     4332.
+#> 19 theta[8]       7.00     6.80   6.39  5.51   -2.89    17.5  1.00     3804.
+#> # ℹ 1 more variable: ess_tail <num>
+
+# optimization fails for hierarchical model
+cmdstanr_example("schools", "optimize", quiet = FALSE)
+#> Initial log joint probability = -52.1838 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>       99       122.653      0.275645   9.62182e+09        0.14      0.3154      173    
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>      187       258.482      0.227211   1.52288e+17       1e-12       0.001      402  LS failed, Hessian reset  
+#> Optimization terminated with error: 
+#>   Line search failed to achieve a sufficient decrease, no more progress can be made
+#> Finished in  0.1 seconds.
+#>  variable estimate
+#>  lp__       258.48
+#>  mu           0.28
+#>  tau          0.00
+#>  theta[1]     0.28
+#>  theta[2]     0.28
+#>  theta[3]     0.28
+#>  theta[4]     0.28
+#>  theta[5]     0.28
+#>  theta[6]     0.28
+#>  theta[7]     0.28
+#> 
+#>  # showing 10 of 11 rows (change via 'max_rows' argument or 'cmdstanr_max_rows' option)
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/draws_to_csv.html b/docs/reference/draws_to_csv.html index e054de697..08c28b5ef 100644 --- a/docs/reference/draws_to_csv.html +++ b/docs/reference/draws_to_csv.html @@ -1,78 +1,15 @@ - - - - - - - -Write posterior draws objects to CSV files suitable for running standalone generated -quantities with CmdStan. — draws_to_csv • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Write posterior draws objects to CSV files suitable for running standalone generated +quantities with CmdStan. — draws_to_csv • cmdstanr - + + - - - -
-
- -
- -
+

Write posterior draws objects to CSV files suitable for running standalone generated quantities with CmdStan.

- Source: R/data.R + Source: R/data.R
@@ -184,42 +112,44 @@

Write posterior draws objects to CSV files suitable for running standalone g quantities with CmdStan.

- 
draws_to_csv(
-  draws,
-  sampler_diagnostics = NULL,
-  dir = tempdir(),
-  basename = "fittedParams"
-)
+
+ 
draws_to_csv(
+  draws,
+  sampler_diagnostics = NULL,
+  dir = tempdir(),
+  basename = "fittedParams"
+)
+
+ +
+

Arguments

+
draws
+

A posterior::draws_* object.

+ + +
sampler_diagnostics
+

Either NULL or a posterior::draws_* object +of sampler diagnostics.

-

Arguments

- - - - - - - - - - - - - - - - - - -
draws

A posterior::draws_* object.

sampler_diagnostics

Either NULL or a posterior::draws_* object -of sampler diagnostics.

dir

(string) An optional path to the directory where the CSV files will be -written. If not set, temporary directory is used.

basename

(string) If dir is specified, `basename`` is used for naming -the output CSV files. If not specified, the file names are randomly generated.

-

Value

+
dir
+

(string) An optional path to the directory where the CSV files will be +written. If not set, temporary directory is used.

-

Paths to CSV files (one per chain).

-

Details

+
basename
+

(string) If dir is specified, `basename`` is used for naming +the output CSV files. If not specified, the file names are randomly generated.

+ +
+
+

Value

+ + +

Paths to CSV files (one per chain).

+
+
+

Details

draws_to_csv() generates a CSV suitable for running standalone generated quantities with CmdStan. The CSV file contains a single comment #num_samples, which equals the number of iterations in the supplied draws object.

@@ -230,49 +160,49 @@

Details standalone generated quantities method of CmdStan.

The column names line is finally followed by the values of the draws in the same order as the column names.

+

-

Examples

- 
# \dontrun{
-draws <- posterior::example_draws()
-
-draws_csv_files <- draws_to_csv(draws)
-print(draws_csv_files)
-
#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpmzUYEz/fittedParams-202203181225-1-96521b.csv"
-#> [2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpmzUYEz/fittedParams-202203181225-2-96521b.csv"
-#> [3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpmzUYEz/fittedParams-202203181225-3-96521b.csv"
-#> [4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpmzUYEz/fittedParams-202203181225-4-96521b.csv"

-# draws_csv_files <- draws_to_csv(draws,
-#                                 sampler_diagnostic = sampler_diagnostics,
-#                                 dir = "~/my_folder",
-#                                 basename = "my-samples")
-# }
-
-
+
+

Examples

+ 
# \dontrun{
+draws <- posterior::example_draws()
+
+draws_csv_files <- draws_to_csv(draws)
+print(draws_csv_files)
+#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpFBtN6X/fittedParams-202307251435-1-8744f9.csv"
+#> [2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpFBtN6X/fittedParams-202307251435-2-8744f9.csv"
+#> [3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpFBtN6X/fittedParams-202307251435-3-8744f9.csv"
+#> [4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpFBtN6X/fittedParams-202307251435-4-8744f9.csv"
+
+# draws_csv_files <- draws_to_csv(draws,
+#                                 sampler_diagnostic = sampler_diagnostics,
+#                                 dir = "~/my_folder",
+#                                 basename = "my-samples")
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/eng_cmdstan.html b/docs/reference/eng_cmdstan.html index 78b56c953..f39ce5cec 100644 --- a/docs/reference/eng_cmdstan.html +++ b/docs/reference/eng_cmdstan.html @@ -1,80 +1,17 @@ - - - - - - - -CmdStan knitr engine for Stan — eng_cmdstan • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CmdStan knitr engine for Stan — eng_cmdstan • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

CmdStan knitr engine for Stan

- Source: R/knitr.R + Source: R/knitr.R

This provides a knitr engine for Stan, suitable for usage when attempting to render Stan chunks and compile the model code within to an executable with -CmdStan. Use register_knitr_engine() to make this the default engine for +CmdStan. Use register_knitr_engine() to make this the default engine for stan chunks. See the vignette -R Markdown CmdStan Engine +R Markdown CmdStan Engine for an example.

- 
eng_cmdstan(options)
+
+ 
eng_cmdstan(options)
+
-

Arguments

- - - - - - -
options

(named list) Chunk options, as provided by knitr during -chunk execution.

+
+

Arguments

+
options
+

(named list) Chunk options, as provided by knitr during +chunk execution.

+
-

Examples

- 
# \dontrun{
-knitr::knit_engines$set(stan = cmdstanr::eng_cmdstan)
-# }
-
+
+

Examples

+ 
# \dontrun{
+knitr::knit_engines$set(stan = cmdstanr::eng_cmdstan)
+# }
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-cmdstan_summary.html b/docs/reference/fit-method-cmdstan_summary.html index d550c7936..7aaf67169 100644 --- a/docs/reference/fit-method-cmdstan_summary.html +++ b/docs/reference/fit-method-cmdstan_summary.html @@ -1,53 +1,5 @@ - - - - - - - -Run CmdStan's stansummary and diagnose utilities — fit-method-cmdstan_summary • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run CmdStan's stansummary and diagnose utilities — fit-method-cmdstan_summary • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Run CmdStan's stansummary and diagnose utilities

- Source: R/fit.R + Source: R/fit.R

Run CmdStan's stansummary and diagnose utilities. These are -documented in the CmdStan Guide:

    -

  • https://mc-stan.org/docs/cmdstan-guide/stansummary.html

    • +documented in the CmdStan Guide:

    • https://mc-stan.org/docs/cmdstan-guide/stansummary.html

    • https://mc-stan.org/docs/cmdstan-guide/diagnose.html

      • -
    - -

    Although these methods can be used for models fit using the -$variational() method, much of the output is +

Although these methods can be used for models fit using the +$variational() method, much of the output is currently only relevant for models fit using the -$sample() method.

-

See the $summary() for computing similar summaries in +$sample() method.

+

See the $summary() for computing similar summaries in R rather than calling CmdStan's utilites.

- 
cmdstan_summary(flags = NULL)
-
-cmdstan_diagnose()
- -

Arguments

- - - - - - -
flags

An optional character vector of flags (e.g. -flags = c("--sig_figs=1")).

- -

See also

+
+ 
cmdstan_summary(flags = NULL)
+
+cmdstan_diagnose()
+
-

CmdStanMCMC, fit-method-summary

+
+

Arguments

+
flags
+

An optional character vector of flags (e.g. +flags = c("--sig_figs=1")).

-

Examples

- 
# \dontrun{
-fit <- cmdstanr_example("logistic")
-fit$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-1-61aacb.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-2-61aacb.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-3-61aacb.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-4-61aacb.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit$cmdstan_summary()
-
#> Inference for Stan model: logistic_model
-#> 4 chains: each with iter=(1000,1000,1000,1000); warmup=(0,0,0,0); thin=(1,1,1,1); 4000 iterations saved.
-#> 
-#> Warmup took (0.034, 0.026, 0.028, 0.048) seconds, 0.14 seconds total
-#> Sampling took (0.095, 0.084, 0.16, 0.12) seconds, 0.46 seconds total
-#> 
-#>                     Mean     MCSE  StdDev     5%       50%       95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -6.6e+01  3.6e-02     1.5    -69  -6.6e+01  -6.4e+01     1736     3791      1.0
-#> accept_stat__       0.90  3.8e-03    0.11   0.67      0.94       1.0  8.5e+02  1.9e+03  1.0e+00
-#> stepsize__          0.75  3.0e-02   0.043   0.68      0.78      0.79  2.0e+00  4.4e+00  1.6e+13
-#> treedepth__          2.4  6.7e-02    0.53    2.0       2.0       3.0  6.2e+01  1.4e+02  1.0e+00
-#> n_leapfrog__         5.2  2.6e-01     2.0    3.0       7.0       7.0  6.0e+01  1.3e+02  1.0e+00
-#> divergent__         0.00      nan    0.00   0.00      0.00      0.00      nan      nan      nan
-#> energy__              68  5.5e-02     2.1     65        68        72  1.4e+03  3.1e+03  1.0e+00
-#> 
-#> alpha            3.8e-01  3.3e-03    0.22  0.026   3.8e-01   7.5e-01     4388     9581     1.00
-#> beta[1]         -6.7e-01  3.9e-03    0.25   -1.1  -6.6e-01  -2.6e-01     4345     9487     1.00
-#> beta[2]         -2.7e-01  3.6e-03    0.23  -0.66  -2.7e-01   1.1e-01     4144     9048     1.00
-#> beta[3]          6.8e-01  4.2e-03    0.27   0.26   6.7e-01   1.1e+00     3965     8656      1.0
-#> log_lik[1]      -5.1e-01  1.5e-03    0.10  -0.69  -5.1e-01  -3.6e-01     4462     9742     1.00
-#> log_lik[2]      -4.0e-01  2.2e-03    0.15  -0.68  -3.9e-01  -1.9e-01     4523     9876     1.00
-#> log_lik[3]      -4.9e-01  3.3e-03    0.22  -0.91  -4.6e-01  -2.0e-01     4388     9582     1.00
-#> log_lik[4]      -4.5e-01  2.4e-03    0.15  -0.72  -4.3e-01  -2.4e-01     3964     8655     1.00
-#> log_lik[5]      -1.2e+00  4.2e-03    0.28   -1.7  -1.2e+00  -7.6e-01     4456     9730     1.00
-#> log_lik[6]      -5.9e-01  2.9e-03    0.19  -0.94  -5.7e-01  -3.2e-01     4286     9359      1.0
-#> log_lik[7]      -6.4e-01  1.8e-03    0.12  -0.85  -6.4e-01  -4.6e-01     4658    10170     1.00
-#> log_lik[8]      -2.8e-01  2.3e-03    0.14  -0.54  -2.5e-01  -1.0e-01     3596     7851     1.00
-#> log_lik[9]      -6.9e-01  2.5e-03    0.17  -0.99  -6.8e-01  -4.5e-01     4349     9495      1.0
-#> log_lik[10]     -7.3e-01  3.7e-03    0.24   -1.2  -7.1e-01  -3.9e-01     4199     9168     1.00
-#> log_lik[11]     -2.8e-01  2.0e-03    0.12  -0.51  -2.6e-01  -1.1e-01     3817     8334      1.0
-#> log_lik[12]     -5.0e-01  3.6e-03    0.24  -0.96  -4.6e-01  -1.8e-01     4375     9553      1.0
-#> log_lik[13]     -6.6e-01  3.0e-03    0.20   -1.0  -6.4e-01  -3.6e-01     4601    10046      1.0
-#> log_lik[14]     -3.6e-01  2.7e-03    0.18  -0.69  -3.3e-01  -1.3e-01     4218     9210     1.00
-#> log_lik[15]     -2.8e-01  1.6e-03    0.11  -0.47  -2.6e-01  -1.3e-01     4392     9589     1.00
-#> log_lik[16]     -2.8e-01  1.4e-03   0.088  -0.43  -2.7e-01  -1.5e-01     3712     8105     1.00
-#> log_lik[17]     -1.6e+00  5.0e-03    0.30   -2.1  -1.6e+00  -1.1e+00     3489     7618     1.00
-#> log_lik[18]     -4.8e-01  1.7e-03    0.11  -0.67  -4.7e-01  -3.2e-01     3930     8582     1.00
-#> log_lik[19]     -2.3e-01  1.2e-03   0.077  -0.37  -2.2e-01  -1.2e-01     4012     8761     1.00
-#> log_lik[20]     -1.1e-01  1.2e-03   0.079  -0.27  -9.4e-02  -2.7e-02     4074     8896     1.00
-#> log_lik[21]     -2.1e-01  1.4e-03   0.088  -0.37  -2.0e-01  -9.2e-02     3715     8110      1.0
-#> log_lik[22]     -5.7e-01  2.3e-03    0.15  -0.83  -5.5e-01  -3.5e-01     4001     8736      1.0
-#> log_lik[23]     -3.3e-01  2.3e-03    0.14  -0.59  -3.1e-01  -1.4e-01     3755     8198     1.00
-#> log_lik[24]     -1.4e-01  1.1e-03   0.068  -0.26  -1.2e-01  -5.2e-02     3965     8657     1.00
-#> log_lik[25]     -4.6e-01  1.8e-03    0.12  -0.67  -4.5e-01  -2.8e-01     4284     9354     1.00
-#> log_lik[26]     -1.5e+00  5.6e-03    0.35   -2.1  -1.5e+00  -9.9e-01     4011     8759     1.00
-#> log_lik[27]     -3.1e-01  2.0e-03    0.12  -0.53  -2.9e-01  -1.4e-01     3766     8224      1.0
-#> log_lik[28]     -4.4e-01  1.3e-03   0.085  -0.59  -4.4e-01  -3.1e-01     4028     8795     1.00
-#> log_lik[29]     -7.3e-01  3.4e-03    0.23   -1.1  -7.1e-01  -4.0e-01     4634    10119     1.00
-#> log_lik[30]     -7.0e-01  3.1e-03    0.20   -1.0  -6.8e-01  -4.1e-01     3998     8729     1.00
-#> log_lik[31]     -4.9e-01  2.6e-03    0.16  -0.78  -4.7e-01  -2.6e-01     4003     8741      1.0
-#> log_lik[32]     -4.2e-01  1.6e-03    0.11  -0.61  -4.1e-01  -2.6e-01     4299     9385      1.0
-#> log_lik[33]     -4.1e-01  1.9e-03    0.13  -0.63  -4.0e-01  -2.3e-01     4377     9558     1.00
-#> log_lik[34]     -6.4e-02  8.5e-04   0.052  -0.16  -5.0e-02  -1.3e-02     3746     8179     1.00
-#> log_lik[35]     -5.9e-01  2.6e-03    0.18  -0.91  -5.7e-01  -3.3e-01     4964    10838     1.00
-#> log_lik[36]     -3.3e-01  2.1e-03    0.13  -0.58  -3.1e-01  -1.4e-01     4081     8911     1.00
-#> log_lik[37]     -6.9e-01  3.3e-03    0.23   -1.1  -6.7e-01  -3.7e-01     4588    10017     1.00
-#> log_lik[38]     -3.2e-01  2.3e-03    0.15  -0.59  -2.9e-01  -1.2e-01     4002     8738     1.00
-#> log_lik[39]     -1.8e-01  1.7e-03    0.11  -0.39  -1.6e-01  -4.9e-02     4239     9255     1.00
-#> log_lik[40]     -6.8e-01  1.9e-03    0.13  -0.90  -6.7e-01  -4.9e-01     4339     9474     1.00
-#> log_lik[41]     -1.1e+00  4.2e-03    0.25   -1.6  -1.1e+00  -7.6e-01     3742     8171      1.0
-#> log_lik[42]     -9.3e-01  3.0e-03    0.20   -1.3  -9.1e-01  -6.4e-01     4374     9549     1.00
-#> log_lik[43]     -4.1e-01  3.8e-03    0.27  -0.92  -3.5e-01  -1.0e-01     4940    10786     1.00
-#> log_lik[44]     -1.2e+00  3.0e-03    0.19   -1.5  -1.2e+00  -8.9e-01     4045     8831     1.00
-#> log_lik[45]     -3.6e-01  1.9e-03    0.12  -0.58  -3.5e-01  -1.8e-01     4315     9421     1.00
-#> log_lik[46]     -5.8e-01  1.9e-03    0.13  -0.81  -5.8e-01  -3.9e-01     4621    10089     1.00
-#> log_lik[47]     -3.1e-01  2.0e-03    0.13  -0.54  -2.9e-01  -1.4e-01     4005     8745     1.00
-#> log_lik[48]     -3.2e-01  1.3e-03   0.085  -0.48  -3.2e-01  -2.0e-01     4261     9303     1.00
-#> log_lik[49]     -3.2e-01  1.3e-03   0.080  -0.46  -3.1e-01  -2.0e-01     3930     8581     1.00
-#> log_lik[50]     -1.3e+00  5.2e-03    0.34   -1.9  -1.3e+00  -7.9e-01     4358     9516     1.00
-#> log_lik[51]     -2.9e-01  1.4e-03   0.097  -0.47  -2.8e-01  -1.5e-01     4624    10096     1.00
-#> log_lik[52]     -8.4e-01  2.2e-03    0.15   -1.1  -8.3e-01  -6.1e-01     4377     9558     1.00
-#> log_lik[53]     -4.0e-01  2.1e-03    0.13  -0.63  -3.9e-01  -2.1e-01     3841     8387      1.0
-#> log_lik[54]     -3.7e-01  2.2e-03    0.15  -0.63  -3.5e-01  -1.7e-01     4317     9426     1.00
-#> log_lik[55]     -3.9e-01  2.0e-03    0.13  -0.63  -3.7e-01  -1.9e-01     4429     9671     1.00
-#> log_lik[56]     -3.2e-01  3.0e-03    0.19  -0.70  -2.8e-01  -9.1e-02     4073     8894     1.00
-#> log_lik[57]     -6.5e-01  1.7e-03    0.12  -0.86  -6.5e-01  -4.7e-01     4496     9816     1.00
-#> log_lik[58]     -9.4e-01  4.9e-03    0.35   -1.6  -9.0e-01  -4.5e-01     4920    10743     1.00
-#> log_lik[59]     -1.4e+00  5.3e-03    0.34   -2.0  -1.3e+00  -8.6e-01     4125     9007     1.00
-#> log_lik[60]     -9.8e-01  2.5e-03    0.16   -1.3  -9.7e-01  -7.3e-01     4281     9348     1.00
-#> log_lik[61]     -5.4e-01  1.5e-03    0.10  -0.72  -5.3e-01  -3.9e-01     4572     9982     1.00
-#> log_lik[62]     -8.8e-01  4.9e-03    0.32   -1.5  -8.5e-01  -4.2e-01     4308     9407      1.0
-#> log_lik[63]     -1.2e-01  1.2e-03   0.074  -0.26  -1.0e-01  -3.2e-02     3584     7826     1.00
-#> log_lik[64]     -9.0e-01  4.0e-03    0.25   -1.4  -8.7e-01  -5.4e-01     4099     8950     1.00
-#> log_lik[65]     -2.0e+00  9.8e-03    0.59   -3.1  -2.0e+00  -1.1e+00     3642     7952      1.0
-#> log_lik[66]     -5.0e-01  1.9e-03    0.13  -0.74  -4.9e-01  -3.1e-01     4752    10376     1.00
-#> log_lik[67]     -2.8e-01  1.3e-03   0.083  -0.43  -2.7e-01  -1.6e-01     4034     8808     1.00
-#> log_lik[68]     -1.1e+00  3.9e-03    0.24   -1.5  -1.0e+00  -7.1e-01     3761     8211     1.00
-#> log_lik[69]     -4.3e-01  1.3e-03   0.084  -0.58  -4.3e-01  -3.1e-01     4219     9213     1.00
-#> log_lik[70]     -6.4e-01  3.7e-03    0.24   -1.1  -6.1e-01  -3.0e-01     4278     9341     1.00
-#> log_lik[71]     -6.1e-01  2.9e-03    0.20  -0.98  -5.9e-01  -3.2e-01     4865    10623      1.0
-#> log_lik[72]     -4.6e-01  2.5e-03    0.17  -0.76  -4.4e-01  -2.2e-01     4453     9724     1.00
-#> log_lik[73]     -1.5e+00  5.7e-03    0.36   -2.1  -1.4e+00  -9.3e-01     3935     8591     1.00
-#> log_lik[74]     -9.5e-01  3.0e-03    0.20   -1.3  -9.4e-01  -6.6e-01     4389     9583     1.00
-#> log_lik[75]     -1.1e+00  6.1e-03    0.40   -1.9  -1.1e+00  -5.7e-01     4353     9505      1.0
-#> log_lik[76]     -3.7e-01  2.2e-03    0.14  -0.62  -3.5e-01  -1.7e-01     4299     9387     1.00
-#> log_lik[77]     -8.8e-01  2.1e-03    0.14   -1.1  -8.7e-01  -6.7e-01     4303     9396      1.0
-#> log_lik[78]     -4.9e-01  2.6e-03    0.18  -0.82  -4.6e-01  -2.4e-01     4716    10297     1.00
-#> log_lik[79]     -7.6e-01  2.9e-03    0.19   -1.1  -7.5e-01  -4.8e-01     4325     9444     1.00
-#> log_lik[80]     -5.4e-01  2.9e-03    0.20  -0.91  -5.2e-01  -2.6e-01     4653    10160     1.00
-#> log_lik[81]     -1.6e-01  1.6e-03   0.099  -0.35  -1.4e-01  -4.9e-02     3852     8410     1.00
-#> log_lik[82]     -2.2e-01  2.1e-03    0.14  -0.49  -1.9e-01  -6.1e-02     4220     9215     1.00
-#> log_lik[83]     -3.4e-01  1.3e-03   0.083  -0.49  -3.4e-01  -2.2e-01     4202     9174     1.00
-#> log_lik[84]     -2.7e-01  1.4e-03   0.092  -0.44  -2.6e-01  -1.4e-01     4250     9280     1.00
-#> log_lik[85]     -1.3e-01  1.2e-03   0.075  -0.27  -1.1e-01  -4.0e-02     4072     8891     1.00
-#> log_lik[86]     -1.1e+00  5.0e-03    0.32   -1.7  -1.1e+00  -6.5e-01     4140     9040     1.00
-#> log_lik[87]     -8.3e-01  2.0e-03    0.13   -1.0  -8.2e-01  -6.3e-01     4293     9373     1.00
-#> log_lik[88]     -7.8e-01  3.5e-03    0.24   -1.2  -7.5e-01  -4.3e-01     4503     9832      1.0
-#> log_lik[89]     -1.3e+00  4.9e-03    0.32   -1.9  -1.2e+00  -8.1e-01     4104     8960     1.00
-#> log_lik[90]     -2.6e-01  2.1e-03    0.13  -0.52  -2.4e-01  -9.4e-02     4118     8991     1.00
-#> log_lik[91]     -3.9e-01  1.9e-03    0.13  -0.61  -3.7e-01  -2.1e-01     4368     9537     1.00
-#> log_lik[92]     -1.5e+00  5.4e-03    0.34   -2.1  -1.5e+00  -9.8e-01     4061     8866     1.00
-#> log_lik[93]     -7.5e-01  3.6e-03    0.23   -1.2  -7.2e-01  -4.2e-01     4046     8835     1.00
-#> log_lik[94]     -3.2e-01  1.4e-03   0.089  -0.47  -3.1e-01  -1.9e-01     4234     9244     1.00
-#> log_lik[95]     -3.9e-01  1.8e-03    0.11  -0.58  -3.8e-01  -2.2e-01     3790     8276      1.0
-#> log_lik[96]     -1.6e+00  4.8e-03    0.29   -2.1  -1.6e+00  -1.1e+00     3777     8248     1.00
-#> log_lik[97]     -4.3e-01  1.6e-03    0.10  -0.62  -4.2e-01  -2.7e-01     4262     9306     1.00
-#> log_lik[98]     -1.0e+00  5.2e-03    0.36   -1.7  -1.0e+00  -5.2e-01     4933    10771     1.00
-#> log_lik[99]     -6.9e-01  2.1e-03    0.14  -0.94  -6.9e-01  -4.8e-01     4361     9523     1.00
-#> log_lik[100]    -3.9e-01  1.5e-03    0.10  -0.57  -3.8e-01  -2.4e-01     4557     9949     1.00
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).
# }
+
+ -
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("logistic")
+fit$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251435-1-7b6c9f.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251435-2-7b6c9f.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251435-3-7b6c9f.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251435-4-7b6c9f.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit$cmdstan_summary()
+#> Inference for Stan model: logistic_model
+#> 4 chains: each with iter=(1000,1000,1000,1000); warmup=(0,0,0,0); thin=(1,1,1,1); 4000 iterations saved.
+#> 
+#> Warmup took (0.022, 0.022, 0.022, 0.021) seconds, 0.087 seconds total
+#> Sampling took (0.069, 0.072, 0.068, 0.066) seconds, 0.28 seconds total
+#> 
+#>                     Mean     MCSE  StdDev     5%       50%       95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -6.6e+01  3.3e-02     1.4    -69  -6.6e+01  -6.4e+01   1875     6818      1.0
+#> accept_stat__       0.91  1.3e-03   0.099   0.71      0.95       1.0   5487    19952  1.0e+00
+#> stepsize__          0.75  2.8e-02   0.040   0.69      0.76      0.80    2.0      7.3  1.2e+13
+#> treedepth__          2.4  9.4e-03    0.53    2.0       2.0       3.0   3168    11520  1.0e+00
+#> n_leapfrog__         5.3  3.3e-02     2.0    3.0       7.0       7.0   3698    13446  1.0e+00
+#> divergent__         0.00      nan    0.00   0.00      0.00      0.00    nan      nan      nan
+#> energy__              68  5.2e-02     2.0     65        68        72   1501     5457  1.0e+00
+#> 
+#> alpha            3.8e-01  3.3e-03    0.22  0.031   3.8e-01   7.3e-01   4281    15567     1.00
+#> beta[1]         -6.6e-01  3.9e-03    0.25   -1.1  -6.6e-01  -2.5e-01   4206    15294     1.00
+#> beta[2]         -2.8e-01  3.7e-03    0.23  -0.65  -2.8e-01   1.0e-01   3812    13860      1.0
+#> beta[3]          6.8e-01  4.4e-03    0.27   0.25   6.8e-01   1.1e+00   3764    13688     1.00
+#> log_lik[1]      -5.2e-01  1.5e-03   0.097  -0.69  -5.1e-01  -3.7e-01   4137    15045      1.0
+#> log_lik[2]      -4.0e-01  2.4e-03    0.15  -0.69  -3.8e-01  -2.0e-01   4070    14798     1.00
+#> log_lik[3]      -5.0e-01  3.5e-03    0.22  -0.90  -4.7e-01  -2.1e-01   3974    14450      1.0
+#> log_lik[4]      -4.5e-01  2.4e-03    0.15  -0.72  -4.3e-01  -2.4e-01   3841    13967     1.00
+#> log_lik[5]      -1.2e+00  4.3e-03    0.28   -1.7  -1.2e+00  -7.6e-01   4262    15497     1.00
+#> log_lik[6]      -5.9e-01  2.8e-03    0.19  -0.94  -5.7e-01  -3.1e-01   4639    16869     1.00
+#> log_lik[7]      -6.4e-01  1.8e-03    0.13  -0.86  -6.3e-01  -4.5e-01   4605    16746     1.00
+#> log_lik[8]      -2.8e-01  2.3e-03    0.14  -0.54  -2.5e-01  -1.1e-01   3527    12826      1.0
+#> log_lik[9]      -6.9e-01  2.6e-03    0.17  -0.99  -6.8e-01  -4.3e-01   4522    16442     1.00
+#> log_lik[10]     -7.4e-01  3.7e-03    0.23   -1.2  -7.2e-01  -4.1e-01   3966    14423      1.0
+#> log_lik[11]     -2.8e-01  2.1e-03    0.13  -0.52  -2.6e-01  -1.1e-01   3682    13390      1.0
+#> log_lik[12]     -5.0e-01  3.9e-03    0.24  -0.93  -4.5e-01  -1.9e-01   3842    13973     1.00
+#> log_lik[13]     -6.6e-01  3.1e-03    0.21   -1.0  -6.4e-01  -3.6e-01   4585    16673     1.00
+#> log_lik[14]     -3.7e-01  2.8e-03    0.17  -0.70  -3.4e-01  -1.4e-01   3858    14029      1.0
+#> log_lik[15]     -2.8e-01  1.7e-03    0.11  -0.48  -2.7e-01  -1.4e-01   3977    14463      1.0
+#> log_lik[16]     -2.8e-01  1.5e-03   0.087  -0.44  -2.6e-01  -1.5e-01   3495    12708      1.0
+#> log_lik[17]     -1.6e+00  5.1e-03    0.29   -2.1  -1.6e+00  -1.1e+00   3184    11579      1.0
+#> log_lik[18]     -4.8e-01  1.7e-03    0.11  -0.68  -4.7e-01  -3.1e-01   4019    14615      1.0
+#> log_lik[19]     -2.3e-01  1.3e-03   0.075  -0.37  -2.2e-01  -1.3e-01   3432    12479      1.0
+#> log_lik[20]     -1.1e-01  1.3e-03   0.080  -0.27  -9.4e-02  -2.9e-02   3626    13186      1.0
+#> log_lik[21]     -2.1e-01  1.5e-03   0.088  -0.38  -2.0e-01  -9.4e-02   3506    12749      1.0
+#> log_lik[22]     -5.7e-01  2.3e-03    0.15  -0.84  -5.5e-01  -3.4e-01   4293    15611      1.0
+#> log_lik[23]     -3.3e-01  2.4e-03    0.14  -0.60  -3.0e-01  -1.4e-01   3628    13194      1.0
+#> log_lik[24]     -1.4e-01  1.1e-03   0.066  -0.26  -1.2e-01  -5.2e-02   3504    12742      1.0
+#> log_lik[25]     -4.5e-01  1.8e-03    0.12  -0.67  -4.4e-01  -2.8e-01   4234    15396     1.00
+#> log_lik[26]     -1.5e+00  5.5e-03    0.34   -2.1  -1.5e+00  -1.0e+00   3760    13672      1.0
+#> log_lik[27]     -3.1e-01  2.0e-03    0.12  -0.54  -2.9e-01  -1.4e-01   3688    13412      1.0
+#> log_lik[28]     -4.5e-01  1.3e-03   0.083  -0.59  -4.4e-01  -3.2e-01   4007    14572      1.0
+#> log_lik[29]     -7.2e-01  3.4e-03    0.22   -1.1  -7.0e-01  -4.0e-01   4367    15880     1.00
+#> log_lik[30]     -6.9e-01  3.1e-03    0.20   -1.0  -6.7e-01  -4.1e-01   4146    15077      1.0
+#> log_lik[31]     -4.9e-01  2.5e-03    0.17  -0.80  -4.6e-01  -2.5e-01   4260    15491      1.0
+#> log_lik[32]     -4.2e-01  1.6e-03    0.11  -0.61  -4.2e-01  -2.7e-01   4253    15465     1.00
+#> log_lik[33]     -4.1e-01  2.0e-03    0.13  -0.65  -4.0e-01  -2.3e-01   4207    15297     1.00
+#> log_lik[34]     -6.3e-02  8.6e-04   0.051  -0.16  -4.9e-02  -1.3e-02   3530    12835     1.00
+#> log_lik[35]     -5.9e-01  2.6e-03    0.18  -0.91  -5.7e-01  -3.2e-01   4846    17624     1.00
+#> log_lik[36]     -3.2e-01  2.2e-03    0.13  -0.57  -3.0e-01  -1.5e-01   3788    13775      1.0
+#> log_lik[37]     -7.0e-01  3.4e-03    0.22   -1.1  -6.8e-01  -3.8e-01   4249    15451      1.0
+#> log_lik[38]     -3.1e-01  2.4e-03    0.15  -0.60  -2.8e-01  -1.2e-01   3822    13898     1.00
+#> log_lik[39]     -1.8e-01  1.8e-03    0.11  -0.40  -1.6e-01  -5.3e-02   3734    13579      1.0
+#> log_lik[40]     -6.8e-01  2.0e-03    0.13  -0.91  -6.7e-01  -4.8e-01   4407    16026     1.00
+#> log_lik[41]     -1.1e+00  4.2e-03    0.26   -1.6  -1.1e+00  -7.4e-01   3931    14294      1.0
+#> log_lik[42]     -9.3e-01  3.1e-03    0.20   -1.3  -9.1e-01  -6.2e-01   4217    15335     1.00
+#> log_lik[43]     -4.1e-01  4.0e-03    0.27  -0.92  -3.5e-01  -1.0e-01   4495    16345     1.00
+#> log_lik[44]     -1.2e+00  2.9e-03    0.18   -1.5  -1.2e+00  -8.9e-01   3801    13820      1.0
+#> log_lik[45]     -3.6e-01  1.9e-03    0.12  -0.58  -3.5e-01  -1.9e-01   3946    14348      1.0
+#> log_lik[46]     -5.8e-01  2.0e-03    0.13  -0.81  -5.7e-01  -3.8e-01   4290    15599     1.00
+#> log_lik[47]     -3.0e-01  2.1e-03    0.13  -0.55  -2.8e-01  -1.3e-01   3827    13916     1.00
+#> log_lik[48]     -3.2e-01  1.3e-03   0.082  -0.47  -3.2e-01  -2.0e-01   3733    13576      1.0
+#> log_lik[49]     -3.2e-01  1.3e-03   0.079  -0.46  -3.1e-01  -2.0e-01   3527    12825      1.0
+#> log_lik[50]     -1.3e+00  5.1e-03    0.33   -1.9  -1.3e+00  -8.0e-01   4204    15288     1.00
+#> log_lik[51]     -2.9e-01  1.5e-03   0.094  -0.46  -2.8e-01  -1.6e-01   4007    14571     1.00
+#> log_lik[52]     -8.3e-01  2.2e-03    0.14   -1.1  -8.2e-01  -6.2e-01   4159    15122     1.00
+#> log_lik[53]     -4.0e-01  2.1e-03    0.13  -0.65  -3.8e-01  -2.1e-01   4104    14922      1.0
+#> log_lik[54]     -3.7e-01  2.2e-03    0.14  -0.63  -3.6e-01  -1.8e-01   4025    14635      1.0
+#> log_lik[55]     -3.9e-01  2.2e-03    0.14  -0.64  -3.7e-01  -2.0e-01   4074    14813     1.00
+#> log_lik[56]     -3.2e-01  3.2e-03    0.19  -0.70  -2.7e-01  -9.4e-02   3722    13536      1.0
+#> log_lik[57]     -6.6e-01  1.7e-03    0.12  -0.86  -6.5e-01  -4.8e-01   4340    15782     1.00
+#> log_lik[58]     -9.5e-01  5.2e-03    0.35   -1.6  -9.1e-01  -4.6e-01   4377    15915     1.00
+#> log_lik[59]     -1.4e+00  5.6e-03    0.35   -2.0  -1.3e+00  -8.4e-01   3916    14239     1.00
+#> log_lik[60]     -9.8e-01  2.4e-03    0.16   -1.2  -9.7e-01  -7.3e-01   4048    14720     1.00
+#> log_lik[61]     -5.4e-01  1.5e-03   0.097  -0.71  -5.3e-01  -3.9e-01   4267    15517     1.00
+#> log_lik[62]     -8.7e-01  4.9e-03    0.32   -1.5  -8.3e-01  -4.3e-01   4186    15223     1.00
+#> log_lik[63]     -1.2e-01  1.3e-03   0.074  -0.26  -9.7e-02  -3.2e-02   3420    12438      1.0
+#> log_lik[64]     -9.1e-01  4.0e-03    0.26   -1.4  -8.8e-01  -5.3e-01   4216    15333      1.0
+#> log_lik[65]     -2.0e+00  1.0e-02    0.60   -3.1  -2.0e+00  -1.1e+00   3486    12678     1.00
+#> log_lik[66]     -5.1e-01  2.0e-03    0.13  -0.75  -5.0e-01  -3.1e-01   4559    16578     1.00
+#> log_lik[67]     -2.8e-01  1.4e-03   0.081  -0.42  -2.7e-01  -1.6e-01   3503    12740      1.0
+#> log_lik[68]     -1.1e+00  3.9e-03    0.24   -1.5  -1.1e+00  -7.0e-01   3902    14189      1.0
+#> log_lik[69]     -4.4e-01  1.3e-03   0.082  -0.58  -4.3e-01  -3.1e-01   4099    14906      1.0
+#> log_lik[70]     -6.4e-01  3.8e-03    0.24   -1.1  -6.0e-01  -3.0e-01   4023    14629      1.0
+#> log_lik[71]     -6.1e-01  3.0e-03    0.21  -0.99  -5.9e-01  -3.2e-01   4800    17455     1.00
+#> log_lik[72]     -4.7e-01  2.7e-03    0.18  -0.80  -4.4e-01  -2.2e-01   4327    15734     1.00
+#> log_lik[73]     -1.5e+00  6.0e-03    0.37   -2.1  -1.5e+00  -9.4e-01   3679    13379     1.00
+#> log_lik[74]     -9.5e-01  3.0e-03    0.19   -1.3  -9.4e-01  -6.6e-01   4126    15004      1.0
+#> log_lik[75]     -1.1e+00  6.0e-03    0.39   -1.8  -1.1e+00  -5.8e-01   4255    15474     1.00
+#> log_lik[76]     -3.8e-01  2.2e-03    0.14  -0.63  -3.6e-01  -1.8e-01   3886    14131      1.0
+#> log_lik[77]     -8.8e-01  2.1e-03    0.14   -1.1  -8.7e-01  -6.6e-01   4420    16075     1.00
+#> log_lik[78]     -4.9e-01  2.6e-03    0.17  -0.80  -4.7e-01  -2.5e-01   4490    16326     1.00
+#> log_lik[79]     -7.7e-01  2.9e-03    0.19   -1.1  -7.5e-01  -4.9e-01   4123    14992     1.00
+#> log_lik[80]     -5.4e-01  3.0e-03    0.20  -0.92  -5.2e-01  -2.6e-01   4287    15591     1.00
+#> log_lik[81]     -1.6e-01  1.6e-03   0.099  -0.36  -1.4e-01  -4.9e-02   3918    14248     1.00
+#> log_lik[82]     -2.3e-01  2.3e-03    0.14  -0.51  -1.9e-01  -6.4e-02   3867    14061     1.00
+#> log_lik[83]     -3.4e-01  1.3e-03   0.080  -0.49  -3.4e-01  -2.3e-01   3809    13853      1.0
+#> log_lik[84]     -2.8e-01  1.5e-03   0.090  -0.45  -2.7e-01  -1.5e-01   3815    13872      1.0
+#> log_lik[85]     -1.3e-01  1.3e-03   0.075  -0.28  -1.2e-01  -4.4e-02   3631    13202      1.0
+#> log_lik[86]     -1.1e+00  5.2e-03    0.33   -1.7  -1.1e+00  -6.4e-01   4016    14603     1.00
+#> log_lik[87]     -8.2e-01  1.9e-03    0.13   -1.0  -8.2e-01  -6.3e-01   4341    15786     1.00
+#> log_lik[88]     -7.8e-01  3.7e-03    0.25   -1.2  -7.5e-01  -4.3e-01   4546    16530     1.00
+#> log_lik[89]     -1.3e+00  5.2e-03    0.32   -1.8  -1.3e+00  -8.0e-01   3869    14070     1.00
+#> log_lik[90]     -2.6e-01  2.1e-03    0.13  -0.51  -2.3e-01  -9.4e-02   4086    14857     1.00
+#> log_lik[91]     -3.9e-01  2.0e-03    0.13  -0.63  -3.8e-01  -2.0e-01   4332    15753     1.00
+#> log_lik[92]     -1.5e+00  5.5e-03    0.34   -2.1  -1.5e+00  -9.7e-01   3795    13801      1.0
+#> log_lik[93]     -7.3e-01  3.5e-03    0.22   -1.1  -7.1e-01  -4.2e-01   3875    14091     1.00
+#> log_lik[94]     -3.2e-01  1.4e-03   0.086  -0.48  -3.1e-01  -1.9e-01   3797    13806      1.0
+#> log_lik[95]     -3.9e-01  1.9e-03    0.11  -0.60  -3.7e-01  -2.2e-01   3692    13426      1.0
+#> log_lik[96]     -1.6e+00  4.9e-03    0.28   -2.1  -1.6e+00  -1.1e+00   3247    11808      1.0
+#> log_lik[97]     -4.3e-01  1.6e-03    0.10  -0.61  -4.2e-01  -2.8e-01   4122    14989      1.0
+#> log_lik[98]     -1.1e+00  5.5e-03    0.37   -1.7  -1.0e+00  -5.3e-01   4388    15957     1.00
+#> log_lik[99]     -7.0e-01  2.2e-03    0.15  -0.95  -6.9e-01  -4.8e-01   4345    15799     1.00
+#> log_lik[100]    -3.9e-01  1.5e-03   0.096  -0.56  -3.8e-01  -2.5e-01   4119    14977      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-code.html b/docs/reference/fit-method-code.html index 73e2c9e27..5dc9cdc20 100644 --- a/docs/reference/fit-method-code.html +++ b/docs/reference/fit-method-code.html @@ -1,75 +1,12 @@ - - - - - - - -Return Stan code — fit-method-code • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Return Stan code — fit-method-code • cmdstanr - - + + - - -
-
- -
- -
+

Return Stan code

- Source: R/fit.R + Source: R/fit.R
@@ -179,90 +107,95 @@

Return Stan code

Return Stan code

- 
code()
- - -

Value

- -

A character vector with one element per line of code.

-

See also

+
+ 
code()
+
-

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

+
+

Value

+ -

Examples

- 

-# \dontrun{
-fit <- cmdstanr_example()
-fit$code() # character vector
-
#>  [1] "data {"                                                            
-#>  [2] "  int<lower=0> N;"                                                 
-#>  [3] "  int<lower=0> K;"                                                 
-#>  [4] "  array[N] int<lower=0, upper=1> y;"                               
-#>  [5] "  matrix[N, K] X;"                                                 
-#>  [6] "}"                                                                 
-#>  [7] "parameters {"                                                      
-#>  [8] "  real alpha;"                                                     
-#>  [9] "  vector[K] beta;"                                                 
-#> [10] "}"                                                                 
-#> [11] "model {"                                                           
-#> [12] "  target += normal_lpdf(alpha | 0, 1);"                            
-#> [13] "  target += normal_lpdf(beta | 0, 1);"                             
-#> [14] "  target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);"         
-#> [15] "}"                                                                 
-#> [16] "generated quantities {"                                            
-#> [17] "  vector[N] log_lik;"                                              
-#> [18] "  for (n in 1 : N) {"                                              
-#> [19] "    log_lik[n] = bernoulli_logit_lpmf(y[n] | alpha + X[n] * beta);"
-#> [20] "  }"                                                               
-#> [21] "}"                                                                 
cat(fit$code(), sep = "\n") # pretty print
-
#> data {
-#>   int<lower=0> N;
-#>   int<lower=0> K;
-#>   array[N] int<lower=0, upper=1> y;
-#>   matrix[N, K] X;
-#> }
-#> parameters {
-#>   real alpha;
-#>   vector[K] beta;
-#> }
-#> model {
-#>   target += normal_lpdf(alpha | 0, 1);
-#>   target += normal_lpdf(beta | 0, 1);
-#>   target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);
-#> }
-#> generated quantities {
-#>   vector[N] log_lik;
-#>   for (n in 1 : N) {
-#>     log_lik[n] = bernoulli_logit_lpmf(y[n] | alpha + X[n] * beta);
-#>   }
-#> }
# }
+
A character vector with one element per line of code.

+    

+    
 
-
+
+

Examples

+ 

+# \dontrun{
+fit <- cmdstanr_example()
+fit$code() # character vector
+#>  [1] "data {"                                                            
+#>  [2] "  int<lower=0> N;"                                                 
+#>  [3] "  int<lower=0> K;"                                                 
+#>  [4] "  array[N] int<lower=0, upper=1> y;"                               
+#>  [5] "  matrix[N, K] X;"                                                 
+#>  [6] "}"                                                                 
+#>  [7] "parameters {"                                                      
+#>  [8] "  real alpha;"                                                     
+#>  [9] "  vector[K] beta;"                                                 
+#> [10] "}"                                                                 
+#> [11] "model {"                                                           
+#> [12] "  target += normal_lpdf(alpha | 0, 1);"                            
+#> [13] "  target += normal_lpdf(beta | 0, 1);"                             
+#> [14] "  target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);"         
+#> [15] "}"                                                                 
+#> [16] "generated quantities {"                                            
+#> [17] "  vector[N] log_lik;"                                              
+#> [18] "  for (n in 1 : N) {"                                              
+#> [19] "    log_lik[n] = bernoulli_logit_lpmf(y[n] | alpha + X[n] * beta);"
+#> [20] "  }"                                                               
+#> [21] "}"                                                                 
+cat(fit$code(), sep = "\n") # pretty print
+#> data {
+#>   int<lower=0> N;
+#>   int<lower=0> K;
+#>   array[N] int<lower=0, upper=1> y;
+#>   matrix[N, K] X;
+#> }
+#> parameters {
+#>   real alpha;
+#>   vector[K] beta;
+#> }
+#> model {
+#>   target += normal_lpdf(alpha | 0, 1);
+#>   target += normal_lpdf(beta | 0, 1);
+#>   target += bernoulli_logit_glm_lpmf(y | X, alpha, beta);
+#> }
+#> generated quantities {
+#>   vector[N] log_lik;
+#>   for (n in 1 : N) {
+#>     log_lik[n] = bernoulli_logit_lpmf(y[n] | alpha + X[n] * beta);
+#>   }
+#> }
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-constrain_variables.html b/docs/reference/fit-method-constrain_variables.html new file mode 100644 index 000000000..96e4e0456 --- /dev/null +++ b/docs/reference/fit-method-constrain_variables.html @@ -0,0 +1,179 @@ + +Transform a set of unconstrained parameter values to the constrained scale — fit-method-constrain_variables • cmdstanr + + +
+
+ + + +
+
+
+

Transform a set of unconstrained parameter values to the constrained scale

+ Source: R/fit.R + +
+ +
+

The $constrain_variables() method transforms input parameters to +the constrained scale

+
+ +
+ 
constrain_variables(
+  unconstrained_variables,
+  transformed_parameters = TRUE,
+  generated_quantities = TRUE
+)
+
+ +
+

Arguments

+
unconstrained_variables
+

(numeric) A vector of unconstrained parameters +to constrain.

+ + +
transformed_parameters
+

(boolean) Whether to return transformed +parameters implied by newly-constrained parameters (defaults to TRUE).

+ + +
generated_quantities
+

(boolean) Whether to return generated quantities +implied by newly-constrained parameters (defaults to TRUE).

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+fit_mcmc$constrain_variables(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2))
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+# }
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-diagnostic_summary.html b/docs/reference/fit-method-diagnostic_summary.html index 279cedd77..c7c0ce326 100644 --- a/docs/reference/fit-method-diagnostic_summary.html +++ b/docs/reference/fit-method-diagnostic_summary.html @@ -1,82 +1,19 @@ - - - - - - - -Sampler diagnostic summaries and warnings — fit-method-diagnostic_summary • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Sampler diagnostic summaries and warnings — fit-method-diagnostic_summary • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Sampler diagnostic summaries and warnings

- Source: R/fit.R + Source: R/fit.R

Warnings and summaries of sampler diagnostics. To instead get the underlying values of the sampler diagnostics for each iteration and -chain use the $sampler_diagnostics() +chain use the $sampler_diagnostics() method.

Currently parameter-specific diagnostics like R-hat and effective sample size are not handled by this method. Those diagnostics are provided via -the $summary() method (using -posterior::summarize_draws()).

+the $summary() method (using +posterior::summarize_draws()).

- 
diagnostic_summary(
-  diagnostics = c("divergences", "treedepth", "ebfmi"),
-  quiet = FALSE
-)
+
+ 
diagnostic_summary(
+  diagnostics = c("divergences", "treedepth", "ebfmi"),
+  quiet = FALSE
+)
+
-

Arguments

- - - - - - - - - - -
diagnostics

(character vector) One or more diagnostics to check. The +

+

Arguments

+
diagnostics
+

(character vector) One or more diagnostics to check. The currently supported diagnostics are "divergences, "treedepth", and -"ebfmi. The default is to check all of them.

quiet

(logical) Should warning messages about the diagnostics be +"ebfmi. The default is to check all of them.

+ + +
quiet
+

(logical) Should warning messages about the diagnostics be suppressed? The default is FALSE, in which case warning messages are -printed in addition to returning the values of the diagnostics.

+printed in addition to returning the values of the diagnostics.

-

Value

+
+
+

Value

+ -

A list with as many named elements as diagnostics selected. The -possible elements and their values are:

    -

  • "num_divergent": A vector of the number of divergences per chain.

    • +

    A list with as many named elements as diagnostics selected. The +possible elements and their values are:

    • "num_divergent": A vector of the number of divergences per chain.

    • "num_max_treedepth": A vector of the number of times max_treedepth was hit per chain.

    • "ebfmi": A vector of E-BFMI values per chain.

      • -
    - -

    See also

    - -

    CmdStanMCMC and the -$sampler_diagnostics() method

    - -

    Examples

    - 
    # \dontrun{
-fit <- cmdstanr_example("schools")
-
    #> Warning: 182 of 4000 (5.0%) transitions ended with a divergence.
-#> See https://mc-stan.org/misc/warnings for details.
    fit$diagnostic_summary()
-
    #> Warning: 182 of 4000 (5.0%) transitions ended with a divergence.
-#> See https://mc-stan.org/misc/warnings for details.
    #> $num_divergent
-#> [1]   1  29  13 139
-#> 
-#> $num_max_treedepth
-#> [1] 0 0 0 0
-#> 
-#> $ebfmi
-#> [1] 0.2395348 0.3195515 0.3628632 0.2240513
-#> 
    fit$diagnostic_summary(quiet = TRUE)
-
    #> $num_divergent
-#> [1]   1  29  13 139
-#> 
-#> $num_max_treedepth
-#> [1] 0 0 0 0
-#> 
-#> $ebfmi
-#> [1] 0.2395348 0.3195515 0.3628632 0.2240513
-#> 
    # }
+
+
+

See also

+

CmdStanMCMC and the +$sampler_diagnostics() method

+
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("schools")
+#> Warning: 104 of 4000 (3.0%) transitions ended with a divergence.
+#> See https://mc-stan.org/misc/warnings for details.
+fit$diagnostic_summary()
+#> Warning: 104 of 4000 (3.0%) transitions ended with a divergence.
+#> See https://mc-stan.org/misc/warnings for details.
+#> $num_divergent
+#> [1] 47 52  3  2
+#> 
+#> $num_max_treedepth
+#> [1] 0 0 0 0
+#> 
+#> $ebfmi
+#> [1] 0.3009803 0.3412086 0.3406291 0.2824946
+#> 
+fit$diagnostic_summary(quiet = TRUE)
+#> $num_divergent
+#> [1] 47 52  3  2
+#> 
+#> $num_max_treedepth
+#> [1] 0 0 0 0
+#> 
+#> $ebfmi
+#> [1] 0.3009803 0.3412086 0.3406291 0.2824946
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-draws-1.png b/docs/reference/fit-method-draws-1.png index 17a2e6ab8308c1f0faf28a415e935c48a7816e38..74c1b4fb58c273eee1af12c160a2dced550386f5 100644 GIT binary patch literal 72272 zcmb@tbySpJ)HXbTfRajxvcP>wCXH-?hG3y5?r!KIh!$?0xNP?|n_MijoXI4h0Sb0>PJ)l~jd5Fx?>#jJJ2O zz$Y_t8T;T3>n%)161;=gxkQ5>c)RB)tL+SdPzRy@qeZM9m_Zh9mR=R9Jl z+-W3+=>J{%D?j+Z7+5|15#_-8(bo1h zOII!!%>RMjga$W7%JB{X9$sEai36erwX{ZUdZO1Hk6wO1dkAe!?pt;FyTA0-mID0d z(UXUvq|nxxiHYuS55dpSVFbcy5Y0fI`@iKt{_ocF|Nn!M{On_IDV(aXu<3zAiKhq4 z&HA{#PHkBYo4w)Ez@q%Exh)2%22zD}FVFVa&s4h1z~3bad!6p5A&VP!zg2T_-_l$h zjg|Pi!ewwEv7z8pNW?|bG$z-O2g_#Nzv~J#o*Pe4(_CL)f4`dhedWb6gXV8d%y{%lC@3%3d`!H3{Bc9i7~qwHVT#|B@jIla}$h z{U+ed?RVon=eFQeP*O5+eYQ~YEYR@RoLD;LBt>wyT z&V40}u9{E#==Yw_o|orAC)toOAwO2DS79Ttiy?Q4xCvDz}FLMu&si00nU*wt^B_bGYoMbXI|58Yl3P1iZE@7w&) zEE(S#`*D&*do7Ea>X{G9&-T+upTF4vcA&}-P3~(qoGGQ#=w4=DyA-Gf$2i?vC>ip* zIojIYrSY|!skWQ0woBe^*b|v^>SL|nZ^MPA=p-$<ZU53+D{^o>suK)2{wJlcGie z$koc}jKf6QRZrSRMi31Gkprs&`{1)}LaW9WKT^|tQuF}7S2cgs(ebR{M==5pg$1sE zPdo3o7#$hOV)s1=mH$2h&vx)Z>Uu5*Kckn!$H&*HwoSBeK3`kfQ9ONlEF*dm700TZ zYt|QMwTZ}yt_*^tUjTITf4x0j1%n@-5BV85c+JfG;SXo&Q+*os+d|}sTn+(_zN@im zZOz>0_Hsc}LLvbClI7tDS@S+-6Hwm zaLW<$ytwui@WPQG(W@o3o8qi233%6-)L&rvFLQ!FnQ7M__CAMhR=ReH{QXJDOhu4( zn3~~4@ENwn;PuanB9#3uM!w?*`W{B7#n3k$BVg?%tXidF@MGYIx0N#&zhp$7%IUQD zUMGkodoV&`uzB3kjK0>_*H;?1W9`g0={CIIK>$b?8z1N0-O=;h%=zwfeRVVxProx+ zwi46Q@COc4x=hVT-`iy9jpji`i@)84+7(Kd7no#vkLB8hsS%r%1hgV7a^LZP$*YQ_ z#hX$Tgkc}4Kp0v7J)$k|k7WCLIbewt2cFy%W9y=~9w=v7nKs}|e7S?nP?3cz=?0uB z%moI;{UUY#@41XP+@~jk){-Q6+|}|&_ZPzAG<3BHwrx=s@A1zD`jh#T!Jw*~iKE{| zudSMnQuan;nABnv7-IWYyM$j0kKFOJET5a5_3m8=`@S}58i5LcH^VbCKipsXPESDd zb8@8M(lL0t^)8#)qqSxBSJ%#r#e+P-s_n7Xd4q_0#Oh<*E)!x`FUpCYnv|z&>ABL}N zwBfSH&|k0dwY(Ngi8HN-45f)0q9Bl}8+`DFEuM0Om(PASFWI5lqc%=kSvjJ(<;K%) zx>AnIy#Ke*{6TmK#Xk=>8A#%(KN;5zT2BJerG6z`sM3B;Q2+gu86Cw9%EsaJbDqc| z&ztl0AXqR678^@JwByunvl*tG`ED(jbM1TN%&WhLhhI?rjzt}v!0Svji1fQL4b4rM z;RSrRZk=Y%c}TduL?}GV@M@VXNyx4EO^HDfFy8AVK8Msnp}ocm`#sv5C0cHPzru&V z=;#axu>vkOmHh-WX>_1p{~hYPlQB)?v`xvF1fekZ@=1NYpuX2m6`l1CFCQPWTlA)^ z`Fh6znPgKeV5Ym|sKqzu5{-_=?sEU`9d4fs7wVVpDzsON*sh(Vnj2k|4lx}sNg%zS z@eCl3I>BZT%z3P*cuq9TL%ZLNA`qDoepf3xrA8l6QN-uG&+rWl7TUJ7@94H%;jwTz z{~e(I7W?fXJ1HL4*yLnhgzs?{>d(#l6JEbt0o2UjuW?;?Zy3bHBGKy+QH{s&b}4hHo%7uPiG4ERcRhi^ofUpmJi9rdy?sq6d>m;$_|2o&%lB_PErA8{ zA@6;2l<69;Dr%opeY_~ z$$zi^4_oBc#`yQC*~!$9?;H5>2SRT#O+%lP+J##{IW~nB&Ie5iaUtFx{*T*%H-MNv zVOfsi-Yr=@D9VR=jXwVWorYGiz)nt1IwxH^?_mY>nsWB8I<3;kYn0pRe()DiaIX8t znW&H`qu#_H*PYw6FrKOeOTmca6kuADA6XF?ko`00+5Od|&V@fNH1ywFH`^n;hjtA) z2_uJI8X^*xsS8yIoIB{{c3jSH7vkeq_?@U^FfcI8sAbmubDC1Bm5#N0*K%aVtbzF!G3GONld!HC5`{(>T9_jy5(i@gplM z0QWahh~!i751u_ZsDy3+B_RkR4U=}g`g;KPd+*nUXxn)r+{jwUr}RyP?jg7n(J@-2Cf)SlIx zpoO~|ftnu$r@c-WrNHXd8eVx?Ys&hK!bfTrgwk(Cc1Hi)Yhh~C{w2^{X*vNU{YX=`958n4KxX)>lX2e0W-BOd$;-`Q%pWb-23t({@7CG~kiJI$J&h!SNwn&qRPbUX%@)K8|WAI{4e@SU>?U6p3?RZYu{f>XF z+xY2^VE#6=MMVRFjje)JxvDcSw;`e@R?@FR(u(PoUdfwf(+MMnX_(kI#cHyKUK*6_ zHGS9hbGGv9=^_B$rT&rZO#Ze~R&aQ|=^>%~2(bmhxcLzx*ns#E%_D``looMM-x`;h zpd7ap&8v49{3X5ad#3?=$UUE^9QaXCfK569WF1VEbl|kloDO6 zO+N9{Uf1xPcJu@T($cf)1kDf9)rxzPqNL_l_bCwAx?x(m_ANWcS*Q^Q|Ez3oOH10C z4D?TZF(C7vo}M8W0)!5(CrzsvO9E9Uu!^nhuGep@h2prO=4+$a4T|Y}Y4N9S^0#Lg zh?eGOIW4ymw_|Hz(pl|h6nPZ}I#IHc=iUa_*YtBSA!gd`+6{ot`uO-{HDX9Gl!u^_ zE4-u{flQ>qL0m+K{^qiD54DVT?U(0Gb|IAwbA4}Wee~6}pXbo_$L%p>K~(zWUBx%* zcOCutk_Vg~=)C$MCI)5zZq2RzJfkhC;lHaVedQ#0SfQS&bruqmqJ}Fu8(kj}o%pJb zRL<6YoIlnno+1W`89bgQ$KN13BYCOfiai`^K8E{U=mW!8g0QduMc{mX1psOLjEI@# zU1u|GT5t>zldrB3WdB25vNV<(B^|Pd-+*=ZnCJP*TLM%H zFRJ4_ImVX;W8~w{u=^^zPG-zAyLpFdq_OD9a?|^vz%R}9qsc26VyqwoP0!7Cj=Nc% z`b6N8ENjE~SRY}T#7Cmr8a0pDz=WqwlOiy5pIZ14gJD&Sb(FD=V=bcP*ZVto$!Tm= z_1a!&#mONUZLJ%<9aLxy^-9L0VMpmgF8<=vw8Zh1t2^D z5O)8hcMDx#9@-JgaYp~F;ZCG5$Qh?PNh;2%K=8#s8%;K5A_vQ4o;mQZ8 ze4^Jj=4dhUhwMW-8y|%JWT*(iI>g$!p*csbs4Q&&|x+;K? zTQ`r*S$vi@v#V3M$#KH%J~sTsdG3nT%3bd)6lDML7X&8u4u&oLnUZonnfizp{Lj+fE6E-P6_=4w=}KjD|r>uWj0)yW}3 zFZHia!}hfoP~470*P#nRmL_MdA?|nv?>am!d&CZ7NIpBvUu%(AnjMVEL27m%`-#k^ ze=agMb6TZ>LKO5qe-?xoIbHdBX*u&r1*RE}K9-v0B-mj27~aG3$f^Cui=C}>kw>Vb z%%U^fOiL{s*^;PXQI6fhabhAknY_5^t-X2wIv{QjAMpj>-vNXr z6;}h%<*fz75JJ(k)JIc_HAFnp{j_rOW~@HzpFbf?iNL_=9jJScS59>NMoFfqduNM& z?4KWNNusYUIB8dqu*=?9+PEFBT0~!5sku%MD`jFrZc*M>K($eQ!VV*rw-`7I{K8H5 zdj3V&*tsiQ$o;6z>>!QEcDND$=%7RQ39UI z_!&2;v&p4(Y=O=C_kc~-=sM=&b19_PW<6>BLr)^@IDsW>bhK5c>}gVo$BWdN)pipd z?=%R;oLJ7XS$$Tg5|mDp8LdWnPVBCq8o|i8WoIIi*WHRkxBLtuM^@F!SE&;s)ceZx zdidGnvAPfcPVx%iLO5OnJKhvx?-)yau^?Vdr$4P#XEw+n;iFZ6E7$FrgQ1@*=PdZ_ z+S*4QcJ|3-C(U)!E-cZBkQgGZQ3-JXG=D=>9r`rt%hRY|8=_zC0m_Vn<)c2l5On3u zVsV7fd?-Jol(b%K7iTyn(r{ov2G5nu$!aqH=y7uNPWU{3D^-guAT7d7bg5C#o2t(w zy2tSVAF|Gz9B;2rIi`>Kqx&gZ^c)4VPDi?vR-9Gq_@k&(w^IxrXRYr zUS2KU@3qm=yGV9i=H}-O3Kcc+`pVTcHNS%&@bU0sy!bVk?%!pk;j2I#-kRz(J6YlQi@TnZCYY;1K}>~$aRReyHf!z$sJ z{CuL?-%Vap4b`7yv=d@ph0e7IWQ!&zBD^f_{~-@s0Nb_Ls7FE7w;~aY1$oq>fvvB9 z?ft6zqS4tBrYFdrpEUR;3_GVrIqI|;){VVqa_j3Qds%NgcYLD+{dfn!8;yB8#M^F6 zT=Vv%JidyD&MHpxTz@t;my|o`avt|x)pvd~;KxR95dxdJJg=`uqFvj6>UVQO$<4>z z9o@c|$N@6I@Jr&-WWZSScHe!@v4s$lNB=Dm@^`$LK*rE8PTjopR>edz$w2D4%lBv5 zlSUpRCO*~8El=4FQQZc|;!~IDi)3t;xvD3EFH62+fyMIRcXJ>{6I|q?@BNiWR7ay- zuMh*qN{ofAq76dPe$J^c%+Q-5zZi|HE(BB2D$x3^Tz%fW4awQ;d^(MHKq4O)xNC2? zI#T3trqn=aqiLKddVZY#v-fmECJFq>Cc?ag;hl|_n4x>FQWN>dMhG)D0OIE8LAeEY zw)`f2a|Q|=*?8`RHR~}3l1TG0kD(GTXzutag{1`^NWaf6rAvyrm# zyoqbgWBfs>wl6g63wv-6?O#6p%@+p^)FeU|s{@#`_#8w}ZPJM?iQLd>m8McLQVVKs zshq%{^=+HqazIc}p(RJbhOlh5z5!t8*Z2f(u(8X}yD5_@AnkZZ$l!qEJ z{5@B2rD;;2zUpevgH94cc``otp>>CQ$g>%y46O7k)Xs)!64ur4lJ2pcd$=-SWIW`r zc6psO(djmP8cAbS6y^64wXrPhn}a8>eSEZ?+{ZK-A2+#%SWyH9=M%!#p^fXeH*Xbm zl}E+8?)isE`;+;zQ*BqDoO~42RD3W(k{Q>br}oV5SA;=`)zLeKXfPTFr@&@X?7O*g zUcoCJPWq;%seb4Sfm~>I;7`3!mqCQa5Y)0BE4>g!LEKS$^)g=FtM^;^+hP}SpG0g8 z1~n5p=+}ysl)F@UHrcj@qF7Sw92ovAnUe#Pmc(JaA$<#9lGtPGp z296w9pfU~5<@M{N8 ziNO%?QH%dZliU1u9JWrRQ@3!3jJ1#ZY3{uJfm8tE>qpN<6tfy6dEyGmk?|)J>zp3h z+z5K!)VlgLJLYceJ26svsz^3(HS7iK`SeXt@DO9|fnEIU?QmSngGS%|{fG%!YAp}b z3n3im6B)NaT9goJW2<h1Iuk|UB|RTQ)8Ybjrs zj#vf51_Bh}&K{eTeMHlowF{{2Z=zWsM}Q0Iz2`^NA^xm;1`$jp~Aw; z)h?{Lzb!}QcI>}(L-t7BK>em%(dxEYFM+WY+&XHtt7B(L~uSP|7C?zzA zemI1ai>c}icHVg0Fc4RK6%aGe^PHdR4{PTc&Pz}V1vzuRdcmwjK6#uO6VVH(yi*<3 zr7sow2WP?&BEYg9AOv8d?q%uWTCYM7-tTAIiUFRFdMbd5ma7ap;IeXaf(>dGfpi z%E7ST243w1cKXxVDv?Zm>RfIKexoFk#aRGL(Qb^F=dl6B&GpN_8Rue3EPXNL|}Vsin_lYjf4z%zFAF8->^QiiLHW6CCRvr=*cD= z%R#hwmUeU^9vWkgNe`@fuMyBWI5oq_LUlH7{$6>&*#Bw)8gz%LY)sjsXiF+%&0bLZ zKSgDRg@^EB9|7KP`qLylG!h;iR3;Z8pv1}aI|urX1+YqOGU|vvYbTjvQ%}~jHr~@v?=OQH$&M- zaqx_b#7DeARShGOE_7&C%sKtuKg$(T{yY)YWx_8_A`puL9TOE^FjZRfGAW@D`xDGx z(g`m3OHqA)Qzyw0gl`yP=zf8wgUYbd@yx|o{l@;$5t4CW@!5&-2(1MC)K$Jk+i?B> zl4k(WC}zL`%vEEk%A#W@*qy&WEKEta zN{M3TO33kv%yu8*Bopp8M-}9lKGJK25qY8#*0}yXC?M?FLB6jv`f-n0hhoYb(m6py zDbRCu{^N!4wDc!xkT|GTrn($=C~=8auv!U*jd`;^I3&^uKbz4h4P!BYM1yjYKfJ-r zPJQ#6$1GKB+lRyO?++D~Wz}#Bxb!MGGvWzr|651kTpgR>>~mvlJ*PUSv!=E?pDj2%{cre3+fN;g4r0UK44?7fy|MW9C%c3lR#22ug_57FR%Xo&t2Oj{ zPnGp=J^yR?a>ML^23c`(U%L^)06YQ$bv>E}BP9GGQf8rqA*4VUeAR$pXDaa1Lgz(vTj(5Pxgz1;1G{L?>BO{$%$s?u1H#4(_l z7xbuaBC9<@cbVZ==ppVT0*Y^bFP*+faARw&N>33yOU$%Sa$7G!?jEmHev@t(4s0}$ zW&SV>Q|x3Bi7qB*%TMte_p)(Yzx)9+Y%-S3`Z#r=;zg#J*T;XGDHgUVtL>RRHfmNS zxie=Sizhey)9Unty+y6)J&3{`)B@gh3$lB;>CE;$lv>Th4eMw8J0AjSC`UXXMylZ8 zRtJKtDEd#}7tL3=V-}{Y^Lb9^T(>Ncub7&;e2AxN- z*LI?tb5ZX%?rh_N_^h%n+;V-AH_!i*0=T&U+d99{_$Hy)?YSfstw58JGJgXBS#{dh zIy!(fc{mlK-{DP2$`j*!k9|LVP&GLFT#J}G48I4gnZizh=65|E1A@#Y+tbs2><&_~ zhuy8nVvNe)S+g>xP$lo8C6=X6vOq6h$gXplfAC9pm7`v)L|OQ%kN&Bnrq&ERm($`P zHqlje!at^0w&)o~y&LS6;l&=tD$TeyNToyu`*MDky%hP}4Tbz2xZ-PTkNDo9ut->y zA;3(%#mSSPN$pZ~pd}X1XVELn0ufOMxhCioTn80M@>9Lxr=q#We(|Ij>6D>s8tub! z>lqtE&B8E7b+T38MoR^&PdAu;-~Y1Zx++rSVla};j(M-P>*EpVXMmZ*SIc@#Yad}P zDyJ~n@IPrQ&t@QcY8_&is1bfCG;q|wC+X6DyUo0p4*5$2rGSuhmKj*xNFPm>gKgFf zPS-4Wk`jA!Vl@l_xMc)qnSbzx*+%s}Wm_&8x^-Vnf@at5)ZCMuUHa!(t$#^b18Tra zE4i3)*it=TU4DwClPNUA@f-xZ;Z`v-cEYoY9+9{`tySjO7l}Rq7?tsExVe~IkCSsu z*!QA${p`g2j>dZj;#L~<^uEQzf=oQNfjr@{7?&*M1|DB&g#2}gmnR%U6{{I{2GeaziA25ju>tIJ?wRH8|YheaAJDSayr$$N|pueG+})4p%4 z8G;S$GG+iIZM8DJ>p4<(#y(3eC)V=xn!Cnm`wt%aLlRsfDg(z7Igwwt+EhOgr8G}G zI$;w8QKR038gm(C=F1s&m{Sc3Mb=43S;c@erV;H;zT`Tqdq&=X1|_H`e&cqF=F^6H z)p+(7LF3?H2X_wtkJW5lISTJ5Il#xYq1PV*L}Tl54td>Tus9=vLY>@$Gq zqflT0>K5rUcWyGpnI147`Eyu3E!L!Di;gb#i`@_gsZyQ$TMNtDIvpP^97i*QqgQCo zEV8O@(-g1M$RX106LQyhrND}S=q?8&)J%X0(ghN-d~hPWgR);_p72@7q56^K51 z=UpJ*TqN84R$2S_c}0`}*L*e`ixgHg6UZGsjSWj1G0+!rQM(qhU>k0^vF7Bg?j$Rf zXN~{BVYv?Qowhey@4&`?UF*CSr{3d>UkXkqaRL-cJ-bb((V16r@G1bp@dYgY9#U3Y zr*_Wtsn)}RF>8>hHfEn)f&3BQGG5M%< z3M9@P#{XgO&!)iAKk#4`VqiSBj-8WBG`cM_(hq^wpf%AYn}v130==;sIu@M}%vD+# zrQLWI2|^_aYW=c;&j%Kk75RV77qNb>a^KUTeH{WKF~onw^R0#KW}(OI{oHKw!Q#^G zxpHA~;I5%qt%<0apD1tRY+9xX#C8Oz#?U}`pyZq{*UO?saoPBd58K>NBWt9@Q*Fsu#+Oz*{R&MjnzZT}O$^bK{O3^Wy_yph8!gS|rjprqbpLx5EU*t^R^R)o z?IP~_W^f0hLl~QMrp_W7mxS#@k^-w*E%juqzsh%e<$NeSMxLPgAbnP8Vnt&wE~%si z)>M7tl)ew0H}%t?eXDpd)$oF$pvu-${Pc7VNk@^6IxK&BrucZ@nkldOU-DjGC5$7Y zy$s_$hlFo1(nB6dV>>FHx(MntPmmb8vN&6iS3Q5gnbyTdE1ySOo^C9FUl7ZEiAM{7V`Ne*93guvq_8Zq6YBur}&Tu+w{2S1XVN z_0usIg969&gq3f%NioS9Gu6udcLVZ0;lBAzhDk?G5h4fJ{T1EJi3-f|efbRy4XEs< z+U`e2hB&-OmbFu_K&RvFi@IVjrM=Z3n5_bdpe>jo?>g|J9wL#H8mSYJ+#~!lUbM^@26?NTFQPR;{)cZX=cPl$L6#$v#hLP&RjlIm!ZZ3@0KMH{6qZ&-z_N}1=-7b7%wJX|4g@=2Q4dXH~ ztLChrx?w!Da}QAMwSk^=l0_VrQt3iC1z27BdsgeCwO5LTp6z!=gK`#)-VOw2p*8vbncZ?h&r<@v_qESN zNJ4PhLWw#!vPS)GBs~mK$!B%|*9X91Jq+49?7L@4L=(;nl*C5bbkR1zdzuI#G{9a# zjmc^^_ah*5_Z=zljhVO;^h;OW#n`Fure34kcN!Z9)EP@o8u$;0(=iw_nYMUp462{y z38#$qViOr2i!p$QJs;sc2HhNN%#YsfW z$DZ{$N?Ah?BgXbt1H`F!xZb9Hv<#3e}#~-IPme;>x zi{g>?M84xww69&Rjkw7!;YuyQI_g~kEYI2egJrH0@u<@0f}uM_|(@ zYB;=G8)W3{KD4L*bBJ(QcpQbErvWVc8|bdhHT+C|tl~vR{J=G*!>A>S__~_LDrW?s zTpk;t!_gX_;){w;G>p}4ocJv>58oID(h=rKq3#_YX^@znu*2Dv){p*)tI%-_PRfUp zuJ=zbA6U764|il$rF#(5kdd7A(d5TqD-=uq znq`U11_@W{{Je(|)ts&~ zgN+Qr%=YBG4{R9P*bO-SAt3w(?R(Xp17k^6#eqby!0ggGZ#ATx%n!nP(6;i2gM7j0 zQ)9Ma(V2P9$jT^T`=5Nm9Kk#J~@33vihM7J#BGRVNK3pk2$qoWW?&np?;l;BrE5`ETQ63x{%#pAFHHJ-7T2D9_WZJrKkyW}74YFf zUtcbVxYPr!xKViIscXWzo86bK_z}TpRx{t?L&H$n0!swp{_nXsY z6J`B*$VoC3ljo8L>;02Wda1WF#Jx$V%iwScSbe~e)4@BI2t`lfLT*KIsM zQdgQ_tSEgBeWj`?FF5aD1FZRGLJCN_aK-MCu(r;OjcNSX>LQ8g9$Q#I>A-fE)Pd@d zBrJ&=;UZXcb92KPkte7fmoq^9u4vLEwSl*cCd%hN1fk>7kvJ5K*^ zMi#!pc%N~)@*yaRtYQ>TPb+$w(m#f7P>14hK+Vw`;m0#Y#^{Y=VT{uX{ILZ$(Z7Co zY6_+l2m3h))D~+db0rUygKkzi%e8Gzywh(u2O58RLT*=fyMC_pMGR6LJ(0V7m<3-DPbUBs5?J?K4+Nvtt+=x#mZ?Z#3$2iiX|P!y#rcgt@I zRdp zyr&8o*%XSvnOw`1g*$nay91P36`HiO=^=+`sOrS=t_F1UqZ=5_?x+ia z9m4N})vxlfr}MHZoubJ*SkLNF$stUHUQN$Pp*T^V-kE&}V>DMpa#n5!xr!AIv72uV zkJB3M9Tn2Q%;?sWpzG7*Ys%pjKi$9>-Dbj`9KWmA{Q=YfnMVLIYMI$G+p7la!imL8 zRxps$apO)OKNa)9zLTB@C;$&L`{r7e!DBm5L|^6|c{`mpp&J@ITr1^gpbP{_KSUKm zT?Q3&DCktbB#<-5WlxpgrLNWr97toFDz7D0(o(s~t*c#M1# zQMm-;Wnde3T>ovl*`(*wbg7vEA~%rIL^k23S+cQ=l(OnIe}FJDqimA=mri}vV@F=P zCDs|V-XbW2Kmhh1py8YL{Mhy$Ja*$D-1ksk;r!=1>pN!?iv584Ql=mVxe_Xus+!OM zvOH-$0SunC8&+fL-4_e{?=%b68Q7R*TcdKjasu?%&J}I4>RsPkcwl3un`_5GCtLx6 z|KQIDt+8vM2Zi=`3h7zDNxf6u^0o*w9euT^g&A%1FK{xtw? zFdfQ3y%|+5u=&|Xyy<<%m|t=d+gT})zs^)^XliOAhVW30_xva=>oTpvS&2ODVN0`k z)2)qo-^21VZ)t?C-JQQu#{O=`HnI!XeJTUg-8nZRK@aZUnLJyGR7$&aWd{ckI9v6k zU@+#Kp#&EetdfX*=Ubo&Y-9z|{Ig>&t%YLDhjuGH=B_`yZ`DcALCY8-*|Iq7iTS*qYSL^Hn(c?J=l5jZy)P0cB<7x*%099EV5?PI8nY)I;*-MO1R? z9rq6u@AA6G9eBM?9u3lYk&)WUD+tcg0r?P8MvrxYa@6pfs0t`Qng zDN_g1PfI3Kws@14cok#0K>7Kzp*`HKC_9E)8*K!p>v6HaACDA@nA6N}Z3&OqWGP7e z)7f_53>|D=9wf^4JOvtLp5^h0w-mpUn5@(!Lma)ZN&9`(+_r!!CZ;FW0klcetEoV( z^_tY6Y>xaMlLEQiUoDA_{xYc5=6s}}>TEHj0BC>Gv22CNe`oIG(gc{j_9DCaO>)4N z4H9sj+LUmAFvefd9Nk(c794f}%M~A?gY0v4!zr-7k8BnmPZDgMppAV3E{0KTNIjyG z=?Bxp5j3MF;9!V@4zT=MZgRC#ZzAUebXn}(TqSR=2KIoyd3PL}HOCQA>s}I$y*i6E zodH=L_@Oo)~e;Lq_dJv9U#F^1QdWH9@P?)}f+A~vrPJ;S1v!SsN3az+^&@fi z+DIR-G((HsNpxQq0|fHo^6y}zf#@qr8_;bQFqqxjf(_C0MDj$3N&bp4hZMa2^&*FY z9%TAO8rT6So+Ae;s(GrQh7lxm%kWB-T2Xb$q?lFl=#9eGv_{zy9wi`Y&L^YAkhKLd+E~Gp?YR#3_iie4Rs^1{djP>THr7j|V+Rx2 zH*@pyjdA{_*o8G}zZ;oOn7&N>3PAXv=TEkcW~b}i^B4H6cMEWn7mM3R(Ro_Sny^0R z7v@;FMR(%&6%v~8Tnx`sV1$9YiTmBOzY5Je5`_B%hi#*_*e}e z!$)5^(G{rhKdz$hUun#~^=49P2VzRcH^^`-pcr(uwsE6Rw4JNrPNKLKV=ZUNH|7ZC zW8W;J1|i57NDT6bfPlS43n}Gd!Jus(xWyxEZ~s;nv9^LH{8#$VKg}PH&G~%#%|WnW z$~IqoOEGIx)zlAhoY8{Zyfb$Yt$rm!u0Uf(1qxa{8bZvD(2Z1O6P{83`qFrhk?88<6^+6&Ae2o7y>(D!?{H-LZu-k~7m0{JlDWE7&h)FYrMYo^0G=ss`);hmW5^)D?be|8uKe|Wq32lD zrsZxQbS$)U)!QR4d8I24w9>sXKpq4=*xpTFb9gaD09rao#T7<~i6$*?Vk5i~0PPM_ zo0NarJ|*to-^!ZZF+GXhYH!Uw1R+ej@@m*H9*}ABJO>iOJ|PBxyf0-gZLb$QJNAz@ z?=guVKH;#BGqXOB2a00hx@b;(|1?nU@HX$fNjI3ZXtt%EQFx2yJ_t+74FG8w0tm9z z3ebR?3}Q&t?=4YudJ+ zi>~hKbIv=~!@##wY3033!Eq`AVcI$a4O~xGem&b&Mt^7Ty^Fy&0q#X?sfM@U``oms zRu0QU%IhEwWAKf>3>CWtsYd6p%8I#{;kBw01VB@wssWL zds(Z*r9E{1C{a%B*}c}yl=x$|V_xbgd!HOYUqOsoztA)>unJ5cRghvZQYDGAa<`R{ zJwL?9Y~?3m4d9EFC$$+Ng_7VfMfWf!n@zoXFY`s#2 zXA3MA{_8+l3)`pOUP*1TTi|aLDJrNv0}ZH6j_yYd-Xh9OxZfHFvj0~LkRDhlylp-F zo{PL13&w?(K=ojP`(U?E7RX-ggdhB#p-9|Ai!HX~S}HCs21({yt+*T&#_Z7}==La*I5Y~8K;M;}FN8ND=TXqs-IjKDlb4n$S7fPGXil+kIQevkOW^qI;^#~YNZ zAJqhY^vaFcxy#gt&Y!*vMq-`?j2vj&S%Ti%f%Kst1Kqo;Off(3Nlo*^v0bHII0%w_ zb5=0El9P~~!NRO(m@0S~{v?D+-M}S^L^k3;K~5U<GT4l!vgNuk=O*gK%9k+a!`1eZ5A+OAg=4??)$5G_I(E3Ryxh)CEDL#%ypik;JYa zD6!(Uw0?rd&$m7y_W@8{o*;eJ$zF{o5Qy))>5s_ddc_{ZwAifJgX(5qPE($S(( zI#gFJPzeEmkNN`BEwW9$E(kLn9i%PYOulN;Gp9~$@+XNLfu~L2%S&fOf$^sgVnBX$Ifd8@p*pU z$&VGY7omy+HM9Hhi6{HsZ~D)#?g@SRbf>&qZkzwWzWe>ty$0wEyfBIznYci^5M}BYn#mw_C8dAzV)kP4d=HSKt>uv)Jg=3k?ECfdFgd*HneB{^=XOpgvHg$e0>~^M zZ#L!(75yL))QRCv#DSQ{K>_WMNfWTT+I2R*KJ6_u;I_EXR6wgtRn5=VW79lZFSFAv zw#i);zn`bVIC>>utQ7|xI*J)u*I+J#n^g0MbCsFas4z0#{aA4`KS&(sd-ZKIcTJ04 z1afl!OmX>jF{9l}=ADa-5|KM_D)XlMWPlGZlKhU*tNV9=E{n&H223oULoZjSXa6RZI1pLCwXyMg zV`Br#8rm@mx+PLi%`=4r1w99GnbJHnRU?e;`$y3+IT~0Jj}RuS5+i(k>nGJ|v_;5- z+^4bfOn1@o2?-~>80~;s4di-q@Q}_-r8E)!^#?o=xUsB)v;i8#Xh~AIw4X z1qrY$iI0fkhRvf>LW>*7nYsdWh#1`lwaZDXuUf=(nSvn>w)py@#)6iDNSje7Z3zF=pU#Z}b)F%+%CnMyM#zAbBgaj8X^A{VWoG+3W2a-QalN$RlYi~?;r| zr`mHf|DpWaliqqR$B2!|o53&$vh3Zvr!UEhAI>hHq-=aU<6E|GG{~JR)9$&ZG&~#P z{>O*yw*M!vXjeizt-WUy!c{R2<)te1^QPff2e;=i_G8NQHSynx05}gHRWG8k2fm8P zZaM}DUcTEziBdE%41T8ecBZFO&GiLscQP*H7!>;@VHlK=@n)pY(RI0ny~!i9P0>Lu z1RdB-Ds{KSPjtr^Qo?&!qkz7bL9{coZad3E%C~RFR4zro2Uc;+YJXX|Q^S1iHYBC! z=jWp{mD%uP`azU$-T)4^la^{3vTmu}MkW~}`w61L!ZI?6v$!26ap7t)b&{FAGNeOj z{q(>zmb7;r&)^~>e$@)_Y08vdJTLz9aq4elOvo0auVa%)wJ6*Px+R$K%!KAhg&cXuT#o%jMqML(wN9PD$^< zk?hNM+&$#`%9ED^n=f|48T!W*ZCgwUbj{&)6YqeyMpc*k*J z-OK;*`GSIOb+gcM&&9!f)!o~Zppz}R&VhUPDD{?jNvq279@Pm!Gz;Ofm$jzvjga8Qi#85soClSxiK z-llsdDI?DE8eGElMMMm%a+tCe$823B6R%VY@WtP5Rc)G5bti{7$V~D@LPG}Z@_H$S zb^&i*YM{il--6AI0M!Z=tUew9iIy5EKzgnK%;{~tNLM}ce6`;2AOy3>>E1@V@)Qij)e&%P^5 zr%#$ZnMPOU`T%z*C60kjGQi&FRWQ;w9}7zOoWbDtCRMbe<&Aj?@e&6_qg^4wxB#h9BZH2{s=1RsLXz?NjM?VCQ-OZomT*OE4>$IzDGm}q0V3(Sj>p4Q_z;S^OMg% zOK>LxdFI^!Vl%CoR*xYH9}CnPf54M?8{x`owke-1td0vU*nMV1t)aIJfiQ3si0Yv< z_x`^78j;=Yj%|4d9OX~c4=dabI(vS)+2CJ)zm7pu`1wU=M947_Pbm1h!pYVib41iV zW%Q#8Vry80#rtLiOYeSvY77s~e=lHgrD0%(ONvH)k%dtOV{>pVX~t!eW22UBm-%8q zxGX2})uBM{y^hqjVMdOop#Iq${7v}E=}(-h?yWHwCq79TMtvk9$kr z3wQovp$I#o#SVo3d+w+c+|yf~l?M9=4K@18pl`~5cwi%Bg$i!^fr&2aLW{F!td6m_ zC-+<09t{yPl7v1@AB3B#(%0Jeo+z@~Rp@9)q?a9pq&c#4wOoDgK}H9`y6FXJtsh3% zQR@tkbzkdOJ3H}i#p~n3zZC2>Q(O#TDYC1VALgK^o4OCqp9RM8a5-Ej zu#T*LU^i_@Tl7%y zr|U*_w{=)Ok{ZK~WuGU?ntDG*#^_sVVQG0W>)ycrxnU#jsjgs)|H*vJsm{ynnoO&o-3e)5wokZe&o9md+n zlJ?}J@@e5rM=_4*IRBYQ3(fZ^x13PHf~-CBs^T1Gw7PwDUI7K9wWh6^fQc-*htYue z0Tz~!3*_5e(IBYZZZK5NSt}UZp3Pgjqb9OAg3W~K4rHxf=NN`|9XFS7{irBS7kE5; zPvFRt(dG=MtxU6O{ve+bySbh3B15oP(Qob1>J;(`RS}lGqiUww2OoHPk)z~8i7Oi5 zn>_E5vGemgTcIU~ro${?@a5R;nOe2pyZ`-u3h;gV8hetW8Ckl&@!<>|Q_pM>cltRP zDgf8&Ef%!Bnj5aV`q|Ru4-JnWD^#7HX3coI*iD~SdtAi}y$z~TGNCFpemLR;@>0=n z3Zh#%p&k)>OIQ*^I0d_(r&~&r85CGMEV_V}Y3h47R0*v5bfP^BUi5bF@{FR=^PS$} z{w?m;AZ$R*wdCTK+r#8p`k@5^d3B_d7HeGgn9K_dIE`&h&(wF{U?n`6sU?r>;yMrG z{VTRCHkcf{{*mW$TIR{U(^lj4;p$I=;$%tn<-&qYK`D(y_)@9;p|CFLY|}^LF5KB4 z5bdlGdNqQ!QFh{`U=GzM?wf&3RYt{JX8ouT@aFo&$h?Jk^lp~0v) zSJ^YXzT+cXGw;*)%@m-C7R~)YmI`+HQvEZO%Tef%ojCKCy$OFYC{X_3w}gBo4R=|V zt^6IKJpAc2v*a*wnc)88{>_(-tAbXJN_M&wUE?B4HR(xeHW*4t&)9b?|LVe3X;>yk zzIY;=CvkV640|`3>~}miPtD}`n*%fr>e~E6yl16IfoT4JK71nM^Z(n2R{`yQfViDF zCFwl|Wpvi-KaYQWytM_jHz$n73lz=x|FK1({A^0Lp7!;?^VC>ypJyyCF6w$oB(RTy zh9fYwZz@AWLn^5Y^AWHU7k9J}0Zzn}$!I5$}b;obnV z^gR6nQ9Q_6b>4@w*brwegVH^RmgF%MBV5Z@$C`B&7tU9KWl4o)wD4pF2lYYuCwWZO z>Uw9+#vx}RWK9p^zOup*zuw&ZaXPDYASoOj!tz(O6)yMS&Paw=Pk7BX8k4alu zYMzb?Q6B?1>aYKyYeUi=a^pXm=a(OV7EH5$#O{254#Nb?kccEM{sj_4V?0MaW-?Aa zL((dIj3ve26}P5ou{T%!F4n$XVTpjWKG5%Fgu|4E@IQ7%m;lQyDA&Zf5^*#yX}KvL z$q++~zDDsZe_W14v{q!Pb=twu=WWCpe*NK^Z{cMkWXMhlty(NzVEmnEpLfSt{>Ep@ zoS}(A_@{#8B1$LDRae{MMbl0~p0c>2^9xAoplPvS^5^T`w0yERWBh?|P-^;IlVaCu z;(syOqL|A7$wn8?V1}i?1LqpB9lGMn;mycn& zwQE1jQ$eD55gciwhFg2N63UUZ1Y;gjQehySlhP4Yoi$ieWz*@sK|?#4~tC^u1jd_~+Jc zz!s?|eWPc0Ls;XR;Ppky8?IK){X!h>1d~p?zdtI50@wD9fyE4%&R%?N5FLvTQD%ug z3F=+-I3Wk)ha&}B+jn3Z7g;{NDbp!IFCoI^e91o)WkNMVZ7IgglR;bJ9Mm=DLzT*WpmYH8tFR6I+CU#(= zi7s+=RxLZFfjgn%@M<Gwx$C?m?x-B5x4v`4P)#@HuFK~<(6+3pLXiE)nM1F#}q zz*Q&&KX4=TU-m6q%B}|x|AG#(hqIeY+e(PXa`WnvA;<3}sI!H_ZNUL=@<`7AVln*j ze)}PZl4j*qTdQ~`4udE(a-GK5|1ejsMOWa8oy&(4sJBmz>e-R|S{E=nZu}=%ZTAuX z^+ocoy*qVXRP7_qppY&oAvc80EuOS+|DEfFI@tkU?sXQ!oxy(dQz1?df-L6&G-_lm z_FI2$KBX_=(e9eW`(`MxbAQ{49!j~Z3zz@q2S9LJ%D znZb|7_&e6O;^071NUO967Q@< z+TGGRaR@Pq2N1g5Et}1Ub@q1{d@2F^p>pEPhTu4(7vgq7xUgRtCXK zJTu!bcgmk^Zp&CExj$b57w|JuX5uHz3Caa*zv<~V$I!R7Z7klXspv8)Y^fHVxawxO z%Ry3;#~EhxOr2^#HJSk8_PMn6dDVoc@M^-trWh~7or0OMjylI1ou5ayok>C9RB`if zroH{P$I)tkERvvClog*8$`_PJW@Z=MlAuNUVG{BLXGgwe*|=*#fzzT%wQjw+J-%Kn z;bYL~S6lh8B1GwR@D}}Gt)>0fLiGDuKE3k8$RZNfxCvhn(Ze{2u&t%URZ;s6>W$A4 z?!+RP5g(CJ(RPdu=v!%cH{t zPNG5}_d^Gh5Q)2D95uvO7~_~!=CzvOaFt)u_74^;aMA$m7?fdA0+iZ^GsOwzu4nRj zSSAAOn_stFTso9bFy0$d`K>tENr3HGoqBe1hIvy4o2X&8d2O}201h`R8m8@*j0r|` zTmK6<$pL3Zd;T33q&EP<)HW@CE&YUI1M?e!dIzc9FdY!&OsK>~yDxy^fXotV*bHWf zG`1U^`SDf~B*pvJ5QRG%d!dzUf^ytZMu&EogN&l(K)isATeiWjI2s1eq^!q*`pSiQ z09bdBcE8H)LfzH##sH{JotA$5b?(gCG!w@U8NcT81sXCp=!&roi8pIJZAZ>aBn+WD z3`(#egT^^fHGisLZkpWr7D9hVR2NrOUNzX(mB|nxMxhko$5)3Z3rg33dS596eLbVG zL|P~*)C|Vf#76jZ|6NX--_ruOuABSVP%hU~7@aPS_G9iDxT)F~&)b0_x0;YYURj)V zDF5V7VElrcLYPia9mFpqHd|X$S!?Ho@Aebo6P!VmZWP||g1?q`4B#%HVYX?;gA&qq zB`DLZ7f`~vDM0>5nmt!vi~u2xER878n175#2x4tLqxt-j0E5^Bp_@Pz7dROP-+*|3 z)MocKu^x3x@8>e@F7j!YgJ5q2Sh0gmGtwCl5qC^I2^6VfDA~y8lWq%WEAu(|=SD+1 z6e47vm9vV20>%zN1CB@aR*evjRFM-q>azoj8qt1Q-=uf%1uQjpgc~xsEwN{Am#O_) zMhDIC7+dPYUqc(!F|YGBu?G`-_kQR~@k`vZwh=6iv@uH$^%knF&J1g8g7hNIG@!68yb+jx+pZJhB9LabcQs6`~lc^M8B}fHQ za%{HxwYKVwVR^pJBQqU)7D%xwjO?L}9%h5tMtWmLNA@tF%g)eWq}R2dCTzqm_sj}! zC?U>6ORlR}9F5wXpWd=?552iEpUz@6nVSZ=B5z>vGfgfMQE?tj% z6GYbTLzW$#63X*`Y(8)6#Z?zayI2Cu#p)G$K#o}^$5+*RH|v()0pW*z8Z>`K3y7*_ zW@inHf*piHZwz-#-MywHrnk+!1SkF254BJ}bqs2XiI@tZ>E82tK14JguUe@!3M@4$ z&k;fLEzvJ6Y>I0c$=y{i6!fFdpdFfqf&c%YtfAH1AC?&V2pZam%LcXPJrp^!A0i~G zICD>0#0-IR?e;npuR11Df>pe6@Z3%xnAlh>&Lb(J15xd$;=Suf;zbArIvf8X z8*Y89ZU?}dI}K?!GGIb+Eb;;Nm$S)wQ>Z}%p`RgviIU`oa?t35zzB)9?na2G1IQqTt z4P?)`1PhBset%qb@h+IXNZ{v^8(+&DmdvidkT4Hy`!iC55bEp? zhq26`ivd^YQu7wMtQ`oE+-Al%UQL{SHKT%KOYv6rV4i_+G81r104MeCMeTs|=nD}_sg zSSjARc`KBmp;VM@5LA{cYCNJQ>OE){iOfAt6SvvyO=j!;6X%wIkbs7YUqn>1Gj2cN zo-v-hKf)F->OCwLJ-i?^P)sxUMn{uBz;ip#(%0x24KOm*pWtyIjilBk-G|@qk<;6h zPlpFtxzNB#+BRdRS}n{p%8XJ*VpB_ux#;T2;w<_xu+Ys$7sK3*s{WG;Z{lxj`|l1T zBdtxzWt27KYA@3XGYDO!v(t)QO-Hwu=PsmlPA(c^@>wNceyxz8?pmx#fB!$J97PuG z>cfa&1e$}33r=HW5x=q@ffraM7u4U~Y_7C_g{r9Ek9i9gbuXv}IljEM{*|7?a_nIs zMI{4cQ167y2{G)-DP5y*_iLc;Yi}q?>G9Xg#o>?RG`zhVcJhsxz2=}F`hdN&oJeXc zfI_F#OC9E^#*AA+)Ck)fZu}>DN|ID|bC(B=yqTg4PT@bw?e!~k6tc3iSSj-611pEO zh2Ct!ZB0pM0BdRJEx2Z}`n2;4*4S)O4>>>z9ZvS2Q@?wITHh)PbQ|tkMr4~-$tEs~*!+gGrs&umV-Zs0-^~Zrn@3on_ar=!b^%^+jWXS&v7Ui( z!naU-rUolZa{u_bSD|eiMc6N0FQxm-tK9pi=ODFXXAVrE53Pb~5|&hNm?sjE?P~Jd zE?~vkm`h)6lspxyYg|{uP)6V;*xMyF3|4stjbau9Dq8uM>w2l2SqOQmf(n)iE6L@G z8X=uts!G(6Ei_c=Fh{1jn{mthj1zdS)wRMr1zFV3ep7I?92v=&m3UncB>=ix(hC@p z(AHqKg8rHF43bD)fd=B{4~s^uC#2UMH(@%7+;gF!&jKet02Yb$JFmR!Nw z8N@x1$Y8AsPnwcQaG@VoTJpvQ##1a@a&mC%pEseArODjOrFt&A5VY)c!8|~`c^#@v zP!~l@w|k2C)_(N?@X*>Vf}>;uQM`{^;^^pcxGcT}4Nn8xusB<NJVHK> zuB$5j4;KKhniIR(+Pco+_$aiCb>?6R`To@fDB3GAcdo3|I0@}|JS%UB166<241R$b zsLH*WV*z6mQ3&>eteN3O=)DMX_KOR8n8ocag5f0afNy47`7@OzHEr?*xE$W!g8~y5 zg`<-huiH8Bi~ke}W96(t+t?8;K@2 z$i7cJi*}dTr=t6Ao=*)MRzMEQa zPy5{v$`_tbHJCr|C|d5J2)zls!70&|4NM*@j6`0|@`?!Jt_a%mZf?lT@sNlaa*8d>;|n0N*?O)gQaCbN+8-`YZ6X6X~L#tV*?ieo$?*fp4_T84I)vD)vs4Wp+o^lwL)?jG^6}bZ%YIG)vc^*s!dIz*OdwC`Mjn%` zTpYeBnO?A^LCJJ;2U|+~=ZoB(1_l7-zARd1MT~=y6!u3^Ut9I~6+aFCy!qF39Jr5u;DZx8a^Bw&F{rn3w!`F5%nLGTW7+g@vlAD?<|c-sW8R80au{3krSSA9`=nX#IbP+$Sw%{lm1+sBU^b!GXYwUSAr8+${o0BCX&$>ipgU~$iy z^MbS zv)Kx@q>ClCP?)@>%oh)mK$dm%6^L4{;Er;GY0thJ=28DqEU%=5l8}`#iHD#~m;{P6 za?Arf87J@oz&42RSSrnCrd7v)5JH9Y-r(UKs+0cIH-3Ji+M`4Nz%l8hQ4+-*_^3^b znm@t1?$+oGl-W=-b@++jmOY_D?6(rgV z@=U#*B(Vtj1P1!Jvp`hsJtEFRqXhW3R`S|M8lPivpd{@d@oQadGXWlLdCT+)j8KNw z&fKes)F8<0hcr z>cn+)3kdkAu$vR^6%qS*A1JF+^&}mcw%R`PKyBTDN`6&>5xyrC;)&cZ z^y7p~j)me}Uu)-x4MbI!w6(S&{^0T529H~d$eyW_K5PY!EH+&KM|>$-pvD6WDq%w@ zJ4ozR@b85(NkM_ATXINz+DKd5HeI~+N#trOQ#+t;5)Vdc9cwI6gM41g(YqMwJ6kEk zLN`0!=z(J~Jm&1@;Agi~I}lP`*{kphT+-3#&(;|u+`zNzC2U)=f1eOaDCM&Ig39wz z{4OKURKKadPXM6@JhIsu*(m?i9yVPARy`EA_A`0ze(s1NTxW0DeizevW?&(_lj7;s zzi5Vs~s;*Z#UIabi;Q^7A=nqTp+UwTE(3loP;JoYPtOFEPta zqD*n~5$U6Y;f4zlP+;8Jlx5~-oE3QsPU+mflpIRzxe-+Vl{wPWd$;hT-0F5(el@r1 z#A7UIfF30uuUP^*V427UoX8>+jIQ(U1}cxgI{djiGDD}Jlu{f_jlvaI!&e-^f)GC( z8i;(njGAiZ$G|EeogzXxLEaRA%mHUAE>8~F$OV9J28EZ<8*r-Y&iSvH$nJVS{YK62 zrOJH6)7c2*m@eAVb6ssU#mhf(8P3wbwvHo$@<|928_nG{4?^3{VRCuT6a+y!H2F-t zW%V-q*(&GOo{BJi@xU^`XHo`u-!~T#1s#s#Bd9panY*^CKsPDWk2V2*z?{;ti^OI~ zTQAUIXTAB}o7kfB!o05JL}J03TfM>F$UrY9BasS_^pV0g6R-lEv^VVXv4=HRnxmDY zUzCOQ^w%VPi@)QhFZ zHWqcGK60#h$jis_Z5JHc#^>S9My>g75%E>u#sF@RJr*q9aNmUS-xLud0%%CX^f9WCHo7~RO%BHlnSS0rGT-L<%CXt|3*&Ff zq0jIydZ1w%e>cdoPmBBb8lHH*INd+>XMg2H3F$rW5xjT`C=NxF?u0_uFs18(ZE$sE zmAF=YR_j}>%tbIVgZHAG(CU1}*3+{drQ8}3If+?O=MgV&b8#`(&Hb?3EK~#Zz(_Q4 z_AGNwRADUWWD+-mZ_M+jws)LUnysJ5^5s)9F#RKCuL}&C+80YAT0q7!%vNh*{+Y#s zG>@wn<+XWY%<_s};P!vabr0ciVUN=6MT>Is1{>Uiovba9Wyn-uQm>+YlK9?rHmlYDpMizx>^)9F4f^^!5C^k< zjJLAf)IQmditGS?O|1Z5`|WkI@b*gf*IBACHIO1_if(8Z=N@pIwJDRM_npX`o5PX?%$HuLe2<3lvy z*Pz^ZN+iUDx0`s`Ibf=`l#t3pan+GP``tj@RwmfV&QT3oMHYY%`Im>&Uwnf>4{yB= zoHIX=qp1G`XNs}awJa59)>{W~@!Nr#9N=MGXckSj*6h-27V` zwM>7_cpHm!57P9*P&@aV^FDVZfuzEN45NpK1fS460KJdtz2wv9?~Vw%OCPRLMxakh z`&eIPP)bEp4d^xkG7<0&)9U-Ha(Wy4UUNcHKIJ_{lTXqBHK4&BBzYw3t6mD8pS)=? zzw9fcet!0(eFhJ}8V>UmV-sn&*tCn;@!BZND}8bHnR| zK=zlENMgm&T6*rGAQ$bJajR2|Wpcv38=m?)66LM{LA&&m*LXSALkkjNUOYDb?@L>t zK+K-ui1|~RXrgoh{vHNiREx8pMxEW4jidP0LmzF+Lpc0kwCKiAIS{bMMO4DhjJ7Tb zXz`W#kxx*m1qg@-%(&4zqIx}sMUx^X8_)o2jzjVS{Km@He#yRV&w1QF^B(xH(JigO zI*dgfu4ug50hZ)((Cs!Qpq57~(KXqMEC6YgxKwztUi!7J@96`zMG}1y*|)P$;uB|6Q5X z#+`U1re<{tLv&X$oJA&VYH}U*DbIA;-E{wII);lwF3AE5+&%M^tek6d@u${mS62Gi zo*YkNz?Aed(-3t0f==3b@9*ZNRi>XtKY1sE^Y%LsjY|Xn35g^Q{vXUUoa{_ z!Ak1#T$@3(yV6MTJP9Xfy=uM4F4|@`1a64}o9J!t?fn#g3tc5P(L3xP@Kn^F!x{n9 zew{_|u--@11Gj|%Rsw`K$n9f}?K;Tm{X^II9RKsiz|vual%EHf&s zFpzM)jv~s8qsGqsoj-o+sJFi1nLK&zZ~U8{%ob^iut>`54-X~PhopK#>T`96!7)mT z#;ZiO=f{*34mmyD-e8+~g6XLZS^I_E;-IV5Rx4JNo+@w26^LTF0qEI&vyH2#51Q_2 z>m~5%>Ohb}Fe7Oj41>JHVo4ROn{S$!NTN@>+*_shzOc|;tE>m(U=%n0h5SdFzTrfV z+-S^VvYn_Jbe4f2Jt;o(RtiQ1%iZ<_zJ#dG{n&Y73DZpg^ruPs;NH^e!GbK6CSyxe%#zAzIIRwdEbT9r3qJz? zeiNsSFEhw_twKl;JOsm2*tO%gKGDUeHJ%sHwNp68^tu=PbhpgBw z{O-|$Hn%X8%N4IU?!__rgNsfJUB(xDyvluY>N8K#0 zwUV*Wap#M*kf2xO-6Mrg2UZ0Wprn0T=?$moQ}4$T5hjZKO?;oDEKW0G*yQK8L2hF> zqh9|P=Ud6$1$(C;Da@eN6Aa*`tbXW;sibG;P#n$y0dmzXVPtdtX%9tqv{^1}#`8n1 z-7qq>$#=G)cS6$;w`C9g#s8qP*f+S?C?gdl`5ND}9ENk;vIvXrr~|F?ClxVnRMb#L z$ zFO2{9Z#kHh_(k$^=ef*;c-l}cX{=DQl{~dc%=*6){Gle`iY2L3HuK6{Wef_f0}hu14DPlJ`eXmD!B8=uN?ELX4#`)jeznv2EzXBOA z9{IWHzY=teU@<;g+>S_CD1F&N4netni0+^A7$lO0?phVg?H?qs`d2uK@T?932GxJLS<4 zatF(~)u5SHcu#HiYghpr>o#4CG$0NW)7`Y$j#6s5pp-t6;kqB)ky68PZ^8cVFGN9& z&&a3l8uuo8B!XxU3m3}3V~2H@pZ1VtU)~!Tdh(QO%E+T{q|q!bjxo%8q{a-BuC0(7J zCAZP~R`dAnU!(Vlg^{4K!!W)!b9F3mpdfVVec6jFHDkS=^a^B22k~Jr*Y6Ete#1%L#tL&e7YoWRoF z&jzjm2>R#=ZS$ip?MT!#?K6fz=^;Rk+nniW^r@4yd7Wa8Mh}F^8@}m&a=p>_(l`v~ z_#1BNs~|3IWnP`^o*d1^M{d;gRxS-qU~c{$@BXx$CQ->GVbhSXE-SCDhOc{l^Hht( zu+rIkDA#emY9T|V+q?ua z;Xuj5mjN`iQwjNq6lf}0y_lGe8Q2z@g)TG_m46Ut-^HEjtFrFzmX+N)$?2S_&> z9y>|YzqUi=!qnNxaiENQjD83Fae{R>=Q8! zNpa8sCIH~}%?EH6UsiPY&VgE<&k5}3zYok>A*h<63qirGcrb)!vb$yndidW*8XJff zww+J$yfz>q(Ywd*$@N_Wy}TNRq!FdZEam1SPvyuaJaQ~pSe|PE&p|Qkt}L^RV!5D4 zoX*bmFDHz*I^1Ndlsi>Vf2;bz<8A);8CsCr4<`P2_`%4yh&b4Fjky=Dfc%+~pj!#b z37S|7{(Q8bJdD3DC`qv~%k^NmtR&xX3h(V-rTH8%C#Buz7U$Xu+f&%d_QBCLm`d$l zxWmQ~l1vpa3VKLo2*-kf2Zwn%->MXF#q_<$mQq0TmLhGgZbR438(4D%!16cBA!-jB z8hw;?fZyyUa&{gH7h?bh|Cv=^5vXfZ`SzEwV2q-UF^MTz3=Ps$3 zKO=)d-1SkmzQF4-QSMUOOgCt}l$W}G5n+~$nH1+?`H?~e)V>Ri zjc%~+qRU^`96L}2pd4_#OmuDZ9j{JRIh0k6il=yJlHLG;1-=cYhxgipVNl1hgCSf% z|NZJrkb0Aq1&LZGEfIcJLd$GZ;fQ1wooi1CuuAsrP&o&1jHA2@W>?-t?=JDc)q+$Q zsJQ7|PNNmSkNklDLc}F5qm+^5p1`|VrAjv}sNTBbtE*;YVmd9YfW<3D)a6Aht=Pim z_1= ztj8tdONwZ9RStAwu8pDU+f7iwV;HA^1E4BbG5glfjJ%R9aBUrrU7}4v&R+BtaPdEV zR}7(Qj?W0_8&0}Z@;gb)6rc(4k9++Ut z+5f2UtC5J+EBFNu=Uron-}B)-s7)DAFRe{=s;QYD>0B;VvK*i(3HM&RQO&EM%L^Kp zeN%~v$jNC{MfLq&D56|H%$xXvTD^0io))b3!2@GmDImZ9uKFT74`2)8?oYGrnFl0h z6fn8NFP|xNazc=2D?;Hx@Hem^K53T}VN0)DOjV8As=3SJ`v+N2>P;sKH#{G{l=zB~ zrlFy#b+euv26V||0+DI{4YVxgyr3@5q9?)apBk&3TpFn6obHXsX}Pevb;zkqe$-ah zb1j5pi^c{p9_;23nZP8n{cH(47+mx2)QXnVs#l`U$%qDbUTJ2e1#*Gy2{O^Z@l13$)&11 zK(d)T?T#W7cm3$s&VIE%7p>wX>G{9svIDfC_%2? zjj^zU_0#|oAY#~HzwXL*fggImvgWEf%P-Gxo7NSiv#b}ydOI1Sb-&(DPBN*uJJon? zS!3k@K$yahHD7Vw5FiL@8J8MxFP6YF);=mLYP#bxQ!dSdHHYe@C<*;CLvN!mBSTFC zDhGUjYF*{uTCfa@7wD}q?_qz{ET>>J>aEqm` zQkhPI`V&6kV%b0sUyEj_;}Rs){GojtN>amCGP&BO--JXG^$#R%9n| zfqqmyg3%gaNA!B|nI=VzkdFh>#+;qYvW*kcC*;@Gj}Ts7Eg4OZ@(AC9aDZz|qjS6z zOQs2*W}C*P#^ZO4ijcrFUhNy~y$q87B6O$LL}N1?Ccn{v=7Rm!3|4Z21QDp>3f_r0 ze+LJnG97CKVJV99=?YlJ(prMQ5ftn+bUfc@T2-sqiyezoY`D?+j(y{ys%Op2&)3a# zb76YbtRs)Mo(Y5N9mRGho9v=o;&x9tHM&|5Mtp*0#tFtlgZHz;{m-=%dDmG7N7pjy(N0 znxFU~jSX=hh}ZC+_eK?eMh3aQmZAoFyYDTu=x|UjaNJM>4i&vkw2hD5=VOd`4$VEWa*G(z%FEfAdrtMB~ zSOve`Seut&+jyWC#VX*Z8|+T-H;#dZu0jzyiOI6UH#j1ypB8l8D;nvna_Cq+b^6Ue z^LXZbYq~EOv`{hsrPiS8)Y&I_hf=mUh3<~4ruK7fR|+@X!VHV<%s8gcazXNayaPk7 zSVroD9pj7Ds7lex3)$)6MV2^4AybG(<3|DIW?7M-&ai^k;}!M(#pPL#~pia)jN9OQGy7)HO&|_i^weXN_%yB zO7%u=)Cp_O^ltvi-;9V{dzuH%uI7`zl45q7D#C1n9=H0bsGCxB7c_9~ir&|jzU&0H zUJ8b^R+3xKaTQ)j<}&)@_fiQaPx21%_4Q(I@^R^nrsOAn^a#Rndo3py^#0TDOZ#KU zariciJ)_%3b@kaRqb6kv*T2;^kgwCEYOd2>)$~`hp0L+qRU$ye$G*Mc0=kEl#LDIJ z=!;lsC5EpldR0hs-CfgBdqS2^Z=jZ`B?^kg9zn2I&ozSS@%5QNhoP2=;)(fCk9&a3875GyN{I*o8svbXjg@gHT zTcgZ%5f;y!=}XBttX8GP-!6hqHJ#6B5Bpbg6jXt$Zuh*0Lh;z1cb~dogm`ojA59P+ zLx3yyQBLlZ@z)E+U*ge^0XXN&zZEeU+3zdBcj@O_Y8$^IqP9eW-Mm9^6^`IMm$VjB z+0=HMq#<6a)Bx5l!@YA|KUEXnd?2oJdNDc zT5+4XN;>VDbNw^!QTr=~3#hNS%GZgzxGehBF&ui}@haZufJ$CKJhlGbxRDb|*}=gq z_OaLaUL8^p8{BAEH?H9(%~9HwaB9=uK5&bbUz}J8@wuv@v8q92yeCa4n7JW*or;c3}7si^xHe!(T^@zv{tRuJX+F= zrSv?neKG7ihB5$q2(7=-Uwd)>FT0-MAe}cL4&Ag>dRJDNM!fey2f@E_ckTJ)Wd)x;Gy)nLib7Cib!j4D2o}WNJeF3dNL37x<(fd{Y;1ecP zc)T^nT3R&R>8|wP@6(^n!q*v<4HdDQTZ=wf*8>>{96v$ej-(T$5B5^T9k$bM785JJ z<^iAW%hu*hywto*;?s@Je(_BuEZANs&Z0n?>oJ4L0%xDG_f0e7DGPC|G z?r(JrIGbr)PxApou06X%50@yF;o_{hK_};1O*@U7u74Mv!;mF)bF33NUomzSJx39q zgM(@B-eFx2Y)3E{`W=mVgX!4%pnjZqrq*I+q&J~t$|#72!tw=Zybi{fGr}YfcqKc| zZYzeJJk!J7wWyAkPc&D4eYB1a;_`i*&Q{$_%B-Z&yx!H>={cCSx;@*Z4JMx>7=$A_ zd}uiwed*}JetEVHlH6tit8)F40;B;D(EjJ82ugtd&s#ngP)+&Y*HAX75d803sp!mD z8mj3ZL3fV8&~06WYd`Dzh(Ync76;%8Vs;n#WWf{eLkwxgb8+o`tpG+8#=854c0NA>!B9cufT$3p>zNX-) z&rrgeiCa!6FlZ6{kLTmMv(#T+rn}Jwc+@ydlc!Te?j+0!2&s^YTeT>fP}aLN=IkGD zNIzm=xSL=4(Oafvt2BE`;No<-Z=g?(&8D|6jb(Sj&6%f46n+ude_yz5qjUv;5Hjv)2Ep3nneW~LO-4k1{E_R7s>)YRu zyKqV#^zkN%dAUT+!$nfBXL$2_A8ufUQ$F?RHdoU&bc+QwiAG+lu@cP9X#tmBE$92a z!55dmqv-kl&;Pw6B7GY2Zonfnfi9RPt zyO%q=uBe>&p3S4eh6Y`aS-1E&F5}WJx&Q;o3zRbck&cuxIs7t5jpceg`)BE_*Y|~o z3?HxWV~#fYL8rgd3;$X78GDhvpQNi!Le-A*5{fB8s=IAhtqu*V6ym2V(a#9X?q0e{ zp=9$$-a0x?p?1b+|v91d0g0kANSkN zgAZ#R#F82eB*Vt%zl=BVCI7wil}IP(82bA7rR!!%T@N>bB4`M)y}kXxn?0lKv+z}# z&yH}3ew`AyI@^!m?{;nr$FKJ%Gs7C!l5Ab~yP13W6d!#01_obWoUcp9Pi&6OwfML# zw7)aGstQEF&$jC}edf>792!wf9LBA_QeAp&{w)!bm!6y$)ymt#^P#7z=+8Vd!h=~c zWZn^Q(AkWn`99xBu3|KUp*Jcg*mrRA2;Ij!7{gGaQ(>wCn&4tPNT3<1X7i(7YLV+wT+9Iq}f zfFSzkl=&YOHU3))!v5l)?v?u4+zUR z)HOj__&vQ~VtX-*daiX(3=3wN?rL&(xw@tXCUQQ{u5$Y}|NnmpIKBUxs#Mq3%2l#9 z6V)}mn4tStYS@IgeSLb&F5Ql1k37X1_c|ZyMg%=q-DnUq{drUTZ#z!oa)hYsa=0Mv z{zVvPqwItMydRP#(D?K&WjQ%HnGyC*N=iLXY3$1b-fEyD`1$xYimL{YzbK81Ri0|( zf5XNR_f3#_oDM10Ha+`o;1rbg22@%2vd ztj$*qN)!=36xcYPaonBQK}rRXnQ=N!SHqe-c6auC!YDY9;5?iO$w8-$m#sT(SN@6# z+|M5h*r~8njEtA)AAyPXo1fWAhro{7(K4#szuur<3D+AW_umboByYY?22`S2smyW| zb!mVZPXxI0w{>GPvn(d%wBmsCu7FYJE*in1MXVOBjfV4`rd{-Rd^(M%WJs=nQ~6YH zbb*UV>RTI|;fIdR-vsA8UrreLl%)9`PuVTtWjNN_&oIG6PX55%PFLe?H-^4FBY5d^ zyfs`@)>`N>=OrUJ?_0&Bl;Zg_j5&k)?m^ZlxmkounhXoeA@zE*k->Q<_fs+2NQNTH zEJEw#%ckBXDvw11=^44C|j`x2XJXNKwi=`nMb@kMKW8 zp5FlP2K~!gTaCbeh3o0*@!5{Ex17xRG;U;u;feTWFl3L|Tx`!@Za+4@S|kq8i8;V{ zv>wp-t&7eJuWS+$5;h)B+g`IFrmCzqQ$6N}E#eKe{-zTXtfero18KX zi_CCL59mnH!L&=!y)iC}8s5kA!5xVEZQ&~c`|0cB`zwr+jzlb1jRl76dxj0DDhKo@4v+yRRWR>xXUqc}$$; zU|LCy?Zou{fhrB2?EvXP78$io>&*2dnQ@cOiz!qh}ShXbux z+SS#UDyoe*j4kuE2+YDORM*!}=4@CYF$A_`N|W{!iE=%5Jhe4Gn>HaPetO%~Kc?}; zfa<`()7L=~#Q&r1fCE+v4u!27x7@oNdLnuBmO^}EB_C3 zZy6S4_q`2Yv?8HMr$Hl~(u%<_lyoB^oze}8f`Zi0Af69oFNbUEkfI4w(vY>9O`Vw z3$>?v_G@d!&cPZPN_A@CDw`}49+a-CzvRq(NPvrjwSXvauxPt^UsY6G+$!-ZID##6 zIAw0o;@4nnYkPnx3!GaS#(nnyYXN@zvjF)+t7Zcm9p>3vwiF?S)jk;8b1*WSyS`Q# z{J-taekj#o%7wN-k>yM(L|Ne zF)XC81!I{a=8%$bPyaI>5t(&glAt6&5AM4iG(42Vq^QH2=Yh zJ(T@SfCylJGL>U~0xbpm6D0Kb|N6vLP*3IG52q$8xD@-}&;K7kynL_o0l1;qV{_`+ zM+I<~8Z>w8Cp$hlN%OZJ;0a)WxsG39Nl!2sy6&B$qm72gus^w@HZ?UhREr?g1+`_q z?^^s`4gN(-`Sp4E!eGv0>|?M$aaF#HtN z;@=cPLibsCKtf^ReA`phvM*_KaRhzvEDEfVpb=3$L5JIm`sg2m4kGm9Gfu4~$--8) zTc{e4gGVprQ9K~IanIOo_%XP%uvhk9EAE1k3N9Y{di3_{O}7yk2h$!#yGRF(vMU3a zz9+z;E0QTi$RaFPIr;PDY1!q7>G`-R4AHV1;bgtim#oA}A-U7MPb`MzzFO%oSYyc} zZSKarFKs!s1?I%lcNTbnYRi6lEMSe2JeaX4rKPDxmq%i!K zd;Fk{r=fwzL@H^L+E{R#tD08G^7khZNK18W+KLXE7Wnvp4hZ8=^v=&DkH5#ewvLVy z-m7VOiKYSi!j^qQXFIKBx~Bfxt7)DfBQVafJCw|90v$LxDF z=V^C(fS%3}2ER`8K-o0zw47B&={LCK%mWpt zUZAib1L4+AD((iqAY+o$5}$Prw(L*Ik?}pu|MCUgj$NkK$^zw!fW3dqeBjf4pWh!i z;uKjcO;fSc1%?i574tFzeCoQK06CO9=~f#KPsS67gF|9bd$% z$riw3+1=0RfRi0Z)!1P+s|MGntun>|{`bK+>DN1Fv7^Kb8-EnFPdiTWPe^BOl+ z>vn;SDDTZ_sq9$G$|#rWn5wv*@8bKVP^+qGo0=ZFuAd zrwf%HTL!R^_c!iF9!uZ?5jChsqc7=Dakp6OIO*^?0FnTf9W4a5zNkpzLmrGU&&1=% z*W3onh7OEjqazZd158%0014n?=f_=g6S}RZ+F)YAv~!$I0WqMkrbfHbeV?4zL)u7CA$H z(Y^(E6=vOvBNm3`L%^IV)s>xrd0gU<8@?MQJyNuIkX=v!1L(SJWgv}bRK$I&W^x?? z=gTq$snYTUg=0xNl@v%Us0Leid~7TmO!)fw(ZZk(2cCS*zkvqROiUOo6JU%8A!p%( z%~5;-CM+Nd{`1s!BpZe`q2GR$W6h%fNw4VyFgcWDp}@kQQjz41+8_Z|pI$WyI9YD| z9mYd>1)oqF>)hjVmBMYvS5p`M=3+NA;G&e9NsxLeaCGvWfF+F&4LSdq6t|zc1(V#W z9%Y4@l@7?8-j~=Zs%e=w(DDs^+-^18ECvnIvXR z(;dJ>s+XU@+c+KKuaENSS3bh(NG+HWCv@XR>Bg?0v36JJj211pJ3qA;N51nKe{=Mf zZ~}Jv<5<%GT|yKy1-G7WIl(InnSAy};QTgVsia7GZa)ij6iV*ni)WB@Tf4oGao^Q) z=K{xjWiX=*J({=i5yU*e5URNV%#*34^-x*L=fIjR7KSA`p$;s&fXBlTEr&H_nA3Vl zz2IM;v5MWFc5*A*=#Z%U-fj9skushtGo^GouDNzBk z$HejOA{ALE1|V7F%;1{AX>H);g*{}Sk{ctySNI~Hu8rna6iV9lq9N(7UXjvTO2 zSeeESKxmnZPp0SVy3^|+rQ9Dh)nv56QXQhR`gD6#MAYxt*&oQrqgcv4sw$--cX=|1 z{iQc@ev|A+XqD4K%jVWgA@8wE(0uj#gm-hdxYFZ^8Ac~2f!?HRI<6LYs+s1}!v%Q0 zOVb>7wPOY^{|-L=Scj8|qD7~t?31VqC657>?V7(HB%cmAy(VD%QR#krur%jyR|)Q0 zf4T69C%fir@e?dXqkN)4~1NXyUskaMgnVOn@R5(f*MQf?3M6ecARcSgn zIMhbwp@8Y*j8Xg}Idu;lJYRN!ApoR`SIu)^e&%iy&R{U}!uDKNQ)*_;HmspWoI=-1 zLx}aHui6&L6+xEcrs43=)J8FXzSmAm8PIy*4nY}&`O?Tpvec(qbHL40g?E5oI3B4N z8aOzVD35A^+pt_+tLEC#@uf$#5BI&K{lje9f0ZM;S$9?CmQ58{DXs$7zp!#6o_`<@ zFIJAmydnvZy8gTCX8=)Z;sZRn-}g=%TBmSFtS}r-LqH{B@W&H-4-|MT@gnmJWx6XSkct&6cG`-a6>02 zH@g9`I?1!^FOOcTglpOveDZl&VY>P;V!bQRb1m-l(d#FGP6fG)QHvjpD3zJ$p-0mr zFW3wvDMBor-0co^>b8#`owjy;cbn#7Iz7#=R}99H_r`)Jf^xqy&9buH^{oiznoeJ` z4|BPc&T!epB~i9SO&2;u`)w{>wox6bCsszbQSg|2N)>}am7E;9RGE~D=`YvR-mfnV zSdR*@P$9u`A7_n0m>4dO-wi)U(QKtpiIQ+_*8Wy$HFS|;wOYB&1^bXqW$3}4U7IvU zuj%2ge}@fOH(hch*5vMD{7~-1lu7bvit~dBD98pdb7sbGF7%uVvrOwqo`o-&OqwKk zT8wS8#t&ZaNS97`*j|s1ePZO>>N*u$+d7Ii#K8n%6zJ{7X{+^UN8kkPC#dnFPM{-p z-+f~?Q%BU3oIBXjMo+#FW3apW%S)Ln+EXqq;f`QIPWx#Y&xFzF@3Gc?5_E67U*p9` zowWJ3bwP;J9ovwo&oBk-5G4=#p{6HXmmeW7*B*E{9e)}T6dlPguPz!t<&OtXQmE^@ zYC9m7cyZRJLtu2^hYyb3J^B!L@Lj02lYiGLwj4pZr}#xdr^7LjHu=Td@~$xPRLNpJ z;gs%GzL~|b_)%MoCx_?y!bP*nZSh@pD2^L&=OK7Sc9!m>7v|9yHI=lvn@x_I^ec%W z_*!vu?^P7&sSqKjIfF;AGvUO19^l>apI=33CaM%{CCoBWN4JcrPc1phDvVrjb+Zi?5sC{J_pG^p5VbS5Hq3Tm zzhDB%qi=yCGEd8`6XC6ABAM|-Y-$RQJZq%hn{<-`t(4BNjg(@Ow*EA-^FLd_Va{_~ zFW&lZE&$GX86h_hcxlh{K?L)hG)J~qZrFrK_TolP%7ZW&rsQ#RcN3dYSv$O+)!9!I zk83`67c5+4T;cr&WxiDh8$o7jbbE-j>S3dgk4oK=V;gUp$RkPbeE;q_zJSDo4#Fa} zGS-?7((_|AGtuo=OeY_}yEtJsFg4zNbclY#QYrtf3y6t=sKyHVqQmM z(UK=kGcXQFzToFa63822gYJkoG5#AWsj`q1{#hp1?O__#&kzUI%@x=C*_t6MVv}o% zw5RBdZw7OH;5Y@?(+DNFX8`?>|25a?F@rLyKp*9^P7B%zc6*MoWwGlfGcO$44o`3x zhyZl?hjZ-=mCv6Wlx!cf=GWF!Aw782q-q`%?oUg%wF@wfz%U58*d>PuhEsHtnS;|Z zB{|q6WT06S4G&ZdVy;Or;Er6yG-Maft^191E3jIz=QVv9&tg$78b0-B^~Py)G&;x2 z_t}j*(EDPV*s7nGf_h`ff5oF|*`sNT`N_Dj zO6+9TTYwR#Wbh5yG)BJc=QOPtbPw?qRC?mZp!zF zDn9T=ns_5E62{BqnWLUEt6ddsEOp7D9Atvsi75w#JBF+Zst*=igXa)UUE@j$f5pfO z6OD3f$iJ`4Xo%Q2sQ+E`{N{7`?VSF%3e!0l;y2u*l!Va1LazS^8$Zm`FA2WgsDjSE zJyKD#oI1Pp+f$#0$awoiFQs9)Wfv^m%%1%LXriG(a8M3b=*8G0#8hmjX%cqg~i zU%dDbAKnE$FQ+PbBTgRy`i%VTEV1OG-_IG_%^J@n=Grs|l%QrZG1ms8X^Ucy%LvqM zUi9*EZoqk#=GN})Wax1Tvy~-5$7gq08!u(lxaHjCw4n|1{OFY^zIAU*)KS>TH{ZuL z&P{XEVx-G^CNZjcy}4=lkmAVklU5U>pG^UgB$9&>+9>4zI0PegP4Swt=tacpvv7-`Z8wyPrWOVWbpPLfrgom4a7yU ziNPo+H8KLW{#@Oy%s{!|BUgy_o<+-bIw;Z+`0$Iv2m}9ewn@Ck2_AnNy=8V!HoM8? zo9iy|-5HzC=#yq8tjt7GdVgm26w~!}^ZdqxjaC;LH&3SHZ#xigmHJTrs>b5N z*;TP&S+M~;136RY8QCd_>^@19a>ODMCe`9ZuY?!E>MVm zjhqnbc=Cou{xJ;h<6fhis8%wL^4DoKR?rds+O5E*2JRFUuI*gEe-D3%ujQlZ)D6;(m?Vt*f9_UK6f(k`be1^ z8FxbU9m_f!Hjk$v+~V)INk`Y_3xEA$3vl=_1SMA83#FpcB4dk3_b2O zw(r?U&rcUKM-IuZf~S5yw~5W0VXy0&RGR13d$uCjH_u%BbQR8FlPJ zZ2v0Gsat9rv+tKb&x7Bj{~&{tIRrESPANk}732FQI$$|lw2Dc-+?Lpy)%o8Z9Ud| z3Pe%W3t?3|BIQCp+&XCj=b9g0R&XAj15+W@~a*t74T*%+U|+y))^{ zMDC47iN_~6eqtJV%^6PzsdEA-7Z<_SWZY=!6OkfWqM6tmr%GpFoT^PZlJ8uavNX!X zhpbCZivY`o@b|^L00)N}+TA>@tt1= z5*LY9J$nb-{;PaU;6ap)UI}mGtgu|BSC|quoN!@1yEPWhQ!+sklD%ll7Cgk=D^MKJ z9ND+CG}q63{)#uA9t;Pzam^h~#;197X7238)&f5fs&p^3|Gan}a_kD*D=EoMH+>Kd z0Kt&%Zk*dpSC;)_TLtyKDv9;$CW#{KWrZh^YPhF=p$9S$2_+EO*c<@0nrul;oOOqD zkoIqV4}UGl%A^_FbP1N{p}U4H5KD$=Vs(zIYde&*d_IkE9)?O3o8c6QX=)C|2)K=g z-tN#mjJyiIh^P6*wq0e`W$kBo-0ogW+*%{0&s-YGu%Qs zLc5;_R#tpy1Am+20_GF>>gY|7q)<`+y1&Mh`F>DO(5M-Pb9)c8o!(Pw3tXcUcOT4v zlIk6{^D`nd2b2>Vov|f*LoZ#Cb+f9Cbz5VaB8>dr=P=N!pVrhn9_XzIohOrKZae(p zcdz+ohbaF|R)UN7lNIo)iV{HwL}<{*)p-IO|H^jMNv}xp5+6e$)xr<;(ub#M5zM>` za4HbJY5SEuGO{B_`f?}g3Psi7_*(f=@7qs?hW(DMYLNU6)*@-lPcIF-^7u`kbxj&> z(W(3Kh1M*(0!vKs94G{g15&9jEbp6CY`p$(sx;OZK)`$OGvpPwZ1*Msil0_l^Np>P%_jwY~woXfp>hR(0K@|HXaoz^R9 z_2c_z%Uk7$EsuUd&;gm13dD9h9Z{MQztKc>s0Wf|_(r3jEK_aIw+dud`(s@KC#W&Tm-;Y){0wn;3TCL(6Xg zpPk_aG$*##1wI>D)LV14?WZt_%}W#VhEVF0c*SXPX0Tf{_P!~Z*1nUUy8G_Bh^u8BaSmG^9gixZHhQkLDsw`#_l=A zz9Hd^HSgVgWQvtqzOW*@{Fu)ntM@v zyIkqj4iaDivi*7&FU$1az#ktF&NherAnu`x8borpmS1CgPNMGcPg*E41lyxx;4R0j z<_v|)&z#Zu7^RE8$GwL-fKlg!0ohcmY@$BPDp3<@XwMhwW^Ymy;!AyuIl(!t!)c3= z*kT4n)#DNlQoX<0J58Eh%nmegGJR1PL9uTfE=B|xyJ%=J$GW>fg~2u{KrG(DEHaTj z&BzE&zn`WW5QX;>fHI8F^6t=)ZRXzU@BIw=u!_q9_-js6;f7d4-LEnxB>zXlOL23n8^ z{T60XhB;~kT@94%8&9>T{bY+Ig@90t_KezYHYr1NXM=0c=M=c$K!n+-(cyDXBSI~X zyd`KDSH7D+&)Zl5AemNJUiQgM_SOB^#+8G&bM5q6WTAJ@9;k4~|30^XK-L?@hqn|T z)>9PTA>?Sh1JSLeM~Pn2^X5X6B4@1zuYpT@Fi9`&w)18-Puu)k;kw!)d~U70+NaF@ ztUE{6|0quNBu#453nvwt3)a@-Q_dKBl3#9;Rj`^zcfKwh zKp=z;E}sA@I7%k2a>=D3DV$&k$WakbeE;5(q9r-JoWi4{g$}wF;peCsv$TLWjs@x3 zLTtQiwR|sGZA{0qR&~pVfziol26(E^s^Qp*b-XD!cYz>2VG?Ij~nC38}1hl`9ICURWc`de-UZza(e<)pIm z&TozM7EK+~Kx6=-wPiL2W|{8x6&BtO=(J}0Kb zz}&-I^U(WGgUf4o-)}>sz4^bvV8n0jb~Gpr|9Vi=LA&`%Sx-oKlAOiC72WU7>`R0X zv3=Js#!Jmb!r2)kjaGgLkj!*xZjrP{xBvR-&UqU|4bqsK2a#4zuWW|-idces<3fml z)r-r1_5g@LjpyH#%1SSklokBBe2GM%X5Bbu6QVP^Cw%_T_{tt%B;jPLCv3*8IMfXA z^F4K};-6}_Fo4agoE4vW4v?>J3cecm?iHsd2ztn6@TLxwBGe1r76xIis8F?|P0N(r zSw|06PAb4Eo(8JF296|H)8*(XAsf-;Z;S2?#QM30YZIjkeM)`dlh*rP+s~*%CO6_I zdmG_RyOMDU8eyxEtTBp!XhKxhf8We*g#>-Kf*msaPD}y!t7xHEL=Vo?KqBGgqfKUN zXsXou>O4q+p#;?JMM^F%PSGO)=BT1CU)m>jwVFWbZ7uGOjptSJx1vir!yerU*7+*s zkA(9a>ameCYtk!W9Q7Hf5C*7{T82SUC0+jt?Yp?k`uBgiG%W4wy9(1saMMSi05zf} zA&ggK1*)tKRuLAU?bE57!4f2!6c@>G>IvNZhckCa#>OWcLT{7a4dA8gnotDqc+BeL z@MrDwO)2-4wfebsT*#1$mm;#`MTewYS-nvyHeg_L=2e)su(RsUi4)K&=PN}WPO9`q zJO;5nxPg@7R^F*}Srsf<|A5@mkmiyRStz)atQy}o!@w^pw?WVD7vldhF>C#M6yoB| zt9Kheb12<_l}$LWJ?p|~M|Wy)v6`5y)>1b3Z-3W=^Ir)6y(If4!c3oCF@nteT&Dx; z(Tl!;cJEZTtmX)kWeSW_@jj5gy${ep&XY(ve47P>yB;yTIgB8XvAd$wqTGO*(9v8> zbdJHvY{WtUkq>bqErPnD(ulNBo{*V7UTx*Y5*87x1vb|iBQWcA88@4jj1f0A2zbX^M%3aU(0 zf)DZzItpy!{BVuqXw&(kfDqO+7YNLQYfkRJ!O2$z-k!1}L(IGSewj~Qr`>*# zI%Rt>@O!hDHnC5OQl~qBkB%!_)m=*XM4IPQx0Fm~qb%WhnSO%v_(fbilrM{YfDt#Z zOp8d_k>mUIIi+P#?agcH>tkSXp`bmm{R^V%nppJ(Rb{q1dcIxHpmZ~=OO<5uu@>Cr z3|dzJ=;tf)vcYh}pm^Q5EPTEx-qCBQFg-eZ+HlTxcCpe^9(d>TKfeo&h9vJr&)7jZ zPWBF^nM=#svL!ZV+H^td55XfI_r%&Ln(zG{_FW+wqh5W1FYkT}%;9tT zqxx#+MC}H&i=x8xLS`We2EMI4{@uzq&Bv)~W*!FsxPxz-yp9e|HE(EJ4T-ORiz@*z zsJ%vpiA+VWhu~_^x;%C;r7J{qku&lb$8kZj!r%VZujFApAQ?^mgRhx|2?<=bd=twG zWNaGLGFSw*7)T%z3V^JMUu?L%T&&@5(y9!c#IE{hM z=wEra4!309Gmay3>BP9fF2+l#(N&g)9pYvI%a~yIpt5cNl@PJ~% zKV#?q-qdAerEq;&cUN|PO_N887VIJlVebgt;Xb7=k{RQlapKMOZcR$sj|OP+mwmn@ z4C?7Hqz%HcLAzP+r5G)@zt2Q2KCrpson0= z2OSvsGrKp-jMM(!j0ctE?6}{otK(Z!`!0+UiVp=4C?K_YqUih?k3r$Cl5Ac--otJ> zs>*~e*vN)mGoXn$AW&I8St+9`x?Q2<<~wkl}atC0wO6I5b9r!~Z$s!<6bfH0@;MOR^uIEYs5=J|%B((KGz56ro1DZ$6v`3Hzq59 zXE$_RppV~MI2G+M>$F}oK7UsZ2kN=vWCfjWwwG*guHc*GfqqILem&p;*W6$8XcY0i z#Qy%@edBMD)enI1!_yZ1|&Jke*7HAtNf+mkAeM2{ULT<`4{K~g|k zhoNnxniJ5u`kTX4mn|hlg}s0%w_MsHgu7nbrp^t#>o@{7G;Pd!q>>4 zS*mKIT%piv((GVFy;Xac7q(%btfK!EEe>=qMi42Q-D~b2pSo3V8*z1&Sj)r+J(}>| zX2S~tXQM$AVmbB-Jwro_b5!M>Xc13VwRdH!r^+JnZvt>dx$v={8Q228Ks-mWE2WmQ zx{j_$upO;%kdjcG9N5x64i4c3`pNSJxxvPv!=5$svjpSm9---m53T~n!&zFLotLmo z#T7kRH2UZOq9fRd)C~^D+8`(Y23UXvc-S>2W zUkkIE_ej(cHT3^1ovYQm+N(Xd9Q`E3t>BY=12a83du2#eob7fB8-n_6vVk}bw>6_3$4MIXB?I`Dx-t%zW zFm~7_9;X^`PK?@`c?z+U1{u~2AtGNc1e}3(nCsN%a9#S7qq@YK;a{^C*FnhA*7zEn8Yhm%s)B5De|n~=Q^b|y zB}iw&;;Vn47>DfXwlGu1V4mCQk7dNxt0EZ%nnKTK)*)sDV+@AHXXN*Y*T2Ow?NM&c z6-eccPH%~=a{AJRdn$ar?>yS=mya56xrs)mV=j z*p{G;;rAi99qwtKFkv7g$VK3ziJf(9Cd%zw=PF4NgSe{ zQ)0l1sIRb#oW^Yy&ST(wr9Q*gfJBU6-)p>&H|>oN<&;`4$CZ?nVDV{EZnCisZB!Sy z?B9V&t!Dr19OS!!Si}B_-}E6HAd^q+hyVO_N%W}E_wRkkY`2e-srQh%V1j3BBaZA* z&ca~i#!6LKV2iV4+MZ7daRxZBp?6Q`Amc;(HGEt^II)j9JD*M<^5}9!c5vT@jLAyp znQln-+zgT?4X{y5)!wYk?IVL)v8I}r;;9ZeP|!Tq zARD_UwBu^+e7@`)TXK}E8OzI6Wxp9T(y31@FAivhwO!G8-sBF#w)v?yuDU#M6}G8O ztE+rt05p=VRv}gUr|jd6QwVuG1q2=>Y9$tU+o;j{u$K8VL2-A0jjW{LvUzCE=SA0D zDK?dP`&L8oAnjOv7f`v-HvnuSz? zUYz9C21uoF_@MVXe2{Bg19^=>LMYYH-nL?Up{}jWDe{A5Yzriy$4C|m@+2cK+DJx4 z&i*nU^ogkaN7C8#T5{9rTAnZN50RZVJ2x~YD|o$bOc}&cCHH+>j4L^T*}iyKlZM6< z?`JSwNDOkzr9pz9EP8M?wo74GF>q<{aotS65#z0l!Nj>uDO1V0OH9%DZY257d_W7N z!C9CaAn>DGG03aVj7CsE9?1Rs;#9>#Lt+ zM>zJ<%sWSd-j5;0%0f?ZmLEm)7FBvvKLApa@C@|(Vj4rMYX=N&)1)qz7S_qA+HjU= zJd@8=piicWoj-hy$-;%==(&N-Hb62AU{Y~6DfSf2qTFtAZ*7PiED3=#T6PX#gUeIF z*kgae6Hoq|3jioBGQdxQQHMKT!_LkQ=uizHgBRowv2J*_23`h$Y^D}4_#ppv{9iUx zSfQ}lTA<}*evzY%o1)N_7tfym+ia97if~U@EOqsInc7Fvp1}bp-J9T4b0A&N>8TGO z@*n}}KjIE0v3{@Lm*IRAs-ACy92}4Vsln|7k6RZOcBqtr*Reh&j1t?*JK|FsOrp2x zaNIa_?L{_j#*syk@T3{o7Y&>;RGq))4D@ zuS4z#kQ8aRclKReKjJ^Io1My^wcLe!?+VvWqqd{=VH1ZiI z-vwX_S}^6$Wb;H=g=H4G6v+9nGM->VSKOOu=d}+x4Ij5@56>Rl7(ZccooPgZQ2(l~ z`VG9yY5-aaXRAq>{pZJhCm&<@l+-R*39jybg>;MXN-J{W8INj!``**Zr1w|X)BP@`g8gn#mXZ-mwFCT&YirLB^nY<}`ECLh0_qf7Zh!Pei0i?rqI4 zdMFMT`;3&zfg*>J366mPpPkPnB>St}ZXmk|Gr%78FZ>Z)2!c?hCk>>!PeWN_i3vz5 zV}8vxpN+Cd(I1t|;X`MD*c7VY^H#_3#glX55gkqWIs?%y+kX)>Ab4Z zotD}ztNl+G#rFNujP3LcJ33s*1B4Q_vQ`t7q-I>c3L-t9@vZwI_(uagL0o2XLI4!O z%O?a7fW7pGg-lio2FqDM z%nfzZXdVZg0|4=~TTL!j`KH9>3*7cg0>#v~jkgVH)uk;Kyr3X_esv3Wz8K)!>Vh4) zlo>SRlFK>7ol>Qkd>k8gegG>ESvX8k*I7E%aI^LYQ`iL=ub*QZd9;t0YhtLPOxy!-pnfi(QE&d zA6u)vG0q7f*X<|oT;inz^)|*6uBZT0mX8lIMkYeN8H#SSevXSEgGVYWOS&VU@}zz` zB)cL%j$IJ1KlJ)fhg{Z}bnjGp%gev6?>FLa0m{NlL9AvQ*EH}p3uJkKfHGBrkx|3v zw+)OVf11C|HLVcE)jCN)h1Z>8<=@TFxUZd1d5<^UW+1|&{Q6$K^u-GlVbnyeddD?L z{=TnYf!d6)zw$Ys4K2_7Gs2Ki7nVnTAX##g_ka>dzzcMPY8!1KTf?=$^TEt5qmG>au!vWcR5Iu{*5u^NY>^==QK#RhqD8{ijkW4Snz`ro4QByq#flF- zkFCB9rR}@oY=8jZk}Z`*{shlEV6|3ML(@Si9b z_YZt5>vR?~U(%?Ph7h>OoULeJ2M(5s952x~6k;f%1DmSnmze8T`v_t?ouwQgA8w$x zWPWoin#TqCT2Q3o2TBKLx)9K&U14Rzx4S#2xVeA4Z~hhxG{A#lkDgH+ zN)V!xV~)xLVmZ~QNYh-cuk`qU0EtQHgYqxqLrbE`>&CFVXk<9I=)GH8_VEZVbbE09?fP_kw+fBYISq9O>^PC+B@NmhU<<6IU^cpJ5f%sy0M~c9rVOr zK-!l@Zhf{u-hyraZP$#Q{qGnjAEj0Qv<*0K z)e3#ucP74~Btt3jxkIA1`VJQpv*6$*I=(*OH_g6_wGip^*W5`UT3H)9#HA5KeeN2WSFke5m+{Jm91IBGgWF)1E^9x40q=uX(9{K6{?t)ZPG zio3HpAg{SR5qcg;$gxS+WUpiOCLl+i2Mtr==*lsAw$9*z`O%Ueup81{o=A|{F~U}8$ZU+GkSX)Xe*|Fb<19HerP5PNP)9)N!#8yrSC5`+ zpDmPWac7kgfo5z?I1JN`h62*j(CdTxG19|cA;pde~8-p$w40-#QqsVC8c zJ)y892*&=^PPfjt^#0F{p(~!U5OhWJ@kpK>VG%D}fRc90%)LM2qmzAvjl@$)!nSX} z-s4sW&CI_qP#4oLE=^jBZO3IzF6cmx9|lmt>Kw!0H3A%hno9|85hG7(U<5Una|Cc4 z?%WWJZ$`xN>Yesj{?>r5f$cT46vFN>S5;(y=zcwEP_@3~f{aj|l$ofxdq$y531lY%~+a)J6Df|S|1x~K6x+C*CF$ZMavs&dk8 zUq8yS=LQbE?RS&wQ>}CTo5h_nRr@5?9M5#u3+5L&<8t^N!c0#IAOSe=zwBZub#6zc zw}ea*NL~rb;~Az97qV^dgxh){Ty8@UHK;E>F2rb>*VrZ3-YW_6FA?C7_IN74YhqQQ=AEkvNBnz8f2vhy36~uB6CqUnJ#`b`^}_I9p(wu>QIwKg4x~_tT*Yi!ek(v#fk*f#9{9)5pFmW-5srq=U_e;*eTc6-9sYEa+gFf<(U zY6r`h7s+bYT%c-Q7AOueI;@^Yu*b#d;&NJ#}QKDQR5kY`#W0Qw{#m(N}LRH`* zdqfKNe5xs_6EC<@v@nnC1m)g2`wYW0>O|RgBKi4i+orkxQ}XScH$-i-0MMrVsC}Tz zpr%vTa9=}rQ-~jOelWM28D_mDYjdeyzkvf~##8o=B#IHx%~9F722-`Y8`s$laDW?~ zP&m;-t{~oHeW%WWT&NL#6Thd;?|#CVX36go#naPH^+Y^K9`5ZHs1*t_$SRU}+%@7L zBMnJzwxZx5qE3goI;=PT_XLWr%E$?3A(bP$SWRdp?~uJ)#!1o2brxW;pGs6pBmJqt zAGgCqcsqio^`2>V`1U3wv4=OiLVe7@Yr$qKWC~q6q)X>a$essCcaa#yW_N#fJM!QF zQw;_(xiPjD+BR?71>uER>QSsaw)_B00{-^G&2cZx z()^8BKXI^Mc<(S@Jxv4GYaY1Okb-2Eq^7f@W6ArLe1bP`aJAb|CB?pcq@-;=DjZw8 zjkB$2)YSAE{o!Cv;~?#MGZG`Zw+TS5*r^a49MLAxvur_ga2! z*paD;Al_?GPMcE`mwqa8v=mrKgz=!@xhKxO=Jfh8IqrC#5ltfhb97Wh0Y~R|Q3&BM zSi=0RviN>0A*d=U%)IaVNv^k&8e?_*!^D&$AG}Ig=uVEGjR*&6B;3<6K470ng3p9& zfOlW_%yn^9B2ZJV&gU-)4D@YNN;Lj0oPt7F>-DL#eE|=98hs+M1-fPK8FDqWtKEb; zDEfZn7Z9qy-|8q^_EmZReERP7h$q-mo|o@*^Ys#1PM-YnyT5-Yh@e8+f}$C5u&kTC z@Hgv9R|gL`$sMOy&U>hI(!#&6syZL9$j~s!7UZcYNo0h%I9y;eQSe+0s(w= zIXHs-D^D{1&d%k}dTEYXQJgEf4Mw0|>W|`9*%-xqBR;zqFFr>W**c)BrxT61<3+^8 zsP}ABbiw5V6?$V!OE2dx4as*fZq8GbY~N+LG%r614N+^{(Ckj_hjUbpc-F=p^olzh zeJTT2B6WAeKzW_=#F02qlSJAPL2=?2@N9s;ak4@FJt zdOmHQtCaIuqd@B@!>r+sKL(RY?|2{S`I?|cnhldC7InESYADyl%^$V#&6|fh{FC7d zxZX)wj(sAzwV;-S=FITpqdWVT^ZNm8L5*B-Ili6*%jEE>b;5gLs3VjoZ7m_x#)fnj z`HWPsY=V#F_|WKC+n6dP#oUvWe1Wbww+|zB+lYE%&0+9JOwN%WfILPz!byua%}fI8or2 z`T6>>xxVm%tLN5dN{}aZZQ5;QG~t2kU6J2OW;qbG2|GRIUqPn|!(QfGI*Qj|T+ z3+GcfSR>E3r^7=<}xXf;EGt@?f{s;@G z3Gk`=9x*s%7O z(+1yH9Qj#a(sh_R$Gdu1N~A6~SMl@Ic@(BrM(*{C?~1$v;GZJz!?Mq8QNjn;@8M+` zXzVjLhnfLJPZ6<=qSDN=XZGw;wz~^qfKM4oImWV<8VM#m(ii(dc`dhH;bQAbHX&bd zv(`VtRNZ^g*Ph;4AJotwhpe4$K87D{lR5hE<=n#2(~`M5dSrkwk)|I`zb%gSzO?RA z^x32Tpln5P<_&Ns+Esst4<|9SH|5p}$Ivaw~? z)BgQ3rwFvL6AP*5Gl>4-90dTuiw8W_4oz7Tqouigv#(l47e~x}`^{$%jdZ1iB5&v4 z!u}|YAf*`3J7yRk|B#u8UzA|=Vo(Hc3h#I^1&0hw8t%+Bn6EgCLg7vT-6o6fbmx(s zO77iVG)Vh=1;^U6+8Uy7lbOkRi1}pH#LFmO|2M^h%Zfca)HAdRAkA-;Epnd~?ye9T z^?gaBy-qzz=HZvXGa_C)98M$8#N!nj60(jdvFdO)2HPE`!IwCCeC(CKVdDudQRsLP zv(Qcu$1}23W+md@nv=M0oSAjNo*nba+i%0}@%e#E?bmTQd+amWjLGlcA5Fn+?2?r& z8&wphc4zI!JN=>c#Yo@`0DAppkP=5&nHw>6bodxE&%&(h)XaET3BApCygIMJHOq@z zbG`^@#Ts~Tywd;b?JdKi47+gA7ZFiG5ETgtrIc=vMtVr;4(aZ0Q4x?Fy1NCGP5~vP zVPHVIq@@{RsPpiBd!N03?DJ#)Idfg3<8}GQQ!DOuueCnS|B`H;+1=O+IrE9|QOV(R z|8>ibdRZ>Rq~sqje*KhNFCrq{7Q2FIkZ`ed+hUrj6=%vDLb)w`%|&&~|I(jGXnaL{ zTWn>rplrU_>2dB{tx$o;<@fKvw|OoBrb|piAhiB77|Wv8cz-Xt$wQJA*qTsvR+9&? zJTmI}cyHaEsLbO+X*4@{sQrpvxa`n-k0bL*P34ZbQe}k zHQQQPAcTp0C@Q{H<`p$p_p+e3vlTQ#NO(Kae~gQZd52W4Kb`|A*Bg_@oqI_ccnlaW zcYEY62~ZXTx@?Qd7&Z{8CIgjZ>F#KYcqC8yUj@+5z^;Pq#C72-{bWRgO4tJ?(O=pb zxD@kOQt}Ot)SE>G08QxQ-ygF;Ne`IA6l|)RJXBs_fU8c%GF9|DJ(SOqsybkx=%-aXD|5a6Sc5)zAi&$^x;aGwUWo(B`J| zi!?PUaYa`tfy)l&@E?RSXR|IGT~$jl*{#T#l@6SPr`p?&HTESh0t?1;59eAecopt} zo`@@kHYFZDlXiEwB36nQ_BHD|01sm7E_fgT`OuX2{;OXMf#wq=@p|2Y{tfe^@2eE+ zaezPbo@{6eeWliHEtlR=DmwaGBHz=^MLS@8zBS_SRF#*`5A)4>a6N)!FeFSZ`-V5I+2y_RWAT639%1}#3Dq;-+vwG6u zF{;RWS+w-J-;VZ#MQ%)9$4}$rseF17o`jZ)jEMH-3cCnK&76lAmLM7Sy`;0l!1DrH zTun@!pW*Lpnv6$vb-)TrolN}W*~<1NvedDfpUI2KYX=Vac3v3Ke=l=^00EV%vC zEe~(7LR3!XT&15o>g0id_sZ*hMbPtu>$L$VZW1^`qq^Fvo#zp}OQZvUXI0H{GdcAh zC;irU!TOiw0)tXOpn>!b#8vqD7|ozwm*d(=d9c9M6>3^7=PZIS;b3L%*CIr z?`Gxa`_XG`YkOLDjC_&4mxcjFRMZNI^A5`LD-wZ z{-m0bV#yd{3JYg-Q`_ANUuE{&;hsAcEd{JQrf*^9_5%j8D3)Oj@5cqZwKPIbZ4*80 ziZ!50ch!!T!kZx0GsWLHx4ym7$X8GfGA%i+!}SK{P+QPK$zv^}iZR>1V{&@(mVO+l z`pE>(v>t7-Rsj(!cr+4EbE_Z(-}2cHM&=57G4wPFR}WJMl*2K4KWk7$>jQ%un2QRO zY&ZGM?-WTLl7SFigo!0$97lRUg1Z~ePcH*a^&(8)*0fsuMy3mZf2NRwi!rvJxgRa5 zebZCz#Lj9tRdoC=Xi)8nuNsB{J@`9vdF9$^V8LX`d^=0%Pl79w4|4)tmG?Q&yi0&s zOLX2l7fj@vZ@AS}X58w{9=1^onoEHIhBA~yeDG^`lQSnrBjYMAv8s=~bYEhUu5o&z00cr*H$A3_CUZl$_PdIz2!7~sTe-P_w8vv>Nd`gf_912?)&k64_zCmaBQe7yHd(4na_hd-maO2hF&<-Oe`Ih zOp~><@U<{=B*2)RldoR3+4?3P<3cQL9{Ggy8-GH)u6Va%z4sD>Z3*vYNN#zue;-?*X2DpNWZ7KLQpai~w^LjxM1%O%avyjKe0lW@aTuni1V~AJJA{7U0XBt#t8uyvZM1hcGd*jqr7973jB^xXfeWzmcU)0d=%Dhqm`Mx_a?{BOnsLL5atQ4%8 z;5Nc~J0HEtAU-VjF&fZ*Eg>Hu(uR{{z)MYx?Y7g%$_Bpt&d!c@08dc*m;cEHxKz_+ zD+#Sg+iG+{Pnyv6RRYxfNkj_AY*?N&-ffau`zcPSJ}~Jevl8qwclB((Ic}VDx9Am; zsu1LHWJ?6SaMlED_udnFt%gTErUyY)EMc8X&8;uN*3AGeexr%>w>mpX{*F$_(i6<) z4s01~!*YB?+$R@}Nkh5xN7~TS;O|af{DyOL9>8D1W@Nfk7<`+jy07I|-i5!BP z95MJt^)g^hW23dDI485Qq+!S|Ew>`vOnVdt^N-XlFbMdPmlx>unqzbQ$83T3L?r-R z*aVV`;zDFhyf~Ye1O#%?>6A%#0W83hqsFlDZ?dZD3Jprc#LLn~L`IG~sh9Ote%+|W z?KJ*w6@0;+a}y&t^a53ltST)xvi0@h@=FDb?i&-gMEg18o4ZR3tMG-CM=)Do=Y5=L zhP7m5P`JpanPSiS{?F~Z;@oRa(#A_!YgbxYDmQ7X(V(}}9FQqbx8ij;Z7zNyVniW) zqNCSotqgO}J4kn&z|8XcTGH^sb*r{s%GuW87Bn_AB_6YpOkibr6oAdf*r(bap1iEB zn>jDDtCJPu*2u0nR9aC z#R34%gghI8{z!FLqlKngMS9kI-r*u(Zn(THk;$BTIMbk)kaz!9+f13flABoE&L0Bx zfCM|h{Qu-(IyS=CtsvG_r2t9?^=tHVPBg0u|6t)0`g}*!; zu(vpc)e~BRT7~bm(f@3m(g(0G8TfM&0kJ?YXsv(d;J8>5&cM5CWDJRZIJ+W{e%al> zVn9VD+0zXr6-r3xs6^=Y5pFEcE0w&(^H6}yfg#lYv-G8)U)_+YE3;f9!{dV>sQP0* zj9$gAwdpdh^z*Uv+Fj@y+s-7#k;g88ciISbV&E5>@A4h~Anm64zCjTnmg`8;0MA$Y zir@P$fmY#KAF5ZWHJfUYKfCAxYkTAm-Y&V_X=CF9-Q#5~6=@~_pIUm?((2C=2)3S~ zlJFwfUJ+Y90fXv0MHJNIX)5U{4{;@92Yy13#Ht`kUIVr8qnXcX1ElIXl3X{=s(f$s zYveSCH$#s3DnG{bTi&U?(nU8V!gGER4-pAd>#{hdx0dRJ3GDIpDfN9k#D`HWpj_fhX`BpuLXDCRlO-XJU@h&Q?FU z62Y51bs1on;viZLI8f_3v z?%=K2`q@)6cLja}rY6fmDCX4>mk$sL>DA`YTSG8u@%J(UHc<7EW+Zw5!fgTga)z*J zB}*->n^VO@(R4laQR9E@KZ;q$= zReNaRKGgP4Xl$+Yt$6_`-A^P?@k+^lmSvmqU+F3 zmUDTcmo)Sg#?RoUwmiO{lRwnli*ffZ<~;*ZL|1Hv@TKhVmv`zqvMMO(xrAosXvt!R z{!Li}y+8Y3%hsZj7LecN!8X=nD&;i)mTP(^!Qu7b!)Hd{V-_T*w}$x93!L{ zyz$a4Z2%}J;>%yQnC8XUL5)x+f5CbmDrq!+L#R=1xq}`k|Hr7iL~67^`pwd_E}L1M#?M+@Wh*k- zz^v8`-bAr`7kbcDlW$|~{&c31yFhc@<_;7cYSmiBSp?4=AP!{9zc4zP1Tc$*!D7~J zrl$yPOS!rhg3*?y5AP)>+-02(XdzOysl)`0<>j!vsOF>;duB|?(oh#$X#2IDth-;g zd=+*eTbUwulX0ft?>26u?~ZOEwTHm@$@F8ckFDomc6*46Y3ax~+8$8^Jk6Lb7%P=KMKlaztOfZ=~5MH?pnfFxYZSj8y3n7Y> zay@qs)&wmKH;1Nle179LqyW+`Yjftk&Aj;n>01vP&k>&E2LkBa!qHI=d`@L6 z^;JX#MxfME69UQIUB+rFZ`t>J)D_AC-Wz9uvm2wdrkHbmCFbyRCOI9y!M2)ZZ9ElG zp?y4YU1H)oR7N2O=|C!?PNGBY=l6&QqE#DV*y=i+l}C$fh@)k!KpswwZ?4xFSwg>5 zR*5kz*)~$%+r&!BUGFjITD^Q4W@S|*t z(amQZ$qy%D?Empx-wrf4(*{LDBd?}-XUFG%)*E|niCW*z8obUB1P9vjr# z+zR|OMuNjo#{j=t`V#gQWNUZc4 zb-7T=+#NgdZ{g6VACYJ3Q{f_2=jz1}kWU~?6M06T(G+L=o%r}spyd9~m=JVue?)VU zoG(rmAB`ZOJCG8`0*`)q3*NZgir1I__n{V>o7~>G0U1}2l&+G1&e!>N)$_boZ#vH( zRD|+r*nt`*Jx9y4d`4)F9YNa5p4bHYE3Cualm}~DT?nSuD0A(Xjak&{}m{X?pC0Z`556Nc`!$r6UE7WISRmgidOXY9!|Uf!4cNE{0&d>IL3tTqc|?`TZ9Wt6(BSFjTnHQ4zs>v(oEtxz zQGOw-t2~qP$;y7OTcQFRk#a5E;uBXFXJf7vQjqB3l;DEZFDIABzpCG@w&-dx1eovH(6Q9{$-$AM7;US zvSErR&~QHp*Q@Qaz{0F~A%(u0{8B{wA**e;%{3d=y|(zJFnonFs*PCImni4fV#{yL zu`>~l{?Kp%Q{T}ts=a%98L&KSo&7U_BaPzGHx)83b-Keke_aF4X)G+g1`D7is`9xQC^3${RMg z6Q)1ol%Pckg5KE6DO~%RMV?4$rHJE!9!KJ^*lS zeWE6?BXOjKWtiLtMbMF7<>f**0~fMkx|IIH`C@z^d^-U0ky2!{Y5evkyn~GOC`{FB^$+5KetRe|O5ff18hNHQ{Kj96HfC)Q{0VMi`b5ye zPl`z*_4#N%DQW#;fu}dhCXy;LRK3vGM_NAWUv3#O8&O`@FxSo`A+mLcU3IKr=RlHP z6f;o!JClnllg{>_qxH}6rtv&CbNE}(5Ug)+-Q9H{MG02wsJ z{~Jq$O;-<@P2cSPb5x}H5m$^}GEa*K22{272N9uf{;26cuKW(SRm5r!86=u}dK{Ug z7@3UeMn&Z&A#U&xVuAy>JV_H+(ru$wsH$b$BtoDxJTJ9%DZP2}jcZ1{Jp_}g??Fka zA=I;f$gArn-MWyJ!s+9tF*K~uM?1%#Azwbf!X zO4>tEcqR1*u;L`E7H}ESLion z^nylc16{t5)Cja1r9aEfUU0(@ZJP4$MT+|&y&fl2bGRjg@>E|^=A1MlR1GQjjtV=9 z+h~D(_X}VJ@=-gy5$=V-C0-Yux7vbW=c|AC-RA!#|03`^>dd0WgN&R&Xs(clM!i_r zFd@#@90$YFgk7nzynuM|Q!e;VSW9XP#xjZ7FIt>uMlB2#Xc1Xi@7{bloeX$|i{TB7)X*(K+u(`>;BJmI{suyUA*%0jUVuTm0A>F< z6CZC%$f=l@VkZUgb=Bc+Nv?K_+D-b)YbztZt^qHRvytchM`{Qp2yw#pe z(tK?GDY%*TfnR{4p~LiXn!yJgo5alJuE&u(^Qz=u>Fz&+7L2`1oHnzq)6l0FhM&T6 zhb~(mhrO;krzqd5?QF=WAnQ)vPs&`r&_do$+%kAgZlnL^R`_jJ(ymLN9Dp<3b*$Ki zH)~UXyk4B;8JgW#Y+u5@gsVk6YO>f7ciU-@OOi5bL)IO$=H;fNe4aEk^ked`9HP6G zQV=Xuj$S%On`Cfu$9~Sw34z{qQ z%vXWs=H>#h1u#6E8@}4|=yt2uu=G`e`U~_u7ypJbgq|IQZEne>8gB~NvF!P3>dZ}@ z*+NX4H(zd_j(SSn^+| zxnMf`MkC88k2IWz?BuBY?c5bVu;I&`jM1Ifj5>K4gaTy z8OXO0BFpQZY;6GBeCB+V8fE6{`7y)T5~PI^lmHWa9r?K6V&;4PRufHXGX}&J9$Dn- z=kZp%)@*omS0p((;}j2wsMI=|FQY9xhf-*EBS&pD zq|X;OakIUa0xPC#hh|b%UC8bLe0LwjF2^3!10VfzaGB(R;zKEb&om1!Y&@F|Iz8Ep z%tGTrY#vTZ-Yy=|a>OassHP~3>c!XMZNmoj9OAEl4)n9tXH&D2>{ScJfJjw)6Yc-8 zH!;&#&40`F5KK@GNO*e-GB*Jh3N>q%*B_qE!d_8wD9gS)CxZuq)g?W(*Z^r&Xb{*ANVG{fFg$`tvX$qeblm65|Fg z0^gmjXZcaWAofo-*BI!1bGz?{aryWn{G z)bt(L!7b5jrpulxdh&{mCLeX_-(PM8@>DWVmwnyvxN5W41Qte_Sh8J*r4Ws$vIhXD z{C9mK>F%$3btP)BsojEff-K}5HhEgSXgO)G^ldn%#Nb4t&q`&`W5IcVFNzl#Nvl60 zX~O|&+4jK;xCi43@gOs<`k zJf4iN8Z5wMQDXswO=mk!Pi6y!gGfTA77GO_T`*uvP8&S*jUCt(1uRzy4@mTNqcOsb z8#XtAEqK`;OY%Soo+WAnbZ%pVpN~BLakAw&oG(eG0fNfK zu&!=HbPU2j1cc+6!dzmTbG_tty%}<6Er5Qrw9~z#=eLl2BhUnw3-nk@@#ftXL914~R9?$`73Bs!JU5|j(>|D7 zt9M0NPmb4IYSQEw88nOzGqW19Qo#T7FAR2aby8`+`NPw9;WggczdH-B2alFjik~dW zngi}TXqLrvdXk=&rpA++j+*cQ!5TFND_2+7_6YS&6!n5;cLU;iZ>m;mvFxGg4PDZA z{{UucISYX584~-_Pj+K_@^VJbP;Mo2?2V19BxyP}XBcg^;4JeMRrg>A&PKfZkHeo7K1H1%amhS&YZ^K8g*wf+eT zfUa-UW{Q*&5*0;;NKi36xlj7+8P2Y#ed8-@o=WkM8wWu1r{tft^N@r zEX_#%uP|TcI;0-fyDQ?@dEZQq(wDBx?!zjy4yM@BO#okYSVm9oL z&+DlG4zxB@c%;(Gx2K)3!S{68oCwMa-j#7EfgMl)gfHC&LE7^j3D^rDGoUqquUN@J z#UQ@Xm*C4!L=vee12Ge&B07p5|6d9S5Z8|=L7`LZfX9OT8Qo_BqC83(A=Es@Ny z3AdE{p#H@mp_>KW8~02n)Q=jH2*R5a<@<|Ich`?07m%+mGkliluFSGXAuf?2p0dtI zKwtxB<2++IK7 z!qc8e$Wm`wsR3Q{S)h8+^KDNJ2ebZcyXZT}l2(=ZW#Rpa`?MP$hVm@_3I|F7a-@p} zQ)=(*?sge$$6ZGM`$OOhSdxZ@xQctc-9MhXr*z{IpQ+Su-3oM3r%8i`)$OnV|*u(m^e9`ttyNv`4{ z*a5(#6h#i>M>MMO4qNBc-fO?ckJ_l9yVHQZ@i|j&W7j}dkuo(W<7m~uy$nONw5s&* z6Mt#URwmH#7=W-Je z_AZv=KmdeC@6K%eESWi57OqFLP#Y|MJ|N7&d zX6wGCbo5M3-d3Z`&MoMvltsUqwEO@)z(yxBg)$5DF#{!$ynokuDYl1i@=Ruv7+dla z|FS^UP4QM%>inKpe=!Vqy%@?y1Y>zdK+t6geq#(@`_tqu19er-JUgZ^&knWy1ZHFU z3)L3~qo1kJ3sqJta<>Pd%~Gud`$^K#3Za=|eTyK#5PwA1@XTb08WKey{`PyK;+Hf0 z&-WuYeD6J^ySnf620)Sx6OV4LS9mh8*yl0uHdKP6{h)2splj zE=*hN9+~UI0A=)M%k$e{7J~17tY?En`?kW9c|>dC30hAiq%7#}X+unDdOZjt-Y%2~ zCEipt%shHf*C#(NAe0EmJDsjX`^}hUYyZDvAIm%hgMt
f2G!oS-jaw=5k{N zb38Wi~C)AD<8@ycfD18$aZM1x8Jf~;CAz9<0JLuE`-aGn!2X@+*dvLLOoZG{W z27EgkUs4}U@xMNFF65*ulf@PQa#HJ+zEhQe>3nn?qwLvDs8q>%mR*VR+(X+JwtwHjS{uITy5#rOD(C^ ztAu+k*V0|_St&%8J^-a^^m6f)+G6kNXaErchDV5>5M=ja73Q$P{^_20?5c|m{uzKg zH`{~|F*rdYcXQe`v%f#xDg@|xLhSu4^{M{r>@CI3033w5&r@W?+ zwW=WT#B<7Rs$~Hpo#w#Uf#*G0t^;IqF?JdcDWf8kd1d;J+WAi+-of?`Q=+B6$TK}u@m0uq3s+Qpy-8Vv%ye`0>=<8Emc}0B=YPFa^0x8RhPjVS zt#L_3I*z^_Fa#~M3!=<5<@tTbagAd0e!S2OIIxv1Rdu=?;#JxtDjzE5l>KyjSoTKW ziW;hcl#qDleq382UBJfnew>p+2|eVWkDnW}344T=R^WE1HQE?`_NdXvLXGlt#OX4z z)a4&XJ0W%Z^z*L!2_RXGoV>=(x1Yv2Dr%CSef1^*RQ2&O7Naw3mrI^&NA8vft&)fUz2HP7uf{THwYlFN9WNXC5OOpJzwY< zReRHKKrEVa=^B}Pr+FSsP7X3IhC^~WV-JVgylbDh4M^Hk8DDK2rUJfk%wW>Jl+See zvD#M6jStGTcv|=-$T=UZY0}7{7ilG1YMP3k34CefQANN#lZ{9%MPGK?3c385h{dM% zzRdG5`2K;G{xLVQ(Mlwo@D{}8^Mi4Am^tEaUarM};5`_iVjw@42n_o#m&Kw#ZM}>S zV-qBH5hwi?DRr1on|@I3HjI4sq1-*q-Obu~CsePG1CxrX_K6eiJUe!92xBK%i3611ZSQ78Jn}JPDWislX7$${ zEy-Es!-hdiZI(5?EZ}6hzHEqp8xyJI`ye$q%2IfF7Vn!|RmHI-WD5;~R?;(nj3YV#eK3nb`0}s;j z>el;iLq>13Id~bHjLos9Lcek;&%RG!)dq%Nh4jg}>^WhJpSppAZu*20%Mj!8V~3@m zGJ6K^eR~iEvE>GvQS15sN=+=o_+JxfM=6;WlN0WL-xLvAn|TQ8b}`-V<4}L3I$bA6 zIQrvVvr*8xo^o%zYtOrURR)f-gqt13<9WpwAX)jf&Xov-Ig52U4A1e;Dl96Jdx$s2 zpxGZ?jp`OG+l{gW`6)jobxfqmM!oZ?8EE~M%plRNV(_K|avx0S7Ilk*qk%h*zk7(< z10Zc4uLlt-R}KeVC-(*2o)MHl*PBN9BJ}odn%WI6UJZcac@b*{H<$b`t?^Wmap6f( zo)TeAxPE}xARkGaPqmt>kx#WES770gZTsj)=RyD4r^>UGUPC|BKqebv-4~YRiz#g6IRJ&R) z3A5tl;dn%yAI z8Ee3DK<$+r2rk)?939n^smzqEcPA1Zsg@piPnR?&(6Ol0-sNggeOzES$=EGq2Usz` z*E>HLPw?!!bn%Ajqj7IRQ|Tr!A*Vy-5*g}PbGGTpnX{(BLHQ{_P98yDEx07)nZZV7 z|82U$N>ymfMs=7{c`-kAADI8> zXlUwVfCDn>O-4Y&^~E-ua$OGXeq};O-)1@9)15Glxc$D~Q$pI7utx@xP8C%1?3C>D zDtuIYwQ8?qK#;zcR=ebPWxp+b-%=VyL9No>s_cITwvksF@lvS8A2I5=zZ3vxpSYcu zTtuI~76lkGGIZb4r*!JL-U;(6bN!kHX-t_bDUk08iDuJK{SA!fZEObUs0I#(;AoY< zdev3S&s8J_o8tUyy_HoufLS2cBI2`AIh~|CyRw?9Z;yw;1wn^P0&vRM#?pK9vYr3v zQeP#Jo2fSS9dsrO^jRxBwcs^6UCKd59r7jo{X-~e+vCCz)~GVZes9?{FDFh@L%KRg zah@j8XPFSyC8v|YHWW&*T2c0>8RT~Pd?wo_;SiFWd%a>k6fKb3xT>c*-Gx9h=+}bY z3xShi4XY-wpSKrsy{ulpjF3;dA5IV47a>5R2V4ZN$fIi}=5>+s-mQ7&p)K^_Bi=ZI zFtN#wK`_)QuP)@VYPoxcyV@&5G%LbG;a^%wEi<+M9VknfNiNK5HMey2Q1y_u0N&C7 zKl`1SFQ!a1TkM}NOIWN{8OI7^=w}W>*;;Amqf37T87CpW`j(y3g z#5f*g3)s;r2U^0XjbbqG;`vdz%Z$8_-X><3rw2j)wq9>U8mqHP5j4KvzX8<}8J;4W zfoVYwxZTx&4A)*2DO**akWw^e{Al!jg+FILS@Z*!Ti- z#8Msio^sGgQR=rIDv7JR#Jae9Rl|w`qk~6tRrovoT9!E5M~b^GqbH0cJEp70gUQZ= zCuBd;d?Nh;$LWm)D{~AR;v&3MvPz%X7R3wK<(eL)WiDSe)1xKv`Ixy-0}mF;)w6mK z<-5VYj_%e|&d*gYcO`p8YK9{M-Hft2uMj!RnQ%9E+LH)zJclt=-uZ(MWHj}Nm zTRGv~r&zg9^f%;RWOIJFcbUwKwh+!%f(`oO41g2hb&wAku7h3s>zxwVzWPoQJ)YdzsXxd^^z^}%l)PGFC%|! z37-pfpK=BeRePrJ_8~8{Ggi$6^~e5|>-Yw`q?ghK?InCx@%G=W zoi7GhNn=$$e78=IR}ugD8zv+EfsZdgJ%(0yiuvpxdpx;QX-2?W2H#eK@by!XYi^Jk z5nm?1!qfI{=z8c1{E^(t98~_d_cwFPPC~hI29`>LO`q(wd(2T(E4IjGo=br8^<`_p zfl#;V)fKEe#-c1(zvAeeEnb)Ga+*Bi`cgJ3=p~`PDqryTNq94$zy6dC?t=L#%9n`?6KtSWbMd#V*jqgS!V z;$xfN&scurnbE@NCi5glzrO9`93@fx(ZU{{im$x7IwX_PYVDQNpwQJ}5v<(8-u0W9 zdI*PCW6;z7%}&<5CJiZqBVs?POsfKPclcU!M996Q9pJHaQ1J9uMzO7{vcaA9DtMv@ z^N`!688%|ug1PlA&p`vpasxf<%3JU6zSG((?llBM+)FhFj~<4)_;>w|^cpy@FhIUr zsn~h?^}4=aUFoXg$AOrMy=u?h>E1KyKK--E)t?h7F<;|TxArqrKc=nLI`%y_=@_N2 z%$S+6y7r=?e<>s|bI?KT=4Abqmx*nD#Ltb)&q2>!R6pmC44lGy{e|71Z!i8n!dPtDuALs6on`KJIs#p2 zgZ0V!pNor&>wGug-@;s@cosH|kX)pwyY3f&$AgU@CO6BRP^^#8?Q-OWNDtkn=f9z{jSh zO0aOAe$=DtSHr}_L~qy4)4)Jw))Za`18{A_CeQyn2vj63;Pw9J^M5yD1`>CSpaZaY zTj?HiBiNL_prHcyMy7}#x?1?^)Wi=oc;4LIoybo#D7l&Su4fKvRBtHA_8*!G&9(T{ z83bISoaQrm9Wrac(xt zqvrCh&0qK;aa)b*mlrYA(no(|m3nn_z3g`9TBe#jDrz0(%M%T}VLhh$7bS1IgUY*G zudmL%rqyb;9~9+3pT_XR+yvjBPh2mlS#otc#3Ts3a3wm`wy4K)WDD~~>IwgnJ%UTMbf8PUJ zF&?Jrj4kI%NA)_9`GJP?99?7IjzrmlY_X3X6oKzu)nQareYiCN4gu|7_J@RTkdmQE z#Dw3|tFD*6ux^4UBCCdkOa0D|)&0`JZ8boWM_(fA1SY;`g5u`py{mexzSVOY2G1gnSBlbU+*umbIB9 zAs-(0htBgurQf+b3<0`HOSdG-=DYrSz$*hk-kldITe(F?z3%ET;k*i+p z7WY$Afh-46A@EHKDi7!|Q~RP7;w>RBq4w}`E_Lm_IfYc7o6gJW>FH?DUAv?)xJCmT zpCXQiS+Cx)-=)0yKo*90kBMEv`0A}mO;I=4jB70v54rzOFG_mzqR;>HqVPZE?E)$_ zqiyHs=d`fM)z@xnY&;r~rnK4`FY{jfMC1zA zKvUQD`P=`)NQ$I8BqZ*E`MK39ryzS1-|d@H(n7WDl;3u*d$;lP!^L9kCd^m zTy5Mb>i%6k_lADHmW_)mPb`~lm2YKZHBrwQwvdT0KXx`HL5w!w!zDH^T$UA(@naJVcth2+lt<(Kwey635q;?0u=7Hz+v9|i^^xjU} z60@TU_tvA3sjjWP<1A!U@_#s4%26SqA&-rhjza(9&TUki50JtTwRY1|Er|o>D4+m} zi;H1jh*&draxfl0H5H;DCLRWou^i|Q!;G~E?|B{FvoxEE7WXh(?4Za0+iOSW{;!Sy z1Z;e>u9u#gWjgt$9L4%g?i--5z4~{pxEhUsi~Y|$_A{?w-hX4H>8b$9gf5RpO$L?#7+Q zt-E0%pR70cJ?(PJ$`o{KZAZ%Wn{JLMxv;NRsDzkY0@FxuJfqvF0xx~KqEg=pkUGEJ zMdt&(!Iuq7`?{os;yDnK4&mwqgeFKyV`JB`0bD?q{?8}L%KU;9u+0DSd8i4h3jg=# y|DEdk-<0)Do525mLg+#5>;HZps{a3R<10$zT1irmJmS-vr@|x^CCbH&g8m=oZmcf= literal 61243 zcmd>l_dC^p{I@MKGK%bSkWGnfMK~0)j!k6GWE^{iY)(ccBqV#DW3SANWar@6WUuUX zzmCuMd;bOZZ};b_OV_9C{d&LN&+&LXo}&m2H6?OV22vax9CE0#f+h|QAshz>-~1X8 z_&*cLncLt8k-4gp0{993JC?5t06$1vl=a`>;4p_{f8)k2?pWgBu;D-z9&5utuTOb> zUi8$TIqg5>Lkb$7iZ|t(46g7&-n}io7Zvz5(1A!_Pe1J}Q5zFa!_?*b>n!_X zN?<0(egQhm!&*?!kwdBFV;M?i1O)p3>#zMvZ8sc5WnT_o&o`Bv!I0FxC7M9%v9_mU5yBLg($VCs4R0*&cfyT$w)CpsrB$oG@-Ynan6ti|H zpAx$lukN2**HmfpMo>!U57_wbKn{m~{P=MwlG-eJTaOv?z4I$D5v1N zEz)0ywh=iQE!JuBpOCjgm3qz?knu z$>3WhM8IF*JTN?WKir%$G}Y6CcPYv`EBMB0K|FEuM^=dgeS!|t`iJlX-lS(jH8Y=W zs|@PjOHw&>=-4YpdLMQ6!JWX3^9F1j_Dw3>F3%4G^2Z&Vjw5D))wE(1=c^cEI+T7b z?{Me*`TGH4%lly_DZ8TdXKKR|r%V;NK&9o<^73@NCVG3w8Ii5^pap9HC_+H5!6d*{ zfsc+k9J0?4UI<-%YU}TRVN12cGjke7> zud3XDh9XSGKh z7zKhZ!*ZNI*QZ!B@84(5yvom-Cq=-%rN|a1**_qhL}X=UksqhSf#GUR^R!QZF_Q8! zGc(74+x?n{0&~HkO$B3X>f$?l6?}E8f66Q}2xJCrM2f2~FHY^7P4QX#%hw;kO#R%E z`!$D?Z~8Mt8Dd-~#%e*J{{x3K9xN35=t0qNPB_PYVrXcn5{@kHOeJKtKjvQnKaz3kTo94t|@M?Wg-}9v6VCg(F~M_3kfU&bpm< zeT!w0K6&6d#5}yD+MvRnOmy=Fbc$~M>$R6#N1f1J9zYMgh zIF8klhAQ~ZwfS)8^cNe;WS(De`dx*LIWI5o7i-#zUZ*TgQ(|uhrATmnSDDg?VkueKOMI#wZ!>zfyP{`q@(Pnlns{a)1OD2Zo}#PzmtJ${ zSSj3~vOe)@;P@|e)Kyk0sBn*OAT@I|#Y<6*Wdk;R&T}aql)O;JX4cOL@4+rKK;y{ zqm{1r276QwsR>k?tEZe5e>3`>te&xPL!r*|RLk*k)bmHyz@GCtnj|wA6!#Ch>~~x8 zATg(MjRdL;1bmZx8x__n$A?5u>}!B8W2^~FvePvSMp!7%N}5vu&^)xB;F*1rq>`hG zdw0%)amV-Re_m24Qkj~@`L0`V>E~!%GGYh&y!09nga5)S)<^n;C0BmAifqk1w>|XY zo!!;9MY~$|rkA5J@N&E+N&rdjAKx9l4g~r4IIzTbkyj~5uw?4=&j`%aSbBUOFoP!2 z;JlyfRIg&(SFW)|x)9<>K?5?s?9nmk8D?7ly(FWtv5~mawzu5DQ)NY*J4X!Ug|TPVW^STMjfdC%K$g;_x0BOm%YyjZ zM9c9_z8)C^o}>#W{@at(&0;u!=h~o8T!4*Jn+Zy?P2ModsjSd}%b5foA0OxO^0N*OJQT9N$=oBYb1@wea$Z)S{<3)hO z21a-NWzU8qho*{p^=>!s>94&Sh^^9gnlQBC8~AL-pekI2wym7RO+bnP>6YUUf=me; znW>sSSdQS-kX^>%2R+41mddug0I%^ymj&Y+8Q+A2td?U4f0LT^v)j(apa29vukTq~St$ZCEx}hO zyr49z8L*jznoRka`26v1@HbXjv7p*H0Tk7EsaNl(s4-jMz2$gHIbRfTXf7Rd6hQpp=fK16`pHiL`@ek9q?T-%Vy%BmKiSXY zK~pTmQN&Q>e{|p=g=H}8N)I{$98($;uP~Q?+BTB9KZ&UP$RhMu>+0`p8EZv*%Izmc zVPXt1O*Mhbyue?snUA5G15D+dP&&@OU%!6o@+4hSkl}`hg@s{`7Xdi??>?r!dK{YC zK~MrQ4&>`;ye9Pz!g|sK@-oSgIS$w#adc9!YvPewaETb~*!jhrs%f3DU)^5~4yfVMY_mGfB^lo>#>nh+ssy{L z(iBqIx37D@h4qhjVtEN0E2$U5dY?`P}5JYG?JpGQyzHR~hBcQF2&%DIV0e2@*C+w!% zb})Um9TefbS5o*TKD+o;Fl=xsQhEHoIwf4pW#m|v@=}`P!V==Y|2K6ACMqBY9aO|> zsAb-}Bb0BR6Ktt$$J59fsR(KJUFIccYIX=zA}73S)z{q@k4JQBR*mPEI3~-yOn=&l)HW_Voj-A z&$pxxn){*K$DH+t>0*0Z&;F?l9=rth`qk2CXffu;%Y^4&d+r}VXvZ!j{sFd`9*5Jo zZ~W+_-^GJqu4ju#=tHPC}SIl?6e$o=Fx!W91L%@$j8sJt7ErqQ}IIX zymPDBc)B~T7kO+{)4R9jZBgrk-;@Zkwa1$C^=7o|IqxKHeWAO3A?v|Lx8YmIwKrEBVZq?~(7^(TfBcate8)BBoqCoV)d zj`EVI17N1H8533X_^GtPkQG{N$-q6FGWW^qeUsYujV`gC@q0$1gds^qQ5L>9EF$>w zBlA@!G-pm>I3^tir37>~j+XiirCcbH~x%XB0GNKm6o< zxh0wUA?6R%$y5<&Y*9hOugGl}d5Dwo%8~Rnk}O3{-tXTTv5wKZ#`{L1h}i#Am-iMUp&GMrmP$tzt`*V!nMj+yyZ;KWI$5<&p}Yzh0U(yzNNpcVQf#8 z%zfypG*Ab5eUqAGW8udeff>$IjBDY%R%Y2HU0u5#H|W%V4F5i9NYu+0Y-fy$QUO3` zXnBD8z<$L41yAVV2Rb&|yb?0H=L_xExIVl3xDs578%E~*oUkbTvJJ`O!ruO29i^gf za6isiKkFn;J%J8Ln5pqDvfOP_nkYvyG@(j;KbkYt*M znPIsHO;8a&H<$Qm*kEg%@O<$uyF`Nux1cioH~z$DX({TABK}V2QKwkHb_FizR3}g_ z>B;=<>WM06cZ7(3n+Ek*n^n*F*IBdU`fy1Qm5?WXI#cr#Ke$Dg1PDbGz-8i*rP(zz z^X<2pRug~B9&Yvd3@Bg$Dg&_1M#G1-*jksi668PU6{a{jLrPw93ES z8Gd}r3qy~4CY{VL_=Oap=m8MptI||?jZ9)RlcTjrdmMpr#n$0pPsJBQX7bk}R>C`k zJa3fs)s};USB%zc#5pz1Ymg@N;oTHGiZ(D64b~vy{uVcD?jPC^r;>mu!I6|XhQm^X z+op#Zk32%4*<5?I=j?vVgh7>6Sk-cThQN#Yo3U0URq~I__#jQBSnnu(~O*rY))0r~cdc{>Xq6+01K# zfF%CTl3z@b29IZ_6COs(ei$k&0j=YdR;ny$28jPraoYe zOu)m6`oUv;yl$K!cCROuj4}JIAy#|&7*w(acW{a}bwDbA`~Ku@$x7i+h$A^b!&;M# zZy#2g;$(5tapI}hGrHbqPTBMq$3A|YqTAS4L>R^nq|qB~!$>ZLZMUvv{$9`HiT#Lr z!L6hUFI+_fSrf5a+ZQLWqqDP2xXo)SKpj@#s;~mVPK-7P;8wP+3Li5(&e|^l<9Uf= z)PFj~IwDf6Rd=GFgkY~a9-++rx)fO$^Q1<}q$WiR^JJr2pMGn--^cyI4s2YL&sE9I zmcIg$*2HrB@J=4xNb)>sa6fL!lSbi2H_{FwQVNIbWc!P$^U2rzy)M3e;1Rfc6=4n^ zt*ND`cOTcjG4@_vcRT5EF(ajTwsc88s*Q)yxxAHO_=uXAluR9gV#XR_a8gr>hcJMs zS6-&*uzbb$hJMKhM%z<=nm2uNNWFg@;TQ>De<06X#adR| zJw2Q}FHRT==r0yzuQ8V-t*mD7CHaT9^>{PoABmqWZ+Meb+rX3x!ZLa(Iv+$ugX5sN zx{OQW2}dk}7b>0Glj%BY$I%h0-z36HW4_JMAQvC~#Sw+xZ{f}BdDXM?9&*7iM0Q*` z+hZjW7Q3H#LE2k|w6hUC-frU;Hu*1>^4jllz+v{|Ya?}B3;;vF?FCe2&+w1Kj)!~0 zzxH(gh7PU(CL4ZOo>MIoJbmpuUi87T?ngO2r);@^##JI23xBMi5G*M~P)?LQ14!-j!EnyD`S1zG zz`^{>OrS^L&&Smcq}}%VF$BcdE*uUtjEo#!Au9I)a-hqsWN6va*QTWJrBH z^>*meE5rnxQN+!G8LcfI_c<&iKP*DNGMt{@6yUmOoh&MAYA%J? z_FyFddmn$h!=lLr0i0)nWJ-G8i&~2drczqxS8w1l^dsQ|A0IYM68JkUf;#2Wzke=2 z+!ViExR6xu7)o|M=cY^=bLZ%6-K*sPD+0m5dj8g42QEKrLnNzbxke z?^eQ9w5`nru_Z;ADVO&l8mt@D`?KpVQng=PAhufAINsB^sNBH@ zs7JZ{PU#4B1~Ch%*In1k>?HEkm##O$M#G0U8V+Sl^vxcD?3r>FR^hJBM1E$Lnc`Y| zvN4g=zon-osXJFyVno&{qPoh1H9GHpPBkZWKr&lMJ>9samXNT9+p8)2%%VuKnT-dm zq*^QBq||KDUacV2JFM44myYwE>!_3lnEN%m1aDntwDR3#OoBM_fjB^gzXPT-3{uCF zk9UR(riQ(oHl5ba@;jEbx8j|iJ;7^D06kr&7Dj`VH@aI+5f>`y3yyF6HZ-KFuvTgX zR?@DTr(SY0}evU1E_-6)<)^RzpD}ScA{9(Ft<4aBpdANR@$OWfn(=jrfis#qH!_@@E;!JiP$sAMNj?G|%Hum-;Xr zMAwb;|L`?@Di6VwxKJTmtCA;fH2!9==YY1KI(B{d2g zPGJD$L%WjaZgU*H3HFPJt+9M?jdV#MDiA93f9B%mo+NgQJ{F`MEMg`%2)E1dfp!le z@o-i?NA>pxR_8f#-rJ18Nrt(vJ`d6$8gE-aUOimrIDba zV2eLVr86^9l=GClona&Vr0Yb2uu0U%u=gfi0>qde>#E)#3G{`6du#Ts!_>_-tWq{a zL`>gM{A9Xj<_&aPU8|7|a1PM)*WNHR!snBdIW_xk;3D@`=F?@9-UCCK4yB0F7oW44 ze*#G2FLQw`!C=Q^E{TI)z5W{O3#+wfuH zm{zT?b&yOhO$iHZY zspfuWsGmX}&OLW&UDy!zSnbD_Ox!uB)hVj$OsLen%-CC&AupCJ_`=8+ARG6vDG@9A z)xa>4Qelo=j&&g^lk!toLgND|-B8AS5A7W_GG9->d+$Kh{Q3o&4s_e^<1@@Fx?cLA zE3W4+BA;$|od`Me1MZhl!k$%)elDzac`G1s23h1o6XJ4A_Ag2OP}z|SiTe2?0d9e! zZ-jufkRZ;ns1HV(hyQ&1{fN}pFh6nkmMcFA23bl=$qmGbgxH05LS^O>WxKpqQWIXf z%Nd4#JtGjYJm8TyD?@Kwf=Yk+A%zYbKZCKZi{}55^jOgk#awIqbq6|5$9N0z^9!~G ztJ(A06+xkimGtEAfut9qxup7Y{a68w|2{403tE5GG0hX-soquCp%j!iZnrL8zh$NT zE|2-@@VmHp0`mjj8POU(+SoopN~VPP^$7$t-R7@*W6ASo&7DDy1A|mBh?l-IWN{-X zQFIDyw^;wE^Nxw$6>@xyY{#u1=+8el{~TEC{1MI|FG|UHuNSvJMI^x5o#QOf>r? z+s<0s$)gR%_%6q8=?f#jJf8?+s$8xAK=nb|i!3G2y}*qNuD|P6-}LvL8%j~#W<(=6 zuJ$siatgqQ8)lVw{&6RdFAKbT$A`T%fV%s_ZG{2wV=S$S!70xe?UjL0u;4Yrv5MCC z&ri#;_@W=XOXr-vJAsW|2B-{rydqT`^vdIJAG)_g*US-5sWBmoJ5FC6upxsS*k7Zq zkGXa!1|TS{Z8IbRw?D);ZZe-}!pw^$uoRVt@H~2rqJeZc>pmIv9cdxEx??k>KgvhtG=0C_nu?-OLvQour2J zN%a?V9oUoJy5Ma7?=AooEvi`v+xkV}Ggy>Lrn`HYYp+tc8^srPObeX|lZ}T+YbbP6 zwAZokP<ZT~);a=fTqU*T?VenTM5^}{FE5m!t;F54Hh(c*r) zA^;Ta!a?C;>mnOZ@6xQCq*?{>mLGpkBwG02S9Si$W4!}4>~xFCo-c2w&nMr5VynbA z>@EXs-tFHAjx(Z0St9~r8=#OSB?o#dj}1IT9{~y;*-y;kt@xD|&nZykMykE!lb!u^ zK;0FA_#_GRtLiAGU!tUu%5rx;gkK0CYmC#rH+_bAl-w5AU3_>|#%X`JIQ6Ub8j;syzhY8gYa zx;~)VFs{~#L2HJphy7E|rR6eikG|A?(!&}Dgus6Ny34u%>-+<(IyYQJoL%7kL4`2R7W)U2*LImuw%&JPm{ODDJ-WwoliQq*vzy0t|afd5xh(06_+> zK&bC~*PkNMwZbF%`AD4|qRB4M6~HRvZ2nvIneO7&`%eAe6EkIM8`U^sOX-(jkeeN> za++}lk7!d8b&sAE%zZHk>-qNb&%e*VT5M#=N}BGJ<{muvFK>3is>WyMW>TM#x2#9X zxn{ik)-_~O%=_f+*^OWzmAyZcnB);#g21(!QxEdeQ{7ZsJkQR4Gh))7O6`pFehifG z+qN~sl!{gg&drD1k3~mO6qq^}$g@3Fp0|QOa#-FoNB7*@;lgzX1)fRGgtIGb+1B^V zd|wc6IaT_9x-QjeR=M{*Tt<$&B|6K>XK=x+(-t+=chltR1|Pr;!h2a`RV5L1mep?g z^EL@?zo>=AMnm|5GuVIJy#)poNL`p0xd!eBhIRVt9F~gTy6ub+8OD#O-?u+y&x0M? zE3Z;v8$?#`)jc7extaY7#iGp%W=c1Y*vng`@~+Gv`W!_s-^^>cHl8r#TV9J%HueAN z*jJpSgVDs=ArWhb;S>vJ@g3Vy@67|)vrRh-4hb$LrR{?s!2n>;3c|ha&5(tN23*6m z{rY3nkOPU6-XrQg>=oz$me)@HD|!((6Bi^M&*d)oQKY6IMogMuWPh3`u0MIc>?$-X zcXnTvGq;Xpx9~hGI(*5ZXS}$4olHrS62yah*>S3Qer*_7EqOWC);8BT7Jys=< z8!kcZMU6A!s$}WO(_;R5KzWq|+{B4GVO8GuLtETq49)%=38>H{g!U1x!8IuMsOM7Y zoKR3gON(_9M9@8M>_kb2s98EerYZ3iwkJb#;L2rG&{ab0_h_0fH9$--2BwwQJBT8q zl^g6(juxrZ{gqhNx$)`{1{W715E71u*{N`8aWDGJChHUc4TK9mA_a^SGH@{%gRR@) zB&Wf7Y>UG+WjSqi*RKQdyz+!kk)y@lP_@9*o%2oWT02Ga!|~bbe5?gZM} z^YZ^5fCLiw9+yAm2VDBX<_oB@HT-jBQa6cRmZKQ87}YGR7Ff7pPt^4?Ui0?8hNz=Y z3!y21F9Vf#om^RS421E9f~SIe&JSn&6kKgpm!ED*>;bkDA8xOH5IO&YrdI#2`m87i zufZ|7=0(-st5LRNvn*Z~S@rSg7=-yS7?i0vr>&Ei<8`Yf|5ah!-zCrvOLgV*K?(9U4 zdQU50(jTLLeGC5*u<`UoaQLti{PBOD=rB*DuCGP9Re$;gt4XY~vw=o<0vl&Lfk7ww zF;H*B_)hjkbR92KLRc#CA!h!C8;aW)zo^*^6JQAb5v8Nqz$jb*{gW^^1@upJ0bFal z-a%YI5HQ6QNXJU~UjFu{97hBiL?0Md*OwuVo-`psM$(Owr&F{0nH+$RC!ho+w#NO{ z%cF!q9eJPkzlfaK0BDhS9e|2Qpo4Ij$~M%D+AN#Yr}H|6iUj3`PCW;!CTk)quIHw(O7w>P*dg-am;h4*KK7U%AXZUz@7jc#u1V4E&VnY)oty_! zYUnV~Kd6^&qM#?f6&0-nVp`$F(#^sK-E@X^SpnQx>Qsf6R7aWv?F7cr?L$; z{}978KURTCGg|<3htc0So#V=^-b)7gcM)s%Mbk1RRz5Bt@cl!86T$k6Vy_Uq3IfH` z-0mJlj?|&afa0hxzq2!V98#w7N(xa626wiSk;>@%~rte2WAiq>tvUh?6pt*EFehJF6Q zo2rhB#3K#+1|M3Yf?S`sl@B<~4vS+MoqtkvSUznB?)Cn|{37S4Ix_h|I`1Ios$lEF zNk&;vw=f1P#A^ewUsLm-JLl#FhY+5SjM=oYdsPC@vnu8nS-s)R)c_|^2Iq!S=SMEc zi~6UZDnHsemh9n~r15vU0jNc)D@^X5_VWJ;S;7E135F zy>mb|q`PHzolJZ8G1@+i9qtGC0zHd;v<*~5AKH8j)*9+|Atl2bJRExTrlz@VRHyMk zy%gOK5V{qZD<^Z!WzSD@!xjl>GY>86Ee@m>ln} zVYcXfR#P=?-6J_{8!nq=kZ+HA)Pzx;Lfq)Np>-+xMZ09JCMptkRXIE2j|(~Du~2C6 zw{DgN<0D8*t?3d+O*EcvPoiC=a{qv$S98oT8qqqUogK@AX(OkpYCGbV)kmm-F^+a= zsQhbabDbgf@>wr6OF)zF$6svc;ZYTBMf+0_1hvMFA6VJSSV40ygaS5n3$$p9kwdAt z^XHbu*bJ9Uu!ZC)Q|0N;QG8+DwOE}$w}wX#w|`SI%icP2{YKEF@Rpc^13LIk83=bZ z{f2_z;fW}w8S?j4;noryX*7T-g6zV*UJ$6iC8qd<@P2`JfJEBi6zB2p%*|nEYEyP6 zX3=>4LG@dk!!x&8*c=g=C>|5Dua`1c+R1|robdQoY}+f+h@Px>P(9Z3F!%p7U zI>rj<6}vh7LqJCXJ9*3NQJCW>vOk>TL2H7NS$Rq1f*I4vVxiByrIq;-??-avnP$yskAqkf-Vr9c-?6W-p6Gv`k(Y6qGy{YSs zZ4Mmr&LRGuxaVKb&kYBYwsZq*n?KJm5e9#;=Iv_ZXd!UZ;i!!Vb6V2~*mZ*q&^o^^ z>k^t^n$im?ciN^6P3t(x9F@^|bhN*o{6`D>xNT>Ya*KEH7(gMl@&g{s%yXKmx69HT zX$c@-u7m!zkc+B)gAj#f4`KFDx8jwz?suOZ1WjZ$HG@X-Acd~!Tv_u|9C;o3%AO`v zLZ-|gq)B_7eQ7TvXe10$fVBbi6A}zrue|@6)u0{jO0;J1`FpSS(@#pir*)GiSku{Z zsTz&derDjge`93~yLTqVH_8f>AA8{I}h)6C{;k*vww z&HY9AQ&1ryveT@m&B-8emIdd#>Uu7jjyB)1(eF{%n+gDEkirby;qHh%^;LggL?t7Av1p+ums<`GR>U+V5yC*o)wk`zn2 zH>KlK}G#^;?y~wJs0}+S_s28$Y$*S4MzZ zfdt4RlDzQ*04B9v8mvYl1FzST#;Kk8?}1i+WlSG7q?P0w1z9%)AF-|Q?ZeExkB-XA zYrE^Nu7Fhyd{o>senpm!Z5C#OiW3iUuFtQnDNh=i#;#orSsj&fbc!8~)yX~nJxoG1 zcrV-kv|py|Jj4>5V;9H{=u$GAp@kjyq+95L8?coG>zh^~$w?3PBO5>QbR#{GTedzC z$^?Qo{9IEh_ThQSl81GDzVE6a%%E;ZlTqa*>N)VQWySN+#EbU*c=W7L{YUYQvo);z z48m^2I^H#${Y*D$0s~q|))W4XAl;1f5vZ=BUXa&)hNosZcVxqr-p3h{GDoLol zWHBgz@7BDwrk>>a-MOz=&C&))0dPVtPs;9(%w)~EFZf{`NbJMtpZT7^6stG6TiR)a z^rZdmd4grTog$>MBOyejn{{GpM1}+ZgxZ<1;_J+h_ zua&gb&#yrn06bMD+|?xUZv^vJy#`)Kq}^?SkKLCIOXb%t%ytjxCHq`8^}$IQUsTm zs(k<@JABTs^^7xWcJ>*d_vVJfJ#90(=07;v-PZ8HO8IDVH!`~P&8E1jo7ji|YnYW- z9r^-S_L>HNOJaj(&JXg#miG6|MVjeJG21hfPZ;FZElMXroN;z`Z$Z!&nW4b;nm(_2Mpq{_De+Cgfn z(jD4~Zn&II5j%k})aaY6+*1eW#N)fHwW5QbXQMV1Kew>mwX4ivz2Xo z9{w}a_b>WW)F$g3c9^6MO2a|Izk++DX0*F^)Xo-0H)dd)U0*#v$-zln*wiC&$15A=f zvMUM2v&WX_F&_(5wgvUc0>r>E@Ku9PW7q7kf7osMkm^+!GJJck)GhuG^b_Vva4q z(%>yn?xn@OBd9M!ey;F_q07Qf%}gf}M5(vF^LM@>y$(n)E#>4+F_8cvec{5Tazl30 zHEWy29KoSMc|s{OjPfDPyx;fuh3?+rma#c)2eS4Ct;Ir9Y^t1e1cYxpo6hN zLxG77W6-Mi_3}lP`6e6K1w=X0CN*vECsJP z@GBA?MU1+dmp1EV<0a9&$^ux-XBjnX$fB77xGOCyl>Idu2Z(;berSM>it!!XQ?A`C zM|%P0>I~Bzy;J~orhf0Z5fD)7&8(?m?Vymo6apb|)KwYQgsB<5H#85(r?GVV0 zUoBw0fuf#v*Ywka*7go|+j(-Gm+BK?k7i0~-f=+Djo9fvq&Vp8IhthILYeHaWx=}u z(P9-UkL6(O-rlbkxzH3m_V~m7-M=ZrY2U&nMBDrZ7Utim&KgP}Add2j&U+4E!)Z*; z|E>}>U$ngK#w!ETbW?0k0O;7%c2mE>@!AL7B3_R6w)@4yQwQs-e>;8jxgTMhV_*eO zbFOxDKM^#p=JPqe$5c)!*Jnt!-siZgQGaP3P)fzL`FF*+#QnR;oRHn8t>*H=P}YRH z1Uo#u5)cx-tkHxs;W#`T^cB(}!H`~EbAPHYc9*sXucrws zh>s~Q!S~Pc(CVedLWCV%1YNl(j+d*;#*@L<+%3M#{s-T%`)@$wI1+(K&<7!Rv-v%n z?9%;}0l_@$-WWIS_5S3OBaN#%BKaL{1YK3Kmi$v9c1KfCf}>Fx%++xjf}j_RU3GW6 zrD?do`Ojq{!4YaKh%W;s9TPG#+;xV){0B-o=_1mT$>UNhVgY;wx{*uNSNZq(C-|fFd7K-bQ2VZ3NDe;vS0?4&4|Mh|GoJ{eR4N>(DYCa3ydaPH&A z*KRH$kS@cA>mUE>HF1FT$?E_|vRQ@-x2bnce?BAks8k2ZMLP8 zc+TDbLzK86{jC0Utlfcvw~TIR|GptdlA@UK z%87(A^rn%Vz?_pFr1x#iJv&+>V;I2;edEVZsHqWMkT+ZDo5j|bk93Bn-hcf0sH`kMYKIz_=Wo9B194}3~BKud$*9Ub!RegZx_m#GyEC#y)NlU0% z8lK(QZpC4?r2h*V!q6UFJkh?qIdb?~*Tb@$?{JX2mW$jbgo$@dacW*Ms2G$c$lM#V zF8tP3!CMD60kThm21L-jf~a9bU1LFG_a{i2g8FGa_g7|p#tA^-i0n}j7X9WmhSXo> z?>}WFenCstF}HYhv>GK;{y0+Q?GyGkCgX&2%c&gfBiy|6&nt8ksEka;S=6!6hA>dg zNnOT`3-o>cee)n39+W+VYLv`qAdR1y*75f80`K>(2Htzq+q=#4g=RP1#F7s7-i-lW z?(01v9I;m98xP5gpq8D0SGRCH&-tc8lJB?YalTq3rl9nAR#^fcL;D{7d6;`&=8r4f6~z&mz1R=RXxPT$ zc(WIdv7?vibEt$dcgx?;gk#q*FlI_w9-#AszP9yiHhCJ45q^w%r5t7Mb}`qorPsuN zqUK+jy$H6+wds)K?IU5I4>(rH3O7C-&i=Ofw;Kg$0Xi62Who^QbUFM7w8tyqC+JLP zlyGa9>yn;V=_JO|#Ue9E}}Af}Jpqc^cqF6-P&-6ZguiCC@l zuPc}7rgsaZ6G^dr$s)i{qujxRK7WyB*v@$N<76!hJHk@~|30nAoBb62SdMVlU*1^l z6j3S1Y^f)E`tn6h^3K}T=XcXqoXduLw?I7x2mkUcUuR&M_`HPetFDNAm>J1Wv8-?0 z`2#%Gr@7MOW2Yl3h{4J_90~(6+~{+S5bQ(t+#89BA9A$#d^0u0E|>tFx9BNvaf) zSc2`KlsZPR@4o}yznZ*L5fi@VV$8qtpOd`;8{Qv(-G#4#jajlS zjjg=6l*&g-x;qH?{`UTBk4#*s1UF5+0W!=}fpVLti#T>I?SLciXL%xJbe!YZpN^Op z634ut^6w=pA@;@@b$eVFAfEH2GG(cRH{qL?64nQV*yr!VZ~vw|j;yC(;@}V4pE{-$ z3sjauxk?INIU%`ZVD0O5xP^sJ+p2M%MiR!W0bWAiAn=PHWEB_a#1?2_i1)~W6Qv0| z?&jrin9YCYF$p%6JPiALkB0?{NUr$EP(l!fb0Ybrt_T3XPH~Zlv+yps))saG$_08T z1%n-FQiP!7PICAAW&q*;iZuQ`VS34KKnZC5I+})qA}u zb?3>O%h<|-MyRDgCr`A>yQ;1c{JU%ZuLBoesay2*cTUt@zeAt}n#hzM=x1`TMW?2x z^n>NUbw7!%JuDb0oJ&e(1O{+d#LzJYrpRGbyV^$}{Gp8pUSj!c5?nRdF%Hzgq=P+b zW+Udg&z4G0k4t|&4<<#Ys+}EM=Yv%(yOE#W^+Yzr^y|ThCPGAVSBLn+I$W639~O3O zTYhl&#fByCHhti)`-Lk!02cR{Oxqc6Z`}i54ps$nMix8nJ)N^=OqCZX=CmbouUb5Z zuX&>pLlY3_-MdeFIinQ49Wy!km0I_Fb{cz4>IZkP-(`OPzqIGDQzpl5K^+SWTH8wZYw?Wa;m8B|7zk_pcrvKC0=M zC)eFl;gM4kH=QgiHcFixyFy#@r`ZHsP7blX$86ezyKjsHF)4=ay!@1$aHQREYO^fm zm(ML`)-AYD1Ek`K6>$fyV4g!1D|R%Fh!idli3(y3JCpzI4vy|Y?=-;e9~KLMk9NmC zuF%2A5I1(HcbhHmk>mK;NQ+j7=-qmtj&)PsB~mw`|1hn%M++1YwC-)?26(&qe8vul zu}(H`CH_hWzqxQ-wD%(mKQxqnHPn?mfN8t1yVy4@2)O#`zC?%3KMa4XMCaeZ$ltob znd8%A4xHmrFdnecuUhH|JDeC7Zp&j3%3ZMo8qn^t*Ro&wX{_XB!rJ3!|=zrVh zGq2+>`I#yYiz~k}!P*VW=}NT0dZC@yNSYL5HO`j;@)JogK|ri(2HQ=4W^VDRiVjzm zQ$ZL^z!PaIM(RdnK#1`UO!8o^nY_GNf?W`{^Lhxkm@fl@Dmk@9;IFMLs^-|{aQ&ew zB3oB`WAFBDl?Nf=yNKdOUPtZQ9%>Ou}qsoY8S%y@TeMxMQAdW*MXAWG5 zGe%T7`o9)p50IfCnZyKHsX4ct z7k69Ct^c%@9onPlEJ-#z(Mdoy+^s9Cj4&8_MJapTA%4=X~i+MI)DB4;F zYC;3#v+aCKa?FkMW_z~&?cyz&{T+mM_CU}$>9<69>)(S%CD_S`UHQq;*`FuR_D+7Q z8k8;R>~NYGjNhW8Wr(jN1s~C~`guV|jP9iP65QiVy_M^iC3OJy=Cbu!<|8NR7x;y_O43ug_no8}g!c4)|hVA~&w-@}} z<|qlX>Sb*umBHQn!}uFH7OCn?D@(ao8Ae#(6s3eXQjb4OownIB&f_3Il*>!EREa2J zp8$OgMMx0~Cw#V4X+P4KkQf4`lvfdw@I>3qeIjSQ#@MHf&@Jq$Cc1~~ z@b|SNS)NW(`LRwt_xu>x?ldq4Ubxv%vt#yMQaxFDP^BDo-Fvmjl3Keb9Bzo^I5(^O ziFu_Ux+fUx#yO>AO;yI#pA!#m-by^8Z47!1W;~-3t$ISSJY}|$vX`~V@ zIY5dTYB*7}H^Dy~wT9dsp_XJwW%t{omPHu+-eFyM6=DCbbX~>Z75D&6&5x#E{Y^@? z(l;*Ip}f!44gHx5)}AenGrj^DADm!(rVt!J>9tmQCrU6|=cWAN=Z74@Tsj`xUqRuZ zi#ud$cbjpVZY?k)$ajZ`I3A%}YZqWb_~LLb!EwjpisNmuWTX5L^$;a?u>4>hf^1Cu za^F+9z3uc!qw?l05KJFHFnzFHUJi5$d}u6yVY6Md?^&XZ)#@OR4u^?B2g^Ne5E<;R zMde_OWVPF6mHCBhzXA@|ezA_H&{;qYn80UH=g$wM1)JJ9#8tyXvTq5C1R!@hwSD07 zAYBmu;uTAx9?zu3Y&^$5CX<8~DB+3z;@4J=E}6dN$K@B4O$b71^mm_q;6Bbd^Y4uX zN+|!`oa5`oglFjArxsW9842V^TMS_#E(KUOdsrsO$;O5qk z9hUZ zW$P5x@o#k+j@{-c<`!bgv-~(S((Pf9i07Y!Rdd<8BQGS1o8U2&wKJa}w<}yy5R0#>>Co0%wd#eFzwANE+ld4E zG1qaPZ*L}yhiw&?dBX;IeovFw*Tp&>D8x)-X&Ldy_jw@)+}Xm%X&Md#x>hCyc zvqKGB*(K)(n)~HVT*_>}{DDLGol}XIw2!#_I|#?JmjE3mSxa?OMT?07O;pKHv=x+o zV3Eow8MljV>p_YXskjL*Z{Nc&-0F+^KjghAiz=2sPAXPQ3qj=EJNt@0t(uY3{n~UP^e* zdG^`;?7e@I{L3%o+gru!Sm6hNhg5)Hm<6T^n0kC^K#5|%S%dQ&5F@wmT=uw(GYCHlI(I0>_SBwGnP5HM@ z(PUMNi4TGs-9vzHO@HSp8t>u&D%rG$^b|Sy&au$y1W>|%ht32+w&W7|am48Cg6S-j zq)kBOcafo5-Y8=uU6rb3=+LSaC&V6D*}vaj(_@dnjLjy3bY7`7U=rT9<`>K;CaL^< zFf9Zi&v~{5qs;4-Nu%7x6DkUG_Yiq@MH}-%5Jk4_`L}bP5Qg%C3pDm@I-76L=U6MY z7s8_ddA66XiB-n>r(nKsz@J@%^p^B`lwP4IE8BP@M9PR24LE;O9;7m#8s=V)tbisX zSYPg7GL6n+lOFo}1%MbqEFn0qdW~fAC@d-bC~_TBLin>DThThdvncuWn~_HaAMy2L zEPs)RVBt5`%b_Tx1cqXWV}u@!eQDN&4z@l5+MEQs08IS=Y5eMMYVf(jnDh?OiZ=ocu4d0vF$B(h%+2f5&AJ?a%T)^H!htoer38bDk0NnHn3y|I>T7F3NB!`#c2 zw&{6)Io2b)9)6_A>`UDW1xp!`__qe^?y<&+zfAxb(PT>j)bJ57xA+Zm4W zM~JRp-mxCj@Lx%iDU90bkl4Vlz z!t|a@7qZO%!Q#AgJqL*5hDb3XeVL=Pn~6=&>%5eHg?%=jUR9&TmHIKp*%<=t>vYVtVeXpnS1p6wT^5ii0MzwT! zUjEP+6USnelNL6ZeS(EpRu`4G051MhV^?6__SGZYR!#c{zzcgRw!rgF`oU~glftYC z={m=>hW9+KJcRx(RKdWd^J>x6CZ!!#^UW{VO*g>onX>}DUqH&K<6&u}3+6BZf8cpl z1kzbqAZ7LSlY}j2^Zz}Ek7Ob*Iy*VLal4n> zKz@qZx_3=G$6XbuG4el1d7g_%WxIe99WQWsqU})AYMkFa$cU)Atk=#~EIdFwqE#&3 zwwTlhJy3ZI)_pafAL7_QP<=&n(A0x6F2u_B(Z98vltHBxN(9w=GmknzPvs>i#r8h- z9GXGZ>|w5DaXCSV<1?;&c={NpF|%8I2R*$yqKeU(J-+LGvwbK_$cD`DZQMCYM4o`m zuyo_X<>)Dr^G{-JXW;jIeCQ98#4-L7YwxnmFn$MI|L*jzL!N?^^Ep>YVLUiDy=T0= zFPM4WH0xXYYPr_mhII5Sk?nSM9a!i`bf4Xf%GO6cFueqP$r~x_TNmgi2ZtvsXhzqm zL?D`XzbGSxqX;owC?R1b^o0=w;;W}kLb8>+PgC<7Qo-d-;qnr8?bp1U`~#z zEpKoD=>vMkM*4zg8yKV|L@?JP6@1o59UtzVTc~0N|Aq> z#pz?Q_!QIU8%=DW?n-AgC6D#-<~Ie&Zvm4ZDwp|7;jumf6)m5fu(l~<_G}5sMwCE7 zn9DvaIr)2CMr=5GZ?(%bwC2ox{v^K@h^E*%%ue~+SYhEm> zwOw*47J@c#B=_@z7|XorCRWlI+S~zLEz{=&$k6QJ<0?bfxg>I=^P0Sq_T(80usxsT z9ai~R`lw{s{v@|@3rx-~AumOc*U-0lfGXz$`h#gOZD#%Cd|BS0)G#$)L**Db57!8* zF6fqaSU#|~hSOJHBFrz)L5c4xZ=uDR> zC5OMxsXR#j-2CG=e!I3g|9wuExMII?x>MEFjB|nmlg}SYA!iC#y z+4#r}oBxgq(?%FoR$Dw?&X~}Np5h!~NdwgdRdQOmmA8h_W-kr(!}?^Hug#RZWTx!9 zTKfe4>bRD$_g(>C3Y%%y+r%J0KX1I<8?ZRRpHV(r5(=81zO8OEFPLj-BRmI`pAPvr zOKG~>fZLA8K-9GDujnw3kuCvz)Sa0R_|7*a2zht2mq<=dTk9cA^=v+7`=y|;J_B4g zbzT_=wJ!}tkFX*pUcjHd49L{Cex#=lXwY=a z)5RS2646Kx)3mYDHJe0WfA;ei?MaLJMgHBLKMRBU2=TbvK8Ki^bJVmxvQ>viG!w~W z?pnU7@w@iVykZcUITs!2Dibx zODcM*qCbOR&gap5U$FY$T`2}w;6gpzZuk3={4xV*zXK8mZ4SOJgr-B)mDkrQAx9_v zBw1w$K!l9jr%`*Ar)`v7BHCLjv~E#^=X^=5T~tW!UQ(_)KT?IRh*#v0K|? zXBep7ab64W3H7VDe3~(%Tly*??=SlV_o5+mJ3l0z!s(-B^%(OaGLnf+CN-~bFjeX0 z#D^Gp!I1`A#muCVOCX%pks!(YYJ~%uTrsyi;41^S5>?f3`J2YzY1mR5q3Tv?bOPSj8-qa_SwgD(2DQ^3zDFE>y1)?A) zr2gpSj{7|u;&+KUcVVIgxMU(Q!J5bU-Ec_*V~5DMk>75W%4Ju{gX0^1I?~3*3;Mx$ zwjE2E z`5in;SHn8J*qC%sHNBm*;OPP;zcWgB$O)eabgQN+})KN0IU^}?9*n+IFadNjGG zUniT(_#McZIA3p$GwVxL9`g=Qu%ZZYwfXg6Zi8t5CD7p+5IPfP^CT49l;~=p#Hfua zTCa@nLg%Atj@LALd%tg99Y+Uu#Z5OsdH(LbEi`?04^r5s%V$&^WiY`mDK13taP_Qh z2|>py;5Tj9 zo+RZxcV2p}-DIH5$#)Ot&s>@T`P71}h7u@n2e}NclqA(QP!nE|T*PW6N%ieK=@*hyKZ=zN^ zm*c|jC#53jz_k4J%s$q?bq4sUr_(pa_c0(T&NFac616D4{1}`8 zxEiV*h*dXQo{^^WrQ%!_>>DBndXh<@?Gb|h#s%CdrF-|asCW4O2mC=9B4x(}WZu}D zXdP1nL6~tg$w#BA4~hA^VKtS-W|F*qUjouGd8$6G!2CPTSRZF>@?LWjb!%N*=9Yu< z1(dk+3wuRS^G&T-T8>P%wrLnD8=yesed5Xrt!KhaHr-I^?qI~-Ip64k&p5m#@{?}r zE-56gshe4yEzf5!qpS1YAWZJVx3V{Ck{W;u#i?|Ioaj$PeR}0)CmnPW;HC+3^QpJv z=_XmLW+;2i6oFpGa~Qxof~vhlL~IwX$9Ub9&|9cpagoW_1(0Kz+58f zR-0-_@Ux6>t*u{mNVo=&YuP(lNb{5MrX89;XQePaYO{?#sRt)6jmg|FoWi4yY_K_f zkeC6DCzy^V0SiyoQ^!fY9Vx!spTHV ziuFb|9JuCh$C3UYMssTI68IhR-xVzkpzvG~e0|J8%2)$T&O>*7qd5lXspt2=<=o`1 zD6rg<5muE4En^`$)?I;ysn(y_9{O{IYw8uh&+JM>-byvo*xfu;#px;0U>D@{tLc-N|UTmVH{rcv0=7Y9Qx)Iu zQ{TeY>pjQ=_v8;_hY)p~FLRE)%~ESs$E?L@Af_o4bzWBJpk_|1BCPKCPO7)t%_k1k z;7@*&g1Y27IciBwcM}Y%fWgH?`K+@lR9RdLggh@j<&hgokBRbXK9U${s0#nVb9Ij3 zwdXFT&jn#=-#+47-4BJ4Z=0P$cH)E4h2op#^)-Dz=plO9(+yr$*|^gB@MeZhv}j;; z-RCMWQq3abVOg|z(eHS5cZh}}8|-n@>oQNU-q()Me13?R8fHdhIQSxy*1QsYZk}$6 zeOgiv_xjBK_^A=0ZN*aj)&*2YPvoak|Ad&_mf}#&(CzWU%H;U^A^be(<461*RiPt( zhgv3LA5bBQ1hc-)i)<-Ipr=BOBYyiOb9^>L@%6Mi1oXb3EzjmtIep)DoXFn4zy;!@ zIO7;pR#Rvj92W>%Nsc6HhGp2@3L<7xo=DFv&v$%>*tbw^-qHg%Qh9HQm$8@fXoFEXdt~s+B!|O8hLi0kfe6br!%btfC~I@7`Zw+P2d#@mghYlcbh zrp1Rxooqf-Hf!qPRvsxX94u;zSbZIKw1D(BNngNC&`uNhynF=-<*X5R``JwJbfhFjSoK{=H4;hqFmiw#tE&Hx#Qu6`$ryk`Ejq#yj`s zLiIf-N$mGB#qzPO1Y3}N%|ffH&sc3n=I+?XoSGChcaQto9vR(xR;2cRT1mBOsC8t7 zv0D>YaweP38TNSyWWye~8s9g=;X`KKJ|D01B9! zF0BHW$`=zYTy}v!aoqmFY;TC>`OJVXa|#-MYOYW-!oUEHt(cbp#57T`a3}PTIc`2) zcN+A1d7o7#oXYOV?ynAr0Cu6;Lw~ms*-D~P;l_Q~_%^Zr70p0XpLtUVVrm)?kKsDD z_Sf->2ywHGt{vaAQ{EhY!eASQzc^_027m?phY>&nobE1&1i9s9_nU{^TUSLEAmfw% zF|V^_gw5WILG(1`y>IL8ZYWVm32l7X@1)B4r2${*66?dK$|;Hm3&Jv9fFz1T#@}1h z)EKXj{keEh%JNmgt?AD~m;)k@Lc1HoMY}!n%!~TOHWh}qaQmw%=TrIC_jlYcZotUm6k$fg zLPQk~S!nnEriRS0kF|+E)J=g3=$qPd1t& zJh@14tGn7(0=!H1m|ymc@J{0+vqw(5dB)mdMz|L^69^V06up}&DiBlCx6$Qx8f|BS<~YeGkKzSi z=K0s%Vh5+lPuKrA=vqMZVxmH51!vHz?$+DQt^DsO2sNn(t0gT%PFNMr_1T<#^v!jv zUhC@X)vjT(yl2;^*F0O|2g2+NQF}j*E^wK4*%f;4)n@dAfs38e7rt(u8CL;Db?4qm zG!>kWf$%wXtF=!G{uA9GUgUEg#p1A2UGl!`@|)(zFBtW0lCQ*zmw|lWf&O%qK8u?jHN-CA9sA|v%@+vtLBfAFlB8`Jbl*+@U)brAFc9eaAfhk;De z3hJ1QFx;SopFPl^CpZdj4dKmyy6itQ+e(Fo;Z=KTw#Xx#1S0q@w}s6-Vx&~{Ld#G& z|JvSUD9d%}_>jv8%KFaN){6%_2i%7AI?CQjif3ci`hgXBr|iKI*;mb0M_VT=uu)2l z=RXJvtLODF7C*2wey`wmthRq=(%jF;C=9At(Gw&p)d$W7WjYI=?xEwRRGc-C)D}UK zU0w-Cb67u*g(@+11tPB5!LIQ`eZ$=#YyE#>#FH)cGguh+O4n|DmZ#jBYrKlaP|T?yi;HQ z4(Y-Cw|h^XQ{G4(SQ((W_qzFESWoh>_jf}Ix+TX*bPZg?7#WiXnt5zpJ+FWm?lj5W z{d|EA>d{y<=a~a0+{8t==l*I0vnRHYfjyA?q;%)FpG3kRL*_oXC5R@ zfeCR`T*(kAXP1k)VxPDY|4d5A!Pr+&uU*WoV`L5QNjIu-=xb&aZ0^Y@0qK37@dev| zo3j+#uZ8}z+3yiw=ya%6Ay_9#Jr&nndmgSZZN?4{_!1l3fY%;(>bZ9_l^?6H+f_eh zvB9%kDlnLTC7hoIQo{F?mXTte;V|<2AIqGsfu+8IKAQL85Rh*U1H}4xl#Jq2pTT+e ziQX;iF-|z znmS+l5?7LoHtW(F1)U$~)iB=k(NjgHliME?FDlDtZ7P`R(&v0s#Xu?0Nz)36 zby-(x)uR>K`^FC&q{nSOVrQQs9#WgAZw632!a8dlKAX(s-^RXBzoYxtcd*dgXRuN0 zuLf|2`<{fwZJhME|BBR2>Z|PA7W}K#8Sw5ByX^5Y1`2sCH2xUny!H~U{?(eMzB>0iS8DI(qH67{C30?k zw|xKW%OjJC9QI#RY)}Avl`(ZmW8af56Bw-6&@!1z#=^` zUCOu71&)=++|O%FC?d`?O4BGs-?904JL6(p3r2Y1PSY}D6l(gywux&Bx3$n$zOXw@ z;}@x-O&Lh{SgIYe87}k@h2-dx@`0X@p6lDv|ZqpWYwl zm}oo%3~cHBXw+Ty58l`PGUF7Ucj3Jex&wXVHoD&9A~ig_sfBOm-fJML+%Ert?()q0+RhdHndt8A#GvhE()QF~HA^;Qnf+6e~F z{2B^!V)Qd|6ppB^LFmfyetE2e=uvV`$#% zm5mKMYwM8X&BmPQ=xFdToiDYuX=t=^5H6g@efM*qOBH+%kMHa2YYN0qwC?VD_slHa z-qYbi1z$s2{ifPn|E+~if_t1Vm`0<;_#+A6MtAEt3D!f;(++XTO!tWn0L$_+|I@7&GQ`(aNH~?Dl=1mXvo~=hhQgU(!OJ_M`6oWiH zw>1Os8x6Hfu)HLbTGxpSfP#|4jlo2=g2F;P*p^N=)(yTfec!TiIPk(gxun&6u&Vg{ zT4i^TPN`AAAP$Es+^Cz2TaAfR_pjsL`>dYD@V3;Vem@42@rLxHY{G6~8Z)OHt~Wb7 zyYXk^Oz-nUd)7nzYDs0+=oJ=Sq=6{U=_DSF|M>HyHt#in&p^gS3As(84ppWCHXCxp z<++z%kPrZ_u*9n@W_nH5LpLW?#7I8{TZNLbWd;)t-A+dVBtz$?xt`u%Z|d z6|jLC_W!=_r3Y7m|MO!bVK6N4zfV-%2R$MGeVKRZi@7%$xR%d%Q*Ry+$!`>vX9c-m zZ|}9N%uJMg{;sTAD{oe`nOj4(Z;u+AMV8)C+9#Vt@4o=0q;Pq8xj2-l?|r?u==bm6 zaOr$+XT{XA}Re|G)chFFegE$ggPW zGJjDD%p#@VdEpb~<5dn0j`|zOo!DLZx2@LF9=>8VurjHYD6Zdh`!p-4Ut?cl@ap->K2^xBcQ8TI{wT=C#)ilN`{7Yv3HUz?BPq#mRO;HPXxzZ_=egtl;ewB0 z=hrif`9nRd$(Oi7OveUG&8+B7R7;JHJ2didfnltEK@WDj0DX|z!=m#Id41LI7AK@~ z4u9WrXZWKmC6$zZ{;PMUU3L>#2-#M#<`$yME$43C5lTr*15iiI^fY+bIXI5{XHTdQ zI`ark3HIoyER5Z>Ho)9?(9F)}hMK09l%s{6|KmG8^c$ea9H+BX8Wq}Q7f&+3yeWNr zvi<#w-OGKyWuGda_mBU40uQS{cR)Mc6tKVMH<4M`J3Krr8yc#S?>QK$axznt50_l_ z76%pIGBUp6I1TF;{{H@b;&h60xWPf`ydE7bfgJmZ)E=a1qI`El8G_1TxH%*%t6NHg+OG?+6Llneec3F6j!Z zXPLjsho4oK7t@9e#SJREHIccco%ELC=}`wa?v@&j?N1lZuHRiao+1*&Qjh+Jd^4BF1TrMu4VntX2eK30#FbH{E&$d`lF^!4?viW;XBYFsD$;rSaG`WjY% zlQW2pq0cl;{YCP3odDTMv{san+1E_7^k265nv#*qvPoO@j&+8_5!6oqIARZ6seOuy z5+V_9??LyJ|7vL;h6>Z)-+%bu`XwY`TFVF6Oly-3$Jqr`qFE0ryJu8Rnp6~%_52M_ z8;JWA9TEYDB#zz&tR&6SBO@)nXI`S0Xp~uAUapRi3>u&`r#DZt9hr)cjSbubA6GyQ zFXd(a-%cQqbOL%BW`Nvf`kcP%r^73H4QbxrS?5q=-+i2-Gu6%zW4bsP``jhPrO_r_b zhb>O6tN_^cPO+PfC9~avg!S@cpAdm_bJE9(%6UQKX&&Z5u0~*Y;QBXP|leQ0DfKA z4>5U*N9Y(!b=Q}g)hb`h@2|1%&L;LXpUc-bedT1ovbtJNw=_~|-%Z_(s`@d zU#`4utd9hsL4=(g;7v$S|D8-|-k7LIep{>PAc*Jh&P2i3+<)E=OEszo{Ff^_S8W8*JhFykhdxHD3cf=DRO4jfBHup+|shDX)H_fQA~fj zpwtN7Z1QMls$sbxgZt?hK%gPsOelL{w)7JU4VD=n6YZWguR4HF!~tly$xR%`hSRDH z0*0i--2$eVp#sow0oJ$4cv0z>-aoB;jeEshe2FB6`i}pJ;mD_ZG0y=^Af7CkiO~YS z^^-%#{TJWr6w;)fbRbFHGeesrl|ord46xwSuazsuH8P7re;kL_oa{A3$N;=r_~a^y z>@;WwPWj500^;f>sadp-CDNb($jC^RB=@P?qy_s1{!do_xl0q``3=6XrTd^cn1WDR zS-E3FjUtKLDN0jhfDE`oP7HoL%lzrSOe9OL7knVqs)9!?Dj_4`j}NT66b&F9S{fo2 zJS@l>gnGv-l00jA1o>GEH2t>)>|*Ckts-lhQYDZWUszHI^Orbl-O^HRRdJCGBh&EN*Q9EeC8(Stih|-W|Xa zeS4`AWmNi~#cX6kDdq3R0EO7-1AeDv>Dea0m#?>68{$QBzf@}{{a2gM6JWiLY1v@w zK2UG(6W!@$Of_(IHwnM(mA!rhCAAV=b+xsX0J(^g?m&%)0xYtT?!XN%mCGC)X3XX} zdPjXgpo6{Xg1M{9sI!V=8&&@q*SX2P*%-+U9I&vC`=_6Q9NSVOSqDg7NA&d$0v9d} z_zb5PE?j_;>uQU*$oIUqnL-D$|2gXou&|u3Z3;*?z0ZUe%E#yD5B~Oo0rh1}sS6;& zbVbrSpg>t_r)uP9gecK6al%k-77okiL5BN@)m!gvzhcKo>S}3e#l^;+wnbZd zoow!t?DWNhtA8%BG~Q8C3OojSfvA|jyG???YU6*E{*isr(WddA;;n=3fo(gHw*BOv zQiJF!-(Kk#MipZ>A(BdosIe!?)mthlUo!ADWTo*Ap9egB-n zReq_y(@d2h3;ltlHuL&pW5ZLwM#~+Iz~W=$JuSO}RWn_ssYNgu;Q#MO@c&E!dITwC zl7ZL)*hVP8A~n)juExXzXwF7)`OlSx$f2osdPl<$^zJSIy<~G)^-Q%@VMH;&=-sUz zoMhOv#)(e^14!2F`6KX~sXJR;Q3p*;k<2P?UZ2uHk74A~jO)6F7Wj_{?CxeVKnxvlI!5#dS=LMZ5nLAG?zF-Ogw%a94 z`Gv-3Kx14`kK;Txmrz{!ay<0GF8{A6wP~Sky)z zuyp`pc;z!ta8?QvO>^x;gLYr^a#K*>)Qo9IRQt{Q%$sKG9h21N$F6 z2C#?4ci{jRH3;|TqiW8ZDEqQtvnhbyIOYGP6(aI;P@M$j_i6o=5(!!{hR1*DBPbi@ zX4NX%-6RL7ha^W%*W~>wSg;l>2w8jyuy2o&d;4m3?f_Tj4#-Q03muQnG({2_5BNf& zU^BigU_KN7y_*5^p`oEur$rR(2cUJPQUC6M)GTa1Y4ssf($Z zv#d5<4;VXNL>8QGJ0g0w$lyxjVs=1!608s4iFWYLI!?1=$Z)MJ<#0gtsXLKVgQmGX z2l&FnY&QV*n3L56ioE%YBKMZK4_O8rt4*f$F_HkFFy8sEji#PZISF`3G)1EkzBXLG z43KY~0B0&s+hR1qE(rye$j15RO#pG*2K>>}+qfP8-Mb8&R!2-QWsrDB1#x^5xjGrJ zd3yRV_lavjnGkp~`eAO*5Za@&Vl9bf%Pk}aAgtDdWJf_L5y;#RuxW+bSv(g@yjA?N zzS-6RAl!m7C>L{Gr*X-M6PSY+V3(-c^Ah&&jSqW8Q!nWu{nK7kFD+q7~I z-g*YmbBi#mp#cC;IET`S1Y*mbo|&n5<{V3H<)KQr=DdCw@dz+qFsdzH95CC)&`!X- ze*<9!$o~}W$>#h3R`(8|HPQU()4F*$w=qKgJ-C-wS|&l``ppKeLb{2x!fMUcE>=g> z)A|51b_nhFL%<1RK_-%)As3VLo$)K$>?o>;~1CD({-*z)?P4?l| z*YkH(hphk@4n$~}E?^!kOk91KdoxKdoqY1liPBf7i_3a&Nv$*BD;s7@h(aflJO*Qk z(e3K&JWFXz20*Y|lHkER!zR)l6-9tt)lz3m(TSa2?P{u z1qs-2BdRVF^?W+iE3hmHWK;Gq?ts^EcFJWy`g|iS$N6d%fcn}*PL)^!$8wn^j=)8V3mhz~YUA zLwmsOT^1=51E=^YDRAb+0NrZzP&b(s;6|jgtGO%*_WXjLhGtSylF9?+wflhJ@@z_# z>OoWsdsaEkpUyy(z$iRGUS7+yKsN-jOaq0e%m{ZlMF)ET`N-M*Wz=A4Rc2Hi>{Ed( zEiF}Avv(*m_zv{UdB01|1;9uI3=IwjN$2mP_KKRAr2v%iXOQnb>%8wp(gL-+?c9fn zzq`9tf=>T-I}dpQ90ihI7TVj}Q!1g>PPMJq*FZu<8i)Xx^-e!ZC4aYTFCz`4UJn{c zwj}Ok(FS03LsM-V^#FXXU109mMa&QhK6f5+zmDuG??c_XMn=Z6VDYSuO&5}JN$$1h zhU2h{!1N~ViE1)1Fzm2+bfz_)EC5u(JK)s%1N{Bs;i4=ani|Vr0I$|&oDL)U=|E_; z0%XRDtfaK`6xiVU&<8=@Yi*cdYGv5cpO0gsk{JyPDgUb$KpyDz%|v4sGR2SP5=6sP z+hUW{(nS2V;}yW#8;V^t9OrM;y_pDL!LM~x9E1Pm zS|U{zK&oKXN!ho-QsvA^FG;X|zE{rX_W^?!5dIKg{t6#9Yx2$(V{%@Y_}=cQkfd;X zYG!u4@i4bMCuw)W@%9y2(y_7~a2`@kH45<%aLIru*d(P9QmTOYgcX+#LJxjk%Ret~ z2RWT?rG|2Zf&@e*P*7V@Fz?(RIWluPI3KUF9&$+K>({UDy{rsE-w$SEj)#)!`&OZb zy4%GlLUq$WsPgQ#Igm>(D%{)U_M2#=CMe(vl6!Ohfq`cLkms|oAdO|!&PCwyAZpBk zLe@{UeS#c9PP8uq*}DmVb5m`DzovA9xgY#llK5W;Z%3p5m7M-(?;%{m`MbAG*I8vO z9+;yIN$Ql}g_`t1Pv(pdB9N6c#()ZHV@KQI9Tyk3U<=nf;?Ix_z}j>d{DY;@r#kQ^%uw%kDkb zG!QZ{Zg+7UGscj*mDeFM-7L;xp;?MjLOVh2M#6b!nsE-?P=0gH?a3TM&5|6b9=8*_ zfa<0qxS#F|<4VQh5|Gg+5@@9(9#|Cg;jDjHU_8!+{Nz|KdO>wlx3NBU2DCPeq=4dG zB!w4g1={!Dm0D=HXWJj4qNg3~-P^1#oct9%Sus{gAWVrfKSp?yribdKy;nHbEnK?g zG5y5eV}#8p9A9@QE$;`o1E?ys+l)%wrbIPT4bi^M>9EObwl@yXY1nMp70Tes;a@%h zAIbx>;^Uwcjb{wf@nHIz<2f3!1PyoZOK2Y|^cI&~P0h?U?Riw8ht1KjOr$u(k%Zeb z)}J<)Lf-i9s#9YllF{b$MnCPTfN+Yn`1CJP_@eO1919G$1mW4PZO6vq;a5+)T@C@% z1t@>Cgsjl8h?vb3QUrbR8@fxq|F-ERu}#v1;b4^V&w*e`je!3Jc;3yHYLrwKd<+mw zgM`fU?fvOZQ*jSkT*3aa5RLT^j&GA^ObK?8u%X~iq+}k`*wp(Gp4^NpC4hzLlS-QzM7CN>grNFh z#1^c7?VA9U9L$&J!_`^bkgA%|jMv5zO{}7sJ&cpPWql#N-@uqY2v-UEVmA%J#7jw* zVrTIkqBM3sda_@GQKi^)cBc~B{OhoVWLBTsc6g{!5t%MiE+H{CcH5_OcAOKX$a z6#Z$A%$CP*VY^>g&8YmAP;BW$F`aq>*IG2}!oP$0YV zuv-`+Pl8+Qk0wnaz4ZAW`p%R>`5k7Z!5ksKF;PD&9EOD;5HsTx2IM=fEp;|M*;EJ( z@IjySJs$ayWg)V#b5$a}-At~enB~MW@J<=ubFsx!&;&P~;Rji{TTO{#I0*VB+sT%# z);%G0bCj7)y9R+`IdKnsN11^K1%{AO$Gwy~`SZ2l%vIg7^AXJ3wf)UxioQb;8Z)=t z%>rLrO#pP=u?w&H0U{Q^&%196byul7I`^`-P?4i{96Ye6hB%+<{DYp(U@$_Q-=*eA z0|*~q=icwf=-RQ3-VuW5B-3;MIkFlto+~qW=w`fg2EfN5;cM~xAW+gxlH{rhJmncz z+dh@_WrSf)vILdDXKMYw_2Ajtd z1v$L|tvk`bR$GL(%mQ4a+t&Sj!TV-K9}!^ABrr;{L)>^E7h>E6z>>B8{$XAGbNDWQ zPy^Wf0!|t-dIiX3k*YyET=)aijdA=}Wjiy`d^o+Q8FET9F7@y%s|~|&Yco! zu>k%~kUZjO`bW_^$7f%eqYm8C2>gthe%Q}Z`W$n)%i%xO1R)LoXc+Q3+`~hp4*hLX z3VoV#IYlpfUR*jyZtKUvb%m}9oI;S&gR4k^&ywEh%Udq_~!*rGz;3s_~ z^~_|0NiB#s7ZcAIrFV=zvNy2lSASM9AXO?kc%T(1u+IvLJ^?luK>d=&1Z1oRUf`ga z^G)dqf~_$!Rw4SD$aC6K=L92>07Pv$8LX?Pw2DV%*{h`sSh# z(u^#GLGsVf=f=m)?QC2qyX5aWoSw0+&P4KU9XLkUKYc5qa1oS6^jlO=Zq4h>4w(dZ zD@##Tb05BI$lNCPsB9`d=7sIl^31A$Iocz$yh(W*q>E~L#96OPxR)VVM|-0eyD2;T zZf3KG0(;k#XK%^_RxVBCikfh8g0o(ZKviS2qr<{SUx)zo#bKpv=I3U3rK0wnexupm ze~$yr&|v00z-*yl%~*BzZu(o3%V<0k)>fl%-mO@2XNJ^z2w8ce50I6L;SykA^3%{5 zK?P_I4DEZi>1Q8D1k&O$JwV!wW&{5~X|We^hi5XSbE?1tX!=+Hj%$6uq#houI6 z)n`?KMnQiCvwIN!omo%TgQIlSX15ZUCmUWms%}>b0RUc`f{%?Otc)#u2k^iz4_z8lbA;9x3!)p^ZeYT}Na`iDVfDhau#X3WgfbuDs*-AyqL{Y1^fJ zdg$)rH6h*r)fFTUXwhA2V=xXnzDCd_iN^&{+@OblXBJl|R2PGqx~?DYr*M6tfMy24 zE?nCcWud2Y@}%bCl%e%gM1LZ%bGl}`AQpM_d6amX}C17XKUc5Di`X0 zq@D%5B9yisubwvTg*k_LX*HRS>Fj%$8;DzCUlQYKbTZ~A=(_}i;k#e?i{}0j?*`P& z{1J}nS3r!d#c%*V@i)`Lz@>pfvGxZTp$f{A{c%t8$JQ^wbGhvCgkPVUH_7_Q!VT+Q z%1`mF%+M6$-n~EjlO?yPnQf(GtDU40Nchi)ntRK_JWQV{ely+=W;WQY;aCp`Y;UNo02g?9?d8|%6aqKDczMXHKK!!{>_ZnK)e}ha-e~TTGm{YqyR=xe?zlF^ zW^wzh>cA(;RUHv58uE^N)E(~vaGT0MfDxDOW+Tgs?=d`OQ{2?(>et-@kgUZrVSB5; zlDu26kcl}z>P;PQ?@JrOWP}ZJxeEX7O$gSCJV!);u%{%8GtRp>b|sJBC0;Bvp3;hX zO;=)_Y;WPfWV$tDl|fL}=)=l56|R;@h7Ydyv7&c%gkKfE#G1^12~4o$cS!`!MfkMN zugEVJGOOP6Wazzf@Hqt>p7}>lwff0mBh$ePjgy!r*}VE%$@`E|PcT1{E3fAKrVf^? zb5DnW2`+{&@XQcuoK^W<>+liYAHYmthhvgYFaC8g^_`;Wj9Uo%uYlLuj2Af#BI%~b z)Tep4U%4rsVde^Gc00b$Pgri4nuR{qiqAZsnzup1Rk&7>0#l>Pu{#Zf241Y_2@79& z5`4^z8`#&dtgG>``QAWUB48vOFe@feRHf736I#_hztP5ToyYvQb=sqTJE&nqn9+!5 z-(@N%EZCT5hS(0KsBh4>CW2wMaWc@W*(_LOdG>e~`>k*FDLN%@nGRkbd7}P%E@|u9 zGL08?HL&sb(YaGfmhN{go_|Fc*O<#^5Yss#JdcYFK>v==iq(vyljc_M4+?P7V1zC> z^xi9tXBo-f?lt|h9OS;Bg}r5;4LWE%8%zC~4>ATonfYA`>OXPfSO$a98&`s!s(tCg zSYKjonW6Lg3ZD0K0-(|0W>Fx0P{MWj=c1=n?kba-Ao5s6)owxwT^Mxt*lih=t1t z|2HCMAo9$>(}ie^UwZr+?zl-;N?G;us2wQzVQJLNhFGKW??t=))|dC50cXULEiumM z;Icb81_Y_vtei>Q1*Yv{eqyC-truuJ;G@sUAPpoL&ML0|1h4ub^m43!EdAr<=o8{; z_SVV%cQymflZ_wPccRyC07t2^;!+uT(3x%D|AVIUj;Hed|F|u?57}8chtQ!2Wf$U* zk=?N)duH!Zk*pjmgvznE?3ELl867k8WQOdlh~IVlK7N0f``q_6-s5$>ulJLHSyRfQ zPIuoBKfii9NWqOiHs^wXws-$S%3`@}(4KKsO%04k!A5f@@Leq$Ta768qi=O?3RZTmAGOgyH$q~x6giX!;Ukv) z=Q<(?X$6Nm2uwsX0M$?r2gTK3BY)fOs(}{8VN2k9zeoioec)`pd-kr)# zKp}+HTkG92kyp|xPxoZrZA^#A7;yVgc(MWo3z;t`)dE8(gERNDv&HzGE|1=gC_Yw* zeO5d^IV?oCUEUxjqX6>R0VLqzsL!LisP$7`E zcCrm8X`zCPr342?mEXTdoPakg-0xgh{vioO=pzH5pJygFEU{=}RpQH4c3UDEhi55R zoc|ruBpx|=b9*yGx2wdUPb7o$^xRzRYJ%|WxV*lx{Xg$ zX9u8&yVkES|5j*jhI(@8q!w<(=ie(9Z*PsF;eRob+Y7vL_?o{B49idi56=AR7FdNT zg8~9hq!O^6d77Kt~?;HMO zsS;Nl{3X<`tjm6IC{jx2@xJX1Y_$Y3tcj0fbb`{Rh+Ld!#qVBX)<~b;uxV0}{DWc# z`B|26+(ic=BuAOtw}4Q7!as()gr;kV_8>of?~Fr`l+Hj_)E2ZMfMw#VzWyqv*^63! zVcjhxFY@-!b+bBk3gLIXLgHhlsECXHGcd1;w5gA@f%V`VbUE^0i^H?a%U!7TE7Mts zE-HuzGXOXaNCz>H%YiLW$I#sBH+RbNo7Lkzs=OdclckZ&5Lm0&|Mk$<*=IQkd(<89 zHy(n@*6KNvfZsxoY0Pk1!H2t$T{yePO+*#=aI|6H}Py0d-r$-Q){(2Zz;`IST zf7ymQw|tz}_mPM!<4Hn3@;&o)GIclp|5{@cVfK+I=}z%Bm+q4TEt;%lqWzL6!tMvg zVBQ{2acK#G^(J)}Y#8s7NlHl!zVr>9UXvSr=U*+41bJOdf>)5kp|HFD&AD=2U zzXkUCq&pQLJGr0+G5(xjlO$w#vqdBOBUeeT&ay)MiF0dUQd7NH5#$}&!P2f?uqJ=v z1Cv9xnp=6_InT{>_P_MNm~mRnvG~*pYF3GHt_=vW7C;djiWU3Rl)b}$Ts{4~o~3)P zk^=Qgk^tV@-nK!gx3ZHz)V{k|)Vs-VyiwN5PsAoP)aN4qj|B%o&Z(pH^0d?nXINL` zuX#!#tilt%t?P|Wcg@VT@1Uyz!4fI0_D^VeG7CEj&9Qo`vQW}%m2d1}&Isp7oqt6K*kP|G^F+!f~Dv8nJk%ufxa;(Pyj^vgnJ zGWpS~{8ktvP+~9~AAlhU>mw|BYQzGNPhQo9JS-;J(eLo{uGBjNQ0qISaqg8<767N^ z=skRmD@;;%``Hm;;Cc>|ZUZeaK_~+1>@Zo9pXHfC>42~5l6=^f%*ppZP*C>vFObBv zM)ZdjAYXz%o5Mfgyj}D-R>!cK7n; zHZ!w+i1y|LyT#u?;>$X~0k&FZvx9DqUb!VLuO)Y8k29HPwC6ZDS(E;Wc~3`1Z~O)ED9QV?(t{V{Ov5t zay*z}Vjz79+{nQ2Z;o5^QI6@30b34YFbjYVts>;cKm8Hb-1Q0OEDn8#CLjstcEke< zW92MciSf8;^kX9W@Pq!WE@)Z_Tb0pkZJ!E5hVB4_D$~_B{trS(nIf8Et`L@I!|ZK!zWolF-{B8}CO%nr6b0YZ5*WlJPVjbnK5c|gPZeP!-O=F1Oj5p+~}eZGsDXpD>Q3HdC+4FH#Q{*@-^ zoZiBqZ8lho={OuUylDNq20C%W_mfzDW&#|@Bgu_P>kWN5!Y$idf)_Z^0W~GHDE15h zwyn_zrd6! zZ(ATB69m8m`{IW*F7JL;_{W$-4c}!i_QB53C0CcN7C?Ea{bFLgG84-FyO0m5<8%GH zAjw;H_e9`ujP-4l``)G(OF&ejTUHZ4Q(AVeK zCzF5iyQ}~-({#<3Q=x`r__?wQCi6W??x&um+sU`H?abF04N$bAqtzYZhj1g5abiiS zE~IA%q}KVr7pvbEM!)4co5c}feOy-C^Lsmj$$W3LLksFK&k6nXVYI5Fo36)d8?|#y z;m2GY3L$X$`y$S{&%)Ej=0533^#(xAs9~ACY+Izg1Y(NXJEj%2KP+qOtU|bDSLSo_oeE^T9J>u8I(SBR?>Otolr z#-m8oJ^C!$FC#Oct+H7QewESI#fot#K*oU|sY4FD1If5Vs)xZ=k%g|}O&!@*lk3A$ zQP+_xpkXwoEO~}4K((1^H;lFE^wfCDshW#Suq!nlXpw;b229oD$mefDp;QDBe=RWO z_|WPFxP@hN4ee3e>6y`W&=XG?s(sR<0+5PRNnk}kP& zgG4)$W%GHVfpG)&`pzft*g}=JvVm#PJ?tvJrbcv|!o&0638=!Ih474=MM$U&Qh{Xg!Ks0%6NvD{SclYp( zs{qv3UMXrlLvkfzFK0(5|Cf;KFN(~uu?cgq+5G`J2RQ@w4vgHUz~dWjjf=`QEmQlr zAvozm109zlm}OU2LQFBg>HcJJjb%Z#cZ>*#F)!!^!Yl0BNYi|YFG80N?#;IMhOq+1 zCKt6h|22N$ZxQa#!3^IU3kPQk`Ien`Y1Xut=toqH*LCKZuHBGG_0$Y3OdJXN62`hF z!>>o)^6$S^K<9Rze1jhR|5|_#7a_lN`nu^R<0#f9a@U)NfJCd$z}F9Shh0MfE8vyf zK)e$}GDmZ4WB26!uUDAq5pBl_w(VXgxXneMbPfS{84+I_`1L@6dt(M))wjiJ|DrvG zPq93O#Ys~|KDLp@!?w8}lXV^M`v7315VWu`95#TngYsbWzMFr~)WV@+|DF*pJtx^cTxZZu^@-`K>6+Kp;*BfDe%X5!2MLc z@ysxZ>zEShe+Sx24~pCn>+*GZ9B)6+-kv5W9$n~DU!mRo1BC?kDo8iey;?BA8X%mb zgOADbHD!!L7(HQYetlk2W91FWQU*EH5wW}7Hf?XLWR zuvaw6$mHxC$>|IIX7m7b{ABtQYt_ub61E@4>f$rF!)L-0-@_$c@G z_(pV2vfHs4v%xy@l3v@jd|o5))z_Q0ysI%s?e}%MiqCn=iJ1rZV)@7{-#-gHu6KIN1sPk@j#N2hCgK|D^n}U4j{rkc zi$LzJbMiOzgM%)yvKM>^*T$oRUs{gMO%2|f&l?{TIXLGt^6pqEO}HuH>_Ip=^1bS* z*;_mX!M|h~cRsI1l7}^CWqH^4@*oy{p-<5IrE>mTNWbGBz__^GKmS}@(h3>`B`RaCn9N{rZJGv`_EML^B1v;C7GpbJm1j90L0##%Se&^S7{#% znnduVz|b`VPv8Z}?Pw{(`Ml+RQ$_8nvqz>j?G@rq!=T${o?wmsR-7+Ry(K4U_fTF# z;79?2&67d%gn>@m+8habDZKp_HBIfnnza$b?=#PF{sS$qfzn`fhlQE{h_L(cwQklQ zo%kQ(@6Kn)F8TIPJG)y}0@yrgcMqO2sxQ~}cd#Ih=laR8nrQ(lqS8RkXLm})_c0WG zkBG&UK0~Z$=M2=szhKNB2|M4RF;MG5F|MOMuj%rkr3QddqP1-knj}HX`lhiz;`K*- zFErDID&UYD(A`M^4W@*@Wh7_PI3DJ^4#_IFxse>8u`VX=Q)4d;p+3+rnkjYG3>b(@ zi$0QwWPZMzcd;Wk0W@|h%rZ3U-Y<{x0?tvxh#u@h--prvp*c995Hn7v>8?$fkHVR~ zSO1!+SA&j&^Ecw6Vn#D#K)iu@RZ8cufuN}GZn4kkg2X!;>N5Ad=a@t&B+J3b&5~2! z%et~&4zki}ib+8c)u<31$Yguf{my76Ga_Kw>DlwRb=8_Hp;&TF=LcaHsSN3^t{DR{U?(^yR)*2=vu7I_{|o^A zzIDipo12a{$r}t#Bj}OfX*LyYTd)L5~9}~BrN0q8; z523exF{(dH9&_I;))M+73sSOT;N5ys{GMG*u)a57y|)Pw^x7}Eea~zLa4Pm24=f#fB)~aQChlqUcV_3W1RxG@ zh;%Hsp!21Rwlq^g17rEMhql9o-cT`kq+9{! zOm8DkuY9X0ZLV&QJ#fDM+RroS6*FY1A+)lN<&)%WlD|CJlvA$zW!|A`z;5nM5iZch zZsVYgt58uH_P^*@$3>BAZKTL8wb}z3s8`3tJJ;!5!ndDePGz8WSqgM0Zf=6?4K&2Hra^FFI$wms-Fsl`Y~0KxVJH>EoMA=F zA-Ee_O0!m;lLJeqU{-iX*0WIvlljzU0Y6(eF#Dgxtpx9z&=btvJoc8SFX{b5VmJy~ zG!#i(uM5+IKh;XAZ}kiCwrlTUMwZke>@(K2{n~2Hahf(LWn3*Ny;@O}TdmQNvQz-% z(-@k?b20+ez`1YEk3@iFKuals6UsLIH+A4Z+{8&nCtu2P8y_%6&Y(640qZ8^{wpM@ zlStkF-mLt*ngMQ6HJbN9&$NFk1vH8pF2y#55r!7>6i5}{lPdfk%u;aGTn+^7hY+Zk z2$CC|9msvffnXC^)IX`NAw50P8A5D$hRxNlDTEf53fZxDzO3cqaT zws6AVW>zKAWKnwSd(!%K+dRIy9YihF`wq20w?n5mn& zKxE!91t8Xe2?e;TeU-}rmM8L1O*$6(QHMdqE!J)9vcF{sFmeO8M*&N`mz|0VG{z18 z2I?w^vgouCnQmW{arQNIu!lXCjyO##@Rcc2I!n*NMb#q(IXyMpgN-0x-W~tV*SowS zbwD++Mb!9j4bV!*7)lvi{TrHNQ#%@ISn~A;-#Tdn4RqSf39mX9)K2&n`{X9BiMSqZ zTeTjX@4UMRmW+Yy?bl5nn*!RmC3||)-yD|cZoFcmYjhJ`n+d)3ve7($C}HU8hd`qU zOroRf@EbK~H*H+)GH4(xP?$~sFyExN`VJiWa-j(@9geoC9}R%BsEroMMAvqd<<`II znBqRa-5`f@Q_umE=)|F-vCWpxrH=Lv{~`SZLt-d7?m=sqIPkb(M3C)LL5g~F&IG`D zf}ZN=S|!%t^nJCMA3skoHDXi@1eRKc35YsDu~9A*yx_))<=TviI|(}_4=yKr(t6j> z8B@)A{djVRRC$sRTHBsA#tWkKh6=$-1vc;q&j*uJw=Y9)zhD2gKxzK(a@BH_$Z9i< zfkkyPt~4m<1s4RVngVAT7>(54z_-m}#Gl%=Yy1tRc!4u*wQr+?d@yhw;M7Q;eLk7w z1j=DLq!jIS0Z5YG#Mp=%T?@^1Mo}w$Dsef{b^ZbXG0&{j3GXpjnQ67sH-pwL%sKOw zo?7$gh7uZP)|e6FPV)EV`qq0RP#g{>rJ4qMbrWo4hNMsIV(b|_v!U?A*v6$>C29Pp3TtX^8(`?EFu zN?HJGeNs`QBUhmfPJ$}kZVHo`5grG-cOFpV@`U)`+GkP=m6z`1fE3I=8Zrn1Rmn%3 z^dck$NfrQ2wfat@uFf75ZC;WWPJr#l)QE$wNZjShgt)PZ(b+u0*}DNMN5UFwkD#Uc zx<;HSqi*HFbbY*F-Sc$7LYqNmf`-J31c}7Qq2fs6l@|@y$K4H& z$~D*Sf#1%o3(aJ-%It}%!`n?G1r#)DCmW2GOSSOw{Z+%I1q!#ch!ttt3~8F z$S%7*`X_X%4WhnDO6adsodmK*&03_TfoTVYW38ew`&{M&Hs{^y7xC(zxAiL_)o+9` z2X$yobgjWjq=0=ad*&ARJ|u|^sKbJ*`~t>bjW*$I9Lncu>Z|Hi!EvkBe*!?05MFUw zg*#{TU=5^`Kz)dG+o#uHqHI7xMKSg~^5{6#Nk0${MfK?|LwQ5>4`{CV6Huc~9__~w_i)SX ztB&hwWLo80@m);pK-IP7H!yDTu~p%2a807TCEfgUWQnoH_=!1))hD44EVBmq_>Lss zyNjbuz}nVoUX#}uzXC7zg(GKpo5T;^>@q3ll6-;InNJ z;aouJVRzYohOQQ$jIT<;t1(?5C1IHj*WEd^BTk9?UY|TD)UNqyz!=nK3hAx!kuJ)Plze$pJGj^uVHqI;RAN&|SiuRH1_u05nV($t zAJ8?|=>p}idc|A^3Zwx$Y8T>$m_cu`k2(;U{1StaQ#nZ%RbKDP(BVQC7ERY5e{c5E z2*HCB%X*;7z??4<(dIt0W*r|fS&)m47(Q_o`;ep)T%v+IaLDj~sCIhkztx1##=w8w zcpyM8c{>&0{xb7F6X}-DhlP!b$^dLR)Ra+_E>BF4}kJOz`A8k63<`Ya`!=5l;s zEnjSp#NG$Tg!sxgpIu_l&wtPOy<0MTj6WYdo(*s`Z!GdJEGjd@Az!6ez_?u!l&>Iv zIgrHG6XjfP`|m0Fqez|G$nqYMSLvZ_3Cav_hGuSkxnzCWy7TF@4VlVU9ug_4#{s2l zWdYf9Typ_UMN*>y`a$akL9>3N>tn~WxmlXoa$cJrgMOOE0122rs;$-POCE}!#I9Re zCG4n5`9L!^)?H?#K5yH`@Fw?T>F>8=m&{Cor3ZxC5nq*s0K+Y{XhTN|vbADgV!_90 zm4I7zQH(BC>X2MVj36N97@7z0%$vj;x4`BOe@&d!6St!~w6(-SbxK-QeZzwojWItuf>s|^4NKDVM z?DM!w62wXLCDk&|N3KLP?kbZI{O80u%{C|ke%=e)?WEgqTKQJ{`OML?1|3=#-k0FUDc?i%(kwTZfa<)zm6#eTq2TOwaqj z>jKumK*VQncz=;-kF-0mO?JtmzcQrKn&CowkXDF9bsEfFalE!Wrm##ji=sP;#jf?b3VYQ_1fz>N){?9rHONqm01!%Sq=M^mZ} zsP%xo!%zZ3*Q=9kb0|B7|5f&D>8PKI4p9!te_~3+C}VO+rHtbwC8_7SqM^N3O)p_`dKm+e{ymqTpMrK=bxeuBxhLG$Dre(EGY+TS3p(GVwOWi%u0MGXsyi=^eY;A3aHifn zTXjmY7(zBy|H>sP26N9#72H~H2SfHDS8tY+|0VRFdfyJ*7W&o@?M!J`hL9tpHR^}h^~*kSY!4L9>*;;^?``l z!n=IiuRn0?JO^_Gm|k6``GW_BF1i6baQwP3e`ad-suHFw*d?9rWof7tIzNXC~ zbTch7SPgXz)yMy?hkkl8#N*GqKNM(65IfWF>prTJOn*M(Hv>oZTkZd8AES50r?};m z!0gRpe!m99Dapz9yZ{W`QDZ!2S7_asf8r6?R+?X~#K9tbwx5ODDh3x!rVk9uD&n3D z-}TrZ>D~^8h4c78Iw#r+C$-rqZ zE($owK9icxj8GZ^4#d#-BR8KF|hCvqvmzmuiepocqgrWh@<*}uQc*WqFU z-so(TSuifv_?UtvD&qWa~*{88hGYygk6YXkQs>u8^LeW>< z_F!~1$dp<#L1R2d!DK1TJI5ZNR^o6~)kO^c`Et{%C8I~blqDm%4pm7(n1fKMQ-VKj zF8_Y@cSH)VFpIk|tz_+8Wzc39d3k&+bY6v`ob2?Y8loGTlF*OV#qEBoq!ZNL;wry$ z`<9no9R?u2l2bnXmX zFngqf09N5-$HuX5aikaA=N{0LwcPU{pa6?1U35Zxq2g=nmD8SFD@#J=LZDLQKDBm} zmZv%8RuFjZ#TmzguzzlPoZ%iCI!?FsDj5{+YM=Q1&e=0Yfb)Ues8^(EWZ zeZw_-_xC&jlD{A+%Ou{%&iK^r9spH=$`JdzuT+;hwupk>qyp^wbpWcOdQ#zt;@zQg zERPO@PSwQYUxEIj`K_$hUi^B2hu7_yWcf!H;5#P(aXJs`(o=Xsl-2v%hBSjIeJK^ck#6?;XNvS5P*@C4}{P=w>^}UK7O&=p6=s-ZZ+3Q?c1V0NWElIu1!b^0wNbq%S^`mXl1 zB&jPj)&R*0T5m&^UuSzg_CQsudCrJ+`6Jt?jONNt#S=4htaK&V}8N7tA+m#`#y zIboI4o&xnrRw}1cg;poA=(=mfrx7V$PR&R)8p7oxpIoS zc%8&oNy&a9+TqQcE&;jtw=nTiop6nyHyWc092{$l9A6*65r7n*3%BtNI8~wyCfl7q zVE(i8*#u4jBBph#`)yiuvzJq$xZnO3%Ph8!7o-j3aGEAS=k#+ZUlj7oKaRzg?wo?k zos`=Q%{Y3|Dugg?eJQC zDdu{ZG526ui%ai6k^Ys|JfucSd93@}+>k-ZrtR=g(+ex3m)}5-x-;1$jDH3MN{Ofd>EX4V2lEKUsfAo<^&r4V^GSjMit8X}+xW@(%RxZ&8nT~T-7j-9f<)59)1w-? znPdNf#(O8aO^ipY_1nY@57MI~e*?Mj%^BL4&^TXVXx3q1|G~I1+QBihzIkBeV~BhF zYb!LaaLtzas!Z=7I2Q-lmC{)z<~+P??7?3%I)lX9=Mjw`0vwR{2Ce6<+N~^?Tvug) zB9Lz6h0X}gC)N|KZ%emF58Nku;GZ}7c)Pe=`pHvE5Xv!i8d)#J91?F31Z@mW8Mr+J zSTiNFq7OP!v%zep&4YW~m_6Xk9ds0>=lPK)RQu-We`7Pro(GM_AL?3PUwfa59|t)x z`t}@IldmbsAo+FMiq;j+!8e!&U199YpyQv81phTV*pZhDO9&P=K(92k9e2bOZp^k( z13idyY+p0OqJ162TzwZ0G9Q~}$Oy;#q53NPi*_HUglA0i@gvoTHXilzHN@KQb+;ZoQAAeiyu!od@fiYU#v^vd0ppp-H~jZA~*BAf;g% zBz#3&$y%A$a#;;7aI&udDD8yC*|D7t$#l1vmWwc-7UkIrqo|$Zv(>PJ3ay#`x?D-8 zh5&B}=uexuJmIFiO<%kKI!&o$hE+>zn?3d*g7KKA1*~wL-oeD4JOhNO_(V@RDES*n zI!zW9kNU~6gJ|25m~;%-ZB$0zHYAgJ-}0>44{M!V2b5oc@BzWo0V+1006`!my}Jtm z535u6=rY;pn*-cbfZ(E}vj>ia#KRIyziY-$6^^T~J3klCBqfL2_aGM|Hy%2k7Bg%Moq`LW+nEz;SynrG1EHSp78Rw}q&e71(1Gp{=;$CJ<$CmhHyX=tu-BB@E z;hm_{cWG*?z9;sNw}T;op%l%q67E{+M*-yCM>KZm5Aitt*32Cy z1IHu`-?Q)i-_x-hge?R!kdgUS09zdq%sm~dSzw*%h^O0J>{srDi34^X?wZ%CQMI+c z9OWN^Kh1Z4F-1}FkeAwakZ-KSdZ&vPu3`?t zR}`Vvb=zi5$Bs^a)A=kz&f25Bx_6e8ZbTUXkB}uJFPiUV0Yo)t`8XW~G{C&ve0W zT|6d%W%|FlEPUgGfObkduZ1X#@_TpuWVH;|F{1YCArrXTW*_l^k^R7fkR z&`o)LcEG>k((=?TK&Jey9p)>zn^}G7&)ikn$J>wIQy<%Z2-?W^12+Ht!7hc@RferO z;H>~z7)fN|8xTVz&Aw41JG6dG1A?{>n}vrxgvlf!QqAJR0uqAu?jZGIL|m(h1BYeI z_gg2Qn7V$yhr}h)c=a3o=iQ^k|5*p@nCjm&8?S81PH*iWdI?EeXc%I_|B;!jzb0Rt zS0+LqBH?RVB6t&cd+2FSAmY@R=eJx-8v5pM50r`|n=yIf*;{~iW&l-tmOxuZdqMa; z@%tumk$hgH`eTFgTeFi!o^K|Sl7Gln2~86hKLxR`-unb+cXzeZgB@?WYQxtt;5Kt< zF{yzUr)x^__DRgkWKOs}A&nVxvFMQpy(70sj>m_{Jv9`c;NySXJ^l}b=;t1n?((?M zX;p%XoNAVqg1^%|z5=JwA>HJ9CwTfDDR9@cT;MKK7XDHe4>n%D|HY{>u z-oDb!z`;!O5H1CBXgiS=t^w8Gx>u2b%pv(QaGB+7bhU%?zC zDgvZjT(himrWUHWa42?p|3R}eRES5Ur=n0~&Ws<~@8Dk;5vqpRQDN`;QVqWI<*VD( z&rJ!-T}@DP8Lr3_mkP_??DZP%(9|xWr zT*E0RwjP*XIH z4>#3t0^sW1w0zAGx5aho&Jfc&@rdZyhw0|vX~*8`dPVwYBt{fNKZx38p`jB`D@!>I zjc!cBl8>7)69XV7vR~jDjkE2_pT$i#NosFq)+CSwJ)_7HB|yY-e2UUYgv)JA>#2il zDfU@3^4R^h!?x9uT=`?^M zH8_?piu>gw6zWB(1sYrQ3D z$+Nu*SbyOrQ+8o}~}HuDyd8RfA^Jk`lw`zWvy0#7Rn zLplK)hHJWK9lC(PlrSp!HNBQTn>dTg@7ZDaSs?QZnEmE{?b6EpLu2>tDy-(&tEu+H zXCr?eN+Uh&d~Kb0e6?~~EWgGuEWzy`uNJg5+OJ%(#cDX#bzR&mLeLh2I?^Sp!2jy6 z?oPg8rcuad0kl~_G**|j6r zHizy(hoh9RQvy&+^@Y`L&}UkLr1{yAV_+MsJUyAhY*ohjYIH4$4-q>7s?Tp`X7SN| zG|8fIyvWsV!$LyXQCT~~^!x`$bV}lfjh*x>zX_#y&hhKU2KO_Tnqh^Pg`_oAcgf11 z!T@qvKls9 zSoq8sc0diK$t(b}SP0VL->Brzgjp1R_kFCmsJ)o4zJ;w~$fWS$jJoiz^f8*ARkoui zAB%_DFk+wrPf8uY)yYMVF7LaCJdI5BLQ1g;c=K>OVvz!EUJ-R;@iUpH) zjzf7upOI;?c56AM!VoV10F3mT;_&Cs&)$A!`N8;)Gy)DO#H%+-W;o{w%55_W`w zJ{h*|v^AV%QG7|phqSd{ioJEaI(sk=PZR4HwBgnxNC)nB*^Er3Y_-1i=3cH-j{OXz z5MiO4H&!t~qOr9AbZi9ux%hkL_c`ctnLfzq``IJfbqzGtcfY`sk#Z|)X2||0ZpJco zqCMv%WlkpSn-Lm8V)Vpl4Bzl5QR=sM`Ef~jDxh*w0Z((W>~D|#ew}adQw-dhGu{y9 z3`L&<=C)>a^3VBXzs%l7^&-~nI=*s@vh$59u<+Fdb{<)R3Ws?CM&rJqP!y~O|2WUr-xp|B|slJ^XnyqE6CTW@Qpa!OejnMTYA4%lw@1AMmE z2=Ggk#~uIB$=cqpQZS@mf$Sz-)zMLgTAsUxkeE$L^$@oU>%lVlTds2Ab^vk~Aa5Lr9 z@EY=^Duu8SA+PLB=as#s8nzJmYF;yy7wjM5{Ta}y+y~88rR@wt#%`@gpRM84K7%Zv z=UAV%V@vf!(Ew}e=lt}eF0fGz=q!fsrw{t|_z;_ygv`-gebZ$bt zDn#1J#1bKnMBSj}e0lCm9jBjP20wB#n6EJp@%VcP+J_8G3Wds(aE3R>6*Hx(cYR#? zEtBV@< zQZzOKM3)FrQNU!9eEj?*D^w+3X0A)Upi|@N>;-^ zxOA6{gRjv_)c3`6%Xr?=`kRG6C-`hwy8Sb9_ARH$(XZ+xXnJndi)U$cMlzU;)Q^p4 zl@DxJ0O83+E2jM)`R9$kF6?a~3{zGzDL znJ9_7yiOWNTr5om0PnZ+5475zlkZLOQ*e^Vq18^mw@U8OxkTxL`cfwNH5wptl1PVo z#rEai$Y!SX^z8=Ryhe{+Y5B-9JRH)nl9L#%I%+5};4gF-8Wy`u5L zW9vWyU;hP^F{|*e=W%LbVSoxcnt=2@(l*QSNztb#Fej6{nizN#X0B##I)tyDPQxSG zCY^U;Q1g8FhT-z^-=Q_K+6R}#kvJl0UupmnkB?K_Ulf>*yd;;17je*+TW{9q2ML`z zu9MyB>k(`Vx$V)mBXF`d5J>N;0d^Os17SQz4?_!imPvlyPpUTZuHxyl3F%=5D&|bc z3{cqt^&`(S=8&M=W-=gqfXQe)aZU}WZv+5)>M_R8hd0_!qC2C*e-hY9>!Hm4SA#N`#^^`=G%A9oCF{ivWZ#ePf-?TFIVK-8!hL~+D;|Hy=_`F`^yS`O z2gw!@By6il`&z+fmCwvbLe)}2Cgxo4g3$chm9_*czmGoWcdkmMn@gu%IrRV?C9^dn zk55}GJ$=LOk6wd5vfU^tQtG%e)JIERsXy@z9Qo;Ppp)Q_0H*_l`#I&QreoSv>)*UT zM_(S~!v{#V0@vXiY(!(0AJN>olaMXqZZVIEzQV?LNlt5i zPz`Hp9un}kj+9b28r+!?uMwZh5hE?{rgx>Ncl#SOQ+hns;b+{h7}YjXt7R(+G>2B9jA3{+T$=q1C6K0!OYv~Tq#)EeeK28 zxkg1sjPxM~{&RrQ6b=Ns8ur@}^|zBjrK+IC*FPtdeHFFL{)jGaq!iU?H$w{@*Faro zCU>Z`W5qtUF8920`@22GEv$6c*k(-rxWhC-v)$kQIRDHU;dQ|3PAQ0TXf?77`L}*0 zvPSj|Udz#~ig8C6&crXkee6@bQ<3w8U<^-wGWGT4hCM8cE z|3xtpzw_8&pYOnq zSOt}{QJwUS|3HMA>97dEiF5=0*E6EbGHDo^SL}`=jI^SBh{|d21oz%IkXLZ75YpNO zO*+!z^!Jpua#S=Ee!u*T+`dO2(!5o=`jeDe%L9}#P=2xr@JtbF{e)Uj3*XX++9HzD zJUAK?EwOva$T75p1?3n-BWqs%Tk#`P(Hu6 zZg_I$%@GPl`R!CtX@9l&OU}afFUvbJ>a39@T|@vy@3Jo4o&9gC(sMVdf(g7Z>SmCn zlB;^>VG8|VY=_6CG6GHiEccZ6#f<)_UE<-%hif1ijJkBRZ0T!pptQxTJ3oHV^78Y| z@5PVbLMU44g z3R$1_-VwzX`-y=Z*ipQ!gj7UzJE%=8P2#TnBY}Ii>cKd7A0bbJl58NvOvmy2(=8YC z*r=X*QI<{lypoN{P0Y40SpvM`!WiY+yywO0tgdnRhc1ZZ!19Ye1GPv>g5N*>WQZo;ml}XS~pzIpkQwE$bCb{ ztP$sOUtA@u@UC)OJHN;bU|&lc)9(B%kIUU5`Z7HKB;1FIx9Q_&)RI0$urU71>fuug z1FyDpmjWXzLo~)uKf2j$i#PITn^M+k=2vqfu7S9nO4nu~*{b)14hDR_US<-rEJA#7 zJ4Ez`gd#*^S-&fjsT-%>H^I*xWbNI_kmz(>-tQ-E*vEAZK@9cc`VC%l(!=9Sa#9F- zK4120xFPQ{*bW<>?Y;BgnPh4MFYFpEomL{O@KTq9U+Jt$TwB89wSA&Q3OUB@E{4+I z!GD->XQZ@`z%5uW@6K!Nn|RiQoonID3FoqdjmWa3(Zq(*p(sft)o(b?Wr4P1#j z%qI!jIChvBNV#?mbn3Bv@}Onby5Xq&DPAqKJ9D@t6K-CtPo_Gr{!*qJ+ z%pdcQ-ODI5=}l~i#kl;~@Blm4;nXOYdx}>@Q|dptOM&Sf`9$UrqiRqT_0v9V|+~7O*hF=!0$DM z_zAGB=poeuIl188Ch-8T&xj{MN|)}b@FHB&M;p9Px;J=NWzgV{#<*ZFPHvY>hZ!H z2z;cfxd|PGZ)5`W-y^oaY=f`xCpu=m>v!L|X`?^|YLK%ZYlJ5AC2*Ajquhqg@%wlf zMe+8n%u;(-*k{nBF1-a;4r+#c3cKsDG)l>_kDl;iNd-ho-FVz? zRb+`BooFsn%mBHuxCt)O4lYP68zOf7ZBF99ghidY)qW}%+*W8iSH~Li?udtWC z;MsEp56ka-Q7nrwx}7e84FHEx{?_3_MQ#A$Op84v`;+y1XXB~SciI-`V_MLfr&>ni z;g_?WLch%(wYv{JNRC9Ysk{s341w^`BSG=asqwCV%@VT5EDKLTjF@QDAl@i+6HUJ! znt*y@8t%1dzP@hGVSCaDPF)TjtH8}QzCQ~424@82er0TKzph`EB8!Y)=4a*tkRj@L zG(Y2K3Ymx7s@5{;qrNvpIkm70H5MT$mGV>vT+fMfJHk?)Up-xGN~_@9erF#?vSnn= zE9+FG?`P0fM?6O%!(b=-LsyNWaeYPfPwli@C+IwUk z&fg1H301kaV^p-Gr1rZUP7Ay9Cj7be)e?awtYeeNg*uITleP?r)rW4dJga6(&TJi2TCuMYHA82ogwW==)Kf?%<$UfiLP*RZj0&ZNM>Q&kDLKm=H_W#0 zdp*zZy1sw>e%JN;{dU=|&-K~oxpIcASczG@seVwwAVkj$*kZ13K6)5`>C|R&g*D~7F+q0Rhl-_iCgPX<=;qsvQ?xo3j{c9Y!8I`#I%t1-{#1VVTm8 z`@NAnl+*t_cf!z)WxuGJKfqVX&uF_^F;CERHd;KLH#rr;wGfv-v$$RUOs*%Qloz<5 zwR6>lm&@lbKU;k_ar$<3oV_@?+SZ7r_#!=ik#x&S5xmMRLpi@ck`uXb_7*gs`sX9P zB+^{K&u!k=NI`BfT64VUrSQ~uk$dzq!o*rV)py$|uAGy`)mgD@l*RS(bF6koex{iB zn&Csxd;9RFtqJZnPM9wWak$IAzH6jBejWYmZX6?vq}^Z6$V$`P!7$fcNpz3IB|Q1$ zZc8|P?1yhFuHdldjVtHlw*KL=AGwZxZPnZ7>RV$qM#S<2p}s}5Si@q|d6`rDj4s4I zn)G!qUGj6=*D&p@`&r@U=m6AnBNN4nxc$cxwA^>jb|w%+As$F^PZI-{kfkqAx?}py z(Ma>$B)1uh$CMIYj>lMzk zH}~Tr37+=|s7JOf*#kB7tga-!-2+XQ>(TvN=9h$B%~Bj%HAFMVudtI6B7~csF0osh z<}cetct;M=dq?rNo4N@LdX7fDoC)wer|OWd9~14yF&HRkxcT-p8RmRg>+)S^V)=wC z&UEw3KIEjij6jZ;Q~#3az1GnDySGd@OOqDdCnb(9_8aLNvO2Rr+s&+1ROxm6xRt;S zkxOXRR@*ARxX1Zx+g$|PU1a$%)cKuavNaX=c)|0NVw$y=?rcS0$B5ithgD14LJ7A8 zbJE#P`97!-Y=w_t+u+^v%>+ysCN#C5<0AKNmKTY3D0mg??&L!RYbe|{HE)58#W!A5 zh%QQ6$YXah-g5Kxm@&EBEqum#)72+iQvKzi;amq!@F|#`mw+M_^s#y~h_nCUv&(-1sv_o&}hk z0zDDE1clysBU$FO@F(L4Rw^@fbzkeT)k`=&NRiS@XDypE2}_`~>CgF4N$ zack{t_o49L;ryX2VNj^38gntTRfb#ZQt;C&&r-Q^a;QTT*r>A8?&;}<+v~dkv2*;m z;hT;wV&xT0M@PplGA9I}Zy&MyjaME#c(Amz)ElTD0JTlX0!o{8f!&vKFXy%gu8~KP z#b;hG1@O)LgYss!uDckKli@2@XHlHrf3+WbsuHk>I>Hqa+KMflVhg3D)`TB5-*3D3 z=C^NWaP)QjZZfoCq1Dgz*1HkC#Q=B_3VT7e7W0rjjHL}*YfBhWM5l* zyKr!NasK!O#x~V(3X#RFHvu9}Mkj|O?h*?4;-32YdV_)8OHl`uhGY3$PaJC3qy}<`a%IZUJ9@1S-pYF8;q{{>|L~t2ttb{@#=^ubHfR?b!STG#F zpxEN$mAC;Gwje613F;M4duFd3f2w;IwpgbS4Je8G+sa5h_-B|>pfwq-9|x<< zU+?ej)^Zvf9St&==(0@J?_L;s6eF6wpSeA~s<5yyy6&R{^jSdVld0v^n5njE4^K}a zW=g&=V{h5f!F<~tH7{&;geVFU2R>(8rV_)PFMD~3_wb4udOha&-i<^eu}sQiLY%Br z4A`|4c&qblP?u{zw72uPpYaR}N8ohq3@g9t?=O_*eJr_hZ#q^2L zPbK_)oO=}JM}K_}lhg2%_}>@YOR=qtP$L3Q$F9Jxmb4?DNvLn_xX15Q@aYf<1@gbC zL?mQHPz;_u95O83d#HYtfKxp5<^g>Wc%PYz@9ddVQ@dD9YrxqWQI6np3x?Dm+W-ec zTca_GcF;Il4frhbN-BT)?ZKN))`tQ1H!&w}9A&KZPyhbu#NXZepFiETJ4n8k65y>H z^5l4*TL9CE!b(7OE5O>Lcm*GSAk!{1;X|j;Et3rG+Yw}hgA@Tr2AaqyxhG)Ieqhf< zz3{fSwqt-Yiitqg#m0RIf4-|~yv2kYyhw85N?;HEbPxhYlsjP0l?rWm)@ON z>@k@-D$-+(7A+H@H1vOVSMj&_P6AJU;5oLL%(0zLZpC~%NuU4m`N>6{jt~z&zZt4j zVy3sz7Bw{vqIx>Gk3F?FF<`?Ro94vIv7BB?;GQqR)kAE1XS0n8Cf8=`(AjZJtVRnm zx%vVt0k}<6X_2Ci)GG7(oq6n$S_%4C)^}rBN?iiV!AWkhh*-d*18j)x>FJ4&j3@+p zO7uc%Tge_{Bs=WEf$+PNTbo5q=I*AgiGI*u{Zk1|-Q~L~UVw?S| z*QJqr8YXR5eh*%7yU{n?)XZAZGfcn@RyUYwglf|L<^j}O>o;0Zu3FVcGs%OI3P(W` zyRuf4(&I98$!+etqaqv-0Gw>s;;iQ`xur_!4a>xQ>@WfRE+1G|qRDyz;CDBSa55^! zoR57X@rx8|erTfSX>lJEk#v(*Ei1k7$}#+Ow*fnNpagGd=Y}!T&)f9gpR-v-<)@Sr z6YfdPc&$U<)S}D%PRWz?!3cj-#^YUjV8UeNt4C1*0JWaOMa%Hj+`_8yz*Te0C7j@j zzlzXtPOXQ&ZCkDG0EvLFLkv z#^9Nv;lj#+ltLKis6EN2A@crq)PCPDnFCc`dGqgGGQFRE2)=U3RyQe!uZ&%%P#f(M z!<5 zHEB&AswP**R?c=T#-kj{(kcji2=29AfJ3HYW1g|M4LqV+8bEXwAnEqN?`5~331FrF zjkpeSa7a4;aa@nSvGEH?3^SJC*-_|o2%IL)<<+FEW&PyLUM0`#WFQ}3&D)Gd(%G|} zXcWnI2hQA75g3Y?y^$G-b3(QIlkL=#vd07T!DUnr)SDsO`{V!ShN@O+4)B~%Rn}2yA zy)G)u5j=Jp+TnPDn}0W2Nise{%W+qt<>ZvRDpuzxDqvT2Nl>$rd^C-wmnx`h@Y27Z z;j`6Dg0qXNBo22c!eg{m@Gml%kLYEZ1@RW^yU83gpK*o!Ayv;9t0W((_ zT%VQlj>OBEA4zq0nr_ol&M#f7yoJ4L;Dy&uAObu8Lj-2VeTc6sn61*rBZZ^8s+QAb z`zd)Bzfl4MEt7j|7DnM$LnvlchQYoTJpK1Qnx(*Nq5`hd(bl1wq*4F2fML z$`GCUUOkI5zL7S~ZkUA*y$+Ji?FjkoxX`O>4%jLtvqXC{&t;~3B=I5Ypnj7?oq!Ps zpTTr8Mq=vdap)2QXixbzIO-{CbSnu=xL|Mt!*5&}&pr=^|6q7s^{<>j(x$3wFa+tX|}97-LFyj(COYc`D&o7_+fS{V0j*vcIA_H)IrGQlbIKhP7z=jQl?5N)mH~5em_$K5^5%5U6pdHssvV9zHqNv zVJ<#oq)no(eYj!A+05B5&?}OkQ&>Loz5g$*^&{yLxUuLW?H_63L0X?lk{FHJCXv?7 z5eEs1=zM`)5DFRzYMu3eJS|bxAjOy-tS_&sqGD#;ISWobe)Ea_i-73?d-M7Y*!d>B zh#f56g{rohMfYZoG+Sx~1TI+$2ta;%Q)^JuF{i09@c6T*7u45+`!Pcyz{}^vU5<_r zFRA{1MS{UlVEb z^WGl_dW2!V&&eofoLkGnM-nq3;a0$qBJBimEiKx8sq8F@Z2d%smP0(7FiZLRo95=j zM~>90%<(`q)#NgP0p#DSVFy1f=>!kDQ8yUIwue}ncS`ljQ6X5d^AoajKKKz?FRo675fbGS31^{%A=E!ho$-_8`=gH7}McO?rQUni5ltZ}y}K63@je z-H{}u1H6J=EbXNbAFEpVjK{v9Mta@k-tr;EYLG{z45bM`&KzA<_wY!5QRv}>3#})I zDr_S%B@}2c%p@&>vU0A#t>r_aM>LE6k7kYlB#>ihinJI`Pyny80cp{c41FrU789(|Vn8}l{ z*J(gD#7bvKcl`x5=(x#vMc+|>og&B)yvfPQR)i2q=N3?J*AFzOqOFiZ#eX5Jf^Wpe zdTn_MDfz0l`Rdc_y}+7^ERe$ACvKOOmBmQiiHWIIt|+j}cl7CSBnyiaGku?nWOcx_ zQtPS+I4@I(uG}7A%qiMChs>S^z@cOBL+NE}Oh*B;w#hjQo3TQ_L+GIoQu?wYKQT)oPw;4!EmZ=6an zTRryxq$cd^>l0ktX@eWJ;vI7-6S1E`xOZNf;h4{6mXNeLSYu;j>mBjqGCXBN@7_(@ z$%%=TBVtBEhAU0>vdb z1lI)D;&;O z{MP#~Up{`g`!?e5f9~;BEPu4VyBJ)T`KR}=?{1^IqFeYf#j>?;NW0y zp-1HF$#rxI@iQ>+WgdtPK&=d6_$~4!?hX zF@Nvh;X5k1?+^c*%^&&h$$vu|vRmB$4SQK{|Ig);HSae5ca$gh|I3BDB{%NPYZ}7X z{tAgVyPx@;hcIPz_B58WCjq{JU-nl4CbZKkj*+E_>=n$q>H1on*>#vsTnY|Ba`veC zi3qQ#pc5=8<9vdCerm1YoE@0|xZ8R%(P*vLz?%Ql87JV@2k7#=~A9yPtf*i8* z81DH_T3cG=R>uiWRbd9+ER8SHy)xGj8mxtLJ6VTXF5OgN**2AI;K-%qasG7So(i3^ z@zI?Va=9Te_k#ixlu_c6a9rMki(+9I>9`$;?C$%Hr9CkBy@K>LUv`R5ad|^_c^e*n zKRA9s^w|@ah5OwDV@%l-qiJ@vhW?Wg-v2p){vX?^|GgKs&N{qNIkh6jz7g8+{We+G zxUO=-LCuVCcP>^8KC9wDKo`?8jO*T*vwoZo@@*SR-G0`~ z;qHd-dt*17q24#1VW3s`pCin)gy~#F)kv{jzTu3}Od{g|%F`ObY$$@6h&aSdn_3$l zr|S~}v)@oWzOgj%2RpO$_9ah>G`R}R+_KL0!k%ZOadgF~VsIa(Ht!8sBm#5sY=~|w z?K|&5n=CE^5td!0!9nvoD9pmH&TypT*%$`1$6NXsnD@Q-kIf3(G_<|Cguso45+R%B zJ-foz+y%k{D6!9E9UHtQ|F4M*Qe%uNy1ilw^})Q(Ju!Sy@F9LPD&SJ@3>9u{-lcaD zq#dVJ?OP?)@xLF$PV zqzBd#P2Jq5@#2pYiB3tXtQ!7RlpZhG5PxTmyN9o2C%Q_f95uyGvqW&S4ZZ9)KIq-r ze7BaNEwoxRG@?Ft9zmaFJ)DsdE6y@HDX*sNxg%0c^5^E|>V7vT6-cu=<1E*TB8M}b zHXniT#ks~lRrD%ApIkf}Yr@33+2kl6hy-4|Hod=fvuM0y=^Lr+rgM%;1fEqO(lMi# zW?yTz)pWbpbX*)$VA;ZI0_pMuZHhhQ%4aNed0wCB)uPg?tnoDu zOZQ^bv25tGQBW%TXFd@aBC#UduXXg}XPz|Uji3@cjwt3X+E#Ewl7`GJmkB%V=oT@=H?T4{r~CM|M$9!|68F* zTpi=2e|@Naop*gPa6MuN8cyWUKRl4SJ~*4qx-?8#U8vZ04mwCX`_%ILQ|D}RRaG8$ z@M#adSJ%H1$J;e1V4iRCJ%9(Tc?T&N!kVs~=dQ0#jxKkC&CM^d`bR9m`z#e(p}}I? zn1%vIFrEK-ywjb3Pkv{+pmVoTlWc?ddY3rSIwj~dDcI2bY&a=lU~P26 zN9IXuzI=o-eKy*`hYG=rZ9Iq0*T>F4416HD8ue7jW8ufWAEy6ZfMpd}%gu_Jz&A^v ztK-TV!0~6DerDV#shJfW+#3ezHekl`m6I|tFa84(^Iy!)-U;>aHJxua#N~~Sjs{At z9*vtPF*Y6ai#-rs^$`zR-23nP^S(YCVm=DFv5Z|6oGS;Xz*K3C1{BKktEHDHEm*7E zV5H>=*Mh~FU)PU~je(@3%Cb&VvWB>?7P*loFQxVOZGVJ0$gFN`Rol6bS*yOPt@zAB3!HB0o={qVobP)Buuxo>@8^vz@$I8gJW z`hKpIAE%T*d~?cSz?`LmrpNMm^Wm)>579)%QF~NpP;aP<6AvZLdh+!2bX;D`MNf+k z%zEegd}=E7YJvVr+WfRz|FARoqSJ3FN+`PC;lF3}K_*Gg2R{C81UnjuL2ZQwij9qr zGl8?tXVwmjTF#2>QAWUah>3}v0p}_NZb8k_^cSu4>o?m*|CDpwULZrv+i7kZmXwsp z*!6xaB;JBDOvSxyfw6}#AMyuboXqofxj!(wOKj%u3xJ>>6`c8yv%v`8dB$Y;=@oO| z4HSA~NKrQK^E_ZzLa=Po=WlVJzRbjN9I!M`Q~kJ+ofG<^$4pm`)UV(t9B zQlms~ML0UI>4fEjy)Yz7UPk(HW6J=ci%@ZT<+`)=GZ>vXMZ zY~ZERFM0yjn1$}zOXigM@11`UhP~rGj0e4EeSoU`r8;#b-;{(jW{X@J}ba_ zi0&i>6Nk%O9+_XGW&;jXb>Z&(K+tV`cy0@$VZpRs~kjAKtUR3N0u8Ej2riS(lr( zM-46K4Ldt_99yc3k`s5g%C4|wRTUNcSB`*FYq^|mDOqr--PzkeZM*{9g^6u{y2!B6 zq}5jYNvrA@b3A@p@xo7u39Ax}(;pZd=n{_-j-Iq4V%ctGhF%6`sgAiUO4fE66`(ER z1=58D4ck5~hNcxq$uj`M2r-1}*#7B7>^4+t3Dx+m)T{{!ET;vn-@H$LIUanO6xNn4 zc{UV-y%BEEB1_O>w#&(TMld5-D)s$1&hrv8cbU<=7v;QuWAdfuEicg^&l|fwuFM)8 zxSSd2I%r(m_x#V*2sf$q-QD4|2svr*yvuC{QrTUUYyF?WkN+9`IZ`Ru>d!9vKgrI# zwe0_ewB~<9^sD$6t9%b1q79$gen74A;-f2j4`_LXQ=mtLOX$ zTtJKp7C0UPt z?MMvYIU2MfJ{o*EvhQXV*EK2(K57z^0KhV8g0sjl$n7K~sq*N|-f%9+*7SHxwogmV z*7)OEl@YLp`QLw@xA)tEzeE)j^9lpbas4(f1EmTxnbv6M2kBRM`j+P%xXbmO?P4*e z<^b6D6*?JlIk`1m;Sp4hOv5$UJ4RIvA6oUF3B3)oP9bAT$zhDOOzEEF?(70w#=H2} zH~PVoTkWSExaOr^GaVaZRTf^Y>%BYy8+HIjLIj7G=h~hu5u&-T=*9Gt9Q&MV>xPa{ z5}!R?>jDbT$xV%&UKV6%=Jd1b6m8O>3HF3DG{2&%%eGo|f|A+Pm^u@be3cJI8}*Fu zCGIy$0OAqmyb1yKCRrMo=#0T)GfEWu?4_UBw|5>7s>!I=2tFWqYaW)TZerKF%;@~M z0Q;GNDlEUz{kN1Qv_1X)mF(8@TRR72Oo-V=p}`DObK^GIwrAS?)mGG6={Aw^2BR4o zLN%TBAoG#YL_X3nK#t+B{ye2+90+|qv$4_PRQ%D;uml@lM6fSnT$YZ&WuT&+HdO@l zGu9p4kNxZ}pXpI5JN1hWUz&F6c_%}tnFoBVebRaQ4};6Lu&Yja{jx?5WFB6viPl%n zTr%la_QU~A4FW?hq9b6n44d$c_1aZN7>Yalj&XCHjh zE%CH8ou*_5S6Z5_R3_P^Xjv;{Zi7^14C}pVXj(u$hZ)ASO%1g6B*uP64yJt-CmOXJ z;A(5TY_mZQTijF%Wqq~8s4+B@)JOnGG&Y~#K;vZI6R~D5`3Tt4b{S0Bm&%nD)2(O{ zRCp_I$@cGMTH|O4>t{3rM(Q)ENcHn^k*T{pk8De}aK}zx!B*JCw<`$JI+5w&{Cs4! zJi^IgOBZsZf*UuviBI$*tQq}!$ExB}Q*(=6rX}Iz-6p!Jk#7kDwB_Mm zsr^5Uom$di z;x5Qp>k=el0>d>QLjw5+JKhwYo0?|{o1-so`Rlr#N9U_-fN8W zuxKt6YaqRJLF=Q|kuLApRypIholXSw`80H!uy@;uVah&C)U2TGX9IR?GV|`22Tyti$ z;N>{JZ#nmP`Lcl)87(SYFwv5#TdeZb_G?n*A=b?>-sNlB(?C;`V&v`z{2|UB4`?L| zgR)0+3Oc`i*xh7ngIqQuhf!N3xQYi;P9z70Kg($b7{7Ul%>lPRkd?nmw>sP2Ds`%2 z&{ev}HTE#((x3-aHO$%FV_07<7KbA|3^wyDgbJ?=b835&A%yVBWf6ojZjRcpShmlW z?O^6<(HYhq7ZRlSl0?;$)lUHTF{~&gSdUB+@6&Ax7gSfSRJqjr+Sgq6(MY8(-}{9`Rf`-nLfZt;RD4r+`f&tTAHZVZCwXwB+P(%M-*USGXLkWjgC zYC~Xbk)(;H+E1|2W>V%uAH%qMlKYGz4HLXMS7^1-3~H$lEQ*y{btDy@#jH8cD^ROi z0?}+51fCj+vUDc(K&io7gPwCg3mW z%r&6VeOrjJ^Kx;a0;M7CF1|!CL5@K=+Y$VI^|psqa+sMRLPEgjSB1)nSC~CT{y|BO z+ws9e_OPHlTpEMpaRId&|Gm9godQcL_J*)kj?$N0uo2dkN1ijIdeWazjoX`_QT%_6 zJXMI)?gA05&?>vHH0rG>)w?;FT?SqM0*t0nuU(hPdEHNJ(5|a$<=J&!&KS?DglgS~=h5RMMJ+Awr-Cyyhy~*K<+Pv$N^; zG^zDfI!PyrEe>iAQ6b{vL&4_GIAbAEr8u3kGC?cx)p`GgZAC+pjUNGKNfe)AOl6#D z>>KFqgrmsXQ#jEYt*8-Xt$m`Y;p#%jLYgr{6H0RAGAL&Oq+Pe(jZ&ow14U|M=s(n> zX>mb9gB(3_01rWLuk_M3t-eBMEHS?V&~M6u3YDcxQU;h@Ul(}S4Mu@19rGjgly6R5 z3042>e9UZ5LButFprW!^x7#K=*Mvc5czulu;_P(Q>0Mi1R$($kZh`ZrGhX4EG@y>= zS&ar#hxmZoi6*_FzB=EK{7jWkiS%7lJmL0#|Y7gus zH_Q|6`>i|Hdz<(u6cxMe@s?}klW0wHFyn{*3Y8|!!lzrY7j}h^A**`{x+w;T3PlncFS;43wDqKn5E9sfBf-=wudjre2w|IDT zHw4Ld-ha?4U}nSc89dyrX`bZ*e`i+Srx!d!-cFvLHGsW-`Bm+hNKop~sgYoXjX7cd zzJ2nx1jTap^J6`hzV;%3C&I(uKW9r5Io>b$2vp}|(VUm=xY06)(Xr(jL7`;%97K*o z>85m;#s?=;5tT&h&gn(lPv1Ft1P1PSnb$OwhfHWxY7Hv&#aeEwmYt#v0624g`D#lu zXO?joXUnc`vLrt~PCTwqkB$bg68r~4{4amPHjSoDw0t>N+3!#6NDP%Obhq`tzyce9HLh_s^mV@`8JN0@9kFa7Ja zQ4v9zdw7kHe=Fa@duKUs0O!@Pgdb;A{ruEP9PDwsxI!KE!|YAVEZs(5 zl!o-oPyx5lY)h|q6AlVZu*=r_68O*TN@7^SZh$svaPdP`W0C18T($t=ihK1cdnpW; zxypX7o(h|HFl*@J}7T`I#6!I0B*5<#D9LG%mRS} z+U8L}iko@6@@he+U(nMPkpqRACLS zgRnnu2X_RJx}Cvt|YJMHR#f zdf#?@bW&d=<=R|&%8t$%nDVm)qR1=s&k>GKD|bvq29c`*QExH0`}NHtBTtJ^OXjQq zMEEx}v#vOjyThk2>xw7zstf}Aa3xh4CX-4Ob!35?V?ez3zH_*J(L%7G71yz5I(umnpg@ba37$t6GE{HE`x zVm0;k%}?zShI+Xd%NixiB7WBS+1}&zZd>K4`RNqL8~@B=)lnjf06Iu)3O&(Tr8SPS zmo{#DbrVB+$##1A31!Mkb#%B_Mhxifpvi?&bkKq%HZ_2WWgCcvalrMc?*V>J3N(4e z2Uz7)b(@i%QrK`vb$xSCt%xR>_u&RdRA**k>(C4)V?9V} z3>&Z&F_0)_kI)=3_J-GwXJWB#-Eb)0JJQYu`J4F0=3Jy%fgC z^+Uae`@~P{`^6}h`sqhx@G;Uh3~_3apFioFCJmu#c|XVe9L+DkcvM9c3XfQqX;gDn zf|8rZW&FQc&dkz5Egu&p4@e2j1U|D(Gs@tGmPibu`m$Gxwi#M~^*_@XP#-Op4C1fp zguqNOQ@cQP1GtF~oIN0X(tDchZ1G8z>W5nUAMMTD$@RA^3Nxg?f**6J@r%H+0ZnU8 z)o)s-VYl)hvPv1r`j9G*hyx|!43IwkG7EDcmBYDlZx2U~!PmKYej9u5ORQ6koG20p z4C{yJx;*{PCQsDgt+El?Br9`o0d zj`1bJ>+Jmnk7fd+8?CYRz*vCw4iT2H<@%?)M^JO&ky9ZG6I;qT`skw8&^ZRA&*$M( z`mATU^*E=YQfzGX*;rkBnYy^pArzIbee44Ma|?`mGpHu49h=bcNK1M{BsCs>EkD6`kW zSJ_2O)436b{TpYiuTG}2yXDlQLb5si{i4)!)5oo-A(u_0uVaejTNgfu8pQ9er)9&$ z6_#8J_YPeK8u0LF9ssb&M;!*P@$RT^<{j=A);wtvfC3^k4&zNpv{WFx7&SsF-8VJf)6v}cC! zY^7G+IrR~G1-IT@XXKLt1gDH!MrfG@%}?xU1kJRh)`UJs9kogzDttlamNN`&kwW%r90}igGN+gTV5U5eQ~NP@4cgsg%8}KMVrgEq`yKE3i$~H@ZXlyjHLfh zFMu^4Pzn0FWroB^gZtdB2K6NvyJnd~!zZBIvV@lgccgR@HdW1^m>8pS++QqAqoT5d zz6lPOc;)%keI4(opu_LXWgNBfE~a--Sy3=O;xg-u#s-a#*Ek(}SPE(7&u<)! zE;7s=aP|OZTaTfG7D;`fe`+rPtte3l-dSYS4(1!3h#_b8AEg%Q(dp_z7&<}`VMAxy}QaGapuHJ zS)E@s#K3|KQtq&!B2flx20}~N=*>TC97=pl2&*)Obi7K{C>AZs{8Wwxq<^#5d2@en zBC%&W=1TTJ?m#1{nFWFb#rHC({ptWC#l$R6(XT!>7PW#BtrWwu&uJn4lpGC{Ln$_9 zMeZwi%1V4F45yKhjxHmVVk z2rp3n#8fwcxNngwHkt-yz+f3gTMV=}c;|p*-LqO(Z_eE_uz(xAUbH@C66<#Nj@0n= z4AuISS{}bU`Do<%^)}E-U^Jx5yR$w$UOk%ev*Li*xj7{dA=uDdT~A6BsuPsJrp5?y z4#Btw=8#*sjrjJnB6Ec^qSngs}AslfoDoGwyDid&M~7@?WwZai#l#9 z2s2(??ph;9av;9`3L5I9*puw!v!+UcZw%c|T5iJU=@bMWY;bW*ac(Xdsd@CMk=NXt zj8&C{OV6}V|4~p^6V3qG*oDJgDE))-StA^zlD>HG_2w%HH-n=e^B00zYf5;i5wrbgOT*5C!R&L5YOhb z373;lz#`+Sij6a3?QwyZy`yzW6~>ZkTB^A^fTOZa+joQY+8%yvb`LI`&QLr+4O;>c z7D!w8=$76Y-IOx6Ne}FRTQnH7Cb9ouB{&Pd(TI7h*jE( z@T9OvFO*2ej^0ytjoWxCH8Iq|qvsvh4N{$i<#YCt{fv!lh^6p?i%xqSb=V*M(j|RB z-hsHLU9s!9~Klj+lmZ{pX%F~SU*{NiEYR{wiBy8uDr+!v85-pd87XDrT9_~Hc{xeiWd8NdvwDa9q)fn<84aSAeQp@}R3*)fU@ zZHQI&2M>|kThXXlB61dlp0S+nNs1XTfupan!y+fcHAm}tSzoJ;nN#8TfZ%ap^XGUS zA)#|4&K=L)tKBOD?}i9!YxT19E~yc4(>RoR!^}3X-zsMaG^TE5Uzi}3IakCvjf1pQ z)3Wl?G*xH?IXrVp3G1&?!ma}B8PzXd|1+sFRJ5a`v4pv)6k6+G_tut4$yc@Dysae0 zd7>qixdzNAre7y5R2HqcZ*r4QsF9pnd>wYmYuG~c*N6uKO*M+oH~zrOkwTwvzpWD! zg}%t5nlTn^^qnY$_k!o;ByB@(xi2IZ45A$oTKi;Yi0b=5Y@pomRC^h02bg?OU&I+s zy95<3isqC$vYnz;8-1#JIQ3RdANEMU;11q**98pIJM9vFLV;vvBspdV6;UJnopvzL2r{?6E0JQ=?)U9}=%hT+u4nmfCkka&Yu4mY*CTa55!EI=+t+ z`;n5HDu_&L*;o<$kd1$)RcVJkYEPdvD6wDp$8`S56QvupK6PxM%Uv*|4S&NT4S1=- zs9D)3Mwt}$A3L2`$JiSH-iitG^;lSsMtRF3um6{R<;AXFcu--37z2tRZcQ;L9TND= z)Ck!%y=L{xtPG+y`r7owOWx0xKf&7L2(6#?pqXTer~2&uw%6^b7c%+-;*v9$U;Vm9 zL7eMm&PeD}nZU)@mnmh+Q@D_f4;$~rqNNHqSE{Px?bqoYZ(qtShDTh@;T%op@C68n zxP;h|WNySCJq=+9F#?ZDiJ>?3HFXPs0*F`ePN6@;%R4OKE8PptYi9jgqwZF$Xs_|4 z6jcF2{)`NNw_2b~SkvhW|MZERDA%CII{;`wa;eyr@%u)FWwgVcITb!llgMyPT-iP_ zNoGtD^tU9EZ+=-r6L?_~b!Ce8{b6X!UwG;E(&_9U<7L- zeKOU2c2ZD3M+L3Z46w(Wpi&c{JoO9g8eM50mZ<>(Bfwm9>6l_>=%AHiqpZKkn6nAN zM{x*jV^Mx;!Ka7q#JZ{)do-qooc1(?o`Jb#pOnd0GCHM_0DR)K^I6A>kSl?fYa}bc zZPTQPwdaeQ-q#~1IL_EoL=VZK97f0huv65ZW!RLfj4DwkJrHk7te()&N`8U94rN$& zr8^U5iW;lNGIFu12pIY~qNAaK)bSBX38*I=>tb5`N` z3;qdtN?NpMRv{LYM#WEbxd}bIiwf*pHA@VJ(Z_bZe2|)*A~T*UodFJ25=Y{KsdY7a z->dHnafhe}|jde4YVboFdzLQ#hBUdLMIZMg~&)6XG$gWCL^*VNGt!Mys{S@zA z!Sg6Q7o_#%Y_LL!zK=@_XHO<32d|e5&2t#>P(U@;)O0!pJEn)^?^|p}eM}rnINFWtYm9`t1rW zN;<%vz4yBRUe7qZ#b_C<()PugN_3Gfc*m0)hmh)gg24cccFi3*M1IdF^)i(TF6|_W zB0t-vT;qmjJ?cz~N1B=zRQ?&wuCNW46}92lO<#>#^}Kn;#-5=SQq_Z0wmvqH8pzj{Nz-QA z`BhEC6am7wBx2DP5Z5l_t~>jEn!3uazSDX^kl!KnVIa@BNc1w*vQ;xgaYP@huUQ$R@8COG$c7K&`4>> zd^vBX4!BB%RbqXS1p_#{oW7~laQ_=*Vn^jeEf#4`sa_qR2XN}I^5mM)rz!%Eq6_Y4 zuN(1w*3gWj^UE2DA*rk0>t*?rRq&@$1+Xffr;h`F!>rd;ii)F-HEOCm!wGyImXBHB?}(NSCfRRju*nwF92H6xjK?P>S58bCeE z56anz!u{^!vHM3$x})+}GRgsOFq)M|t)Q(MqEgMKORH`BgtkQR6Zg#BWD7-oSho?x z2VetkxQpxyPH&DBdm4yJTydj$r<{|1LTa@aS_)|mXBbMx`Tn#i(HiUOi!jT`- zjS^?d9jcE&)u{0-f7k57ZPtr9e4c5Y=e8|Ew~iMj&B;uR%Sv<(rY5{^i3 z8sDR4NpbS-AoyW5*Ef&}G$6IRA{#5OPo90&ZU-NaM9tE5maBXO!{@0UAgUq1c!`)c zhLApsphQRNZjsId9}@@}$UJj>YFoBX32<&clq}ds0Zq;Gb_mG^(;OgkVpY*Sjy@Jh zkWk2dHvEQ!!(3v*vP{zrrGfGc=FxSgDby1;c$I0BGUlyNFTT^AklU(bWKUit_Lsn+ z*z0WlURX8IuVbR# zS3f8}L(5ul=BF$tYjp|eawG#h&>&DcjJVqbwo3ymM{KH% zWZeyao?T+<0^Q`fOJ`gDRiA)6kt zoQkA)G>%UMlo8E<4r;X~pb%!wvaMLmf!M<&}WJ$W@d#D~Y1RD$^Tj+dRXl9HklzMb@toIdqaxN-m9<<4L2>R>^g?>#<+t?4V zSQXl+P?3R`3(@+ML6}7pChVL_7qO_sxJ8e=r%xrg1eEO)_RqcW)N>8IYZSUuB1&Jv zjv_v(Ivxc8z32cv!TvGbtvWkB@Y@=YxEUNUX&;0a178He<1U-@uepOakPM*&iY;Mr zRVBD-?aN&=1)y?g&vcUA)dHb})=23(Z{kea#)_2XOifrb=eys>ds{;DgvP__(Gk|E z)kdQ^6h=go)!EPfshZw^+L8ek*g~-lEKn#hGn70tz!8r>IJ;_!9XpTc@7LAgx>B5g zdFybmkux!2rm5EZqF!kZkJ*-1+t@H5W`#!@o-$x|?qkm*E)5fMPxh&hPAi~ zia5x1Gu;P`yQhKNb+F&3uwh!p)S#GVJ;)dnq0(-I@HGlhQFoXv*WsR0@(xhV{XH0n zY}(U#D2FpRv(h)d$++t0=x~XNq0ZISm=oPzaF3_{y_4T`!PVC5EZjoY#sKWgz-Amp zcYka|nyHC&>M*nY^Y*ASmKpuZ*1q~>zwg<^X{ae`|HXk@^?vp9>wnTFPA&43S`)lK zWDO+d$8q*(^wZgw2o39b=hzP=CRHA*3AZDFvRcq%Ncmgr>mIbaX+}KGKorexkBCFt z1iKZch93{=r!pHMS5bd@UuJJAuyMe4shc_>H;Fv_ZF5Q%WYaQY3DF7Glz{n*C?lV! zG?i}x48CDeiEWLt7|*BeZMDNDu{)S;Mp1r!y^{5A9>SHmhgwnTC4M4h!23#gBr4SW-4(PWj)M_pO?o1<`EXH;+MS3}q z(Y)#^ou|b6cD%<1jo+4mA*_GrTysVZi$>W zzEjelME{h;W^#7f$ys=4^kaIE1FC-K!_u7|PxXfxifI`U{oFL1;;_6SkQP~ID#w~G zhvfO7e>rU2p1&+pY6xhP@9#^@`t7Ocp9a}FX7O`;UZ4+)ZYgT>RJmMU30q~i#LQ3^ zNkoHa!U^J=y7a~cM-Ek*hP!dYuQiKG7PVW#3iABW#rp7TaFmwXxS;&h<^$^A?>pjEAnYPAkUo!Ymb<5Gb0OexKAs?HRN4k{ zwj5a1*QV1{-Cy?B@r&q(o<5kei}CP$UNAl&e(DW$#iYLkeAF)8G~Z!|y)KY=8}0m_ z+E5pGbRsD3%Bie7Fl5p`I;po%-sJNg!kgN0*4otUC_su{RI zO7$AYL5$zab=hhGG><}`9JTeu=G3zEfTAyP-FyfM0s!OMi$^8qrMN{_p|k95=bn2~50`+O$jC0lm;=wiCzcZnb58;ot%KTlIuUU1QTc!qa-AaLh^*-(xNraYM1 ztlD7HwC6hs6~3@3c%R_z^?=`|yvMI6cD1VG-@?&!>k2#S~Z#422;51 zxU@={?WCWc$$v^;l6!AGxrTi;3Yu8IPZMRc8MUY`I3HCJ*CJsU){vx^22x959swCR z`cA<5x8FotszUA+)sCI}W?={Tn&+~H%ym@aBlEuKyw0(ROJXwo+x-NhG+#$KFv~$i zOI_NphJI7zc(8Ij`C1J4O$9JON!oo%1bACNlqJ5cvqt{$VQTE2l~DsDQ0fc7Js0 z^hc3X+ln2vt9Z{Zhw)G+W@{6m^-z!vx z&xZu;j6}trQcP#8>{krrpPagn?{YTJI(;G6rlaedFhL9j=}{f9vt>&~QT4xdI#dbr zuJc_z$v&}eS{X(h`S7U-r@~coLyF^`F-1cLADN6zzf97)h9Y@o^WbJ@1C`~ z{tVvwC{Fie^DwpSDAWiZy_Q@IGL`@Ip-uB2sPW8Xy#e>Pgg7~)3PplQCAT`9r9JO~ z3XdwuakOQmJ=I_XOddU2SGVZmp4;!$dH&|IyB_XZx8@FMx20I)jEid|OAd>+ed=Lf zO|T#t(j!6toC(Ie*2S176}2%U9x(oMBnkE0LW}1S|83>{_SE*B(kjUmh`}_dw%lx~ z!!N2aAx{pU1Zc!XX#f(5%xO#{@FBmkVv(e(!e1P|3rcw7d|j34RgZ0(RhW`ccpdj- z#I?r@^YX}U|E{_{>zaJv5yzj-M93hmw*k7Hlz8T*rKbdjC4B&Q_#{^LEpiwrBr@Vf%9H;(QBn+ zc`4PO9d%#HzC!1Xs3}7F!lXG|9&5e9ie_34mT(n(?2GY$eS}MMe1kMu^TPHQOiYRF z&M2DN?uDb%AtdfJ(rk>d(uGNvCd+S=UbRh=U$+BO%|AM5w+TwAI9L8rffmB7)dK22 zUQ5#OHK$D7rcU@8v5FzE>_HGU<87r5slr;;qfp4XoxI9eLz+S+cP0PBfQ2 zy*rteKO<$do2Fb+W5(RYadj}1(3$(W%`n2+@NQ|SBR9u0d9Yjp%p-8{C|%0Lv8D2{ z@u@X-#Nlc~KHSiR$7K1ZAy^aapv~e!=tr~`%GIQOqt$Y!8h@N_4r_Pa1U$^zZY}kD zpr)CgLC5Kjv3e-Ugs6p-G?N>1}xyLm>kX-Alg z{c5f0?~^l;IuNj(9sV-OCVFwXu0LxO*bI4B)rF?$kE;CGk_Omx+jcjGL0h@%`fK^~ zJ}}F){<|-dfBYTb_)Lu6t73*WV)(v#ZTE05rf0y^ zIE4K*8m2rT{tvwr@S!CKMgwN=iZ_=DlkVQuhs4u zNY>CF%Xq~woPCOey!fRQ@%Rn(Agz*K+TX^GNhSBI2KFQiZ(L8L8wvR?`=7zfOvi6p zH9q1mQ#$VvU2oo!OJ4rVaJa;GO;ojhHPyk}B%0!Hh4|zsH5Yi#ikw_9wX5yus@o1a zEm@3`<-^AgqaI6Gl5KPPj*$ulb7dYQH^+AI)Y_RuI&)vr`vm`f@Wz_>G$XfrT;ttY znq4}Tx0#>$E#xtu1h-={1 zMSM$54Hk%5bM#EfaJ4d5Ki0S>yP|&^ut}EUpcJu!?ozIzbdILjP~2Tmgwca z0$uq3{$GG54D$+_DStVwz`a>Yg-Sjy&DhZdw=$k*jRgh9pwu!JI9J12IulmNdIk-g z0~@FQ{@29>XN?Kb{yThFq#6z_dZ|fQ>Cdw=B~OY=P{ZVqh>K^U_ve)cEOG;{B2eFX zPumhH5|h7)ef}pdBuZ^}YFL1csZFkXPuQk1+ZBpg8|x@|QlH=MrZwj7aC_W_MxSt% zH!WjcCA$DZN%(|K-{|G|%g4!CKg$Unv~yXxDRx)$zj`Tky6F+fe+nVFoCyTN$d8gRpW1Jq}nTOJ?xKJ>h5aYfWd@;(ElPz-Ip@&)aCr; zE7u>y>MVywbR2z2cWNp~RC3anz(4K3x?1+F)Fc@{+5TbbSbtGtY%1>Gy->V-bw$|* za&;s%HGb*x2^|tv1mH~D#9lh6DdL%s~gTbrB_=VN ze?|{^i86v4XW`o^n{;wC;rMeB5eRiY6&Z#BjvvB#8|yPBytPvYr*9kfI>5gU>>dbS zFZC0shvAUsywBp4L3utP8mgYi1-Dm12E#a1OF1Dl<>*A79ow{-w>dA zS#tMx^xiT3A;I*>H^M0Kl0i@u6ej?n8DZ`T`Y;dd%jSdq9DcUv{eV`G$-s(*RGMe& zeBN9!tIp*Z^ZW9FBMwN$`6zq&?^k~2DS|!B9UP3Uh0FH|7(jhJpcDtS-^cE8c%Go< zPwE=U$VT~|HUHH%$W~ewU)W59IY^saC+Tzs=L0CljT8pW~G z$=FTxlAW_h{N(*7l-hT1b&^Lc_%(v|Ye>!;5s}IHs@-e+<3CK2OCbT}qi+gu2Ue8m zET;8pIX1@BkwNw(vo>qnc7GNNoI5<*-H~g4%IH;`KBMo)FFk`Gm)II zLzn(pvVBbH+dJWG3$VwwPlm>3tiigVWm^T?o8I$XOL zDJa9&}5180}Z5J57-pkyNsp*4|t$en4CnO>)Y+Ez*fm<6`OFe zXLQ~>N>mRo2jcE$W%+{yGjtmWomSf>D zgb42(Hs5jZ4j7|;e!chRm$cm=}+g!U}jMup%_BtS%(?Ak>*8}&mL4p5+qKjiF;KFR1 zOQFzP>I(SY;k-n1KgwbBeW~wN`M)7P0I^r;7QDGp%@pn90IbK|$PMFnJR%WG7?is@`^Q@?? zXRuUq^nTR%)J3b3b9Van{O7Vcz;)VDZ$> zjx6y_>@p3VEn&{RP~UF)(rSc1#lE5OFFBTK0oab%PJxH)zfS1Re-g-PoPR@dVE2Vr zb$-n2pv057+S}<#uJ7^Q&-?fLV_1u|X0bTu z?ALxh_j5n?>)gAbPUCQ@PSFF1&1ZR|N*uluxdl;BE|awKj-~hB$2|0e&xvfv-eDxS zT0iiQW_){EU=RtRq-<5^z|z$PROl>E_M~1!JNHcP#%>%5@432^$;aZITG7E72D!O8 zuz$qi5eRP4>op)*yL!Hv^yO2#`5f!#t=91HUp5z60oNWV(d(y=gu06+aR~2EF)nQz zZ#^EIg)=SET9ZzO2DQxh)p~Hd&+vVG>dR4MTtj%8^=}qgfB%2#IQlqvz96x~pc`ek zGi=hCnpZnhS}Sc*<4-$L>5U5b*d#*D-0})zxqqOsEdthqEvO{5XenE}b}x%D*DyC> z=5^#_U%dgtSOFnb>cbRE}lh2Fv*q zRxov_vEh-rF=fN#(Tvl8>Pm8Ph`3}&)A^)&PwWx*MvP}>CJmeS&7Ee3&CqRNp_0qXr92gENY*`e zHh1((LKz2DsB_dtuB`%vV-V0@$dF6UQyZ2P{Z6T0IfHh+RvmjeEkEV|jR8^MOVMxt zTM+;FYVk*NddwGk#sB*2A3bw-{)z2T!8+M2@g+3q0&q!$F@+nUv?#64AQDR$PHLWX zIfVEk^+PeF4()ldw@!}of9(A?T(g^-{31T?)r+$m==Gq_co)%c59wY~F)_+7S?L21=qSX=bO<@6m?*FVlWKY!Y& zjexY`3~0(%yLN^)l`tEajvbpk*!-wQ$#notgkCE!X|S=pn}PvTiK7r$zgH7#{ucK- zDJz>Sx6!#qbmV*VFJt5-4k9$QxOU-6jA`>EM4>=V-S9CdQ;k}7xJ~VZ3+lH|Rv`(r z?GZjbq2<*H)PQqvXg#zY=Z8aFV{Xkj=+dqD6h$HbN3Rc6NBy6i(ag!{!?s{u@bNpF zwS~#gH$6rKLy{3w64%VR@b0)%Y9qP?h1|rC(epeM=S9I zOAK@Is(nDM(@6xEqu%IDN9G#Y`_BT|Hnw&78yUqokUf}M6a2)ZHB038=u=jPgx7j@ zI6FM&-?pvR(SlhvzbRN69kX}7#KXJg*_LdS`aoL!kM5PU zXT|6(F6vw<;he+*w`DW=o!)FKW;u6Mc z-l@Lr85}bQ(TiXlUyko8pCGnkaMxvQ_05#D;5w+!SzpzdphvWEB877hLBPOLuCwP1 zbsrc^M5RFH`uQ^%iAh9ifqM(2cJipXR&(Ixs*Z;NZVrJZ5i19}u%~|WjcYI~wx@z+ z!?v4arTi^n@wXjV|7n9n@7u#IAJpF7Sw@+}hHE#s@}*tx#?PM-Pj-@HsKUe%Z^o*! zyAlfaXTZFvGTXq_w1^-hpOAPZSR0#2c63Rqy3x~)G|!1#0v_l6-)x;lp$F$SH!hS2 z(aJ=adiln9WyV&kyr#}5`OrG>u>wBXdml9w2hA8XF8A4)I0>8%JZen0DW-(?2CSe- z{HqJ*pNQRQwQUMMescPEVP2;E&#oF0YK7<7GxDX5KF__4_5R%V6MlzKPv~SvjtO|g z)wNDTYn2}pnDl~Id)QxNnkVp^+e!H2+)F0jF%$vOj3R*`K+ps{Gwr4i)4GSSlVyp%n99-Ri0SNru=ih8Aelhj)@5-;f zkd>Dxy#9j7?dKcH150P&qmz?QBpa@7@3NfFEOX*j8sF{I2fu@_n8bL^7nNZ7ha@@^ zX&<|N>zdoof#lx~@>@6Ml59wbHo=@Lb8IoR1SYF8LE#6(+fzZN!l8d0gB$A7f8PC< z>ZvdyF5KDm?0g=V+|{;i)K#~aXVq8npKKtih?ti+ntC-GJDWL*KkBVs%C>nO(WCCT z*QM!Ir|D9sQ%6oKB|0s&%O@lABg@K&wFF_sb9)M*Gr`Q&7q=;ADooXI$jydUjI!cB z1KZqe?gXB(q35Gfkul*oEybP>EdlD0(Ajd4u9hEf(GdrUiBL#=1$Yqe6a14+riUH* zsrD~vQ~}9brOSzWM+4jqSEXdKWHMU`F}Tc#P>J}I$|%`ubV+Do1#@S?iH2w!=tevt z9QRgkc%1zW+LJ)7pad)KYZ0dAvbX4;N*o7osoj%xTiEff#W%Fl;K<%t>BsKWJa^BI z?ghrRZuS>G;PqVRsw_Vd83WvXhA`S;PA7*3)Ew8`p(E_mKsWcQh4k8)^=iAf+MhcY zIQjqWk*dYkXcthtg18aXee(HZjtn-cv`iwN`}t4}hS!ocwyhEx#o{T6Hf3x+RqWVG zdSk}5kS+G>b78z6mXWN!KcGis$iq(-!IiMm`p*%^D^fZ7udZbj%L7ubDa8 z0!tEW>O4S>1rjLpXm0`)rtv;DP`ieo#PV~c%{R89tgBI8)!plp)XM?U#1vqYr3Zzf zfUe9@SLUr0LA4V&*%$TC-v-|*buTkpoFu>Xqa}|mi2mBJ*bF)X+0*4R41T9BjR7@x>RA@=xQi*JHW0RaslaM1 zMIZ9`c}7gaue%Qp_;gK`?~YxDdoB9AE53Ph!#Xw0sTDbtVxVk@tXl)+AX69r7fYzA z9bp<9U2iqJc$DquI?Bg57K7ZMFXF>CsVT#W6pv;3)$!ytyTO&y`^^sZ#?f{8`x4U| zNYO9bM$(;TxE!US-vm8iBdn>Xu0(&u^=5VF_wdKM^FV)qrgwq<4*71vBKe{9u(phl z*!_oNw@(Fw+@*@gcn3mK&t(6^r4;CYT3u-xSai~cE?~Y4mf-rt$GGsnJ1fV}WE+1H ziHiPGUCruM8nXSI^bWaig!~6?1C-^>jiv)F<`Ju>4gd5)`fs}*EDDT*!pMuTLkzxU z3&+sVqdFmWdPTua!50&pDQwT$8rv-N2lUlv9D{|imulK)aZ|YB!Q@R-NpFMI6Hk2757)gyRJb;uP)iC& zNYV(`2?7R9ESEi@UGeXa>IGRH-N9URNkElUt@bG}o>pb^$-CLL$4rvqxRClL8H9EH_pjKDw%=uPK+#^AJQ>Ll_rcj4*6!0WyG5mru}9@zQLEcw29lDkU55Ce+RMRZ$tIp>XCw zyzFs);kLo9@zSa3IXAR|Ss9PO!u=z{^_Tls1W#itjs}hpkpk{)6Syj1ZRx%O+Wwq} zd48&oe8{&^#*nuc13TKtWh#1O>L9(UA^kAwrbYVS(kakN>fV|0-^w~*s&mooaK!Xm zcl*kaVGQirE!{3}b(Xc!>Ui_f3wlDlDH^Esys=fo>G21MChiqF)H&h(N`P`!%G``Z zP}=H9oJ2};y5INJJ5r46)(_eTqdjB7L&_v9;stWbx~X&e0+)8hfqAbkwHvg&FVLqV zjRNc2%&Tn!RB>PXy}4Dj3%FS(=l}FH(jnai0m!(M#RR{GY_-SP&^h>d)+JO zxd8#h7VQoi*9WP5+%%v~Po#Bbjos=raY*+7^9oPN@G*V7U{3rCK;$(6`>9Fy04Of? zk#&F68j7Z8URm)9oqos7G=$~LQM%^#>)b3TUC6ZX#w=)V>6e{LRdn(laEckh0|kXG zsBczhgXt^fs1mI2BqnH=iqQgS5j$5X)@`upJB)CsKK&2zC`)RwaCy_m!3fd08+I;E z2Qx|UpnuzOH$O|evi`fPpaFC;1JMg2>tLf%Ekpfg!k^6No0_4|; zGVEPO>4i*gU{Azq7*{VJN_TF#-rIJgErz){lA>KE#B%uD@vp?YSa}6a~WbUm&L4TRd_2q0}mgXBUXjt+`J-~AV(*^@TA}! zu02Qj?<=X=f5UJh5zzSps$MxDsnzh%42bZ1JdBO@V7eF$d4QxoI`tZ2sBS22vAKpJ zM58m{hu8L8w4+)?HUbdSiPCHRYSnxdjtm7VG{1F6RmdT~TFf%ovhsAh>y8fjbwW5= z>oB}{BB3xj)BtE1krur|rwX&U8or;3)!w0y@XTy27m5-3xoIq+uLEsohhX#O+(O!k z{@6q7?s*SP|8t`8IF=X2gF9?!<~8s)&*TMe<>X$tKqhIg&_GXiEVs|W!&LgEgbPnq z1qQg$#$Oty-1tcE9KFC(hP<~1LoT2B9wq>;nq5_&@;n^lvEajb$ron79ngUW#6|Wp z_yvmdxUNTp@B!vS(SFwSz-1n#1SL$Kh1bi_n}sDDrO?H|cLlg}PSRGXc5@4MUF3dJ z#SrQ}*z2YAImxW|^!*fNT*O*Z{(&kUgaHHXF6MzxLkxz%&m5{dXz}_D2G2FvXi+Kz zXln$ei*i8ER7|~8j ztEP=}lFUqzJSsciYBO#xHu@m)WPWip+iJ_9ezDVG&*VFvOi*f_qRO$B`4)9m5n{r{ z_VUa=;ns~RM;@Byv$^{!m87oyo4^P@U$`Hpr+%&5b$DFX#lP79ee?D6S;6MkskQFq z&_;&3@iP>u{^&|D(2wutelg5al(=OpMh5{}aEJ1maZy><|0 zcBgq(MPH=Va1lVFs3CYbYKqfDIQpEAxfx6gnJ`kQ*6Chhx20{?>aRXL^VZOd%SvT1KsySw8r{o^2u zHvePN>O|=gwwhTQ)f}`+FYh|O zU{JFR$NvQR$rj%B*A-&hktr>O>`A=x8LPNPxBzBKJ#aoP%&oD>@_ijuMr<)`^mCy-c>wDfi~W_a+8&oIFjHB-Th^%|(oWsd=GU`sx_hXVV zZaG90#yf@q4d|~T(s7u)KNVeo`u`f*3ZKaV9%KIJHiN=S1v6KkLD1#mOrsK$@^QX| zz>=q9lGc`LJ}uLFsdl6l%f-$rN84?3F&e5jVpXuYBFWm?Xk0-K7r%uJk8@j6CtQrE zhj5iU=P>G}PRQ~{33?^S#@kRqN}sJGQO2*)E^4-ZIKm2SC>8nI43> z(!S>M*_B_7id%W#0cUQ;n3vQDB33b_N(ASyfNJ~qZf`n2XQ7RyaF|qUlwh!yxa0=R z6Ug)C4!1k>9iW!FnJt^Ys}uNyCExgZ4|fC(jT_f35H;7|`AsTjrO?L8_PZ))C0rEK z&>0?BMCDl7ruPfl2B<451(67;aW-Pq4@j#*zx@zaef=5@YJ6u0l%IzmDBV39yjJQv z9DnIwxrwhT-QU*3qU)gP;b$`C+@``B0Q_xGX+cfhY1!^NblWnUNeGBU=L8bKcftJj z5sVavktO$-m+y3ueoNKu8~2eB)mJ~PZh9yooYr=&t@<*bOdl|R8O|tnH3Z>Ga>j~v zJ&Z#Zp9^>6a_Tx(+-=I$ifaSZRXA|b!QI?UNH-@Bp3a|DL0Z+X|6OZY|4GhEIY9Lt zaa)OvMx8|Glj_6IEH!=^B5&o+3j^arjd2{DuRDOg5dp|7V;`*_j-{0o=sP>c(dYIl z!M*20dNR~G2X%4BZ$+u(aMQZtDhjp5yY18OThc- zy^GDqlj)(DCG=2R-LTHq)Hy2YTAI?kuFT6!h8Fh{V-g*_9j0KLMvqTB6>%-}n8s_= zODTMKM^xaN^00TLR@g99x7c#CJrj#F!4;qjU8-{hV}+d#>I2(z>PQs7{_w1~dD)H; zh0&_L-#>lk;6BtfXXv{bvqfN1un$xGd^3H}u85a!yd-f}DvbY^lIc?VlGIL^C-=iy zLBQuih)D}1n@&|xCR@p9MhU%K;4@?v|5Vx0G`NxUn8Q~>)d%z#_?i+gI}jwE#wrEt zJG&kMRaMy4^K6R(7rm1rMU)81$n8-QR#&<79f#^0(tTy-75cX(uS5|QfT=)1{oW{Y#NA_gTV=LUYl z?23}_(b2yK(1;NRNwrx|ir#<0Aw!{7&&})zQw9}Ny!`q?C~|PH*3?eE!{`rd7K;@ zTH2)h1fK&!ssWC)o?ShYmoW{So6mIdiQy~$9}56!2z=k^hTbXGRv%$Hlq z_WLz^S#%lHnNS$r{=u145M~p4S$CIDg2a3`PhHN`wyPT4tHj~r(}@~hd2fKMNt|17 zdRoM?DmMfkV&{zdOI594BLnm>(-dt}@a%0eS%SQHxk5~uIx^^8>TL~Gfp3p|vAmwn z8d>w%Y%ki^cU7m2#yGi6nGTGg@BVcC*#Hj7Hw@sAazY>w(v%P6WMFClt5{61Hg72oP^&`{EdwXUNqTr^DplUS;bC>w`un4i+<?Yqw1xjlqizUJqqUM8(a`qo9p?AXG1&8)agViiPP|S`;IxYmKh;^ zao!9o>MDP?L(}RMpByX?ZlvECf<>1Ff?FL4W^Z%%VsJX9vXfM*=b#S&<{b4%N|Y^i2?^ae@damg-0D2AWR?$lz8dUG&403fMud;{_&&7SGa{ z1OI>^;^Pllf>A0zh0~OgiLx(v%;{~0^J6Dr-3_L3LG9zfp@Qjr&{0F&9Os@(1EsFfu!LN z?kwic$@~QPa91T*hy}GR_d_Wmj#7nSS7SHdhVIinvQA&nnpp7S%t8XX?qDeGoiPzP zS&;PinK3m2w5+JA!@%8<=RNBQbFYZ!lq7HYUT&&8cITHNmejrMtEM9)7ts?o?E9;| z$T`|OKORT)kpw0-O$S(F})r>Kj|Aylp$tZpSkv5 ztytqbPiPT^uJx=U?LEQQr`7IqEYmdH&7wnz^20%0+65=d-;K(<<*WPH{l1RNj@e@; zKL{lbg~)K=gL+>HnQUaB@^dYz#5*g|RWm?4e#O|WX6&+fgf(qSl5N2W6DBS40O-c~ zxQJPm+nh*ungYFYH?{TNaS$Q0F3>ZVDMcdza`!uMLky*IZ>A}QgRC3#ybbA;zRPk@ z>Ub}uS^P0kSE`fZAzm-PR7NSoI%en6{Sb4uaI}ZUU8Mg8&QNr6s(kCej?B0Ev}yR; zP}o|te}9~`o_G<_(s_3o?pm!J6Ah~V%Y|n(`w`zJ6U#2)iyWdCSBOxta$*wIhp?PJ zzuDEFhS^_*XLez~f4LJ`<389cDw5)}dqm_Z`T~V59CYanw@@7s+7fxWtDSRR;46!4 zn%1x_aY=6&2}6oQA}!K|;(@N>9G!F^0Si#0k0xfG${^29->_d?$Af~u;$UFecbi@Vul z#VPv&d@qZMW|S3kh&E`|o?AyTqOcwbr2`5~+_D3L(28lEJGtDPeJOHxy}2 zHi|s2!}do+%k!8QEEgQ14+==cJ5oo}W`%!t_BjPbGk$C22HaLk9=z}vJV`b2bh5P+ znc|t}E@q~|DHJS8a>J_tn8w>&hr=~OeNt~ul4gvzus1`jhFmGnCUTWpUOvjQUV2Z3 zWBXGjZ!2b{!kl_zx1O%5O4JPRVIh2Zn!Ix}>#C_XQe~x{nG0JZfyBw7YHUwTKm5}V zW;xBLtm2Olk^CBwC(%y5R@AHk9qUK4aBoF42ERQhOJ3)j1GYM%F*(M$bphP4VbYU&PyR zpqtkWI%6DPN!I8)w;1nVZtf@Fh_6(vIBQs|x+nD-C}rW7xDl8Xvt#?}jvs?W0Co^# zKc`=fKEWmwT5XF^Fg6NuH@C1Sd508-rTDP^vu4Y0RSp?*@#x_XSl#=qb!_h3Q`D$F z!mS?g5$H3{yv(YR=7>_E8>!g>J=jf?@KH1DR7>6fQ}xACa3oy@&(6|)(5(|qKmI*- zRg(0D;whZv{$xLW-a`FTw;mg#i&^^q_l0T|smPDIeAJ(sjTyN{(kRwlGeX0@9wwth#+$`3J zv_n5sOSIyKrR`aLiiimIi zEa{}Of>DC`Vm2Ll2vhH|vRM9`Z7G5Qp((q!%uGpxnfD^0kiq*CV30vcrfFmBF+n&Y zF`v`+;w-@i=b|xww#uC!c64E@j7hp-Fl%fX7v?}GD*2sVPdkVN-QJk$>ru<&W@D(>X) z6^pxER_ZT;-^G+RUS@ep(A9Cjh?ZJw%gb*^oq&VQFGcy{Ehm1{I#)#>)Bms&=0;w9 zS**=9+ms9FTyy}Ox@5Bk0CDg~Ku6T~8m#o$j_8sm^}cpTKt=!$ z(Y=|SDx56bVo9G>QAtlPQf640XLX!$MO0D^AgjB)eR4=|W;+C)$YaBeZ2TL-oHM^2 zX_!$WmHA=C0YS9Stnu?JR?7ES2N0oM3~ONe;g-4bE)ls8)YLmX33Al6qOby$!RTtl z%+XP{Xize+Mm&f~rlEU{y~?+1@x$vARgu0HXvkIl4jZ>B8uVba+ub3)-<`+?yGnSV zf$+GFnTzFA)a*D(Cpq~3*1Tt4CfvTw*r2pSz#Idz#U}E^Y%CweUf&X6a*m&10wq}d zNj7d~Z#gtr$*MLds@lFOsmGsbMU(-LET``|3bw7+plo3oG+Mn63WaV z?Q46&GHO@<{W9T^B|Q;79;D`HSuMia0-X7_Ec65Y=B|dN2M2CJs210j2(7XGC0II# zCAJ3-<63%MEnQqt|D{6tXQ!H^zSMq>W!*U8bZ)=Twfh+>$BBNyjGI2B<74oIJGb7f zXFRJ=A})^^8yByXi5s7kOFm;-G3$!+0S^&xi8V%%Li0&J$jQxr+H35doLp+S z`qW;ccpTh#kTc!IRkYTXGwdVA4Vip-7j(8zAzsJz661dcUXN8BF3FIq%bd46Wu#LX zh_C6T=saq;j}J_G5n!dT^-Y{@%~DB$$5XV;fLkvQFN|Yz_@#zDpP2(Ox6a?}U-UnS z8eP6Q3}zOrjcCC1`(ds-7TNsi2Tri;Q~5OniFyc-Z^w1x@RUsP3KR9=@s7At)cub# zjh++>nB`z)U9lkFkZh8rRJgCvz0O-LQN0Z{C_6+y+Ic78T~P5Y0e_-Bzk<_4w{J{1 z-X@r0H8MdYlVa{X9G+k9S=fuzd|8Y;X_Yy(k^R7Z4`0YtNGX+5!RmR(G5hu)? z^;qNOJb2~9=fe!!#WLR?xexgIpWdJKY+lxXnXK_`M+f#o3}a}DW6~GtioacEb+uzG zocHj?j~^zck##NPo_t7OZ6(T`+b&>d3~fGiZ`F%&*$W>;JY$G-Rg24&-pcm$7`PQu41PAzlDBQ$rN-~)`9Lxry1nhv(`&sgX)X^HtnHsj=4pLlQozA7}uc% zL$3wnUnjjR8W9|_wbR?8-|;^slDzi6)=Hps8>~D&^Q*zf#Cw7?OgUK0U4}EdXa7t0xa#sHsK~Qs%~{Ieo)NP%OAKsSbzMs_ z7Ba=#E?ujg4WIK$@#e$j286`36=FkZ-=HX}H*U7n_tJBCpWW!xA?_O4Q^|q{1+B}A zYTGu9FYZixEyNPUVXU#K92R*;HJ&G6@d3d@+Drq|0>yR7Y*8|7sd>lR$%dyzUw!tU zSqTuL|0V;!?skOyrY1Ksd9(_(6o<(T{2w3qa1~PUee~pn`8C>%}y25 z%#6ZLsn`_g)w9(nr)%o_@gySxTT^evq;}BbSz+j?8IP|U$%&iZk~n!tm5D-O30^+* z4WxS;ydoOG5nY9NwkFmQv;JGtP*f)~9NSgzGcGX+sHo~!B%0t&e~3&-f_LXk_=Ip6 zyz&{_-!Y0ToaWvx9nJV}Q(3e$feWeNK1#}Wvi9vI37Y*eI1vj9@-JW<#O51#s(?uD zH3-g42E^u+ut=Lo?tZ;b&2k3mw5A^ET?$$dDQ^ofFV1CiIT;f)^K zC`Tt!pda+sO?UKQUr99Po!FL-pbNxv$ZF@^FER%{QcnfHLnJQ?E=?AC`P-GrGI0)R z)l5!o8M3ldhY`0^UV0cmLr!RZswjE9k|md1yU~5?nF$2-`|Wr5fBPB;5D5zYO|w5D z=c20%Fy6x<|A0oO?S%v17WL;t+bI=dA9OFY4uFO^^Oa#PuU3Da zdBJS+A-$_=w*QWVmh!(t1Ef(!_USgOs=8Qc_H~b(W0#X#nriS%+>2Y+Z{>1*9dSf) zqR&l1x_Rp`9TLO1P;IkAH)uWw-fV|oCk%_~dNtDZt)_Aq+;ZHt!TpmM7nbJRLl4pu zvs_iz3d4&#m<%$X#66xIIub~63bQ%0=j8aT%kz8|u5M6z;p$psbfu=#5ZEbqIs#8m zKAePnWr%+m=%4N)hK$4~*2*I$F|hKn?vVzY9FK>O+#k>2oMRR{1RDO0LA@J)3zoyG zKS?v1ffmC1JsR)&rH}`Y2@BV{uVo#>gOuc(8e@*eBz2YMs}o zGETJt^A1OM8^^fuBrkT-v3(wUT1Vz0p&Cql@1j4qu|iyGU5C9K}Ama!i4w) z=;uRNFhdL5UzY+bEnh`QvqX(|qaS3_CDbywffFt-79!fbnyg6HXi-APWnR$>$$L2i zg39A%u{mtwIT7uY@+IL&?h<$0n7ZeK_~6$TP37)3BQL#myMS2%peNvhgV?) zN*D5ot50OtTG0P~-I(4w#K&+-JbYoNv!sA$WS{2zoGgFEmPCtG{K++rbQ`8)3IZk+ z?5S8;AagIxXBT@1s@U}rc~lT;kbZL5fg-TB=%*N9s@@1zkjy@zj7Jj70=>T-lNco( z3p-p#3?mQLsQL@_0GX!cF^`d^syNGETb!}QEM{>+Q)g)Gmkh$bN>zhX=mwVDF(DVSUFu9pOtn`r`aMNFqSG;|25K(!(g zycqr!b9%S-STU2%m7rUPuY3$=cWa!kg`u*K3Dk^m$Npr#+7_zS6tPr20cg52Lx3mnj(v~gdMGAh6d}FJU-V<-qvUm7FhLwNj$*UQ!I$v z#*qUGfnKYg1S@=9P0D(1oXVeMu#{jeMOBjyMnS(ul=`SO_1bWL7KlQpj(mUCi3e-( zhZ+qt3p?j+c%K;gT9HpNdq8|XV_7R!>-v@`uSOt+XTNP&RcU(bad+prJ=FNntSkNS zQd3dipkCL7;NiSZB3WpQIc@6jN~nRpThP*2;Ol0{w1g(_RHTr__ir^wHRN=_nDY!QF@)r?o#*mfzOH>q5;K9gam=8e07 zyd+tHi_ydMcgiM){!}xG#6ym`m}F<<)=ajwT!WPHL1eJTz)G&W*J%6no6`bbc;lLj z=Hc2T0Y>D|;h~8xyBwV*u?==e_>jWKq};`Waj$Bpw_4HB70Fvy>*}Z878#z}UG+Zq zYCU|WJYt5a4wb%`=r)%|i~N7Wm7DT};sl9U1zdT>j<7@GW9&^nb^=yu3v1iQq)qe# zXcM)1m%;Lnsgr6 zqP|(_{>gCbNpHha4HvMKexbVKomRyGa)^^bzIWdHHGJ{!mRq~>y2cOruP_91Tx#BX zTRh*(z_`FVt+BDMH_zN}jR$Nq1XMVu+M4(t?Kja`$0UvmUN1eU&P@(&DMYt^Y0Q?bXWVK@?Vx)}U9i#cpHe;nYSoX(dMyjbK#M_UT#&KurGjrN) z&i$Fk5Q93&jj@$?fl~3&S$YZQAb3bDeSNKHsh|>P9`i=$HHKI#k!(eEP0Lj+CBGs& zyexldhBbznWyIZK$IiyAS0pk&ogGnu<}Efk+uD=>vyf*A>qb&x-5&I zLH_f_J0D}tS$MFDW0ITy@lOK>Ilrf&POT#ThHdYdV31fV9sSlb8*XF5wg2FQucl3! zC|zCMiciMjq|$e@xt+IogI`u_I{%W3C{R*tvO?D{dOeIOd;iqxjVmxZ&sePkj22@( zTKy*@cgcoKNt20_n2eE&oA77+ka}0Q^#IcdzF(Cb29X{iZblz*bHykxWTM(zxY&6G zE8*OAs{bcC?IbeDkABQ~2Gj2vkfw-hmU97S=mV;snC5(0In1w-M=4qHnYH54zBT*D zKINwh6b1iPCb?`}pS=H6D|$CeTg^CXD(jQaZOXmdQhsIT!;>js}M>`RoKJdze2+b1ps-Mw+%Hh*=+SC!e8$<7LKa4E*bE`#59tmd0M z(}HcV^!ln|bf@=b>Ya%*T*2W{q#(vx=K0 zh(fEUYoE_Yq>858(Ib{?R8yC13MP_TIVWe&{pY-p-y;Q|=b4_&ZI z@ockSY)4cT$244-7q77Wqkw`Qo~oMf!4Chac50Lu@^jGnC?M3(Q&d+}avT7RKF;0G zPe$)umFr(t))`@M5(l@R&?A!ydg|V;c%eDdGq&TCd73JRuy6;fWUz1uN%z)*Jk;cr zl{ED}gnfZ@n`#tqo{Hu4wbNpA7sO8MLHwwBM*Gl=+-`OCcf-t_K2qwwekAWX$-4Ze zSML1wT0ctR`AVQPg+oZ<8m8K?ow38>Ok>^^Kt~gv(>v+!dSo}%HFw#+VYy7(6v_b( z@B27th}=fgrXu0JW}K#F}$zz@9pa^B$#{V$D52kAXH__v_B=pOeVxKaxV z@s&*{bry!V#LFdH+L7%zZg#-lv$QNVJyJPScxowKi7@j_i7$2!TB@D&oZc2^@Hk>E zV&=e#&hJbM(bm>w&WaglAW_qOn}WoyFHa_N8$m+$@QkAT=6nTkzl%JgYeueO-5Psc3a zVRUBbA%~@$2K8cN{Fa3|CY?_sxKCBG+~nl`a4lngsgX)s2q$8>a~*jqA9BGfrL z|Ht{2+Hswi(zUCMsh*-i8bUI42=rY%-I|+31{hGa=udhF%yUK>80XnZ7}=}gI7a*s zphQ_Fa=@$Ve4{eKd{~+=#H;|vm||VWq6qzeNT$?yBvjZ@!^43(ehgV?AVqZY^91m zc}|_$V~CWJX71IM;V7Q-zJ$SJ$>oeP~uZT!r#i|3%o#FF=PsLQJbpZmX+w@_lpj$xnixA_Z3c z+jXzlMdrgjEo(b+EV-lW*UX_pM1qadE@4A$7dMkoz6g#aO8Ns2f~s8L-QWM63&1`+6Vsqc#9mX>9*gWO?@AD~Hk$HUJ!p&IQns6YF<^sMnH}G1p-vE?ncHql z!+Ug_sx+(TE2T2C>jz%2TQL*>2A9DjVtCrjcx8rd4FtM*794X%LX`EYsh_@VZXd85 z?kBwO6pQz1sPJi??5kb4Ba$o4gb8Q!DkLSc6_CEpJdqBV>(cV5D10O1yA$xFU!Yi} z8}%iu)ZPc)Q=3~mb~YABOk(|XhN5}xn7WPRt z%!acboD`Q6Q(BsZdckTjI6cZ6nW?F_v;|$T&zNjjbwZ&vdrF_|&_!f>ZW3DP*G0OS zqGUH*RkLvKeiaDs+KYr1Ccks)6obdZ*ETbD6_($2KfTU-`Zhh?XW@u2@!jq%!bt7?+F5Nq474xE*9EflqaAu9lj{k8#CW5D`?K(vi{P(D$xyW1@N?cM!ETJw zt5&?6qH!dgZ0Zv@x?p~k#>s@mAL zsM7|FIGCnxQuXXOX2vH8ojwz3qX&H5WZ<>O*m4GWIab@Psew|}iX5e!N-&JBa8YRS zfC^MrhLbuQ?$66^)56Hkc{U}PY@qr!_2!nrn!D$LbZH6a_oXSShAN=mrh46FPQA8R z{#DWt4yIv=%{QF;_>MagVhjGcf4Dv#Y*P?A50tm zH#zQ~y__-j=YMD>gj6^^&Vb+MA7**3ys^Y)Wt|Bkaan<;#pOu zBng2*Xq&PA^az`gwCBO>Z?p-a+-xImstsQtJ~~!C*2pD3`aQwZXRLw2q}YZj&^OqU zAw+}&h0BySMSBesKb%`=H6F?%W8)e>^1h1O5P12{s^d29>-9C@dMj`w=`XK#oE7EA z=CgaykUNg9t+Ou;^7oEy-b@xa50w>#e$Ev%E*vh%0#(0FzJi@ovl^Lf2E=a*s~UO@ z38HM<$Vb&*tKSA2aND+$-!^{AD`#g9g`g1C0r9VEc<6-7ZGA=J?GyDRqD%9P$dHU1 zGmx*`yxY7&Q?nv?cu8MSUbhknI?99HVy}{dB~4<0^zET*MV-3)Bn`Zim58M;{zxyXP+%Y&s^E!b5weP%)U06f zECJiL--c&leN&iAZ*6Id;>si2-VOBpMqH4F;p4lzlcz!VeNfkmLv!a>n6_qi1W!n+ zP0|KYfv=(t`OPMqiVkypsfn%nX=jBn&rkIsuV&_*vHZ8uqLjvOwbq=$ zFQxAFs<_asRC~>2vQhPi>Kg*ZC$Yx=0EcDY#zJ&7b*OpNY=S7qyjdL2-Te_nKS?aCxNxXe(`(*+@^J;62R$rT z=xBQXKH6 zZH0%~(9~0lq?TG<2p}0*R7tpX-`k4k?zP@TB*17#v_d(lc;Dw&$g<cisJp762DG;{GaeE!H|j?PB~_3Rq(nNTMv5ZsNQr?XMvfYYqd|ek5z;Ywq%@-$ozgLSA|TS; zF{K23zt8)7-|OPK;2&!Hem?hopL6bWpgrOm=RTH4xZm@KAjB2$7~Va<5?xtuwJqI5 zc5Juqk6IAP5>P?ax~b}G@ZL%`^WoBCT1=TK3BEELzD@IgOZV!CYZNuVss`mhRICd6B2c;KN70c1#fJL&WZ}6-?*59i_=3SWEs*Pr`g}) zM}tK$Q}w4bJTX~A*HuX-&@7rF2hD87=G()SGWp}#QC z^xZ_ksoWIW)0}wHE>mP`!`|T7Z*uFn6FUL}heA;3z>BMBtpBNn4aPrX^EZ>zd|<(Hy#;zWDDm+86sB$E0RaHfd41t6x~a5Xd~ zi$Ne>OorU;MPS`p!<`p@*CY7@Wk&Qo^1LYef4%RKtV3+-K=k+`BJ{~rpeMb^FO1!& z+DYgkXyy?Jdf4dfVRk8ehb}44cXV5P(}8wHPO42py7qGkFj>LY0IM%7)|%y~SM+kv&D47wy$eX6dTMblR-)~c! zuDdGi9CciKWCq|@*xNHC?sDU)#)X9NU3IMT(>(kFJXYXThOpc(qIH*V#`r>1$U05B zIq;Pffs!VO$oQsm0qZ&7bN%9b{>A@Mpgy^Y6yL|Kn zRl>LTpqXI`MQ#nLxo__!b*{%{VuJ2l*JVME)%!)Mv~U51C-Ox?iy|23-iUi4ju;2b zY1@u_p=0juVZ%kkck%zk84WyInlNo^9!0gd#4udhSYf$@1N6uC`^G~)1z7cpzaD*v zuD8ivc~dJ!MGKX|AzUIb4sw{43aam2z`Wnh_&OoJPVD|ykI%90`S1Hg}{I)yM_tsgtlL|eUoogMY*B0Cf~{g>!JAE z4=QA5VtNN_zI;`FvG6vBj{?wJ!*41ZGkM70Rd-j5Ox%ZMZpxW)3F*(wIpmn-sx{Xu zCfbuA6$#%a43qApyJWG9&akWhVD)h|KWqv5u4?b}qLP6o53bdZd*{Lmob=KN(WghN zEXu<1a2@?Oz?{cDx^p$9p`rD>$c48|TKV9fJe#X@JZ%7A!rRNW?gORGr_Xz%&wm?P zf_tT*a~*|zUG+`Wr*YX_@El|w*{v7F&fNp7@=Q0@5?c!eeqsRdVkBXew8%<}Gx>+P zYi8tFP<>EhPU+(WQ9GtW5X%F+`0El0L_(`&O_W|~M*l8b;IHDdWWo8bkbr;=b^b#6 zbXyh2!TIRhHR$1>3{eq&RbVK8#m|pNXYa{smzz*(gDz(0i{UDy=GI|{f&-{Fnd*Yb~X1+ z92CJ)bMo7rwetTe2mU`CypXP;k;Ov#KPm)YtE2}4Q4i_gBsG_|31lH!&aWRE?V7ec zG%jn52rVaWzkO7!xkIT35bh1k?&`Q;MgmKB?N+n22)iUTH6El03;}&F!(EuQo+ze* z-Sko_Oqw+a5lkS@gJ0$^zB^?|7c@obdVd$nLy71}WU+}fH6ZJrNQ;pCR%SP@^pSaY z?(=)2n(nZ>s6|_cf&4aV)$z+|j`l;|0jv_)n;s@GPn`G{eeR&pKD$Z=$Cyo;CkjQ= z#W)qp$NXc=Gmco44O5RVIBpHl4RMxlfieY2sgw=CBy^QYd(6!mmD{2e;yN4T_oQ}e zc9aZz(sJm~Z@ZT&**)I?OpuU7GB}hyKT*kgi-KBNwQfgifi3}1Yi-GboL9|;H6qcz zsmHydR`J1FK~#Extuv0q%THShl9V5p8Z)X^#9)V05~ZZ=g6hDG%|#-|?WudLZNf_*ja=i76Cqo}VK zuOoRx9r^bhoE<=P`&hfrd^OVVFOTRfh4wLu6h(ET_5Y0%P=)G0-hT7We#Ekfu!L)q z`g@r205(DjC{msMQq+uO3Oh4DMz|D5T?Z53TKBW6rd;!O1@L^mIcvzpb)0!D`(d33 z%cl4cxq%|FZf{|3+(W2c=8Yv2&%OE=f1 z@reNpC_zHhkdLAd_+I6T_I#!HP|cOOTOO*8X5b3(LEyfQOG@1~wbz^ zvp)D04h>CouYHvkg4$N=(>?HbE9?jBpBa<+Md>l}GAtrRIa|pP5Dk2LK9{o!-HiE4 zMC6w^dzkA41sfC2O&dGidWtm&J1aq-zScBot-J;@LqgU_qq%Nm5Tngo6m zBe*JG4N>$r_x24UJ~hvJrrimp- zo^imj1N=}l)(xPKBK>j2hmk@!8S&AnbuQBz*Fw-qb4pD~&8&uE?gPQSPSX*DcQ*n&qyOyZ@Ao85dE; zp|R-t{JT)+rv-SN$13`Y&n1+Qvo6-Ea&J{EIbrad|G!Jg!42eo;Jy~!#d5?WY8 z@ee>y{WVf#Mx;(Tm z0U)!zWpb%7Dw%0aog7IHTwOAiPd;;b-cnIFb_7gT>qky8-j9{*r#;TQ<*oOMzNmim z-m@QmFqpGD?xt;MxT2NgbQ6?m#%C)_LPR6)k5?J`X(!AF1O?#jKxa(CI8(~je-suF z2^-uyqRaR^*|tK!$oU0hNkQ;gpIh?l znG0^Y>WxVh3YEq_;W6Bxaw*NP(P>L#Rl@D<$MEd=uA0rYvK5m+;cdO+fb~^?YplI| z|G&c*aj#cn$s9O^<$!#CE{IfD=K(`_SAR=eLk-VoSd{B8d4NtI(@8Pxro$frP~a*1 zaxqXiSwrHu3P@Cy6;<21g#EZF0CkNOg|}f4`@W|L1uI2katva-%$o&v*FPW-ebfEV zj^3dLZ>5U^0F6psj>KK-4TA*UkjRGo6X)6Brn&s5zBkA-_V=TOvlUtjG#bjaC1oCP zJSc_dvpL`uLdE&Ufvq&A->(^t$Or_xq5KzIv(^^JDqrYYbOBvEoJ zCp0~U1egv3na{;>SNKS#zvs?_fSZaqY5VB+{2W^JMr3N+Xokg$ts{lN`k0c^{E#ID z{}&rmX75~a%B!qQ6`_UVTa8y6mS8Bw!CR<)qs<3p^{v6^p1?$>w<<;P5<+?l>^ z3tRi1K+~=D&QaxyXPLJ%fGYs$8osGDQ}5?GH{V>E790Yi&8)ndIroYmAI>ox{O7K^ zMO3#!yjpsB&w%+v-W`hFthnSDl*OoJL;;DN%Ym1J3}d0>c~v=O*vOG`n6e%nb4+Jm z8`~Y_5Jlx>LL1xfBKH?_;*}oJSS(88`i~zYo^R8Kp$6E=($Y6XcO!cSM5;2;fpdUG zBoep{rYZUWgl--D%4};G_ub0RJ1^)nBHvf>;q+{k?^BtF>q3!osL!jiS}PGOq}N*c zhDj!SzfC*<$vc6s6b-`99Hg%mbO{YF`c@nQ^4SqQ_gEiyg-tci%U@Y+dh)`j2%nHA zyCM2{TC~Z1@L{`fLEP1Dq^w0Wtw(pSJ2i9GN8jXbA#8+Ay(AAY)NF743Bko-B%i9*RRA3DgVmDW$L2)sLO ziWy+k&+8gZNysweJ-^zFe@{hiALh#|=7vxusBg#KUz)Pxpt z8_B$)E(^R$$6o0=-OWc8@%lhbj z3xM_OhX!RjJP7M@`6LRXv&;_0u(|&y{qp}G71Su%4eui&hqyv@y%x;@@d0Tt@65j8!t-|M*_(l+$c`i;K&w~O8%^s`kEnT{9D@v192;^r%m80Bk z1a-73{_)etrNe@naexJ}ASK$Y>RatQ?GUTi`VdcV4@kMHoL*Z?ndGPjrts2eHBZTK z@1F9CjFO}|Xg+t*^BpUI`{7-Pxwg$w7YzxidNRu%@kfDI`{sbWf;Rx%u-$I(;~Ep5 z{W)E#V3$`3^!H2e&z`*rp2h3gYCS zQ_GT*1nESoo0bdq`xC^PBwqWv)%J{W^%J842!Qq3aM0d74H=(Lva0@C>%*s8z5?(4 z&xX$85!-L%v~o*Xz4gD16y4>eS$+L4c3D^-2W z24?n6YM%#jR2&XAtlMV2lNsa7AG+3*B6+H;fN%2;kT3FPu9~l_oR9PUPsP+`cI1&- z{R$FLGvB`S;*RKem3EL$x7Ch1%dlH;25}p9@3$7oH^U_TS3?>nn5LAnM81fjW1$sG zS_yLQ-8P08&tuu?G0gq?=jrf7x5g+M$V=`XREs|G|7!s*u71DZqX2r=nrUy@jJ~=& znW2m8(tpg8J5Nr!K9qHWm!2;IKn!qYd?hBGI2Xg8Omk#4$52Sxf~yR-J8xj+6@X` ze69c=cM;w4dD9$PEC8%sg}LW)DJGBt#@McG_2>BCnTJ9ae+0D5;lOt}v5EaMxJ+3y zv*`V*drURsV9MUQy?q9>Cj<@~W?S2Wd@NUe34kTtE@$76E&sm_B;{X)_lHwl>KME2 z=h+A*xhImfe1>(Y2U%=&xV4ATDfzR1rX`vG`|S1!M<7^hDBUXI6t6|>w;rk7EY`~T z$1r6lyq`+iF-+^8_r0?tV&+*!-Io@-5zN`f1mBVm^c_48K=2l^LeeHtfzN|CMf`DiZ5mqNsjIt>F=}T99kxt`iSUn4 zrB^#KoO%%7JTLPaQ(P|Bh(^8nMV)v-bH&BZ!0I-aN)r9syJk`?Da< z`0H-|zl#a?AulXTZA4=Q4)#0JA64@qjF2~!3gJvX;BfZ!+OeTQzZk_i{F`$c)9dNp z52|PYK^84N%Ds|INnwQN`-11h9ZX&SGN)|N^WsyR7Q+tVq>lUm>%hw^h}AeE?mm|? zVeAdB!OEt6I=33$r%tkf`Sa)6sty|WzZl~YBUJV1I!)Ht?G$8`>2p1PJjjot|NZFT zQKk=6ZxTthf$X~mlm+!VXj?5*u;?I@7sk3rI=rWuN5W#AZ|&+s|M~`ym@496%FL>{ z-0A3(HFoV8V|PQBmxhpIJ&vp%vFk8~sOtS|(E5La6NdVv2tO3{^g2N+3!6jyt;8wGlxY2_qIP`(tNE4eqQ&EC! zqL`y?0-V$eooiDwE7l}l^;F2KRj@xY6?j8hHrrD)Cno8f^Dd}#ucOLGTPQ6pHn+Lf zCLFa42dG;59?j2>3*B__V6iI=i%+qP@jp25qCrVm88^S%gFr!g;?O~^Vt^bn^NSj# z8TX>Hg<991@DUy&iE}N9+S(hZ_4s|46krwIel4j{I0<1|F8$PL5WO_?r2y_EFJrb; zO!|zKo=Xu5(GG3GR!vMcHZNzS-2Edvvz@kAd#exDi4`K7jPR<+WZV}(djzz%4>{9Y z4bWdFhZhd?IO^CdZzjw$JYx;|{A~G5*Eqw6Eqla>46GDS5v58b$Ktn@to~kxhr`X(4-&kKV_Ul3FeLGuB>uNb*+mZQu%M)oN@$ z-d7Deq=eGG%B8v85%Zb9F|=z&b1?Em|2RLF#>+y8*`?l-2R20X)qvTW9U>Z&LrZwK zYB&ao(=2i0GEBNDV$U_Le-8`)MwMhAe=3q#F{^t1Zg#t2LCzTQh&kJAZjGQ`bQN3l zqGg}18!5sX+Yu!r0zP3@HV}n`%Y>7mQ*IvSxQtu$ofD%VapPVv#xGRnJYdQdym!XsVy1iRSq~4F9~=Q zu-7d(ixdUq-PS8wM@t|$ZDM1ghCbbJ|Ba&3L@QJKJeRE4?;SZH%dT$? zYgD%gTao6+1T(|6ObE3*V50bj>_Fxne6P{~`k!JbH@A$=kle6@lkACB_diY`s1mI< zOGGvh3>D2D+JjW0+f0qMOSf3EWuYh*BX!&e#bIKZpc6SA`PJtV}@ zAOWgM4t4=L-o>W}?I;^3pZHMr-m)%N_D{zs0;G!wRf+@u0HRZEzEyixP?`3ChY!9i zt+q))VV_HALj?kPnfB24ul6)PReKL*-!&S367Kujl9?H$_7hiY_jyFjkVMj5ffGzy z;Q=|lbC;L+FSWW|rG2LGp5s-@;=z%0gWuKhAgUWWEWlq(vxV*XGSOJ=SUue^yU(QH z0EW)#H#O|WF=&GqrbfP*b};@%_Tj4glPbE}%s1~mC{zf}ExTC7k;dvixSK*+_?dvz zx{Qd5BZ0z}Rr&LPPVr}&4}VAd#Pvev6yjSMl*pd>3*tS3kv!}bA1150Z0nt8tktF# z8F>@BI646wv$M{sa@^wdg;NHOf>hYRzes37W{@0xb6TIN-q7!G8RKLs$4f{}%l?y3 z0y9VuqwBn)embu>2S}T=j33WEKSMQQHt6d)KWCFWGz;LSrJ027e2}Q}^x_(oxd~)c zA|%f8C+r1>G30@Hjv#%yOQzLs0S9DSGFHckMdyLDrj6^l7xIi0!|9Ib=(ms6C9}@% zXeSqZu?~j2^=wlP=IO3DX)*aL<%xr~D8_R+PM# zS$J|8Zm2-oP9Y|lT$hK&rn0Bbqr%x?bJGRGNHxdfE&c4@C47)9k_G8uQbmx^vsE0U zGuw68yC?Mat?Xt$m!L+#nZtywDxR6$!gDO-iO5s36AHZHcAZeOaKeSaie>z3+{yy% zO!Cv*aI7!LMC_M4t%t3u@)Vud9jQky>D|x)&_Gili`x#jb3Ijag(4RJQi`)!~2RN zz<+fv(bwk8x$&r+c|6lA{9*wSG1;z(Z3fu%u*sVS70}x>;n@U4)0sdT#stBkFr6;f4E-OZj`YjvmYu-zxp*QFqClXt?BYx(&g)aQ?7>HzsQuul`0}iy8 zZ*d&lnk%iZksJM7j(xw{ht%gzO^c`loWDw5EpG$ByY&W*+2Vn-~SlD1L(;s`vTP;l{fQIjlhZrlcdxpBMStIow&@?E?=%)FKU7uK=$3}cd&5Cxb4ThZ$EF-IwljZz7|c3Z+#cX z1EEL{9k5z)YeGSZG5j-52%sRq9+>3Tt_{j?2rjEl_lhl;pF2S1uvoEs;u`^7zskJ@ ziF&r=U^47UNZv&*ZdS?=@Nh%BSy`fvkKfpXKJwJVl%%IQr4Jnblnf7wrb#9n6s+A70=WbsabRPvzJwEEO zq5bwIG$uHr;?YYmL~-t06?I57hs_Fi6eCSQ-&-s5HST^4mn3kHvh3&h!y?k2fE(QxO%r&6mcdAiu?sU|Xr%=g^%6OP>O zve_uwyy=RT?AoII`g=?#s0m&wQO3R|0XRj%gaMPe< z98Q+|P()y0h(eEH4GxBA);cGQRCMML!qilu)$o1a_}(#>7?$J=hh-UqDRP)?{Zlq($O93oG+LY4joR4O$QZn^nOK3P>1Pdp&+ z%AZJ3|8%!Mr}Tg*#dGawwAL*h?$SY2Q#U+>=vCv%(I<6pQty4{8#5z>!qWjf zGt}v%{!T?qLEy~U*E2=3&bBrN7tI2MZp^LXa_coS~^Qcp#o z?VhfVGFv9kuuUUwuI)bJrvO$6>0dg63m@wyP?=gBO-sLDT>Vz@-vJAB{&w>aBZ*j` z1k-BI;y*ndm=sIZQf?W#rT)JrCt$|Z-Iu2{Ddp=AOuaItFl8_ggzoohW$e=|nb z_u*3$e8b>ddv`O>yk>trle*JS^FYLRbJ`)|Wr9^&;#S1V_@08MN@2~_g@BZAAln^o z631bonAAJk_0B7X&UdNa-t&0t(zJJ`t&>dN?bzi&)5df3O_F7lQ|CHDajuo)z{oS2 zv2V{i{UPZ;=8*F>jqE$>l;~F5rJdC8s#aWDlv!D7!K6zOf&3`m1afA!5%edNMPgj`6-2+u$!M zjf(~*nwi^MGCauSRc=`;mQ{=Z94hX-{gX91J|#}zrf~}S3JRNy)Jn)I7u}lqheR)e z)9HKikWMAj$LtVp=}$a13VqD#dmba+L)KG?JkOqqyl_^O5)_~dcZD=3u5ml+Wy8g~ z|9s)p^4Dp7sG1DJdskpF;nRsax1ribUwfKV=y=Uow}P97v+q98F_tyHyxr0I&E-dk zCs9_&HsqMoy2U;)uh@J_auu#JjjpCkFnPsw(ZJ}FSH>}fD7)vOJr_FUoOYjQvEA~} z0rJ+?t2O2sbqiwmcybH9SLtFBEgfa9?&~2eIc)`v)Y?G}Tcf9*(UUMwFtk?%c;Uh;4Ri(Fr86=sb zk6A~55k0bJ&l5&sI(^bak1_I_WqcEKOqvO^4%^ONQHNXtt1VoD0&LN@?i!PNd%brB zbkZCoX7lgPS1bX^;rgX$!QZ(= zL-uFWQhRDzB~E`XkS~prz!I}Tu6zA(Zd^qEO?q?J{Poj_R_h%`mY7YB!5(nuTVut9<=L6;NO8Hs zrG@0a$_h`W`gG)Cp6~FsKrdjKJHmChh=$1Z*>&c}EW&(1ZZtCDm~#bKJJGx~^0oTc z8M~<$6@Bq>USe+9;&v)CHp*ro!-uJ~D#u>KHL70BFeH7>1(0&%`5tRfu(62D9y(m< zvU|?%r&F2jx30Oi*whg2%KozH!-tLN;KG(vWh0AKk_#da`k4FNW9#`PoOib>6H(h( ztq!>O$%8Ci<%0t=5B}I>a$HsZVzXI(+^BeJX(i&L(zvH^tEeSSK8fC_JYO|cUijnU zPEEHySWgZ-Gt$J<+`ZW;SEEM58aJI=Sh(s0>=*Dv{)_O>G&Oy03tb^1RrH0(3A6?Y zn9#G<+=@vQOoBjDDSfN)C!~Te)8b2l=<-(Y(=~Zv0*NGDCwgrR6T0rhD(2ri)8<% z4~@h<%3LZIO;uwAD#yB8?|e4kQ|#IMbL8lQ zAM^VeB4*@|8=K~>0Ie{HGS!bs?;D+cpYLFTo(g8ujNRnw^fYchE5wUu9G^>55~$0J0jc08l|u;M2a+VDRh7`e^B7I4tx&s=T6?i%b^Ah4}d242Edu{eRUiGC;Ek z{wj)>sl|05@WTU7^Ho*Z_vt6zS=0?1pYpJXE+CE&GbyB;)RCe$c!<_L1p?bq2=LW! z#U32=3n-dTKv7Q##=w_rXGwVgi=TCHSt<~px-;{xjjwl#l=XMf8Izb{nuYmcZnWD$ zY)WtiSaVadp{RiW!-s-?2OtK|XV8af0%v>6kHbF*{A>3-+@^96;xRm8FtxqY zeQ!VL;*>0`kVQwi!qc>}SJy1G1fSYG(i7Q)TR%!KW$0*{_4hZ}VOHRuaoC>G7%0*$ zk{73zGT{Lb2q2KSU#tXs+1lBTUGi9dJjI)>eK2iv^|(4){rdOb{J)ci4~H%mrP)~wL*tB1?2$1oL|uZH$J z^`D0vB6>Q9^K3NsJnD%0(#i(s4wbRKJ7T{ilC1`ux1W$Rkt9kQ`VI!DZL?(&cLtFy_qhXA0FRe?suQoNGr4U8Xf(8SvLHxHtdLimZ{_)!Z5!M{vt*a zQ*pnLKd{a|i}9yF8HV#Uw<4^aF14ZMb`1#zEqc!&i!|ECjXq_q?#qNEicghRvd{!K z^{ugGP8Ch!V%lBTEU)kjfC@ejQQoeIDJ1g}dN3^{1GUQ-R`Ss96F2-SQ7tJVn48Zr zqTyHH<7K*eIlg?+lUIK2%dl>!3NoOAYbQ%z3_<2?eWjmnGz4x73{j{iFYxL7mGyCB zcH2vGJ}c8CzN6M&-thW#tvDz!`f3vOoA<>d)<_xQ#upxqHOs9NBUl022Ji1cj?Re# zdfffyTi2Fl)wA1QvWKSQv^YhBZ!JAsG65-l&TFx_ru6KqqWn^V@9K+>W7_iAx7%Boq`#~hpcs`tew zsVYk2q~@O+81Ah%w@q0eXm20AyGq<#_@vAB(>W zksA*RbMGVm+oBx(?AgJ^7#_P}$#CpUD^5*uw@^%%=W#zV7{2p_iwSHrfJn=Cc6-ZB zj&0^7qjjd1G8MSh;J@9KQ8@HfsVg#4qJM4oFaMzCWQ|$}SNn||ICiGF=R-WfwQb?8 zNbpNEH>&JklbllKlgL#;%kg?yjL()^rw7xxr?y={mq5u@-5#;p5|W%MbSqzJWg~Pn zx|{(k6b~f0s;VSXvI%L33;fwpyQ7>yLl%K8e7Rf-I8h{t-_sB9Y|{h~ni=MXdNVcH zNDjhc02DxPkkC}O{?82?kyX*b?HM^JENAPcYCI)=KkG=`nk%?F)Sw% zQPgesV@>Y8o=&o{;a4)J!Y_-kL%>$a?h`zbcz8z>`=LjsxOgWDiZQ^8ONcxcLFA|< zA+%I;%iugP4KyQ_36CQNM@7q^cOzmnXRgqt>({=NizfF-I&D*>&JzaY7lEOR5kZ^d z=};+M-6t>D`#o4#A@M%qC%{8gXr-tpHNl~+qOPjHQawOmmHd> z?xS~UQcb9G85f7dFan;msdoZ2FpB*Xuk&wC(*YH)(8+A>gLznD@X}6IZ}%#!}wpsL`h-j_PV6%ObD`bgtv6bFyQKQ7*ZBp>A_ zj@`q&N^Pb6>iOW>$`KV)j+?C({e@>=?gI6{mbEY(iHrzO5vLSiDa$Vt0IBA>IF8hP zHzY>`csQ*+w<()2L?pjT9pL<@5z_Z zltr*aiDTf*aR_3GT8r&@Vh-Ev8S2*t+@l(ktpVj6MDuuVa@FK8;>m|)tN#k^w+{_v zqzl03R_TU%7PZpEI|Q=1*NfYc7K!vxjE}uThh}e`U^?h_nl~n7mi~Zyjs@(KzCOf8 zRR|cq++mc@=yM*H$kOKgv%~u6Ffi&(a~JP#7XZ4FzW2g@WMSmD|HYBDllZgDe-+pY zgt`-BuHPP>IWUY*E^91aeS^j*aJ5{k@y+?jt>QPC@G|gaZaPRbC%G2n9Zsa0FwxXP zxHLM17;_!$*PS~|a}Ny>Z}uB!uKMatc+(x|y4AG4G+4Yogg{0Fv}UBZ{~YMVI=((y z`)U)oG#L>2CXSI_x9^2_1ND&MYXf)DU$fih2WF*+RA2JEkTC%38auVHQxA3Q?061) zS2ZAG#-P=f$Wcx%&f4N)5Tc#nf8YMxQT0BrEi|0diHcM%CZf=Rt8V!EK`a3>;5BV^ zRHsS;N#+ye(>T9`UxzSVYoCO2QeMXN$wxs$`JZp*`Xn%N*Tso6{s9z$Bbf}zKV;r! z$1p@)v#3smCj>WF3=)Cu3_aD(P=wo|6Qia9pcMbquX%TasV-k)kLeXw#XQoTG4ClD z(Bzj=X2lqP@XOFQa{se@@Q-p5lA6>s9oH^nB|G}&gMCgM>JeCOcFXN zqhoHqVm~K9d_68cB^`mRD9P;i`s%3->BbpkBmAQQ=(|gtPDoi1O+u(z^MM?n+xCKn z2UMd=$o-llN&vT{(82**RdAdw3AD4z;xz+zfQijRskZtcp(#U8MVM6f1M{TW9TVVO z>#6~xU3ffhZiW%Z2oz5zFjUa(Rsgw9u$p6-L+ z{Tf2%Nh~|H0-u2wj*uwu0>e%9>>J`{`scwPGqr%-1M+pv^Krfi3`nE|iwR+Wf{mJJ zMJAoH1l5;6Rv~a#DR}t!)Y^6rZaowqVPdA2aX!0Qa47}`c+_qj-|OjftPKVu*BjjPrWxGfyn$_HMF;uG8o=-vfE3gNq2 zK^XwH37ujQmWZRPe}Bl(1o*n7KmQa2?+Ho(`s?P_Bf(b`?Q*9wm?EKEG)E~vmKjP$ z1sRv(3a#kccR3RgKrN^94Mku#f^b3b$C3u*>}EfzfjSMKMW3<{2&=#s0iw+@ed^Ed zf|rR$K7#;b@(9O$qm^&Txb|FvBY)rGKn2Rs)jWpO_Wq zT#=r6G@_{drH(+!^@-Y?9Hxni;0P@K$E?Sw`th~s@*y;F{fyIKf*)_D4QfcbiF?SbcaNa)pzlq1yArSOUy%9 zfc+)C*@@rn7S|e#%*f2@Ys+K^$e`^JE9L0_=BbjD>AyhY0#Qqbzjdy!9;Mp?b`u?v z5~juH9tn5J@J2l_Lh8%~T*IDdq7cd{n-6^CN8Ux#ceG8}%t!&x*jW*|sX4Lh*KUnJ z(tZx!VsPZM_Zh-0Elv0KEs*i2HEqOH>+$Evc91bIYEZx~@A`B%^FHcWK3_=uG_>#H z_)Z+jE5N5oN!U=RJAZC>Ikk^8ek6{b%KUMZ!QFnr9TV$uMCP-wU1qMAml(s!0aW@k z7d94QUvcoPL&%RAX9k-}mziMjVCT<71vVZMenZG0C${+h$c2G2G$|1tQpo?)EVJJ6 z*dh5O@Y*q=fd6V^V}0fO9N_75oz+$p#Ykag4v8x=k&#`&Y1B5etDoHt-T1JPbNTNI;PlAIB zz#bbkw@eGBC!wVL-k$`dr8?%%|N0~yRRZNRiy^gsq}(M4!XviTD)garmsvzh_C#q? zoMVKt!nNIFO}Fi#as^X8=pfriK8uS*NP`7%|M^x0B1JNIAFuVS;=8%Q6GA2*Y8UNR zg)Cb51g^O3K{`yo*;XN40ARBF{|2tP zseF5S@+cvqwzwFJY08eXm3`)t_DJ~=gvNt4NO6vH08)P`4BQ5^v{=pu_N1y=S^X-u zzJm_*I1Q&$Da^m6au(Gjt`(P0_5bhO@y>Q?QdR>(Va0$}hy#V>>M%^aNVyggP3l+XE9*NeAU+(H|>vt@<$IF5QPHi7&AW;Ukqv19VM#G{m-fgK8Yd*`o$2!jV_B{yPvOZS0*4$T9SNshr5~_eIE#VI9_rX%|i|( zee^%jXcHH$)(+v&N>wMtz4`Zb{CXi>SuqJ@N=56T>PLa8HxVIa#UJEikP58sHu%rK zZ_KW+@7Jf2%4TPCWek(fRjz&P16;Poj#p4<3aF7ND4a94vEHhB-rEr>r-zoH;_ZMO=L6j(jJqWT)m_TObA&Vl{@M;+6Ee!Pl*c4}WGG(QYo8uh zxnwoPkh=7L5eMnm337CP1?sF6EKv+b@`Dv?tJ>QQ9u+DRX%iW13x=3+IP%1l=sy82 zo6M{aPXoNngL3WNe12^!qpE>PH2@X@2~tefa6Ipr}Cq)}=p=Y7KCtC%&13ExlQ3OcJo= zEwYJirHbe3Y@7FvwYV&q)RHE<;!}6+$o`UUOB!J~QJwhrzV-Qk-TKwT|Bc=k9KrX0 z%pugm-q()}VlAsrWJd)fDCo$Aw5=^QoLw+K<~;*z(sB_`c}E1>{QBmWP88s` z$G`1weAk-^3lI)6of{=x-V(`FO+5y9w5(D0t1_hJz9$IOH0uh6RZ)n)VCNoIj#W*S zTz~6I)APHc@9)XrZAHJK>mc%JFFz`?-N`#+%@xoPro6P+$Lm-7BMcplmy^F*W~ClB z%JD{~hNrBk%7XPiXVO2WHer5dPL!AUIm@*|xE1S5g?U?JKkzjdSpxx*cR%K91UvR7 z5aO~e^j_PBZxX_SR5iEW3z z4bUw3u)$TpO1gLYtKI}Z<|PKp2qURF1O5Xog39r9aBY)BIv^1k`=)OE(vGJ(S@7edj7r2BPA$w9dDCdgZbm-nf`C*)OLy^q zePwvHoi?WT`ZzLuE=<&h&Au`DSKn}y445DB5y*K0YxKvBL(<@JRr^CSuBk7^Ax!;W z(>`(IIG%U7DrBq4XZMkTc^CyMsNvl4NjJ5PtLUlCjVEjsF;A=npNRX4HWH>a4K?hnaq zlHFvpg+`as`J@et4}?#iOoqp-0%0UJ*#;e!Nrw5V(OCK|lZ4-;5OP?eXe7yXz2^er zo&Q$c2E95yV^#Il+*kh6zUQX5nk$fvR1-=6wiX|)CZ_u*wdWzL22w-mz0dp@8QM`? zo3mU?Aih~Pk8wx{JXQQe--CNC;X38ujcRhExQA>SC7bpGEzizMR}VnaBjZE7IaK=9 z{%=njTgC<{ws>N(nnZ^zoH=X~Rl8KjvT@cqNpBRO^mC|7SMS@^@hn1d z16JVf{2!jqGA`=v3-?E*Q;-H}lh8I*EE`b1AFfv8fmxDodNq#VWbKY;%2RUO z+W2_#q`*0OZEIDtbR9Zwp`5WdL@3+~@8>Sg62~I2>kn+oc=vxNg87uF8|>X&*X|ep zzWK)a&U>%g@sHJ30v+O?+rJ~w;Qq4Pm24R~1IEt+s0rVGD{6p#(9a0wR!P=TNK63~ zzi=dYI>&*oRd z-3-lHY+@h>SJ{NeF=s>quCS6dpwTotBA=C!i*OL3hI6Y_#Zjns>|=lrg^3oVCf?kxre2 z(r*@;yKY?LQkNNC_ND||{$Rua6FnbROh_qKKJKfU{wVm4MMHa2N-4oJR!`=O@6Kti z3{5~hv=O*yfWp?$vF@AhA@x@|DkHogK1m^}+^$zE=EqM+0l(c+U1!b{o^NXnY1%^P z+FACm5wr0aWIBVgYZqx4jE-17qE5B)77cg_*3>G>$zBRNDRn%!o|wj3@oe3xNU|mU zlM5fLs&}>lf$QKeWWW*B5=p7F6d2c`8e-AYv)%%A}C(SEt4IKZ%76D@50P5@?3^+0Mu-0_}K&Oe@|T#Z>Sm~ zC9Nj#;3V-7JguI2jQ9xZX9K2wUsTeUuDns0r36u|TFIETm(@yoAPQixRdy3>n-u@= zi$KRU_M?vVKH6A4+waY**y1J@hkCUY@wLYh(uk!WG7_WbF)TZF{O#Y+EdHjvh}7vEMAx|-Iv;)<8+%{ zq1dTb10KasSyF%_8=+q$*m%L$AuBtIGnHX7%*e$x*=9vU#_Fe%k($;5pd1m2S*u(h z`id)=anr+Da!pMgnLQ0$Bi~*5R*f(ZTv=qf2|uG$%`xd4w^}S#A4|COo!TfcrZh~A z)Xvl-{`?|=am2kN2y*(m(N&A_4-RFt zm5vQhXM6^IB&z8>4^JAYz1qV2eYioy=(#~{^aheo36Vah+iHuxB{eI=$ z6@=-AnW7AbEo)6*e{hgqoG${XK@E;jejO_x?aPM=sr(WL#YfAFjvENzY2O17*pmFGHzqa3rODRsK)Aw1%oSuuxfpa~NyMbSkBmNS z$gZrc+tGt-l`r>M*PB1CJCJAS!dJ9v2xA&7^NmQuPJ;uNHqEBrzW6BONf>GSo#_D? zscH1%_F9)g1@N%5_y;*kqcZTQ-;4cm7=<`&3zWu0s(9*-N%W+m8v)@`SS1Qf?G=8 ze+=PQ!R<-TF9BNb%Q*zS$qP3aQ*ou49WWCx-zfqz#}G&yzKT5^%(8jJdyL8=V25NG zkep6mL-2t@P9sgMMqFp6DK$5mWpgV+^W5OfIT8sP_*)@{AOmJo)0)&(q#Es6D)C8$ z5{VUVTQkui7=}r^X}C+zLk=tbHa-E}V0!B2My)Cu z_oOOu<&5B;V^xSnwFYXR*pa#IQ}#BSHBDG{;Ju%w(bM7&m5*I_OCU)lea>fI=9t@x zj;duI!mDjFBge1;X~2?!JCB`{B&DYsC45w0)knb!3d&6>@lqHIjq_eJIkI&S1ZL7j zl_;qkwg9i>nuQuW(7vAl$awA&+`lJS%#9z|P=VH)Tm0Qt`wzR?iRr^(^4Pzti^^0g zxyH_Kp1V$^hGE-fI@`1$N^Qw#q<6LAu3^`YFXTwHcOg5m!a|1&??^Eokqv#94a{{u zbBki^DrHtrI&`^bKKUBK&4_L)Z$H1T`peA{&D?N}i!|?Aum}?tb)ianW#%t4&l{r~ z${UR>!y}rjN&82UDRfs*+H`VG>GZbWe{6*23&m#`ndrR z{~UnOSFAgGOle7GTb40ivCtB6IZyK!2LRWh)7j&K^p|xNc}^I|hi{kHR95xft~R-C zO{fFrSgH)GpCq!j>n`6V@}KFWc>3p?OuE`iFIP4)j@JTjerpeJHhY-Rj({rnF4o)X znwwX?SW}374eMi!(Vt_hXP2Bazrj$X8d-;kw zo*Ji)eJAvJLEiX600E8C)Krbt&A4!lF2uBqXU?cY zaq~`p;h2A((iCUtKHi#Bmj^dJLyt<$PUo#nY+|CHbI|smzK_h-YAe(Fv6?hWy$Qr7 z1N?t*19Ej6esRgoek3S`colJQzLSk4Ax#`$A)s>_mbV|dQaxFECPylg?DftOGpkCv z;U-ngFD8g@6fYJQtYhg$ms#r4Bo^7!F2qPA9s1EqXSiGbJrIk2bc(Rm;$| zG&ElI(>uD#ewaIBiR)|;dtI}IMrW0i{f3G~DMdTp9+~Q82-+ zBkMJJYf@jQs)ZK`%Q)=q6QDP22V_8)DArm%b?<#jdVWDK>N|tW#Fv)M1$lmB6rH+A z?Dkj8S{>HBK34B}28`SF(Gxc8Iswnh*W3YfeYHkpZm$dlxL#4B??*$g1Z}lAoDgp~ zl>?wAyFgGqC?(i-7;8$D9XZY!hI%Djk?+EK^DNH%cH{ygFwXgAeIYf+GS(UFD!H-h zEx*FanfJa$l2>KD0~T@lD!@An+?2SyykEn;q_I3A(uz#uu97gCC6gE8Aup`NVTKT zc*R%~zqsz7{p7EXzW!yUQUONCfi^BQN$1X)?$tlKMe>Hc7Y!_n@F)dcxQK6iWTN<) zukf?q7X-1}9&L%fIZAwi5`TWyHhCu@YR!^E$#nYioziWBi)a4*gMn4fzBsKyM37gzWbz6btAN6l?T# zw<;9245RrAD+*^Ql2{o}jDL2qykj0J-EjlOZFr>7yPM=81p}FU z4hg9z6AtHyQ#FlqTX>=^es{j$G8G%y3_%ue1`|;$$D=i`v@^Qv1jJzNTRx1+sTKc! zu;E!+k1rZ5uxiG1_eoxjggKtiu%(;k*O?m&FE69#I|f23Lh}=L>RX;vGM_z!k!#l~0~&U>S_KGP;o zh|ezJ_kO9QO}}f9IRY0zoK;7EOhdO}h6s4b4nQYNK zi5pYZm^Jo3x(utC6DE?#u%=?mU9${2o`4d_cjwzh-qOr9h-X-H(}vXsOp14VNmX~b zlE>Ip7k(?J9je{fZDYbU>TmxF-hJ9kB>&U%$7LLXi%5^ONF>KH0ZY{I2jVU}w0&dP zLX))Vw#Wb3QI&%ROaxw@M@qvD?ed~FP9{E#`>A${Ikp4P2ZTxHLlrAu;W|B2KyFC6 zJXN@PrzTiS(b_(e&GB4Gx^A}Iaj;~DAnBhaNJtXi4}&Px`2I)Vhl8`HclyNxw&6u3 zh8DzF6Jog+;~pJJ{^d@rw&~wyDC~F9Wg%>~=@zzWg2qw{$$ec~B%}h-dJF%8(+*I$|?Ij*O*0iNsBV91RwWcN(!p3RqcVjA(G< zG(2*63_s-S78kdv<7u z-^dh5j_;4Df5A48NEgL(q@ItGFA=5mTXDVOiukOxQ*!3Ra^n4r(NRW{N8E%%Fi)k2 zw`|@l4Hl&Yhg5`ZndzPuNQNA@oj7M*-^MYxoO&)Q6mmzw6se;pd~5m`i58wifM(&| zB9L$4zltO#`uOBj=8i zTIl3mR)qsaip`fvp5ExaC=-p@D07;L2dp`BKB6=p8kg9|Kf5-d>1Jn6BXP%Zm3Qf~ zFAc2iKB9;MG6kE*JxC_sL+2cQ9zK)Qa#!{yo7f-vBI=EmlZ7;5}yrq zXLcs&KZ(hQ3Y-855gXd>5B(AIJ`4Mp+_k~uKJ%*2oXY(B)i;ydxa|rb&BTX30#&4= zf-{j3fn?)Q7Mn71%~4%~sWdI88d-w*Pcd2haGgQ+h6VqT$#1Tie!K$?BaFD>hI8C_ zGvu1v67c*YfQdCHUTXLe`V*el9zk+y6b|mhwto#GE9}@Boi}f#lx}~6l?{|ZZVJAt(8SO2sQR$^=jlBlKm7k(07k*TLv|3#8g8f}=ck}|WG%)L zhAg=4Yp&N_EjT;;{p+tNE*dREe@U3=C|6MTcGn{h{AL8(#UF^hp8;WZ$}!G{^R6YA zURB3Mc(i2cK3Chxa-(x;9&W^@Z-*8-y=QD?;y4W-%ImQVZwuaRzg11q^fm9 zd!@MZ7rGBouwK~jv*(2n>luFIj{#s4a2?qh@@oYKq-HEeMU{tDbJSe}=J;6FY#!QD z?o}s!wDCfA^wCcNS@8(ii4M*S;jR|@C zQ^buukEIYvSbL5xwI|s*BwKCJ@g-V|^q8f(yh38;W__gNA@@fn``@HONeqi5vQuw& zn(`eeZ?{(0Q;Kps`6?~ZW@6jjE>ZCiRpmO-?AB%YLo0jm|2jGA|9@^0fdaWkUnv3N zf)y2Otv7q40w#F;CIm--5Pff$*cijIqE_YNSZKe%AH~9r0alnw-2IGdh6fQLIa1S#Y+-c|RXPSXb zTWVKFb?#r)X`?-L-TxH%tC%!78bi%DYsI3QTAw2&L=P2ocf61*Xrt;3(V1Qpm@ zjo5xC*F-%o;`vlFc9)$@>_W(gI|5?!_0CC}V|!P2b>oAW3#4DF^_}@)m=f%lcPbrs zX(4P-m&kw8-1rPGqON6NX-E<4oagmxP|6ZiXJxH2%N{WNl91a7!F=Wn3+pNdP;l@U z|E@!ovl!+O0f@rtcmym{@{yS0Tt5jGuBcKUwm9l3(e1W1TbOCeurDy1bKqf*M#Jno51m6g; zedD7T{VChAPA9`uHvRO#mK-^4o7Q$soforlx|fkXZPSW^vt!6}5=5Eo3PwMtAG6Z4 zLbb07Db=k&pipxv$=CrChvHCKKq}g z$T=R-j0&=4)-6C)v^YrSjW@hk?J<g$Y6wX&_a5W^}BNB25$Z2<^8Dl8c=- zc0@9B`@5uEK{Hrm^`J(cSR}uy)aHP7I(d0iBBS~q!qj-leBEb^^oPD4Vcs#RZ{mPQ ziGczaA(+gpFu!~e)fZxVhcLFe&UHfP!tBCsq2I(+d zj_oXGo$Vr4i~IvY!ne0*4+_wmttBhDUe__!ekwZr$*vtM-P#g77Y*%vv1kWC5KlK! zwK3!EIQ#l0o-Ww;so1m-pd2gN#%h}Dvun8n=yR&v4$19z@4NFe0~(=b|CJ-cFRz?1 zR{t$W^KOxA&nvGEwYn*VRAh0R;Iz-$;}JiAkkAZplQ>Ebn~M&y5t7+H=MMiE1DIvN zE|}`62Nxy0hmTLSfiH1v$5RDmRe@=MW`yzZt&C;MB_3~9JRgq%Pqi~GSSz_+ zIwCzKd;Su|d_rR>4w{vE#gn$f<7fF))$&otlU#->kSI}OqyHQ%(KtMwX02vzK&yxk zuf+u6>Dl0Lxt+v@a2zauZ?-^*T+ZgN7~^Wt?na`6fzNEcS+vrbDW$*0)Z`njrM@1I zyw}gWl;d}_SP9{@O}}b0vd!P}v2)ClGLbjG?**3pVhC%mivf>$fGzEFcxxBb-tN*$_I#Ca;d2R)j+qE+cLpFp0}hcl zUb@z5yk}}ZKj=0l|BPpS6zkmXSx$%8?Dx(Uc1W|uoUJV?<%Dr5&^0xQ{w6khjL|dx zv}HDB(}j%;4$p!s^e-@@0(9K{%|77Ag@6V}z-Nh)Amx!SH%|SXvPNz#uyWd%_v_+M zUr`{pk}u9J?ze9wm+AVfQnm)`)}3ZjI#n=ZF%xXz(kIRBDMs$(v{QEz{qKeGoHD(1 zl4ddjGul`y7jzq&Us(6-X?q_X=>}=rao2fYu7C&m)6rAJ7q76JEa5ZwxQFEV2^gbd zxsAFGfmgm7I(obFn|ZeARwOB+_VSF?)?o2Bp~baru(j?pOmn^Wwk&DN>wHl+^_THk~*lYNE&0lY|+Ra239n>8IE-vTd?9u4J9>)$R3DMB+wiDRlZCZWgG&phwTq^M!g7sv|8X@e*y$a`HdXmFoM7 z=jyV0{nQBT>gu+gkEr1}`h8|!xG76taXHhCH`n0%U~^2X^un;n&$DIcd|+dxeQPnG zyOT`UdblSEU=ruv-=AuACpBY9vYhjb>JlFGGsh4h&{?&6Ls!u4Qk@zemV3D9L(*Hx^~>h`jJz6&l$jY}xjfw1~&sp}|@)LVa~ zpa>Bz%*Db8)Zn6_cfyW-#xUj~&fpvUSw;4b_$c12ba1;B%CuH&;Oay%&M+y>kN0xzVd+t)M5$kc;M+U|<2!0`(IKiRRd-+Q zKfXEWlfuor_B4|%k3ZXuQqz_0zhY^CoEXrS>wxvWrl@6RT>7Kt?Hy zYWT&8X$NF3rO<(_;;51~X5VPEGUibHqlj?*rWiJL-tlQcD@%rM4JXb0QDBshAV@x9&g2 zLg>^Itvd!jk;bmz6I@ETj5YtV`%Iru@^kJMp_pr>XxT zCle?o(V*n$^So$~?NprfApr+AY|0a68uUjLA<16JjaF79h$TVK((-y@mU`^^Z zw|)B5mng^4IHJGm-fyY4Ng#I60f;YqWAwJxcTTO6;mvVIjVt4QyQ85BoK1EZi?e zS*26l^q;A6|7RYj?cd5=y&>UMc6n(wg7h1{YiAndqT{*n#-5JZpv2}1^s8pYb3StZ zuGb-2Wtfv?{rknqs>`qJ3R#WwZHC8|1i85sx99vwyP|WBa%DZblk_DF+TdqHSBC#Zx(ixF_7M^`SQ#vH&dO6aafz%l; ze^2&veSTEIyZjr2sMB}0MnxIfCZT{FV(DR;z$>ef9yH9*@H1pCA~Oi(Zrh6BB=Ys` zuz5Mr1nZM3<5OUuvgH5j6(bs-N?7A6aCD`PB~bB8+yp=L_t1H)E!U`FmNwC#Fd*afI}KcOaj#9^@m&1hsGaYN4xgv(dTO zJQb&x_jdw=$RgEFdquzjr*?99!^pdS(tS%1s0TefvT9#w5q2U zWgE8>@CnG#C?*#f@&F9yCI9>MdO< z9@SPNo|C$q7O+hf6L2w9bHj`MX;9ZmNaM3j^EU^w9On`RwJnKX@F0mMs5?1ST5%@8^Yl>Hb;Zx{lHOp>v8*hN}J$u7;lnK>zYJ0+JL!m zhf;Lxf{jo8{}iJk%GPdQq+8)?Mvi%7|ebshY6c-y0ID zoc;C@s0Av^62m^hvIJGI?e3pdm);9k;(I%5x_-~fIke6j4zh}A-%16VPfiz4uuhK| zd`?<>qSxpsq>S0m6>%@l-*A@=PjXvBFtE%A3Z@eMDqfLTt-~eNO%3LTpd_fIdb0-3 zc~$OfaOVvjW!04jyU5iRI}#1guy#J}uGpyOJ+3h6%C4Zwb0*s1i8W5d;4N4cQ}Ux5 z#rqVran4%!rSD2IEA@7BFhgmIVhOxJ{nzg#7^-F8Q|ja^$d^V=4&u35^65X5kBiH# z@Dp}y7T5z;0e1{Ao=y}a1_NRW9FnKD<%p810%yIPj1cmJ!@MFz!qB!Y-8CD>r_Zo? z8CG{UnlY!0YPnm%HfvZF7)o%Fipla?8cXiSW8>3=CR-D6PA}R|nnRlGMPjDwD&*DD z3W-$1q@)ce;)G=bs=e%4#s3&zs|-|Ao*pRK-u|&4Y(E4-Z&zm>8|S~M>@I}nyjuy} zCiSBxT*G3-NHR9b1Y`QGoh&MsAZLCM~KCVPb zUju)mw}g>m2$An)EdBOfrHNAMx>D)&$?`&cY?X&kznk7UnBexc?DwviOgCFQ6M%F> zCUEu`C%(OwB#PDjh4*d>kJL5hfpO=^JNS4}y`!=3G=IMsYY1lGK!;DSsYbKmZYjCZ zepi+|Tp(Psz#eYd z8Qe=;(TYO7nip06!I24@vWGaTD1u+-Nl_?-JWbT4O%^1p?=oWo%#zA+7~pihT1Nx802ut1SN4oQtyN4GcKp4ie%53h z1X?vQ_HG6iNKZxi{TOhFnU<1O)aKYPw%Mu=2xzkKp=yK8@3dy?%Z_09Igi77R(JpQ z&SfIg4A}4Q&@fY3W72DSpN6T_(M*AAUQ7fuYd|2N`T(7TUEn&>zEqcy`?+QviiVgn z%veoYZhmTN2r2duAHy32cBoPY`!5Zq++ntlTXqO;{ms2EK4y64On^O=_#HTfOdMf~ z_HIr<%M0Fgx8>S7*9~xnT>snm-~Om2@f8Ca0P`o~xVhhlGyJSCR}bfW=CE>xm3$N< zbOCw^Ryv3m`QBJXZT+Xoe)S5zpO5Tz|86Gb$Qm0k&>Yn2Kh>@-C)07{4W*8df@*EK zS(2S}$E+ISw<|%T)}EuJjO|w21sDUm6-1B12nvFgrba>YRGI8i;1!pIR2{# z_|W9o5qMH`Z-d*4OZNnh!n(d*i{E|nfa4(QJLXHr9&8`pxZ#e71O4UIDBy-drf-<> zDu*PRe`=Pa;}!f&3A~X+w@c>|yPos4+vMN}-uw2wwB@-$CJ&Mq*~B^vXuMHW8GwWO zT{UUA9lP%iLyY{d!P2plcaA@<3HduN_PW!-FIdoPe$^IQFf zRgM52k9YV4nr@TT^J>>77idBjYB!wl?gq@cYQWA4CTgDlESbNixvn$tagiEk=EB@o z;N2}Hn~vG9{5#EpXEdsZm-qZJ-|nBcF&;%df@enOFFx14LIE$Vv=UDzDuxj^r5<&h zhBoPDAUukLsDic^XP>bml}In&nkax8*U~)*Bo#Qc`O>NtQufDAQGH;0ly~*~SZ{aQ0tbWB@me zUTzn7+^2Rnqooi{q@+JuaN@f>s9rpJINuA7E3!)K|4qb#6(SQ-ymeciuS@i<$gZTu z2^d)?>LqGZZB{??QQ!hmlE+N5YAN#({-6!BW(~5$;ei4sdhv|1;hzQjH#EO44{my( zzLc5Sasx5o(BpY^p>AQl>HeMeyN&u$1Ey9BF-tSoG=5_}eHc@nA>>JJ>EURY03M?D zgiZk@LLfgPUbCy<=DKy-9T+Em>7An&GyO|Qj)mOt?rXLYKM$S^FO4@AN&7sDn`C$( zi%419h8^s~NAgN*EkCBIMP|Y$kH>7j4BKi|Nq`*HKX`&pk@Po7^s955Ska<%BOV$y zj`Zs5xoZq3cS}#cJ)=E#>C38^lgcDtZj~o%KcZ?Sx;<-ugn(ugl;=5fmk#~>rxG*4 z6>Pa%UG5ZpSYtG2Mc5029$!&mw6nf-NcFPz3Jdcd{v^`l{iOPNA(=Qk^H~3z#7iPl z{LE3W@)V1Pa;$%CEva$e18+TM z-Q8NO8Vt0iACLW;3GMuif~Uq73_o^v+(PSJK-Ly7D&sT$HGRI@5*{q_Q@_Bd@VEDn zqwOhuO2_SBkYcGke{GNze;{5r#8^2mzC5laoR}+ZaN;!K-Ue%mRU^F?uMCTa?BxRw z-!|PQ-1&K~PD3UdDLCN&b{1c7L^aF*Gx`2`%_dw{`YdzV@y|nLPmS|xvw&HL=jHv# z*F%%nQw~nF@?++c_*887jiMPhq5)uCAdgr?lakb#uKzA=qN{x%ecCv*w|2+Zn16&O z&=)Y0#iA45FP`ja^(QrNtsb5HJ4!;Jfq~n^RkAHSV?eEjj&`h`s)!3ez@T z6jEC{PaDc^2$tfa`p#5Vn@0(71G$3O@Wd(g8tJIYk(7W0@VoHuz!#1X zS@xF}oN1M`J4Jo7CnLt;OCb?xg<%tiDNQH(=R6C~*_{6N-~$z#&j~~xd?@#JrY?)m zLwl`S7N_ZNeaW-Az24Rp^RAvAjy%-K7(?`)0`K=>2x2OOHFggnW2R5WV?7naQ-+_{ zH))p4g!6fF965NEfi|*SB6?YdX9+6Q0!E_*qz{vS4_=A(nEwJ7Uv8EI{`74ObPcXE zVvCUY?%3TjD@>XLv@vD8%$7b>H8beQ&(I#P5oHBF;~TzOZkow1i)nODYw0Kj8z7H; zE@%977EexCQNGT;7dv0aJxqtgynJC=9I5yG);bNp&#EZhMq%XN%kVOJfj(KNd~xooKThH5cXcs#*ggn>tI)( zT8t-WE7N`&E)ycmdLlJ0(!0ZLt)4h~r!VN5^lfg=o|Sfin0Sni9weMimj&$n?)O!1Wyo}fE}j(j`6n4|yhU9B zA_9h^muo}HH~CLZw=?Ewm|8(t1)aZ&oYVM#cr|EKQ`M8Xez-_+1u$(s)EZjJzMr#D zlvliqIP3;S3dZ0DPX3l62Mm>6S)Y=f6ssH^jM&YaR-%VL6BG}(8^bQ=J{Q`fb0VEG z<$j{|O%f`dr>#{kz@qr7W(l{;gVrZIpy3C!UCiNSVIEXqr9Bmcw*vBTJs6DiQmf!@ zWBiu-)~@0QmmHt_kHRqnW%;oJXT!IKAtk55I7I}p&ecxg>|e7x!bPjg$vbQX#<+fW zZDtR~oA1MgZvUjar-#`0PdJ5<(ANtquUO+Tj5x-o0*ZXCRjA*`nIdfc6f}5h02gWc zBwqq;toEyx{+5ezdsO(-Vo>&Tlh@utAUV3+e{^}(4<79{9DLzQL?Z9zZ{}f1S6#_B zMCy#Q?9zgoWwr4+Ii%>h|1+2f@L{BJ#4bg*w-d=a9ul?v4cv;Zg5ucfQPb) zo48{I*!h63V`<%)gJ!TZ@D*w!{k0(x@ww7ly?`$|F6LPAde|U=S)klwL9<&0hDk zb&BqXZc}XEJ*{}GCXDxh>?$?{1d+mTv-mwpYy;5vOkWdH^$t0!@&h`{HgFp*(x^v3J~U-Sipsl$<_Xw$c2CnIODjNR}o zb49MVedWGBf9teu{hal+1GHk=G7ama{DvQqmYJ#3s6L>(7+>$X+)g*}6dyqHL3xFn zPBZ*fat8Db0(J%ZA{2g0o2^mS3;gQx4cHs#i+s=24^2Wi;=>r>@Tdz{BX4$^C!LZ; z=s|FIw|4M0zdFl{A8K?;yDvS6U_3>h`g1hNFTfiC>2R?l)2Oxbw`#hg%k<5^O^uF0 zZ2A-;vhx?sjdv+Pay}+ho+SJp-L|jCRJbGcOg}c<5_2oXsujVerQz7KP`f_3kO*?Y zc|sf`w|;76lGx~RU4sB%h1|19W@09CB_P?`aTk+5$1<&w>(8qq%I4*j*aTHLub~Yn zXqre{lA9|?H!+Y;&VIx_8_tWFFsEPx5Qw_7y)vu`N+2=<(oFUfyG5Ao6v5c!?KLP; zJ4zg<+K6(j2JDnM?y65oSm~Vf0uwah{^g4shapJ0kK2)*ZFxaO0Bf>PUNxo!KcK9* z5_8fM(@}bi+0E%wg4$*n6$n1#bJd;Q?U)Tc;K!1Y9=+URN4c^6C;~7GoX6PZmFmCT zpMUGCAuD+6VM(?zPf_W7df8U!V(D_WnOghFw2-nT;9>Nu$?wxE@IVrr{9uOE>UC_4 zD5SUX(rjrb`!Mw3>ShTDS!%mNwu%D&d%h07jI8VBS@m{5D-)sV>Q=IWM+>`c1u^0O z%LR}yge}u7wsivRTwt!XX*IiHpbsY%u*T?=7a&Mg>&hLeiUkf98B>_Y@THd4{piE^ zaV9LB+=3&zz%`ywN5@GMbv}3ge)$H%MfZq+5Trg?jG50) zTYxw5wU6;HM}7FQxU%Jm{D8tRUruF7=;jJ7ke(Ktqz*KZ4cY>RHoq^PyQq_1gTGEf z5NOEYMh3$gL;L16|HphZ^A~B(ZLbqL{H2eNt4e#1o(K-N zeY{58)a-RvtVY(<2AAD(%RryBF`ZD{o`s0Tu`d3HyT%(B_=*BJ&VYm-Wmiz{yJ`b&1nE7^MkF*gfeGd zf-3Nx+V6H~HNz$$@^(%LKV;ia@ID%e&BD*)6HejJg8idqd87+GzmNSpPW~J+QIB8G zr{sOn*rHJV5rRNTd@qB67OcbjgEY4m?1Z&xG@vkhCO0iBzmY;jO2J^&&)SxVPy_ZO z=aao{Y!RcsZ$cqMcRGP(O z_{N4=c9ZWt5pNht4PQ0Wx52E$WX7qX3$GB)t8@rxr6oQ$PA#A_AZ;`zVbtKCZ2}LXorBMv7{N; zXh&z+Kw9fd8F5F^fOt7R`g>xOSzhid?5v+EDo>NF&2EV!jm?CS<0!QM^0M-c?kolo zP=uYyf-IGp<620hBuzaUN=q^+=FmS3h$RWAlt^NyQbcEB7(ZkLp-y*=;>93}qgU&* zyG8FhbU4^)-q1iO=Y*p_2^_E1X)#-TM=>tK@)M{mNv?T=Y?H8f@8(fXN{Axs6+a?>JW*F=U-9xOaqHt#il-Sgk}-=YJzPNyzjvUVrhLLo^D83qP`kwhxS zZaFbTkqaUAdA_>x&d;@T?)lm@D!lPVxqcIk7V#sm5mR+@5;qjFGitczO&MojaP!oD z=1XG)D3-Jxa(PatR{wrXaBzu;*9R|&dAYY*jnUU6_UrFO2>lYwqivG0Six^+*ej!I zX8KPCH?mBnK`eEB#ID4EyhEm%z-2Xd5lpf-7TtWr?xS;rFyu3gjg+;{cU%(LD0R|I z+N52-gryOcp9oD%Y7KJkzi$_N1D-sX683 zDssz=*Vo-7pzLbnKhM*+=098GgmSpW34oham8-b}5H@L-(b}#bvfMBMsG~L8aof4X zF!ER*2deg!Z zHtRA|n#{Jd3`K3Vrgq71&TK(?Zv1>8n1~lk37F~{A~)WA+J6@+ypJI`RwH5BC?2_a zZ`FmZ*Yxs>Y2l<)HcTorg36Z%_mRT>ev{dk0`huQT9D{D}=C}d0%AQ5Ts%xgpB*0o~aB&Yh5NzQJ*3hZjHK@HWkyi zJO3nVGylERFtl1cgYs>V(mF84G&!IXTXgg6=oiGIi<2W2;A4%oW=UZ^FGuysYps*` z@u#JKfJK4H(iB0zzbZ#jX!rEGQgZXbCovj7dXK2yqz6FMSW zHXTQ*j&4?YE*+n>C4fCDsFUaHNu6pn=oyqkw#5q%y*Rvh-`_1LTE?uz9=Uyawi%HP zj>~sQOCg`+=Ah|W(0&S`uXr4ZP~pA0N+QT}>W}OJAdDUHjW2gghPSB?0sTn7kYcFR ziCv@-_MxP{%U4lIn?uB1*ps|#fnqxH1^7qsd3*A^ye#kMk>y(=373}R1j9) zq#$dq9v1xJFC-IhhmKM7{=F-%N0HC#y=y+P_$x0ASS7}&SPHi9KHDu&ElyC(pYM4A z8aXm}KTkUFcW1@YhuBB@Ek(;1ZC^3#KVE2zd& zLRS-$tox4R@VJ8!spQ&t%f&V=0|TltqU(N4CeutjA}L_76lJ;G54`r$_*f{*;$OHf zWIT-34L4xWp@KyJQ%62+pAonFA3b%WfIwknEUDw_c*eKnw2pK4llWP&EdBFlkl>p&Tqo? zpyA`a$$*)bd&iK<!`bE4Jj&~&?4L_PBxMRr_EuI@f!MbupEa^Xd_Cm;vp- zoWPK~2j8cSV|o)S^|>fN^Ag%cEp?71sW*pzd^3X5pI!&&uFwU#ih>OldmY76TWXv) z5|S^qy0Y2L9LG2868~HVM5-O+(_Aobjw2Sm(djx-*19B3u#k@7?_%MuTM=-**xzCc z)7&0?vN9PSQhHexv{DE*yd{FEkJPa{BE)=-Iis8|&S3$cEKe34Hzi2*f&-QGnq0Um zSx%}j3da4>C>;@4OU%t$B`^Rf+;Nlm#P*Yc~aQ7(0A;7y(xDXqc3GW%%Bq2YB)@Mh;zOEHuly zjD984(=nEmBF8*X85$>f^Ca;f8*~O(EMUV&Xx9C~N<(Mvj5)k%s-Wbfe|DWBp4z(K zfU5{pg@fv^%M0&t+8f6ASM7%;j66IV!;dAB z>vya&1dR0xy$t`cDu0a!6@6U5CC>zc5wbzt4U%SHgB%B^q5 z7VmlsR0eAY*va+6p9U)!{PEPyuh_cl>86!HJQ?dtG0PAweWQH3tQgoZnr0h?70!Q_ zKH}lvtPdJEpef?UDM^;6RU-{&TtB@^E-!x?7%j}3bDVIAJBeRYX=6fz$C8#RQP#FI zY7+ObhYG(7mzPDRGlt$ zeD{lGC6d-nG0*woLpmDzHAJhon_XV4m}lU?Qzakph|pLnSgs51$p(kxPh`+BJ)~3*yq-t7Ly9UT6k$#%mHo1RP7&)L=Iy4Qh zi0T>K#DA}wrF|FCe1T6>5g|{hr0O~4*39YWUegbL9_a1S>+ip!SNR~1hFcgftJ`eN`N%Z*f0zv z0&SD8ycY~WyoqQmDxr;9Bn>p=Y8VdiAl zVsHdtPW-dqC3pGgTLG^#nhl<;@^&zv&^;crhS^hq?vz?({{tzE7yd&kE|aGtvHKG;>_P;Fu{P|5Q$x2TBwvL$3s+3Y=2~AY3 zMH@N)^`p=KGc0{ zYtUUX(zx}Ph-p@I--)&s9whvoz&rA8!g7ru7s!lZsSRXm1Ef0BrCf! zv?+t7n@(DShI+}s6*1sL?HzEA#piSRYciv=RSbFmaifX;nBm-Xo&e+s+7XE>ras{O z4$xUE&)eOTj+NzJx>YOY)%hL#Ni8pllseVmnOy^DbAzG(L(^HtMfC@5UyzoTX6Z#b zrCCtABxULD?oK6^?v#{9U_rV;8brDTBo|na1|_7P^S|%=dAF~2&u7p1&6&C8n(uUM zIcf4uaYe1>3&#?<8~3%*)ti+45;tjBCD^D6Iy<+T*kxA`L~(TfNdEA+x<97~3yYrY42Ey--tE^5A<&&XoNowR2 z3`7Z=K?M!MCuVq+_L;Hq>EeObHq>ho`|>!blKfNTAeuD~oj1?W1x+Y7nOvMt zFob&hMld1o*TR|Qsf#PQTvUj&z;?#*)+v_G1H7qD1dJfvi#vs^cUlQT#LmGW9>!`; zgVn2B!%SR@B$FdVC;pmR65E$e&J>U!e6gG^}R-lmWY+^GG<#3ii8dRGAFA<6Ca%PUvvTd4%s|w527*tN@;D6`T=pg?qLWJ_)pB$uaNH`CzIj*iDW0iV< zIS{SIoDwDF0Iknd0~$Yk!Qf&DUSC<7qm5`L>9`PEs8^}yWQCE<6lMD9J%pygcVhsHe*|bG~S`OqJf3zeFm(HLP8>3%M;*zA7z~w~r&-6i-##;a4;| zDHH&I3_6g=&CJ=5(7yl%;wGqc%KY~p>sbTLheXEOgm73q zEeFS>Rdv4#lOmo@2s$a`Zud1Iyj2MaaXkCJWY$SOOWMU7dKjg!?F(!%> zP5qWOugs^P#Tg3=A>QoSd(`eAmE|ZijH>4IM~Ixvj5b&z*^m zoD=kh&Ln`0c%5^c~GUMKLqknG_rzs$Lj)m8SvKv9T`->tH;?Y%s@{eF?-J)<#?!NfJJ zT+BMi&YRHZzVgO$ggFLPC!?74eUpwrzf6GSMrL`s5d8DWUAjS<4qne`<0wgQaE^Nq zr{AgDw`KtBYCT+48Pm;y>4GAM*cMH`2~*pwUr_&wB^e4aPt~!LEOKWw517ajiysk~7<*jv$2(gx zukUEMR8zY-lkh5es@X16hZZ92mACb46CV9A=X*HZ%j}kBPdCNFt_ki$CQ~oVwypS? zuu!+Z4S3Vk)B|%G2aIKhZTAasd^#F%kk{D&d+S5dnW}dB6ENvp%k*U-mxlj8|HMGA z;QyU{Vo3fT+|zdfl67~(FTJt*5H3kDsRL=NIgqbj!qE;@YfId$5WWKt6m_p0o1BGJ z>sEe1l`l7)hFD?Z;tWfnW+y4bdhzzJy?eJT^xf-qD8PXaH_hC1Y+xN6{ML>iOUy~! zD)Z z=Ja;Gwwgb`{l~d$SF2cjma%=(>R!5e5>m)hFzGMC)+U)q>1*1<*Hce(>85pd>p%Q; zHFn*#pTBxWxqcN7*bi=aG}4AhLlZ<^#?~}|d`!9d+17ZCcWWAAr47?ttRhjVq*=^* z*a6@27TeQ#UG3|W9?2>xVOa0O@AN97`yKhhT=BZ82i)MSYXZz5|E|p?f=mj|6^tOa z@kbkZ5tUT13B560%}`qAv^~WQ-0yv^WOdiD zH#qF->$X=WpQ(*s0T@>%6bKh)-U71~DnT&da zqDtZ`%GyIM`!NVPXKYReuuTBZ%{Xfm4#c|@mT5(d{7nbrR1qw09p*IayQldC8N#R? zXaW*_ir+FpaInV;fH7s*K&jpm{lkxR&}f@CsGk%pQG5G^`a)|iAt)e59rSIv!)9Lf z?KepEJ_hgae{~6>$}b{0I1Op&HiCse3RbUZUfs|b-4`Ke^`tzR&$?b17#c+ho3uU( zVnd>!wVyJ0sk{m#)a^50wxya7GX;~6Tq|-sALU5V-BcWRz{Z0f+HpF{cfYBLLY4B& zn1aMJ4-6_`70{rPdCR~#i=KRFFcO{6`Psl3{OQWQ3#Ud4cs+N&ot}o9`g569FTSbQ zUhD0}PEFx(LO7y>C>L9{U3Aiex(+YL{w~!KJqgAHMg2_M@%9!hI=kPvs~y0MlEF}# zT!t{gslF?}Cm}=8hEAwi+zkCq4t?rN&TsXiU0UIta$vXqN`^z7d>@nZzMF%u+kYfG zyPU7>Vy%Mke-izqG5O8^1_zT0hJSK=RQ7cAJu`5>rUvkeE63+UX`%{Q=)$d*u+XYP zmyJ}o%QqgEYhYI!0rz{G@i4fQtni;RqnLUP?O1=E`gbb$|G0fyp92NXO*q?UJ0N#s z+xx`GMJo4x1}_+*>(Nt_jf91}Uk8SotWgBbxYQ7|604Ov0-C`T+Pc2MWiI9m+9?OJ zva-TKkdr8gN5Gtff#^qUs=i4hJDJgH?D7fApZ+iw&mnqyIf^oB9sNEa#kk@ynp^B& z=0to(dYt7$em2eD?`?%`6lUOx5w%8*z1aTDgg)Rh`uU}Y(A3R+&9lAxg*74|U5Nup z_T27GvMb==%L6+=00lv7cx^_dmmAy2#Wjla#YA``^XYM~Z-u5@Z`iqwx&p^}IS zVTE67r`_5?aPI-vo&vz19rzI4+a?)QQaI-gZuJekrz?~gT7f-1Rv=or!~EUE5R)wf zPgfgjJB74dejOTlaH;~67If#N3T7=akwp3{aH~9ir+Zf1k8KP*EIY4oluUcIoO1Y! zrk992ltf-EhRO^n#|fP&GtBRF$IQvY>yFpl?FRt1|sg75d>;K!)Td`HtS zFo0rJ>zVGNm>MlBQWKY&rav4N6{?p6ZwAy5?$-lt)C@FA6gLpD+!Slk!!8NV16kS! zu$e0b;JV=UjN^cmf7%WzxUABfG!WFT>U)_v$mV~ulRCIolza0-7v10a>?Og*rQWVH zrR5qh!O@&((o6c4ti8XFzAS1Hh7pteITr(tC$~0@{rP1ZCqSZD*O-UT3E&tPf3t2j z+3LhkB-aphJ}}f_6Dv|@4k|rtm?Vi@WW}35ds6zM?)i`LbGU<|x+>cxy_V^E{(VJI zpF7nd-GUF-2p+vI6T%HGYM41zm#XSpyely;nuqjKPPO)RoI=POl=<*?9phd<>487S zEyhAogs%+TuRK2SsW^U@=im3`YcC?d^ZUf83!8RCsMIvy6eNZ3S{Z0&3U-#Hs*wnh;-&HMQIQt)?4f=j*HV2@fR;1%K7R3U!c5g*+w_DDx?iSnS3}#<| z4ZGiT@#uU%^GkKFXV)jiQ#J%G)8;vOOSl{le%Re=+KFQ!6(5KWxsQsbS)c%3dbf-$ z=bH%QLg1wb4p4|fc5W*6=}a+wV$bydI~((w4J~jWB46r2g*|H(WW1%w|xmWr$R}O&0{etuK*xS_K?9LK=Wh|}E%e6icS*VML zP$Yax!MXXLSYq05Q}$mMGpQiXdThCBLNEv#en0M33Op=;ThCvN3m8pq6#)%_KaTV2 zdqV0y*;10f4)5?7z^FAN6g2{xqn%l9NX%Joqj|jRsrQK$=+)o()g5V7aTdM%7^~fB zvdfO;uT_6h$d&E;uHx!#c=`o=;x}G`=PFVES7+C~id7>yOQXh4On2s^q9k`5dK4+z zz#faunX``Nr=$nmRHe{9fZWnb?k;5bI+6LVC^L4|E!ZmkfwK=6%SZm@E1sFx0=l!e0-_4YrbhKBB;~1VM7rW%_SX1@;)1fX-l;-x= zg%>GWjLaaXzoP9%nhs+n;~-ld-s)f%nncrsAY4e)=R!BH+H54w8?=o}K%ud&t(9+F zJ!~B%O?o9t&N0FGz6K|f-}&k~s7!GoHbgPGGxh&_0f>qH!(3Q@oCKyMw|VmvvUCGtoFwS71{<#)FPS+t z1VxEtyNx!Ip8{Hv-gk)_Po!sA=^$gU=!F^;vY(=xV?Bd2)H(2Bz0p=G%&it@0qt%e zS);=$WADIA)xh>f6LaK63@U-DOgVbO-|0nXbdfDa)dUyL446Cz>e8$priwT}9VTsc zSGfjW_s;>nai^K#w`QEgeXWH%;qFoXI_JB$ppCz6&7>Z-EW?3A56)bWcB7XfFvpVQ z@FlFijVNTx_70%H%ITynk}UwK@C%CPy3=^8NAvF2L%*Q9{q#N>2JgS@wg1@65@qR4 zaNzX+(kA|g@|Ek=a`7VGE;tbiEc0#e6^JfK6f#wx|9{uP9AHb)u&B zHti?}o1&@16zPjzVL5BH18s=2F{#np2Aj;?Jhk^!9E8&Je1D|JCW;1{i>6F}QQT_| zt+}?f=er4T`9mi5z6>X40UHfy z2PQj*e*91ILYMY`0)|A^VUD#q#ozUHW`X?(kox;q)xS-*Snk|kD+WFO>n3W{ZdUtb zq{$Y3hy@rqJLv#X?<|zw(vjsdY>Jjkw0-5qo3W5LUWs@BtPA3z0&p~!IUy{=xWp#2 zt{3rL+E`Q_n$(&p88~d{bJ~3Bjy;kp?QfvCy=nTL3wWc|7?vt(6>LBWC;VsG_byM3 z+%Jog&g8Ih=~^$ZGzc&sz~LVrq(=MBD}4rjeXGUCs*g7H6gJBUA3;?Z-j&s0=6-K>rwWJZ^5?=F(hGtF`C#<<;0XE@5d5I zodG|HTbFe19p5-fwGojB;az50*{((R_L~jUmv>?LA){(MpI>4P$Bg8!MGVXwG6_X| zDBL)?d~lIS&h&4a5+kvd4U*w4uIO$hg8lsivpJ((%a`o?K0 z8vc`8=y(2cQL*sp>t`QNrCA7%TtxFaXiYX#$$%0GcYy+oI+gaP_(RA@=mEy-I)U25 z_l|nxe?3uea;NbmW~(edec|Y9C?{gTG{jz&Q3*9Li=5JVBhNO$&+(~%d`C{{3uo3c z)xui?;nK*r9bw9bR1B5WQnjS2jTY61vor*S95pFgU&VjB^^A(-@+7;A_PHPfM^@nI zOBq-r%ym-56j9vGEv1pK@QnZ&-!#AG_Fu7CN-KB2@{6FUAE3PhNfWnO@T5Y-l5M@pWosko+ZWoZ19Nms9Y)_tE8(MB7fE|72E$ z%50bPi^5+uGi_L$iw3A8vv|z2?&mZgT?7gc8jCOk!_w2@W~nb28|bDhEns!&9AOe~ zU=A3o_kGK*82npcz2dSmNU}p{+XL!xrLUBOq+njAy`5hH(p3E&j01na#BcOiLFCW{ zy;U>SaHc-`O0U4=S?o2TrPVZT-A}0{i6dVyhbd@{InOn3q&@k^dCN&fGwtOR7C)4c z56en%K&YQ5KL5CCa;Pi_#3d%xfje>9H@$y(a*ryZ17nTgnJ5OgKD6hb9oKyZs=`2>V%PIBp4GNE_L<1c;TOmsKo>@ zdhYgNx+$noV@r5&nI4>apt~_0p%3852=HB<6a*F-Bg?4+z0#tgCXIRQ4FMDIeP(nn_fO?-hkq2OY}4~Ik>%Eqp($;^?u6I{Kl z9DM^#_UB~Mru@Irj_~?J=d^wTWzU5BV8f`NSI_p9NgdZ$BPbxuH%-s08jiA~4;IMc zXmo5Nm!u^jLBaP^tpM?5;& zV~kId^;H&BE<7)54P4< z-H2fdBm62Je|zB`TMwfSJoP#4Onm$4l!qP=Tq=Yyj|iAf3d=xKDj|1PlLyM0?0uDp z^K~C`2F^16(q?)%ZU$-11xy}C00wsRhj0KQy^C|k{n6rPa}+LMO+vOtfp}P~{M8H= z|D+L?QR!G4Hj!hX=xe5B2-af`B#u9_N-0ZeI-u6kOm(&AProoe$^Wr@tSH&DC~y8> z58K_IYCyMwa-=aw&V@C-i0{K!zeMnr%D*BjSBggA_`=TXrg|*4uV-qJYQRCZA17zw z*lXYNvj^nal9#~wOlI@*`=W=8HE^`4j@36v(IL~Ns&UioFXD#{KOga|qzZ2!O(4bH zjvgfcz`()YFw27dvsQn!6tQYZB&TViIxuq!vRn;o zAFE<~e~k4f!F}-Wakk4ozberGz;i2A(P7=QmxBvhrxBqM6|kIvqJs&hE-Oj>)b1em zCH3WfR5F^5VA!|%_*5U+pT6khMk!DX-LO;auv4rqX=R`Zb>tr}MubS5x-Di@vH$>x z-st0Ew#}1cD&dE1xMGs%3?zcJ2zsNanku41b>a>H;bRQ6A?YJsScr|Hx&&CQ-GY>a z-S+=w3UIi=w?BS_p|YW+U=AoYYy_UxWWVIn&#Jl2FxEE_mSq2fwsGv)+KMDONerFPB^61V}A`YlgV zxpqR-p`UhwobtzdR0TVWUK_%_qdzjW&c1C%nSe49h>7CsV>&mju!?fF%EFOIo@2dS zD+)NMRHW7@)O2DuB#keh)(r{oQm<Q{gxL7HkaTynRb?OEUmmhx}TeI0T_6kx#@MpPs1rw28 zY=ey}MVc;9x+FG>7BCgYz^nKm5JRF)Jl__klM^35w}SH_`^hxxf&B13m66>tLR@i_ z@IfTdn9M@r=$7A4+!^|v5Hx-`n#&DtPN5#tN5jS>#$K! zeyOXK8N2&NU-IG_7&W=s1|#g&&(@`$+GXNT`4v%9rxEMtpzGq#bE)zg8!Gr-SaVd7jq&p)(Y9$?U&L8ruJJmF#P-OjF!R4~eipwy$^3jap zluT<&y)tz1`rV2p>3Bhs42B7CqK7_MRrOCB)3#Rrk6D~##FVZ zJ;6l!JUFZZW*d@Zo7D03x^iN!<=n2ilPO4E^hPc$+Wl*p7yzcps9Y=E3!Ia@kSD!hC2U*Oemww8wLIf#N_HR%jzoc5nQ z1@vDl;VH2c=~tlDT5*v5N&NsfIeVvQuq$Y6I@?LbMC&61%a0a`5bw;lEgs~A->UZ( zo|%we2vA0DA1#+rJ z-6CI|Lry^PY==4anM`uu|I%PS8KRaUMD)Ok={wD@Zr6)qugIGgo4A{+8L6$QX!9yc zzxUB&OOS`l<7BxIGt@pB?|1I=%3_TV8CW_G|Ali04sesl#Wz~3l86NP{T#@G88QJ7 zlA9VQR>q{lb;<^DtU8Zy!QQ|I3&4ZWZl3e;P~ir}P5?_~|3!fNTNamb)W(MLN)l^0 zr4M5%+LE6co=%t+o;$sNXm5BuK6!OmP0Ri;abOtuEM>z3MrEm2>uI6y$&1zAxnY~b zZu7}z`j}o!Fwl5xxof*R4hJ&$zf*u{kf0DCWmq`TV?>xs6;L%Fl8dZJRCPvSnPi?v70sUGLV>VQt-G4fR zPMmx%(b{Zv1RxO;e+KrE~H;gFPx`0{Uf@8~~|MF-VDcNRPq3shAvC8dx&w&?3ei zKFZ(|?2`BrXW^`{H|Km4;`}<4OGKK`-`U)7n_A75vioOXi@cf+sd~An!d%YCd%0yS z3?)T_7df+TEu)00V+kJz9;dBMbU^v7n9e|J+Y5^Lj>|?r;kai%Bj(%66yk3yTZU^| zP%E0_cydBulG~1Z-1leVn3iCY43fj{^N)>=_MR8rWGdzU@15cNzpyBz=ZUe9;)u0v zG9f1I-tX>`TE{EEvz#o_vjgBr@9thdUYzVVPin1E0C(upuBxqB*ssKP))VY|b@wZ< z<1=}w5w$|Zn+t-G9**1Hd|4n^L*4z*XAbuFHRd|3xwC z5XJu=#(*@L#V{(IX(L4ceL|mS{sd+l9h+%EjXz}W2(ObTx^vp)>g32vV?{%oxk4@h zd0VRdvPR{flHz8&bGxplyO^=&9=g2Ppk|i)n<58+>URKZ#)yH>prK}~5wWQBudiTx zj{1QdtzB*Z;xUB0jO(yr0)L_7)dYV76=b8Ro;X4dE1?PRVbnt;zBW&hR=j(&Su{{? z%`DRI@ZQKyPT!Kbxk@A-iM8CJxv{4|$}uE0;?IHkFc23%m+>l*QF4v$+1DSAi4|<* zU#vr`pYj^{E{p-#fo)^DSBbFU3ByJ zfzs4e=%!#w1eemQz!$4MQ62Q}&vtTM67tuZspxP)CE^ll)!Uttv?fhkOX(@P8P>B4yvR18|s(R{8X12(KBLY9Oi{>nMALsSL6A8 zm>vxD;l6od!NJ%PUWledjeH`%)P|~S2qS%~fpaamYKTzl>^o@6+vRK5hW_v?W&sWlco3gSOFVi^mgx zgr+45TYVBZ`#COl?ZG8L#qcWV%1*v+Zfmlcd!Tx;+55}wZ>x#rDk;SnHwh#44~P_5 zJQu>EjQ(t&KkHw-0_ngEKZTh53ZPEc(qis!Ed`*KU*LV7oOj_A;q)BuxiguSsV$a< zQW&h^e8XCfs!)!I8BjdraeI%f2zjwzY^Iqbglf4moqcs?nQdrfu9t~ZtpbkyR<8$T z%FF(C`9)g>t}NZlwSogtg5za4tRNQ*X1iV)R)`eSJn!v(>FL{>bSl=CXlEQo)&o)c z0-q}J|M-=*p6B2i(emxZSKv2?tY^+zuXV8)ZcIb`&vZoRH}Cl}r&(KL@0mI89bLoxNLqPz=z~$``kk$JQ&CH&iO8(sQfCDU%kw(EDB%z-R zhY!RNRI{E$YZev&-<(8?HU?GtG@Irt6JeY|qd}C!uE@)cFDQawA467u#(vyR3Mk${yldBT@jp7pB9#+5h@9dkR=F;I-;+=E3gY{8F_lGIzxs zLqjSi^o3LpPZm1t^XnfXM6mS8nE{}H>$JSsJ5?o1s*8rbmFcpVetCPbp<-9M!f25Ukv`a}LA{kx2)5wFG&9SDPptY}@T(xu_jpPy>P1+q>+YN){Vecz2 z48~r%-&v_1I&731Th>aEMae&#cx@#!Wj!A!VwwfXCkrL!Ue_E!)@@KD#wgXZNPFL~ z?+>9UZ6Z2Q;avLA#_t^R^*adTBtR*8$Q}0kAEj*de}M^SB^56_20zks%)ju5Z79H3 z^ZH$GEc%9Ebdr}%dtUk_@PW2#2h3_Dj-=SpwrHwAorp<>nYj>8n~}|+IfsE9tsa~( z5qLSGbc<{|rPQ+vRL_1LHOm&!Ci{p7jouC6XSxL;hrsF zeA$^6wrU8QzHVEO`AOX!Wq1KDE@pG8fZrB4k!^c}=JCwR{(d?8{|Lv08T zOHf*ntw1Xu?4atcFidU4G8xSowktP0)5N$GW+*NJq;~uq*?YXSuLKfMwx$og+2Zo# zOe|BdW=vkMHQ~>f3V&t=(pn#=4JF?fz&1FcXpcHoX3zP(Nx!BfanKvj|4U_45qZF9 z7FMua8C%9HFNSjil{QU8>4;Ji&IY&q#$MqIx81#jMZ)E^#eD0+$BRtwM-!SJG&7(ZrlHVDp2 z!FB9SwQR=Z>aE(~syheYDv<+a?J8>CK8_iW6B+S&s?OOfm8 zA!_^ccsfqzW|E89i8}IPDyhPVj-n+jf~Sy`mWNAx0FA?;qi9@k!<<666e?3jn&5z- zk}ujkQ%mZqJb1+&_pZ4P$#L^rVArmEna10M(3hjEJ+x!zKss%ZZ@^`+N`u6}qlHxz zSV6`^N)bMY^5W^h5y{VjN%M56knzF=w&2W4~wdU2Z9-017 zRa5#xPDPq^pq{xA8R?(2WTPcCU1pan_E?5-51#FgyZY?m_H%Ve?We4~m?B%+Ia;&HyVui!5s~VzRt^ za0xi8ktNk&Mb%cpYq-2NfNBzLoh%JZNj!N5R4q>LC$rcZo~VMK{tmE3rTuwT_%l#s zO`yhMMB(mDjV%K%7{O@)rRHvx@ZIYTUEX-g3JPdIoL3?3>NX$Oqg#aM^C>S@%jof8 z9zU}3J5FiUZzd2Nv5m2vmw&|LeslXu91hpNQ8Bqgx?1oh-fsI(47J>q5c;8zl|(=N zn?WheKk^IGZ?wl%t^Ojxic+HCcFMccXEGT$f_@G^M`Rkh+&)L>YWU*4QKUV4q<~mywhBVuCf_caz+*JN}QU_d6$cgo3`*4ias&?p*7J zjLggSap$qitc*BJ1)M6i4sqV7lo22vX|i^U#_E`$QLFU#f<#kUhWCv*VD0<+zVuCM z#?#$eMB`lS9GrQa*aE$g@nOXs)F%=Lp!xIas3j$l_PGLzjf1#*AxtmMBVP0AQC;55swauOI!18uW{D*DHsSAvB z?1=cNZSRlvSCIn>1U9AagRbhzA%C3|=s5!$U7v70)*ImgLck&uMfcDY{s;N3PwA*g z$fAEiokM|ZKpvKe$&Q`r`y<@k4qKC;2My;MfiZ(%}Zbt+WB8Eeh@Gs zf-RC~4YpU-zaD){Z#~-4?37c`(FlQzn!1zb*HNQ#GE&7BKT{K?VL2>VQ9Swax-?z! zG`j!80c)vnX2Zcg*N%P!o=n$y)9}ov#J6~nR&tQRG(8&pFteq!SA6=!Gy92E2QdRh zZPh0M+Zt}bmOd@1XYhJX-|&~JY>_14dZh_WIA)=s*()+;OiZ^mhoH?d{;`GHU60ji zYV}mK`EKR}*weVE!iMiVE$_5zLolEBl2!{0f;~1DT|w@c$Um|Lwua&z4%OTkum<=n zxS2Slihw5TtS4&!#y3V+9i;fhH|~`1b$1W0(k{JFE+XuI8HF;(I3%h*=VQc4o#&#t zJ|i9{6l*&kD?&K<-5rAMe>#acZs${jZ$qq_FzAne|z>Q=B%%qvrt}QFE-iA zB~O5Q%KF?gps9r(QlcrJ7ope;Tr*J&VFii&^&jWR#vFPIS;XwyIlVlW{fB*1@vYKTxu`zqyqag~hAycE z^2_wx*0tB9=AIkoy~an)EWC97&`tC)U3NUy+PHBAkJ{7x^M13Xv2Tqp4Ro&l3wzFr zXZ`E>Zu)TraPT*1iIU6!03;zLay_=*cSPRCH8eBo{29JDXe2ExgmyI-_WBsF3FRQG z6W5imxYj2>2#v{*v*KJl!~jU)l~9 z?L3+cL%-Mugkv%?f3hO-1?VXq2>;bIS)iVQ*|ycfMnO_k><}7F;@LkEogFOLpO4YU zouBKxj@jCh`py~W*K(q_Sb*^VSf4rJGG4wY-1{Y+gi{`S4zfL?^lk1Ez}YU<+(Kw(sd`PlvrwRCU-}jnF`s5vxnlS0O^JTZ(x(;<_F07CiZ&#`HeoNCE)^YvgRqx%_ zP~enqmf)}qN#q~tuhArNMrSAal(fJO#w?>_+!Xwq-7J+nT*bL;5u8ec30CS@GaqR9 zTzEdOx=bPxz1n#UKwrucpRLa<)wZ;>26=0Gb~J;2@YnnmC}#>vu%nOIbju`)l*j%f z^Z&g7u?ks++hzd9dHIogbrC5yC}O`nzn0K;ZQRtoD?HVbcf?6M#+ooDGi4aax2k;U zWht7y^I_9t6<`hLdk6X1z9Z@S7gA`tpy=8zp3o-#6ms&px&c}D#^mB4k|FxNjelPV z$h{d}Dew+~%S>uLr+F=r-vXLw$c=i>%xwO-GEl*mQ>l_QHCbKV5!L>7>WyI#(igbU z{DCl5!azohP*3j}Oh{PZ>mkFyTSwv000h=^B@jEXw*@Iaer`d? zkywsRzM9EWT>!x*!m9mV0PtbQ4Bp!-P(^x)CmcvrA7?_2HP|=d;i!2$iy>;uR}lSXKlc8&{K}|8(5ZzM^Guuoa$*G>ONokVnDud6D+NOx`e4Vgrm}~x z&R1t#XUzOa6Y==Jm4~My$z*xv>E*%%9hxL_f->=Ma|UJ2M{AuA{+ylh5NbnJsBk1j zXd7K53O%vsBwWfYxNkXIkJtNpTnBp%6$a#F|;7z-!y+nuS-x}R`aOYHQVx2g*>(dj5^c? zw)GZHQ!1>96NZUsdv${~a2&ZN4H~>64js1%b4(i7gtBXC2U(DPK4>Os;IHt2uJiN^ z%f|Phcp3R~UKmYeoqUcRLPXzXc?y$HG6ps8I>^5sabP_JWR?c;?ji?#lZ0)F?yR2@N4i2o3x{qNINgyF zhd}9^IE;=}X9s5?4}++F@m5*(`q9fLnpVbe`pfw2b`1ULL74ph=y3-V`<1n$U6ZZY z<3I1n58`n_qfXsbdQb`Px^}$`r^BsNSHpx$RF>B@UPlIYa0@bYG%Ys)x*0E6V?)W2 zfr0bw=H$L1wVpeLMk$|ev$ehtNX^9feE{NS7lBupwAxF6*QA)HcEhy} zDbY(+2xIgb{5~T9Q(fk#v+cT}nDoz!(xf69!Ewa(eT9;yoic**G4+Uj!aU`??>u@$ zmn$m+;V0>|$A-OvU(c{1t2B1!TW;j!P~$2JLI;tA-?cgl=o1sO%cbBi?FHp^6Sp=} zAR+B~?)CPf1<1jrG&*;Z&+wwT#QQSlR^fpal80{xHNeDmX={O6Z#&wbjX3h$`9{tA zaP*^Fkv%J@NBx1@daqENZtj-RO)$$mDjVOVWld+y+D-lE1WWxZf$r{g-t_p{WPPF@ zzN{PcqEwJ-Xj|hgkcYMJj{NxlF^p-DHzXAL_x`GeQ6pCAaqmZsiJ|11Ww%hWvvu?H zr>cXy%+NPYe<_f?yV=1sNj#1{9Zt(lXY2WF01juTYfuc!>g}qdHlat;=$F88Is0gt z=;G+wQ~jOP5`AeZc1EBs#Yj@1ADzptI&P~Hu%(_@uk3XXmNaeVR2c`WE0!e3c`Jdi zW`zaJred?JRQ-Dn{JeB}_$*KDa=wmo^cX;Y#O;VPk=9%W`50UTIbsha%cWfGSWK8R zw2Sj(?@=B52*Q_hrniywIWyIaSJ9VYf#Y~H~C&EkA=W!{1E!158TcSij8b@lC zfU;9b0P9gpX2as`1>%594LV2fzHHxDqRX?Y2I% z`pnH~1wVW{f8e~?u*>@W;?ZuZ_*{heCF<$c%%$K|CSDmCSZUZP&r(>NnAu&{O8ds)qp~>(7N8=eaakEEs2IQA>v=te_-DS|k!5=kfQz9k{L?Qz+kj zNiTx1)IN+Pwu&#JfhA66(b`#0F zJoDZ4_lF3<)nP*USdk2?$l~VIXDYH$R%0SvWfw*@K ziePJpo1^8_z+uh7nN3{bMCY|eXHXpQ1PWQT_qMS>RinO!A?#)0G}SA7V~E^%0|vCo z@H-oDtarM%x+^HywY^3n)Zggr^3HPBE~?Dj$cn^Xkt(u|6s?RTem0`vP*0a8Y7W;j zDw^@#)N+U1v$P<>1GE9}jb?N{!7@& z&%tPO>vo!BOn1mjxb*(zMk4M%Ns2=(tFsbf-`w?!<%2{3Z>M@_-#ie2RZ5=(PSd`) z3K<><-@cxF)(aGjF1Iqf|5rM)X73V({xb_}rlvJ^`YFh!(CC_^6F{Kb2}`Ta5q~)r zh&jgLGH&?z75JiLRD^_U0HlLGumCQmlYl%ik)mfV(sF7c?Ox2Q<%s9g&S#!tHn{$y zGsC1|J+gYo=3CCU{P(9-L4<5`nX{lhKZz;2atdo3C!XzLqCKtok;VxWVAaOeDoh9N zy&nSyHhS*ZWuBj>W!jGC5S%9VyqC=xm3UV6cX%*`DC&KP0Ar>y6q zOl-YOT?YKIDX>ijZPF|#u#RCAFxWWW{%&}c93JkZd(G+5r+zYgzrE$eYW|P*{@++8 zPbLhaludue6AW^U;xs&;eQp=RwOISRz;DgXH-!MPvQ9wofV*_raAYq486~dqjJQ9+ zoYADZ$U5cu=?&BCj~b!Wq>qu7qpjXpR_>`a2nitJpLiM2$c(eV(f@A?yyY~fxnS6% z=MaIrAN)ZJg~;)3+`L+iw{^h=B5%*6l4Q z8{Ui^93_@7U~zWZh-{Q4hwq1vh>h6jdbl^+bi0d?1iP|XjFv?AJ~@qrKjg{&ZCboM z`t>!NthA0&d|jKS_x1wS5%*5?oji86&8QE)j2xN5yaAAFw9KcgCJ-eGL{pL?TooFz z>;eYn##8;-HNWh1A2yVWXfK_}s>Bu_9Bq=&1b5GDf(EyeKdpJe*8Rko+$X0O0lZVO zwk!qODE+H z1ESLYdWs0OfQv&-NFqV=2&nP2Va7lNO4S5497avRSwB1sQShLXR};6hVTXT!sQfHc zKX+1+9udHfK|DKyoJ+(B?1hJCH_n=e#e=o{-iZ7Rw5*?GVxG@0aR>y33E*RgCp#|? zdgUC6y<84@O>xUimKI;_F_1;R3oLM!Va6`~)+03^*y6O@@;3S6jCJpzqYMs-bMG(b zbf=Ao5{Su;2ew5_H1K0N#0tDM9J#cfxa9Eb`Tdha8ercxVJ*{*$QG5pE|}DNR{K3M z^xk3s0iCM4O4lKhXKOr$GCLR#oGx-;v zm77!DQ85$ipP5olxvpX}w3NBn?cVRPbh&$*IgM#0HKMuI-eojHN~z4YtwD^XLe_r{ zw$hOm&i$D$WaQ!lN}%rN(rhEett z9^w!m*LSp-r752s{H6;z9o{uL=!4k{HsWi{KNMSmZZ(@ zhY!~To|0YSA(k=oTyI<&^im?iI{1XQ4~5TZ3rqz$IY#;fy4@pH8$RgQjy|8V;Jm%4 zsW`zM_DzQ~#E=$gfr;YhHZyQL4_K|=|E>9*!-}h+T$~QpC|JiY>0b!|id>&71NxVF zbt>{iAho%T15WI8aFUD=+}8L&SV;8gzAr)*etT21dk~2{S>MmNTnU_7F1Lgl0{0^` zH>dWn{ml!%uTOt5Mz);XOeG9^C_bJjrSDqlzr4+S6X@9AUGkyzntxs5-<92^uRHV# zV%#iX;AUHA7OYj2&`oroo@(0JWD+n2Nw5o?XqoxZ!JrE@RQQZh;j~6sHCdESuVpcc z|4;x@*E)GeoY_g%WqGcZSbqIjKCPl;LPS)^%|X__1MfBmulV;c^9yyGw;%p&$pokW z!S8BTM4$A7d`^(!4NBCAtmNHYhAI)WozSU96I5`v3U+oNtQP85yaxLTqf+g0;jckS zFCWL%$>QgdFI5$2Ur2_u>{{8j)qvuZ6f@A+| zCi*d!*uvi3C;ShFv%s)&fTcUN%A7+@mLI{#CeDUz`vQyn-R=jXIL&V!OZC-wC8HN( zN@Fgz{vw6viy@PJoZW0#$4y&lwj8v^O(gdH<@P_F0=eOc&sd*^(fdWeucXMuvT7V9 zyf|~KU*G1={3Y>=DTXN`9mF&ItI&3SY2)t1f*er0hAt6 zkVYCD5RitUl^D8HB&3xP>FyXph7ypDp}RwBFZcfclYO=>PMBEptmnNmZPeRjg1r|n zl4CRf@Fz8G0k#bSHt*hUKg&e)Bv}qytZvg;*7Rg2?AfqYpl`>@px#KmLj)^?mcVE9 zjKpyluEp==p)XA1t(#4^Mj;0&dwK%?DUYw15g;wu;UxKDIL--gQoEn)#L)zyH+yCw zOJ@b11|$O|yVAoG$Ls#lS6KFLDP`4J)70|U=5`$@C~$U{Dq#a;YZ`%3LL zSzwnzm~a~Q<_&XE;f@G|c_1GsWF0|8C)pm@#d2y{hmsQ02LE=O>g3$w3uo*3Axce- zT4p(Zjx+3}wp`x}QLT_}34~S3sDLW@%-<3SiX+kqmZY-kJm!`ZB>l~&JrS>p4asGZ zIbkmcU+p(^%;&wW&r}|r$@;)ok*W?P)7bZdLu7p%)X7!EY)7feGKZ_JOzq>Kf|I&# zdY7_>D{XUi>N+O{JKsR_wRZ;v6!9-alUY&Fr?U+OoW(uP_a*5Z9pIhqUrCR5$Pmbt}$15KBHRA*E1*A`Ts?xMN=TnW`$qJl~E3@scL>P>j&Ou@N$8jgq# zZc8>DURuA1`%^vXXcC>!6LO7LswAvQE-S(4Dl&?F40@k{d#s@Si!1t%K0|hWW$Qaq zf)5-AwEFW&vyGin33<%X7;}5*4PWj$Usf+n+bz068Qnx~13!=2(~o_A1q)2>{Gt@q z0fSNMfWBqYWWn|VLz$PmQGT(>4<&-`+L_ScXqww z9?R3Kv#p$#H-qLf^k>s%O4k@)ibll_>N;~sFwu;-X6rULKMCWL=LIj>1+Pq9y|EG0 znAEAy1W}*9js~&fvfDar$0)BootY!b=+R&WdMIPeuKc~KE%kd-Ds1&G4k|w;QoA<5<7c;Kna0@?j5pO%&=*dV{T<^EO!W1OB!qvA$^wrBlA`?O|jL?y| zNxZiD?69Nnp2LHsRyw`vQ7Ndth&BjZm(B|Rh@KsoTU;sNHeGskQl%L>qG;{jABWec z)y1`fikU*(N$Cp%!CX3+Z7f>(mb}fv)YgVayAJ8|uPxFPa4F#IiECESmoLlw=8E$_ zfZOoKUNCFE>G$ac8|-hO6(~$xX2!%Vfspl?07wBD?NL;iH1Tz)TKbyO=>_D;Hg#}a z=gKa$EQ!1p(X^_dUl{tnWmREc0S!t@Vi(ka0(q<8^TMi5P&}d_)|Q{DH7MhF{2sts zqFn!aYOi*&TPCS8ntIWX2o+NIr!Us2(r|t1bgy&T=2f8(U7>5+2-Yte9^MQcZS=HxD9SIai9ppv8XrQ)b=G-RBn z-akAfy9fH|Og4XH{Nx#BYQ1yI2o-73FNp2Q-=2eOQzk>X!_P>jX+36snpR3oGP2N9 zut4o5UMF$&NO_aVG=DjPZE#ujAxbFb8Fv{|bVKvxo`1T?=?AGER4pZmFck2z5Mp?x z=Is#ZV546w27mFEKPhnMr$;Ow{awiI?Qu0GMC^Q_<0{!pD|`PZg#`bEzwMJf$=DTb z7&HZc%OoJX&DdLd&i4h+D0yShYho_QTXn33cVf_Apjxcx zo#YTGi{J?i@(l}Wf7A$plL!pHp_~lX$~l60=~?@~gt9~T2%W-X*^20GO;NS1?PliF zwuoX^b!CZp(=gHmCA7LraC^DYp>JB9?L25mX_DG*aCV9!GARZl@;L6ybGTVV;nUag z9%(zlsBiOmHV21>uFyo5!gT|!daz`fU3UHsgAP8OsIBRl1>?je95J7BHX+1dOTUg2 z%!1BZDN^CyYuRbX{ypa1jMrN|x_3TziyW4W$TW_!(vk+jqoa+`IpIOyC21spv$mT% zNJIBB<45U)Y+rW*%2q{n^tt0Dy*{sr(>9yoE#R$r({SM+Es;hpt6{xg72yd;t zH*hsnx#Yw#B6k7PDfR4`jb(mh$BC2|1!ekPRU{?08O4cFmgbaQq-jd^yQ-688afM1 zTUP-4o5rkg)D!O#G=OLA@Lo+wLH9FFI+f~(W?k#>Z$z9A!b=PosB8xjTC^|384OZa zTYqD!3XaNK(O9StlaS?(qo?8LeZ@++TIaDOGo}uPsrSt2 z4t|qz=7nfAyjEFH0F#$kOBDOhyQ9ZDeC5s`A*Nd{*yRjD)?qbV%dQ}uyUHM;Eh-|Y z_~C{&pD#iU4!^JNaJ|Z zKzp6?Q?#|BhI~d&6nR4Q8mw_85_zYXGaoc3r6?z3Wc~b*Jw50w7BZ?J>Z`2uI6psY z%pBHPm*yQ3D=ujs9mPreYQqx?SwvG8ML^Y$&q#NfW&?|IfZOvXvN}XQ{lprNaX3PX ztfwg1%j8N4b&GfwA+JcPnu+{rPw$hNz`?Ra)a}|dPx3UW?OtLjE)C1_an_hzq5O8(_CfV=4zCD=J0Pwc2pk@Lmi-z;| z0rM^l+|ryE!Exr=@AjtZJ-$YzQ945%Ttowh&9$ajDKB3^GxagVDYz0!dB z@KDy=)PMCKR9u)yl>u5 z>P#^x3Tz&F5KByg)~x2MN~KATcswCbc3|4y)W zDrX^-wkNO7psJmlNF~UXF*zNde_BcLJ8eEQ+DZNheQEg%F<0MI{n`UBtmu~}cUQKe zW+L3*r}@>DXdPUewpGaK#MEx@de>?=#Au);yszVXBftX~6cz6fAc?lF*rPKxR*7B+ zo5Zj_`@S|?-`2F;9#@G})2xS!K#bJkXFn$?BcWMJl?y###0&HMFkiHI0|xK@sD^iF z=2=cYkH~ZH=C5_0V97-nXg+L9;8EK{_n)j|a7cg&>P+r!4$y8$Flsz`{}L8=u828+{-jy^$&)OI3_iBMBXFeS>!daNIit(Y@8 z!?u2|P!W~4UgwUhR5_NY)#!uEw-{>1EpbEbQfwrS+=fkD{mLhd2`-1!7+7 zBo@)MIE}iQxF1>go(0&Wu-A@ch+40J@2H>wjfpWH<}q8fIHjPik{HP6Ut+JWE9RJ) zDS!{Zrd@sAJ>i}C5)$Ky4Tew)Xv&|0R?Is0gqNDLE*qVInn2YmzJ3r!>gQS26#*s~%GW!1Jas(*4R zYVYB;i_D?SYSu{tU#H+^bWtK!Gv_L+M(Zq}LQAOVmgb3Wb{0=~@`s;#E4|&9l#!_z zoCxJjbNrH{Eot-4WJ{26Ej>(aUNGv34rCXZ1ZTk(v{7b@%#YEV%*(_oM=gOZXxVwA zX10*UuAM=?j{vB*9x6wp`J%!9&ap?2=bP_0->~*_`#Dn769`zH%&^4eWhuGmwlGY$FOB{NRSxV@{Sy$jdZih6=S*B@ywz}9IZPv5AOT5pN zI<$=aHJ2#YX@%E7vm!QR-iy6p`x(#e7wWySHV*w@A2IR|6=gXcw~l?X{#9!p(Aio> zr~hyu%%9nlyf4XQKtrIMzJ=T;{{F^^3SAny@{4q|J+a$1iwV{|VKy~oIAOL)edP8E zB?R`?ppJzxvZQA?$dGzkCkAb1ZJjYcn1m{b0Ao!>rw$&U`(Y=-=W%I$%9{Ggujqub zX6rZJ>SvWRw(7NdX=sO~$qA@w)YvVVdJbS5_!-%kTD=3JEXGvnoPjMm4Qd*4-X9=}H zoXEz}>IX%3_y;P@c!On@-GUyc9JJ_$-y77F^M)tmb8u{xW%g99RkZEbCCZ|B*Os~> z5KzJmKTUqMmZ6!DJIS6cs0kq+fWz-}e8%u>`Ultz-{{vkF)UteI(+*YK{2U& ze7G9XaA{7)aLk}y6Vc;3aHcWb?0W(B?E*w1Y0_<}^kzYnN7BPiJ?tdB?FC;C4$Y92 zx##DUaV3IrIa;}uO*I3bH!AsU7Imey_`T`l;Biq`@sIaX&wWua->IJmE70)eM?+ke z#ul#YHIF`&$)~G-BmzNBJiM^;wc>>@Ud79w zvfSV1a#(A5qarqbK04_1yBWW62yY71QkA2{jWTotMM`8ARrcZo z=bE<9Hg_0e&xfXu=eHll5L(TL=RaNuT*Fxrv3oibd7bcoB1FacWo{^0~d6Jbq?9)zqckRxUa1q&%ZrSgKqaF$h!O z=S9bs4^w7FD`JjzKV^SsT7P%y82xxl)_Q!WK0Hijc-zRjgi;!rgb8)B(}`zF;w#+Q_C|tpk>T<+%_2$VGv^r zJZtSHKIL@Qdc`~YTS}8ZePJ{XhZ%a@MLilfME+Ia z0oc^Z!OONsBwl4L)D{JENF!0fKvtKs;A2#SS@GqMgUb%$m>H`Mp6gfyk&~!&X~hX2 z9@?JwyVN;Ji1Go5bLvPeF6ZG_4UpOyiI9`p|#5ea4vUs-YSo}VtU#lrD%hlB5r zdcLP@ue(q7t$^4?7Jzp8)g_*9a%y<8WZ7lw(^nA9_hmxoOGL3iO2sSHL3SoL-{j88 zTE43DUnShEG4{Sa?`r^6;@I@3C}P%pL^Ag<2#@4>eoaz*Zh2EjdWYCMi>$DH4Pizk z+;N(h@K{x+o&kyP%8;F0Xb+DfEoY+dMoP3^sZdPR!e5@l5|8gn*>UI-f5&X%DA(*4 z_r9-8Xbng>jt&>DHgwgy*!M-iQrzfGkoEB9oq_hx+0TD0HY`mzXb@x8&RPD=gzRS& z3fkZO6-B+DAyw^pMN%yu2lFsK4SLuO4hnqhY9^o7tIo*{Azxug$+nzy2=pwzxpV6a z@-_T9fU&fQcYSB2wpc{3KFN~P?eqJ>lbw1^NX&Kwl=us`Z`B`DPJmBGQ=g-y`O{10 z)Vw`{X*)T~6UAWSr1bb0V23DUxPp?v`X)6HplGu_-_{nDSur?gS=#HtFDKsL+1u)= zQLhuxS1bEWN9)V0iV-dU?TI=a{WQ~|3v;4x@f{A_w4eK^u?NUF+Cs>k_~!d9S}hWT z0yObhemu&Crvw8x|DfJrV|d5iN&+)+*4xxO^9~6PeEIiuE}-GTwXST&@Zk1+BPS6g zH6a3RvUTNTBTmEqj~IN=E~6IqS(`RrAyO>CEJ*xQ5*VC$dhQ=P#izxOnG7ygkqdrV z{oQ!F4o$W(Z12YwtQ__;z3K{{HRF@$(qs=QvLMI%u?t^zG`W1nUEuKt?1G#;Pt%*y zo(A4>)D@9%;Dg0!cI9or38Rd}?8z4>ht>>Mh>bD+p+7ffTOF^%NU|P6v1!voI>j%5 zX=^W9V>dc`^K6(UMEVB)wf%)ueXOnIka@<6Q7QvEA|KnxU)evI7t^-!& zq-9aiFaNUE*`p)bL=~iKb2gg0chefSfa7JMzemt%HtXNV+_#SNEbZ*kK|jf)VlfJF z4fWuq!VO0MkD!olC|mfaZeya{lGaKmFwaVVtCv=KWw!^f_vD~13HCdSxr&Qo>Wi4; zTOI)xS(gPyXP@>8VD(ZzjbO+9HW$}?_$@Psi!t^g;gEm#Dpu=umx}^|gfj1i*Obl3e4e^9Hsp~MABsKd^#-RV4P zaL4rM-!%?eMS?e=S4z5z@X^uww@#vMVLNJsW7m?EHTLvKpXMKmFbzhW;O)I|cW!na zEW3Hjm=KtYZmiWSN{EbK`#YoeFVuQ&kjyrrkEMCSG@o zx9|Burryt%pN<*1QExRbsX$!SkrrQ=TW6D`k%jww zRA1tZ{YsIO_9LgQTZ7e>W(w-3DX%?J=BnrFew%8&(JM@}Ny058_*{a6NY2Pz&UF#F zH^vtBa*+t~skhZL*(Ge&e={XLYCy-BTLFZnAr-ke{^(JpUXjlrxKv3~bVWQRK$WSi zaco&YsDXuGl1<*mYBq2$lFT=6%ofYdlWs|SQ&l$m^mztWRK>C5MxY{ERX~f9=(dps zDXzrHwjtIdH>-cYOU3q*VdV%!ciK;ku}ox{Hpv?*lj5o`3q`SC#mg-%jRHp}rlESU zBb^!)Q!x#oi!_}L^(i7=!KV5NH@w=q48y(QrJ{xQ3bXat%dXBzHL`wf@;)X6UvRpq z1AVh5-y@d&y#|*9w5=4n!-mX)BQRw9m4e}( zMB7{vR!+|#bKreX37`G@j(e6*ytptwS;U{DNezuM>tdtc=4ud)(4fed;*ul$XH|~g zq5;b+sjN6w-CV43(sEA>ZFe-{Cv%WENbk&HjUFqZb6jvo=^xKD$cYnGehPZfL2;~{ zVvbTdZBO416x9*qYjJqZRHhyUSs%_Lk)QLWje(c>PJmyoPop(5V2fs- z@Bm~J^x(q8p$Z2V4s5E6-VnKR>Bx8{UDJkPR$AWQYn!oy04+*g~fQ~CHFxy*?A;A2nh&(%ki(g z$qwf;5KuUppwL1#}!S*{{F(E)%zDywbLr>F`kD9+1`oTm&YFl=3@XujOyRig3C= zzM>~pWb8N6&t7$zb*WuAcGN#R^J61?ww!(IHPQim0z$<;15gB=W*S48(zqA{@a4~< zC;80>ERmhSpQli$pr>6uX{ zb5eO}T=weOxK&rnEyih~<b zCe0V0L1!lMYi-q>zQV*&w>NAv1d03|G_tD}rc=G`7L&g8mAp&~sy|e({YQI(Msko# z>;-nTG7@FiQ#}%oXWJjYOr|u~bi$WqjkLvSo!nL?#yOdzrH6`9j{1Q9(|)%=L9U3G ze*%KG`Xlg-cP7xvs(E`_65TyU`)PvF`%39cf1^2(Wy15SC1Y6$v@yNxu~?8|qehr3t$JjW#w z`F7t+;V+*b9x38m4N`3pbol5n--X`#S+5;W{a9u{Z)!?)Z{>|-jSm8sLof#Uj|iHy zxpVm5$5BR=)aR93QWx$QlS<$idM4c;pnW?3Ce?FOys_;#GN~wU?+?}*;|d`T?!0ng z?Zl1uG|5X@J@y!6J)NcVrN@_N+>g(o;ESGPrWB3W@ zF3B!q6ZJH3`rDw7%6K6zJa~>)rO9O;}3~na;s{z2{xr+Oq6gzdP!geJH+}JZ9}j zT_qLsI9ps9Mu9vzn-Kr+hwNr2eq)iCG`L|sCzH@vmQXJ5O%+$J1Q zkbAb&^`A?q+2E@G$WFI)mLhxNXz_b{$ziL>*J4rNV5!NE-d@6MM*>WkIh~C%W>5Yu z1$P8`s}|r#%Z+n|xW~0+a~m;t);{)n@8>c(@ekwl&2K$oz;w)%YWzDQzB~MA+RT*Y z{PASz-_=v)5{8by9{tcQ24JYU2yjH2c}}LdH-EcE*(*ZR?xrxj^`_lNa-a3EuT(Os z@6@8K7SD_L=Z+s`Qs`e094*t61uAz992~rQeEcmzq{Xk2gq?m7^z6yDT35|C2Y4bP z_WP*RQ?e-4$yx>?Le|40TYn$z)8ke&{xzt#bvElr3D`_ndr=oLax%4e|0=t9RI4kWb1&C)3F^7NJNZ+oar7L}K! z6;EdUcwf6KAZg*SQlJt2%ED&A_=cBmxV83|xah&ZFk)5eP5~`La*o91g#ND7gcX-8 zezm!JmUFN!?KsURKMj0@IC1de$ww%ZAvgCoJ-^8ZXO7mE++;W-ldQ?fM!A)k`GkmI z4l(F&tG+$?cjuEr#HEMA;(PV)-!kL&g(?krLV~_4=koD9i+s5kSAbj1&T`^E?K!+8 zVBts18p@2%4ix*U2?IWzXd?s)#bKG?;wP@yLQG&&rA3bkeuULk@ro1lQx8>Jh`Zmx zmXCzl!&sux;|v6;Qb)P8j^E)fa*M@sj83HYn(omnE^2cR3+&A_dViLRNwBdWDt&&_ zvl{HzB2R1k-DITMQ#ZEAo}i&{<8bMkqv38^pxBXoxO&w?+p_%GDaVx;>j!+&hfB_R zf0+AA+V#d>dE8<`HX1hkTu9hrLuVP+ zAA3Lie)6@WnB6J~f}}2ZKsbgh-g1EohScXf|dH4 zHJ9*wU7sm%{D%f!K4;cQS4h&%`8_AAs@!Cn;W|)-Jk+k}q-y<>qOo!XyO7FkZ7~0+ z{_l!+8GRz9-uL}EY$*)la5UK_nh?*sOP3HPg_yH_pjy!VZ^ow>0O#$`8|{gXYtsfg zy&C$$#lzn^8nfEppqkb~zF<^?EVQ4?NLLs+N%}9N{HNSt+UIMzcmq#yAf%Hb9JW(= zb8)eI-y9-4^Q4E1U*luM&nG(_9>ixj`9g2X$H?ng)!+Z;kLJou-OuQU%4l=#Wyc%* z@HC6pK}LG+C$C825ek*461FcgY#Y55t^kpns{f{GO8Mf__Tz+65|;uNPBvC-`1xP+ zl3{|nYA!Y-a)!(5!q-DGfBR^a*KZ9Fp4)!ZJm2P)e9B{rwlQVf#5cPFHA@JF&dG4Q zY=4|15Oi?{qSF#FRc$bCOh7*3!Dm|rDfp;qq?yKnQgHE>dgnlVa$LCLSOD|Q8$d1L zEYgNo!Ff+#8&9eZhd;%Ekv#7795lROJClyA=XYD-G#X&P4Wx^6j%xcQ_BfJEdf<Rqw<-EGzv<{1`LmKc|k|mku#q zcrZ5c3(=Xk!2FV$K|cbn&O*8+5VHPFD`(M!@}2P?@Xe%_HV51LBw_<(WX;L`(!n9Q zMm{6CVtM?Hf&nZssw^wx@Di#gS63a<20uo7=?dyB>wJ_-2<&n0q`;mSKqdEZh-0N? zgEzOX`ki@8KhTck3x&Qqj>VV#7~91*Kj$@3dX&^M+lJ>QC=4>hZDQ^h&(2HkDBo^GX_Z5on zqi23!Gz(&EP3dH$KipjJMA0QLH9f;zSkxGhF{%EB$NJr6g6%XwDRMO1Uh_MBEGh4B z9Q!FmK0aSCMMHT7r(7^YZPhbqyFhnX3GY;7rv_78wBebIx3@W{7Q=kaJc{5pIYu_n zr@#8QW4v{sKeTtJ`kk2rFDQE*@$srrNDLN`Q62cR zw;@e}$|wFl@lYm~#;7Tv>e9XagWVTO^*;VX0^rI0tdb&{*7o{V8Ba#aU}1qyFvc(M zeouwvy049!D4quhgr7?Aq9_0>n%IT0j1jP&K$6gmD6f2mat^FT75K3R4Zx2gghmu- z(L+B=o6D>nv}-*|QInHN1*sMeJREA)q2%Rs8k};4Gl;O-xqNkHT9XMb1?+*fkNGXazoO#e<9H<5K0A{Sfx1C*Sc0j+ z%-FAR3+=Frx>mW$(zIEx#~(g|h%qO$d3t)1gD30O91FCkNv|1Et^5}g{o3`JH?7uF z)WGi5Wv01}cj#EcWo=k@U%Z$+8GN9v`JMNNn8&luE4dgFF6d9=)x5}1wsSva*BLK* zJ&rsN>{a)6AWRCc7XMW-xvG_`%3mwtNm0yNYAUmLBONdzzpa7-y|ign|K&Y4KlI#6 zCC+&NK`B3)93ZZx;g=Yns!__J`E7A+`7ogaK;3w`#fP{KWFL@oW6HLR+JMSH%O2!T z2uIT~UEI-&aRL*g&e0E;^qky06s@N)y?#M{cC zl;fo_oY9%tD1x91i_?Bt=lUIxsVHw$oRWQELT?@*698qf$1LT8_cD)0Ll%Lm%3-EB z+MKM4rkdM>+<-}Y`ryZ%5-Bc+KeYjVemNI%ZvY)P;JYFR2z;TK6~t4)ARp&m5h5SS zG7vw0hzM)@FRS~N=@c%&jQ8PlrC659D{2@lCtNVCeo~OXTtU{G8uXWUPEH$smX{Pf6>wd|4(^Aw-P=<)o8)_OwXJbTrDc_?y1#qB~8N zvIYp9f&w@vl+t*f)KQ1rt? zKWA(2W=7L=hpcMl*)5y9@wGU&(rGi--{w?%Ey0KSn28Z z*6zn(EdZJBZ1E-Tic@nk?!QGBa;_JaN2q)G+q-SGUy_~|OxTmJLAsj5oOZND`Sj}y zr&eF|Sr^@ouXR!>Br+!@m_i{h zJiM@lccun8QHx6(`_=o*SBE+5<#)3bu!-J2k+hI(zp?9v$DQP;;yL(41@&K%THxb# z%51UKq-r<2JR3_48lKJ=no~tnoit`a=ktky5XnOk$^vurC}OUPi?&;sX_a+5TZ2y9I)seGe@xjLy3NL#{b+Sr0TW9s};P$y4t$c zlanXOU=A(df8z{=W>vhOQ;hCAS+cFKt+S}GUGyR^<(FDqv!qBk$kUT$Ahrq z)(6*hCk+{#Hda&E&Vkc$pK0g~x1Ak-AN^I9vBc-PZQ+a7JC59(DUoOEOPao(d_fG5 zZFg%Pl%4(>9 zHa;n;wZ(Tf^z#GnD5?XU2rJPB`q~Z3vx;a?3Vp~8FIXE*?M%Nr?ZR~00)vmPuXln& z>;eASjFch8^=&!NyRyuAvfQi7l7*aNAg{!w)zN4=O-UlNzHd8Dgt+ze+1la%Y5^J! zT>Ye0m2WX7hQGxF4i8hQdQSd)Vsx$Jf)Em^p13o;k>bNX-Dvs0#I}qDyf%K+XX}n5 z`TI+|qVMtgUG)UcpZ6OL_&t?qH*nMTpwICxbbUz=6H%?RMi!R%1KD?BU-?RsC(nAl~4&G0G>{mPJtW z>+QBG`_J#U`!n|yXi^0+;GK-h_l0_f6%nDRO)h}`1?m$4wX!plsfA;1Adju=!>y-x zvRGQOVqdimc?BD`e?Av8$!k$QJ48+0>NOM8kn{7VV34cxua0z*4&U1ThzJp!lTceT zG2P<41CgimY)VF*#lQe89`w&KBL`L5O1Uo3-%tbEY^%Oiohs=eYQ)df@P#GS%FrKM z_(>mM?YnDx0fF86JoS`UGAePi4guZ9PId8)v4~7EIM3_e7|)+*6*KW@6^F_7vpd}r83?_O zIA8JZwt6mxws%y%U1z8H(#Yp`f_WaW2>`7Ao7N{|&n!_UBHkd($Y@_${LbvXLPBzw z;KRndIKi~ygE3>WTgoyD)B==0zOf^D%>^t&IDj$I5d`gQxs_-sxZd>#A~W(uRkT>! z8psU5vr=Le?!R|^u$;8$VRv_j8JhgpcaDQVTY9m6K%momkcl@nNkRDm>Row0SF{Xc zx5QjMTQqwQs2@$y!B9U+A4f*wMS@PXDDBTbb~Xvjp&Ns(MRpuMvqBLVUL_LZ#Wb8i zFHb6K?DNA9(2o{+y1!HrHVHq4hd?G%)I1ULI1GkQE2^)j3u=;LgN-5nBCp5=X^Las zjn^LU%#1SXef6uUy=@BW`NjlT$fYb@c5QCD&dXmT9N5>c1i{{e5 zFcgo{ze1FAK1p0g&^g6h0iVBc1P5~D^Gw=!CmZ|9;qR@I2Dg5r>sSp6Yle6&{j49x zIeZk_87^DZ>%V}N;I2W-M01NnIXaVH^ZBy^a%+Y&4W`Gh- z?L=uX-|ohlDHkXP%tBC|QH~}X?h)r(r-w#7L88SXRc*maS-K722Mc}-#0jRvufrQZy=>{?zJLyrNOtRy+Lzg$}~-70!#&NE)q7JZZdn~Uu;A&Tc@f?93ej~Yx* z$|Ik0=3*ur739G44nTcmob2Y>zrA-5v3Uc$bOyQdrY5O(P-1Q6q%+T+*t`|~OSgyn z`>Zb}0rvBio{A+HfxggAc07LJw;ea_xGaBi3jv97L5*nkMb-3AUJB1`R|lxdHLgLq zSem9WEFD69)C9l@K_|zyxv0!QL~WZ`oadBf1zRi_tBRd~ zy)h7#|DGbY-jPTc&;~iYnymdL7A5ilMBT?51=NL2a49d+h+LR=%(>SGAU0;qkeT8TS zd@TKLd-@~ND?+1Z)?c1lQCAN*0=Euu;>#6^co-M(a@lg%$%)~b`_`H=6|7`^odnQ} zFQrynubBAGpa1n__YrZpS^u7|8z&O=hZc_M3a79rZ}N9aSVv2wVks)xBQH72Gzp3M zGTHvkjr#VNK3#E0!u&ix^tbm=K7q>WYA@Zpw?R*Qb|OTl1KHSaElAIzZ`r_2YpZ?m z`1t3NWIv_xXMbbuL3W#qKS<6M_d(*BPX(Y2SS$d5qH93Vb=QXkxJb zXbgKzx4fhZ%hQR=4=LTVn*u_`F)sUQw2x&F>9Ua8c;!L}^89j*B`TNxfu- zEPl!*bsol(TwzB~)s-2$J!(xg^k_@b#VVfctm3Ir|FuIGtCQQ~(DAil!B0v!VeUbz z@f^vY;IdZfS(>uTF4&P2(?hKCT2305lPOq^rEA$E%5}2J>5$?KQa(Lx_bf+;wqUR` zgq0|W)D_BUL6%r)`OwX@akw)Fv-AM3Z&aMSjOMk%0ITid;wOCqD)c>Uoy@B~^0XS!V~sl@WXscsIw$|3=BcUkr4W z#_T`UrR@l?4kOCNM5k%cE8~NVQu+3Z3-saML~fgV@ZN@noM#J-FCgvzSY+n6c&-1S zLn1)R@Ovmx@xdfEN`^BUgK`gEdixHJatBb3nRrTn|59oq(!=0izmfb$?q08NNKHJ%Uc*-|Xd(LPax9FY?*AaZq?cCCIPVcr0cuiHibs0NNx4&b1@ z3S>Ro7y|}kx6P1f?w(DM^pQ8|7j0}Rx_NsR%Der?X9GZiT(896u~h7I5?Y`arLlK* zIKpB?@AfaV)ypOkA$7|=-PLnOS#GUQ4c5}(QI@LwmdX+-UtVRn=;DAGblW8cfX5Hu zF?;z%n*85xL02u#+ulloz*juz)3ut<62(T#Hj6x)iE4??w@vcfCw%7NG<;IcmXh`% zX{^+2rL8TsI!NcV1qH8o)eGX2s5IUcd3F<3rZKdn;lyv6JY2eC`aPWt&xoEmtE*u2 z>47?bu|nRcK=dDS3NvQz-KcP70zjgB@b_x?3}7;=MO6oZm6M>!P|rVb={P8$`s)s_ zY@2lN6}uLcqC@KK*&e@I4zH*Y?S*Fhm@;m-yNmwwO}*jL>4kYsqZuHaW+CI!);jw; zOma`y35D4$@Gc8MZrY)gg|WpOA@fv}kUgZ6^LVv^XcW7fm?dBW#IC_wTP5PDK@R!g zRFZ`h2|I;J9)XO0r^Mz9#(^jddA_H1eai&zF=TwgEb>^!bY@9Q6FEQ0;)wDGVtjg$ zT$_Nv`-Mr|=9mY3n_`BHWd*4}RLSIgy)P#?;NIU<&{Zw0q1`;{p)OsI{%r!I4RW^tKX7(b1Ngq%6~)b+*2cUvBk= zt6a3uH5B7nqVgBG9Va$M*QEeURBPD*JvZjG5x{n)CKW>DGXIjNvJzNiDLE2YyFAea z2=YSwj8q1lw1qhuAtW26=WzPglU&*Cn(wsWWt}=pCXKbH>=V$etNg-*Bu$zZ`cH-+ zfR~)vvi_Tk@v5Ips+Y)6u#3vPrGI$J7jbJif5>rPX6Wreue_Mh`M8=F#tFodf_yx%B+A41M_KWSkcpv+x32Wn1=~ zE%$`3B~W{Fpjo*IV@_I@Bd?zc*&5dd@W|Cd;wgoonPnVXNaekWvelmWOli$4UaySJ z@_Noia@oa?*2T}^7u8v&$lDs}m&(7$z*6pQ0vNBO`INvUvf&@!SVAbJ8iGKho7BMY00+Hf^qkdzjN4yl>h}8 z9e{+woAn-fm-hnQ{CbYp?+>)9X9DBf3}@~($77&)BjF8LcT-X(%Pz!E+ijpJ z0t~%?vu5*3OD#tKZeRaz#haIpNq4I*Kwmc@e?((TGF#}1)#9nmw2OA9pLcymh3ckf z$PzV{QFrmK;cqxdXW`F~5jkM1229MFKF+ec1vQ5qKIvWmAD!~*^AhAFpriF~&Py6^ z&&gxYtdM1W9mcEd?2i+6!A` zw9)Mz>+Ro7ES7=9;zRz%`#oZ7bAPe+QI61pTrT30m(pJm50!WUG%>nuom%Z?s54ZGE@^CoNPO6zCd%ucbn?$BPB(BKtK+=3k` z{rX2l^xdBr^#14iS^nmo!%OV0`+FAiEZ~UUXi>lc$a-Kt8V=Tgb!(FL6no1LG`HaS6FUM(QB%EdkrK6DAQ@BO=KBSk5-k;)+Jh_~DY1HQXNtzG7a%J z7rJCi5z2Z?{E6G#}Vwe*x|%V8c>#u<)?qO zF0A@DsnwEJN0*@e%t$Qw-;c_9>9+VMM%#bk!sR@}Hsls>XVMRGT=n|TJ%^M~Sr{cv zhQNMsBQf7&wq7Z9JG5s-d81zQQDgFTooVj(Q+SDp6W_FCOic{wAXh*vn*`C#@pJ+| z6)u~<=BDVh=e`q%!Y3`-+tBvWnrH&5qP#=nZh-*EvAyv~sViV@VrTzS%0UoN9AdVR zssk+5oQC?mWpUM>y4fff{jpj;Bhr&&yZ$Y$zit8r3>9b%_8Ek%9wh?7WiyYhp#s?c zDJ11XSU4iaNlyfmS3kbSbDM`oh*i8)@%wyVE-=kTC|}Lqgpap#QV2vQU@RwvG3w2q zAULT)1ACX)KMMFd>As}LTKGfbTF(m;az;ZefmUmvJer$)M5uOZp<{N}6+u;u@*1Gg z_gxXt&eb1j9YpTtz|&+b{C{{l%cv;dsNE}~gmg%E3rNENLwAP~0}|4m(j_U44CBxt zDc#*tA}|OeIdmi4Aa$Pqd(L|nANayrESTrn_rCXa?cep^t6q&Oz8)`l3Lr}pCc$7+ ze$Ld6fB<|3dhBOSta7hLaL;rJ|FNi0d(?;KPfs%i+S|L+2Xe#HBDbGc{C!m-UNJ@P%(3@^!tO=rscyW)}YljA{ohfHSeb3EI|A1#lmQ|VCw{oKh zIt$MAG@F|@Kh7QR4FR}syN2jl6!!3Byv{$yGcEhRvI@Vt9lNP0ml==a#>UCl;cV3F zT)yR5b`0n~J|2ns8G!P9w4ceb zN2NC#ID2o2s=!cnqTz$e2z-~|N4{|$gqoy8m2aDOho|MJE3N5oh)NEtopUw1>3`b( zx4@))7iAq~W{nbHtvEFCdpq8oe!nSQ--*{L^Tv|PpwhVTpRX)TVd+>bE4VCjc z+D+fYp_%&0=mqW<|QDqV~6=F_^xTy1K} zUAs>)`vJ%pfc+3C@P(c{k$H8M-dL`ovZ55xRXl~ODG$TWm!=QGw}0-8BpT^rb`2m& zmIqXkNVg2S1N@(FbAf`m{&9@w@Tw>;fj_I-7y!7$)k9T#na?=NGQ6fN_RFiqr$ueb zV=sz#UJeHilL?F!SI$xy%Z}F?rc!UfbIBH`6d^lm-cf|c>EmQ%VY8pTrg%r!3?8zo z-~Ne}rX<2|Ix}#6A)!R2M136}C7OKnHa+%0hv!&wdr&{|%kz;*UgKBtI!D+pUyF5K z>@m-S8tl^t;fp#ydj;ExEQ?5})dOUlAx-f_Nhy9q7M;Z|fiS*^$v%I{8+Ii~OkJ45 z0q?Jg-PnX9zMucj6iS3tuh5Kv3>k3Xfp-+YMZEs~Cx@`zuhuGfk?+ zGyPqYP#H7O_ha4a)#uEv!Z)zOZ-rIHZ(jGYYt4-4Ma+?L=+iBnsr@Fb8N2$u>fPFuZQUe zZm==M;X%UnTE&{xB40I^{h!otT|d2>-&NY^a6o&uL+*z}1LTr0FpL*aL1P!cWQcC% zLuJ;L3nHu+iMTmkobkJmghqngwHU~|gI~|P699Yca)*dDLA7twews5j-uS`SpGR&D zCZBm$7CYn3mU4-@Wopw1j#tb_rGkTH4^tOVV-F#@tm3A^S)@^mD*lbZ2z}hMUtb1R zzwUwYo*f=)8CEz7>&aFAqgiMx)oHX0p*}O`JW84nRJ0G^J~s6n6Ov-D(u?>X2}X|6 zukc%5)}F{K^)A&pcBphmllg(woSP2$<)AB@o1bo-fO^_jA-0hypR4Ww>e~z{Zr}^0K=uKwFD`bFre_>rIkF{zx?QHOBiAk>XFt=n;Ime?M(7J@So&b`e7 zqCkqO5kxF3()^|vDL%XooOWyDNF$|r6s4+@2ozx07Okuc^k`TeiQLK#=bIJ-Om8@{ z;9Vp~kf_ZgDdC?hm9?~D=i?&QFHfzJ$^CrgUUAN9tSM(r zspW{S3`n)qiFiY02QP~ru9mai12{iqC@C|LGfPUf3rIWmg#@*%opki5=9W#o%LYaj zO-RFaw;xQgL4`OxQpRJ&6>0L&;k@PFm_bR>KZ;#?>{y@SFYld&bkbYEwn@B(iYF?l zlt?1gU-atX)fuY2AS_hrDI>T`9e+ixS1*!HRKkFIOSjHZf%y}f#)}h!MwdcaR{TMK znjG~-{^?gD!zoh}vRK6rmFwmiDG~7EM&Dp*Cn0@lxI#oKi+ku1pV?QgGPp)Lg)QMP zWpFhaD6Y(sT+KyYHp;%g07+V2Rv=6$;^TQt`DPmT>*W(Qq+rFza4)nEhY?VsdoQF*^P&HE%vk;aB8NeL4MyJ5J5|S-2 z(FA59>b)rG=rsfetrN?I{Jy1WV!tsEVxt;=Jlln*T*w{Lk&MRrYG*RGF1%~F&+|8N z5;|JWep^_L7)sqLzN+w%n9WHF5nSs%$sC#ZP?Ht&hhWmMkkF6WPV9`Rmr4m5aBCS` zXgOsk=KTF{PPy|?89#a{!K+_CJsshmt|2Ww|5$%<6_2jLD*7ad&)_^!6JSq|$>ViTBUD|CLRot*;5cY z|BJQglndsdc&j0=i?}R7#0q-~17__EC9044SrNZvmk66KR9`%&l52Y}s?R_qf=U@XG<~EvElQEo!*tL1@1xKWh@6gjyXD1pes8eBL1f7UZZOO*TpS7M5l$p*VQeY z>Ml;Ef*GSAf{TCO=xp-aDhdb|(z^<*;Q^vE9+tViCA{c-k6#r{Wb$$o)8-0QINWzW z9^gLgo?Nks({kfrAfbK3ZoFdT0~M$o4Wt0)+X~P3RgO&w#e6x_Z13Orzj7X_OwUsn zw0{TmJ&ypT3|}=!U-wuE|8nObZ2XOQ7pQ(goIE*Du6HePC2-NbZa-a1?NFT*-{q@( zH3v(O$K0{+MD5{m$nY+-K|?CW25m41{nWHW7x2d!BdIW zGEtNmFx2)FqyF9SnLlYaTwe4$Dn}=$voJ?>E!tEK_JkBy-V!BOjmzn&4S4YSJhS` zmRjN8nbkiIBM3-T&eufa&jKByo)pX^x0|g^nSEhZ84f5?;1pfx~J%Cmh z!`C9|w&vRCR8vD$(mQk=DgydIakIkk$MYFZ)*Im0Q{1jcyu)Hm-#7$sD`h5bdXK1B zA2U+Gd}N`6Skn;71A8xX9msqa<109XHSO?_snF+ty#QM5RWW}V6EJ)0ja>a6?M7Ff z`l}A+nqDyw=TyFCWZ?ZkPisNSP=zJ3hbD;={P#05YmR;MDA%;c5$ullIC4+H;M;To zSFj{cH8X)@GO(Xzff6CE-k4n37no?3wS_u>`|*R`)p?A039%RzAEhmGkXALWqR+-I zISUKo`I;eNY5bC51KCHfpCZ5>IPf|CFTiwp3IHc2T72+{dlNO}|wU+%38vCro zN4LxYL?pcK&@c|Eloz(!i-8piBr2nOYm8_64YQYn7QIFl=hXUZU}5(vnj~sDrtbtz zw2)pv=_`Us#PLS2=lA0te);?3i@m(s*jpqadgz(Nyc}Gfq*OK<9YRH!{+mi^ zu^~c3L3+YL3+3?s3yEJWjOYV5LXMdQ&2B#N9oD7Nb9>5qN5LRQ0?DHJi|@a`M6S9< zLgUw>!EnJ|4=~3pUCDdFSTN1}O`owPGbs?RP(Omr@eI{wJrn+g50azXn23_Evd8%3}z-YrFd-=i%-wuI{g`mhRAIZEzS;fA)Y#2x9 zrJ+38Pl;jN7*m|{ucxC~t5~_p~sQ6j48M?eG{~Mw}5+sGMZ`b;z>B)RI zcYPWdAOk5|4CTA1)MqJTlwFRYS6}4RZzif6E=l}st-rdA)7*ZRn~f`Zo!V%LYU^EM zlFmoBew{`Z8*m0oI?^!3#KB|RjZ<48h?+jCv|TjPpAX^jxKdM(gPoPfdN~L0p}Qi* z->*bwN1lZmi4#DtHI5ho+)16FO{b%ZrGoc;LN*Dl%2Ejgz7j#rd)UUvhOR({~_|wj#_h=!tE*$GX zi0^qLw(0h>PF(rqe=?iCDiu2>c>!vlP`JS&wSa#JPScs%A$9w=#@8pO-kEz`pJrOj zTVtrL_FPkY>-kVYTHQqdBtNY!0a|WO73lRdJ=an(@$d58Y>IO$;;D+9Ih6o}$PCKD zXDtaEG~N?6n+Xi=Dp0pr%2aY~c+on4kyQ~ey{NrLZIY4G#tv$9%ufkpoajzA> zFQD?;|Fy+!FJ_}BFR3}?PF5AoNMfG^aurJAV$YnlP>nl?xBa{bU4N>YLCY}Pw~Z_Q ziK;S`SWwW`TnbC-*8yym8H&PK_Bfc=%Y
38@C`KV~M>y@>O9Dak_jUO~?TesYJqz>zO zhLLXcUBo6GEwM@Lv~|=EiDathSxUJ?y*-H{bRWHtH+eLuGE$-Eem~H>&FS=%$1h6# z_Y83x1yQ<*A;Ooo`E z>vQCgLeKegQRU0G}9!QS!9m8gcX zmScwVQX=Ws)89+U-`FJcFw(~(NQKRuyuXXd(Wi+9M?a#X8xFC$10lFm^Ays_&oW@2 z9DQ)KE~UA^+xM23D142yn8~$p=Aje(fIE7Y9i^3Rq6uK!fE8_g+8Bcw2pF9d{YBc;FLCy*yF9gM zAE;`B9$&00S4$-h>6lSzg=3uY{!tkt2k^P+tBoB@NbVTDmTGm;ync;HsGh#dR%`TQ z63aj*^wEtO5GKKuFghAWOwEh?KTbQNQg9oxGI}wviE14O+mjwcme6j|p^1q9mf1nW z0z<+|&q*1Gb*O$qIu|-x2Z(J}SIFQ$das-+i}}JDmM}tQE@MAzA!c2lcEZUk@~-85 zCK45ZtAEy$!aj)@Vk<`|x1BKc4G}N{w{BLr;dP4)9~B{Cc%O=>em|+e$a#}7Tre4m z6nTMU9)r(H%-^EtleT$-5L~Dd8O(S#_ONp<0^)S8tp=Y|(5Ma0b(9<|&W1CY(bDV5 zlmNujE}_=ah!pv)f*4qPOAzLt=tNCH(aURAH2TJi&zv@FGKGWd&xW=psxlcx9b<=k zvtp(T4bT^ograls)-bMkQ5JGS?3yB%MCIu1em6Q!54r-dQWx6sAlA&o#e7i*+`j@c z31hma^&mWTm6LHFWIY+}l@?Spv?J{AW8glmx%=KfzNHQj3(6}$c2|T{0@2w%X<1L; zFap>tH7HRqff6Z%F-X6?$`*B$;-XuxZI%(Vu7IB$jNmz!){lj1WBx*)tsw-m#R5Ai z@7gEzou2v~AlAYyN!SG4i_FQaf-B!5dFY3TTH6Nza4T?g>%Q4Slkgmj&@NY2Ex$|X z7=DYo0C9G!+uIFC2F{Laq+>RJtB#3Fbj}1&7cpbTryxa-Zd{tWMdkdSe8*4UxP@B` z<5&9MmW7Y~Pt7v!4TGyYuk)Nwt88x_5D083;9?d}Js4NfLOk*z*IQIKKr|cNX^2=T zG0zX@Eys@!39&-g1V;-%G;u*S%lc`z?vJiZ2DDD%5gFNAnIs$;La6td1K{Cips0a9 zEjOn<=c#dVQ}X#eo3N2buN1xTEM1P%pWRKj(sy^cZ6vP`c+-C#&}fJO+(!94+1)IG zrLYs@9skXyGtN-;Zr1?U4lf~PSC6*cUE7)4;k0$VGIdGEK=Z!ct<$vLCV}Cf75I4T zTP~d!-j6Y6Km4}Da5g$SC!UEw-NpDDEBP9f!?{g3B^^7j&u0wb|Aw7BYw$!jG?Xv_ zeHs_5rHjuG+;q-Q)%%+%oX1q^$JA1wZ-h^7WB-beCfXi+UWR31#12A(#{ny^-+rw2 zdu?^>Olgh8zvQYrc#{KsPXFTK&9rWaUY1XVZ2uciwKm$a{cxCT*lAV%GiQ$GgB<)fSNDPyzAv8f03<;iyxIj^aP`IE;1YmHKO(pw_OBO}NG z8r{D3>31yhw-`p-_Fyn=x6gMxI*vKzo9Eqy#EBa5k?P402Qy9oer-YNq^~+oP!R%f z&_U=x<-L)Ec7YbwP0@I27eWp3BZtYK>wubzs%R+YqnvJ^mvpCerh3BhMyfrCCjuWxesZHya)B%$v2wxRfGrql}`E&LK`{P4Z^^BDbW0ePxDUhzF?wm^2;Vg zTWuPL-$q z>}07RTk@=VZ99ksm5)%VJ!u&6)iKoMtZJQ99iB!LX=5eZcSy7Kln)mn4fPj zgqF-gWnwP|VhF15|1k3^@4K+3;$vJO)MO3@t1I@(PV7ZxI0M9!29uc4x7k^{1Vjn9 z>j9JmGXo%$?w>W3nefAi^_`Z%gbSsnPGZ*VMByA!%)0FAKl9R^w9yGP!w-h-`K?l* z!@fJW@`>!2CVrnIFIz2FMI-JJVHtAsMdrem?|I|REPvuFR265so{G_U?A&U-VUTG= z2PUy~nYC>qwi5$8Nn`jPp6pD_$_EN*c%H;&3aYZIqOh6BNtRwYc4qXBZ|MUvstmE- zb_aD49GFGtD#?QUVrEkRRucb2Y=B2WrPF&JF;OSr-cT66RS26(w;MX6l*K`45+zWA z@#mj|_d4k2-}xDp!Y;t6T;f3`DEbExVB&p@iT9R|=XoM7vZVQkM8cftjKD(|yJE&p zS^}Ez%+KS53?%C#Q4#>c4mAE@wf6jc{fBP@er7LtyxlWe3OLvW3xbaW$BsCY8zss=xL z;8cL5tq}mfAz~oG->iGqAgsaY2|DZ)VaPhJJlr3@AOcOpe{s$eGx5DYbZ-2t6o+AZ zw!wp0bdqmD`Wk{vs7w{cXG4;#KyZL<)G-Y!nGb0_ zKdm|Z=|5&>zPc6`Z_4bDS}MSp-_&c~a>)uqA;>(F(Sc1`3;G_`3%Qj?Ll=I?sK=*4 zZ8=X`?L#gxSPw?I6f9V3i9?bkLKA-qY3R(ZQX7?~S~x8^Z?!T813SZxmN# z(=@l(Ebj+c+|W#86plLTaH4A3!^Y6_8nv-CoqMl3{5mK9&|8>L$y&Jzhnwmu+ZAPM zLkCz(Q?1ITiR8!(N9NeR3P?y8Lv~iH2o2a0ru7gRiD}4)FDLGiTozM_1SqIr&p`uC zl+v`jZ9vGUej&Eyz0O~CL#BPxa)%?q6@Vr_D%NSpP=XPKwdCQDhTZUEo32~zk5~}B$mc=54TPWKUn1HPBp$dqywnX%#GX&*GG@+Vb@HS8 zt6Gle{_WCvzx-P6@O$9ud+O?og+c!(CyJyS3kKbw1@*V}y?nXa>hXIF2)Y*z8l0Zy zwrcR#%@3MlL2ZK6UtO?lIZIdio@??o3fJ`V#C8*h9OF{viQAkwn2LMn%xnyfJ>{6q zoGna5{u!eG#jmr~erz$HSQV;22G`m$lCHhVZFxS?oCwu9_>p~&hgD^c@IU4(BM3oG z5}iBoBb-eN_jkK9bR0KzUg;a6G&kfXJ&$9BxA&1m(?SV!8=rXdn^o&`vp_(? z|7c!j&juY@RAQ8x0`&HjF*EEW&$>HOkOCnwaP9hfG zJB=?Z`4lF-Qe|8M3yhEt*s-ZH1S_62RbRey7oXU=_lHNp>D`ZPKe-bW$pFUiaX$Z- zPxvz1XoLZR+AvluRxe3bF<>RaRg9wAGfuut?rFvi+1&r>y!q1$h=CTX;)+Aq_EF5} zg33i#hyac&FSq zoUNzHBpTz^QX3gIXksQFbrZkfTddw0q=rR$RHdaTe7MJc{B5*|iN)?l*F{udNV?2H zgbXH%c!AtM17UYL;R1Yb{rb~JR>~5Yztqpsi%eD)a;WA0v|l;8vEl`A7)l;cjK4Ld zh`ckkK;KvTV4?Xp8b1$WGRRmCGam9aN-<|dds`Xykeq=6-7p%%K~}kAQu&Au&LdV- zl{XvGjRPm5<12DCj+4is7%vw=_cwg3Gj`*+@5m=^jRU0ufjzR)XJ=$WvE#pWO!7BNXC{+e#-S54h@4cMH28| z)xME!2_O{01Y3O#9aLp1p&q8kuMiEk@N$*fi>rd~6|m0ekH^V@6P|AHq9(zf_>Z5X zly-T&Q8dx9F_}=i)Pm^|*N$KMN_2xmH?taHu#(a$J5OGtb;0f@@4?O%W>h)FK9?|b za5OLAI}BT>>ooko8r0PBkOzFpY}^ma$lTuQMVOvhF-^bH%to>AGmui1|KOMB&8meJ zu&Tr~GJC6^()3{c$w;OH75j4;O_*}J>bm;R?M4Qla8`L1#P85~!W676a5b{!m^IR& z_FcasjlMlKWhq&L9Vb#NOv|FEKRJk)IN0XynF_&)ebl?oB&($2E)12NB=xLFq~sFX z(&(AF1_qmOFl@earlvv6^P~-2Wm zYxkVL-v}HF^7;JX5jE>72liLrX7g$M@94{YLM+$qkHNB;9RiT^`FNNx9Y_0v>P~>J zzXB3>e~n>VyW25t&`-snq3xH=Z36Ry_3eMZ4_{xa`u$?#GI-wGlE-wD`TXD>5RBPh@AA1y63;-gr!V20w7Xyr0c=tyLA69bA)MY|Fn>KW~yWBc`7?zE=&JX zIPo{{@JrjQMmQy3v=yDTY#c9s}-v_ zjoiJm9l_EwaQu`TxbsY`;385dnsYq0y4$U={i%_X&_HG59mMugI|YNKDxpWgwfza9 zr8Tu2=5L$a?D(6`v6*BoUr>YtIfV4X=8b*^e|FBP^DE@J5Fja`IZ+gLAfyDmuwtWx zjVm4U9$1`7DCE~|5^DuI3m`G!d}#x$<(+p4HdUCBQ@vAw`|{704#f&TL`(Vaw1kMRVC;M-Iv^`zIF0nt8&;;}Iz~e%ZGLKnYbn#q{~0F(*NI$6n)8nxQa5a z60<`5B0!lX#H|Qfq<}E9L~Rl&szMQEIK@2Oy!<0dW@9Beau~yTulkTS+;Y|s1`T7z zrHV@Us7h}1UB_bCX|z0PsgagY3LS(kdg3U+6UjS-t&{NCmba@Ud~yWPNMiVhA@&!m z#-j&)lQ$NAimbPW*lT>UnL7w;@sH1kDC)|0hvc(Dj5MGDANeS**bAm}(N`oy>u$A| zB-*#72!Wmyz9%9NVfp)}g56(r#W55Dg)qw*^REbU{TvmT`tY#B&j(-RxeGBLZzsnH zOU6qe-%i0#W^0PsJRu}wWQ5SKCD8L(>yLJm4xbQ=y?1ZNFPAeFnltrWS+7{FAcqKO&eHk2&S84ijc zDJBLd4`QW$1yyzHqN&kQufyFsslETz|M#Ya_or;)eVzg`I%qVW2^ib`uk=El3}hAq zZG>S<%^F5}yvbyo_UpzTDiAeB^sXpIiw+U}Nz|pzV1i8T$G|ukwUh5sKpv^z;rE|> zld4tz0aQ5LM!9!;@8xw^r_-oH7AH%8kw3qc3gx}CW>wfrx_&uWh?~O%qT)>(w1yV3 zDaY|@hb!pBH1cR$Mnkg>TZ@R37n-oLB!lB3EZ2)t_YkIo-Bc=Z4xzNCW6q}1f>}%e zYN^0vR>%)A;dx5~xAV1UK2-NHQW)YmHF2UF(f{^{Wj>CVeqyW?uh)aM8pOb(d3)H>|;mclo zgEAq+XFs05CRdzl3L)iq8X8`Dnbfie!XK`=5RXHeA6ZA{?V20TQ9ynXjBujnaLa7M zXT6bv`t_V0ZANa6+a|XLjccv{s@cC1*!Q4kDWG6K-6o~3Ern(DR8l+R9^UTziw$un z6*Petc!T`!j{6n*KUMFzYx{+3kmqIqi~Gh90Y-X8)CY#X2fsc=P5)l=o|!YUPiht3 zyLN+Po>M&wg50!Hx#E5uEe>75+?$nBf?+56FRn|~LINcRx0z3wy+I3Nzu6u$2TtfX zeG`_0b4#B`HF_SF$13;XrxEg0tJulI1*}6FkwiO5;yyrxaj?R~{0)IVjmi?v4m&cKE zZJaK;ydOhD)jVhNqIL3d-UVU~Ug}S^wjn(3;|FMUz%->9Lw9oGN-+hr7&{tM~jSrJ8VExZhB_PZmjv ztvRGaoF}{ORg8xqe_P2$oR6Ov7xTR_7r3__KCLW!h;O&I_6A9TuVy|9+cF<>q)zl2eP8lPJP{gF!x$ z@RX15E$tjgIoRn|jTWm8Wz7@$hC+nXq!79^$!de9Vok#1AR0<)w8^06Q!=}V;w?m_kbJKs^KIg7jVh@F%!-dQvvjERTEyfF2Aq!*rOa@q znGdF*Ng~=`OF`v{_<@D@)tww*cAPK5alnI+7P=(9T?ghfr~o$h(p)EY8HutMX%;>J z7_5#&*MyZwPLCRvB#N#wJ{d?t;YFYN?z_OzSBz%ZPzLJz8lsjEW|Eo*Rs>BB;m;T# z&w?aA8X}4QyDvT)gL!s+?+S+48bxWBzzwJDoV(pE$6x=I487ie&-y|paxEZ?ckVwW z@hFP)PHoCx$h@H?459Uv&wjM*_cMzb&9M-SaS@$li&wQNrlrEAw*ak>zAt3OW>k2m zu&YJ-qFkiWGh==ZDRfuDOe6P|i`kEoVuHX^!bZ8)Ge%AqO1>H(t$L8ljvmus@VRNH zj+*89hVVzfksJMN!;2rPSc!zsYMiJm8qd`f%7=E~gG4WfLI4+sftgF1klo!q3bB5& z3W*=PyhIRl@GT7;G>X(z&UK3yl82xy{6|znMFHu=A%;`XQm|oq~r?jf8AP>_vqBtx;te&zWLNH-u00Sv8xNdWt|cwwWp{ zj}T%NZ-=oYifN7xH-C{)#bg2)03knFGg(#NFGgo@0JDp)&PnOq3JyA(n2Pl0I>D;QTP)3Ui^>NPNrV>y zk0#d51ERsI!6OrIhR}vlj%m&hMPKm2=x0`BFt{cUiVSP<7`FYj>&~ZRq8C%3hglhy zZLWEXhmRT-asVFN2k(d5G%m`xWbH!~Cq@P<__nCJT(X%Quh1U|V-m7%{Pdn9BX7H$ zxJdO(IiQ9(J41A50Rs7fgEx%`_aL$CN!i@G#zIYI6rQIqlru9?=|-($Y~rgoLO2zY z=`l-FvWkg9<0aZ}$VoC_c~x37Tg^L5|8S4j6a28y%m(k#aI&Xe{Tn_vocXvPbqFnF zoa0ELkkX)F7PCczcwWQ_umW#%8W5Vvv7bbIm8locc5PUzKau!l zH}T`qZ@~tlY~=y}F2-dr(XtuiS+Cl>bLKc!C2Xy_rSBZ{*~h)p3_YY%n-y^g#`iHl znQ`zxW&SFblfu2!V}UwM_-ZEth~wIM#wu4pn6O+c?0fpYoe6=+f{c6rI@wt7w>mB|Del3V5}3pz(c)1N5Xw$!hMfZ zuTh^Ks{gT6E<0X1_fq^>Mw5K_u)HCa5wJI^cErT#a~QL$m$&XyiC<{_YeLWCTRWd% zw;KZQwxJ&WpNmxEB`5dv)J(kTpJFkS-yCcG@$7Vo`e2~Twfo>pQ^LL)u5?%DCl?R8 z4{4mjxXLEo^Ul4ekGuP6v0ROIG@tx@uC-)YA<8+`eoL|k;Z*9cV)3Niu3yGW-8KTc z_xzl4Z>zYkn%0TQbCGj5GM4=P=QOj2&X;leZtu3Zjd-?k^X}t|XsCIuY9Ac6>V*aA zTuHdw(x#F!1nIUMq(!bp48-DijaJ8|*K!tdWTTqjL-gK8i5zDxqJkDVT4{zGi@3CL zl^MnD5S~ti@12$t?St+m3uZT{-%ZcDc@R}v%z)2cz%S)Y7T0_fFR`pD7|oi~N!r9^ zmN_Af=-<_+1jP8Z?5SeXF7_Vt&$Kw!mPEtF^#_|QD4pa+OLKw_z!O0=UsQZ!4bQn2~T?EVPE@4=KD@q2$dg6E)GI&HVRYz>X!C}1w+a8#WVpH z;&>w=(p7MY%*)L4)?5|$>a_FXdkRU|h3^7ile$(6$}zX>%FP>22@oo1UJjf0eFdMn zXmL~`0Ucv_+GN(-us!dF9h;Q`Ig5sQjy3?^4B8pUDuOvW5^JO|ooPa7)(g4yn)ko8 zoo@4bT%|&xtWq9_wY0;%kVZ$=s<+14%<1wjRMhZrjt@kQf6lQaEIxBuN{KX%Ug)M| zMnc{IErrT8h)dudw98-c3yUkq{kfu0_?C89=$`p%8Deuwt5-P0&*1Z(389&wkt|Q} zWEtS~7b?8RT_4)yqa`K za&AL}oXpVIHX6oG>aV2Kw2C|cWf|d*TuX*IG(J^75+#)+xJlvk4fF~I^vm|Yfg>9X>&$m zG2IxA^k;A7ewOnguBpbpB?2wH{Qvb(ez|jbGL0T|{n?AjLjt0huMa=y|8ke#xCafO z;$LDe`HC*uoycRMTg!o8I(cQ7oXCxlBL83k_cbHGqpDO>kdgR&#fJ9={DuOgK#m?w z|Mj}b{s9B)BUnRJf{QtvPyb@-jY}ReCr~CPhMa$Ih&^-jd?~m*iqT9f79aBqu@q+a z46bK#YR`p9T(W===g^6^?sxhmFc3Vw_Z)HHHk8drqA6p%Ob!$kl`gvU)3(eIa*N;! z`9GS@Iw;EbecMWxu)tD+z`_zscZ0apvh)(t(v5VZba!``Ac_*wCEXGqx*JhCrQYZJ zn|WvUe|Bfsy`KBN&ht3dlKtyXwPgDL#bMTZK2>e-M;dw?OzhwmA?BW>6@LF+2 zmPjqoTWD)eInG#^5Z#GJ3m~~SQ)GwTKnn0sW#IAsSoPGX^^1sB(%5l^wT(3Ne%WaH@g5^v<#U$&ME##0dKo-P=^2}EV5!%IM8Q0Q-cQJ5w zJ5Hb&%HL@Ui5}s0Kx!{onTULfC_k+T~r(B!<`LQ^Z7XLULidHr_N8-LRon@daw$u zr3yQA27kWJy6<&hFKwbOJp|GlzTCsS@a;{o&S&kIBV&ODBcq{2S&P^s+N1Yd)W<;m zB-fsnxHpk@RfPWAF*BHtztR4B#cQp3?fl_G2NJ>dr624l3j=~B8y9BXX8gcGw!kd= z8Bm$hHCU7aStaevy55ybT2(lH4CDgo|v zX{3_$h@a|_1UjFYKHqm;IrX^dW`9@$D4S16Zmlr#wufO_soe9&%AJVoiRSNfJO%!X zb_FE$AzR*)lSJq>uc$gcI6c;oPdJwa(6qbHoH%cId$(-FiMLJ|BCbvxdNyQF&KjJ; zU)+{266TMW*Ssq8ZL?_(Sag|^|9U*Zzm(dBIH&cchQ)67uZD*B^k64sG%377;ypez zUUhob^@hn>lZH*MWfIS(W;?Adl^pxVXCu}cayf*WB>s!m;rR@`zJhvP*?s(dbWi7D z@yFlXXD>x?fPJJkfj}yWE5Hf1U@uU!n(2JKK0W8`>;t(YiaTD#Qq@8K5+WY4qx%OM zSrQ~1YdhWY%Uf7NWj|GZ+IFm*s9JyE=t48)-bCM3gwxx9@oe2<#9r8|^F+8b)me%} zvzSXzL7a^4RZZk9so8^opnIjtvCr{jHj%1v_1NJW&5vt#9#ec|@+LYIam@g4|?^QPR;tqfqOmPU_)6<_o z;>?%O-r#nV{KJnPdQPk^&-lyYsJ#0coqn?A=YaGX?J%w)p2cX%hG}lWHN*)Ih@RuA zn2#vU+Rz1<>&NBiQO~{aisgNZ4si_x-D5sa5ATvWxF5 zVTM*|B!l2e{O-2!!=2CsRdL!c5<)<&3S_D3vl+#EUd+fgux*hwa9T1UZ$VJ<{iR|h zHt|XK)>dq-dT5cgO{-zmE)9-cld`zdxqxU+Og90X^+Xb`i?T^V?o4_o=~X2jQ^}L# zq&3$mDi(V|UAwaqeQ8G2nzxj-KUL8$(#j4GnP!b-Lt!l6%b3d@RxzErztZc)DpP6} zD#`zIKw^0Cf{yn_i7`I?h8 z1ZN2SB{2t_oSsenmLTy=UaepMCYBL|8^oBrPzkpyR+nbzP5K_M$BIpGtHlq>W?T1g#?_CB7 zBtnreWS(ZelsAUYIZL>q)LRNnI8(J(X3KN!#88EPX?l~YU=a)pSBzr>j5y--dM}5W z&xql@vhg9CD3-rqbm%0|w1w=VL^RbR;iT{m+GLpPnVu)kXTEEBD^119q3!@E}$R>{ikr1{1ULKU4wEFj6sAXU5|Vz&(7I zhFL`9-u?5}RO&xBO=&4vJM=kE_SvQUX^)YV6wTkYg@^h=OFR12@8U&%w{do> z8G$nt$q94EVtSO_ntUPw(w$*^BOC+!^gL5Xs>ziRR|%E#OtcD=%BeF4#8DQ%3p{!7=W?l{daZQI)DA zFw>B|`;_6K$fRFRC$!`t}-2N!$=xGbYPLjd?P;6&@tNSD6UH+>@af35t&| zT5|PePP9#r`BmEWPkQcB*FGyPIZv^>VMX$BB;X-3741pp_CWpzJ~17CG?AK^Z`C1ZAy z)cJ2cc!0WFZF%v7n9-#r3EbJu32qC1EoP*d&GfXTMvK~)*f32U4Er_#p^}V;)adf`CSMNT)m|d{x-FppZot5}S?YG#ckXsD;Lf`{q5{2uLxOmyT8OeFlGkMw zTgq63mP1HF;_&urM0|PC&vX1R&A1IOFNuRoV$I_~rZn}yZG!ymzs5?6_a{<5XQ9U* zj{WWz3LdD|o*u88o0_^Pd;S-Q__#swG3ecG+Pe+wfGg`Kjm6EH3vSKx#gl?n35!*Z z=73qZ+Ev#l_c&A*xur}YZ*sf-$*xA9+RDN$a(6rJt9zw^-;3)jah(qX*lId${vRG2 zpL+6Z5{ul{f`8GowwHJyd7&WWQ z8+j2F{?1cwA1Zc^Hr{5)+$7>8?QsQu!t1+Si7sf?bG`=Ly@&f>urUCmlS<6!!pHH` zdTveR$x(^m>wAlyhp+*Kq481+p`l7y3I>)hJAiq4t6}R@Zo;7E4=QUkKoiAt(|h1+ z(mZC>g*{*(CH6NGUUx)OHc6RwoAh^`D!1^{&iv=c`fQv)TccXwy5OQf*c+cyG&h6e zZ8Fd&Q%#oyrWcHf++bv(#r!|9+zB$7hxK_+(I|Jf8rk22;zZ+<`Wkj)G&qpbAnCuKEpT?cLYs8PGHUi0DL9|^jROntm%6=w(+)l)e<#$S9J%U#Bd3r=8t zxUl!?hpF?`Q)wnimV|+GU|I?k#LeZRfltz6A-aS|R=6Eb>;L+x*<@}x4 z{exNSO6?U18QQEPF7R-ZI+fZ0C7`8q)zec>0yDtzs%m>Huo@*2FOD9S6()P`)Ii|ofmO)%-+O{_0N^{qv)|x2IU)L+TS72wUj1N?02vnhZC}rei z5GoQH`IFu~O|x=~wHA!aq%Fmw=z8F*Z5hM_`XbNHTi(v~E-W{3`AWvBw*3BtmUK&y zo-+J-LNAWWoCm?z^1qxrhCH9oB-eL8_E9(P($wRAfi6Lu757*PLP2`h`EgvL%iMam zgmZFepb{qB=cdZid0XE+qgH{;DP4! zGhcx)MhIb7pgNYUDkU=mpLQzS=7*oYO0EX6lfBzXd^=_>`}~Z;H~y=WB|+@9qDUxK z5ID-NMdpIbiPk6Qn0FCyI)Q`1ke27fVzHyRj@rt&ljafRnu4;mZ|aEHRz#uddu*r)!cJpIDO)G7>YHR`ec?Q zrTmk9v=w~LBNWV^ER%h|8b@4&&uS;OW?5dt1}cX~bs3k4h`yWrK@Ne2;D=J`IlQ8# zK81Ye=REl#2UB~amK^@zGsa=3V`Sodn?K#0|MOE@B}U$qbnbd>YeR0mzd@#ay0@>22{YyrvZXkfgVM)~Mv?|wQwl@ExNLui zxA4(Ch7=Bum-a^tw%|8Wo*J9E_igp1_u0j;u`y4>8j|W|ES5w&JZw%h0MXDCc)#{% zG@?G}+v1e+oIXR|Lp!#(tiN(8l-MQ}!Rn;g(dUV#mo9qGjb|lSOi8+N zZky$Egq*_E^?E#*$pK~D#Hc;}SW)b*YRIjAm7~Mwh(wytzsL4HoZpysBUZ$DemK7A zu+8EGZu8t4saUBeQ5zV0`rYnEYm5OLI67vRAF8h^$a2O{ho(+}VFf%lW~OJCgCrIt zSau!l7naCIe>=S|=}q;n9ki_Sa`LaEPnk@>FBI?i^T23?0EpI3n3f4Xr^%bhx*{eUGQSeJv^sqd6 zG?>HZ_co8cpLKd5@4U_FybY@n-Me|b?wq*peERV5`1)hQoUvcr|7Kj<<*WhsZ|AQU z6V)y1_XBUQeRi^UQqI%sE_vdYrHcdr*&i@cmhLXe==PBlP{{h+LM|M>GQOvNi>gg_c6v`3h1P^(4T&`ko4GRU?Sr-e2~-xaM}%SUjq~ z4}SBpUUOP}Vt|GEM}fO8-R-R~;U~NyGyan!MWyMPTqn0E1=Z&)vFW)U@o)t@MpkXJ z{|KwU?0vF5(=7rQP{#pwQNTa< z)5J-gr5VM4S>o-}X)@`XzS7n6ANihy(`vQP&OgGrYvuy>#X!SV&pjKYHrr$4(+^qp z9#0&#X*zM0?ov?oVvyHy0Z)NSCa^jHnhv~9cvfkqOO1vcZ)ryDub8v-G#S!N%g`}o zl5e~9BORJVfVI*$YiPu@M899Xfd-;`fQ6mQS7n2dYj?BZ)*(DbvO(Bzslct{w~x1P z_o?eht$}mLk2Zhvh?W^|I@>P_F}-8W78O};E#L3|K6FjG|HQ5c>h@93&hwGcd^s*G zX&P9}1BRQ(6g##`bSOeY5n4=&Y1=|~h73EJ3>2T`v}#N8xZPYyLUgi=8lenq$YdVb z3SCdmrP&R0pjiW%N z{KFt;Ln#-y74CMt7z-lh!YEC2&>v|{(5c2!)O@R)Vydr@KAR*lSUha2S*cjbj2QvS za*NS|9?Q$y3^HC%LfqxRSoc5zE94b|HvxxHOI7pt%h3ydabZ^hc+WJ{mc$}9IM8>I zM-c7UH+_Vu_xZ+bB=mbp&`WcM$>)io^q8;9n>ZOP?y+j=pFyBicEP-oUkymPmBu!i z-q2gV{{qsw#4t+5&>+5JlEO--*xRp_XG>d{hoi`&-y?eWqu$dOJ$o(2MQeDU`W&Lo zWy)T@aCp7s&ls~+t`lv3Yp9yR42AXkuf?T4~~pF0wCW1`%lXY<0SjgSftbTtMXZO~9(Br3xuWlXx; z=8PrG?Zd~{5u=LaILTlrSdk4u%S2YD0oG?sW1x58UE4cX?U8nu3?U*=T!$BWB>KZQ z)!)Cl2&O#W3uU#l7@RXhS*kkw#U#Xsjwr8Y;!Ijk8z4*5y{&Pjb{S-iHroUAb<%lX zY}dOF?sN82PMx(q zKZr){-h9Vx0=^3Wg(`-q3HmW6i1f=Mi6KyZd;bfGRgMD=<+u$K{UEKGr@+r_7=I>M z5RSPALrA1%Jg$O;2TWZ4x$>%ClHTE8y}J+jCu%_*r{5 zLm-HZ$DLP`!9gM4(<pC2mT`5@|l3;-wAwAj!Lw-_0TwIpcM- zBp$bNGxK#e#uU|!@s@}AFWUm{O!^Yy{bp59`}6W@{Nr_ivRc5;Qoc5?Huz_zX}%IY zV6N1J{LMZL`Ttn}+aXoP+wJeKggf&e{@FOL2Cfv4zh&}&>+6Db{J)^z`5lnNP~TmwE-LkJpYi`<8rN$91#}Pl3T+QpOaE7~>by|uyeK9-Kt7?li?EC2 z?hqw|NO?cFl>S4$_0=KMKRI`zL6K4T`wN|%SVSrCeI`IE!ei@07;n1bnm4cB_jwsV zW4{`i0Y|qa)`#4|r|8Rj`eoCq-d*VF8V!WeDq~!9GuufPXu-UL{QD z`dcy)_@~sJ6SM6lyqg8vhR-F8FYZq!)9)yf=H4HF_#52Hmp$inbkdz7c$`N#(}xBj zQA?Mk%k%lF@oPagIlMSgxzt3g(&Zz^!(?GYZy)Ch&;l2$4_(nmJ%+4(w&1eHg%!(5 zMYMaAZZ|c&I~z9be$RB<8euObab5q0uaYghs&Iyf6Y57gcDa(=?XxO*)YOXYWC*C} zYmUwXI{f+%gwCeo!o?MU(KLp#=^L-#nEZMA2%MfNIO*tWnr%2|yRDpc4@HfCG;ice zWVh)uwGn}brih4xHy0JWVs>6?eYk-H_NJO2D`YulT zM;Y(^k!AY!n!i8rdGx9|<*nr_c$&N08uL~5zCUO7;1aZpAeV*PzJL*<;PN?ixyma4 z{UP~>$zWA7A8)=e$oue;24X=A$mq(tDk447Rg^EdHhEc7#^_U-o28E>%n-NF)tKD% zG!G)#SO~bLTAwdqSBX^&Wmm2H@8~-IK){^CfFjl2IcpP;rebiMgOtf1pv^fEGi;{7 ztJpH`LshJDevhUw#aGWjUGY$?h6vUy7qD4dxZ|^9G^zOB8E_OYH6u)J(XgsC+p*oVodi3VXSRQ}rywxqyfPg}#N zt>-o=uR0f0%W69+J@>AbU#V1iK|lp_(sMzUpjkw%G`BPstrV72MSt@gv5x-Z(79Ze zj6W6VJ-IGwP=$g^6Q6WsF^V+Ml8?KIA~7uW_;XZ>*?@-DGg6&HmrO^sEQ_($O1qLm>vu!Wc9SWlD>W^viREf*k9sPtEL4`a=a1>+JYojuHDW)~ge zjK{Mk#EwV0aDWNkTin0_SumIAI&o#T;S@zUu@PUOy&^9fTPIenbY^fW#?PhLO+*k> z$%6Oge^5L!Mc(gJG%$?Whs?q>Pzf_z$Sm!g977ZcOO~D`NFx%A+YwQU6PoL(Lj1l*A4AA+tTi~(oZLiU>}>;bhVs#lEXGwMb3>6ng( z%}?90F^&oNKIh&jYt;*5$u|_?|NQ{mci(LXu1nRy^2M04Nfo%Ik)``Cyyxu=!=$Xd z@IBX!Yz_7JoG3$Hw@_o`BmW%_Ew>!GCVn{beG9Y(e#pMahYoWV_=0;G9u( ze9&G&C*k!~yO+}CNRzR#F(0KQtE^j)(&d|*o;~-nZnz-qGLnM5Ve_gRgUKM>Xx+V` z-xEDY|JwlvuRPYFw{z`A;n{s_tN8u0{oRB>gHOv#-4{`Lil+_UwcIHh0JsOJn~u!Z zio?8b7gH5JV-l$fwfVYTXFZOkl3XXuE$7-BH$VRO@bVoCHJ@s=_qg@D5+<$_Ft^M1 z@4%wOWs!NcO){W7ai$H{^j&zTA@+m^$J>7cX*$6f3CuJ#bN;SBc31h+ zrzg$^za5wxLG> zb1bP5hYyo2KQ!w!FLS@#aZ|SaT|Y1^WxL3D>3cZN2Q#U-J;M(5(X2f(AUC*91I+J} zZ5B0DB1WFe%K=wJ52+;9pYz4th8hCU$UT*#yrlHbPs6rW8}h7?nm7>egT+)j;A9yP zu$NoKb5q*fTp_aJKCjDT%a-`Oaj0!+*rpo4`cArvq4k`~pmo!%!PJ;~)FJr(=Ov^c zyzO1h>ry-K`pc%g%t8kSM7IUVYSMvAVL9b2Jp-z~@m`+Kqc9nhzhI@07unQ}uZ<2A)oPNKwd(H%Xzz9Lhqz(SeQV?T3|dG<1|qcEzIFOe`^+ zIb-QWv{F;nJ~(nr(7n<~wo{Uw)BvhV)DG%R!Zp(kL@-3wDFgPv2Aw1q_FS4>%JfiW{ zmrH~#PuUxrs49U~@wNs8T8^U{h=h(zDm3<8BA$~Y*FQ50mRNanfQ`U&o*rdI zr&bkQp0AmKnUJ7%`V-i=)DD*Xg6NhC%*efwMUgY{0=8EK+Nzao7=qqCN=Wdl!H&S$ z&0W|GPFa7jm^{BGBs19loU}+N*2P)&%seU`elWVN4BxY`CyAWj75?-UkujI^^?ZUV zB=Qw*!EI+oAaHSZ{x|BGwL?(Uy10WZ?ghGO`D|NVVJzWve_`J0D?K7KkpC6=isP*?7)zq>#b zKVZ#X{KBx>Ks}w609*XU669B_1Gy;OH`UE{>jqn5*;Fo&7z@#zoPuN)gRI2@7IR8& zayp#Ho=qo*6XNqMHI+3ji3W*>mEj7@R>NURV#p}aqsE183_NyDS&)t(6i=FUM^dw; zp?zId4Gw-v*5a~fp-^0*?K{&^0`ua&J3wOz`h~`<{U)_JMH)F45<{MX_0Ra^_X?F$ zd)q*j8%GO}{>EKvX$X1D8CZUIVaVa9tbqbVluS#Mr?b2wO3vvfey*=`f7-+)M18k9 zVb{(GZxQf5*7b_FwVW?_2{dl1mAe0iWLqgev(5*awO?RyW(*j_2};T95<~6+VWx;A zDBka?-cR>+@_6+q-DAb^l+Y3`{Oq4RFoL}ivY0Qxe3jzV2q7+*@nJ4ouYJqAyF~NM zZ#sA-ZOdQM+5Ty|!9m;Rh;}UVy)W=&rq}$ed&*HbODPY{FKutFbNOV%X~laN>|l+a z7blRiSI6~xiYIjApyA{3w~a};fp-B@=clL9esfmuTDxgAH%nOT?OypD6z*O$~HYtuZb*1DH@w zw0ij*`d^PfO2QeLfQ5Q_VvrxUpx4(#>bIMiRBZt_5dk<|H4lq5rwTwZ>fK#?qtn_h zAeKtVQmyd6Ze80UdR!;E`MdV`m%>H&ZOc^JW7_Y!FoJe`hdZ8f6DF0)N4 z1@<3L6xD>CK8>giOyttxvvpvQntm`@5;b0&qwx>(F`$%N>OL{YV=xxWKH0DxS(7}P zFVLPa+y#JAlMK#&HMa_ue?F!G%y}031YXSZ-GzXHWFP4L(?g;4Y*)a)A?kbXr-^}F zE{VH4CKWxCmB8d6N>UUqUdmX^*!rgiAj9xrH^PH~x-Z5(G9M+Xf|D>xH0w_9_fEF; zwg4EdY4UJA&GWz1uRk`nvjHKB-1CwM{PDL+!Rj<)1!K#Yi74l`BN08%>hQsOE}z@o z=a<%a>SoIIv(p~5-GL{PHHP*gxfxBFO;56>cfdwDUB+ZVj_V(78F*DsBDvi*Pgsty zLhU;XjEPG~5|}Pu;#f;WCD%{~vJ_B?b_-G6)W{1iARDyY`u+T(sXQh6<2R6>;o#Om zfuI+V<@}3f+^*x&@kO*TKa&QUH)05Xf1AQrWGRp4j@mL)b(E9F33DUI!7)97bRPBM zIO4ZFBVo{+k`^jUmM)W6UBtwPSfC?qKv5RxRf&kK+YwnsnFp%h}fME0~OwEA=Le#Mg zak)deP0Cy9VrXWFo9#(tacNgXV2tQkdD%}<1x5z(6kLEB)28b-i5INQy5ko8*yKQI$eUCuqVF}Y7F1oCi$y1_w*~T!Z8&f^ z3W1{0{6RQ~JDKUag0$F+4PioE|uaJu2r0AKqRwovBq^`=z2*FTr7k$XNo`GAX z-{eyQRV1j^YU;kY7Mfma>LpE*Yh}J0Aq-PW1bE-n#UqyPxJh~=Tc3QPkg`%%W-!1T zO68+(FkSL2-T#H06cCIatk+SF-LMe+wl4`|4%ON(h9c2d!lC`VA2dnsUhwLc{!wiC zDTmLSl#HFh!v_h2Duzjy#Es*+yJT^w-7aHfu;40Z`SdeJ2;&BU`XaMV3AmEmN??3k zW(z*#1~K?q8?C~y*Nu8GOy%rsDg8gcLb!;Lpwcb`4&UU9pdW4yKb9OMSQ5o@22kj5 zQs+oY#Bie48#-wRF)?#MVKrybCd5Qc4bfLX27O{X3%4!h!SrpD`f%aX68@}vD^E5y zrwwK)5nGo}i7Cb0!ey>|wkFmuTl&0gk^z7TJK?Tw8F4v)<@Xc`VDq0b!a}cUb6vc+ zZpPJUPU$)t1|I9zL$)~cVw9_*OMJ$j?(Z9MHm{zYpXDm~VtqH6(e7XX()edadFK5LFv{v9o8E@CX2%9aS@ zg%QXZk*I_EG~FopWZo(TA!v^|TS)$l{jRW&2d=8?`jTU`2dW?4UwyaZ)Mx&BoO)65 z5c&7tg)U(^6Lw0Fwo~&w!Ji2>E=2+)9^sx~w;$o}LH4PK4f3har#`&9wj(F7&&mNjSJuRdi3NpT1{Lu zK0f^iL<1HuuPW9Y>RHQrtxj?sHg-)Fg}tk4O^g-V8tRP_CSP*C8jq+r2p>(*boZn$ z8>~;HXS9wHRH@7q5{zt1xqT`mh!Z+)tlC8BHaB4A16%bZpWV|toa zqdi;e{627Cw|-3b<>thD?^+LFaqIo6 zJsTT@D1e(lAoa~dGUVF-#HRbp18&)n^5{<+*_qsG4hrg#4c$dEqxBFLCF4cm=DZIT zkmx}&^r@kR`;LHt ziBIb?m7Z}{0X`^afl8h(;*B-G%)9YNcBHN+A-^G7IJ zFAV{AY5U_XK$`mn`|Dq=KPq^}oxDXJ)TY^s;9%ZwLYsAB4uBKzOvkKmGGuQ;{8pT$Y8#~$2hY|2-zf1 z74%8nGrk6R*;BNI#m~Up86kl!O?r_S@|%1tfC6V$ST5LkW?WwJ?b-ThB_Y zj!cHrGk>ZeRpvFtY5w$UNs5B{ic@j$`##~%okMmEhO+yQAdrHppelU~n!UJKV{uh^ z^9dI-^E57>j~N03Xie9BPje=KsFRBdCa9}xg?7l1M7k)y7-v^m?EgBbyAH;VACxxys0bYm9() zKi7STVrr(eeF!*+)dXep8?|Fw!Mj+aSy)G2fQCq6d8r(fY*PG_f{B4S_xKy}a5+Ro z*GU|iW`h@``q6ODknbB)59r*3VNjQpxOqExk4Pv)ojA)v6}0qx5|QgQqrV<(UH1y4 z&s$OKMF>E?@k3F)iAr~=TxVGM7)Gj~RkL8GTmR!fLKJ^gm3HOgermub#`2OII~(r6 z#OwZ0=%@a3Vc3hO+lu#R{L~{4(;lnUziBG$m5$wl@B3F{yL|ys>vUMG&y;+8BwEaM zRD0#^RKwdyh}yl)SedxkytHN{cV+=^8F6PyCpk&3lf*32ABlJ!|432rjtHk~P}cR8 ztdjQ3s8n6j*VIxreEl3@-5gY%9d>$7khov8kKl&Gf0IO4Zsp~mZI?6c5g_bz^g6t1 z29Y&_(Le%JI}4gL&KJjwsuTQAEc!l2V7_)K**NMJR_aP;rf4;OS-++=u7Did3no;l zyggnODHOs#dPg83VZ51Qg0j?Hz3ctq1!i;c``YaH>tuGs5B#n{C;&^a%*v_wcG5P& zV?#+?q(hY=Cn$%%ue%zEoqOVFg?;4e@zJNIeGlx({I?3)r(&j8OV{__tyKu%(;ACB zK6X<9#%b-K^kyyhB~kyA<>krcr5f+MG`flJ9aEjJ+Cd=ze4gdm8cu_=bK))cF`!wB z5h-OiIskSekTb}JdFih+eEjN+_?PKldit(x#Z0_V79koxXFs&4uUy4XYY!N6HZ>SH zn_qwS+QOz|)mrU-a`AoGja500{JxckO|iV}H2^a&l zuu=$E@)OHwx_Xf^o3r};2u;ougmV4y8`q}8!!2RQ)8s#odDckCluyEB*4}>mzzcLo zVIFppG5?5FvEVum9ri^j>YhW-Lam^kVfx^me_AdA+|6E&<9-S0%6KQ=yR`a*t!U`R z*&l1THZ^-cSsSEs>JP9`NIDk$>Q|q60oxHc$VgHsB>|@SxQ=8T*X{l&Us8NQ$M=ED zykv1~L8tcG9ECPjgU7i}pP$$Rudcai637iy-%7r|IZ2&jy zX)^2F)Omifn1TFdGdAu}KtyNs%?4XaO3(G!Szga3RwMMyWNTN;$d1Rz->#5p!A4K8 z)3S|-(IqS3G%1YSOJB({Z6t0=WR2+sLE(6*Al=uSx~1bQ8r?~4GOVmIY>YUv^sF-V z1}8?UenYV~2zW6Rp}B4(Z5PC2rks2!>hXz0N)gO6A4G|+FX(OW^3h~p-wcOfk^!Km zAV`|mttmB=RkTf&gRam+)TkOv`%jO7Xwk+OY$H$@q>?$ImNBVZ{v|?2fh75t5oje* z?<+&jWFlhHV)tvJwCv|x`dD6m(!66B7%ZQHu|Sg+T+X*FK2tBVK6(3+dZ=U%KK&07 z3tZEEve+K~I5slJyu8fbtn0y{lTY`tcGvpd%EqE<*kw{{#EYC&h}?uY5d}>RhjmU@ zZAun*wZNg{$U4nzSZ*x=EiuiiFPwntMIl~jwd=4JGBFwOynAH@yd2Q_Qmr@@+h><` z#}-)%Rm?Edft_C?o|_5$MtKj9k+}?xzU7pzsu+N|rzIqW3ShxvM!$2GLtk6;Vxv`Z zYaWTJhzWz5(F9si8P_``8e^$G?kx=1nDzUMwai| zw^`jLS)S+Ntq~CscU~P&$0IPe*CcX4DNtyqtE@RUY6gigb|8aK1$^&$F-$?{0$T=A z;2ME@L6o_2WP%J5oO#osna9GR%D`otw6ZQa1OvkS?`xT{*1UkzD$jf=6}m_f z-_Jg5M6?Aq3K8lWu7$f`?JA67cXV{ zYo`sv*Imrdw^FkTs~?+~Z58yk!;B#q*9C@VsTKrO3$@~()u-xc84e+6=?`+ZhPJ3| za%RoZdv8wu#eEaC?|1mu@1&r}l25S%Hho7NPJ0JDeDP7;`2Jum%a3saoHmy-6Tce3 zKex0xI{Ysz#@d7g(uxGKt5_+kcdDrMyvUCw8$HQtXvX&CN8-}@f#uIN4OboS5@Ecb z+N=-5m;zUe^t?RsOCldHA>(H@*IL!Kb9K_-y{oUgmnGS2opt}A0Y!RkvhS8 zWqMlKK9wSWIU+vSBq+=igz6E zAAi$xnfF`CkEA2|{U05gJUyT^)aHNjUkLc{JrrwVqs^iMHcvoEF+{4`8G13pyYW$R z`abNV+hP{kz1o_{`l{l{LtXy)EdM{^W70H|KYtNddxG3_W6ce3AEu(tDE?hnu0?j= z{qKB27hlx3^#8K}k0%iXJ<>)24Kd%YZ_9U%R$4owRD6plfY-l<1Alu!iswaScf-_7nf)0e&LGSV}W9sxithShE47pzDT%Z|m72sfSc?CFto z$K%LVevyF#y@6^r4HB!D?q5U<$83;edIrCo7-L9c33HvqavM18hCY@x{orZyckH$U zNam;ES@9g~Lvq+LzYnQ5ki)Ty`x5OhMmykHE2n*33fAaj)mRxWIF%ajyuTtKy0e&v zp6eD%u5S4NA^fP|y82~L)#S3@)ag~o3oa_H024l8v>9PO8blp>vfoN0B^ zr0rA^nIc+0Ppp&gVP9rA9l0KFh~|9V+c*gSuMJv;aHnM?WqJGR>CCiLE9&jF;98UD zR*rDSea+KyQR(K@M(()L>4}|^<{V|3(o^Xe(HL^47ot&!l*7y4{+IinlRb2Bht0*# zdBD?ztu!kondO$S$X=SX0-L6iOyAy&EfNmm4COJ~Mk^6`zfUjwuZR&VV+5{&Qx=6H zmD%NEPEvR@s@-3REgrpw>$9g`iX0csWGQC5Dr)lx>K{0)g4r@+Ff#_cOJrbCeEf#> z^WRzGIEynOp@SvsLBN>`8a68+x2w|+j{H%kqn=3ulL#ZZ_%IY4h!(jVeY0uy%{6_r!zi92 z{mzt7|Gi%+mY#C%X0Ge_69xTDFFRoqIB^wSh*zpMZxDsraMmCNV$CkX=cd>eyfXGOI9wHOnq8s2wJ^$o zAjj4|-=LUaS*y1Sm3$=Atg^!gM*@b@hHo^i$&cz3lcIj*l5}5Y-669xOBvjvP6K6? zvAEJLO7KyX6{QUrPz?+eWhNmS+|Z^aOCK zTp+Tno0_brXHpD(HZG3G6g8#Xgr8W;P#Sn01eUDTgSZE&lOgfUPB4TMU)ja+VZyx8 zqClKWW8|9L=Tx+A8usxLo9)0U3OQz+8s)F;Prz6!)_N0}1j^DqSm;fVdgr7?fI)Zq!`eCjvqoV|crg!w zB4c#0yj^JOMaaNyl8vQmbV3q6J}X`-Ya)pqkH;P=q{<9#1Ez0!sl&OFjQ#3y!$rN`bdVLig4lK}4_&gYv;yr2FywQy8yz z8Q;f#Keb74ESIauTN=*KdnBDG9P+dx6k~|TA5N=gf9KR*6Wf76rSQvM+XQIxgyJn6 z4GH0e3LLJwOvc;5KE)OaV9vgX{N#Q6%{0KW$xlE=(c63RJc64WAULlsq=Oj3X4|g4 z$4t$~Lg%7+Qh5;NlH=WVx#zNPYiiza*MOms(tP$S8d&!g4 zlys+M!;|4j0?7Io1Eu(MCIFWss%J18Qlb%r(hYw<(vTOO*Rl@OBWrg$r~<1hcbBv$}{spH#k9hI@A<`)7_dooJ4%=%D3{#uuf1cGb5@hnc!itHDP>G%p7Ef z0TCF{lZi~qmp;kQ5f5LwBWolg4>aKy`O#icm%r7pH<=O?qcSp~D&dhe{U_uF!K}t$ z@2Ss-Hg%$JKe$|OC^QH4Stv*;o0V_mD(g9pCPW9ri5RI?a!-c58Ew0iq8RhZnJ|F4 z`RGTBW>t+2>{@vK%%@865ricqFuJQ!V8syl&@5x)SSWDIioJ3d` zEvopJR#5sts)zF)2Fi%+|m~D$z{icWs zXngVySD9X9#c-AlciX_IT0P|X5MjuuRK87y^2A9Z4H*r{PpxKFsJwmdg!$>Lj5|44 z%OglRUk`$?F#o-4+&5%7P>iV80VOUtkfAY=D(*5gL;0~6_W9^K6RAev(WA0*qy-Fm z!=l|kN=P+}(>1WN`q|a^(X0x%m*)Ktv32@tA?THA=x5ZbeV=&8gr6ZiHtv zYHq>EYS%_%j7C{O0jsM|9u{>zWUOzm(Mdjmu%w*KPb>f@1{X^@mODzbN=FktRE}}_ zi{Zt>x?z^28&t{Pg(M?2Yu!h}aVz*5k7TE`tLA?MI+q(Q=ckLgu(e7KTP4M)cZK*H zV9x_a4$R7Mc`;uIk~xu>B^LyE3B-MgP)d@~b#b(9X?^NYs44nB8&oopjH-PmZp$m+ zY5zy^>Z0<{&OM`pKMnpaOILJo&7?$aF^&($_!#b*7g10a-pR{ieiLfvb6SVF)v}Q~ zZguX)C_iqDD`r8}voJQk%wJ6vvKz7pu4UK#v?-Yy6353oVdFfv^~_F6|407OzHR~f zM;Pr!_u%gq-P_G8m+pY2S|Ln~j*Hc1kEiStlfv?Zcu=BvTCwbt*|InxZ-;D(M%u7wWLQPxQzLH;-Vr{y^Njfv}yFo5}ENSI%xwLPD9fCx%K9M7=yz> z=VX*u_XG~L%S*vX{GW?$d8Wfcl&n$uZS-Ti ziQ`|2sOW19^Xl#gVvpv+k4cF%^J~q&)V>u}^r;es<;e6GMyEwD90qZvCNF}RV(>!< z>pb@qPBNbcw)>oSO=Q$`6nW%^y-p4L$YzIN4qVKNp;jhEp zJ-Vdekzt39vH7LvL37;lq)vgm#_r%csNPPsA@nN@*G)1c!*!X2R#V|7_m>lyWhewYrI%T1mQV8d*^N7XcewQHnJt5e z^ih>{rx8y!pAHENQHaU0TsU6tvXZ&^Xa?`BcVgN-90VZ4GHj;Sn%3s0y$LYvd?VVS9!F9ban76&c4xd|(2m46@|mvE0RF z5ax-U3-QdTCV={Av@yTqEuhqo&40!N3p1t`xR(Vi@#W!BQJ|ubSzK~j6 z1ouz|tHQf-3!@m?ut-IU#1a5nWUak4I!kZkSCUJ>zFFPPP1%8c_B`K`{0(->5Ft^C z%|-q=OtbjnOOH>u*PQ2{D8VyES9H7fI6}6W6IjgTvNHlGyWrhL8m?>+F$=i15(#A- znZk74FW11GlAW-aP0+|s&w2)7E#*`hHH3F=C$=Y^ogp&eK~>Hv5(=<@m7sFFN~U6U z;f#=S6$5BX%t0<0q0(hRd+b1PN|_ph8R|d+$uI&z^8j;5gpFkxB$gKL_Wt>*3ufxs zjpisl37Ej#q;&?++$6PU=C4=OR+*v#MSCLJL&k+Z^_hbLUra z8HmVP8ts+whcKuC%xOzR(}nG1BCr`nHoWc=2!7EG8YGk;YA)k;_9IEmMzyh($C#;+ z{LkyNoY}`C70AFCdOJ zr*jMk1)L_lA6SGG-yUKEs9}N9(yGKkpD`UCJr15(CDc(j+KgNRO7{-_4lAthM#tfB zetf-s%@Xr9e^YELr-L=Taf&&nw=Gt!OANl31i^>7)O+%Xv>k0-&{j?neGI7OiBph z@o>+eXAzC;-&3-Uj*W{2MH+glB3#O9)&uL@@w1V?+FTuC<<~!QvlSm4Twc61#e8ZH z&cflPv=Dq|A6RuN+$R;(UO3)P z`StbuEw}q+wcFCmD_#HR=NVq*FM>#6rdzKsPtWU{FI4czq`%WbYw-XV=G#K{rKB-!6`v(Im0RiX1>@yOGa{1aKGB*i~%jDKlZ?bl)TY@{M^#f*ID zGS$Gt3Esp!+C!E9m;So}KWBPY^U%i|R7)kSW+O`j*4YaO6@I6n)b6w-70KRk_5U(h zIEQrt;QpcLNrLq>9tFXG=NztqVJ^)c1~>SYm{W4F>r?iMX=#o%u@gaCdQPe;Z}$!E znA`#ud3p5fV+vh>xQPAb6w_46-8>R$VoNJ)OI^kN-*s>er3j4zsR8VQIY>tRS4vi) zrm}z@$4@CM_qhmnkoCbt&8zLb& zjVY@00#Y0&3V#4U$s_I34F}1%DQV#ZDh7iB(NV=H@DS!SC3`xTjGBn003x#;RoT6b z!e!>KviBMD984i=1q=)<5EhpaA_i=BC0^l zSg*&1jo?2-lo9X|A1_}K7eI1b91{O+PtG}xG=Lu<^G!7~=b7*}?Vty9fsxpYC6DR4kqNrqe@j7DPfO8^y~Bn| zmKao$1W#MqRxrW4cT6O{t)gz$mIpbC<*GhJ2okdYA)dP`WWYX=0!qy|Qwb(i2wGer z_NDR4*D%B>;bW5q@P}gAEPO+Y_u56`=yAAapMDM{bX!;a=1D8hQ1S;#e{zw7w{=sQ zlEmidiYJb-_)A5Q2xZks;r?Jpfh4Gz5qkAGp@ZM>@q*X~D#Zv`lGj5(S{oBE8IJ_$ z7IWPvDJlPQDy#_I*XWuMo+WkYA|5qsGm{#Yo1xwTR6BOy7gb(T#B(zf{rCsLY7(Z) zV?JXuYXK#ZQLshF_Btf3Iu8GaOR#`_6A*qK$%2_d);B)@w4~}_MoT5wklfBGt!CUk z0_%`gpHImHFXvF~r>dJBib8M_9nV|VJcYgehbk`C)|J=bT(yNWs)OH$DpqKK zCC}WMU@!WtNP#NC1wUjQOEu@;Le;cg5*Lj0Y~10im?%@nA; zv;XL?=`P2*4puqYeXb9lmtW>(P6UmFRN<|*yoL%QxYTjZ2kY5=R*CtSW30oLtC_`Y zj{$OS9NRpe%iom~6WcwLqsQTpEBJL|iD8@g;I0+x*~TJs5$vzvMC3ho<2%mKyPZ19zL-+Ig7Z9 z%Rx61W4)is*|$a(LeF_(7Hz9N2=|%&(M?8e53Su; zvU~Qb@9*Eg>E!xwo_H|L^H-YZEfnL&r`J7|SB4t1?*G$l#6E_P&N}DtD2QrQPJeJh z4egBD>-N^ewdbwIW7i*fwB4M}7h^cS4}{%KFTkqcM|0LgBX_6N9?s&L&>dy9mH(68%bGGAE;PPy1@Hdm3H$eLZ5{MRwJ)aq zcRGnlvOc0IP@5G1vm&Bn7N5qYtYKGPzC5t_Oi}DyXbwuJWOpkl1;u@KpEx;`J1t6~ zaqf4FOUJ)Xv(kgj25Ib+3Pmc;rB~fE8lC_nY@5#ZOPsy90wU*dM~4 z5{e~#{z|+4$ro_vq;11+I@KAe&2UdMjk5x`6L}+l5o#d@uO`1aTWbhHm3Hh{B z8-mAnq86P_h3_60tlt*{vBH0lfeh-Fn>AuBw9+%pKh~*OzV;Iqk01EPID5kX(k29e zsS`&_^W$mbXjNxc;eOc9ct30{GJBV{I+iNiiYR7BTKPy0BIHPSv_p|zoVk%EhBWSU zGV;(9sW4%?*gKyXHCrCWctvzH;9hw778P*$d4W-NW*3c?K+iJdo85s9Ndu63$lx>5 z0Pr?DsPGp8JD-|F0{Qv-mtSgRvA+{fevk=x*~{kLKe$9kf74AI&uW(X)QzKVh{qLC zR#q%46P?7k!V-iFr$~89vnYd|Y8MV}%3&9Tt&H{%&jXN0$Ov^cYkm11iadC9RD(CX zIzf_-DVY-;txZ~9nE~%PBbFb^v76*$whgS3V340VxKqjWUn_p}KIMC~ePW6U$5+ox z9Xa8F;4UQNhq+)YBor|F%HZ3Y4I&rKM&NS0(lb@k;QmH8wPN57US?8o zk0W)1a)&cw@}uHYXO_T!16+`pMo9(-P3m}8&fI5 zpb`I=oq=Q`VEQ9t;#`S}{c4D`qC$*n3g73`2m}T^POKH)qN9jj{55%eorP zNe0Q<%w>CY2JXNTso<2LEw?e=Vwe>P_b!ICD=MQrK%!RuT|LY)c(Eyga)yzPn?@k< zj-E-*P#!l@7a(Y$)fixO^dl;$jg`#MqBgU^F-4Nuna<2Sqn3<-SbAiF@KX{Y0u(g` zL}1RJ{h)VjeTOZXmM{dnip5s~4}PJS7h|k`4;-A>ZVveL`-3tx>d3GQ8~1{4m&upn zIC$ULG(+3D`8WdT(nako`imE9To#Yo^PabU*?>-YQH;4lUUml!QRZ&rvF=YSn@&#aRlbCk_c8CpNXR%6KKnI0C+}N^^vz$)L6WPSf|p zSm->9XLum zbr~F9p`^Sv8GPd4_i0r3Oo8gi6>a4>_r@15e1yzS_TWbptW}BRvwYJiMsFF z)yZ3dgz24d+T99>>wo&N!m+81v0;wyIceCtwa{vCg9K`c3&9!=$lk{n4Xv*#9*yb*W_x>*F2C=dPdW*zo6~+*iUTvd z?Hx_5@M}(OjldNT3yR~`yQ*T#ya?6<3y*z2nOVVZsGY`6-^2R+wtJ?EhFI$UIwLea7Cp<~guOHWarz zC!TM4Aoj^QxZVb+C}nt130QoFUBT~Pg6P9-mEGHJ|eTVGPvSfL-7#Uiq3Tk98^Yjyv4>zU;=%6vofC9vMgW0WMMr5NYEUj` z+lMH@n%JW=cU*bkxl@W4aS(gr>#!n4{+!|yUk6-5&Deo$yyb`}1PTV0$qp_CJT*Fe z=vK$A266?f>+@F9ILDU1Z<+xXAik6yJ*qKF zapAs7SP}BPh1ubN^Hn4-*;DJitaY3-0*zaF`apgJB7?#rgHz;2GN3DJuM{bKtLGLq zhnYNySHAdq$J9Ay6z~!elWAFHk@F%0kXOPq$1f@`Q;$&JHi*Ch6v5*mh>i); z#G2m}lR!)i_%!&?KOJJDbbSePH7be=A>tJ=yAc&r2bS)*2w!XzROaScQa~YFrL!!! zGl1euV!fS)lhhOhUH;4_D*8}RQAUI$t-p~V5FYo)3;YJkY8T?e$#sa3tS1xhMGE@G zVtE{nm+v}eP*zHcmyCxun;<#j`E$ho61I{?rVav;HNmuw#}LjnjWicaj+YuQC7l3pe%%Z(~3>Jtd?^2pICuNehopLy1o6$A3UoypEq(u!g^ip z$y(e5{ifECT29&Yv%Mca{1fwsj+g9hFNcjUh(5K$vChgAFnDk zCeRw6-wODd}j+<`JY*$ z^H;ZajuRD0fQefA^emOJCd!W}8skRcF3Je6*~97zV|`6Z4n=rtB-dac zYpHChd-DHr)!k$UdES#STK{rva+&)1YjbB|wS61^)t=qf3hVh|C3lBwBkcgzu(rPM z$+98uG!#R!=WnIO`mMiny##)%XnIF|!}n|F^U*MY&(%wLB$eTb4Iaa~F*A?@s_ACvq#g9av#flx{NC{Onh(cS3?N5}1+hA4w64<$yLrzv!^&>V0<+Ve1QG zy?5{}SxwUn0ivd*i!FE@lutA_qp~J_Dk6SW{87H#j9HaT$%sEAm)^x={`TrmuP$c; z_@okpLugoN*t)~-PuU6B3p<^K(c2wl5-{EWAwDUpL={tnrI5vPuyS6W$Y%H*RF0Le zE?yc455~#oD{P=~6Dysz)Z9ltoAGHZSGAP50c6v2C(6f2so}FeAtY zE45b2k0LPpIs}!lvZPu$8U2a}@o3A9Z?I8}E_zb@JAe-iw|B*@=>@zLp{m~%ZCC8^n_2jC=z+A49n;*Urtc$Z8QNf$Ar%q# z2^i9U`cz4{Vsd%$a6~_zlOadne$zF~caOwUm(;V#T8_fyx#MfYqxsJAAq9xWA%4W2 zL@Of+@xhXfur&~q9ATTqtU=NkZII-x@0FwExz;548e~!$NbmLs6az+r$g-kp0!EcI zU^V0cvZ3u?S$gT$q zOy%N&i0XAZkgt-24Yq0w6dno?TzVjKpK~bdpBQ6QfZNwl2I)wq>>9||GFJcYNhAZ- zgq^v~Wb;i0rAVXA;P?So*{GnSE3iH>vdXJ2eZvg9awIN3BTsZ|?5@TDe=*eDic&EV zYaG`#F$+)igAOl}6mA5bFG@s4^iYf>3D2{jMKbgVZ5%HVVtK79UYjQ9VyIq=c%A@N zKFAp&a8XO{lW(L_AkDQpN@jKaD3~0Cre?(?fN!|3_7{X{W4g4tZb=G{SsdB->>-s= zhm-l~3l$M_03pB$oXT-5RtmJvx>4HUkk_Nv)sWG2)oq~=aBT{Mr!{3+6jPdw6> zzA7WU1exW~GpPcD_{$1Ljv_MDM@9d{gyX)WVea_aAaBm~I|dG(k_|~A10}j4n4k zdM)tp8u44&eCT6NKj@jsGG#9wzJJ7kD%X`W^;aUw?p9l9}f6U4tMOqnsdi* zt#%wfF?yu#=`Uib9FXS`o|x|0{Qw(UyLsKg_f15y|IfaOi|v@wSpR{ zW={LOc3XoaK8#u|Zqsp+ZWmWh?dA-ZDNO}M=g3*^7;V zY%iR{zAqAtg_@tCjLCu0#aV(wT3FkDFxRsstGN&|U311)NowR4hSRvaDkh^56xD~S~Ja;)Sb zRqb8*PJGHqrnQe5s+v*P*&^@a7orGxrUlIP?6c6Rjn7*AOmrr!bn2a`6}ljtDHu_(^Ph6;NCR zHFUEp%mDG|C|QyN8$(;qb@8aU6z(Ptfm8)Mv(03>#%OMGR!xb_6zM{`(Ke;LYyTK3 zlRau{W21I?ann>fMDX~PH8}Os3#b4FVL1G!%^gH*+5#Dkz4X$RiD!u|;90m)N(2Db`Rul`-%q?ve zH#aAuZb>2!l*B+P?tj|0UOH-`(lJAcdF_-=&oar&29Ut5_Iom#p%RLWJMV}DB>BsU z~$;qqui!eAteD-HUqVs3Igk>DmaFTSi0 zwWxhh6=udJY#rmI?p~CnYuu00amD{RKYtnB6_Z#0gU)iU%v4@kT};} ziyfM3P8`g>cAWAqvhpxxM-<2OKfQ{Q{A%819GgOrY<`GjNbYKMRIqOU@1b2cX<+Yh zuYFt#2uCb?z|r_g zFnmw>A~tzo?D=$1%ZQ|z`rr`cbj7r&r>#dVNB3xCrsxKD>HDwx@5*6_Pg6Vt^|j;% z+M-^m5)gF=zMJ+1dmDg$q&LZpMKIk$brx3|DDoRpPtF~H6J7DWwrlcvv^6Ez@SIIo zgOY>Tp+?9cNogeu#+?o)ME-;8GiCMeBk$knOtPSwZ3`>!la}pXwaLu@uKjaS;TNrI zID;`n9|a4i<|D_#;CGMwm}|{N1oq7x`sNtP{EXWc~ex~ z%X{D9&fQt|+H%_LCmN!njG3sCh028vF4JySfb{mS!;49*K$+N5+W*=`f3;p! z9~)Y&_=^$LloM_stw3gcXs?-*7Xqq#Nnp~T5~MlV?ld3l!KGIf8cMgjS|zLv_r#p0 zn`#4NrKYV7$H|=FESilR&|BP7q8ReJ1N(_6x6|aEtb9ZrrMErRqW-u(^X8^Sr-A;0 z)XqI)aa6Z*LvtT<@9%|e$;U#B!;YD}?hRV_1ZTf@EkLs`BoH;Q845 zXHC&~tmQJ}r-Pa=mwRS(7A-=K-mVSi>pp*J0w1pXZ6}R8-TE4S=VqVvY@|sxV)CYL zE%09dt=oCKzZ{K@e(3f&kHB@h-y6GJehAyuRs8fmy*)SlFfCX&cqZxvcNr&bH}X5* z#jpchBL= z%DxBjDsG%OtmtOC?bc;*oHdg9MQ;un&J>eDHM>B&t5mw1f<4uwU8C?0ji~PB+m{YY zBQ()3u2Q&s(aSeT#vDUQ^^-lmi*xNG)tpu{m(p}i2Mr%kWpIF$G(6GnCD(=q_o>;t<_lI!P>3gx8ma|#eNmi$#J4eGhQVQ!J zg@PuGmwf^2RO=fq6~C;VW)w03Vknqdi48y%TZN~*UsW5s5UiqBF?{}?0n(VvEb8fToPxg{lJ|60e1Bejw` zLkM%7D8Zxh@4@+vvPV~>L6}|idtXVH^Pvj!*OFC7(O_j$?n8@FA2z=00EM9nato{! zcTb?(_+FiA6@w3BSyYT<8L)^?TIjPn6Ej8~16{OBYLw}o->);br>u&KC_K3Rn1_yj zuZ6}{K>!06Kr}Bw`OpfL!Y(oaU|HlZtpaDI5ff90hURa&irjK2#K{OvVx&Xlow0$I z+fmv?qJlxU^xZ@S386@oilZ>vbnXTNChnx!7MSaN)lf~5#Q$5%H+wA(SG`FE!KeI*) zU}Bq$#T4TG+UUiu?Pf+IbHdX*|~l z(UF+~)Z(}wS!(wET!=-sXd8Iuv>1wwcEbJm_gn}0@A1D?laElR6X}wzaV5AiluAR= z_IMIO<<*2i(8};x)cjwuKn!X+b18-rX+wc|&FIfP2R(z6VL46i0R=N;?vt+S5UIlk z;Ob;nFH`+2%Q|ji1S#!dkaVJzg^fD&RNAa@piJ&xBSVSC7*&)@oL_qQk_8mPDqbQK zCzqf51O2V1aQ;pO+Skvsj6TuFos23^Pg6TCVP)YGW&_;746pmP!w>c5_5^p>7;=G7 z?aI6NMQI$$Wg$KEwbW^=xhEcbcCGvFGu$HX&>5DzUJ3Cb^RoN&Qm&fp7)BBp^*hhnSaU~6AdgB zqv-^BZ~njMIW!_E%y;s27$(PmNJAia=T(%{BfsW{_T!&rRC$VNv4&uaLeFojt!Tx_ z>qJr>2*%2oGp}PypaJ@TfxHE`9z@iU;e2z3G9XlQfpl;)z`M_^e*1bnH#+A%KDS>G zmVfgGo*U@cI`w}j*&+eGi;CaQ3l(E%o&GlSg=~s7)ws-s69cP+1 z$x_&W{|fXzHUVn3DpWvplfT8H5vOGTjaw6*oIPrHak|!HeW*_jo9yoBO-{-fue<+4 z)MsWUm`Sdcv%kwMa6|n!cPe9kU+}rzOJxZs_jaS6M!2FvRdN<5PB0zzY3sCX-%R9L zE7)rl<@KR~@$F(Xtx6k}h-=7M%f68?;}ZGdRc$i4iP^LzfOoKT!tStTrO|+6QMi2> z*FtfEZ%8RalfpCqy}EZa5Ch6_Sfmw&J07d0APsm9|-PL_(r znv(rpL@Ah|3>JZP%L79*mMkUK3>((Zbk12n5i`pCUL9GkJn+p~@<{vP5e{eBk`r4a zSHW9i%dSs0W5N2@ecv0YX$U~FY3;c}%tXTt#AXAbBP1mHueY;i4f%Wo|ER+l648_5 zQFR52L{ZFeM}C=>gi8a*m)x*z)Xe{&=>-$pN>4&8Do4mt#)?QUL3ly*NbwbRW@7_; z3B9qhh6rk@MP}5c)5C`n<$dk3S(M=QV#eY_FZPd&+Dt3wcrlL-z-r|A59ttfSQ#x`n-LP=P3cy;wXld26UM3~g-F|H*dB%T{5ngd!;>=M$7>{YU^%cI35x0QK{fv) zB&fJPPk+Z}*7&^|kvg_wrH}uUNIiliQw1{MJE;y%j7MS?@>q?RH~VT|&r$8}MM3QW8RlR?J(UiQ8W9oDjyR(NMRCQ^acdO_vPH zdT4SZGLp4*+Jc&{I{Up81IoHc0%nBPTGoYVKwjk|aM9z$$z0vNR{_eC4p5i3z5i(F zrMj)p=~p>+X2S*>B@U&Xo|67fWwPk$6<<^jw~3FDH`2u(#f^Lgh2Ft{n~ULN$G&v}h|q^$-V%SIlJ znv#m7W-_S7-se`wNB`PgKAK%2h~d9{CEP7QCYv2uATGY2dg|eDD6^9qv-I7qP5ZiI zGgIr9c)JnMzZKvQ`XS?=$glZr{Z$5mTyFgSY{zC{j^EsL^RsrAZL;q(myTowL| zw_%;^RjPx>6+c>xR!2L09N+QF@a*=j-zQP=^!Gip{S&1xSNxS^%i6for+$)hJ}UNF z!>|aHNV6eiCZBan+Y%sb+WzwtWq4G>wc!JX3kuUkp`RS(x0gQQ|gz8+drcn z2j83ECp~j?{dD!uw9@f!leotQeepEkAQC?;LCNu1eFxQya&uSYF=;Glz^;u!56J$? zqYBz!?TXOr!!=QSdTz9#G^&icZ4rxn(BCl^=d*{MibE8ylIJ5LNRBZR01<`Y8f#)G zFRg2!p>tuAxYfV<%J(oG4qYbD>&)%_kwRM{?LB$@w7MtR_t@9%6!W+uIZn|zYeE} zCrvT=vnoIs5u2zp7Lt<7w_lfI3Uw94ieY#bOsS7UGI?rA-Cqt9ygo7i?>7C6{7-7K z8Rs-CX(M|+Y88dOLCcMo&!0E>d510~W&eXs>z%=8VSY&50iQ99vwLXKP|429QAP~$ zBMB-9f9*kH`Tu$W2u_h@in*;Z`f>geZsc9p8MJYJpdxk zlpgiwsi5(#jk?P-g}~t45SOEpsq_N8s?R;c{yTvQ*F2ZU4SYB?D55QAx*dB%{hJ^fJ@m}+0URKRk!IAJecR#jip!VILhdYuAG%8DTDx{lAI$%FeizB&a@;y*b!~`l$vAQ- zna+)n1_saIuk&Lr;F-}yHk0N9Xbn`>r&Jvk8%3t{r6zkPOR9Mq>tnSQT3 zwC~o)#U1I!I%KJpM9Ic+$MO!WXW$%Bfdm-;^2^e=NRT#(zEnP~wJ19Eo$1&Yob?P$ zEqE5=MiNZ!Ac(!Lm2F^j57ma1QB4w0W{4M^MEGd~uYv7!hfpEKqE#yhqR6DVO)Z23 zA%84VH5_)}f6uI)NA_)lrX0vgp))}nXryxyF_Rw6?d33~s96rB^CYQfqntU{ir zG62e_f=J}x5@~qAeTR4~9O(%q&tB*u8_7_xQ)mLJVR8f0q2N{4QzCN?p4r|=1-b@V ztab^Ljyz($Dg*E02^DUupHAHmrc=i)2P2m!3us1mCxe+yOb>P%X%7-^g~SOtfJG-$ zXDTOTfK=Bgn6Q|I1L6ve*o)kV?a;=mkYzPYYLiS3m54yhXD&5cE6l3zxqf~!v}HBi zHVprS2f+Aa1u8#oOM@?!oVD->`1|kMcnofdbT&vMLIlbxWHXS@73)kO=J~ctdz1+b zdu}O;Q>b|5MUo5XndoNn<>(1oc6|I%z3yN8fpEB;dQ!#>^lx1EAvzx`d7uiYJ6S4yGJ%kzznU}{m>fBddZca@Lw zV(!o9G!rdCW+U(jF23TgGxC?6v0Nuo3NLZdcegY?kSja;hnlyxon6InexmdEVt-@J zvCNano*IsSr>AKfGG8A5&d_MqU7ag7rI9i&VdCAt$~N8F3UeAwf3J!qC-KDZkj2S~ z-`87t#>Uhk!Ot`P3pU27Z2O8q20(1$puA|8V|U7F41`8gIo&w@5I{bMpZ63m5>3Ye zz{o#}P;~`R9KPn-M;jxc?t@eLuI^xgn(+F^&s1N;;*xRD4&Teog)M(5F?dU{Xp~lz zO};x3Qv_ch)n|Mg?~0jHEGG@J>O27(`}yLun+OWduLE}}*jm?j znaD0UCx;uK~e-mtBZNy zi}LxrTxqiJ#sXLtx^r(b)iqQ31l>{xeoIM`{dVS6#_PJsUvLQEfBy?@~*9i(ehljYdm5q*-dv z5Mi{@esoQXXk%B<2jcoDwU?;=Lbp@eMKH!%MZFr?OC6-e^4yz`;=IH{ZJa^n0RYfI zspS|v?=}HR!DR``(Gig;(RD)nMX8cNhRhIrTgqBa%IRND5+^|pBh==+U`xsR2$+JV zx>305NCjOeFSnH&P*{->O9}HkmL$-Xd{{;<&M8t<);YVcq>4Nwi&jQ6YCD8pU$OZ^ z*};Iz8%E)%ZVIo>W@>m;B8<@HI%qmu&Z6&L43&>5%boACe{P5CJ z7$6vbJo~mCB-NFVAHQ`rN#`k6=u^yRq(H&x_@jDI``8(@&|zHM3K<3EbA(h1!Hj7I zVBLvGut#won+_p%Jw0>AJduZ?sbS9}UKKa-;iv)Wa_YQPkJqZDC)%U=HupN>%)8;;-io_FKPQCcU3 z;Z(wNMjfrX1ix1kJ4_3}D4XXsW9RvFePgt4LK}jjO;&)Ta5Wg2pD|g_dJssQnSc*r z4N)0^jf^HsF<=Tw5~}{DRRrXZ6gT2&+P#e=k=%w>aKY#;6Xg*0Vx~iX!$t+qTMT?U ze#5B9!@Bl^=aBX$Y`ZPY4M8Ev$hy0d3b@f4_rrA2W@#8|aSwuajK$((N>v}EuQ>Zo zfY$#y`)!U`V^xbgDjoE!oZfeT>Db$jwfqQN^jqKZ@o(Ct z;cG27VNDh^JxfoiUfa$nCwLzwXV-x*MQ0wMcv*G$@xKIDnb3SRIil?so&bTpbPUe=6=|fDYs%$;Y+9G7ZY|Q`JdvLDF9BdJ1MGzOy3Qvpa zT)uGl-4Gk(J6pgyxS>$ZpX(=wwPM3s2+rVzi?H4OTta~|yhCSety!)bI^cjn^TUs} zfF;*?Yy-!~O}mT7gk{X!6b#87t0wCi^rDf$pO4siT9DpG{!<4Fp0oi?nXETh`&x^7+Eifi%U!Ci`n;8ubZEw06- zXrZ`UNO39dQrsc9Q=o<7TBNvZ;mb45dB@2AWF)!w+H0+A&dHoIv3tjJ?VZ}Gwtc9Ts|u^Pa+-RnZd!`+}25^)%+;4z{x`TemR&Y zH5canv{Q{H-7Rek@eT-Fy*#*&u}*67|WCJsxd072FXDFLg6-c`3=OJjt0HFjBFw-F85m#tFK0RI z)Oo9t6yQQut}kObM`%*G z02QRCq#|ZGRn8lcTnJa?<5Y~WRn{;pf+ z85R;)z1=&LqSuyD4i-eu$DO$WcgV@(qB3QuEEG}DV{hp{rOW#UmFi&6fpsu}O5GgM zAFQz}g8PcOifv2c(y7EK;doFUg?#TF*I+^79V>vy zQ6Ds{QD0C&lYTnNW)EqFP}R1FmKmRjMkR zso2_gO^D5a8$sEId+|^p<@I$e%(}XTV2k}vK7`Q})TCpA%v~91z}NH)DyofaIK*Oj zp3BUJK!XN^TtQ=d9>>55XlXC2vD5J zgx|lnsNud%ut(bg7bEBBKDB7Rx8jeVq1M2)jg_WdJ5tmX4hY;fP+N8p^6se^W0Pw2 zAE~9|$;k6wk+#q1W^()IQe8TpYbw3!-iGVw)HirR^O1zpaMYaRe&Z(2d#a?zqp(F7 zDz1o2C%Ia`WRS31`=PJpK&5>5pepd#b#k#_w~p}XZ^hop`HsK&jh^^;6*QacFZn-; z+QOjE&D*B5&i7Yu>(YNcojuF)dG{1dq2;c68cTmaS``X05xr22P%5D$M; zaJf?G@iuXydAT}>(X6HM=Mgn&?DTr~dH^QSg$6f(m*OZtG8@?Zo~wk}_rbBr`el^VJJI2UMkCOcoaZ)y@`Q%+dASfVNd3IW{%-?C z7?xFW-K|K}79HqW1>+nTPjWzkV6ETp0wm&%inMey!Dl$u0EnM~#4uHCjn{b|_SmPft9wCe41 zkExX9FEOYfqF#+Ep{MO`MaNjVtW zU;$$XB0Sj2VlIE7hgd^U7`_h@R~N@-7-;jAQ#Do`$Q9@XPfisoWvdB2nj=ENp~!`b z9~oK0bnL2KrRr93i_Dba041=&jEf||x>=atiK$eee-j=z-}^^>!w(-(WWX4Rb~M*=XqYCJj;EslkK8+DvVKu4F(Ax%>V;SQJ(5BPyv5JaCk|@#ELWr zpQu~pV)PYOia|SJ6OYL=KxO6pdgh2PXJkoMq}VOOcv1VR*uvxj@)-tv+J$t!#g>_x zHttYH!+cklrqra*rCPXN zOuF{wLWKj^NXM!?2^30TWiAtsWqwT>2L3y=bMh*xkNO51aWNsSpD((zeh}R2^=q(h zIVDud*NMl?*gyd}!%t=2-|0Y&Rr&(w7-44fY#NEXf++)&^IHLC3)VOPdaoy$Q`!}L z#?{_}FvC`O5gY&2pMO01<@HS1w$D=J9s(mZ(Mkagtnt_G&TLq}hkou?m}CCv$b}8Z zc{XAu6M%@rS)j_?fKcXi?A0k(;f*fl%FO1P^R@U`oc8SZL>Kxu(HYSG?jPVH@JM}c z(r1i(p3!w3DLVlqC@WjAAnDhtb3gQUuG8dNR~G9=OY#ruW8cvBKYbvA_tQoryoB=Q z?H%$osUG0<$6P>sHs;@pufA`&8a5n$I{M7?adc(|smp8Nl@w}N+)j2xZM&eD6q99QR>zFR{a+goFPc{i zrw@Y$e5wdY$Z3pN1oPLDrtmGcr_UoUw{Bfc>knPI?z@X)^+)_c_fsPAz3Kb_>nqW( zKRbFG{&cJP=(`#agKdOA9yDw(>^(wf)BXfxs<9kBhQY{wX7=m3?UfC-E9t&j>sh{7 zt3J)u(T4Dr3pGut84adXl47J_rIdg;WNlUL_x;{O`@W7*3&x~uW4$~L(w?Gt<58Ya z3w3%3|C#Zc-do7LbVAU3+to&q>-Htbe#13%W#3?Kl&77WN=>AB1w{X|c~G*~5S!2w;D7GY`qlKkS_Yz!ogLcJ zBY3= zndie8PpF0K2PcdvmrqY;6K4yjw9hj|n20uO6VWCGTzO9S>%I|@%5Kh{5QJcNCOZ00 zr`V>BucpQ}{gLhr8mO9Ck#cPHF!OnEH52i%i8pUTTr{|hID8K0w?2~p70F6CP7)Q% z_2@o|)kEB3n>Tfelun?bclC4OK-!0wjyq-sM7WQ|Hk!Uyc z5UsrbTmI6ycTl-E(9RLgaESE01y8e#nj89|16}KunDMB$^YtWcUGmlVl)zKtmntdI z+JxB1X!4+GJ5fLp9v9cny${9;0Uipv#GO@T>RZQi5!)G?HhcFwelqVG{`wvL0BH9g zNX*6!{fpuAr9$DyL*=RKOR|swP(0*zFK^2!Qww{U=SEa1Ay*j2r-$7Y{4vBH%;`_P zb!qVgd`0`JcX>*urtg&-bRRG}Ww%0~u1>Y&J7{DpfPkqQ+@q~e7mZ}=@tRECJVh19 z)tJ%b*cBoZ2(^X6Ht4TWrN32L^s1ICA>eg6db@Zt2$isZY%=;B+i%t+b@PK5B_g6| zRy56+)S{0?&PfW%mQ0IIe5M4<7jm4#l0$v=LZoc*afF1QFm;o)z^_$MVOqZll%_y+ zN-8s%B-oy^p!9YSrt<6D9h%vnJKsr&kwCg^3{KU2hClF;;^k6!Ap+$egFAyiM;stGX-OLx2tzAEWmy8o0e$TFdbMw_RfO4+37MFZcf2XIXY3G#PD;i| zgj$XD?)ThJA;s7?PF;VtI)}Zez`PW|Q8B`zp>%gzb|5`Y%>u`b2JAKwQuJMb`ww_3 zCYynuMTs`Cz`10Cs-H=5DvbsZUGFlYWI`Tz;ei@e6Fr*T0E&u%i4$Q5_Zg~kHgkhD z$&d=zBdP_zF)1}5X= zQvy|H7~chIa)5t=1r1p0`5m1sFeMfGYA$u^IpZEs0am;5)M71+_=dej)uUp>AbLMV zaFXuktG$ge{*a0?4s|Y61)uiMifTkAEmnI`Vo50!b8YlizNUBA9(HA z{|{yIyQ zE2RTd7|-XixrWl)FAvKdD{ss04>}?}xi4M<8h6~K&(16c|IML5#vCy;T~MUJ77_3W zNK7UtfVm{zf-ld*Poh*9jo3=gYECRyil%i*zNRJNKd7_$VKpDDs7uNg=M)^o2|TprJV#Yxu8brQ|OmTLXwzDHX6uUnmN&BV3M?di#0 zv6{5BhuN4&dpG40w^ZBbwCc~D5bG64?dfSIQ?o`A-o56?d+F;SPPp5 z`dk=J|5FH}g#^9a{`cCC_bII9*qh*e9kf5`Q+S#*pV`hGEr0s&y~5hseh18bmrkmC zGi@)Sa4=3>vB&--(8N$VVGkG4g@Q0AD8(*xgI?Zd>Q&vQUay^x4^yT3Yfjt`4o2Wd zU$0%9xkq}X(&fGhe3*it;!Y!+!ztv}Z+V_4BEznu)s&aI9N)ieokzSp+YRrHkNlfM zyVbOvFce4y1D6avc&o)STa1q-_Na!gcHROz^1HKGJ@0^6kXVrk*J4 zb_eXr?N~I~?!B^4*CdH;Z;U*xE)@=UabTiWdkwUGOVz4>ZHL8=5$Im~r}UII(1mQl z+kW0I4l25L|5Py?q4e=8_UE~W6UKBcZSCR>%5X|wu(1)mgvUN&GAU_^px#Wt>Mi+! z)26|^g8AnG-^i9!BahLEL-l%-zGesH$1ml6uLOHlaysoo-Mv);spS@)@qTynlw^2I=E-yhhYgNiM zjOug#0pcIQR#q%TSRor94s~uluk}?KoBW)w3cZ+94%}@fe*WzutNEQg!aUz_Ux9=F z#VXk>xr$xoZ=WrC1K7q&VL&WUqqXNbaJs`b2Aa1bx#{_#;SI5((|Pd%H>w?1h6ril z_{dihi59ZF7SvnqU8io3O0Y8gUuEvg^CoUkP!A)vyozr&JJ40?B4E|6&7<-9{qo8{ zzEC$6_4d9#UBb&~)t1fj%>KZA{nqGp?&Et#VTU4qn6<1f2}~ES-v~c<4IP`g?X!*> ze{JhpVk6}e$UTtXc9GITEi$P$cM>%mNGRHwQ@8ypu+KM^1|#51Y}JSy(}a4Z!RkXf zx4V0DvW`N1p2XB!4a(8Pb$DeQZM;<9Da^N2H=h+7Huf;;pl@LX3f);oCqL)_Ac>=*bwFl!omA2(8C{h zagvz16r4oSoQ9si-U)hSE+QjE-r5uJ3;0=%9BBhkp-NFmc`bg3eG+qSnMU|yhn8~(!fH{c~<5$mAqMWRSu2{y`g1S*ru zdh+zi?T3?FCGY`b;oB@P>pErb!R%Bph#i*}X!n~Kks|$>Ww%UpaMBQg!iu5;OqVxtWH%GIwqF zHYM5So( z2!l%_SmYHFV&OmLgOPy@aflQ4J?ica8o&g7gGsgC0S2QWnOT6@^Oc{1QF2un~E4fH2s*7+oY=-h97v=D=?`YSuDhpLzigm-PwjKOn)PC0oNB{Un z!bDU5Lg;NPm;lXdjws14nXLj^_$fG%a8DvMk*8Lmnn?;ZwHiLsne4Y>R}tEJ7cl4H zf8$`gY2zB;ITOmu*=rd>A>&3M*{6t`v+~nMI-lG_b*E&Lp{UnG04R~@hhfD<%8$f zr(++N;E(Me77icae`l_HQ<@an`;T43un=I^$$dy-UX5tizRzXRvn-1k&)CDr@bt|h zkPlUX>dbj-^D=pwn@I(eta{T?kGGcDJCVtSn<~)H`?>4BB1j2W*!87;``*S%9Q7#} z-{>QJQEM=ZRme8dHKDXJcwOSyr&IZ9XL7r?E~i^x0c%a%^laa6diZ^6v+-qKP(WYT zT*UDPR{6FmopHmG5m7|R!-YQ`%WxoeoQ5leRbv_2Q26G37Q&EiKgYHy%QdC9J?>*! z<3>?x_5?P>n7+ttwaulABNu{qPGv$rG;x)=@W6o=dOK}Y0h%Q<$~~d(`5X`mi2?2P zrjN>p+)gywQqVw}s}$h{T=)0O0(u?EmFm+D>F>hL7j^XM+LrtSsNXo5t>edi9)3I( zd^*zj%9D%YCv z4UFZjrZ(%^+w%J?Z=>g%VO8TOV<#3CI-lk@<>OHNR9tDl?Wn}OW1x31+%NNDd}@h; zn2$NZmg|sPJXUj8l1F0!!SUGT86Vx<-H){6V+fld@Y5ZOo_?cP621=jZR)S7?{6gb z##0S)P_A~&ljl<-5JZe9F4*8~-e8bdEMm0YxObmAg`DE+DdS}cQ}X^pvU7SAa8{j< zCnho>QUv8|OzWt5k(HRscR>9T#wW`@*0Xvr`L`G%<@jqO*lAr2D~p^CZb>d+GFY8T zGI-_W1TSu>gxdfy!3xf&^YQEYGkKrgJg)@!?_ul-Y~6Z!TG(raEkCKNNEjxz+WE9o z`ua$Ir!}9MPh!xf$FL0tilb}ylAyg+`@NA~CG*N_t{?^~h>OSy25oo!ryq|nU_N)= za+uqe167dYy;G9HkqOm-u{9=BSiU2KMBSNxrSWPnl)Ys9nx*GKmalgMctgXI}ai*gypK>vamzXKBJtd1cvr zR27b~0<*e|an9ng93TS#76FUmqgT%ew-n#0Wi}BHDDhQqMXKVr*r4;pRpHpgX~8$O=o$3Jks!@9E;QlD zgtlD!v#b6uFfkdII%2|b4W!^{xz}<~%6w0}_o`4P>M%^HM3a$CH#X%nTHq+2PmpK{ z#^%2=t~bL7bAK4wh{>8LCCDl!7lImx9XZ{CZY4{=z-Wxe2+7Pu*q$Cn2NtcnB=w#f zdz?=3&C9bK_Ih)&2Md#48ptnbO0?=T!axpkuPF~10oDVdi6JFy8GW!@<;H>;kBC4F zVEqAS5!jxj37yeHH19I_;0qHrpl7}PPWW=+YAkPpV{+&1jh;nJ$SzLy8jPW1rU14T zZzN=$e-4<@3^o9?wYu%^raC3D^j57JZg18HS^?q?O^DIC_=4 z9)6i;9v;S(mNw|*PwBNAb)T6{%?(RiYrxLBWAcG51Y97H41Bz4M94L2_zw=-To5=!V}|?GtQ};pOBtT~eARmGok}Vl z-Ky|V6&WdGza57h2I{N-mW%oh}>D=hAbi-V3qL%&!;qSYmRX zRJAF#5fk7<+-gN=g!L8315XbdrOwY08`K5HN^XH?$QZ+Q6-hRiid3t>BhkyviQ8ng z@2KsxIJYok)|x>03Z4Pm=xEg+Q#$l{ODiF!lblJ@i0o`yArb|MBl=B>?fS7Rpw>-# zHO@FK1cLs&6J#Z= zF1Ew!FqIkdj>@?yqEOg_h)0-FUPW}de5#}i}q2)FAz-;1XJ`tT~vSiyD5ZtROr1X zscZ#U(tgKBvmtY1`B4G!V%HJ6_}$mR6}jdeTtIXiCtOSlg_6-NM?`i2O(ujWbUf}| z-XNBC+UtM4&EBnSKT3+&a20xaZ`5XpmDvcfYHVR%!oiu)omp~r9l!t zR-X*E*bhb3XeI1F;%aehTx>~=Uw4li#GHV>?SJpj4ie!0jXF^UgNngK8S+%@N|upX z=mf~s^-<+kV^bh$0#u)T{pOg$LQ9QQ(kb+?x-W7iD|)n+pub8o8`XmkQ84ss&reLt z+~UOY!KrdK*c_oi6-3mUo2X};EgYB0M-eKBdZ{n)dzA!*R0I(u= zu4|KD|3rWSv|$xeN!f{ApI&8Rsy8zHN*|y{7&0Ctw-8w;LJ;U}c14EK(lA{Xk18wk zTKF;*$%4KJtVUS=vEXt*5MV8J3P?ujw#uNHT?>ufZ+@jqR0O64-Dg(FP*Ft4Xp`#i zQ8?k99fs^V>`{a=91wBgvdcr05V<>$OoVl)C@LRaq;q_WrGi3wG{NlV9#$c$iuLAr zBv@B(00rHgk_B|egop$s8XX4uW)V_Bf+wo1++0&52rk6b`SQd?zECNF06H0HZq-^= zQ(h|288MZV&}t>vE$4D2q0iKIbF(KuW`|9646FP zh8qM^(f;+6ILMu?8jCgzs#LDLRVVcACyssg4r{?}6f7B+1UD?|$2U)9Ofk+$yTJl( z({|KqkCC0lL}V%t{mdjGLb&MBxqRqPAwzPpC_XTB7F9&{LJU=l>^?7q;O$?Ai=k0I zV4uT6%1UoNQBlR#)lSsp{%?HZZQhxmUb~_$EAU@AZCl}ZO1%_=yrQ2TlMRftU_ya^ z`QFYN><>rZS3}z;OvC{toJYvZa2qx7ql2(bIl>~as`R}jb%LQ>Ot2TZkIT6=>zQ+% zD4l&)#D%L`6A>rnWEDhoogrW=XD+5E@bprBS4H(?T|eQF=B3~N(glokM}QTb=8^ty z!>5~eQN#5#VCpn$NeQ5&j5xEO*5CWvlES8i6SrQLIq zp?3m&`1DV2&H~!>d;jMC!>KLlJwL9q?wW;bTVrl8t?{*lQGAtOueQw~sBWLF+15&zp0kc^||By*9;6em-3_K0}Dy)uNqTd|g-F=I|vx;Cd2M zXa{ha9_k{#T(_l*BftK_!)JmRbm^{+Z!O81Jcc;HE6R#C%t^;dDynH_o-rxB8=#$v^TV=uNZNHuox)7TApxh8M9JCyUgx-BOlUfszs-hCz2O3f=jniFp6!fH&Q^l|2TZy^l^ zEvcBPFk8RP=4u%6{K=L;hKsFSjB;4Bmm*q6!zl9EdFEQHK8H#6O|Ii+NuH=Z6966d znzM`gd>jaa! z7zJ9Fm#Q=F8|ajN5Tad7r}{tThIP1KkG>c5<#%k1(|Jdua2Av1aPYARX8g~4b> zu%dvjUJVMF$UngO%6$wLy%FhzbRY}tlx6Y`Tr}Uf$&Yfn6x`0A93#N?4n@>6NGgmz zX94~HDkUx1-VnFZNLjChLZSB5m;eU+DZ6T%0+^5OjA22gSoyfK7EpI6sq-}#u~LRX zwG>Mnlmh5BWo{|8pyB`-^GVtfpata1q+*sUZG9n+e%>hAb=9z0bj`tIkXRH0LRm__ zsO5$!VCG~r_aWri6zE041-Wzkz59#+(J;;^Q0Qx4M@AVY$iz0EZ)_B;l>!5){BBnS z$$oYJJ(jJaGZ%J1oN?`^Z0?U!WU9jbBhKVvCNR-FeJ|R*2owPcv)M320;#G_xe_q1 zwAs;h&3)x<3`quLtBlZh>YdJ=3Q=j)SpoqK5Y2)ha&i!*UZ!l~Z+!3uSACLQj<{O3 z`T)gDg~Y};ra3{Y>=4)$5=4ze1?m}}xKS#}$5O$*PKSw9j>g9l26UYRGtl=X4Y#0j z;f5zTev+KCvRB)x5{!aPk$`dI8u0ybvvl zeJ3-$&r;-zUc>;n`;G42dP$zXF!mdEkcygcr|0IBL%)rR#QIykoPP>?ruobYSe=O^ z3J6n~>6G|l_KymOd@7&qPwwj>ClzgmUIFvBcSQ#!Bh_u+R(vAtH*8L2>`!th#6IZQ zmA5Wktz;_4sR-(-&!zHcF+!Abo&8njE@S;;Yga|zL)`rr&jQvj4_U?A`OfT)TmN3{ zDj9txfo#ZaRd`G28{|4kY9a=J0k`ReyB(3b#i%cyL6fz)c*giY^w9q8yfL=8IIw;*--8Itd|0`F91FWB9z))vmiG z<^uDJJ-a-vi-qks@@PprYzQli0hEXafb1$tY|4fo4Bky@k zE|)U9>p!T%BnG1x0n1X`>U;^oyFye|Q>)-Kn=& z3n-AOTIomm#Mm}c+iVc8=H+GLP_hrju%#e^*Zjr4?sb`g0dQN zF^&cH0I);7m?Wbgh~KH0Vo(;2e5z1Dm1JK;LgeO}rBJJXO#F}c{F&da)ah<#>3b7a zo&qu&(B?T8YQPjzJWvv*Nh_`XJAL(&ye>VYuf4ZD%MKjdLJUv}L(GI#q0sh7W;V|+z6@eU3i9XWq$toCm-?R}2Lzc7 zt3ci{F( zUFpct;zNp##}YA=YSld(_a>8A_LOLkxnI&T3;p(X#X}M~9~T_y%^gV$rI}Ju43QR9 zZeoF+@zHxJmtT~8jG}$h@X9)jCE|;B#)0t&s=PW8p-V&TatcakJzJotZu2x-MB2KN z>D}89-Y)=O{88G0>!j745-Tl^zx5k_%bO9@1F1y6?ZKp^K#eaB5 z5}XlMl;8q3%FOm}>?EtOoXj=<>=q^I*(u^ZMO-85Pn2y-_^7AfQ%dPDva^1ldS7YN zGllCxzsL|qcU$)0$7w5Uh?bTWP2V0jGht2++I&mPtgzo_&B|G{_H9)d8CDobi&Vju zhe2TyKtCm4>L^rhpEv?oaS6W|n_I^z0UyKo8)n+g7J4FGU4C^k9wEX>L0lc(oQxsB zw-rxH5Kc7VFj@U4?oSni z{O=`NVAP+Jl4@I-P^p_p4)t%5 zCdapJ%FlXTIpxum^#Pd6vkcYDLn!UVV+!(XR0D73FFw+=BYG9n%@)Off0iJwkj^qXtDpGAnt+bM{i`lajS4b~kmmeh+*twhW$dSK}f*7n6KNSM-T}o#dim*wHgZ$mB^g4-u}vjEUEMNr|Vmrzj9eCb7KiaB^`P` zK{Fs0U39;%8Bf>PGv5EueEw)USIN$Ncb&rZa@)w!sGVMWQGWu#hrv-c zBIvos)ynnXbXZ!XKE~@n_gSr?!t9hql_EVK(vui0T|pz}f+KI5^OUpzpRj64tS3TD zm8jnVpl}@MNU&2Ed6*et#fHPj#hX;$uP9k1qXk7wg9E&VI)Xk-Aht}1>!*4v_hfTQ zxcbO!ienKURy9#-)NC+FoJkC;tHyx{IwPk|eg~_)95?&kpQMF9OJGUgPACf^BQ2>S zP_nqv$qbMRlfaz7rV@DhjZi=_dhf=_XV}H>)vnpy4R*cBlR77~ zuHb=J14jeDyoNYNTnU0DL2LwwpqUB6AJz#j>HHtxQ%m(*PgJ2Xw1%d0FoglODd}-C z>L(-?NR;P#$6t$}LU!x80lnz-KG{uTKQvTk0Y9B7G0Og;Fd8JqmB5hc%%$~fMEu)< z1z!se{X8!Ohz=lBA^aRdj;EtbgK9+oTW$7BnrK{F-w95xCpsaW z-puA%A?GU&Cq$f}`@Gp_`n$NVKjc*roiL*sV}&CDA5!qX+%78)?pyJ)Xfk7rnSw)f zOLrMd5w*xe7M>k>BC#EHiXzw@8&=?s2>yGWaPvllS-BCiSs~~zB~XYvWHuQ>USuY~ zUr1G62`%Ba^W0iZ32B)|Bz%iesBnxnQ$_J@q{y&D!&nlrn39k5{Et!bch{#M+-S3&zU2)8r=eLrezK*`GMw?<*rng7zW=mF>~iLO)%_v z3Sf_kG4n{uhGr%Su2KMDg}Glb0f?TjT}Dh>qI=1sZSC?Tx`aze#Q!!vdjbDs!pBqK z>@YUh*XUcZ)oL5Uts6sXZ*3OF93C;{ms>0OdBNPuYQEVj)MKayH>DJ~(fX_Nd)03* zOIqpY@zr;_`35v0zCGm$hW(Sxdm`kiYpS%>62f*v{d*div`x(uHGUd4XbodO9tWJ` zBuv!f%1bzL5)vr0v-Tel&Zo)Ej z4{w;{MV&+yUkOaRxHAF106Br}hn^!Ng3@Z0h}r&@1< zP2|`Bc56)+ym^4A>g7>0pLy~F8gs{@TwV|Ro4h<9UM~ij)E-V()qEmG^kDspX}O3q zt=5WJSK)H%7!AQ~^pA5v9PL&Ab>$xy>tUo#-4feX$5L>0&*wZ0bJ7vr;#;guuRq+R zR7{lSoX@0?-sX8rr}sXN4(8e}_ugMbRs3i@xP3`LEW28FTM#kT^NgptT<_Ki`}2$G zi^39@a(*a55SYa?{Uxo*S)B9sSn>jJd|gi=Q!A^iL>BlFx&%lQ24q1t?vq*}8 zY_XknKZh)avFKRGAX>s`S$r4E!yf}ZMZ}F7N!Z+Rj$&i5ie^_{Lt8Xd5;CFA?r~G~ zi|Rb#Un@A1W;}$o1#N~V)9bV@9e-W!nt#7={AHf9S}KcK4{S}K?DRpytQMIrn=7g1?vdv-Z%1!2o?2%^bV3E5vsT^ZP-)xk!fQfRe% zfz99>{pKBL_meje$b`x8;VWuJdO?YYn2-I!@$W<=dj{QNyT1(LYNKk$Nen=Ruy>+V zt~2I3SaApkght|$A28562l;F;Uhb?xHcra+j*>@ZwXaXpVDZ3cm6=)M~VnqLN3UA8N9C`drvVrd-WK+c<=}{ccFZ*UAP9(M^&QS zEO`F@-FQdo(={TEgHuqttpm}@#W=mBw|}u-b*Cl9=WdZ$ziCZEJEJiE^cyQW2^~TZ=rQ<&h5;k;R-|Ef^z zYBpOz9Y8u@USxe!14B~=Co$lx0IgWQ^G4EFM3Zx&AS;3BK@)wy0e0$2au<<$ml8>` zJGKCYci2}%5nMj_4#=xJ7(?VI8z`3x*9L-myYy_aoDZ4o7_Ec?HjHQ%6StIr3l<(3 zk_aaEIx4ta6kGDNeGHihvyD^i3?(_!lsNQ<@hQ!++oT{`)i~VPu($L*`h6Ca-ea`- z>G`ZhGTUcY>6oM$EaajwrF4RZ^qVbWdG_!Yq-th}oiQe6wUTq6xF)7SdV}_>*G+nj zLj)E@9$0kDp_d4wAD+Is54?%mV+s4@KCkb^sRz!`N*k#NNB4#mBj2^jnXSRXC=OWj zv{?A^<;U2fNa(Rve&>5>?2>^tA!G2!QESD={A-zv7++m2y_mo;DzKAv8IF3s86~4L zKS{VbxuAckz51w@b$X|TfZ7p@GEq$Z4>snqSmbgmXTtYHdngE|ZEUCR8$~ZkviPvT zNltUXgXj1LG9Ok|OIQ&CP~iVUFQ&+pXrAA&E#x8lhH{Wb|H4*5t8@de4%+v6@!fs%XKe7r&uWex}{pM8j^Ut!30rc ztQZSOO``kc`*v^bO)jLr;7iW0C(~Q+r(J_-EiRvz|3tW%wT=WW#k+eh-1|onFE@uc zF@0u*1ua-~M9Xlurdu6h58I}wEA7M~3(FI!wv{V3Ae&J36zX{MMk7~Jq^}X+ZY-Ne zI}(pF2}*1nYz&HMa^&-DFtONM1f9(7@wn$~5C z_mr=DmwDH!N4@d4z`g8+Vz}L4!y)2&#+CQMGrKVcBEKe4(;dS3sQv47OQidwSl)%r z1b#2h^R02|CDhk`(UX|<^7PZm!s+XAWcVUatJ726d|&>Lr?ZM`s|~mHj}4aLDv*Sp_zXJOIx(7L#M(%e_0oPWRDvx=s5{7UW9faB%#iRc*%Z~4fmrKvD2H0Y$etWHaMaVhc)fAtP;Y|90A z$A5$zv4lv|dbGe|Vz~9lmCM!d@zD+!{+yhz7PjeFPUM?Q?4$k9$xopSfHX%x$9wuT z-3W{0_Fue#R+ay}X*gKeJf$zJrT2$yJYimxu*Wu+L{_YaAZrVYc6V==SP6=B9eX$3 z)sBVFqk(qTL|oRSMEpfu3=|UAL+@vIEcxd3=lFe2#nKtljXTHd)4J(DYJP(uQpTRN z+42M1-~PNszMoOCKD1jEwlQXFBdth1eM1HuSA50*Bj=$i$!E?LUie(p?zq#fCg9bW z3`tMmBi^)0Pt1)L2WAs@{g32wb|Svt`#t`M&yTW!T;B2VY7=Dfje+~$EZk#9{ims8 zV}S-|4mE<IJ;LVI4@& zhD#b&W}*Ke#v^H96ew6GpF#WK0uyLbz(#a7@}rj1v-8zUh-81dDHg5eWNS4f`2bif zhvK&Gk3%f@$&iu^;8O~>2Q95UQR()j91O8}vTClH1fk}o*pcCTDc`IiP zx_Hy+bD{3y{*fSt4cMGeKU^~FBV0(HYjFO2>X5bgW&BEUz~rnhxgL|^bv_%$`@xvZ z%S;mIjq{@5xzQbWS@C>|)k)#aUS<=sPD?CJ*`Ki#C;_p$iA1p%7+EQX+}UFHKXHJ& zCI0&0u29F%w9*OLjr}z*LP1Z4loT|14M|a)k&yD5*l7Y~q3rjgS#f1ihM#wJ-f0)d3rMUtA)T_@)^tsAg zh%WE1a`|4+4M=V5wq0%hLoxoRc<=tF7C!}GKt2z9&YSDSXKrw%yCPB?L-aaK}j7<;ru~G5xJJ%@?xy2g>O48X=Ek%`(Mlj9`E> zJand#Qy;XGmH7T8%Pr(QHkd>R8yiB9VqlxjD2ozh>bv}33IX5%*WT(d8J>|Jl0W{P z9a{FEB=zXEK)MI|ZMSbOPJ$n@uwvh*W$U8S9EduKPeAVwCxExjQV|zLi@EK^aXK`YQz6urM$WWd zjX5XPWjvwycpM+>nf=W{TwvPY(wug-zn`d$?W9~Oi4drIiFH-j?QC=9DWr2OY*0|d zK%kX`2L>J1Pmt6N`!);QQB3PiQ!Kxt^7=^ujFfDJ)uS1@@0%Y;GjugVyK8c9G>5&; z8?u)pvGKIS_U2(5mGw)1;cM+J0Uu$c+iJ@EmJy?^?SegRHoC-27AqMy7yuyUqvV}^ z`S|FWY;58p;rZs-ahSsdzCl~P{!MfqH;TWCk(lTdH{KISBI_6vkp_Xg=*Ey@U*vSx{tAU4bY z)QYswPQ!A%(j(frFIW`l_lK?jr$R0^3i_{3)7%byFT6Xy+kZ&9#Q`CyG{^*AnVO@e z2H^L&X{`>d@Qj#uc5XHqOpACWEz7$eSgw=JZ=7HaBZ>i2+-|Xt5=vl0Y5Q=J%6S36cj4J#UYrXJ91^70e3CR!I zZyPG0Xgy9R4zH-oH@D+AN}owy6A*jTe1>N{VS}xkUC;fuE57cl{%Y9mNz2(gIICdo zYf$GPI6O+4oq4w+ycE@Z#V3}qT=DOo+NHUi)f)%YV0z zOI}zXxY|AG=Pi$q`_>WrFE8)U#b(~k>B{v~obb&`4^E|*)tYl(yJFz(=uG0x9te_| zursQZWVx%QA++_uln@2n302hKbQ2t!EBa{f3(NF*KmJJnc(CFsNT-~&$%hHz*X4cL zRhx^ujlPXBGKv@af_z@97wEszOF2bW9!>AuV%-VVZCh~(vU%7MlzB+bHh{V7UB8e_ zG}(RO<8Wntn`sc-ZHMZ=9ligHem_YAmOA_xNy}qlE&p{W>}rWdqEueP#F>&Ul;lL$ z=jZjK=~}3ZvZKIQq{!qIYaMOxOq9z-sO0KhX4}dr z*n-)(P1}1%yT_X{h4>vhyX`wxeG6gY7XGu??iY6O zwm3B+`{7N(x6R|6U(QQ8zp!ZEF1xgKVxz{mSRI9E=h99Be6a0-NvtyECgPQ7U`~jQ zKV$s_v%-xOI@|=VgQO-GHdY z7`dy(4WrYW47X(rX-7ikN+FUk*%U)cr4)|lQQhuKhZHy=K2x`(bKT?R+UNbZ2fv-j z){Fj3-tJ+cP}&zDZR7#MLI?~aFOdE{=*yKhNFb3Mb7AL1)U6<6SKX@N6JqHB zLf#vJ{^uN688y0di+q(@75xK@pTa-3yDNoX;c_RHzx+{vvT41P)<)?c91QHrg$r*+ z3sz=;AGA838MzMea(!uZQDP7h|Wec@^KOdcV07<5A5G}tm5{v0ewCNV?oqR z!;>HW6Q^^JH5Ua=WLL8kFRgF=r1^!36ST$?!#7KBN;aK!pTu7IDq57!U$~ zvKvWnpjv%80eBSI+PooZo8ryNhoJo&qrysmr<|Jpd0(;E z{|^0`KSY(mOcsjU1G0)5Cn>QjRh0dPLy}tUGMnNy;A`z>Y)N`IEO{NHtF7im+9|Z- zDLRjZ>Iyti#LP6u7m!P@u8OBkn-}5KRRQ5SXx8t<>7_x6A@DRHF|;;`ZBN|+KU^Pc3Q)xWIBTYl_-opqV=QO{o^W9t zsgn@cx60rDELn3H`G_ZP{$@Q+nAfta+-GK|-E(ME_SX!t#GxzSf>!LKz4U#Kp9{*> zfFprMxj2v{-%K>VmXIqwbt%K75sn1}F~3XO`-fg=)K?jBF1|}exp_TBB+QCt06|nM zX{8MxpGr8e6a-kT8IiM*9<~lzPt-EbsDKFg0#F44EMj2-pCZw<>whZ*5@YGnorL%-Kc*Z0@4 z%I=!Tz51gy%iVtpseh@b<3?bjnz25G3cW=?XpHS2F3c;0!;SE-i{ zyeoDGbvK}12l=-imE!dTRg^u(Cf5ZeUVpzV`PWXnl*RF?5>YH_-kE**o!_e8hQcG# zqUPc3EE;Qzx#Jvm9k%k@s}~mnFl&~~y_JxW%1cF&=+QAYT1iG}J${Yu=qq;Ts&`>e z&$oJ}dM}za!QusPM&+bf1=QpTI%AkdE zQeS8HdgX5R%dOO$|6b=Y(w3jMq#4v8&!g<;f=T-R&Y%36&GE$qnTZf8;TP9KHrN)T zb=R)$9{S;};{Aam=-GVa;Db!WO^)1(Mur%Z)Ma03>ewG@*QEz$XSg3W+rAaYiar|x zIXOe$|9wLl;o#tF{26+GQER=PSKEn z95o(eQKRGefeM(0Zm@Y*x_>2xV>U!Qv1u{YltByVq!50!Uc2zzi2qs>Tdu$GXZ%>f zI|ch`TaRbeMdlLolyWg3_qqTg7O>s$BaU zM-nkMMOX*2S>)1{kuQ$!4*YLf^($qLG&*AJHVd>H!ArYXLmi_Wk)P7v@J?b=JwwHdmcYa@7TZ~M&+;z(3mB3s|@;bS8=k0e>3|!UU z$wSf8vFXC4iixST1mb9=OGr5R_AYDXD8KP5>b9stSN@5am%cn^4Bzum-%)SNBMoLy z^Hfd!LP2=E^eV%$MjAUg&OD7XX895`iaZ^z5h0cjOLj=r7(qDUsThjM(EF;Gkb--H zDW-u8ZwCU9$^Bv+k3ALB;%SzOQX4fnX&vYaZpWmuKm{pASdhDiVZ(!r1Vp){?KCuT zxJHP-??uZJDp4Q$vrAgJk|zV3?X!q=0dA&T(sw0gHh)7EV^TI9r}5iJ;>t*9GBp+MKn;RDu~3~&Y&KcKc zA)1-;I;iW)>n{DRbO@Fv^V$E$l^%)GCO1s(7uh6|#-JOe#vU)6AqxvtpKtuj_%`-_ z{%qv?5velau=`#M_RTzK&q}QCI%p6P+P%6R33?eLdb#NDIA~H?$^iemC7$Z-fJh3V z6Ig$mXiDt~de1*Q_UZI+r&}PI@Vq3S335_N<@&WAQljj7j}6UqvssPGKCXztO0-ob zGGe}bJ8wFwnEMc3ue0=#J`VP+T)R)S(x3C?gX9eDT-+o0k#Q&~XoocuQlT?FE=~`{ z--OJj$8eGX%Xu++EwY)X9#50+jfo?7e>QNRM$Wm7i%%1SwosIA05?-sf`MLVolsao zPAcN<9JO3K7%>8CH`2K$TKF$WD*jtAQGKOA+RIi1TE#3nGjSR)Koez`Qck?FJ{(WGP@ZpUy# z=a?t&@C-Wk<33hvm5Q*9mn&mH@|0ikaa#k)5 zJ5);FrI!MR@1O)Y+SK4BS`x5I`cGTGo+bJy#?bTOZ7N<{U2HvWJWvfZ)n#iYZt}Zu z=ee}mcj{PuIY+3u9q;z)dogj`)YmH5pUo2D@2#GYffYZ$Ub2#qSCnrq&v8l;5K(1r zKFGN+^LQlO7eYUs@-J|ItnopHzv~`n|L88SPBLuT$6zNw-xA)oK#YJ+16z6d%Jvvs ztR@u#X#IoNLi3BH599-zn-_p&q}$M1lHdb8bKGjN48_vU7TZ9WaAMR<$a0P>}B5C z6zro0_Gw_j0vm~-B{{~nVGTMcQq=dOYw8XPW{9~~0u1+9_{EX9JosZwO&thvE|4)q z8=k*FuFvZN^GD_GKgxRp-bfD1)D>$~O?6eub&gD9Kvhhzqh)D&T*=QQW6bGEoriJ` zL`4hs$*<|=u9@iw_gWu;K6kv|Pgo{3G`bMe!L$Hm|9A#C;0dQ`EohNDHMFFTU|g0H z29}?I0_B8s+Weg>4tHSrUDYNni9!B7^p?krANe7SgF;4`rM<_tE6+Hk;wQU|Nl~gC zX0)x~Q>d|@BrE3cu0P&qV#+D<;LHR`HI5TbzUMyv=0u_Qov|bPv4N5-{rczegXD4F z#Lt`~lKm|3X)CdVR!zyPs^stD87W&`3m9NL0~4PW4~cwc-Tx+&-F9f)i0|jVb9ZV8 z)1VouB66e-%i@z(p=A-3sp`?B1GxzBBiD&G!Z@QDD*wjj<5hoAEp?M>A-v#2;XeLQ z%G|R|tdLFknHQuH=Z-VmC^!3 zk1l@N+Sk!ZNG2{5nVze3T#;<5IgZKJ%-L}WQ(s&b$?4l~*9})j@eZ4*A0?}wuuMA{ z0iF{Svtv8tr8?nmfLKg6X2*?G)4>W!O@MpYc!XJ@6i$1S7j2cSJie(6qcx6V+%1B+ z!eZM4ah1i8fKmOHIFO4WB}R?&AXmg(Jaeg4>@bvBb>N41W6GdauDmlRANd^P59~iO zbJj@RhI7@i8;0`f4BvzVhAk&I?m|M(D|J0se&3lJ5$_eBvwvqFD`+h}#sN?(MI?W) z(aS9djOaMe$)`+spZu&=l%BS7`&$l*qAcV5BvZW>S`Jpy2gSNI+mXElQiCevM=B-t zZMyU*bI~j)elY9IFp6X3w}fsUA4gc#aQ$i*hGHp-%p*$(!!n%bemKls)gHRnr`VWY zAgX3~j#@dZO551S5*Kg)0ZkQD!cZcNiQ+WxqvPDzi_M~d9mJ7~EbL3@CQ`qz?-X0% zN1;wiZ&$Jhqae?z(JzhJbL{2PaGWxXAZzzxjb*j}G16|R>pFO1=#)|3DuA|EnBS4ihHe|o4>Ecy`A>N6m%?xP8$PP7>zb^P=90u_&F9Ew$bU+8 z*12dbw)PN%bD^I%+wPJ0p}Oi^HFk6tBy=*-qu=T?N_iuHFjRB*6^cKZWny2JL@BID zY-bVqwJ?^)H&jm$)c$@Ri=-0K@dxkjIRrFUC+zn>orC3TA0K{S#f|Ziv*{5!41U|s zhh3~2%2Z&ZhX1|eSpW0(F`0^cUtLhZt=4vcKJ4JCq|=7~)HgF~Bq#Ubly5saP^~wJ zR}eAcpa08lW{J<#rsng9pQ1+w3#Y#j669Yldk)z|g#yFwpQ5UD>;A->b0N+4b8bItjD5m21A^p42X7sLQge?6mv3sLHhA7^?c^+$?hX-z zs$|3SwY&DZ0j#KZ?rW*o{_dSt%Z`s&q8(wGAZ+IHe|6!vH3bz#xLZC}cM-eio|mWe zV`q+2afXNHYwC6pys zb1q*-v(@t`?PKNA|H^0$xvbVNbe(uLq}{ml=CKmzrS>70v(;v;%x1R@@NMzZ^9~4lVZH+zowbgvhhS0%~yD~ky zJVuY6l5!3f_f;d}+}+GwS6Vm8P@Ej~Q;Nl$uxaq`$$<(K*swS6l}&MiZS!X04|$Y+yO6K(A%mZT-mpDJc^*hlVX`d@l=* z4Dq+MIAL_{c$4L2&Y1%8po@kKDMHZI8%1W_zBTl^#EZ=FF65|Na#7Y4KO8@1V+NnDjDaHBi5jSG=O z%*Ch3l|`Wuz*pmaL{%!yt*cBJ`V~noH0E&FZAJ0^_Uz?-nbVrW)@eR`dek5fG&!7d^mAc#y};o zk=Zs{cUMwLJ}0SO+B4`rKAme_jy-Z38>6~P=kK+%V;#s_=7TiVgbi>%wz`Hxk1)$` zS{5F{Sh|3c0+4|yN~J3lW&_37aFdCC0C}})b}bV6_`j6V$|WqZKtdrFu!3Gd>nta- zkl6kk(?0GSC}U_0H{YWAkUH~|d;I=sj0pDTf5F&jvt32RCPy+ z6oZ(xro?2*wAl0n_Gu!%@Z^%Wns>DbLppvq^~5|^=Tzu-)yPL!%sWzA9Bq+Y7Y(wE za!zy@^Q6EyB^d$?Z;r?UaiOx4k`HCT;6zU(0x6#O^0CJg?)cf7c)!cy8Ek*@4kf#q z#Y_Ef6OdOTu{TC>iJ|$pIl7Y_E74RCjF4B7kL)U-0*3pBO_nZ+O|55|j-qbzH?fV$ zJAXnd9U}QH);S+u0L@k2T#Zy5f|j}smB*?3tdW|`XmB~M6J~w< zDv61c;oDMG*(#293`^t%J_AZXl6=Xn#vVcy&+);w0me-SlmWcI6a{EZhn}XL3NHXE z+|qa%mMbmXN>lqlUZv^ZyYNZkc#xrW94d(&p6<~8NdXSsBo9nO1;xcoKbK2HWCugB zTqoLgv0)q!w&#ZZ&*s=^AUwrjY>hXz%aYA{TL1)PJ!Rjd3c~C zj_kaES=W=`)mk7A+!A_pQs_6vRT^9~Ls>jlIfBitO@!AwV~&^?fCX-}#-Q5L@q8V> zKc9Ozf84zdUkOBmB?FO6)bcoOQ)L!_@A1zR`lL%@zNATw`di9&hOa0A36TS%m()0w zGo=i#;ac!Dd`-E0Y)fJQ*EU7pyzIpgk?-Xq(S zwM3~GVOvC*o2GYP7r`bg#_iI3*6G!>?>^5SpLy~+!d7(A0Yepb zp=}7}1XOjm&g0e%(DUnh!VIF0$ST@FbnFVB)k$)<_AZ}PLur(>j7U81qiWN9MvZ#f zX@7j|@7-u5)08p;Ja_dl0JD7;rK^~}HUiSi$un$BIYB0Kty zEqCp&Qp-cfnsEG`4Q>j$xkQADdI`L?2rvKt$yMC~q3tUmMxk!{93&=w##cwx-)s)G zJ-p34(w}lFOfaYY!QeWnLlqb1R^@%vali%@eJXLOR3_r7(YkiuL*WQ=aAZtZZo>*m z)1BJY?%24F)U0+TO!V&`);0~*At$g_ZH`;^CC(vCnkt8KNSqpfQ?Xq z3L8gSXt8xdF_#G3&+>1YGcPvuaZJUjQhWPj`8?X}buQ=%H+{c2(pe^ybA-B}6-h9? zs1vB0_IgD#^NR+$xxqS5uPnj(w#>8iwMVofp`?12=JMhNoC^xk=Ml~t5LMw^RjcnS z)NIn<{rkZn6EmNBxvgzmz^-j>>_f%^oRlu_Y7B-*aoK0a2#?IX z+uUxed!1>0N5>Dm68o<=xiD}NQN(QuEQ6%EbANW@GLBx*Q@ODc!zp;49u(9)eD9;u!{HmGuv;Ocqz%lJY-0|)s%{b` z^{cQb^mJ?(chX3-|6s`q_Lc-PVJ3=>CDW`}iBDprjXo-3$5z84{NlgzNxOC(f42Hi zbSc;9nOP*9b6f!fP$wJ|qD;cTSomJp4j?&j)Jsrlzm$$Jst8Afu2>ys9g&i@f!E#ey zUd>p08@~N`*|HP#Foiswhs(xEGGH*Sod13RnE1!lYhitH9iba~pKjJ2YSYbdT zI%h_8sM~cGX+qTwNYY~g|FW9p{#ZWRj~j^jjW?p^6m_0Pedq~|>MIv4!UrkP>$^-& zn=xi`2dzLZAGe2+sE;SG8K&~IwYD9#>;`>`x<{=;XDwZUhxb-XMsp;U9h69oNUOns zXj#+>{T825LLJiKuN@3aM%e_W1ImNN0hmfZhy!pA+&!ENe-AC)BCyIpW>G8}@?fGw{UenNv4Og0 zp^2$Qh=x&P`;*QRqUq(^LprmhAu?!E$qp{n3UZn}+#XV~a73J7P&y9sH-193e(c)$ z&zYfL86j;bzy#j-6HQs5nuF~k16||c(h1|A7xt4O1YB;I$k{T8`0x?+OpxGx>G8sR zZm3i#1Z2;m+#>yxmqJ-9`}h*jw0?iw^^JpP_S1-U=ceFq`FX7}a2ifNg~#cgaC;t{ zAT4h@TB7zbLZ&5`4r59-kJNGDFV&iAgOk@pY>@b`@k%*s*l?#tpm2tSK5ih*4OKmV zsBm9cT`RiUV$VXXYGfgryS%uD$=`WAonHVPdS%(jP%sW~X{G2`QU<mC zZ-zj4zNW#s$-EYG*N9_dj~|Q%Wj&}O(la(weq=>N=IJ=+i1D%>M)ose#F&O7j?F%_ zh-&WAWXywV?Y{n>0%oXj4tJWx*Uz8-6|Qgd+Y;31-u&+1uw%4$j$a|bpErqRE#s8K z>p61o+b>A#;s$#9)wdy#cst)6emQ#d-7c9blph|>DE4B3Ft(W)3raU(J4JUl*skyz zC}nPLcQnUZYtp&FW?fWEHpj8ZVW&uj|?Rg|LtiZy`19u8uJ&)cq6svFenUZN;nMXSY5tr#H5{z;_<) zGLUFs$Nq3P7i&8VlWP(n9?co!!&$?xzx-Z&a^Ud#-tfP}LQQ?S_-THsE$8XLf~G!N z_4aqg*(EP5nxP_tJ5Yot#)CI;gJxv2=1kaL=1qyHBW|Tu=>6QMFZM8V&e{!h?bopQ zK88yD-G&h(^f}h$RojS6{RvXIz*mJ#&*Jqf1SYP7xuM=ViB3Z3wS`YOTzb9!XJI#u8*M?q{GJnr5a3uHWlCkq=O%|Qj1lq%oBjwF8QAM4!C>(B%7-tHQ8GeCSOynD zc^vE!BWJTW(*m9=H?K7UM^ll@wWd(gFHp>U&(I5~!z1*y0s11+N9z6kWBxSR3)hr` z6DslF->Sa|TC)`O=-i|sa}grfZ~SRGw6^P9`T-I>>t*m@i9?`|zw2v{n5jCBuJf@9=!wuN8LQUvfZBFH+=Xu5KC0tV=% zAfpNV9kTIK$JVw9E3xIc4>Q44`1dNauy#?U{sTmQ+LGl0z+Xr-cy+*V>w zs5^Kbo=6z{8KV;gMLgL)2{%SsOcSh{YVE~2oT2#!SO9m;aDEA^Nx?m@s)w%lGcx8b}3c*lKGRu(mcS|JR- zJU@&b;6^yE!N@*LPOINITfEUMN#jlIDrD7MS1c@I*3WLioMMaZ|WBBylO- zb=?5?Sv5yaYDYMp~+A^AWD&3TQS`49PIYUL%L5J?gj!*2=J2VvVL7kZ9r z6H8nIJVtuR-_EX<<7!0XQS4*ORIj9?ed1L=d20GZ3np}kBq?hf9sni|BrehQ0NK3R za;1iq)X`p>52nTGOP;0xaKnt_bC~fpEvS8-2Gu4)#j&e5rV{m}%=GdjU~GB)d*<%H zOlvQyckx*PBgA2X&N>+oqb|NBd6yo8MyP$27Y)`qb7xnGPliCDT@M8+tILu`Hr|Aq z=H2-TwD}&sh-R!g*6`vFex8_ z0I?fGg+ajzw-(Zl@xr_UB$m%)w4v#EVCnq1>A&HU&fF*2twr@?QompfgqF++!Egp( z^)1|q(9u}rO-U{+Dz{Z_Ri*x6-&Ms9aRn7t>WNVk8*FQfF$h1d8k*-ykl+RRn#C2{arzFSR2jUXA zbXjoocw*Dnej8Qug8?dh=6(OK=GWG`um1lvUa%Hw%F?vAJ)bJ5Id9|)+GG!+e@9IE zlc!Mz+UoD$e!eFfNp}H!D;*){cX)@~{*QMIreJ&%`BCCZSdR-C;wZ!SAV{m~Iek*(Y)4HI6j7hGCND*N~KVV1h92iyGv z)&*Xka@?2W#6uaZLzs#8?R3+O((v}>aE)bnzq6QD zuI4khz9ilOJP!4o$db+-+mps` z`~vk(UOIMb@yQXu=*+kPph%v#|6Uc3JLB~pGKFYo&~*Rn_5;&cx#^_AIdfW+l=gQr z;gYDfx<9{Fe6pO#)d5vCq;))l^i5mhEgjgh6I_ZhGHmVPw5pt34Irv~Ys$hFV$`Gt zcufm_zDOgZ(<=@$1mco;J@r4kLOui-e5=8fc%4|X2aU5=P@Cb$iO}=pPn#3<-h{`^Cvr6=F~y(jBbL9Kb2!7Wmb^{>DN>p{pri?YUiHg^Vue0xH693$Yi)$T3l9 z*fo{pW3YjCWX>Xcp^_FcygXnTgBA?4fwGX_={Qg6t~;1yA+VS&l>KxR=$=QHY-IWW zY5{PT!h^}_A&Mv#m0chX+SF!QPH{pwo;0k}pOMEbrQcrqx&jmAu%j9jz9WW~F{3pT zpm9nN+pc}~WQbkY1&`YIeWpg?U{6Q{BWf&PZtn>5zJ?b_*J5^?8Y&>Qgr-7O($+y$7snK}3lEZF2BcF-#030~^%uu8aRR?=o@G32;lesKhra2Z_iB#`LxM%FL)bWlv#h@+3L zNQY}MkZ_pe#?X!c{4cg$6Tgw0aYyuCPQLm(-&y+bvw{Ah0dVFs2e{SMQ92?}Ph9CD zG**%|g6R?UN#U!3L|fh_50pdarMUaFunagT7BMI;Ob|IgZX`LUHS(BiR#`D1Nj24? zBxlab{F}~&Oy&(?NpOB;%Sf5)MG8f7=`nzcK2JERM!01NO6-^ZKt$W_$2S!ESll_NDw z9Ky~rRbO6%ZpBMCT2d3@Q_|UO?cF&w&FTd^;1n)iyXRpQ{VId1d@~8c(HOyEa{qSX zyNykq_21myRz+Ub<&Pc&53{czf#FB;x8eJ&XryIc{I6qBBOJmT+*|kSXz<65kNT<0hNGmCgi-nD*nE*(uvmCc^Hkl@Z#te+ zo4bbp!O`X5+;8B~E1&B4_0FZfEXHNN9~|-de@8VS*JQ{L_*0`h@u&Eyml9TELW8_V z8u~U=MNc~T8LJYCc|)-gcJ!aXEtx_>1!|;nqKfu1GH94Z!Ta;xwq-v424PF_A`2to zLbQ4j!Fv|NeN>l#MmxdMWCI63%THH7Y$hH#>C+n8ge#guVemHk=AO4~%=`d5;A@Ix zWh9KuhhIXQ7Bv+x20L9K8CwE5+W4e`iu@;?e<|XCu$I-Q?flPJ^vA6vW$*jJRQh$E zJf$d+9CgA-6O4HE?YyD>6{IirpJtEH1O;jng!(CtrDZg&Gg}X#)lXJ==KfXU3L?HO z`@>P_za><};$s#V2wVGx;Wqv#zeoI?-R@m12$b~oHgIu*UQg8&Hd`b-AzUE<=<0b9 zL^~V$vS^;hn|1x9F7yd3AmBWe4zm(C%?jQh`ui+in{K*r{(4z2 zB6sVn4RoqKwRCzswcM{(Skj#EEYi`~D$uGz8~)=uXwhxuvZzK9^RYAgZ`3Y(I=-aw zH4Ye2)ThIIv*GdIXdxCQMZg~M&aVu}Zu3;8X2$XFk>>T3sQ}uqBz41avCGBgyfx0O z&RqPw0ZFY05Yp=EZ?@d(!GV`c5d(B{5L9KN4;OSTpYU7#lbl4m29B>gmdjyBAIxXf za+o*0`D-q$IESZYU8IBso7qw|Miz{ar0oOf%O#m5dW@S96<%*7gV+0IP_N&1Htrbe zV~B&K=xJ{Gl?h3Dg*Yu?l9u~sE&t3@$xJWTb#S=(0$Ec#!qPP{PM{}1>uQSL?!3X0 zoK2oEmw=1Gp}py_uYP=}jc)cKZxVR(l zF9)uEY{bPoM00Hz>c#?j9Bee&-4qh(m8OLsm1)IM;)Q3maB%Bb= z%>Q5-IgeQT^r`7wa6p{felpicnsE0b;jk>yS}Q84%jP4PmfE5*g1k?5(?@$u+AmTd zxV_?kLm{;M&i=zYaTmuzB9ahX;>~2Vh!ZdTPxPN8<Y?!c&eUeKZJ8XN`IwC#S zB>&;ma{KWN1%jg%PV4qP4WAkTDfMJZpNU+D|0g@ z5Iqh2m_aoIfgprYlf5rGRo7ehSD-u{liWyt5K5@-qB*6@nz3RZ+qwJH*&!hq&0uhX`(L z$cGG+1XXx20kK%inxYb8(W1MotioqIrSuPS%Kv_3qp0mmHZvwO7MG2u{RiHz-D|tC zkA$U1f|1Hdb(he@hkT-}v{D34n&M^?mXAPbQ+|WR{hPeTsF6!&hBHCqWHZSMn2lDPC_G(vkVD*%2v*~8 zi{YcOz_+S;ukVyo7BFX2cX|bio1jIy6niC3cEz(_!nv(tIRj|Yz6A1qBGcgwRpN7A z)Qe=O`Wsp5YMX+MMegDuo52Vv^SN{=~EDPY$GMvnlNR3`0t#W4#Qv68_ zQ)9{^Dj7Ca6X?nFJ|`X{i7V<{TzmJkSK{;M&2@Oe&MPiwA-0Ehs)C#KDPgLhK#4NE zdu@WFnp_xGJL0m(>&vyN;FUR14v>Ke8)8nb0AGZgrgZu{vu6FX8KTm+V|`+6Q1c}F z>i9l_l?-}3st#LrR*avFkGpVQX{v5nr4sa#Y%FE3Bs9v3Y2~@lg9tXG%x7yxJ@?=D zzM^##8|d27E5kGqMl%zImDR8k z-0$C1;%NI8)t*?1{@RPe94yX6Svc$6b@EVgL=kigMJ1Yo-~G(TpE<+Va9fI<-$42l zxO=l6v$jEv9evC$)|g8M?bm6fqJd@4H_rB``5SX-MAB2BM1z*|ILoL>Y3O_ z#lh&ZA{$ua_IN=|&<%BY5{~F}DxC2-sy#ZIqx(I3m>wAYumx6*a?Ml%@T3Mey zA)NU7MaQ&-7^#`FVr}b_9n9qSBIsiDE|9K}`0@SGuKvQkLhv={n_$?zvey%ijT#yB zlT*hZ{XoCYz0J7gdpZksdH*NVu}AUC*~e2Eu2?rtwuij?B=Axsz5W>NGJa|ZcvQDe z9#Dl@xQ4_<>|ePI!U*(3r*<q2SQ7)c+}~ch_C{x$t>t;`6x4G?89aoY(0J>1+dNQk%jSLj z-QCGwf4Y0`FL!K*?ZqZh&R%^#_@%-t^eN9x=>3olP;%oRxD1&XC=R8EFT4<<>hlq| z!S4^~1E7tr*=EYz#g|v1_xXFyblox^M)J*O;Y)x2Lr0p{(avxw!p@bFM9J#y4Y``Q zVd?K9&(ZXI+rMh>FBhRT4aG(Z6UNt6B5A&i+CR;hG%7|W`^k7cTS9?eA*uN{4#>hC zURGX6&1Nv6v4ruuU*PK~dJA|>MIgW?h*?VLuiqi{o zS$_B}(o~6_NG4Zy9sS(I%}|HJ@iPGrl7SyFpckhWS(0OV9qRDQti`hBQGrSPkoSpU z&{{)2T`D9c=3x4(g|V&_RKQMWDffgILOgRY%ZY)vB2&sMYg0KDDhUjq2G^CEVy6~Y z=XkP|6f~--^5o9@+F$Zx&FY$BjASp4L*&J$^MG}&FuyZI$B~!9j2TCBSV2TWasTR0 zkwNN@f6(buNYhW4%Xw2$c5Qr0I}Jz+Qz!aMmyI0%lvhPb;gkLm(^Ws;HLxB6o!R~@JC+00>&(Ga<^>^?}mDGx~m<(sI*dp*< z@e4bq`j5pAMSrk>MSO5RX7Y^1>{0Q~tR8ss00{3|I%hNxr$U4)KSG>13>%ka5zdWc z{}%{xVeo6FG-Qs50UqpY>Z+JFj3f)(LdejNMNCU-aVVb8ol=TYENpHjX|66V5*nn> zpJTotS8c%j@ab!uy(?AyO+)M%En?92*@`TV8?K17DkpuCl=z>}W6Xm9aD|(MBCR37 z6brhuM_I-cE<`LkL=*lQZypFC+=fij+)sp!lfi|CdyvqpktU>|wtS-!2B5`*JJ!r8 z?c{M!=z55(X*yjCA+TT;0uk@h@rxZH4j?SkDu9pkH-}S4Y_=M)mKsRUjWaJ zW2&gpGNvB-oQ(H#4loIs4xG{_N`l(~>;79#;9~ zw7#Pg?3bow@Qi$`M{Q0*4wG!!II-U-2npig{D7GJ$`JI1b>2;C_CRe$B&&F>1=I0h zIRL0FYN2_g)y)j)BZbKaAl;F=J>J~E-GFJMzkFxDZO1RxN6)_9-DJ{^t`f0ce#SEz zNgAfuv)Q$Ma?4j zwqsKc(Y6t!{mMC>_xNYO^%vrc&;oAM{viFoFAJ_<5NTQ?apa zWO>q_8HZs!+}Q&fr|;X7C$iE}A757Iga?nJ@CPrdRvM;GpN@+6Nh_}}y|Vvkh#LGL z`ZZ|wa{+y>pXI!MZH*~B+*QuA!MSzm{QGlHPqX6RlKj7hwCHGyS2HL50$qKl5&JpK zTp~=;W~Q zT|jEd9rt(b%pY`THdFr6r=J6K>?g++>woCFDiG$RI?qD>-3(>!r(W^Xk8W4-;|sEV z^3Ry)xiQ=5jT8iaX0l*C6T{!#EHm2`Q@1ay!ucMIZO~1_pG?gAUP6_(Ccp21^kyj; z=;PMtg3P45RFXYjI)(u1{aM3(k7F>R4CpB46T_FelQIDU#@;QmhwXPgf_H1$PX=k0a zh0b}Ywqfsw3j5b|zu_2PN!kD}>5Z0f(!)t(hSzh(m4O~+C#x;iH~XSr5rQn}Z)iM~ zU_e~>S-`Pu_ZGRd8mril=hPO`=Qvo+Me`Pr$jv8`fN`TJA}#mYv3Mz9>O`P;#pB&I znKsPl`GbHnwG)SwaLbeqDVs-ZwrObV0EmPswBi=WA)I6;akBKf}n#mCz!b=|bJiw$(}rSjB^c3iXT( zI3`b1pQe}F)30~0ARKZLt{>to23w7Y@K{kt_BjyEf5SBDa{Nud}J_t^Spsqmmpyl z!onG$ncc+W8%98>2pK3U%Pr4)GRxNX@nyL@EJ}C%9eq;E=I-0a7}hHiJtFu>Z6ydV zHLojuwW@$SXD_%KV3Lw=z{TRCNWt^P~<(XNB+;_{F6pL0qfx3L#ChpuYm z4jh`s5oXMa-GP&qynX0eWAjzJFglJuKjD>q)goqjJdPiR9C??GsmjKpJZxVaBIb`T zBPijK3xB&0x^{+xEZd!FgB8lK&x!{ilW1u^Lo;H_D@oVCRr}qwg~^Rn0~nV0M|xs0 z$p6TMQlcZE`MBjdMe#&CI5=2OafEH&|JTT-$?H6)_xbOmRmAqx7elrJ^k=1I7pPl6uqcWEf*O5?D$Ykk!1xvG?M#!4^tKXyW_(KbtQ3995Al);>%E`@E zfj^(&%I(js4%_*QKkgitOWhqQh;zrDVNFa@Ds{6=eEI5$)41F5d78saQj3JAbc8{E z1@ZlS49ERdsMkPmnjAt#np?Dq9MSr=^JUOH4P(Ao-vS{RTYN#r#IEB7w%fve_TmN3 zW1OSc)z_;hv{AZ|GyqN1)27vBEp0Tc=h5=ltWf7?w2S_OCNFQU0NEG*6Q~E ziWTY1QvMpyJEaag4i|oX`>pM5&!CR!zrQQ4PF(*>qdUH?dqYG~x6j+|wr&k%4XfAr zmnHuxu%17_^&N*}NU0IWmJ{XPJ|%;Lk-Fi=m6AD6S6bp97OIkfh2?INiTcj$6)%JLyfBz9O`DBb9^o97`7Zw_^zdAlFF|Zey+mxhKoR%0bl7$HXxw`|hlSjS(9`;#1mzKEDQ(2vA+|V_LQzs?CWzr@G3CI1cWJHgT8Ed?fgv4cq>j$NrUH*xPINSxfBfKSXjTd zUUme3UhfCS@5!Y!b>Pp}z>16UYe5}4!bMysVj+0J2}NpjA_Aa$2}8Oi;u>OseG^%{u+ zTJ0TFGNai2Z{tZ0en08w`Q9BB*XRAw3lnKd3I@-Ds2z!3zywSNJ$YYJuA$Z**>qQadMYvr&Yh1>E0kmu;?f#%-v0bsh*}%Z9XVob9K6 zScpv%RppL?qh@som=v}}?Bs~weS1s{)6Kw(zsW0~YVq8(Z&VZEQ*dHeLy3%Ti5TR! ze&QFyDZwdv>ey@V@zaU9l2+LmT%xKb^pQO*g`$$J40Yby?Ygu3NY#R_4I`M)n&q;P z$df7D2E}NF{n*EHvNu?MKcAR1`N{NFf!w-y|M?<8xk$_K+FRRY8gyg>uqyXnDDDHD z&N-`996IX#)H7;pMo+%S+ZPE=^c4~{L3utBHXlbp4XKuVla3eE=1kHh{Ds_0>V07}fJCZNnvdI{>+7asMXX4)(=&n? z8xKD4eQX8=E1Rz3X_8ND#q%TbVTLD$Wc87+{-}oOW3SnWeFsLig3n7MjPu}K4kIag zV~UJJkM@3Weocu^!G(syEID*Oa{uN6tBg(^5b_gmp^s%zaQV4MM=~>xXOU{=dQK4+ zR7d5~pMurKKhQKd<-%2nhVGY#+CltUrkTXQGTAKqYT?~`tGgeL(@%os1vf)OvrW=1 z-yBmyx7Y)iNOZVCim&Dg@W=oaqb*X54Z6LrPmk(R+){FVB9_}4d;k2s8q|Knf(up1 z9O(5Ig{8+5C2JYVI=!_rsLh&E$A{8Kl0C8!iHkjERZwfJ;EcznW`PowenL*nJP$j~ z7#ZpaV+K9*fKqf$ew}M!jY^6C*)*}pT|I#r8=%6U6r!p;x|~)6NPw! zv!#t|Uz!^E$aQ10UP3ao6!a!OFMqw&%&FdKaKMQds^^5^M*I0lxX=G#D7ioz^z6L} zI}7~Qb0>WvbQ4J-@6iJNP@XtZ7X3Lbp(jn(Rg{o{a*(L z*H{T~_a<>c(LT9y8qR?pDXb)to$!U4WE3CzTWiM+1sOfRJ%p*{f{KBc!|t8q%>^~3 zPC;*nVJyu&NyaKG3^?(MUg>Nvo@4DRfN~~!Y72aU5DXZ4$RzI4UXc5?3ng|}a6m@R z{`sSK_HUZ*K{W!(_Sc<;aaZke_!9wq!;f*F26~9>V1{39uWn+`a^$#kJ6to;@_Gu^ z-Xx&zbwW?x?JTUEUIF(eIV|o=a<|)i`M3s7DOQj9EPux}sGK#38=4)ez?QrhOkXdW zl)m$D4>@F*P_s~?k`M(+!n%yQG08ILrQWZ53r^Q=oqe)BX zI$a^W$!)*QhJC>!PC5w-s*d2=Z3ogPN6WbED9WB}T)z>_%RJM&X8FyTsa zXzz|m?pj%VTWM@3VJsrOkcWW>$+uAAObM z<^JEn^Hac`m>&*P9+VM*BfIXSd)yo^{0r0I#v4K02hayMf$x+SPY^pzw)jKSv05_7 zy_jk0=d81k964l_z~@{~MNKhai_^wO+&;VUXUKHtDaTo@W(XkEE`OE#A4h!Lhy8+4 zSFyIPy|)nDMT;N$L8EQ?Eulg?<7c$_uC25@ilB3wORq*0eu+3a1um0gdlC$I9-{F% zp7uyhY!l19Z7hb0!Va%pmQLmkMt}IdZ%BpU=1$ZzQSxcaTkCFp%fbZBLhD3O)0~Q& zyD?a)2ejyS^#Y|?0cRl*(>dk~iBz`||Ca^8?DlP>p}uh@z0v@V8})*Xu7eE=84kI~ z{HQ81xcJkvR_ix7dhqvnXHlEBuj}XiHSZ!p$cC+OLCBXErTN%*T{bI;KPxyJRdP`i zlCi0|=8R|6=aFRL!*Zhqm8rnmr&2lW!d;{c`rQxZJ;eK{lScPSftQo zUL|7k=SgJD>_{cOitv=v8%=l`+`hU?guV>JjF*ZHEIuXDY^5l!SF={SOh>1rEnx0k z%#iQANp@`P_sgY8bbLx3PeMl6vnN%uDQK0+%=rv7;KhI#ogQw=GDC`6Q4CSmIi|{n ziy@=lZ4Mtvc*@IfdWnJ0=)}w3#X_k!py~vK)K*6nPI4YyZ?Yyu-q8Fc&eq-66H|Ch zUZ+<&RVR_)kf{=(D=pgwu=56k8u}j*C<<#)`5@tJ>J=^~F#&4%jBXG`u^OZ|Kcr0YI(om*HU@;)Fxt)%rz>%GQ{CpTGXDQd1lF|Rg}*kS)h z7Pa=ON{hyZzn{lN@T3T$JX+stj;jl;y?qWEfrgJpxo}2I*$ps&@cj^iFX)IdbFj{$G74Ur=BiDxNn5!XCSO7EU=+AeWiZpb@5iNZu?#ZIqoD4k+~r@@Z&(8TR?$1qZ?p{sOHAAK&e4(ofD1yZ~D#fucn8uWjp+Z=Ps)iSfnnv%fcO8*jutnizm*HAZ7N-tf5bX#j=$ zN|mGQW~_@(h^LIl5G^wMT+MTkIjoacZbYFq+$QJUBd>H zPWRua9v(JT?|elAtDWp?(qDwdirH&NpCpU;b{qJ z%PLln#20kbstyyw$?{)3gQDG-0EWB>cT zWTQ%Ee#hpMHUVSKZ9>rzH;D{LiC z`5=FevYJ)wO5Y_Tf%`)d7pcj;Iq6L;aX`d1PlDX)m>B3+5Zqq>9B5vjr(kYNTuyayuR|IV|wNJH@_L}KLdT3YJ`D+!{36h7N*gAWYp61 zk=_OL$LfuLyKQB1{oX?<{u1<-P7(cPIpgi0tN+3;NMEDnljd;4OW)BvK{w%mviM4K z0|{eKZBP(Z(s7V1zHD-Zve}#H0>S6Ynn0-N6D?Rbn|wbalsDSsPti?Seno zR}Uv~$%VxD4UN4-FVNAk?`H#j=5K~*-V5a1RQb7T5fnJ({Jdfqg~Olq3Yb_Yba)D> z5~&wqbBxlG>b#+fM+ioewakDGn0xHf^ z@X&O_6)7{kUY)OT&$ia?f$G5xiR=G!SB;GROVj!}5EQzQq)85kKLLjt4{U+G5Z?kd z#u_%&hRT!wJ7NXq1fQL0jrs7j4Bq?7$K#>8?bS3H3yxU(DLxy78;GKyD=qkhy6;Ezwh82OwfXk1K)hHS=2~u{RS7@f~O`W@?;w8ZOU|5lnYnn zOHS#bgx2F82Vny)6rXxI%P&E?Dfo6@3y(aS007ZvZ1PmD=ovoFBaPWNw~B&bgG~~q zV=a-oTOxt5J{oUNlU8QQ$5un?yN=!nUgK@GprT@V88%zCTyafh$k6EYS;BS5Xy5m7|m?V-SNN1W^o<0aB6h&BuD;99s zX|p2rs4%wvKNY1o+?~%XQ-0XK-6~2)##G3URFVIw3imszE=s(lBK}ZCJHYAoD|J7= zpLC{26MtC4R=qn7!=4UTW8n*GJ#zu(LRfEkwhgIyK{EA(n|9E z?I^eP@z|5|Z^!5MF;}U+?(-z6u9+B8ZydE3M|fh8?cl@Rj# z&AQhV`uT|t;@~JHSQL)~K_vmfRQRUX8}mvfW->*VN28v+#)iZPWJX8$zdzb`ndyGG z7bapO`?y+xG1_U}Edk>E;&}X#O&UI6_p51nZnn=n2llZR+~U+klbIAN%_kBWf^S+NFKx z7CHU2$Mmnl+op_|fIVKtg!}P34fx{m%|YLfAnO_tlO1evWyNwe^T+>wM*!;FDhYG6 zv?=N0W$M3fmCzPYNARzyEMt=NMT->py4>5=t>IbEC41oj!&i<_q~NqxfdfB#!GZLo zo9e#{DODi;SIcnL)`al;;oV9b=w)+rKJdBPxpYL?|8Ds-o?n_t`nk0(ZC?C!snNuf z7jhnAdU*42ey}CvYW$S8L4)rZC<8w2>pZ)8|ur=%XU2}}6CzGVWbpXgdSkxiZ zLSI7H^AFG4|9`6#28fC+04?H8L)bER$m#WN6Zwbo=~lIUtAFhlW`L(b=f2m0?%#OK zj{Zu|tInCD{Vz>_J`8zX_QH6XKRkii{3J(8swn!KDJir2~{4yvJvMlOBg z$RVOu($-hg!ZqQ(ww4h*;)#a|OD1R3+5n5&5`{Ot_aUn^@9bOyc3zK97V028Ti2HG zKLFvsXfhm~FC;_Tk!Sp>!pdSyM3W^3!@;CF*FQ2}A$<65=1XpzEdcv3UR8EcLjf7& z!zjwj!{q}meit1JgHaeIU-$g)3!IuN>>Zw7fz>kmdda3L>n?Eu+*h4;hTZ)q`>e$l zua8b-nYDU!n5w5>z@6pKL@aYq-h>7_HFU81M+hV}cUF&wW{`@F2aAI;|UOFnOIHi%LI;B}(^u z*_xk}5cgAuQQChuNj;J|_4dM$f~7-sLVK<8jWafr?G2No=S1Y!qyF2#QAWvmIHc~r z_Thwf0dqExYOkwN;2wSddL}1ow&SgP`nZyids{C3f}iaS39h@R+qcI(L2541R5h0% zu43f;m zm{46)@*qdGT2$GX*@ih134uj^r!ijDQ)1Hm0Np(OD6hRYB2vT#!BQM{`BLU`P>v>p zGouygiE3>VhG^7q5i}UZaQ*BuvrSZODbc`in13H=EoC|%j+^?DrAQ*n4Qh>MRtmdd z#ZBZSV*a3`Q36f0rx|!<|DJE>U0<+joeW$pl8>}w-h(F5UD9-%)0WSfzIt%Cum2>x zp#-m2q$qv!pJsg_Mf2ZC9Y6pRSqg0gC%S4=YUIy1v9vcUimO$6IT_dWdFq@}sxuz8 zo;pWa)ek8R$4(SC+BX+22nJu6(b;d5X?N&l1jl-%pR*H27lGDpqe1}4rGMl z_W{efgiJi&lCHDJ$V-Rl)R!G7YdP_to6xWn>+Q8Tg36(o_deUl+OA^$$|Ru2dheCx zpGBNu8>1t;IL;YNi#V%Z7@M9uR>lfhH?p_L(#=mZlidWAyr2-!72tlnt*l_qzjq7Z zHpK79b@=?!;dO7G?1*8cj6_iYij#Q;iwRtj!v+q(m08x(+>qz?Ml^g!s&a_R1tR*! z%U^7$^_+bs-K}56%qJX>acUTe%#&#b5ariClyjYcg)}MEIoq6+`jpg=AQlr9D(~Lv z>eb_l?-Tpv$&frSv&;}yc-y&L=XVU<^isA{LWiaf_d!Iw;kafH*N6R9=dRGeqm1pGYpjih7TrAn3OZiSeUy8j;$(!JY!|u2DqCe1$w&~FG zJ0~R*{en&sHev z9$$C@HheB?DldQ>#Qy(PrI5calN@uDov#qb$L*R{xE186>SC=F2J5m&c2yI+b`oj!ILa-kmA; z)YmT_sxmVESoABXTI%kdp=~};l2Z9ye=XykiQ6mkSpd2GXjvK>@L)Y~wm_Z=3pugX zIWG8myZZWHD~mTybHs+9xQ7@e0H8l!=?FemEwmWY8>S_`{UK>A8V$S;ZN(pbiNSlg zaGm_oW+#qEgr#NP$>moin#c1*vD;Oy{y_DA4u)!H)PBcw@B~ONX~m54B@Mwlw9-e{ zFF)Lx?*=lvPMaReAKtxc?p%CrsG)&kqhJohekrfDl47IjaNa-6g{?$hfoSnM4V;;PF8of>T}{mf<*Rh>-EMJ>FAJXcZ! zQ!^p4otWc>-iTZ0JJ6_P)A&}aS!KS3>u-xu2s!yp=o4zT{kJ;0xD%PtXK8G* z+w8$R3?XCXvyE_{SLAf;wODBK2j@iZf=&Em8JJ=bcd@?23c0=(jFy@? zMY4yElHdaSRdM6DWSKHdtJ<)LS`8VyZXcMs$0_1@>E+_jjXNxEeeH8~BpRhOHt?Cb z1b6X2MY7WF*1X+xr36jRY^|a9#OmAOihK1|P`Kj9Xb6e21z0E{!$JHectW$76H?>t zvXJpj?HQ{?A(!^f;Qj5IueN|6Ys|P6O}!Nd9J>sotn86QW5ag~SXyK8fiO?%G}#8$ zxaDBS5coWdoBJJCZ{pUO@{s$O=i^i@iy$Gm4faRoTaq=Y#*rtMdScgP1T0|v^v8dS zC)sQ1)Ssb5=FdZ9IT&S1WskVQ>H%ucj5E}rv9c7y;yWUT>R`YC437p$tJ}^n`EojH zFfsv)5hRYbN|==BEWOd59nVYaZ~|HzGSU@)p?n+j?s@V|`fD}oA;%=nde3LB%WlTo zq{&oh_5htYF5w)57d|q*0w~+m$4jXGvRJNiecxe$C#?{El0ZyvaBP@$G0=jj&Y*65 z;>ZM^$H3!$e8M*nC={W|K2dD%5x!2*D>b2b$ytL`e2IR<#-&1=8ugxw1P`vL>#V+{ ze@=3k&kvY@%2SXK(UKU0Osl!8Ne=0@WneK=Yk>q%iTFfe9@TjF7#-nGK#Pxb17N(Z zzTZ9kL~Kl6RbTNs{n0h4|IIVY_N(gy-{a?*t-a~$Jh1*745(6c(vU&y&23gJyuLzI zHH{*bVpOG$j$WNJ8q(UzSkVlKP?_KH^ZP@LDGU^1!#G(+07M8bR>-b*TPIoUDk;({ zS?z}LshK?4*)CUnNWLj$bpj-2us1xCxi3}{{genxwSll|sbDtnQ?n_N zYsE*xQ73gxtP%=AB`#jcTmyQoC(CqQUGohLK@*w8I&LP67Dtl#;*lVU0vhd7Te*NB zIs$OL8~#L2U-dguJ}$tuAslk^52JCkuT?2g_*B}=x7u_RbN+*glV4n>Oig#VtE`CKONpDfd>3#q@uW(P<98B^NVkV3Hnv z@xC`^sGK3N{?HhWyy7SN-(TZDQhg^V5G27XLdx$iZDRbz(CY{R6IWq;ebtJdQjmDB z3x`9VwS98fWq4RC`~)~N{l4AYlS1&9cOR%>(3AO@j1}qmGaqWBD{^WHTq_E=M8-vT zjRV6`fhZ!tA%kOjc{lOjyJeaV_j8rd^=D?QJq;Ng3*D+$Nfz^)?fya=bENJ80W!yj zH(6Rp@gO$BlMe4cs^OE^MLhN?2|ck6 zylUB=HJJP$&cL)JvprGY7agW59;`w~Yqt4^jh}K!LOvm3!uzOn`w{?L8P>ZzIU88* zw4+g#4Ms#r9@{5I8$)k^U1BK*V!75V`@@ywsZajZrh{o0-Q^6&y28-$_{nx7u@3*` zv-$M0Pk1)&q?OjnlF;n$Y$ob$Jzyn8ZuKHIU%{BWvB0 z(W?xXkfvfztGPqkh*A)$l7tM~e9p%8{S;CiY^=?d&_m5aUE3k0BZv2D;#r&EbY@Qiz3m~?p179TY{5XowNVBxlSde@6wc1 z#b}PHgK)VoGC3^QbMRE;kZSy@TrAZP@uFNDa{V#Iec|E%0Q6|bwoa5 zoA(Y~iiY-r2O?r16(ivXWC34HyhSjaGB&=hy0FiA{LHv~?Zgq|d-41QW@rDDUsSSj zjE?>2^j$?))Ai9DycmIqvGA&Rg6Y`y2 zOGCCrCREppQg59}Hh84hclD_2XZ-8A z`q78J`ha-`@BZ*248*d)5aVf512HgQQNhQgxoE@%_VjXROSUa<*}fs7$sXKVY8ebx zK+7m*lCi39LfD~XY4SsAEEG2Xnty<`=#ZMQ7qFWx$$l(q+>e}RvxMmy3)5DA?vEK1 zcBGw%$Uw1?%;8LAP`Y$krM4c7eyiqK>wm*VN=k9>#2%r+?1hw{2;rv$(pDZKBc9VB z&78Peet&`maOZk};LhO*TFu^+n(HO47In}sLFVE?0-fn-$kR#ZRot-61}nBmNNYht zcZHBPvkDpBH~IwK8sT5vtvC@^DL=R9uBNl7zcAYu)be@5UKywzkPow<2KCcA9EHvU zm|_K4pG3m@R!{jYhLCOs=vMtPnqpEQO<$;NH?Y@v9;is8 zkcBXhSAwyW1ce@5!D%9zgKgu~^jHnlQE}g!N*&cJEoEzH&A;7-|LtNkR*ef=b(JNZ zE>fDLhb0%Qmc}My>8q1bD1=pWNV?L~`|W0Feo&&0m1QW$5oyayF)P%YP~Y660o
=FXnHB-g1a&s`Nz^xIIN8-LdpEddXP~Tfy}Ed^ ze>^(E(Q&(csq%*Rrr*?CwvF>OxiqqE;&#WWm`A?^b601|H?1YDnvMbyGeX9uZv0h8mXslC7O#B~ zO=ChuSNn=psJ@TCzwvwBypL^S($dQK+tGS zlp0$*JCUy?6)kN&YJTHz1F;FPIA{o zdVWXDCf#p;+zGu4Wjy=(@9wVFeB@r`!I!hVZzc5enz+t$#!on^XF2}vnffc;zl^C{ zmWl6w4|lgMe5Sk4-MplGTl;@SD5(Bf%NthJhTlz^g5cFz-QKNFTNtf}fz`Lf=_>GD z={SkMv|5*I5;b0VnT%kQ6u>yo4z@uT|wH z`A@Jv5j!JH=IAGteMiT8b9{;Rr4I1ekSwIT>@4w5wPT}w`%N;6B2L}FqpiB~algRx z@73$n?xKs5w+zX=eSPkQe*#;8aA4PJ+8|x{_)SwePfMtCPe{l5n`YL?l=vZpBCLOx z74`@eW<|5xCPOAECu=2Gc-r3p#C4_H;%I6FOzbucoEVpZj{o{5A62;0-dnrgeaO{C z)BMLpUVXS!|GcCuaN_nUvoZuPIgx_NyGKWjLO+b}nOU<`f>UrLG}WXi4MG&9|JyQe zb0bxCo-Ks{OzoK}FM5{e_oP7Nuv>&cm+{#!>CB(W_&?zv4y2a@U7RGUG|=>txwDIM zxog(A;DEN#Az&qa27c1e?rpDig3r{Q3qPAx?_H5KI=Q&$zVp}d0DpF6s1lnBCJ@XW zzp<0vLjT_F3C@9&Pq|#msGcFlY&7Y$oVD$bSAR9^^C-RxeVuhdaquf?yvtz;{~n9~ zeIt80f^i7Px!FI1MsWheN8=Uw{Z5Fp7;m5rZ5De~BqW5jZ3ut98zsrt%MAS`8TcPsjKv+qa-iuETOF=|&pXvu6Gnri*@Bo(C=+ z8MAQrUKI{l2SBBSZiVzG6M&a0iByMJOlkIQ@sX5l=44QBqA8aU|A|w)*e0nXb9EaM7-2CDG zQxt@MN_r5s>1@hu%aX+B-k>j;dXGweXNLuXzB_Q5cU+?SpKSHW&Ehc*MWDQj{>?+|BsqwcFrH2-mfTlhGoKFWH$?APNxEV_ z145lO*n~I#sV?!)u=3$yr;MB$!UY@U9c0rFv^}nA@rQdhFnP)f1qrpW3?SToGS)C# zlNqs_cuI4h2!hL2F+sK~QtXdz#Yz}hSZvY_O@9fdMy~HV_MBh@fwx2Z%1m-L-S3W3 zdS^meBbbeZoTRk})KI2`ZSE2^m0DdH#?K%lUhXP6d8W4N~KReESo}G>es4Ai}{WrsrT-i7<*pc!~Z<($vY-3+G_iO%E)jqKw zP3p+8U&)E?N$9^@)Q{!q`LW13m{5%fVC22~pm%7&bA7)Tt>|j3{Y15voHHQc$B8Z7 z_3+XhH&vHJ&asX~(*2IkE46%$xWe4_Nw%m}Yrquf2yhzY9+?pg-2uwi`0&tx78_%Y zXAHuHrUW@e~(_SV*g`3k$MGI$X@!?K2WQx-?{d*mom%9BYSCDcI zRt!;B*RuDFa*_ivpf_)7aJlWXHt9V72?JMf_s)BL^V_(-tCfwa$W zK{ljuV=mc*@o0_1KXXOMudqrWZefs8d;2MAfy+s9Fc3=I`umFiB%~zNILPL>3q$u8 z$(^aLe3;kOr65%ZlI~dcY-)OVVtVuJ`=sw@ z&emXkkEIL5+s>Wky>t8r_P)MBiyE4kq;#IvKPnXKc#o$@Q!fW}p$ZG6EfS}nlRJE$ zvnd7m%zXZw3e0t2^IdzlSUW1(bO8^uk;#W+EpPQh2R#Z5A2|>oS!mug1lL~GJqSrM z-Nf*VOSjNpzu8Dp6m}-k&UmVGiOl65>^0c@$Y9)*)T~YC#xREnY_dBK11l#bzo{O5 zG$vqzppc+GhSeR(lw6Uqe$eY*+Q8F!#gLhsYC0OHD$YRd4Se#7`&+<{%eMv z>u^}OKW?QUdqd%i;{g6+C>9aeyC((B+Os3*Z=jj-vD?3Dtlx*X=YqhAAjXFQnv9YH zJN>38*~du^w0uoCqUhPQZ+wLRrPDzy8!&*hyBqtTbeBbO?1z~jkvK-&*ULdg;|ByP zpHTAdm28Z1n4-SB3?}$P$wktw8Vh>x5{9=N_m`54p`81)oo9dW@4l@bRF{Z_OAQhW zDL3-H+{`%b-)!Hpm?H(sP&>JguHLTM>x_3ak5ZnJo$!&p7%TsT?n|i;EMURdr3ti> zq|z%om8Mj@ILSDhm*5?FR0d+lQ(Cd9VP%;zSvyc;X$XzC57Wfn!~V2l)h;S%eCbylO_a{FOpFUAEB_D6&X}ci@b*7)^iY&r3zn{9 zvB~OUY%D8_q#6-Y#p`+}sc6H18aL zy+5To`*rMZ;c~sH&PRX$+um|P6%ESTJ+qZxT1{Im_f-N5%gWQDx$49I@M1_yH-)KQ z&?ym=ZhB!G!{D!f4Dl?d$lHJMS`W;&l}KQCn4l+vF`*VI5IU!+DPB+``BdxWW?jxm zk)hStp70~`HkDl=8?FwHwgToehd1UZAF=sxIwUp$77a;m%BkO89!Lki^_J^D-FJ2K zS(ix2N~Q9eVGtpMSLwcAqHS>H!eIo0*5;~l;iVdAfj-bn#i{RXaHmGN zzCo`5iAziY$W&zq8aD>{lIv#85qznxO0RNBQc6BWn2X;-&n)-%SKw5orNjb z-)IWOlI^){yIzSg>xaEe+U50i+KuWdBVYj{{MpK~)MmNV^eF`95uakAwhn1%IaPw< z(V73Cc)!~eY>?PF!?@I@13f|}gKylCih3VeO607+4GuZJeO)aQzEQq9Z1f&4^4OKq zh#*X^RJ!np6^u(gD4a7gm4ya%K^n@r6o+xc8W0@acaLoNN8O(x24bMfU z2(4`Rv66mzBpW(=(o;j&oXpMm2|wg4VAA66=8q{i;MG49lrYOL+T5iBg!>ZLdP;bI zb%#F8$ld3$;ogWh3$0uS56{#%n9$00$S8~^OmmTZomtt*R|h*1^)HM{KFkyr);S7k zTpRrT5@_R?0F1c`WI46t!zXJ^#NLu@A_|S^NTaZ*M+s|u@4s3!Ey{kp8LnXp>3ZNV zLR&5wU%rqy*H+7cdrR|n^|FMN$W`b59dc)Emj1;q+u?s65ic5L*by4sGiU(-c#HwT zI-;jAq(uC-{QsJoXD+tq$d3VkN6podsJOLmEq(Z=*cDF6pYb;*qrTFjd72Q5ngQy5 zo7#^YRT(ptK?mf`(z6FG^QMiU?@`}&$7JvO|Ki<&k53xH_zRQ|^T%F>u7a!SWWI(3 zjF%szDtec49TsXig&f_h6VBd+^=K?ktwjS)s<6LVS)g#v?p=BDpNoBFD%we+Ov}c{ z{|@k?{_TVkJP|x8lrud!nJF+JbuZ-2p8Xo!@*p5j!4Q%5y=)WnWiTp(%u2AbBCY`L z^hrF}dPLc&%1wiejaGxl*gu#k+ejo%(K!J=#(}qhIo&Jajfz{8c`%3H#6WO?BFpw) zp)6g%alm*r)L0YtQ4w7Bo=#?A!mD#1a7xRyo1d<9v|p2)9ljB-XP_KB%hNX3Bq}$Z z{KS-n#!H^#C6D6cP;x(!{SSC+P<}}cd%J(~sI)S+d#g<~ZV1!3c^?p$@r)XKLiv+F z15{#U+T%-}S5q^Pt>94U5}n{u7IVlgV&5lXC>s#)E_|gp{!4whCD{ zO;o?ij&B|yQwTu7BWdbSI6`0j!G(U{G}rNBRpgB1Ia6zKdWi{BYefuSNg8wELSjIS zHL#B#!776SN&0V?L87nP6V`tE;w;SS;;I&j%69iQl?BvUFVpzhlzO)^3zWG)3)J=8=Eq8Y*h0^mK|3AS>{&0NYu`N(i#$-l}d-P({?tttHy>>JCsF=q`6Q% zR$I1P7HaKzJ*TC8iF#>t+aq)t+);s4FVQlnyn;UDu8m6G4!oF`DeDEFlQF+18;0 z$2PF1e_@=87&5cL{e}-GiYXz zu|gJ6nKLtr0{2Mm%jm-wXK(e~@2&57@KAhHChtcdfkBw$(uyfALxE-9@)(vH{I^$HO<4}Yd?@a$7FRW>#p;!@{G=Di9>0 z;)ARE?3cuQyH%f$k`{)VRqJ?s+pTRq!YzHBX;7V4QVFruemDe1O;pYd%I11Mp61Rn z#NDuP1!@{IPEm&cU3GMgnf)C_&S_ilOb|-0_@4n7B_EmKo;plfiVaSMzR)ZqQ!A>^ z)=(_{Pmwi2kt1F#eDnX&be3&Vwc#38Ne;SvB4J}PAIs-bq^`Hn)?HBupGI!$FHRZ{)xeVDt{+v`1D>VNZ zeqe4{s8@1!6x=z{2HGCq6>S6t=U54LTMMu_*gw)0zNL`|Gz_EjoP6UxnyPbn#Zkoh zS8&%BWVXB)9=z3LJkKsS%*6dWb3=X?8=mx@>8^H^^&I(1!p z!KMf@u8LG#7+NWl*=H^O%g+`*OJ$6GeW_YxIU?3AM{wU`nh(-2Ske{(y$88PTDZM~ zqCF6O{}${W#~e%y-@F@NJD8Tcx?4(;N!1nhZT0_e7LmUNF1mn|14k>o1^0kEN9YSU zQV$83vuw4SV%Ih=-#vJ%Ax{u+-z!<|Z+x0~EtE&?dDnEHcdq4g=`tF2AN%EK%YH4^ z<{sYPQK0LgD^k;Ylt7CtlWHr$tYIxT_y(wK(1#_rV{F3{2QG{HW!(4KeACWs zzd9U+@cVb#H0xpb)GRyFp;ihEdL1Bx$b85nMDv1kOaHW|&Dbw~z&8Zmy^eldTiX^B zm+gM+r1~S#7Kb)HM+JF|xCWKd^q((WXkZgq5=vAgbfD-T@xR}ik7lHKk?U0w1r@Q% zXHUKdQ>kzmxhVAqr!gkc~z4`#Ne&7Dt=wp=J;V8<**w zl^k55d_B^Upqw`(+_MiFz{+HnY(z>1cz5Gq?2*0qD&a}!rny`a-=^rxPmcz>hqtms zbO|GSjiTaUhG;;4YoQG0N|45lM1##xNR8FJj8U)j+ugH(w*ag`JOY`T#i?gM@shs) zxCCmObTue(eld_A)&A<_)&8?Cd!vb^aT7Het_@u9Zfcwty)_aomDm@z39o zjkABS;X%xHnCZiX0klha_=kw05+xl>-?*lwn@`eP7^GXZLo`qM_wm$na#lgeKm8+OKk}5cFl4dm-UxA(ziFd!C z7?Y8vB&O&S(m4|3WXYAro(!+u8R~S<+#M2p3j+qaF8}8J(PZOeZc}~EOwF<4YD^$b5#dO0AdAEyxIYjmBR0@eO1V}(B*b<>M2qYYcF|=#(haYm7m4wl9A&ZPj zloq8Lm_(~;ku)*jDTgiaH|sC=Uhdd-G%a+I7locL_6moBf6)xpOZg~M=x3@|f!9mXSl|c@uwX%?N24{u}s$vPYxdq2J4Kgs+=Q3X3 zmEV7;KujK16DUlHo_Fq`AQDnwior_wqz+~dksHfACpHE-rs-kM78dbDO0h7a4ilt4 zX8fiVkFM8`yN9*@CPbOAFyede7!vle7J|OdVA^^f+QQG_p8#?8Mp*@(>i(*x0mZ0z zwwVxOL1?1Iz>2@O;{BSCv8FF;+e5hou61g)i_ZwcsM#iDAlSJz{0DUmO@nJyAd|Bde1LO2@5nJ( z81AY8(V|Pt-Ci0`F%Vk5KC|q9RCj71F&N%-rNaEMa2lt&Pcpc{Qr;dam(QoKzFIE; ztcK06&yj$1+70U_AfJ%Tr7N5g)O9E2O;{Pcwa#pFVC>E$NMb5Zi%KJqn84q#cQMCr z-*YMDOP@Pg5miODYaO~>?{wqa)tHM3=z2T}Wzzx{^I#2&B3D`c-CrFsw}N^Yg6jz6 z6-cjh0`jc*{{&aAiYbq2QU&DC5ziG3>_x`^zy)tx9W33lrvB@gBJq{5M_cFh$`^6e z2pRbK{6njA|3NJ}g5fy!WxcK;LW2NHJg=~?7#t1jFWlL-9Bn&57ikmtr;72}KeE+b z?|R)+`TVorwZ*tM!Sy*eZC<-{SYgF~yd#3|1w4S_<2jOSf`$p9eUOYB zw;Z;k^pe3v(ib!3;<0ccmF-~<&0^-lE=Wk)XdbBb#G}-vq(u1nYZsw#7^O!gZG=^-ldl`s>|&Tc zR+E7w6PV(qX(WXYXi&22{P(sxf7GJH4G9`ixhN$XV@l^r3iCvE*H8D+S`2>ULX-}S zNwN8ArQSlhC{&O%8HL6wDyNtrDwjqZuMu;s8}7N1K!HVN9|TU3BU)WFMJlZnwJPAE zl~l4$tihI(6h~XE$^T7Wfi6p*G%ab3p^b`OBpDekiA9~8Bq|*?ol#rrRFYp8Zo!T{ zQYt1Flay}#B`uJ#XYjH+1|w|Nfv+GPhuYHxTe5-^o-b5ldFdORt3@*bHT4LbUfeYc zNG@0%xdIDhgv%veotC7XBVvH8$ct8f`i(Z#5fPUM|HQ_~ca9@ReO5}5bMB-K#&`4h zh0{gdH0GM9?niO~1&i&18j+NZV5aYvU4n0}{6}ZsO~2rC{lY2u!Avsp>#KTT5}{rJ ziO||%T}fpvR;oP8FsWWHVaMjyUaUaz^?AVuW4U)J_gnt)fg_AX$zoALGBJ7a83lBV zV0Klj2uzqQLHP8izi*^>#p}mRdkz|9^c2D|Vbh7h(#Qb-XbMAe0n3s@>q#Y&;lq1! z@q>tOX|+AJ2uTnV_0SU8U5GLPXRoi0CA}+8bAR-=ilubQoItbL6tnbI(Ck%CcT@A*gj#s z5PH;F=`_)F$Ooqp{EES4Oh1$!?ecY>0002wQ5QX@MCNsPIeNNoG>k7VLPWxnnk%tV zbo}WlL%o;-?M#apusM;$o|TpEf49SUT^%k6xFRhA{vaiqK9I5|S(n!|`(M0&Yu=;0 zC^I5IT&;nMjOWXrHO7ruA4oNc%={Wj(0xp2886Z~Nyk)9^L~RF6x`UcbHa_bo`l`y z)x2%R$-zT+@6s=0KgT9lL>Wg`bqC+o@<;pL+pS|K^_sWb(BqwDE5P{v*Yf#y`u;lCC8N4oE9)~45J2J(G3lC`x zjT1KkN{TLhdrzOWnm=Bjj~{s{vH;ljxaYt5B=gJw#REe9p+fXRBTb^$vnV`XiQV<1 zljN>4%|HPL9CG_R!L{h;lZmX;YMQ>GcF)_eHoHCIXx;zM0^DdH@TDjD*Y`YKwC)VE z0y!_A_cV0W_P@5K!fiUQatqTEE5#Dhgq%*CPkMMnIfpE0|m#n#z&?;2T< zOMc67RN=162NE38rYX?-IPIZWjyK*vFLgXxiwqIbwK%IgAIJ#O5&%}u?yzr~kITH$ z)pi1I9mUFsBe-McGeI?Q>TodI$UW?#Z~vlU*}@8e0rNtWa69yW=ggn z99CIliG_V_T5QPv#O@mvJ?vUeUUH=hWlx}1u5?@R<^`KTQ@h!iLnYf_` z;>K>h*@z>_&+))LsMdSBLj<&g39r{3!VAkTk3R~5AGW-nGLI3Y2iWvo}XKy4W~E02lvBr~z8xI;k5!qhY0?#akQ zm>9qs+ux;z!8iCs#yq+vW<99-dFqrgWNCq4#So)N+#aZ%t*MFu21zhuJUeC(O(-}Q zSC&3HAt{8Kju1&+u$t?%`Ymj5yKu}B3>QyI{Xy~=IjqDdSS&sA(0N!x5<5s@Br7|G znkq)JgqV#CH3;Jam5JFCoUj?GDt*|XO}3n3WbCetx++XXa)T<5p>-&Y8e$cWLW4L1 znM%rNw6!Ge`7Sjsz<`%L<#T1x--#ip$4|Z^X?D2}2Y%Q(dP5n?Ki&uGLc(v5TF zvtnZ{*s6f`;JDHqF=dPNszcwffW`OKt(}lTs`Q-BoMuNY5!e|26kj>dg(eOPnqAn7 zWk;hSZ{nqV87-o|jfgmG=(Am$8hJ*(X1(Ad$`Nf%mkC>vzC^r9kpy1b!f9?OVavx8?%K@C~tJ znUPdd`x-o3pAU}SDa%n8d%k=BZomecANH}3FKr*?^!i>|PA!Rh!gRhyJz>*_^%Ri# z87WC5*1Q4nB?Y&ttO{pC2)}zi3IA@S3H*i`+4D|3Vb-+QC<7pchwGigM=h&{fF+Id zyQL#sH?h@CdkE=1Xg7HpiOXI7tGWL&nqqp8yv}>du|k$9T!zZ%k$@Rt*8nb&n3gg) zdh3$QpneF#{S#g!$xRnoB!;WP*(^9r9wN0eZjksaBu4PTd_8EK2Rv9=Vz8tx}o@;utx=phRl z6!n{_Yn!#}C8^2Nos9hj6(u(F7k>X&B2~z5#eg0}09ay(ipEZvOkX%=mzGupn+mGL zg@TxLkqSQt2U=jEmf;BpYaJ<4AxeuXyaizdG327Ue^-FqzCFiAag=F!Q3R`oW)?nt z*jEfeL*b2i;HtIYmWd*!piVuGRTRlB{K`o$>@n?VOjHv6gFRtKYP_9C)Wl4FYd4R* z{3$z*!ho|MGPbP9rix!n+Zxhs@!P22_+1LmWn!C*y+WcNo=_|C!`uv+@_hEchs)5( zN}S@468Fvd(}A81=GK_> z{5X1PG%)AY%fR;eS&Pf&pux;cUiF91|2UV=eZHz)OF|mQ z=3V($^4wW0g!JLhvEA1WGS73#Y5vL0ZQcPaz#8P|;XRPJ7Hj_9TrS`DV(`DNrjF*YTN}o(+B}9h;mPP!fKZMSvj$vkNsnEBJ*TSALiIk4}P{SO;d$HL}@ zlS#FcP*jPoL#G#bkZ;~buC0`QJ(u;wkdiN@XpFAa_O`^jRLwb_$>;hg1L>^ zfwkrgM~3G4$>7wZ)0Hv(acXVu{JyU<6kic^xfqkZY-Kd^+$%%iGG#;}uY?qA;)El2 zlu$pSVyRf0Le;FbCV#z`MJSkm?~-r+g``~b1PQ2W?`S$aHG3s*62K_|n#U@f4*BOt zZ~4aF*tl$)hxIiYJ2Jf3)1YIgpd_p;7Tw_BB8Z-?bmcu61E-=c#kG$R2dzxsrz%=1 zGiiGj;#!j5OZjy)ZK6sgY2o~^nP49LLS|{NgrSfXq7MVn;2~yJ9aW-gx#<`b8tUXm zJ-a)b;^0PGPod+#8q`o^S$&PMm6V|&JCPrz`;zamBX2)lUe>WR#pwAG=+YmdY zq%bZsp3_HAo@`%Myas&kg`RI0L%xdKrb*vh{e_*i@S%uEBXLtMSA9pn_Fop``3G8v z@4tx^iH@M;F<8eyG}vSF-k{QO52EBUl{fSyRInhrvKvaW+ad;fYOJTG!Lc8}=jFfo z;>@}N0wupCZrQ}B*Th5(=VnkQ<=I68MTcj-Rs2AnBgDZF|HIsLXeOi2eZq>kf+%Ze zN%LY!t)x76`m9^&wG6&uVkV3EMPdsve!+};F)-4FN*t6PiGs~pHfnWJ<_gj7^DG-Q9% zqqFY#Fq;)1uwV}>gQKWRPM!KfKy)KOvZ z5XO6*aLPW(vRcWVtm!3_U72)5?IF6{nzhNP8GhdlB zu$ROkmc4t85*9%Ms!s*_7oeT zxd=BiV1@K_4HM(365Jl4cBZid@8YZEBZd!!Klo+n4KOL+Ojh0_@NPAjUHlH~E9JIO z3q)o`^_dt#&_QrY2@5s@y#kDTrU=@Exd2%fRF1*H==;j*SyxdMDe0g$ATE$%I3ak6 zVCM@Hb~+{;r}a+&vPE0mH!W9=i&lxY^;A-wV4sy#kugCh80n~B6WZxGfy&Qf5CXyY|s6{!(y zb$cFvr`&Qz^Px8ec&`m|vztq6&2JsJ_6rOm3f5<(Fwr}d;wTK!!jnbgxcG7;2CW8h(2M@abdjGbQ4PIgX_gzn9*nv&=%l5a@9o!xp|9h@* zR5iQnUHU9s-?Z-B4W}CLogocYWuG;lQ2+beHfYdll%Kr$vH}q4>i2h|V7f9w_ex#R zl39vY)HwGZ0dr2kwD847CbYM-V=lqey1AyDUj*0CH=WfDhoXzH`H(}T{l32e%y_h> zc)kP_X5$_RJv+%4-Dsp`U&rs2*FC73N#p}v*`NL%3$u|#dZ!|@Uo+gNpxN~fPBykX zcY0K89`Wvp@9r?@#_B$tdA@l2yNN5fsEIS(4&0m_!gu+&j3+9Zpd}<~jE*mpL%iwc zo|-6x50>V#Y0>J6aCf-|K0OZgHzuWi&tX=9R3DM`ye~=`{mTdgq1=f%3W3B+A2O9c z^Q`s#Q%|qbYrcE>YBfN5Haq6h!SPQLj#$w?R?t{4DjXr6ARB8AD zVd)kA6gNNG+z8G#TyMCf_i+q)7xY0-|8v1 zl^@IgzyURDRYruscd0ll>({%Fc;0N)x6vdAGRRa*XYZ_Zv0X|zxHNN3Ocw5a9$Y40 zuKONKw@WA0HB*z%5KKju9P2-on}vEaoDz)bGg{B`LVI)8bFRzG$^2ioV$(KXE#VEf zmx=I45h0f1$S!IoXh+@J$-BFlvI!7&_#hD#NoP^G5k|;G-={pMZYD)a!{g8_MjLZ! z$%F@hr@D zq!$SZGt{b>;ELlX>Xs4bbAvqe(meidC*(t~CSr&x8f@;#@`s{J@=C0*VNt>mEj;rw zC*gVM^n40Pu9i_dweozp5FLa*FDJf;x<5P~m4Ye`%NCd8^RL-d!-(%>(Yw8Nf`No^Y}ale52sj0j%C7V z+#f6xo(Y*MwDp>dbSUhYrsI!C{17HLh3GGM_4a7HTv)z8O05MG%o_Wmy<07f(u~;M zo!Q@K7LhK@>|*jk6oMV8rQ>MhopL1>6nl?`K}||fo=o&X&tJ>IBeBaC^$2tva>)hR zF0pBFlJRms3h5Qp49_AH))hhIzp3|{qk2+rYUi1mWOGKn=^DhWLoi^=Nya`R^z<9f z9p~Y~93j!{DSZ3|xP+z_7*bl09?KbO$v8-$Il;n1T0B&>fhB(F{N3G4a0(3vgJ4FZ zrK%j(%MdQ<%q^bUx4=?BJ6_T33%Pt44vs`2dLPw+fejnoWTWjVuS{zHdX!x6{|}r96SMGRALvR#EE2NuYL1?0Jh|=8%#h zG=6M-NlQ!+JK9?iwPJ;idSsOTVBL85bZDaK_;@~k*?VWMzV1zGnL3V}6*u_IHT?0u z_K^a@|I}p6Xi-JfMMLEc&2NtFOgh5eqnBwquUhR(Cj=HYrZZ&12v&})ll@+vw_^Q$ zst<;Rug}Ku)^I(iT(wQvp@u56^A7bE5B)*hlWsK0WJJXot$&|0>~F&IG;F~ogwbs1 z>sR`Phx%WyU>o&VvjHKeM!$SxOt#t%ILY+|$n{Sidph!~Y^H@fSDl?%E#u7F4>^xt z`YyN2NGe`%)3Odu%*MMO*FpcD*x1r^Jf=|aD{|cf*7`rtAtr|=wqC%_zc(^ljqkZ_ z2AO*&XaM=Pxfsj7QjA*_ zFiQDvNW1sj|LNK5PE_1{jhS2EzyVTZ&(A$W`DHMQ4f)xBMC2=7CfvzI(}4|7W+@jhN~3qzu2bA6{Y9M5cMizncwueDQM?&##}k)_ zMuV;ti^@acIjfqJ=X6$!UaqhP_S{)62pGIm57e9fK|eB zuVh{oy{0uOfx=XnK8LU0M7P#J#@_mg#9hs79-4%!cS`SO{1x^0Y)X29%}%f_*Qn>{ zX(Ghx>OyKiC)?$j&Hshy-&t{!iuk6)FXG|$>8z?!-VB$nIHmoRwn02)U@KMmnf%My zl478afU(@CGUj>4>D&;(Js%G`gqPjsmL$DDLo;OoxHPgU1wSxQ5_W9>v?j}-^V7hB zmk3`G=!%y@|N5xVn3F!c;@zFvwusOogaTDb5;dB}=ZcF`mJ8~1J@i#$6cc!=)Y1A$ z5AR5vtkQk*Ac5)ZF1~z(T=*Z#R+rCzYJg1ZT%|5(py~<}IL1g+7HDH;<2~O{`YD}g z&c9nRp-||5{?uRolWTU~&k%#_Ao~4Vy7!PbhKp?pAF|o-C28Z>DwmYtT%t$k={@lxOoKT+>dem)x$G)=`e4X7%;xLt{hNPO{03xW1*FzJgirUIdWH3`58FK2g zfEWLP{a~3_$^zAO!X<={!Ag^((*99y)&X?B2xz-)=BKsw<>=qZgu@vt$V!>Tw@RC! zbTT}V=N=o~0W0avEZQQl2%0HjJe*(TEOcrfNt_mzJ{LKH9DTK$)x7Loiy593&!2SO zkrl3!l0*}ELT|!kDN@DF9{J0V^;~#YGUIdY5G5CT%A}YGQuY_xN?v%VO!W5}aA#Uv zDH@B-?ce)zz@#nAfuNWYak_b=FRJ<8h^NYJEu<3GiiAxYu9FunHXvCfm?~))45hLB z6KRPVd`R>p)cPgG#Qv&zDpw=qH$IGm-CJG3f;3%pX#2u1uCPxbEmgyCuZ29?z!iLo zs+!QZ|E-^rVF)Xq`{{)|2UlZ-5+evm_{yau$QN=Y%YoA!)#9QeND(;9FFMP2w&6k7_U$3!F4vQf(hIrd34t_#(g57lF`;>xsPWckX6_OVT!lm(Gf&a-TcN za9Agrm(L>B#Rwf8sb#`0Xoe@IAT%ahi0|L_@`))CR`3~U&K6Np8;xk|PR)kiy2TBV ze?)i9CoCG`QlLf9FpZZiA2D4K6&>4B!Dmn{MPo41n^8~sgjn!K1YI4fYKhGH)rnqC zf(}(+Eu{B;<`b+}8xu=WLX9`8!b%ArU83o$h#k14yyaf)>jwXJ4;m?wOf>Lp$TZ~H z(yHM~Ql#;sn)J&iPC>7i_PT=zvPY2Zm`O0JEx4c`^lU#%W@4Od4^3T#*X7f(O46S% zDZ@|CB2%fzVH=ANCBS9t(VAk^e)^7swJNF*n}mS+mV<+QOh-&K!Lk(9)>>TQK`&tr z0?HD=yPt6}1_r)@JdxrwYxLTtk*$pGTIf3ccyAvyb+dgVrsfJbJb*=la)>F&ezu*O_(2i zT`D^X7Hy$aRp9ztel= zBGLQfe^8I>Ez`mF8iR>f2YM<_8$ucemdDo>uMao#o4#dDe`$WV&p)3C4DR{>_iSVRvoGfe5hK8e;*Gw7rLc zuvhEWzrX#QSnoy>0pGG6~~(I(n|U6j7wJ@bN5$Xrrs+K~r-q`txd50h`i5 zPNhJ$J-iO1UAY{;^hA13a%~AF4m+I>Gw+n9nEryzAbA53IFEKMmTA<>U`#)UHY#0; zdeL2g39-md6znq=(&o{+YtJU|+GJqk*?HaW+{omk@U~9QVHSKwuJwFzcBRYh6*d!~@Ytfn9ykXk0~*rHXHTo=7F_OHXFjjk@PvS*NKjZom)tRoXYa>h!PP zw?Fko@ziTl#Mf!Ku1tvJkrr1DAz$hhm%1vIa-ro9S!y}^m#_JibD2C zXTxUh)G@{Ft-Wu`*|{6rb>bzFI;JSB`6fkEbg;eEK1+D6ln7!9c)9L&+n@DcSJ2TC zWe!nUIB8{25o;W#0b6_}FdOnGH~v{Ri!S}?$Fi5?q=QZ-6P7|m^C8C@bfH9PI=~p< z91HqZ%N;d`1vC-X>&g*gSoUk9VsSZM2(@|Ai)#D0KIYJ!rj2s}It`Dh-x%;Se1Nu= zw8~6jqT|x+ZmT~#)4AXKpzhb3@pNzAeMvG8ml_6yFpLbtt6ki*q6h>^bRE>-()el7 z5n2uq9;HNNyc*@6C>>p$bWH}^l*Tw8;7ug!h*H=IPldINuL@)U;0!8wHn;491p&A3 z#_L?n1T4#7V3aJ%Bor^=$cPskQcw!(_csOiV0QF|qgb=(;N=&geff4MM(?6D_9u`* z4!o+)+1%Y5g>I`sJo%Bvvcd!vD+D;-Eterfu5(hN)ECh^*;&}XxMmfD=Fk%}Q3tw} zMXAX!QN&3-R=t$e9r%@Ig{PrC3LjaYkU=yd#kGNrzAwJkOygsg(!8}RAEMPLBT#dh z{w$&Ko?w)(DKMSt(aNEuK_Y|c!cflXvEPcxX7SEbh8R)MZ`U*2kOY^4!v-mg=?@b0 zQd-jfRwP?6KN(UpVDKjfvStOVM}`(K?-?xsBCxh8p`OvkZ69+< zsX^P+WCS5Rbr3G1ev%J8f^uX4?5dSM7Zag;$U;AMsfS=Tb_h%C(HE-4ocb{df|gH% z8dHO(7Oo%=hnyZSKfF{PE4b`of?D5bv_4NekdQaQS-B8<_%2>WHlm6+@L_dTkslAq zi3w#-hsBaf1MQ-BXIwHkYf2pvi*qJq!Fotf2{xj+st2W~Xo)5+l*{9bhKCFCOV7eB zd@ip|aceO-k>N=95f#$2q0Z>8nQNw<`ix@?;;)OzT|wW)9jccUx!V5a*sw(9M2#Zr zgtJWO#430KMPa&j`T_Xl&4hW?A)SBFevYA~BLeF}Dcwlk99ysK_U63f_saZzeb=Mt zoLIkQAsQKssr9z+#lOawX=*+2IX0|Nnab-D@wCMr3N@A@xBO5mz_iGhPtZ0qNkJ|P z@5wjqIgZCpo=4WDDi9}sRr;upv@MQlBz*)PK`n%f2^^6FB0Tq#sexr!3YG{Cu>Qa&MRA&{~Q@b!CF{uL^CN;lMg0a=o~h?Cp*Ldf`FoqsN?+`n*Q! zW=j^(r@0_F@$Q^1_pwcyF0ZqP&8ZA;^K0IW$O#$Xb2p;rx`VUVpxu4!a`mbslk;=a zj+XL#yTA3V0Q!C!2ZhKAM_t5uR@16b$zy*$nKDm6ooVkl8%ob}s~!3H!=v5yRKw1@ zj(^MdLQvV)=QSE(uldV!|GDeoI_~q0$5Va!mqQQ7jsqLk7Iz*A>tR0r3H#ZN)};ky zSkD0m;|r42$8kH}>6k5DUp;^1X*maLjq6Ro7HYORFr~%Yz3{zvXe=hA@FgXP%KkPi zw<#yG$*W8R`#amm*FW!y>+doKttWS)(AB>EQHk7=*it3NYOdib$9J~#jfu)ig*K+F z@l`(}5q!44ZEGMw6*pX8T!?7k2#{9M^DT~@A*4j!)Elwnt=0E_qoiL&ko;aZ z0}qbL8*J2(vEq)pCN|bo*Kr!UuNYMeG~Ff>%SsA%a6PEl5o%r1$}fxu4}D~f zb^6O|5Dc4%Lm@VJ3sSnqeftrP;hL^^P(Gt{RXtQrtz=B|v-ElTs^krfl=;|pyOm@q4n^wg7kMRR zsFId5xq)l@;@yp!B%?$Urm<+03L5bT>A>1}={yQqx!n3(cbE^O{WOR1#TfDMl2zy| zp4t$CDr`D?ypd)M?=fDm@iPrpJR)=y(uybOmw^m|6$wqnF_wJNP@GBWqmtbY%~n9r zfxJTGR@Jqu#fkqk{zAs+GzX_MnJVc<#N%l_Ve!)2od5hNfgIrI#798bVcxV32%tj7 zndu-_1+$F0IL19x1&7}fj3)RK0wuswl2rNc;)&& z$$r?>1NF=dd@8ff^G|=MAcX5vk8aT|<8NbUw`TG|iw-izOrS+yhZp-KHEcFwLT@rb zD}rZN=k}o8Ni!M8lrfStnfj(_jq@hEf5gdLV`PJsN5GDm`+9kLKJ7Q0WD@(GpzcdZVuu}=G9|ECgf6=OJh3b{ za;>E=jlv3E;-QAN!!|hr1L<1f(~v(h2Au@ z1dS~jysD>3>Z@2+!%^!Zd1y_pCm`(I`Iai3w{VkrK<_tZ;=w$`@m{(gIE$YD2=5Hu z<3|fq=b-ysSd1@~?z>F2^Sp^+ckamXXguPxMLf@$dnPz1rwZA8blYuJem?wO<-S3K zki}{}-`jmX{t$EK9ea@SZ_B)?8Te!86laf{hPysAAp{#c*ZAYHE(RNCI$2!FmqH}lS zH3=jSV^J!&lQn+IZQao*+Fnv^;aH>}emCAg*G?`mcW&)rfB1Fdyfehd&y<+Hkgzis zxVEVZsiuI}`VO#R&1}bO@oA~o2y5h~pyRbb71q+T_3ZI+COdxq85m3X^q*fXqmL)1 zl(AS461w^5Snp`2c|>J(Oh-m?VAbUa*8}ZF^jgyF_T~sphKB;nP?gN0`nay#@X3&& zNs+egDGfvEbsaLt<9gqGA`yEgKaYRW2e#vbaXHsE0>Ht>10*lp)V7b7om=j2$eE1U zd0=qhGVNSg{>QRGI9hBv!O!wZKjjiz!MmG$l^pzZ19nkZ>0V~?!|J4q=k$6IW^-=d zSSTA*yRFqiFsy5xFK|6)-ALj-O1j+BILs*O7dl>PEmc6*{`|491k@beUpJkN*hbNGdq z6uXg-weN@pwUt>C)CF%XM`~DL$SlK4)H~tydEwf=i=-+578sMkN$JEF)M{ll+Fp^K z-nS{J-&AO6MyQY0gbNTkfE`!{?pdMGPEYX6cIGg>NR%7r?f| z;P^;AdNffY%-zFqpxDy0VNyaxpdO$Q;E3G6h@!^sSV$XIQealdjG~WC`%Vs9k2T9m zACpK2*IB#RM|&xj4rJ*9Dj5kGFGtWW`>AfUfyex z6Pp(GoGw2OxOvmG-DJ2|B81-!BO4H9LDbTx9*}JN1hxeF3JYD#S?WJi`%R%+YLb0M zN%uOKWs!EM1xZuO34uYM0_+HVO#~2aOsl15G?7Z=iMfg-^U4Fy2LjS3#t3q9*d&mH z6=bVXRUs*#XJ)aBB*ApMfG|%^41i-flw%Z{s%j$VW0p!Up|T+IGX!RWr?7$s+fWKW zLShvvaR-tEFo$rpCve?knUwH90&i5R*5Hxs!BpMn>+t)Cs19QWoW!WHa^PpTWq!{3 z2wbl}uA;*$&B8BOEm7?!2!8IHZMf-p`W)_+#^?n$~ z>6i8>A$?JgG40|AwKMf!Bkh$^7d$X_G8UI6M-|$!ov5VtnC?*Eapg`>SSk=%C6bnH zWUYg$b7?yNn+Lc#c_d^6nb}aYGg4VQiesb5Fa=%|*0!O%3Xj_LsaqOQP!zVIwWj=V z%X~qWLB%>reo6V3*EuquaLZNnYeuFFORDbRhTryYIVT`GHHmgG`uF|8+UQ!JSvV=% zm-B^1^tSW{t6WtaWk1`R3|BD<9(e@vh z>5ili6HFw|!Bx;=uz3r&IQh(kWcX})j5Hlw+v`J1fJsY#(9*8QDSfDK zv@~mhI7)2hI&QR1ZQ5dJfnNS(bVfe@(Z@lO;b)AY2wme#-*7IbBytZBDAxQa>{C-b zz3|*5BK%KW{#I+Q)%kh*GIr2<&@0@hHkQ`#18*yi;qpvBtiQdVxVIcOE8QVC+OBTO zT_;_)1~PFduq>F2U%nDFU@B7JRq%Os=>;5fHIXBlSyjw+f13<~Q4x5R4s60AFen5M z{=M#Iv>l|+nD!EOoAx^`{Rr;v@2e+iUX>|lwQMm}tI5*&GXiMi4o@tTW$1hM_JXOM z%vsa9Lz8B6Dd}t0lnB@2L-uUu9JttcujIs$6QmbbwYZfVValvX^AbpcqdNN2*>b8x z<9-g&!ZY(Jt`x{9_{n`4sNu88#q-=U*9&n`BNp@o-=eWg>YLe^?8=Fe02-FcXoljS zmKycv6K|jmO#eBjW%*?3CScw~GW*hJZ6+o}6~N#%yQ^Q;os)N^oRInMI-M_uWvnpB zlYeow2^F14Pr_UF@l(rf=pVpJ6@q9_2fReSF~%LX7QZTQgVEZ8+VasnR;zdFlUF8pxRC!jr0rxkKJq=e^}Jh7h!n< zHZ(fqxNhu$!Y5t(MFK0I&;JE@rQ>o8VB(bTl(7NAxz0GWHbDU_ojGN79 z=Bzsao*H@it+A zoNLoZ;PN}kdnC3+@8PF=D`B7JJs|KJsGP;D>(GLxSA=EF6(4R$%NO##zZ^#vGvQYV z3|A3Y3tRZm_Np7_s{iu*4~c(0s#&0zC86xpOEIs}#gkJ2Nr-0p|iV zMZ;T)x^Re3ij|j4?M-fEgpJX_lXjzlA3C?1P84U#+MMx#qls|h*s`aX6k|{6TzoEf zg!=^4j=V|FPJkFNHM;lkApt)q!40Va$~2#DYn@LLYP&O0lDrj92IvEM_x$75Q^teiC9<|LSo@zH@TRd0&x?}prnQjdO)a}^)fRzJ=FyprHK#Nqdp?@k%JZIjdI|j~_&TRanEcbUz|wlZB9}z4mv5S($M2>9EvQJrpHa}K zj81EV6yTCQZhhZdI8(=j`)`yHustds@?#FMlTNJwnvOsA{PfubC?QlC@f^Oi(@R|0f?-=-PGJZL6 z|7m`t|GoJEv$^%_Yn{+byiR?}+?Z3R4gVK)Fv79lZch}fO}_u$j_3JiZD*scZCf!z zBVz!S+^%c2*)I`*=t3Wt8;1Uern3x-stw!qE7C}JOEX9}62eFfLkS4d-5@EAAYB7E zFfS25!(nk4x69m|-%pi&AbmPZjJ&X&Phtn9WDnN`u`e)F|{KXy?&m^RU{ zn_o}`D&1xv9ZTfmGAikRczK_6qc>N@Ef^Z)aW_d+pNE6c^M8|m9`p1?OJ0MjeG_+d z?~eE6!e^r@SXaRM*N>kGj|0oE61L4ICsbn}SFqI2`)k)5o3=J=dU#=eS9^b!0V(NK zNlsm0DS-Dej%ogS9f83n75|t4C3I*`LdI^vrobFR)l)w!Iy0#0*25vD!9kcza zD@a$;TzxA(Q|{5%nI4pE$6`xyA31DlF|BuHl&c&{&EG1G9SFik>^C$o#CL?e!9h%2 zOcCXoi6|>;jy5Cf+e+$NIB%CSNe3IVxrLhA&#v2iy2s1I);ELNyzj|WZsp$i$SOwS z24)3Ni5;JHB*Wv|*^o|L=}Xl%9I$wHv2pyD!nS(@*%q>xKd|4s5`2|~5-guh*uXk^ z4LIvcn5Q;f`gDs^6Vo)_bbNGeEc4zziu%Il@{9UrbNx;HuLUEDBkt_)NtaMpiBhjG zZ`a3nr->&&VJIDEr;-k(2iZ+eUcOYcvjVvENCH&on-Gv!#`_be^GFl+H zh;oz=)J`xX8G;~DZ{?^PlHIV#DiU@;Lbfn7iP5WGu1im=P9-OzmvN4H%(W-eFHW{}uzTRKKnQY!L(C$C{ z_MnmJxoP%bllGYJpN$PTe!U<_6TQq|`GA?rA(ITzB7n0S5*gsDeyfOBB9aAY6eVeS zWo%v;6CXYa$=migD!d||q<|wxSjq9RBp!Jg%LfOrdKi-(L==}-eH{Dct&sAsix8RgE4 zNhTP%;>76jV_WN3Xbg*(vJ6FTl)`QucnBS~bp!rV*Yz|6-p; z4?D>+_1z#Ot>lm1=s&BK5ipa=@H;7~u~=kQbM;fg_zRWPn4NdB;q9l@YIT0DVlIoo zou2eqlATv?)x0YdGOz#pu&>fMw%&PIZAq*Z3l9@P@ErRT=~wyud)s242^jw*jdCj> zk0tb|$y5{#7f6cqMkmRWvI3UCX9#(|H-dTB2K4-E`=_NZ2j$&7KDv2oolIM!9WRlX zJ{R1TDJ9%qWEox61e+T@uiiuwFwHNWwHL2Vos9<5+{cw1FYEs-6WS+FefhOk{A7cX zC|9~)79js5ZY`7WCwNRQTb_%nUP)UP8xz-5MKyTQz7kY+Z@>u;rI4NbP?I zp!&?Of_-uDg4tTBZf6ye4d^)4D%q|Qvai!_%eKs%#`mH(yf?lSHq>(JPKEW$HQECH zoqf4Zk7YBGkD+G5r7k3#0pVISnrpl|Ni1FLE_AV5-pi^}E+4V2v}E4hIBmvf(vIH# zn|y;b#`kl+>SZx8a#ZlYnvIR)ruB^7&fsEwU8l;Ns@gKSfs^HcnOV)WV!aS+yy=8=5*%X)lM|hzK$UHc@kFufakkSg~A$!B4?i zhq;hLX>lEBA|X$|7@{_Ugs*TQZyr75y94lB1#0tAYq7&f5Yd#H+mQ^?11zzRDW{mV z>O-Xxj>eQKw#EjNzv;XTI+Yxh4diNhN?8L$dA2 zXs)Pp-3avZ`SS(gK`81A8Ac+iAJ4wiV9ETX+$AvoZ=NARLE9}PHSjO{pQ4nQhZUJU zg9>ZZe)ky*_!NUyAPl*KUPb~bu-U>PA#PU)s>9^Ml$uZ@JLd28)gBhrPj2>XVuare zpCi>qJazU(moF{I4RcjuzGGXvovB3wLrHAMSTPZI%y+96iXU*{-xPAc1oNVEb>V&z zWjJ9kwk^?ZHG*kcdELPW*axcqpiyY7mAHqbXnq_OPWJmE=1ReI-k2YsI_bBg=SPAn zvau}m--LzHuMlpYS|+x92bL+n*YYs0N4nS5f=* z;s3h}(3Dg+74>`??kOgoc6vHUd#Yvh@%E(Mljoy8{cKHA|8``@gav0eJrzl+-$olS z|1maJwJ|*DtRpk_cI0%bh*MB$&vfrx|K#g`>&INMPxERl3aV7**+sF-n&pcD9xBKP zHB%=|MSH@fl2ONx(o{NnQ)<4J(z-NCE!@LL0-j;R3Ed_9W(NH#5O5J zT_1CMe^CP%_e~?eJ{K9VvP&K>8J#55O$VHXOw;En6yd6gBw2srqF;D-7hH5_Nl(I$e;0QYWL; zJh`^#kxl!v$Ip4xpXVaMaz|zU0k%^CPE}H8r-Ls?g^8P_cS^B- zoUo;jE6IR*Jtd?j92=MABnA^%AJq7oIr#h7+uz&7DTmV@CXB$^O3WXRPu@=Ad&`Tf z#XxDc(HK7IxYk)0+u;83yub8i7V`R`Zn3Z1bfYs|fzv()AO3q3KVu(Vs#*VJy)MqE zTKZYi=NF}yU!uwk{t=IVr736V%Aq)p|GC59|1>0H73h$EE1!=Qmc`f32dyvO_ZDrx z9IcDZrYdP1;JKX=n3FChQmD-aVNX>z)ApVRIi6+EqdcO9s^_yu@)0y+JpMb30cE|6 z>E~f@QTZs+2kU=}>g}Mf82Mk#O_D`X(ntB~U8eh0e4aNM1{>RPb$))kUe5NniMh>r z-tD;=qD~4UQ*P+nzx*x4A;Q551A!eo8HG)=NiJs&^T^WSSS_q`jSp+=?tmDwTcz`o ze5-t=j_1nYz+ ze~-72eB;765ffX;EubXmmH$mwcS41@qX@WuV`oJ0#^}7W#H$pB0PF;f2`?uxnkySX zo;!I>%WW6LD{G6%vC5tbHmm>rx5va{Tpee+*gXLLY zit)osWQPU!Z(59am~{S_*dag4{YV~;oT+2{vc~#%2R50D24#C>F+o{c*nN?LK`oPl z``S^-)@i{;LNpcwF@sMJ0sBXI%xBO~?l@&eW_n64n41)(!?Z&1 z>zMsAPb8nA!a~DT>-yN73+G*ubhwBug2^oLz*qfulSnb ziwri=!#7W2D9EDmMNJw@i5|iC?r3-4h3@h^N|hXAYVJYZ z?$YQtc#G2WDxB28Ai*vOHXPbHbIB~7L^{AOol|4+PVTRN7rRU`%Xwiw$E$eRd9@2+ z=~uA1lQX+G_gFcQm=TFV9dTZG6(=vIM>yI3gNkjA%nu@J8O2cizzt45<-eh8Q~iYD zWM9rm^neara=PSooZvhE8Xzdw34}7w0-x8d)x4@=ILVcouXXUl$8(~?Hi49a!RwGe zcYk=j^D$*vW^@(AS;=a*?3O2S$?}xG%<|sKM!d{Zv*Gy`qCr}k7w6#ylkgRtrA9dTa|t zf%6)S-pNw3ZMRRo9SGQDgzQhx*x&o(s7q3SBYjbGsWEYQ;rfX7bf1k{%4gQ)cZJSq zYpmVeWo(=go#?W-$NN8b%>=sa?YW2rqr_%k>nC5`ytW&^Ubf_<4unMG<6^g_yyB9i z_}Y51*kKcr{^q1T{&Haq-}mWiy5W!OMI^^9S^ zVX(97`RQ3=dppaMCc{Fq^7iK6x1^V~@Um^RqUfkzatx%CWMkRQ>h+YtvBDha3{Vi{ zO^SVgk4fGXc>cLJx_oofV`DsP%>J}@&%y9~$HGMXLPF}@$<$%@v5(Q-%)8iX@IVkoK?ev29(>l(`@=i=^+5aOMQGZoAUNk&$5IC#k@!fK7Z@zm(pWKlE znBYs||C&9ohjAYV8{g^xbxM-8ey=69EysIJz!{kW-mu%hIe42M=VXZgaFZ#tSywjH z_SL#)J6V#xWx;Kh%}$f=<|ff=#YD2Q0`Gt=Wtg0zoYH}gx&oFm%9n_mlyPb|)kq1st8B5e^@JXjYXI?_u^~l!p-PdQGuRcqs ze&k+W_K|llWk$f5bW3_CyV1DxZ42isqHSLaP@gZn21!VV9eS3VwFnVX7(SDk1>KY3^P4w7a)-t4$ccL^j>87_j@poN1^V=5F6Ex=Yd_k7D`o(bu zmpHur?5doSk9A_RubX?qo$Jsfb&;g+un}qND)vZDQ!ykTMT-}rc(UszqY(tQM@=P1RHMtZc78!VsZyM2Vq_NucDw$f5)tT?Ahk$~SD%{$;@E36 zbw-wcRwza(m0y<-)C$Bwmi(pRf=V?lRUdNK*K&G)aUf84`7F5KJ*|Ll7-0ZNZ%={J z^D`&asA+&n3)v8MsI-d^`c~+Mmo|YJpn2tfrFmjd_i~cg#h^Vm-r+8z*;ahelxU}F z1mp|%E`a{Y{WfXYzP632o)H#s#b#k|^MSUe@s*+%K?*o> z1!xl_I!-QA3=mX9jbij-w6DvPDs%<7)mHMlB(KpNt0;5<=utmWI_msy>l8+Ls<7S`oTw&KH!aXV4~t z8ECsjvI*<==4F|QrQlqO!8ClGIKl#0`fNu<-2q?vnXq>wFaI3YG2`#QI`({uXKGnC zQf)N+!0lt#VB#;Doas*~zOXI|W>pI9b0QW-cUw`Te}4YlI}SnVwi!H#`1~JM4*0Z; zZy|1EK4BVESrQs@W%Gdh%N2r<;d5eMB>4$ScGuF&wam|y$p2OiccG$el0d7Xm1tL} zIazqeUxwZO&oQ6+C6p{;MUju*7;5`Pm*Ha|4ptUX$qxd5uxvPCB#DGP`HG~5EGxOF z5u*8P-4<~Fo&8N5$8+`2>BMt$W0uQqe^+4+!zTZeS3DS>x7ntWN%Gd53NN^TTgk(4 ztvQB5U|cV<_vC6fqw0qZo7(gn1Jv9+5N>Ev13h?z!jnmJAgj*kV_83D$-c+NQOD08 z{75i!zZsvkSSc0*(fj^{9jqlc*BUeJ&BzD>lttdnP!TS6Ep8`Tw(xf1mhHRmni22n zeTR%C4L*>k8T+-GPb!S4ZdJxq46tK!+1!7F&~eEi&}264ZDEm(B{RodkD|XaMhQ^L zMqwxI?fvR1@4o14H%P&h@hTMB*HuvEIhb@2#EF50ybq564F0;3`@5WcH!A%OdhZ6Y zoM|f&346`PlgoQcX0=MZnc(+tV%(*8gzY2o1N~19^y8F#ml}PC9`$>N_FlOC`*p~q z_r~K<_FXK311J#}K9gKevjkjDf$FB8t&{=mn&;c~+luw|WUlAk#dV%c`p7ur+J`;e z!D_?o+MjmV&;-rh7X0RBSBa%B63fcX0d=e4_({v6>RyX_NQ0OIw)SxRw);h4Ywvx!&)w_Ou(-$WH-^AoMwISkS9ziXArHI22ck-MW=pz>6K% zb6n;YYiSVaMcgD8`TkGatL@Qi#r3;HV+%MHCQSKt#mBnA0psbm+TK8;ELlMH)A{u1 z8Tnz~ncd3$^_dH62;g9RuW2XswKngAZ&K~n4mIrjT~Ww4`WAOtCrSEDegAvE-e7&V zw6%S*c>GB6^rYYP-pAVQaZ7ss<|Vf=|08*=@^PC2D6lQ#(|-|B0JG2H3&0$2SZ^=y zo_>|bQ%Xyd)iQT}0yz!7cZ=P2N|S6G>-JwemRdSF4K<%aJgkf#?91V?HPBx6cL#O~ z@;3YlcwD)l?o9)1N*kF%^euL`hjMBHgTuS4O7e5Jm&i_3ORXG)+0Q`e_9GSe&t1s( zJyto2AJGQ+-@i-DzRDQFI4Yyws$IRup9X%1jZbQok^XC|@QwRxG9Y1nsHKwg?6(eH z+-R_)RfKF^(tExVD(E2z+vondPpeQJwA?CjGF}=OjpzR`0@8zW|CZF>7B$>hqh&eY zX(32$ZxmC8n1X68bg4_grziP^hZSPB%#Z9bvel8~rA$qe`2-~@~PjU4bxp$~gB6{QyazdS~Y`rln)eJbC4m^^y3JDozrQ{ij$*@w;;S7KJw0D-B+ zA43c<<8_5upp0J}TQ&4O_ZQih(K0o5>e7@);ewu!E%>qB~5)fipdq_-1h zSZY%pjZ%GF9b;vT@DkyJC?*CEXXC)Wzb4s131Jn(A7|(q7JK2P9zTXz09jhkV2Woa zQzENI^Vs{Y!IYTIp+x2$s13w%tHwpQPqnM)tJ@zO=*OT{{4D$bUf~fB>Y6z7qRTMz zim9n|Va9R1sh{;IN9))`!y`hV8$cZ9UhK1ID1nt}?m$HvntkR^@C+2y+dAhF)0R#Q z*9HSv---Zna)<^>BtrpXI-aZiE3#A?Ml^yb5?Y;b%$M=sqZ@+W@2#;~vY+^a3C8&Y z)EKY{BBfC+KY(B7cF)5RjEb1?wfTRowY{VAcL+6p@kU)cCOgbSu%moEfweA=fsL1= zmFQCb|M-=~TvTAngal}pWvm>7f~vBv1A3rg8)Akl66ouMuKU0ZR>XgXqp73(6$;op zVVE7SHKmC$O&e#0fwPYUEiaRIV_AZLvb-yXtXxnAT1MKuAonzC>@-Uwk9yLnU{ANG zr>+&|cmm^gZgnA0D)C7Qh+E;~|22PYvGgmuE>>=+Ks5OBpfS=hgFPyqvHjFQ;4aqj zo3sT5XG?Vciii^xyTUU9>&2D`^nAUb4R{kqR}4{-pDxD)8O-_#7D>OZW`(FrJ{t4~l6+7Y-^wefiDfagMYO<_vHq@^L>+lH`CS`~z z+hS4C;)fBZxAXf%0;j&#PY05^ZPM-Zm%VN>Ka+rMHoZ3hO|#O2*i$>GljH;R>tf90eimv!l@43>uqT2{NxrD&eGXJdxG2KNutYxYgnd)jp5%JmvzN za0wb_^D%y3{IjXqpotWP(jq!ZINj=+t~+f0w>Q|f*d=TGI4P<6w7@~`cs2yf9S@wG zL}ODBsF>?pKY!^6wE-zQZj6j)DC|=L2wYE;Jm!{52mF_z7tOWYx2N}Y_0@K*gvLLS zS`tkM^)qVq1jg^~pY~=0jivT){D|wDkf$xxd-e(T-sxOf2EQ*VEM4LBZI74yzsS?! z(ZT^Ebu3DnL4%!rgZo45M}}SDUH%XQzpCx5tOK~-di>UtV&qiVzp}f5?QYEU_oLhC zbmSh*0Z!dz6B4a6=Qicu$Bsx*B`p7|-7w=eikl*9CdE&{>9DpxT7!i&KECy%{?Jiv z^pxJv^aIBubF$7dt*1w$UMs@OcZPi;BxFyHH0rpJZ!si9lZc97VmnZq{w;}PqbjGyx zGgld$_FIWE#Rl?YaZ+)ps>DuN6s-Zm>yJ&#l605L3Fy=}y`vs}UmP1+kZx5ESHIE= z|N1!t@<}7wo1F=bWPM;?@s65aCvF=p;L@bq$w=i3pX&~d3~O3R?jh^%%x-xZE};=Q zyM*Dz0d|kWKO4mI_J7tTe_#Wx?RgK>v%#U(bfbXf#*Yt|*Q&Sw672WD=TEb1JRN~^ z*RdW^oKFz9YaA}S_ttE9=G?_MH(I^tUpIp9_osQtdIzKOKL446d)Gc-`QU-`WAMbQQW#| zH@wb}q`*swAZZXZuwt@+d&w_(C(rR;iDz*;_>lx z*L!gCv&l(Afj{4yUBi{1)QTGy`TTF|;ywXyJ8&&NXC$97xIFnH?ppiqB*Bxkj*KPm zQdw%nV8w5GGSr~)?*sB8Afn;sh}S16o@&d3t`s4p2_J(*WpU9%>P$9VNxWW zgBqu5U6aRl6R9uIjp3;TtMKw(bn!4aJNeSHY0B6mVTUhPd*2^TJ1U%HO+DyNJ}Wky zYs*X2ELiWjG3)WCkI>9Qn_A<8Y!1jhm!7sAor6@@hBtKeHnD&mTQwG^#MNS6-;YMU z!@4UF29p9(Uv2)b-D8){zE0gfMyCf=1&4b6cRw})p3U|-a zXj;0KBUK-NP+vKaWGBj}qA;QISu`tsA8{4UXJD6dpnrUR9}Tl!3YNd4=+Y74=RY%kx;CH={L5E_TLrmDdWN9IlL_J<<_@-kKWZX;bK|)S|FetDLdi&^$h1rpHVSD4Q#(j%Y|1bQ3e2 z(9LLe@95b0?O4JLVt+)5eXr~f)n;E!WsN2yWI>(#sH_=|8LOwT)MU6y{O|w-a)jgm zne|i^@0MFcuyhXxJtH`mwjZ@3`|1{a*lWP{Xdqs_es^o{51nS(lk;$o6j38PE1*R} zCj&*NpRJ?JQ7^@0MhH6igyUkUZm|@(bsuV3ZF%OcPfff|FrSahsGvf`-x_5yZ-tc~ zWmx;}h$_QNMsTlc@gi0lW)5aNUvliiD^ zDr%<0j}WvbhB4q?Ak;3_fg2Fu8rw<0uF|C-gTS_hl7(99o~6m8lefJz`sf968G+|A zVGhssM@?!Cv#E(l%znAC5z_;0sodUaQ-aMVRS$Y0rmqws&HG~d)FZLhG=;JxX`aXB zzk+jwBdb0R#hA4Gn7M#omRS5lcz@+bGcu|jicTp~Z`~A=i+C`Yn(H$?p&7+N8eTu_ z$qjC&_xWCHFW=~IwUL%>vjDcqd28y6+dozP@_Whd4fL}ya177ZS!F;^2aGOwJc&OU zc2T_7OcaD_{zC@bUa*iW{!9ZcHEIdygr6sUO7Gh(@qd(HVaX_?8UUZ0dH%Vxw+h`V zH}Ae4hO`uTkNLJ_;>91x{H}l4v16^K1%D_kf{i7lh@|d5jQP; zEzeuG&~^E>4#?vG50N4sIMtkl4peI2}yo>KakBbk(d3H-sKH#*aN-lKVr*$-lal?bQN3Tmo@O4ZkGEZyOX z>$Z$3$>*cd_B<&|ZXJKn`D03v6_~_A? zOUvz;7KpI>&FZilzU)me#IY}x?!|23oBpflcqs6Klb2)`j zWOCdlDrsLHu0*z1ANw59+DeJ)pWXZmq?hEp9nWKWs%!Hdcqefr?JY~;?QRLz0PL=4 zaXn_({fp~Bec{x1{_q+XJzZTxR{zQC$S(*3MOH%LeL6L2s@=h+%=g!Mt)x;vR|xu z6l<)JyPk@xQmV0QC#s55%xEN4L}+zWY~Jxt7>cE9V=bo=LYXFGJ>iX#lSqV#?%${^ z3HDoKf1j(Eg^qXHP;5soXcx}lJAaREQsZ z_TG+V47yn*{!XSwR*PtiL@sdLTRocthX`w5V%@?mSvzQdgio0+LC{@@tr3r0vOXl9 zs9X^8UY*QKoUj9O0w;OzDE0|E2N(O#rI$FkhS`?lrJxjj78Gh#Emp|B8xAKp-wqg; zN$vG>gX)r@C0y#HE38-EclE2k-xOk+{mJjbsEi!9_Aenq_IqNgR|%%gZJev^1Ub@- z^A}}Vhpu`7WZqy%WQCK96r*J}(Fam05Tk}{hT5yVZl5?3fb`cT?JNmSP7UP5)*xbL zjPPrN{1{SK-AGCM&-_)Ro=JEl!NRb=imW|+obX}TeM}F`+3;m*T>z*e z^RXH|sl2_i*yy!5X-Xm$7XSvNDyJ|vYbXOE9v;x<=C>y&SJ-)^OA6q=p*fz|tUT}T z&5EK5obPnyOd$Du*g=?qftN&Ost75cJU92X4Oj7=*uK1rw|ge#c!`DB8$QOh#Rz5i zV$OS|KsXUOG3c#XU!*;yM6_=E;U%nJm4Z~=B^!agPXnx)kuZLJx9WnB3q_zrg+;u+ z`F7dB{2_i4AqP)AxfSi77A>G1tq`9B{h~%WFO=SmS|phD5UzeMmxe0Za+L7%tD-QF z96tSVt-b59|7yxi^`4FM{WnMKg}3}P9guf#W0QGHUnRQKlK}oD;YckOUX7o&#B&gE z#1QVnJ8q-&v;Xe>0RUVT=W5+{Qu2JPekmpQMr-gM`fZJqcq7O(Dy?5Z_KVItc#YM& z`sV+v?JRPNsoLG!tS40;I(t7E)Zv!u2J;)wwW5Z*m^C>y)MrR5`4`o(fmH_7Go{?XUGmg!?k%0{?U{L$m@`+YhZP7gk@KW9*zP!KHLgMW4-2{%zu4^YHmLUy+2RVQMfJB-JYO|E?!mmZ@w{UP%&f zMXNKYb7_|wOtbTFhnwYKvorI<;e59jF`j?d-V*hpNNXp^r8ZJp!>uS+Q&*2xpKe>W zkb46(QdecwWB)m|C( zWn~^CMVft=oGB#3$4$S=mS#-6vmFl5;=67+a<2@aT%>NkZk%p;k}>4Od0OekuV2~@ zl@;+Wi%VI@xa+Q3S#7*i82tHE3rdF(H$@)Pq&^E{F?e|_oplc*M=QdQ1-;a$I)g(Q zpf}6hBf5S|E{nlj#QMN16QpIn1f2_`=)d=ltV1#rZ6{iID`Oz^e#OO)Thi@SmXr$u z-@pFBwr~00I(KGN;QPUywA3)1a-FHDYB`xld6!i{3~LIAJb9sl|17~|<`a|i<;>VHwcxfl)XY8o~F-Xc3&CAIHfO@i9 zWx8m0XKIF|)!Qb?jE^tmuV;q{zYfhSf5pdAL&glJH-D!)r}5LS)hF|#{Mp^e+4(f> ze^VITbxZ~E56R>hMu$ujH4xMuRD)L`D-=b22X2w^(hM=+o^P*-4j}5{<)# zH2<^{yfGw4(!n+`yAh`3#ZEl)@^{87Ck>sDEf&*>YoqpVCXsMheZFbI0JbzmdiewD zD|W))JRPx~?)VMQf?eI%2*+0q7)hy2Ib^T!6tQW(=SthD6r*JmfC(053uHd5jFA-d zq%g+8qC?zY?9)~c9~gs36zyu1Z-9ATgySnL#RA5@h-*3x@b?obnJ>&vgbXsTAjOdk z$@!7%VY3mKaw_~>BVx_zKh@PmxVn>>2EYUwcM}*At>h>m7PI)S9g5h#w8DxXX=#>| z+3x>YQqo8ma?c)!dxX;^SIMxh^_AFp*K;;lS$#o%eixK=3d)vc4l(#Kg-f>Y%KvDg z&W|2?jkglP6frc#C^uk(1^93n&0fn#w9^KF2+gD-2pdjg*cl@s&P@Dz(7Ydeq>9OX z<3t7jkw?0P3$rLf82PD-pePD6pPjn|gG;_!rYAf^JX90ZQ}BVZ^TF&YvlUq9$WbSZ zBsr7uDL?=9n%bA!vHV_4j`Hk)g&R5 zel6=$n_S2Nt6CGi*Ox0Tjgx)!7ztI@Ax)_nTj;Q@J-g*v^{}$Y_Fn(^+rnA?Avj}D zG3rR{ZJH%VNmagJkfmWYy&y~MIm+)vWo)wd8~F$Xu&9!4p{R+zB*`ZM^!cx$C$EHg zM+4K}KY22W!Nv^A?1RQ$2@?em_V^5~+A@$fil%9|6EZ5RTTT58KAm8!l}Q6;iYB_O z98828ops}EjM+8PFSHWb-mo5zy;qy}Kc6cou;|ewjVEEzV0OGFp4`z%>9R{XU;R;S z;29XLHGEBvkkCm6nvenwoh+eoQCdvJHMv~X-D-I;jrQuA8>uDRFE&y5e__e4?W@ni58B#wCsAdWK#4Jifk9hPP!{IAFwTle=@^qsusGgad(u z!Sj}8`g6j&Kq=mL4yT}Cn9~H*IGHSnZ)>7(x#sBYpc>NYcGG?I zGTSzdsxK0Yh4wRp&L?90DVF05ksT+o4;0qDU>s<3|JyQC5p;ZjKk134br>nwx4V;AdemNmlEUo8F)Z%cs)0leXt8w9f_J z{?>pA#4TVQa-!|vcGvoogaN69AqbhF<$#xzoDR|YB%vq#+9ZWGxemi4c>8k1 zrP0CbnbGaAqK>JpKE|OCJ{Y|H&uU}Y&l!ODs~LG39CKA(CB^Yz6!N-=c%?BpUH03m z*b08SPfeG4ow(+p?lL`m!!BxA>ww1u>GUxbH3ZJ(N<7lWaekFR4xgz~w1jg5daDXg z0We~yeVwt!y?yp=UElRN?V_M*QdmhDWe?oh`Lz$yKPJI5#fp_W)f$ISBQ4%t)1{FE z)6q&?FNjBx58j7CY-vkj75+tv?Ew6z&*7knVCGWbID>s8nPr5{=(B=U+flPKNDqm3r;CR}tXr?J&se8h$+|3w3u!Wta@*Da)@s;@*c z%|%3~hMibb-D;_S@u`FA8yBl*;}&2hC25Sq^@8eYN#EYau577Wku&OvsivXRGlo~Z z4~c~0jB$;lb8F_=Y;gmN7`11zDladI%(JJ>#71p z@~y!g6yj>>j#-lmOdEcH3{9L~7HD%Scd)n9#Rx|wlvsHLx@7Q%GR!%kbWya6M-XdJ z%3{y1)DYJrH`qd8J8LAS158tl@n&GuI9?Vx-JC%KiGl)?EBLIB*@c&Qd|&_Sj(mu1 zRZ+$ob*#Pcird^2dt{m3m0PbM?4YM>&~VL8eD$krNdye32w^6grUi=H>G!%Ouy%^* ztmiiQFFEG>Q0QoMjf}JMs+#mlGG2=RR7C0k4wx405I?PRYRbgM$m?{c9 zo|N`mBLp+p9X1IpP0AToRj`D=m*r5=6^S6><;cS+>5wMkOeVG7@8W5z&(kuktqkS@nfxxVJZ!r{?8;wuA9GQ;5N4Gm_=i$mhYy zArgfJFSJ%PUqfPuSw4^dtb@Ob<${E_(<;@%j@!-eadP%P(baZ#Mrq5my ze}83;xpXPCJ6LPDWSpELEe$o3iNr^}a*FrEX5U+)MDw3FwSCkjkuabH z65NjmX|~8&N8t>SB0-{&si*|`^9>2|$sPpz;0t;QdSg{T{}$_(ALnm*8M!4{Z(6jiDDi1rX%~AJ!2|JM{IVM<&>uH)2?$AU?S1&CBGKlv}^h-f}L44Bovk8 zMecS!zhFT=Cr(tZ#cw?F*a%9HTTU_l-|}33st}XKDF;uy6P@GLPuRxk6TGoW{$Db< zp#~9tz=zW3T{R^?)%prEn|uFvK^`6j@>M}*L>j>b`DJtLa zTdZzedq0eV{uEi{-&_Y(OJT`Ojm@8o_GpF>dCB^yy;KXXRn$cZrGf3 z0Jad}J}i!JZ?(T$7&(84x6R6GWj0V?A3lD5{)xIKUT6E}^j=79w3T04Z9XXAdK~Rt zz`xT64~t3PMk%KZ!N|!8FeWE#E6;PZes;8Z$v3KM#V=6=t#iZ6yL$WgwX*uzjam14 z?drJVlbDK@b2V*oteo|`uR^D%rK7sSX!n#32theHEezrNKkbS@YWk4ZxtiYlxjlY= zamXV=LVUEpdFPf?sd+&~b&=1C<^acaK|MC}<94aG%MH4|}Y*3S?taNR$??ebLPhF&Y1>^KJ}R zQcIUnEY^eR2dw99Z)=tfwoFuzW6JIPs1FBP-M~peK%?oTm_vxEY9I4Y#cE2HdC=)# zdf?qQ!h)W1zGd}bFuWFWp5Cjp|>xhUC{PcgBgPExl1+wE;c;);qYAkz#IZKC%GJ|DF zg%)kF+|(d^$qPbn$UX;?1lh(aljSPDW^UxI7aME(mTEQE$Xpomy*bnB`e}IjJpxmg z2;A{P7!~YFQBL?RI+B1!WT~1RxAeCu6^5NeX~*fqKhhR*3e=dustPfEo?qP57uASeObf~c#F`-R)-skj$xUjPSi|HDU&NCwbkR9qS#K{seFNFAxu#lfyo6@ zkfm5z99&NI1}aI@}b6$K`h4#WhsLtSdBtIViHxWC<{CP z7Dmjiq9{wL+YfJ`0N({ix2m{yLNj4ShVwbi0Zit|!SIiYlk?J=8(1mczh`DmKXwuP z&@xlw_T&et+p!Nm-0D-Ph=~4vlr6>$R;p+AY8fmnQj0XLdG**8a5Fo!xu~pf%RL2T zWgUJQt?)IUI}iIuEqK6(Ti16x&oj`IU)L7$Zw>o?+ei(TrVjo;G@Vsc8*JOG-{QrJ zOYz_i!Ci`L39iMpxNC8T;2NM1q-b%M;_mKl#of!!_pi0ja|FqA+srlKt#){tumvgZ zFwP1>+L$s}P)+-WeryKgC=L{7LS{tr0Mv9OA+DCf&u3a$RJEBas=&Ap(x#50s%7B0 zFq+HQ^LkLa`3FVR`=~&@e4Ccu#HGC&8KFw%x?y&s;%#s*@^<9-9!I0c1N;guii{zjShgy|-)x z*4s6VSC?+7UtT0Rbn`gBi1X7aHnqiD&NmMlPL3YIxCvi^!dxYL~kub|BE&I8HWR-n#CA)1d4H0Ke~FEm-c<)&v6hqkD7EFE5+XMITmHUt?GYt%Qb{2Hw!3$rE6c>4=47rh@RpVtJH z#m3NG`8%~hG6bQ0u&1v4tPe)UYE{&^=GGH7`k{!+pjnXjmK}m+#{H7vBumq44qtRo zX;ayB@guj8rEB-)|4Pi=-;J#f*uVqZ!hLwkfF+l(P-lNPt`5S&1?>$SHqJTXpy;9c zMZRwxzY1CLT;zSab>&?W^?Aqc@^T&>&9R7%x920vwDYENJR1}Cm_s7ZsobQg^+!4V z1j}$d>x}EZT29B}|_`(9EzhcH_{hep2lwr+;#r(k_ z6ShHxt>tFjF_KD}y^6nbQmc_u(^<`6TstAHHW~Wt5dY-sr6hEIhosoMl(y^UJeopO z*QJ3(AXYX+TOP+wwC%5VOU&!_^&fkyGJH`>)M-<8E z*prg-n1M1;Ff147tW@ziu7YP(_ZmhRJOYf4e3etB0wAL~z!_Fy-9FNMwLT9YS~n4A z4oe76paEQ4&JwEk=|$juG*b$gq#Nb%yCW{E0mfLg^$(E3zm?F8SCO} z6ai#hT2F`A=0P(U)G+3MXGQaJTzh-=({%C=y=cB7sYq|scbDwh)x*)?uD3BR`tfC- zI8_0GPMwO(xAeMkLko@i{GoC2&xw zda?(Z!T^c-V-W(Bp|qia@R+@^8o798mg^H-oHXrna$?wS()Z?1SG?FL zlsB8|!AEbL_=E|ZaCGT@!?DCvbaPw#k|db{ zq_|Hwht~)Z2}maZMvGe|HERtg1R4e6!lBbOWy^5i}kOTL(C+?32(*q8A z%fm=E?BFd+UrGe8YM9HT^(RM`GT+s)JDA`j6vINEjpHsTZa#EEB6xf~ci1QVCxcW0 z@3NDqVTvOLyE0pyP8W-6CF*!x3?d_c=MsARyLH@S{LQ&^(NoB#&!4AJ7-uyN_SpI! zhHt_IZx7^qAq&MoYuaP0m^krhNCRv&H=ljTLe;4TDNF4^1_LbEe1u(UgUVp92{R^i z_n^vbnW)~Ut(wPd;o8dJl={VC@2Yj@yKb#*r=Bkntpq9)Ya~tUbj_pYM)}Z>|IIQn zmHlYFLZ1jws9-<03WaW)?8Qdnwj!^4i);Hi}efAsLO)>aF`G+4Fys(}& zgpXQ#b(=NBb5kEF-K+UE{+fs4xJ2AAb(4Wpe-80CU)tN2x6AX{18qs}jUEN1REGbK zi%?xAel}I@ce~)wvS(Cw0yEPL*41HP9C;q>dA-9PrUd2*beq%)Tkm@Y{Ch}-at1fH zH@U#7hi^yaOL@=#>WJ?fnw(-#+E%ajR&C1GJ;#j?EEqIsxCD}mmD#_97mQpUEqfBZ zOkZ%*zM-N+!Pu}@$hx6WiTa`I{ohBbp2M~w)gRvdwV-xSE$3N!(q%m7jqP@5&om3I zWG3ISnjM+v3;Y}8$P%o9sLEyw({&?C2Dv7h)cSRj9JL080W}V$9rpyTsjl`PcYJJ^MgG@#h@iVD2(QIyL_jq%Nb;j8NGks;Kyprw_1G#h8wRZ#!0$_Evz}#FNOE!n6qk%g#is$~FlgEWZto+)Ab@Op*7er?d!|CylGj zLX%`7eZR>eoSyLU)pE-xFT25iZ1E>-9Vxs?8{4=RA2oPr#j+9};IjNxEx48?3P33Z z=!2Xm0VuI_@S{W>e~I}$a3+J5ns7tl4g%#UK|P958sTLVAB`#CWP%>jyc?Ub7*q2Z zT%xD-Bc^}GB(;u}b~-eQ7qGyy@k_yQa)1;mLJVomub)H2#hv)Yc1=fMA+Z@H*b+V7 zlW_L-UjJVUU?~Pop-Dvo%hG@wsCTEhRH#3Mq+mKy%Ogo@WXu@5v7M@+#3pGJnP!l_ z@ly}UJhem#8{s_`+LUJ%k1+np6&H*Bb(9z&MI+y!qlQN-Nz)zyYM>FSA)zx(LKn>+ zT68_EonAv+$pVF4q33(qG5VW935W`3F~OV5H6`GX$@_G$W^^1;NBJywSsmBOnuV zQzF^FW!(ak{JJ9+zTBOgNR7`##{-$jm0MRfQ5`OQTgOO>P{69gfhC{G7r`_@wApah zh>{7ufIat&A}ZYhCuo7^lL{f^H%Va%KLv`YYH{;_C20NfIXYQcy~ay=-FR>d(n!9P z#eFUTVYJJSy1 z`rP63rBf?6-Q6#7S^`%>tKzeg(Ya?;UrL5T=CcM5Mdlmn zR_b#LSWTefKpd`vwyc8VO&Z%hExlZPge$ZM;X6~Bb$ARM0NTid$Cg{u^1T7oq%tuW zEzMIz%KA1cD>zUWJ^pn6fIEn=NV&JqjO(4LscQzm;lbU;DID8nvH64c{5U2njyu&R z(c`zb4ZMxyMX^mPiW`a9bXVRR%?8;AA@-$P)g=0G|HeX03Csiz&pjZu#=?$G^kZ z`EPIdhKpCd6#o*$o}pqppWZJ&b>U9^GkhNHdR6>)tNMN`)b+kMjQj6&^51EmZ~Fi3 z>G;YTXXa#pW8Z3raqRS5d#Yj8^4&%>{o=+G&vuN)ZN)}-ru~n1>^Gy1I?1g}$t|k40#egXvRF5cwG=|Y>>aCR zgvkP6JO>~mcJ~w0)x^^dMo1M$bkE#U25_UB`QJgh{=Y?voGuIAclFz+v!Q)adWKGx zBZkuG~r+Yo06`O z3y;PzjNR3@aEIvT6zh|$P*@G@7OZMS&dt&M-+%a`J4>_ZslS|zF^;?F0BTkK?ZjD2 z=Bn#;EB4L!>AM-@==b|jyylzuRxH1Z>7+V~r?|L*bf)^A=e}~2LWa))mlMS4p72e` zqe6!ml6&T(unk4>{(R!Dotyl$44Y#kX_Zf7G7Wa-cXH^1L(d9vC0hNz5haRBtLwy2 z{>FCwaM7*y!SK)-Y-WHS9gJ6@|1)LB0X<{A51O22b1y6igQ>`3zyM)8#k)Rjhn~4S zeMC>HlR|~OYaRqtwB6%=&hxI`u3t2@Dpp{^Ak$2kzGDT$Z5^{xF)6QZng##DaPIY9 zh(-xNK)&ks>KN!n@}a+p7?M6dxMPWmRG;xw+CQb6JC=$#9LwJ7W^%V3;y; ziSFKc9Lz3~HAO@rpBz}%itEd%*bXCq(iip0kj+vL{?;yfd7~}z`%~^2(G@3Ga|if! z$HiPs{@G@ymC)~Y8>f;cW@Jyfnw>G!OiaF=!AV{%8zLx!B6E^al1T~27ZKISc4LNz z#3-Y{>MT$FHN`!-22S)REeb$EE`<$Sa=T#gza(N>VMdx$?u1|o#VE8WEP!(pwX~`W zCN!`h-(Ft?gs8l3}b+llq z1X#Ic|3T4y8aHymDcW?)Oj_1K?QKC=V;gn9@HEw?FsPYy2?m@mvDMkJr6mjr#=b}< zs27KU3Srqtux#~pj@@ExY8MHk$i&bwQoHLLs!*vy1E(}ON>}X9-O0jIe3DKYcLJKL zfYi{Gqc(Vt(;K-+pqwFK_NqgU&FPEzxg)A*NyYFnZ=@XUbTMEh(I&Unh8aVF*kx0X za+^z?Ybll>GR>Nr;Y$*Zw{RgqMsft=VTKO+Oq01h_WcH>t?E*3yF2{h!SHlJ;-Di6$NfnM7%SE*hmAFQxA%I#LY3-fhh2yr1hwo3+(~5=br}|vuLw5p70)6JavRII-SZ?6jV}H7J866McSGg9Ki&EyhK9#zfSZ+#Z>1rj> zNxeLme#@~N@LfMKdiS~eJ+Sq*m#q4@EB0@JfRTt3q z;@n~050yt`w%7NmhyR<`9)81w6sKW#TA?yvb_S< zT^HkM#9UTb7^1_a3S1M@g+_0LCJD1M!MUtY|DX_3HPEHdMegOaZ@8>7%&Vp#XhcRT z|3|*k+pW=fdCr@fK&ZH5f8Kra)HV2=L(mr2yS}Vhrz6K}9fNDZC*ajwTKx|_A6H2X zeAi3Io!E&$>zWE|dig}&h2bG&sP(!fFIft|xc;Y#0_IiK%zUc2;;((XDktyhyXcs` zUY#UxoT=+2ol>YNuG=5eeL0~&6^!~5{*h{0THZfE9&KvmR*b-mlWwM)ju)B|*`oCl7dzd2bS(3Oo1U7Y>uIb3Ty>t%eMSaR6*sX|O~% zm_1SI;wlgaF8(nV3sE7xkVxDXqAGR5F>J*p^N0K1Pd8;B*DL#1Xi*uHl4~$8?Sd6N zWA;8HwNR*dF$$w;$jE>=Dh*&oPOBlG4aZ*$TU{_d$FAZNO=)7Wq;=*vv&GehOVh=1 zL{zJd1WIhELa(kmP!OEWk}SuNidK#jHDle%h$SVVuBbvl>zUnWf{p;#H3^wT9Khp0 z#AeKoiI%iBb#q%6&3*@AStwJ`x1+IO2pjRXuOXcuAW&^-isH>7n4-ANLS!Gq0Asj? z;zhU$_{|L12!(9F8Yrp6_e{B9hu=bIVhUwA1fscosyR(9Mf04wMeUGFM32TGrE3yP zC9UcdG-DkJG{Y&R8Hfs%evHD?4;_l#$N+z@#m+SR;2!8` z(##%kNK!HdQH}V%&Z1p_QuzxOs&vmw;Du#7n$iLe^oiFgPHJj690Ppi5H6cBQ=Ij)hD`Eq(NYex$r{M-e0Qf`z%5E+{mEv@L|bH4~4@t6 z(QEobl)UF}frENK`pqUYrghY+maF=l+al<)3(x=fjEG|qru->VjN zsdH@6p=drwaK$rlKprUDGFK}|k>fJoMU*scj-MemNo1G+S<-#mU#{gE!) zLL{K8A*@f*xuk;f^j`ZwE|y;8-EeTzdNv_?US|RX_@|XPNbvkcb1SEv13iR(#NjCC ze20(x_=Qd>gTpiun;JpVRAzCj@ws3o%thCRiEE88Pbi>m@oq~5J5!BVH?{wapsCC> z%oOV?hu?Xu@x-vps=MkzF2`s#o6i>WFNA8Y{l{|6k_Thy3nx!b(HM=pLo<&mTAYH1 zK?P30FB@Mv5*3>QX+Rueat^Dp%853dsxV?_UOT8Puh>Z*t>w^12 z&6-Came&&pwHLp>Z|eqVkD_B@55KR(@&!B+G2|oU^iq6ao%hrKa5{Vd)nOdXV_m#c z7D%e+50xAO{xlPmAGE1*`80HcAah0K1sW4+>U{L zM_v={4hf&1rnl}F>n5nfesUz)r1~@A9E2Gi?LLQbcvUvMZd3fD;*<$^Y5I}9`INJn z&wyI>`}b3_*b{E&+nFQG4Fel2|6RQQvkI7a*!)3Jg!cpX(U(%F0Q;>Z`)y1)P<>65 z!K-=Wiv7Z$vd81sVtAapb>Cz*k)6xdy9Goas8 ztokhvEvUb^go1^yBx%tx_Ls}Jr)cv|NKCl4s!X17y1$kHh3+W}RrDdjcTyFFkshaa z(@|5Q>NBAN8=!L;Rd{VZHNaZw{YgYwFM; zm(^LlW_MxvBIS@npIM*~Q-)>uV!Z5{h|hx8c>T_SL_bus*B1sqJeybjP%k2@8m*QR z$TE8>sEC(zJqw@_D<~$`;fa@qt!2q{&OmMMLXR4#+*0ScN-eI!2ezwqYqt=`&un}m z>sXY6G69ra;S>6JK~G*%neybUtg92a0Q8Ajck+wx3VT!>y17Y%z6WK~rKtgnXI}vv z{4uV%tz)n}pixWXq>_hcsk5!8OViO`9Ak9&O`qs1e67D1-i0p z(N7Y7fnQj1X{+Q=77GUkF6mUeSP)*ONp+U{Nn#K1Whu@xjktay!Sv^{{iXS@b_Iwf z;ZPwBQcz@~F|tP+>vUTm20Hl#E2sp$#Iv{ah^Z(O%JH)EyD$6><_gFK0CUF@LDkC+ zoB;bTIrzTkg-DEa&Yj2 zhnWoU&^!FIEb%4JaaV)XiZf%;!aKhVd>|8|gJzA%jVLAC^+Qj1c)fySZJcF`IgKdb zV03HZEvKYyf8)iw;Ztxl5uEZW_9~mRslQ@uk`It-5s8?rp@zXGBQ1E0AxpvuIQ;Zl zNeqo(4!CT#m@fjd-y=y?ix|#gwM%EkGI7m?Nvb(@*bD)KQlddt)};x+ZDvYfXk=^< z@1uE!5I7}R$u5)>J3<35lSF<~zx@HlV#@Ni{)6aaq%38Tgd&PV*E2OfB}$_#x1v<0 zPLU{o!BT`&s?5izo813~h_Ht0ACGwyWR$e9iX@)Gi4Zw;-~hx413xtiG2$cpBE*S! zhloJiDI=E~+xLUNiO6y=rWzS469YAva?nS)XH<`_n+F>k2 zBL?wq(V&VH06*9@%2W8Kck}7-cXN=fhks8EtovQqHry+8tM?yt2=w@`6y3WtGNc3r z9;gsJ6=ii!){TKB0M@qFbAVp@!iH8--_lff!1W>L_xA8r5d z0{6yKloK*RU*X9*sysTSi-e!KplNzxnYE#vp)@0)n#t%erm!&nCjzWpnn$VFzHWcP zm>tBTe#<*`7PqEb*(|_5r65WcC00orkqn7Qw*mfL6#FssRCX8=L$(-KLgw>5oj^67 zx%-siaKdD`oE};d)2CVW>)K7(6UZu{@4=!!I^K)j;?Uwi`iTuW6puT{qUFwa!{O%$ z?QYG1xmNA6`ReJNnJFFdapj;3Ai}N z&wQPK!UMh~(Oi`+8NkxSTH>x(jRSX1suk;^>_30Kj;yG#Z7tN_Kd@~PUneq1J#3QW z{POb^fx<&{Bttv8e_ve4x=LHf!~dsibUHn5yyYaFR-BfG_dqkt<9P(3IS%=6eIFt3 zVj_$gJD7N#O-oo$Z$rO{g*X1uTH2C(0vU+7E9vr;7QO+t9NPHq?lOi24=f&Vh(ux?Q^53~2%K!#G+Wv>ZlP<-Z%TcdqR5@NaMP-}Xs19|vZ;#Ds zcCTw@1GO~<7o@A?vA1KgSqAY$5EW^1dYe?ENb}3NqJx|W=9D!~wp;q<-xkG>&nL(z zA2$69yzTUtnortC zg5rc+3}dWxH*a;%r`>42D=iPIf=p;yaNHJ>GyR9v{xb`>_BNk&gRpz*aHqG)71OBM znqomNjOS%gq0MV<{1*Z_aZ{k`HHW&m<|@w*B=~h#k;+~9?(iK< zHZQr2u6-FG2h+(g#eLm0ECA?pFvURY1NgV z(x0*2sW_Gb?-IKO3NI46B{f0TPBI*&%m=g4$l%{cZRB#tw-PYCuA$UQH>aC^yAt0oDum}MD{0kSc}s8O>O@g%1;h|f zN>Y%G9~S>%l0>MK2xIvT`KDu3w+W5XAkQq|D;{DI3x5@JvHQfX#boh~aHn3e`Hh44 z*%gM~*=VJoII+Bg3*^5u{S^;*hI&RyeuK^G`=7wD7{ra3l<;IYCEVrs^8#5Vt~i`o#o?-qnV8Z!21T57Y^L0a~C8 z?V*-8WB??V?}sai=X1|k7bj;iO)fQxC?L3`j^Pr*efsTT;DGd9LoO9s-z}tyhe{~_ zi1t2RWPI5Y`Xs88)hO*1lYmNvvmtGZD6#XU@XV*c6N`#nA(j_? zVtPp~6HSo*s8W_PO7tsZ{}neqmef+ba= zKon`K8qs`%k++@;MOl20qv68cT4IzkV|!N4!s7^~?rk|BNl}kFt@`zi=`s+Q2z9={ z&CvTBtLET!Hh1|~e%ytx`ovIBMZU&rFiW-iIZq@F%}AiMdC7i>o9tUB*RJ1s%@QVe zh-`RYuolQkn3af-IGKqR6WwELA?N_j6YS+Qf9W;t(y{8?vSwO;U_0*&3^L5!a4iDN z0u#-y z1$I}Ah+h3Ls1RzjpCoiVXm{o^e6LywbrFE`d{bg2HDd|@Y*dcrln_~~{etG!u{_x-H z#Kr3beSj7oV%NQ0*8rb-RDIt6*8+4sDRONnna+0+g`d3JYlk@~z<0Q>9{%}H{m9d$^XQ%B*4?|~^FlyHnSuIQGUefD+F&_8p^&xTKr{JO zsy3!K&V;+DQ{&Rr88SxI84(!OVGMnF!KK5#Wu;!D#bi-etMUs&cRy2&AKKYRquRfQ zpKtIJ@2dM5vSkX(1Ghtn;uu^VtI*UtmleV`DF#Zx4i zsh1%vi81wdx84qxxvf95BByphsyZl_DMOM&h6My_3b>c*Uictc)i@-7MFqjAze-lD z6vEf$dd|C-{}ECLv1TuDR-)2{nJ(!;xatkTcfl1b#H7|A^hWz2ag@@OUW!K~GMQVC z*70@mhK|D?7z(x?InXH^n%PgV3LhSUHu`JKd*3Y>Yw8}pGQ`^=x45|ok92|9zlsW*&SOerA;G>d%6cA-ixO`O8QJ`{1NPU!)%S(f?@K9mZF4OL87BTw$TNLz3#pz~XO zlrW{4Jbl5oeh$K7K~e1Nl9*`NajDtf7uUn5}qfIez))TRgM= zr&{fA+VnY$GiyNxAy;_x_`B+?$I*18`z&_Hu5sa#J)(>>Ds=fyq+Gc-VU!;Yu1qQZ`=~j?bT$=%GP5 zfLD?sU}N#bsgE-9p^gtiUkm81WB{Rs#Giv3N^hrp;gT2-C7HyFZvzM-J-*aqNY2?r z;NVcz}_eLUmM6N? zE`^$yK*+hxe4lA=x|Uum9i;K05PHl)Z=s31+Q@|R0*TnP3AGSZ1|6TIOeKe<)A z8#mp5F_k4VsHN2FSk0Wn5_SZuY(-`{MGqgDXLM!7bO4;Jx&47Bf(0j2t$+3qdyQ+p z7rM1JD1nDn`yVS~C7sp)e0 zxv;=4B{d>xvBIz_NeNxmQW-O!JOUS2d;y}5NQ6iB#tkIZ-9a~Fbi{+RI1Ju?4S?oc zsOjx3kNXQ_3`IY!Jk=#9r*d8{eCXCd``*Ib<^6J2c_XY`+AubYaHV7xv(3#)>mR}q zo*>C@4a{ARIFHX4t1hJ}iC9pfeYg=c`**>yzR}Ei;3IVD_(|y8u)_Pvt>-KR=L83| zv--dbu_ql&u19IzaA#T2cTnZ|4VmCjxEdMGfXa98O82@fx?Z@9Ium=_8(pqM?Ub{V zxip@(SR-Ulh?-qWDlAyG0DJ~J$~D&L&9?56V}5S z+RoP5Qj0J)yd=AE59@lSJ&;#=RilnsnT1B?F*JwNKYF(~WD+>Jrd>VTj6a;SH+#-s z{T;qENLFVaW*(%5O^q61JKsc)THFnsbq@1TbWlG0-hgmC+L5#Qk{fx+Xv&D`I8a!S z#)|p;v%;;<`rQG9F0uc)e?i;=s*}blu59f5_9Zu{63EihkZv4wBOWN~ZpG8k;S(vh zA3b!~Tg%aWrT8YY0Tn`O4C#G}OynRzJJ6gs3yIcNm7QeW@XgONgw?s<7e+TBgJf2# zf6hX(7DWf_0sx29RxxCo0(Qr1!QTQbp4UtdVO1*6H-v9sXw|U`JyK6^3drQj49#SXVm(a*02 zeob1HH3L>QuVJnUU6e+%q(-51-xVKjYlgVzCd*IMD_IcAF#lyLX-Km_rUU$6@}(sVtysK=&!_f_Qc zf~ZEt8z}w-iShY^2#`pmnO2`Y9bPnZGe9A(d8K_@8%Y7(H zBvqb?n~ZykOdgK^;Xh~Q4>Lg&B^r31tCvV19eTGfU3*!S$_hZ4evB;FiTbAlnSUx2_{Q{^`P&?t*ZBX<+PPY~56t_B zA48t3zg6JxaLFIhe_wLW*JW88W9nz-qqn|zT=<*hoOzYkH1BFEQW}Ho)q0~5(#o1%Nt3>u4zku%m+JPALiqjaQoRQp5T7H znUz4WjoL5zuLQKS!$diRd)Z_%J|4t@LuN{M;5{; zuSrU1S|qAXHjz~8Iw$_i+0%f0Y6L7e0Dr97<9{7tYE{b^A3|v` zgHGnz&dOytvdvR6jZ@I@f<4JA6X;5gsliRTlsuhe_Xe&Z(MedO3aH`vB||ZNj!1<` zHfybH%eQ3yt24L6v(8~(;TVyGgWV~XG2kr@e({m*_IyR1tUx)EOn_z6HCb6Rf^5Gs zx~OrIEmmZuQ5eOx6VBL>eYEom6$y+yysdFDXB8P`|s_mBk}X+UyE2Bh-yW3ntJntEDlv_5n9w zTOl{LiC&d-&~H+}8=#;#I(O1G_oaG@vOazUcttMV3Vd`zcs@eOlt}QJHM}6&N5%fG z=Y{rxiiX07wlg5rX2ry~?dpGuYT%Mv)q~`=p=2&>#Dob@dEl!HIFRHio71*pLKQzi zRQIDYdB1xI(L}e7WfCR|qQ%xFEy##Ae~&D@nQ7+k!x|n%4gKk;?-xm`!sHjp;$fK8 zI)v^SkE(Shu8H}2`CZXY1lF6e>&4<3I3k5!l-j-AHvgkBsci5*&>@yz4PSvK2Kwt3 z?V)*EFEpS`eek>Z$3tI-U8gqD&gOMO&?l=eD*P?kE-Eom z2pM+up&t_+7|YQR*ahxJfEQzd!!Sns`_nizou-Tfjn!!U4{umMkS?^)7BuTU+&q>W zz)N43WVm|s=o5zvpBjXP?kJ{16-KF^XjA$)GY(l)wcBVHkdc=-187$1+P_+{u}~a_ zw*n!QOSi`v)o18>`jk63lih|4uepOCcb}SZ?116>+l}y zJQoIMIGqOqo2rTOolXyt!urk^wfi|QcJY7Fp};C$kqzXm?pZf@`i7mmRobrCl~-a{ zUFI+lDiw`_{6@3L;8EXjmw{u8Q3+Pq!m*#Fp&Y~r-}|JVk}rwGtUpe-KL;8;;WMEv zQf?`Z+Fh@(4M&^zBj@p^Sha0%!y{g^Zm>;GK9lhPdYbv)$YjArXV*^0Xi&MopfvB4I@k?ih_R`cQT`kEj7m zA@)yc(%au3&c(m;i^q~5&-CY;U6aGZAG`jYbWJ$EEjV(Lzk6LiZN83gURwRQ zx0-xgw)@>N@V`B(fAQkA2E^R@#a_z!9L~(ZpdMseU#*`}*V2 zB&^*$zv{RN3b6r(1@usz5cd-?Mzgw$1f0J==-yo`kA|FmwunA#)DfG$E8a&26ZoA( zFrTv<8-TAQp(*r7!7p|?7`fw9H`nsxtl~%KQ8$a4< z#a21VHMv82&t}H&wpq+t+}1L*);ion#jBU@>y2uyeMOf%QclS6NlS2tyF_dK9!`_G z1>MqSg`^YhVznNz`1J+I&*2Ei|EyE@AAS=picy@%e9|al^3J4IH-Qx>o7h8x{ z>GBYSmAE>c>?H*7L(j+zH)r~T*31Jf>y%^Vf@uc<1dMiGRZ7UJ9v}z)GWi$-)iW3`0JoD{Sz@vxz!PrQt0#UGOkI_ z4^bfskB?CX5Bt`qWcaI1{~BT#4VC;l06PSayEN>Z<|h>Yp(g&y56BE($KnY1BLE%- zng(Xenhm5LZ1i$%2!Dok1a2`G>))=N_J<104q@bca(1dwmqK3Ojhe8d8l05jYARD; z7$O&~)WhK`87gK=nWfk;$&jJ-U9nL?|EidBdT2c+Y4 zUVPQd-=~*Rq$Vn7;bk<_u-K@!P4q~jy18^@5$cuXOEugMkA1-L35?xw@)OhFr5q3z z%UPk4*cIE`pN)KN#t6UVnULFiU--q0Dg2!=j>i77iH$9_h+Csm{&i&vvvkh^WX+pH zuh2W;VNE6gQL;fE9f!}orVt`r;z=8|SGDuiAbOnF9*iAnUZvBp_iUpakJyi@_{#pgXl#B3k18V72^=3A+0A zKdpw_1qH+`-NQw_`;wpW%^~uykw>-@F91L`peg&=Y%qr;Tv=l(zf;Z zV$a7vxvQ*ERVAN9X{WC6S_TlKQi6*;VO=JJ>QMNL;%JB6Xl=~L2W;9!LrNsDR<4cB zkx%y+3RY6>6rT6MewoF3Z}pKg_CiO&FBz7X$ok+fnls*$l)&fw-QF7>_8f9hDzSI2 znbIHg$sH`yLt=d5M!VP8``bj4P0ZqH(+yj5%ew8lD~eop>6CoSi?^)4*9wJEA|xv|R% zUQ9!dAviH*LWm8rA-AsFx7NBU(pKTm{E=dMPgT8kFWkJsQ3?#$pXcnl4rHcul|)!b4J9@H%H&3qI+tEg{#0K<3(StjC9FD&zz&P@@f;gsU<%vsNSSxQ35v zKRZ92i&NC)dA$kbg^V$cJP@|KZa9bT44AwrQNS1Wr;)I^^Surz!5R#u6C4Zc& zz`@@mon16*KC@qCR;A_3zHx zi>^CMVGv8O6s#cFPV z+-_+mj@<5J&*Z zp)M5jNeNN#V9{{38CbVzZ2Q(x!D`nM2hVhrbJf5PAMe=pj@M5kzw*{8N0W~xFk*~s zTg&~+HJ95BZ=Dw0oM*JQ<9i=oaZnc=)Ryz*hTBIUnWaxgB0oIeux$fxy*#5-k+VfZ z+jX=yGR_m*ZP|opM*oEnxO&7D7r+r>Xd6+idyZe2-etiIn&0mMp>X4TDlkj#gkKevuVh+L<-O% zkc}#&&d?@Di5gby^Wr)eESY%yA46ck!g=Zd3g z#blIG6^2ncLTv&~9~hSFV3RYvfJZs^H{;&5uXb26bH2CkPawF#^V2$OSbHX%2LgGs@< zb=*5&ux1;Auoafzp7u=p2-n=maU2^Bf{3%x);Bsr97EzuajCuFLf)AHVR6b4kmJmT? zNS%uJ%znHa>F`z&h8BWN#(C3WdrzeUUCimDqqhkoB&9@U8P@yMRv$bn2r^>;n-(zl z$08vV2#?GvGU@TQ!w;SiESZ9vhg0Tt$>5>gHFPe3%uuSpONEe;U^}8v1ahL3i3=+N zCNIg$62G+!E-jh_9Ku>O4xvk|%)lq~Q1tuaJ8O|K?9n3ug#i>rsX5Fg)5&!II)NUsd54d2d8ufA|rZbwty{7Fw*Snt1 zNzNA=9$qDLHthNF<3jTKq2cA@ivRTL;d8Ig=5@~h_T^hVxZ3e((QvitXotjouJVNB zx_vm}bXGj43>X%>o~upI*`g)Wl5u62l!m*rxG=QIc_~pIgAI_ER-YpUx>iDb@P9+f#BQsmy{0%bx%C?v^UgDE14#-C$|@ zfuq{+zkcq3UwU=U;kaPEPl{@lwuSlV`O`E2N0WkSsW^Mwap!pQ^l|fX$$VVW4W4b= zGdPbBsmO9%8M6KNw{Ck5OU1X|yX4Wza<;N4ozP?g7<}NJiA@5?cJ#|MipacM6be);u7?i`JN@)Q3Lr1u|h zc(}Bj9@L-7+ut|L{rv027y8B2&%MF>#DM)Qk?M~pLw=kdJa_=W zZ~o?Q@_WDcdmJ4dA%x)K;)1{a>%Zn7|M4IBgFpBK{^BqG;!};~J9qBz8^7@z&;8ly z=_$i7e5Tj(LjN(6@}ne?_j^$O3%$^fC>5i4`T?ACPyc>W7$^)OMn1UMQJ7?Q{HO-! zAa#a6^Y5Q|jP*2KkM|y>)Mq|+j0-`bBa6+BN%hlR7+kG(j7(%)$p5Zm)va01=~(ft zdpph-nu8y~P;Ab+vA&GMH8xK_Et$jdIFN@$hO(@SaS;jak92 zbG&!aaJhnu<&M*ROY!vyEg1TaaSo47#=B>aX@*EZae7ca<(|cu zbAP$y@wFu{a>}yiU?y0!9pArqg}Mvl@?(jUO!CIf8TT$${OkQKU%FZG=1I;EuUhV1 zcX->;Hjxj{0$HB3ZY=9Im458nj&WU)DM4KrO4IV;#e(^G#H7v<5|)j_yO1JXdub(b zG16PdV!7sOv1G{8fOrUCou_Xat`{wNZkW_L!FzHopenHuWbm9!424k?xk5+-(ol_Z zyp3$yp5=NCQXz!ED2W!H<8j6-CpA~=jt7@3hN0!S*1UE)g6LDh-QWqq^WWaT;Nf-4 z_0G`_F>US6;rswG5JF1jjkKjQfnR!cj&+jNwx{ifPwW>5b;iZUvTb@Y6B-_!!ysmPu*&_0OLmqhrxHHl3x+bJ!PMy?l@no#WowhSP(RThokJPiszR z<#YG>-g(}6xaPx)hSn;ANDRzeOJ;Q{z_Qk}?JcXeXWLtH4PEakMm48LnswXL?ABm1 zq|6xQV7(&-kkU|=5Qh$e<->>P)W-ALNzLc)%n$;qEOD579}&)h>=Rn%pyKQ*nH(|p zObSgqc#KYF*5Ew04`?ZHL(i~Vr*bT-DGHcOOHO7v0>RaCOOac!tGNi#sM1Q<=J7XSbt z07*naRMuJg2n5n=5egG+GO&z+;JIA1`?REB7(BV~%!IN-1K$ zWm08)`Sm*}C3)|I2gz))0-cu#Q(=9;cN@b0%if#BT9)1Ap1;-X_H?(6xY1=BPCGHS z47hOB1F8mK3dsyCq!NQ-28>{Y#E2mS512ATvW3J5&zPY?CCX2(qNsXM592obXl{Q` zyW7>)VBZL5I3k^qiScvd_@p~I=iIaR*?aB1*ZTh7|NFcj6rrODzUzq50fOE-R1%4i zA@ofqpmLn6F?~fQrR4j>h+4cxn^nC1zQK-2YDky9oLydcd|dgEyOK$3X2wXmG|A|@o+6E;3T!mIbF|=WwW0TtD08GvNb?a0 ziHs2+1=bjxGlNx%h)4#d{APKIlzZs3M2O+q$`XZEigjxlr$kPmG4XyF`b;J35^)M7|-SqS}V4UD!N0TT? z2!Uyt4*^)qiZV+GF|fZ~V0zEp^M*WGustqFCA4jeRs*@Ss(Z==#>FKe9^vR>!z>qE znWhBq3Ba^Sd0{VOs}wlXv9XTrNlu{^&RG7xI}2VrsR_~Z;@+5pt&ERd-+pXPwl~e0 z6bY~2U*O!5L`Hu0xeurG$~D}+I^y=#$B8QA0&kyGynVLed~NX2k|&B;nXy03xVbxG zf0hFtW8QmRa%GxuWjX_9gGyXLmJatE@BJgnG3x_w9Bp{#tYT3+T04wit)NT=yW@mg zyCV)Jl5gB^I6PNe+4|u&gVs9g&JtwJ^Vq{-wSDkR-2CiiqUv-#X@7_bZldjy?h23I6@)Bze!d5GIySuwjwaqVn@r!)%i(h=~^($Ad@CSeJ z2efU=*S_{O{`PPGmf!yE-+tnEeE##F=kuTc{1e~v-m_RNo@!s7;lB;CEPH6f(f9qs zG4u@2@NNX_!F9}+fveLed$W0G>5QZA`^Vl}R}D#YL}#Bk#zKr3(_wo-*L9R-`P7{V zlO!-!mbZ`Q+`jho6N4BdcaG=0w4)d)_te{aFTA{$aJKNgcK3v9yJL1I#pQP3d-vwt zIqMmfBkH!oxX7Yvfxx)bj8h&vZabrdyB9SI$0u)2*&hkkb>NM|3r5;fsNG?~A0vI= zv#whLj+65xJ_J%N$g&Z4FFKZu!AHsVB;$P9Fe?)hEvcItiRWY$uwuh2H+BX!PYk?$ zT7$QI^7-Az6g+$5gtzOK?;bAr>k*oA%elTYqR@hq zMZ?|m2I~VmnPex^RIO#Fuw0vr_}H}(Wh&nDnez5&&1;7%S~s+LCrT5XyDW4(XoZGJ znP8maY*o{@6&}TnD?3Q37$t&>+AyyhqVmWrk_d%JhVUdQ6543Fd$Qt68F=yPjL+WQ zrRzG3_0)CCX0_pH-jN7FA3M@wj8c-T+a=?sqv}1a4J0_4+F_%lcLFH{0zu!bQPCrk z9FdLaf?~euux{`uwbpQDqUm?X^dh703ktIaL5ONg<6{_)q&bpx;UZ(`x4E~_bV~j!injMv}dysQ$x5WC$`MPIk zT=3#f!KyJ-P0!uq1*0q>QCzAC2bFk?5s^q_ma=IqNrW3a1v}#u<2-F+8D*Nxdz>8n zuKTv;Y%VyR%fVkb2CQ``5jdD?u1-^me0He{@3^@;!oSUQo-&{Ntzd&tTeS1_y9!EI>R|n2!Y@o zUM7Ge)d~UFXfH#q6u~*97UWu@Wq>%`i_TUJZQD^hh(eR6lGa*WtcXG4OoxvkWkg?& zb%Y`cjj;jM8>CX87;G1mCdQP`c)EU$k_wfL>DmfsT7(qHBtvUWktdW{MsIt9HLSZ1 zV_Ozg%i+ZYghE6^nP-$mNhH#nf#7LJj&(J7bV_xpCX7Q+T?i3}M9YM(H}uA0tVIY( zl4^u=gs2EgBBUbAGcq-zZnUT8k($;gEPBDF0~Zw2aZcCuxX==8Pm-2IA%-@0B_EzMsWY0U;pF^+tGnBmogW=O z3UaM!I*T!b|83hDs@BkUJ;qvO2u$L18*KJxqUT|NY8myk){c4@9B)?%2+Z= z2K$|(c}3TCWMjo*6{+irJXM@7J3e+L;niy+)~)4a-Lu}bEL`Mx*>P{)kt@XuS4JFc zr(D^}@wTUVk2U^{ot*F8TcASEEDetZzx`WKb&j{sYED)in>IZ3t`|-cQ%`~Q zH3O0Ed?ZU0lR{IxPsdfdvx70_Va^P|vNrtl+e_{(ET@&H$Pz}`V1-ee$l=2Bt$Q1; zO;>E?g8e*l_xPOUI_KKX=!xpu7$fJ)HH%fvwW;FztRy-^GKo?ytFf(Q(mus(vUE3Lw=2Eh5K@ zinot9{Nl&2QfBXe?(5F-`rSFfHt4kExhtC8X+n69<1xtu>2AsEM=SciMR>#QgU9XL zKI$lnf^Q$SeD9=T)dim0E0Dp_G*2c#w@Qg5bk1>Wc zef)xdV7YMX)~z4O{`|&o{03k9+ShpPwbyIyB9Xc?+?WWaqbE-ko6b=rf-?KyRteX(bFR%aXI0Pns$*~VbUxDe&ub?9f`R6`RIUwSg;n#`Z{YZ7YAomf!*M zR8tlyi^g$L8D2b?4U3VcrR!S0edio~N$Q;}DlQfqyikmbl-)_m{>~Jo6lcqZBuN-$ zDZL+(y<6LZcXVBiwI1UDZMk#a4*rp4N|L6`E63Tg2)?uvUaNaPMf=}JZ`K8-au5K3`%regBDS!X%IdyM=9Pi=ISA?3K%=AM~-p)AY>y|`GcBci?B4@s7se6MKJ-G~|NiwL(V#HX3 zl7XG6V4ROgl9an=3sNbm+Q|M`a&?knZNvtN)-cKy_ZBUOi-vIuH)biHduhxx6WqQw zVO(g=H$C4vY!E7^e1eMU0bGc@aa?gx2ev0Ay@OR3z&Ug>gj2m6sg~r4?sBBMLd+I1rS|Ei)%Sf(aXQa5kmGaqF_Q?~?;YH2iqG3@-sxDyNW$;sQ zRIW#;lw3oeBy1;|thbzW9o9HnCqQKAOrd-WI8>T3neE_$Kxx?Y5o0@Ia6Gr0u|F+o zt>>%?XgMOGqevjwjxIX-P5}eS(R4k*^#~Njc!FS1u@XGd4IzeFNu*5Zt)OZR&RJa4 zC@Cn443P?Eg`&)J&ga9mP?R}|PLWdc`rT79t$6j;j6@6errAKSxeebvT5+*%>0Ck| zI`V92PX~uUkSax?Bm2{g7p_h@Uw1S%QbdJ}0wpY~rKM?F?%IGE>N98&DH6e~6r?)k z#@>WPNnSr(fge*5CloEI5SuTNC`qpq!Of3 z;Izhzj5Nz=YeNXb_%BDATn%AlsTOomU$*K8gw&uw1w(WO69PKXgfQ@u!CHtQ)nMU~ zD1pMGE{}|hp0`g{h#0U@QH*8-3F~@N5wW4YRH$i)5C~|k2a5xxFMYv1T!yo0A+Qmm zPVuG(-xCmsFerhe6zH@-CIvc?1n)rvigf6>aMt63qYsj(GZM)_fJ=SpO`8$opyC#B=<MnjU0=qxl#hw+M_75iHgj0+gsVV%JRgESVM zmbf4Wj#>=LJf{*OlEwz+%N05+=xn4cQu^L*8w>U(DF<5x z@0?lgo-YS$91Vx33yQp8XC%40m9wgQ?i??1!BQlWD#`fHQG*iD^#&OYd-;Uh2M1({ zxV(-Go7VDweYc~x8J~P%kL_{#*!}fMNVVi(d&JSY=B?A3&%E&7_JB`|7~!>}4X>Tl z9IYJ7HjrltsT$hUk1Eewi-z6P4jBW}k*0S;aIBAzNEl_3y-CUo`y~fYuwQt9-g;Jz zCB#V6JO1UJ4R4%V(k$iUH)l_5A@7Xi-g(8k^IRPZUfxxlR6XB%r(G-r0r(gHN(S;4ppf(6|T462;5cro4W* zVw_c6-_D=7w)iOHV$<`+NyT{;xP36g`}bG2PRopg?Z`XxioK%^pZnO;iT3_m!nlA7 zJe3OlNl2u`J+!uZhG+ObKJs@2Aq0Q(H-GbvQH0MmZ#Il6Rn@&8J^*%3%B-4?#>PO&MQ84ee1Dys#%#bO3a|*n|y!ryxg>u zI5Pa(SEvo*4@0@G^ekjQsBOJ~v>aOEw4l-W7`J=nP z-P$V|<(g@};%H^LSZ#Roq~h+vk!A(GtMI`x$u$SFl)Y*8Lw(m~Mat>YvZ^fyTRGQu zOE%4vTicQ=(~M2$S+x#tI|>~sOHGny?94`JrRc4vcM+{LMWTrsmUYklvzq8TCWS<* zg0Ac61DsZ#v<)m9N0|sZ=b2aC6A9i#NoJ#*WotR#^i-{5(FW3T%>VK~U8Ct7r>mB# zbsR1lR#hMsj;)y{(}VJFR^+77k*0z3x}xtL>s})8)cugoe7;x>3s|KorZb|&MC9u+Ot4Iv=cwhDR&^G48gmW#?!WF1$xN*?&yh7ee{l9i9V zu#<3YJmY`;Ao8krzoImx|p<#=32pPR6`&AQ`6-LZo+LP{JWdR6?02OveRzqIvUlO`c_p z63Ny$k_yMo?Hs{_WOsqq0zt~A?dgr7RD#dl-sROB-}mV|*vdIw8D2ZWx z*|BIHN^8b>iZq^8(-LDK%bvVwx6UJVLLgz&8oYCy*A0n+okKo@mNErRp4s z2)Jm#gSXufxRs1(Ed=jTLK1{$)%9E%36#k21Tw8iWW-5NnU9bRRvlXtJro*jpl*63 zV(=9oSW7ZB*f2yNkSfsi9ZF|d?`e%2$Rh_@I=q)8LXaebC4|xvB_lVsN^b1W_})p) z-IFEL%(GRNJb&r&7h|M%o)F{UNh<|HKt!-<-LPnjflNorWFVN6WLU@<>!`bc){-KN zct5C0i(IlZDK7)D#6ZMHfsK)Q3*Ww9@#20t_^;Q7<9W-v^_;8>CmYMAbIdo6b!Rc& zP`8%e58n}xNS#9%&b1JcsMX++t2Hrz0@|h~ID@H@f1qqP{^b}$$&K-QE7@|M5-m;r?_Bpfa}@+3iADrjwtAQdtW_A#zq5D_G4PFjqJ-s7xgx$Fob zV10lX5OE0NlSxXHk*@1VlN6;D%i3YxVBukXVATeKkmw}G^flf(WH!Qwh>wuzlwi$J z0CRyPkwousrX$M|ND5r;tWWgfsjN5Qbe=@;{;vb<3m8| z6z_VpxGX71eJN;(*vroh;gOOk1S%clx(b3p=nN4e0*{Oq6*Zc{#z7}3QbfjyXOwzS zB?r3`4lf$ae1i>;WeL5DgBQ6(d>>(dgG`tY)s(IvzDEaW}Iq_ajZ8r7n=?j6~XuHj%K`hxS;Mj zLJ(w0!j+u~Wv&rXa8dOvd&5{OHb(M)e(#K*dvSXRKP(as7cJ9q!HY%0-t>dbuXd*y zhw}|*tNy7xVt*K7g#Y@^g6|!-+*^9~w@N;FrFiT+PgfoP<&ASzbx$hc_VpRNvjQog z4uRtf!`l}PhYQ2=dl{d8{`+LD50SeU4c|EIIhxnpKd+cqhBuGbTr`5+*@%x_*Q{H| z_DDXq?L5`Id~J)n=NnFI!z))3{=5Bj=z?l3o7&(4XaRX5_}HGLOs5>L9OvEOBmG{_ zyeHvw)v!BKT$%lNJV&d}@&CTD;5)~LQIYX8&y641U&UZgHO>?EX9fTG-BXrz%m4G; zLtehV%ie4tV3ibjAE`UX#iqqO$Mq?!+l0oaT-|zlTY)^$>`royRvqu0H@te|KNjEc z&RNaj(sE;OIur+<(gksMnsK&nI9wW*wRwUq#!nW8jvw$ozW;}x03QNU4wh`s@C@(9 zN1ot)8hqgkU*IqP;xG8>SHH?1{m~yi^4?~%;p<=jI$4(S;)^f7?|0R8&F}oq?=T*Z z`OClj%V7fhF2or5>Q}!Cz%Ts5FZ}3!#4|j@Gkhd*bEo9iPQ$Y8xOcJP=I*1D!`(^7 zR@riQVVD*vT1)EIu&Fz4&J@>opFHVnjAPyOyl^AsVC#buya3yyl%IQH8(8q>dBvh? zxw<{NY*!oh(TB*oF&r;y3LSa%Amis>+$H}S%+fkdq&ApsF zRovH>RWG>u1HUzdNNe1~c10xxi4qtWFBPLhxXdV%eB%0qPdzuo`pDmZ>x6NUa{I=V zD?23@tBz$=Q>z{q1f3aD!;N)p>W)p*Ij#^=4xuax`ZS>^OLitDsn+-q*)%lJw zrK%dH(=pHQ6@w~S8CDt&=M9^><7hs#a2gi~!4F|x*S8;4p`==IbGP7d z<@oOX75C=g*0Dl~)@%%JZ)sD!?6Ic)=bQ_YJzBSu>($5rHH(O|6S`fkScoq}nR z49dq4FDKCwr3QNefU`}{$)cm}4AZQjHh~5~whm;8L@Py+2(%3BjB`Ho;uclwsGA`W zk7CtvgP%1|2^WkWDanJRESgh2j;?j&NwIE$1DsTK20gR`Ex z_dK^(a&?w7>4tG$B#Pd6+TIXjKr4w(6m=cBbGinlnYV%)ds`%0(HY0J-7$siIh;q< zryX4bQYRRpX*xmMb#zvc<~gxz@l8W-A~6_5GJ>eUcqE>t3uq!r1kNgll9Ad3oR4(Y z;Q}a`()NKx@F-$_8|bA*)xCFxznGZZIM zC5QqxUCa5hW1J-%&6}ZEQ))o6tb4|p=Jxe1X5+_ITroygjp5$1$At=wphzWeoh(>c zP1O$swUnAH7iinDt{r`6un~gP2qMAu2%VDU<->49>pY$Hm-Y^W@;8KlP{r`N@+g^* z=*UIY5|jWjB+m;ONt8w^Nl*#K^xzD5N3<4&V^XG+$q<$$0HF)E#wjuyf(tCFipDz9 z9I9p@BSQ#-ZGy8PGvwv97h@!NhZKWlgOmyp0*MGj2~B5+eh2|YfioIsT8IIyGs<#0 zSagViobT=6hQl;&tn3?3IGN~p_%s3Zc;Lz0egeLX1eVvmpo5$IinFg$;TLP$qi z8l0D`8bjSj@=Ri*9)cL1L&m_WF(@3@rYT#~G3(ASuUiV`8RwE)+ltTJ7_+P`eUzN9 z7g!e&GUfS$DF<5Sa++RmNc4PYy8>RpVM|VrmI;B z$B~UVKNM!lM6y3p987bvMDfkThTT#AfxUG1W(D85({Q@#{zDPGAx2JD9h=T``)2+` z`?~jn5Q3d?&a!PeTlYWl1aGP&MItfY429wkI0zc!DD-e$J;O7+8$T5z;#Yt5SNY0U zzQW)C{onJ_OE2+-FMI*5HOI%t{OOJi{~mghL4U%nREz{ROX|4XTJ+ zd*g=_?^Fw}ZRON`!{MSLPb6K}b1+rBbTEBlimtZGZ;J3jT|4nlwy zf$yDk^wyCj16et*4C}_y`-pXs5J8Hen!Txe%0$r_&z|dBhyVZ}07*naRI2Hj=8~Op ziV%XPw`{tAi$esBowQf%6QJ!hNiJ(NrG_^KsRlmUU}cHjeqS;r{ss6E$0tf|F&>wQYGxwhju7n|l-H zmF2YhK2a>CWPhvRX>blZ;7`Fe@ZW9oTe^SOHQFt=D2`0Xl^1BE^E#z!>H&u*qLy*X;DIuCxRDuQ-0z35k;aoTer+>!=`luS=x7ub4jT@ z4$zZ?YdhKZ7vw?=;Zr|I()+;avgPoiVw^%3Br+L}(Q^kg(nJjk&DL;QcUW%`qUPm; zE$Ys2xU}4#H{9I){sQ%t>l04bJ>NcVNVR;dRkJgms`uonMmXrK#aqknNFs!$>O4~U zM}3$X6kt+HoC{QqqqBkQ+ZiukoiHsm_Pk})I&!UuF`(qItvlxfnID70<`I>Z+}PiG zm>j=q3_Byu$FEJfGOGw-&b)Pm=;^HBeBIJ}&`Oi4j4~0B4#gH!MpPQ_9g!haXsgip zC^0V5^*u)`i_(!a)r_)~5F*wEoDZCBYMS0s6eR&Lq?E-Fni!=)Yp6CGybpx5BuP>P z2H&>?kszc(D?x}7i9job(h_HSf^A8pXMYs=*_%^#CMmN~iZv77I6CKIGb}cbm+f%B ztOHsp=8dEEk{Bfd$g^Zne}_mH2U4zW97n>jJ67zD1)xzv)w6Ql6tw3<0bn+p;JTJf3-))nsJj~L1$S0#A$c(fsiNRu>$2w0y5@iaAOOjR~l|W@V!42PK14!%X zd&BnjgsQXDePC;pFv%^es;4s+B_hQrhahPC9v>`i^cZU)Sc)tq&r(_xUM2{Y;%$#s zl04P80M1*8k%%P3fKVDCbD$e+En-jb9xo!THxL{Wi&PmxDU_0U;}K#|hIOVNew)%b ze`%Q!1}R*G!CoVn!KOpxcn85*>b9flEYIzg>`iixE;dxPBnnNMYK|{zcE^f$&LXZ4 z=m=RNkwV~nAQFiPbXwB2Yof8lBqfGRHMKDiBT{M>>lPPljLc~5pxQ06bhwX2LFIa6 zbkv<^oGS97pcobOHt^2r z3Z(?OhO1-A?l|SvUP+!PwnhoZ3&|T7EybG)e)-k?%jbS1c>s=EbYuC;lcy$}2avcOaOZT*%XxNa!@1)}9i8#iCQxL_2U2!Fz_>^_Y8sj!5;FLc0wG|3nld^Yo+aBq#M0wQ zm~VRaOU2&g$D?L{hK~+@DhS@PEcw!xzQiB@@gMW$FMpXYfBDOgy#Fh|@+4mi?fu(T4BMlWU;6kSN6VIzWyhj2bT)v1GELc?WE^bek4)zOy)pP@4hmhR z1Ur+Q>B_KJH_Rp@PL?gJHqh8enkZ7O2d~5koeM0Rz>)I&>)jQ>T0Z+i{?N+eWYx23 zBUffQ7fry2$hgcHl__m+5v+%n!bDN`5fM_hCnL00Snp}tZU}r4!+t)15Q437&U(|Z zYAhFPL%FB9H*Yzdcg&kW=LVlek78pY8w);Bn3I4z3_(D^s%nVVfc3m_TJVpr&H451 zoktcNMXLDpb2I+&wK>a8k4_bP+Yj@VR1XT|JWa^cgspMO$$ZV7xk0ev zV5j8tV$JcgdxYQ}WtyLRVGHLMym?+zb(U*8qlay@#s{p2Nv`QP9cS|m*S2zYcM7sR zXI@#BjiFVNBGV*VT`GbEU2o}vyCkMG*SCjjK?oS9ij&2Lcd81dB+d^C8t*0VoVU2X zVQ;MXnU`h{$=$lM80)w()m)k7>`pR>@V(QHEJ-<9Y%rZ5bdLSmm{G2%yMR&?a_`wx zJz}Vi5E%$UheSw;P6Q*J(04VfreilLdFfz+#B=AOW79w%HG5^o?Q2`K#&cFVrey)q zbFyq$w-%)W(OazPaZ!?I8P@ceZm=FmlLW04k${$g=pv?VD0Roju1@*I+q=ASV~a$~ zAvxIx{`Jl|oi}vWbGYcZzpyw27fVa;28D5=EWPWP zI^m!1R;+rDcb;jk*czqiAL8@bbcTyn#j-KnJ?)4gAasFBC6$HP48&Q;J+nzpra?+g z5DM=jQfY)v&*{?HXkC6qyeVgB_(92r8unDh!<}#7LAH?;OHg&&e4;$9d%$7a7xHI0w7ak~^n8XRD6uJEKR4VI?ImUY+rcx6gU|WP#3dW}^~^<7DaR z+n&}&x~M73lIe7eK;T29YC94UIIS$T>9KCkXI_{+@*FEOjZ&I5+8f^CppjWKUw)}To@GZACjC+l*ESkABy08(2b{`LUi{eJg{^y_M_oTduzAk zt+S4|7Y)U+a3hAJ;3(%_CrPR-~HX+@%rnpBc;SYHo}(Iyd712nCeYE{j07WWkZRBjzV}0QM!ZL3H z+tZAzdB&=4N#j6}PG%)nwk9m=p1X@4<7<3~T)iZtukValHJ1C!nj}#am)k3{gx2(| zYs=}nW79=$-`HYP_1ryPGR~k-k!dEVh(vGcy@lXObcP_%+aVZelts!^GAbmAEXY;H z)oH?(4IHfuXUm4xMaoPOqhw>mpbo{;_dU6a~o@n;VnIX2tl zj2Etq_{L$!vNu$9jfxWKE!JAp=dNUDEV;RpgV^woZ#A?|FdMyhAtOe3<7mOM z>DV3T9PIC4d_amJ?C=MJ|CAD3+n;iDzUF9U8087Ahl`CN4)&rCab=eCKfJO_o-8?B zS-yL3L6#2b=S}N5TQ+QZk0h{@1^=iDA)bbgSD1r7b)|EQJx`{A`WsLEWZy&cLNx^IiA{V52M%C7|PLL!i!381#V?7E_mI!(ia6TbTCF3F`gg{+G z)mg4@d9H1bsau1V9f_92n9vy5w3aM+;C(w-we;4KY00D*v#KPCwJ7aCD$-F&hz&7# zl<-IqvB83aGy#X9OcXb^QeJ&-=Mk^TL`k;BIm_Dc-TNy})}GVKvTh+tg_45afhEDZ zp0mZ0(YT~YQg+4#Romi(CP^d#NtQ~yb)2mnSEngkrDkXP-igVps^w%+vup!4CUhZ1 z%g8IYu97B_<)&rb7*19_Mb#sOBLsoI^wy0cKy(PH2`GG2kR&8Zk!X!p`qGjlVC-OX zpp`>}K&D`vNF0h$E|`qQOh-9QZ&N~!`~5F}|bgo;^(H-YFZLJu|y50<}aqG`J!?6FXiQCSc~LPU`L z|LncTl4VzVrTJ}j)IC?Y4vBzFq#~Ov8q2zyW>qszx?d#Ug3nR!(bm?39@L<%%}UV7 zqDV3mi3AYJBV1f{^j1CCjsOBG0W_;p(AU&G;vPSKeBZm*UjMf;P4EskWq2R3-r^z% zZEo&?1eMki0T2X2roTt;W12+BQ{qg7zhzu1b!ZdPrSdt%P zDG`~M?4lq9N0f%nLZ&r^v3NX0g_eR4(|R6*#f|}dGIkR@A+0$sN@9e7q{`BJRCJd8 zFksPmE10ZB5YlrSJuwEXPdJzG>AO?#5CWb6cyJR@DTFL3a)pnET&1M^C)XqWc*!(2 zh^RPONbW8S_YO-ot;Oh6m>1oIvl?v-S_+(uOjAN%L}`%HqE*EE0q>@S1~CP;>Bw}J zO7Z-J41kS{LE?QnRGjCM(}M-8`3z$;qxG!25hDj4oy;imjIrzKZg5Qf`I-eF0mQ+es4`tIgQKRNEyDM|i)6l3D*Zu-EFPrmxSY;u9C z{YdZN!Rqr#TEV=`X~&68JAMg&@F@rRSJ$4mP6hkX@qFVD3K|PRD9WNhn(P&lNJ&9m z7VNu$MU~1We|Tu4J>ohuAmxf0tH9qV)zuK{y&sfwsMoP9W z@^sy??c$9=5Lxd#>RfWXEO>Nh2?%0{SRZJHiKkZ`>(0`37Uv=oNhSjam7%T*l!$a~ zLuNFGi-N2D1yTe=a9nQ(q7WEWu__h05}efqtGeK5Q8PNr<+|m=>xTd1uOD-=%z6Lb zXNM3WMCN72gX1}R?Z_K+dGU?qHbTH^p0gQ7ep(y;z==VzN0WI6`p+IpIhg7a=GPt z-(h6r-q|5qDXz91buO4!hGB9zpN5pR669JVrQq~n&XbpGUaWiaOfdQMJNaVj=CIEA zcRxPnqn9mDuY1;AGA*NWEQ}?Sf;Z0=Y}<*;E~Pr_Dq%Ecg{B=G-RPNoN~V`0r|(DF z;91TxPUac&!Z2CS(~BMJE>Kq$d8KKm$QYnd60HPAWhk9f6q2hZux_^0h2n?zYTi1T zvAp5ig%~+nlx*7ZhQVkNM7;G}kEzhFC`)9Plk1Yf1!7Prt#Ll!gU64K86{GrQE#Or zAw>3rlZ?@T%9v#att6$A zEUJ>WYiYWXebX}-$LMl|cg!-4P?9It162;&z6a-6-oqvz@ZK}^J;vlWYodf0G1}n6h>!_)qh*BX2_bQ4AtIHeC=~OmAS((+ z8yOr(Zz+om>n5UuDARf?WJ<6G9}swih&TL#gph!U^$w#T&nvvQ5FF0AgmH)pghH1U z!b5Mt5$T;p?HsF8Fw+qm2kKnlyrj;O;b3&mJd0el4le}w$T&@jNh{S28bTl`AUYu8 z156W$XfZOPOvW&IG7-p>WVERmE(VJhr1hSV5DdsqEo`D25O^YzNJzz%&J%+};n7*i zY*sX#WmuGdw6&G)PU#NG0j0ZPNNGXoknRTQ8o;4*5C%rNyOC}IDUn7(x*N~)f6sfq z&1Wv2-`;!ewbxDGY0dxc_G^u^^j2$t?)kSQGU|B|9gM1LrpRf4Sm25x1wxCS-x(w! zxy{6rH$A;|lg`E|MJd5KNe+E2psqLWb|h^x69A=p=qH^*^)f;cWQ&{Vbv5B zj(r#n#EJq}JWch2WHtjC_^o5&Ks5iJ%i>19fsm*oxqYRT5Xvj6(`kVOENc*Fsr@%v zmJcB=IyzsdCiZA9WC;TYqce!d$}IJ)t8v~G%Uk4yjG_Gb_Jp0pD>m7TZ14~}^37Eb z!92#ff_)x-_qN@l3AFI7=cY%AiY|I3YtzDdiJ?|xu z0zj34G3ZChwevxhr?e?K2vl9*{PNROpWMWQkL_A}_#gd{#rr3}dl#>A;Bm*7*TKL$ zfb`vujeME4p4MuFG;ade<6Tt5d$mzBq1m;pr>;MZc{6Z@u4z3WZn{3{I0af%~9x1p$;%>7# zAN(8x74~9B*VA^dt1y}qf$G|M-7v58(wD4M@n1OU>SBTWyZ;Ky z#%oq_ao(-3bp-Y7G0u0clk-+DHqu;_NvrwQL*0dU-L|AzCI0lrF^(-ZG9f8pxVF+z znD#_BjOZpCjCY2&edY1EhIek5f9i5Ru}IzSZ1(`F?JN!~*nJCY+yVE}F*uWo2mTi0+&v4uN{hlVSgm0N@ zMUM~(iM9w6dzZM_z|Z9qiBA+p+&hQp@m}dMRv*Z!S-BgMEB+>1n}T?8C*P;4uj`dE zGpHJiu|`!+%IxzNcM&BptQOg`{APX@sN^$3qZp*$N1z2XZk_vUYHtr$^kCHJe;&pF zh5f|g+582oI*DP>`J_{7;>xwN|H66!*D*(Me?1J<$O4>r!B6>5W-}KV$)%Uk))51*`q=M$3)mML{Sm*@w;GR(u69H7)2?!%wUyF8J=(ZxpaZRPxtt=ijG;}EBYD#NSl7N4qN z*o=nS1z2>9pim*}Ko?*2;ogvk`P}XepX+v&f_n!P)MPjn9=0vk@BE8KnU69@FY1D( zNQlKG`#;wsr{H2f(WA_xg=u20f~E-DdhU9QYE~m{XW6W^^?Ms&S#S;nwo*C! zw(py9rmW$LeJxs_n5A@qP1Z>A{p#oC1zqL+`0rs432xd71Yj(sEpHMYux+x;^boOm zbiZ>#u#x$$v;x;^!U^ovpIDR5?O?T>EDLMl-EBu)X`x$@cazbS-DAnJu5IIf9aJr= z_~Ugh|M~_fTAlrAc|y8d8nV(Kmv8V!#-l!d#0fn0l*^(b0}|IhG5pY!RCg=uX&;KB zTP^oHIeu&o+Hbz6nXBv#`0q8>?+(xYbAR&}<^H2_mWKVVSMyfi?)%mDke&f&v$#0A zPQP*g*<7xc%@yq@9v?g=}}a8Tt4-j zfeTVMEu`GKCqFu2RbrbszB=Y1@xg!ne1Xm@vZmmfz7?iCB3ntR*gHfOO^kb!4c+pa z%(2DR7Z|!YcBi+QS3pR3mZmy3&Gs};^` zX2(=z|KJ=zE7fgri9{~onYNM3@@jw`>#dyCg=6?kSluvGmW8KQgChFhb<2i}RGUqj zpJd>)du`6`sr8he0BGRznutaQkjj(Jfe;sJ`#�qP~6jEu;S+y%7(2wUf#HG+~&n zzP(K=Lqvp~7O0a|+SKMSJN8QbJ6C^>dWZH$)s3PeZO)Tq*;DcpI_T@(kjJ|>bS8=H z_*uH_+=f-Ggdsq-h+G=x z;rdzAZlGmLhqGU;6J?OU-kvR8M7bJcRmL+la$}&%@*;^ha{|jE z#;($64%PTIj}X&?q-#qI#s*X;Wj{<@KsQS|`9oh$TySPguq=n%hizr|r~+FG^tuXp zRh6{YqhJg9e%m;+Z3+AaJtGE!7!@XEyi)q@xaG` zA3;axRy2(22iwXmF0)OQomCqO?(XwbpahaHh8p_h7lCMHLV3nG4vcFjxl1w{B@KN2 z+$tLg?b~(<3{0%Qz(h`R~Lk)~A!x-rR`W@rbC;`t%CME2o za?5EBGhz0-d0x*ZRCJxtP>gWpg)F@;9!QN%ld!Oc{*COb z)Okpk0=?FJtrOB@BhFSo0n%)vbuzBJj5&P@6hSr8Ad4C51#7F?3y8lWHpcpbw2LCi zUc`^0$6)c@bR7*RaqA9;fjhXtuoy#@KZjS?aGX$RWR3u7-Z;C~)}46L*-h51G2i#X zEQ&{ROCRx5?4xSuv5TsPP-SuReh^dg{D`TLSr8{YMOJ4&w~njjBbg&Y;@@orywPzL zY;0vXzpK@MZtQe!fq{?HoUtp3^?;w#ZJ}FZ=puKvb!W-@>pr*F-H{dnzas3TQ0e3C zUile*_msaUXzNb$v^-wo&{iMG+|3Sm*GU0nEqCC>38<$$2_^?3ejM%)w+H0 zhg$dn(UjP`ux7kjl3o1ciEa?RXZib}vXIW>-b&~H*U0-AyX4O~ijHjqAI1XtVYFbO z$e4f)Ta7LzSey5LYTFCWF7k!GRAhQ)?Jew##hQ%ggTXfQs zRnz2BiQSyg@|DE$N=A9D-f2zZ3}^YVOPNmHFMn9r#smwG1IP511;E74brqZ=XJwmr z2-|b5wVLHNx6RGAjiU72O82)DJxxiD(LH^7>d<|ImRPQ-tif0=0~7B=8|vFVh4>MV z4o<&|)iD2h+bs05@2*|gxwMlbjaHopMn3QFtNz#-)t7YD$fU$t?km4o)I4H%NnlRF zRW1uM2I@o96NWaGO+3q6jPrHFpUH)si;WV?Ha!7KP@{CO?^+QQv5Z@EFsEDme+yC< zyDoI6?k8Dc^7~}BnaWQHHk)5rC13!FFEEMcAOzl^UNMGyo1~UuD+F2 zar@#Du>Xp)yLji{%=<+0WvKEi&1#5+(e7`fwEYjrljsnVYRbxx>&chpqFX0)CwFrL zJ!YuZ1v;TbP__jcN%>!;-5fV6OhcpFxw?1crFhaH1c1%9qwzlUYtW&dU-Hpd=?KyX zVw~5*;}iYT>a+hmUP3qSyvF`Gf6Wn}7x}I3fBO}M_ziD!m>=Q(Lu~iM9O}E6OZ~4G zae1?Lyn9ZZRAWU$4^iPpu<_UAdlk~9I1y$z>8nbosIBRx1RPj9dcAzTt9asT!lK`X z1ysI-DOd|<{;9^fCJQ@Ck@E=OGyyGveep z6yMrzZU2wb_eRYy(Zt~|Na?kVkYAINDh6#94hzzxg6!pD>8>4GlEx=m7!6~jnG{64 zYjI0L5S??Hn`0Ps45~U+bHiuLom_o%Ni}Gjx z7aj+Wj#@6hiHXY>V>uGx6Rqj3NL0qwX?a|8ES!|xJanawj#WSq{_|&&NLK?!<(bc) z$UTWC#EzN}6KaATR#h7mf0J12Q0u1vL5hTrGVkFqAj#*jGS`fI^t$WUx+qn>O&;-+ z$BC7B&7xx?Z~a9!*>{yNO!?zGB$XaqnpH$ZFm&#c7Af2NQdy)x3-dF*3?ZLR-HDsIyxQ0aLWX_t5$~jJ`AaleJo+XJ$o=14s@ig=nTxY3M8-Th)_@|5k0)!> zDrG~8b2>@MI|l39+a*!u;SFrw5@-Oln8(Hd7Sl2 zx^Q^Tpra0|)Ml0cu4MIVMm_Pb*mUtH8k8tHa|9E9(oFvvU3MZls!KlF+m%H8DIxKWji->5?A7zfwvDG*2;J}UZgQYZaI>Jg zz{Y*IS}ax!n@qDeQ;9~w9)eTPdgvj}S4UnCslNg3zHP2Dn2iPHQ8OuB`Q=l~AM&Yd zYD{)1F`JkuNouA}TG-R0*#5>J>FF zc4}&eU(>qlhttP4>xK98m6m#nxp`XCp(#J>DKRGMNPZR<3=6WwhXn;pTzNCkhJp30 zRO?)TJ9Naz`Jn|p%1~b9=Ciqqm@w>wTH%BHCz8qN1;g*HIt?=P#ha_skdR>i34b9i ziZRVK(RL2ScjBTPSyQTLYX!wc;{v0WecR&g@icVeojXZuttm6Q`Oj;L=FY}Z=m)0^ zjOxGSeYRqVURIK1Jixr@v=BWcq%UJzDccVbpW~TQ40Nw@ffFYdS4K5yF9t(P zh+*9USGVMvTkY`@E+HL@DrVIUatb-guOGq+uqnfdk1m}53P(*+_T$E!`kckwL!5;l zXmjeUt$j}%6@}hD{#}alIp!IGY$l|L;3qm<=LHQ2yx}ZhdEQ!+>br{&%d}pPQeHq$ z<=_}a`cfLfS(3y55R22>Md#YPUPv!s^ol%R{QI_U>yFunx=cOKtzoZafxC1!cPk^j znW8$Fd4C(8N}lR6uz^l>t}06&X}%1KA7*~{Q{FzNcPo!2#~v~RnIvL@$PLwvmSD^2 zk#H#Nh^vcjnwW`@n>dP2;^|~_2mlu+GPCO7pm;JoIo0IrNjSvJ1_?jO_Hn%zQ@LeP zFz|FcE}m73gi<55fv5KKa66+nJXsKW#;oeb)zg2LJKo;!PQER z3kU~$Vtno&S$Eqsw+8msV>9j7qTeVNaE;OKZP1;IXt);I(>d&R08N-x^60LuPjM(X z?$BQe2e(||>Mv!g4kIdi@!4vv@mxdSFooGAZJ-4eULG>@S(L3huxexsvqm7_BpRg7 zK{5__it~gP4C(e2HT6@002V3Z!i)Zq8GJGZ7C&t*DJ2zP!=Srl**#R&Uz4o zVxo+WBdM5*fmuuMLNQ|uh)6)h0&A4{kiv3pcG|cX=IW2-SQFvnamG>f>l8)-=(O=* zmx{{xmxA+bKO+z?$%U*(2fG`?>Bs;~J>>^BHu0i5RskxAy(6c_8L3o;G@@iS?iC|| zA2#JwPQ{V|*1Yt5zTu5fF4+PcQQnb8Tl1NhyBl?k5z125j!K3uRAnZX5q-obyMQ%%eMyBMZ3 zQ_?UKROV14o)1}`UIkfyl7pn&1bJ__>;9Zw4OnvYT@6zkib{S~hJ;^kvVOe^ba&s_0P)(8HgCQp2l)dB(WXrNxm%vEdl0Q!{>dERll+Jw0G;^M)WN{ z7rxXp<~5XKyVmtncyJhq_{w7%SaLBdk`yj8Y$l#>eB2y~64KcH`2Nt=bp33AHf-xj zj56fW?y9nS|3*FJ{F)FLpIoF~-44vY!B5()MtDkA6EKnIz~Y_8-YZ4rOM)v-*KAc7 zgtKj_CtTjOHQxufz486u6@h+t{fz24n@bgi<14+xAhTBiZ9EBvpExopdySl0Jha@J z%Iy5dczNGehz$ipaNhpFB@{zVJvO?DOo&ImcF-79os4W3is6|?3>4;Zhsd8rp zZo5nIgC$yr4NunHX`qa!X4Wn5!>56$riFG1gB)&E6v2b|E>O# ziUl3WWfo~II4PiHt3zzLzOi*D%-g->SufmhUvLxW)lzlc{kpRi57XpP8f4$^{N5@p zrU8-sqL9RvObVM}Z*o#vxXS$*7Cl6Y?v;oi3Lwk!6clZl=wK&%xhOB=%hr#zN0xeR z;X%NfbERk2bZXuBeq{D0oo2JpGKvgF zo4I@zq&|E>D7tT(!m;N;I_6BzoBV%=NpkX+Dlr7k6BRu=S3&COz2AsR{NLI(5IXh^%#luVe^fY6C7^&0t1G@0`DKny zK@dfOG%A12)h8P7lWtG_1BuLf;FD4JeTYq@O}nPeuR4N`-vJ+-MN^ZNH;6Y2Cq-VV zX?z->x=H848W`feISd$k&-Mm?YQ{^Veeu8kx7FvXz%eb2d&m9HMT|?h97Zn%Gw19g z2Onx=2Bo%t3^9_X4G6fm(;G5HpRffUjYj3fY0~|Anrwq=B3*TIVH15pr>P-wT=+#xNlSo#3_;k$GpZs=#LTuI=67NXvusflFnc8U zb;Ep17qS?A5sLPvMUUo7Zt+MQtF%p5{%>3YgZ71VH2fGvr}8MPnr5Sy7_pj^;~b%t zMY8>$F9^jFFr*mLzzRNnF)&_0s=-h}eB~oLKn_<|-0Jwj72iLaq2pd_Z#YL853c7D zPl-HOeV3642Wf3*D~P}Wx3~DoViFPyGH>LF zOy|^7%)RT_uZ_0V5vb`lnNvXa!BKw|fX>WXp+23nR;-WiRr~-*>(mPvFY%Mg1Z_?v z=sKxyuf}Mms=et26CzWr2l8_#g|o=1=O{IKZVscAwM4YPE1YD#C_wT_N8KeSo@(OW zT@ID=tp}ZH{2`v17sQ2Ybd4TNZRb+LZPo z(_;PG^!~qv&-Y5Ud-s#$y~ef!I&L}wPUz2)<)pFLu@(NB2hd#UETBFK)*i4amY!-) zC!|j+-E(~PKrOj!c0^qvpnU8HCH)pFy^YWRt$-yXV93*q9CBf0wj2Psyq@Acb4_Lid`u9!FceE$-ySz>lR8H4?C|T~} z=~ftkw;pzUyFEZws;fFrrE&^2cL!3%>tY`FGzK;4xO{ydpOzMYxq~k5=YQy=r#Y7g zM(3-x8TP-Y%UzK7LE#Pa)r-3G#)k%{JV1r{N0@E$mBv)n@KvAS^F}3&Y!OkRciecK z;IWS(m4jaY0$k!5k1j2mL*?9ozS!#R{gTZ>_p1A)0kIE0pBx3$oPyROkbAwdQ$e zE^87%cb)y9~B`Sl-vO;D}O|5=k$%0i0m-0y(4{3N6bQ z>wLQjTghVR0+4h7>tLNC+@p%pZSm&Fo3>Yv@40?f9e78t@p{5TTm z#f3aL30DxA=@n*_i*c>pY|e$lHp~O0=!^5lpYilrgKzyLEiE$#qJLZFCD(uZO9#G1~fa@6TYnNsRizM6W1oeiX13g zPLI2OGDjp4tWOlqS+@WyU=$dS%Z**7dI&2{S!PRo6(LLu4Dft?CMh71_)8(YFM4Mout~IgjvqLQS55FQ z=}Jn#xq==wZ&&-s>xlK^GUi!(uE1q=@z%e3O)k`;t@HXk!w^su% z+Ike!G3XNtxIhrqsaz7GVvniqm&nF$#2K|Z~ zWH~7!YegXi2$)*-K%7jSWaKh8u&2#8MfpOQ+-;KZ@H2#rMLb)1p;33Vo|DDcULz4- z>s%nkCX*{!+8)IvPghs*hk9cC)9@aW?65BVA1=1Buz`xA1N1@rB2Z`@yRgwDFsYav zD^IuRr?Xho8R!mRGH z)F=Uu$0&|bW%7#pvJPVh&-P73H@bf?AgX2CJJ(%dbA6Tj4A* z5u?DMrHAVG0hCkiI!XL!t6&(18ODYcGefT)tBWT8D#(Dbm~ia3S)G;i@ilkGvc${K zWP-34rDu~Yt7S8{zVGDOW<9{C5v6SRFDa2qs4ZR)j>P+MZj9r`8zmCO){9LceH(*- zu_BVWu=}3iL{Hn?%ZFd>zCf(q-;%=ZB8$}gyVX?^Is+I699pc`EPX*WJtG| zvYd`;n3j?0P!lx}6Tgwgr+^bJ)q+%&DKa7g_F_u5DMT~QJJCyQhQdZLnz<|WxL~V3 z4xHuwrDJp)vLY51YINhy(JXQUKlM1tmS8O<P+C}{To%Nxst#%67YI92ZHf@~! zsymov9{<*{cL8`ADKV!#Z1oED22-v8yBW|m8BqoUuRC3S(;@3Xz>UFA;3EJ7-_>>0 ziAIF&0qkSv-d;sNLuM}h3!`*PIkQeDZDA}X4v&cunTfgJwyZ+}-PVnasU|DYgzq1P z$OtmIlu9)xKE}>xSJC;lF4fy}=V^}nt1t`2ts-oB6sPK9#<9bbp?#uCL?SIBuA(I# z=i{f-3Q1mxZCYIAA(WDz#6+KuQ)mom#Krycmj31F{>j$l5&g?`zGrov#`L6vWz$+f zqbS+7Hv&Q2Mip-L4<6aIh2CfgJWnh~uf?%(ImZp)*`P6rCo2nS94QGq_2_j4U%t8^dI+g&~pQmrgVWC z)2y9t&0hQ*eeYqEUT}3YwzW1_S>!1hb1fyF#dnDuIL}3F}F;^C)`m zP1&8F%1;KZ?(Hi_oE4?X^%C(3DHfX~@_s6_K;=`xr872{jNN?fF$fEP{ERen z4_#+VZ#B~ozCsUzv*>A`PG6)y#1F58hy7Pjrg*DvhisW1rr~wT&G;GB{J94+uEM~C z&m!LXiyy5epv|iO#O?u5Y`f&MaHvpJiwo;?0Nok~KID%t4Ct5B^3V4RRKdRgmXiXz zfFR0_5SQL?cw7>%=m%%?>u_5sVICi3Im^~xH;1e$Mm^EHxD)L5@LfB-!a#aO1t@tn ze|Qi^g%1Zy_m8i|FZq+pFCE3mDZ}J2o~5o=D@N)Ly3QYtHXuIZ9U?zbzG6rhvA=m= zw|#tlIWvv#)O}C9Pjxr3fp|`7{68;%YeYUZUNmWC0gE<0Mj_wNFCMG=kbuiR4H9h> zX(9p^s7}>TS;c6;ROfDP(ENvgYxk;tn$T|z`&Be8ttLZ@Lj%eQ0^lFNAY-fR% zID&)g)IJ;%y$#BPs>KC5iLqqjALh8&NP+&+To8PapLB}{xBO7rK9fIuA_aR#Zh!gj z<6cC2#p+lOs6m&}oVM$Z-f_df`*>yjDdom5ZOPG%{wWWYq>SNr~ z?cWx=)sL5;EKJ?6fR?s)o1RH^&umj{T^U!k0@p|pjRlO<5ob} zqlasWu>NY-sZqu#-DXqVQ&M9^#)`%abD}f)$s+6Eh;gaEUf6>cSsPX5kY#}SRy9>W z;;SK-jc#k<6FeR zkD_6y@_c*LvQv1+gGQP$UpaP#hvQJF`ZBRstniw4Nkq;_d@iE7w14_l*eId3MzLTE zUMOQfq%1kcOG`$3?gKEk7ru2u#XRSDAR}99LRadvhojbiq-7quM zVk~glM&~TYzcYMIU+?J}N|uCJKmv8tjM^~MKesb-_0K$(`)J~Wmgd?tQyjC3tH^Be{}8q_dBxkdD+*vP#;ai;}D%h z|6R{j)2_=D8f+Z+`ps8_3)8FbI;yLw=K0%F-)8vi9Q!wa({UAkJ-amuX(a(7glQJ1u9obE$v2 z@Btu!L(#FjHWp&oTF93RwiDXGgPU8ktkJu-c^ft^6cDz4HuYD;2q? zqGg_cfMy5&JZ zFP>94##9v3Y(KAu?+Uei^>f#)-6vu2whw$SEkt1f`8lvahn#0W`0^TP$VtXz0ZT4u zs-}4NA_uKlBK|IQj*BM6KCAu!EzuVB$*y&k)4Toi6^}D@4d!?9p;P#3gw6doQ~G{h^$436N~}hDgpp z!otH-i4y$+w}jOgd>?!V$8w?$sKx?u09eWTtbtHw0h_2oy0tvR=^(KjC^`at01F%DLwagM7}52}hK zeWR7uskUDs0mChv56SK{HC z7z0+BZg1d~)%u;baAj%sft;E$J_&Z<6wja7``r;V;2)n(7;8p>JBV6zYK3xcVws>k z;0!8OjhlO#*2EtHWQyc=7N3u~p0 zwIWf_SP&~rwz8vVBF9wt8i_70FGG)DzF6^0|A{Jea79zDzgfuUAAU}F%-M6!rIu61 z$?w0sIzuHRr6{^wjAH%Pb^({<)AqlorKq8?Tq*yy?QSuVB1)zeJ0sF`lh=4;#I)4l z?rRfrjJHQU==qg?UEUS_I_RD(zlAQ9qx=KaOpwzqkddfmlys6Q@^yLGW=<(@{^84+XmM zw}*j_FI?vk^fAxJ#$<63SZ(%#c>Cf^0@4YW!j5qJN`3aiwvPka0n@w^sFWhXdZ9^| z)7{WE)%v?s;H^9UDpP!N8+)wI^r3k0N{y3}XSzTR_y=n45&Ha5c94sa&r4mE6Q)pV zrmvJVtFLo)7uWARkNj#W{k+|J_r=47``W3Rn5f5Rd>+5?Hkk-TsguW)u`Woo_qusc zPa6YQ^_gz+5z2Y?O%F893p8W)obcE2sj%+VE5n1~v_ejhWZU9J7w_$?4?@1+c4QzX z20FH!ULxUj;wI@Ny!2vHPM^Ef%Iq)zSV5}Jyw|*>FM@vwn{ymZnPt(zqGcZ_Kfp~{ zPM@m?l-L)j1nFO78?=XN^1OduAMsgUf2?bp=KnS9c)fB60U1De7}9LIQScuxnK$Tc z{-gss(cJ*r7RsGzYq?jbXmqr|z*cxR@OJx#AN_M;Mxh3~;{~fz_jj{S96;_|*q?%c zmtSqZvp1kRUMX9oGse%?-5`Yb0B#vMG!BNw@a>@8h$CKY1O$8*ZI^N(j@ z*sIUvJw-%Ft0J(|PY!-=E2pdBf2dfu{#0!>Ve2Z~MYUqPcAgmNC(i=E&DkQ>JEPTt z?-^Y|29K+Km=CM`hMEv4Xo2ehhmJAd)F-fu_HG2bZ==D=xcD*||EvnfUK(ragvR0d zbOZwTld-M3+iUw{maI5of^qx8=z6Z9w`jZsd%{V;3_vk(8cl#C+v5yLH1&xC*<-zx z(SGq#=X@JV5u|Nt?C7x9krnp$ej4tx<_unHimLY-Ymd$zdS^(J`WU??=Kpkh9Avd{ zHH<@B$(cxoD)a!W&npmO$@AJa?s~rCJ&1QNbZ)qW7gT++09Wo<7cAW*=gjk0lCI@E z2hPf<_PQI_=IRq{Y)lV(i7v-eudtl)6HU>vFfb{`SA`Si}>wG?N_BmUU&i z9^XN#M9DWjWXDd>Kj(Q$_f14^EF5M)>#bMo8zPtZV8@tkQ`4X?iw-}?d=r%t9`{#G zg@ywafHBmYf==SWDZ8lf1nMbfzx0joiWrLt{?co^ND!VF$qF#LBt*%CxAuNRGKFeM=7ok}7MItG1OeUFo^i zxiR?*l^3ZxcBO#9l^q0_^?yZiA#YLhy1dF&bGphRQzm2wVKRQT1|=2|=o-U8u}o*W zY#&vBwt`st!f$-VW&-3AH%8oh_v;WH0x@_97Wl!^q4D1~#F4-LySz9;msgUKYiDOm zi8(x!GCJ1s8zp7QWt^C8MKp}n&c+~?%moq6M9vSZR*v`m%Ff=Q6+b)bIgclmMhYkG z1IFfPu__mF(Hv;9NDT9sQe}x;o%;jg$+1po_=PL#QAob~p@dQ}slKif8ynii?c+2H z+#V@IboT}P_C(Hmr>ih&#+^n6VKoJ!WL`Ol8^N?+hyD!fMQccH2)7o5nO{P93aHyw zEgb2w_FSB&6%wWxbV+l9#@f6ns$Kb6BHw+>WbS`4&+t8FmIXAn^>{kf&z$2YS#HNSQlO)6+Uc+Jq_A?WsdKKz)plXrW zFUgtx?k*Zyr3qxE%VTgtJHF(OV<+P8*yF4=if1~wbK9eZ9*S2cpfF(e(idAv<`JQS zJ=B>k0-5DiVmPc<-_rDopuj%MV*a2g1H(NBE#3w=QFz16Ys`M6rX&~QH5|bRh^YWe z5#Z==aYPk5k{0-Ik^dJHAJ4YJ2K8YDvfIoQK-INDtqD1ssy6;`%*a?SEGkZqln(p7 zTfzEj_QnxS&e=a|TAM@!#3}#TyT+HUR;D0TNk!olZwq-kVciLIqZg}G5Ggk{T%k4I zWmRHj?ABnyXV}L_vD0W-e`4MMF$AU68*1zY4k9S00e}RvfvXnwj^4n?KcJ*&fC&`D#S9GJQTMc&skwJ0|$J#1P!5)z%{f3oh?oFpewCZ{1! zb7Typnt1O12O`}|^hJo_H%=}~$4Vv<_6$a3$$n1?p&LDStSji8+lIU8?Y*pt@Vc`-^A3lo~D#dpg|X?Bb!r_-Q7M@?jdN#B zj&aKxW0N$uZ5Dv*vwCCQpQ3fy^3NInc3%G54e9)P5b)u*al}v0LC|6%J+?o#P7%^; z2Ba*VR6PY&2?i_zvKVk3L>623!%QLvYL0E3y=mPsE`UCbj+V-A5jC2Bysb|J!dLO> zLd1%`FL20X6g`alK$Go}XYZ<|Zn`Q`6nZxgsbd{Bggww_J%LX<@j=L6N}_7%5;s5B z#qVUAh=b^WzyG$eyYE4H_C4^1o;@iM1IJu_iIvcCHj#s%jA#6-xoGrCl#@ISe&;d& z`Yj8p#p^@_mC=c)#!dhp=g90<^1r(^aRd!+{%_%`CJ7fB%~|AF<&g$ogO+56krX|3 zv9~6_9j8@w(nl(SpR(nw_+Znhka)-U@jcV{$g^Cw%c4BaVxn97;O$H4kPErI&n6Dy z&@F|Dt+~_z^{F8PIglmi)BPfnXZSaPm&%ZzAr~#1R@xx{-l}E@&w99$a&I8gEThu! za95+r8$vesR8`?Pdmi6bS|cupI`FI~)5@wv7+_)1K=t-1i-*irSgi$2E(Am<{~na> zonG}HDhZvEo@wWXO9Rs@e$o2*@pO}|3857{pAyog0ar7|lwe<5I8n+Z5iYzI>O z`9y6Z4LwAw_Y{Y|**YLW**(rXn2~}UG7Uts z;xM?ElD=Gpj^&xVz7Iax$R+qnDX#PV%!GgLP3J<#KeMR&84=bLDYWDu2yc_=pJ$mr`2VwsIox1H7~(6!K- zQk~#>=P5}a-rdKgjK{{(+}`i><}FKP5DFDzmrzUXke|}{Ckq_;watg9_bxzD%GHE# z$#NFz$_wrxMQA9NQDL$&GGe%#$*vd5_Q%A%WI*L93VU0CqJz+>P`9N_=r=-^WNEZc z*S1e+jxzacMT8zZ$+l-#!{9pf1QhpzqtjrsZaNv8ZR=kh=g@^9^en8!s)BKp6j}F< z>~}!Sd>F=hp|u`Xmye)A6piisa0VUnk=rXnMyxI3*Nvmhv=nG#!ikG+(kSs?h`5;; z(d2(ZWCSK0Oh~qS>`b%APqoMtM1_7PsXl%>cfee zyf+QuUpYM(LuZjE!mWEhCD$(f5)wZBSTal!uQGDiwPmJ%+A!L>>UD^obcM&u-PojD zDk8ydsPWqWpmh-kIIBsc*HSKCY+&jnwW|a?0WtjwktmAjyh6zJoVzx=hBx-|1(O2v zPWQ0;OaxaV6U-(stsP)k>807GreEy*&f&KNCp;50)9!wSJ>w| zXq#57?TXJWpDcuAF13vZQ3h+KF7cDwfcNSedo))?ZNedeCS z`hBulg4LLH79a@Rd!ukMGJ97~j_D5XQ%90jAVli-Bh+``VMwA2m11#&Sp4}rG$ZHZM*OJX7K)lO@Cta>);g5CKaX|5zORy z>Zpa;qYW%b-y{HbcKjf9FHP05^N*4X@9-5YsTzl_%j9Cu zb+lr1#LHrBkRp~JR4w36SyD8Fq5z?sXM5LsD;<#GOKVE+uY{0h7`{kNe^t0MH(28k z;7_%x+!vfD3kEB)tmkcUM5gzz-#xZ%RZDj50E(&~0)zkjFmUS5Yf5a0SOZU>%csrE z&Bs}oZeRV?iUAQ23vIgIM@}{A6ABNfOourU8(Co;P!B2~1R|4@|5OyxaSKhveBI@) z{2+y5W6WRePG56i90&yQ860rS$1qxedjhq)IG81Qmnb8TsGD;L4KV!|h{8Aa&r;_O zmy;~72QQ}m`YsxGpN7;L(gQuaw!(KE=7Z!_6|__3)1daD1j&r$`}fN2Ck&9U?nR@u zMLa;X5CMh^pJ6CosE*7sK4dF~RaT^U{LMuv>nvxQZRGGpIn5+Cs~%`3oQ?qt(x|JT zsAs(m&<^2rj`t=Fg=HkYpAPw4b^3ZMzV0h_;B$o1Fb)PluDHsF#gQwid8?YmyFVE= zN0o8;)npS{NXHHrf+3nnCPx8~Gdu2B$y4$vW7tPP@xuNUY^9gc5nw{lxtVeR?FCk3 zmNhT}i_Q1<5skwF34<1SOj8by$|)~;sfv_QqgRb`X(%VY#FbSjU;=1s^w~HS+LG6i zsc|-kcE4kE;(!JYjAg?phiO-|RSD1TGa48>prOn6!I7yo8Gi>TwK#(>h^UxalNY}v z{~$=_XHx@F@`)04$YDgrW|s&=#6%FB)DY2H05Ptix@2UJ${2)UN?5Vui1h&32wBbu z5@6W*ADYfGEXp?A+KQBPH_|B}-7THc-9vXu3DPz6Ff^#ZAT8Y>(lwNV2n^lb4d3&A zd++1m5C3Lh=6X~SN> zpdnME9iqbLMEG-%?8~qP7#MF4{~#jJB{|jN&86<>V#atW%NP;7U?RF-#zJ^eT|};} zyxqeeDxtpwMioOHc)=xCc`occWGe^FQ7#~>f+zX24UOg?Fl3eHNRA1v6!!bBuz9Fw zddTAG*8`XN^pL7_*_FhzR?Ielzay@>%gxMn(on<3?rG%cBOreVAtY~ackZew1AG_i zj~Gg0_zEnpBin<(xTzr*x=|VfIg?nkGvwG@Yhy~5#jMUG_2lw`b*?7pcpye(bjA%O zX8LHKs&PJLCWe^$oPfvC8dbRB)uXigbnS981MZecg>v*Y!~G{kpJ?qlrPFigZoITV zm0O$RoUPLxp%5M7oE45j3nK`DKaHbeG9<0Y2;PuF1^2?)1233qW{e66BDyLgjFLXh z*Sb+37r;e~^y8SEsmFJ|!;P;LPInv>+^GO>^wJAk{mH~vZ+dnle)TkPEgOo-fpQ4q zNjRAXjGEpn^%rYQ9V!C*$ICW$g7w==${3}#W~VZsf!oj>Y5QJ+>u5M0Bg25|9dqz) zp>I)9b1H`pa#@zznV7cHhTpo!8-3q?F zpl)>&P{7W?lG$Y0L~58lPSDe%-HXCV&!Adzk!2?M`s%^l^Cz_Wv|PItmwjQAYjMek zB}D->jPcOt2x57^(S*iL`S<=eVXje z?RB$dEq4-&Xs}rb#~T9AXI(2pC1!Vv0JsJPO-R6khvBbeL;+trCGIERg|lmW=#VxyZA=1mU(`cHi;hc#Tx(bl!--?ac?;zS5{Xe9kXo?$=x|oaAB66?^XPT z6Aa%595-y`r_WOkTUR01?~M~ck}X_CmV@*Y*wX8M-f`7h@{J=56>@>bH?kKHrQyB> zV0%HGY{0fF&c8YSokc0k40-3IPexM$U~WUeo<69#+Wc4 z>8BL^71%Rz8fg$kkmDh=RTNW8K`t4X)*c7JgT>)e0mmtSD%ZTkejxbEgx^VLE8Ask zp_O7LZ81%7QG`4M4_?}7niGkORszpUS4C2EBZcEM5mPJD!~oWW4(63k@e%cycknbENp_?!Rc+*CTb( z=y7|Zi2^yn>=uj8T4{fEr;P<(^XT;Wy-5gCJaMP6<$sU>`GY9p<6OFy}Xw6 z#qU--?yw#*1!uw{p+#5J$j`L}WJo5*tgnOpoNq_%n->M4MFH$u_ZQ{s=?PF8agFVV zpi&+9tVHO&*T(!Umn(_N@U{VVbCrFgQkfT-B6~q5|0i3vcrF{zWeoi&4c${8}TPo$Mj;~BV z?zeqCR4h(=xbP~*&53XvUZrO_K%329h0SYO90}*ma2h^-t6o58`3zcU4lKtn8fzW`B+NKu}TS;=T!IQHz8ZkO3n>~fMVp<@> zEZmVv<&~iyFQ^v}Bnr*^9@?c$PZMmo22k=ev$Cd2CPK*q71r^}Uyxha=;v?|hbG~i zeFK^6^G8^v$@{4%XF|C)#J{T)xKqKeJDweSf+g*JJR0|RQ3ut}QMh1zVrX#f*q_$y zGp)5*T8kwur6qO?Nj^5b{x5%<-2w#3x4W~;y8kXca%wBJ!}2|!oJTCWjdEu{*rpaa zK2<#}{f|o0wEsUU$tCWr<&5?6UkJu`ugiQ))UFCJ=EG$vrDQ9Gf1@uz=rA`>Yozb{ zX(=Jq&eO5zljo}Nv~mi*)?l${pX~UH^)NTl)bY>Qtxd@c1PO@N4xbHzyK@vSUUmF4 zEO#${{ zV7%k6(s?)-Nn^_*A|aRH^_LV|_7K;j8Xplt09t(V#DDUZwrw7lJuLyI(B02(BUuO@ z?{l@&3mx4 z;Q!v^r8P(&1XB5Y;qy2o^Lo@Vij^6e4i5>Ym!uo-8GF$7a=Y{+LxqLjOfm;1f^4Z$ zb$qyPExAztVJDwyK)(31=DW8~XZ6tCN}&1YGR)0GxbO#7cADA<78`Q&eVFg|2T@hr zvyWplRt>>kz~!OcW(KpVNKFAs~4}cr}7Nt4V3CEeh3Kb|EVeqp;(ktKL{r zf-0&3u}aKX%UCz1!e)3J`zIQ*7|;oO6v?E?gadXCD9~Pow*naLKF6g#W@$I6u*qGF zCFmyI+}`HwFMVdHEk!ASIq!~B3r(|Cg5_+=_yB}I+{^fUz8^auqQ=69y8ri5EPOJO zb(M?l*jg?&QuW7sRgbvMvU*;T6WPydj>}{WS`LsaVKIVF(xhb_Di~p&Z%C1P6s?hu z#iJSznw%SAiEL#Gh_Tjx3S${=$1(nnNG2#@3@}=+qC+iA+cuKd+jMxTK9jPY=B1sU zDUe7#Y1mN@d=ivx?FCbYKQrohRG6s7gct*r9C}~z>Q*d` zJ%lUfeHb3~0P-w5VP83|QW&|S_2&-O7B>ZO@G zx%a<7{r1#9-`1ME^J;={FdfihAmL4}x2Fl5O5=57BFjw;`sZ$1JcUFc6^|Km(-)jb zdeE%(B5bAC=jDty4W9lYv5`3~s%}9nyRIT#=wOTv zbAm_!@Mtk-{msSyx4dRY*_CSs-^C-5v_tfsy8alz=o zAnP(5P>fyoTNW=aiUv^)NsRuj94JKMjj>!B%;Go(%74-O_4=}?$0!C5iB6y>ft5Tf zW8PpEdWth%TpYaKs6+?EE2#Mexs#GAL(V_lzK~S#s`|idSyf7-sDSpzCwk%dwanxt z!$MPOBX{=Lr!Y6b3hSKSIlyO;d znvSH6IfV7mSBF00a)V;TBE9#@l?l0{u{^-b z@cdDc?X&hW4}%7?_nY82do+(u znpT8Kux`z`**20vZoy3J#;MgM@z&w4%(#QZggf>GmvP7q2p(aZtq_v9ei|e6uYd*;B9(WdnCglN1V0#_>t3Pkkur`s&hqkdZNd zdM9KYCVF@v}nw=6ImFsbf zBd5#;2Q#s40ipFzl!=BBVKa`}fwQTEw@-I7CLvyk^ApRFhLLPdK07^`)KcU-b+MCE z+rdk}LhZX5wqr>{uO@YvLMkc8zJ;Z01wV;+?(dE^ge_1AeoW2vQTS zE1V2X|6t)dN_u+Xy*gyN@1txga+2M<7#w>&Zmw8Pji(8%~|uP03&lfS8)#B})`vu^^#s zv?qQly^Q_~t>7Hcbgm4-b_)rP;Kmo@*LC5e|$XNS*3 zJR1bkr&IopTFh8ZuxgC{>X&>+p_Mg=+p!*l&LVeci5su8go}xX#OM7)V0%e31=}k) zxp)pO<5Yn7>vz)V4xA@kmO(KXP{-`A5D?!_I|qDZW+NFhr9hiq#nC46Aze$O?L4?Z z@)o&_639GNIV$B9xNwNZ5D z80al#|E?=v{dI)Fomw~ec{r`YHCX_vk~BmslBL;Iw@og);UqMjrSxLNbowx+p*4ee zj9EZ2AwI%rxT_f)6muTNQjt?+G;9g$==jH1fdr&%KWf0JG>Ae^XZYZA*es)I>+L*2 zoG1ZM!TA=OUom*CiwCJw-l7x!J~!h7i(bauHAJNQsW2h;9}6#&wo&HVQ*)Wkp3mtY z5mUl>d}PD74WQ3y>F5(CgH+cBE%6<7o1U{R_F-B7B4W{N%Tp^o&Z2L)BnzjNHb_NnsqL+%K;> ztprrBw-c6rx-FKDOJ%@B0U3Az9gx{t0S{mL+svbWl}Yl6ObZx6*^jZ9tP@e*55(?W z$?s;#63RhWxU8g)bqI-kuP*YN`})-FIc#_8@es#=+{-X~dy^!V@Q=wA$B~NviQ>TO z(ymPEz8?4qJtFn25hAt$kwu3dWqQ_F&{z3V%rRty1ZzE<*2oW_JnhW@VY80S-t*M` z0{*kN{6(aT{9)V9y=?+6xv9O5r^zGbcMB6j!5)&6X#Mxe&}CsaCgk5ycIMkaT5~;r z-7=?BgYsB(G&G=TwbTGQsOYT$6*!{=L+N#>;Oy|`XtxF>dR*wUa7#3<#bK)h5-tw& z@(rI#S>H?ktwbqOS&!CTkxGf!)HC+o1(5t(53@mG%c1-M zz0wcr>;psnejB&y6Po5n0e({_*6YLdNcn6oO1Pe_8dRE9jJ{uMXT4ipvz-QuKd8m$ zDVJEVBuIAtJ@v8?k<2T~D5m7x-cR}?cXp+9n~0v;Hg@v>D@sW%9o!A#ilUv1pNZfP zymOM^C7YdaRXLZZ&^I>=hg{6?#Vk-RZ3j2lg1pprukjjrAuN6O6z{Ofp$N+4Eknb z^rwi(IwIvDdM*)@7fJTp&gO1d3tA`EuC0Ilke#KKCYdTCqDVznxi0ibMQpWx z>|;qTGQFx=e`wmS-;ahPgFR!(Je5RihJ4_`#)K(*UaTs_&oYgC^`2s?Sx9v#mB8XOx8)(vaD5+)yeip-YeUkBSdv6=v038H@=;3|yZn zH4vU9+vWAhU7idMX|L*)LQgLN&Rn0Saj}?Iz&Q6;5h;uOtdetp{ENnPj z3BSE5-)=kaW0htfsr#~z7PvormS@$i%yv*NzY~u4iK0DmwfaIcTy5Lj#P<8nB+_2_ zi&WtHG#20~#x1N2bsP~V-n+pg8w9pIBjqqo)!7A!S))6vU+G#UK4jC&{64=agZ zNWjS&JRsRHR?3mKu7xkkMbpObtEJiFh~oVXlvrNtl8=q>^VXF76jc=RS0LrKQT$3D z05eMdAVh4d#Z4B2O+d=1UBY-VKNSw3*wlF!NfQ-DK?Ler^m-cZ;_1G4NGc`2>plRj zXT@r{E+v?nA%qo-H+@rvFM16$AN zOMMPdVc0wuXO;Vh$S4Wx4ILAmzcrS>dmOs;1#&JKoUV9+pB`U!ic%3hb)9+r5RFg2 zK5ac<+^Q~eh0R;tgcZv4Fh?WR79=Sg3E5y@C5oP!niZ*UTZ5@S9NL=H?J@(-?)LY| zCxo`l!{=0l0!CV?j!}(-2qx4Ztlg}LYjgrWiGk1BX}s@_)5el`Nm)xAy_NN#WiL*- zX}_&s25uBJZiD`uJQXn=i#r%Kim1pbP_mB1vmzIHe&6wg^NPwqGjGUvB&|Q)J+-qD z*RYj+_ez8meEVR@g0cGjTo;us7DBLxR4qh3xBp79geg7M$Ye%6sp)A=;%>|GuO{2+ z5|q6C5Nak=Oy5>n%tlmto>s18>$K}HPUpX+rvM|a7X=cKU*%R>I8$T^BG#u&r!@<5 zBV^(tAF^Clwk+q9k}lm=dd~%B!Th2b9!^Ybytf$>fpo|V@GswbS|Tg%I%?S?EzRv0 zkC@F}nO)TKt_(7rD;qg7RY%$dx`5?^S0A1cB!@S1So{x#^-r)3JExpgha#Kug098M zAVccz5n_gULX1U?;8a#a{oY6-g_#R+lQvo=>i%_r>?q-t7Uq?D+i31 zIV}=~TtbBM)riUxNxj(Ag~y@uQR>pfm3w=p(A%TE=j$b+(D;!$@B_}X6}U&_nkHE8 z`p(;+U>k^TNnLKAiHGtAI?Xfe`6+=iBtT?JWH^21mXyS!(5WFQa}G_g zMiDh=eEHLJR)2Rr&viw^5f+VE;^NgwuYH1hqv2ccaT9w!$+mxALjmKVJ=@{QyF!=S zKN+@?&O=>xkq_tE+9!x!k>|VNoNa?wKyOZW`r*BMDm47&n6()F1WlFnxprr@ZlhSh zwDSHPWXA6vIotEDtGB4)Z{``~pWja*f~BF&xf6+`{BqPCUQ z;fL{2M6a#9kRg|ad+L;jXicqFy@;9`SFPwiFLD3DUFyP87#d%HefpNV7DX}X%KqWd z{a`#=;evD={?q9#6#iM!CT}AsnPg}A-y*qs|M{rIbD-NiZc^nmdfCcR6rEg@ZVoL$ z;S{-QsvZEwQAZ}|uI0dBmO9N%z+qao{5cxlz(MKNKiIz)st=r^}oBxenr7$-w& zUq{o*$Fx+uKF#RF$HEya0@UYFBs&WH8x8{eOheZnwci5iQtL0dRGsS4D(zjA9nf3RwtCS9KyyUu5&;Kja5;=+7Pb~HW$ z52Z46HzYnhTkGhAm7V~fl2oH88@Ykc$#LDB){MdljGt3jOUf2RjLf})g4FGahlYan zX&2Gu<7j&|84xAs8ZUTSE^t50p$MXb6VyK1Il{gS@tujyxP2n?nB}X$K!h}G6lG?{ zt^ZT+xVKF5bT&UGXtOo4zv1=>;G4|;3A;9=F*?Q}Tc70h_~%5T7cqu;$Ai~sjo`mX zm%%~(iqc9%Mk%AxFs&;eV!!r21MnO1l1K0)31*Vv-cQx$D2sp=mno&`ovD zSBZ|0Sw`U>JHzv$!c*nQUn-<_o*%o|imoz1qy=WTgl$?GG-B#4IB`4LY}_^s416t1 z)HgTm#D0$lx6tt&Mgj`n1ch=LJ1B2g=?_=ZY zW3!x}CGHy7#b4710`{QRU22-{E~b_^pcUlrwa%K=Ab*CwZJJ)LO&&{MTrYQ8MWhy+ zSHCd843NG;IdVZB7}jD1r0}T8pTZ?@Up)0adL= zXtQV$V~pCc@mItwk?an3-jNXD{T-Yj1se~^52QJ=w!-`aid;=NRic71VZ1tEl3=y# z1m4WCk-R3d*1si;P1*!}GU#D{DvO17pCt-}fDV>y{WAZW{eXMcbqD=v%0Ye+zoOO4 z%dkRmt}bs5SrezHrwGCv>)wRRzT5GFDH%Q11~rt5BHuBLp6eumV~Ov#Q|Y%f!EGo1 z*=ahT7<>v>9vOar3cpUa2(5mNAX}v=sJWfr$5$P=PbPuP8ypilGKu75sBw99sh4nQ zCUpgw_P(Rt-CW!oMK|a{OATCRNt-Kocue*rJ?q3Om@hqKhmky%Ptdo zmukVphGMK*{gI(w_DM#&gS0*Fyvh~jhjWag#|M zN>1pwvvu|D4t|Q`@{4R3$p{%;b@m3SW~hQMP2I)9`NXpv5hA (tj4zUO;QEQCq8 zZ}dC7ujW-(-z`nIsLk|eVi<>#v4*5 z+f*F`l=o6~ONZQAVMDYE;myObc)Qzj_;jd5B)iyY(y4DXNSMyGy*nw%VXu>aqJ-5K zgiw*+y|E8OA6`^dM&R#LiN0ey(hjyY;;9iBg?(Y?we!W1_43YY0p9I~;Hhn4-kZ~t zmapvmeM5E7ym3{1S9m!mIvsaXbngb9QKE0b)a%GGVRN8}Vin0}z{X5z?mZNXm`-l( zTr|eBz^HmR=4D(dhgZC*XBRc0iN6n(LJVFSM>>D2H57A?N`<)APZl}h`c0V^2wDVT z?)!WlMn$};ZtkdeuQzNvt@%(%D(TRzVk-nU44zW!=X5ordFCJw*P7=Yvgx> zKFGfq&yas7oE5X~o2;@7^LrJ^^;5cH-L?;XWB`>FN>D}YTRoJq!L4@Di;S#c!1wJM z!tVN0zZEnE5*Kj-@cru1na$Mg@M-c15NEIDZcLctPU@6)tdfyK4y=!&_lnz+TGv(? z*>7pc0g0>6$CUmQ5{371Cmdp#Axn>9C`MnRHunjvoBf4@S2DP9K(Xifs^@)t&Cgdx zg=!tN{PM7QoPtLP3?uS~PsC;$pCK-K{|2o}oHV{QXbv0ev0x^9D<;_~5j?=F7F{tOrY`jDWQDpXk<6K`BCSnV_?@tm zH`MhuN@q6xGwjvIlqh{$Iy|N$*TNo z_Y$N*ir{e|F9S|^fjtm*FSoYR;8Z1WOZM3~r`hB{ z?x5N)T!K9t@Mx=l+O3zZqEueQY_BhyF0Vq&<^$5gDBvWP@;5v!M3|qy`D zbWYZsCjR)fsU`(CV-5nZ%N*MUhY(l4yHL6Ph>hOxrubT?p&G>G>(s_@|Nt&B_Xc-XldTMj1Qu=k+X&i1~ zBVt9CU1~IA!$Ipj0*%;y)MmhQ&6;oOXqxrePUr_GLvpgon{nh(*xISqIZSD=<>%Kz z|Hh-;yHpy9hqdb&&4Oi9V9^~2q^{**OxYAezlSc}EICN=k7hm+HE5*%-%D3OX>hAKp? zRX7&Lt)vHZ5$Qm2`!YF_-%qg!dhQxN5Y|C^YPxL%CYxRputjK3B_p*MB_Kl5*cNeh z6UJhuds$4c<3Tn(nvxdIQtAx^`IZp$qtSTC$hiZ{hsY|lu}Wo=Dj%Rg0JscuymG3H z>Md!@aNUCMhWvZ8p-A31q$;m?ah93*9#OdatIl*}+;m(b`8aHDwKtNAv1MNUg2sSa zZ-a7fG<=%>)oe4qm-g8Fug)6Rzz}BD?K}&#cSqPVck@g0(zvsYUV~p@YQ9q-luP z8n~848$c>NZX81ZBSN!Y}rP(N;O4Pdd^(tEniV}0=HApK);`Kf>O3BiEEr^n1zOk0%2KB&Ss^| zE_gV|?LL-(;wx1z_sQ7H$ofCQnTIC1@m{IR}?m!o~ zRd!MV#h;CIce*<-FFW+SDq}9UkH|EhtmWe5V?`+Vzs$ZG%7*zD z3Kk7ad5Bvqt&$@rFXWN&7Q<_0Y8|{sOm#@Y#(hQ=g`hAu;bFx}Y_$;jzptqohX*np z#MO;{(Ov{@);U`-7(Ey`uEP$|I#<`&}W0|w>~`1q$I#H}7mrTADq{4cuS4A$Ox=S*F1WWJ(Kvm-3TG2P>;WET zJCIP|nAz72BSMi1DjHlUT7^hbcs{iU>)AIY9bCD#DLgkpj(^eRV@CWavj0IF*J@|+ zBjh7v(Q_VGKf}EsL=bkSw*B_H8N64$4(Nz)Ykd*&%3po0jo!Sj0MxTLW9LmoNW9(`A!Uuxt4`8Ti!G_W6Qw@cQp@Lg^# zw$Fuj+zop+=I+3JG@RHFrlsdnKZ~~kAN@A$e|_}Lv%qeA&SYX{#yv*%MprY5GUREv zPtoAja**)-Eaz#r>Z~Fv;qcMF>Fz{$^R6%#?yl{w2j0hyiO&h)wkFz}Ir3&dD4ESG z6+4bUXa{qK(0TIli4u~5OK+1%Dhl^bwfY@{R!B)-u_Wx7lz#aXYg!GckTc@udds^! zk34%l-!|SY8`1VdaouqisTANO01-^AW|4gsINV($yTN-i-(RPkutdgiznvP|xp)>Rbhd8K z2Xx!u-pr3uM?YYWb%oRcj(ujKHQAMW+=UvoVSgGe_RgJ8=_I~asH7(jM%LlK(#~&M zR=EuYm*k{G@~&CwhQ0gOXBhO)3Gf-rSp0vm3aep9#hWhr+m&p5hjyHMz#pp1Z09J= zmwZfJg1DLh&L&G0NNnq&*<$g`()^@Tk@KB9s<}HgOTvWl?`6?4N7eBIKR{Uu@owGr zIv%dqWoC}$*u!E%kqrcEx>$V<&_PP|J(gx;Cub|Mr@*<%;--*AFISksCS2#Awo9$Q zO;b_rsF+s|ccVxtSBtXY?L~`vO@OZxB*D(D@kWd~i!}U;_a4m72l$PSUWG}+K?ntm zxje)b8mN12DS0w;j?cue^SOM|1vR2EiDTxso{E<$2SOC2#W2((hK$c?!~2&J`K1;p zZ;U<3PNWxfI`y^q7aiM3Jc~VMJV#^Kw$arVzn`S#$5*};-}5&879t=-dSUvz<{(pl zKJ2MvhR2;#t`@Whb{5a}B|CBZpOS&W6zwl0uov(wNIa7?7KEahBTE1~J$q7E4wBax|8YY93!|Bo zN2A1Nnt;2N*zTt%p8F}!R%#Kouqmg9%~*1(ky7M+t=x3y{y%Pgk+G0)R%A(k7o~$G zQ~8Ej@R1I4Ipg<~k=dj8R9bqVEjSYBz?Ihk>x^s!YIgK;%dKbS(MLR1m6-6a$;c1 zr1gW@{OJnyfMw|c{E4_rh%VBcE*rubCkCfpaak}SMe0=?| zp*%d=xYqr=7L7KVRv$1d^{HWVZVp;n>J3#7)21Fs27D6!EtSiQ=X6%4Zwk2rd^_pd zkLWOFV+uYd@rFG3XqXj0`vKUIy(QvVG{JmZwOV^XZ>RzABJ;{>LqM*QibZS|4O3-m zRv-n&-W<^rBQZKT`I{*UN+m!Z1v$XMF+dGILFZs+kHn2-1s0BM^u3S~%TFE<<8MEH ztHUeRULWqzWrnv6`QKLpX`PlO&5nP?!${8JQ6A`pSkLr2AojI9n}P^Jf_Li;399g z%=-crI1R92rfoa9cAOcs>ieHk%YInqD^m{n1P-lGR+a!?XKuedHS+6EIGd;dQ;9(7Ag##b%hweUOp13})_3qqkv~8c^`TP9- zQf%3Le8NK?{4=JnJ7S}oa}Q!IynF>}EBXXY-K|l6kH}~yqjc!XJ&F}_QXOwE#oiB@ zhymKpLK|-KZCrU^k$cK-{qGy@;1mhZe=ld1)y-J93S9y^eNAhw#iejRuGgok{px@O zd%yzj8`cyxNtWlMd)@bYM}LE#;6=o;=wQ`t=PMJ#^wIIK+)mqz@bo`GiDW0u{6(~U zzFI?wR|_j%8?0I&KQ+%i&ClGUfA zS@H#h^%yc;;=y|rZg?(6-%xck8BuncgX5_kXeSV&?|=9K$%`X-3uQus!Wv8kLl?Ea zhE^P659LXRagB&AmPtFJ!0Ep+%we!G>2Cm#71K>D9g55G3^~WJc2)h*^qwH`(SZGy zjk}fGOap42O2bB!1%yZA{7eHU5916|Wb6>5tP&EEnXbd-F1nji7Rsvfac=={I_SUb z;I3jqb3u`OuhZ2ZLgdE{GS}iVb|NM_{z|ljeZ{TD0hFj(5;oq-b5q2dzqrwQI$l}g z+GYIo!S)E&Y?Y8;bm4b~AECa-S}H4%AS%vtak|&RYTuEIzceQei;d;?bna8DAdiTD z<5icElk+eXU+1npmhh#%Jq^N)7poj~;COW4o?SoI{*hiwww)qd1d&s5NB}qaboi$I zpwD%)UJ&yudQU6G#}SEIX4It4YgtsPOArM@utG;cB!adBFd-KO%;r~8fp4uu;ae!3 zr{IC}BW3t`f<$*~!QTZR7y0TT$QM3I&+rV!wxLz-mAujsV#NgV_Q#*x*u~$PFTe?M z=#s0H;dIEFv68^1GJjwZlan-Bvj)wzU}SgJ78%$;#T2JVjecW6V&ZVQaxljPo&O^LO7tuB3>_g(Emj2^PasTp%jurqpj(eh zhjK)DbC;ZC5^`rfcxPz{=aXzboX5gXhSg89M(L*cXX4c`vf_IT)EPoK{>Arl(?d#! z=v#r9_Iby(!)Zp@a7+cNjf5;X@0dCw5yW1^A)rZuJqnY1F)`uZ`j_TA``~b;rGRtG z-;2dJK#H#6(msKf%@QN^{8E}IT-`n(V8&AyIg;xJMA-WGx^fTS8B2`8umO9tJ?b|A z6ql6yJ#cb&C2h9Vpa-2^*ngX%o1K;ipPA^CAbWD; zvl6nhAcB&jEFYW1p~Lp6n$Z>KkltQj_$mfJ2tszyjP@SJuPtNlQ1F{r z=kJUWnc2#opK9uMC53$Mwo6b9CIMdMmM^>STi=|@VS;48AE?I+w&tIX#0Yfd!}n_$r9TL_yFG3KIMgSU{}*NEUh-(U3zNPN@^F)?X&TyL>9z6^ zHw+dUgs>?y6V6>M>|6$}z%T&@MgWX^_G_fHYFrv`21(S5;Hz(Q5b`3inmfD&pN$27couAv2B~ZYJQuITBtLhMD6`auKR+s# z=D90y2{N?P@mx2g>9($c%z?@JNk;wn(X#b0n4SC%g2vI}EO4i`#9b=_zQ9Q`sPA)S zq(ebXA81Zc0}@dOu%SwLfY0=Qrdb4?hcL?zc(>JrtO|;h>7R}cNsq34;e6tc%9o5Z z^zDd~jj%iY8yp(?qRU7;2d}tV!rl#ncb6tk{5Uq8&olKo-Onpon$>7rO@zeDhn;P0foy|o>*o5R<=&6vx<1B_fBa;|9_H;n6OI9w+kG8Cd} zu9Ln^IuboP)cn1s zB*NdPMn@ME7-FDxh)XNWD&xR_CmDcel%&Wms<$V%_*=hKdE-R1+DJSdWiAzVG=8C^ z29)_={~0SYc_C51eETJ|NVTDulU>#>T?c99Nu*#bM}G6(IxYu{i~*p-i{(6Cy6+!k zr*gDLN|SlNmW$V!{|vq&m4&^6_i})zE;X!;`=dhsS2keUR3;*!)V99Ti4Yp_X;oqq z1}7Z?XgOLr<e2(*8{5*iv3qHJ*QLZ6MpDa=A{1T?enn66L@qsU|Qq?3mPB zw+f|1PWINK({0r+%3ip(Uz2VB@MFvjS0UxsQ-<;T>ZKejj)5k%zj}=Wm-q|3o@MG> zZHoL%IXyfZTG5x<%v}{@GBRBB*>DjtV<71tLzhuS_o^p|h=fz3;x=!XJaQ<7XDt?; zjj&yCzQdpN)KB^j|J_eJj9uh`5+w0`r4lE0-gm^t3aCO>5w^=cT1MW}pGk7i6gfw? z>ajspj4?enV|~(a6Gb-SCJ`(b#hCY65OQGfz&pZ%4e^u19IeR|W;%r}5XrFBqD?78&rC3`om{iX~%g2W5!BfJsF>(2L zH!tK-?d$Vs0L*WVR0vo<&Fak5E%50yT_A$^;SpDAy%n9#|GHASKW;zuaK-4DV}NBb zmWUA?PmrgdR@fo2-dPWP-}}*}KGkmaI*;BbZ(T5-t#1C@Y$|0FC>>2Z>`AkA{AWpP z#RP%3&igL9edNdu?-vIgg=y&l!=4qlQxSOE7Vc z2KJe)u2oJ$ve(Y-4YJvyf!c)N5GG#ZQ73%@ZVvWw$5hX*+Kx7nFOcGW)B5^CJ zQU2t9w+ukWYL2iG|9=Ws83+hgd#DOcXdt1wtJ(haWCC$kHet`xx8^tJ6dL_ zLJmED!q=|89p4;!XTHc=gSn+H^Ofn1vO9fE_0>UJT#R>rI}35><wJyR{Jl*L@~vU1(stSS`eDS!GF9P zbIDxTsZBl-HZK$aUI2H5l&$`wTZFK}&CTA=_13MWdH~Uf2>rYEe#76$1_U(p7;9~v zy=|?RMb{uDZqE88}+VtAbM?tRDOriIk41S%Py$N#<7kknLnMz1- zO276uG{my9KS;7sE~riP(~eA9qk+B=mpTYCHFo~#MuaUHC$H$9EKsU73X86}36Zo7 z4Y`|UE}_r7a~l(}H~JZo0b}E(|MET#xK0ZM{=kx4k;(-5NPfW#gNFBRw9K2h04-Pp z>=P|*#Y~D97qJFB4aJn`DQv{tA8c&L`S=1zkx#8^zk^`Yq;gu&{s|L^r9jXLV83H_#C9dgz5!GPjzv%XU=QWuv(IHndw?fR$>@R2yV8Xoq%9!B^NxJ8h|`sqJ~$o+l8jH6?Wd13ZtGMvnc zH_3B{2W}PIr`V6jr~+(B<#W*OtaqCFx?b^Xv-w=K@K}ax=JI7nR-%%V!`_5h1AL=e z?bq|uCZq53TYfDwvN83=dY56noxrfZpgk6RFWYwuvc`0G#FQrrpF*wF+=l7K8&Hmx zXH@%^_rIL>+g1v?LN_&7& z+ye}+lOi);_A?|kY!pd0V9JhF@@up?i(2zBn$jmW=-liNEN-*$l5KmIGK5ey zi{v8e4mr}#0h&ld@$dLy5Qob1mVcKrP>ec9nUIpRZq%=(R4w1z%2iNJe_4w zo8K3$f5oA=yEU}96$$PVC|caL5FCnYk>W0aqQxm%w74{QaM$2a3dMr!&42FPJCm7w z%uL?*>~qfAYp;h{iA{#Pa2(aBe9}1|+bZk!$WD+!yz(xt{Pq8?6s{AKcXu4&Ga?!q z2dv+;6QeAxHZ(p6d9<)*ky(sridt4g6j0Ju?~hXm1u{9$s4wvIpbt3@0*$Pc_grUf zL?mis=rNA&wi?l$Es+F@e2v(ti>)fIRhM4h$$MV@uet2ACbf7x-l^3U23-3E zTMQSp2NI?}jF%YOq4S-4dF6J2x7(@a=9je1EOkEl2l`P+T|+|?>3H9|i!?W-p{H%8 zo$L-)<3`KnRQcpQ!owrx5d_=ll6u#1<9mv<(5wFcGxG)0_wv;lOSbW1^e)D+m@LTcgm4wmn=P866#<%QgctZXLE=k;VoRkvM zoXm#AaH!>9|59A}^(3Fsx7_)tn0B{OC9C>;{q!br6_vs4dd(9uK4D=3fO%xI7wbXn zlK9~g&b}{}515uP_>m8+^@~?1=gTp|wK86m#OWHtD&ci#u*WScYcL02MMUpFx;jIU z1uwY2aVmzYCPV%Rw^X8(+!GV_uXe}&S;587*tQcG?<59mXfZj(<8kpvUOMR~S$f*$ zoC09WMZ5pf_Qws{khaNZU+9o9$$^-XYno+cM*l!=7u~Xx;x*JU_|a)()ecc7v1k<^ z7})$SGyhkjYJOhHks~j%5$&COiozgnm=Y2 zWGXl=uGa=gFA84C8NIf$j++g?wjNvGGQk_|;7<|kd7puox#)c(oxKK=)r7g+v2_CN zJtv6qrfny~nFI;BCQ{o8{|%B()E~&;6$H5@W>_vC8YZ9iURX(P!50yf*i8J?Q8)P?N>9C44gEAUu;so zx2wb^pT;s0y~?GD{;)f;)^_b*Fwo#0B%{OV?wLLiIVgJ|NX!buS)}qf9AO>+qAd26 zg-&3T7ZcXWRW3Aeyjco3 zYSR(sswZ!qNX8`%2U_JV23~K+YWjH)O?%>$%bK}JqvNIZ^Qusnfk$kzEUMqtiAAJ- zh$MnQ${J}ktS?r6u6V7j7ZjFk$HAzrA5fYb{wi-L8+<(;4eJ~|$10fuzA|lu1}Gf> z@|rifs6*!@7<4@GWLTqZl4>hPW8bO~p~QpP;=&996hm#Z zTxXRc>6G=XzmgSyL;jx^;9m}D7jxG)$;yO6e)3Y(PW`M2-b5Dq$vp==Ge$4zT!3d; zsROPIC5Zxdc%em(zGs;XHfk-)6kp+gr4RCKvP@plH~D4bL^?KxkpsNS0zTImIt%qH zPIrgvd{Eb|(_pqEahPcg$*WoS=t;+tQ~T9Ok41|doHP4Zq2##yBfNPQ64e}X?KKao zAC>i*S!Pl4l=VkU;9a)RLA)yLF%EMTS#!R)Y^6CUC@R)ww>kpz{Hu8Vg2ZX zBd<=ZvL6QOz%;KM$wgYnn;F(kr;N|8Wp}&4oo6IEq04XZcpiJ@8MBG z3_L^X&ka&Ld&?q8lZ!~Oh?}WFOKBU|^1WRd&%|f5TH^-5v-;;7XS}pz=(p`VuRhM@ z^RzVMofo}d_e|~o7qHlsnW{;Q9tN9v3=i(EY_fm#XkmrF@M-Pc?;+XIM#BCj8Jk5))bpL{84PfgEVd;W(3uFg*){-Oe zlB|1k^dz3w3%Zlv;Tc>3s4|r0;^qPLJdR~xBv&jo>Ve9Kg6ieePF1qUVQJ1yfaXGm zyxaQk`Z@!mK&ofEu`#UYc|Fw&epb$K90c86*X1hcHb_HDRC-|Q@#PJ&-W2a(2d;XL zZLq68z!Cl*y;*Ic4Sx(M_VY+jGs_>GA7yDBWcTbpMBO9~2$a@3I2;hdHTBD0+Grvo zdwF9}6U+Ik=#I0q&oDBlzC~K%nq6XCUDfXR8++vaBDCNBuoV<_NK3!?*~=Bz)G=V; zFp~#fbI5C|R6-WcYLTnhX_g_L#eTvKue0b6xLqx$$*;wgYU`NPIQdTxcmR0?UP^yRk62=!L-?0V`0Vs-ueI_zhWi6F4UojNK`?fm~we0hsqy^l7RobatGPW|aF1mMC+Kxzv9dG-6Bt&yG%Os+?+d;U_$!Nex=dkA_og zGgQm!=WqD43eEj{1;?r@xqsaINO^C{ogeXUaFMmHA}4V|&$l-lV~Z;sU?D`yOE-8N zn_8td!W$R!CPpvy2p`F&Ocnmn@F&6W$qU_xxn$4_FRt2-pgeTP|7LzaRxkCAm$|F5 zBEG!BXuCqVtg8`gEBg1duUsg-tO3oerKkoaTDol%X(A_%eRcy6>lv+!tJ-%`VAO9Sj96avXy|u4N8=;}{i%5KY(I6XG^0@;D=rXDbFq9x7Wz|O<$7$l zt}sCKU>Opb3Co3HCtx)Qs>!8t5udooEqSeqwuCF9~OxJwNi2!jjI zyW}#xrO^pi=euc3LWubt8d01RlgFxwq^FpAb2`Qy(%Nci{)%Lc{$rw4fmU2{pqFF? zB6l48@wU$Um?dA}@ACMODYj)?Ba_eIE2W)yf}}|CHckUwd@?Y-I%*g!wMxaUU20$I z=>%nmv{(kko*i|vlTtbg(SY6mvGc$8@jL03yDU)YF)>`9EHI$Nw+Pk)U4AkK$$v!X{HU!VY;i z)v?)0Hv$30QXl$OGO7EsqoSx*KMft!EsBe0e=pq}Q5ZoWk@Z_y309tjmbEm$dF3|= z;T|C&!#0HH{`%G?ed4c|XdKL(5y8i&H9Do=g_Gd)ch_*lW#ACw+1fXS2k)51%spiLO%zEollsJ22y_ZHs!?%(2+L$h2^@;fKg3z zd+DS!3v!@fmJh!&&n~9sdRVw^5F9p0@!!}nPnQOPbZ@{*T37*T>F3gjfxF0AP;ZyQ zBi?2I>Fkm2K%B0eHOmb?Cf0N@8aajW^FBy*S_LiF-`d89<>_$nsc7qa26%ti0Zc#4Znmi;)3M9)0l%qd1ooZ}nd-yr5@ag>;}KW@bvlCN#@pR=$~5Lf^iwLJvrP~N z21UU67xQjTHZ}8IART}aEj0B529cgYM1-2zlu)FyHxQd?Y9O|AB1PLnoWZ00hRn)V zQpI3TZ4xa(krE}AyTpk)(RK)P$a;pRGkg_c;^G+m5D236EyJeOKv%>3u`oq;Y$WF* zAd8l+$oy5>;j6vzPdZ|P-&?w}4C zX1#q+sRoE0@g`I;3~|_hd*&}(4oEp4q%0J7Jd2|nZRZ`O4JbE}`l6AC5~rTk2y<+p zW4Zpx&X)Eb!vwiXi)NL?PEh71!JU-677+fX?)5*|vV$+Stn3cu`)jM~)d}BGJruuE zYCKhMwsjwFX9amhUFMtqTVfcb9Vdt;O%G`Mx^E&f$63L`Y@UyFL6mX(hF=KFZZ9;lq*V*#Mw`7+@s&A@=Wd((Q;Sn zl`95Hj&yZG$h!nLr5!eylpn5pe34reX&@%LF*6q_fJ50&R4ibTCk$Ma%1Jo%{fT9W zt6dgONY_8LAIPiwc*HQiR3IOp`bES1^!BoPx8X~Y`djgyrTQNdjh2e@lo#zAUx>7e z(gCSq4fR;GDs93`{G|d$m;!^E!gwHBz7N*k3Q%@eAZ@vDMS=0t=jk#hL;Xa+@wtYS zyt$;tudBo=)A5m;wNn?bD&wG1}M;rhjFbN3rglz;93P-j-65 z^r?$<&(-@rJc|*b`&O&<>~!xCuZ&cy9f^8{cv3t+d>rW$uojQzGyPq)M-3{#IL7I^IX#h;IaC zANq%U@hqtVM9v$-GO6dh@n7)glPD09lT_Hu!Xw18Akx1$V+;>4vZhi(VL^wPs!{uw zfo})r4yBYP5u`y>lS!u&K;sOPYSq}5Vm{r255y93ZZXye+;f8vcW1QVr0=?A z^U?Jq^UCTBPE{HNcFmt|Z)te8Tsfw~80eCgX!;)yCV8-J<2(g_;C=%jH$tMu5AWmX zlxBD7cE8;Gtm4AIgCDN*G*iAi#(VwztROSpDW*lpDnrd)?Y!aEhUq@cSZkPpmj_-p z1{`Z(Vzro%MGD}9Tz+z{HrO)K8kQKkAU9kq!mA#0JxF-Upl7V@v|G0jGVzmkOuMqZ z?1I=%Y@78h4%uP{wd1|OwnrKXyJqNj=R#-Ww|u!D-jMG^!(f6B;C{@w^aA+PHTkIN z{JFobd)}|P!EFig?qNR<=eteci~UFShU8nD-vhOMs;cvSUg)#Ph}@y$3p(3bEr@aL zYFJC+w6;YalrkKWsPH$Y^i%d*T#~+Lpxj98%JuQeq2}jiw$>G0gPr?y56 zfB=$BTxiM+&uTF05cz7;KUWsk3Rs`VC$CeAdf?AH6f|b0)ZB|K4{5lC{v?{JNeC{~ z`RT(*P94t3)GqhmEl!4u5P;gDW~YIkd?+A}3@ZO#T|fmGd=;3$Vj)h=>y2$Y|3>fR zM59n?H0mk-Qnoh(K z?qECp6Ei707FM(Dk7r4e|8@)tOBrLWOB}IFEg@_(0LIirhP?P;7o1M|W4pC) z;Iw#`3u|^Fv_r0%&99lv;777b8mp4E*urY;L?}hW#^Z0IC=@pF1p(Yfv^s5XF}Z2~ zmYPrExJxwUXh!^W{qYNeFjH6~La?6SMhm$ce8Y|OB~Y+6Jiyap1%B%nP&_}j3MYA0 z?7mX&rsi5=32=V>gFN0U0q^0|(%L^>?Ym9Vm!Ej1)VB?+d54co%nt+R(XEZ{7f5 zj5I72HlBD?)9*j9SG(`=QETDp=`K?v6IZq`m|e4dC_b=jh&LhZSKOTVe?3jt;sTwGckKAxAmbR8AN_PSbH?PT;K9|3EbUD+wyriONa!WnU&TC zH7xN*_pHz1r`zO4n(RN3PBA|5XVfj_;NSLg0(+Kt>HAAS@@6!mJbVicYfS<*mfr=# zrQg$|)2dtH5Aw-=Nn=XF=MQ$LKKoOlVZ1PSac78(n%^l-V>40_vQ>4a#RT=fbv+dS zlfy53DiPK_*5V-Mxk~gmnI|*fk`Q?|81w|V%Tq?~im|Tdn?6M;wyysz(9>N-RHglS z5~}X06_}~!(3oF2eL1T^)?eZOcj*0f!mXWeo>6Tkm|KH`U%nxw*vW;bzZ#~2&nzkK z-!pJVkU<(I6~N0fLlf+ycjg-MYxi#Db30{2l|ycLr_1t*hj@M(c(jkxI_r9*PnrY2 z?9D_bO~deog{$;;Rfyk+C=c?IQROu-EuW&#m3b7_{trkEI^iJ>KDW5nh?PHsBdiQB zoZ7bI^f_7*O1t`GF>n;?tx~b*>2X_oULI5l(x-6E2sS#nyeV4c0D4u_0G87e8c`)-`DHn&fLH&o^0c zVvo{6_nUY3aXl*&@E}AVQ?xYy+jK~k9H*u79%ShH8j_P+$@gc17%f&_h89~9Jhm0s z|CT3O0oteeF{j%ZzC>0h4C!=o(bK$AxKV`)-IJ|m<%lw`UfR8{i852bCv@2DpjozIy@vs z?yzohVd?pS@V7Wgpl8;_^mH;n=R@O{@RWVm{1e}{T3>fFt)`#c#zq`jfb8(lZBKq1 zp?FP|^FZW_xBRmb@Z)P^=UdG%nd)KLn0`JGs?n@P4fMa7lzy2n#dgz!N3?9;D`+B9 zb}h)y&+L<6&I1#dxAaj5v-%=n+M@7w4SRjOLZPM>=m_Btt0rxr9g#M%BDNOktFS>z zz$LcsV#tj;1!{{eG?0=niOlfrl3=75;+AGVTkpvWIYwaZy@#$a?FLLzF2Zi+h^hIc z+dQXd;S>d4ku+pOp9z>zuBR+lKIL7QUXtWn^6NY2wdDV>JO_z$hK_K`aP$RM82)P! z&*IqT-FMJW%S2zj7raUQgoWFvLx>i3$Gd&Fm)P*HqN^Vhj~Mhl;3Sy(5KgKJC>H3w z`_jIw?M+v>5KpW&3ye5AEBk&P)^L0pXXIb~qrSmb9|LRp^K2t4qh=~=%5K8uvHuXv zc7jo@B2ckHGymHuSlY+iftFTVm zIA3PBE0hQIPnITBo&nnEi%hg|73)sq4gBOe_$zunL?qQC%o5bOeb>IHKUqszN@Pgt zA+Gk*w9nPt6}UkydUOUSVIKgbNaS5k$nQBAJm75d5&EoL*V(@QPWTt8* z915$V!z|v+vs$5mr6Np+JhRqFd#)$x>Fm)^4g%;SNl`#jA|w?&NS@ik^`4hGhFJQ- zQYDWWS~7emwB03Uj^Q9TIP=3nMf`y0lMx_c&Id99{~PtAvAd_|vG?*MQ(I-5cEc#& zugZA+@6FM|X4ia;%N2Y^?fd$5j(RDd{6E40XlqiG#7ePV zW(ZH+MmWUa^pBJ4(%8olan)rS`v%~?(|SXdQRK+D*HK}$@muS@KU`)_H~FSWDG$=< zHPV2b=))DsD#@*LEbfDM2>`P|3#@d3{i%`Wo7(I|40s)kq6Ag#zT+ z6&wScMjtU)roA@Id`VbAj$C!Msz%m7;!jmWA{di-RZu@cvj$idvD`lz={9m3QSban zCO#ZP@ym(-FWM-zh3#gjSa=G5Y<=~tyS$CdiTCX7dqc!*;e%tG1;6!k^KOIlJ6&jH81iU| zIUB9XL2b=e-H*%%X74MjWp#i;i}|1cWQy-0XPo<3Ex$>Ipyl`I-^t+$!@fY9=5a&h zW0X;@ATuN-jzpGX#6`p2yBp?{(^Bzn7q_%@hV^gEh6ta@1?0$n0|2Q+>KX?wS{`K| ziQ9=rA9ikfx{f_{#kULouYp(V%=)@@!ZEfT)ebvwJea)jr!vEqZiGJ5YGt8|@!$4B z!LF34>_kmb9l~v$uwLX95&sP*H3#tr<4x0WumfSDf!}_3{0OkP(s%|d_tYLKdw#{8 ziEaQJ-~t|szZf!)5La|0o|LWePNyocQ?b12_XK?9$)PFq6yuzm~>1FKDsC@Uw3}YeL(~!sEk6uBG$P2-dZdwPXlh zPn$hbgIR_^261wys#TP}Gx8>WKdhip{Vp>@qdnnQeFbRuQJ7_AoI!DdF(P^4<$PUW zhZV?Ra}(=ZaPPF!a90=8tjzFU%hp!#*wDK-&viWkqZ<;2A3MiHa+Cbr#0$ZA3u3YD zmT3WR-;A6CtTM#dSBcPKJL|EQ;otGP+g&{u>ba@o|0Qel(hOj6Li-zQ@Cv$DG@G6p z%O;fr`p1KWanL|h6b%V!*-+46nLX7&>XX)3AmxQrV8hVCfrv&Oj!;Hv2DCu zh>nlrc?16)`KR}81^)J>?6G`48ATGEV<$3z2O`0*``mzPh9^d5r#R^|5&?GsLM0Ex zlextyyIWfi&K8qVOvSA>8N#&V!AtF)%~e!G7she$XJgMi;l~rN3A;LPp!4u{nVk*? zCCe-13jreM{~(+A3ENBm;;#d2%jxoM#(Gk7R7BAR5NP;exaf=NGEy$sf-~S|x3u!!U-(01$M_%%BkE+%3sJ=Kfax#tTX0B5MV{yUf|ECF*#PGQ}9StD~Q+kF;t zg!6Kxd;ET!VX7tVRHuBK4JPmQfju#(2KCKkZ*w2rgq$IRt{rqhN)=0Q!jk$kSzJ9oN%h=z45C)eC}p*DQ6z;7(4;}~7KJ*`i( z0dE`{w;N$}&GoeEA0r5{n1I=z9v{Qex0fT^kA5m&jJkDd2D&A?i)wr$U`~%HTW_c| z8ZEQ?>kP|Q4k$A;=UfumplQv??!2l-M3!d;eQcfKdO1WmmS_jbXpj+f~SNB_I}ML1C!UgS6(|6ppnx7{XX%iF|cB%nSHvz(S(MF=jYI2 zRtg(u?4Gyn^_iXFLBIc_ec6j6>+xS_)aiZagK>J6EOwBB)d1LaORXPO&976jfDV;^ zjEs4^p+kf+l4}0SZ*b1SCoVZMaQV6kV=8qF%N0^M|er2M=o=~G676KA;NNAWwQB7!lcc?EHRja;_C3< z4B`Gx#M52$Ud#9QTuYEL$9k|VF~+XKc&*D414dwAEPVRXXM@f%gcjs=Hp4HBAJu~-TMjIu$+WYwdw1(hz83H)x-X9vX z2K{rrYznt2-Z3HEr$d$?OJkqbiQ>$w6x ztBh=Zk|_L!>+czTNb6^6ooAhE&%ATKFe`uun~y9*CbZN2wpX{~Q;JJ2_8Wy)^pxEU z5pCWb&kL@0?w`PuM}d`P#SM5`72gP~M9HT86s&vwm2_!-=J!%nRw`B|7YW4^_W{(| z;*0H#JEIJ+--bs3aehwJb2#L8u6@tgw~^coFLqCg(+;cr2hZ|O<3#rRNVtXRG@KGI zxmN!r@=iVNv(v?-(k94$3TK++_7OKHck&UDp={(5BNa@OiT4UhK@xoxe8rBin~OM& zev>+&+QCqE$c#9)rBfmj1@M)3D$F~Y={i=J@^5Ns52FbP zFgX#BqQl$OA;*>s8}ZzWw1D?0HNM{U#f5;_xts?GA@8Vck+QqD>34arkGJ#XE|{R< zX@Rm6c1a1?GAweVtdG$q8-%^vgp4>Io-?XP#V*qnT+9 zGKB~#P$EbdXZf4avlHI*c)yIUMd=2_0cNAE8FvN*SAXs_A)oVtnTuy%ATU2hIBz`q znRzwoU5bbJ&+36>>;jzmOy?DPm?^F~YLbQX8J0L}H0E%ROD zNTgZD2vcp9$4u?OH@JRhi+S75hd?Cnp@rU@G%#4}v#xLOg@rmK^4NR)SFjMi z00Vu}-J=r}iC!u8L`<9`yR47*|HEe6>WapHWz-BhOn0Ul;CeV<@!97K@$k8*DP{h3 zUv@pj_p%OU3pGImnj&E3g0nd~*?M7?)Vw$>$Oa@Fus5Xi23bX2iGP~^^~~24fGi%x zE)UnQkmbgc4k2FI&YlnW_1VJff7};UE&QN?nD^Iyg{_f9B|`b)XU=H;e4P-bFK})f z<5Cy;Kn!yBe{Xjtkrl~LiK`JWzdJ>|8ig0CL48*1;4ZDEH_(zGfk@s<_nrO2Oyb0*1+ zC!LHL90)G@jo-CzdzhZP6}AbVe$0A@NH-)WMo+!1*&XIFccP47c||ulQ9F_vEzL1# z*t-QNb@9n++4iJ!gPeZbY)m*b-!572O!A3-FNoqiXQDml)-6AgJl;7cKCm#_w`1l zXz-f>%Y&mZ*i~W!>o+cxi}oFT9wC7;7Pmxmsg5#xTr+#gHYun`R6V(!jEc;G;^R}A zYRf<(g~Eg(AycV*W&Zk#P~;-}yZNw=0PE+yq-)KYbB5rciH4uiz4gl(P0la}V@;|! zd=zH7H(p=P!`k^IeDuOfi{A@QQ0a)us46&0(61M>{UKD*6y|~lG$?=8ZN^1OSopbe zGMpOU(r8jk^R}LGCKfOUFsE7Pc7QtQU+jb@L&0OSu(0vW0T6O&y{ADtPOxesUPF4O zx;gJUDDYl}cqVVRM}Im4Zt&ll)Ctegp=pN^N(||MWdF4q>B|#GXmY-u47Sa(dGfi2 zojNaie|SF-zd{B798sO(*{Sk5`LCI7xC- zaKL9d7I00WJAR3*il#K7eZ&o=+OA!lo3-hrS+dbi=;aS7B$jwS-i1^b;2Krvyzxa z9oyC}_o9$4JS(J_T+ep3H64ARZ~S2vsk{0qaxf!ZU%!1Le3aRSWxtZ(aB(VFrvOV* zfha320|zt>fPEX{i}Um3cbP02wGJQb&L^~W$h>A5BR22-6H?vyVYjnU0l2PZoSio& zrw-RjOmR8Cm#duR2o)X&l;pVUeB!MppPop$&>M9w9M<`0F5}mdG|E}G{Mssq)ztn0 zj4iR7rB#PfWKzM3;NG`&OMB=)aF!ObM=_exyRN0gN#`p)S!zvm8xQf9E>ZVM&~d!J=S0yzNt9m!ik1 zh)v-)aMo^<9>0+l;By^`>}2@a7&^a(tVS%e1`kDeb#L?scCml!&yR|WE|ALVy!m89 z4Mui7sz-B&4vgvP5;oqj;c86$dV0Ibd>5Xjo_F;)G^roxjVxKc=T>3vv6;rOXMoJq z8>`Vlw~WPIXE|qD1_lB@H{jbC0Yig^N@_Fnj7~G%ta@fc8`ceeeA+xgfPb%5^0X&8 z+r7=dS_i`uhe4vpTh`4w$(47n3t3l$hgQAupBAVMK_GIrGrF=yG^4AS%QY)ex$mS3 zv+J9E_js$nqK?!S3p1o0L95=?K4R+pLj@9)A&w2V1V09WNS*pw*5~+%5cGhBR2`X7sjtf7b7FlN|-^h zZ|XX71}rA)S7?Sd3`n2q=NrW({@#@EMG%Suhj|y=TLe1|$j31Fmd9N>zPmzQ;`cm^ zKYQnW7DyKtRlXg`yP?s28PZ-AkrZ1!4_rr@LN2?=Q)IE|kCX1M;vDnMbdl?nviwy~ zQ1hNX&ue-fF-61vuJB#D+5)ejwy8ATkp|XbMI&@#&~cbOtkKf@n4df3TWONS?CT~ z+xvikwPl*~yDGt>BAaaA;Ae|r-h3^Z|Ct~gorlsI1qZQZuWYSmJb*a2%iys?z`oth z&?#lRxl{{4d(9Y^ia<`!Pj#94gY_Bwy#|>}<#MwdkU->H&nj_UvyHyc+_?V1=L}(| zi&PDKIeNNo*gKx=#CH@dFqGiH)0}>QiRMC@l5~wjUwl#$J>} z@7W&C-IxI6u)E6b<90xuLp6akwk95X6niY9k{3uO>fK^x3 zk5QZ*M$pFwf#O3(SBcTrC_&Z^nAE3qgPxww372AFHftLL&5;WvQQ3f4`}lHJ$0E&e zLSIs~jK)zhZehU%)%e3KCmP8fR)c*mrP;cy20v$CkCr`8Bwp+6G`P`5SLC7$IR$~j z%tfgXHlT@Q)gqlgg9Y%Mv*l&g;3uc?%K1Rgm$Pv%Zuq^Fv>lg%?t3*s6^GHvQm_0@ zZzmXh_endS#*N*aPoTgwqtoEq^IF9hC&I1VfHZt^;TrAGRNS5aj?kP}kuHXMoQDDg zmnCKRzn^8Vn6R1^l|hSM_OEQ$Q(RS>9NXTj-Hm_-qm)pQFjR>xce|8}zCk{kq4?Qk zP1RicwnnLAVL2CGOt+7@lx&dI*iJ93Q_6lMYxQ?u^$Q%U^jjjJ(ek2(tyfxA-3~Rs zzQ?r_~>@TdKbA$qE|MYFFcQ|reV1UV)0=eg6Xv=V>b z|M8OC5dSOlo|q+JVvK274aNRHyt(`UC;IM=oNx5vTGEA;nqM6R$}=CkKH1pVM07;m zH9z!FEc%_VP-Ajeby{VBXBtWwypOP%&q*yU2Xw9>3fz;QHH?`gIJ8jnmw% z#d1lhF3G?`;H9&#Yu%)%>+NpEW}F>|%gN>Wx_%w_uO8x$Kwnzqsxkbgn?WkVT*~PD z4m?r@S(BpOWL?It^Xn7kqPl9|5y3ZM-q><7y>Ob}tVn-&2)cgDXB?g!SF#ybOoqzR zDuHP2+9AQ@(=8r3a@q0uRx(#JFeP&I(7oBkHE>mvd~@9qcvg8o$jFFxTU#W^GsuY- zAlch%;$;v3T=#lEM5JBqX?X zce-=ex29g%UyC9}#}Q^Viwu%F_gp@+$wy3&PrqI~krH-sz5yBuigyi16RcCz64gPC zd;?oPh5T^~5NrR7*R~zYFg(A0N2YIbE_U7=w?puXslAim*^EcUFLC!N{8ebCM>p7j z44oooKnsUu(<)LuLRBDY#);v=+zLY$!>fF_;zINn-am`Vv2 zR$_m53m%2G?eCv_F}uBx)fwSlbe}``1?FEIf)YrPGfrKHx=aJcGp}hS+XeTZP{;jd znhO^b`aZLylvhwgDazTB$XN;`9J8)bw|{cTOjBP}zUAPs(nriXhn_;0+c)(IAmXj8 z6$#R@wUpy46B!Ww$Q2{@kM+XtF(?H3Y~+54sUnN(@YllYiNv_FQGTD6bh6dJ!BfLF zTFoC&%{(N7)%2P(!)o_W`l5TimZ)xRbVo+`?Ocf^%pWtGKc&puZ#4>fEb9oysZEqH zxC?14Q2~!NdJJz>v2XB|#DOVIMG825%`o>ln#SmN6zF!K5q4LwCxtWNJ-)F-K{%K~ zqTvG#eZfI`!r+PvsUK;8(^XJ>vpuT&uA7gmd@&vvuG!2r=;`9)7%48Q0R$QCgOERj zxax{0Mseo)SGwulK4t(w_p#&dY2{! z-Yp(ieBT#~+wfYF|5i~8z$R`jqf`UI2NxtLYyfx0n4ZQG-oeyeyrb$*aB{R!w-X$P z^|n-F>kc;!&JFVM9<>_0M33bhUtH!SFK{o`OsA!8`mQ0=e(9Yjnfo6iKOASyh2EbO zgxd8as@xFAzd&XjasPZqH?9={Gcm7M1PeDnJ#?xX$ifqdn2%OC-L@M&B9V)CxhM%&vdj5QC2ufW% z#wY#7t?IST(AVkK&hF=?^^=qHh9I&+HS}``apM?!KKJ{;$)M*TB$^|D(xY=Kmjy4) zY+!=d*9;3ta~s$?yK^d)+x1gxxhWD+VwEyanlU(26CRZUyPkT1C)}HT$I`i3f4(Yv zxn|t;OhNb>65exHeefY+@+6YhqV%%$tSNL6G(gwEn%#`Ss3GkM@>XYLf4(ppez`t% z>DtAAsrM6|^b@EXrWpyq4A*zb3Z`L06aMuO=-#wH-MNjVCpuGW{fd0KxI;J3UEEFB zbqly1#V&83rsd+7l&q}CW(f)Md9X$I)uA@G#qQWX!xO`6tLXu1SiUQA1nYUq`p zTf_cV=O^#fi|3?Wo4e0Um)#@$kKB!+lb80;=}Ifz-D)9-o-r*8w#iZ)8N)w<t&x+q{YM9Fs-BgLm=y$66Ul(}Y!DhOeO zveD-QsBLg+fQWC_xY{(=+t@;DQedpXSMHa8#1DYeW0BJcwHPQ^ewlKtXn(|SwDn;6 zb&#H)nvvmsq3R|ZH@V=uxOf4cWjY|Ox?I^)=6gZ{I|<2v1|&pb@LwL~q?T{u_Jm84 zSVz}^b-A|ZOHT9p_{PiKZEl=SBlqQAw3M&P6ap>H!2PqB?0@$&dVknQT2E1VVyQ#R z-R5HBnzqlu3bg)ca9a^w-Nq$KNxwDzv#oH`B=a@edPFQxnZ6*UIKrJ9;Em42UnuU-H7_| z$<)q-`stn8Pr@P^Fr3bUP~2EAf_p8i{WcN#Klv>PjK8b-wI{*JpB)SPFP+r4(ED?4?dUA1Utb`;RyQHc!Q zo0vN|ShQ{^vR!WV&4*zM-}A(ng&y$6291jD_JL)W$#ZK|A%?Jp3MEoBfp!?pf|{DK zu#*tY?_Ovo%H&61vrNI|{QP|o)&!LYFem}}q3HOxCZBaPFFePivSKU-r>>@mvIW${T)1TSn$ZiK* zb4hgwS%p_~U~j zRz+54jXpMqRY#`jj5< z#wppc%^EN|4SR~dtunfxZoydh61k*(EM#UwvZhL>ssDZ9i*2u=TFM3;Z(Rk%zsV}c z1HpNHzhdb(M;6uJ)*?pL&i-jBr!LOL#TLg2ilsfj&WQ~@O5G>(a;zs;*QY{wO+0ka zhqzn!GkV^gP3=LUeS3;21G+f+F$^qq^Pd4donn?%1?ThNL2*|@R|o2KI#BK3+}g!! z6PIx-2lif7pZc0!H=zi0&c331YOx11A*9_%sdE92)m3n;WaC)Ih|0V&10gZmGtBSJg;}*3{-wZXgO%?)kpU*~xgO+c@HsO#gA& zgXPT3eSj=P+v%zU{g9bUq}3$3J2Cjr6aigzg1vyasA{xbY{*StR?ceAhWO8ojG_gd z8kWnD*8l#o$0VonXb88~PPQ_1X{|j`Gi|PT_mE6a@VlH|!VP5I2Uht`{2u=We1sSx zC1z$W8fs_vbjAzPRpCWdT2p9r5U1qv)?&@&&b{Ze&))m&BjYBK0o8vQ8ztiE^V$j?A_8Axjj+92kjhhq1}37MC^2fkqfFZZ_Q%PvN~_XEBQ^<01el zUX(@nl^iHBE!dO%e3{n{yIBS;RLOPC*ACoDl(K^eiTtx!$t!w@KVkq2BVvBm^q7p4 zQPy&BQaAuToAgMd@&ZBiM(uKHb|olvqqI)AI=NK~<(45ImP*AVV5l!Vxlh3%G>!M> zIV9;dfmxU;u$HF}`13}ABBhuo(T4n>KOd?XM;rmjtIOWna{T&0{B2yHN%t1x=7*#t z3P9CeOq_qRaoyUqAZ|Vf&Hf#VI$Q7lGLgq4Zt6W=*M%1U5{EH39Nmc8{!Wm!%zyBs zaT%DW(_H^SR#i8P7-0fK9HNKqW?$zVHak)))AzP&JF`|4ZTUiK$4$)a1|2^xkm2ah zpvgi_w^Zfkgm#)vIwTwWU}T1>sT&G_|70#zX{_JTKPjOL14G$I^LjsBmc}JE5a^C| zORS)0S6`zyls{E>xtfpO9tkPoL`}ZfF&OE(?LY8NA{!MYA6eCLh*5lfa(CpaeQt(r z!+luE0%eI=J-gK%URzt&K(5c5B)6TI9krj(2%lib@-=k29>?FHuqp5N%DTLq*hrtp zoO&*#Ec$$56Mq;)U3rfR^7wfox_jMiNIjfs8hB|MlkR&_l(c+2-ZOStpy$XzJ1CjJ zCwOWy`v4P<ZOP1^IiEL!Z$D%_{;9e8v|;t|S3&I}(r>;XRpIifO&{|{x0g8uWW-flklymGZPiYD zOZM*-YL!_Nw5H-G5V}v$2RBU60SgIruCg#!=lA^x?a?+ zwE={ddGO4b^iZ97T(9{T&Gql5mDIPUbvs&ygRqojg^eRGJM|5=E2Sl4gzRB@d9|PG z#^cm-QgEJ-i4E_7%LXRvvx-VNF+zs*d#lHq3YOm72|-se(+t`UJC8r)t^;CMxs;^s zKqK_HLFqxy%RO@*B8iR6pnn;0bm4aJEHj9!37|xt> zx8K#3bJ;N6|MOrU0vS4V->vZacw4rA*K3N@bT~^5FJxIZrq6km%|<@bqw^t((fxy+ zhF&B02$qe>9vbsu(q6=^wrONb==~!>^`D^q`XR@8zcIs&Ufu1F%;iYCZurdE;eZBx zt`S+)=KRU0D!D;M_fFW&Gv1US*)>Hlc%^oj$M360$M{lkTo-;0G0Y~Qt9v$ihg_-3 z&oiBW8SnP)&L+q|@KSl}>d-y21qPzi>aCu2G(cZ6+f85pRQu!FwDK&sx9M0qPHVDi zC)_!2&8n*xO*?_Ui7_>e4OoYMJgoT)j&j&Q`WiG0no7%<)Qu86`4IbU;qD()ZLaZG zD6{8N-ug(j&U^;tY)V$Pm~0l&S!KaX+Eh|A7IMEOou=lIa0Ai;X>Z|b%ZXW|fd zT7KbpA$%m&z^pHCYrI`r3wJ$jT=<22>=Zl8OzyYuTYKJld9+u+JLOg3=b!Xi8}D!X z7j135<>sF2z$><}?V-q^#5yT_vBlq=7gG9qZUH`nkLQP}F`9iJW`CrasciiqK?X&j z#K1WyZX=GKa}MFw`0zt77gg9nJnvyO(1j$|mmgp!=j2aZvGh}E6k3^g`w>pRrGmq0 zp5RtYPjho7w=oY1^O)PTkrUtW%F_%h6DyJsX4b1ZWM=QGd%AdjAFPO8Y>Ji7m^`@P zLf3DNzlEJi!i}{2>1Cv7-IC1CCbVC~8xl1A5nH9%_It8MW9Zk%Jjc$8lWIrxpPnA* zL;01Uw|uS4XIUVHlPfji=g)S^@f}cH-9{D% zN-NSwZR4*fi4O)aDPyrhIWH~~5wxFdK%PhE zET%4mZCW#a8c27SLZ2s{i{1w#Pfqe`n0mZe7}jV!0xR~m{Q4n!$wub$h~zN0o!YtU zR*|^YCea}0*@gcxo$ss1bGfQpx3%#$(pFavlTvXl>1uwoJnpqM0`6EZ#NOY zC!;<;M>h?#e1^lhXf$VNlpc?1fHhJ7dE-oafq20 zt**J>-i~c)Kzm{d9i3hZ|*+~4VAN2@gS{6N3U}wM-BbPPy-;5k8+t510=ty zf1QVm(pT73iEqq;r8ON6qc?Kjx7NpCEt0LpJ(N#^#cF(gbra`vKnGh&L*S|}{qzgX zcYI=jU?Z4o@Stn12>>+*h)_V64eHZT@miX5(0v_E8w8^|i!nvEkP8YQuKl_XYn zBkQ7pqSHE=^RJiS;qtz7?uO72es_+nbiM>qt>nleA^h!6K|~aOz5gmFoTw5rgjKSV zi?|obgnh156l-o`;YK8XrJJ-IR?~G{xT(KrQ7<&cwwfT#|$VN2&*pjaYf`Z zk3l&d|GrG9(v1~KFI4JgKBdjm-Jee7(VQc;#DjYK1RxQ{^QlIireZh0sq1U>}eFa)15;@q+D<&CXLmDSme{S+ z3|JsWc!F@qT8i{DNg>NDzz8R*j&T;&`;a=TNLnJ5 z*NG6`Mv1D({5`xA5w#~#Tc6C#p2(|O+=v#3yybQut9l?IH7>@4STu~{m6vbLzpZ~w z>+V4@6C*JbQAh@vI;%+-ztv{QyGZF}N{i9+mc@IuyMc174$Zq0@FS#UG9F2J+MO9!9stQM{^}PEwjF| z+Zeo5zvT2OKtP>ixPP<7+%FMG{a-Vg_v>>^f!Xi9tekL6NqUc&C({rcfB~EQ!MW1x z(Z(af+R@n7iM(ilDqQu~bi#BqsU7RBwNGT2<~Y6>Qp?!|KahjfHVIzVoLxZFx2omGR&VU9sEh%2x-zDkSD&`c4qj zW|d1`WH)l$j)8RjOdyZ5E=dKe=bZE!VoTt?If$z|<#~?5o|PU|3Ubjj2fE1hWonUV zGyl+cISa02zMVmV3oXngfeF~9AWv3tD)Gx~3f6a`$+0#G;c9~?>^tX5{lR~CvV;Bg z1gY_Ic^zqSlh~-Q!QLucP|Twot1bZLX%oFx#`x8qxE!YU`wvE$Y;yYL{W$9hPB(J} zr1JIqK%!lQaZZX+x%*9! z+9Jcp2)PgeEF-*1y(2B_D`gBxg}j&|*>6j9KUTB*K0LLgX9Zi#elx2+y&}dLPl+yY zlvM|-ROI;YOtpIA(k6{0l(-^bf)2*riinhMH`c2nX4@~f8*fV z&cVP56?w8Naid!4pc;d>;f3gtS=&_oObS>K^11d^n1%^HWM&QvndK?B_Do*md)bK5 zMy}@i!F`X{`s1oT1_##fp;NL|Vj?Z50EJRzRf_xeW-5CC9ofeD9U8}Fo;k0>{_(J{ z>|0+8;8@sVDx-;G<1u)yO^Mk%X$1tGNgs5nF#38Tto9XxX=}^pp3CnPqEiRZVFtL8 zrV8+d{7Mjo+W0HivgMMM)dU83o8BCt+wd;c}{Sv9jKAl7}>;mt+WNzrWu+mkx@Orw^- zFGpAkLz(%6+b6qCK4_vc(=yI+I;Fu;)_ zWiL=ljblmoetVykK8m%$XYoUwAM|V#3uE3S2<#`=>*@UlH|IkPkkcNd}r3kB^>DG?$@iiIVqi>DD8KtPQZ#;|p z>khCbdI%<0oP#2?4Y-wYCIX+hVAE>Z5Vj!j;rrvaGj`)cf~fO?Z6^ZzuU$Um)kkKt zNL8T9!B8sk5ALaoYUg7nC~WK4;yJ9!K>G|&TEvc@f{8^{MrR4vpNHNfF+f3<@veTh z{{Qh5A=4&hi?VO^PE1O6q22H~*oHX3Hv{>}pi1om&&b3@o9^UvV0GiP(+i_F*tIzd z1IdmQ;s~zo3aI0t8c1&TvZO-(V}hH%M^}@O7`1`Cpa9pS&H>M7HyMHF)UBrO#N|8k zEPLKo9Tvo#pQZNy&KyHvexK~w=G557Yn-Z_N!GN;kJxi7YHeDxPn$Di;PO`Ch1fxM z#RtABL3Qw$jofUC78&l^S{CA(gBe1(aA~`T1P2O$`x!hRTs*(Y2|LhQU_TI`z7mtR zB>t3sMPArdRR{M=cCU9qwEmYb_ND)R=Wk+F=!8s=U$>IWSNqI*=i8&k0;g^wdB=Pb zc~3~jaPhFIq2K53v+ky$=LAdYctP&YqptLMg)7k$xje1Xhcf{m3kOV$OITIVsnC~w z5A9#VhgMwakMWVM}L!zP4_#UeP^K0c-(Sa8Hje!_F5t^4cW+(S>&1GWUC7 z$fDhp!ek|ksjhrSi74lk#!FEi`@?9E=@i#%LqOabpM2@giaN}$75(q7eVMnSQ5LtC z-JM#IUJvIzFwAS_(fsDW&_c4hnxU8{MJiR2Mh-I~I|geH1WYjC@LJOO*u8jL97n=lrW z#JOFLNaJ#HdZy-*j&S5I=k5#}t5h(cYzzLri)E(VNK)i|Th*WR#Kxq)-n#N-YO<}8 zQnKg=+G@I3)tTfN7W)P@4pa7uXmw~KiSP8bM)*|NJiC@5xpAWU^arY6gd;nGzcFHn z(;HoU5$rnhiE|Vi9@NUwXfHg&!lqvsLXpki&(kn$CYSYE7#-9a!T{FaiRM*OI7K9p zD?#he7Z)*Eg+7`0|J{t~_6$r2#*0R-1Rzl^x4IMd6%1#64Z@Gd-CUP|2z*)s|5}X(2Y#s==X~`L z(pR$cA-92>JF4LvJ(k(!v9v*>Cz{z-=@9omXYS4R7wC(GT&^;cI0Tq2#sr-}L;-hy zYfYDT2Uu_qDcRl{eEc?meytFKLWm9LmvM?Hg(FiVI;^PzDrdq98(0G;5*?!SG+f3b zKN)31Vl+S3W)cx%zp=^SShe^bj>sF{54D{dgD-$98>=)8V+7iQv2ML z-aJQUplOhnfztQMe*D`_ZC{ZL!$E5fuD1pjcX`eZ5yn93(70PS}k;x zP_Plxnq|QAc=@i+9z`%CH+og}BCkRj-3J?nO(%+m!T4$~e-} zylNv6jZU{iTbn)>3?+5|=OhA<^yhsJ{RI&XMcp*LeG)HS$ybm}3(Wz1<~nI)mxUog zRA;+WVKHrcF-QUiTKv!VM%v@|#yA)Zw1kV)x2^8uL(Hsmq*I2BjLEga-c*P843ja1!Yr@>p!C_fPW>O)QY1QZ`}xtO&GMA7HY1{E+k_zx0f*yWC$C z)>I~}?g?2wHK|2$Dpll9bq31;JQmOvs&BS$1B4b!lKj~WRKYtEqf<%^=(C)LGZ9D6 z3c8X;OUgZHkFLI61>b{WH`oAi^xq^>=m2lmo6Yt5Z$rIRPC5XNbXo|BGESTKtA0~# z9J|xh+_V&OsfGpZOzmXaqj0Ma>%VhbZO)wfiN!zdVWz~SbciFtf@vW& z$qS;b-GwR#_NAbO^~5f1uJvdeQoR(Ws3JQyPttqu?xX-j7Ceap1Izrh6~;w5wE6U^ zC)Q7{`Q}-bTWI40Pfk0HpP@Iokwyum^xP(qu&zHj7;%+L^br$CfyO$s%I6(oT#zDR z(f1^x%@*|f9Re|5e|eNo*PFdp#jFatH7@&|u4%C)p00LOD;b~LS#rQ=os#BTzkATu zg_F@?4@ngsVJU8op1iK@-{erJkYiU{n^lFnO^Q`3{*d-7c8E!p6~TX*XT|>Y7q8kn zZr-pEpri^>T*5!yMFa3}yliBJyZnE%$|ZGXS${gek_fK_=ru<+6!pCFGE!${oX{v& z)_Y6%HQh<5@-V%?;3y6*;-CS=chI$JwDoZkw=6!=`s;KN-6(ijK(Edo>jrBho)Y?1 zEw_ZH56ItkBA4P{t{|nDS#8!Sc9o*I9Z1pvUUV8@Ht2DTK$7;6C{5L>YgFz>T!ikA z7wYn{jBuIRB~~+Yn0T5qdBz9`SCf)a6$u-g7gpJA{$2h0_Xx1)V^0_C$^^qhPx1Ui z|F_yXn6~4s(-kKGB`4wqA%{s+T0$olCecf`C-)6RzvqNK2qi2-VbtRsu z*UlL=^5EnTUBq@><46s`^^uc$+m$)IWvEG>Eff21w;7oh7oy_B* z{WLB3gn_RMkSSj&ctijW@(%1;;zWIzJ39mKp19;qB#6Sj*1ypvgn~BNm3Ho)xhSRE zUHU&zD#sU-U2cC88gsiKZK13c#(`9es+uQL6Kjorqipw#t27-0b96giMMg_XN?C7g z)d>LTcoYjW|G83eUB*&6_RJRp4)!bx8^#)q=(es!m#q1uZ$(O((TJvy5tub3orh9w zY-@5W1sN8jITi;Qyu%Ks{PiHW?;areUcVUb)wVspz8^QxBT*r4UIe_eu;9libI^vO zE_8gJ{P7e;c|X$dWZ^l^(_SULDI&TKi7{Vls?!4B)j|+AhXKTAcmbHmLUsYiS z%mBirLni5h?z3ElO|K1~pD#v`=!k>_jYjh`ne`3-lf*}JDkZ{R*$dk_xNGcS#bke? zE$e~`o-^5!BqF3qgnh`={N?|sdafX&|3hw>OI`LN`_Dp@Omu-098tNhHWBRnexE@? zzUvMp3UyGImEgT>TZ#bLKgf0(p+-RXJUf$-r9fiu^JVW@J4?FANU zk6ebg-=y+*SbLh=-r&QNLjgJ>9X?w+9RGZkV$5YcGHGB({UFwEC>y>gn>2%2%;%>W zR!1^E71ChMFMn7TN=ekcU*5U88y5rYp94+Rxp`FWLdUG}+H;n8vYohYuZ~CG!zkYN zp2T#5y3}pUw4Pl9Z(j4P{y+xeOr18> zETGM}SGB5|{uZ$(zPo&xx9B5gYC66k?$EJmJ7fK^sBF2~YfaawJ}mEu-Q`H}9p*&3 zc%D?bDm13R!p1puA4UN3EJtt`f&RIdf{bRPbQ3!cJ_W|IFiS@oV@6vA{N+}q zv^oFns`VJ5BM(=eVkK!qT3YM#r)ac2tv4iG*&Q|4)#DR@1>Bsu;~L&7^eep~Ko@C4 z@#IQb+Qn4QT2;MQqDoE`yl~7+(b%1mZpBU=$=jeCfv&P?oQkZ{Rb?^tOiO8|+3;9; z%c_)?U!~y81ujmEKN7r)&Q=E%r7DKToZdB7^^0KcCr_S6U~zqOcJv3Fj+%Kv#8YL` z*F<*_@xMGUA#=EL{WN=OK(@aX@t@q7waL%0tOBMA6O*;Rf<}4W3|5+t$S8ayMv3zo z-B5`@;~^H574Y3!r!0WUsO23+{5~_NYf*AQATv}WgFXPEm!+rg&^OO6izDVhQ${{x z^$V!$b*KEF6fO$3hnAvkQ^H=JeGE5DcAyoclb!dxUD+4j$J_n+j>XTWu2Hc=o>OgS zqh+bs zEBEUCwr_6(U{9D2Rs%%#Vlt#fmxGBM7tkX|o9JhdBgzvyetFQPvT^*znh}{rbnB8D z87DU?w(^L+Quudu@(I+i|0CwX?Z_i{@BYeHNw1lGx1(+p1qWFAck|buz)=sdQPczl zEuY5K0Px`Br5EQjMD5pSxqL26Z2eI8K_n}}q9*EozK*Y$5$DNt`LEwkoK0x+KW9@M zT(b*h^ZvYdD`x38E-&f((950-pf$;#=jeqFnS=8zw&%IL?#f9jn>$X?ygoiYWI(Sd zuFhf_9#Zd8cIbkZ!;>o{$ZAt$&#Rvw#o`sc9so|3o-xHJzS6()x}+z%UbJ9D#*WY4`qourRUR1>U=J)AesR&BMgaS%;-gUOmzuCJ1f2;Pmeh0Yb;Fc@M@7jm5nvPZcOC;FTRSdF6+m>JSXC<5W zd4K-SJomP*0j$GcR6)_u>s+*E*s&8Rc-1#`KM>OZ$wnSbq1eHE@hEoTwUO?y0qjV{ z7w{fK&X6lP@87ivNmRsbS)03ejQ>H}Y(54-;n?-AB*+(@#I?V4DDMa@fO67sZc1B>2>;yDFTXVQtBOh}yq? z^h0K3fGK^8cgLf$Drf{5lg&mc9#dKDgiy9s!Sjv>y?>f;Xv!ot=d2n5n4dTb0Pm$n z$;JyFL45OhInB%uS`A^xp$gxahbb=V?Wh8bq2iGA(bv6BWL_sT2I`D}|8L7vAK8l0 zK{0Dc@{vBbki7zH!AoaYgihZUJS z(TaM3&G$0^UC(;;teUK?bAfG+umPsH$$PugDv!vyd9}MB9-lrfx}&6N(`_8?41J=a z8@axuUYckBy1rC)%nrrs0 zcQk;y^2;#=5Trw^WqS?;l|~Zl=Yhv#+1UMD_lF*M0*YuX|EPl=F_V23|REo3)<;m3>vlmKUIt)gR`Q z*q(|pS=939#HEr=hP;f*=&N=({QfFH&4h+mga-M`=aYYYrI3(Tt>h8u-{>XlF=bFT zxfq$njIzJjRR0gvDPrIF2a8c~s;_2uK!DWxT|dng()6j}kG)R#kG;NfD<7~ws-kw@ zz0QU19Q$RSspRn~9{eMUgW{`0-&#-ojJ6nLPN>;y%%5~ytosE4B!9`t(APqC6I+$_ zJHUL=q*stxLH+BV*brNVC`>1i1(?eXvTkHiwn1IpKZMGZKcpXze50KhH54cvF5Qo!U{$kDh#$GBU3jeywK= z4A1V5u3AvdkHpi{XV*lO;7r3e=w{V1q8BHZHjz1B5DtrEk^dwAV5>{$6}_RB*i|9~ z6W_#hK$trRx=b4%Vzj%#*%|jPTZ=_nCQn0Vl)dv?Llf~v7!F1#DZXTMq3)Y~HEBuU zJ_SU*+#V78*sD9rUFCP~sc9ltv%ne;50kAkM&alJq|raV)V! z$>_Ci2Gr~2db^R}4JXWX9@A#0ThY;~oM)D}f9)}6-k_Hn2=Y%!{7IQGn_PehBz_8b zb#?f2Gvtggy>iSrD2(-**~|%OAMIUfYmNaT+;Qp=D40Z#kPfz-2QVXiH|dwOy6SL$UB@Kc%i{b7Y{TYqbDKi~B%Q=spK(`|@3qw(wNVhc z9XDxo^Pl^PYm58s^hYKyV`udzpem z+>(#m{aTbzD}zzU+b|>`A5<$ti-{R&(FR&3H0o>buQ-L&_P4mA+=Zj0x-Humwt{bC z6kfH5*f)fEb!@r@HM|nlp*aKBkRug)-Sr!)Y3FBBmj#Y)j{Yd@IgD{voTWt(M^6$N zr-5kLR3(o8%*0Q{81g44uj%GxP-@Uy7MWEF#z$eKSVU#ake4h^o|;%s)R#G0yjgJK zRN43A_WI~dlPw&}td@G?>NSWvh47rA@h)a=16SuJ#|8*g50$0Lwqlh+qa{jp@Q7`z zd7hiT+sAbCNIsic36vJ2uUiRj`X`C8IW&#A5x zd3m`rC$;AVTrZ+=gQNOJT@R|MD_&5x*SD0}SnurOz5inKWrSl+vfQVdXMVbRG#U`m zOXfVm8uqa)zYN&7n}lm+a83z0*C4wFFPqp?{!zVhHe>scvt}R5jsJoV*f}--eek+0 zGt$8%$sm4+|6Q)*8=GaZ3|Xe);*DcsEI`WxKGL9BNn@`bze*s}SeV7?S%<7>pm zvz+K1ITeV&bFX38tf`mQtgwo7giQ*67Ah`Xh`l|CI126;;5yi%3!Ho|yviSGIb}%wD@@-$6a%FYRGJlM+`6J>g3L-)syH zV~SZCaC?lL>?8K~{E~;oJZE(FSVgbBiwugFDa24Z`xyK!$G=HC>n>ee{M(&U#%Orv zl;#t8A~OpQY0K%ueio}w2#Fd@WKh&3-`5(w`&zl%_=I?g)ph5WAb9c3?gMU$#o^_@ zqrS*sf8TB+`$&tLfig;@AD$T({Q_?xfNU@+L2=E;ulhL`#1`URqC&MTcAN>lE(v^o zI$36TC1$Xy<7zELYE4Rbg!6X|CcZ^^R+dyc#x{##6u{W+NXC;D>5(oGzijkzE1O(# zbnKjYLA{SxxWzibwMu-AT*jb(i?O1IR9Eo|HP` zM(q>vJH>#U(ln+@ZZ$?~?-!0$UtKsOcn;d4Pl!^Yzw)nZ0G05 zehL)(t#S32N%YkaXz8mSPByep$%ZbJ?SKtq3Sq7-nWVA}iZ0L3`vv48>&}{Sdw{j` z6JwEy{cj_NNpJBiWp#0f+nBhvfH|^8LAdWM@&j3LsaR2JTefUS-WqkXK_uU}BYiga z4*=Dy2PKHHtEQ8q+qTgMnYRd*8Pf0Eb@gv4Y$vM&;?C5(xOhzu%ie@9&R!}EZAhyh zV1@m>qkDrixjw3VpKkIhoUn75j;WZL4_ZH*n1KB$avl?;p_q5vUxf*T8M&PVHG4kmm zf+ko|aW_OITT)tb2xnYho>i@&IRd}y`k7i;>mpT2Mv8DS4q+roXR9Qm4k(uwPue^q zmakL;%qkGa$_3qX(A*Z6md3dm8vBbu6$3%ji3S8|%>Y6TR1>|ZjsyI%t)Mf86MWrDU!jSwrK~?6 zA}qF>Y!J=lWE7AgP9AX@-_@14+k@lo(=;x`?sZ{ZJg|)z)sU@i>GGvm_y2vcZW$ZhRyi5hkG2WxDijKQY)hQ z@u;q?vFH-S&h{1Q!&z*mB*c!kY!kIgA#XEwpw*CWTS8prcw~>hw%~=vzsfnyNur60UXzPTG3X9& zNX=WLx484G;;DGsz^zwuNE_9qAEveQ`Y6Ik%hra+0ir(}seI>oCsoj@;7r33FLsI3 z=IzW?5v^4H-Lt=z(yCY`69l)%Y4_Rq=3h5(&R#$w9rq#?VJ<4giUU{UyP&oRdgWRu zLJ-9L>FMXxszU!#2Dj$wR#go`HabN94&0{fe%{X9B=irnwf5wxqAJIdQ`gsQ<{JKj z_vtjTmj`8EW>tfH%D>uit#8yUo<^;H`E?4{jB8?_!kf)!yWPFBdD65qRJlDu`8=8b z8|ZG83Oiz0z>>^XNwej}Yb{ya&`WssnHVjb=Sv1A6MKX#WAQgZ0<>4Rvc|u?t<0Z4 zaSzSV2M!-EpGbGi{oLD+4gWMWcjy5YHwAu+nL$Fd;X@}BhqN{q)D~89<`p$x4!%1B z+nYa^?zrWAqGE=4Sc6Kq&xp}rC#B;13(V{yj!E|$S+i=8*$SK0fv$LgL~B$XVrqUu zJmPDuEfknRaiuJ}`!AH~F8a@Iz^qs8qKfMlJ9P_02j1}ckNHB8ZU3zaktt~w(43sG z#!3PUsJ>WSmIixv`b=)!f>A!DeRF^ABpmVq1RK`S^h%2vp(eY7TUkNvvl5pAcw#&a znE56alYnR^MUwz-=+W?vO;XP=G)3aS7L)KCV_$XK_3ap#CYlh8`vTm;2 z;xr#uWqwZuA2%e+pHccmr6RUs&|xZG9tP@nnw1{eJnnmkVF8l#GVN=@HoyXOpf3Jv zTC5iN{|N9DhiUme`oyP{#|-AnQ4p6~<5yP(A)}_`igDppX_V^bH(Ukd`?&%!rvF+p zq?~Vi>xYFqgI4>QfIXcS)pD4)Ig?0k=TC5z9o*E)PsPmm%(2cNoWj`{ab%HSV&T%# zM>e2d7MGjdcKYymZMI7)?|UDYCvCuf=2T*xY)g)yz*DVL~N4uh<(F?XeZ-l=bqjj=S1{LAm>%jQnCUaZT!SDBIu9eu~t_P z>V~=SoYmwRLS=Dqd(OuO_$)b|SSA}}mI%Bx(F1)d2q_o9p((o532PQOxDcJ~PuydD zqER|yGx^!iPvGU4MzU5^gF!F2fi~V2^puGpNXn&QfyGx`h6E`NH#QQVoPmdK!H`lFVJ> z=OI7;|F+=zcE~2Mrr9UN_qpdk zR?E+-X(GNo1w}9%+6CL%1rjOzhY=Y zZ<7l|dzwJs8=GW6(mdC)XvbvI_+zZXeQosw96p((Apkoxx$YX!b!Wa=5y)bA6vlOw~2=G8Wk(k#j~W-hMDKoUIjkuNE#@IqXHiuwa-)Q&Z08> zpW$C}swYY+zf@^>16w&cAhyDxHG>tv23)>#Q|xAsSp@Tn6V)&Jj~FqGxbS$ND0Wte zi@LixdQ(2du$F`1^tj#JbAn5Ow(&$Igyr&P7u$wV1(Rq(Km7-56<+K|x&6gcKb-G3 zQ*k6(T|&Y|O*`*W6Bh{BVv^-sQe@4d&tc#8`J4`oV}zg^Bzt}S1uiT6N^$7VU<18(=ta1{E#aa0LtfdCoMJb2zD^9?IvioS zWbGiGH9-bsY!}}6%XTwivgoQ>IEW~X-8A`fMnN&*yIY;jxV9KJ+zG&{51y?DR}!Dj zl1d&uj;lKXR7JK$>c!Ui*)j|dtn|&+)OyL|nC3mVCQ{6lDn1ej%~@i@odXRl2obX0 z6ej#r>DWeDs?eJ`u|s?R@yp@JK|mc~hQ zQ?(w(Ox1CiRkPf7>rZktsmgf&#hM>KiGnWbj19!m1Np0>Er3$y(0o8FM zs;cYf|9&vtACT_6f7b)95;skEQdcMjrIIOj*!mFH$=p0-M?KQ)7?}RXB=$|#+JBf* zv`5i4IeY}MnqFUug7P1AMrSfR|GAHce^utlyC%H5k&`>%Ouc|mL>jm_L=?_vTCjgg zwGupeN0OQb%pH1T(MKud-C>*OPUyWdPmm1AVc6jmh}%8)>lgBX`=aj>KmaXTx3YV1 zn;c&sFFb2fx&!lQ5r;1$yr-JR#?@fBue-w$jf|5 zp8MhBl527VlSBUZCM+;FsU?Sc70EbBR27#t2j-JT2jglz`K5PZpR{O!ZpEy(MC2$k z{`^}EA1yNISJ?}ltjNYPzvjszgKn`T_64`P>Y&;#)2zt<{eMQ3&zb{>QGpI>U2>?#IjLBY{-QPO_j67}lMiYJTk14tx4vyz@QAVVJSW!8a2 zg3h4Y&5`GtWY6yB*i^t0=I2fp`l18@Qh70zO!3TbW+bMsCi_eVHl(xBZu7OTj|5IsSb_q$+q}k01|hEQoaQM;k&Tn7^GpUI z4z$Lh&Gb;7o8O+A=4m^AYDjpdGIQFK@3mUnX5r1GOUAV1`8E5mFbW)OaF-9i@kBx^ zk%b@y`W)0fGKCebG<^)R_$i-&c&UufOQpj8ByA+Y1bH*CF^I%IO?vd;nJFViCwl%TO z=RJw_<+$M?{qkj9h)AKTvbbUwTKOBXV&in-&qVon#)pxe!HwHU5Mb2tUg;@n+w++) z9Mv0W^DK%RK#;P~F-&!}O%7@fdFnT;aQU9TJ2UiI_cRy$^z1G?v0YA4MqPZVvGAYh zE8G|}w6(78+eoYJr;%(8VU-k@KWnn%{UG&b&S#ZOHEo7Pv#kB!M<%c$*FGH#Cq z=!3Hsj5C0xD_$z!yy6WEJ!&he?KlF|FV=&iBm;2&m!Bb&=E6UppKVuj)!MYNJ7W)> zXN~n|;@iW_s?VRSLYk^~{Sdhrf~r1|BUM!gzO=@odh*8^ZJ!RKSKo=lOash=uARD5 zAl+*Zpt(GB7PE{NmA;SPG)R2`)WyQC z8k4EDO7CHbSLCEL)fLWu261-83w-#e^VXRH)`_$iQrL0HZNW-A6M#yPW@Oz9K27su(lpFO%# z_^UQpM7i}nv!;Y*!}7xDn$0MSG8ji9U&a19YJ~CO7fL2jY-fXxT{y+wBaHl5R%m2Ss7Fev7_#Y;$-;IWe-lTxJ;S7#50M12uLh45efY)_eC|w6P zk}d&A9*g~n(SEuG$kpv+4w#{Ao%(}=oV_VA*=U2I5~oScyN}YB?^+m|1TezS0y13T zl*KMLcTrXUi5#&WB`hBqYOAM4SS`FHBoAFkFSDCR+8Wf3a%UBkWGF`DZhrW-cyPPA zbWH49&(o6pJI+1j)H-**Qe8uo@N7f3@`6_)iAX+4ag(dN0isOmt9VK!8kFN@(XdJe zhfm`9Sib&!bNO-zsnoaC8ty@u-OAk_7R2V{@dEujvlaJSD&rZMI`H6g1t!*$OFyGB zFJyU^t72gEh0~r&BeS1Pn0{Hz&@aw~OqVM0nY0yE?KtN!B2?>el;+wRxX%a71B+PZ zY4`hQQEB~RsGCwYE}&;;r|DA>#X1EW{j8gK04d~}JKb!*K~_+QvL!wFce@n5q};&k zxcAixFbIS5+8o8GULD~};s*#obW@C+L*s#f5FVIF>8$j;#$?1dZrBm>7K6RJTqwkf zSPhtjDtb}BalM2eSYE%=KZ#6Pu_+nJ4!DyBjx|>X?wj8|ji{?ysg_eZ!cf*yE=|PL zjq&`~LdvK?S%DGZzn+?GZkO^y$FQ#UUJQzXa7r$3$*_O`648DE;JqQ5(4fixg3s%b zfcXXsH)j!LopgQ)SJM^$T|#$|ZtND%j}Mz4n5VJr^3=C(>^D&&R$^Jhh-(t+pZ>3m zxF||5tfai`{}J`o;dK9h|DTzjZqCt6cO5Z?=^A!)cei1V?l#?~O?T%G&M-aQJ=3vi ze(%rszQ4cwPcE*@<>K{xJ|EdUq^rbrar;SiMI@q-vH(|XSPtPwm1CL1?iQ~lEE;kc ze>2QoFo&9MXElo5b6TV#&4uU9Tj=zq!p;TX)(9Y`dg^XWf7qO{?)xlYaK~AG*7`8C zp+x3kRIMq64g`xC;n7dC;PA^3VSqoh-vt*hkg#q&=#i2_f z?LQs$CK<>UK-gv&(3*Fs`RVf)LK&EAhiO(z<`7~#6mI&CK&gok94mtU!m>J(lh4gi zJsX936>Jz&#nQZK20h!XL9BM;M zv`AM(-5QTJ_FpZ)w>Z~_ITaliWB`&Wo?lyz>Db)VRc#4wwCopay~&$6Xt4M!nv&!} z_#C$mJt&QXeZ|Tisk3Dl;!?ufz3yUSW?2L=D9_|1%baTk%?sZkU-sI6D>w}WzOnQSK9W5NzVvi zo=`YVJ#3i6ULs?_mN|Cqx_$&_Dz;qH?|(^uR}38&sjS85f~(3)aJDU^Ty6c*!|8CZ z9(~NUG>Hk;ryqRUK#O&Fd?5@TyfVw5Fs|`wFZUK1^%hsr)rw|*kKFwFF9i-NezMVb zhup>=`mU|{U7yC_PTxJCbt3{rs)5ZUj2x}k9Uj|>5i?1W4O?8NanB5=m)SRpZqIdy z?nlncrq|v4?{4xK{t}Bmo=ps4)dPFo^V9#*9}T+`d6&?^+I)6lLlEaQs+vzqhiM+P z>avtl?T<-{a(m4#di2~Q#Azv~yDucyDcjYHzr%=s9}M$p z*sN?!Nru83JZ|#7^}^m(*-7)O2xp|NkfA(W{Y`}2q>HNbh%KIZ6euK}DXyUW8ltM_ z`pyO~CY_0c0qIg&0z`&`gKi;iS#YO;HEy$MY%+s*A+4pr##j;jpli61xgJHs(sWcn zZ^W?E5k1NZg&ey2)-p+AM@_xJnJbrKQvEN00PL@+mWal!6JRoWXpb6e4OR7A`Jm*R_LzY>1Y$u5zB{A;l0l&6M{yIo*YWJK%L< z=NmIqFH({=l4gpkEB#XM*!9kQwfVD;jiHC!HV&5+i_FxSo32rFV*s@(5EETWe5mO1 z-1eONtcu=0wknRRnl6@Y?_qd(81c#~RfL4j9_3-dra{TouL|eQ;^1@)4g9&M$>9)k6$uGLFJm77zGN!j!dz z54*O2*K^HiK+OpoPYSpvmH9$W(X?-Cj0sVx(9w%R^MWTaL#ek@8CB&U04Owmi42G! zI0&*+qc&gGK~_^t#P+8N#G`!!TStWBFIt6tWc=(VpJ_mR6WY^3P;qkN)z&$`IkV4} zmp6W1T)ZB2EAmy62C{Ont@GK~LC1n6car{!2)aEfOAmry&P1>J%X7J;u+pI?SE}IH zjJ77duUB;^s9nuqwO27B7h`d0I(cY>hH>(Vi$!8s^u^`wTrXF35H;nG@lvU=k)W(d zP!$xLl#v3tM?gsDmPyJ6n+(`no+aao(CCMTRan7cXXPIj9N0yJtjKc52`KC|>_%fN zR#M>?G6DV#^JaN-JA1C?!(9pOU-r+=<>%*sBvrd>#fE@-gja6copTxgmibJ04df!O zCai0VrK@0nv1Qs>-gtgfnC|g;a8t>k=K)LR^Nk0ffTIyEZw>I^iy2A zG~Gd%;n{irHJ2NvUsTzba_k{pEz1z1U|pvJ#(t2QT?r_YXa^6cL*|{jU}GMsn@XdqLR};wuDlZ&9t=Y_KlS`heXTu!cJ5MI zoDe~Yo}Lr36Q&o*PJI=?(SKPvR=&jV9qlC&X3WWJw(e?9WQ1`YfZ6dG!LSVUtqHMM=73g6S#Nu<%yrAO9fC#k*aJa%K1xk?ch1{epzdT2Ysz4!5gEqUETi65=%e`bnXPJhOV}^eE&TNWFWBtE8g?j z(~j|n3iaRNWiv@KIQc6*;^PwXNS{7$EEZb*j5h+gu3bDlCIC@k{YBejd+#2EQF0Bo8u@numnRu~rZAJ}#Slj94 zjQ&d;Ip>}|TpZt=Y5@Rma5}w?MC!NGG=9TCt=jQpj`6VJ9i4mePxt909N6O{jn%Y& z2o7G4h(2Y>?UQq9hl4sF?W@n$!|`0~%Mi~`2@^V1f?~$fyF^0QqF0^xGM2*n^N!(_ zJUJzdy1B_|S=8%Uc&HT|G9tO=vC>}yB?*dB3hM==Fhl>jT>|0wCAdnM_Kd4H&|xk~ z6{vy}=ZO!iN(ljXMiPQlV8-BLYD|a^lA5xza9l5>s8dSpHP1Lv#&Rg7NiqB{V;)>l zqg#U=BJXm~z4HhMM2e!?*?hKNS?tbIq@oTHj=5S@ZYr~ZuF*}?q5f}=e2IJ{p`<@k z;h7Yx{1A!P2eIXqNY(ZiV)U4>t+#Y%<)HWJg$XH`T=L&%1S;A0J@Av$({4p9YhSj< z*YZ+JwDWA-vLoDq@H;UDKB)=B<%evP3Q&$?TQ^9`{Fm&D(QkF)W0Aqg&=5Ao0S5K!wLenolHa!9Fi>@U^H>JvbjCbaTCj7R;Y6T+-31I0C zWW@`pw1yG(hWIL=nb!R_Su?%`ql%r~JAIFK=IJLx!8grdCxWx& zWykTlfs2hQi_d%fT8Z{@LjIj(6n9j%Yjq=vbpl>PRGlF$>>dF*+^ZHjoTCzK|MQSN zffxTri2|%*X+CntZR!V!#6AuJIo~ zeXzE#;?!jq(E5aBmM^(Q=#8Wn#DPr}J`h6lcNu?OweL6=C)qwTEx5G)SE*^vz>ZHG z0S(tSfUwm>J=fC9uoXjP>wX-bLMb$VVmE%6q4_oVLf2MjH%miEn{(s4HWoqtxR4-G z-oz^^H6;>In>z%VcGy#Nx;CXR*#y-8C42Jt-47BLE^X9Xk8!n=cptZhf0pB&&5=yWB#g*GTo_^sy`aVsw=ES(rVI3!GI`CJvjwN0NwB{OFid41|EVSP zjs~Xg8uw5;2g)M7i4VU@c@iPeTKZ{@E+ zeCx{SWSK#Bdv4X~Q%`QGMpmq{!?~d47>dxbj55aNus@}x+^G9O#7KtZuovH6{rV`f zFw`>WJy+2D0u4yE#$PaU8jlqL6hBNvmp@#&eLM{@Y03VWZB1_Pn1#88*Q_U!pv-u7 zl9zBioa~0rzTVvjY}*Z|?$70;z(I<0lZ~eNl%O_mKoq>`@*Vkmbe5&ACP_F0W~4dw zdW-^_oYsz5r;fX`(AJ)0{Dn`0W{686`1?2*+zqa>LeAIdLtMD~8J0qLQxrbbP zMs%tu{5@+Ym?CH;I^XPUpgAp3Vwt15jyS7nC=%i!Uu3oE9{?MBGaV$D`%K|OHg1<4 zYj}5}3WpAn0PeVZ$5hUBLbxM%;##uBeppD@7;4~v6~X{JeEjPv;wpc%OSFWCMpfiA zz{}?C(ISa~=&+es?yAWwrT9vO6IFjUC{hGch=%MPEav#r2rpr@dqHTU)GuhZ8Y1OM26~35NYO8{5v@bXl=Xyn#LO?(S;P7bB)<{LMolS8x6@y#WX?)F=il z`lr1qf=!BG=FN~I{9%1XuFyaufC`1F?g{7T5~utDd~HVTpp$hnzcNt;^tSey-KoVa zrpv7EnUsY1``-C_B{pe2-SEfOm_R@};#uRLcn@rLqdL>4v%;pbljyK#0`( ze5&7Ei$^3lT?M#^y?lGbMrWJ|4Ie%|&%0RS{@rcoBS9v6ge_&2opn&^*eE!+$tEdu z0laCkB2QlNQ2`z+Bo4P_D9;ZK!KHhHgVCaBgk7WOt6g^9ZuPc&9DXvI+CNB8N`ox3 z8aY{q>dmJAd_9Yz&tLWLEB9S>EG~X`(VRVAprS(bmh8r&g!X+GUmiW1oh&8wSLB!v z0bD^|nq&vtX573O?i>g3RPg6-gCp_fg;M2d~ zW3sF?`ShL0u0(G%5@&G^-(n1(v>|RC#)d13Z-2+CPHy;6p^5r&YFREP)_|t}b@h$k zq65-3I?lnv$z%^Y6sgZzX@!F5w0dZTo*AY6{MbljG7>+cW@BD{at+*Pd8c`C?%Ep{ zjmHg{d@ z^b+l4;Y!C7TFFX+f{a1qDL?Q3Q_P!J0q-A-y5CD;D>_!Nz@SL1`GfUE-(#)m%_6`a zk88=kq*#@_aSzd?ttye&!YLByS?|ntHHq!li-DU~mmL38u{1wDPc@3{3vLXZcInD? zYvL=OO0yn~mxXrkeQkKl3Vj?bX>#8T$`iji&c0{y_qV?5{;Mgpsb^T7%m~IM(T;N}YJV$rXV-Uj^;GckQ0=c3@ zN$?bN$Q2-CZ_52-nZ3dfBzeQ!ayWbnH z3!nmiGbi>Ug8n}k;>ZTD54-3VfnH21<9sa#H)(>w0oMUErk5uaaNgRl6iRD$HS<6v zPbmjSro-**Di2d94^!g{RY;AXz@>^b4OTv-?8BdKe7TuW{vT6X-OXl*Yh>Vqib`+) z)a0<5d3aa2B*EF@D&^LtKme@eb-y|^ej@)5bUZImoxNS?4LA7MsLpOruk{@~OM6I{^d$ z=bC{@ltr}Yl(gUtiwpYilXMQw|6H)ostzyz}m_?5_6$_Vn$d&Q9mmPfPb|5wLTI{+Y^XG_*x7Oig`D z^`UOpjLcIEDk_{hq8w+Ui5lV=Nth-h!&%9|fHm6*xNEJ$I1+<0jz+tLYBe&)4F7-v-wk%3=7)rcNfb{S+zA3PnT2V+v>8+ zC?V);>ncFV8iadpnl(%ir$|q#Y;sOI3cZrp@p!?{`9ZN6N^plPP83&A0fN0yAfgK` zKE~>Nd8^3CaWy+yS!j`VYf&#(Y3!)-#G{>^wl1N2CefSYt}dT~{#JfScYyy@W(nq`_j$B*k&j8EFdM8FBtDyBkv`Uw4?7E{c0 zY9^n(Pa)-}qGD>3ED_TZj}7@U;-VpemD3GhMR2U@e%R+3SL@8|EI}Bxr z_AAEIv1f^uL8Z*w#~-jY?q@{`o|MZXF&D@#e2W>x`y~~w^MXK(_Dn3lM$mvW?Px8t zL4xu#BQ3d~i%(CLOUB5JX{g4WE%|7MbA>0vbSKuj0% zyN-BBf&XfnI3zvCHk}lk<-*tkM-lNgTtKt&d~Oy-z!OFQKE*l@I3r&-yx!*D7#<9l zWS-*uW-3KatVHM9g_ql)yLdf(G;=xZc{vF})Gv50U4Um;OC& zmP3Q}QqR&D^aFc3Qe}41ZuA5==pTv9fj_N9h{*<9eCfH;)9Lyfy|Y=%h(wHvS+Vt| z4-^Y0J+dq^1o9Smle5Itvuy_%G_h?99v1V$xCutk+P4z1$7Z_! znxX!b$`Ujl2nxM|2D`FnI@_hC?(A&r=Vyx+MIL;!hPD!UMoedQX%Ial~Km~_dDW!W4{E3(B}3wzWQ0O)s6Osf)wzVdq4etW|(FxIhmzQTUe%2$59Y6 zaLr17=+Z*{1X0>}dP+=Q?Ptqh$F20eZT#0TVL4j)fA5yG?dWNK&COlIo`d-UdFhHo ztdqV(Zb*Z=fwSGyqb|II7q-MKF7Dqv{l{*mzKY#f$aD~RNY%-4tU`psJJ)$e%xXwD z)(xD_9`NXiWE_n&NFIm~jj;i~{BrhiXYo(1a+VtkvbNtFrVX4sjETzc^7;rLCu@Z8 zBdGGJ66Aw}52fmBKzrE~ER3UF429<4{&W7F&D_Wd)uGdsp?hg5V>fFG_l`LOcBOc$ z+MdwvsQt@e_-Dkbu}IqiM_X&Q8-J9KcgDFyZ`+AmPb zDmO&BQY6h&a%o`?r2$=0xTDhFja*a5{tC;b_=h#;GqX;{jfW0NTfoR*0rg6%Z9*=qd4fxY#PFv zjc#g+tbb9@m0N{;NIY3`fZ4CpMbSQnYjEPX+Q$fdgn7PY$ywO8nXXGatIm8pBwqbn znFIgOu+c6aa;vPVBNm`{D@HSV3--R5p0rUh2M7I=Gei;AjAV)8{X*l$JCIA#to|R8 zGuXu(xBOjL4fh!&rv&ES^O}Go|G_RDEkgQr8HiITtMN8VuYEbXjR8-Ya2TGP0W=t| ztg0L|!INK684JM9%_NCXO|VBRXjSNAPeL|3pvs!Vo6KANn;Xrk5k9Nb;{JSEAxrad=tPemr;pmWb=W5kE)Hu0F{Wv8NKO zERGpXiK5efPeM0GQMb7%!}anG;M!H2AXTP^Bko_o8ECG0PkFPjjZSzWnI&MOIkBvH zWz=(h=sD&KwCzY+Ecj$mg?^8BDY@n5aiQd_t&ssxG=*TmKT-wyh^>F1;@<-c^u?XI zo>w?9;4gkv>{iq%NASpXb}vc&lUon0=e0HVHPOL7Q|Df%Y60ql*UJw39(V+|(RZcp zzeG`8`{3Iw?MvM9_nVjQ%g-)shRAA92N0~mw2+IuB7){KTox!kCXwQ`*K4@);4W*Zi4 ze#`R3A*&;OKIH?2E29za`6zuAo95qh_zE#@)@j@Vn!&d~Oht_wRgzvD$?_(E#$evh z6H3B;KlNmjy62{wc$-$_I>7u&vh|&_%fw#uR%=4KSjUl$VA|4KqA&~fktbE^a(nZy zdSjL~2g=gc?459N8f@5b>pYN~jk75e2v6S#@ESuW!!?uDeYeVL1CN_rL-LMVSeMq^s7ke zDEkg0LV`b^f+w%fiW9iSeP@X1rS?E!a)bR%#yh{DxI&-M9NRkqTSpYcSiqnLR5AA@ zmhp^1cc)P(9+Cx<*x_;G~OfkTYC&IM($GV(699j4JV3g>d?2HmtEw{X_0c33HJkPjwUJ2u}h_lnPCv)NdjQ!xr+^ z>^{|#L(>%B*`zlx!t3WN6~v-W?+vQmmGv9)v01X>U4gXsGaI91al`WfSi`+mKm8T- z^XaLzSRRN!dh)W}z%e5|nj5{1)%|zNpojt)Dpo^=tuJcbU3w~Rj!E~$j9H>_#MFYZbY`2@2^VY;t=qC?9 zF5Pp&x&pX&U_SjqhdA4JoN3h5hg#%mJ|QjPqJ+aCA+fe37nQ!E2JCJ?3G?%f_cc?% zoCltpvyhc;s#y2J*#;VXVeMTR7O<+^Ag~NJ=mtEJpY%itcMMzzs z7;|c@sOn?v!09*$Y;af8Z`44DQ!1|8ZQT$zDrz8wwBgsTZxa92A_&pd#tD2nrdjC; zfe>~6fbl2{j-~%NofTs_cpt&VwS+qz*zKyEV^QoH$l{#D1ShDzmbgz6R+gdys{w_R41V>d&P3f;NHexE zu0qwTj&en-O426&I4lg%A_ZOrl8B~`BK2?{iRYv*)$r>#dMN3K(UJssC{G|Xo5zg4 z>tC_HUoQ>~feSE}aQM=5w0~z`ddhTKn_7~ZouOwle_ic1J;6VU<9e+4Gh%&(dS;P0 zp=R!<;Yj}&{{?kX+-~+Iab@8ad^Hf*yVt^hcv|adM{AMqjD4w7pm^AstoD+`>@Q=Y zKRGEC>#VsswYJ}5lQ|N7I_QsN|4Uv*+o-NgaECtC$|%vZ;F4;OT2K@R{D|RW`QsUJ z$slJ7L`C_1pu!crK1X06Byu77L$y?-5%qfkNSC4pGbB42x!hzdeu%GLc!;) zGN@N}_vIY01O2W?^xom2Jf5|8s}JPrO5KE90x$+O6^@Qp*1EqYfxO(5$dIH2dz~ic zHXYr1C$?8y^9M&I21+>OOEP|(hDV22e0&7?H(kutz(nP6d^o>Jt}lNsW26|WLiCwf ziXdb)X@IuTb?8rph5APeg+6YK#T0CXUvDw`Ompea7`(iu^F3ZY--tbL)Lkb(p9}}X zgAlVLZo0GGSq!T6G0+V5p#KUZSL4>q|%nq?6I!mARn=tKY6WnU7JvYj%F8=ymJDy@Z3ThazG zofQ{B!}~7R%Tdm|di1vtFPkc4d8kIx|1K9g0K=3SDe30VW{qGsN%4_M#^2ODDfGE) zaJ9Z2Q~7dM!(~*%3k5K~u$++~kxEPhp8PQP-wd2PmsaQ_-F}5ex$OB3o!pIhP$>V= z=^HhAZ)c<42A#XtFg2HkWL_0EbhDk09|def^$()P855%j|d1bJ`Ia(16k>K3c> z?KKnIi0pe*e6PQnI71i;L|6N*^*c~Rk%%5dMqW-by*HYA|!GsYct z=KtQ*ulocIaf~B9O?|Ai4<z(33Je?13J4mlxmq{3WUS@#h$tu-`? z07ULObkMPPZs~fDgzP8q<)5RmCXYk!!wHnhITMu7pad?IX3t-(8G+q3L~2}0s@eji zm01$qE)I6mzpiPQ)|gFs(E}RX5F!!p9?T|wF`o;A<#KA+(tO|^Hy`a#_QBSd!9@j# zVg2xTrz2DRo77Zqndilfg&xkA(_BdsWySMS$PU|B`Li3$nFt?#Gw%pp-?vOws*Dz0 z1|N2r&-~3ZWkD`rB4gZ=t>=D-!44%K5C^t6h@s$*q4;$e#3~WB7P}as``y}$DQBFm zMhJTw+5fri*JDdIn3kajIYWZr63b*0RFQsxxbaJ}d#kh@8`q2H;RmVvr<{_^nr= z#WQZ!K#irsH?-k_sHmx{M)*b$ii?lj+>A#;QjQYn)0!=HKY~>?!xet~DcxDNd|7pK zb#gT4dkhBbNaN|m#`)zMZNk73juf#>DdWf7>S5>e!wJw4buQIZ51Eh9(zBuw?dkm_FNM zsEz{ac$54n^|){bseh6se{?YG?qSc7X~_%^16K%rG)c5YG+V! zCMpB>R2C~uU_Q(oYCD3nX9*WOg)Axv=Uh3CtCZdTt>_~kbZa!#6n=84!IKPi6_+`b zryuPfkMxz(1PX{|D@-)!yaBvMiM;H^ypNK(xfS6s!z*%2{Gtz^8G-tQZS`I;FuX!QhsiZCk)Z0es+q`k;E*Q|}lvYv3;kWe>+@PY7pr$fol}&Choc4#(o!Y`Rw0qmG!+9$NR-!Sy~|Cw~Z6~u+%v_-%7f;j=z8gRPs*U z!}pt~o~96`E_%=@pQ@h}uEQ~e&mak8sB*w`#P8hOBNnWY&2nkOVX z(9;X~uwrkpRBAC8uJD8Xa*;oJbedc5XKd02)5{+XQ27aSK}7|msY(~5C_$VT-kBCD zsLG9rtDBV*C$1Frw)$z7*Nqc-ed!O|a6_qp;pEEjTD5cr_3zBs7_i0rXwk6_yK@(@ zh4b4R$aXTdO?kLNA8A>(KP{*H4I)9dmYFd)`}gvr>$ii3t;-mS(cpkv7qsqneU8r44F`Xa7|YF(#k)23+uy8cZP-2$0Ln_!rLF)y ziq*_@YzyxFS4_z&0ldZe`cu?qIB_R$bo#f^A5}OBl5h> z<=I0wii&8pv*T;`8C}T|AC#3u`VI=E%-Zjrx(pnI50M)!r(n%yWsO8D5`?wkY`P64%zaW6}Y zV!KY-Mnr$Vrn!|&q_}ykYb>0t8i`*8H0XCXR6CZ^JB&xMMt-(WJUvXpWQZdi%4zaq z%r=Rro8J{B>Pd?XrYWyeEEN{XDT)+|0y~oV6&n{m5P6Q$L~(01>Q&dPV2txFkKP>m z(o=!0T4#0dw=;a!&c~$czk{sc8Nj4zcF%FvYpBeLKoG_@i;3&I_ndgU&ElboQ&Y#^ zWZys1CAr7zk?&<^J&V{cG^MFj-~3rUKNf!isHpXEP~H#hu%w;p*D(w+-zGCLOHfn; z2Y8~9;rtdb!WHbsjIIis5gfwsCY6{znm5$z*HN%2CPoeUaNTz^P5dJbq0lk0vE;XB)k^e~OXkL7ei1fg!oHo@L)bO5mX?HW`AS;8P1= z-VAIEWS{0cmSKoVK7`!g9u%4dOmd|>o_%s`lB7TDo?lnW3DF}@31auT)y){-fkFus z0Gi6yIsc|_iedAlsemt99ZS11>CiLm^wtJH=|)`9KDORNk-i>_zMm$} zMjy_9cuF~1NMF0@_Cl>j@u_y*UwTbMbsm-ejA)#x-2t2=%3$^y^qW2o9-EKRMkx0W5 zVmYqdHQpZT4Ha0~ErnHYkxXgYgwl=@ki^9(dI*%oyIFl;vki90)T_4#PGv@#pItdO z4lA?x4czz@+FAKKi*|B?m{R;oxB3srSKVEX8Fl3ijA*^yS!)tcI5z=AqF4~MSAe0% zdmGu-5p0}8l2a|Jh$32)q9@v+%0pfX9={u&_pHpQRT`-&nPC&};cKtvf1s^w*tTKL zw4~)CAufvvKkS-UUdYDtAo2w(o1L6m*8bT%#*!8nB`;JSFKQ(A`iSaW1)du^H_Myx zq@+HMc-VVf)W=S~=YdwTP)L{EC%(pbBOCR+l%FXh9^e}=V$Bp^siHEQ)nmDozVc?t zIv^*Ik3CK4>3AxMYuvtu6|zXq>#s6tw;EXKIhN0 z+5qnC+~0Vz+Xcg0XT)kJYMTm@%brJ8NkzIpw6!V6zdd=lxaY=W#6&kZ*{0AD;T*gg z(VF^AF+(B`{`@tHFbL~zq!t_6tyHcr0~Brz$@lp@jk(a^O7{L{D6xRZ^JF6> zr!(|0_U(Ua@1RWEi=|7{GHN*NazL68rws@veQ2slfT`3A4$>0Im=DYF}qmDGIa(oXdgwWUFdDj@fBkb;A zNNkzsAX6}60~p;!katG0mw(z=CA>FqO&1)sjPhj^sxto(<_&fWINd!IK10@!k8RXJ zupz2vTCRC|tTJU%IGJ`nJ6z^#y(*zB)d=!H5MaYH`siRMvq%QRMXk3; z3A;gxt(aO`$?*-&bzO44e7*0bmUe^{(&|=xnm~V|EM0>cNBwRk8vq{JM6h`=ZGThH z-W~ryk_6vjO=nai;l1k&Y&3TZ{M?ceH4ox-|zPyVt#zmH8ZRo zNR--AkR-Y}ikCBRa#Jl8N0Uyl1mNM>uGZ`ai#~n9tlLsR`{L4Y*gZ}TH;UCxbOMgD+xVX19u1kT9Rqkrbmgle|zE7BaNzkUwO0& zZRE@RqLA2x`0&H1Bn`Rxt2+h={=eRC@YK9?#-+|^}o-$1*r3tjWKHlp%Hvp z>fp2680@mc_?Wl?0l!~c3Q@dis;$bkqIpJ8m1E zlXDdliX!Mj_K!#?5fxR~!|Hvbmj)PuRyG=4mQH-x;4Gm(K};9w*0`5$)8?fphm+fa)wKj1 z%O*)=P2U=U#HQGAthN)m6(_$j;dIY)3Ie*jGOzGZqv0<8obl0YJJsPBOO<&;w_Gm+ z?X%6P$Zk{P-9M$9Z*6~F_Md0e8dpU=0y=Y8rw);KusQ{A?>f63AIrzLwm=?u)Rc?az>jsWkLy~VE03VsK>n2&xp*Ee+gx1M7(=eJf<1PI#C z)HI=ekPxnahhBrC;-v+sVYr%S;+ z7<_zFRW)3hwBYqFk0A;cTZ6CBAVg&b&R{6U{Ng0iWDD47C5Z@7#G4yTC~eW>cr897 zb~|=lHqJK~woL>#hb@6(^Fk50gU;#m2 zctn|qQ6-l$?n>T%7A?j*!lM7^M1mJ1+sH05u8qB484kJ1u-ZK4yR}<&GqeHb#zV+@ zdP;o`ml8>KR)ah9%eJ072+p0mlUQW?N=lUsph5Dq>&1S2&vz;tvE@G{b+%k55x+<- zE0(Wq;G{?5BWDdZqqsQSCa;t#sCp}UbLTO2s2SQPJj$E~P7pg$Ch+(4`Y#Dd`Tt4C zXRjQWzR{Ag}YZ`l*}cvNz95c^bPj z|D?Ow{LE8l8O>HAui^=?3Fp-6rKRrLSWiA1yqFX5cW~}$Oa>%@2VExW6ic#~5!Q^A zT^r-7iaY%bpJ5-#&dyx=(9lAJoaAQ6f2#c9yLyr_4xU>#Ce$TxJ2m+R)2wIF*DJEg z#<8cYIEbV}IHbApD|tJUoI4xm)mO0YWnWbDV-M-|cz+z^Z#s(S8w$z5eF*evr#s__ znGKiBN_yu9b%9W=iP$dJNQ=Th&G3T%nci>UMrXPCXv7D+7A6FZe33>=XbQhB4~8R0 zJx6Ct+g)QZ$1w*jwS%cIg|5R_mPZ$;S4JE zC80)CO91?X7spE)NexiTT}14o3?;PJpnSc=m zz)33HeK&J{+bG=l;o`vb`}&&^>L3Pe$Y61qeO)=sE6T zvpg@0^N*-d}n-rPJ(J%xZmA1>uYcbfM^obLr45AQ|>ifi>grJIo6 zao;`P0+Bx29H~hwtJ&P%DkaKgSY8!XP0<-GtTp97?~sM zm;D#I_+p*Q3b_Zd=>P~pOYDepPbyK-S5vizwbGm1AlFYJFeFi&+GAIS|euF0Vl5; zYAh4^R-c<<%d8DS#Mmc0on%#zUXOjK)nP7x^9csI&PQqzb(ZF=fZ!ka1}VZ zb_V=fX6^G7Uw7r$Z`HS~k%$BcQD)_)303$g+dwvnOw-nu*-E8dqBj(@S)$TxBRu3e zaexZ1*~J)5(VF1Pfv@*a$0D&=DSg=nI4L^{Brb?)UfuWQpr$+^>xKN3(Q@xR{S`$& z4EGo5)X-p4^PTECmk;K(o*`{*{W+r}Si;|WQ$A{u-sac5j>Azhhx(r_E~CeXjR&(( zLXEuS5GhN0`EQ4GgVoO?BWYj7g(`x#YZ%Yo&f*YAF97NZo_c2_6Cg=)Ep?oZ({$1T z|Djx|-XqFeiRga*lIb^npW)mJ;+t6ny<+D*7(SqI8S!eCW(7{5Mc!5ag=6bq;?rb} zO;0&1OixX~K1p@=Q;e2LV&a2kbWZT7g0fBMWENI{iD!WSG65vX`eXL>ziGb_ZjI_L zOtLt5nDckEGtWq9DnxlSxx!^6r&g6T~@wy zvjUmP7bqDsqY1gV^|FWMMMPZxY1l&f>BczmhyMJWn}EEhQaDsHY|t-P@yRLv3E5{F z3cO$H0w9{*3|K0#bN->wYI%C!o8SX>eY1P(Q1|0a(#d2UIoeOE{P9a9?dD#J`SSS5 z?A93WQ>13y2x~qa0`hf3)N035CKchk!Ahz(+q4vagjU8qO5hVz4wYG#d zS4rja=8x>7IsaPh#$tEg7UXAVurm7J;<)5{RFBcf=9y81k~}m{Q8zW>0GYK%Q@{<1 z)9AYa>yD}OwFWSp-XI1}iq&(*aL#70i0ELyjy<8r&Z+JkMnJdNoP-=IN=Am#evBHX zv8&=dW>a7A{<_KSQywfjBuhfOTU{XN=>3AhP-uyX`j2pe<3cAA{^51N(SxY_c!R!| z$|%o+5_UPwx}_MY_lMyWtWtm7o$(ie-_Fy=FC>TsG00(rS@o}fmG1;+KzRr?*ZomM zeu6sj#m2!S40^Q#w05gT;swZLxrQ|fgF^HoYkB)mGlEKL7U~-tf2}yDRwMOqH*=Ym7#cjRIIlqeC1;L}fcKSDeSK5(IE@aMnJWE=og_QfIjR&y zkBF9p&fsKY-NJ@{>gkPIWVX=dt1O*G z0Ta9oAw45_SK*p>CseDaz7H72{(()o6D?I)NN{ezdTPjed>EPkUBA)4O>V2Q@Q^W2 zM%k~<1&fw#HyklMjmue}Ie?;jvDMjMX_n9ybHz7oI>hSGWH43@D|wZ)o?H)XTVXA> z-44*66}*n}gY9*5BbS^5{ez@^?;*aT6q)A7U6Oj59StR8?mDZbIp8>pgh#M{0MU~rADQ!hK&l9UXapS%p2tZBhM<%eGObhts4uE62`*$gvb zhA_<(T+LjAj)j=_T~9+vm2~ZQr+#sgtEQ&7Zo%yeRg zIn9}BR&XarbNh;6?&f1PAQ4y9tNgJdO_lU_qva8X=i-oSb*x#HiAVZLhn6t92G%=+ z$#W#{ZJ{JSb|wq5OkFoN>0OzkYFjgLA(^?9@1C^Ii{m0-N_$vAhp{u0!|LyRq%0s>w7d%gcCm5>m(O>1yNmX#^1=;MBSz6d-2+^E{D?&#K_GUnu zRmOWi2EHAU8Z|C+8oeY??wFoSw=bB%)xbf5=zeYh0n&SH``-Ej*SQ+DLDmg_2oNHRz=ETag;ryOs zo9u1xZ?tA!W9iH7ms>@JLm^?V&zB^&eWm&snKj&7eqL=gBn?*_QOcTbbq7j6&hMyV zd2halI(N=&7c-|JOw{|?nNeIU!;}`?R8})7YS&eGO?1U3!mmqf%g` zsI9RC!kOYu9Y+kZqrs^ToBOf+XnB^$LIHs4X|$CRPei&Ob>S~^RUDIV-%!8Z2qclx zUHC6HUBE~Gow@$^Q%7apHOQO)1NYbBT2!j}UYBU42*JnV;-cbRi_1?^e;aMmcnaWg zu$i`jB_HifcnSbrIfm|)Sx=|SSQJcS)+eB`2X8W@HTkCC=|ikA0)+V}ND>@J)5|>J zrD;xQoMyXB!UJW+O39e!O`8lYeD1dh1g$P2UqKB^tK)JJkp`qe4+jKLLhp);y~Kob z6TXT}KbW0gJj{l#@~e!D(#g5R`z9EMLdG_wlL#X2#bh1$`D2z=I0Y}SBdGOiUh>B* zxM!6wN#(F&!pohE3y?>~_^2VdG7VCa7<`>kuUaKJpR*4cW$#+wCy9;_`1tr?2Uf_; z;!ERHAz}8P>p61o3%&+xNC^i57P*vkF`v|FbAEU4V#r z_ReOSfeMR#=PXy>i0}M-j-kYWjJxh*fFcgcH6`DqCRXTa z^o$AO)A_AY9{Rl#OVm?^^veqxcgvRK(Z7K~J^c1Zai;alXGMdaRhuWE9#Kzq_H?ca zh`nTfD9Tl0U&i^5SvY${tNqxI$OB|=p&v(iwg>rwqHBQ_z{!-F#m z%qA?iArQ0KQQPh9AIw;;F)_+l9A5BMCh07$9kwkOqgNWAAtTVwnz;}Z<49*E{CuL3 zm@IFcTKPU>91Hv0`?|@hm~~P7BK5d86cmwQ%P|iMf73ome0FT!9t)kJfb~cY-r%H| z7sVrd1jGssSf1Y1)I7`z6I=^k2pT+V?~%l>G1N5IFguTMh%W$lLy)c{nDXb`+?dx5 z1(vScmT3&|vMTA{?&Lk@BbLj|lZHA_LkEzjOGjR=4RJ@I-`zIM!q%d01k$OQ0ddTu ztzyHnnejsM!O@>ekQzi|yjgi5O;htUhf8Gk-;`>IJ$#<2Q3kWRIl8`hduIgwX#&`v z_6E$aU^;P$#w@mV%F^0q*J19e9deq-+C4@~O;-@8v^JL6G^jW8Kq}IKV`Kl1&RN6* zbASGi&3a5}NSB?L`P=UaTW!ogH+fCmhjBCRk?X%aBjP#t_Kz#IW9Df<_BZUk`)50( zHaWWIwUq8I5ZW$Ek+c_G$}0=C<>mf500bO&OnGV#{a^#9wwAnUUz6E|hc^53^u}A; zB(ne|5;vCo{Nzfa%`*&?DDBZjR-pNqd2WP4!dYrv^=bB)j+5KcRk75(neaAXEDs=Ca2Y?KNJT_tp_puB#;KrcFC&AfrCif$WdT z?rO5xF)Xp^-qXX`1Jb@K9~ypbZ@=6o1%Fp6C1(|AT%+k& zt*0kV;a)Ii5Nxrqf0F!wLN?9b-;YI~mm+s;Caby8U{5x@yyofdJ2Y{B{`z%%XlyS- z)CA(+g~n^1v+%nXN|`6NX>P84frWErrL4P~5j;3UGc=gKZn-xP zXWoY_Ad@RsDdPt`C(rEoeCn0_X%0=~>e2EA8r@@{)lKj7gFt;+;&JFnz<&dt6rfwjLZZKYwgRvneoYzNSNq;gR>XeY4Nf%b(O;QX+biyTUOyk?p~U_ zTX=VJ?E79M-l_QllrwbZCdiWm~ zQ#kAQp~=q(g*AjPH+^*tCG!&Vis;;pz8ca@)E4DE6rw|Z=Y@p`a;v=Rs|iSv*@A_A zTNy)ids4rv(qtGZNX8V%0oAq#L6h%C=vq?{9;Rw zf=K+;RpD=_naEyC``dG%h=6zmGEm#oOZRak8n9Jp4)2A7u^pq5%F!9EbVe_zg?Y-xP(+DtHvzmkpUvn9%amj=&tIXyBbDG$-n z(Y}0^H1bfCWL|hl5en=ij?6@FmBYu!)M+6*< zC7}hSqLrc`iivT!K@aE8Ph@H$l|M?1?$lF6l_sQM+D9cpkM1jv9_CFI?WL*z`bMr@ z=JgsSOlt=gm_%Q$+zn^&Xf02v(7(@uauFqyxYAXL({h{{72~HFdj_qZcu?Qt1n)hm zwr(}6Jya)vL2Ot_Sz_dqFVO1Cta%6(BPoz_l{95a1q}~0N`wm2fYgYqXYN<7eM%!& zuCc(&fhxVY;x$Szu+A zLfAUDSQ1`twSAxvw=P-Jp!(Irrf9{)B|7pE2;F=4y?)##vwcbEf=~BR^AGU>zLLlm zmx!CO2(o{HyMn86PbNZlFkEHAgZCwme|mRZxvh!?40f*lueSVNN$o4`{9hP*)pwX6{q!;kdb7)OXC)ZOQn$_R#km;bK zqBjYgZy*sSnaa5u8MdbBB&0oawA_?aYHfKwE-L-T=o2N`6uf6l{`nhvWdQ6qm%rE6 z^rOEnI+-nOLNxcM)wq-C3~<55a=tmRy(M1~Iy8S|=`^}AiEt zC$GayU)wPUqwHIyUsM*c+F5Co)t;*}G)I<`yhjc3m!yVq`yIJOK*EvPb<}KkfJ>q_ zE6VY^+{=6xk7=6uDkGdbR+#3k>!s2F900N2 z+4L#`5GzQqT7vi~zE7W2Tf?uwog3z~UXdg6i;c_}QV!&X;TggL45@b(y1lfk)xrGi`Jg2C4!L15atf71|JJSh(>#VVn4yA_PuqH- zI!-P0ox8#Ag&j4T7~S3m8+TF424k4Bm*;NpH(rY}wA0s^)f2>znPSVN$&`YytrXw) z0_%BR+GE6rBdiI=Vo{p6dlgsvzoECg=2@i~CpT(8jvkDBjZsf|v!1wo>z>X&>*b}6 z>TQ>F*A7vVPouSyv|uAo@noyc$$|5y<$ODudowePepKldg0%m}%3^pAEGXNrkM@r# zgT?~c7ufQh&ANFHG6mC{cl3_Kj&383H=^Za@Z1ifff`oUWn|VPiOhMV+3s^tBCkI& z`IRCMc4IP=P+q5efJrb;YL}8^)9)I8r`v@G_4q1Q*OXQtAABskjl*wYg~9 z>6#T<8Wmo1Nt(Sw-lS+%ZHX$#)68*VK|UfDtQm5W23-ZkJo~aqkRN)30?c&Pz}3=H zJ4CvyLdS$#MFiSc9EDZ@4Kc-He&6EUdjrR+&{?~=Hxays_&KDJd;?Hl_$p4v3c(#} z#$SzCbfDIs!ZTa@JB*szx$ysMPLtZ8hCFVMMXg=8tFgG5-KS=yWQQaCo8WMp-E~O# zBbIQ~@Nh$f-oZ^9l1#m|_qH8lGO2-M;d5$z8siHZgcQWpI%<`ZQ?TjhZV{uKs2?17 z%P_Sxf83ZOtgn}6_|C;oZ^7Yd%uTOhu0KWUP^q2SF2#h;5M=X>PF~Jb;!G0LMac;6 zT0lqu;9>b&Y|#ov8qb%BDSH(M@3WOz z(-$}>SKgo;N?~Sh^061b)~7ukf4mR=#wurE|7I;|Uv6@3%E#pqG$22_)>~ul{6+Kt z)ZdjKoAwbc=jwI;3uJAEQg}!I1vlpljr({30X4A(k=I+yOYB0-o@VNqMt7)_=P+x{ z*688>$l#}s<40Wa=l!!)_nZ@PbA3+0E%7E~SFe!b+mlQBE2gzU$v*4M_jxTnK6d>LXyL-HobDa%h^@K!Xj?Nh1U%9jm2Z|V^Gk6!!%(^(zjxI*n*`gVWM+p&TT$5 zKmiu51Ts`O824v_o2cF_05QFBh#V=h$tBtLWB7+Xe+0ECETXTQyl=`8V8EGZDSy0! zIY3lJqhFRcHZ$PUEWxALUfDfbduHRa-sVw7X*uq3>6q$<{&{)&tmS-3sZSgf;AMYD z5CIao+}sHUj}&Y@{r!RVk;#j1uLU?VdIjUtLb3;M9X?)mdOVk+fMQia)I`v*=NcJK zpS-&AY}}fEe4Y|e<`(tZzLqQ4aBtQ8p5I-MfBhk|C8%q0Iih)&$X?)7@NtE_JXJ4y z7PPNhw6^5Lmw79LgI;hV+)(eNr(>{5R&kNSn45bkVgR;w|+mmLfpaubI z&vK5ic+A1^8J39vsHTr?Q)>qF;>=F54`M9@!sh;0RTBJP80#1H?5H>*-^LJ9jnE3O z2KU)BtCg92HR6*u9rck^OqKDk#=?A#Ln=+CB7v=UoTZVw&3OJ%o8Kv&49|n3^UFiu zFc+?93kO&$<%gxbtsQxDx=Fnj_om%BhXhV-fiv{)68SM+0f2^@LAA8G)x-eF|8#m< zX8Jd4Sg-XxgL`u^u?IjMbQ7di2`P5;Me`%7f67uhy+w)#xR?A(qs=ef*zmh?rRa(lb!#0+iJ(Ro$fRuhw=MxZ zwvnOFBcctr-3kMax7R)%hZY7*1wBuQo~M5;-bq-WjGS_hQY%ZvM?S3VKzljR87_R= zAys*kf{Hm}QqAs`;&z|;f+dA)WmK>Px7+(oHw~Dg$zwG6S$-uE5ukcP;_2af)wH~= zGSge>xsoGE%zJtM-y60)zq+A?KMTaLnpV7s4uyYq>TIt+afy#Ty9Vmqi)$i*+a=C? z1D3k+ONYR2kt2vf^&=6dVJtJ?0)+DlH;{}HwQHJ#jaEG)0Aaf1%8YGkuyrY((AC;@ zNFQ(+leh$@NJ`9BB<0d5TfdLhyEu?zlI?u{2@ymljcvNx!Mbq2!w@Ma+nm-6(sYub zw6G0_24sY$<`g{UCmjO%nvVYFZBsk4Vb;ip*_X zMFE(lM+2-#<(?PU3sN-=ho={U)+}0%_+9NSCe02BtNy9WOr$YJzZFnI%rY$3D#;q> ze-;R;;fxWTwFO07Ub!i1^nqUpu2r_G{~oj&!-o4Y)=gVh zCOl|(4&63OLz3E(^vo_AGBW~zzc+1BmZ^l<^(o)~zZO7gZdby815vaSjB*{rnReU& zJiub)m}JM6bik(yG0zx_CH9=p+*F0KKzgx#B?TS9(5TPFjZdjH(>u(k%KSY&yMuL9 z+d}46X_BHN?l#>cQV{`CH1HThD5PMDL6N@)^fOi1=Bj*oIBBH8N&gLj;f)=m(not% z=Rfz5Z}@<^@bB*M@vePRwujTI$KJaWW4=G}42XP~j=6yR^$lbXY(%|D(otf#* zh#QmDFpOX4hW`3ULIr>CVEVz!k7$ZW%+lw2=MvboB3taBW&wB#XUo5 z66(6eE!W}x;@%ghl_M7;s`EllJ+2&r7p@i~7n`M>mF$WwGksO$p$>gTwMD7D1bd~n zC65~w&LF+AORvbBJvuBFSa0&c5Dl}Pg=rSg`ix2-xd;q%vOx?BKk}p}EBb zq&?}{ykdKakALfH?WZ`M+P3_;n=@*TA>S1}kpGUd*5y zSmm%ZT~vLFbL^RIfo%GkK><4Ua1(<`y>TBQfH}@c{c4L|lE;GWYTxxnl({(1nw?SJ zq~K>XKM8X~ZUJZY-`f|K1Z^vS&olKy%5Sd-T)6fLN;XXuU zqI&GF*?APULh9!R)Pb~%jN>Q2{|L+R*vBYul z_kjiUbs}?Hv&71=LhWZi=NhPV z#|ICMmySq@H;^s(F0ucfIG0o4Ik;+=pM^eqbW8zPrxg0&|ccz?adO1fv%QZA|oiGCptb<~>Y$0vO`5tB(-Vc(fc$MH;E z0bN@h%S;mMBGY@eW-GQVHG(ID88E&0gve?mUVJcTY6N#{i!Do{e=V(>ji&;(73ScN zm|h$9ZoibXr)C7k2QOy}hTElQ4xa**^O(HLLLOk91Mx*`EVIaTTGhK+98B}>6%3Il zQ7o}F`TL+Ai^)&y8LT`l+aAro3R}tI4x&6y_5l@YiDWbnduC_aC^#HqCYXYWNGZtjdPt zI1SiFXnUc_2PhS+B&3%cJ`?cdOEr%Wf23f6tAL^slbgr*BO(4Y{=lPYsU2CsH9(fi zbZp;6cr8nBo`6>%EZB+UBiKCNkfL=WvLw7;k}V{S4C!zYVElnX`+P*2R`?;q?rjB<`mLdA0N$9{!#L>&Ao_-AiKT3 zk&I1_dw486&(UmP3}*^1w;DSUOmZKtfd%RDJvbuF_-T#$Np=>#<8;EU6keL(??ZFq ziK@~Dj_}_6ZRIZUZ(9!9a&kZk-6-Z6aO&jM6;2Uo{(m|Pkzl8t`2~`4AzwZKNs|HL zyxd=YpZ{oN=bFJ@D<-OvbD;QqbF}rxLgvR@wE?WXY}5}+Uy6FWlq78|N!Ed}NS&*X z8QI(qlQ+XeZzn|fKbQ024R-tYcs-5P1O^$t%|(qZ6Ma65x3{RKv@c?s}6A1>t@%2W1LvLL+`sA{P;bd?B~jBQEJPX;l-yn7y<)P+@Fk zj?Qzdhlk5`BYl$uYpAXwCl65UeMYbUIk&9E|GaUI^oo(^cbi@XG$_D3gyfW$(S)W8aRz>q2JfaN(#tJ8F7)6tq&p4W4JqL4OhO@niroJOB_+ z28_OW*sMd>CqW7Y<%>DuW2XBohtU zk$P-R>e9~9jNKFJ&-M7z&TjY1W_c{N#Oit~5_0aj;*s`DVNrD!XCv>WFx<~2C^KDFc3ar5WY1&go za+?>{9W$NlZq5>WdwF5*5LHMvOTX@mEw|qxT{y$g`+xu1Jy5~A4eKEb9;Lk<#R>6z zsGz?}OM%dRrHgBy+Lk43Et{B(D?@elNneo%C#wx@Z5X|X%7AlKxx|}~wz}%j{!Egq z*8S`hvZUZduq;}lRJ1)_;{`NF=NI2!d+Y&mDRYoGTem*S1@bdDI55C-0N~K@6;F+A znaf>yUr_>2*P~#9C4eI!1ND&EG24)8yW-fg8a{gPN{pi8z&-h9><=nCDZ57LOx9c# zXj@p&_+}5iy>+PX0@85MPe`wex_s?M?uQZElC5T`!%O7w6FlK>RR zfASD)H)Lp(fGqy)vd8ti=(1uKm9-W~k#a3^-OHrrla}}5X|zUUnwo*RY^jS%(*V`2 z*mOboYpe%ha$;WG=!e^%2BdSK>0Qz*0LDlUPpfjAJ)y(bnRw{nrW5nILTo2hTW+Lh z!Mh&#x;MdBx1%Zt9PR^~gnl`|Rv(ZFX3OV4=PId@{29##{oM zbHYkb_ituCo&49H+WEhpkA2hJMNiA|pN~v26Y@}P@j<{A^wVZU`E(_D%WGA` zc|Hn-zYDL~b(5KwG7H&;IQ>UD>&yeGG@mCV39e<7J+le4r0pfJ&+ZnT#j)_n!k}Y6 z2)jnzTtZOxb*6e@%a-#iJD}B$+PtlD5f5>T&GY}fRSX?~Mdbc;wT#SGom@H$VWCt1 z4K6{xV$>_EJ=c7nSM_d5)R#{jm)|wtxRZmBD;*~x3#-6(BXmLO37mj*7$tKW0#PF){unYC3$n;7B?Di z0UZoro~mSm@6W1Xj!P%fLJa|v(ozx$jfGlj^`uAJa(p%xTbtb*SRcaa&9oOa!9;~x z&J*AEW{ksib6KkOs&z@(a%)(fJ~3`zOW6;uXOiV5WkVxBVf}`XHsYr` z;X$^p+Y&LPouxb_jT4nV4uXY! z?sB6l1aK$)p`Pu~$JpoC*V9~0oN;z}w=$X(gYM0{z7iwrFopgMq)hPY3c-aI4w5c; z@rqMyQ7lQD=O4}4s`YRt?$c2ez^fa~_7QWz^VgrMBijNimz!{B#)?Y*EnHJoX1uzF->4vFT4<5Os=X6`P8xc?SpFq>w znx)%8M)J`0zZ3CNK9X+Q zXohrEw8v?+EgSJ$*Oq->a~|%(JyP!9RQ$mck{yTbYl4~xsaF>7{5hiKQ9`6hMiL(% z3>Nh=qe4%PVV^3~iE(l}ZCNQ$YNZ3A{X1{jI`!KgA!>p2lJvQuMqG2Jlx0OB5SrX{ z05?kQ1nJV%VEH#L%t_bifr~&V-@r@k@wvg%6K{0#H#DTIy!Y*hmCv&EF5c2uswX>PvYO668db!Qc%pH^ z=bNMvQ6VR7ec)dYEeA_Vg7qGUev6~HCEFrK$Gxs6{s$rsEAO0Z{>hrOeXHThOX}@psmHIuggws4NQqP4|uo_`n0yub)9{ zhCLub03g2q3WJZfbak#)@~k+m%&O7W>}jhqLQp>fjza7!scPk`JDjx5xb3KLX&W z!p(3OoGXXuf8+M@WFCa5Vc<^At>Non`SRZPKPn<|=OcV+1nw=`nmE^C)3RaMoJ6YW zCjt&hUpG?1d(k5Il^?NWS}go-P&g6EeRj2N-7{gRo04Q;*zdVuna&e#^_uS<7VSXC zgi)RlVw>pp_$lB6N!TO?a5YjTMV%nQ&e>&8?yypLQ|N)Y8~+s*>`8d&#_8KJ86j;zKuibodmkG%TmcyVD1l(e7QK`@fQW1qCKtS8UCMB5oH6^zvu@ z92o#HW*)dKy>5@?_XjfU_*m$(Pu|(i&Ak1Qk+061T>h$p^|iw82Me);`j%Y^KUGVn&2@7p*ftY*bN1_mNBkb1oL*n0N`qrZa1;ecKUh?_NDJuNQD zB9!Gv8S}fcp&Mfo0dm57x@{`)g!{-zJM4w!I6nK#jq_oQD^$CK-OtqSJWnfAb3{}l~Y)qkQVrT=SBO9wg6x`eI1 zc&kI3vmku476-Fx+B#PHAh+2R|7lHaSX=;*RU{e!>Ju>npK!UR59zVza;|^J+kVwG zD!3EJz7w2Sw&ke#D^H(XW(FF zv>HRr4eI-e0=1hKB-1uSQ2#9ADDc2bPn=3ZaZc*ep&s|V&nlJW-v*8ZR<_T-$Vj2rY#@Yb4mBSqa#kG-s|RSNG&&YC zg@c@`e)%9kWZ&E5`s#J6H6PIf6)}g1=2`EDYQ5DK`%^4V2!rbBDC(`DW8ls@mvpG!F0-^L*M z+NA&q#brLIAtiYdHq?a+9dZ5rC1!OlUUDf@^`{=NLT{LavrNlX)t;!WqI7TxVtm$j z{&FIvwD`Wxu)Uv+I{S6Zs&EOA5FtmPaFGEdl1vB+@BxK7(n~}MxV3$gq?B`Wg_1_55y00(LCnFZF&pudW`9EtBz^ELWq+($#1Y z9~WnpE)OBf@Gk=WJ^VJ~|2i&b|2U85m;8@%zi^Sw*MOfB3say0Lu znA(or${1#C)zL#5g;$T+);+@n11e-TAX{lb;vxrShqRJ3!nyIwYCq)?FTHD&C2?nC zb$G%ZD}}uAlTrio^RO>9t^^NsxI-$aiI7v zsMV!P;des~eYFJX^gdvxqZQy|Q2E0w{A-j`TY!#-AekvCi;IIv&x|~*LS&|?^8*Pc z82rYh!P_T;$>!ogETcan#iLf-Y;IpH+B|SUfzt>vw1!Kl9{O%NWU;_(q0zm>PPk95 zcYFO@!{ox19`P8<<0cqc+>;dC%nOW>rk#<(sP*T70G){f2(feLjf#l@w5O<*K8erw_ zFk}5?pS*%+a#qjHuW+QROLi?cv;|y=bTqE98+uy9;EOGV0Dln$ZM|eO{gO~2QqxSD zd7td?!I&x=LaGg;uo~Zzgv;X*GTk*=c^UPier6lt3N3GvJO<(gn+SavjSLN-UMY8> zC7>){XiPUL<10?eDIaJu;<;MS1LhZURw$b*!2vLZe6V(F#9P#6REtk4Q#hgeCI-QeEDZEmm<>S zq+&_<;$N%o57sD_*w+|sYod`Vh#kR`nlw!ai1W@XQCB*5n8%>k$-Rg!SigIm2Wrwk zpyl6`*~oeJFT#ApUWm;TK~G35-gboWO{E7nEKJRm#JlhAVDrY@L~vT|2n z$4vP1x{okvfiXrc$EJn%jtVbUIW@ms0x%z~;PBgFB zg2j>0V)c_{+y0@iU42brnye-iS^MXy`9ggq;XsD|e4F}`5?wpZq0XBL)L{ruvt_xs zvY@_O-f!jHNqp2y1%qsZwf>lcasLkqUyPRD->vs%3(dsYnn;j-5FF$w zqbsTIaG^-=V%+kNV-$>IE2BVejgvXQZ&j^_db=w9{xL^5SJl|cjNA59%N5V#zKC)X zc~@!=-_pFf;$2%o40FB#l|3g0*c-c-{>`-RP6F%0Mt}ca{Sj9tR-D~&PvYmu%J;8c z+PFXuRQ_NsWBT3yi#3Q(A5{vG&jGy9ph7-h5)nOI7{6_f& z&h-C|PkQr@He8tQwIQmP=WyXe?y)y9^;G?N+zl-f%PS{wL|gQ0GC%|*isnon zQTzqp(+d)lkG<>5oKMlFWPI|+y|+VoWm<|Q3Nl;COHk6uQ}NW^Uq;|^$f~C$GmX_I z;s^CS(+r?X8ZY8tRXD2KbJ^DNH+4hvYSd&5O$fX-dwG`kdET&C0xvv3TK-xuF9$gpTU?F_nSZ`b<0nfsqBWjl z+4N*4tJ|3p@3eZuS8)rwWcRY~-XP(l%Ft&{YG6M3nNw$z*G7`p_tKkKIh6dm6t+h) ztH%s9C4M(Oubg_r=n!^a2JI!bJZao~qj)JrxyAkDvvuIwLJos6mgg8cOw)5&Y<*4I zLf|x23c-mx1^Eo6NjgpW-~W9tVwzSy^;boMPK9Qip>ycz3-`KE)KH9%@$5Ofa=ShD zO7JwiLYLK>P*GvY8sczT%>~XJG&Z>}6MWOEwoJ_O>vRTiT)>)JaL)3Fj1vMlTOtZL znF*}O^iL!&^|e&s`54oP`LVANE}~F>b+qYIe&r~@$`KkY%|2pQob$c_Vv6M5tKtT9 zPsLlR?~~Ejg#gNl?u($28koiyok7g%q3+)hmxCYy*|!+I1nHAMR-J(X4Ym-H zN(Jxm6&zEjq0u|>!4$@{2F98lM6IR-l%DOUmbJ)QRsg!kPRTMU!QUfke3u(dr?ji5 z^TOn13GKz=nMA7#7`vI1+Imf{?AdkK)SD{Duw<+_-%-%{(T*C2WxInY9-my5F0!OJTI zCq3g=l=+}CJzhWxM6aGaJznTSuQMUOuhIK;q^1~_STCP$_j4l7>0VT<58hi>HGf2< zD(feO^s?tKw_zFFl7l(csYzx`9n^$Ztq;E}1d!xeqt;8SY32-^!`=KITe{Mn?3f=V zeo8We-(KEESG>|pN?gvp5o`O+-YTDB)0(Cqbw`Mci@TgOw2i~vJl+>BK?+XY{x&&Ps6hB}1y8gftk;y-`$=#|N^SL~IxIz^naJ%8&+ph;^ z{tRpM^?sx1nIz3t&fhA}f8i{$TGJA2G0)a)ghyEU^t7>NCmt(%9-A~j!~Dfi2yv6? zaosZ_JRTRX)?0&6TW=yBd`#K(rW$z{Z>fWt0ciybC?>)I2#xM7{YWxVjQ*0=30e4$n*PAhD0NfC0| z)D<1y?e?Jt+e0`+iqu!xzMf|zY(VDyO2ffE>rvP@W;?J-*+d1A)%P*H&=~ME6y3S( zZjEyB!G`BJKYek*zcp<1BPNF9`pvvLeCC_ZAQ$B}0i{>T<)-LHfAdOi%NSKs?Q@@& zrY4EyJ$sMvK4z?Rr$A$Eg^se3T9g#?IaRWAY_1l8vni<=iP~3cKZd4U>hyoF7!&@# z79iwf!LN0lg^fcNwF!_6N^SW)qY9~xxfDoeFdiNfvc49R(st>v*7Bb8mQ9oq>)%)B zmiN$Q@y1{D8JMrP)mNlGpais45@|w`bIueV%Nc8!CZ@Aj*WP3|l;&e6bX@*la*{@c zXG)5`;)pGbgy@OxSU_J2rYAK!GP%5uUOgA+zNljA6SkOV`ZfeaVeYD#{XI^d0g6Nc!g7Ki6`c9)XBxYZWd+8Cs_ z+91zRC0j(*!L+i8dxjO^-|UA7xo#5OLHs>f5Gyx3^))&L!fx*LlT-G*UnETtTO@6M z)yq4johQ~N6k8f<;F2xo*krL!A2_nO*VVete%Oy61swTM%`L$|ns71Vrqi@F1QIeS z<2_t9RR0uK7Dc{b1m(bsdjqJHOiz5LI2Tf!a<>~lPS-|GAO@OWlV2&A6+TsWfmg{A z%gVfw2Z-PD8khUCAKgd#ytt@h8GX;>`uZ z0-NcK6v)3T@!t(e=V(UM=ChX4M73CM2kW;$k}vsWQjLhmHVREXZg~bb+mNMA zeVO8prDW`GRZaPb(0iRgf1b>ZqCcG}A8A?mMC{s694(> zT@TTN92r!GsXrnCXYl~W%N8uNdet77?+=D|$@gmj&DJf+C^LPjw6<0W+jBn^o=2E> zXd~vsEC&A%qFuyMIbh$#uu2zl#KQ5iTkrql>8+!pUcWccBT`Z#9U}tL%utewaEJkE zslge#gkk7zBm^a;k#3QOp}Ub7hLG-%kPZo{`#HaR@Ato13ln_sp|?J@0#E zRAwgo-s7fR;SYIUv-`_40eN}nLI#RLzwy%Z0kJ!tM|}l{KtSDccP(C9hI7N9c;>G& z*{#O6Mq=T*u=V*+tcNWRq-q<#=y7xsXhc6(e3O=uDWa5_Utd`_jwYO~f*zMf42VaX zDX`biPmq+*SeD?}!Qx?KX@i>BT=A@s!BdOOmpojSUy`6O-36w`$mR8Gi!E0C<*FrK zm1U2oBa`#%T0lX8-c&2aTLS6{ptk6A^mw&7Te^1VrOtm={uAFWa&-VXXurI6aJ~!a zy{9o)COkG&+UAo$)B+lCu!{Ry6nTOh6vo97+q+oRy(PA^7a1kVxExx#m+o{JG#*-P z_sH-fYJNPH=HlFYVe#|V0z)u#k_$A?>d3};Te2TkQayWd*{{K2`;=2Mg9h*eibwXX zqXF4Msj^xQiOh=mwG@6D`e!8<0}MJO;0o-ecgrp>`K$3mU!)K8p;UUI{Z7{gHd=~yiTeBVOB)* zGN7iw$oyEnN)P)ck#+Ex9LH?g)P>xsfvX`P{b)!(&tqvS`N?&L=k-cUcG1s*{+|M_ zv)i*k$@R{)_f4!-UTvX0K&tO?Gpw%WeYku5K#Q)pkbQs%H}zi}HFfgdhDnU zx;H08AA6jb`5AEI8m8@%XL%IqkDiA4pP8))etJEU%#dfpV>6yRbgjl4g%r`)5*ex~ zw50;hX=sj5)l+umd=Jd21d)~BgnCz176DYR_9}Z8& z5Iem(*?R^mwX=1U@^ zd?ImivQ|SgQCz{)HhJIazOnM+siD0!=h@$AHca#FgGGTIYmTcw@$5NUN0Z2?^-x0K z46#gIP|+bDiQ+#AM7gxR;YwVs$0~zRN_+Da)Pb|(YZV94bn;}&nfbn6?}Rd}=t|7! z#TRQUO9ALTBQ}iv=}BKMt>M8J%9bOC3xFBR|J4fO;cgb4jHfXY;X&c~4XD~qr*P0> zz><=w6BR1`Q3#w)3fgbO#>3A?9**Z~=`54J0Y=0V)vNcH$9YH#MxE7e8}#DzyKg^18(e5M+Lzo%i(ETF>7)J`x86 zXom2WTRv6`)QZbJ7a*yR)q%L71pZxqVleMC!_cr7U3dB0vA5nzMJypS~4Sv za7R;@G;>nmlii4_iV-iY0Ke?h!!Wqk$%dm6)y&>`-*6m*xp0*K^`o3x)5PSxKVXH} zu<_7fbjY_YdITYO=)EQ}tLgenis?9E{Oa5eCkPjc2HtrvYW3=%s-8aRK)%}xv8K-= zCMI_R@Wvr+ygLM<>ob-s`jQ#bKg+y116d#pY@tefEn_=(t_a5#tv!;wX zvA!?1)5Q)v|GWC_^JH`*3&dF#7+ALdqc%Z7Kr4~d$7l!;EYbxZ+=U2S)e|_-N0WJUYH{L796xH<>w@oQ|m#;_m zq&TLbmil;L=@;K-RGUT~$%TeW+shR6Ya;soJKid22D*98;%n6tq64D@0{@FPboEf_ zeGU$0=*Tb)@u>RuV)CDhhhYGBu~0MBa!G3n()Zy|89(RtMl$ovK-Obq4E)P;7U0(s zSaHh|A_`(@BIVMWDPvIaAEJ^E=-bTEB9Y<@r~zfpU5 zT&pHi6AV!-qaU@__j^Dl7^cUI>P#J0PR2$i9HL~r8Qn@Mm&uc>f%;!W zx(Bgh_+KX=@{B6$;14&?M{(}U&DnG%X_LHXEn5E+*1V6^1!0{-A1o{$T~20<5loBk z&Cqz%n;+Qs|19%r+Zb^v<+W;#VhwA}89L^%N=k_I(WCUzw-h(B@U3BKl=5m{Sjb{f z=@_Wawr2M@H#_fR(HNmnr|e~R{_d)V(D5vf+I@#hG(sbRwj7!*)o(kTqfk)}hj~QJ zj%J-m;vs5iht}F-1O*He7w87I)CUegm}{4mf3o|UzX4tPOzHEyPLT<#o}6w) zip-=Wct9vKStDH5c@8R7FlQkrOj43r6mAgpq@_<-&WSSps>=o)LKa&mcF;v3Q|;+n zdA~#lmy>QGRmLoI%4BwN=f`o&b0&d$m%U^V&`2raz3M)Sg6wV$CCogrCH2j>JpO~0 z=#L2PpGeww$||dZP7&tUa5MZ!4TT!Hj0&G&_4Y^}mTUOdg+7=TTicp0thT>8K?d3@ z3;e$NXQ&F#4{PSLT?HV@s#FA9fUj~ntS~=nb;qmg?q6r0U1!rSx^DP*;zz?}|6B0` zMZQ>GKy2{{hn72goF5m_AN!N{h))79wGK}>aCdDcuF3Yx+GJdAzHJJ1sEkFdjzIKn zeR#8ONlLj*oUBvL&j#4gD$b@7CO2e1Q)S!Ti8^XLU*u<58&~Tt>kIJE(G^Cp+o+jXMKHf`0{cjYBY?#lh zCtisF8!oQ-m7gtlvmbC%C+1G#tI*+01MYor*w_WBN1bfB;R5Gx*lES%_cwpxGDy(D zUd$)Suv4}iTax%dV%e3MI8*-IV#7@({0(s}iZG@eE0;Q9FgkQzpCj zwEr?$d=cLG!5%uSfW_mZrF=cVHsU^fC!-KtNPG3Otm(&Q{U^yZ(~aucQ@&POI}#lK zS6FdF7?+(?!=dE;6oAu}VZNXN&vrO4OHeG)17Tn!*H3~48qOq#heCIR*b_+aFX35P z#sTd@@a?wcKM$2$5aUH2&xSd?fYR-3nrt6hK?Mo(6G7q>CRpKK)D{)ZxO#0M)Ed z=JR=NxZyeOXt&Yoq;K=Ek8tBcClH22bU&u?hQm0=zaK&gF3SODO>A=0J>9$iW*g7L zy$9>JL@pk08O(3Hsjm@-kT1-JBNI6=E^)f*EAK1{x>v#7ZnojY92k}G;s~LcaESeW zWP>PK6}a;tKZcz9v}LWu=5tWFW2p#plnfmy-AQpr6<+XYp{?XzvKhEWNHsn#UyH4v z5W0577yqN6NEudwL#$?`14iTM$ZaFeWFJ4WeQYb$-WKR;&?cokfRfXKQ|V%+u{6H! z<_o<|DdiHalc}?;-P=j}X-v)WDFlFgaRfC8l&%B|3?3?XSrF3hA1oD3xs9-9VKo() z4zB2)o%|jNPFyYkvV0p#kB4@4ql26_s!;~yqIB=5Au(1W8k-RPs*iFJu+-qw1YiX5 zLaAMUVAO?FjPt>(?x8Spz)88F%Wb2e>kCR>)td?3!i-TaE;Mbj`4Jveyt6QUT9dC6 zI(k94z~_$tP$`wNndUYkM|rlUq0aMHitI+BaxN@>72FjH*eF`jDnWHJU8!Ou68DOp$t6|V1YT;d2mzv$x7A5l1 zI>9cdYt`b|%_cy&AYWh?^n*N#A|lPNJEt(8uoaIp`x*UQ5>6Zud9gh?!B+D%%G$LI zL)Av}_L;k*haZRINhI_IZGfCx)23l-AiF24D&&*v6plyAQmztcrAxSD}anMDj=s_X#QcYcjv z8O@Fxdsta4tTg6Zd2t?~0vP?Il=Jxe=)@8=pcmx^ftjaCBJ7Oy@p5e}>|)V0GfDE5 z#!NWrz2moj<7Ij$zx5b>(z?kD`mp~DpBInkPj8JCkQia^ON$Brbpfyc(EO#o|2@ed z*^bs$rfH4`F<9pj7`4*ucoCn_9TvbL!3SiOG)JkTm|C#R9u*k%|t+MuRZl0W3^Aa9Cws!FUtPe5JgSl zwVyOcAn#R}k2#JWxXmfSHBK%_7XF?U`RVPAgtbb@fR137? z@!fGjp?+Y>{Oi;2gGN0X>Sp`(*-FEcatFWizZlnw-y%ay+A0LKaXNx9=gb?kExwtC z)*s-oY98t+{2_O8(#!IXTeEbwgimD^-|~KpDl9^^bXZz3)Bj)o%wr&Eb|BXzb!u2>vT?Z1SD^|kex4bj)JHdxgtx5zqV74d zaotb1-rMfbTMW}Gdx2(2%jT?b!6Czey*KU9B)|03?>b7A{dx#^H&j`~8qG>(5Sl|=*!d*ffAkjz%KXP*Xlo`^dV9&Ggakgz3JDavi1PbJmNQUjo7#LGSdY36BFIn(k4WN66tFH!nIS{Sm64w6Drwe>j~5mZ-(7X4 z=i64{SvTDr@NJ#^nVu?~FYNwTAl!&S*t#;LlLlv!C;z#l|4O?5SEoJl84>s-xuN0V z1Cy`QJ#8kgAAK^&fpH{@BHCPD-$fWDGl$z2VKr}U_eJ3a5qNL3vnx9iQ}GOPSqgZ)rTNZUf;WZy%!}uyTls{+ zUGeVmQawRVH`akfuy2$Lh@Ui=Z~D~ed3qn=o7V9p-%^g>v;Gj@YjN$>8UaC30N4#> zO?8%WIJmQ^adhs$adx!(=xN*(jd87Sn>U{_m z&NBSiR`aM?QKI$7b4Ru$_9FdWo{A?5E_Nm^sEPH6v~Cr&kTmXyp2$$w=;KAbq}%IEx@y@HfFU6Yw zb^^ebDab?cp;pUyjiFe0x0eEZ3LwH<0Y09$VmBkF`D>n+a{I) zp7c4hekVSRPMJPlAGyzoNV3tQ>p^Q)R}Pc6Xh48N*PYi0umJEWyY4zXEXA2QoubsV zSl?ch?K_#?T?T!`yDprEZn}bzzM)P(;E46wAk&RCWkOKQ zzNTuN*su$9(nhLTs>g#ybppS@*8kLVEbm z%lt9>WqgnI2s@_)@@X3+#f_@wUFejG5>}6oaViOY8TAPfc5jC8HM(j8H~MCN(bg+C zyjC}u@2bo8>#ppe6VKcdCK|6c`>M~iZf2Vdq_p$N0<*)P2m!3}=1_vaNtLwoEYo9h z!cC-%iHBF-a%&chcU*Z@IS|7Bmy{Jf9tsfrC8y7D9xo?uN8HSo+B(V)>7crF$b0<` z_M+cWi?mCmoGde`SJ{QZ^o1FKoWY>tCpX`&iA83GFo7auvP&-RX870C2UvRMM2NNa zUqPQN0jZ>V)zgPLkSry(hAMiu>AsrHl8L1I%RiC)_U9D|U-?5C-rYLa?k-PtSW&rt zbUoxbQrE-l?)W?4vi))Ut|cao@n4Sz64Kv?F8f(4s4ikC0iZ3((u0ZjVKkL|d9V0y zR?Bh+1F1@Xne8hkMoy_ef0p%C;kDTy=R;!&JVHf*sP^`2>*0Qfft3I~ z9KDkP|Cmyv<0FfK{rygTJbiN#MRa;g(dBIUP-{I&{D0(~O}qc0;hcdGytlCHF}*B4VL5!D63=0>DY+3k#6|q*XP@SPcZv;m=cq zWcv1+QhSSDn}0~gAk) zpI}F{3zYzVX_iJs6x@RPU^RWm;G|)P6 z(Tev#UP#Z`<$1m)J+M8UmcP+9fX4yFzRY(J16AuN0H`w@wS2*}bszlbj}1E2l^N5i znR9fgzmODVIkL64|KsUey#kaK#`gUn3~?cnc5=)nc%SIpCl1rrkHLJ|=cN*OsYUJ; zR|wG@WNuGBzOqd{aYYDJfJ1K|B=V4}seE@HG+3RX;1<>j5(qCWcZ8^E6NA4gYKeN5 zK~tv&#Jpy@v9L-8l>BLr<>BUwO2;Dtn!5 z?nabX<1MVN5CcUm9qRQN_Oo)M5B1WIW*Fb6KTSxlAad{1qjV%2(f-S{kb>&~5@;PR zw!Zu=@XI9U@@#I~s}C(0dTy2#VEm@J`PF|TX{Jm))%fuI@x4ex`LVTGYUnIQn0j)s zetXk0NoX$*5In@SPA<<^_x#^C*?qO_(RQm%qCgLWY3tSdCNcdrJb6+N&PJI*8 zuaa4YplQr`)`(jkgL0Fr%*$ULE8Xy%brk2noZNqM7xNCh`HCjsbSz~au=#A+@YU56 zCb9z=AO*vp*8N?)yW_@Rt*>W2Q4aj!=~!d3Ks*Bq73Vkfh}yyb8g@!D&n30!el=M( zM$qm6oaNKildaqQmj7@r)YQ%ACjbGjL1cq}PU^I@(1uhQ1^n2{%-ub?Ld`{?$wS%! zaez63T^;&HJmTDWzKJ^>s|r$@__9ho+t zTwhB(27x}ikR%&ED3;K|BJ{uaZ;0IWx^7Xa*e08%{(ue1B?*Lla~+50^y zi4;+b@$wb5Y6s zf(}>NJeSyl*4c~EK4$0$9O2ybLx~7F*TF>w&ymRd^|S0>eKMbi$(Xk57D=lBkdq`< z!btSeePn;17Im@Skwfd9&Jpg@9MsJ0}Fk0GD&6sp{Az%J%m8=nk<&35|6RU zgsQQk=lwrAYASFbkQFrE^8TBHi~0JZGYy#K@}}=YzKxe_n>->8vzqp_w66nw8d6*g zn0T&E@G9*4`lLJ8XzfL=Z2?p5>pQhrit*o}MlsP_!=<4L9wWOl8{|GToJb7w$&n#( ze0>E8!AT1+>|rn;i}188zh_?>Zh*>7*gB#)RQ@J)) zx;k?3X%u~BVeBl)-xz>!meLnuICFZMCQiu33q|n8jmy3hsjJ}{OfCLir8?BFzseh} zSRkT-r8p-b^Qc=EtT_!}4et0AaNUfeJB?g^y0su}O@A&cyWCj^uGu*^g~7y=%r@vP ztXI9_{H>*Fv0?-dw1`HM*#pD}FMm4)cJ;X%k_5HDw(Hp4@$xcNq)&D(WCsyrHlu10 z4#YUiy%aRS{5J6b7tW3RkA2e>8jKuNDxS?G$>hIo7d(HG>vAG8X!592cqimTYm^kC zKK+{n!9*kFteL6rR*O z!nL*o@7gIgGhboZCh9AO#V4j#09w|^u@9} z+X8I#3pVJinUYewG&4YK@(PEERwF;n z4)R2ANd%f-S`4l7I~Zferg^%LtJ%0o#XOcE@0rhRLGz}Gde#V%&>!fpJ!BENX0eHj z|FY}lLkF121p2;gNI=a$`Dk*66Tj_y3nQKP3p3r;5=$+%pE-PWCQG0~2y4V17q5;X z&w)$?f(Fq?N`o|ZLI=_CGKf>2iZz9+KzDQSY29 zi{6C)&B$d_-;`j?=^YWcG+hl z#efC}vmsw(<0ivPZ9gxs*{1_}tgsW>SA0N=*BSU0~)zC5@fhy%IHEU>GZ+ zCpb;LGW)#VbQvA7^`^LxqXe-zn8Ty25F2HQIIz81SRmOk?0Lau2LN2ThxGP#fbJ+@ zj=eho3_SmL7x$Ie`_p-C|5LD4Ec?&Io@#KurD@Kw0AX2Uy4JjSnJiLGlh;f`Al$63-)|v;4zDbym@%RFk!+awUR`1WMS+O2 zn@n<_>!Vx%tb-RmGDz;I{m%H&hUMfJN|4~Oy4FcuPb`i!;hXAk_}mvN46mj0>*dY4 znJsH0P2@nsvut7Tekg0aXIuaNe$f!w9a9oJ-5)=)WO}$n?*-@tq0Qax3P0T+Bl^y3 zi?t-jvn4eLEf?B|pa;KX)+|VuU0UclHI)M8(1Z0$%-RyOYeOq_HCT&@siL9zMaMr8 zEd{p!04g2`V=!$NHs!!~k{iS}`GTu4oMh+p~TzC^*R6*~z4T1&TiU z_}th0zPZh1<~Zs#ir=`N?LdFO@p;w5txFuJG8ItIr^OY)q@F~6$*tyj5s3T66MOBm zTxxUfR=^%O4La4;&HYWpA{?3{7^3tYSF`*`GDbL=iwd5!I*0FO*xr=!(WGoQ_Ptbz zdUOJQpn2t(M+!%3Q#rTt+d*6BOj4P+#Ps3K8gTZMXHw@M>Ih@Q-BuT)ZTDq!yU~Js z+mR)`C&{)s|9gEs^&1+f%k`L z>620A*o;=eu!+4^;+B8v>k$|y6yHo1Cn%Z{(RRza-ol02L@JqdN9hg#Go3VA->1+5 z@vl~@>v1hQgb&IgJ9whJkn7>90G5?J3{5u_vmXB`ZT>d7?;LAJLpw~von`+L;8rdF zy?ho?aj_{$uUG2U?V=VUT%`nQTS_}|O>psIw_1O#=y5aCkhHO$LH2CI*Z!nhfdN0- zvgOkzyxL+Q?xIs`#qhrm?9YUC%Vx4HKx*2+Z)Jqcw?2{4nd*@JN?EwFP{(%w2mv!jwPZ2?%tcP`%zevM<_H;Ju6gJ1)%L>B-o7@akNy72aWHr_y)~A3#D?QLM+IrQF`~cb zo8(t2lb1TR|DXpx)PloGb!N^_WA8#()d(fij9$rShVvlAf+7S`T|Rjsbi|@uXd2CG zm-p(muP$n5JClSfEzl|dj6O1?QR)&N0XRYr-YY2q6hQXubupZha%T&OpxtThn*Hg^ zG}YN}c?p#I)w5`T!6lBNw!Ben!s^u^`eC_O=J-l+F>YW`N7}XoYft-5XGB3xLmCpYQ{@zeqAh)*V zwr<1^?pWzlh*;FUuZ?(+_Btj;l&IRKGo$s<^wcYWzV6Bdr6W?T2oso(atpG0r1@z`{yu=C&Rq`bmD-k-CG^bo1s?vn7&ll$^@#BF0|tm zG>i-{C$>fA*nK9aiO*-_RbFG0VZf1_u919*b#W|pkWVd!;fm6AWf$4tC<2b@O2A_o z=#oll03(_vpZRu`uUj^t5iy+4;%K2AMCnbFX)c(78ZY&>)c;kGTbkp?(D`;!#Le(Y z6AhlQ@(IfQmvyF4BvIUWHsze{nJ<|YQ2WLyY30?*@H_6*AgG`NN;#gMFb}DB2adi&by5X&3>0bFe;y4pUGq#fpGe>Yc5j=(4qqN(T*1TGP->rBYH2dvUcG1(=P^8^B4I=8vrjI|~v& zbZqA&c<)*ao|-!`0JdJw?VFP zVNR$LlcHo0Z^585;F&3irSdmnRf8ICx*B$R_xF*9-)gE#3(V|^8%RX`xD<_M=jK&| zwFCrSNZ4W5fGZ%O){G(90_yj70e`Wht#TAV7w`W#b_K_$=M@!FHo4A|MRSkI#&1yr z!_WYjK)4mjmISa>+s?g!qab#5r3lC$IBI;;jJDS#IL66!5?Kd{6iz`oy!K`DzW$A; zG7(wbTvn;ve<534#PJa9niF1kQGU_yUl)zQX#;x^B;g$XckaMf%#8sJD_ATmQ>3!* zw5iJq5ai zjgpZfGNJsgBUJjT6WO#)xJheT%h2C(0XIe6{Cplp*PinUH|LxG%%`uwgO;VE;Q104 zxd~b_3HcTuJKtKp)uU@FT&$k8CMdm-<50pZfo154LOfZ_7{}!^6hJM7LytaG+dHRP zIfdK3>q+e8=c{jQU1W#woM^>TQ6mD!|Lj;1fF?{@*ZnIZAXZVxZ%~Y$k4t8>kOKcg zZq$d}H!|}o?;eqsmtL0EJq5lovxPy^i3vc$(0I*$S*`Ga?mP*x`es5F>qd^OgBppi zTcveu-9;NH!?(I393pO-LFJF^9%6Ck-PCA0= z^7#e1Qzl9>CXYnr2HsDvpSxcd^u|4o6)$;1jLG&n-Yo|BH-Z2vhbSuOX#THMJnZ;i z(FpQcPmq2c=gn0m$0*b^#Yl<3u$d4^jgfPNRWNK6sjb3r zEk5^(6Yu6-Q9h%m>rJ{pY0>=L_SrJ7H^+Y6g#zBV}}SelADFQLAY{e z?<&D7{NHTd=N~W=IEpqj$bpRR6s26Q-yDkC3~Hecu(W(=XR?Nsff2EW@qEeQq;TqZ z2Lcsd%BT`cEUS`?+Hjxv_l` zO+UVfq$EP;#dd0aVRKTq?jP)ws$X^sF@M*2o|4ffSw_W79@_Dxf;M!#B_Hf^qDA>D zs8;H%Qu=7tUEQDo9kD9 zKIh_juLKyV(C76AmGH8ce}`MJ*J|<82#pL$ayR^py*aUY?BSqZm*wmc&^P%%7}p6|;x=zRS)-Z^7MFI3YU59fljn(I2*%H`m( zOM*aS+ui0iXj-VT5s=m`C0ujTFEX!9!LAegAF<3BlFt100%3$gHJ1!3Y@+NK_{;Tp z4_7|j3UZ-t&Hk<$3KYob>RI*iLjJGI5<^pYCWHU`NC0Si?`~ZCf;9&Yytm4LJRE_B zrlrlhPN1X+N<3fdvgB6b>+59u^*H9e7zI~HI$7lW+(~J9nCwF{EgJN~MIh`nvToku zq}(kv;E=IjsB-xI*}WlGF~jsf)Ap>7+bWJX&~*Gncc+~R{qaF?{MMwQY*c(Qwc=QU zsaX!`!_^rQT9Hic>d8~`INIm)NTbsgYX$f>!Ux8C?)*A?`L*u@;+$LFE-XFzN4>sR zEN1q%&v12Wp4Z?Mv;3oh*4sxBvhn7tL}}_JY85C)Nwu$A3=_ftVhW=Ms^M}Hs$_gS zkct*HPCte9V&un|P3|VXu^IWq_oF99JhwRSc0DCu+J_ufoEM($G1Tq|gcEHlQ5xpd zyr-b;fy+o~5$d8RlMLV}$|f4c3E(e?bYC&cXd17sW*b3;Sj*oIu*%Vy5Pn`y(NWGR zVZg?P_>cC$tixJ(ey9Mca5=v)NCM2(yde8{LIT4YPN=5Sf#8Z9%dtD^@Dor2a{Vl>Rl-vMr2U=-2q5}H&fL7_v z_QkJ4L!>s{YbV~xNvQT8x}k+^BWTh>coa&c-0%Y!Ta|BZ_R?sb5K@vVJTwM{mO;E1 z$@_!E?3h?gtVA|ebJkY5to{hBt!GTXieDCUGCDrI==XKgp=jXCG9WXOXp_aVo`I*D zq+-an0K<#bfz;@E(Brh*l?AMr1_Qx?P02vb{p2JQqC41hWE$IdhTvQ5=&Z$XWWiD3 z?tdpG>MAdN>wEvyuGB_URpgaYL3PKD7qWVOVublTg;c+?9UcFgYK$h8BP_Rbzx=zC z5~ld5xuo;!miLhG4E^O^i>2nKT)!wOnA>P0$vB}HmQvpFGB4Zy{~Dc^um9gup(4e_ z#KbS2GVS{CqHOKJCFBFpo*2J=dc4B5IcQS}uDZ+=PknzgAIyEB!Gr%aW-;o$yfF%J z6oVs7-chiqopW*rWGaGygpy?J)QpteMr$3@I4Ssm2`rvMGh9v5uan|7vZOgsP)qab z^?=R%d-QHbqXW8n&ANJJv$(gm_B4W#wqPF~QU?e`|KYon)I?N{i)~*@-!$xlxv!?1 z%@5KO0AWafvTwd=cFpGTaXJzw4G6joR7p<+jcD;C;iN%)i) z`R1gQO+{N$Wu@srj1Tc3+VZ)6-1>S#x6<$kLVZOP3taASxY>MAOVD1r`ACxRyF+ow zjH&b1mDhRt-R#u^GGjoD@)PBupf?zo$`}1Yt+?naajdNmG+pbPOMM?c#@-)Ogd3u6gL=vBe(R)p9 zFzIWp%vR3F4-+@E9T;!>x(lnldqg7Yy~Vp$Y}30cyl#<6xkYU`neadgd%m>Yl(L-1 z)gV0}ZuHakEQnYl4}u&;HKzA`BUUe2UUu7&V|G-PR3W$e)!2KtWxWWt-E(K$y8pk{ zqEY;R-ThmTnu~%H21ER^06dXR*%S!`oOt!Q-+$W=xgsAJOkI}FJTZA!-8xsuDBPu$ zkR-|YO_hc^LN*YtJA3y=eZ($fg@@jST-dVz26=&{_RGZSo8#k< ze!%Y;m{%XPR#{4N^Dr?DH~bW7tfg+#t~>BTUenzEd*T^~elz6qnG5@s6E7528Zz?O zpHiO}^NdNYQJQC`)^DnL8T6bJo#14^0!bVB&MWK)0c@^JK<4z^5UelvTvtL+$*vvK zQ*}BJWahWe783@e3^hU;>|%_e{G^fzQyX4A$^FDc?C|}MY;dLYUm%V3iS6jP7A``_ zM~E6C+5|n#j@2gXz-`rU#xN1%Boc<=;*tO9Kg$_HT5#eHW5(5}Wu)46O)g>T=kFW7CdM?#Ij2PbX=D5a1qfC2y+t=Ibt6@idCx6s+UksR6lKnlY>u07O9QQ*rhhJTJ&6 z8?|ow0zv=}#6;>K?Y!vO=d&jm=piassRG-o-Ow`ip!+4uo~rFS10ho1_lO zAqQe25-3=DH0~B+6GbGnah9{^+dPN6`u-J49pS19IGJCQ+ZfAxw7y{G$8LhEoox$? z|F010{;KdM!hN2_@u@2EEiW|gO7eBm%@r5N%@EbZ@ztL5;qa4bcgDKuwEMee{j{6i z3>J?Vt;{Ixsp|)~N!M|_f&(g{=!5~*)v5O#`}<$`%Snv>UVD8PPc4khjV!e_qc%_( zcaSLcXz>{QaPv!%%v`ekxS2`oH*J!O}X#dzUe-r}AA$K1n{6)v*`m7wh2 z$+9ICX_aWo@i`0gQ9o|pDW`wwI3~DE>3FBE`6F`L9>?dKY07Yey!TI`W<74<>wQIY zT^GyXL0<~oNLuzXq|)^H+cybXRO}&jh5I(3q1@f;GdaeBY=tjW;Vr^|DF#PLrZcLf zJ1Xi+RNcVgGer!AO*Cf@gC((CrzFdwqg@=fwO_0v zU46>sGA;5G73e7Dg#?MCgoTrIC&*sI{yxRI@Qrx9FM6W@Cqxb2`vjPD!^m7B z!ramUW2meLqofAzFkg4Q*Y0A*ka!%zpbtj%y|=f57wXDzA&#Uz+>z0_L!n zXM;o;nuDEemgztDG_=9of`CY{6HNkwo4*XN8P`et_>*YD^(`G$0qc+o2~~+*TC{_c zaIs(s12h|EO`~}J{7*3_h7;sw)_2+%wFlm#q85+=?u^!}wbarc2BR2q> zy+^4%*X>LURIAWEHqoi%5=Lcw7)7gNhf>9P>98(rG44DZSLt3$s{H8$1tO25Da-9@05b`OeHSA|f18BFwNA|*wCS*TIoI#^ zf+Iq>5WVUpx5ld3LLZrn(`Aa~JCdgI2)lcrc8h!ba6&jh>taSl*XIAlmj!_90lrgD z#AjuyA6LEk@l*ZAttKyK!gQFc{HiyR+}c3xf{Ye4%mbTaD%1WpbIRz;LL!UZoGQ>g zkkq~LlD5nf>rd!wtOna@pTyKpZ{unoJXHW;wA&c?M`cauf{Kf)3bMCU{2SL@e44u_ z{~Yxm>ea3t$n>-Zf6_D%1h~@5!T;zJ&6Qh_GZ+{q{`G2cdaATR3wpp5xS?g-B0%cd z6Fub@QN)ExGD*`5ixrJO(U<#wMCbh;YZH-u^PuCv>xK{i&Hm`s*ESRpetm6zlG`B+ z@%M5G?p{qVfBeku!uZMCr~UHJ8v4M37@sx);ZXv)KPZ_LpIAaF zRpZO&Ez-gfS|)%8A_{|QYEQi=qK-D#oJwWmt%95i&%=Ru2Rm$!wBk)iePh)a0Nat| zT(!4nYHiYAoI6ilvt)F13gE(xp0Lwn)!!#rMsu&66!*JNulB(%YH?OwWT)S^&adr= zZ|=6z16D473bN)4I}>aFAEMqeD$20iAAVFox*H^vZjcyY1Oe%C=%Ks28>Jj~6J9l4VM+AG0#s|t9>&8Kk!4U$FLk-t0kx5zI~X z>p)ICm|Um*HgF__A0;Tr47D<num5`?n1Dk7ZAzMxHvy*Wz)-7&cnYZc}S#i-&rP z^+rd?>xo3+cxI9rA-~Y&FWBG0b6@?beUTxp>-=P_0N?sk)CoVhN=?=;Q}3%haa=3@ zPZA-S>*Q0?jd$_j82>i+OI|=44RdN|0*cEk#S7QFk~QNmXjkX1rJVT&uSZ^*T^6Cr zo?*@>G^T{n=W&qW$q2(uhSyEH4&Z`0`g@NIQ5QhSIyI>x_7?xRb#@c~K$z0{tk1=| zj&epedz(T2^i4-TqqrA%r2Fb22N(0vPAhm}T$?(eTE9;s_HPOw;jzVjYsy4BXJk+WHV5Pz)at(Jb1Ca;Y_eial%Du6&X z;SQ7_wVJt_ruMZ_4D8(3@-VWFZG}J9^49VkW2SWuHH74$mOV8PN$$!&yKHR9u21qx zDs?d2>eEpmWdB((+sLz!{ft+8)smQLeZuRz)F-dVJmE6w`R+Eu(9TEwz&VTvn3U!g}S5 zJz!dtsijOz^IMIMhVn_~c-r_+w++QUslm^E5x|LML*SNcpV|vacEYqzQoB(MlFP6B z1$CMB^W>rIY=u0rMZb6Zc~(z;ui(WOm7@xzV7zL`-<-dW`?AeXKvV--o$b%$Q$+{I zv%l-~ocY1!nh@+(C5Jiint5cr<92R**WFMlSQ-+lwj9}O0Z$t;(ybroS9xBlm|Dbey@ynqIl6erj3^J?Ey zqItlNIh47-{zd34EO|olmi;S$i}CT{Fw&2iQK;UJ{C0Ewu7&((;pag>?Su^+$DFn` z)9t@DBOv?b;%`FVS}x%69YS&%@?bcAbj2FA_0Wdrh=2}}@m}T7j9~y&;G^C8%eC^2 z+5YFtTJ{wp`bJe^e;m$doI%Lz_z3mGoFs1q^3e8rph?v(V+_2ls%k1Tk(PK8V`LD} z;dq|&JbRk}1-Yh|Vqp%w{Ga33#q+B=@Q<51dAL*1CB#i zYE7MW#OKdeqs-fpls6VJQuU&oQtPOEXf?&d3oq2g)q#;TNL1Sg5^es+$O~7+R=ja9Fl%+djMY z9yY>tOUgcO$#`%hQI}Zfd5oZXA?ejykeJlXViH7brblRl&Q{1WP&D;eK+-=YkQ3i9 zFRn0bn4oQ*-0K(n37#ZlXNXP~&0xhJ4>gYhBp%FD1@#onn50^bb8yIk#P8e(aJ57( zm$755Qb=8W_d2A9VR$E~Z|^Pos~4oXeWGtP}7z8Kb2o0&d@`W9ruD9;L;YSu2t~O~ToMgo|=@%cH#ZmE;J6 zyYl0yxtzh)&cKh!l{B=F>`p7HckTAy@8>5n0|yJJ1SMVAOTLq#Mq^VGjko@MQW?{M z5bLO|%rRNVsv&n;-jKi$vxmCd1A+l6n(To7qpM-Pm{j*L0e>bWqz8V*2nIz7f455V zSReNKu>#`z7~ao~xIB3kT_E4ZqVUB{8Y(2MU-S%NF^Xd^=qvWwjy^P(N@S>dW_wm7 zrL&^qf~Ac2N*2S#u<9MUGshPP%0K>Y)LsFJ%zccsIM}ecqRC+0WF<=iX5A3M--Cqn zqvzlqxDE(`JjfC}q!=T!`nE+9hk<7iqu+Ibh4bd2o7F20QxJ_lWO_ednjVGyRcgJH%ad_9 zNC9XM2#4qN>RbaHgBbSzlQE3}hRD!3F9O2{x~_lB88y)t6sgrB{BjCyS%Tfbj`SF#-A!6`>4;(^UyEV4OSbjDY1Wl|$#-Dt4gL;`=2X*X{NDi;y^x6f~$MvWwz zbe(|TaPlYHDpWl011@*1V944vHFBPrvO9PJnu=p9feR{w3OuvFOfGovdjkiLk?Xe*G*-LmU zIj{!#z7_R2lfFqZ{lF>#-`QZ`B%n-v6^BQe0x2U;6kC?!JrC)e5Oi9JBMghZZC3aR4-Mwm9>J?36NEb z1J8AK1!oHJpK9S$8#wnNDV(^T7$lzXarzRTipiBc?G$t14YF_?uYj9b8}Dkr|0rhh6gOLZsu1Ew#ei6wZn!g&i*>2px0GVZJe4X+GAKdGoXuX zY$H$YtN!VU07(o8H0<4(07rQ^caw$X1FezcLpTbT*9IC2sfnYc5bKQRa@ei4GO?yJ zUCdyFR}>%w#;rfxtE})0y7oq|EeLVNtC&8vlAU?$yIHiIDbT#(7m~^P$>i)n6UWy2 zb8W0Yn2KtZDvXfc*o$*xu-J?yUJZeGN#?2N@gl=<6y4AQXqrqpZ!goo%s77K}*-v~I_51JD1eLs_q-xDhOr=-3>0YD<3r#j>-Z=3O9{hohrL*Ys76e%2j3 z^eYVJUA=NSK4Uj{%&CyM?=$RWP}r&SSF!;l&ywUVknM*d7$P!Z%X@GfWmzkE&gVe6P=vt%8-UL6GMX~*_Sd~xj6FiBRR2(|us$W%Rm6+T@|ENpM9Z2Z*#kxGtYA<1w@Bl>tDf)+u< zZVv1B6!4nH_trhW^AAg8Hlg_Edbq>#iNgNhev<>e2??9mWDS&V1<6Rfp2V!K5T=U*^9uGWku?VmBcK; zhI(B{iDt!_3#=KvXcz)>&gPK*NM2F6x>qNOv#6ds5l$-xi^*S`TkWp>Ge_QB z&x@6D{a4A$p>8RA79W8Duo~YvTzyyPouZO}K_{92Bq$?1E{&AB4gcBztYy-_NJag= zd=1KZk&ei(B#Z&7E5f9b8o$0UAxFr1zHJB?9ZG{~wDsrr);jV+orpI%Fp*Pm8hJxE z5c}?RkKUsCIFAl6FYW^dkN{j>wqP(wW|BsJr=8X=|H+Ht5F)$2M`D$WzcQBqw9-mV zN7g-?_Ug@#hKYIXQpOHW@IiO$g{hIbN{v->@4Pza8UXa9tUF0AOVQC0G{iCSA*_Fk z&iXa_e`dZ}Yr286s;2G6@EGPp{-drCewGyC^Tm(Y&8_TWu62$DzGE+c%lT46nY+?j z6it;$nIXi#`$qB|?f0O^C)gKd9W13TH}pA4^rY#Eiul1=VWZFY(94um*)!> zQ*>~})tkdtJ=A^aS;?e1`hfz1AMD6wS-m1KY?|n}Npg+^hdt-{OJM1JvJ_wVcI#f~ zK2qEnj{a%{F}*B6=S6<0iK;H@|4*? zBUw8t-|+*&tCLqbs@lII%)5yH+Pg(LNx3NqEN)=#e3Tf1!@Nyai)t zIA$!FG2x8VKuNg*iYB~1l64ld<1AO6LVjIxI>s(=?>-g8cu zN0RuJRN24qHKgKC!QX&dY74%ajOEYB+hS++5(}Qi?e@y{F~Ja+<$c!ur7|8 zwe5E#9A&2pyi82vXK!q>J-OK&WR;&^7CzxBV|2 znEn5|4*G28SGBc88L4EKBVlP(yKqU1!SeuUm}Il{1U<55L=AF13Hvm`QJCvGF_njzv`F5E0+-{8YD6h*ify zl)*j8u={K6b1Hi5-kKzi)cXg=oTr1Cc`MQ&&AP4l_+o&l@fPIdRgQ){ zFZG;-UR7d;K(%;c+$?R7G6^moZD$D?GgSAhd}1yhFpebE<+G3!@;Q${DzB*r zlTfk;ALS(>8t1{5My&8z+G6SvzAc;eR{p|YUJN0lmSsn64Uj`a>$8mgcS>dh>n^>i zsu{r@qcxRE>yvo~d~tkx&0BjvH!ih)9?(|ql#~!J{_G_ncnUm9g9k>}Wy{Xj7*a{) z*KIi`;4@RuI)VJ}iDKS6K4H)*B%OF>xK7BlabP}MO9CW1#hdY(4};Jh{H|DhDRS&C~E zg#6WP=7UYYH$jb1REg%By?@3;3jbCUR}TT8pkcact|_%ubUq+{QDT5!9+dXOe9obF z21te}ktg1t8QHYvaoG=gpOG=5?&pD12kpOU4yk+0wLe@wlqI2^AY;EC=`dKxB&BS- zVpH-57mK|r8&Q;or{AmZ`{82;$~AvR#$J4qrg3?4?!7?brtn9O2Qz4wUp_}6lroRm z=FK)!XH6SYS5p6I6Fn;E!l9fX0Bvmh*(K-?j6x~4qJS0xbt-_vtY!4Xn{SZSwzDtw zw%j8CrVC*UV?{}^p$jz*A{KHyOr)?25nMV$e>wQ0+$38|qpm9n?d-=!|4GbZd%soc zTx~$c>|d(SGj$9d3kM0>M29&afDvlGGeJF@iC&{osNz0{TZA@3-5JyoLZ61@Cm38mT?N6AuNwTv*V1&&^x$FS-%w|h4y4U zL_Wy(@=px$DW{X%&p!}+wyGKPnEl&qkLRzi*d9qMOwL^ydiylnQgaAG@Gl-cr9}+= z7UoSv0e#je)zDPSI^5g46fQfU3exf8G@5U(OkLb>jF|A4EKK;|g2uZE{$M^tFi$35 zB}YK_EVAbE2p0AH_lMWte*3R2jts&DqPZ&svlkm^%PgmU)o9)ppnw9JANkLC^5&

Yy3S}}&r4CiK~<#{y!A>;5HNcnNJ{}3*o#n~LQ znot%VBF;NP*VCQ&CYn8YbsA&`u_-PnE)-n1!V9Q12pKDhq0Fi=?+eb?LcP8ens7+m zU7GJcRUoIsd>3$-*+1=BOAtV#XD?40^K_=^BQIN0J%Mjw0k&d0kAS+;|t_; znUcQ6@BQk%_tyLs(R*p|df;CHAHIj%{vW_cfZqW={vK}Av5xi5`t;L@PyCPHLKX)| zc$B7j_U_|WZ<6M~`tlQu%U7`c9H0B;AI?NP)}8wKe>f+__u(h|OizC0p6?gi4bAKd zn(xD|JxVoa|6=rqe=GAde*pIFkJ0V^2;D4Fd*H89RKuhWH+|;QT%o23(sE-L`YRS$PI} zjb`TpHqW3>LK*q^Z@hk6a@p(#Y76q}edI91`TKA_6DHS?W^>CpHuIEWPVVnR${toO zfL(;)JZa2m`9O7a%J5@Wh-=&o6jLUD}#xL``zkFwzT;K_s$%9OHzWa`)i)q*c znc+f@_$7p0@^BJAS)${VaeNTvG@%`7cCO!c(QTr)My$mt5^mt5QpN|cGT>(earhu* zqcrnW<^ z5tq@@pt(TFxGvzPrLsqT!gniNGo|!_G8D?#U}=h|qDB)9=Zv&+r0qDf=6cgG8}XB#w3y)=B}^{h+YLHq#Ba#`KDnQg z4&O(ZL6}6Kt?rvwTv(HbLxd?MH#px=`T{=TL&CHHPEyhyX}QLAH!wSas8C8r863W? z@lr!@&Oks@APs`3;k$~uFauSoDNAoPy?o2yv})Y8iVILmM0<~mv2tBeq%mTqj>Jbp zG9a~}&rE@W6{*Ia)N&>$;Oe4+cLLflE>7YnBTcu!yb>mfys?U?aK0u>qvE`P7ks!` z7UrNhhJ`S$1VX`+2$M{lX3E;5#Q_n@6v75o!N-K<7ON4|n%v!<&r4})D%Yf*Tef`J z;v8NK9Ua<7Mgq=xOmgiZkP$>Fd6zu;EqmADx{B~XD%-cdE4p1kP+YA4UB>#Hwl#k( zS5|S#f_J5IiJjmZUrpDcnsAyNE`o1w7PVcoScA(_xxT)7%LX+GtDs40O5at~ zhPXm#L9?T5BHk4ch6%^{_kgFQmAkJ9P0^#1d-x#E^BG1 z43ef|N>Z$oNqq~Jh}|L0ydbd$3VCw^^NNKj+P{y!cfd(}Tob2PaN#)XWS^mieiW|R+*ua7+S%pcq}$l&7RH1h{&CX078cWft?HO1dw=b!(#$9RTg9qX@}KKJzNz#s9{ zGgI8^B*}NE6TJ0w`5M3dtH(IZx6pfO@cw61b9Vv!W8nV+{?0OADj%ZRdHF4qk!iV0fA~Jq>I%c^Iez-JXZhWK&D;A|K`ETY zw+DpTm3Li!*B)FVFMfm}oCm(f56a@boA#djAYu1$x>=+<{~YTJAGn=4=6k~Q!gidq zL1RF|E@jhTY>2xT-blDQ#W7+s*ixQSR zi>vP^Z`O>f7x?MR-{lK$`uorR<~f`_jQ2ZW9pW<52<`0RtpXg^T&v@@8W6XKCP&E| zH1Fb?5%Eiej?|wZPT%0KmDEyl#B70UD=xqeUMA;G}0Jh=N^(Lxv~aL2Aoeg=O}rKYNU*(K~Ld*!B6gXTD!+LU2saShqof8KJh?Ow zc8C*K6Px1D;>lwjmA7IPsjcWLLTF&pDQV>}QJQXnmIgP@DGp3Vn)!g{U9=?P&g*rw zS`=H&>KO9W@QuT_8NYoWM4|W=H#wr+eU5Z=A0;=WRfg$+rjb%+pd;;xi*);kXx^!% zP^4<6Yrx4!9(O3|fH-{#rmv!Q9_^Rp{4cR)Lt~$yxcd=3NnSsM?+Qids71UVa84^K zql9fKSU{wZ1}Xw1Dxrh4c9gO;OS5d*i_}7-0OEyqXT(o8*KE-hrJnPAK>I+>4wokNv1CN7L9@rl zfcaW}R!8KkX_r^7rnM*Aac{WAVRc)B&qQzG18TXL)#4>r3){qP)4C&-S?McJ7iZg) zcr|r(thfUgGm-@-H6g8KK>PyV1S}naKZ)r+x(N(v7vGI_bUz226U^5VuX@hfn!D6v zRVq4^JVF^19Ws7C(#%IpI`Fb(L3qp>uxd0_q=s%;!3DE`Sg3}WR{PpSLHb&zRtml| z%Fv=3$wP~G725y>b$~R8`1)QJLGw2G?gd;$ecUX3th`*UJy+YpKfEDQ`NyUAG#Jgl>ax zMl7`$5i3nS-bA--lu*mchC1ofzm5E4Nps>= za(;|{{S=*?C!Tx&-v#{SHJtA$iw`hn!&#!}32faU7V*;_Egtb6A9}3#YJ8Vwn-uos zu>lfsI-t3$MTx!#iM7Oa?UCjRA#N}16`I`(*wra{^M1%35+~aPzF?bK{Va>8WCsSC zJZc5!2Ex4ZpPjaqTdabFC9%6qUf&PeRQ`AKZNl_Ti_cE55+Qtx#(~O@q534`TH1ya z-~_Qf#JPw`#K#5rv3|dqr{tzCg1G^l)s(t}v^qfbEY0p32|^yv(desh`@46te+6=f zrAO!wF9Lsl`wdKJe*FPLJcHl;7MmNV(cIz8V`O~0Ax;nPT1MOMzfNAAhj1THJu`h* zQ*gt@D|F{xM`sV#3778Plep>Hd33mfH19mEcdTO_>o2eBIp;Vy_b&8a8oa*@{2uUg zz|9{F|M+cgfAfp;?{jSM9_!x({nW)T@x}9rk~c6u$hf+n(A>axsLT0uEHKM?hxX;AH{Hz5ub=*m%sa)E&tPh{ueXu_0%&{#6L=$o~1kY zwJpocBjUEnoi}fU0zJ+{6PJwoAgMn^S)Ihq-bgm6FMy})2i^^`raSRv#v2_Qdy*_4 zqMcl%*|~cA4|a9*M+cdn?JqnnEkuI*>TlAN_*n{gzhU0i=RN`gG?q@ zZkf9PJ}k(2in4?EJyv!|t9>-jXs55?rmv&J3>^-jyIxvrCNSAyP8Y{j@vRiL=AomM?sFMA@7K?T80oCkz)* z|00{^lcb{s=z=hxf{cXN5T`GYHt(mDb2tr%d*k>KKM?!?=J2^?w9_DmgywrtCJd{F zJbsVpUjDv)BZOzpq0Tj*kQ5@+3tX za5HS2VX}ks&DOw4xFD3&;&9a;U9g1s8{izG zJt3?iE-7QAXoCyNWbaC4#11o@Q^YG~69fltL6|IY;RbmuXc}?!;1Z@rm|mydIjqK& zUy}1Fv~-y6aSMp6yW~`H0HJ$*n=b8YQJg%B>cceqE3Drqj}FxbKn4OHLO>=pv1%2D zx*azbtSI1cT{X}&3qC4-Qp?vw3}5PfnrThSNdV`!v+zFSJ3-lnaX`3%VU*3BFw0<5 zUz7A(E@P-O z4X90Pk=W?b9B|PSr(3Raa_B&v{1BzA8+24`-|*s*0C}_J@eYudvPDMJT9@8l$PY^s1>ibg^=SJde zMd=eI*8)oEj6CeoAI_Pa`wpRf}p}VuR z_08(`E|B{lK{oH>XMXb={NbB^=YHbX_F;OSVkg;L`Ei_|;iN&u)vkhJU_4q9+DnAJ z3&b}qB0G%GUdMO)$Z!Vu?mZnR0Dt(Y%qKtl9C>+~rF)c}^WVIufX2AoBc(ltqpy(r zJEyCTb*y6@>sW88_tN0~Dd7FQVeY<@Zq}&&`|r)>9P3!`Qq}bODnIvYml(>ExNwf# z?%~`769FB`o9|-jGOEw<`CqxfUt}`$Shq{1LoB{y!sIRGN5n-|eGB}>7`&eI_}OV( zcXit|4)}zT(H!u;y!CzCYiK$_=}!{hWbjIfm?W-l-rei`t^c&}sZagieDVB(aXiAN z=Shb@f;0)|6D0>sTk^0VuU?=GFEY9GB|iVD+n?uJeu8G_LE`j~$=<7shfh+{NLo#R zg6~#{GsKNJzraeM^iwoP@@5a;tZ|cTgmy{F4XP8$Fh!aTWpy4a8@#(fH^0oAf98kg z8nNAHNT-2ns_wqEp7tqu(9G#-wefu%Lnn(-Pe3AQUn!kzcoT6=aYC1Y()b?re z1C=X$_$Z(F%(wW$FFbo2rlQrhCaIcxG}&E(T<~!~OAk6l@RU5qxvr+|+|~fNi_9FB zLft>zwiKsb;oC$16jplO?N&+X2!7hUds)`4@?=@{CHJ~^D) zHpr@u#piK`W>LksZ4j=r>?&4_;aijs*VD9CzP^v^6C~zz{s>}E5@M7pdJbobD)JXHW&~ZsbOL*I9JgaB`V@3d}~ar z88&&wThQ(sb^2bb<1tcQxhmWJerLO?@IDQWl3m3;SjGoeZ+%ozgEDNL)1mTuMXFQR zd~GgZwHcB&HPlUfMXNV%of` zu3@i8lPgL?6p0fi0qtE4#@jNdK{4?)Skz@qA|j!>(4FEJYl6)yS6B;+lB_0>K;y~^Q&K3xJH<(T?kjjR$Yz$>xfw?gdl8#G`?|ywZ1>PILwRiCNcW?hb_%3YAS)J9lUjO2k zuL0Ni$=_ZOHs=Y_VOj7xFr{03=2zarRd`n4krJ0NQA+mT(wWQ5kyUB(o3m8k(a+>O zKI{?auZ)m+zaggq%QI--_`=k+8|d#sr`_rn^H!N-=FoD?=YHv}IlA9icbXC({MD~9 zP47a&B?3!a2#96Mw86?XOrGKQfBBn*(vN-0Gpz0-w&&^gzd)Mi_}K~vtKNvxbk9>TN#YbtD_p>nM2# ztfeH-6!A@4#T6$Mza6PpU-_n?zOQvYoo#3jgs^pFdLa!hVJ`T^5t5dm9cAhu7jQi~ zw&ZCC%llP)3I(i1yu+CzkFFvCXuwVpD`=WQsqSDbqQ#NNk!G3E;_>Ryv65?8Nswf0MOE}-5IUteERVh%;jt8}jHVV~20+_LiZ=Tmqc|4H7D@LQ{)Qg3LT>87Wp#87Q&S z+N>5&Ng0uX=o}XuJ}9uRNS(>!%7p1eP98N+oGYP8_3+9KSsl(NFhwMzBjj<1vT?W; ze9WMI72JwrTrCg;->^kC7)3+fZ{Com)b|6}IiKinHcY)^xM?)IiWb4ginoYPP&G6e zYKH3ca#PB1SdLgxTr*%qJOyDO4|5QYc&Scb41BJr4|6rFD+%MuS0Osuc3)OxLq#u0 zO_8J2^=+P_6`A3igqDb9Vd^8guEMsu6}}yiO^f%(Rk&x3$q1o3*PQ{r-3B0TIm8&Q zN%fqQT1}9LX z8+Bf8rFQfW-KJ#BZ_l|$x z_o9CGe?CX*A7+?cq?KJH#@d9|k&pi7ZOU&vhi;zyTO(F~>{Fi2jr;M<-F%J^-Zq`ny@*O@a3hfIZ;NJC3yk`0uy>{_J;ci_Yq-zMcAC z|Fbo42KxVAl`lPJTjBH_?guRb)}ahszy6&|<5lS}-x6l8h*WWYM7gzM+FT?okKhfi zj|8jkC>N=Avy7fzz|u{OtKZgk)uzMe0pWwc+7K5Fhc59y{lfUBPJ(>%9DccvYgW}I zx{f&Ug!wJn*%9OV969a4*inNt0iLJm4a62We-9=t)A|Y$dNg-vo}s$NyQh%&67z$n zUpKJKhdS1J=LX_qpYjarb65<__OFv>D>gR=e1D!koyUb(lgiD&0-6HK3FNOK*3!)r zCi}$Mi?q8pzxnsI3E#dz7;b?*j1Aiw;_DV2D}cGGXJLMzFx)I3Df$7Vi$?C z6NckWpn<%j-_2-by9Hqy}xteBgJ2gNn4v5GK31mi@0@xwGYpdYcY_M_3E637yUzy?Jfb|WgBR=k6@!zA}zeZ>pTs+1# zJ+4!l<`%NNP8!!t>&v8*U6^K+Xh;Mng|rEnf^$-7TnAN3W)p;vk2yG&?slHwus`JVwHHS#tj6tg-wW?Pr1y6yfiI4Km1)xmlv zpfNk76}(q00iilUtr!KZOidAdqvYOGM$lZ4a0|^FO74+yAJat*YzrX1;H>fF~VZo1BJ z8FaXVXV~sL*&Z*!2=lSxFj}qLcu_nK*-1>RP#yj*p@42v1jPq@tj>Rvia`*m&;Xp0 z`+!i_SD1ko!&C{OkSC7|rHYkPdAr19FRLq-YDiGK(ctyo%+Dzc8(Mh%$nzfm95O4l~p|5_=p< zsii-v>rAAc$1y?t@LJLD4Jwc4D0&~J7clWi80oqRVgavs5|&4rW~3yA^<}&VbxPS} zgpSeXAhB}2T|vCyBE&X=HKbn9(Y6OW2p-EF*nqhn84fwU`hIrq`!{(1%4eyWC!Ick zZwum{K)hh=5NFRJ{wi8_5Z~5Ny3SW@Nm%1zj}AL%aVRTPg?zFPIb*WH$BC5hVmi$r zUnI0ov2*FGlhWe6O1<=MrKzL2bvZz7055~)4z4h@BaLA@5q(TgYbHwE7R0r>BzEZ#Re`Fh}=0w3D`|F^%3+jLfEbyj~@E9L9B;g~X< zbOHmBRXE^PagZ`5Q%TKssJQ)TzGpr4f9htxsIhn#-jt;-P>>S1E@T|B3SO2 z))z=8e}yYw{Z4J+)-kYk4gCCDJ$jD!Gkm)pFWtJJ!~|(3y5>a;Ax~Yskp?Ve7GYr8 zEE#S`NV{k}M%bXG#q0#(Dca@daormyo)&E|O~^JK`*`S>S&O)F8)oB(?QPP!L019s zT}?d?J>HF&Z3yu;x&Rv_QVbtoq~rtC?SW2|v?lo_%{<`3L@9|fHLnqPHMLBu^OZ~7J`18u>-!No zL>hpM=m{nCbnQ#b&R-?D zv~_ZzOk0+!hv`%(Qw;<2J0q4BOeeGI@HRy)<7|OtLtTM9V$LJ(&RD-9RvqJ^$#uQ> zDyTW-*6rajUI zG%LbLo~%04n-Ml0=$hHymoST9mnkXYVgm%q=xcDE2@*EAa054Q(dlp-dS+XENsAar zBQ&!O&|y|L+$u#I6G3xj(5B>Yp$fK52jk>PYa&vA(` zS4wWt@f0Zy=4XVapxgDT>JW2S*sc?i%IHni;2Jw67lZ7;3q@=ATuJp93bCN2Y~A*m zG(_mPIF@vTYraCWyhRz$Q#K27{~)e`&`spzC{v3IEzUVa0;O!5R_;~oUjXqq=csKq zt;h)!%#r#%jD+(mR0J(E^58&c7!6JuiaU=^b5w5EvHZMC(;OqO^BE$@%VBV0Mw!l2 z(gRq20I@mZCYqguwFkffF8B4^Vl$kMh}+=YNGfw+RvR!SBBio}aUStw#lN%#!&8P1 z&9MgbML;GjtZ}l2a)`AK4Uy6xvVQJ|=?=ewZ&#ol8#G_U`U6~BaF)?wPQ6A7INdM}OH4O7j&R+kir~&6HX)qS?L7^mq?_M|`3H&f zOuPRy!|jJ~{uGZx?KX}hluNjF7xI2}u$L9h>>3hZH4$UrcB zCZfDO=cko`x=cy+vp$ZGGGE?DSiVpL5c9on{Muaf@o&-l!T-o7KX3YA~rCx82^bek&-dJ*Slh((kE(Gy&Fi64LV`Inue zAO5v_KpsW7fM4zrVim3_6AUN()F-YJ(&PN@KY5P7X`%PA#~gol*S|x%{65pez-t5TJq7~ z{4BrspTF>J|L(IotF!tWtM8`3`$ejI_kRKYYpR>(2q7{4Q?O zS)J8c{avlah3ll#m&n5v(t3xucw@2f)ulxkj;^4aBfJJa`s+Uip{>Y=(i8ML{@~iT z)Fl@NOh$BQaXYVwcHQ2y3^(>DtNY0zSJC$y%2Y4{XvWO5y7@5E=4+Jl<#&`){GBX~ zfgld3uSL2EJ(G41|J=$RuLu{bT^nEJcDl%r_cQezBxy;e2KET@U||B zeH>}#&!fW?B)tdtEPwLz10VkN7m5A-j7R5dO7&Y*vwVro(IMsJ5@?TiU!~+FUqCam0Az zA+Yldn@G3!Qso7HLm7fj3si4G^V)T&ZI4*S${e?QlYDeUTA!-|FCfnF?gY1UicK@j z=7?7If#tnA4(vy0re0L0aj@S$urJ4_``xeJG%nV|e$;pw{0UwRf38~bvdmk0wD3+^pwFYlh7q3Vi z>vlM_Gpy_|oZOFZdaNWgImSFEGNIE*8m)#dWvw^)uEUC9C}r%>!J*}ZX0ZYb#MyI9 zn~2#tFo%^ElUNU_nMYGWk}{nx@Usc04ikr^1(sXVW=RB7f*{ISAt{7=u8MICelPcIk9dp(Ud8*C1 z&{QEIk5&5hP1%m~=PCrZ3eMxhh`1gJk#rgCPz<9WEG-v2L5M?qkeV4z$$Qo71+@ zfMUgA33Bgnu?Apip*rz%#R-TvL?DkIQNb{DvWj|WGDt!~Myy3b#`=V+$K@H)D9-1K z=Fn|0W^tI*=a*Ci&YDgUZRekc&f;LRlg0SJ}PmubEa? z7^e-><{mr_Hw(BZgpeUeG0w;WsKEr?G>@E zHj~Dd(4FAKAz?zA5pxlp0zNjBlZGk%#k?JO_Oq*S@{5*OC@!>WSTo zkA13PI=v4++h@N21*X#(W|t`AP8G#p{k(BO-rS4GIb2N4md`L8AA%{qTT!MSH-Cz- zcoKK(UbNq(JU8Qmzxsc^W9&ekhJ1Yc=bP(y*&+mve~7kw7&p70&}t32?&|xXQ%fGM z(5CyD)|dFmC!XN+J^z+J`TO5F1IJs{pI&+|?eY=Yor|>RzxbM9x>uKP6Y~pSqi+ro zc?7+c`RG3{lqVh`>^(%=U80#C;hQ5|Got2@+%j#>lZSgr!v#L_&n|HK$KT%T`^d-7 z6T^=ZmhYuK{Q7H-`8(yj(acY2=8xlVT}0%)tWTAn`Q-n>@BIARnaDn?vpTE41@+w& zcme(^;J@Xy*oCXW#n*Z4*MPsrJ4uZ{tFt<*|Nqpdf3EPM-*_CiKF`p;3)gy+_J{L5c_9DeB?y*1yl5;tO&D5Z`@ zhNeeYVdXMr9+5W|!Cg?C8!36DY&yb0BIP+&t`WL7=Tw(5Vw*kE=2!$mC5Sl?XGG}kb_F0;}rLwNtc^SuGIaJ=VRIceAldKDdb4H#jkGJiVGCI-}vEs=+q>T|e zYT#XT_%LDww9Ju|Dib&~=z15+BbIIu7N>8O-*x!T^BUXv$3EK@zM`J+H?U;71`%bq4Z*h2mA{DWr1;jgiRGNh_ zPEc|+aJg2Ipn{7T(LyPXnqOAOz{})Fy+Emw`s)1 z1g=zp(i=j+Igf+}bcdV{Fd1mNm#T0yM0D!vd!iB7tSVBWG5n&ZLVde!5Zs15Hc%Rz z6HEjb^koh<3bsX&2wGXa(bMc?##M)saluvga4@`2#O?*mWi)w$FR1mDL2;5W*HzTexS`?(O5x?>SVU@oqSUXYb->isY{dYstqfl8YZJ?B z-S2r4@=*V7_l$O#>KV)mA&h9~K@E|Hyjc)Btj~y86BonOoyvLYOCN0oG9WOR2#%1mPcggX5|F~E_(=yZ&pYg$Wy|PGql;o z`D4m-9%Bz!R)mTxNZ!RYyA{2XW@zpzZX%v;1%1F3ygS87PqGKmVL|TiM_R!Ktw^P& z;vifwnnnGruwstJ|YxK4mD&_71kNJg|si=r6so4Kw&z)$#nen8Vt%l zaegn-y@=~xpxwI)VPLxUApL3=mk*FZK?XuIp)z2F%FOnuHagoC!SzV9Tk$nU4K|gG zU)&;Z_L(+!u01hkOaloMgj$$4iUG&``f@A-#RpVN}6x^*{|TjUOrd4 zi7xd%0um~3|EJ`VwWjMs^oyx;;XBgv3Iay}-CRPwMZ+HT}!i zbE=WDz6Wu&x$CwQT+UYkeLB4YdJ`XCB+gEVcYlfXjdwG157Xb?0RK0ycisG>ACKG< zE|AL(o=71P{EEBsbw2TC851T}2YseSLnPg&-&{7NT=_`lA}Es=yp%(4sOsKTqiD$iJju(Ib6IC zm0ceFP~-qJ-&J48?FBMkn!X`rq$;u>n~BtEzEX+U9b)b z8J!f%E!c!>o~BuRg}8j4+#g{16237)y8`v}M^`ArfZ3M_yH7%yA$k#stL@FFV$+g5 zJwP71dQ&}~qihZ->jQGXBu(c*pRc2X;kC!gbaEGYeGKUq!Uej6r^&01k}sm#5}Ka0 zSy1|Y()t4F-QoyzWhCW&i2$MqUT@`$~8d!vBMICCz47g^(cO%Yups6Fe(cvW2vFE-;Ck1;6VO0O2jtBGF3j+)Ljr_QUVi_aC=#qLc4zgy zaT)Q6xabiZ*{WFNRj3TlR`8t=T0^{{DO9(b50&TWh0xab*rbNE;kTh><}26MRk4yH zRXEEnP+7OmDIyNDfLaZGGpbN44qVs@XX>}CT_LZxI0LFMs#6V9vr?N8&KE*A;hK!I zgjzk1s>8Er#Wk4WEE5(JVV3G>eyT_h%-8L-E#$WZG{?FQCl6FMbC`{U`E{C| z+l0nd7Ij-$#+sUn4`~yQhpi_1BYOVorab~1`h7Jzz zD_hsO>Tqvn$GG_lVFtE`NyNA7x{t#e7q?9dRmu=CW`ubL4QOR3plME-X4_!8fzXYJ zbl^SCPjziDfsK?DP+OqWdBk;yultY^E0UssN`niIX2)okYg{u>QXmg=TqA@n`!@A8 z4bgic&a5``+PZ&TQ}7W?1En~`L1>}bg?7hk4p|GKGkjOYUqdL#QF@RtRhDori1d`n z;hST`d&)RROLfPJlPcaybyVZ1Vr-Syx9hEb3#^LGsfN{If^)VF?yJmH34*1Dk~+lp z(R_efZI<{L>J*ACFuy?9tt@L3OpTmu{N!cgo~SPfSyB3 zTd`4o!Z!)af{RmKRwbaxlZPcXK^~95C7jzw^C9Br;O6+aB6dex!F|5vmTxHQb7*P6 z229p;-!R!CuHrq$)fLbK#3v-w*WAqOF;xZYZTvv6ZKJ{?8nB}HkdWX(TPmNTA@?)1 z^n~s<;#MHniS6gW9WkYUMf>$nGb~p$T}OBLH0|ZDZ0g9A>pkB#QxUZ zOs7g3zR2(Y;xoMOLwDn54-<9=Vz$(x96?~aI+2R_{TASnJ_zL=kk{cZ^Bo0O(M>3lG6Q{ zH9YjMH$3tC@0^Ky_@CT^+j&3Da>nfZ8wc`!WBF!H87?6i-t(brJpM1wa%j%#to{zv zciX{xr@jdMG|&N8fOEhAya0UnPsC?+R%iA9o$3z$g5fqu9?|h#_|t*5yM?qp=Cs~a z4ZEcN9#VgZ7{vXru)Op{b@gq_1GGKu?oD#PpuhEQ#?Z1k2Ke0D+Lq7!$_XF+^=C-a zCHkB9F}v{0Yrfy^9ub$HXS{hgd3BzYceb3#C1o5)tG~uIe?*fW=hMG*48QpH-h;;; zbNt1F2WXj-hlt2XNvGRzwQs#X{Y%IE)F*CIrgxK93&PHt_TWWU&)tRf58#{X9(}zk zsl`zc{~UP|%I!yy=C5cM&*A**i`I%T+_;z2d-}~+neRUd)DVauKR~m)gYTZgta;t% z130I&`&Ti4j*u=>#ser&%o^MkCjOn!8eTCAP<0l@rzILkxxW4Wu!bvKDq}D zRT!y)MpJ{oN(;K7Y`%b`zu@=&(Gz^|w=Q6}???RC!T%V3em}uKOUV%@vno(VBTpS= z+6UReg(LEMm$bfzJQPaqY33f$`_O#ARCXyxKSbF1Dmn>i*r{PP{@xmv5eCErojOXI zq3IgMBF+A5ZxC&2$ipsa`U|`rk=O5Mbayd7c%0KCr5ruPaQi)kAc)(+WRB3FWe#OQ z9s`! zG~QQ{&Be;H)uHk)yN* zlxD?2WXr1s*BUz1QT?2xA`LPyuN+hJ6A}uHp1f(%Nhl=)rSg?kh`uszeOFn*T0#x* zQa~NWTxGJ}$rcnttkl=qoYF3PH0_}iNVpDujdZ*9Wv_QA>hP|IBxvJm9-TU@G`LuW zW>ZfYt2i!xE4G$u78ftr6cMRRU58f0i#JR&X$aL}>H=OW>sx#RX>nTJ!lr@HY$(kx zVv5R$6AzjxQ-@hw(G;d=DImETY^8&c5XqG_514q$co!h}))D6g*Hm{kUMiP4J9L^e zt=sKMda5{vBIwlAAUy;Z3*%{Rs`#dcm{rkzyWSMc98L`9jku7Pg@4(8mYZ(Zix|!; zG7~~ndW5tVhFbyKZ4ozB2B~(GTxe!1v~RFsiRDO`duZzWZ@l8lb>0YAJ#Kvs_5-Qw zyHF<$FkF}LzM$D*UTR=hsk~elsg6nd%B@{qC#?^$d>$*2)HlRA_*tng^rYm$)ky=W z=h!>=lcx3>Oce_Sk&KJS_~tg+M$+mMB{lV2**HWRObtyL=My?CnbvbGU!#;S5WD9Q z-ov!|LBf0=?{3k|p2xYg6|oO0jzF~DMHbXMvdhk%jLtL35VTRCQQu)y?+^P)$*~7HM(ra-^Wm;ge#>d)>Da9c7hYQk$^T`mb^FrMe%8bj}lAH96KUo@P90$mt$zbsy3mQ_>+4X87*voevZ=bxh-a zMJVN~%nqJ>&Gj=*X?DL(9xszO@1ow%R}299(E&d2$&OSWVLrc)#o?E6^43TK7cz^>Phii#m-Tp?k9_=@?=t30 zFn1R|?$ho5)jN7->6Z{a!Oj!A&%I@~1Ju{Id-xU)CcV&B z@J1nnBtY^4bl8rtO-lsKaD*bn8(;WleBn#;f)O?akRT;p5)O(MWFjcj4v13F43dBW zV4iV`+V_d_e{@*ZiquTsPpzkpXfeSm06jURrUVw_q@(QP82;Q zoo3=Gq#{+xrZ&74r*U2i^~P;KQG|&SyUU{5@??yZ$OEo@4Mog7GNb{@sOK zPBOC#-$D1TQgTN+xlB$El81jq(Eo$q{V!j-=Y9X^?=@8OF)Qj;M~N9+Pc5`MGd z!@qGIpP&9WDrAV|DN=uh{`3*Hm)`)%G+R%RPa1~9hX@*IwhpI)s47Fqn~R_~35)0G zZ+rmL?+}`n)I_4XIvU?EFl{gua&9P_bBvpTeEceU+d2$A_#A zTD<)zVSX9c>>$$A=UIES9H8U#BprycVkYVrm+N4k=jZ;>pJV+`Aj@~rZl6$cOX@F^ z(hhRGhvo-yVXLNM$6YjMykF7noZy-zpv1P2))CEfHrIawX9eFC;1r}IbZdNfilqgm zNyvDN4TYSySiSN=W*5FrY&U2THn$$7^w%(Zit)jxu@^d2-jC@+csG&&wvIOZyTLgP~B#IydwBDalRpJ-yqBm7>^&qH)~wHS=n;^j56#~`dIg8xu*01 zD?7D3$Op{TM5AGZd7)&GtPH0$3F-Wl0!^Wob9tp?N8Sj|NgcB`13Uzeqw+FMjXYR2 z04cbTKqHn7BjMYjZeyC)Vw=*?ox6&ur@wWHl6SD&6SrQ$EeFPveYBjT^xK5lE9B!# znC;ZZvQw-qDaoVb28|BWrW(8EG1{8+j0M|7@;Jxq3ElbYh|i2CJB%k+5ElsTfW!_< zGpuB+tZJe>B_y2SLXV}3XugE&T0B|}+8GS`bX++Z;A3{yBN?|zO z2ZiBq3)i-l=^Fq9#IOWDIO3w3kQkLKZ8b?68$+pvtK#r8D7mt3T?DitkCD_nV#j1q z*Q=#iP0xx{elIv&Q{6llN=>`!WbpYKq;yem3EW(1qLNP?B`JCeqg2*sb@iyL08&fZ z%oK?DsKgmHVVpxP1 zlzPr2YY2GEIkbd3ykjelYUm0j1zfck$s53|CQ{W^<1ZWL!89XDrxbUr$H#WbLx<}} zu!yt;jUC8HNgdvufM_jAl#GuVaRsc@`}3!WDIy+CzD{oRu5wEORacYi*@Jjuv!>ZQ zWjx)*`+^Ih(hS^4nB8DFIgif{ieW>bv@NUBH z^*oGiJ?-)a@~I>2)&e~%)FQkov1tS6JCF+dZT|cnB-I zXp+j(#VEP0&wD8~1>YIY`{}loBw|xy9C0=E?hrViHT4Oe)8@t#SXdj&%|!ZYoDy*ANo3 z3HUptqk;j(2X~9;g?NLq&q5Ts5<}zkI{Yy{KY(Ga>^t@h{JpJ)y#*_1;<8Kkm z=lR`Vs^)PmsLg1KuYSgfg>LT~^hZ6G*Oc@K=)-t#=&*;0K#mv*$_mq4`14}^cRJ``-U)k_;zx}85w>(moL%G zD)$U(QdNf~!{GyzK2kP+O3I(-fBX2oo4X(W#0C1pAERA9f}35$H!GZ9<9xwr$(wyN zJx=N$;-`P}8RozG=Y0BC-g>m{t@NYUp6AcPGx%nQ<7}xm3Q|h3iDf};R*7p(8QOB-TgKHrZhwK z25Y_hi^Weq8Swo>goWT|BjUb;@4i_LkuZQu#H?uP2aPb)5{lCP;5vAV- z-Ntttgo3jrVe5$Gm*jp%=|?2a@bQ4W-lf=iL^`|=7!E-(RfK|sghX&IPRYWW>=SQv zJB9wJ@-mYXoC^>XvPVK?JnHCBKzt&!iQu)Sk+Pqdo<`_am5nNqoLkDcsAI#f(CuAg zI4b0_gYTF4tu?+|W19te{RmpFQ_{K0;4L1fyO?C4$InNoVM(k9qs2F#Vu?COZJ0G^ z>gaDhfO8{5zlG2ex`b;Y&MT_7BdrU}<~0$Q*5m<%7I6VA5gNg)a;?P+LYc;*GtTuC zOX%3*T|#1y3o`_V<|Sns@eW;s;wU}DXh=6DW1ZmVR?S+2)Ys8!AGH>v*=u;9QqbaX zLC~?9bj?;xt7aVIV9dKvcZrM^Uy%%=q`}uQ^cKVvs3w|;l9HQvyIRw8D)^?<^J9?+ zazq~H=%AR@=ia*pLNygjUn*-n3uX~q2C<2a*-m9yHNEJ()KU8^q$G$d#KkeDjv|Yi z(93RGcxcI~MP!Dih@}=I6N1Aupk*YjXO!IHo63#8J!+5BidrZ_$wFB>Y&66Rd2OiH z(m-)SNg3Z7zAgC1D7hjoeCsP4HMcbxo}?OT1W-lC2FVVySw*`z1yRI%#477lEKrJ7 ztb`LB9>IgpSZOd?oOd9KpA|HBh{WmGD^eh_~t$4(jYTh{`bV?a&A~x(?B@aEq-juu_ z$o*qTSmRtmLcli>%O2kt5BTO!u(*`_ok+`)ghtK%-0n^BQ4kr&7~H>5wKDyI+ZX+z#^tX$)IB=MGEksXo@fe&U;eg@RBUstNHz&Bpqp3y3;e1b=RWtz3P;Jn(i>V`o z>zE#6=^=oY{FYEJ+rxz*7i& zh{qdTdyIDp6-P=7Y`sev4p{eJLHf`0&wu%a*SwHUu(U$Q^W2rZbV|4PG&WvieIOva z^*y5upgM8US9Kc2JwtSFED65zyoM88oHM(5j+}mkZtn@&y&H6UZ=LjvArmh@$NK0Z z4ev#cdOq~~{}-SAsexpV(1eEx;U#jv3%)1Bdoz0D_D%BX=`p|Z#)X%TcYuUvkKNHKgE3G?GR4Fvf103R+BWS{#ocunB9XSh(%?}4G~ zBx61xX?EYYJ=2-a^nKAgZSV%*e**sgoBZFSRBXeq1OFrNKLGcCu{_h6&h&4uYMH_{ zUOzVU;ZOXB`1Tyl_OthOe{pe&Kk9L_OAJRB-f+KfrQiF-Lq7b8e@Q+%qHG=|Z!R)~ zMGMd&krK{6D#frk|kMc`xnqlxF)YcNJRQRz6+OA3ut3 zcF1)6i6X8i}>(UQ>gnC6>1_-C$bi=Fi?S*zZMU zh_$)6Kj05PcFdD6IlgpZ50yEzKZ8Cw z88S|3zzle3<{9w?%^;`%*tc1>tkXlfA@7Yb#JlyQb{PZ3{;%L6zHq<&5==LxY? ztcE>`xGOk!8GMVGkX4~nQ!Z}x=nyHzQ*>67kSP&{9*G5!2En80BDwD{ju7VfMIb2(^gJjW|(Un@FRh_}rx!3y=5xhSNemaX~55UZ(0!C_8nV%1sg)bl!(qk%-LNk1F$vsVb6BM(a~#|e3m zD7n_uixk3QQ^%zZG&#ii%3?K#h*iUBiuaXVx6RZr3T3PXY~l<;KqO;QkrH0PwBkz) zn-)EUvYAoJi1i*f26P6^wi26Ek*#=4Yid=K!vuU=)4ghjknmk54;8BsXG)ltTI4q2 zD@1ZNjw5J&2~>>gia2mRL=1k@~w# zm;#AFhLRO$S_?Hx!VwvAiQl#|T#BPOg4bdI)aFNgj zLfhj#4C{G)oWbE+>hsx8cm`@3Jah%rp;}GloCHK`x;!YptuGdHkgT1~H&>rKsY`d! z8wEiP^2QNbp_KYO*;q>vGlsajP*VvPhT`#*dMz*^ozBthaDGLzn-~r|oI4_Jzd~LI zOpma9lh}Q8Vvs;u?+`@6DQO$ghO7Y> zu3?N=zCg*haCl5Bs$}@YH<0y3ID8Mj{rX+W({BF-a{o8Ga0L5o}Iq1_)z) zE&DZEHfRczu^|uFiT)Y<*8f9jk1#TNx{q0p#{Zkm$%f|SankG@-Y;=}LD7Px6H7=^bb&5GE(i0TD=cZ#3V;NmJfJgEnN^tMS_ z#rM+O001BWNklpSMIsVAN6Q1T2jkCy_mC(e&>Av5231?lH# zdX)IW-{Qlc91(sJaX(159XM6Q1`@~k*)_W5D{pI1#&*Lxwm4h9oGF*h_-bKLvrn8Ko|}$FdjdMxGluZ>&4C>4GE4{;BV3GUd6X-oNK}6_qV|U?jP!% z>*xJSshB2-_m@^a)0xinebYN>@Cr4Z`?Ih6_haD8R6f$fzz+fcJ@6j`KS3RJIn$ZW zbf)i3ANxe)TG)ps5@z>b2++=NkvBUG<39Jpo2tyOKj!EE$x{r+4`O8>nmJB|q66OF zB-_8|e~>Re{1DQd z!*K=Sy~M>XAwEr7J&r3cBJIH&*uu^m-QG3CXW$7oUYIx8yuyefZ-z|HHUhhm%C=FVMs{^3u}jIYhPyizDKEMc$l8F`Qn*!ZToo zrE`qOk05RY3J!uOZv|N6V~?;;-aLl*Te#Q|;y~8OxN?YYk?BAjMop5YBSUQmFCJoL))F$3 z74bd3SyMI(a6$-)qTx>QP|fAKiQihsnBBbKvQo0giGf#KvO1owP_mmwaYuA?h-Bi{ z@nq6vC}o=ecUX=vVMKj1;&yJZIoKzk&MCCuW+3O#tiT*Cxh76!4L(qEL%wmI*p`~) zn+@dk7B~ZGaACm55yaJEuhfBb zf(kfS$U2#KP7z5{v6o;O@}`>J&I>wL)6Xq4!a#^HZst=W7knr6*t%MVCRjDrwDL&J z1cKt5Y65Dh_y%%|v8fDD*I}9g!O0p63A1fP7Q{s+Zz5?O5brUo+y^h!$P;y_XiZ3F z!8Jx~6wM8J-Cz`gSMtzSMz{)PtjA2GC%A&2J49Sf^o|}Aqx6P(M>>HL9o|DR_ZqB^ z(ku$4_Y7h=5X4?y^+M+EE5y(GUItZFfK(NAfyQ zQiD&myv~Il7dI4Z2^KL55^|kRC`wu{(J>-UFfp9Z5Cf6}q?$u_U7A?elVd-b@J<48 z4*jMczgi?U;SH#7O>!GiQSr$2T6!D1UZ|aCep>X=n({UQOmULx+J{7})Y83L6qs-^ z;e10;BkLNch7LE;uuECZF-{OaGTVBYaW#-nwn$};WkDOgV|0zh6SNljX|akN@il!c z_`1L_G9^#zadfR!>_AX*_DDSRk97vN_!vs20l!A`(^ zb4m+8jF=K;E1KmsTuAh{E~9yCTI2xq5~|y1xqR9%S+&Se7uGaiDA7{xwzINQolR?!N*_XXFu^Z z(n-fD?`LuD+k|;dpgN502fxhmbLYsXOEfL8!v_$XfgOV%@Lo|prsxeqJK(#FrhUi{ zp|%T*^>v_uaXqV_W2^@fp}B_EhDmG$+%c+GAT23jz-3R#EjfA6g0xq0IufIyLtu08 z0or(oSxXtNko!Y?ei`8#%*&r}vi6K+8)A!>1hVFVe4S5z{LRBrpZ5X_qa~ZqM6pwY`T2$yyE~4os&GPM;t)>BA3fWyo+Dpu~zCy{7@#cf%;S%PS zh-*PS#R}37jK`JWhMz+ceV`tibEub(}Ik_Y&jm2;37`QSx{K_&0oLmS#n}`!zJ})YfWU zASEJZI5$8zrkUNu(kbKhhuN&&MclpqrtG-aYdku~xY=f0JrDj7!s0Tq+sDT>Y4tF^ zJx|;|!0a;EKI8g4vU!n^y!q-cHY=NPw_y!>2{Q`+S66VDAMZ^za4zgA}0$`~@ zG!o|c+ zK@GV1h^1zVfd)hx%2=H+wT$y)P5S1Dn&4dJ;F=~h&1y2!2XI402n-${Z{dSba!*MM zLNr{n#yL+(9h#e(*mc#oQmrXg9OUF`a?^pUn9N$XHr5pN?WClLqo~w#!)HP;QkH34 zz9w*uf|Y=l$+VRk`7DFbEEQFyj8OUl&f&Ystfd;x;%id3WJqh!WcXHzT}?`AGORbO zB=XSI4Mi{;(ebnzNbI22k%tb7PUUEoky;lB^wnf?kqHAa9rsZ1!SJo1rvalT-7JO5 zR8|8wqFNB&6PsKitu!F>sgw$0Tg?E?6&;e2r^u94b=AN_YGM=v@r8Dg8CC(o<3qwn zrDz~fkqt&hi(xsUjn%~uF`O#QfVHcpd21L=joP}5=7w?YD$>ZO$-rGAZ9-F|tRpg; zn8 zMKvyKs@ZYpkmB4_nCR-i*@W*9DfMTydM=s=3-$FZg|a$N(dY3#kxviNS+0{MGDoc; z%uXodVJ%53jOd8XLLPl(J5v!PBK0+p zQbTaHa1onA%rYgePwIZ=ilisZ6jy5U+E8-OtGs9nAP$ic*BP;E@ZQMkYq^&|PFv{M zQpTrfFMf;onk1*?iVdg)Vvo7ZZ2M-#x11_v*uytcndLxS9MB)`k%xU|TL%o|49$Vz zU#JPIy69YAz2rkreLFW?|3*sH7dyJ=BICq5b zBCdTI*WIcJi1kIraUU&T!@s6f?QgvFRsQ_4!}5rfM+r?L&TeAu8tu0zR@3d|D~R-{ zo>ZJc9>`XR?G2hJxaLy5c349m1HvI@Cp5D)X}yEe5xb*_IxfHxVGVi-M2Yib@(?Qv zc^Dxk@_LJs5Afl6e0Pm@_mtuI0aE$_B4F+i*BpV{ql+6%-a{ELuv$HalyCB5FMZ{{ z0s{N`6fd7Xgfu_Q>c+#&_U4o0o*m(-M2?5=DZ-K6G901 zu)<2m>iQFSxrm={Dsx)P+T{Mmuii}cgZTHh9RMHtwHEB%#BPsn=gaqeJH=@Co+l5FpzC+T_h9f= z8UO6>z6iX?hkmVPkOh8dDD8;btoZETf1`Q&(cke5M<1lweVlgx8JgK!^V@E#nXQq% zrx1G_edT5~Cp5m{_nbGJKO}T-UACq=fO{5QA8~r^?gY;peix~H+T+U%M>i1P(d|7< zoFCDh|1x>CAg3$%7B5#F`Vja0{8O==NS9ZQe&} z{(ZV-AZ}l~+uVKavs@Y^)`(c?9dU7!@$@`83|MW(TGT?m~)Y82BqXE z$tsVr)2W23RL&+s8EZ*eHGUwEij_*O_y;clu?*%87|hsM41da-_*psmC8WX z$s|=<)C6KM0)l!F1i4*&CXR6hS|YwNa6O127-_7DPTyAJSfrkNc`|oFh;zXQ!TDUz z0bigf#w$qZ5jV$`1M*m4Y%x8LlsQ&BDG7O8;2MEaOV#4MU<%&3sjv$$A?J!fFjLZ| zE|!FDB(#d^Z9G8+Nq9^6*$VArO%51>uPELY|$3Eyr=r-G&ht_est zL%3ajRuJ=)F_1QjNI~;tM5X|4tmqpbaBhu@P31?b5q!eT<9xwMWE>rN5HveV5_}86 z8Aa>sTYx-WTC{~`Rs*MI)!a3!3pW~Sde^&JEN9g7q^P0A@YpnBZy32Q0F@Cf8PgWa z4LKW{4oLaSU@f!oF2w&JrUQB0!BLQ=&~0C@M$E84bAt;l z&?Yu|Y!IZgx(Jr*aWSjBVJp?(6%$I1WU1wHE`oH11nTh6-p6IYY|^b z0;|eR*D~Q>B;&kbr6p&fSw2nZCX>lc=lyVjyg8uBuTERl&s|^QL%(s8@#YPh_A-9H zhvgr{=@q2e!Z(WdSMhBjEWeG&n%vLPaY4xsFb?Nv!g-=M(q@J+5W)+zyVolohI5$h zQc^7og@BGr#DrXS5HWoBZIDdvFCw7_cZ#{5xVXyv;xm-7VYvA)Cg*9kG7vEt5OE-1 zr)|GRXm2u}oTu;JM_O(1XIGXy`Q)G6v&s9(k0(C*i9e-3Y7u(^Nly^ZJ%0x?<=P_{ zPRaQp+_<7YzJT@vVpnM{K6O`N@avT~3nUquSA?w{)IopqgM|1XVK&h0JXgyF@6r)* z%aKnm&>#IxG&w@IqV?D7f`s2-I*`U0p{t9~q3!YQ3zYQ*tbdUHcx1NwmAXLi8kN-o zD{Cx#_s^XFrb_Z0(m#$nd7O56a8D^-hf#Fm@#i_Qjrdthv;7@hc#GR>Eje?cXLIuj zN;x1~d*9nT$W`3>72NTAVC%*`r;bfM!}=1X|2h(0`0h72)0xin{|CL32JeqgfB&bv zwYmFRCEyo<{|5M#cWR@~bfz=?HJ9@P*jU@CzXw@rAg4Pcf8S4q_7Z-%ONfW}#ghdH z^AlX_k$92+>bE=o(S5RbU$1KB-a{Jp`W52N6++k3%#Ob6@84EjoRZf^jI%A$;AvLd zu)Ou%Z}=KH1Wij)`XVl0MR_miy9iBBnBSslR;Uz;Mzr4|Z}!Q<$awM$s&i%+E;8G@ zN;5rf>9_G=PG~FYL35;}9h{34uHb`VJc@U>FjH(Otk(l}w81skaq@0l>=3(&Z~g== z@&>E}Kpqyvt|tu#j)ub*`ogHe*R!6SOA>Jq2-5ZoLQ1VVK$7@E+3_^lpP!guG8a1L&gfh%< zL_#+r(jziQeZ-N0j5GtLI}ijvUn8Q#S)t!-qdDM1#Lwy%c=0-!VX5+YOAF|f%#D?e z>I-7kEGk1yI*U+?rd$WEC3G1h}on!3zSq%Jm1_#AHcY(Oy=nVR(H8lR={jJ zS8I4T=~N=8MlNnD1v5l?%&l?ml-%#41LSm!FE1hfm?#f}SH{(YXx<^4(4-B%*}`gky>w z6~=>E}S&DrN@PWi-%~L>kyZ~=>|2EJ0)BQK@kV>@6`JKCs~b;% z2w%LirWTdon?C*VBR=w}f63->Lzz8H8s3BVitkSA=b^rZvPU{uVEQaBT%*~4>aL{s z>!k?mgGU%QFA}&0>|>jUV7^PUJgLn3w~%v+??z%4%6JYd8>}=04Y=7Mx;i0mcJ3Iw zP%$pPyMg8h(0q~c^fK-C-G<|^*{2M*Fu8ZM+V`r@{@2&}*-t!6UY}=u{V|r8zWrSb z%&hOQ^^&x{h7VgvS7>*hL+&Ar;^K&J5)$WNdp!B1yRRhoAN;*r{LF7Yiy!t`-FO#^ zi%);o+bF{3mB-M{b#(JI|Lm9FdE@*{XFAhgirz_s_j%xbz*q0%HctV6=bhP{Go9&7 z|5oTTAKUQZ-#VhTh+4qO{S^b0ew$oQFnjBi=wFqbEeU>(yN49)o8CU2f}3-5vxL7Q zewKBY@ZCA$J>Vo>ueiL9KY5IDs^rZcnyCDL1GJ0lcg$U&j<8F_`4e>L$g2l2IfwX;cIyBw zOGHyOIm?n%E`#mW7WARVED|DE{08IE-(WnsMO=5loIgN0Mc*X#e3!DT>1cbC#RBn&bKsc@t71*(zOE_ShZVW3 zQEj2TW)bD~lxhx?eh2U8NGP~417DETRoTKTA@rU_$X6tCo55FK#gI>wr^zJlrwq6x{lrjWfNvLen;u%Nhj zL^)ZZWfx5|T->02yj=B}$^aIvMv5q|fxHQ|3{EnlLP?dy>0(X&IodkDtseCagjkK&62Ju2nv~RB z)3sr7j4+*_;*5rd$SsbFz>%~(yPqC)N~ z1KBw$yBEWpOo`4~K!?+c9}j|a9#A||P4?``*;k$gl~G*!2rAW(OA5LJM!hysC>w#9 zO#E(lr`XIzhhUK6giJ6*Jb5+4$AVi7NK{-~F(Sk1jMypS3Z)M~H8CZEcZPG|qE-g! zo!5rZ>=-&g8S8m00IncWX}lx%zNVc+T}aYglj=U^%3BsiEMT~b$#97@%qquN2ZFnW zYero2DEKAV0?YfT?xTYzr46Gr_!zOWM8Yms3@oF$$LxT1+2h&`Xr=ZaYX#Hb? z7Dov6{107Sj6ihaP1Hg)n~IKHQ(q%a>UoDZn6B$Gy|{d9~RiT_v7j1nl;ghf; z&R;^@fS5zAA+IiiZBbVPGFnzRcMG!{w9At>%oEA%7~5<}>&Niz7YVOpu7<^T(Dfd} z>T%q7gdmkOt1Wr8M;_)#vxSw2YkH(Kl(D7s3&ce_e~N34YbvR& z(HTDBTv8cJR}#p2w~Q=_~tKGddc1oxo>edzx_xDMy^` z5T+0&Ys#=&Ir)~L9MPWo>L!i9d7YY(>AA09`91W*k|%!s1wQqmH@^V>>7P30iBEo( z<^J0U>xc323_eWjLQH;&k`p1Ep*{UP)BB%$!{n|Op}&2R(kI5{iKR_LI~Ov(6>3iX73+g7Ox{f{?_%rpbI4r^>9h!Gb(-$itJM-})55H&Sv9!6M?mx%OId83_RiZ!XRJ_$Ri=@h5)wZCAbv8ZIsN)n1j&(l001BWNkl3vVrEg{D1)xkoWQ!+5xhj?dGc zx%A3Sl=p>#5}Ot3j~RQgA1mk4cx{xN6@)yJiURmOeof) zci~e@9xzQyT2v^tM!a!Ptje(PjEr-lMdIEI$|G=3945s@ij+LJVp`>+|^ZE=kwwAF+vLu5QK!o1*I+N5(MrU@SlVn_JN zC44+0O^!+3!`OHVH`^!7I?B-2(QQ>+nh>XlWPcmd3~QkD3&wSzltkRVgm-H+EzmTn z=5slXOdMq#u)KpNK^ci*pyVl(Ij(IstjYMw{xWT9agbBkq@%^BYG$e7Lx;+=vXVkODO^C40DC(OLdx}!}%$y6LK-U%uq^221LXA zLh3iKpaav04h{yvO#oL5#{9%yDH8Lp8jxliMydyD=(&z$2R0^hEnLfE*kBA^#fsFl zs3Xpma?6uPNA88#TD@sE!I@FCROaqhPh6}KFXZ){+*kIfrG%D(kCllkL8>tx1d)W8 zqnJ?Evs&;KYqHWs@Do6d)C&}==++{eGB?3`D85vFtq9IJR6U}BQX1Umec7_B=S66s z#LAs@EoCUARb{7&C|WWhI(+oi+>3d$&o}#W6f`+fZ-hWC>XKBsy3}$(N~su<&LNGf z#e}Ya3g(U6MKlYjM}n^ykSwSfr6`uc8Msym^Bx-$ItEIP=+OpGVGi4bE6r85vNwmH zl*&ofKp7j`9%cj4vODi9AHS8}T9G;u7aZ@^Big18!lXVB-#&3+O(CE5uz#hy~x=z^tGe z$j5u&r)W_uWzss5OCU6B&;i_Sn%N~>Ygm3qeUF+5I^?>D5sWxZH4(jR2$gbCOYUCw zus7x|t}aYjsmWjGAv9JOX2!aoH0y@FOr0Ds!#hK@fVzseX{hIjW}H8$pJi-P#6yGG zArhCglfZB^r=(L5zDV4?bk~HTZ+f&_P}~aVj|uH-7n8-cTaN*McF*U1H+|~I z3-B!PEKmH(f}xoch*%zRWf$FkKRG?daO*tNQ&;ag#&4yeyYVn-C8WbIlct|T#um$S z#Gj#^KmXca*zDVqW{8Fjzh6*2g9K>i`=r%7$a)*3;$QsYMaJcV{_s3Ae;N06Jqq=6 zf56HX{c@XTamd)s==R^ic=RmAJ;~=k^wvtmzFB?h$48#{9;BUE?S-^>|DX8Q-tBI8@!+W z$)3kQ@n0El8`JzKCB2u}-XzQq>-&HTGPI;&m)!4>*4OBdp26{F_bl=|(TPs<9jR}n z!TWy!|2yz6f&cm@e`k+c$abO=o#;g0y!wM@aq+}--Jv9Vwogwd~x>$X~6lnjFf=N*cHiWdxNw*#)bFM`3tn$SMG{7cv)qLtZ%-P)L)^j zp5YIF?$u;Vy{Wmn`MOh&|N1-d;WS}#>yFfl_z}&SynX=OF1{_q=|MHK`qMa@p?-&w z2VB$Rn;zG!2~N>*jt=J#C%El^)5;qx{RL7V;WkW9Z;~c%L~O)#L9=rS7e~_ZfQ~Ig zzk{ZCf*Ub)!{Pxt9#e)Ld~;mK zyY*fi>O^R6l5&Hl`^o(lrE`=~anp#8hHJXYjV%+zt-!Bv;kYIeEhD}m1V=T|>+$4# z!Uqp}ycy#zwV);xT-(?3qtWA9!8Mz*DpP!uYS9#Hd^`qGn#B#$@;-7JaW=)Z30UDK ziP9T7PSMFRp*l@U5eb2cz1Wy$>NnWU}qR=kZls-~YQw?9AF$LEsadwFKT7YC4 z@bM7q<}l333Buek9C6lF?xhO0o)EmNY0J`3=lEe9H-%;f(p2WHD?-i!OK7)hIh~JS zYG|qnL*Ldkucl4fG;h!h9$c)W^kgM>b#bDW_bCz#8$s0*TSFuew-q}UED6lCrZh)E zd`8UU)T5)4hY>WOL#B))B^fSc;w<5tIJKUiE% zQ|6aV?r4;;CVLHsYjRCpzMS}7wS?B!M6UTwk~`y?6~0ZBVN(CM52P+ql996zn@ns| z9mfw%H9CBcbETH7nV`|I6v$mmNeX_fxS^78%qt?JUE`-mlztZ*Cs=kHcC%IkH4H!i zb7=8cfl>@{Js}zvJh82q3l~?E=*g}jcMfrinc$*8u8H8#jD$tOxe z+@iS32;Y>WX)27p$J7(3Xb2mnCYuX&0bpW;SX0qu6ox}bv#6$f$xu}AuBO#|WSeB3AN3@wjT6aoRKhL47CYB5+D9lou{xaJLGSfrjRAaYf zTr8L@5%c(X18oYExspvNsbRQ%mbj=TorY32j9mbVr{vB1r4_JhWZS3=heC*oxTYqp z$J1DU5z*%vRuLV)K#|)#{)zLHq2v3mAKo#x{%05Wn6&rf;wjVxLOe^HUU{vl<`l;g zH(SE&3U;Wd-@_Fpr?bF~y7-m@^18&Mgv|8x4azVf4^ztE(Y%ego-9kG%b%DL~8L>6Qed6>o&F(GI(L<=apYQ*F!W&+T zy^$)?;_DaTuk%A6Uz2Z-*!U2+`yN7jfHiBp6!c}R)i@`2FEOs4!@1A#M?bORfBub0 zMXGET?VzHYblgG5y=soTc7xIp)FS9k#qg9|)9V5n>+eo>Z=(5m%HUYvzCb(QCoB%? zGy&jzhoc(l-K{C(>VDGlIW&EdKlt$L4CZgKKK0|>#^k+%EgvJzPvefaK^8bsvWH*{HxjNOIWery6Urf>J-8#yTXI9`PGhM>^%8OVEaI1p?os-?EI7Ow z)g<>H(;eTS86KfDXOQCyI9Ed;%7BeWl+_nVJj3dpU;d_#u@jx>MBg@jD-GTy)!h9J z;C}($j9K!Nz<>T0ZqtcQbfSL_`lDZ5@xvc`5?RmbZoU&c{WQ(=-tR$*Fx-5Qd_1A7 z{)U72<$uJjy|;>6Qc_15=e%5U@>Z1doHAS?q;TPDtxrboO4-`YiLO&hWvH{@Fd#yh=ifRUYm)(PS(o5b}zg{*3M*V)jnT_)aGC zON7O1pNB&j4)*8|9wl`va+e9M@FO4pZoE6>=#lGh;^K4**USj*or!HfU6KxzocCyE zfo66M1In<4>NZZMH4Q2)zFFfZuih}ZxUL2}1Hm1mvWuX^t(VZ_bEIWP+>*)eKCWHD z@sgY$Af!iovGRKb&=GOR zn3kG83}YQ%&7m3uUI>1TiN})1#WlsI7!G8g*sMTim>5VPc#lP+4BL!H0k^lti6Nm> zBg77H%}tPMXpXBLRBoaE|q2_V4m~)s;>ltBywwsarL>jg# z3pRVi3<+c9tHz9GN9t#|m@CUtLM>!*iO?QbZsYL|WA8vmnwj8ykEWXDbd6v+R&=vZ z_%L8OlFP=lHji_Oum$y+uo|61M>Uy-ifV8HnuXCHLLNZ{qBATB`Rfi3ah0)X3VASa z86O?~Wqh<*ZKQYA2sV{!Hg?sB^fBS0!&-wfBH1w>&#>N+MbQK$EpdK8%N)WCtTxq` zaYQz$SagDl!^Syg6-yDeN5-SJ{?DkJqAVk{)x0lBa00<^?Dhhdswv?;s2Q4tvMJgt z1++Ls6Ih~*3VMuqkCsV2PAuZ2LVbNRB7{w1*C<28oFQ>Ue5tRicZf5_Wk3h3aa(g} zE_gpsKdpq&43z9PBmaT}Bq(N~nvU7(@SGYxcC=d^xCy0K2pcqqfNwH+>2Y3&bE8>Q zCUEIPHP9Q6^O@WzuJe?^LvS>Ufpk1&>>EN8@KFfUV{E!6-`K@@#f3T*V4aQDMPrb$$w27?skDeIG}}fQ-DcVVuobEbc}^M|G_NVc zG0pUt+1aO&0J(d(GEbW&F09ZJQEkw2A2BwtI3%xk8M-<1t>-9Z7t3=>+Cy}XiBj?o zR(yT!v!Ep~t{f%~*HjEe?i_ht3)Jl8L@o{;tFh@qO>|>`ca-EwgKbjag)){}=;jM< zk}0be?;V7?h>`n(Fo27wC_)deC$uXFBifxI4@MbZ07;J~IxcSm>mw#Tq@Jgp99t1I*66Kz}=tx{={%k8$}3q3sY!Xql7ObNVAg z%hTxkNl0Jd(?9+6KiIx_E2<%aXZXlRzRI6n+(YuI+H(*KS~|qtZlkq?Bn{%cb zAYLZ42c)4#b&sKc8!n8Sl1RZUV%edx5BV}-@howASQm4A!0`=mTyMC&Bk(u;!2kIS zv+euIx`z)f8F1F4`4+P?SNZK9e{;c@6P@Tp-*Ni;8oZwb9{vW8xd_|<{+L>%b}#J# z{}%X(zvnic=tL(v5&XeVKh5L6I%ia{@=nUu4Vu|?eEaJAkRp`j7Q^xaWtkZco}`pN z=d=I*ABR`_jWj>HPWJnh^}9hYz3Jv#`1*$98}~G0pZKLIqx|>8)9)ot60^O(du{SI z)D&?uJ4Wp){Y#JH#^iruW{I4SS$yp_KlbrwIIGXxbHq$a zgAp))rx6?{hpc)JdJZPXq;U@|0n58MH^cm@S}WcWCf#f7?A%XKw?dr9N=uv^kjGPq zcC_=GjCzh?{UFWmK?OlHH=&#*52uN-@{7fFG}EuLb@B5!gj_(>vA*^_kRKr*@7A!M z4mfw~mG<-3%SuCc@OIMj8N_}0l~MVx*XMsG@nau?&qLA-d)C>{Tkzq$H*zx)T}>;NNoi( z+hufT5m}Olo0V}_))cPO$bk6*;>fGVz-DMkXfB9kLTlg>A`KFi;Mdgl2j z11(gDW;I6Z!KRd?wUo@jx21AmQyst0siw5rS%y%#!BPa*6v`-+K@jh7p@T3IWJ2*q zw#e9RAu2V^E)n0f_*l@~QicX{gDW#E1k?w_Hf3m=w611Pm~7140YccE2U1hKC2tJ4 zXgE=Fc9{B_-ZdjmGifd4;?O~mD7d!6$DYzpDPuu(g?FHgNZ1ECjl^15CoWcuga=XG zjHpj2slla;guF??1zZf2z9lcg@_=(j85O~!b_(YO7vvQ@gBCyqK`1H+sV2_VAz~XY zYC|zcLq^5d?Q+zDLMKS$F{>$3EmqUz3T14oAt{DthZ_P+3$9Vz5YQ}y>4@f$GENa6 z36si@wX8U5I#-e-B|$T!RW+pJ+@XV^3DU|c+Q1;?m*KP(=b_n9JtB8g{M1Ot1qmBI zwOf+Q5^_Lw8}t@sxJ_)w|PCZvWwZ6&wi6!SC1;>0Z^-pE>3YFp<_!K zJ*B7excw%DE`}>W4y2)Swq288EQlT;;dx*MG@C^Mp`>lfdV;0_8=s_^{{<=UkoMn0 zK71IPBt#~NQ#7x!yxy?aGwtv+D>aIZr19V3{CkLt9oqTVa7{;d>l}G?9}*Tcvs&Qh znl+{~H1AYaacZ#s6jmC{3>OpAGuK()e2|>O>zv6d({zYJh$b#3phW+N&(h{_9f!%)dhV+(0+@tyMX1EF(>c`n_|v?Sk+|jU+}~y|CX*hi>~HG z4Tua3@rZwO`PLmN=`4{T2cS35yu*0-Zrmh*TM(MtuW+T_{I*NIMa zqW^~U_ceIG20Za5ANzxE;xPsIt#5j3PIRIZ{p0DUKN`6)*{0;S&R)7TpZ=-uq?T3V z(?9oDeBjrXTW$RvOMu5HZUE`91hbbRn*FXHnB zy8Q>xEA;bBt$5H6fW`~)PR`7F+WMrn4AY37%a zW{nd?G&1%(#PvnWct2NG=Xv}SpS$bvsniO0m;LL$FN{ZLvGFV>3re0MxkJhXnAPU* zuElIYY!5KASMOkQAa@g-Ey3@QR_Cf=WgfE~Tzj+fB>Pk7dV!lA6K7W#_Rj%t!w*~J zB2YS9bFkSStE^LBXm_5WyFH<|N5IVp{z=;TrMvIb^%mXjM=8fIQu330`omZ5x&L3# zXE?unfT?>wHao!5H4+cucX}jjZcqhKUvD~-rwk2wI7X-}TyY(^1#p0m`!!K3sotFD z8D({bl6Gq1+0}HPB?opB{63-&p}L?w^#Y|EDQSccDsDB?!M zX&rBFHm04p%B&=V6@u%K<`C_7FpCsF;!46W#IHaf#7&^t>VY*9m(>&*wn@hm#8=*> z7Pm1_Q*Gs*3O*EE7*MTds0bUQ+3>EGM`aUyfU*YH7!m~SYI@pCu+G&)o&yclz!4N$ zt&H8lR|Ehct`Me`v8v_~$(UF*ZcR{iwcKh9Up%Z5E%YU^0+@jg-V1Q{o&H84>5tj^)QF{-B|Hp$VE5Ssck zCrB`G4)1k?Ven`QXfi@GfDG8UO(_m9Q`7;@)ddxU)ETl?N~<+p?F*v*cz(I#5d84%nbO2El9sR{6p>Hu2R`nasgeZtX-FRt}Podq)UHm}gL<47G&L zl2kN7)cSoZ;5$Mb3ES7m!-|~G!1^>M3Is%AU3@D`OrF6vOTtz|*g1ew62@G?5JSMMN^6`w|Gj3Xqy)v|Tca@}xX1XAS1*O}gWRGCDsUhkR4`ERl zS4J7Zio?ZV%B2EXDMxi_e+|suhb$m25<>WEh$}~tY-}6JQhAjI`&xmgT$=` zlv@nP6UuN7E4xUX5f}S}$ssq={>IOo*L>?mI-hs(hodolDF089bMI>3kxoa?W zl(dD8=YiLDF36;3e*d$e&yo8%B{z(#0PsK$zl+FnAL%}`G4cQD=wm;V_~0j&wCXFv zJP!EyHI(Pk)PhEwy*$-%r-|G5TbQlz^6k4%dE%3v;qVMn_E4Er&VA^Sw9i#{ozI`} zdQWtsZ@2!w2Je3cei3*Y_=gy2{v}?^1Uk`)PV}9o4}SCvW{+_-p2Pbse5~m(9eW=C zwc8M1;OLQOzhk&fzxT6G^TCf^W_|PsK0kmoJ4mc404+UMuG9Dz`Rxy1zH74a+b!{5 zrdvILbMJ-FG2MCNkvv5hu02fZ6M6MKfAHZK?ik01`Z~JiVSKaA{OsS|(XM?rIgIw) zS1I~l+Tjr%|HQL*n5M%+mk2Li#ZUePxvwLLZ$>3Y((yU+(HEJeS1;OYD&jfdIllit zO=yQ1GkG`GHZ-#tlf5Tt7I*HGxP6l}HjK9)#`*U%+>U(kqknuynpw&U)s8VO2-N42 za$vaiD3&h}raRT7ECW^y-;9_|5MLj%+(0s0ip_MK^vuyu?(zD|GmF!%#<(*Qd-bNDz? zvPXv*xj#+mTgLUXGxfZyLOmC6L1|1{a zt+x}V`^0viq7f}qbev$i3vP|2LrO6;uMpqk!vXjK=SMuHG8@Z2uH0n4_cU?7WY~X^ z<*gg!^mpX)3{H(=7eID#vLO3ia8OEtv_{KN$N487-@Jh7yYT)L&Ejf}VP;p$gsk9! zLP5KHL|o4*TIJXpf{CE{u`ndU7Z5n`!@Z63TI=*K_-qA$2feSUGV zj|(Ht8L2ykZ&!$CGtS)LnuMpOT#>OBEh6N;vRa#%X~_|}LjV9E07*naROSisZGA0E zs3y5a%se>7$|k+5(99HN!wW7Ivk<14&?;%H&%AO$9ZUi5@J+!I3nLri*pf)0wp1#!v%*64j(g8YN1aB z)GB%0c)8v-h7^=idBT;;`|A3Ph+rk)E3_9#SW#?rxZr6fpsJNqU99p@B@w0_Iu?Wx z37&Dez{iF-ACM+h6KwSOxH0;*8tSeon1ZITNr5Z45$`;Sh^fbUMZywuGw>eS9B~z( z>~T)PCorKDMe_vRq>sZ^#RaHUyo0!!z|KQUUV#=*oYt~0yik%=1KbNTRE&j6#qQ8l zwyg~w#iQA3DnBi7vD9+F>_BqG7R35|)f~=O_c%@*sYED{dLshT7_QAVNjX;RfP@hn zTf~n@OjvPf_PFUWdN?DEOGNhZepMF<+#X?`a8}Ur0IFT(XWJZMhVv`#=J3V#81yC5 z>J*G4&O`r`pV2vAGFBwWY&tK{(^MAnsYoVSq>(L`m7J4Kw*E{x$QZqN%g-Rr2$ zgrx%=t>O|?QIs%U6DDiK#cDVsi1Sc{(L64WAUAPtM9XE$_#940Ogth5Yy@{ioL^(I zxX!TJBlquuc!AgyLh}@h`@e9f3NfC~W^ddv-IKSiV^ z7l1xyO3zNE;Fva7uA<=cX`aW zLkm4-99QIYUqzp2M)VrF-$h5Q=dV*t9a@^&MKG)@!~WDQ(sF^Y0M64Lyo(TSM(9KHA~ zpZ&;p#y#MPPV~=Ee_w<5@2J>?cfp&o2TpXN6a7=w6QA_-`|roa$B0`G62c5=k8ysi zE4b-CN`8#oJ)<1hHF|Lv`FW4{%urrsBTFYrUZ+`{RW5(+zup7%Uy{N;tS z{KzNXgZtRDjx5SR84q#(rF-HH{;~AwpE~9PzxsKyHo!ZvamIA_>Rn6B@@hhVe37y= zhQlv|{rNpj>mm=*OzvZ{b>%Iwa%0G}^XruLIm-Gx@Ris6{4+m1@M9l)inKn>>h@!7 zJ@6%*yzzyesi(XC7(;)B?4IU#?vbMY>`(Um@UQHm(+kYD?qhQHOK(`97n>1(_8FXi z7`1oOAFY7@m#@FechhTVw~zL3L(5CxVYu}uQr?bh7x;JsBjRF0hi%HZgKq*cc9oHc z*7R7vL(xoLEl@RFP{LwB(^$up{S^Fwmfc#QG&Yb{b)^60?Vas|`wHu8XRtiMY#Tpp zZGM=L5=i45UF|Ur1?~P2%N^DV-FO<7=kXR%3T8dIyMYcbA>)_%y`O)P-+RMXq3TmV zzvPKu`V$6Q;lm+1{t&r)KW=>%=N;lVMxW+t#5j*h#)l4MAdgqjw7_(Z_yAIi(6DMy z7zc!Qj-MXkV^08v@qW}YflPK!f-9797xWCp9;``GEg%UkODqTcv!$|D=k{4^;|24D)_j>h~$0(2$n}IgSnc@6Iab|M_t^& z6tzrF9pVjiN*-IB^n?hMmfTNqzTiSjkqOXHMu*8rh&{*>?0|giN#!hNbC9ND7@CY( zswv6r$$d+h4mC{*Py&Vmik8X{)dWTzwHGr=S8)xN!NtnYgaJ1%%ofJ5oZ@30x6Zxc zT!sz8#5qTp))Z(pUMuPX>i@735fsr}$998M=5P>n6kJ2aCJYUw+X9KWiQ?N#?tM*Q znnCOkueFp+-3A@uYO#}Zl>_alSc61O4sXWlowH4HS8#3VO)dRfLgAFcoJ*bo))x##|Ui(Up7YeKu)#%O*G&eL8R0| zEqqPCI>kjJO2JQ$0YP;H>43v}P8LsEox(N8w38>9?fnUUag41`k^27*@pH;Jf$j`W zRt(1%5w8`c67J0?cfP~>ZPNNIrEKBC%Pe4{;BYd-hdJ3m!V$i`!esXg42S=gJYGPY zCoL7%q|N>Y=Xz9u)&(KyFrTnoYFb{ZL5dVKM{G7l{1WjAatn-(R4p9?9y(%hdg*(*si>7 z0W?vDHQF8HX8UM<5YtES!6Bg_J|T377F4$?6Wtun&vBCu>y(@Z^kADX+3>$Lf`OI} z9VV1?og#-cv+<7NEU0{bvQXHene~{<KM7A?M}{a+`;5bMEO#;^xmzvD5l z*N^?i9@?L$xIM&8>+7C-aJNX^RX+0z_eKi58GY(!4tf07Zc}F8i?my`ljo4wk%v=g zo--IVBCF>;fMa=c+n?1(TUzre_w<58t^4f%-s{6=)Y}MbN7EBZoP|U z`yB1oCBkAK_qvOQl3M!x3*`7O81`p8{_+39r$2mWdDXX5pZSG>ItvRQ{OCpg^1^rF z+WToIJD5yx&OuHTU84Oh9{=^{Xpg@9N@>TpL%;WPmwEgX|AldXpv@Oa-S@mQCh4Wr zgyXPHN>fUIh4J8PO!RYn>gVrmPDjoV!w!D&!du=RKikK}N5RhUdV?3>w?6zdKl-0e zk(ZIxm3K2e^&DY#cYH_g=B)M~L07lQM^Ev6SD)nf@AZ#O2p{`y+MS1qi|2`PsNP&puBc-;ItJ`M|H9=l6c@)zQP>{>ejrDqG~0iRcb;MW z7@E&Pt~^^O86B78VTpumoOId?NV0HG2g!WnxFmA zPpp8aHeavm10Vl>Bz_mJ5p>ukF0N3zDQP%ID0BQY;`|YX1O-IqwfL%=5au290ZW3o zzLw^7+gKLjJa2Ft4xGbMz{CgG!Z)lY}LmLomz+a&MS6^?AFS zWZehbt0SH=;{Ax9R+F}j5kd<(Cl6CJPmy3a9}ubOW6cgs>KuQl$2jjXUEsSBq#?Ft!%fbZPVlCr63`-*#3dPg2I{t>MotZ;J#lf2(Gr?~#9De} zC0531v#M!ZiHMcT;w9mN!wBe5xwEFevW;znjtEqf4KNYJs(GslK7fQx5LOHywDKAd zoHHyGybblWC^e1dn_Kv1h30^kIcXI(sEG+BwOAKwLbn}p3>)sU;QoL1-YnL#^gPdd zzJFN59?n>&#;&d=o86>DQnWzHmW@n?0~sV;Ee1$CSBY)NCS}Wjk((e0a?!alkRVqA zEUB8LBnY`85sM%x>n2DfD^8+;6fMrp?q+whtGlK%?rE*{4__|+-Q6r!^-vTkl2*N! z`<%1)I%}`J_FC`%e$VS6bzLz*yRracd?HH&V#_uScAW<83Lb($HJRbZC&k9wd)CNK~BCdSl4*NhgO<+B_2T~?yQ%r9s7^kIJtt~0-*qP$RTd*EXU5fRL z1&oph$V)>Go^CZFuer&R6H-M?6LEDJD~_E*wFNRFT}hV@lY-#DVT@ozF+L-MLGwB- ztOYRg{_{#xI;vi42$<;sut7yF_XPNEa)3xkc4I}=$mE+93XxKvPT}k zHWu?b_bRG}u((KEokLuMb*7x-n&VP17p9~xqIrUt6~?Y`b-=aP(EKzkCa7+ZlY>}L z4CUA{_>rPsh@nSwE-7)Vh*TJ#NQ07tV%!Qb`{nbN!AJnp;oS*cIz`rVq_9g)iqW(5 zi}xUs(Hu#fV5?hH{yG_@pIt=k9=1L~wR{#;-hdkm)H^pA<`0sai-;MMqoIo|#e2sb*`8Qh6@X6*-Im#Z^a`<^Z9^wiUV<5xYnnR){Q1$)e2A+!qXm3Ao8u(6Gm9 z{}R=Bfve}FP!nT~_BE;BL&FA(IKsFo#>_Ewhp~`Tjb;aTF?fghF6c}6Hj%;^%02kq zb@|E&%ukWiC93T`#BPA`WvC%VQaZubH~G+$FE9#!{ujcNrRo`m?r&kLcVRZ4#@2IO zdsNPcsahMnshmH3PYyNR{#jDE04RQZqu?;8v9+LagB;J0hf}QjJ^c9TQqnlzraS%; zlJ30j{BhA|esRVxTzQJH+@!nt2=@Fp-ZTUems7f|H{wc+$H>} z3*aiepCd#RM+y^Ccn5Ly5RZNQIl}or<@Y}L#^-?7l?vys{0wgMJ=n=9#v9MvR(PBd z;{GNYwy@O*;~qt$p!_t3r^)FEmizp@PrQKP%Rgc$zx%qc{~+}jHFy*7zVF9|+}C~G z*N?Em(R=ZecT!JxXm`GLTT$6}%KD!1>1SEir@_37ZZ`1qzy14b1La3YkAKqApZpZA z`U%|h61M88n*+q%M14j>Md&t&-A`gyXBqNo9{>2C@Tni-3iM}w@g*)_IU>y7$zUF& zsH`_x-6y=w9qlzQ_;Qg@!Pc#zxKa!S>Il{h=ilJ@{3e7#4z`C=Kjbt^RX zq~xhKZ-I`{B%vcR^7Ado}Du&B`3qsC#4EBj3%HfP9Fa(FMp zq0$}9F&q-(a}4um80tTv35ITY7UL$U)?@-EFUaXSzx&UQc;#-i6pY-xWF-7b&Er=Z zmdRjj&u2fndiC$#P%5M=4^VFfs*V2}@*fbpGgv=Cq(O1Tq|-8bZrlP>onZVD4XZK= z+qP8GtH^g6?$NVabjWZD%`>1l8aYitu4Bv%s!2)OjYfY&M$n<8fQHpJI-jEH2GFoP zJWqT2TX+}2z5Y!@j*i*EBZTD(-qug8_Y?4DoiiJU{S_;wBrS>;-N>Q7#+n{C znq#Vn^$T*!Xx=7<5uxirTU?{K7OL?JoUF@ z9HM4z_zAx=ku-HW71l6{T)nibT z%jZ5;U@|5LFo6_WblDapg->WGqxUM2)Jrix3?&_z6PhE|oshx=<9mEN!%vRDjxfW3 znI4iuMSuKuLO&ws3bBFMSp=b;M8ua8D?`TmND79WZNVHQ!IIVp8Ce&Vm1Dk|g1Fbt z!8u7`Sr4E|QBklSqk>39CSgp6^DD527=MCxkhqE;~}$KXQ?O8q3IOHkMYfObjN4VG^ZXPz3KVR*87BE57jBqLz-Ysg{gF% z9L*&St%gj4<_1j-VrHNnqFd{Nzf$y9#_`9b4)Rbz3I#*qGAWJ<=iBzLp$WPcH9f|5 zXxLzNxJm3Y*5AUqtMv2JG#hsh3}kh5Cy!$7CB$dy>2nP8DXH5grXEe1xSHVG8D=yf z!-PJ(2h|3{4UApg_O+;{FL;gAgH@bA#5Gq5i!)GHq}U<)Wvb0@;!MY|*g^-#FpIR? zW9re2usBWX8oJ>ED%hDX1-At1lTW@-t+)Y3G!gKuDgSJUS(PT;e4zYh5-Qihce@K}9DOLJY{`C)evhFXcpMLo({FzJ%SmrOjn`-(3&DPWF{%n}5h} z{iAy?kbdmf9>s4wN;TP~KK(V^cic0)f-}qE0h~XDZx`71DN?^f8hX%-u|017Ffp8? zJ3hq+|J9%HZ+`j56R&Y!_w{3QkTk!9nrw>%M*v`nwigSx}_;im7ykIvKVK1=o8>C$>dH91DZ zOLRvUG5&pYGsk0}__sgQ)#rkX_%q&r<;z?Or$~%QcrZQW8K3>g{#)V*Kk+{apT8vK zX8n87`ZSjH^(kF)3|11eWL^90Szw`!Yz`1gMagjFTClt8$Av7 zQF1tiu^wBEG1cl-#$C=6ayW;EGjumk<0jvt*}TcH+#<}k(ew!7J(sUokV9O3joImY>z)$zlm#Z7Ym4LOf;(V0H5hX{gD^VGB(a)y8u`Nymvq z(y)QJ6Oabc7Go@~Eu+zzJUJ$;zX4(~?1CMy*@uco3w0zXEoPDg>d7H?{MwD8Wyr6Y0wy`whS6~kP46)LSX2psA|DEgkr)v zKPPrgNz{_jJia>Sn93o`0+8TgSV5DDT`crf^ET0dUPll3=HI85i2ML zV+$4`4;fJ<$5M19)jGLbl`=<7h1Z;;Ff0X4U`)o2Qz^;{c5TusA_i++Ng$_yA6;jd z)ug^zBS0)h65{5iSV|nE!bVnEC&V?;FD;r&0iT9-IFPiyaJb_}Y*~K>`d4sVPN$J_08w z4PkYG99pW4!}aIN5tgUO{Wf_>7`vq2d=6obsgE(fd^hC2Ar8Cfut_&Ng-DC57G;5{ zYskEa8^2~AH{hA7zC@0i-G*8iF$r?c|qAl2f z%?XuAT(+1Aq*PM&Vmho}kp_V*r93lwQn01uFm;G{Vz&wT8HA2%v>*+TRu>rNH}KUQ zt(Lbak5 zC&&n)L-Uv%!L_$Qw#mT|k|3Jt7v~7wA>r^DVhv%qjO)(h#w{jAh7*fegYgZvE}ZI| z5-CJ51I~Gn`qc$DRf`?<%CNyC6m1`-GSkoQm~QN|y7d5!zK8Ar9{*Q=@I57YpLinkp*w#_HQqr`{v776G{CTU2UNzfo@`|j+`txNI}vJTMQ@H@!{X| zzyH!3C%Jzd^u!aH$3OW;ERTDt@CZ3xBrGoC+zH})L=7=CXunB}2ZWO^6ZkX!`itM@ z^XLC#?B=`hQ<$Fq>gyubs>w0cXvX~J+fn;2ZXR_!@x;Hor-`l){7Q{Bk6_0aX*R!p zo4K12!v4c(cpI)M6K}6vi#xxfKeqxmFF^m`Et#l4NPYgJfzNZ7i`h7=SNYJdZIRT5k zxqe$Z6d*@K?6w$Ymk9G+`u%`uBO()Qdx6?*Vz8uqKpGDC(8pec@MS*z(Py~)E~d)s>G6;6&>z1aJAHsra|Usyq?0Hu z-?e&DzfU)RJ0HGM{_ndfr$-o#9-!KK{#Aypk%-y4h3EE^Pd7UuoNN#md(_)E=!ZSX z&!T*b6wcF~JczGn7`Ffuu`AnOM3JjQ$-KYi-1 zD5=oYzmgK|$RS}4Coq}enuwAy?wC9{(lEjJf?Uw3gr$X;@sk0`wj{$PfG8;dAp$FW z6NsV17;vKjr754iVH;OF#PsF8Plbn>!Yc-D;h~mfZZUUZJ*n)C)gfiFk~85(i|4VeI-g{6ussZPm4 z&ZZcIy`hg5;~X(YGzV;R6HFxMEnKxKjCiflESRdpwTGl(2Dt&dQ_ASFM~$M%k$VLZ zL@mCJh*fMfXeumZF^a7Usv;*KuQ|!47{4hfip5qL*A^zQiXgoftOM&bdef2P1feJA zC9XXn4^v{_;=C>CRg)02B*ZXu~!dEvA|i<2JG1q;juahooUlj;FwkF>bL=PnUy2Y(PyW zXTN44_ozwujRS`6QPOf7<2F$1Fm8dRBd*2_xtyI{lh*IE!H)te0nrRtOeMrcETxew zsm3pW3xtzXq%dNb1#GnoxTO^(g02Cu0*s`@I!atG#@NJB-c zuM+0x$>|z)bd|8&B&@cHaf->+x|FWRRVVoIK5jH4b$f^{DR>wm?gR;?bWx0;gChDeGGOX*JZZM{i>mzZYC#M+t}yNp)RBh?DYj^WoO?`l1ZjopO-%hq zNV-VO51>X6*&)sysX8I&1(x0-y@?+`U#1ZNab@8QIl?xH)Sm$zkwb-X z8yM%nuF5n~s4?}5*fqpui|P_%R@iEWh8dU@nnz_@vqSR+xqFD!>KS7Hbz=G!qv{0+ zmyzZa>MXu~j(&9-KV49@GfcCBY|z*ex`(Mo1r;Oa1QCmPL%$r8LoS?NwZ!fci}?d+ zz5(V-)YEU`MlP<|9;-C8NFP-w>i`P{yae&4PShY=P^}Asujj0aBW`^Z1{cHrWCwIOq znrTMq$>|#yqg0a`^w;0~+TCO}&OF1(wRe$*xADa@-v*w0%iHMM^Z4d0zCFZ^ zZ%_aBM`uX!0;ZYL?7mQnKELC2ZGA{i+l1j1bu*{ky^3)G)khc(7d-yS|8u<$?(4qp z>-(&~(gyEik3IGeRQ0!V&j0lM`Sbty@bIwiy6%5TDP6sM`SMSE_OqY8XEx=BKm6f` zQ%ZmRE{~83s5f3Btlow1ALR0r7x^sr&aM4n>jS@1Q%_sc;1JiN`G_3n zbn}f@N8;bIx(zg)Aaa+Hvb((QBZKI5vZ_{lYLe}fp0aP^DCEaa0*xaK_J@D!J?d>+FA;$A>ZPl_Al zGy<*AydZYRc`~n6qVOU)SFfLMW+@d&64$U5zlS2mG z$5bz3T_kJ3)>pw!aQSV7`Ad{iKtP4Ze&ZY(&yr;ams`|jR96VMsO801(IFrF&9e;Q zZ_?~Of*S?=^jWI*_LOc`M}Kq%@$V(h#ys}1F`xM-U*hqPZ?n`hSTiQ(ZA5+sO}mIb z_onYnGd(1n+#(HIgyjh1A@?Ji_8Sbt2{D{w&<%|4tbwP3n1TK%0VkN|W!f`e!d7=C ziu;o-hS`Ip`98z^DQ4Hc2o!Ga`>*^V4;){kTfT$X>>zHd+!&%r>IuUw2-jhFiqHMh z^Y^?kqj{eck8#6!w2tWDk-FTFs9j;0;o3QD>_fC<6^u!U&BWCR`b~5wTxW}x#RxM{ zjWRf}5*T`4Ec*<$jaX-iA!D3JoFW9^0OJNU^|op16Es=KJE$|@4iR1^b{Em)5M#)}kUAiy!rPmV z81kfy+9xecU6!SUEiVum7)~lu^w{I{iYj6*SZ$34|_N$6S(f^j{u zt4Tvkj*)8GVf+fK9ZG|k0XGWt!w#uyaZLa>ApIEQq@XggAq^IhQdfm7>EF}=2AW9$ z@rFfVz*?(_DXCq}28{(rAvTcvTvCszMyU`c_~{KqEgIT_-#`$fC3z}KtT7mG@smWD z)%2mpwP3x&RuS$I;zQ}l}A@qNqbv^^lNip+QJdf3OWM` z_|XB#F)0Ibh@=TQpU2BCYM0=9Of^C(kA(B&c!uTS8H_td{S@nGx0RaZWJ^liT?aYB z`7vb3T|>w_IM-sU6XGx?bb_@lssT8`x;ZJ&gb$!UyDBeJxhZ%2n4gyH|C z316eyThZGTMHDYUkxw7wG-;0F77aS1;nK z=MWrO64rH8lOwE2%wGH{qP@i9pZp4+`ujh;PK@ujN>9mk9{*%ZKiekcnk1Pzte9Uv zSWm+NDwyAR2e!RHyZh7~9VoA-jM7YAAaxJ3SiFOKF?dZrjqzJlUW?yzeug`OKexpEvow?&~jK zU${6%+yrsUdro-*xON7|r}1&)2e5VbuKxHGu0D%xx;wE9Zy;lp(bkLTaFNtsgdY=w z_lYMAUp(_5>Mmm39%374EkYm-HE}s19zR4i`S)-6IWs~(KTGJJC)m4>%Dw-J=RtWJ zNiSgP9kMM2EY-EZ<&Paeeu?JgXZh3*lG`JH;Q2;LP{A;M-20u{KA#zsQk06($7-w{shg&HR_$a z-=GRG2+lAXZJ~BVolc=V3^uYl{Av2xE$ZpF@#EJY-JI+` z!{YjgxOfn;4Ms-9+){78fJxV&-y(IJkUg3m!}2tmUuHNtrrx^F^2Vd6>2ZDuB$8HW zn3B7J(0_|Ke}?(HzQFH)D!LU%SVWu!Q_T)I*?mf^Ay$erIHf) zr*G*{yG=R1R2BuQcTm+GIgK!s6pOeynwO+7!P$Lq5&Io~8su*f~v;iQR-W)PxnNOT<+LmQtiMfF=Pvut403%J)P9{qY3ib#mUu`VpdIR5Lk3 z8oI)4t}{*bG@;u>`z?@i{8_ig9n|~eP?JL}T-$z!JP2v%(O7|*V0{Lg(P1bijvB;C zF&C?hrc73ctB=WglN1)@+@P}wap?)Z2D5>YQMv3)E6_yF;6@4SJCFepOX@087$dqW zrnE7nZW}@%sROMFi?m%KHlk68b4#^Vn64sH#`Tx3qtA z9Ed2^K*QfZj42$^ zSW4SmZ3(Ll@(?k$K`RF+G1#)WXbrmNZX=%92^uo@0xxBd%SMdQa>< zZj=gW*Z@)%zsR1fAoV^vjL6{>*38iMfVdo?;WS!#lnP7&Il=fN#2@18{lesqWAf@9 zX+yIy

zx5FH zIHK|u;_#EC?m_&h!EfB;m4mgIz7tFpz#M{3Fy4XdNNGU)G2RZuE|cRS^~N=tt)A}a z4D@@%)m7qZgA@(@7~KlE!`*wHS=5BU1t3~o1BU{4~>?ed>r+2sHD=vxbQC?kw(u+8m@@1fsB*uu3d zth@VlyhuVt*PQFuCI#!sOQhA?==}xm8@%^*Uq2B2l{I)(^+^#S=lnn4`6K-wMC4yz zyLRor27cqMKIi4jmq);V^_kCn=3jhYH}<~n>xU)lD%5(6y{Cna*Hi_JbwsVdzZ>>F zD#;#ovq|0DJNwDhbBwD{yT^a_f4T2qr2c5?u}@6-{P43>(~Fqq9JVqTzXDPj-R{P8 z%RTz#7JA()0pW4_-YR-#nP8i;M~9< zq~Rumc{`7N;){Id7hk%|tK6=${vy6Uk7@QxJ9%v3E@@AIxK7lH*`VHenY7v@$0?!P zB*&_>Ui)%EYOGT4+(1nL9>uRZQpyl2d^Lyh?I;3YU!^;F8=CeA^GBfD0-cZs2Wsh4 zjj<=Bw8B+_sTXLhNohh(iVhoSZpdklYaF@TD=p_(a`dH1xjMnu+q}GbFX%<;-FM&{ zgKe+VjK1|ME+O@vzT3t&50jRAbhE!nvr$t`_g~lC{Z6^b0<%29*}si$FLCX&RHJLy z>X2aw=x`cSSx9HlxJ!4iMed(@-RG93quKj9i-U|A-bv&^Y`TWCb9^tjSnhA~_!CdP<;8^W)bIYYLoPr09O52CLDK|{EwV0r z5`pFZMZ)|n;(Wn!h{x6fIC8WYS4OpSaD~5WfC@~66al z4x!&d?9*lZIBlZ(bEM=kShCDWL%@wGC1j>URY^l8tj%^5O&2O+;Ud8CM81r)&|1(fW z#Ad20qrM}D7L8MeuEK-~IZ*hyHj#5v3J!A%$&z#m;>dACf70NGC9X-tc_2j?j>hCv zBQ}+DO&iczE@R?R$N|zIq*z$9g9_Gz9V=#?f>Wdpt(j<2KB_{ z9%ei*sdEz$QPSWGhtwO$6`GU~w(#Q~)g@w1in(VMUMaTI7O2r23nD=*A{lEVVtaC^ z&~|~X_KVpcN8~(4!vrLf;wB;yXIF%j$+0BpH3~VFCu z8e=JBEsm6L6yvm*|4~pCB%6{xb^?}?9u5tzQF2}*CX$kq!CHeGMPeVxgCR%IoWUeC zR!AB!O%w!*9tAX3#2!@2F{*un=3ChM2F5CyEn#_zG;E`31XQTRQf`)Kg>&3p zB!vdm4V=A++CV=%h-O1hE!OpDjP%`v6cfULae{UmWCb|FRfh~GTd3}0e2t%Wucpv7 zD%lE}4dMnYsidAgh%=KsCL#{=X|Q8=xb)V6etRHY{#D z3{tQW#zs_LLc$=P9=mvQye zpj+hVu)d+1%*e4txP={k`!(W_P*SsIz*iG&^D>%yGLRSN(R@VGj4o3#WPL#;nc*QM zRpk8gE9)Krq+tqij-4D+Pj1m?j}95k1#I1e>2S>}D1;hecfHW&$U_I(VofA0m-x|Z z6GP$@jUzNpSm+}>_K6wBz6W2Or`rBHBv@X%gxXD124tOx7lb@a!890uOpY5sM8k+2 zJi<*}dw^|btoHZFxucqvYt?$PPajXw9Y0w1QTG*W_52;*StjoyeuK7s^XsbbmLDDC z-GlUI_n!Y2;M4!;h|3>)p46Qs?4PIJy;`md)@8SGImK8o_V!W#92!DDMf0=N?dz9s zS-ZfvO?p51p0@ADU0Q8m>j`yz_vy7alT1s0N$FxD`EkQnta=WT3T%&c7&_n znAa>eqtOtfeYMcpO%`x$vF!1YE1oAm`aRxX-q(HI*X!x8q`~`v4}4%xME({q{O<4m z?zg`C=f;@-M^#bP|B-tzcvbbk1b&q_;v?SIef>CT)NIR&ysJw3*lgw_i* zOvtO#D90FkbjN60vvo)g8-&#^xwE8U zG@z!zdP|sBtv-M#as z$JT*SeFL+3liAguB_!2SvJtPN47O0GmI^s8-hsPJRKwM$&xkwQ%lo;0iwbBL{aGMUsLV%x7_>=flHVZOwSmelPns*Mvw zYQ(f)JTVBlH#lER(d-goUO2qAQA9S1IbsKl3y7T)yQ-L+(UZCf)+<6;@QBHl0uyNFCNzDCyk2sH`Aiqs{< zJWqT8COev@c`E#uJz4$?gbohuHRz7%eG|Q5oaMX>GQG~nB+1#52sY=bnhwAD@@<85VienS*rqFd^4Ted zB^Vp9l`qpJaS3k8uzCl^6vKL0)-=- zo+Yfl&1m~6!fKD<;s0mv&4MjE&oa&DpVqX8Gv=sF&4UMIgAGN~5$+H@ z5i(ALFGSzbL11tZuI8d5`mQe&(f4&hGy%enMyTShq^`l*sG;&$mpQZyye?ozkmJrwEIU;=3%g&a^;l;;Sg<}4$!Z{ zX!I3+c@okkY1Z9c#&AKlRmoY=0Kq+rMa3g7egJM|y zs$)t*iBj%*SmH*~$&kcfH+R1$t2?wSIk#WKKa%Ui+q$iPM*5*Nc&)WR2_Pa*aflha zzrVjVnM?rg`NStavGK((esSv-Hmj6U|2wd8=FFL&0RA|o^hdw{`@jG9huz0*-PS(| z-OG}fhb~`G-=KNv)+9=|%>(e8hQVfh~R&zF>Ie}kzO*Kd#t1_Jo+0)P@@lFC-0$HIZM5AcuLy1p8Cj(=yDbGL!=;l{4?_d4%v+~LAeGq z+9GsUA+9ql*Ky5jKbw;;A&oFrVao3M{ROo_2pcH<0>$X?gh)BPNL(DJ-B$SaDAmek zID{*v5sL9P#dw=4K22y(u(~=6=82&s zERO++;O~U_7PtWkcVj0#)#zJqn#N5Cdb&@MN)n3U>S^5Zn;Yi7uNn65L%9=FE6JvZ4}Cc>nRy+J+z7x z4Mvt2wXl#3YWGYeP-Ou2o4 znlLDR4m&!_usDhJS1Bgv(RNM>A0#XvM$7#uQxU^1N_!aAP{k-0gDFr&z?1x#1s@ZFVC*I}QaYOi*f-6n0VV7Sey0VK|Rb ziDZt0{xHs!NHN8Zz?L3mJ#moCM3xexbw-x>2&%@qh*J6>Eo*WanF8^Gs2)=-axyuV zXr~X*7C!fpYMl)cn+r-MDbNkttP+b$z_9A zG=6acf?%pWv>gbcK${hYWsw)2^ggz@mXq%aBo@SBf^{{qA0goc;x8krr`tV=q)XW8 z`Rh|J0Iu1gYy&hch7IpoU=jLUuH`14XdB#V7^RF0o3?dH@ z+7;~hBIS62HGv_fjJFa)SgsLtdzNB+vvWZ5f?|xsA5ue5g@8&49pB#dha?L` zw7>P~!^9pj%!$h>u6Si(KM!%JoAw%{z9NQz)^Gd_5|z+eQBL$Bx4f;}x~=b1Ka>Wq zh}MawjqazJp2yw5+C`_jh^x@vA)k z_#u@bqR_uvqlsvQYI>IC?kcgn3#B5qe&w;mz5!jK>R0XCXh?BNiZ4-A7f`QHG9Sp4 z9?Cr^3gW`E+&)FKd2`mXHbODJNVoGe(I3URXQ`SO8M^nP^eJqmF-|b<8tQuq2Kn9( zLD`-}kFhgEZ4kQ$2)+d~L6^%nn7ck$(sB%;AZd-(9$U{*p}~xo=%FGE$MOBen=;yl zol{^pag!a2W}j}@K;jC+{8+{fq>2#6$l@rr*ay>L>IK%%NO6O1>kdZ8^Y^D83etQP z@24ORC7Q6fo7jFGqyLtF|29&yH`C95?k<#lAFg~5(hlwApP)?>l6E=aV>`maGo&RH zI~2`DL^cui5JQ{;|KIg(J@U9^a^ik)_v7p)x}Jb8P*F(I6a5~4TU?=#?{anVM;z^z zwEcfi9PVe>{~+2fQKmNc zez8W;U`vIUL}+uNmn${JN*J60&Nqq zkHj#CSR+1>{5Yc#sve~iT7-Zm^%`M;q+OI-piPUUjr<&v03|U@QL4{)7!$Dd6;g_5 z3&yn>A+y{BmC$aE!scYJiG-mi4US>9P6`ci7%;_{6z|43htm5fr;soPTM-ro(J2ZO zG7~)5g9NjnOv1X95fM?5Vq!=IzROHpTRBu}P^LqhlzFNyr&V=Gq)ri9g|>MSLurL6 z656O-8rEut9ZReZ4(kNbM+^~V?E(MUkd{+YOcc#NwuTfd`nExWN10pzs0)vHi>^G% z3`hz{I!8(qOmR9dg6QXQlb5d>yWf*3228Pqb$bl+)A-p6$$y8~e}=BCsP#`%HYEy! zYu2!1!P-4g30HeUKY_SP3S%(&^|8ieM1eM$cN-ic4x?IP>htHiC?qu~Rgx3gnSU%v zp|y~tAhzou9VvF0VwV*2*!r+nkT6)aNiqv`)uU~^ZtkX75Gk?sMauDcy7?N?Tb5gi z=GfDyr18s>7`KTXUBfT$M465_tYFH$8+vu7>M*gwx5tScn5qS>APHJ8&~6*jn4#Sy z`kFX6l(rfBp;|}@B>|x%1%pta?H)8aIX^6>D03W9iF)-SgD;5vYJR_&0Ah$kgYTxK zAtB0BR@+3skr#!A3gxb$;x$r;bo-kWqg{;ixool8%3SEMK?*167gve>KpL)L?0K4% zH=cXOp#eFLEmAg;Rl*iK6sxa2uMw57%^s%cXt(YIW$=B%xEaOx1x82z3QR}rHgF-K zl%t=oV)`ll>@>yd4#ngwG1P>9h2&3wG{NcT;lxAu>kL!-k!<*gObRSC@nFp zzWF~NqU0N*zeZo)c+uze(&jR0*dlZ%fmaM(r94Og@8AXu-i#00$GA6^F-s~T46LfR zejRpOw{=@@svk;&S1ENjfQa0TQFIP?Kq+Wiroa2T2y%n>lO^l zlf=CmV--d7iaGE5nqlV@X;=p{0_D&sB7uG{mk{ZCL^--b3RA*-9ammB{62d}@SC{0 zp=w?x2A2)+FeY{r&=!S7=?#*s)9*iwvM*z+Yq;??p}mJRY(jsIyS-mH!QhBKi-3onmlk-J$dnQj0cos3PKz z(e56hD9%GQA=G#A(cgTYuYCG4XP#(4{RD1wKejo6t@g3i9@_LMrSL;V?2Zzc!{{NLkJU>QDwHg{s(6s-+n_EDIJRM>(3PI@r8cexEk^;Z|AW1CcXbw!!R-yDRZt@Iixk>NW5#JESK&#NU zLcBw?!5EI=YA`H^T(U-@P#>+*4=>!vqy~Fz#h^HKQsPbX%v9;UvS(E?RpqQ*_;u z;t`Z?QCuNcf*(rV$cW ziNTUWi^PuR*mv-YDQQ@z59_&T)%Iwe(MN=i5RW5^dr;<2aPA!Cu{Sbm0Tz1?W86`c zZHYlaBqLq=8eMd6IL8V7I<{1#Sc6|8ai3!KMfe^|R!2BkupJ$Ho_6~rX*fyfRv3n> zl;dq|v(2zKkdnp}J*He_RD(eD6Jlx+$ws_{lGGOjuhGkZ=pB%p@`bZZ(v8nQwQM7C zpRioP6>kkiHq2K^vWNC>y`$pw`h!ny^VsMAUzYQ`i2J9Y+(g@a%v3m%IfrR3(5zfU znOiWTlO}WL2G@p7RnAlM5?#2Y5CHZ zzO)PcIq>I?KKkhY=A8S#MdTwQ^2d)p`sj(Te)X$2Oa3(?^5Mh&V6C-P^;RCrZT-;e zXFj`*P4B@kPf$!(P^zLq0|VaAIP*JOnBRJuFa67xZff*?^^+I)_!DP|)jIvlcTsM> z^ronWVQA=g?jE{ci0o?z& zMhbcKTy#qcTQuFJLr1{hEj95NYP60kcX0OB$FKlKt1sjCP7>xXvw!a&z>Ai8E`0TW z8D80cpE~@uVvnPX4Q%<=PrQri%lPg%ERS)h!3$_LCwh-`6{_lR%?`?Zi($Vf44cGo z8m%I!A0cT%j1_UQ0y2Q?qhy9FpMBF3IWdW3Jd}UOa&LyH2k86vP){#l#*NMxolX`=$R+;UbhEx&Kh+18_vAH`mq@hH~)$0bejw1*UfZfGSvq9VK9K+A= z1-pqZN0@4n8H1@pQh^#8;xHw(1p_^(i0`M=9B%N2eu~iz#w`f(fPWW{V_c7}ui@+* z(E$?|q~0Ov1R@o~a)d3epj}JoSJ=P&VRSLYx^+xF!@4{sKFmtAxkNqL!uLOdUmoGo z?h*dS|LZUK!(Y7hJ3KmA@GK^kA$pYms z5M_tdnVcRr3yl6IffJLM8CnXcbt?aXr*zZfN?EK338CSO{v9cjZo!% z)tyAi4z{|IN6*a&SMQVJQN$Z!oD$lVJXS86)SJvWMIcBvj{FpDb`hx{RVcri*}2wZ zYyf(ccYwy&7PLW|mgp_RvO*<;?{jF-HIZTzNaBc{C3%zCrY0T~eTK|&Mv;8xTMIU) z3YkHdQ_{LHC$bO1WZ7vv5m$xbNaoCe$Yrr2dP%_ zv8p1;`W&Gh;ai0+6H2d9RM+xkg$@`yqn}lTUPwcM zj*jF*CbQ3~Xf>da86|EIT zvyFC zGrNzpdhM1287V6K`~Q}Em^2j6q?Q}Z*D-rooYwe@Jmu~d`Pd)Y2 z*U7~Xhv^rAfA@AD%WeHoN%4u_d4QxpKsh;$u4_!Wi?W%wMkvt3DWtgvzdX**edbw) z-}r03`s+7Xd{BZs@f6)|gYZ7m)%$7IUplnJAw@-;ZP4zYBFvsArf=}&|3W82hv^Uh z*9B*ucmdxXr`oK}|<4Dw-DIw@;fi9=i-)#8m_ZYjs{^qyUtl;dJa&##6=BBE~ z`*b_^q127X&LmzT4m%8s4azDTm$sZ?HvW!bd7LhuP(E;_YqS1DYV%@%bcMdQ5LUzUx~)IQ<8VU7Iu8$da_C-AveJfSl7{S-9czG zE4Q5NpxvC%ZjfYzG9_9YFbbs`qCY}tD^i$Y-Hnkz5{i5uM)32~XmtdfM!Sm?&CA!h z!-keeuaTmK?h4J3 zb9C2Ek$gg_4d58_YagOo`BALf0A8lux(6{HtcQL?@at$hK`}zbW2kZp_6RENkisT* z`VGp-v(Oi$cAd6+0Ere|T)QFOr6%2Zy{xu|(v+C=LpRqZkR%ggk!S>ulU6&@P}FP3RzXiFAOgP&C;PmqhX{ zlKPB&P?k7Uq}b)hxG2$Tzz2^t8rP(Z6^NN1i`KZx6S|TROVq5&g?3saQX>+8h{QEw zzn;^@A_w3>jW!@wlR^n`g?@GdUG8D+KI^ z(Q2RM>zsOw28lf>R1C{iWDuesaAk{f7m>KcryecGiSc2q(I}J6(V`5v$`E@?w_9dT ztP(<}30;++yHaR!(I4V5Rf~3(lu9I6v=-u!P1)o%vN&KZYDG~8Bq@S7D3#DAq71qG z2a^q44W8b!%D@>ix#%1#))#wyK3(B^DC$=@H9$i>MT1Ux(qS_(2 zgiZ-bm+;FmN;zPYFyF(iUIKl<-<=<&o8Jp!(0W02>>1*0Ldyix2*uhns9@>M1iujc zYz<}h@vMWLP?7-uAhpAbV$iWa3Jk_Hren0gCTkJAImCsKR}SR9!Aua$BE zX*Hx0?F8ed8(igiqS4ic8Oh0>}yi8XeCIUKQjqJJG&1(I-jviuTwuiiKulX zLF^TZmoQZTwMyzoD7z#`M(HT6@QW!>po<<`wfVC&Mqd5WXtN~Hp=Bi_Y;>aDe2K38 zQDT3bIMkS;MY|=j?}>f|QyyGfiYBE3WD8p^5hPD<@DfGyomaTDLqQx$(tL*}KYiot z(l7qcTYT)d&*H0N^w&;N-=YcF@17(sdxqsXp8V9!S-5XU`FnrhvCq%xciu-RevH}W zhj7zfBw7X{LXTd|(C!Mw(8bvps5@yHo<#C3^DrY|PH$3?(ZzyPgS0wtobbF`J`eC|qPBnS+bNzmQgJHhP zu>2+)_UX5LpI`h}TYU7he*=34^aI3UO0|9wQ(Y$bRr>v-RBM;=WK6$G=vN?J!q}H6 zUN>Fg7h{Iy2Ho--EZm!vE8o^_-PYUH52e8?BF`QCT))Y`tpI4P-@0M(Dt+y1Uke}m z*vCGll==&$)JM39!FyY`^^WN0e*3*p{}|=;9*Xe=%GGDzu!>)CkuV?8@7>K%jS1T& zkNnAhOVgAbhY+v`u(FA{Xu&D<|FvNujyX?0Dix~)KE=o+~^9{okfp&WKc+~@XJ*~^&{wZ z6Mb@n$3Fj;eBl?~-88S*5@j9ckWtKAky0LI9E@+hWfwrG^4R%XNJM{62|=j@y5>gA zv%WruSzKmleuQGUdi~^E=}0-g09-(PMYlMDpAR54u6ZesfeojSI3tA#;p*Lp9v_T( z8z2&sCk%6x-bRL1!t!}$@88Ss1^EyB+~-a)Y@b5ABWP10N+HQ3VVA@uQujPh9xRSL z_HX`-rJSam?jT*0(@;`UH2cI@BB3HBOLk4JQ8ee!MP6`>L(MRsAZdxx1Ks{6p}QMh zoS>gxIejDI)0{u6O}A&Wk!) z1)7u3(7k*g5+9)7-UXg|%g0E!`$5X-gV?e{JB@M$rry48?!H>ac&fFl47-~Izd=~E zl~}=)CQT zogxv;WqbetAOJ~3K~x39^3@*vGGq7h#b zbwTo*NN|MKVv7P>y_nx$syKK{Cs3Xknv6(5qm)8BkFi4RE5rz*5^2z+!Dd5NQcecN zh(w6dp%_rA#Z^7H9)^`{#)&}=QgIS0d?-jUAXYIfG*&dO)+lT)zjFayri?K#5Q0J( zi&6s;4{!ou3X+i25+!6qJz1hJuyv0zeKva325C8^lL~Dk+FSzyF&!Ze#X*94iK+Ln zZXei3sVjtTl@t^bJ7A0CS4mzI(lw+{#AQ!tpGOxJswipJzlCx;gnpg*rC-9>nq)m? zIYZ0{i7HPF=!6{|>??s3bWY;-26RBQV4OnPh-*4Xj$z*tqd~>Y;uP^nio~VGl!9yG zL4lZom9ROz7 zrblsL?rBZzPNB8NluKN5kzqMT{3dZY#XyN|H*y)3^2B}(B|z6s5=e2T;ag3WaqMY$6VlUrSYDqD7*@)SehE%#PzmOOkI8 zDbZDjEqe-;2z^2D1=@z3bVhR0TLqNqQMN@7Q;gFHbA0EK;c49HAJBG&gwsemio}6r zDpI=v$su8tl%{Cipk0j)9$}6$O-4jSiz-HR`~L#vJR&7&7-iGa2(FNW_h+5c>aj%p zlHf06^<|PjLK-$n&9wv8@+;rT(wDS*r?7SfQwpXSGIGQ2q3j;$1*xpj{Xl9RF|6U2 zCFR<+oKzkr_#vV6Rm#b?acj>L7bB9C^jAE|Oeji4y>TI%oqdDI0x^N>^UZq^Sh_V6S9fLRp#t_Dc z2nk0TET##RqpP5MQa{e00qq!FADC-8q3vE?WDDb*uy+GRvxlD<+N zH%#)TsOWc3;`i6++Ha!l-`#ZE-;d5b(a`VzIJQ}VG$zFjLcc;Wx{5ZD6sE*@k`RvI zw~yi`&k<*<4D&Vmy{A$AU-RW(d*_(`kNk;6p8XdztMA9whUVxu-+(;Fkvkm&ar-|)T{M%2z zJtV?cKKT+K{mmYRC2{2rhQ<4GG3KC=aD?cOV9gTPBc&Iy^-GkmD>M$v3Cq3v81|mS z>2L70U%HKFxUJjz0qLKu!F%S+nL|@TA%qF=je}qBy}`{Da2!A>b?YM#H&Bf6@2s_? zl#1_nTeo#vH`UKPaRS`Kl+$}CH=d@bZ#>Fm#|v!pZPe8VAU=rs?u?t6FJJ!U86SS) zPdPrnh*J0C>b0D5(TO1@O4M|F3I6++8&Cs{p*xf*`-Okq^0D9jHi`$(<{YtokUrc=XxFjM=QO+olr?B$ ziNhwC9n#X!&DTlceCFSx@w@k;i{liNs|?*5gd-?B#nyWm``Qy@ZF?dPtVU#tpWVs+ zg@+lR{>B?F?6i2iZ!qcvm>N+Pq5?_>lF4jeA(4CseugsJnCd0caFEXO4}cah7x4Q> zQ1l2_5Vmp6**vcP>ia`zjyz8a4C0+bq21==l9(i8d-X{d-kky=9X7je}K_--9PJcSE8n9ZN0Ub{?E0l$4Oq5DZ}c^vI5uIYga z6$=tD%rz-!Y$3SOK#B@Tq~LPFnSeC|xQG&efTO5T&LRs%41xqfksP``$4+lB&Hf;onRV4AD}z5wt3>h=HuR!31wsEYR5zz#%N=+X&N1=)0lD~ z1;Ni1fePESXq`}ci8ilF;;NqDPa#Pm@jPbq4eDe606_4U7`8rwZ{JJwCARRSxPqkh z%)r(W<07et5G&G9g4)9rmqCu9%xO~SiJ?cEoQ${i4DAlSyOKddLl_GDe1dk6`llJ% zvnaiWZ;zpNp%`yvmNc-ub|*%yW1Yb@b7HIymSqRDYKfsHw(CgD3qo`aLzvR-jwn{n z68t)GUZLFAapMbU1U;Gqb5i>v-TZwyVM|8Xq-0QZxtQ!G3d<>^JxsMl*m{Asy?`4Z z#g})XDM?o6^096eV+M-q0`f$9*Hz5vVL@Jba$S_5_3D&kKy9?TbDM-qI zxYyjjF18873Nao*sTyTlih3KrSSNIeZtFDF#zmaFhHpoNZXJ<=7;047!IX(^?@q#^ zLHu*no6lW8&+PX%@XIlA_8fP}vp0PmAWy12_SwIn-&LUQXBd8jrn!U}?_M95&xNfk z`1uKjWsh&aiBEsYSKr}M)bHuc6AfB@1Uq?%qA4k=KVj%jLO6z>Z-5Ipw?rFF^2dq8 zUCj61iwb{3xBMnPKE)sW+M5@#y%kNKd5Cg)Kdy09$DV#eS>x+fPPY)1AU%lPne)h> z{69Ag#rz=kF>{w z{>rB>^Q%w%5tsJwL*x!zxlXb#6BZ9*yCZ}`aN`Tqlc&+ID*2VJqT4%;?^f`8&l3GN zdGc3Z{61fW+q$jWx{3bT8oWAud)jPPY z+q$jS>hZ@lPs#fzruR^8yl_)<_que#Z9GQ^@5c=H@YrXc=L?^_F@XQCKjndM^T^}R zuzunQ11pFvks(64#K={i{L~NT1o55J7e2YoqrdSV=>38?Jw@`T>1}~AZRW800;0nY zJEYkXT^Z`Nvy6_uc+>5g(lff_yDVRR5D5>ma$+0!=DYHo04SD3=?J00@zzS@08*@x z*s-YJTJkvQ78QGxFJCDhxkX7VNWr7z=6u!l0Z7FQm|7Uzhu!ju~8I^5{R*O>DLFwD_mhMzx#Qi3w;eB$3di0Wo|wT@90|37>0 z7G&9Zp65N^e^`gTckk|*?w$h%Ck_FSjE*)XJ5`P&`65mrb3j+6;M};9$Tr1NROM8; zDi^uKT`uC1s_ZB+kE-}0sS-+&1c*YU3sX^Ixg;49C6NR{5a$8r+^2n7YyF2W7yl-X zGr$ZDnljmMRo4JBy?d`-d+oK?djI!(-T(~q6=H5sZ(QfI@4+yWtXRL#Xy+Z=x31>_c8hm)#N!=LqqP*AEn}{`%g#Ml}6s$>Qo6#GXZbLTrz7*D%$aX(7#y6i1LwL2Ag{z&EcF zR~a$q5!nNI24oN8R){;qR?ol5kknv%#bmTMgvkQ5>d9RFj@=uO@|eBdr823nke>uE1Hu z?UR!S*6G-kvE~THmy$AdSR2t~%aO=pk0bXaGe_AZsG@Gi(T;o;(O?NW~zmJER)WY{=1< z^FwlIum$U&86!#(i(Q)y#%7EisHgj&a}Yr_kvd1}3Kv*Lf)$D>s68S}jGvJLRt#V zLLQeAQg`&W6!ANiLQM`M(ATkI!M;dNU#2@c2Yv_1ErNyAqh`QNckrXDZqzJ?89C>*Tsic^@fFQ*Pe8=aoF#$h%jo4da*haF)*od(mVK~6``*c_CDW3&7 z5SLq|<%hw{L0X2z-Q>6;$LqMNBBw1<>)#|T{Cwdd(7~({3P|sk6<^H(atw0B=~(6 z2iJ-7vq;>a@7uC3xkd_Oa!}G@L94&dZ+-gk1Kp;Ff4e4M{V=9JL%sdA+h!Fv8~dcy z394{6qvy^8caA*xAoSFyF7n714(VqfBQEa6wI{%wSQj8hM3=--v|EFhaY};_o=6zdS}zmQ?Dc_1Axxdg>n^fqt2v`NF-V z)!m4k!uT4*f({Jb5ng(z@2A}q+FLX-zf%Tfbn1Du|6$_n0WN&z`47Al;#TVQ8YUc& zmiNHcovqf#`529v9Ix?vci3!e@-<>Sq+4v$Ha9Q#x|LRk+vI$Jaqr#u{MciekALQm zxO)~c`T)z}Jmz4)wuhxODGQQnQh$=vD{;6+98OYiEbz?{ZglhGI%m(YcNdag#*Z%{ z)eqAjoX3n_rKv3{b;+lto|Fz5=2!VUUwn=){p^c3Ef2KXL3VUUKZ+Y|G1_yi6Fn7~x0%;RlIsEuK!_gLDc@o1IaTS@fh?_&S49l8qMi|+k-guQ5rlq233u1lcOps2}Ijm1=%Y8)_jNeVTZJSk349C;Xl zHfWUrUy{Pc7GukTh$RPzy~EXSPb`TkCZejrNJNZ7rNNjvp{sx%Y4$)j$Z;F7N*)X< z9-Lxq#VSPt5VAQTh)W?W7xMqpQiZoP+Bz%=iTbp={1?wZm zd%{A{9>f(SKwK)e3RIJh7!qMoV=IsG3G0`rHl*H=I}2tCnDV?T<%Ts^q`pTPKn-TT z3wDf%kmD)BQZVK^u6~vzW77OV25rdsZmRJfX3ZVW$)Tw!Xo6vmZ!R&+$Ask(arGQ0 z{VN>EkC3xP_a4N#`>@psNE2)&h%+GN^}-7#S*(X4OW`5T6Q+i&f-|`w75bK(4PjYf zouD`b#YjXQ_-aM)U@HYWLN`Nfy+%Luq-Zf(6~i-|QcM*E(t6#(X2eB|O{g*0%94gL zakX9cqfvZ2$8Rji!P6gZl7|h9>oN6pMD{SYLzAI9+yok+0z1K`6Oa=^M7tHej*&b9 zv%<73CVAAgh_$FbKzFo_vwKL@k>e>~6U32vhw2KEBgg}*g9kBgjA;#Sv?x4a>Vm+r z22zc&o)n9TZTx_y3h;Ux4&>C42S<)8Y`sF<63h+I z7F#K%+Q&#aSFaB42kymIQ{3oJN&POv>?bg818XB`h=s2m5-L67I>-yi4r5ysOYAJh z3U-umwOw12H!-pTC_UEIXug3TTR8j^^72IjKZ9}(SNDYeG)4@zS|aWm!V2SO|I80N{)X2E@HfbtFI8#Hpp^)&Gp>+>rGwwOoQ;_=rDfI&kiYjkR-XhzuwqZ zk7~rbcV8okTVVB7VxJJ*B!wMR1?%@v^E$&~gw7uXy$56iq{7%y`ETZkM(VM}n0pu& zyTltA^jmaCUq$V|<+tvz&!Uu?{s9kv@j2}BF2rslQcP6U6;}7ia_PIm+iQR)(&hou7c!9NP`$zok|M_dwTc+&=;S$G852s(uQ% z-Zagxd;rfQBJxYB`VT(fH1M&G^>#@0$Lc-^{mwt! z=M$g5L=1P+%}(4_0mHX7%mPC9vzJKAtN3Pzu-F810-%g$cdroouOjvVs>#Pm!wzn|M~(*L-ZV&4 za)jk3dAUgn*T`}O-x!>1!1N)=Bbx26V5%D!Kx56&?GRS?Viu>l@cC1m{mj$vLN(aD zAapx4&4=;rD{mOfs+PDKmx8uzixDD$x?a%_J7_pbxA!sB9?`|?jM*aUNyJPrt}m_I z@1fXlV$4;t=|Of-X-WB`tgb9*wqGElgvB<)@~>j;7Mk~o-LNj$xkVyb6288{(CwBM z_SH79!uz9=*t`SzYQTyEJG$w=KlYg6PtN@?vHA$@=x&T3WBr^kOt9YJsx88N7mZ`O zgL8~FhqsLZ-gNVQKyieA2Ul&BG+%uCV}`{hA`@)AU^v<#rzyxLZlu_DUQ7z<39}kE zwk6q$MQsDAMRbgtYlMWHH5v%?a^6 zUMfwbV3Z!5OiwUutBGeeix|GZjdXH;7 zQWbN~HcD8n%X_*=jvYBTj4fqcKrxJ*kfRWm6=D_Z-aKAX@5s6EuvIeFtPpJp3r7ko zj2+W%T_!9dDV#y`q?By;2R zIa06KI+KRFq&@KrN_M3aK~2V(74)VcB|;6Nhz;1vlzrO^t{qAjUs6(M$!SSFzDB)OjJeQF>26F> z?P#Xo!qf{)+jI1-v!!^;FVSjA>b6MfZp7|TPhLeN6XPbXGMMTBTg}MHBBEqzFn&q1 zb(y%_p_`2{&g0rSdDuq1!`KC?CFv{gn6T04Ny(#WKrmQ2z)r4`;yKdd_laR2`kzGP zJdy{*>zYygIwF0U`pGRA2eQCu0O`aoU%2`sFpy2@1UVbBa9lH_3{uy$!bVCWvC$}hcVG8nf428hKB&O=R(g0V+nZ*H6!KgZb@(P5jRpAvgTHJ4&iTCz+q zCgR;|DElZDks6@_sfeqSs1|Hc?s~!^W4o{N8~^yu7S7+Qr~bh;;Jp{nzne<&Jk|6Z ze){}PucM7Em=4d|C$dblBo9ACSe)n4U;QSJzsKa%oyu_%Yc_H1otHpM(_yS2ew#-h zt9bme@2zS1yZ^%i_}-Mhy;VIzU63m$!}f~n=lj(Q$@?AF1AV;$>Q-xdA98oWrK=A9{^p~{a3)x=A8fP4cgQ zuDSD6t@u>ZUvA99|GeTUZq2v+e$~^T4m{0U%(_oQ{_HN+PVx1fUp!UqK0|Xa`DmAa zUj{Ftn9n`KaBv!HAEfpxqu=ZTAJE{pl zyI=gz(DYF>T&LfkAa)nA5tWR|g|(?VL*mQW<`TwiB6=^?R5125o%KW0f7tb{>wSH$bI$98 zP5RD^-SOYLMHg7wn`0SMQHXHSiA3hEbq*W|hZ+N!XI^pL(tbC;IH`lrw z$To6?5hc~&rcRux-X319m1(V);EJ4@MHO!)Uq3SR&2~Ak&;gr&YqB=l3D$}Qk15Qk z$wM!Q_je(cKe;XmihD$}ot8P~Huh+0StxeRS=Vz+M?W-k#=Iw{dH&^^xc=FJ%E@(W zjD<}lWehAX8EiCN!<`)1g%S=ckEj$333sC^W0{QNx)5=t*5nfs*QW+exUUIh2s@Kj zrLSJOI{tqz!0X3!;`=GUq3ZT}4e7F9ceC0aV~!jZF;PrI+_xvSTn%q*4TX8H77T5k z5x*w)sAI)F2_xJnWlO@X)Gh9zdP>8KGZ{7%x0CZpHZs~|=#NNHU!(ktr~zTX#2bjpY@w4FFqM>Myw{3`VU0DY%LjbEW4Cgh<56uY>Y8*=N-hOxx&jRnqr`y zUJhX}KH+lQc>w_Fc~EaA?06W%?ogjJv47m3TuNe5Mc{-9mg6Zt_s< zONqOMX>Wh$6ks@TU@hbLN|OzHHzK&k9>rhj{Fgx$O^2>gr!Ogeo=WVbYqR9BJh(p8 zs(gg)JOKYTZwE=B!13vm6@S>;JT}r{NX()8WsMx*9JypZg)B`=Co3tAsJGIS*nY>f{@P zHODxyZQv6Ukv@-jMzA{N0CO2>H?#GvsSbq|6owl%9Z2WAKc4nDm=n0stq37|uVnXy!H zn_B%<>^=2XHN|bT6n?R$ns>^dKbZ@Gzf|!@c1sVwjZLz2TtC+>2xOj@?XD3+TLkvc z%TI(s6GPCapT8EkeU|v@RK5DW`#fPFJFKjV{RBh*k&pfX8;#6w?bk&k-48Z+GUxoA z*s6xb1^Qo`A!oRLkIa}?TkmDKUy~<2Z0<`v=sw(MT!xSbY!flS55Pg6^V?rV?|(d< z4XfD6+T*<6Aa##Ym8~CA(AgRVsMF-&{T1##b;=%Y)3$Z&HoU&FzUj)W+^=o?-1QYV z(V*;jrb+vFple_WS_!!i+ryh%$GQHt_U~Kh&5&-$sjlaElK0!Imto?6o!LDmGCR+| z4a=ti~82q zkbgZxBM@|1;U%oQQ#6y7>DcI+zdh^nr)s6IhJNZ$R?+F#r!NdnwGS1_gdmHl3PB zP2DpNjY^c@r_Fm|0lPwuA&G$F-iCZ94~ns+VzCmd_tWittksKLtI~ZIvKqr=CtvL^ z!}a6N{>oO=J}??dCi}-etP_tNDL3~E%uc4B0kmn#8MZe-HvS2tp4OJ!%8|Zp+uvhj zg^`zTVEUQVK3v{BPE}>^*b9+bSyc@ZR3|?D@7Aa4eJrUS@!qnrZa7eaSF6v}!fkRN z)~Q{+VBcDsUX|FhUGVog6)md`5kfNhCr^BR3?Sr&B-=FbQ}l;GFJkmOaj$xlSRXh{ zfA;Qa60Z>4IVI#wg$?Zs=N!5p(tzBOTi-{U^)8qIXZahs-hhmsL6DZpVw}Ii#nx=I zJ}SJ!JRw%+{&lNNF8q~+F9dVb?5Ug4!t?b-Qtig{<+9{H+MJB-gOej%al`FqWgS}c zOzI+GbbkYvoClXb0~7u8ZX1i@oT{3AN)hM4&t-$PD)rjbaP$C3Jt1Ss?4ma@E1LebQ}a*l9kWd$pin=aMoT!buB&+ z=X;lBF}{drHtHu9wpNO_ug+L7pFX8FBS|&+PhmTMMQ|@5Lxl1%WpI}%D(14zYM`H}z7`vIN1FiD40)MijxC&ct}^ zNKSJOYb!E~E=Q&R_pr!Z;cdD8UH7AsNAl$w>ocq7V5> zuV3W*U7U*KSMofj3@J^BG_o^-;Z!FJKm+Q$Syb_83L~{?f%}7}Zet-0n_CqTkuy9x zJ);`~wel=z<9Clz-rA?Mw0+7l+0EWvV$>>`H#^x^_*^E$00%l>dx?vh7(Ve!GpJpd zYSr_If>x)U(Qf>${xHGuQp`7@m|*BBFo4NKyZigE>wRC+IE8uzK!pLmoNU#Xc$vYmklbCDM}Y8lugNda`7xaLZ&I z=$rBf=)u%<^6*0iZBK&V%cvK8rQ>yl#8tf9$RXJq`tnMTPC-7SW)*h@Jupx$5lui1 zk?fSM`$|9j>-wwU)vln$ZV-pd{YRj6LcvBk#r+;*QqThiuIHxzG3Vr-Kd{u1yG82W%{=2w5Q10U`qKEg`+TB>!lZTNdbfV&ZW$o<~*?);$ zb36q>rv;dg=aI{e==DC4(}+w-i1e1pq603b#Bsn}t%7B-L{mbf_&^G!`B`aRcya0X ziKQ4V8_(ZtS3kQ?cyb~G_Ql=5f?iOfeE1CyilF}2l&9_!^fxf!RB38 zrrTdbe~}Hqgk`$xDicNiO21`cma?iH-UMeE$fW2D&_S|y++=DW4CF$sSR<{e zAsdI{ZKg510y?V9(0<@mAu;1Grk%a`_FbvcoAv7?@iuUfa6xsUX!Y7)`hmBGrG#0z zwPa@|yCMahEWfl#Nx1ap=KCX^_-O+OoULTjr) zNC%YO3^aVPxrbd&&EjS1f?n1X|OUkGuhsAp{xXzIS91KCC6r+#20+4ivf{&_uGD1z|= zH5FeWN9+4N0* zumt1CO7#eK;e9CRHp*@GseK4?dcAo1FC-B9lAsTarJyt#l<=dFnnVLRmooVxx{8ru zD0$w(^?RWxh_qCx(%d?S^aY3hGJBLo38NPBS4L{C$zXNl4X(ZwUQ2G?1Uj24w*1?B zb^cV}fEXGaZg&?0+h9B0lPc!WIwPs0tMO45SyXiyxLYLKI1OH)*9xRd5Sa^9@Fm=2 ztMaGmx}i)ZiLcm=_0*NR#y#7cSo0f3i;yX2{>TDU`}f$=7CvO0nRzASrmk#!?J-jrT6QIluZI=M_C${?d?T|dr=-gg?p9DI;{EKa(vV;fB<>;Z5kTe_$kxRsmp6gw+eFClew z-rJFAY>hCcpku6Ed#1H~$++&=*zpW93pl!=#&SWz*U!NB!69hEafmN(@4 zu<+kiW#FH{KHloO^jvMHTlyB`X2j(C^|cu7o1*~-K{dUyz}wjREbXt`zob%#a0c%4wHso&dPTktjehve_?O=!Prn7| z>IHYUg82VTHu0UhU_s4jA}<{1!(Aa`0A$+dWH6>>Z-Ve-N!W}%Egl&w3jCby zN+B+q(R`Bd^oc9Wnq^7;!KX5A<~(ch7iF&{TSW|HIEsJLDfqt1;zwvr2GQCmOW!1g zcZGX$6t}|B7W*TImDA1`_Tpc#5GZ|bdPqYrLgJt%isdcG0G%06^Ccy(#I!Y?oxR|N zo+~?7kr+sPvy$s-z9vMTksiDY^g*1^nMf}C8KZ0rBFpg9lK&C*q~gUygWt7k$ZPnemWVJUD*11Axfo`+xZ8#;XCY_IlP;&7C})d`53D6;Xltn-WAGLr6k_UB{`_h9?vq^paiG zA%Ippl*CzcGt)&cO}&=wLS#|0^3Z(?r-cQ`H=qWyIqI?X5OfeMB%l4%!i`7%6&?ek z=C0JVc8<)6qy8Zg%Ueii;9bm`Jn|U-8foX@s`jZTAilbxP}B6OBMI&Wj{OEtN?tI~ zR>XRgaX%)4R1M1xpjVGEy-~jZPEyrVCD=||*v`8~({wp9rFV|j1cEFI!CQO7Co6xy zX%@0`wzl5%I5i@5H}T@Yj*QrWIYl2xg|R7!$|5`&ZqgP@sN$Q+fmft|f{aBN>tjzv z^h`1eh9pAw{Ti{O`_mbn>eRpUB*C%Z=hvu7rjOW;8sV1nsX&~96k${VYauUf+KdDu zm!l_A3F(bw{!Nk-sP8fu*#q+AwB+%jrPINLAj){ZK$&W>#!8Ae-uEWNh}_bG>rAJF z9**(y)Qdw56YW3rDv@Se@tUI!mPpTcGq2JM3k6@7og=011hS|quxh{q%AZQhZHPJll%)k z3%$SV+Bpt7f@}xTa+A^EHd$F~;p|ZeCvU;NI}udiepfs|1e|HNCj#j>BXD+2tfj%8 z^5UFqyZgU#kaKE~TGzxArq6UD3VSHPeDYba1KD0agz4*m2odNTE7`)Ht$u#tAJATa zE6)TV21ri~u9|kS;IMjxrmnGNSJ?JIIsnX>DFKVZi@_PcL^Qy!@=EBL)={^Ezmvse zH4ny<$#6HFGHY7p^Q9GK9~485k2C+osdiqItH`Q&Zt`t50aITy2%377)Xr(fZdc(a@&X}`X7j5-3qU&~@Q1={G^g&#V`9@(g7ESKu z53MR;*bXw~ZfmX&$(wAlP0_CEs)*b${ol>`_rm|q3GEHPq4eT(|IMSoaL6fli|X@t%&H-A>;?X-`Lkg3-M7{ zNSx`Wn@-Nnb@&N6@afQrqCbG=(G>WF8H1F82P0$|lSCxTx&PTI9-I?gqb7?hsZt4c!IMclY|jeShb$xV z%0s(_LjKSQKTyAR((mdO)xG2XR^f$`zJ}YZ zat;BV@V4+3KZ;G5-XnIM*bq-=O@Bgyo5w$QtZcq1Mzxh#{T+Lp*!*-@`Oy>QdbTem zoFfsH3FedO1d29K8BB)nTtxbks<8+XXzn*ef%qzcIb-N5I`vvEKF!wI=6psA@sK?j zO^k*Lg!_^U-P?$E>lJ-s`^77MVY&pVP*$&B(n3fJfUvhF#Z;Fm%}urim7AlpaCfCQ z(S&q1<1Auv|9^eBwS>=qEmUZ3rg0V<4ljdd&pl%(FsLXnmXr`=4qwg}lV=Rv1c-8* zswpU@tq0XL7R~rfl_U8by+3iTpX~QXfM3Q^&(&!tJb4kGPTW!rm#`j{;Z2;Ju^DBX zqv0elL)Qcn<}dJyKVaDZc*N_@l(`@-xL(tt8jS&#aFIzqYlE*S{{R%LD3a}vZ#^|k zi)^a@@m^_g+mxe`@pg`^wAO(!d4}d7)(l*sNX`{SVK(64E~qus()XJhXbzV_Molb_ zq!oljSv^)?UAGl1ageZlUJRnh{;2RH=5-*$j;llLNE+;#AH^du#T3lgK*o9@8XX9o z#TF?9CP#~_np{E0_FY!*MQAgV2K(p~@`OfpG4dIG(ETOMma2Wi#ShBhN0I!0hFzk* zvWdl24?NmP#1FRvXpFxeTnyrEd0O$416T3Xp0F8fk#UK4;!~a~u=D07ryxc7#A0cpQ-?Mx^XSH#Ufz5A zdK-QG_4c8~FTVaS18EF8Q1VVzZq3Wrs6fJd{ExxJ?OiBY-znx%o1mX?Z*!~ML6;7s z^LD!JxJ>CL(yg!jzTQw>)M6zlIP}*dRrHpK<7uXP1rc02=7e@D1K=3}jathx&?L=# zT~Ztp4xI14M229tqKw?Y_#R%c3q&!^6JwQ}8_J_b_)owBHf3yahSxE>#i90+w6z>tRNO$WF}0>n{?J><|Kee{vCnd z=4Fn4E)+|uV$^SugFDhj|Gc9rp2Wb)q!3K|#+nNpY+W7F*knjlt#qe&kB*MDw7sc? zg8O-}RcgFhkF=(Rit$Oc+ym9#{ruXpsyaZInPa3c%PBX&;{#|8dBo?~Od#ekK6%su zTS#eM?7JD@?K>wq&7gfRqR|mdbjk%{=yj6tEJGZQy#Byo&hwqH3kLQL~-u?hejq=OFq_^7lILMo(A;p=49~CV_uVFr^7SC03=>{>YT8$8Odc=$nKC=>l?B*Us zkVafuNumXw1wY|2Y%_t!OMh;WXY^tHs9Hig&k5p6C!~o7rtk2Z6XqK>9PB9uQ{8$NW;ihV|1~@d`=$b?g)lWT$q8#R)lAHgn9800x_R@3tsQE$p)_1t3W5-{|7 z#A=LVi+FeJ43&f%Y4v?f(+ESb+y&AV>fQ;OsSicWz+bj6Fpfa^hNy}*Er2{COQ#>O z6EgCeCYLdZkc1gHiVLa=(a$0KG>y5bo3tn6d16v=W#g|8SG(QxgFY>BDJKmY<$TKo z3Q!sg2i%Hn)ov|&ME)f20r?KzBMx=|9M7;)qdL(7|9++C-|bVDk1TYc)EI(s@rVEzfvFl4DN*zA zWb0`Z31j2n_tNibRj;`nj?|}9jexVdsWAmxeUJB#qg@#s=PaR-GqQ4G;*q)ST2djH zbOJ&-!p#A%$A30y@PEps$_PY0(1KlyJebR`Y^JeCQk^_L1YAh~)D*Q3c4|VXne6zg zCbF#jC{X*vE6mWmH2tWZVAL%9+=|=8YLlqz-%iKRf&7~%KBw~KH@0^6`Kmej)K343 z-crYS6F~6p=A`$2vT0rc?7VX~{yy2nOl$np5d9ss$M0ad{?*N5E3IHu>OQ+#i23m= z@`#6S-&3s3K<tm&E{QjWYmzNSoZ-)YiV7OT zuBkhJe#xQLiWPQ7tSk>8l2Uqa0hvjEr|d25ybpCBw782NK>s?rJV#XvT2w^>S$JpG z8Z$RO&hWFSwBY}f&=X{Y{(mok&(+e?-43I%yJXO)!w)R4)5Dg2tGkQvB$+eHvbEjq zquYg+^E+QYTwDOHJv!gwcZwOmU)`}g?@?~>anEo~^&@@zSBY7&MC$zlF7MmQBpEr) z!o`5Ax^R%+zcyWRR_lKTvRnydh%^^m} zZSv8jt&*jU$K?yJCK=ebarPaGyc_ZmiS?$c^!c(glKLi$k3xJr#gItPhzjZFG%#Lq zUQ}nd*P}BsBnBj*bM)ZdyGJ>eUA=DiLcuqqG85>Ex9bYzyBZAECN-%oQLG>`90S^& zu}w7aAfd6p{ippffYqJhL)Rtc-^`I46ukl6391MSRIpTbD* zWr$4TN~w@UhQ$4TRdYzAA4+>TLCY9dR@=!jpr~&`H6ZQ5hnot_2KdxoVDC}^6X0!Q z?;iKJCnuR?B1#BQA^`Mh^8h0T;Hd;?BrkYPS61R1`<8b_7zO^Ajx6GLS(G9aR~)}2 z(UKo0#;S4Cz03$wq0J{YNk>XUqiX;1Ozfccg3 zxP!YxC+Y|Sq!d^|70z!5sazbQFQeVw9Efyi&pL4U;QKIC{O5M&>VvHlm z@>KrANR<4Qfw`8aznt&xHyiLOYk?hAXbNtM-xeP)C|c==GvvtOgLddL(sNq%zaK>_ zJ#RM;&iJ+r&NcaC7aVLB2EG$aW5D}rbPt-;gWqb@e}8^!Z-8G8@Zx)hRT5W-Wcwtp z5LU@NX1xN>RcYVXSp4~oF6j+*k|~zV=N3#Jc#=bPF}66NHAUGF)Ht=^p){TShH!~= z8=IcNt90^-;-FYPE0QU%Eet5KIj5G$)sX-4^A`5j4!{}cP4?3!h<92O(v9hAX*yU| zF(^+y&*p>1g~nQI6SyR~)!8dkZ)tB#l~|azb8HvmM|dm5304w^4BMs1zXo#Ix(O*@ zo9H9ui8eXOXi%bX412PZA0oUGy{)*1!{UR~6frry{r!Bk+7ZcWvjVb@xnzIE35C7z zuS)-Ml>a0>8<`OHd8MY=|4$n1AU_WI^)W$VEHQ7`zw#g=Ai@PwL0ndR*jJ;1Qtz>; z!O|Z;$#aGw+}LZ^Fn??LvR!GEC8`-GR=SQ!UbMu$M62T(Lsr~QtQAI@I6H>ip4J47+ ztcstS3KP#?-|MzJ9?9I7y;r%S4tZcb%-*FN;N4i=a#&R;=7OlFvL(HH3tp1c8t*{u z{o5*Tu^01Ce0etxCr&fsj!2g-SSNk*T(z9hD$>(j4{TOJzgRB5S z_PH4E%9o#=H!7cI%x4P@C^N`}1l-axU;a&dl^?vNa%resaNmy{D7YazkAFYY$wFzk zbnLotc_l!p<3l%R)pK1PQqXvA=r*ewa?6)UXC@_pS7{#6BKz^vGZmFOuPi5}7~%66 zPmh1NB^}D4e&GwUh}v7g8n}YSE-i1^MuT}GZHsg&1fCSUXQou0g2j2p*_kjCI4>GY zBk!F!jJ>sb_vCjDsW}<3>Mf#!<7?Q$?wVT5_gyQFKV7zBs|ptsr($~AN%;GY{t@qD zYgMoc2LPc-P9d2O>y+ZFpK~nwbBLkQ1=7)VkpW;)=@_Isa^VlAg;9boC8s(V` zYQHeXI0|k~J#rxw3)OA#rc*ApR4?5>n~L<(pS|iLK61!3&v;-y$OqXV?-Ju|yuB{i z_)^LN54FbSWdM4H|t=f8v1RQ{A@D|s**tM+&lAUb8enlcq7@M+etL9nxRO*lZ?vAlLVP5VRNOV@!mS8g(w?4=c~eQYm|(Dp!5H=upLz>d!9hwNwKkDZwf>l8Z^suKnc_H5 zzH0e68?EnJfK<}J@QXaBYc$5`TL ziewCf*VDtFFeMB%{Vf@Zb7-)v4a^4BB1%%#zFMY39}e_llZ6{4U{f|;>J6R@r;M}T z>INqn*UfEAk5dQ4x5{fGKdtM4NT1oq$L2L@lo>bJNR8Nqv*YDas3lSpCx(?#f0> z&g=+sz-(V`#hN>$)P>u$R+dX7{jg_=0BuRm*jpo*(dCifmE)eOZ<%GxPfU+n#hgEB z3C;T|dv@U~h#mD+7P5N%@bqyG^SbM3>F73H)ev(VHY}Iewsr$_drc_fD3x?ZRV-7I zBO6-KLWHBM=frZ@t z-)V-{X`6oJAroYl)lB?v^}k%P`K#mo-`u?|HHf@2{Q6J+hwLL>_209{XEh+EZ!&VG zcVdmw^2IL*PzMC}qV%}Er!=T*%CwKNJ84~~s{;ceS?4!ao&D(l2g*J+xXk#%@syU^ zx?RPq5FU==UwN+3Y>?f*{`%{Or0W%{+oR@=um3F7{@(B|Uy`Dv@t_5>9V1#-RUJLc zLAfqTLmZ#6fCtGaTmtEq3x^=;-HDAR(!Xc>SKdR#f*dS>l0Q<4k3Y6+ZGZDM)9>p2&*J%N}47n!q?tT^!kD&AJP7 z$DiKV!NafV^ra*56?}L;d#T+TIgk>NII`4@X--#H8`bZ^9^#=y&jDRYMiK5ZIyrfO zkfg|rsdr&D|9JH)1;I7mQ719T2cb{Cl}Xs|bXFKsFO2fc)|Ucj&X$U|3ADNgns~h4 z_hwOipa2?pWI^_Cr-bysg^Q~Vyt~>*@MLOWjQNEFB*2eCuE8-tpVDRAE6Vr!h9T(z z59F_z=J*9=KGSJ+JV$Al_KX6#-%|$UUa9auj~EOPju>YpK-AdJzixwsqd| zlOOwX0ztc-!+ZbMx@SBKF&p4DygZsNg+<-zHJ5Pl|FH^+AlmRGxRN}dhbnmQ_$O9f zB7b4Dgd{UwmHM!A=cPmD-%MsyO7_DvX@^AAVlHCdN#gLT+nOtb{A<3rk2S+^75B8m zM>Bq4Zgq=r+2*hV7|{v)hj(ejxz@{XJytRJA$QQy6JEtee5TYu!zFN#25%K4t#{-N zMk6b6`I8n!EfdQ8km%Z!@1ihQZDP_)?$SJR6OkJ3-KB*++Ocy9@Y*}s#Ql2M1q6Ub zWaYE{8i7}`zAthtV3aPq;;ip(<~qliN9i~Ic9AD^0Q&uSW?Hlm?Xi;9Nj)V}LXD52 zm^a0A71nS&Y2YGm!xs)pQ&)$2=q|hHOtp%{!+KTi zj$X*dcBm*K!7#!0H$8ZM3jM41iHm@@3DBrs%S$9+*S!!Xf?-SA8b1gezO6YTWVuNQ zop>4=EZliy$$Ie=pu^lh))npSN|})fw9BOCi_H-0Bzq;W-RQ2Y;zytGMh5bFOKifJ9ov9p_lkG z;FkN(UreZfcDv_Qg2y+I&fpFC#@K+&J)LG#a~L!bX@lK@sNgX?lWZQ!fwm~`*>6v> zU-v5NBz^v4Gk4`f(bhd!l*bsF+;#J(!f+`05C@UpH<*5cJn@_$HkAg?n0x7lY)R7Q z#+oHh$B|3d$ZdVOLhuew74*=E=bKnRg)t|4?m0{sj-J z!Vy1tnrZEE2(VU}Pk-t%X7qnJPIOW4@SBK-O1^jaIT zrd{jk`1Nt;b))lk;D|J2s)y|Tf4Z-+cQKTD6U?A#XT<5xri*CxbJ%rbYOsoJW*Hir2BCIE*9YM zvWYF(N1ljkUkCwT{w8ylxf#kwd$VV@gL#5CGOd1XO;pm6J{{b~nBJop?I1jWzYDjZ z@KQc<6RN77+oiDULt%=zofbGMJYK#1iH%n2j*hu$XC&dYs;Z4h@#@Y&fd%e%yj#RCa;bpa+K<<{ndiro8* z%Led+b?%9coC-i-(&RU2MX*AbiN?=S5^l+fAERsf0s!t(?<*iZ4ocp4_R zC#2EEdsmk=obFZgyxbuANGPXUR|R1O=JD10>UW6W2X7j{Ct>H`d{NmJ>qwexD$Ex` z3}oJ|{>)3W9F7H)2D@wseU388(Nr)ZE`{c%K35&LR)V8A{B=yLDe4s}kbdV+O7>31 zfK_9~?RamBa>!@(`M9JVVp5b!(h(gtHebE(pC|QcR9uA*$iFnA;Gxx+qzf*EH83Zs>5M_h-bL;DeLCqCxDdD{?256w*{G8}L$`ksy2IjW+`l;RB+y zTBt?RrE`Z$HBr`w=h}g}0@_1$@xW-U1ZDRh`8EsOF}RQ<;3cW!vcSBYdLcx8&lC4W zgxkowJ_Z%(xDxs;UVI7)Lq;&h9}39d1dj83sX2#J{5Pgu)DJI>hyB-|P#`jPambwo z1v1}-L<1Nvd`e_7jry7e7aBQ(UTdfY!Wf6et6G689_%p)Z;mmYiRK)6k8PLOdzjc< zv3FE4tvBg3TeDgsvp)Nf8nZK^D=|AS`Dgzki|XzpsULILMCc}N*Wc0j(vRS(JM5pF z9yyTFU%h(FFBBXG2grbnYRvp7UH4gvs?}o0BAys57;}O!crBS!z{ch6!jO#)K$RVL zmI9d?A#Vt!ip+K39sChK<0gE~>llP~0P8Q+A6TkpSaEz~?$KG*tr`#4?oXdj$3`yZA0S)+j~zJ zb5)l99nprgYLDBefO;BeW)CFblPf{6QCh)%Q!9-#5BW~mENnqmOG6WB`9)MR%FMN~ zlp(%Cbrg3kb+*Vc;qYjqD>3>-?kRlRccrppN0 zpczMNE+Td%wCVc(fvmF)p|p2nbn2J=@E5us5_ttFSmeq5sy6G7So>5fqJljZ&Pa;} zHj~wKow3WFB+Pf&7Z9NRDa2S0`CR_;Zjb&NlK(Jj@|(f z?XCROZs-lIZs<;^eMeAM`peIDEN;@g7+X&~7&g5_9$74;I|$HJG{E@OxUc&UEF@>i z(|ky%H8W3z9+wLC%%_VY=RvG}AM_kXByp zG=ziA+mMPNPh|75@|gfSO28OYL6QH;R@LoBFRxXccrqsQ`p=7(9ftw4u&<>}%>6n^ zWfrQx*>U@|UDHrZlLU?~Q-WO1@t)b~RME8f-A_vXvKdK%@utPdV-m~XtnaxRlyJWb z#Ga)=FEV2BgS1kCCt1 zC+>D2IK!)dtT{&g~A0r?#Q@H~eo=hF0wPcOco@HOmdFhgUD;PL@JP z{~g&*Ul)Pud1?~C_rQ`WpgB(|ckhq@G&n!635Z=cRt#n3N{Yss=|rRicI$pVVTT5j zesHv*AaW=laTQJheUA#2HlI3Rq^-KxS#-XmHy44AYs%Z^s%=e>qmxl<@&}CO?>R!k z%Mxo-)zsB&6c}a0O5QKKa6-|?FqF*fKWw{XM zQQ4thDbz=)XCGbxN_9g^KFj_Cf*w_i|57)4z`F)&?pUEBFMn`f#Fb0;8e)Od8KwG8 zWEwBwOZ~fVL1XY9G}CNBx>A-kx1xDZ^y1uT=H02-CrR<7!;rX>cew?R16CWJp^wg2 z_f}ZfcQ03Vc4dM|<}E)b63j}fWu?sBF6D`6m79-;k0*@JWy9A(o*a*ZpC>hY zoRvTnr}Jl4HjX@L}&7JaEr8W0gOyFz-@`y6S2m7;R({})&d@>@ZR|2eQ`xC z92eaXx?YkH$2pp;Qo=&5G5n!agQ#d+57w)QG91X|4C%x$@? ze{_jMs=AMf{^+PCA0YHHYEJ7}Uy^iiP%on%kXQc|%L%eCw4i!xWV$4-LqX*}`xfa{ z9HJ6Og{bYO+&bD=%7y(R#zB3-`N-xP`##X+SIxz1^{^Z&BqzwcU8!z~%}XJX#Ks%f zoT$VRXcM#GR-_nFV?|EY8dai*s2tn`&Or`HQT`q|$*KKF)6{rTInnKHHU=bKWw@3K zXOZdY6LCDfCL8yF(L^?4sys5dd7*?N7vwHS%Jma3+fivc!6H|V^twAHuz)9tsU(1; zP8AMknu5m9xS{RKLs2hNbS{VbuquCuBk||_^2Dh7Tf;%P9}0ev8CwGxLY3{n-`KoE{AslU5GOXWXlm)1`Px;A7&+6o@?OYme!=xSHZy7QMEFm< zrl?6w?j$pgZ->>(UgmpFG0+yT~x$ zAl5P{CItfqS~uFXhm~ClRy|h_P)X)6%#8?64y!n}3DVtb92!dJl%S0z-BO_rH`KV* zqN=f!!=sM3iv;vcNT8w$ry`8|55Aq1kXIGAeIqH%$fKGwL?(wtMqojM)sc?9|ERh@ zwXl=gyQkKOQbe%v#5-w<@_#0e$pDE>i(&DsQ`a)@cNUn^#csC5u3E{?i!X2Yqe0wBfl z*}`vkR5{HPg;Nmplj->X5B@*_zx{^oV8GQiVGS})@iVZk;+iAUU`p!Thr$oeEv{W) z>_{Fhp}UA`MD2pQxd>Z70{k&TQknn&AOJ~3K~%p0ESG!K6HhhISoaEc0vgYd#tWoz z3nLSZ8$edrYEytnAdWT0zJ$1|wA(L`)G(fGle#U$jHrnimoauAsYTL&sd`M+0u{z~ z#D0RGu0XV4E=>h-C4cknqCJ%!D6Y8DcCoJEGu@?(pyk1~UhXK|{kCOWb z8BP*$y-P?5YXi+Rpt(lu7T6X;gcOMD6Vh;uo882ifb}CrYRI1Pc#mHT)Sf4zCv8jS=lETjp1X@bbR5LU%mfEL6r2{ z&oUnEBl3RQ-D||YM*GX;co8?dbKkT|j-KK80b*x1T=Ez3)0c7ny&IJVEsMd;j0-0- zprl|w{^=h@CO=NS{V>(m75wxb#=hNTZn!~QP8p6LVe~VSRDAef{o!}SQM^|bv-qog z_><2wd3YbQnj@wn5in`RC3Bm9@@s*Q{^l-zSQ3T@XnF7E_BWOKZ3th-RHwz!r#jWC ze&}>+@Sf^aKP>uNpMDs1f0=sr1ZL9HOkaCDM>YwmpE4X@!j0z`;+&8F*8k1tKk=Ol z5q>Bo%0Fdu)BWpzzMFEYH|4TJoGD|$Z&5WnrkWiSqJyC!1&^yD)(uECvcB^GX?Tnn z{yJhBZ1u{0&z++ub$g_70(uSWU!iK>L~)pEiG~?5%$ZD{!^w)Fe}FumN7IxX&Z4oT zI8_HU2hbWb5sVWwFOYPXdiTZ~+na=3`bn(Qh_x+$>hakB%giwq|%^i646QV-$@!($ls$gxAU#<&qTeGPQC zm@ryn%mmXcHbrbgk_s^^Fotn;j@WmITVq6UbwYDnPym%J$yD*g;e27iCR0*)s<>vs zaIyu_gSMOF6pYQ~J=TZ@Det=}n1R@NQYR=8XA^n!s5Feri-=lW^%`mh$TKuqjLD?v zh{F_T`c0bkY;nU-vPnp4$kC8jJt-@&!rCS4lP8JEF_|9}qu8uTosv6?$R5$|6a1`< z>{pr4J5t{iBhhE95sc56%Hk)PoJ(3`;A-7L5*Z1I9&M3s{3n*ckl|)R*E_8xb2znIlTd<@*Iol1s^2U!!Ei z1kfd7?_&LRY*nTLa<&^bZdW+dv7}aG-a|)&#ua(khO|RYdu5>@kHmfl;|xX<7%k_k zv6!kbkf9i;whrW#Lz5+DkMSGkb1K(AjRR>k#BqX{8nF@Uav8a&Og#tzvL$OwUceHZjcreH}RqQRhMjrDi&ZAXq-U!E9$L#q~I79XEw!(U&78?#_kb5@-LpHYZKM0e~Ehb1b*^7 zSsMD|htP3EhlLn&0m&dK8?a;~s#gZDR4tO_=A&@iWdMtrBY?JeZ@ub2}u9ovk4P$o> zR$Ij57YXVA;rD;Bo&$f3Uf3&#?BMh z4rB{A8F6hncZG;?Q;gfCY9mGx)F;%{ns9Q7G(Nzv8fbQ2F83Le6DImHaA*;I9M>F@h~&7B9zR&lTiYQn5|%rJ)mdUclKVrnKA@_; zMm2f&V&8X?(SWE&+mh1$$TJUM>kr_!A7-}yJkor7!`RJMxXE+qohJ~UL|$C*&4q{W zl}h^h=H}~u{qtWt;$y#kfR2wcb`z?1DC7H9Wi2I)q=q~|>hALSUp?)?IMu06^+Tsq zgZEUY`eD$=KDCece+j$&e(LS3)LZu_bB$HZbcvt7$olvk*r(|3Wq#pz{tdtPci+9- z;Z&#kiz1N$je(Rs{vCeqT?Dth?`M;q8Je4KG+$)EhbUw&m&sQ)i;>G z25hsyOsvvD4q+k&YBAGbslJyWP1xaD+6`H1?lah8bigia|9pcJ| zR83}6HU_Q^#83kXF&X2QWWPg-`=ovs{1UMN)l3Q#{Pta9{{UgVPkW}MZ(MUg=#B^{ zXRy0B-nOlp!Ej7WD{?sb|JZwzpi8dvOz-*5kePS*=bEc95Cp(MqQ-V><3-!}1c+~x>{k~^Gq zAiMZ@io8+wC*VTFhlmxWjNm2^rkRo^2#U1h1}!jx&9U5KVyH@OyqAD?kVl0@Buoq! zrKXXk;KP7w!L-9tLm8^MtE19vk4O+)@@Nutyohs3)&qh(q-0@ST*L=ZRcr`|Zzv-` z#ctpYCmRN^T_mMvu(SgC978`N4}zy);^^=HAh=!PWJ)2RX-!%8l)OzDcZl7H=?Wbq zv7O?&5u?R6X*zY5~(&aR%y8mvHVFY^3BKA!Fm5<@z2@ z53wAXT)K;BAsuex;w7|qAS1CmB$q8phV}6j7dw2&Xvw(PM#7LS6r36^ZW8YbHiAip($>e73}i%NP4ADjbg?K%Qe0E;VO0yxWJImQ zl@aknUHnKpm>z5jBQwS3Ft%tK$wS7`)PS|DaAm~#T+_vkL;RAG8!%UaEmBL&0L&_4 zBTjm9-_&zI36{Icl}?2+3P^c$4v8exj4}iWLeK<$h2d&K6h~Tl$^g!LdkMtnR&^AsfECuf*CEHCP0g0q^kj0LeC%@Zsckg>4D z7zyoJ<+&R8wh$(XalHdWgWr9CpYLL?U1qyX2-}Y%vP~K0nBOOcIkXPvRv06e3ONO` zI(#|BO}4gr8l-hZ!LTj0P(WxlAxsv>2gV5o5x|MBFV5M|1Mp zS4MdZIClgxB9bY?MZzS~ZQn*hVK}*r4hP8mF5SUjkk;Fjv{#wx7BOw{)?ndjBt*K& z>xAhWT-bzfZj#4B$vc>A6WTR-n4sx8<5|Unr1c)!Cwl!NmR_R`3!Zy^hrawhn*9&a z?z|2zk;bup-}Q6MrubIzO`ZCPK9Pql@^V^}>XYG-xr+&9?C{YDlYwG8NP38iJtBp; zbqhUt938Kc*7t~$`rR963!3phI^1A5c?S5icl<8?pFj07pLjm8I9)RtZ`8s~zoMAO z+!WDEXqiD?G4x+W(+h~*p_x673*VVRsp3!ye#|`g>z?7{BgCx_Fx!9mUCrG`a?14J zWmfrn2WB>FOe)qrm&Um=rk;LFvUZkwAGM+p^H~CBS8@R<=X$^E)1P+y8~c93{28WOuM^+F-2FyP zckZz^ho+|)PmX{;hHp^{ajtXyZBX*31YMxR7Hqw@8@Hgn^VTc4aDs|GIz-yRVu>ak^d)@ z^(D}U%=Z3*)!~y^`CeA{{{yye{HsT;;_oDBbG;!S_rc%A`ByQUq3H(bA(|&BGjLLy z)U~aJRU75LK}7H|qZ+BLTp6Q|M2jenEtEOl)`(kxW=ftC^dS-plD4qEL%N=B_bz31 zg}nYCso5dU-b_B1VHeyH&DNI?*5u(DS+aXyj#8p;$)O)Uen0WFb| z8gv-ZalqttLi-xHffyD{_Rm-!UdHkrB;LVsxsE*N#}V%k*}`Os^AjW_yw*|nGET7^ zk!Asuy3fTECWW+W0Y_*LtI_8?W|6i%p+B0DbsOIZVN%(-IKj7FHDpDQmjM?X zK9Ba zu3JFq$m=N%!HOe~wfsv=DsNQ;*D2P6n4x1pf^AB@40M7}aCQXkpF;7B{fKyvBUVI_ zB*YysF8)K{b<`EC93#mKf)R(lm<1GsGO4x5&dBN5G=u{Z>s<##~>!BqY4S z8wQ+h3JfDo5=Oy_L!?2>A#sIT2920?l+vNeK~akKh&FZm^%UPN@l8*ho`DlAb*Qy9 zMQ>okHkvOXvBQNCCnIWX76@8|!Z7TBMcSmFek zBp9Y3VuV=8YuO-sj1Y~mZLH54P{VseyyBu_u7FzQpQolSOpW_EXM}dZhcm?Y*a)~K zae5nZ1L8K{7jJxI3z!9hPn2l*I1prt+8lG&G5`0_kciz5p*z9(2eou=QyQfRe%OZP z9LsCQ(<#&4+w^&Yo_-K3m+;{d60Q*&kMEYi0pjbqV5untZUwB!{Q*Xc z@7_$*nkYUl8TA^15W6?GQO7Bm!^II7mY7X3ox^5AOQaCHhopQ!>94-!^|f1f$tNFU zZCAm*_)Ys4e*05j;}g$6pfAr5me-JG9~WBCIhq4H4!Clg?7zhF=@;4k@)Jne!^)I5 z5DV``mZvzoMPbEob{RjtN@#k*^xHR#UC6lkZQS}QF+K4eZO3<`AAbE+{^a^skanB1 zJ5Mva^m65m1J-veAAS(Ci+I0_iyh(=&HIGbKo(g0VRl}e@yXAA@q7~TT<1F1-zA+J zyyrUC`(2;Ec%8U)oj57P-M2UDeltyW?~s=Vl=d-x^p`I4d;je_k5ZlMTz{K%$=~I$ zKgN$A#q@Q=m2dr|jfUd`NGA}_=(azPlZKpY2_cpU-y-?q{0Wu@)TVFyKR-Fa_qQlz zLOS~(Fu^tYXuLa*6HM2HHwtFdF|4joQUbd}oZV$R9&z-wE1=g{-TGdp*Z$qRn!IJ0 zBEG@72Xq&{%y@bOEeE9ael@POHmbWYkcUv&oEupV7`5SEOAD-VaYdzBTHUYCflkSJ z2D(PVJ**s%)|YY3Na-6=ZfWSjDeeACtRHrar;lOkX?70Lbw?g{kn$>SdW+EYv|D#r zE-#br{YT{0h+0n!g2XGhW{P+~&JcFcQAsC($%R*F=7;nr6UurI;RbnC=vC-0++y#> z|IOj6Ta@)jSl$01&GubFv&2tM$!A;ir&l(d+BxKwk`)~ns2t+MAx;-G^NM{KmpgTo z-M>yVxkEa;K^`w5a+T0z!t|6fjEL8o)`pq_9(PgQ!$yxtEjkivsg!dLb8E~Qlm?-# zX~&c(IU>Hrw@N!J9v!LbTv;K2kBHbqwEkh0qDOKq&!{ zOqlg(?~x$*IMm`Scb~EbU0)=0eMLt|L|h9&iLsysv@~cEO7fI^gmb&-u){b!pt$Qa z?M0fcTQu9R;ND0tZ~hA1is`k_v3^*{!&8XtqdF4gA;<_NQLKRk-hl*xs2hW<8g#Bv zN*}O6YB?8BdRJ5EE^kum10r+0zW@@*VL|W*xOPMbLCaHMS8?GKG!Wt;(mq64;gZ8m z6V8p4+)~Oec}$2uhO($hW_se}h@z3aTweERj%65H&PaT!lCfk})Y* zF?2L6511Ox9itY&_J}*L;r$YH4_rokL3|Hsj`mk5Lx*!79|974;_Lw~B+_aJErP^M z6HX|mm7%;IiPM`HPa!_wngKNf9njoWmhu~z8CSpxc?^g+GzGLsE%^!#Q>*1~p~p`O zx~AeOXaSE4YcQoGrN7_ex*0wk0>>b`G?P7CsHsxHFf)V#k|B6dt;cY7IE+&2^h3#> z+#7LTk8{+61@f>5oMFX~>}qP*E2052kLiGGj`3kd$t{}ZnEG0{X|=3NYzg56hmprA z<{hKors0S2@dC}{j5s|ek2`3YBHCeYlV;ZfqJ#jZ7tnl0e_Cp}T)B>GdP1Y9RrH68 zBc>zbCY1>s6(2p=gk&A#@g7V^d|PmB!OW2dkM{yTqU$a4$`FAteRFY+FhYtzq%Qfq zRpa^!7CbIKz>O0uPpNgAZ+>q%Khy~oy^p0iWpxFtMf`PenJ{^UaN!F$RV=lH^+!po zXDX(~pMpAq&x~@2=_QKp(!>TA-#X<|vPYcNGASwdJtYOy?l_RZ(I z?hIBJ(XxxFe{@~pLPnaN(C+fde?I58_~r&`Ek6DNpZ@e0`STmQtQQl!I>t%QV|vVI ze)dfs@5g`fKJ>@P@Il&g>us#ucT*ZBl+iJc5BQx=t@w%0JWFhMiSzf~snN|2DM#1o z!{zt<-0xPO{%;zJ>BES-*#NdJ}02>E4Hscn#kL;?^O-&KUXwT=*&$ZZWy^ zRmQUm&@|M>uqV7cIOp_@{Xsv|Auh}@%{oub1QTv*d?y~g@8o^GU#4F@iKP!v#sg{s z>P^!)OwVZ33#`uyc6ybxdKRn$`5H}k6YrE^alm-;EP4I?l+@uGPuDHM@8kV8(k^MX zzEaK4<&3nthK^e-Zco{|{3Vbzp*h4gj`8doIn3h`WMo0&)5}v3-c9 zOEnQ~0%CRa9&jb2Wt)a85E+QGOe$+6M0~eKVxg=qk!*%@ z1{w)nEt$fq0aQoR1t%VBHVNAS#_|v;`&j=2_}MPF32}P(rn&nK^4%%z_E*ruU62Nm zk=#2%bBb#tX{kt>Q3Ci-d8Q(YNM*Nr2O1of!OYO&$bCdL;~Rq<(ee!Vrf!1_gl2+o z_A661>|kjJlMC3OsCwW5v-u&uxkDbNgjl1WQUbo*#cTzBiIomzwcGe0DBYh36dWZZHNSWV03JyP!(PX*r|;p2b{p(b)i$o-_Iw>8!> zFcq|T#00SxO%7=chr>iGi`IIC0W-m2XsJn6Q>~wE%`LRwL-PdDrkegEq{SA^_6Z^$ zD>HK52I;_8vm8-$+`_EKibqqA^RLrv59sI_m%E5IjH?}>L&O6K8$E8apbUlF@8E}s zZ!!)TdymVNp^kZUtcV0FY%ml;oPrSW&M*p?uf=)>&acRniX=%RAr7=#r<9?dBjyy5 zd0o6xz}FNnT5u7{;8{P2_||Z277>b(Mvq#H_W_C2;zAiQX-LcK6oU}%A*Q5pAJGo6 zhxj%#t}la5z}0hI#{j9NlzwB%ACXfvKBX*(b_tn0Mrb=ib4(eoA+l4^8eNZf6**8e zpyLb|W|%doUMKYmsloX@e0P(QFXH?*esUM*GI?ylnu_WO_52TAMsSp&;t*mOaS^-~ zv?wl~5~eGX2&vyeWWeParVr}ECpAfLBFHzK*IEQkdxR!q+@tUmPA4!%XhyzK&LxJS z18JZU&?V;Y6S^0fZQp(LxYd0Sc^ql>(DDJ+#~#1+3UaarT_L(eWmeD4&_7B|YYufm zCYd-#qe@s6v3F%2}c!Dx~lxDJzk4xfYL0Mj=(4%@Aad&C=zJd#xI5}lF z-39j{ID=1qc6FZQJ=eL;^>;z%2JgAf^?sMgE)uus<}dQjx7XiXaq9$s_!v%hzw7Pz zv46Tp2p=IeH|V-enudUo(Yzwa5kK~so5c7R=gi;lKp+3bmN@+=DLh5jTtJ#_a7|5@ z=}5^Z*x3U<`71B;x_;?z{6hV^)yIFa#qV4IomP{>54d#7Z~x50w~a-%>%XM;S8)FO z=>1i?{oCKgXx9+zW5z-8p5l z|CgNHc>>RMP-9qJC9P-JxJG|)xOj8A*xZJ{StvR~^C>Re$8`%@|02VgCy)E+cohjdNOxM<$3U}lNbxVRzB_}m zjoA^wDS34P&DXJf3AM|h`^2r0Ztq3Vh~_JZ-@$iK`X_O6M3~*icMFEoE6_j2;(lhj zf0HpU5WPs;y^kGVVtw>DThl*BV$bg5f5hVU>!jg(;p}^nu!qDc&SlI-v_!<8R71QC z;AUVM9j=kbtCXV$Nb{#mc3&XQPZ@6B050PE1n&>QA8iuOBUVyHKQu!%CBqpa6`f)E zQhi<2;xk0C(xRFuxd1(-C@{qbN3*>I3#e6;gZDzoko%6- zab8*RE@B$M0bVz_4W%R@Eo0@j>Oh>HQ3g-Wb7Hea(TvFo>~BZ zNgO&#+GlxZLNkAbQm#_MJ|QlET8wD6hY)a~;AMfikyNIbbl@&fR^Ln3*Rf$2UmZZ4 zFSsDMnDJq(EaEccQ&Pt6^mdwpm~Cmj6}u9mSMevS;Ey*p0b|SRCeCN`xdn=FozF@ZvZJ+1}#%8 zcjTqRbVrlg6M@ZfTCpaPoFc3Xz8Uddf^kAhQ_zvnuIfG;9l{cm zZ3=UoYjD2Q@;8%Ow&qUol0XCaYl&vmZ%wM?*oF~w2d zR#5Qm$Zf_KB~8BdgWp9z`m0Y6e&+pK=zmJ_haIUuLrZ6@KGCuQxEi@$+kb z^4I=fmdl9pEM=V2?A;``Z*93>XWvo!nqp-XWd70N4I4!`=S`-+e#U-~09oNvKS? zwAv>Re+v07!|8{R@?pB|oA}*_k5CGV5#$ip+{UygEv{42J&+@s<|dkZoIfT^@4fA2 z1*mM>m=JjxS^Nn@a~078MCv)JxhJ@XNOz0h`qVvs?4Liycoq?Pnl^r!I6VUSbHf{pR;(Kv%iVKI}PmxmhlWW$L z^-!6_VW^Ib0d9@Rh>O$8oE_WBzfFeO8YgF@95MR3Z$m?zkCi!Fe0>*8D?@aYnlS7I z-}I<9IE~;1=N&mg87HU#W{4>5e2w5)KbVn65SIzl9_JSLiBb9qWo)s$hdISrz)y{4 z>j5H$X@nB+(LprGfs|XqWLixWww3aca(I4JP7%4!es9jdA&tiZ`X9A z1T1%$wrCMtNtCjOO9AKC_{kDo2XqXK+TgR|yu(Mq38*@964Fw@GU7)3q+%xW(BiyD z8jxhDHE2IW;v8W{Grxmx534!NEy}P%fBJ!1?35p4IBAer2##7L2na1s+6|^;4FrTW zl1H4oN7z2Z>HgAS%81|2_}P85 z54h%Ke27S=gjm1#$JLDSY!}~5aNZ%Y&}=Pg3X()|iQEf1Ofbv%PARs9F9S{cfHChO zW>~sFP7S_zBp7klZ*U(2I(7_amkEu7MsN{y^r$&(5F|jD6eNOcGN9}VU zvvZ7zC#`0H;9~%5&>_?SJ%Z+l^B!a%54rMpgOb-qUKYd~-S!$IB0g3_)G5}mNoZH> zM=3%cJtgfv!nhDeEIn4uK|KRGL0nr|?9BpG#Y{0WP6qt+4A-m?6V}I9FlLOy)7bjY zX@cS0Z1ZQNncd$krhUWrjRw%JE{=r(XL}ePwUH1GG2Pot5Clxy%4T*w-aW*MqvQ!z z6z8dv3XOY&`p9FUnJz)MC}oCg9@da~JlZT8@1w(t64x6l!`Y(=lDBSSQ&B*Y3DZO5 zVRIdRerY8{22r<9OW`6W@>wD?KK-{`TI(ZE3m$>#O zrGJR#tAyqR7(u(b9X7{of@y;mhjSUR6C@ns+!omqrej@*Qw9HaBU6i`m=eChD7D;y z4lLcbKkf7F=r{knhy3_2e}S@yEP00M_D%fO!*|3rXbucVHyD>w`qM8l*q`vbo9iL7 z5i}t=z&pH8-%3^@nlfu5Kk@trNp6>5Eh>tz0{4K^*Y5F0pZ@DLPOYr){)~2hlgWh_ zQ4yTah>b}6=H_*~eN6863BwJxUw#reM>U-5T<3aU>)hZy*SX%;5<`A9W1Do}dzEg6 z?j(Na7jFOAFFl5YAEDWLmTvnNasQ>aaYN($l!n`kN0%Uc4{jOw@n6gQ&Od(dBV~Wr z^xQ8@Na08DyC0*OCZ>B|eCx$-3%*Un*#gbCSv`1)&V39!6@KRV|BcW5El%=&;`zsM z_Ctid#|X_1p*_a=73Q?2ABQWX;RB@c8taqm{G?P5zELEsaHgn580X}&U|hV2<`?Pq{*2%FMBzt2^8#tH$3mWEdi6{A@b3GlDEZ6$-cP^CkNxTbS9-KOhAgiWnqz#krT{7JBV`{auVeaoTz8LVev{Y^HQhN( z$kvb-2M|Y`U%%xBx`;q7{N{gs54iWP+xENvL3&DQP?a~waS=34DTAUTlysBH_6zvtP4q`uY@y{PZ2WUd zzC+Jtf}etTv?N51@OlUBU*&gyZs56JpD-N#$4m|slY98a5x1APkgyamn{HULHTC5u z_sQ#hOb;l1g2_#~t-Gwp8>Hn`<`-V3U+z-Y*O9QohZ$+uCCry7!r9?7U>~H}%Cx&L z5+_T_xPzu!h}=b*2k0X@_cP=@d4=#@EYCtxciyaWQ8_i7vaWQ7ks zMOJ9(C}~pn=SI#I)Gr~`A|GL>2D&6l@@P(*J>4KeEk@D~6TwLWKVYdLPV**ZSI4Qb zLM{R{s0MUsNy%_tYeW3Z$bCmz2Q(F6q@A5XE@<$GMl?g|1I4Yt6yFtma20LfM}#-i zbe1DI@8Ci$EA#OXiDy{3h^8G(JHq5EHAUQdH1AM`3Co*T@$HED89qogPJKfu6MU%S z?_^IdQ=C5piMZwn)sZp^ln#xVGPH=dY7n86QE*;qra~UJ(d-Zr;zS@6NKjIPp9pc% z;hSBoWHbxDJ;BUy?J<*sFCxbuBb{9YzYqQ;y!#p@UmaV z7aiu$XeK9&$87dbMkth$V91D^V0nUyLx)Q+%)uOKJt40>HjKFWh^4}Cw2za!gsrbp z#yL7(#Z8U~`wziA&bZh{i($4z2n;8ore^UQGjK^MMre9`6etJe9^&jSJ|3ZEpQ8I% zF@|xV=@jYSXppQN+}KqvtPhHo%Kg@4XaesA34-&2SuG{gVS<%{i-)+l!Z{(24Pi3i zn?M-@X0n+Mnu8jK1M4vr(#m0a6Bp*_uuWP5UWwD8rgBv%V~3T7tPPf{nT!L1B9ZT$MGC@zgkwsX)kOt7&ZW({9Whr8 zUW!oq38oRT1*YKKZCqSpY=fHvTa>fMFnfS+-WfOGRgum1t2)pX-rXVVDZ1DvuMQCD zkf4;bg{6x~81eBIt~sKaEEre&HD>*Fa#B!JG<$qI;Mx%?(<1WlEotgBHk90;{Sklgvw!sq z=D+i^FY@ER3_c|^KS5kwAxsVtzrssFB~q5#jA@s$en>w30u%Qq{La_m&Vrp`wxsX3 zn0%WH)^8|{4K}u@U1Hw;LwvhWj9WkhuApf^(+N9|-{BL#@@0Per@s14i{`#tb=?)* z^Z?)8#dQnb%Uw;}xYwKy+3$M`3zd2*qVRUF7!Nv&RYD0V)%`bEsu<{}r6x z!F4b3TR-(?0m<+E^o#uXXLlh4eZWl*W$8ysO@egRWU#8u8l`uI1=ShpJ7IZ^1C0K@SGe7-?(JNPCO+n&4j5~+w{Rl(Msbl%In~#VRhy9eIi|DCc>MwwRv=28osid$QHDtL32Dx7 zX#)N#Sr0HA-W}5I+{MTGzLpd*YmhJ_vo^>`|y0zK{S zF%oK_qLzyE5Oc)#klc5SCwqkL1>%G-aiE!!3eJt21aqjzieKX>wDBP(2~&saE~;0^ z!wjTiJ2Z`0Hn3A%h-lxUKBL+2!BG%Gm%$oBLw!#*GcLo1zbmBCFm;Gb&;lYoE}RkP z57B-~PIGeKBH|EN7aqtHv^4nEfjV@|lpKIH-XDXUF)kH!n>fk%$pYsSpt$CQygr~L zk7c1`0TiUixg$h?xFUpC@YB24`T*n_u1l1~<%%8mr?{r_U%#mlrJg)&;aQTqhcrU~ z{}jG0NIXTg!?SqwaMK3M`(SfK8VDmktT0`Za#I&OtVI=?m?~@8E42WvMavW&0hSns z^IPByQvnf7V_np@7R#l6CQ)#58}ko{@d5quQ%K%I{Wf_FczJ=ibN8KJUL3JyMztlc zw^3Q(C$BQS^eV&21@gFyrU{HKdANe}fo6V}X6G(0Buc-Hk?Mla*Gb%fYC=<64c_Jq z>93(_hi>O3PKS%I-lade#_Zx(k$8%Oif|Y~$18*lk|^aF_`8)gowv#THsbE!nzyoa z^Uy&C+zH~}+@5rD6KkYUgfGQx*SX&B za(W+2#|-@*)Arrde&48JwMQOLaQfaZ3g1w>{t(^H<1||j-__jxdNq?1%KRZ^c#M=k z#NSo}@^6EF^4AW~@+7|7X8ZCNzb(eH!8D^zQ_6W_|Aiz=v4r8u;INqgkb58(s!p<`d6Pfhl9I#Zt?U zoV_!TO^nH}e~#fa(WezzA0sVp;Nme7YkJ30i>6(2D&)n>l;sOd4*u*d=H>tLQ-{F+ z$#cIjVX<{Ul#Z+p;c?pe_Y)>3H1k7pzsq>`49*?k;|}h>Zi} zdw6><=cMtQKl5ci`PthHr%&VL3PZDt$PwfnGzfm;XfC|OWcmQ{g*?niCkuwv6X@~) zD-&D|J$!RYj4N`SV8fht_a;e{vh-;BAiDZKPKHaA z{sV}6260o$8Wv>-FDHcQ>-fn{oO}Bsn1qpL{u-kK;Q@sKzkHoycWC?vXuD%H6*3nn z{YM$yCA$4vNXUfAL-HDOzlEe-d|V?zHyqI(RiO+GR+fb3KCmWEJ$ZZvv3(pFv!)ul zen4z9o^4vwiDvka>v1VUS%b!PH3QcH@|tXh?|Z^zfoqpETgSwB`evFCD*xA@o!usn znQ`?)xF#aOsE(F@%RbSu!*q^w5eX;Y7O>fOF7_Zbl8~DaB}<5*kIO zN?uLKs}9p1V};~_QUV&Lh*u>1f9$>4uVv?X=J)%iHSFn(b!udZERLdS%Wf-4cc;@3 zf&hsIQtd|QtG+SXvSc{Pg@L|j{sG?vhMJNHw-F~77SMJiTY?P)?xfwGB-)}VQlvPu zSi`9^?_o{f@aAG|6fKDBPfmdK=8It!nc~xYHXqU zE=MgG)0_AH6fZnWE$xumJlJ zRHhL>=u_}&S=^@b9?O8BbdA;v&{xTZm(V9Fn*A*#%^0j-A~c6|vxxXR&_!RYDA5t> z4NP`~(l@!sh^E;;iv&T;i^)H$f zR{Dh?C8QM4sUyT36*Vbp6g~G@#Pruy6)@&nU)%{G1rimK^8T)w5}~z7+9xYtLEL3* zvC|hkx+7w{fwX53p2kgHxhb9j=qF$F6qOWyaT4OoxQ(kQ9Vynff!-KSQL(uCvnW%L ztz1Ui{#qZq5ripreB=FJmxnN&6T=FQDaLk0ZAe{-9X5>4{x$R0RtWX|RD0)f!zs$N zXq8COAxYCM*05#Of49xAqP3zu`dL(Ty$K;y$)`^PbjDSMiG#y6P!> z_tH&|!TXjdKK;d2OtwL!AT)($$ztsOlPeelLu7Z<2z%xqPs+ z)k$LWeU!cZbiyb8%)*OzVDdGx$#-vS+hTMTKm8c3Imag+d-_9Xd;LR|ug{UMox~2W z+!kY5=}11>I~aF_gxAQ1 zFZ2aRD5}E^!s0CLe4BRhJc~{;o^j~a{s%Cv~q zs5BrY2bo7nNlFfFgnV>~MU&H2XP0F=_o3|tlwI%3u1rUgKuijmTa=XO@jL%`)NY3O zQO|F*CA2w7gp@Sd%4L%J4&6~ssPCoOxj?ph1#KI$;Wa|-iGD)zljRt&C)Ln3IVse* z(M5D|m3+KS43@aJ3Nq=9DN;YOuIc%sVo=KWY+V(v8?_R$oJ%~xwchRUWWGeeoT zq{dR;I-g`Y9uFk1P&T05KIjplJwaOsvZ7yHO38rEiNWG4kFhgsaRuKz04DDn$J8pO z$j~fM_})ZL87Wxe(vPuOL}#F_kh~#Oz1bPPVDmsWPAH|3I6_&EG6_vW!Wz*xpg<`5 z&z-Ho3fk>fQY7>Pl)Y+6k2rAptB)bP0*|(u_Gx3#0@oV2muytQ;YM%+ajbCW*sa*lam|}oZ5p6m|*QDt1ixQGq z@?*idfS3klmkS*c0uBiy{9+w9+()}UC7e_O=?Gm!g$gm(P@%;)0j(U_a2IW}o;hn0 z(R7&XRjk#dv;wg}#4U^76p|9!R47xS%r>sPj9#K>v}*}%L1>29!HiHJET=CFI8ST^ zV=H{Kfi`hj^p%jPK^)y;jnH_ae~HnlzaT63X?8B)yAj~9*{ejaNbLqT-$%kKqQ*Uv z!&cYt_A!hR8KAHr{c~uwz^IC(a-?;{WKd=frs)?Eq9cS7Ya5iAEfc##B#nDhy94nH zlzAPy@+?}neF8p|{h~t)q!wlCx1`?RL1s#GSdb);4=<6GZ@>R)(@!5LW3HRK$rmJ_ zgFeJ&(;FUNPL61%2c&9)Pzc!y08OY4Nm+x= zc5uVXXyxhrDTq^IIHH^9=**#)Xr|Dt6Fnro%V2bk1T=@6w9|7vw_p9;JDRJvFU4nn z`yAnS?#GQ!VBG{!IRH*t#7{}hE-F3G*MIHet(mAw;PGeh^G%fbD8_cU(ezz6>3mLo z`M*K5B_F=P=;ZV7`uHrXuqU5|^dKqS$EUxz%OCUdEx+npQCi|YqSxr`_P1A}NI{s7 zj?=rxI@Ym%qIGQW9_v^?>H3pj-RCo3d=*olV6l5A<;E*FEopeG5{mZP1ya33s$b@7 zzkd4(&%vvk*ue_AIC|FtgqumZh{=y|c8$r|4StyPIM$EIZep?t`RMjX*7Cst-O)XC zY<>tk_Wt$A6Jv}!+b77@Z=IEv5Q>$H#Bi3lerhqKH=^_QBW!d_Ob&wU zlU7PdqDi5|w>jbPK8y>P!48?*C$d26^C*{*4W<}V(bWS&y+vC&LdqysZa(Q?Y)gf-F#`A$B<_dW>@@1IilAutFtG zN`@51kcRmA8sO39GTL6lWYfM#tk?rOkVp#4^mVHd5s-+)7L!Fxw%6~&Qc6&t>&gHC zAOJ~3K~yv;IRu5)ZEpy5B@!lB+hPaT=&BQ_c0_6oQD+FM$4$6xV$~95h?Fgw9P0*X zlSt8nUc(eM+RTxd6Mfznv?b75&_;S*bSH#P;cG+c`UNeU2TZOoSwx~nqD5qcQWYr~ zBzcq?BVL2bP_{wGQD5R_TXde#`4xtzpC!}-LN^8}NFhVgkgymbszvEB#+^Z%3BEax z=v9&=M9+cUK1Hd97#;9_9Vo#(ePvkF57@P&bd2sAFhEj58b*#T1wovQZfT@r!04KQ zA~jkCB&0)Nq|zwe-QD%_f1dY!KkfHv*R|`~&hGo1bD!f^6N4ZLiLz3+1AfhxbW+&K zw*9Q0(cHr_yq4jn_DX=PDY1a3@~uiKfg{vT+oDeo zhE?tIswQf?caPx}r#Y3`vMvgtgE-p1?0&t{{Gt^)oStXT2RA+IX^ED_Qly#e$hXqj z7Gd{a&EFVaWAnZhnOn-jBj3ezmVmf14ckd{Ql<;LQzd_))C8X-Glwg%Lf$-HLZP6X zg^Z~K!{tgoj*NCEV8}YmC-bGFFq!5wGD)T5p&=Pc`%t@=(#}vT{{E64^ zoP(Y`XY=Bm-U~H>QNjZ0PoG$d_79Y^ct9ipQRYjIBxBiP-{(uc+jLmdr<2KK3oA@0QKP(L#e8yjV6Y|WVP)+UkC z9j*ou>h`S1=8CBH9#ejfqn76JJ6%fj@vUN$5(@kF6kF;`2cMFLAt)Oz@HixYZIS*_ z4(=t#j;85TePmB3>}<^nUaim!0Xi0r&W6l`dM3Q2uhEuieN6I+gZs9k%~*TvwJ)%J zhTB}AWHZ$n=Q;sDyKJ+4L$7k~_6CZRf|Q{Tw*ol+*%o3c9 zF1w@#_LZ~sX9?Gejc~=-T#1i9t8)3 zbqI(TZddo5{r@XdGI{vr`M(fK9K-S<#?wCH9WzSjLeHi3#Khc*Y40AVv44!I`Ir4& z%f;E3ugvd`0hS93AY>nJ%>cPsYsFshGAF?a`bpV`o{M912|-izD(LQF_p%=nXU@&A z+fo*H(6Xv(_-6?xDOKkBh7NkSM!Fr_%sz$xZy}yvCuxQXxkuMMpL6MTH{|FZ65VcH(6Vu@WyvFdP0t@xm)7lFH()= zW}$R2AZPYLCa;NNmREyLlnETW*~U4}WS;(G9dE2AMDY;rP2#ljIsNkbq2_{l$U7f7 z=r>qSLW_x^OOPtTl1nkEGK6`-IsI3@DiEAcE&}@*?WvRkskvdvD;q@pX#x31v~NB{ zxNCeaN6kGst+$n?*ArFzzgPio4K33Fd{x@b4VHIyiyzlcMv`}u zt;n?UR(?`5Qbu9^Ma6SpVLU%--f%3`t4hp-xex`#(I$!q9+~bE7!0H9 zqs9#1T&-VLjxh1~0=+>qiFNrhte}s2wW?Drtz7YF0FcAtXn+>Bpvpq+^NgaDlh-WQ zNHYz}`HJBeN&i0BK-P9*JeQy`JPJ{rx*6Iv*ZUzZ!Nr2@oY4Xa zHf2rCPR?Y{?NbF4e5ND@hzW1D%4jmw>}l3uNBL@NPr@(mXx`hz zSQ0=eGRM$K9Hzr4Jm<(>GD~jtFRVrA&#kvQc1hlfRtWlC3nOh2z6_b% z1Pee3avTQ7ZipH#Zcxf6$5^ABehm45O(|`Wyd>x!2aV(wO$J2iSfjmThqDzn#7c0~ zqtJjtMyrU2&ojuk6TH12|Kaj> z4D&MR%P83CO6igtC0hd%YNw0dGmM`kviC2G3$*Rcm!HtXG{3(>SbRj13~vL9nk8dE zIlJUD6ZmC?^G$?>Vi^nv#8aDiYuQCJ&YNE!D}S&dBxPbe2?&v2#>M)6kpmY$9tDbq z;%v3XsgPd~WfjlQ;RA<%&B@j;Z8DA8LS5T@up&2mQb4MS4>a}dO7%l@JJ>b5!I2{G#zYTtuNvIg44@7O;NLFVAYe?8UI ztSmd<-$DNH>)a_0HJ)8D8h^1morTDiM0h`$+e_(SYkliKgngXI`-j{2b*}Y>5?`qF zzyEgeHf||#-^Zf49NfSYL;ggWBjd|XW%7Nth^XN%=Bq`rN`7htTaUyyxC0i{d2A4> zu+_GtnbXz+t7Xx!A4}9dxvxE#$!FqWh8DXN_1qhGrEdJ;ck^-OYH z-R@$(cLc#NMvYl^LfRIgw}aJ}+l*YEAU;zpqHL4{tujP~)7#5Tww#KR+Tgj-# zr00J%2XVW+XO_=-^Dy$s)5>$(KHx&eqBg;kPra8-681UwBS)dOI-ypp+IKPSX<%RQ zUz&sXa+GuUfZ&)I-<@6EKlR-aNsX=6O*KKaW3QP++Z}4CE7=BLfs1PE_1yf6Pl9+$ zl&$U=Rj;`Mak_{sWii$1$)+(gB?3|ZYzn3+K9jrrVBLE8y$Y1-hZT(hjK>~Jtog>g z#yU!4-?ge=t0D}1MG9loYd%WL|D@-T(C50UxV@>~PcicW>TV}m$&e>Hc}w5>1i$!5 z8?sGl>W{=(ftP!O_I{{<)j>~72NWEwqCGyo8`d~$F>KUTq@Q7D5A4vC1_EIdfB=%C zC+2_rPxSh{#A|>mPC9IEgw{Q@e;NMp(j&?@{(hxPq35$%EUe5Af2)dOzm-`vpR#vgA)M=kHl~0{_te7yTMMVywK71sERYNJhP4nE)#byT+z5WEspUv!jA!oZ z_j#FYuV}KpMo`6wHILPc4=Aqyju2STQsBx;TeCjhPV6|_RNze7~$+`$o>#|^WUS-NY>9|;09|dW?ky>zR0;M~6<-NGt z6v>Jxgjd-CB^Vtk3QWgOAxL$2_BCZFj)C(Yt)7c z^@YrZWxO%ZU-mwVm75T24B_)qioZd6PSrWjWY0~OmwNcriyoGIGwmlPrGU7i2SXV* zKMLG(7ep^yj8Ry2N=37YZO>7(!D}m%OG;MxWfF#Bkg-8-v`dxfK6}wH;+pnrrjXQ{ z!lS+){KUZnAkCQK9KgUtZ10B4(v7yE;w)#GdYMhhm(m=@SFTqM-6nM4l15)CsmtI5 zq?@TI&Bojx8k00&+_4fM4>u2NJx}~gNFJoCmgv(kEvZsu{O97L+M45?p^Kz0&)XfB z!zg)S&w&hkE63)*>|AM6Zs?wN(3*o#lsahht5;R~y;rIP{HL|sFs-uC+RmzD1)w5; z$Qj2lDACFH({Nc{2X7v)a0C{8n0zQtaiCN(^HYo8)NPZfY_RwXOc-={g&M&^W{R_g z_jpx)YxLW;^^(!KB|@(L$<&3L>($6s6-nI)hukfU0y-BlCMo;{?4=Zzo$Bfb+#4IX zu(OHWJV0e;qphmvSBp3z8f}lZm1C$eYP3NV*IiE_AzX#RLCLApc{v>3F4Stzr+Y45D2z8k(OjZLE zta!>$ko1c&w*ZIzn!osyr(+AmRh5<^lubiAZXXX4>OOg#vFHQ!iq`gI_!$~!NeT!tIK9oVSO zN$Vxkz4wJ+o~CD-W52^HgZxwsbTq<5vsZ7L^7tv|Zl6tU8a&YHW;?BkA+tvrwID|_ zqX~*-997Zm9Lf1FSi2;rzrUL$l78WLA|;Tk72YS?I=|h}R=WAnZ)Wh5M5M#&T~vB+ z)B@XU(vj0HyZ-!Rg%_rUrGw=3O4Z*$;({E)93RM9N%J^sNr!jm6l<%-R(b}I-zOFG z$VLj1l3}i^;!&B1aS#DAn_`TubsGDXTW6piy|Y=IG_Z!lfL4JKFcY__sosQ`R{cb6 z4b3(2kJO9bvav<%@pe3MCI+_04!>kTccr4II9rc0Z+@M~9HVKAIPE%XHhZ6PjPGKr zT&o3~^Q{4PQtXPPH?JtALBr=V8j?sL>=b^@%Z0srtx|QCUJgxPfY7%m_}JSviH@ZU zoyWHZ41Ld4rZ5D5_NqClS5yWNAh+ihGVZzuSPE0wk>PDopE*Dib@ z*gQFCbgU~>3;tHnf6$q5kSlYvzrIhp=Je}vucl-vDBWLK>VcYj6E)Xe_sDixCdv*N zW;$NFg*p(`I`wu(i9V%1t_ID`o>NkV@H!wRBQ#icJtrmP{~m7%wuKyszusLKSRMDM zJ8hX@$!2@TUfDRQYZADRHCOiEO5JmPQ}n+T#r6gd)@MyJx0kU-1NPrdxl1v#u&o(@ zd-{LMkE7bIRwVvVGmc!*@ni=>XNtYSPSC>G1M{jgLq4w%G~GY|SiG@i1-VTH)Y0GG z%inx>Z7pR)bRgYO7PeO>7<7BO-wAEZaq^X)@=a~uvSm}5zT*dbrS!f`oV)}}aP_(CD|=;(#Gz!G+h2@}=Q^?0 zS;}3-$GHY=E4{(i@8E#SLHrNhjz3l-uc%>WUrO6mP*BC?tb%jQxm%f>56d6qf{wP# z=;?3}*Yq_}o&#!h$GBJ1*4un$*!$q7MqNIPt4-d$akijvmmo|-X2oZ$B;+5HTnL<2 zVq~eCXb?ffnNSV}bMz{qr#Oe$j=GGZ6Jz<8u&;4ekOc*5V#PRfwqlp2N#zw_jx4&BO&lCUCDj)9`}{{On-?FVF`aTK#eq;Hd>p174kn+*$TQ=>iGcqu zaYFELT&;@TI>oyLJ?%4AuMe{$T^HeftB`Is8HziCFsT6;c(#x`{<7f*Fq*IFLGt3j zTkz|p-pZ7<;LO&7#3$1igwNQNmWkG+l;N8jv=h40>Z0%;7=Ad8j=-|eK|KdW4URYf zGV8q0;V;|UWYJYewMz$_!i~T@qp>`#IS!3+kvw`uU@|I1QlS8N$!6ES8ra87t0kbW z_inq@gpVhnpPY0+re`&Zc7}~@heipd_Bym2_>bebPt4$B@-jgTg-M-)V?+sFin?2k z)(2ICI9iSZ7e!Hw&y2>e;C|{vsXMq824ve5@a{1$;?iqPq+d$SvHo%2ofQYDE<3;R z`iU?j$|+%h6cWVbth$NnwmS@ZBBH<$#W09){w1Ll_mVcFApKJ-h}YofruIOyIm;za-bS8DewKd-6N!wnyD9flq9j85YZg`mexC<6EY{D9=X) z8I*@7J^P3rOO|3O1}~PgG>~7fsMA`$D&+I81X4OH;(>d*Mk(032e=x1>~D7>I?s~Z zXb9pj@TGVDClsFTNDQ_kn4W?d_*ZT8f2J=a0Dfim$}fh0#iXNX95Wfq%fiYOx+2P_ zZ$u*Yl2$A7W*Lq50ZmMA?&(|K=5Gzc25=ie{%_Lo5p|wt4cXOiT#~*8cN7QRKw6I& zjMGbU>Do@47>F`+I*Imwl{qVjEAHU+k#Q>cz@dNetC>9gKHG_sg%UuhM=qAX5l&V$ z+q2COUy$UC?as%#W-H!8#{_EaPbfBxVBfjE-tge!O!@Y|+j(loFfV>oK#t2@E;Ko5 z)9QgC_WQ7!d@8y)P30R)>vp>&ZWm8g`~Hkc_W{>OuMBq@X(`Bb8?Wy0B5Nf6FM;aU!<8uad#Jm-Ftxw0Wsh^2Ml)K}uF3qdKyvwt}7 zuy(G74l-RGrvf)=*1hmD!+o4WYo#JYD=a$xlDd()LObFp`5s}b?NM!pvlp_-J*(e! zy*u%KE_-dp`LTu6FuPdh=$QuGVS^ba6sU# z!Xa|;#g9cK(m_Z1q*L!aU*35Uu^SL7=srZo4zRW`V3<_%7!;f9Y9?P&y*MOyglXVtn`K~%As60!j zZ2cAE^VuFMs!~(7f22v{<|V`I_6d)nFC&@~U#h>_(dIme51hW-c4R4(p62R50eX@C zyYoo8QQTV;(TyR=mj!LnfKlx?VZ%@CqhM!}RedjScS)urx=LQ|ns97M4x6F1@B;kOc})b~H|D^*wV#sz9;yCyF3KPs_Jx<`r=*6h=M2`ZO= z1=_xAZRYLEm1qs;fJ~+D{xS`JuOG(QYt$QwNk}mUvH5c z=6z%zXF!~86hhE8`&OeS>>h=iQ!}z+@P^@e8a3m3k*{VvnV&Q%HR5p;N1|) zXSEqvI9%V3ycA0qQyh>61?37&DKHT%Yy4PY6qA&yfeHP%sh;N~Om*~|a$9JQ^r@Ai zLln@Bj2pC;W_gXX2_N*vJ0VOW0zAF+LjZ-!!`|7G>af!Sck@n(B**Sf&HkLQFdBFvUdaSrc_^|1jUXL8ccJUEdv2yaVeC^ zjljT?eVl;XN_wwq;Okd*4vk%}4JOgRr`O#(=ElS!1gUjE*@CQW&MvPw0pW6XQF<-( z>YnsEq1XI7V{b^5nNidQSbaRU*kqH2Hi{z!s<_(&*?yS<&6XH=mGec1vjnh)?aVvt z+or2?L*QIgu!@+%Y||@AQg;X&-MKV9tKd{ksX^}pI^;~B&53p*pQ43W-Eh+@mb;tk zN7gvq^9!_#e{^Zq-->A1h@K6NFe7h&3qDXi+wKs2o z134>+t%vM)khx&!ZF*u(H*pZI#QW8QP?^nzB=qX4rq*M+Q+@?lQ8t)(lh~ zDe$&s&w0)&QK&6*v$ozm{2~jPXcn+`bl`J6Xsdh#x@p&q3Oq~a$hqxa7=9Cm+~@Qr zUBXaI`>%r+8+LD2CN@x)-O|@Ju;ww%~t+-s}&=h zHI2HXRC@x>KRPIlCK$|KWM3p;f*)a;KzX zHtLZ4O)a?9uVDk$sxtjkZ?JD!fbheUk3k9No6iH5eeQMdJGC!s=yf&77hYE#pO1Az zh%_+s4dv%|0=Q+*GuEY9K9Q}_^i3S(n=~lPHAn_2a;$b@X*DS9&Bbd(yH3$Hn}uDK z&3ro;Lj4Y|fs>eDVHWnDVsq#=5r%=s@Fl=AN&nLOqS+309)1Tt;XFr;L}RK4IaW(|L?wgZ>PJfyf~y6PZn;)&(U( z-Ym|TdGvYa$_A0O)yxvfxCKV}dY;;<-!%*Zurl|^8Yw`s-Bwh67cwvNl)hmL=NZM< zsv%Zcqhw@`OyP;HwrkHT~P#}zqO zaHx}evYu;l{Hvogu1W@qe2zj2R?gnA@ehqx&k?uGrgz+}=R41tulnV=8NL*n1nyn` zd_!(GLsZtCtHF*_L?%QZp|o{>LfeEyfQfp+-MR6@DibO#z z1i;~#;n~E?iT8XruFv0h)L#ETFThpRErM6nP8B^Ex-aqEB_YM!AcL$)xLdnoMHy^J z#t^$qmMP}>mj~wc*-E3=dB|nop@wr=s5Y!Xzp^|1*(o=$oDNJ!_wI=c>rro>_dv(c6cE?yo!iJyG zN}H+|^5`k|YtD*e&I5D6QuhJA;yx!!HV4$7*e8~=2A^%g_v7k8zWo8l zhpUGibTQE{>iP@yXA4>a6{T+yqB}{Qd|I!6Gv!>>MOkGjHtwdqu9HWOQy_EBCJIOu z|D&WH`92!Y8~nBkJ#i_=HC^cqMyr4P{vzG|>B6lq;r;)GJoWz%dBK+v z6TX+4f|hge$B#qjbSh6Y4_PLT&1%}(mdx}2RPiv~MT+z~r{LzSU$`ULTf)AidlqIk z7dWB4-D2ojO<`1BZ0~y=D7%;61~oXyQJUdT!oZJ0DQ@w6om!{y7MMFm+$V19J`99K zwVg^;N@}+FL$amx>7NHgxI9n_XQ^!PAsH&#cEAxL z$?hjA+n207GHAPV!e7w&)nuvy`oAQL;=z;llmK^d$<<^vKeVMK-P`a#M06dT3ym#4 ztm;nv5K3^dtm>28;FxjS&GG?1BBz|8_8_`s|6s$yO`^x`HapU0uynx@iA)Qo8TvZ1tbd-)Vf6u*<|RdyC^c1PE0&nc&!S}9mbBA;lia9O; z{5}+(i5WD3#*I2 z8mRfB%lQ%dsm&cEI8;MKf@&aOiU$Du8qgB?wc@kn}^$=YQyXd%3UwY$wE4Nq<8Fpg#3MGBCC324IDxIkX z>;>;1@Fm{Aimu}p`m{f_79xt1Cl~+Ma#!JpO=giK#A#xLM5|n{NUhQ$MWp5F&PLELLAhzx zN>y8&?6IVTe63Jnejbu!(J-V}q>ZD#Ns5=R}%W-e-DbwjSo8VsqsQCQ~`Z?+)>>ME#mtizpQy}b=lKLTlLjrG2_ zs>=DMGkRx%Cf*h-GspOn4nz|-KKt-wDHH0O7913mg=!@I$kMrPS+m~?MNoh}LVKy+ zoIQg+_?St&;gIufjT)Q~`e%M|k(MzsI6)@QRfrj(QcA5A;U+yfPc2Yk7ejApeEElk zdD&4STF7@npH3B$-gC!1tc+gWoltL;Zz_1ie{Jeij#Fc3(-}9Hwlp-~{FZEdM$qN9 z#Pf(w_o%d@8IjGb)ImQ@1spM_%bw0#6oE?nkR5sd!gzW+>Bf4<{QegP3#$g+!maq< z!Y}9+V6eQ)mRQx&zSkNtzDjkSR>l|fgtTb2ZFKAbCiEuVOvd-V2~PKHv}Mxo zlanu<;h%~*At0r_aBem;_aD)-T}aBcc(7Y0OgtB#qSE9`VVx(Xby_>M9Wh)v zg-&>3-L+ctc4#B`S58*bZTUHV6ts1!tpQ7EZ#KKd>fKe3H4Dx~NdvAvWMBT= zVbVrbEYFyKUdK!D`4H%uH*dz(MF&5-cvjH%Y3O#zqLoQbH91F_hHusTNgrUxU--_t zmpJIvqOEaXw;-FBs%%3n@}H15Q7!O!XhDcwUrnL5d}0>2+#So~0DZd9>OXig5qu)X z?5kzx!%)x?UVBX|z7R3b^)-v)mY`HhnFtHrrs34Hs!6EI{cs$q!eMO+!-Hoe&wW=O z77%g)o4i1pzY|de|BHuPMtkX$P&&WNTP5Hr8WlkB0)ZOC#&44J>Ar#;4gWzaNPU?xwJS|l)UNzP9w&*4mk{%TU_#20R2EmU1 zN8K;$B(8R0U`$U%EWGL~T3Kf2O;}|Sn`EICiGAo1&VY?3ED71EW5xYOVNq@u&L#Rz z3+$u-iB|_>Q-V+Fcm~yj;Dffb5K702Q76#t?ThRib^Sc;L=&@NLpXhwpJ=LZ3=Yq~ z%|8?~Y&>7@_P9DaICyib@!bSK-`-C#8v4z?kF08fHcNCLE=kAR}HJ{L?Yp^s8yVB`~51V*JZ`8s9Gql0F# z-yCR-wgEQo|(O!Y>#xk?s=w>f%-`89P{PV#wk4^dRce?@6OF z%5#LSQF({anvrWr!T*i8$_$Q%ImS<&xEODDiB@K^^4UZNSIa}ZQV&S1VlfXY`+(r^ z)J&yDXER{Yr##8FN%|5x0csUwb>#v`Ue+Wr@}#LVlFhiZX}^L}_&~WIzoEJZoYWjz z=i5!OO*eKN$Zvwl(}~Nrw!VkACli5!J`-HQzi9@g%h5`rUs+FLYDzx33QxEfU=Se8_LzmIt#+0z%a{XFaqN#h)Sm#ujXI_Wus{ zake@L@!4S5c)J!VXk0%>S1TJ@p_AtvJwlA4yd+|wL7(#m`H@2wRx;~Y8(Ti*NF7{C zpFPi*Z#nhnu596VIuddDv5CGeu%}VF-4&|!iSu@SOoP3?2-qVWnm^yU+IAhh+`T#& z)_IPUiWS9r1dGR6^hKBulYy=-#aG$sscOa*|0goHO<44S)OVBmCFUgQ6tF|r4>26jU8E7r8~;|-JZIhbHRgG=E+`S* zlEnOFgx`t~TS%p2efA?(S)OPj=2_Zw2%w~T$K**lotfHlk4J8oD`sRjIhVIM%=dHV z&25RV7D9QtL53(X^+qtBAnRXo9&^QXmJXX9S1Iz`SbPT&wQaUj7f&)irp@UjK>4KN zjB#{p;wW7oDWX-i!`kTq;uI7;zs=CmI*tRlwmZY$U3^wl4;P_cQ&X0VOvC7~KHiN& zBQ1~NOfD4DVWN8$?%{5Zb3xmHkLGB4-DAgsb6U674y#HHxgff|$eK8e%adT?4Z@8F zLSt?-^ez$C0s2g(B@PsE40ZtM4^&Qh(5|STGdmsZ$#0nYSG~RXl+yRaAac-MOCXm$ zL$>Np-;ofxLi~9u%*G!FQtMB;$NV}tb<CIgUB4jva6mS`vD`@#gq?K8? z?N)#G9h13^MXjQfZqjGD5{ymUGOGHJV-@3i8`&g^Oe%KVLJWY^H{rgNX}AutJ{0>W zM7_9@D2Y9Ak!uy%XOjVA>cCq#PpY5Azl^o6 zakDcfcNlK$NSi^ptN&16L5#5L=l>DeD`%}DwhtYkR+j-c4HtChZxFbqN&9$J6%2Fy zB@76pD$HIk7LflfG4C1hi7UJ#_u_fe-8ubd0ea<6F+X}oTom3C1kJPZbY@UTM^ZAS zV5hK?57(-EqR(_DbHv&wo8$%Hj0onH;RYD5+S+VB!@zi4H{^Aisug}#R83w3 zD_^UA#F^%|yh%2azSZgsh?|g{-X$|5w24c>H2kBuW|uZBzI2@lcc3H3;R$OD+thY) zWr|HYt1YSB?=r?{Gvinj$W>q$dRiCGunVB>(Q{4paX3*J$2xWo= z@BMX#MbW`Rt zrNdhUQZ49-2l+Jon+KZ535}TfKK~)s?||g4x14^O3rizVVwqEx{rmxWu$)Y&+wXV0HI8gN zz>nC$f|YM=Z%55#4%ZIm7Rw_BU#6yT{nWG@+fmJOJ=+6^8f*qUrBI{wT8-?HIaXiJ z6O`4Kg{sWH{&L6m)px@n1&(U;qj3{%7ks;<8?>Uv5PI8H!PvEx(~%&j?6&!`s{Ndy zf-Md7(ehZwvw8~YcK{s8QDjh&^Cgj(o(*ns_1iQ3(ZW;S62|^6CQ<)@pzZi$+p7HC zj!!UYQ%Bap@gYlKj)P;o)pfm<{#V69o*a{f)xN}4_%%`gEHDfMbnRc0`k37xb9O~M zu$p2dF<6e1wYPY^Q(u5D$8oh{cOx>jy)b6h@>b!Z`mtUma52FGSLw^H-uLFPTH>?) zp7#XER*hp>SY!GR+WeDp(C>qJ3$9|NjE0%gK*_blpD zGge;`;%R1+!Y6=ch17)Y5&m=eBfCbesb~00=Q#?oE2CT~``@%vSHS$lAnwz1 ziK*LPKAa5+ov5cV%x-dOHN*pdFZa_0c+e`L>v>!}F&xcI6PCYyv2l-qT@v)a%;mTm z~#Gzd`25JJ1Mt#(<#RF49=Widcx4Wo%?^N8Ccr++^sFe%wBpx~dvl7~z> z^Dem)*;{?Kw;U8F3FnpQ8z-4QANlt#y(ozqZlmBN1K^r$PeuFTTVA8jU51N?t;5|7 zVc#I_sjyi0CRz`CLzlO|6JOO9ntX_?@rp4z$gL{wizSFtJGA4dsTStJMK<SW1%5C0MOla=?~9Eo#{9UZFIO;-20EcnZS+uep+ORRdW@=BDJU~tCx zi(AnoezX2U+unXz*b3QJ-uX%WLLB%<3$cMrOihm+H^uRq1PUnnHCF0}wa?oY?c;#- zf6Qu8R4!bHtu&3_*VGX>RUWwYSZk41ofE^ZJbw)FP@Bi!w(`%+3uqPg@_v6-G+D}9 zXd_ssV}7`ON+FqK!hP-h`D$r~+BdV0Uf9W_iT$$d5ZyrIyICo>>lI;BJI~Hg=JUG$ zrl3T9AjIJSI}fnkQvCjE;rM6phno)_|BH{!f{b|0|2;muXe33KdwxYv#?hJ< z(eSmN^_EdtFdm+JHdx)PIG`@aSyqvTOVPm#t!&$^QCcg|+fk;H``1Rn5xa+D=IsO0 zNr~RV=Qsv0G&5ew+|C$ZZ9lx4J3qLaZfoerYCEApPraU9rEHaf*Gdkru>$@b4d5!a z{G!Nd!_)BHya)owHs^c~=DvED6>v!^Q7@A}R9dG)R-L0Pczrv1a5o^PzwBis&K4nA zY1&cGxaQKSzfYJ%@4R8Rp;x|2xG~qtutm)LaHbh>9jl8v7qQt=T6HCVy`YsrGhRK0 z?qA^^0?ZuDUEj~bb?7uDm;NYk!!zN)tpPdG@^5;C{pf5NQ8GL_5I6WUnZ1C>d1-Ax)JuBj1<4Mjln&hlr6*aT zZXSC0g%le02Aj3y-(DiWvz9CL$SPm-MLxf^3d!;JtZg`8Z&8+$NK{s7>W+AI2#L z%WDo%YHRWLYb1TT(Mz13l@O1ghAn zj)=a7RjfawP{PRVg z<0YZlzN(VznbZVifS3oK1HEtx(~@7XifJZ!ODDgUn0yvq14rYpC9|;ZJ!z|q8%Q7a zeu^SfN?g2P9T=g1(-7Uf_NJhWo55)e%?bXgT-gRXZ=|HZdoudngRw2*v6l$DR3?7l zsnwmV);y|)y_R2?2tUZ{AwvhauqK($Gz~P!9Md96PGKK3J|t8+eo%0LcCH%EzH@VS z09uUnH?Lgy{cF>p*avoVAleYqLzL%oN@j?%L!ApG;JLxG5nLI(mVY|wj3RoEh%sn0X}hM3q{t&u#J|uFI8bh8%h1Tj>|tGzc_Dd!#{R5 zYC``~FWkfAV(s_Y|L7bo=KF(Mutq!8b&QDh*I&FYwijr&bV5s4*1u2Q9an(`5y>;m zRrNAZD+N3U8Vw1xY>2zNyN373Kak42f0UM=$E;&p9FfNn=b8Z>lEP>iIs&rled8ws z(&ZnuVv17I|5>Dcp^{|2|2;k&dBXOe<_w)##sgZ<{O&hhTP_AtLCisG zw{%7a99H>b$Ztc7$C(R6^b^DK*X%()RFy33@ZQc1+ki9ePw20&xzlETF>Sud8!xk+ zyY|!4s1{JFHN_y3_8^?46+S9%y9ef6!3y%Y;A1tZEA;_SKN_oRkcoh$xoQFf0;7cj zF;-^|u56CIaEE3in6l94Hk6$VJ(oWpTOH=KzeG%)vc`;3dU-?ui*AAm=Tmd5r<5t(tlsuJC1R=bExV|57Uzs`E$=@QAyTjf+L)@NxA9e4G@ph;>F%t0%FTa$h&;^D6yDK(G1% zRhKzGR&taY3^V>%nuli$mU87K_VNl7UPcs6{TA-Ri&J=yHewf6;|XWAlgYs87lU6} zfMpf5e%a6{5+_?F+l3%%qbWVq`EpgQuDG0Ej7thoZ9XH*3=@Kejpv$Qe5zh!=0O%* zub6u~$hEIaUrxpS(Hh|k^GXAeZBv0?0@ee+^9kRD?0qQMN(MtR67jsl2F~^RlBQ!e z%?udBDe-VL`1O|UiU{UZg{U2gP5FSh1{$lciM>OsD4c2uZ@N4%Z`>5|%ly$@Zd$=v zKVF1i(7dnN3+a`=*&%2?ZKP+B7lRPk#Co=$m3IN19j^mVtOJa53=?8S$}5n|+Hr;i(gGF@LvNjieT#j{5` zjHedDqvoh)$5c58na?fDb&9CU1?vz6dOE4wV>jY(P2Q7~mmu~iEsKCqisdAeL_I0M z9c{9Di7v&AC4w&LR^>g_V|%4Yh=U@j&}#?*)**c*1)-N(|fZMLF%{aN=r*;}Uw6MkrV^;iANE?xuKa6$bEi zdiqAGNN4@HVy>bk-?skZzamqvobL&4>$=Ol4e`F{gs%f(s-Qjv0E*hKB|+eA!KQf_ zwv^XF-y$Aza|T;3Yj4iH19p;sVI|5&?DB`pYA(oJlz*zT_-ET180me@ z*e8Eit8aMlRQHR2XjJeGz^j>IMD1~6SYuqK=o{a(2?x?N8mxxT$6@_PD?ZfyUo6BT z@%3-@-z5__>s%Y)r*Ur9Y4H3cdV}9AMRI({vbEgCM=>-1;KT8-&4ZX}lzna^S>DP#V`}b$gUeG4y%7y?6SDvsB6&6v~Z7_>{#L0vFrSir#)#ieh z%}-9w{&IUDI~QJY^eB{kNOihU!uE+1XUX5C<)pWJ^lyyHgr{%3e&cr(=+@-uPh|W( zB_3LTh2=z?$JxcGq*U@$!F55*J~~es%u>Z*NkjRSPYTuEkX9RvV^}d#k}7jrNN370 ztJ7)xFmWx*LJ>-su}-kJ{GwLA;s4O|l|gN_U9`o$B)Ge~yAvQ#9E!UXDN>3SmjJ=F zXen;Roub7HEfjZm3+{UJ-tXRjnUk4wW-^)Qti6}zhnUb`;dcM>mj`C!>l!oj?oe!z z$?+pEgrrW5rAMU@qI1%A$o+hxXuciXOmXEujKc*U<__$#vj{RavxE1M+kQMVsNC}2 zKjCPJILLKEKL7G3y`oKwfhb`3q&kc*U|W-q0}0>RWBtx@A;*Y~vTu;92tsuTig&ul zpZ`Aam2gCTTnLd!TTzv05)qU?@~zG5w?A}~=MqN09VSXAN=LSDk~7Jm%|l~c#4|vI zBD4f|3HxGbA*i-M6QA(t=o@pwqW7I-E7^-UdB3RHtN6pyY5$N+<ObZ&wlytM!mx80I zk|aU!6M(i6Md7<8mVB~cWt9w`im@)yxVmOC{uD{o2bL8YDf%zrUdANZP`6melKxuI z53{0K*U>hj&w(=Hh9Nylbp%k}yD#yYRxzYPbKpau9Ii7Nh(?c5F1${#&hT&(_)TFi zEfx73X?fjhZx@?r&0bT|$qcw?+xv7D2$BuwVzO_7956bi0di=Ff*Ezm1Ty{>zJVDP z0qb`-Q{z}1cn!mg(nMIJ;(B&4bFhSJFhWNh!_2e}*|6 z$ZH@#YBYuMtQLQV2s3hW&8b`?9RIS`S<}`5;q=0oq17fy%n}6q{Ssp}O0A~(TbG%5 z&3Al9H`v_oD~HLMLHi{rg#;tMQZq6vL*f>`_oUa(>vK--ja>6jD zR%*Xjvwo{;MxS4lb^ZIcLA9c43*xA<)G?C#w0qpPC-J023iH5m<~Q-m)I`ztAZa^x zGIe=F4kcmm?Cr!18?!tT$P~J3Q~0UYC>`td)|eu~gCt2RIUZP&&G1NZ{Y04hFL7N{ z+4PT2H$waV@=;*$(ku6#(T_noEzRRxZq4Z>YWnxo@(YFgR?|7T@81X@D<$*yKTv(R ztSGwT8=3Xrm9DN`eG*tWduwuvUO?{D{2CaMJpP^Lo*#10O67m75MO+{%0r0sh!woY z3{AKbq4O3t!-6#AG zZo3<6T$!9WG`k`aU&Pn`$-*?fP_h$v?qOf+djg5f%q;5}sqUmBBzn+<+ggZE3Mi=P zkDQZiTfV>QQMAeZI340pKR+YJJ)imKiFncS;;`OxsUiQ2k9@09M*JsrL40p{AY*DQ zUbZzB)`?n0ACB&@XX<*_MzM}V^)LxAugaH?(nj@1a!L2^7@iBdRwSPE+t7JkC@?cY za9*9=aYAdKCWpsC9jSrUFxXBh9EW8|X!5;vLgwa>&;^ZPcyD$nhu0FC>3O^hztVE} zNj=XaU(VZJ+zcGW|1{o+`e(Y`Gh{~JCg|uMy1E!nPF3hpax-3?>r=?Er7rs_Z|e zjE-VM3Z*~jHiTbU8%KYo*;j~fu!0z_pgsj6+Dps|k{wRNh|7gEk|u#B-(`_4At2fZQWXq<>>SHXqFhuFovWU6JF{+60AN)!eDt#fG`=B^HD$)YLuXnyecD% zLi5pO%?XV&pkK^HYXh~0R$0q=MZtb7F^g(~c>iynhax(DEIUuN+{BSs`*?83cH%l% z%Bo(pZJEhCS`Qv~7yGi0Sx!pl_Qp}}A|O_|_2qrzFx2ghfi?$p$~KhR`ZJFZotsR9 zFq06*IYMDPpdttxMkr@r?){sZiL<;4U5=!AAatqABZ z+3FH<^l<6$#tqt$G@1ioVAIYhs854f)xBJnU1^!PQG8YVbpNjGI4%l5oKDwz5k4S# zCF@elmAj^7>1_YGwL&B8EY9A3LV`J|e$7!F9Wq&Fzwf<&eJr8lc8pAAvxg^Wz0vPF zf2NxNWWP@}?7eBOiE1w&J-b9G!>JF9yY&W4kqzSf9+_*qocSt#)_@N_WfS{2u|r`z zcCLnzZRb$9<;QHa`lo+#E@+EY>UoE(l7;kW|2puF~cN!54I)9wA# zKgpn_?meaAY^sL|Cg}f&-@9KT|C?8l0-rRWCW9p6tH;)31Tm)~Ckw3m+oRE_ml%rl zNjSPy1vvp<3+r#0#*B=oGGgh7&ZICp{sV07*IGR9pxX>zp%?ff?pap;ehZCcJ=Zmx zd+-XQgo_W6XToEhBl6}?AA6tFb-lPt&TJM=uD?`!2v<5cfy&a8TI1_#tS;4vvP*&8 z;e7foUxMJ^mX+GWe|S*`1-71p)f)X-uBusi6f^&XcAl$2k8%7_o#W>)*mV-(g_7ji zm|frMA zuI}{hTT`Sbs7N1=VV#*t*s^=2T%mBe1G;9^`rE!*c$JM$NS)g-66LV0XE}7 zBF=jsd|kNMI8Lzjk3HL>FZwxPR8`YGoE@_^RH%Kv6*6hvXrTHt3J6w%>y|MwK-hzV zIfWBG3>>s>v6Bc@j+Ha>+WVAW@EA#l6Ow#lw*_NN=b{wF`mP96=AZ-yu26-LTxKi$ ziA9=BSv#sKraWOy$B;irtZDrNK!BK*%4zbC2nI=pSO+F~hHs!u1Uk;a8T%!R22z+m z$X*Y_Q#raFS`Fj{@G->miuI)$H%`5OdLmHyE_+mwSQR{Wmh%2gDba_d9y&%5jh$m> zU#it#Ex=)rr)W=OkbELl+V z7OXCb%yE-`AUDMi-|{pw%`oOmfgHIDHi-s1QC+73GKz_8Yl0uB>ZIh(m}MgFekv{z zAulQ@I(iU$7WmPKZW@D_V|7U8Eb4=`)6ZZpLzoDh@ZBk13kB>rP`6qORjw)e6S~%S zaFq7jsJxt0WDQ#&v8+HcAHxTGeeXJ)e0U(@U=uj|Kh2AX6KQQ{4D<^66+ol}U)@bm zD;;3#GH}o|!JW#yV<|YhU<#VaJ{uxslzd6Kvyc5=NPOXA@fT<5X=OW0cb$D1tI|r7 zZuYqACQiN(CGAk{+AHO+LCT661nKmWIgTy~m$F{y;v& zp7GtEIb;oBXg^wNMoxG^%P_j6!nLttm*f|S^v&p|F``1oW3?1{4AmStN8-_AAQwPs z%FsX@Ywn~3^H#5};YV&QH7sqvWTusnKf*k&M1(1d1u_D<f=&dr2V}Mv!5C#d<75&$i1RCgCU*=a3Zpe%rnCbY)Q-F9{)rN2KNUyglV2sgfoWKM+m)fPs`6Lfpgw6hhe{+x-g;$@*nIj3%z>UP6QOC}nsB za*IH)ywf|?l0;oZV8(_oIY=6=I^C4`4VkBsTaYU8A$`JA_RVu|>RdavTRT{U<52t;NQ<4G}Daq0wp0 zhD=S&Zk>=ZLZ)22PNb^wl0W7sJj3H-dbD3@Ftf=uNvn_Pm>2Jrzlle=Qshg8jooM- zD`^)2EmxJ=xtMx1a4eJHf(aBAJhr)UvffRw{*PU6K(ppOk;7{3c~UlehgQvofkHWJ zK4s$Ox7OeG1&Sp3<1EEmO!*C{XNiaf1W|}p;RJ(at0<{e6M-O1m)C|3`_hX30XAMv zOaJbe^;%Iy2i1iF5B(1rp(wm}m`1nggtk!1;293Qq5yRxx+Sh1StMrAc_>T)F)%;3iTuxxa99YbJOm?g(pZZ6->MJr$fJc3^-b%Wjk9$|0(04 zzb9n40XA_*WD8n3VuTz zE&fD1CqBFEF4bDw_&2steHK(AZ?($lZ({oDK@`4HpF3y62 z&@n!;tjJGt8M%g5Io#+0rIDBVY%p~GD#sv#HdGWky&>bNf4QHrJGX)&gIdiCRJ)12YfW9QoAZVS+RO|`u{)N612 z2K~z`ZEV%=xf2PQ3X`H`pCl5h!6w|Zb7{$;+~fXHq*W{{^LTYA z8I7;HI5i*uLyFcBYpz{IXufzsGD=&%vE5&OtOKS7ZLlu#4R1*cW}DKkHlYbmDXKyX zX~*};_36C`!?(luOsxpdLPiKz(3a}}9Ew_T@nLfV^t^FLJkmJ~>urXUjq3?~{`CEH@5a!ti)C2X-rbUCGLZmyn-53bb(XqQoTQw-ZEQaWf0X{ z_~>~7$mK^o@`Sp5YuNx4Tk6@>uN@$4)Secgqx*e$3c1_7q|@ z?0WB8{u1hlP!TT|o;5E4Wu@vT5M8h+{e$?PsCckQGf^&l_OZ-tCQ3ssrZ$=ryH4`c$6i@p zbCB-J*Yo~y48z6z2BLJ>2JsP?1%2jxOv?^w&=lc);jWTq8_h5#9pj>6YL$Z*r_o0b zV6EYsnY=ipF18*t2#Sh^GG=3TKC7}Np*eBm=5I7&@DIgJLDh4*=R$7J z_{s269BV6gf^jRgHJb5@s?ZR0<9%bJt+Qo=r27GLN zfnRp4NnC7crtoB$I;^RpIhdDBdk4uCZtakBh?&NKIi`A`=q(zcKlW$R)CahSB(cDM z@>9k0+`JV-RkI|*h{Ro%DsT^-Sz+Zst2*}kX2nc0nB(C}U;nO@(+MD{y77m<^6KRK$`91@C3-v|nwR(>oAiNNQ(#3sst4oLU^ zKyE~m-xfN7Iux?0%oOw37)(){!o7fN9mJCkx{JT|sYXZ|qjIJeL>Q{@x5$jDH{?S& zxh?B4#Ii%V6AkeAV znSwSqB?N=Qow<{%I)8WH=PcQ^nXi50Z}pxXXvwSX1#criP%n{;iC&DDlTQwNjU`W(>9}Ld=Bw)-utUapFc29lq0CSgu+WsPvQo>w zcIKJb*C)!GSGF(LrGeP5TjIt$L$9{OnF+6WhOff41J?5!k8V4(PQY;{d7I#AB+V#r z#4z^4J{w+w2WfJgqFp;nKq+BUV~U_$K>S{p?|Dgj?!8f8yL4NOBRpD}zI@P@_|uL8 z0}M!v)MzTZC+}1~p_f03)!H+4^;SA=7^O_LH7g?u63qvA28RsdW3z1d*|+C+^(E zHcM)Fs~esj)mo@Bb+N|e$qU;ErJ@_&oI61zm(?W--Pi~>+W z@Ae~PFzJbm>EG`M98>giIt6eS^33USjS+y3X>H)<1DH)&P5Gii;1WoOGrR{G$ndI@ ztHcFn$981sICaDxnfC6QrpjNee4Y^3{yQ;P*r-S<+n=WvF7yb^EV$yzqR@>`KAP`& zpSm)z%WZurnHqoR_?0IJSEXvpH^hktRuc5_z@)f5!Ao#auip!OQhfgHC6|{>o0FOC z&6|WfddPsKCfQPU@-<2^NAt4mhODSlUBd8g>*okQU<&uR7p~4TyRC?ZhF9#AXH>)2^ab;wn?Sk1XH<^r^OTyP+}GWt-#e?8s$UROTni^M{PCv= z7?jF+N1E*gk1sMA#eYYg(0b;Mr!nNpzKL2Y5|BSKZCqj=P^)|f&cE9p5CsvyHiai1 zWIgh=(32zf|5`^=nz$kw6*3;h&xwFo!$3VYIx?C}a3;9U{mBY^Ar5(9wwII%C-e`X zPCh|$rzlPcHAU$}s7GxBDI|<>C++On+cCn96*^wAx?cQVzwb$X%CYOZxZ(QizNRX1 zweemzkifYIUguP&WKRJ}o~A-Ut|H*|uytA-!+?pEODY>&DO(aRwWAiu#??KeE>+Nq zL9mM@Sx~eJox=IZiXrnH;0eHzAPCP#9v!7iOr?@duF#%Y6t-Erz7tilghg*_ZMFTAy} z`|(KgU2$_i>)Ok%gix(g%^RMb2iuDYq6}vO{HKn{{cn}tZ*yRYxbW1b=82niv3<0= zhd=@R==1yRuDO3hP*ql?beF7zLVm*~yH{ntXn5!p~}dT%gAEP zm7LxVH}PSgK{Ris}8ZX#E4S=OsH zuH=_f$+iOrMt)XnmYT_LGL>`(ES=UV#F3jYKHx)!N!R_nufitb#rs zfgmruX(c~24~CfXKA1in!ALD?FwZk)w8o+y!?{!TTkeVFqlxDK*8=o=RPit+hUVuG z_ABPc_7Qab;e8)Hod&gG*&1icQlb% zf$O)0ipH*Z8}x&4muY}i)wp3iD^Z47oKV#C4La7iW5ihLT^lYda1=XRJaxAJ-tyBw zYhPFN#}OESEaoV#HVn&Dn#2^fc9i{z4w%*PXZtIK7Cf>TAimfTV8(>%E z&P7_6*Dyf6UDd#U!4UF8UXq6Vp;Nv-yPVqhCAW&N(W?#k)QK@-(|-tHm8~TI{X<3eygevoYhE^ana}R zC)sxs7+%^I%WR>59|%I-v7ZSQTouJd71?Heo^Vkm&L8=u9wPOV?j-WXV%d``@Fx(8OrsEs}+pYDJcHP zwD4yjCN%+DQ>zuKSb-->b!Z9hNMbLlieSHQd|=8i7M4^429$5UA#iv4Uo_ST1f65c zizAk~PwQ1oO(??&&Sv-s*T=S#l7C-Qc7pE?5?aZ>y2&SOw| z;I((NXP+sy?U|Qba{Pp+1FXI!6Z)}y1wM;$&DmGcFoKZ365w5O>6q|=2u6hLOWn2t ztU76Q$4?xsPS8h{lXC%m?-?XCF!j$K47<)VkIqQ<0}cR+JM){EDz13hz&cn6S^ir^ zgA|Z}AgiuVs!r0lyav%(oK{`jVBu(s*SWQ&9eVRN z^^?WJ-#IzgI}Py^jwb^LfdkgL=4hVCoJe?Yc{$le-FyfC>R*K{&iQWzWpr87 zUQlBWOUh_{D>9J`QgXW;ZN0{tA7FhlnL7DCxL~JwGDi9-P%7`cE1*^ zsrA1}ck9uf{9Mw9A2+xu8N+eAR@uJa=JwsJ;khRXj9vk*uV|kt|5=;wpQrjiR^M!_ z=3fdt*5EB${%{KWyaNmJtE-QiUMsFr>_@JCl45H|^|@H3$X1hKCVGqnM(o4?loMkK zG!7xnbG7$N=|em0bM~bPc~e?|>lSy8=*<@Cpj3-^ItPe0X;HA1?BpE!j4jiSSanO} zfocDzBtRgL9;F-y9ivJps$33<>B!_5{#CkXvUzyxo3`=s=^HTBk362{0q((Iren&! zD`b6L`>DiG`ErbT(?;y1dHKeI@10nT)>b0O$|;1QtaJhq^EX-&{ z(f&8{tbKk4hr9$35>ew=U0X$R4N{fdom26YT-!>sy%RrhVI-_xAY3ctUHm_k9ecNw)e?GuWG$ zbyQXtq%5@Lf1&N{MR^mxzMu3$e0{N{Iqge>!nD_RvNe#Yo^zG+#0nZCbS*IR+|WXklrn>dBS5(m7^$Keq;j%Ok@NMmNB;n> znyDEa1tv1(F2+&XEXsNKp(o!Khx3sM)P%V#IX3Y(b@oeP^3-};6^QxiabPCa zIZWm&O4{nZzyRe7a)w;m;rtwn7|hDOCg}ol(Bui(qC73 z>EP>(`)hsza_&tN8uIQ~xSyKqE&fxKdRZ~Nq($Ik5ectLJgY#LShSF&#{%NpB*2)&;FG)M; z_s%a7>PQn8lNnG!<89=~qZh19r_~9pqy(sTWMrf?2OqSqVB;f&Uk6%Skio1yFFo#! zC+H)q38UX-ms}p%b?7JIMkM29k+N)mStmtkZsDWwKcOgEdBy~x7C2)#TKA;)GMr*9 zK%GY+9m|fqL+0^)h2{h?qoZeAA9^3`_!&={*h5RF(hSr~s@;+9@r1B`-@dH!AC`l! z{h-Ca+Tr*v*Ocmc3+I?P5JCUD%ua>Tg(PHLxWlGoin@hv#0HAn~pmoEr{&UC}{wL|?<+kD_#d2_+SCnpgX=(In6(f z^9^Z{-tOn;?C#4#bBg3mwBP#FbjAxEop(R*@9)T(TmH#OqGM1!KBbo44OgwbGId>u zCndZNdpj^8eZrL$*k;#vHC9dU$4jN3CDzoltT&Cm*i`HN^5c=^z11Zt?22XBnBiNg zf!|L2QIQ9z9kUqsvW{C4bQdbaL!wXbmQX6y^c)XbOL(p5gGl@0mP^Q*RpzTwHzLuo zP8Ze2XzQZ~h80N2O1((2UTIE!b@>s{oWSJKs~13#v{f{&v?Pi???g>K7o`@#N(5;O z<_XRyjFUfra)!?raaA--6ghrVgfgDU(3jIQ(+RCp-w>@2JYWQfYBKd@Jz%R zZ0+)3UcdfYm8Ht0+v*7}f;!6NUXqjq#^=onZkJ|zg8Bp=FCFC^Smp;SR9cRw6>1pq3`|}Fg<%1=E6GT<|^EYi$o4~L*Vra^~8%!sb;C2 zWNFAFfzgiOEdhs(Jc#T|G0erS{HS66kl(JG%aROSIy3az*w9?BBw9DsvWWSM!AyKH z$7b|4jXLbhtsEXGSn9 zGwhg0Op(DJN;NeH9norL*Wv#%Bf3*4TOgl&g6e3c9F1YiLFY{TCq#h}OFMZ}+5eK$ z{~1a9S=fKed|3<=9Qg@Qmq#1IZr6G-W>}}fh!;hGQ|6Iq9C+Jrqs*3Ak$tN4@^ovy zQ}<6gp;?k>ZdL8)IpUz;{^taNT)e{x97I<~KY`0an(P&hc;iZyZ*joPBtqes?wal) zVg|>`Wf225S<{3uBumSe%?%Ig*SdrbOJXQEVNN>5dHIs8ndN6b1F8smC5z6?Ox$o**69DV zKhq2%FjeF2(x5;mGZcJRYmJ(Verw;xTxcH0>0d#NPXH^ zru?(>62q*Kvi!&rS<3q?Gus{SMhq|UOlF#ue6q^fmjavC(C{YR27)Dr+e~x|c^1fn z^mE|v4$LCSSY^zV(YAkl2>EkrqyJ=enN?Wrw8)%(YVc<8EVZSNQ^~!53%lVNvyM5+ zDFfp0#!jGW>;A%42w~?0(+QzA=7GGEaAxU8JVJH-ITHxA8(EwunD79N7&({P17T~L zmdSDskrOCx=Prv2Tq<@3Ihs3!S-vsjY?CLn(F=59sE)i53W!~@3 zP)1kLbwL&AdcbFB>u+Uv5WU~L)x7ON7e>x0K%Sb=w@5efx@N0_2m?vZ6fgr#R ztwtss-(oakfO5RGFegFeo(t!Tkt4-`3MGWlD`~#j8};}t5Vn`LZLT?R^haww}e2b*CSlp!cMl>8Zv?S zBL@E<6#s3!s9XQ|K%@-r+V}p+pYw#GaD5Bp(!PB`^25IH^4F=+QqoKW6-mhLDaw@m zPcj`8)&>c#8m#k&rI^mkG16;kqqkDFASHls%@C@N$lNldu{?`3+E6hxMf)Ll4_JxO-tt?Cs5_s^r{g#J=+^f;s%B9$XMXG#O1{N!7FLR4exR^@`{IJ3 z3ACjj^ax|>8c}xcWucv94yT$F5^w;i5>A}v?MaukJW$__FuO`)eGto-z>S`E0T2y| zr`LjF4f7fXI1AstX>^XNVvZ05NvN!u3O#=W@5>tViU(HgKn@m8CxUut;SfVxbByYE$h(OJ;<7%rEfAB7Wv(lmLosmRU;F@8h5Jff15DMmfZ2paU*Z#2 zTlajf5`=k92ujjZ;)ES;{!wg8bU%Ae+?2r6dJQWZGN=lKyCU2Og^ahhg4zai-5%!) z;WP+*o##HRy>1F$@%^KszF$T%xDGrXae()JZWlTq=2*iM0>&tZZc19u6J+sNu^sjz z+h#9UZB3l*VSezO?srQVVl7<#MgGW}wa~T$NI6 zkOqm!O3f>12ZsGRBeUc!TI|mK5Ew2J5ML_genMGgr~Q$IzxRf#yKv`?NmoMiZT{qg zT8-q(Hp%Hp4?)1ibvyBpv({Gh$Gda4=wXNTgLb2nkFU>+@I@qwbNjo)oPPbt@e>9W z%4KQ_D0bI9aI&lac!WRoBkDp>My<~qbE;m|lTy+7!Tbf8N;$i1dvApH&oA+tE^-r2 zIYm_`3fYgJ3Zn~rQ0fi$@UTg>DT`rik){+o;zRM5`P>27t}hJ_X}Py`(jwulw<$H> zi%r^Pg!ZW#!?@2b>7mpqRw1*fTc`%lJioavtERRK;Vq0=>vO-?p7KiuzowrN7S%?8 zLxdXCCDhHX6tt!f~t z7MT{7Z^b|Q^8B1{d8QrYwx6AV_Iso%h$BK7f4$#oUuNkT$a8>+#)-1^(e9f8L|fGv zwRKQ+h)C!IKxKhzh|m7p3cT`Ixp2&s1)~ZjT>&CmK*Vy+V04z~_|Cp%;)b-oKrXW> zELJbZ*O{!@QR{=RrqL{YCbqshAQNwiikCn}Ruv_G1yzt)NE-R|_3Y(c(7;^q*0({N z6MAm?=6U9DwA%YPxnwVXkJeRKAxv_=!yWJrVh}fek7z5ghzd9Q&#;vRgB6@2Sc-WU zD1eduR7K!Q+IzxUNrrG2<6^iP*jNLbD0GmJQ<=j;3tkQ@reAaNtpoHXfeTT^u4pmO zBZdKt3C7}ShXO&(z;lwhov-ZD{YCGs9BE}8eKRWMtfGpj7~=TK#ijoMblj)Rt?tW<{Qm)RO; zcb%la0789q81~;U7nWAM+`O|Ef4=N-v>;^9q46W#psUH+4;e@@s9v5t;n9{O4M{e$ z>rTH-YEsW|or79NCqK%-1rO3#GIjL|uD02Zd`v>t+J9xS7kwHHEh5@mI}TBYe<+Rb zk|YEQb@w1V&J&*u?VB@^=MJU0}#)+w@3_i}LQykCMmh#RycULvT7 zg9RtMpc&AkxM4K+R92euFj>MlyfzXvEwmxU!|H=(L{?3&5F7C_n|7ziPr&pbTw6iq zQG@zfDM4{AjboXySGB(Wp!h1KvuZ@0x*d9U9*R0?IZ#&r?=kq9{$0p_gt4@-#KQux z`4xR(U!e1v*zK8g)Fxb0GI&@cvC657ms2BTm}wh%@t)Vf=Xm6{Za`|EUqI@7$voq% zM*4m%e56~0g&fev_cb?yr38!eoAA@UbWr@z>pCIP0fFyl60HD&x7sMah$gna_nR7~ zZ;~fb2V+0ovqMK@1U{!v0i1(y8}_IE?v*LLb5i$EcwZke7Fjm^b={|1d(2cK@pw?sej!@-4B4JWa*Y zLLOh?`3i!M!orDb(YuG8K0(r4(y=TF+T2@uW0uF+s0@1578O*mlK^{h$V#k+j60$=Sw>`p^3(+b*opF|1}}p?~?vr;%DSE1yW0Bb3m$RB61Qia|8^NyjyU!m^RNJ zmD)@IE?wjSs^qgtzkH!un=TTe3zr(ERkxvNwXq`J|G zHe3ayVLbA8s+i|#^I`?Uru~p=FfAZJ{zB9RpoQJWdR_8IVTK9qAm3RYn@HMRK^%c{ z9378hg1HfkrW${N7maHY6dOSxS(50|7JzJ!w_?i!R2C;`KKlX6L|Jq$Ai>MNcJb!J z9Htw7JZT79??W|eSFiQL0ttd}2Gumh*0X7$bWw^G@RU=a=e+Zn>)m0IdBFL!vq*gfw^Ah}cc_ZKoI@#x{bRCawc z;>UdV==xG%6qEpYO(_duF_e=n_G1(sc%;p}K9Ep%e^ZKMzK0gX=&!SRaq=w@0B=0~ z^my<<*0bw3?+;P?D8p>=a+5Lycnn;=lZ|eQ*$#tcr%inc|HBQ8oGWV{3m(Bpp$05= z|H?wB8|IY)C}B3-afbxOHjA@h7I4dc*4|GFRa}@REN7KvkX7xU%^RfvA|NP-PU_d) zd$a20<}ZpVjN=TiJdoYqb#pj)Nd=SUT53`-H@1V9dxa%OHiVKXAX2PxQC%9*ORZLy z*Z$UTtc}s536L=ecOG)AliQqY>~A!HKPr%${uJh}{MzHRbPt^BR1#j=GNgLF5Z(1F z<@rpqZ~3-nRsnpY!FmO4voT`3nG*@E`^IhQt~saL)Uue^@sIcB+J;H6`Q{Z0nwYzS=?2|D3YbV{|oqHc0ktq``T!0~{zYkQruB{e+uE6olGBkN%Cb91aJNxW_vAUzcd(?DO5JW&k#6 zH;<9;58qvF+oG2`G~eEhceh`}mz=%^(Brzxm4RXTxhF_jQoN&e)qa|%kS;sim-7}q z==EJX&UMFeNmn}*s-!CdNIr@6;>1+YGm+afZoEWG+$Bf^W^L`Kf+qpbghoMx#*#xi z-9hK2Msq&dbYENKI}Sic7GPHSWZ%y9-Rs+V;K*B^dvR*rakf0Z1U-7HXD6?vj{c=e zO7rLK_v?qI+~+fE*@(hYzoi})aH8|T8~dgKiu{Jm&9`}%ecSNvsCJdq;keBc1jC}pJO=vd?z%|996s!0N4;=GMC?boOimsF0acs>f!#p zEJOc*kR`UhJO%N}_Il+R7z`&jwJRCC;i3tYvDb%NFTvs3OLzB}{TS7^6bt*k)}w+X3d+V*+b5$SykGEziKr+#du%%u3k<(cU+ejh`cN2VmgtXp zx!qfOGs{^s#s^-uYT6BvGNAH;`lq<8GIxS#1}aX@D6qYXHb9PFCK7dk07k^J105yl z_Q~MsMn2qiA8AscmPDOe8@%QM5Z}@}AZ2rY3KM;xx!CctGb}cOQzo>U*fpB}a8s_| zT36AN|8)U)Bcv*~_PmAl=NTM)4pfU5;#^neP1I(%NB#RgITvV5)1~sWSyO?kGw zc{N=8rjHynBA%}r#z$Ipg4#nBL{~U-N>80rY#^#6)!He8ZBc3W=}xrdoEF{g{95qk z*xTIy`I`Crxc}{M0!s1@VS{!W!6T9!`VmiFS}~6~do3EWwKPz!=8pkSYh1nmuLTgJ zy{GNjD>~!2Z)I1AvwU;S73SgII92-H!>I%}xkfsc71NjepVf52f|M|g5boBVeEo*d z;+XQVr$Wi_CV6;(&WTg3C~wtCzS#gA?{{oe9De%G8wM*D>qdAizx!#`$Uf>MM_Z!mON19eL{%xq%-E z@4v(y?uoTlPvt6CmBXqA$U+6!Kh4$uf)n_b3GFEHZ_46))PdFNjd$0W?BAGM!@X}B zMA{6?blyJ0LH<;)pW&+8h;8)(7~3L29zo}}8#A-S%{@EGE#m8LZs)UG=YP7D*CYeX z|8y(&3mfO$c8yc`tyZ9*9YVa?d1IvtK_}p9^Hm>Vd;deAl?0^7y|o9qJ!4}DAKP34 z<*9FU>v^OUe$N3)D-rpt`y;wB(CWJ& z{b}DA8Yo<;Xh+LHi;ZgvLb`N}X?_5zS10Pysl=(z-9LzP9Q(3ok%cNP9!DJXxiTtT zzt-J;;7wb(4cOc3&e=I>@_esf<8)HQ6356)#y~2A$_Er^w6}e;j)uR3406z z-h@VUEgF0t-T^fYylbM~(NAj14ME>*HWtY_`^LjMm_Zy6TV`$mll z4lwi#-OLOP(%mo&Fm#JZDIz5jO8w{M!7)$iacM%t)foN>s z;7VSbjq;)ITNKqyiZu#$Dt#hKh)`!(S0Fi+{Z1WyJU;vo`q%UmEnEfV1k%UCb}mN# z1^p(nwDGD$PyU#K{YPs>U3zn6!&v%z@sI!)BlBnUiy}iMLlN zAveQXRHGYL##;*8ktBxY)G|7b;z*&?VPv}Y|FIqsSe3^~c`~+QSt^_Dqg$PdnXN+A z*@5nig$}_$R6Jw{tNk3Tx59o2)zqS3>k>?0022qJKjN!!Tg%CvtiRtE@GXz1Q?T)Q zUmPH$#KBfUQ`g$%A9%?)R#tBjKG4dVM5YFzb<~?9sk4B^L&lrpvYLcCr^uL#(Qc7x zi*f40)EQ4=!#PGc1t(ozG=Lstf>oj9NRas*A>P@`+o(|qf-_TI zB)pDTeqKWIyr1fDLH)@~(*^O;*ld>VEXa0joK)c920nI6;c4_fOT)I3@4s7WCyK1c z-G1X z->kZq`SD-*)DiKu>&Z$;+8==z0o!3)GmQ7t10f{QhpK>3+V{{xw*LC+bFvP8@K>^w z{naR}@r7Re4IH!6kEm(uy7MRfc1Z_YH}69YjWN%_Ba2xJ5=}YN0hN-TqzYQJP zekuN9Q7Cyw(Z9vYEkF5^sx%SFdV?xxWj^ek?x+3Y`;V>lws3P((d*;CKi5_1gw6ju zw{&$rjn06wnC4$FpuACrer}h1gj}?sMujWB*6sY-r<%_6(vlJ9@9zAlyr8h1sD>Si zjYTf+vdV?Cxn+?$n&82s=Pc1m+hpVN1+$i&Q+Uh#ExoZLmqVof-N$7*#wa@{l#rUM zf*ZN7EMigd>M1Zr5g3(7Ocqp5o%>&??ww@7iRs~s0{?U8`bxzEvi&q0sScuxs&1k7 z7!Z^N5kf)5kKGZgK#QWwbKB)$$Y8Km8I#tUcm$LVAFm7ZpORKI|I#U7=?VNu@Nu}T zt$sh%F2vvM#&g{D;k**h|NPM_!Mib7_cA|;deP`fHX|Qo4F+=s2?bMLRcQpc+ZbHn-z0C$A1ths`t~l3_vpmAChN0&3gusi&JOOs*NNPb`(^IluKlSdG-~s zhY_uoP=c*v>jSGs5l7MJkYxjy^1QE$472O? zF}RTlmiYU9nsFA(0vUIUx}8H_`SGgJNny-IOZ;KGV>HiALCEih4zdh)_!jolohqWY zPoNQEZ2EACxRK9SjUmor)jQ9twWm>N_JPEmSG1;Ily{U-7La|*Fq#Mv1X%&y+76!} zU2~G*t!x2SE`=05#z9fRP#2mNPwpFT7hYoc>f{}&Srz=;Rr!}VNy>iEr!H@_vs6hl z`?iVWC@3QtM9j^~DXiFyyHwyc_tFFC;LC<@MVh3o;GXaj(gl^0M?A^h{vkJ76sJxk zh=ezgb`#~{t~}gG=QtT5DMa&`l#*q_1&U&mJUE_L316y z9l6DsZObLgcyqg^U!lW2qAzV zyGl$YTZ$>3b4088W;3l_p<9Ru8;b6AQF*B12)a?JYI#@I?SzbbhSj_wQk%{(g)CIr z_KjmT(twF&Z4qa2%fMH-8XpVn%c#ZIv#6(raGDQuZc(}J1O&jIB$qNY^x(eVe;uBgi^rG3W|44MuZpZ0!Q zID9%Jx5Ize4&*xQbo;bjalgtA6*FyBUmh;GkQvR&tzhc0_DLH8uob;%i{>|2T~WLH z7R77PDH@ag!)+dTJ)N;~zJ2Y`?t(Soe%SWDzyb6{j-AUgq842(p7px?xo zmg{n=s+MGnNDTs+d$nkjZF6FP@{MqQEIakR73=t($Tm7IEA)gnJ_>2 zt5|37WTJfYzlCD1;Uv<34^7PMZ5PtT3Vs=GR}eMNV5U}=DCJayfUKD!dt45Ox!j}M zkmqXLChRa%@}r#QKkWZ=N+DBOj3P&zuWh`i6#Ixu4-WR(-IeT!Mt>~YGzu#hFy=A% zzPi18E$DwHlW#n<{D-6DA~p&0v&4DXP!eDyU|Ua_#i`aojGk_3+2%Wlub%Y1*&hncqRG1_n7Gwf=Id#tjj$}W#+-|bD- z&KA?UV-Ivr+^=OGOuhLwkwwe<%}LX_f~!?Pdj;$~->xNiKpULjO|P`hfC}Cs_ic$C z`~4oY*c4;-6O_gK@@{~PwTW*@k*jD&2&{)z;-}BFMA!^0I4g0TeeP^H| z2H5nR_rh?=9<8aP8f)*PC$rz)kP&;gQA)cJU?cRQPr#IQbl(}91?|hRCrlYHDkM~n zB-IaLgX~f~$nf`p#z)u`ZeNr8&pFT@j3~rrfC?2U-Y*`!FS)vH%CvyZ*ZY~>i zaPERq&{v0I?nFPcq!oHJpOUSxRS7VrLz`UFKX9n-*a~b+0)*l@gBl&5wgxM)g?mx! z&ZLPfN$p561@eKio+j5!BpKz=2ihypXg(6p{~^DBQpgTu1X|&5dS9uK`zdU&bsUnn zbfi2j_*(ro5)_^`^g!$?jvG|&Gck_i`|*q0nbMvk(wt#43;U2%&fxUg*=dbFj^`$A zh-Q4I>$y_qkFI;DU7P?3^r;{+?=PSdl)C8Z{8!Xh^UMeB z?fIy6fDpZ&R#PChOa)D@o6=Yl;&MV~7C+l`$di?E&q^x%PaU&Di7I+iV>#USem60@ zoNqks4ce9|eD6)#0h!vNQQ{(rkWVklxcbo)1nAWlYdN5L$!%%wNyvI0J&5TQrUqGB30L zH{$P#k$9v@yq2xt`5SHQgdrEgCCLIbDY@19=~Ctqqg>|UFb46ZcbY9{3EK{-;zKFm z>K$|UdijdTh>6SPbPjB|y00OP1i{afzB$zEd%rcjwpFha&j1wjCh%CG;wA198|Ff% zdsV|VW{BL5P#o8sa&=A6wNx;cE#+TvZp&dIdYE=7iGnSuO*E~>*Un> zAv7WEs*BCJjI}yN(^d&I6=c5f5}}Ni&5~1Ux%m4#Qp?#*LY?BP4>-=BkYmKwH6bFD zUJ>QPuO>!ITy(pt(4SR)kg&9yt(I3xRS)W7cT8HnU&~AIH z2F++=ftfP&`a-Yv`yUFz|A7jh$2Z3Hi|7-Ky`D2M9exfxw?{Z~dW!V%jZJv|VAZ>3 za+P64`P0ApWTDc9)lS?u zaDHS=`5yFQs0f0s?!5AutK9+HPgAV>Zy6|T+MYD&@U@1Pflb$PL~D;5qyfq*T2l)B zsr`)-M$jy)2E>K3!Y{xG7^NTpp0bw3(fHoFmDDEHju(o^5Q5#k+d6a4oNMnromi`F z*b&C0LN!qeN$UJeCEv$JN75wbSEz?yZYMZba9AA;l65eU<+FF8Y&~~8;0?P4BZN{3}qy#qJ3~Vc!jP?$3PBSj7>%LahiT>$doB72wCTX)g+!X z?@yB=3OuH>IU-PN)S$p`f0z87(|{apcw08HMj6@vQIli%RQ_&=PD)n$6`1@(b97hy z$~-|dCV)8P;1FEMP7x*Ym$pC*xOUvZVjZH%pf8>1r0nwk$ZW1VV|6Eai&tL>Xr9OX zN+0uDxo$k^AWESu;L{z(9_$K{Koowy@=np^1I@>?3g=J=y+dn*1W`pH@!<&CV=cHf zZeGVNYML`sULj23la`&0#&W8n+IK$mUdVJ$G)YNsacPgn)}=jjRb1`3zCU8LEpuu=B&eOb7FewFED89Y^283yH3jCy z-Kk)Ipf=X2!s8gtqK^2g;ORN47UTRO@BRHiC`dIFRden&L>K*{>+f8bb&~EPo?vGR zWBfiK5(8SX)rS>C%l)E64Z~TQcfr1jZfsIzM}J{(3Xmnt@|vseB_O-yuX!nVa3<`V zDk9Vnzk1-X^vNdaC!mK){M6N=^IzYKbmi}w&Nae6>HjXeX3R)_H&L=FB92V@?Dokb zxTGq)%y2v$cz~7n4%Ccj^h#N-1OCPFbq#KmQ2w6p(a{O`6;hJm9XUiWFVbtoGVm^L z#fJEAJCF5Bcs(UJW9q^3(kaJd z9%{+!UDM&c&YNZa>s9AX!v)G5C0#nM$x#!y*gTv3!-@G*WTA&OPo-XH@5qI}|M^S? z^W0m_fSA~qy}caB8c{>fBpT*|jPl$_#>B-!(*wnU3Aa?socOppDxL0+cb1m^&Qy$i zbH9e_M*h7)b?>IN8x1T_TKO?+$BkU?bNoSo8n$7h&ak&#gE!O1AET3um&SL8ujjyb zYXu=6S@Q=yDLt2Bg+#9hcpyyB$u}5~^z%>lgY!p?7|DGMwrzCU!%sTQp`D|Vt7-JYN10~pa)wwL`_Da##9Vp9a%TfJj zeW31TkaxV@$9T{B=LxfU-t;1Pe!f3~lNk*sL2zoY{#ZNT_N&xs)(LlNY4hHxr#t^( z{}t&jWYL@VoYwtFpCNy4 z=C{#g4G2r*LBjBEpe|GA4TTA7nzjl*52%NQ)ATi^qQRLR`>sf60i8VQDqCX<2Xfq? z#1W?uH^8FH^q`xOUYQ)(7TxPncH?YT z=!&O$x?0*~75Ij>;kqdhmg^4DZ-XE`cs%5(z7(d?a9lLA40M7mCEL|QVa%+$iyji! zox$xAMr4B!u}56+dND|&bm)lUNc(f_QW%u8xk)?GS(ovS49&MXXiv2np`}H%FA`-P zL1QZ+(r;eDNQHp}CCjkH5RlTj5{p-m#zLhPxi*9*+k?>+Q zY4e&5>O<~>)`(FO`36k`UxK&HtR_w)j2ho=IT?|-l~y%eiyJ5#$~)Nk3g|DU_jbV% zuTWoOwhhUtV_-y@_|50Hk>$7B-&*0b0XX7f(=%#$8lzc^s74dr z%AQ`AI`)@pRpxtpZ@*nb<*zIcTNDH~oqMUXO}a`Awq&3t(~X<^ZqIX1-@7?y&1XOT za=YyDVF>e4ylBODG?Xs9nXfd(JTgJ9`5!J-tPI^-;%evw*9l*i@1jTiG2Pb8=%WF= z3UxJF+Pqr?S*L_iA_Z}mx(A0K*6=5&%dQvS;V2T;YZA2`^s1u7v#SB|)?x9soAyHS zfcxE`fCnG>PHjBkIp-67Zul1Y4{S^FkYB2yT`)uem+IArz;^_BZ3w~iF>nAzO7OVU zj%vft_Ok)MWn!WImO#|1BP*#SY})(RcJvx5SOIaPPH`g7@0vrs!kN5q?UuW)E<^AA z!^uF>B`)F`)oOEE=5z5uJP(RJS3{d(PjLCu4l=u+3p@HnI0u}!V2mv6zQ^?+9q*U& zRCVIr82V561)Mc1jK7;dQ#&kn+mF($lMl*5?qta#9tf?>VQZ3_n_#!gRmN^}=kffZ zg`~6tK5if3cx~|UKBrjxI!@^4!1UZCz%X|17lk^XsR^HyKk*D>FQUHjwllE+)Oeqd z_A`NFoExt`fQ{q0Z24hpl<+HwR9EbIN3jt}!^lv1_(^P>e z9+Cn`I>}o6;4n1L+a;S(1?n6zmJnmQ=(oXgZfJkfDKnHw?0^H%z^ye7$zG84^h{4Y z0~lPypsSPm%PHh%@b;empD?ec&P)Gyf0CBI>26c354#%S8Gm`gb|kFh*8uo4OXAL< zpsY~W{zg!#giml{x#<0LW@h5WWz~jYUTVXtoccz;YCWy$ry$2;y8cym|H(qR`F2X# zof9M0&Hgj0IU9JuhiHcziHHZ*flL)vmV&+~8q1(k>}dtzUEzvHE};eGfrZI7A$Bjz zl_GMpA2FxWG^boO?fBLeYT;7?S%l&$LrN3V#zVm`;3%*lgVNh3U`y0DsV+!@nnDib zYu(zox7qun)~D={LWM25BxB?os5;z^1JN8xgVTxy6@~A?YIzNRMtj~)&yWo;D5aCA zd6ITOTenxu!l$ES8^6~B<-!{@j;2Qb3B`YJwWBb$RR^&+= zgY#r9#_99NGES;~rIsP)SZ?g4BbihIo@jy3NZH(`x&>T0QCh4}5LQ7iHOrAHgg@bo zBu=hLQce!O$e2^ZR!}-J1K@J7I+{N5-u?MnyJ!QUlt;<9%`_ zb@O~Ej*NqkK$PuoF)3M%(LEmRUsTQ?Zepj%#rO5K4|=uvSM~_+j*Vx1^j1w@%QT+F z&MEPL!b3i8S!cN&bHhsQFLc(}%aDbNjH(1ru?sLe>?C>;DAQglrmR6-wFpDe{$+=E zT0$nj#8_M;B|>7ja=~6r5xjgWogb|=9=FYVALY-rVvwvWVv|Id@MJ7q(4pV`_E$DU zBTCcekQYJ)G@>rmX2FYfTaiz^xJZBRM^GxV?ShAmGytDI^_A>0Dn1^cB;#} zgbqG%eV6Pp7{0i^NIfXOx?x|#v&ebe!G;H9&h&J;Js-~Z<{wF>GA>`;Hz`PGq}+y` zd!^_WH++arPURvU%I}D~TAW&B+w&O03yvx)@dal&FMV`f(MnvQ8r*ZWGJEs?4mG9! zJKQ?!TfO^|+M1%6GdqX8(+()I7?;avTlbdqme1W&UqJ_yzT0<;OPjFX_*q|V`R4M4 zlO|X*cUr%aGWZ3XU`IwJQk6a<3qO@ytXTU@!6i2GlVe)gjDCrfxhi{_IpBR1|J|aF z-IMv^cn8M6zx%!E`__$oep zu|SA<;sNTzyD*+O@aMIz1`@SMet@6~S`@aIH>OqK;L^RNg0;HurYb$%&y@%Ex)t*H zB+z<6bG;Qq{~{o<<^OsCg8QYbwcxd$xR$~6@;XuvWu)tAOKOsL&V9F}2^a59QV2c$ zt4e7r9?>CXaiGwAIf?Zq%Uc|s2MZ;-=ha4!vHqW0B}Qk33dZSIozC8ie@*Tt4iX3H zZ8MG5f5L-fzYgT37Fdv;GxtgbwWwmQqcgsZvo`?hY9r6F$W%GGMPw362D-lAsMM3^}m zI6xMorj4ln9%rAsQNNnK*Bm?uY4$MU#&?zVhsUs9YEn%8@^>F(;}O{nV`%xrp{`S$ z>)!yKBU@k$!X|i`+EizP6K!Edj}9k6gpTm%pS1P3RT{*GFHa_vvK^YKs1rsgqcs|e z6vLoHyii{fiC`mC54rWbaDZMJIS3Z zgTDjg{;6;zyt?v3+Pmptz3dbD@C4Gc{$XQojs9tFmOz8?H|E5DvTpvjKFj+q<9vV5 z0CiV1`i!a>#3-6kiW=f-O&jqadW2T*k{~ip9@x_M@i)cd&83 zW_BpiNMqEW1BMeH>EyZ`S?>LMF;V7MAG;XaF z>wh;)hp)=HT`pflbSqj#y#ZAw{WG3pR@O5Awrax<8r&eg=9Ht6JN1k4zfaZOUMu<1 z>sGJZzyGw4tO(#4m{ma@(zC^bI+J*w6?%1o=VWrsVcGQS@hUvl=1}mgO6MDWo~4Sq zD!>Z0b>VVAkodl_*n)kqcbYwS);GGK)N5ykxAnfT>~PoL55vK!687*poY9TEeC8J0 zrfFbSC}%f&{QfYanw2|TjnI15oAJ|}ZS_V8@pEt0b|@&pMm_ONZxCl^4D3ieNSm=O z#XYCGO-B1vj$Pz$ye{$4U-zAvUF2~m>07L;p8@s3V2m~rcrftsQpAY?jp6C@?c`Jh zNRO_an@Jp3Kg+J4i4Zx<-_T^TPru2zVa=-VA}}S2aE%lb_n~GdG%ods)ic^!im*0KoFDC-5bMRWZ)^ zpibU8Al(}I(MP^p+_f9F2!9o>?fhARmH8p?cv7lPCIa}>t^U{9#fVz}oN3_Ov=stn z7cRJ^YLUF+r%n8Er7l>X9q>Y`n!Y_9R6dHErv0kIBCax*_;_3y652 z;Qq;cMw3sc`6zV8UtfCm&;4tN&gy<)^#v|%rl`T~b4 z8Qx3VagbJB*()Bs(ew=QJ>ni*vH!}2m{KN|X>r8LeTShTVLJKIY*m4mar*n;Ks6d^ zqecahBI97iZ}A0GS?0La==$FHT*@|-%huVG!?8)idD=`Gs3G2K46p{-;^<+}{#>kX zU=FBm$-~?sM4QJEeO2_A(r zjdbnvu+SHoFg0;t5uzcI2DWI9KmjXY(}IHy8hc&{eHLX1uCf6U=ARh-&#^(>^H_Oo zlzniWpxY^Xi{gBdj?cud$^MPq{U>wnD&i|GRI7^TzB}p+2?#f3R!s*FG^_jXZPYiX zEGF6Q(LUY0^oqt%{dvn0`W7@K(X%zr&MA6T$d=_AOkSk6~! zxt0{9>@i~%@hez*-|ZyAP1#@-v-Q!k~S-aE`t_~nwfD%vl(HEhgzSyEvD zz;(datP+59p*;VtG3(=}Sjge=squeK!R|nguMf>=EnJNg}wFfgyJ1LFq5pK zfX9Es|Lx2$&bPCaIAHwRz`yVOxa^Pqrp;O?%uT(ZuJpHyQXtpP!z*JYq!(|tJ;X_Q zY3+<>#2B*ein)F&PbUtYAyT`^cs%R*rqUZqqA~;;pBJ{i5e|JfAGsi#Ssi3~Wao9A zVKe&~$SW7rPoDP&0}2e8f}tw-V=Qk&KndOfj7!Nt_mqPzDL#i_w~p3VfCeu ztJp6NMhv&_;K&IY05`*CNvgM3I8B zwmx=j5!*CDs`ilMk~1kX4ZmrQn+&&x7`549#o2}8Z%l;}5=vv`g6R%8GSCwC-q8g% z)BsQ$+@hQSIl(va)ckJ&GKgOw`9DtzzUHH-wSExl1;=NhESy?5i>sX& z=ntN91AhQ75uvpfXEk3%dN+zAopOAaI^5D|5kDrvg@MWn}ff zi{n+@v%H}!m?`)YEqIOF!La(-y_s=|GZP=iO`A_c4lXa$pjh^(K36kyC+bt!dsV|2 zydNTtfMUggL}8LQ3>qaQaagHpxVH^6OoL}+{*9_=S?feH^Uye9&1UUJmvra*s5#F~ zvGG%YCrP1v&>)FGi)qxix-BT2Ix87d5RNc|eZ^9%3mTcL_xhnS6PO0-l=}Jx-xcXb zTQS1w9mV;qwW7Kej?UO(TReo$Kxn@d&r61sU6u~+6Sj(*_fD|4ACe7Fdc|Q^^2UPa zjo|M02c)g8E=-}Sy>l50^EtE9vAhRLB}RsyU>TDdviOXBrq_^$oGKkand&&ZWk;<0IY4CnLv#O15* z0woWLnHc~&GL0&&#@2Sc@21FvNVHW+l(&NeZ4{bHl__uF|!NB{voK zt4_ratY*lyn!&D~TFpb~vwEANXNDxk z{&S~Q@1GBr~V)@)LU!pZ!{A*La@+X@UB{J5K?)CCk`ToTjn zo043#M{sh6s|FfrPQv=@^~w)dqIf3q>E2qj;{`DfTYb%NNCV~nauZW~NzJKhaJ52q zz6F5QVYoqd9HANI?IH=j#!f^VTy~aqGR&?LZZVer2wtRDfEFtM!_j9jqcm{N0*JJ;X1!JI{> z`4orA{8~}p!m50uEi=XK`_-?IU>6|P*QOxef7RX3fT8+S@^V-(HY*Su^u`V*=oln2 zcU292PSQo!l9Ebe=Eqzop)+>dD;^`#z_@gmAF14rDlC(E|O0mMO2au&LF?Ytt5$zl~#L z>hd9Q?-!kb6L>d8bZrPIe)Bq0Y8 z+z=qYOyAaaTEWTe&x4wc>UQN_W*YX1zlslLJ{^o}{V%lxg-_0Yhe`uBle24jT^|3j zb&$Ig4~X-&5B$d5Rga-#waygmUApl=*afrjE_cHZjJcCsb+KzvR8GFM;{LRz{^6DW z7Y_r72QI7o*U=epcIu^@aA(DIf;RExcCEFptS}RKNx7QqI>-g}Cn4^GBj`9edi7`ddu;u;Gg^ptm%vy`WhfS)#McO6rJ@}_e z%F^lsrR%WZ{Y6vJQe<)KH%?Vac#~gxp=Z<>8@v!5|MY0{(IbiDaUI)#Q*Zu?Wvc+5 zfu&2X6dKPZWdG)jySeQ@h1<6V_kcdct+%MEy$r?}xvieBL^8E&=Wh^a zHbuV5;4H!yD{hbrcUn^DOWYy^8H2wiX;~sL)GO1)pqFRI0hZ71+OA307 zPA3Wx5k39U({P>$&DpCPfxkKIpZ#d3f9>r~xUP2wYa2j)ZGp%obW1wtWrZ5Mz&dI+)72G-hD2dTf?`5FVlA26Kg>a4trkdy3VVPVaa7_9$J%4F2aYB z1K9Sg`lQIFDd5Lohh2eoc~m{GV~g_GjPh>Az@@-X#kWnYEwy+6y(h*88~w9f&AXj; z3c1WVvH?`C;d|48ISmhxb|&}${rdPYqsz5v_N(wXaEH8V;m5QKD)uGzz5BEQ6KhFLNDg zEb(#mzn7IV;qw+fDUWymHjRMzxhm;`uqD`Gi-nWP!5`go(lyU)rQ+em@%!i3?!gP~ zvD%447IF|+(`$NZf6(jaeYO?3iNW5sK{Ukh#-d;P6bK`za?HqQ$ZVH{&UX^~p`~Z^ zzQk%;T_CeYohKw7be364QQ7w`SNs1uEeF!RG8gSIE7#fTu`A)n{8}9V>)dU#@0{_Z zr4wGt^dD=(T*WU|2mT?YEJ|4k~-llsRFtfhaY@{+w? zn-VHwl3xw_)pv4ZA!ZDIE^VU(fY2-f7i2*Dl*(( zywuo+qK%q0!TUb%YFe2!6O_K>Igx*KG43Nw3D@8^SuN%7pS+eUDt)1_C9hDq%4Ijqs$MF5l+WT#gTMw+Fs1a}yHzB_OG z@Ai|e|CO^y`R1mcS}ce_Bv*Km@@OupswnUfgkQ6&?Kqz%2X_gqnSR6~2Qi*mX z`YRZ=@j9Lf&iX=&MpO8iMuDWxnK9^^()*NpQG}_3@_l@y2)#ZfzP=;uj2HlhV(Xc@ z`2f7E{<2-Y*|Yef_&f5$5p2vaTi&M;{94{uoqV-2UPoSC$|zpjyoXA5ei9#j3Y-Z_ z%Y7+XU7=UR<@m8Rh7q87_#Z+DGFW%{)5E!&stj!hX`4Vtco`mcm@^uT&*U)iuXT|H z`0N-h6CDZZWF5Tj#M4{&PqQv++agQ=@0fF_UUx_#+3Oh^+m;&ao;o=}3xB~jOX=(S z^hf=f+$_M>HogOrth&QaXqrPeWm^(K!O1mhXC@gH*Bi0o3Nm*~g#LK~#U3L{=nwSv zq0Rae9I5rGoO`6)-R(mNNW`NVeMT!J3Vd6=!63+~=|fH57}GSm_c%ZMHB_5-bOXo> zq^Xi^7Dgn9{hr%s-!;o1NGAsWm4PA+7ddc)=-maf{MYu%*z{vT4k62*`*2duj&q~x zx#G$6IC&N6$FzecVwXiLG8_P_@%p4h?96L#i~F63C5QfM_0-L@HNp&O-K*g)b5}yJ zB$6VdB<;%mChUSUiA46wR2%rS>=sm|^u)d8uee2}w4W+J9)Rl1?Ax(wjamneX9Uj0 zzzQuQy3%Ja<6e0)TJq(NiJA#Q&~%4Po!{$_FXv)Tjd9JgmV0FfKH$p>d9|s0DHm0? zcm8$A#By_)yG|kXqh!zBm6|T$%S&)Gy9#oS-pf2_XSGZjj|!gP2%^T z@iZugmT|9okr@Bkf1l0ft@aD_UOl=AwjepcA}S0sx*M zKyK_CRkqf~Fc8UGA(`@~Miu8?%ZCHnht2V)J2B>FwkeHf$^+U-o2EM84(t*VcG~>$ z)2+Ll295YDhfs*^Xg?J+I>J9p*B#`I7GYC&j=$hr@OAp0;eZK-AQf($?4gRRxp4EQ z8+FL7rMKQ!>58*hE{kqVQ!GC0oIE|Vc-iqAtgW5UMI!hGv5TFHi-4fWRQx=`KU7n zT#cxp!TRTyd15d)X+(vN3$4D+%5M|5R(U4QNaPW1U1ld<2aKWTSi2+h+M{fkcUL2V7;9vUifh2SOa_OV+n+7`xp$}$# z(gkZKRtezdU%O=8RNE7{!@fPIDcR>?k0|?1T0QRqn$;bK&X#T}aK z6U?b`cd26b$DEn6!C$_>6N8dRDvtlgbF?+Yat@zR2e@3vIwgJP-GTo1o9$;k%PgQI z0anEjkX3LGgd8EwgHbgtS#4ki#E@hx&`=acfZEH|kFr11J(u>Pk9Q}=gadAE86%%i6lCCxgL4Zgon3ZUClhJ>v}y=rbjh*)L;R>t<|+`u=@){L#XOMvW>i z=prPH7nP~9^L%aIud}X~vpwAA!#b4|4}W#bEmO0(mGAjIum$a8n&`BKOR|>l)fA1+ zDw5z=$7a$91(HmNHL^j8@b5LE_oPgIBL2#w^tt|L-I9?1m9lwcq%|vz62!}YV>Q}y z+X?goLW!2q?$p+h$iOa`wVm1d%K2%*7}l%}FnhIAus|kU+gocsaIX8)`1*`{;sIOq zVjlPN2}mApOj4j`8*;#>gDkYb^nVT`{>W){~cgQkjye*@*|NQtr%D3#d! z@(biF%1~?dI`6d@Jb9Gyiz#mPecJRNz>*idM^U08LY1eJ_(+y#o;z3H#)0yt+sZYx zQ=hO>+!r171j^xT02L3(!$MuR+&)_!S?q1mz4lG_RnPMg&AK+e(8|QepN$^L0>Azz zwMNf@f8m;3(%&eJxEFd^0Uc)&yoetVOQJm&jj|3U!zpNx3HG#+1t?Xm1@70L?kBtn{sygX-INtkf)Wf{o`}OvnTeRAy#E3 zvJgYk4vK^rn-fw+*)Kh@v0y(0FoNfqUhtfG7x56RV68JC|56ihC?|K3{VC$QGU1X* zr5@K>SRTG}f15FE*8W^_qen0~9exwdq5Zh<`QhM9dn0|=OYXXuUpH*zg&{TpyJGKzQ5PrO&k&-u1s;Zn=l+-L4(1oh-?k?HIo}4NA3_sik%x%A}c7EQ@Zi=r(jH~LM^IH{>ytzO*6%v9zVV3r) zkuIeAwq?GnXObuU0h_UGigJ!?`!we4cjY1Au`DixwQaDLubDm{j04Na8B-yq#&dOJn&>e-50jCtnJ9(${wniNIg>Xs*Z zkU^a0J;<4uiC@$e6V5-4UcG+qnTe{%VgF$)aEs++ zUoZ9IDlqvKks`T!0`uMs#?Qc+MI#&~J*p8OqSfmAZ6=q52$X0@^`VHG!$;fUXCP_e zn9k|pK#O1N^nKPN>yrng1zC<5Z`JC}8low1-=~1@=v<_3PuEd4Sg3a-La8cDfHBzv z(3Hx{AfsN~Hzd;j16a{x49+67UQH-PWFU`lkw^YtUhqK^`Nj-^<_T`N(}G@euw+Rg z9&{Ibky2d3K+^Z{X+1pYnlx=SR>cNDqA9>46^00&eDigbtd#;tloF5Ob~sZ@Jb9(& zRMJ)+-sHX$01*=FSez4`qHxuB9krwF;iwq~~6 zc+~_HRVc-NjmxZ;;JSlfj4pkF-A>V$H;vbW-w`ndo2g#@>Ixj`P zoxU{rKpQN$LwwBK6}Z!?un@p-=%3y^#j$VkE0(%CUlsg;2`W`mpu(xa3exNHCZQ&C zVN9NZvOf}8lBn(11G_~gzh{(C=4Px`RE1Pj9!gFJkTcKX)m7L120qf1*n?&)UfWm@ojOUArpM{+d9Aak?~W zGtmOBlH?afi4Pfl&eOSjaq~uCWRb^(AGPHfn8YUCsPShoD(E);_eDNlZ8)DlS}}m~ zhnyfP`z7={SGMujC3@kW&UBw#5M7@ig|x;WW7!4cLc!o0b+$2BYzX5neVEPoB*jq6 z_OVbu*^6txc;F?5XLIN%alrIfFN;*a_|Dt?hp5(PrZ<&zQjaCljQbzDa10*#%G;F% zzo3l|6ot=LM=|2$%e!E0P)>FK3L)&T?D=2hI?b?^oou$5#kpYAq0fh8yW0`PwG(kzsoBXS~LEanFq9#{b8iudJwCGO+hsR&IZXi_>Fu?%Q zNqSnKy=OysH8-)Hx}=EOgn`^%ba|)Jcl}|f;rKMb$A6LE3#rO~Q7CUTUKun5yGfix ztSeo-D4Jw@l|yj4h`XHbvd-@x_Tq*jvbDr4j zP_Yy8XT1Xa`=ksMpIR?2*22RcJ4I}5fN%X2;|~I)8R+^kajf7JjTGK32WVHBY5v%x z9)M6$M9#XTG3)W$@b^Tp++Q?1S6&#J+4k2f4ni3V`DOKU$w3~k#%T>Ct^aKwN^x; z1YTrRFdsPT9Mp9eqlA!Yco^6SZkq@JrrLAr{xRC|MzDdJJp#cW!sZv1$mD{9s2Ijf z3xrcbQi$QjhmB=1$uD03A%~kEp|QB14T?O~ef8)PYgzLy<=wEY8P)#ok!Hx+M!jlf zUuG@EfYKZ>Fvkmn+$feiK2B9Lh?n6(Q=1A@{e9z*1RV&yhhC-KKfdnjfWJy&>FmM= z0105X>@zqe(_CA#BykKI(3mKcaSc@7rO(oX+Uj8K`oNE{`;@n;IXVXJ2QDwjXBxv# z^d7FjdU$EeAENvdfA!G`1C^h;dJp=4%;gNqQo?uME55kCHA^*seWgz`LBF`Ir#EJb z?m;l(uLM0SlgxOh^dZXYnG;Dqw|jriy0~U%(EXQta|&&;Yd_mnp(eLXzw_rsVqtuu zeTY}&Wmj%Bg*~%$=4opxcu-<0wCDLR!*I>RiqwH|hRefJoY^JZ{h5W&fTX#6? z;{vl;$6P<*QD<4HA~!4;r7P)7&EO zLHC`d+XOmxA98EVjnQf`L)y zSFhekkJ+T2ok&04<=RC3`~4-w@&9;@I z>BZ_JIQKpv&y8%H;VV6D&xRqRY7wDNI`SKb`E)*K^%_<+<9g%u07Lsj(ue=%HWxhS zaUK4@xxfVUTr7pIr$_#kPpA^qHtF-o!<8V>9Bs=3+RWG~iQ<(lajq%SfY#T;BzMdd zw}m)cZ&#Y>vzW5|VSxL~L*)0Q;sMJn_;+|gOb4;= z?muj{=k!meSl8ncIt-XE#adfG?Alq7i6>533>Eur{Gia}j^52NRONnHMX zcWS{gc;*d1z#HFr8z(;jiF?z7iNRPeSAwo)PdC{TZ)@7WVZRfAXEjh-S47VcW;c3T zGUoo5hn7g988N&fH&X+&M$obh<5_xX!TcXY$bob#c40?<{|WWs+tcyJyFB#9HcEvt zJaF=k60ZZN7rQKzJ@%-F_XRz4!goSYO$2{-9CPJjX9=BWnF3Xn#K`CB4(_S!`zesx zt(lrsG%ntuR|fVTyO_?A&~iE3nwvVXNvKw!VZ{I6y-RMow6t(Rhm~&QAbp=3c(T*{WWY2 z)z-feSJLSyCXW$94EBAnoCjh5j{ET&cFbbl?Jd79N*=O|q-6T(0BH)F*dCv^@(6mK z?R>BbMt7x>VfDa07K310HyL1W#Ox!3mbY6 zXQ~g`vK0F>RjN{NLXaa~KHc?Tq96Sv=lIPV{$WT+9u0Q=2!fk2PL2fxs~+B#XG#Gn zv&mZz0?of@WNdb8UW(pP{1Oe={U%5h>_L@7C%YM#>NU0gZtt;L#)mdT-UcWe@<-_) zkq;O!k(xy&+J})Anv9z-a4nSh_aEU39&D+HOdsb`9&3-*b{exe+FP=A=CGqo)LD1H zK4GnV9pFn0W{|au6Tnolhrea{j^7cZMW)=Epf&)|g`lDe1hv_VMrj89l7Q>a);x>q znzV;MJGa%#_*G}$+-4|#F;d86Viyfpy92n6SZ1i6c+r0))1d&eRAsZYQ0LE}j6(UG zHS)@NFxcz;yiHQ`X{~eaPYO->13|csCR~}>KvmJj$~(-J5ILY>UM@~$m@9} z?FfwhPWll30o_^FnyGEN34;8Vz4IxHk~7`ng<>fdhB%rx*`$=B&>bhAYsMIPB4ZEO zCED~_Ki%i?vv<(&oZ^#3v`SEf;r$5My5o}jSJuNS@@w@f zGC#OEs@mRs6Lp5kb3PQjFLGI@+Zz{!odkMSKPGlQnzF4Oq8PiNJvG|@2)m5qGg(-R z{q-Ie>qwW$KpY`x8LZ~~&@k5VUblQ%K7h`(KV+w)|8{s0&)ylk!5!-zZ7<=9T;>tv z5UP|4@M_dpYiBzkWj^!nd?^4__Dgq}=Ijt7T$=xTAGhis&nA^CFs6J5kXr6OpGg{~ zEQ9#`aSLC#EFHju-=g)zA$*I1FTmz?z>n8^EDoMYqa9uR!PhCjh3E&blD7wyQJVT3 z@tEXlaBPX-6|UTsiRBGr?qlcJ8H%Lqe1AS_Ez#+Ft2ksvw4n@-^){-yrrx|Ho_6`- z`b6fDT$sH7A%4v;zmCPRQ;k;*aY%7;!K7EHue^x8${+ioW||j`#lX0@-lLdfwplfm zN@fCF-69j^fde3T>mZphVXWu%82_878(3~@zQdTx`Y9R7(qY1iFCDUM!aBfGWFW&H zVOxy?XR6aryl@sx2kAvpf-H|3t3oCN( z@|_5V)?!j;83r!7YA5?z^&Y&=e6Rjt=52y?EP%_CiD)agz>j`An zn;_-z6^0%)8$?h;%o%p_i19Rt_tCs4wc2A~U_^c2&Ikq%2X>@QR+!93r<{KMf3TXE zdOEnU&uUl()1mGxABP5o$-jolUOa@KST*qq^odWS4O zc*3RP)eHbhjBsXH=yz%!(lOPdP%ZmcWXkrjYW*vHqX>4i8V#HV_3ya)4B)vZ3b63? z;wq=_-nq27x6GX4T-l1|LbJlee7C54%?Uk*qIPj)BAq!c_<|ah=`)8Nbs}5$6QxAO zl9zNyIlpbBZ48;hgucFZOIoN#j;NnDQ+zL!hq`8zJSH%ljCQvEhW-%-pFM?qW9TQO zmq#!Sq3mqurrz0jvgcu1^{$@S_kaynUNT)B2UzLjtTpUjxe394eK(r=-#=)d_VZ_) zkA?40l7`a|t=-1i^Xz7~ZWF)CxzgSefxbMQD%%)SRgvGY4@gh=veK_FetpUL+Lsh; zppNIKg;Z%<%;U^unGyz?!QH63so|x-a-y{Z`;Et6(w|wB3C;NA7TW!{Q>WsK0x|Vz zdp-?@hR1hx`rT6Jiakg=0D<(y+}6XzRR`kI-FN}>1D8bF>B+TEwBsfU4blh!XD*?m z#bO7g7g8fO5)tmk(KJVc#s5kj<^;0E<-Qv`G>0`>B=4hDrfmcF=+iM&V^TAhgl~Xi z1_(l$VqhIj$&#a6_VWPAxM~WTlA56;I8v;&cA|z9%Sg^R zjOLg(|4R4ei-38O3ZL}CO(A>{-dJVNH>0YX@!IX8M`PtM;=JkqGI!h7?K7qJO*JTXO$A7<^9T*0nD<(S;U@L?yB5ezKlpz0B;gr_~4eI z>+I*9f9ZO`#Mv4DLMPy@3=2yh3PHLkWt_veUWCtKVb7s5Mx+@h1XBY-FW3I%mLUZ$ zT%JqMWWWSRZVo)5q}M2h4a?bpn5vw6%qt2db=do}Hmf}usii~GPieIkr6!yN>OeEN zHm5$)O)Y0sa^Q{OX@$E)6y2TD884u+bb$pQ=M*f(5k+VqMjHUyG2m||4#7Bdh&CTv z-Rs+eIFAn0sPDr1)g(5*U1h zdaRszsjQfuIlC}@H{n*484dr}S0CVG-k{=-&pK3Gr{w@?BpYVs%Y#FH)7Uu;bBET8y4qQXm5n@-zdH;kHr zH~rpirP~3KpKQPrcUd-XhJm&QQ+2jt+$+CeUwr*2eD3bp|EE0AfBqjm5URcC;eV<^ z>gL{oy{4dhBpbRECzn{BH2_bdGU7zsw*EomBFfMatw-Aa9p8mX>74szL)6X@(~YR# z+v+F37YcV~Sm&qW0-+y0sfDw<J%Qpma+M29efIE zxp?VWqgL+@O4rMCFZ$E>3UW&$;-AAT8U4A%<7kbWu^qA?^G=wADvd5Ff_I!=_5)VF zbLVn{5k@D#4eKjz(aN*kvhsJS9EdWo^47_^5a(4<<@1&Ls0&)?D1;J83t z=6S~44v#b5BYaXO0A{zixsmL*lS)l28RY~f@Dwo*CiE>TR+^dZ6HX~&f~EVb$f5#^ z|5_3PhQT|uH9O^pp~q-nkXy`M0`QDS*aIGJpS9ADf_IY6XUecg;02LN)$jc%E~5l> zUT19JHehXAgNl%!N=HOpqLFx()Z=4X=`n9fJjPL)g+gx6Vf@&c$l7HRF`8Lw3h zxfsD9@rMu`M@*lDJ=LKFQ=-_#?};3CIA8s&y?9r5wjNw0?iZO;;_9gij3-pG^2cyte6zJjjfB5U#i_YLsHVtbG3YV zB9zfMp6A2yC8U1`AMe)ie*Sk)J_KUA?jNuwO;kF93Vx5Kk-}|p)dG?FM0mEs_wa)P zzQAnZw}SAd&V#rKP?|#Qp;O)|N_U7%YnOiV0Jem`;_~ofxJl$=^oUYh46tTd*ND%q zejIx(<>8m}_0hG?NC6=V#DOXVlqNdS{jEM#wXb&D(?hNYi-j{X7PNKdB!l_{&F{rM zYUIBz_KC^<#I26J7E#sfd=74{PQT(E9P})3zP!?J6uJrW$IERTkJ|F$EUPZ*<{|FK z7h8t@{18Ih-q#98q8pT8FqhXy4HL+|T#Jk;6+kt4oQr7By+{%CZ8^|g+k(cOeJtoX zz>CbVgGjtZ=9t6k3E*|k*%1_q5=Hhn#;3`>f{29DshKT!9|@IXpTQbq5KL%yzh9%6 z3MJ5ud|M970#(~94C!Trg0=vQF>n8DCh%6tG@m45itva_hYil2SUXG@=c%E`0)5w} zuSDHgO8k6~d_0O3?fLO0?#frd-y}DD*)=VnMY$|;Z2$(bvv86*f)xA8keR8j1;Hb+ zLdO(MSYyx=35qr@(#qp1{t-%J=a2{=m)qBy%s8i(jZc#O`;!A2Ygqvn=b6m(Cp#k8 zd=ERun;XaXYWo9d17+Go5u4{yHUcUNQpm-~JU*H7(5BjjyuAiFZ39=hp3|k`Ad8Qk z(%eckle6VB1SbyDgnxAtnrIzVA@3uKEx#{6Z}!z%iYPwTv_xG5Cy;mHYZ!!B6Noc) zS6yv+DfaqC>U;;l@U%{q{qxl+<-6H6-?aDXWnLv91IEH$j<~vy?*%(1U>sVU?+tL_ z3hc8mG{vq)1Q@9*EBOW%OrG`xw2aYs!Z^+eTs)1Fp4;-)Bx){TC|EXTGer?!qq+pr z7}+xAOn}d!N{K}>%rjx-Iy-e(pr{7bdUVzo#~LJo@VCG~nMUqV&ioyaF+T^jV~pnG z_TY7uNW8=E#1WyIO69`d#jt|TUu>GDc~Wj4orCsQSQ|TT6#oV6xvzfpBIdn#r67i& z&v^s)zy`}wVFHY%S>_1y#+wH=KDiysdU%Ewd4uwuL`!T^tH*V>qaJ)MjruPYOXt&9 z-Ll$*`;bnC@4dv$ZR`L?XNFJl*{JGqzWv*L#Hu_Fd+5y=?c4f79aZaHmc758`#vC3 zweX*={n}99BzANKhh(qL1eJj+KBA8rSW3%1G2BNUp|ciQ0D9}gvm5Fqz|m2slJo!E z>HE3=yVL9Tnw>*zQo?x)OWj^XW{Qy@TRc#f!w68K`w$sJ0XU@XC6QkHC*5!JF&x(M z=+ARj-RTi^@z)hY6(AXxM6HhIUQx$8e$k^k2!3x+Q3u0gtcseCfW? zD>h#8m=bg&{|z*MOF76|<3`2A0yx1yg+PkG4{05&d&i53B$sxj&MM2Cv9>|rv`cH$ZA+bMq6rmTEyw>MW33fNYrrfDii zrDJHZBJDLRenZ_d-P#ZcobF9nBXkPrK;VWxjJ2Z4etV~6xUDsBxMs`^$VLmT?w@iz zjLbMBRGG;eV8l~bN zH)>fl?tHC68JKhaV`=FigFLQk$N2=AY7ZItw4?$)8_p)}*e0x#;G+8P3+2&-lnW;b zUtQ;;fpqeLL7*SexF&ymwMsT_t_mS{tGm&>Odk$s_^WYvBdP5tSn_~a2XDez>#K_Q zVt3j(Btwaif_5!SoxKBs;%GtQSepf>S#?U>vUa7>;LHE`E1*VdIta zZf+9;jhC@!MEB?KYM(Cu%s&%vN_o#9^BElh1HTT{3E=?oGR>i)cH!#BX)>i@NU$Dm zC?>D$5Ow}bPEI(rT7=AEmwHPdWIKU?$esnaF8psCU*!pzl zuX8L|xR>L8^q73WuV_=fk147+C!%D1GWR<7PWhYW#9lam?AS8Fqd?CWg@$cjs?v=a z3rx71QW}#JqVv1(6b~4B^NXSepesI0Ao{pDc5(fwht2XGWd|;vhV$Aln@U}`L zc{o+3VP;XzXRT6L>HtF1)wVy9bTRgLKVuwR8dkbf@3<)loQoyj$cje3NZ@aK_-XV$ z#F`=Jc(+qbp8bk%#RzOF;ypRYBYd^?^9Q@H@23wlfpK=*WJPTWc6G6~9lKMwuPMNb zlW#w&Lck8=eQ8)d7|RpaoFU#r7&2f_=48luU2a6%A8U}B1`?7%oG?yd4R215{EjtSc#@fra zn`(ZmbhL)R0Zu@sD04oQdSGz(`-c}W&%^*E#1>`Q(6fbs;3;<{Ti2?vmPkGtr620& zTmjShEAIxxBfaba(=3djn$R@p*WOoQbO%< zylGPI_c;T@gAZCF$gAK`ll%6F=~gtG0Tcn=?SNYB){iVbp-kiM9&4qM#+)@FxeloB z=VS0QpW|vy^M|ZC&d!(NV#9Zdnrw{BXwI8f{V@vLPoi*o<|o8h^+N?{qLA8<>@qLR_Q8k`H07RjDfG6MJZxKaa`Up{IO& zH6~OIwb+5MtjfcQTC!;3jyH&i*Azw_s-ZYX>|sGOxzhDXG+!)|?xH)GG-;!#5}}3e zkdH7SDTkMp;a&jznIK9}R2-=8lM37>k^m1zip)jH>ij6?liT zoFZhBwHK|7Y?actlnnkcFH^iqx3qkkz@sWCUYLuSxYBjBmbmU)U9KIv8)eAd&acjv zzm|dvBe(|_BpFpLC1R8?bm9{aXeSW>l+r6o1$X~@WMgERA|cBlj7^3iJOhc{3arF| z)*=4R)`a^mZO{(!>Kx@|+V6NEXf$2qW0)D%zi}h@Eq7-K77u>n+lBtfI#SUazoPfP zA2WSos&x)&-O;~F{FMXS>>VkHLPHbY)4sS`iy_!Mc|qwL3w%@MCNmlpF1>T~Ga z$rjRAn%xst)kNA%bCDUi%afQfbyo(L;#C9!eU78#i|zq@@8nq};dYGU7_{SYuKhx} zj}yvpUsu-iv5b?ymaNBP8NBNcznK!)^zL*D&LmtK+U)3eVhyc12k6?5dj)^ISxrz| z^z{@@oZcSq7pn55NSOfO;tqaA}Sle z*=8k#`10Thq68$S|+_n%5H?Sb3a^dHyZK8ksAI9Q!y~5DigF z&s)6jIX&cUS}(Okd>cTT<~MvR9Y2ggtDX?dZ-A}BtkC=V{>G4@iSkiGi1SGs?2tK% zjm7UeSXcYfP`c#nL>Yl=17INb@&cB1DI1>8i^QMm{N3_2vNTZTSf)MgEadO-SMURL zl&q>e^DSoBHyo#siRwH9#XmRCmfw~v?drx|7L*fuPWKS90YBaN6?|4@O3lC8aAx~ zc^xKqXxq*2?7?cvQcv!z*6B4>Bi9?~MeR(JYyK1vvVzxo4`i&p{s{T_`bElRrjM%w zi{S{Zwer-K)!w>~r2~{u1sC^hy9%3Qpk5n#+_X11sS^~S6F_|$r%h(&=gmUxxz%`T>1-t#qpLIlI@BoCseo3=_C1%8-9b^}oV3zd-PKw97~{m|E^Yap=syCU+dfYQWRaj zOr~NV^+vjKNEcah?=PSHRb2@Z0`enL?s(f}W*(@50*G|#dE?TfJOF0Tg{5R-)r*33 zeaMPon@l|@yoMov3j_L?@nPFjfu>XbSBEXPGbG6_o>FEvmJ_+`Bvd=Eb z?|aM+UZOo>V<07O`aMy1mD$?oEU8(;&DLDHv!SQkMzuaT@^by(>r@pKD_5lGBUhRw zB3wYnpUH)BXR5~G`v9*jITG?{8dm@m|&2j<4mzc zcFEwRlxlV>9CEmVI12I_Md5sC`uk?JHCYZPTP80HU#AuZHl0%DxA~55N`i5vZoc;J z%pNk^@)XvJ0;wP>0jR|^jHH5IQ}SP)nd_H2cdVl~F4MHU*?j`H zeN9Rg+rTo}Oq`A};EBP=rfddoKBq@Zc&-3wSWHA6cqY96vDZN=8Hyg~mzs+e&D z?xc%%t|`6+_QZ1Ow$m>)i=KHCCsifY&2lt2v5^{Hw?R^k?E(=e^W4RrxE3F$%+$V* zT=)DnrU5UMw#W$xLb2F?l*x(vEcx&Z!I@1Jzl!^;LVfTs3hGhGUuM$dB`|&Pg?8=O zzr@rb_LJv5KA8GP-GfRVIAQqz43Ye8G>jtQe^@xxhqiKCN2$Je3M;i_!Ikk$JpkRv zz4rH+I~X2Lh@g){35i3yV%tVS(S0&b|7O4 z=PRHjl6jgLJP~$cQUn5&$&Nnf@JtI&gTICJ0GHiF8f8jbCe7_#r5lWxu@oh#3rER0 z=d(Zf_+hfjW=d_4tTi!T_mAB1v9j|&_d=L)ZeXRGxPUbA&dT}<>KU0a#oj3baa`%Y zF8ZXCIFh8ai?vt72lTwx%G;;Hmw|X~?K#Y=5h~ae{l-l|$wJ|%2F34O@G99R+xw8J zyw}K|Se}C3`F~8ZcI@DD%tz{-s$r^tey4)(%)bN;WC0C~e>*!ICAB*gIz1FESZnlmm#`0yoJb2NH{vx8aGQl$h@mb=F}K01LHRh1W-N z)*R!?ZGuH{1Tf+*XS1z%=&NhDcpD;f$M{6S1~mPneoJY_UAj)QtbdOZP7_7+?LU!R zB-Td)noqwm`<{kM9{q?qL=eD{9p)@<&B3)L!0T#leZ}js&Uc;P^>eX1Ya_fH*9WLt ze9u6{GqzAsMPpiPpY7W??`xT3zSqq(yv+=)y6;Pm_@FoiJKLL9&U!0{bh4utS|7!()sv&cD zEe7rqR_{i70l_@Qt=LH~3oK#G(-GzpRAXeNd$g$EOWwzm-rALJmG@qYC6d%>gT32_ z?j}HfBo}#xe=U;bo0o`XI6{sT2HgF23bK8w&eXie{=k9~ZG#E$tPHolDwOnduq`)ZK>)3xUkTG>XmRUo5;` zWQuJp9gxPplIyrAS)7&F;(vO0A#t6)nkRr++!iC~)AKuptP zknVSS7BNNj%Jq}1eGUdgS}nHBN_x)90`19^0%LYS6scj;+*R_W1zo!_idWwJ#7-@_ z;-vn!Qz$~>(Ff_>d5q>IKs0Pzy7I41-?S?(Q$#$no{Bhvl>D>WpExN#g^$~h{QFA4 zbzI-paD-&rKj5C+-0<@@v2bsKf+JurG2bww1_IKb0jlDIu_>L_<0Y}-&Hd%f9~<0P(f}oI%40cHs?9?%U zleE(uN?4voUbLLKLjcDDR{$SyFY=h1q*U~E$AuJ#U+U}wAGia?p_`LOdq@K%78xfn z0V4*S);deA5Xr(x1#yo-&?vnIp*-P~lp06NW(h4By+wbPJd?XP_71Fy+9(zoNP4A; z_RXpFqf39{;{@TY380I7#D_!RSjp6X`dZ|9vlU0q0PS)eF1H1(iyB_*d$eizN-ff@ zUu#q;H2+55u}{m=-|WRuxc_t(*&1ci!v;qp{gie=AE539aZZ)tbUK$r>w$goc5pE)){Y1*+ zYyfjBC~A2g6$}OQ-Pd<_>O}hz=+Sc&^mlN|65)eABQ)$+l63p??8j@3`X3D9mvaZ{ zQEQ6)z5V`usCfDC)XsR@RgI2otSzx`pO}*>yy^2fk%W6%Cqw#;KDSpf_J{ly1F=M; z{Gr|`MBs==>Y*2I)K!XUASxn;J>zZyrTcs+oq@bN z5=_45Do8E+a{wOyVpZ!ZY!#ejewSOAg3}U+UZR~yib`Hq7s5Fl>dzjM_TjWBH0m0` zw}o(inp@6lGy*-Q2RzhRNZi8<;^UEo`GYSq2C&3vKtre4?zSf)6M@TR&a16G_*{9jwY_RG1GMCrTFmftj5+#WLi`4=`C4>K7A! zr%B{`OdO7f32T@wx7*Z2>5zD=Gec)}^$ReBZ7me>n*3s6i5pp^mSLn84qvs)m=#bu zz&=4Z@pA}!*wvgX1?ZNYzdKGEDj@o9MXoLjVFQ>5nWoW5Nar%V*;dgg3oZHohf4`M9WGqMRG_bnX5gd9;>}Q*KetSy8h4N9q)^3 zQ*T-(jv-n&BbGSX-wR+_^#Rf|MaXM{WdQ4%z9%3%^(oWNg|v^la)aW$k@b4!dDwEy z4owk#p$kl4rQx`thm0*sKFW?;F2MiLXO|3u;j-`C(H^ zc5O!?O5oy2A2f*q>cp1}gLQY4kg#qw`ATgPXW=1#o}?uh#;BDA8!ZII%+bFN28jeO zA28gM7c6OjUjpUBDd}{ClZMmYEkM({0A({JTLBva<}klDV@MxpoQwFK%$}p#Kjxtj z0hZ0u9tw2_8{oYo<4Xu`fL@;;epRTubKbYy^=XNey+Y;#K}&nQv`*o~5~8HrO$fXS5Gt0mpmde-^HY zRv&-32W-UzD{7}@j7bH!GVoF5Cxg&55m@?-QQTj@pve9N9ireN@27xI?RO^mqsu!p zQpsb@SwZw)bLfngi~s6#eDp?>4)&%IMO6$_GU3zMQJ<${`!4Gc-s8@;LC4R$>&t(y z9`{};U%F0=ISP=QGUY7Fls0aG!&Epu>{Ec$fO=dF*HyN$6SPMX z@-wfJx4hp^IRXNKydqt6M4qDOf8fGjnmw-uM^=}%m~!9s61-edfG!u%(-%=o%9Iky z?GvvF>CmCpU4QWiqRX*-f5|jWyfWs#STfM|xL=R=yyHGDdG?FsaA2ol}KnIv)42e{BBJU0sM>1(lm3gi9v&kf~|w?VGx zIq6E3TElQ;H`y7`f2d8t2j~W$(&r|Kmje65U6BOjyIcU11Q3^AaC`y3l3}r|jb#E4 zDTPdUDIwEF4!kL=HCX>_^>N_cm+Lcv!|Qh5nnal2REpeOZB(PE(?oveOPwRyJbJpW zbga(?E&h*8wnwyLTuDC*fAFd}>*$j~M6xAA-jjt5QW%?a0qT3_aRX$E$#i_pf$WZ; zmI3mZIfyx%{oJr7&Zc`4?~vZ)+k5fM)*$Go943UL<&JK;W8LuZmgb+gQei5^vn_v| zGC_0(FvDbzF_g6sry-&-25yeS;;rSJbKN8(gB*+m(!yAu8&aCjvHr{cOn(>c`iIoe z|74T<{cc!HZIl*TN*wQvM9+SKm$A`Xwx8)P;m@nM0k7kD=q@bbbYHsDcu0CRqt@3G z4|+G;GiWNgn{cOHKDSW*{h982uvlnoM)#1lk%PX`YNC)?rgh*L{mpRBmBy3}D9a+Z zcLIhRz!QlJWp{CGt!Z-f0A)ZgzGVV3b-{vI9y-JYaw?%y#I`G>X~y$R zw^GVMaNt7IgoT(r#glaDEgnWNgP8V?$6YyzKXjoiT%Va;61Kw&@iWTDA%}Uu-03HG zNTP*0G6&-+%CHIckJ*2=#oIcYXEvH3rdr`ztf`wECQn<_tRp{daa?hhm!R!mo6^6z0l12=Z1+7*cY1!T%&NoKQ6};-6v;%RBuS@O#-%dV zgoBJH*`zX;?7n&4Q_1uTM@a*05lYtyt4LVGRRpN-`w(6DFreTJukUcW7yB+be%Hj^ zPYCCS;a}SoofNcpCdkLt<)R7$L{oJ2HgK5oYgINS!0Gm7DK@H+eF1nNy3c*Gc#^ca zXdM|@rT%f(DHg@SHnooC%+xV42kBOq@Tygo^rO>$jIFz){rxyBZYtqC@atr_Q;GKm zy(lNJE`7D>>jfj^GwQgU^k>{XU4Tpi{Id@Uh;2-!*%=C(TWd z{cI&C;iQ$qul^L$6UOmn4YJwL+($cV@pbgxMrrSOX`j8P5B5H64V^^u`}v+&>;V;s zX;9m8$gvpe@xxaa#J$*GUNNF@j3XvfpNsN>{%MQK)_-_gMT-D7va_EZtxY_s?RxU zHBV>t15sAj>d8>Lf2;xnc=8F&3fDX+zQ~mr_0CEPh1Y2e3x zr`g9Pfq~7OD7Viah^2qanTl#u&|t9f2Eh{rp?+v&V)LxK75O@v&1Wdr3B{@7#G#N^ zQ+%Ujd8K^vz`&;YJPXPM{0XFQ1@p&Evw>%klaClkcV=)I6ST~Y(--_J(2E}x^9d9V za+8)`_GicqT4UJ#H~TiTLf2_kS?}9UQH%f%n({GvK^Cx}fLA~H;F0%~_^Je^W<2ibjaf4-n$;t8wMx5ePYPqWF!y1 z3p+c7YBX>T|41HJz{|0nn88xsR+()cZakR=2ke7E7*sY@Z7Rp^so_HFm959T`OPj{ z(Wvl9<^dT1RS+X^Pel2Qmajc}$b>C*@lNI?Be7gEk=B9^p#TEJDXISBih6nsZTebc z$}-;k8h8E2DqF6;UL0$=%CW6_3IE#eYVPGfCezoX;dDm62;{l?>d31&EfkZ9>a&I= zJ6|4pV;8qm@O$It)1^G}d9{U|(iNyZmyZpld;uHMQYjpzfz9`S!Y*xibK;3-hKqqS zE_C80%*$E8z&Iu3YfWHH;8p7~K8_2l9B?{fy-g{~(S7?vrM^BrgA42tic+ti$6S9$ z3wmxtQo(#p4#nDw8o0q!BL;Jtr?Q)cJ_-zka{t8VjyJF1FGvF_GP{=m%!sOty5jsO z3Vl3Q2C*Z(!|X&f%V)HlcP+rQI)}ivo;z}#gcDa7B1B7sLo2PH+>AO@M-x*^+>h}0 znN5APp7Za}N>^AC>h0YhbhA-o+&p7lOt9OBBoP#F<9HETQVctp+h{rao1f7t`;&wq znV)qivG2(zf5>G=|2b7!lzmw_IRpb2DZ2UwPxaG6bGlzXgbgfFciX}Oeu(73UX@BI zeC8RKeq>7SRYVy2N7>i9S+)^Zd(Vgr1a5#qox$oE=O*pEs`(1^hFrM zY2~1o)T{M7GlkFcmMiil;*GRd7IlSVkwX=OJ4+JBX2Z#-;*cxwztaqn(1D}nf#lA> z7D4~A>TnBN*|)pIO|ziwYEu+4`+(Hv|8g>R+G%=Pz$Yf0py^JFm6oO?jhvV6_+E4{ zex+!MQrACYe5MSO)vNm~Z27K=k>*9I2SV9-h4)3&#u*RtKz4tIs{Djpd7^(fmR*o{ z=6X$}yIRgC$2mu`U#?XueoQ+0%cAOxXq2s!aZ*wi1$>-)1{w$~ND$5D=70?+kZVdtNJ6WyFsbSRF>l)ava z5Y~!nhgRtB(ZWUdPWY7XXm8d)1J)F-DTKU50AkBM&`*dZ^T2NMmEWJ<*(@a!>@@RC z_aa-f@(VJ3hE9uFk?>;Mf4R&0;&nvBrdnLJ$!ubRSTsozMfO0 zA*w+)V1hBf)kL71zh!EnQ}d(EnZl@%FMZiD&Me3U3u;E^h7t7Rayk|+4pyz)@S&`U z-?FX?LWc$!(cfLMKyN?_;k-$thARoS4tb-1RCxt?F3( z3-jo1I5}8q?iPCj^`g)>fY%ThBanBQU}}jl<;)wnzFZoJ`8(iIvFml;9Xm2uHSvbL zfmX}$kj1olz}Fd#ab+q(?<>(={9=b$anpj;`jBP7D)tK-E1CsvB7IoT z^x3r?Q4*FEeZRnOWJMu2wa*cVPZ1#N0mK2kPX0y5wp*pnZ;FlAR2!8gfLdjuFT{S= znztgeqr35f?YrXPfXAoAD8+3SWeaODlfM{M6*wKRJz6WY6;!|ma)5((17QXS-VQ^ieTdCVNIiq;>kyDD8H($x6$2!p zFy-SUzq*iW7KoFXP9ztTPA4FHV8$I4lyizbb!cVzmTha~I&tN^$vBz_>n~lm(P3BM zMwF<^iJ`#OAZB?r14^iIkH(b!9Oi&g+ z-j%TXj_0$!Y>SK~l9;4OBdU_?Zep>Vkl)N|Y^!!6P-Y%mbO4W_rO4nqQW&WP-KDaEj3f*+o_dAI?=a^0dZ$-KTsBLMZanqa&levn#rXEn15Y<1% zZCbh%g)K(N%DZJ-*;0?#)LB<0l^6gP1quFSl;k8e<5^gph_a-Z#!K(N1OJbvuMCT- z3%eE=Wat^X1cnX)>8@euE&(YKknS3~yBnk%X;8Y6R$2s+M!E)&u5X_A{oY^a{5jWk ze(Zhkwb#AwwZzkmnAmXp38BI41@g1s7%moG_w5afLy?|2(zC1WH~0F@z*Ra~n$N35 zIAzZNT>$KD4fwS7u2hFr5AO4Uy>H<|9>@>_K$4NUuEu z4=Mx%1y2CzGo=zSxkvQe0LCixseX|@ZP=nMH%cLru}bRGMk15W#~_ihC_36p%rHTn z5^;u!FTm}Yc-}A{3_bY}LWSCyoAvO)1r<$~+I^cwECssA(n{wnSB2N`Clpa z`$OOB(DUrgnf;UQHO5|iX6!lo5Gfw^byTN>Al9)K%oeNTQ0fAr_Y_k{pnHA|;=}za z%OX#n>$epJ#4)H4~0BNmn!8cocZc#A5dOA?slB z5?@uinVbf%c!8Gq{GGv*pwwVc3IV%SV;=8EKZPvJE@aCPetP@Yw3OwyjE$G<jemWHZU$F~xd1BL@EdQ@h;pUSV%Mmw#}h8};qN(P<} zUbOyeKLyg!^_&w{DPk9$lmRiuGS&JR2TC>hJ1*I_0Y`snhbEl}9x1qUrODA&Y3Fc_ zX6Od-1~KV!`}qox3Fgr5CvGEu5Hh$ux2V}Q?QB?@uDbH8?5e1X!CP-2lIO3)s?SWy zmfhz^ZuIxB@Fn^G9v1j&H~d3}C1-ytMH@b(fi%Uab?*HlVd@maSRr=Pam|*(5Hw(H z!!7*>_zn96QQW%Q>{4M16M*H>9sg{6c-!?om@0YsMkJBP#0BTG9FX26RB!a>RD%yX zr3f$^)BqGOk2%buRe0zvKH*CJVSJ(y3q1Hh&1%R{8!z|=DH`Vlbz`)n+?_w@etuQ% zMxb|GcCv<2BTe+Mnt58D;Fa28Q2&_Zi_`}koWG&_`a<3Ly*x*0#$aWm$4l85rAdMzJezx>v1Jn7bixB;mQ=cP za2u`|iDtdlJsNV-rc8d+BIH;7#xM%_nvRmVDT_WVge9Rk1yc6|Df}%8*8u4v2W`S? zG^3$W9;NwpU0EgtSQ~3`q6+T_U%KG#j+R{u5f389QKsUOBkR#((s^TIqWyGEKDJVJ850=M*T zZ*FmWJU6bHda>svR8oW~lt=0p)w8jzCc(=DlWW2D!PDR(Pr|N}86UK4)l>3{S3SU+ zj(8o4bW&Po&B4IPm8t;PLI>vuN6lNI~%Q0PjMF}%?#pg{rW80 z&s#uqKAPF9-hOh{{m1z#z}@XtDjuj@&PUpyBCMm2n8dIvB8jodzE0Af z#6>vdo$`yjw9eV?zHvw%xC;3JKEKtARwF!iOv`3IeI?J52t+mfp-K9x0k>F#b}0Py zJe9-n-sH;}n&N|cl)%8o$Jt#4*X1^}kL0Q4r7K%{L~Q_2hg=0(ScHbx*<0EYMxcy4 zW)>yTSdGq)p(7|6eMS$EgW}Wyr>6*mMzc#kYlW!O;(*T26@i~QG%3;);dhN5b?My? z)0xj1wtke*m4s>H^4pJ;e)CdcRCxxaxh>`t7&8nvuhL$-xJ16DF)Awgcy0sJffbR3{cq;Q{27%p3$2rD$m+_`fcHQ{dgs6On^r?ahfEgxJGob1Kmu; z*dA<7ni_!Pfb`0yqky*2KuCizM@w$lWV6{#7t6wEgR+We0mn;=)?l+d24neMQ6OM7 zBk3r~;;^kavQ`T@PgWE9L*raG<(70henvL)^Q+fExoS87$;9gvxl_g?>kzk2%q{A- z6ZEj$4l4&WAc`~^WQsoZrOiEy2qnlUmCR6xv;x(#DO`iZYPZjcE>)138U~vC@WCjc zo8egRb1OdGkfY8n?x)z129E@;oUr#v3Lxq-b{z@m{wFfxlG<-=n^uE})7xJ(PsjXq zJyB8e4MS{0S13)}di7Ba?If3zCVu{SjEP307ZK`W>5xY=_YPPO{8+KB)GcM~BZfW` z&UFbvw4>Bl{+3AvE@0OZ5wJ0V#?FWvf_vnR~cE*LLvK63Q~A?dX3KgN_uisi$T4{SV2L|5j3?6;&dtwK?$Thfmy<^;|%t zh9y`|C+xk8fVZKVmqJv3(|6ynkZ-YRy4MPfzcERmhjMWKb^ zDt~pPaMmr+z`HI*Fp_#g-IqQz_>%M8dz+V<#RFnWljbS8?i&o76yLvwJBRHh1CgHod8N3U#{DM?|aok;;AQ0j_xlFISKZP2O-dT*(C_11qhKJwn+<8W8PaX z(h2@)2yB*fmoXwwvLT#f>pvYzpokhaun_6$LMDN}SiRJU9{k#Jc1g9$=xpZ7NC1i0 z4@&{=4Kt1674hlk6mN_0h^x-FJuLROf{TM4b4yOHlg8=dnD<}x6YYMkX7tOVzw@i& z8%c#p;y2;i5<{^~1r>d*G~$4cv2RpBKH$$(za}9AFKm8kjxpl7=Mb3thoh;EVb1iz zj9>RH`A2OdojCw`#^TjHJ`)zl5>Wm@V#)1bx4bk@_~JFai#1M zQFGmCEXr56DA)TZ{*z1J40ey4{GU(Nlk{^y<4jT1(yt+~V6F{pVzM*TLiM{rVc$by4LoTR~l2=2a<@`u)}-1X*Vv{&S+{a|)O~ zK!yyni(W47=JHE}#&lICNUMwR&~ug*E9OaS$^KB(ni}*gR$XNElHPXhs=F!a z7Vf&9Bbhm3Z5Lg zCIrp?A?-xv9@0w9htMvc4knK!j7)&8$x6t zlJ(J^$Iwa;?IOkM{Mknezj&!ILAR{nNCewu_A2oi5i$js@a8~lmJ6{m#dWc6j=kr{ z{BgPqt{z#*kjA)LE_M2;-uzrKW^Vt03n4b+xTG7*@?cZwEU04*UEiEyi2(C`e>4RF`R@|EG_ z2UkS~HIo@Xc^IQ(5}gHFkq9uE{wDw!+;Mcl>>iDuzs?VR=e{&_!#Y~t;F84v+o9x4+>A2z?m{QJPhXr!vb*r;XgDonUBJ9)rOhO=F!I_N0q(Ka_Kw!TE+kFf zPv@p(asodDReH9gi4KsOD?C4cxF81>04^|#f01wgGO%#9vG2E55l~+J6FkMKQTssD zd2uTAZ<{hUUp-*b1wkoUq+b=>iSg(D6wLN578s4X5ljH~(742dP?k_xpC9oT-!R*G z*W0x!lgU(*@N^(Y?`0YJ63q0}gR2ScVF*>f&!d_AT6zE_*j~^ccpl4s{>81EjvpS^ zNYX7csH0S%Ibg)JktdQj7Vfh1xS@s{a{nv zkf0HBhSn>o156;~eW)fYG)>4XV!N(2=mN|`1~0G}6~AqkXjsK4e}S6+5naFip!dF) zSB9a5U|Fbqjp_7g_>*tc7CCgjFAKBOnl1$g>>&jMYJyjNdI$+hAyr(PIphH@q~%rM zmtLs##uVW?z_L1nU-;`rA;(kT2QsB*Odf;SH=`~?Lef7zOh~-68ksp3tjpfE@#XLX zYi4^(cspiSz-Xxga8n(M^}Es1yFz|uymLzra0;4eP30rSbo#z{sr$-Q;4GKcA-CdS zil9m{m-5^F%NC^Du)>d#+kak@T2A?Gb))C___a;XqLNf1(eunePwbF5&}%3^qOLD? zbxD{9AetDXmYmS-QXBOrHi{P0j6$<@@9Nerrt`)2I7!G#gc6e?lNzMCe~jwy3~HN zEEBl`QvyUpnCVY&HNp5SB*yGhDSj+~T@%9OSO8JK ztUfgIf;-b6c-fk0-NLYh2sbXI=2;O}EIl_?2Tujw=Ns(biy~vmx!={-^L%T6$z}j$ zo(eALzkLO-C~Sh~(3eWv#3Cr{&>^`{*^204wT&1TzeY;Eqy8PP5<5ZDAH0tOEt0Ys z-74MJcOAoUU(mzhH#Z$y(ajlEn8=WUY{xueeSI+|8>`- zvj|>Ji9WGZ?LHL<6!&DFAICzRcM4>H>Fb94t<(_9zqa<;WKRc7DCR~SyQXdVL&Ljo z`gvBVwlPe-PfKd$;>cY7{O=3OHF%xW9!E8b4vqtbc`WCv!zq3BxSR^!&ZDA?wn{*A5_D4e|4}b16I*KnBy<-oDhsZ zP1|^3-!#SZ@Af;{o9#XJJ=HTa)K2KPVh%XFFo$8uVq{hLHifQA^nJWGn;=PBj!|RM zRP6gsCL33kuD}5dATXVGZc$Hr9gG0vfJT#=q=tm!@`gxhhUL6m5YaT}e|OnRn}8d@ z4=h=V{$=0(q~ofz5VKZ*lBJkq!^Str~|r!K9I#&*Tyf{?4y@Y>QM9C;GAtN*^>)SiWL%+70#k{G$Q#+sCt9md4xl*7IW4b-qfTOi9RtTFW?N>8X|I zgs8lgl4%x;PugN=#$)4j1{iS%zxq3RrKr;s#)e%O&YKFH(!H_w|G!bQ8 z5hG3J8guJU-J-sBIXlq-k%S$$en;f+4cCldPL2kqYL@a${hS!XO zl-Nz3Zp;N8pBbLamYaluz`&S_*#42Mp3mANfTplfOO%?O=L2 zKU(tt!GG{T#1B$6S7bfXJi{3pCz*}Tsb}%6W+lDWx>Db)qHnE*oq9SyC-T*f1dtl9 z$!o0z*VPu=?P|p1kc&;q>Vyh(v=Zhy_Is8uKJQ(K-IvYvo)jhC@Dm)d6Lph^-H`cv zf-vRNRl(N5WF4!}289{8-w!Sa>T&j^M9t7GG?`(yXQk_DuM}cVz@3V)breAo!byeN%R28|hDK4MGN+m?S)6 zOMvFO1C8G6xm8cB?T>FfBlzN@|Lisx#8|!Drz^#S#W#o3)H~}Ec9oAaiWm%s14olx zb#v zuW}Sqv|V*nvx5Mc1@%$3 zGo12wsCs1$`B+R`Kp+u{;RI+I0DZgnvk-&u$4G?+S)G*G83)B5t`I7fWOUu(lHmd-MWeS!M-W|DIGzmiV*93vS%$)fQ zvmOMW5tEy$!@*8mkO7yJD@wh#6MSmw5>Oya%#+zgiXn&y zIqyF?K(y$5L3E^U5W>?>g{*me*(+wGpOry{}$?-?2OO>8)*Xudzu_Km1~j8j)05&MoG? zJxF(CK~c*1IZTSEm908qVspW#OZB_TAzphN$N2GCY}0t_b0jYwRX{=)Z#?G{hP_CF znivn~bO&8jkQcaTDVl@!RH*BgG|oG_q8F)mwu0o3=&~=K@SXi@9%?u=d#Nqd`ePgS z%dc3%kvESHL_W;NOLshf#X@W?naf3DTSp@B{$6w6_vX{i;?oz4e%?={XZEDO{MZ*d zO8d{*F`GqPwpxN#xmD}xbRbjPdIef0Ui^_XRh6)c0-`ngeHV8>Y8J2Dp3o0l?2FXFd+ z=GZxL$}knrC%Le!5Bx)EG}8OT7VazI`nA!>r6^er03Cx8# zXzmv*mVq%1aKTl8#Qo_@zw0%lHG{iHu76}ZPt@CUi2#D@d-xs0W%U+soqT&!dXxT3 z@)Ac7<>t6csVxCNSM*x$ceS3Dp!JE!aPSR(r8NEB-DA+Mh4_-|*^%5k(~x zql_1A2DKIvMB;NdWew$U`=v|k19Ug5TtaN5e9C2>#Q-0aR#Wx4CRZGh*gYEt|SFc9tcY~DEm2|s;zE;C-sp8heQWUL2ofFvBq=Du)Zz7Zu zlH+<4bTT@D!*u&lqotrA^hV#Ns(6AG?GJCx}!_4hkHgfaL2 zvU*-1ta%Vm8S7M}K;Qqtl^j;h;7y$Vvs@j#{awMESr_ED<2)>BKbZa~5RxUm&pnH) zlinR}&q?3m3$%?u?#0dS^|2}4KI8&%>XAP2MESyEATLMZ^t%6_^i;{Y;LXW%%M(`Y zf2VQb{|C#@h@-F!6#j~>ROVaYha77#2Hss)5%bS%{oq4-|*IhR~ zq+Xo6ylYFCwf?YX@`D@`>E%9hI$helEO{6S$(80J%JuV}WKXT;T+|xpK_!_EfQ>WB zMM{<+4GNYTGYh5RN+L0W(AIG2wN9~)P8BuzVh)8@k2lmu9t&2UyBO^X?SM!{K7|sn?Eh zyI1|^QtZ#(&}p|lYaDWX>g*3)6#3z&dbd6d zW|Lds#HkkQ9$0$0Ixl4;#rUeKpJ+AYvsj&IC@5xqgqv8W`vPz;rVEd^qLy%ZBxFEc z+Xmhj2ur&Jus8K$RuT?;g!RN$t}-d1v+ZkMdeYCU5uvxN4QAPd7!UEPJ6r)Z{VXxP zXN?976@k92nQtKvaQ}thPcdw3I4k&sjKA-ilse}YYoya@=o1e5{_?kC%WW7F5{GKo zxwU74Y!z6ggey@|!a-PN^+Yjcby9rJ0uTqz6yN9uRSRSK66`3JP>pZ*E+OiOmT(ZU z)^#5<97C(V_#djb-o`l2ox{<40tkn+LKjLuPdQoPSC}k{-sDCVah&7ETVz~>VVt>Y zA5K6pmw9e}ss5^xxsSi&RLvJmJxu#&KJ~|K7re&;C&XsUMcb-`ys0f78ucuZQBDBh&}EcJE4+X(H;JYI=DE@r)Laa`bkOVCdjn*{ZUfH z=gjxvTAlMOwU#t$_SIrdcu#b!-X02*Q}N)(uIiF`a-FTCbfg2!zsj$%G50#Zsp^)c zTc{9aOgDJX8_M<0YxMGqvwz~|a)5QxNj|lTCZt@u-aD}ur*M;p0_q&V;~JP>5~fpK zss40cZ1?C2O-}Lh{P0uAQgC}+@G05oSX$s3w19h=P5A?2x3?6ubaMy~WJX)tT_?0! zpm^^{CHsoC#4h2o{2tltQ35r2DxZpf7_8;t3pQF`C&VJLwU7<7dHXOk_V74E>w;u> zT2uf3iAkz1-ZdKh?K!%sHR$jZTCSaBFG~=59j39ag24-aMVWv&%1jI_99#)iXHjvY zadE(Cjy50l$arL&V6+d!(RHxeyc7+hZ}lpqU(KZQvridclV4s`6nRuf=0y$hO}6w8 z!Xdx`x;?^Hy{M6wY1N{mnG(aj0bfVf))+RoM(~Nn5I5C}t3v+Gi90{41^wGRm>{Re z{r}eb?)1F>R_{doF(m|(A(dQJsdNJpcPEZ1SBk}Ud~Q6UOqNcfoLsCvy>yq78=rB9 zzxW-T34dLSu!ul6pqJqCMj&TD5Iiw%H7BBHbsWW14~G=GSCKG68b}Ni(1(gxwTLb@ zj$ezI76T^r228x(b-5@FpzS_T;&7lmjb^H zyE$2AkZ|OZ53>U~G(He=Id+n6Q>5LJ(XGNn?s-+-f|hYCz3`!q@@%0TxMDHr>CtoR$xW_V*yvevZ}Y`pdDd$o` zoF}Bjz(R{Y%hHd9Hdl6ja*j;WYvuNGB&m{x+ud*SFk9J#sJz_%`dgI^q+h^ifoPzZ)qf4e4+Sec`n=%57iakj< zINx3ay)i@UE^OeW7V!!nj5OorBjzITi>nq}$9$rqqb`gBxzMm`J9|_;eM}nB7&Abi z)!2ccVZ&gZgjwTvZrA|R5(@iomNfmA#Od)tOcH5}530F2BR31yf2b{YO(Fl}$I_W? z3SM=*x5Tr1jG$l{z&Z;?0X(#S04Mnq!yW)f$asC8Zo-j9 zRCmM@Q;uv|tA*qXiu&5W@gCm)3L?Xf#tcC5(28zeau9OVJu@2v$%TwQET129}9xT%ypeTi!zn5`E!@Zfu6DQoR z<9+eXbWH=v6rHVd6Tsz>&!_*;)XoDcg)cv8TCt(qQT@UTJEjpJ{1kW>sK|+ zb7&-W8lJNQ!@iJLErnONKt#Ywmtg%69rzrBPClDtEc+Tm+phPy-62==(_G*mTARRJ zeJA<(9jRt%y}k7_&fz~;h8#Rs!#316J6t#YdXAic)wr5l7hA&nCKtd>^3NvMczz#|uvPWwjn z%Yv8{*V60LdgtC9dGXbLtX72ldUYbb1$^Z9o_atnzSz`LeQ#XlGbX=y_*?nuX4$Yw z#ZN09mRrNM7dL({*#*9d08D~T1J76rd|-6lkT0E;K5l!%1j!WZ&1j?>Duk)3<`>kK zrg3o^a}aHZJ^;MJ?C#zLVq>`la4*y+EMVy2`~utj)3KJbrot@8dL` z1z{H>0@d%-4yg!i)H`F7eZ7KyyakJY;y3+f^#%U!MobCXnbXe9{dzkGqr&{%A$~1u?ss6u=?LKustpE zCPxN12RkFmS};4OV)%mFPN%WOB=t68aZDX5jvQs?-zMAYvz;L4Y=#>wqhC?=|E)*- zGkrRL#np!MD)%`Yqvw`M;7@#%Af!Y@%eu>LeQNy}wZD5z+f4A- zvtYaTi+pytk$q75kCVogRpXFrgL~1283)P+Ub_&PAH{IY+VV*i;1jJ4B=1V>R;Rm6FcHT((q0Knlz0yH%B2FCDsk zQ{Cs%Rk;~aDsU>pkj5BGJ^$UyCXBES zuOjAa=?Yma{K9GD$(nIZ`}Zc|%6I3xJhC_v)48FLI`vx+rn&MeggcY$-iJN#$B75) z)Srh&MyweKEfX1$gk73la)zZeB#v*06g3U{UL$nxXu|J6_~HJ-9=ZvR<| zoX!F4B#J>GJ^x#cUnrvHDf-m^`0fn*KUfpG_9TC!p6OgtU3)usoUIb7xcB=fv%pMIP$!axPJ+a!&#c{vu~+64FoAmu1M^iC=#+*%sLd(m54C91DVMp`UJlidY->)k1yp` zerGsLbPd^Zarb}f*gTOfl^J536bkt?n8T^{7QQxs9ZP*LTxZDK=8LJeZW-f=7Udej zcnb=lhxdsBP)@CY#o?uiaq4t-_27Db_n7@@e1dV5RVUf~yeudLh6@qfbigynX|5FE zT|9Vwd1fAE7iG)?48vdM-Vk8d6%B5Gymd*LRudH)chdWv}hnk9ohQ1ApD@ZV#7{Q!n8n$f1No8n?RRNJ2XnI&;xIlz#^PL zymYi-NQmjy9CoE?vh13V78!`r6st#yc;r3JX}@-@`Nu*EdOGI0xkJ3&G?t349y%(r zyOou=TqI~T7$kC25Kox>0))8qgOs9MzbCy;DrNkY09GNSA%=~;AWsJhceOGG=^+_z z`$qE20tV-o$vB7=m-Q=wcWk0_uzdjCeo0Ea>@`d=VVc&6^3;iYhqg*;Fe^ID@V49H zr?*ZsEPcXJf1Q8Dz>i-YLFptS^1sD>neKmE&-AOxMoHFya|(Q~Lr9nuD5TBBgL5k? z{-RH?bzq}&I$l_ZFXAVri9!fJd_jivyWd@-`A>D4jWLfTYc`xv9RarFwrRJQTz4R zaVx9H$(`x_~eyotBMzPEl^wgfkmRxNaoHmbN#mqYKzig2mMb+(&SrO~#l zYwDmPw2O|*jLh4U0i?ey!i20>2UYvP;zwAg7mldfwHPrdWi|0}?T00>=sx$Ky zv3j5fvYEc0@rLM7;^O9lXys)J{+0vG{QQS3srzh?R7)Frs)Jmaf0%loz}ZI>8iQgu z3(NizPmjHZ`-tL_xc$sPIw7DFXGW+$v80$_mk#Ed|L%~eTu^q#E{MEvmP)S$QwT`h za)fj(R4DrS)<}oGZX5Vp<`dwQs-+cL(JS>+B=*>8fJnIRf};E8)}wu=px~ZG@zAmO zZRqgdKv)GjTv3X)^xIX%n{7M`5k(@#sJ|wDE~Kl-OxMJX6`o%OTm2&@Tlc17p0Nvu zR~o-uDTv^1b`xK^lp#cI&X2Q?Rz2{LQ1#sM&ZcPCq=u$Ekl*M<<)=L#6k7}kHN8U~ z{pbapS8m1*GB2bM2}DVPGXo{)qjYs%BOstElKCa00MX>0=${^GJ^U37X=q}+EP;0B z#A=)Vg-0t(r?qnTe3rX7bprA;IkXP7r(c}jR-DOwxUabFH_i7aC6z=1`2UJDCC~q0 zdG5__SVYJ=dzW%hyd?q^@BwN$`j2JdIL~=6>2DbaP3X-xhKanWMS_J4&HZr=SSxp( ztvSXY=Z_io->L3Yk4ED%m`+I?WD;Ifo`R&4iE&7W&RjNX|+ zhR&I0MT%6edf(o*-$iUbm9~@>zMa3VPcHkJiFE4|rZe6feoHv2VgS~$1^1F@r7waI zCbS+gU+^#iTn)d8*a!eTA*~FA*l#m)ns%vzrZMAN zIBI;mIT5@6xOC!sbR93=N{I-iRlqKSj=4=2rcakAB)#0_SrfuK;#{Xpm7zIAldQHJ z@Ay-(JORYG@Qz4GxuppWa$0flj3SWFTsg`>yAzlzQM90(D}-}v8Kbl zO0oL={)UG7o#nIF6ZUNi<5}l^CJmz!siq4lT`1_4eNXo0z=(z-23Nv{CGP2`ML#NI zLq=OpZ6fU(`-J65h-CXCOB&C%Kb@;A3wo^K-D zJWAsc(lx>>%I>oj673081~*j|i;tharqAOvHZ}I_{>A_y5*u zX$m{bpE9=Ls2n}Px(@KFEj0PVF`J|xOX$>Ex%3kWk~{OAY4(ji{G-5}Xj<{$O1>Z= z)83IE7I&b!d1;@rBsI6{5GVj0r0|1kkbHh^*^*t!fw_&@ zl~-t#53%8(-z309&ZMPOE3EWqbHl97-J&lT`@Q|#{YhO!xD#zT%t&@wGbjLmP+eX^ zGC?kKKaCh2YfX4JQ6q4FRh^L7yP*CR8e4R9XD=7OuC^0kJAbR~dGg|;;wM(E)o3qX zfF^xHCsJ{Ehm&(WOuQb0QZ5+%PX1f@cb8DikbQ!q`Gl?brsqT6!7-{)mk?#laj*s6 zM=)g4h+AIlo)r6GzxSE@-_GZ|&ZxswKr=dou7TQ0L?5?`O@TQ<%g()V#kb)%9*|&4 zgT+mv---~Z3le;*_~~NN7RT@hT{10!eM|VP8o@GUPD|6kTn^lQ%E*BBZo|<%zPeA$hQF2U z=I$t!muh5L_;mTF@Ov6?oPMPyWK?mjZf4i#?ILUV zfEPnlcbjyk`{t4Ebqg=%?Kz3j>VT&SV;WbTY-b!?6hELdT+j_DoIuJdg0rix$;nea z)mw7N{~to=-TvRpUmTpvCcj;(#h*OSfnMiK7avow_-Nw81c6d6Wjk~g1q1I?YJ`T# z*1k;mejK+xSP5tUA2&c0Gk|8^mgdGPK%DU#d4GCVSHQ~2!NNBT+_xYA?AR45jc z@T|bU%BKDO`wlWlbWJl$T7vrPTWSIC!|ztM5Opb*v{5i*pc7EGu)#t|0*&yZwz37B zf^CEFDZuZn4Zx}eW%}Ggys|wa;W6wwIvZ62392{KF_SCay{d@w+VEl2m5@tAPdURroZ7r3$eGv}Ux;$tkbuWN zpCQORF2NQ}p)A?ZN`X<}zIZ-?sb|A)u|UrovOKrm4tVc>b!y$4`r*nh?a{!9EHsI> zR2C!BoRV9c=)2+crN}5FRJVl3i}vVk62HaI5mNT3b%b0}{Soa=+WXgP=t+1^6~G-n zx_2GWWe=u)Pp`sF;gnxCX?4FQ17ozTW>UEdYX)>#QNq(QO4E~L4H!6@R*n0x|LcW8 z#ijYAVLmbx%MX?G|lG_k>gKkP^(Y_ce1kuQ1qnBUtTZ zDz1N21E0;XaxOlPOkMntS(vg4;1V`Z{8HwHA<~iABq=-F<(*Oc1w1tSD(%+A2!5Qg zyAd(2kBbF6UOF+uhZ<<;>0tr5Kq9muu#8sI_2#O?bMi_1*A#mkr5Jjg9|)t51*a!2 z^~8x+YO~J8QVgHFdWWui;c1VFBsW7SJw3SEeq`jrVVwOk2+X>aHd~YhQ6=KXG^?KT zr#8Jcbt)kv`LIs#2T~hf+nes!=1>G&Z}YJF2mX#XqLJS38G9t)FrfN>5JP95!uqTK zki&HfGrq3h)yzWUEz^0~N@-4FjWS=OcW9NMOm;WU)1L2}*)-%;S99G}fsY^F7_*|Gucl&1)l}-QDS{87|4?)0E6*|Ax9GF%BU(PKPtiPM z2KpM*^dmJkXg+$~SN+T;#YZQ@o>!Xu%+Z@9kQ#y-sCD%8cYto8oqq^$QLO(Shg>Pw zhR!{{)bph=m$QWIEyVx>2zRaX-Kt6ybF^V~r3DLCM3r1auoNi=c9isp9+vf`YK8NB zcjb}9(~;p9;5>eug3Dtjia)I#$YxwXKLnEwoR23)AU zt^Z2-ixxcgmN#r(D-2&19-3FWgY)v``vSCK$3L`qAJwQmOl<|daieq!bi4vUt3&Mo^+R*UqpEb9Ae3t-RHPG(hIpA2qd0A zMiNj%9yI=Tp-jr%Jdyok3@c2UrQNdM^CZk~9E!IJqVUSEdaMX_hb}!Z9(}1TiT81LKgT(|VVjby8J)9S4D+n2Q}t z!fll_f0VN&TIw^31!K$j^cj{kpsRJ!R!aYbfSw zfx}nNRdsIJ>YjPK{PAYj=x5CtChP>8-9lMIa%6@)ig!Mpw#OCCmjiM#LC`$;U`RHTuoeqybO zw2n_rN&DtGqVU;qV2Q84@6U-o3hgVCvQWMK_f5*~RxNZx@>x#8R}yDi?A}%1zuY`0 z@(O#Wm?q#Qqd$^ea1YpqVVOW!CW1BecS~}Isz2T|g~z_LxcIjC0fR0sI$~1u>&uzn zAOowA$^a@mfXKdlnC`>uK08+&gSwov=SxWwiG$(>g2Qq6Mg=I{0dPh|XT5$)KFngT zIYdlIv@d7x6$1T1a3cXVK7Q>HK0t)VQXfZZY~mRRwWUIV_U5%w=urHJS#>iWdmc;ObgA)l3pAI9HH*y)Sn&cWs;=#K5u^)Lv#lg@m54% zzREfQ1>ExR*cRyoyTuf150s)28KAFtqtzzT4KjL4`rr=?>#LiLZJSD2WeWXkKHR;1 zIWU4gC$uwdfL=tPHfF;~4R zV7MrioIlbLDD8Fd*93Uws3bGu0{n{S{C3@2*sMKt2m*HB85uJ~E1g)N?QZ`)5vz8& z5WMjF{xYjr5>N=y%Kx))D65|!o~Xxtum;MM3^R4mrtrJ1y6+|pl2_6<>?;H3wL4f7 zyA+|f1s)4~`~KUTF~9r=Sfv*C`3LXyKDp*Ac&gqlD(gGy`-BfppW2$=sdw!lHi$DJ zV^lx*TT4iEXNh+#dWJ%Mdp~vZ$e3%GnEDM!zEu+mBG(ce{VV&llN@VzcTO{IW}z{{ zUd!=J))$*|^MSpc%+gh>=d?^R@HRZ{9-j)C4gHnNv4Y3-rX#zFM4BfMqeC>Zu>!us zUTuf+>@moeec8B|T@7u766Ba*L!Jms= zyI%S2hJCYbSQ)9N4O^B>bvvceY~$OE{6I3~wY;wOr!HbV8U|7j61R5z;>yLfB{1}3 zN`{i?F(e%Qasuy;#y7`~+Vt@`;4%5zLOb|F8&5cf0K6MYdBIb=S9Ijz>fggWn8S}T zWLd`DR12(4`bgKKM5~ZT{I+ogwt%j@pXBc9{&4(PPI&q)iyyU}<)yWlI5nFy?NNXPVblGZtp`R+{aJ3jnALwONAR8vp!Vf1s4I(z6(*iNI_ zywvpcm(Y5}hLrhnA?(aEDg%W$e0ZNFEeirFx_mae_w#|M6PSPA{oLM(zm@hRQ@1nXF()41Mf+4!W{>I^wXe#@ic!`jl_wb1>T-{@-Pv;pPPn9Vo z>r4~){)@2f@C_GQiAtRfN$#I=`WS?l-3 zxKBY==LB({n`#*=NoHhH_g)b(u4?qMp^co{0;6Y0`;g8%jCEMqu^lau;2^TsSBJ0z#z9tF} z?w23ieOPi+vEbN2aCY( zY{4Rhd4Es*pf(K?2wE?-LihxcAN3jP{Zq!0Hc&XrRi!vE%;dD>o2#=+;YOpL?~>I* zWP>(gWt+dl9Oyc-C9_gHU%x}15hB(wjpEDGl!~nj_MW;?`OWfH7`-QrKiYMmH~y`Z z4LOl&dQ)^g)4&xLy|auBb)KB(ezR3WM9=b&OTiN_H&9e5HG%E~uV+Ar?#r9P_jTyv~)fiA17d-&x@T!*xe`v{4tt#a7<~Ra{*==v>{JYc&BX*l_H_G&$l4=05 z0XhBEya|3m_Y!3#d^a7tkVlvo<=#~7+CtSix4hsRc>?6rj-M@0OvDO)2gpp|Xl-Y^ zc|kSP^@2lW++ZKX>ht3C&%Kx-6ta=OzT=H!G33flF>Zw(R$nDlRIK@FF6T2vFqo~>arUpuG8jT8We05J4>?K<9C=6YC z6B%F;W>!S!IG4JHV?DYZdhSs8(o@e^bhXj>8X~dqw1Ch`-J82)^f2;AjxU-ffo@g+ zdc>riQCjK;2zk9pxrEZRJ> z#+p!q!*6m#d93Cb+(I7;(SWS3DzcF)7==&LBR|ifSb`+4MC;{ZU3Vz8jN2jGmtU-d z8Q`O)6bkiIR{+89i$Ug?T3GbbxDm1s)Vi6P)4fLs@ufq1Q>%?bP)6yiWJ>7Kl4XK3 zV+dpcp>r)(`xsJ6E(L0FwK0(oS=;GW;I>@VEL^%7jaK+tS`T4D z%ujx)?$3!7?uE}k8jdOGRtGB4ek`MxxmnwPcOdJp)<}Gym$;mk;kg zcC%B-A(LmWs4QXJUl+k*z|W1dScPaRzM$nCc?iqyI2#yN5hp;+ zmHa0snjcefk3~yn?FBuo@_{>K?v?3QGIR2~#>LI{0Q9Zi7}{QTpDuDkk~6tT8QG?I zMHfvTVpEDrK2^=09Spl4EE3Gr09;Ux-?FUd<50SkbO|bquXc9Ab&g^aK@cH6h$ad0M#VsI^%I-0&`mbeFqP zk)G&HxQ+^}>aLebB&H;Q8vN{`B^bpME`)C)yib@uPB$x8=%1V_pq5z1)_dR7)Snka zJIFnWzbo_B8B71=<=5RY@9Lm>NWsUQg{N>TCS3JP!y5ZgJKTo({Q}8=9bl z&r>Vxl`JqF5k4{zvRCzFnt!M4QDu?xlETO<>JaU?;B{MACxJ?wQ(ZQhV}qBkg@n{8 zy__B1zOMJ0zgu+6Ioh}TWWv)LTGJ?>&kR5LK3iO3OsLUqSNAJy3a4!3WLoL!THd|9 zF$_*nwxQ5`CJe5r-`eh7LZlaovq7~L-+uvT8@M5EZOC|ZwKK`&o8N`1Z)Ki$b0_1^ zs+$sE?@War0hq3o_65UzxaDN(?~Y-f{fz$n+$Ze%cc-JD3O4ppEY?SN3SQ4)SzKWl z;xI(?2qV!CF8U#R`0h4FOd%?5o2uT5EcV7{i`nTv5T1@w+B;YgQ(-BPo7hu z{_EumfMTKRkX^8@*`?m-PWjIM6f@)p4Skg^8meh8SS#Z8n>KAEQ`H#0_MRef6#fzQ zCze}EMfjZIvCxTZlpBf&He}S11W|4@;t3h99cg|gNuqoxE;lSDRAWIsZNY-ynGSP&yf$Yc{-)j4o zATe3f2i284qD9gEgs5Q=Td1Bum1HG(?)vWGIxFd->#l9cAeu7-F?~If#uML^#5ih3 ze2!RXxht1Ri(|t$?yM8-)~vSeC(QlbZX+sA`01cXQA-2MUqX8nIqcn>YGq)CU^4IX zf@KEiBT@;(G0XRf8v{g3H|b!`CRB;nW8U_>p%pfQo&l6;NK}W-vcBA2QHek6*_3_{ ziPw8#nJAJRLS}r=gj-)K`2%O9==F~wSIplr!?spE-ExCf_~LI+I6D4VroH*p)qLhk7BdRg|8$!TsDk-3V?;Nv>yw;)KM_2=sG+vIS z=|{eq(HOIcdt+MB6;GnF>4cxw97Xp%W$a$Ix7K8LjYLY#{@N>U=got$+7g-UuRTf5 zjccsVON|!`=m-kv({5=iP&5wm8co>)4Nxq}L4bdu7WVl5$CTu8u*i%ND=uf7X#oe` zCtz&Ws&1tYkSpb7h2M?vCtCO3!N!2XV&;Mro=aElB%ZOtO3-m&7+2dlfNA~9@HujD z!?&O{*eYk$v>No1boXNZE74yI-X#WV@ud9a2oKTW4;1$oX~|7{gl$bZ91E)Y!HJ(B z{GPfML&q`nh{p^(8x>WbUU56qyS;5Z=CvnCFEW^+pB>-$9>jOze@o+P=aced9NCSE zRhiZbnfx=?=Cp8vc)y<-q^C0swX=zO62_wm(=C13c8T4NpC4>;)+@+3tU4vN^M&Mn z2!_n!3V6IUqMpfRNJwieD{6~I|DXa7@Y0Y=O_9;wc$d~Y|i#1Lflzwhx4{uEb(gJ zhtqVbYH@*o?metFehLDNhL@6 zrA6(;fSf7Qy*qCT(6xn-4O0a9N|)?iiKAueelavEKEhE49g`*A)@&1pl*c(PJ_mmFXT~&6`*c%V7IU0g;!$Jle)vdx<1E1WW)54%nHYNS!p{z$N1_SJ7 zOy^0VpXtJN;!0BS~gv9)JO~8*$zvh@-u&k~Lce@=UDGP2= z?d({Hs^ho(1OtjR-i~8OywGZ{bIMzbl+FUt;_0ur8}DN0hsb3}b>xL<>^wZD_nP-@*GGP@KSh*ggQJVv66=B; zbc*IpS4se`h}XIePN5RMK$#(gmg&qVt)z@`9g?LE67i>+@u36fo}<^LCB1sWXX~-r zNp%A8B+ZAWYq3PRV#%D49hqjAvyUb&sPB3Z35AJkGi>rmGs$ z^gKrHN2bU#@(nE;6&K(%c{>7c1Wz+se7(NzdJQczEI8rNeRbA>(drxHwfOF?pN_7% z=nnlYKnp>7_p{6zOL;S4fS4+64*;L_M(*z4W+%p4Wl2x}YO;RMdI&)7kCu8-2y;-? z5N|lp_rkFXgaZ#dUinI{d+&pHS?Em(RT-1&&N)Pu6NUwP$x%#OwmSE~SBDZXUeg<{ zo(cJwYBco?VE+%+7ilgv>fIxzv%T1= zz-;m~?a3vcGcp>UG90#)0-|_h2gQ9uV_ou!E5(va;9{!t<|}$yv*?69v{vV@gGpCo zemK!pMQAJFfW0KNOi@37%s*>?9Oje_%2FEhck+q&B`go{{cN8)?x631VKD)D=5F?d zh}M2AFDu6+$mr>1kDt||Gme1p&mk{xAiKQ|$wYL~&SkpTV>P5oi#(0Ur>}aC=Rm!} z!gLrgnu(>c!`lxPQRV%BS2VzG!l`wnOO(f=`mVQ|dsV;N@M#NSQ<8?w$wbGxjr=W+ zI%i6B(9R}aorvD2z(6c8ATG`b`Qpe*nggd+L@p)K;tKt1d$Ynv?ljd5{FCy}=S)A$ zo|aze@25@x6@039Y8a6B?4UQDcg!I^!o(z3KMQ^fUnjxT_UE_t1};h?iW}M-al6=? zGz!BZ;+Bn6+-!KJ$SnCRJaQC7V^>|8&6>m!slsqM!PEAPp>(yAn{+BzfZEwy`11+5Vp?8Xe-M+T z>%15c&POdL@}=l=U%Az%roxV=BCwXRr-$%$Q$mM2G{U=DmT(O|P7mZon;2AjH9#vz z;>bnYF)DjA6XTN^{;(^+mZIpAE%D;qXSP->R3}9J%;(Atc}kwE8ld~`eb2R<6TMs= z7XlZhPc{UAX-6@EtH@5$4?)?oT*LsfNM2qwy*j4uHUeKuDrwH*MmDBedYc#B*1X|I z9(J-VVV4ESL5+P#c3H+Tn!!;&)ag*l6WfQDjnIEo zrn;v9#niMVq^8cyZ&>D7c8J^--r6(_;7t_Lok?{3!n0@mc>t%_&?ncHv`V4Y@*ujw zMMzsRznr3oXy~|t+mA3pUbPI3;LN(E0U|gWYf8Huk_}#L*1xQJuF2TCks%>9}=S#Gsywcz? zTaG_rDj8vi9{tn%-0(E@eRB$PCGQcMpvdpDNM5bTA;{~i>}}*ljA7`kD0&cY0_0J% zd#QFx%PTa{8S_BX)pGc$yL&hvey7=C@m2^E&*#qKVV=8{f9RsI>lC9A419BgIpK%HSa;Dc%ODsvPI7TlnL7Z=4aT|`EAfgJ%Q;-JW{vaAMLNicX!%Qiz zjK}dJ*5PNTZ=C;WQ^$y~0Hf#htv&Vz5XIE&>ctb^d{}D5R@&#^wZGE&-CS zxu80>^P%X0PonqF%jn9Deb3w`41|$ES*({|0-cdseYow}ttYo(El%;!fXxjWa_W_k zS31IxmsAsB1g}a{iZOueOl;!}WxIo+6-Qr4W_WBi#M;0S+vmPn+0Xs6E-!yO`A$~X zX-!%%4|f;L@eXI*bT|KOxKr+@Sdu5xS<}KHP`JtToGy=mtVpbfc|5Y?p2jA-Y}2;l z`BAES<$JZ24|p!U@thT*eDfL89TZ@H>U{O31^Kmn$;@0uu2;+usC}Kxfi(gcpfE*V ze4x$KPpEm?m*y<{#nXbYK(Ow1h0`qnoW9S*9{|VN2y^MJEA~Od&EYZ{ktClNTEbRD-5v{zT3qD`9760buM=%12wzB5A*o-K(_Ru73=V4T? zpK`|TWXwjpo+Dt_3uGe00)Z~1Js8suLKyiS5u^JoD+W$?3#a>NCg|WeEt^uq^jY$Q znH)FIHf$msq}Qz6#_cgiIlT_aPQhU-;4Ei8^Ya&))xUCb-{tYv5H#&i=bi8-HGT;& zbl^KPhx*L>M1{6%&SGm1mJ%yp}6zUyu6W$hAq^p#D)is=!{ojI8n%ulk*U|q2 z7z<|n7qAj@#s$Wt(8~@}5R3zIN^KNZWr1%&Vh{bq0XWc%z&WILq=GyA>!NmfkXt0Z zeWLkNne{1OO2(U>{iYVSaTh9Qv1PV!1uN|N_NJ`4ENn(m!i+-#lSCs} zKDAHCuAXayjvASEsr@KU)ahOW7b+xc?jlHv5>mrxAN%g#*DsZcHfMrnB%9SC*LoA(dh5ZR z%;1Hs7w|o=Tp2=pCg`3=iYGoU7fnKe@J*#v!!4~y&2|p)I}sUeqvby7TwS+s@ozo{ zL;9~yg0|i$y^A_XfQuq_eCWb&s$660-YYElgw>s&9-tJ+6BbB$zEfQjk3OWVXnNiU zWlevVwaxe(>XAvYwAO;)`O2LW@p}*YaPFnp&!rUvjC4Ps{t+u>$hh#viX~xEP08n& zG5&_>-7>rV^^vA$9qBi8jMYh+)2nQL4Jeal##|$}|I-2}{FRWKq8dCvaO~qxN0^@J z>vf?8xp9adOfbQJX0+hxmHaVV2?s&OQZ;eqdSA7tIbF{JMZ2yGk!d|a$fvND;61p+ zKQD$wPFKx*IyvpVxNOsHa`k#i=FTvWoImN|`Wn9Wzv{t_V(R~Bjk3XUD>wi)al36C z!_6_L^2s@Gw{^WexH};o@hzz{WJW2BAdKFit>v%)@Bm@F6~AS;%;A#zj9B0t-L+Y5 zt%=O{Dx0*1q;6Gj_^KCh^~b{Z{?_?-XMyXUTF*;j#Kl^?cCTlDxXTy)-ou9H;fhwf|#gXZ5Dwq6xN;o=`?kc6)2{VD$sRIe_{drpkXQj z=Xx&=*pXkFEWgU72t=`xZ(Ao?>0i;Scc=@Oo)%Odh!;EWAQ^J|RY3c>53K~> zxgKqF=jtED;g^wIAThSXaQ>kf=ta!ZLo7lAE;+p_$NiY!OI+>q1|w(FI22zzZZ}>c zt5=xDSZ&Rqqbd+m{dwc)F#NpMZ?YMkYfvau&iLaWn@5FV-H$1P3AEr^WKA znnZG62#78CehgBaJlqoAE9g?TXK~4;SbRVOioNbW#IxX2K0NBH496X>zl#MkJ35zq zQZN3C7fC~(-$q zn>Xg8DBcTG3y6SVd@GVOe_FBfk3C8OPH8%2Iw+1G49BP(xm(&b3!lWsLzZA~+DuUV zn6=QtI{S~A?30RDXxY47V%rg==E(brFdQfG{{EcEyeZA7x_vFsLGjU7zqLn9^OJIT zZ5roWAu;7ODp}9yRMR9imF%%W@8y(7u-3!Afu{b=h ze9Ue#psI#*t_I=GDl8V{#ccXM8AFd*WAF5pDi_N%YvIfE`#`_LL9wgMboiA+Y2d*i zXGh5FKr2%|QU4R^?2T(wCH2u|xX!7c4bF>w&oe65cBM}W|8YdqEM`-i$-u5SQ{HY4 zH+PqwZfk-4)}g>Rqj2-jyx8B+>QRJ=YsuX3qVo>=s?Ft35XwVPX8_7-YD*pF18Sq!|8oSD{&xi3 zwpRoXH7TB0-Yq4WeqsqY%)K?_dQwZ`%oH2Y!vGq2(6u?<8mayj;Y4S){=~ zi*lZJSdK?R6s0p>garnwxAZ-oNz;|SmoVd?f!&W!ZC%rMmR}|T{*<$)GGi_blp;~* zHGyC@5u&u##DiOlz~?2e{_p+ZqHVpG?wxW_g(hNL*@35+-!TrqsWRqWRawMxYKo(mr zxxy@3b+@fs1IhPnJps5E25yId8Yb&RzdEC-UB8~k<#g;imZ&`MOkf9f`lspJ>WH-~ zN@(wFNYQ^p1=Rn=e*ITy)6>(6l0z_R{c(f+vWxv~?d>u$%Q4Foov`|PN1{BBbn z8=Z94ZQ5#tK~9!ap#(rvNtL}hi$?p5PS@Uj54g->|hk?W~RRJwOF4;-1>+M9za^Za2-di zZZ6*DdY9i{`z!X??r>sOX}_5^9T_O!;t{if4P}h`by+U^4x9+`Lad~qb&TJ2@*}GK zB~UU;N@GD3S+IZTleogQhX#v)tGY_UG|Tm*vJooXGfweZR0Nxl59Q&;pK1;JlF`|rkte)d&k(gtnSe!xywzGd-^_z!K%`95FKN;c(Tg5p(*>WV_~FvkyJ~ul)g*rLvtwf*2@cCv0~% ztD$4KW+FD|a#IDKkK8c8%KK2Y(oL&-_P_zcrLyd`xCz6B9U+PP=|dZ;VU>$9y*;mxr?j`r}FiUv`# z)~?)EIDp3a3p_^yMZI_2DSNf4V{`S3{&5p(<+wwEZVt-Pp9+c_F|*GMBK%Xsac@e= z{z0f#iU0rJvw{Y$iVYnV-0!{`*{D6+|c-Z9gEaEy0I!oWxEiu*# z>*DnkQMC)kte7gRb{8W$B^Rkq=MRaUaCuRL87_2$ps`BG{qaFH{Tm%;3n^sTJ-0w-ubI4j4bRX_E7j)b11zd|E>Wvw!j*)D5smHP-NApVM^~hC_>z;w8tv>A! zUm1AXQR3s25DU<1W*et5bArkf(6|Y1sF;8t6}0-aa$)H;%CZA^=To~Si{s3vHt+U3 zL5Z!(_hzeoJc{Q*ZUy7#l#H1p;(T08Gp4$DU@RQwLh3(x-PVm$F8C#GsbvdMx}0TJ zwvX@r;%2h@PNb?1+d%4_sFaPTmt|qU=6lqs2SV5DufoH)FD09uG%v_*^bCKN>-t}7 ze=H6gD0Mx3D?hjZl1~+wI;>vUf+v%qQ+8zm?w7D!63fL{FCPNbXqSd{Ui~3T`Fwx8?%gB0jU=n8t4A^%|4s=`$E@{zqPwP|3PLzxc@s=_1?DtQ@N4T`wF+xO*k##vyrvj_lOnS5$ zUJonH!O9{8c)3)b4Hy-Vt&AqyZAOeH3lVDM{{&MXxjQR^)YX(WG+%;Ck_pJ4q<2$KF(&Nk5?VU4pX;v&DUq&FY zO(bmSnB9~H407Jeo)ir8W}nP?_jQ|}l$N;}R+`fB1k?LTbBY@FFFbY@ap2%bDVy>= z8l|q_(A7-wB*R~e!nX0z6u8(+RqZ8$WWL#T(c4K}ZHiO@8KnxbL=1?|BbQ$K@2{ ziuoOiYex(j=wg%;*P1>{GV|^*wVG_D|;RJa}8qF#(K_{d5)f_gMwT+z$B2 zTuFt-f+Vg?2x#44y-~dCpK6U`?bpxVlxc`eNOdu=zBP+>L3jy-p+)Va*CZHebQ&^Vw`=S#(a3r4@nV>J9_MM-f zGvFEFt+8|hY4el4$V)^?ePR_my1S#D)z+@o24Ay`G2=I(Lq8G1MZxO&*H)NA0&<3Z_mA)iah z2X<|^4*bGyUpuX~J0L&$A2<``kpT-kE4H}0?qHw3D%`@vKAwIqAC?acL6}NP>FFi< zoX{CFJ0s+gZ`T=ENa1C09weR0FA+wvrlh}=K(iO)i$gRa^Q$nEeGpMWG0Tb$o(ptnv;wGO8YG{rC~S9eKTC17Cf=D!~HNU zTTmB6Tr^tY+cZE_c_fO5(+oqC3qURrAn3DmHB4_(Nd2LoV6I6ftbLH_lgiSo|DD)WId#P2EWrFHk7XK{AOh1jn1Y0O zVQi4tMs+TEm{?JSn}ASa8}-S#v*l^=d~V7XwWp@{&Orm0MJbwBSgA1-@-61F40`i{ zOB_mj-yVd$LGs0XUJw!LE?LjD|K}(4Tf8GGO)gxv8=^k966Rc^fFh$vww2LEgd(>l z8H*_tkY5SDUgI+lq2gI1Wh1BOFp-?K8z$cSb9(d1sJDWs9L-H_p=d@~pGRIl zsY1%$gWkYP=qSJ|FSZ?_3hvD9<+~=JS`oB5=8Et!!uk~5^A1g_^H#3z-wk+BZ|je< zW<~}7_{Wn_uiVKbFHr=r+^j2k^{P0onO5STVlKna=s{aj(qwS6ngvseGqU~2M!*6G zN(l9?iS_1n>#{WA>eT|6;RxoF5jDAyHP006$>q#!kX^C-atT)F@4*N^ixm>cM* z1G&OYIh^_8E9+YQ5<||SY@xq(W9{bR)t@``uo2pui(d8t1kj-q9CWizJlz@J_T7sv zfA9+qaK2^A)JON0cvR|H>bT#f)3%{b#oZ=FFp!TyND;E@`gTAWC#(5ZV!#`gR0#%n z5hQKiT1+-eC<#H%(V`5iMY1#Z{LAbDXa6;Ccr*f|PA(!L0UKYQQT{rw1Ju zDAmrzWFmPqbN_&kv*r{!i!Oe*P@wOH8|;NRSxq0#ByA%^C#j%s(rtaUQ^V^1JRKL0 z+ML#*&T@T1yTopluWs%x`@R28Lc(Xcl{(96zL9RTf^=PxTxwt)&L)`UScQNWu1?n( z$zSgN4&Z``3as!KoFnPKcjsM5G;}4yUD)ZIyqQgh87Nz1f7cP|Q8G|JE;c;XEtLgr zQcn-KxnJQ@KBgZUxpTGO;nMw8D;JDHdm>iRi$usONoQIzrvIbhT$K-b-f2t@Ce(Zu z3bx5u>f-IgARF}4qR(5FpUU5Y*_}^>owQ6KO~NLJ(??Al%{265GxHRUO5ZM16&BNr z{6*6IQkCxi?t%3Y5=S<##uL{mp(&lU1OUpG_2r&d@ zQ@f4e&&Z1JRM;sc()XjI-l@tA5p9%MZNEvV^Iq9k-i~# ztl^BPKCiU>Fcy!PGLX6|nY)%X6uwWwrRun}Af0*6T{`?<1A>;*Oibd#?8!y#txfk4 z+a#(;@n*qoq-FumVr-!3y3V-Eu5L$LaixmySTqU=wF*b=hwmj6_|aH#DM1dd^W>rS zw^Q~t-DNnXVKg^zJt`9w?GMVG#7|*VlhnIi+sqOcw8K*X*lR7NghlzPdHje*-wN~u zOs@F`EVm`*$A_wvs(SX>uJ1neKPJoeWl!VWw#ejG#ECJ4N!iq5M2y4z5@Gh{fSGkqgyV*N!{fvpE*u#qDItwJ&9CV-_|R5r)GIuSX_T`-{82n z#^C|GN}jcTzU8DCIUU86dTgPE@lYvkrR?s75m@)AC|){y`ja#)Voq#0xz37!blKq= z%HnO6mV2SvKpz|UYoL0oBM_V{>6P*?M|0ByiG7*6e<%gJAu-{*L`jIi0)pF_5NhlX z+R;bSeKVE9rWLm5YcLE}l4AoDMVVbMEP^lY+adRyl)fbzEP>vHuhyjDUpU6u+z0Xp z7IuT8h4we|2Ru}2Y>xK0DZto&*SYCzV$ibm)FfqjrzvC?rm?9PYi^FL(AS~3wNuWo zg_^Mo^jK~R4eAzE=f0<^)%JQWSr;(uFf^NC_v|h-O)|JR_j1=({tXJ$95sAZ8alHx zT(q&@A#~m|OL48Q^sXMiOm8~?simv5lvt}!p?N;7(k6qPDS1BKC(3tcQ77Xrs6nKs zHp#&iyy-ZpR<``nde*|ao#H!v3iSbHStZW{uV)}J0eF;My~Y@_qwBv}Pu5w#kJhGB z!P%@uCK3nv`Lbfj5K}pyA}FJ}&+~8u_letg;9duf-JAvpY)0O_xdoQ|^&s_X0s`z~ z{ke;ku`OFstguAyCN8&pbUOhPNn+2mor?M<(HN*HniT3yb*6E@F!HdhF58FFX0_Lh zMsF)adfC~FTjy_oqw~tsxIY(X;J4nV z2tmCn#@`%W9Ov`{&BU8;<%cN%4Myj7Fy}Od;QPrP`>^48%jp1!b{y7{Z;yGCx~)UZ z-&F4ji!FkKrrMzFS>H;&nedn+o*$HD%JCr`DwI-xx+3p3`g*!G<7r!4P`2kbQN$u{ zctHncO0TOaSW@x53git!1?%w-^%N@!+i#yone#aCld^mFR?U)z*>*ICa}%+aC*B`_nNGw*XY&-leOqig--=TMoi5BI6NacQG-kpFimSi@a^c;$ z*#KF!TA3fH!%K9ccP}#v*oc@uly;U0Du3+lr-}-KE98_qW!{zW8VI(S=I#Y&oY0De z^NR~qTV4tn9Fz7?THUS~HOu8xiblCY6UqQksJ*cjRUrlP$8cKN17vN2mKe1O*NyM$ zgLnUeYDFpz!3F*#{Ho?_^+`-+?Tpb_`cw&l;BI?bWdX7@u6{V~{W2T!E%+0GUkml_T^QS@5^i2a@!-%HvZ;4I(z@u=cMLQK;V-pq znp1znnvM33@{RzK8+dFo4<1qi_%=f=SO#RV*?OgP1(e+jvqN=xS}^;FmUlx;vPqI9 z>0gRPRCoeY1{MVJSCs+$n4p{3WEZaG>%QG0l#`dfcJuu;@`0b-1V}BwFfHn@$fF!u zi5G*FldHv#J)?Xg4rEl%*tIqG7FSH)8go5uvF~{c$6S3$O`80icO4n58p+7I68SuJC zv6YzaH~!??lcl?(&#gygu0dLfRx<|E-sfpdgJD{zdcrD=+ZG~~c&D!o@HRIs3gns3 zkx06!+l@swt@4rGj>T2Z-qwnLMNr^|A*coZ=xoP~l-|VALH30|OX@rsgVKPlksK`) zL?5`bHJeg6&ZK<;8z%fZ-+RYq>G|dh_Esk|oFRa;heakimUO8 za89^O}A*YH-{ zxGlCIU>#h;M@d7iqaA(To1O<@m+=i2tU`$GcZlWt;3P2Ycsx;mwb<-;5o;u9@t(vy zINBs&BBeZ!#{sSK2yGJ8u1+bgG3(0==z1q-nF)`+53le@PN()wG59r6V%hU6cbAw< z49MJ#F7(L1$K7#?Al72iAYm=6sFwNj3JRrBYCVzeSH6AT3r45?;L8*G7y4r(D@FrA z!suH66-+|=dwA05PEOAb*=ay7&DqqcGNIs|tA zod__Zp1QolR|Ag)@$LTwg){8TfI4W8)S#V$PVgg{cwEZ-M8-(NyTcumw~2RXechRp zZ60A$A>7W&-bI=uJcn=Xtzs^_wj57{k4qZGo!s#b6KgimTu73h+wDRZjl@#&df6gU zSZ<~Q81ksPZI4|WXf}25gX1K(Urj`80ETS;w04o>?$amSLZHdi0**HpHn=}l+8nI7 z?WD`|Uq4Jmgp=$=?8YznlCJ8bg4(4~UxoqDKv_=MJE_m?rhYLzdA?)p|I-4rMf?cX zmX@SfH1vY2*4RwsrVsX}#9SafXWPx5D;~9{>%1atem=$UJ;m@%q;$5hgI(AR_K-uT zW;u`{{2-1=DV|6>?=2sNlKSGJ?6z535V?mL3pj>d!2^8G5$q5*54vPuEO!A&-7 z49w+E*S||)Kl<%O^0G=BrE=*@I^RT{E8{~ZgK%uQz59q39tL4H<{QG zLGUQi&<9Ov4j)(hhrV0-otnrR!D2OVRxusq-hDO)pCT5O6c?3V6M40`41MA+rU(+%b9P(5Ww;MzZ;vX1wv)3bk2u zJTvbwIu3NUU`naI&R3d|G^;Ef0i2d9AYgv|Izl1>!}dajc@KpXc*|?golAY$LSZH1 zdpQDhEo)CHjnjJ!XB?mjz&Zv19fR~QH<2v7(#rjCd$8R_U;Fx;UMYT~p+D%P0c)4O zyTs8lHei_9+8a^Rwi~t$a9dm5`OrNdk^z}ED`=NvfDNsbYaL%sp4!igCj-gC^q1no zanax&2&DC(>G$&=#TS!lu4%&u(vCs#EZvw|+O1h80Y_O#=aHfx{sv`FIB`#7+9$4l zu-~}Xp-?ldj2S$aZl-miq-X8uO%n47Ur5mE@U38zeiX8(KjkO&!IiSX1HYHikWkal z=`nrbm}q?SP3+6O{}%u?ojumYY|%@GS#W^;Zq?a)co~WQBmC3H{@%ewEKVcscCMh( zf&72}P>cP4qv+@P<~~DP11B`etI9!8IwQqb$YM%~X!-k}I2y^XwwfL@Xsz|9UlcD9>do250rp1)($U(tET^Vy2(l#wgYE zn|?A?kHP?#P02`YVJ#Entkel7Icu?O5Xrd}e{}&)>N`mkOu_zlZrnX%ThZxP&ID3X z`$`ejh+)f|?e$`fv30*Aiw(bt2l|?xZb}e(eh9Vea_+XK@;;^D4#SVHgBe8?Oxg&w zk2$_#PhFzVBBSl3w)vrGFmRPMpD=@Ii1#%novkQ`2&w4Z6WC_q<=RCtvVX_O^^scE zkGR4XqKrNw>`qmh&ZRmXAIpso9|hm0GJdpyF3D@#lm1iUH~MaJ4EMcuI!ulYORO3} zW^w4j#V8Gu+(KjKeL^0YEWwX67{UI6)W@k6&TI|H zh0(mP&6<_A>^pDASlMbAPGHCPkGT5(+AE`q|EttCm%PPAk@x7dgE)hEz9bk9)rGqf zI|c!NALys~Q2rz;R;KO1bEKSW;yv`s7f-Ilxtg{T?3_5cEyCJZl%P9VOTa?xIYJ3r zx=4P<$x1$^+!6t zZn~HIWllKh6?tv&`Qz#_rVo5crv;RCdb$I7Al#fN_y%yv*1gL9oBs`+`W9LYGdLm6 z^o1vV$^b=~mKZHKw`OQ zg&wcBx-R?ge$SYif(;lkZ2wv|x?q3ak)qLhZH}^lRYS3GPvxF8wJQ9f{gOH3$K9q5 zO6FQPO)#n`!Jm=eo2bvo9fN3P5rFevQrs+MZF^{35T@BjYQMbO2UOlYDfuO{T`$uX zBK6+5hoV~`{U~**wo$U)k^BQ+Rb6Bzd8~A$PJ<^l7Zn`=gA;7n5p_BcnIe)<@;qzkQ*`U|@a8?|XvOLTG@6hu2Q+IiRi$jiy(cN3e5 zZhp36g!I6uyn45(SBc)9OxGcShM}aQofozAe)un2d0WxMTPIk^ORC;W^8hy0WV|}G z@HAj)2qc%0kkWiEbx%Aw+NESnT;i2546n!eA|8ECU@xcb3GI+!(4}^Gt%MH-w;w{U zNyIGEH25W;```=UOg>I#YgVSNyi2egX!ctJI&1P zD`kqnp$F2uHQjELFON^*_nH##95{qGNN@6)HwGkrdxVixU{ zeOnpt%UOK}Rlp08)QvMJ_Zz_TGES*(t~DUl9T`kWpi}t@Vw0rwR5xILMS5uV5X1p2 zTk$~45|6MC$_aUbRzp`YRJGSJ%mRW**Q?dZp1o(Lr>Cb!qXe3cAR45oncf(T1v{H?hmmCzNgyfhOY*b@UOkL8xh{K4H@hjC=@)m=5P)xs;4!d24d+gfI-G1_rjbnSAr^VC}&OW-7aQt8H}yfa)De@6FI;0^e01bY2NLa zoj|ZnVa)mIQp6PWxUQA26(rP8?!ks1?q=-~NiHOxUv*^|aXK z1Fv6Bs*#1zsVru{d}tEru?*!hR^YUL^UBY$+kLn@FET)f#L(AYnNQcyJhvyBj}_tW(r|6MZ|_l zCY@Ob!pb}4QXG1O1W`UCwtG{wyIlXd5~bnqe@k69CbgF&Dc13LEBrzk*R`8rv4Z z-KB)cN|dkdyB2H4D6L`~T?@YYa2Hzai3*L$u5I>D#yazyM+aK^SBZuiOOaMwQTfiI zxXi?Ttu!3xMm(Y&@&8RX5J0lfFZd!&3#r3+&45=bbGJr0#=Yv{mTsKBvwfP*$rEZH zaQvd@U)ooECKe6ryh*g`rel&SKwo#zHA}?b{UPK-(eC@B%9mzeI6sBH9Fa5E=mpUV z*)mkI7kg$oh=>eB$$a1wYNe?2$gR{$MM8b^ptY~OkejDL|K%1<*5x=|#Uv72k2~sN zZ?IkZ(`s7VdL?&M$M!2uopANdhxuIaC27TJ6dB%yNITXWp9HL1hQF%WLnA1sP=TgF zD}ngF-Ku>v;L_7`WX$2T#>kt|j1RkS@`=E#IO)byZlsdp^|%@7@{!4urt3EsXT?&; z@SS2(I=gBPX>Q~){_0q23JP3(2k3dF#VCj^QhjvNjP;b_H_+pC)3<-BuVu_=HBL`c z>TLBT!a&Pmph`6NVxgjcS$^^3|89VH8;w-5r+VjBLeN`ZAJXYLfBYbCv$)a2X1zkx zT;f&_N$N`QegIp$=ts60bJT;1ARs&4bdF9A4gFmA?)UY93l^#%#x$And}QHyQFW zw3ex$T{G^rKif7(L>(A#fFREK5>`>qs{VZN$50Z;ho+#CoDuh1<7wF+MaQTl2y#7XS%m!oj>I`dVE? zgu5It_@o?)7Ymjd$SH_DNSH;7TWbqrk^@MEUF%LB2-=pYI8 z>brI1=I7SY=a)BJfnZuKQ=={|reInwP}i+APk#jZdWELU?gi6jQBJP&y@f={Qb@EF zffLAe?uo7&Ph_hA*mfTT^p(?y`j^e^-uy4CUqT8auiy??6KX3=xIY)}O1ca+)?IZA z=_8Vv@~^(vx!fX2QKYlPvo#&~xE*5+6wPk*;Qye|l!>?2Zx3-M`&kF8da1Nu<~a+K zC&`$XU@EhHQ-+(vw#({H3nhymUDQeFqwt=V=h3p7&3x&4Uhb{ghFZQfV2j+;(cf0- zYM!(r7JXv=VvpR1E%f0eU+%;m|);N%#OaeN?zRN@PKwy+8@F=gzWqU8%UM6_ep z&pSMVsV(K1akvV3#qgD0JNRu%&VTTC*Kz*; z7}HnG9T3d@cVZrxYe{_&%yHn|=L0$X-`$+=d|kG0RE@8H$@?~bEAtWbokeHZqrVzb z09Q?PZ@=Y)7H*orsj@dMcPaArip)Be*FiGV%@`$ONq=!SLhb_#_#Tz8DxHE|KI7=F z1Z57+)t}Nvp}rOTMvyXa%>gP6)4UZz0H)TadcUW>{d~nI2mIl__;J-Ij@nDpkh8E% zFHZ!IE>D2wtty}lqZ&kFUN|Un`uRwt<*LW)BgI2S`@3(KMZwn(=iu1lTN6>&?M5Gp zG3Re((g6m5JAMJVqKC%IdVz$^8{yJh$*u^Oa`wLk8LFxRO}$ zXc(a6omQJd_K^ARES=^X$k-mklVWfYpW>l=5|fLIAY0T}rgTE~rXmn>a8RG2H#GXZ z%-Edyz%nuTIMGyPcw&XNas(Pl11 zn*u|MJmbGsaHqnJ9KPC!_j-7-oBBQDbBVb{FWM##j%)-HL&0SRLP&ZZDha)5Ktx=Q;_v$1UQrEs^QEf*R zsc}LAZK|=wGu2uBAoZ|}T7?{i%0-^Kf5=_X;Qs^jcdD{ZMSFuqtWtbA^Sop%27m-A zG|{>C7vrr*&3l%Ar!{3%akT7csfi($0Ivi|+4G zj=B&XJshiykS;!7Aq*`>mtboGuTNZSQ<-8lU&zZ=WWL+yPc|a7I6I>@cge{!+2dRX z9Y`+{>SE@9JXU9av#w8NZq9tu85{7YlO6Pf?U_n7pu8IOa-lvIc ziwyde*W^|Yla-aqiP~ulCb8qsHBxn5Y|$SHhHXL({^e^Dt^YE_@3RQs1@Er`bXecn z3b*qkLbcyRl+}$^2s?TmJ|m_G4s6;ze)4tOtWUKP)E#9=ZQM9ropZ*kxqy2^84ih^ z=9F9pQTo2#46{D_xLIf2@F|}m4HmHKlhAEIsIwrWZmOa$(wt97qA?1xagwNMjbOQ@ zT2CV>iSyPH>dcb>jzT-(#w_QLoD2jT!5tA)P^n{&A(%BM-6N6yGQ&`xy&}%WQ=D4$ zW*X6+Ds9GoOcH9GNrP&>Yb0#T>H-E@8{p@y*i{O7q*+=N{p)VRn1P;$JjCdIK5y=j3pXg(i# zwqs*yynEbzn|#_^IsNv!#Y=eeMTtSFSjKu_5@V}}R8y2@9(op~Hn_E{869g#1Gym# z$jq5MDq7-yNzVqj#HSH0j-C5#l>;9Sh!LdxXHtf7?(}?jMcyj(opv%?YYbN2-#gXJ zWn?$)Mg=e5Zl<7yd#?6Z8sCc%Be<9_sC^8Nh)K?klq@UYbmBBarsBKa56mRec^|}c43K5;eH<82Zgc12fk5_8fcER z0;_V6GvpPW?*lu>Y=eK~Tpk#HIR}$viQ>j_eP-!8@1mV*dG2L-U3*$C`(1ZkZxrU` zx33!_dfK;RnXcn6nUGt;Tl^J?q__Cu{qYTg!|{$jLNVMg#xS=cap0>m0ENLwg{PK= zc24R>w&j@u z%ssuzcO&-C&hq&mLm5jf--58AwVXqNE;tJ=?16lEBAF@>_jpFtez$U^LUDZDt=B&u z(;WZ2oH}&XqRFEQ(jN6=fmS5^vFRd@%LY>=PUCZ;2uD47=$72by^rI#dMS4?|6B3jU>|*}9g3h7hB5$u;N)$~y`BbA5Q#T(ta+j91|F zND96@NxgVtuhz(`hw4Jg_L@d3yYoBGnqSQJ zG#dq{G*4%zx6h+^8#*T}M>H&DoF`qfZ15R9SXpN;hCE(+)3aRGSf6||ePdiUA--

`Z zQ5}9|!G*}sV=2$KGcx#b&RZS5;1lpHUP<_wWS}QkK^w0Ehmy)?+t^7r&Tt#g@j(Q@ zzGS~x<5Kt5JE|fAC(=BwvyPiuD(|2A0xj|58j4STPtN(btiXHvmX6#OHUMo;9};7X zug!tvT<0Z@@9&5r!%u>nW&N!$n|xuwrfXN*h8GJ2zrUmBhqpx2f<#lqX7Apx?jkdV zs6X8mTnmj$3oMU>zHAfr`?;Z&YdZvv*n3B4-ijr8s&0ftd5L`-T`TLflopy@9~hPJ zN~8|+%d7$w#pQ(#&*XHG=Qmx{^0Erj9mfl@zZOx*EbH1w?-VNW#;t11SK_Jt?%_!Z z0te>bgOyBP1iAeD_IG*VjnnlgE&?~(0v;FSQPUo(LD~PI#Vpx*{@WUT&xDs_9i}+# zqaZ9&i|#zXQJSfkOzOv0TPv!I#wd$SZ=C3(h;yWcuK^ z@o~6-WTg619?rKp&#SMmNhVF~F*3BeaAmfWwp*xQYk8zlB4ZZfj}Vld27TE`MLD&r z`93=9$KkQx{Bbz&nGyHsDTgH;dws9Xl5BJpPsvUuXb{NdPFEXd^MKVaKHS6jU2JRHMqX^?)oRYE`73GtQ7tv z&C6`K2frcW^LwXXv>ZYypDxKSZcfK8_!ppfXHk{LUb3VApx$f$x@deFsr=MWq3dU>$c4n|p`n=M2aRvR~{-zm{ch0ZWQ_%KQ$w1~L0#K$5PfByY5w5n| zaTGtmVvt_N8+dAE{X%#_ogB=Gdjhpj`}kzG zouy?=2Y_fz&x)F>h>L68V48 z=kYs_pdIe6o3o>4h%P~YF^XtSwlst2 zK|T?MZrgZ;^px~)pvE3N`0`=y+Iql&U*i_k_4jdif?Rgt^#>OB@ibi4h#=%`W=YU@ zQ6<)7xq&`;e_p{CI)7@fAv&6_O55(!CWU>xm0env++bEfvr8>Y1V_kh2n4Ub~= zuV)TrQCzj_u8tHUq)~B_R(r{t0!k(1vvlD!jmg8a2q-ToqT%xde{r8-{#OPaR9JT^ zAu|@iK}u?&RN@;Ol%f)$lyzgQKX8F4<>RAYyW4yLwb$S33!3}}%+*u5DX9+l{0Ef3 z`kpR~+q9#RXOK&W*-wO1&5Q??j6_=^BHof&;cEqLpgZT|>R zqZ$aJ?n{GD!k-qvdoFj_Z2zS*8=F#SC3*#YKkd&EgJ%;OXg|o*vQ`G|k}H#DtKdLo zEix|mW**jno2PE3B8%H{|I}Y@Psb98sUYoNuPE{2VRk6wf0jF~UN3j_oaS6B27ByA zu5tR460~7gmbEt^cd2!+xu2b&IjId}15+2wjM-cN>#{|L$d5Xvt;^R^~!#A6*@(iTz2j`LB}w6IH{!tGm&Gaj(&JN!U} zs~!%vKS+TcU<^wPGZ9;xHTvFbJxPxKs8vwSA=-Av`9-#t|kC<#fbLxPC4Mq0dK^m^N9B zjcdJg8jq?rac-YwJ{RfU>O^n8ajEumudaL91VY#>cjG8umxpszNSFAwu%=<1a0Y%@ z4>7-+w%h0vGiDFxWTQ@3c@H|R={s=202}J-Ak&DSTu&dKcQ>mug@WBZ z3<>3ZD3tm9JxhYy$0+@-{cH}q>Z+i8h)Ey%`SQf_c=iAt%v)z9w^KtT`gimCEU`TN z0wKV(UrW5tzlL0C^kU$0GkBC{=QK+QO(IkVVkz%sj$ z(q|?iJyeqQ2WRp*=IiQ0EQ$EL&iWv)u33eSIrOLxp133`8)3_Jlk|}b<(C-Tq1=;m zFx@UEGRawEBFV50MyB-hYPpTKF+~%2NMwisu%@%XOZ^PO;O^NnHL<5>;^}MNyM2?! zCF#4UtasngT^3?SV1zR#9fbkx0}*f;0eZp7rp58-D^&W*7h!5`ma%hGW%1zQQP`72 zKHAAlQpcAOTtb~D45dl*-*stxI<}kqRKT%OJwLzHB+%Kh)2ZH$<(z8e!g zK8K1YQGT6Mb)?VeJ#ba%(V?t2i5YHk*mq>xpZNOwgaSt)^Fg$x21Ump%2erFCw3PY zX6Z1Go>4}x3>WXB&fjn5@%63bPd$ixx*`3B@4~2$EWTZEMSSH_DpgdCx+oeQiTe(e z-hfkzfXDI?ZC`ngWM3*hFxm zZL=1xML8maNTn6EL#6|4YK8Kk-RnHCxHH=DQ?Eja|bTHgK-&Sz?3<%wBwk1t=^4r!a%u&z?1YvJatk3b0S;Q6OG0w zW`*uOhcn{+(lT;A&Yf6M1^E&RzAu_JVOMcPm6*jj9f0Odv8su?9`nyNWi&4$7TR6r z4|b=w|7l!Te1%h&{~CJgcYF{l_t;vp_n<_Nub&hbnR&(brat}Y9ZrKq)xybD>+2MZ z5IO?-ydpTU70*xIM#Ll$+BUi+g)T3VB7V zhrQ_X{%tHYd$<=vjQE0<(b&kX;?=4BSFLJFrTkBqzss0AgyG)X9#O*wL~;Sc5`Per zGUZq!E+R2bLmn^e7q&L|IJ0^gyX#mQawqE*eq-vC;A=BV_U0&w51q<}Pcv%WCMer< zZEYggPkQ)9UrFV&!GX`5Jx&Iut4eNjL1}j3+AA$>*K@AKY;(7dO3hjb}&c2)loQ2v@hoB|k z!=>ShCD|%xxrSfjd0j@X1J)!LR?t6`4_$<-w&99gjNKP0o(EDzEDHD*OzY3oinU=a zCQM-0Z_4Gl&gbF)BBZ@OGvf;*pzpn`WRz$=b;YeIV7 z^m_ES1RD$UdF_u-fuT0Z2_ZRAXR`A@2s6g5xS)lDH<@!XE25$G_7_81F_hk0?gD8j z3OVsB#_HO4wL*(sjMu%kTsuQNLd1|FJm~1Bp7NiIVkyEl@x#UtR&{-IhmR#bo`)Wt zW;K1FiTP`x+Qw`bc=y%~bQr1_{fDB*5+jYGC!0nA@NTi$}YO z8@NI6pC~mB@nI`?^^b4cho`G`ApEQNN1o=F+!gb?ZPzdL$;Vt_RC8jI+Fr>m_oj-4 zxCzQkE}S=G6s9U*n|Ke+i8O$u_$ByaOUKs@4L-Lr&|l@5(88bcYC8lY78&{-e-@;a zUu#ze-aw(XG!ZtE-L=i<3=GoKf+3Bqp&;r2g4Ewa0RFJQ??HlqeR?(2fkg3a6?I`{ z^&dUAUtKtDk*U{gVjIH$PD#8B^Xgwxx@?*x`~<@HR=mlilItf?rpevJ+&`i#zq}^*pAq zCTl|Ysa*_BW^R}(rB>(-tRljo$JVQD$oG=Um)IlKUSp2iKW^2mJQWHB69wY}Zg*@= zt?yK=lu;}&$T>No%Nyt=Mff|25Nm|}0DSrdpqNADCgyR3ip7YzkaPU3`|6Jg*8k+| zP2!J)|1{I;cOCG4FPwvx=CO6KdIOnbEQ=#bGSa6a4FKtlmHWCZs7#1QsNh|~i%$p$ z<6*qhP+KV4v^AWJ^Lax>7Q^>_%-eFQECVDeYq*wh4F`_N%`PVJRc9PHWj^A#6gNg4 ztuC)6x?s1Pv^=&(9eLbv0I2A*_WMsWKNOQ^OK0k6jCb(^O@Lhogk72a>@lX#(6E!M z^@O0;5HJtXDe{783^sV+Q)sp?a?16m3yM?(7EvnPG4#**s~uQ{=WOxHW%B*q`X=;h zs%E_#2MRPE8_3O>&;nlG01UmnT$+R(3ejEjr6sT-i*?n$aJuk7z_P4)eMpi8Rz{!x zGM@h#k<7j*-0gCq6#c8UB(T*n5%$IbgV{uOML2pwUifTkVBV9#4Z@JwNhk;Ks@Uq`& zc)!9|i0}BTQ{B-e{P4*ZzEF3@!1~?9Pd1PjKZ@5^-Jlt`%WNW>GN2EyVHFTD8khBe zQ;CPFpCavx2IP}fC-3Lt>lAG4>F=T5D!1Hn`~YyppY)AdUKMH^%i;P8yBzme8-3*8 zBP(euqaUrf0oHGWAotZ-bnd5VP&6(~9Z#oxp37}axKb~@Oi5?!ww@pu-ei>-zI@$-d=}@uz7Kq~Ku}BRlof(*l%ldx zncWf*afCuiXNN0}T|?W%0u)c_;;`;QQZ;uK4L2>z&ci9e7*XAOYZF6X!Tii8^k7E) zYk^;m@hQKkLv69D2tLiGX#NuQp@QNO>gj17iSg&TlFAyJek9G3Tr?^j^6JOMYUzYC z*<+cn``K|4>d`)sX$)nZ$lCqb8e}!N#1{CAQ|WE)k<$%)w29;uw&-NSSSZB+zye@j z1XB_lIL?2IC@Q(uk!^y)stGvXnoVL%{A}oa{LyME;dzXOMIVfR`VF}fUkd$P((6~& zRsd~A!9&Ma9|n`@p0@(9KA7uF5`U=I$m+tto9Xw9b{_aUxy(rrd7y zhMVSo)IMP|U^$q5=DbBiS(71CUoq!_V~AI1^Bc?ZmLl1l^WK_}8K}uqFq4X|)RjQb z4~I&B4rlq3QP%B1|HHr&4;z-Fq|yoh-PPbZmSVxJ$2@{>HCnqjdJ0~AKf8t*WlD0j z0Uj(_Jy@u~@po^0-h_bC>{&NkEJ^AWo$%?!1XZZ`(^>?^+QQ^c@3v6-EvG@tkNhT7 zzuxuz{@!NXZ>r?VEeU+^bur`%_PMU1((4lpAai~YlJO(+Pe_pFC|8kv1;>`%AIXUY zCDBuwG49@z@zP9t4aU5~fyK9wOJuL-BV_LSu5zZpD-+G99bv> zQ7apEIk>+5vCXabI;RCNhbpk|KXhdHaJSAlBN741Vo{4-uWzEq9#=KHa8xMGPk&QM zy%Y9wGV@b_ObXp`GX6imfunHcvm?w!_UIhadJ;?hW40Axc&RYmhzh2)mv%L7oDY2b z>-ank3l;)z*685he4apnCJ$~tU_1(Vyj7Q-Cn}zB%)NNe92MTLm&inKaE)Ma=6vJh z3erBc^ATsPuwzVX!C4-c?&MlO)7aW79Hb1=X}eo?#|Vq70{36NXR~Gd(k`P9h#SUc zKBUWd6auEMyP&`cQLawmsBUSEj%_jhN!jLl8cV7zoq_xs2YO zBff=E3@_Yo{LIzII*pBqOKzJr$*+xi(s-CLxXX!z1Y>K(70YQ;A#vx2W}`QxanILW z1BFsm?oyh}hQ1W#7aw7+jsd#1)CQCBGkE?Fy@4#n64A2_)-5o_#;&UJWkG@T0+oz@ zMk2;?k+(M)y|9h9fo`RE&}j4JQTPHVsLgY`2rDuQk14@?s>$L(!_}cSI(<)I;(t%~ z&GDQWxhGGjFfEvP0lLiP0`vTY^C#!8JNNzbjkAr_hsiEZqn|z(BX6M zAWaIAsqws0i&V78dT2he zy%mC!>cW&}2v?Rpr<2i^<|>DG_rv$dxrFu&h_SYZPY$1V6Ff@cT(=>IA@U5R?qH@@ z&)EAwR%smQ9uvCx)h*aO;4vDdhn<;6N2gGtybi11^QfHGCV-6uz z-f?Q|qpRY4su@kF7d$)T?dJ}%uR84vd1$-rN|B#?Z1wA!`K;Xby~PKwX{uvVB?(r| zMhA8e#88AIvZB~T5PLwn)KGHm2|=obZEwAM2eLi%b&$(Ai44-3NUBsQ3<- zb7&M}@rdq(QlbR`?9 z)T~+IR!*QAi!mtYAI#dE$nDpwtq!$vaUgsYE=tu5fii#Q5 zvqXtY2d%Xk?O7IIg){x|WZNgSQhX-RrHL^r5qVkS(TQ$$&oKnG8#i+#1~9)(0`v9T zAw3aXd>uj(@pUq>kGFYp3*l=dgKvZ;Nm&ocU@ds%0Q)=vV5Sb)!Y*v3b&{_@Oj4?{ ziflEl9LhW4t%ASB%^gpf9`l(`x=N=CfosgjTW;};xhK}9Dolw3Srn$a>uP{}8mU3) z5Lp%=2KV-|ORul8JJv3GIZo|7B%|`1o=G7479TI0!OV9RgtK;?FT zaGT|0WcU@GKDZIM_3gJ&e@eh^>N5)HQ-C=EJSUpjERw&^eGdL|{{)Y8gh`aVhAC6c zip+YG#ch2>s?Sde)Md5z$>DobyXc-LZw^5C*O|{IpKqON@)f)+h!md~1s;T(X(c?6 z@=S!J)qInrcHuuDW>N|*^t@^3?3nSIbbpF#d>plF$WL~PgupF&4dqlN)}7V(BwsNm7aRV)W0miIk&y0p4&ZVpVT|NaLBNbHk_n2zvWAtU)e_tg_9IzPOT0JaL3%dWAO` z5~IZu5*rrZwz?(%y3?R+Hz$@Q7y;ObtiiZ&ogHpBo_}QG8D$zoaC(3Lyek&G431zU zZRcg?=tq`4>!xw3!fgQ_7zo~qi^85$Dp%a^O~vXd$evzlj4lU@~kR`oDYHsdK7hzl()j?16HOV65}3%nD5tPnvLIV65iGr2a7 z$(!R9TVoT1DwOiCW$o~1_F4_|vY$^|pEU~-B(bo%z5M0TS=@C5DL6a58)}d9VEa7G z`_ggJ&*dlW_{BPcu`w(5L?qmEg{jQ=cYE0lRg+h=Q>y>6LPEX`#Og6}wv9wtPly%) zw`MyHllff$e1cJ0BpOwq@g&)sPuy27>z*Gfr?X{O3fXk_^ zyp5u)%P`unSm+y5-M*D<%R(x!hoT#!x8{DkMgkwhoV!tmWaN2VdyU!QFBwHOsTJ9? zd6We2j+Q<3qy4OUX-7vHZt6&>em`0fs?(Y(&?ik9{gr&ZLvi ztjs3e--}S^p6c&a09>R-?It0;ncY`Mx!~bCvwsV$0Cs=Xp1N6K$9y7{`*%h~7hZNWZtsWSe)xAf z6Ts)%?#9Pc1!i9NT75LxAU~NxcCBdv8;{58x5Yb1G-&pZsCamp5T1{XPn zpF41Ahrb3&afJ*Wh5CN?48y?8@aSawCC(-22hZ@~5IU0ofF~(v#UmuBeAH{|&?HGk zh6*6{+4t1OXAoDbnTm@mTic)QPxWfy(AYOMJ(15%l}{U6fy7`Z7lDSXYbtTr_FP%0 zu5F!Cgl>zpp;3&=-zr6NbVns5;Yi2y3*7*EdgS4gkF}HKy3VWs*3hfg* z=pq$v0l%t2I&{@c<2&%I89mC{s#Sn>8vU`J@s~Z&+>k?{ov<^l-a|vMH?zU2Mb#@d z`<79%=9`-{wZpl3T_#Q|Ym-Ljln&aJNo?=BypGG93ePlNdg~4a{G*pLPGie&mm&3IpFD|(a}?>!_Xj57Bdi%l!Q-^Z3w9;^6^8k1J# zya)hU!|WRWq9H9E1`z((uxhx5#93Xt1+L-Nt;U%iF`glnISVDOO;F&8i3GkJ2en*{ zl;W{8P5-P4;lY6DGJd*wvdmzQ-^MT?G!5IcqshvAy_>R)q1;B35?buALo&)7^vHhU7khH{v+nhN2( z$iPJWC(=qR@RUoJxNzvM+^%3%uTQ(B`n|1XMKEh^{J8bp`>T&Ty=NC z{ATK%Xn6#Gts!MxWp9u5;Dx)Su(;z6vg0MA)evD^c_)SI2sl3mxbahbVT4F#r^C2` zeh2DwcSBq#CIvJSE-f&qP2m_BsBtFKae|>0EeNF~7SwVLH~fxsCf(g>DDkKsL-r3# zGbdf%rz-MB42)ViCHWjZ>~K%zwf6%b>_phw*iQEE9RC3`5|Fra)Y(E;69ue4V3PITE$W2{GF3UA&e=0crzZQV6e^D= z$-xcHXRM6_`iY;2$KxXJ=eNUbgb&B5jHJD6_!W=w7usMI_ONm%bz@qyQ-X<~lNRNe ziau68U=w4y@`_GAwa<%}Syi`3LR|H%2EifayS z!^4Mq_iBwG4@{^VY`=;5u%}NeNySz55Pm<=9SIcAF#nxLshK;i2IP&|V3G6otH+CE zTIbu3T1lR(^sJQ~>`T3G)a)7qzO4v}xakL8Ar!i%&)kKQ>pk`o1*Lvx24(ILS*sfF zvfAUS*^N3pbA6e{ohWe~1Nd$C*sP?hE)`xAf-VsRo`ow<2K?N;X9X$OUQ*qz^?I%W z<<9PtuUBM5xfOJeSu&}zg=+%-?t3X7M$T+ip4VhmJvfqQy422J7ZCR!GG2H0)a&#R z1?g9KOS9U#XwWuCHA2$)AW)yxAG2E-6BN}-q%+(YxWku>bB z@k&o&Z8QCi5ojhr1uEp_IDJ!LFEDVSaS?2=YBRiOsZ15mxOuby%+fN3Bc9;KfeoKv z=kwoq9!dKVtA5oT)&Ksv(st&;rLpe&7+bHcONH!uye_oRuvZ+bV#i`{208}G&H>*K}lq%00F;`SDW*EY4RDOzkNi#kgOY!+a)vv?aWpy z1;yGi+NSq53JSFtvHd3%KMla<*D@RUGrv+>OWWdGESiIEYD31zmppvTbe>Ix-qx z&*ISX@$jfRip|~cFo0S-100#^?w#%)nf-tsvknZ@B(tDqOEv+mXau4caL;Ow<& zeZ1y<9E72=(Jy1H{Am%s2q*j8~ehQ zST|k9`93&D_FHs0bESKNa+^xh%K-q@4eaIQ)%U$1DuEiE}C0^dkUKh}$5emHFdU`%Lf_d`E+rh`k-w$d4A&-or+$km)|6(>0 zskL8Q@BVOya|inSs{ehhp>DpJPA^4oE1$1*{LI?Y>&9ilF8=v9=57+t3(0COkztu8 zan^za#HeSOCfS&*MO@gUKg|;`lrK0#c0`W;UP3D`T)*rZioS4aHzHP!g=nJi(ezoe z^2#p!==iNcg%l;Tm*2eJU zUTJm7PS80hQCM{Z%!Mgpw(8ki=V&P7*wcY2HU>>=tQXZdpPn5$;emg5${JlnU_B*8 zizADEFCD>_s|nb=@)1xUStcyGASAQ$JeW)C_K%oZ`jD{J>iQof|2+~8*7CKCSUH=a zu{!o6o~??A6uT$PnU5@P66c%4nU@`L-fZ%x{dlL09F^T%$^_cX8j_Ud?>Mm(p(d&M zS)v(qP*J5azfQc&)Vr!hs5_kh{vwbO6^(PcefR}*HY$59zCjj7&vT6yIf4w!v>d&u4#7SDemUmuiJq&L69qmWW>*K$qf@eue!4_} zx_;7>odHNwv=8&VkUdWGCp4cnz}d+JZ=~W^F4}Tn+C<89+qENfi1%X?_CC?SgmJY- z0)N;69-jM?a*BAYXo1y$kwhlbxzO~o7z=l{)Nh*mKOe|WMz!3czNCXl!yX&;VgQe ziBc(4S#hcQ-Sjlf!5{pvzlKWq-1OHf5n{YD{?bfWPPQ|-Vy}GB5%-)4GjB+&ha0c> z^IFgl@8WHDvs_zP&e5Q-&J_&gDRSsu0W+7Dw;?RNvr$FX0hIibEse zf3GyunW&RK=;-)LC)LF1kAO%O1gYRKg~|&as>jz=kTbwWDb`mMr?Lc@C#k9i6K ziPlJ;^h_P)jhQiuIT`we5847r-!l+809j@AG{-o=q$QLwU%BcB?2&WVR8eO8xD7IB z(qFaDBc8^-Xhbm@Ya=KN-|#n{fKQuL@DYFbNOSYRS4*2Ub7duK@m{QUA1U*PC5b*4 zo(9)-yvH-?K8}Bx1)T6-hM!k+<-I^g_Oia}PX{MmaN>;9#TGVD+%415c5r!BQ&M$` zTTb(s3V#?Q#1EDL96BL`lA>1-$9-hDKenrjgQ6?46wkJd*$SI6rV*; z3hgV|MlPVy{uxTvDu^)H)5(l!KU=9tm&C4k=gXZYF{cdVtzp9hNnUv=LZ)Kv8L z89_xwz=EPwOK5^1y@R3%p-S&X>0LS@5D-)Z1SFw%kRrYJ5|n0WL8|oLOX#76><#{B z_T3LVvor6^JG1wbUUJVlPd&e<+;b6Dt#0AQfc^pWGt0w8h`Q+T`lK<+1dq>9Z~wxI zeV`Bw^r>mXasmdNOF5 zG2-|oq1`$Oh4g%PBL*6i7?b&MW49_o12|QY@>`7%*Y&!Z$PS5K+C6%*NwyGY+680U z!0OhTD+?sz`;D&(uQu(z5v|*)B`@#!c`26jaN>kS>cwsUkV4s(^Vy`Yten@7$H$G} zr%xHZOGCT$tL^Tr4EQWd-&SFrSN?AI5%Sm$7FoW}aclRo)L-~!2aU$n=tyz$tG_t3 z&5VM09)7!(B){0m?xgPA_>(%c&=AklvB)oT*nH z7Rvmqa&&iZcT~Q#!yo7(>>leC_!ZM-O24N}VE%gm9tDnjf_?N&~hO0tz{}O%Gq9ft>g1Ft7q1S(xse9w|H%Wu<`OUB7%R& zCHrrE1;CNWbWI6>F&}lCkOgslgGM*Ct{+(hbncHe&@)^|w$5GLPBOuC=W4z*U~0-V zwc-=6QW~FkU++DA2c?P7PQ7}_vZ(%8z%q$>4E4qMePsz^c_VrcGs*g z{91Wc(EX^-f{Kg7)b8V>2et<$E(+p}(W$)BIyn)U5<2QKDB=D4+}q>vw+ufoO-3|ZpyoUbA6FjdY`iW`_7y}*P9Fm3 zvi?Z4r6govo9v@uv>KrcaR=C%TwEU)Y#aaCI*Zn33Mq1l6zW+b)G9ul_~;w}q$xaS z_AVgbl|k%86%8a?s8Hhuf1ekLwKl8xguTOLv+Z41UrxUG!2W4-3obQ)>sk~aP+ghV z@xm_b6HP`U*Q_GdQ8o)^_tDr}x)lQ|3=+0T=&#y0%Z={xmFA4S0w#j!G{Kr(^Q}PuU0dXgMvB9Y1P!Lr>!LIj{u~p<-WY|Rg9?+k$K7F zo4C>ECFWGVNg?A9PKqE_ZNz{UQiY+la?%1<*@_%EoW)7{8y{YLfit8QU5h*pzvUi! zE7G>#SG4V_0#L@dGU@e4>Im5A=3lO&`QGH1Nh{ir&69nFsUAtTmv_CDF z>VIva@vUpD$*gi|#z*J5!>E2XsY{r*N@Rk#H;OX||T=UI&T-wYyk4%gJq-3g}*tTMdhh@0?Xfd5&h zK&-1kGR=&ue-Y=6=C z%+lDT3L(s*15@(0ceD$Q_PsgAb}R7WdtZOevu9LA3|v-)Qwm3>%v=h~XpCNw9iU?q z`m8M;=p1Zk<0sDjef^&5iV&NM>4v=-GTZMvc^%KgNeu9D!`b?^20-ejPS8%9tV}TE7!V z#5H0X?HxotOd|23aw(8Z(=%vNXB@bF;A86~2_KgzrX4J~hUx|X7_hA8TF;ND)c;|2 zCvz~*Z-n`?gTT}joO1jbm9Gag@GTo{&a{BeMn@Xwbt z9=;~Cum2G~Oa}W#QtYMGHC7<0^@pqCVrvzu(2w~dawb~Ee{a`=*krec?|W*V?ub#tET^)X$ycoa zlpfG$kBPB*IiD3#pyaiYMoe8|aUl36=)MiFvGC{K6jypG-jx^L>M|L41C`l-vC9>^T=QG$qkw?iT}wJT zBGMRl{@2U(C6l>|U29FGpyKsR;v)~2Jd1`qBE!-PSGiV_`j7d!5eXNw^DvXEv0B$ z87Yu?F&fmVtY#$WMD)cM#X=9KA)Qgeksmrbr|sw5XGuVfsPvDZo_Dq(H^^XJc! zxsk`Xo|>o6_1X7@^wtP-tu)4o_#0%zbsLc7e(JadRAp~;7m`^f-+dVVK>-;RG6+Pg z(D@ul>AHLV?w=t`$W8KGRiIx>suI3T(BztutLK6hDK8;DJi*t2H1o|mpi6i zacOC36}xxOz49WL>k`iqShvA($hORaX-I>|W?uA_J5r5uN!fH(b#^YtfJIcr645)epM_>;>pDXLY7O<*iDiLH6}f!;=W!}CI6RG^3sG-P#G0C^q)t!db9|+SZ7C=y zdnPRU+Y?|0kEpCe&XzWN z{)~dHvV`)k3~ZQ%B6Z4A8m6ibhW&`bXrU+tXqr-C7dUyrft^c)-D=W1Vd9pUKNB)} z@9B9s{N|HW3=)pY6vw;4s@ z@51s8&7Pll;pLy7VqCqKeS$IzdUrYc+TgRJB0zc}OQiBM@b!+(YJuQ-)hwSg!P%ft zFp~AzhB0vVBi3Lc_gp+ehTz4E>&ngoVw}B*aJv*=l`Z75r8q3yBXr92g0xe;+W{a34s~ zh%_EH__is}3)dyrH(WQo$4U%E!W6tq>t7A6wztNYL%(pI2*9csE>9D~ZFJJWguMSR zru2}tKgayKws)-aK4k@m^%ztz<+jeyw4w1vilQLJiPVXZR;%Hd>(#y%%jhT8k`SKz zEs7J$9AA3xaWc-^D~dvfypdh>CQ}FBw_2oSfzE(`DeE56DV!9=bRvwRI^$NgQBk;u zt1BqqckhqviIL~br)@g=bfReA2Xjl-(UwNl&+N~g@>Y&~3O;$iA+Y=&IK+(v<2aZ9 z>t=*dT?xwr+oT7!Oj$MC6z+Cq`b;4e+8lL zcl$*Zl2OeZQ}ma!J(ug4NZo{9VD!qTK6Q(2H(xA@{I2v74f{#?t?K0c0d;H7?Y|1X zStL#2S-q1WpyKzAk*%GPvUp!A=|NF`TEXSsLK0Ew1)f9oCTMyOay>fDhA78s5s1GB zqwtbsTBJ0)+q!eRvCsaXx0My!Fu`u1#e3F4Ax4-RJJu4$%a8=e^hcR4(zQ^QGSH z3vY*SmV2x$jX_0vc2)W8)*iGd0{KG0_shSSn%IxLh|yt7f%L8V-RJCXbbczc}9 zm+Um;4egKjqF^J2_*kNV3g^ki$o~Aqs{gS2#UXsn&~p2U`ZD>Q^t&>X%itV`2caC` zeL+b_F%4_i#66)jP!X?25c7N?J<63< zaRE2Uc_owKN&KU|cYSR(iZ(n2muu(7nPY}}NNtOMPwP1`K%9v+t{YX9kgXolOpR8N zz@cLsQtfx3C4(7x)^T3~9>AbEa+87%nI z?9&Pqgh}ws zm9L~m+88V((iY>K$q-=O2`m211bt89q7WH!<37}#bU~uB+zI3O)E^!f^8|**U+@}T zf_Ir8uZ7&VdTOksaUaa7z0HATHjNUT9(>R8&@pEFd&unyqi-ewmxuP=jp049d9xL- zBieoDSL&8$D+m(j@#RMy67WwkszeuA?)r^}4r@b@-dg9u$|cXaP@ymomGR6L&@Ht{ z`<=MndGW)I3-vp+r@ym%VJqLyfL>W^c+*V<*R{o)eyXK8QrROeGuLnf9a{35Iun(4_+KyV0n7B=>yFR z&~g)i?lSx`BZU6-Cu%F8E&ehI(-UQu7>b1a{K(@R7s~0gPKGG}jx7AzzRlfun68mP z{GE{hF6_+yMKNeqr3w3P<1b2qtpprG)5QvQspsvNt z9rX(>N4c|EzPMx}uyO2+$hokydQ;99O-^V&3>^>tH@#39FnvRWJ9g3sZe_LH-h>*j zhwhg8oE(J0rz6F<>+w>K@2!X`@oN<53v0_YZvtt>{NZX^yx)`ZYjt-WX189el&78h zrS0+7?nu$##pw?+rK{y+^TPd)R9NnNt@0n?4$823O!XqiGa}|wgaJZ6 zd*<(fz?xru7Mi3n6~FcG{OTJp$87t5>=71*}5Pk`s}D*Kwg$ z7j6Q2JyAdF{&8wHi}k9}n-s$7l0p*-^uEW(VQr|u7iBwK*m0##`sGe zqwuinpLXuSXXv%=b4o(%{F#ctjz!=CZqv?wlo{@}?4r<>ev*Wlz8iOXHVGz3vU|wy z_At1qp0u5yKhN!03nqhyKgsF+PCrwuTt9DVz~a2STf_eZ2&nFn^Vs%XiGXhR8Y%~KO{)IN#Q!z}5OWv=<6&wC|4KWXt) zeSM>X-921Qcbl@c@n(8*-B!KI&UF9Uq5!3B+ruINNm(`s7U7;7HGA$F>(*;TzE(D+ zU6MB?EkO}2GFrOANxp+@!T8FrTPT2g2 z!yAIY+hG8BP|az_4;vt>M6JTcl|_#PU+j3SA6u9jy@M``NqmeVo@V`OX}ahICV#^J zwW@>*H@fBR=+V@604vKPb8_Y~;Qk-1AJV*%SYz8~Y9TL7&Ua(U;PIw$0=h2GMozWe z{^LpyIaki(OzAE@>KA9c7RfAIrT z_taNE?|uE)oMtnn6re3sc+GIzAO^}WEBH;2L~;Ka1-1YhV#+>NwxP7MxMo%n(1L8> zSDsKfIkhj(@Dtzc2y#e!`7c#FtM#t{dxb|T7%F8{ zLoW3r-7+Du|L!kQcv?uZdV^vW_;~9;BDQ!TE+t}!;gk1cD4go8Q{;o(*{FA*KCKHe zK$~a`vZ5WGHTh8CEKa+sefL%~BAXs|ND(e6q$ujdDys8>)@^3x2FO&u+Vx%n)m+z$ z13M;D97*7Lrte|P(^u!o%8#*3AD*OYe&i}&EH@(K84b4*3fZL{28v#PPh$vj7VSLS zm+k6Ozy=}lU$Rt+*#+|=+Oa0Pre+)X^^@r92l0;F-sbd&+@~vLm)dN6gbhItxLc@M zS^n-XZG56Q2e`myfjsZ=xWc2NMm3RB?BEbS#9^+g{JoE(N;Nib<>N*Ikj#A>zan40 zLCP(y|H+L!m<5-8|EQltPb};3=<$ejq%`uY2}?5?mT4)Q~ELof?T^q@=;BV z8uqDMBQ5gX?10IGBwqVKYToAp&HUdeN&6$gf765^3SuoU=j7ZId#M%mu6!fEAU#Xo zBU5P(iFVU9JaFbRCj2$!?5ykCTwP~OnV521>=F^3!X@v@9n|lxj;LVC6uOj#Kz@@t z6Os`_13IH~Yc%|g5+#yJLF(-KtX_%NK_HV6wJj1*gA;~5da z#%&M?wD{sDIcTp()&cftV;j{u5Qh};*&B7$qnX5_LxNJp;YO_9|AMUIfKy$l9m_z$YFTka4kcpmp1#hYS)dUz3hF0(=0mD zqtf#;*9ZEN#UmC+viqzh!3_YK%nbSE@7wNY>__VCC?#H`@&K2B>%Vm)cL+-ygOjvqO24yL zfiBLbYIiz;6`8*70S4(@4Th7M^f)N(d;rTL?659(0V>Ug>u7WGy{vPS7}#y-5bxLq z%lqw=GVhX_IrtG$2hAxWnm&HcJIvsm+8yPLW7-+i{?JfyxkseuPf>0?7KCd*cwOA207}aXMl>|v3UYwicPSLA2#!xFCBQD46z@K z6A7mqb$u|XnV_DB;{Wau#`#Vfq~mlhvMwdvk%-l;=<5Yy;zdsTdK@pCF=keSSsA51 z5q%e$g3n2xo;C8tz0Zvb%^A|mp;GFGdAnJ0MvabUlsVlb{z!Ap1CZDBtqrf?T8$w) zNM^;8s8WOlPh6?H_>|Xl+_^T2M+X(Chw{s=Ld?$=91sJ6EPO_knfWz;XhD15{Aoa- z%Y#p`6VdMY1@bChRC6y7S>KR&s{np_NF52SA1gtmzeyDM`grf~kjEmHvbO=(R z$s$vdy4_uhC|%AN-FZm{T7Gzm8_?p!(%+#dHy_c^wSFBJO`E*j9HFpmt3<_YO_qf> zdUIp_x;}34mhSWG=+2$RI;kYwl;9ui{Pp6Jm@rFx46C}J_S1FJ(dLzMR&}j57~OS{ zmLLr9C{3v~D?QrgR^561f4%K#^(ig8&5h+ucaeZ-IGgFhEIsFD@@v*3@<$KkXz3~~ z%^eV7?Lg=OykP?tR<3qH2AvtY!@Zqt6fEbJdOP{b&sOt|2s*j@=#a3LjIQ5{+}V{j zF1as_WAx!JySKpRHOE=aWW3PhApZ6cMfx$3Y@Tkdr*rK^==l{=Yi&k!08Bq5@fJvn z`h1zeX!S95hLO0@-d8rZD&0^6RY%LgQC4-Pk+_2`&5+n{3l7*xWt&jUVpp6jubYia z4}Ak~R2RIzVsSnC%zT{mq7rR}N5Yih2E@asyOJ6e*O^xVkd%xlL^#8zIkJu&rsW zWGBD2UdxYRlv`SF7V()_Li{3lfO<^IlN>8{bzHjiMIgx^o7KykB1c1gV3eGGN_iw_Vg8jwkbjJV3r1gs^=Nm!W;aby&?CdjUQX8@ zY|Sc?@NUPda+q?Va4Z46;X_Fd>Wn$J0GW|Ec~!ywe*_@@}KPy;6VQwL@NX0}0)NEgDmjg0|9$rt9}X zm9b}Fprl;IQ7PN7%hqalV8~UkU2Spe(I-++DizTOJYzeU+E(lcV(qi3!#-e-ur?!(O@1N9d5MuaZ1H3{%wvEcGv7-Yh<#`d)tlMDp9b9Ad+#2UthQ&%$s) zH4#ty+?q#GZEYG)LL;N(I}S8P=d#sS=sRBR|M5OOFvW*m12xUQ*`lN*>UbGj$oT2m z{{0laD6ecpI&WNnNE~MJ&$w?d}rrSS=3MVG#DWV{GRGRo-|3X z!>0~3o=nBA{=SnlP?f1yka;g&QBjgs!0Gx*~FzqsNWwsdjDC7zB6}!I9qaW zbiKyKCq<1E)KpL040YfKn+0aM+qnb71_?71;g(kTl8~_blb*t%ZYq{$8=brnbkPr( zK_CNOqR=`q#s*I2v=tm`#sYC|wU&$G4p#Uy4Z2Vb*}|^Esgc?*S1*84S&4STDYbtP zrlIB7(;MHsORJp|ZT*#AzM({9|DZ=ko(u$%_zoDL$W@O}uV`LJ?fYw;5uqG+9 zT65bK=;yJ>=w{_t!iYH6OBp~kO}y%nIgyqKR`rdq+|t4z&@%fu zRVsDIuJHv(>y%`IWbML8DAA9_eGmCY9D@kins(I7lNHyxW)u@^IHRat7>BI~c?B8; zy3vnyC&a7&E72!d;dtc}TWgiC=IhIYF0}oSh#<>C$nN?~qap^Q9JxLL7C!JZaCh2hfeXN}FMo;&6IUNIOTlE26tT@muf-e+Kv{lN3K&8!p0prl50P z?oCrvRz47;Y^oo$sW5;aEr__SYJ@R}4Pk{r_7k=H?VY8ZRE*+>WB4AD;u{Hq_Q#b<34#p{c6RF?tgNi~@@w~mU%q^4 zTV*%0gA9E=R|JfHK!)!jExw^&X?K)1Gc(gkMK8Ry20YIhm!oh$f0}{QSZSvc9%y1w zp|?SXET2Q&e3mL`Ke|T_G4j|S)JB~gZqKExdNBGO)t()%6|K99O4T?_)Jla5+<%t( zIWW)|X$I>uZKAJ11D^i>&9&$5i-9<4%yk=|LPoIkAo9T#*8}Y*QSDT**fR=tC6%;Pi(`a;f5A+5YY5BU6JrlS^VDMwe=Jq*%&7wV|Z!omW7ipChqYx#}rIdjo5Ff=r> z1y@Cx*NCA#RjCOLYB}|DlQ68rpRoZDW@UfeROvEjrBRRmka3gZ6Fg8#pAL2Oo(n7G z^nHrfgut`df_)957yNynJMVT286)TQ5d#=lxqiWrm{-L8Y43)_P9UR;H-$9zuDD88 zxGeS9H#avA1M_7~I7o?zih60P1dzJ!du^7mklxkjUV6I5Z6YXxA|+z&tb|Ej>ejK~ z^uHC3Q;p$I=lNdC%1&ugw>o0Ss<1_{Bq|dwF7tCDB+)d8@d{tUnOKjA#@DbgI1lLU$ELez=g%26qSK)cEJmA72-s{Lsi-sc_-)s{*Z)AuxOA4qsdv^?f=zSk}qT znnTS(bhknrg-+w$x95D*Jm=f=JL>x4j{u8MlKnzLdP(zbV3bjg#+d~$Y?;u6Ykvh^ ziR`@-pph&PV8$KAb|#=muZ0K#TMH%(KYhz@a)0gr5O^z_=i93rQz$|j{sgC;JGk1G z&Sn%dU$~j($Q3*rF;VB?>ms##!9Y=!TALN38%@Ujespfag`C#{jsU2yd$&ii_GBcB$NT^>(?k!k}g1_1R!cT8g0N}uMm;uBK6=j zKXb>3MIzt-?ET~W>!0L42Ti0@;u3;_2jKuQ-iXi&J$r2w-vxuFkxiFPIeJy> z!@}F3r4rPpB+4~SUD?0MfY?2+PYzM^)ab3&~Lc=kE>|7IgMrHm>yN;P^x z|Fq@KJ&;rq81Rr&$-b$aolGjuu0W5#CbP;VP7ehhoDT9r!$2pr^O!tV^A8sE?0K+I z{S3upzyp7WG;*MeD5_a z9u<0cJeBWZGX^kJ2gnYcjOyv1MxPV76Z0`1#PM62s2r&&UObKxb(hP|NG#^A8N~(wbq@_$ zSy?e)fv#EP5#Xbmpr9Z$U|&*+)?H{MD+lB?g5Rb>Z{Y(OW32p6GB~fJ^t4+m&&>q4 zIg!E{JS&@(l?87z2K%)NP7&9%K5mMe@R%awi0a5gjEnQ|@CfqoI6XtJ1*cc{8FTry zmVP7Z{BpKYmwwC+fUp0%mqI3x2L#+a$sl+PG5&2HhBQc+Cp(D;672xC0H6yZ2W*bp zj=_iat0(Z{VF@eSrY`rSBIgSr@WV$-5|W)B<5j5ukOr>ZH;|nw-Q-Kyvex|sUYu{H zw!+SLh_ZWoqdDW0NaXz7gcDOjJAfIz^82vB?~7lFLKrt)J~xD%9{I+v^zBPOfdIlX zLr}5kR{3G$R^dR>#wUviP?>9Vj%s2XrO;z1hj`3!p(0ks+m!36iWGtK0%)1I4}w7X zAZ<&pqEY1(7ce?tLKfjIf)+cDz6kPDSR_Iv9RqI2P7QbvP6i|`j#PoTsTffJ*>Qn1 zM$RiDq?DaqxZvMQV_2ach$8oPJPIs@poy-Yik$xgrmI60T6z&iEt&E#VuVsfZ2w6d z*mR(aoQwKyuAP~nm*3t|67p3T|gg{Ohrx5#-PO9A<>GAY^Z}ECt}&hB#gw217;q)f zu=5#ozFCq_W+P8^5ssKW&5A*Ak>buLac$2=%g@^c9yHV9CmCA_^!8>oX8x1*7`zy^ zzRX!Ol)T$e)%E^##Q3gic|xX0f)c9mfYklwC1LIbiK1 zzQtOwqsq>19G4(wu*)?l0bB&xVJP#)9?d1Y38Go+FvdwUhf}U8D3|DMfa~WtR4ZKy zy*F92B$hO9zjS&Wh=JlcCYDf%3w%=0o;9Y}g~T3CY!%YmBMUd5w6dFBiz$jV7YEpO z9-&12aQk3NSW6t98uoDjAN|~1>xZ6&_9gpxFDE()b?uCl-Kj5w${=tSwWmsmG-wzZ z$RsXdrR^r5LkJ1v3f@(SrKa&@a$#cW_5_Z`rt&FmkEU5|FMd$b_|9pVB2B}T+tn#z zAVb8;el*8<;3CZ4&{gza_iu#R%dFvEnp%=WX)XXJ+9FSl5P=A9-{di>YKb+#47B+5 z`2jda9_p4NHRY2my!1Cr;c%SX@aG8tctp{b$ghhv&VZs?{2RNr^yGAVbN8o9n2#}6 zVSF5}Gil@6WDNLdVslN5>pdd~BpyyoKO`4L&(qZvr8GE=mVk}d8sV~Bje{d&$o!NU z{D=eM;P%8)S==N6_~jx10YD%O&OXkF6~%|1ahW8vZtjMGQ4uQ%Y2EW=h_7JiaU!lK zkpqOZ9Md98?zJE7ZJ=W7h5W@qr=;h>Q{ZzHRM&p)4VDtQ?i( zxkj2m>bcMwS&r@pfU=l}hv+^0vQ(}1G`a~PT>>VC*ni>l>*QmImq~az6JhCt+ ziimCqe6b!!5XRi;Oz43Ufon)7`Go!4Kg^@x<$tiBPb|1U2rSo|VN(Uil zGFc(x*?fkFaCWq>vNN8(O^CO_ax=xXow3_h - - - - - - -Extract posterior draws — fit-method-draws • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract posterior draws — fit-method-draws • cmdstanr - - - - - - - - - - - - + + - - -

-
- -
- -
+

Extract posterior draws

- Source: R/fit.R + Source: R/fit.R
@@ -189,249 +117,253 @@

Extract posterior draws

log probability (target) accumulated in the model block.

- 
draws(
-  variables = NULL,
-  inc_warmup = FALSE,
-  format = getOption("cmdstanr_draws_format")
-)
+
+ 
draws(
+  variables = NULL,
+  inc_warmup = FALSE,
+  format = getOption("cmdstanr_draws_format")
+)
+
-

Arguments

- - - - - - - - - - - - - - -
variables

(character vector) Optionally, the names of the variables -(parameters, transformed parameters, and generated quantities) to read in.

    -

  • If NULL (the default) then all variables are included.

    • +
    +

    Arguments

    +
    variables
    +

    (character vector) Optionally, the names of the variables +(parameters, transformed parameters, and generated quantities) to read in.

    • If NULL (the default) then all variables are included.

    • If an empty string (variables="") then none are included.

      • -

    • For non-scalar variables all elements or specific elements can be selected:

        -

      • variables = "theta" selects all elements of theta;

        • +

      • For non-scalar variables all elements or specific elements can be selected:

        • variables = "theta" selects all elements of theta;

        • variables = c("theta[1]", "theta[3]") selects only the 1st and 3rd elements.

        • -
inc_warmup

(logical) Should warmup draws be included? Defaults to -FALSE. Ignored except when used with CmdStanMCMC objects.

format

(string) The format of the returned draws or point estimates. + + + +

inc_warmup
+

(logical) Should warmup draws be included? Defaults to +FALSE. Ignored except when used with CmdStanMCMC objects.

+ + +
format
+

(string) The format of the returned draws or point estimates. Must be a valid format from the posterior package. The defaults -are the following.

    -

  • For sampling and generated quantities the default is -"draws_array". This format keeps the chains +are the following.

    • For sampling and generated quantities the default is +"draws_array". This format keeps the chains separate. To combine the chains use any of the other formats (e.g. "draws_matrix").

    • For point estimates from optimization and approximate draws from variational inference the default is -"draws_matrix".

      • -
    - -

    To use a different format it can be specified as the full name of the +"draws_matrix".

    • +

To use a different format it can be specified as the full name of the format from the posterior package (e.g. format = "draws_df") or omitting the "draws_" prefix (e.g. format = "df").

Changing the default format: To change the default format for an entire -R session use options(cmdstanr_draws_format = format), where format is +R session use options(cmdstanr_draws_format = format), where format is the name (in quotes) of a valid format from the posterior package. For -example options(cmdstanr_draws_format = "draws_df") will change the +example options(cmdstanr_draws_format = "draws_df") will change the default to a data frame.

Note about efficiency: For models with a large number of parameters (20k+) we recommend using the "draws_list" format, which is the most efficient and RAM friendly when combining draws from multiple chains. If speed or memory is not a constraint we recommend selecting the format that -most suits the coding style of the post processing phase.

+most suits the coding style of the post processing phase.

-

Value

+

69#_(iX7nl*Wd7ZCl~Gi;(c+k<6x436=GseDnbF{v&f`!_so)we=8(`% z7CLQZmLKO1!S-g9PxyhSRbzVUB_xWy#rxpX1=d&bvVjGg*}B^f3aQcFQO(7N!C3AO zZ4>#qg)={#zZ8?wE70>YQt}C-i+OuXM4KFQ5#BjfXJJiTBrKmkhWNujTv@`@uK-LC z$*zhYzVwDP1C%MZ4e8q&5f_tbtY#y2tl4H3craor>m5{)w24}MQnrp zTOhNl$jUFgVZ9f1b#E_%W#v1qXuGOkaNfjOVy!&~_kEbr zh>{%5ksTt#Kya1JVlq;SQ{TGVlJ4W&6@OI8Qz-Rp*uL8N9Gl<7rPvw{@YsBOE!Q!<;$^iHfOA<)oL z%;xM^hrN5}4YzHUl~tB*J|(SmO>mz#DOWMHlR@q}kBk>1z|v$uTU&1%@l5!Ooz1;x zOK0o_TS#$--nhSy%TK`+tOZ>w{Zo=(!f(V2#iKFSus>RBDNknzz8Vpk2#YI_ADW3T z0)OAS#^8UKk%f~wrPaHhr|Oh)^3a*-Co2Yr4f}TYj%(@@O}~pp7Ijx}&xr38Y7PRd zY_yHRh`AFXp8-R$%1n-VvmCF6#LU~}z5VCv+J6?0fmdWrFo^RbF>2rBCd7A_VrIM& z=cNW|_y2hT&=c>W#rIGQ`p5}vMWN;2ZrKd-9oV!-W&e>+VAv&-fEX7#cGNsd?M`tR zY~l$4QywK!;+tlu)+2Zk4!NZOyGejT4wLNDJm$#AsuKyX*B89_aa1!q+ElOgPTf17 z#p>ubY?F_-Pwq=XXq4Hz2T{;OaKb9s1f=)A)a|cSCaX?SIE|nzEBPa4=+rp|-sop35>gyflE~zYU)JU%!9i93^ie?9;|0lQ1WbeD z)rdgppFs?JM(9#QdXcL~+c4fs^~0+-g1T1A)H!6q`TCcK*)ot7wY~yVDvgL|6Fhx* zYBFKQf0?=(=s}1ty*$FrVgRS4ZvmM1w?$l_vQAgLJAvL5TUz_GgiRoJIoHks42cj} zmu_*dL{QfWziHuND^KBH9TE9w5(xz`Odr<9_AL`@EL$dnIQlJ>p5H|dIW zd2Rdzi}FkjQX*fkn>h7uYd4cyrG+9-6GFa?c&cyTHeca7J|s9;|FKHjNe%Rw(@boJ z^P!vSF5MV*4|fsIXzWaLK;JRRkALs~q@BP3T(54t9?4H9XPVds>rYK@w~VP*VYzH< z>4=Y4_LL-vD#%U^S7HlcoSS2BnI4we-DdlG!{uVEK{i}eJv)aWu939Rba%l*r33+1Js#niTVKjB_vBAp;j5`=I9XUACL;mKc3Z! zh1Yj?F9nBlMkXYc4B7l)l94X7q{8+`ov()4XpK#7Zri|a)p&9?8RsU36{XZ??KMG} zLq4N=nZhdq9ykV?zt0E_-~L&I&GEC5*3emyGBH3q%>;_|K6kVnRxQFS)x_^I)?o0n z%_*jagy2pvlM>c=`LEGb6VC=qdVC;S@?N2_F2AFwSt|cn(SGg3039|aeT(xNQH$*e zZ_zh+w@h&6h|JfkkvmJEjeFi6m}HWM#!ixoRKXd6s*qt--Cr?Dbb0mhu1o*q7`S#V z?(byzLAJsFZaYl=to4(m^WG0Q0_KhA=1?HN;oXOm`GtEZGa^x8)%QB6VHs2c9B+Pz ztEPqWcKd`{SeL*QpeYQ{FQphF2$`9;v(Qr4E9b5;{;p~R&*{BM!?*nvc5E}#`wQWw41XT!miTo7Hls{tiHbtNHY~p2Ge(yf6ZG8ANcS!;=ynRKRGiCK zXTFqm(_+UInqI)VS{g2hSp_pHrPg5;As}_BBhL_wQu@38A_CDQs@u-lE3&NFZuNbI z_`wAy!SM9nX<2W2Xu^*xS+v%F`;#FnBcJ8I8;(n&xIPR>q7VhRY1N!F)X3)DJ6%Jb zs~1dXX-oDJ^O}uIbt84y6}D89uVP;oTthqd%786#5H-xN@SWSVi=o)0R3mGm`!hfL z_0iQ!00JQ3nm4DqRGI-d%e*+TaFt}sskPM;YV?vnx3a?hSU}_stXW*yDvE&7xD*>5 z&syLq&U~o}hI}>IX}v6QOEL$9CM%*g;P7_J6^JX!Ri7&LDVEj8*|DKOF6VyYrI@%) z)L8qHN}?xlY|rd7b#%9@bT2Caa88|e+CG=XuLYP%N9wk-V z2%9LC{((3*>%AF&i$$X@W8i4>7m-BWa4Vw0iI5?jJ zLQ)B16X?s;tFhKl3jMU+AI!%MJ({z+HRSFv3_Bo+7kz04eEU(|^7(MyNQZd(nd+6l zE61KdqUWLk*PO3PO6qx^8}2s!)3%V=Y3oLOi^JZ9h~!TpKeNY~pe0~_hde$^x!W`6 zQdug644af9gwBF;t6YE@QLZrt^g{`kN++RKB09YCcMLKo`yQose1ylma3BqIhtk?3 z@tnV`DU7ddJqi!}l|vgVzZUwmFOGYC?;4@f-5ZENFb})-o`qJ?RzYVP+ilNgQB#JL zNm+pehFV$O5+aer=~RqhZ7h=PhRoW|E-W0EWSYD-N-)0h+v{!av;KBgvhC(pb29hn zNpI}5qRxwwDvHZ~QKOwCm%z}!{PEx3JO9EK_Fh6*ouo^;B&sV_loLE(423#_M|}9w zc)z4b*r$-7n7l`-HBC?dIRyFFYX*@HF+lsl19H`C&(YI~XFghu^Zvk$hJ9;YRRPKs;6>P5@?J6A4X2SMa`ywR1ZVh#vD_yWH zrb~B2nuMaDc1_zN2_iUzup)&9%eE;C{vyQjogJsAmKts^{R?q=g+Z%m=RHJ^!s41j z!)*->9u#uA-9~{IPNBCb5<5^P^-)T_soik<@G28V<3Ji}{Bo*ukNm4#HoKW<{2Clhx42%Nf{Fo0~I-0^3zg zNSVx`xKX{j`%~hB%!`5}brhVg;en^aDLL;dgkoI{{O`d8s{KK( zvpTJ?qwEiV@uWP3TI2W7uY^=B0siRw2h0{dcxteV&XU(Y64PO1)hK8u@fSfv0Y;#8 zV)9_2x(~#(Bsjd_bQacGSBE(>)9Uur?#Hilj9kiWvr*B4)xvPIVBrXxjWD`MwL@Gp z-^bub@T%9mGDWkzRsEQ2V;xY4{Y;!%scG_}B?YEqP^-T2tqGO?if^&$dJ>&CdYSSrR{Y)TdpA~0VAuY!75$4N6&@Hhs(kvK znD~R`7<|^gF3U;+*lcH&Pi7+}bnGaw=5jOZW)UjojIy#RtG+=mEjBPssJ|bznp^A) zja&H~nszgdU&tF4lG6nS6}QZs&{mi9GjDugYLCp+^!2GP))7f!`$*Xzv{+d878P*QO?+dBGgnpo zC0h0|SfM8#Gvb(0(URp3Zxju50{R>4?ifqbwX7um{gx?~UE_nu(+uj687Bm;NTIgATO(%7WI zi^wgz;!0(MO(!qyo57{Xs1%n%xY^F!M4WwqPiLN`OR0Ki`_I>~Hk04L4)zAI={RXx z`*3?v%;<;Ir6@WPo-Vq`8LC`&Urg6*|3}_8Yd%AO_~$s;-}r#IKDOAs)BF+`SBQc5 z07h46GHf8ZzCYE;I`0R7H#qoKlK*9Nsy{h~t_Kq-UQWb7z7qX-t~`a0KA`ro=Iz{J zFg&GaJ9XdbKCRSlcVVtPj|DT9pJ@G{sqi6ttlZY#Q?Gn-n2a1g`Rb77!})fVjv;i!Z$@Y-d1 zrz7*s=U%zjkLunwEb+zK&jDZB@~c z4#GlfxT5T*URl;;_Ju^|QyVu4OwW^q2j0%%%c^Tm>(aaoq3D~4bX@G^bEnp09#+r4 zayHVRd@%gES`HvAQMaCO1nb}OX%HpSMoamzab8&6U}S%&AV&NM=>r;=HMIL!XB04I z)~4CXuW^V+s&kw4B`#TnMj-|LwiCOLI?3iiPhoTaIG{VUbLtbp{Y9_s0z`lO& zkS|FyfF^b;W((~gJ$5@yHB|*8zt)(LLFqm+HhrKjnXRLJVa1!&4g7_In=hfbp?};X zog}3Hwe$qebjJ3Pq5uxi)hLaNuR6bV^ut=j#$}3Z=B9O9TE$i0-&Ya83O^T7rKNnf z;Cyk+(hvPWK*=UwO31*d8ZM@H&_NR)R}%~?$h32_ZjX%yE1K*;obGSSh@ z?CVWq;!$^$8EtB@4)=DRb7dBV&VPYIG|rn3S*e`pJ5^*xSu$|?I+)hNKZttlZ2Nqh z!052Hq@_ZKG&xnNoBCik^xS9aRmh9H{EvjcX0av?n~SQKY5}sOZ{lmLC+Y*U2LmL< z%nqF4$*s^C(oJC}kbDq#WMiA@mQrnMF)B}sJmN$^qp8)D0Qmxi#rIMQDK z7V&LIam#F*g*jNq=->?)$S)>O(yr0GUWzq~Y|=V`2^mGMMI$QaCrnNwU48V$A3N2jj)Utw_-{#Y1!<*NO6Zrck zr&4djmcw|EC+mdLV@5jX9;BF<;Ru3I)ZP4_RVmD5{K~>^@AW)5;z;Sw=uB!f>Dd)H z3?8hDJuPFuu3_4RJ0P7{I`5FOqIHy6KEWOGzu_r<(v|2g<}M^jrOns0 z@eVBg8R;4J0^kITs>HXiW^|n%JXJ_pM`m`4PoJ$`7p&;>f&;STqSk_ptf~?ARXcNN8>R!cmAXclxS*y+6gp~_rVQ~i8 zYAQdTS-qmUkOK1M#7esUw+eYH7Hr%37G^ZU z4w5PFx3{sF8P*zF&QXO; z(l=h^k@n$cUz~!+Jyy;712hhQaMK9c?7Fpde8tK6j~35zby@dNGI*nmViQ=E14K6e z2kW6nhRtD@-*_Q}dAc!X$@%!i779}%!r9XH{07r{$%(1kHkffXJW!5Qx$HV-y?^Lw zNs!`L0eBdf-Seh&0S)4(p8(mM4;379Sxo*9EwjF!;zVpNvcm`&Lnr5KyE=SJcc4hwcrmhZfz)xbbM=2x(b}aV; zc1{xMM$airnHAO1#p7^vLW``JUt(q%RRykBPQxtXv|Hj|p}03?&65AkiCqR|_#?ku z>n8@zO82nEDav)};gHWG!?p)utOnAM#s!yBQ4v5~_epz;V;0x<%t5L}d2+jsu_$3aUOYzng(T5IoEI?{>pxq}O)hMj zm9i*oMW;R3C}na29Cy}KBD1@LIy0-<+FPz~N&m7*uhn&LX*z(2ksntCa0=vOfCPI;oYJs@0GbMQ=>v zj!mEd-96CfG}k5bx+8Y_ttCcApiVyC^__+FQidPSi|7}d{cH;(%=Q?XRwPEYX`QBS zOk3j4(Bj6}F(;`*=3ep_g3i@%10{N@oRKx+OeCn&dL+ID6ff6zHa^XaT7yJx=$b!P zocVwISd(PILt~NDwY_?hUHZ0qn+FejUB8$h$Bf=u z+~?N>=G1Hd7dE*Ozx8skwF1HPF*^aS_aeQwuI{Sl({DFyO2n}7H^NkA@%nE5zIfNQ zS~_+G6yuV?&A7m{TsF+2s{5y?=VdqmM4)9=xh5l7w<`Uji(J{CI@q|Bz(o-_lVqBy_6-K65oR+wrEMD0TnkT&V7d z8V{x;0TsC*DtBZx`;2+wFxF~bnTb^Zgr@owNJQz&1od6a=PZDFlmBh2>Gvk_7j3W_ zI^a;^A(_mf**CKLlO&l7;|=lot)yNm!jzJi2g9_WkF2H^A9|0zhv>B&G+hrpK2--c z^p=Z!x9F3p*a)vrCOh_nyH7{<@!P7oO6zzv-}SHX!`T^K0WY~E%A2t{$i z+s>3{2Zo_$4<5I|{g_>#f?e4CDs0dyLjd7E@+6RzT3QBlA(8>CKQlR1vnST=rmJR# z=rlAVCFG1VTKe_cOOFm<`3B3XwOt@Jd$EZ%LgoV*&tV+O&*UNcp_+prMh*SVdP)qJ zBY+NA992v)E~X=Yt47MOhCt;S4l{4T-z`zu7;AV645k{#J^$QCX8Wa+B&HX*j4Bem zazQM=l;j{|cP!tlg>ARc-Bi*ZduKveC})}n4Lzl_$23?5QeWsw&FDD!QDYbBtbf%) z2JT9#XrXgF_(xv?XNO_eA@NzozlVTO1>oOCs655R|IXHWQ0(^xCoQk?Swj9s*qga+ z6KggotD#8sqZzsIIZ+n_bC54XtL=XIB>e5wYk>ZU^@x`Tt*@SBX6PZ`UhLASO7efJ znRWkT-JbJvS6Q(-0Yvs^BF1qSs3o(L}W1$!Y;daYHF{h ze|hj#sv~mk(dFj}hb1Ygd5>#wH3U-XWj{?xs;$j-WWW0Qr|-n7kTyyb-}=~Zs|{Iq zJFe`z8Jla+;P-2c0-s+BnwxV`-A)ZuM-IB}hm`j999nIWZ^X?u`s{0A0A0VzU`m+0 zfvWK;(O23}Fjg-KwXZvr16e{YV)5;yQM_2s>mx~Hy)GlWe_RpJAr2J)5pG>!MVe^6v#$Q#p@Ia*o&r4_D9a#-!9eBpJE?X*DaKr=ZyN;DD zt+UyyU%UbwQa!4!d(4>&Kp3<{T2RYS$CK$Blyw|0Kd_j=;eq@YcgQ{b-vY|h z(uZw|YPyVBDBY7|>z|*E^p0k$C>S_LLk%(cNzS(=XMompi*IrQzrMvu%`=B!Vk`Fu zIW*f*SDPwmG5l}!Ptjm3b@@owAZbg*kGYa^jedu zRcGQ(5r;{l2lJSS;59BzUEfjRJZ||?sXtw3djpXeD5MG`M*-&eD3QtMGZYKqHqHt( zwtKdV67m|HdRzS2%miYQ@oADMS&jmm^-y?=dxmD&zx{0L-+-;faGyh9| zh|&FDmiaaRlbvmng$yLcdhe#Ih@YE>A9dclg&Wmgsz;DcA*xu>b@=MWBy)8K6H!Qy z;_t^F0ITjrHikx*07N7EFH^AK^TY4q70w zP?$1Mc_zB$TBVbcI#iG}1}<{!wgS2V8alKl8?IdS@YQabU$J8S_-`6qtg^$khHbE^ zPvlAaMqPVMAdG{X>%7hqdQzKuaXzwWE#mLe^j_1SCeFV%P_n^My$)$5E@flLgQMm< zftnT7RTtND&W=~^lt$DGH)`d;8;_O{^-ExTj|lrAG$qx z4{Rc?1~4vs(yC{!n#TFB?aMq5`;It<(8mc!0!~f}STKR02-Vm*us&(@O)#EMchuZ++_D1B3kb^d$$F-Gaul;%_CmpI@$IcA)7F zdqi z$?k2q83r&PCA4>^4FjIYn+H-wmFKeM&wg8xn=4tew&`<_QQ2T>Dz3`NuDkZFZZ%)Y zq(lbDVdH@H-V73*N@n_f)5N#XP})osuh;idhNVQ&T4?+|dzW3)`hx|_c6eX^cWc7` zc>&_njSS#wRqlHcZn(u1eFefMY@__i`5a>;C9`ZT23hO|sq-0iSJnQ1w*hTFHsx&0 z9qLp|{djhOYXLp(c)0tt#LqOsNBa4TjWs&G+&H#0@s2{T@8a5=ZrJ`ljxD0!cgA9c z`MNW>f)=~6-mqA&0YV&BZ#8zwQSCqL@a}}@b^a{s(n3FwHOpIU46k(H^-qOcH`g>e6dXlW0kyJulKAR0b%Z zLAmMXj8HDaw55kJ0I9Aqhpxx0a^#<#`yGR&v{?sCZ`|zAX{eQLK$VMCI`AAPo05O) z2TT{aWOx3zkhh_50eg|DuOT*a1*Ta(BK5?85WD6pYX-l$T>fErsP%`c_;0*FxY%dw zlj&Mti9)AuHzPNySDfS}GpbNMrH=S_u*zBRBZFymWKKBYVCm}9!-&~s1lHsBa=D8K z{C2=u5@OtM{G!5Q{w{2RwzphR#VE=np_2o6TRu7BNedUW2Smzlmvi?nfj0vR?wZA> zy8skJrrh;M?9rwJL9qrYa)riwpdgV(f?}hTjzyen%jzB1>sv6qiRiN(4496_KAyK@ zzrT4TD2j~C|LVkq1)llECxAbQ^{j@cQ($F=id1Nt4IpCME=1CJSR4H|0Qw~@T-b$) z1!ad#Ln%4Dz@>1%x<;+@nj-PGg-T!#A-_Mm4YZ29co8Sd2n9Swai!zcn$y}205w`J zz;)@TpiZKpDG`}^1X9*)Y}R7r4v4b^qGqrsvPihwU#~2*m$e zssQVaGKv4Ll$?hhsh)D83BC)+uyR&xK6GMbK0~d(AZ?S3220!<;He}muijkIWk}h6s3GZan@0fR|XUKQN7@in+9f# z?>lExp1;UasrZfF*4gju$qGQOnGAND0n3X2GM0@7GWBqO0zd*JQkAo{OZ{eh0kAod zR9(Sej075E|k*YNWPG2pIYbxR7pKnaOsn2tGczJ%km- zFXR(g@sB~nwf+90GX2ArLWp#NOGA$9ll4F$BEq99n36t)YQk?FRXbgzWA2=XKlu?{ z@ILP(G1n{zzfwqd^h&Wc3R&13a=GwBzX7rdq?H9e!$E(ULwD0EJSoe6uh%idbVO!m z*siNafT@6h+aRq1%D|s48)HeI1JC%kSqt=T0}mW0$!{^1X(D#gyKL8|CWaviXSqV_ z3hGu~6L-+T?_#+zXK>cA6=&+Pj2htSJT$9)OO8R_o5>UOP zJT5EMAwMTQH0oB)6g|W`G^{9#e(hG5E%0ejCeO@xx3bD6rUeD?Yx zrpXH104x4nmc4OG;5vdzpuUAyTnR%ZUr>e7su?%A&WTLTvooZ-*S#qncV=dQ-QQZ8oIj~ z)Zk+hg@9sd^CXd4jnmVO3-velWhW}#i?qT;y~|71HPbrBxgExoDJb%C#GE07 z^EzC-dcY1{xZdwQXJM6dD-mzNc*m}A`SMvnhcP_Bwejwa`o!;3KoMgZyV-vEEHLf# zZr|a*kT0lKZ+96@`t$Muzl&0ARMQIJoCM>#aK4)O$f_ujNo+1Xtf0Am8{Y5Kkl_?y zvK&!#@cn+#k?IOCh}jGnIw{ZtW{@nXVZs(Fj^65JdjDawSDrM4jDQJsosK?0WYVF{A3z7DA)FDMWsPR8l(iIyFn0? zF6r)OK)Oo-LFqH3bPk;(9g3864Ba6h9Yg1LbI$Mm)|x+D%e8>HpXa_}?|toS0+b-V zhjdnCFgjffbo;jjntZyD*3`jy&HNwou$} zEt=-=-Gi*UEP*HtH3_XhHlabAjJz!THn$%J9?9|P@BegrVl)Hc7s^CO*9OHo!+=m<*mf+jv>_qO)4kXkf<=T- zv+MdwRlZLQT3}sw!jB^ssx?QH>ba%5AFELbd<%}lV#)OV z<`Qtb%qFY0DqaFjbBaG9a-Vz$nbtiwCtxy6kGnwoc1tkJYpcaZp%2(E;82OqBaK8? zAf7YgJ0I}4X*m;Je0_AWZURO_d)5+W6bPR)ETss~H}*ME9?m~*+$FG_&LbAcPsk<$vVlHcE3DZ%PYy;w>Kpaie8XN56xR+Wl(#!QJE#}8wyyDjao#zR zWUMDH$V!fQgtnC9zvEYE<~=OSfOj zlbZ5XJBLpSg;ASQ+4d+E>n9^@G_ZjRW$~8y^dZn-R#2&U3w{wwUQZ5HE;C!5W|q`F zr4lFU+);V!1CQ5F)%;NE{iE=~b@Ne<1)0@5xrgCF0MRgfH0veua0*cCEk^w&_bv)5 zNi403sst&T5#jl+XEveE6eESnR2_$xt8yJSs@2o(Mt5yP;GO?4U`o*i6bm9aUJpG4 z>_B{TjCoFZ;?M3`VzXsSoW5k!OwA5$^&^UFm` z-5hEoL1-xiQ+pC9MU#gGai{DWbh-*&-^3FK1-1yRD|1-=Q?z5x>l`Q?pa)veK3>^+ z0c=^9%p8Kb#I`mCBDukv)sJ8}Q(;ofe#2 z%l{SA3n@Kj4F{0wbLZvB>L$1@mkwl#sy{FCczJeD>#veIP{K}A4FXIi=l3(oW;a$J zN~`|dG`fLJA2PO~t^oFWMI_2wbz@9=*0;H;HJevVGKep`dly8n{Ov*Xa}G?#BPJ8z z-!AZQV(xXRN-*D^6G_g7Jaq3i$#rkF?NgbKQ~fQVvQDAs3R9IEsHs$ZMEJRqQIOJ7 zG$VvUS-reW@Olf!{WVqOOP>IVl6A9mbIB<_chKuoGYY13?7N2xuq5^^^<$k9>NiSj zeX;&wzEBgMj|@wFL6#sC6ZWiKs-yiPh!w33@o`bBVNPS(PWXCvmr$?f=}Uvs8R7D5 zym!-I6?TnY0pSA+v!+aIzF%CIL@z75t;W!gbHuQQ{ab4?BG7C!URf2)5X&h$lJV$T z(5a}5>4eIc*j4G(vAg{veIFHh{_s@ZC!#JX+cx8j{(lR#Aa9MMRSe~~`pR(kVl4I= zuwz-n7b26CuM!4hzWz`kXB62WAWL$^Wr-n?wjvam?!+uF;Yh@lELF;75zz z=<9t6j-~C_nPIt+%JXYT&9jax*G5@y-eU|zEFH_mmL760qAb9UQz2RUi3nt4U;4v1 zMKmf8D)LXuBrl`WaxqHV(euTPyc|F^IOn$w%70$9ZBCON#OzfwNfnL?Z?l_+Y4R<4 zM|G5dxfmSn)xUz-Hku=YoDXkvse6Z+HplE)r%x`WrGUK?R`XWpLa_);0$R{vd)S4t zz{}C#3e**N+?WbM?_x#$O|3UEGhU2%fVKlyK%n`|hxrZQ2$JOgC2>Z{`E>KH>#m+t zMso)#n${SZ);O4Kkx8;92@L!SLCcMms5Bvi`MKAebQLJ(sl=GuT`yaaPp0HBJ-Krk z+PLa|G$HKnc3(P3sX`f;G$GS-<~&YxqD1v`jYToPJ<_GxKrzT-Ax^clxjxhiBWDo{MZG)ZrQWBt zs_qs^gpv?X5lM7xC%nC@jg~6;kszq~{9|0qLU5b?^sCd}UQ%Z9ZFacSPm@zyv}Uzm zo=sR8(OCSxCjy~O;+*bLWi_aa`-(OcQU<%;V`=d8Xax*gRkLQR!b#^4ozOSl7p7_*Sze227SJFZ~8z){PICGb6+_noAlL+fujLZX}2LlqnMl{UxPuA@|Fwv{T5S_#R z=uKLcop^MAq^lt5$Isch_{n1k1n@pMS8-)1HEuZS;#ZzL9P)OxKdTl3+%;0PulHp@ z%>2!X9l=pP8F)C$#Oa1J#INJvEBZPOc@luBc#fQBPb}WkB9QWS)c9G#7@m`sTH#?O zN@g`C8aA+(%JL}$&tQAfoByLsdn8O?t*>m*cSdckdVAdAb~dxTa)N@LgW>birQP%H z(Fbkq61Q->pR620Bty1Sl@ry*W%e0n(i7$Cqt>0cpo0AIPBzCunTKN*Bdz+-=_1(gy+-vQS^70`|l8-sJ*^(0U zezoxmYyQ-ahxX8}FBBh2IPzOaNkJ11t{q*zMN`#gABL3?dGoEeQPcLy^9wJqpcQ;xQQc_Q-OKKPq(HR>wrxHF7&3A2HO?qtME-Fh#!+O3wGWwX3yZRTe;K?3S z(%Ue(Xh=9#RisbLZuJNzX$V(X+?Vmw!_1=igFiOXKVZ(To!zg)>^c#Qz2VmFHq)*G z{XH2^HQ{bH-hDqWy21l#P>xqYfDLS&Z{sT1QFLjYC0$KM%`ne_kV&al$#TaMlH;U9 z43ie;Z8C+t=1HZl{e1^d^QdLB4Lu95gd5A_lISJU;GW6O^~tWBQ5QE+r?1>U43~fV zZV4S28pgyLDLM()_Ohh5%WZWeTX0Y_41~ZQ0cMe z%^snYYCQ66dFU?z_(u$QGfYY1OKWe`HILD0bZi^2+t)rH6}UYsXl`kQ9BniiUWPqX zK?NOdRxf@vqvfl}O@S9%iCL-wy8GwHWP9XiLnE}=YnYc&F`f-|MG%r1?kwSmW>zjf zXQrSMPbNjI9L4IH*B4wYEX+ZIFFBako!j85{j`Hab9FU~m9~vsxl+Fful@*3Jdlov zuPqsI3^?~UL_a?{VTj0EY63Qa@)HXO`f6IxS7EN$@*g}L_Da)P{e(Gk!BkF$@>C`v zb}%Lruf4iVrG2U@u32M1FTxa9PHowHi0$At#o0`<-fp*0szghANP{my+r!g~Z}7s$ zAqvQ>q!+E>tLY6co}Ls)N3b-3V(nO8Y;tnW_}GwtbK*s~TD5@Q!AXE}V$4Nf3wxBp z*y6c^jieK#ri(iHaEEbePLO?LGKIgGuK!jIaj1X4o8tJIbm z8V0(r{dIBIbXSl_N92T=7{_@b>ajP2{MW3J38ab3#K?6vgzFll#_ zvBm&6@ro()B$4!+R9IdiV0>vfZ z>UEQ!SweGy+I1xff~~2X^H-gG%arQ^8p#bn%Lj>y{juMDOg8v6tn9O#G9LY0>q_N9 zBgvIdJKx1>8Gj@H>LKIu>~4GjB6s;)awx7WwA6wO7HP#bpJGyCeYjSJ+eNAul{UU7 zxKg3AE=pomAKgCSewIH3aI0truDhui3`96xNxge`0vlC$5oAuc!qvBIP;~I?XiyGR zr%`%8=40ZBKEDRp+ zB(UBylD5rqB5)gODb?A2!vx`r(r2MecdgeIOXC>Yp13LqEcX~#UpDHn`RCW=p@2zV zh>Xm`r?@E5MD%bS=pRBg^vC!^^0rOSwg)~A@o!O`IDQ9@N7sMGdL4wQ-iEIT{NX_j zT~ktFsV8!`wr9s03(*dCAZ3A%8-)LEdnEqVD(38PG>&KP_oZo*Hm&YI@Te{sW12C? zsqb@uNHr_l&D^TWmFo3TvR;dCp%Dj@lG%{=iG+=(TlvFkzXomE78tDv691^R!<7{K z-u$J{M^!0s;(R)tX*Fkf)=+n`PQ4+OnC)!c(#6W;?FGnPpXYsk$m$(B4lh5>4e9Qs zLA~9U(A@4#kZ}6uam&($MB+l63AmHxkL-l(dncZNVh_$2NB%mdX`8SAOh--(gZuW> zt;Ti;Q{yi~gD zXi1!$5mVKx78&s7j~vYBKH(nbYuk6M?TH9S3yycJ>n`=Q`|h`GJ^tO2o^;@OGy~1H zQq(L9uCc*u-`OPMPF*Sb&Tpafx-exedFal00<9(e#daLVQ{3ks*7Rj5?FocvliYY_ z(;T8-T%{vI+S5a9HI6zHf|_2TM}t5nM$f#Y3F6p7N<98oi|J1XksU|g1Z$$;VX#6` znBG$Esp#5X9H$H_wCU2*2ix|W5V^EC9CRkkyiz4~E89Qx!aLeERV-MiAjiBlX_m#S zwrE*PQ}{e(xju^;oeipZuUaV@7Dc7@8h3ld8~L8N_9r5%$*MXd4&y~a4CoTNnu6J# zR+E3UwaFFRBW2n@3K|3Nq8}{_qKDAGMXfK9ZR0D(*`O zTa+(3U->D!(e}p_^Mj!}kvU9m+yV?ZC~hK}yu_izfsM@zwWw>^Q2kJfB_gynGTS1n^{!1V-G8g6v7@9;L~5WfGSB4! z-*yf3>_2@)aDXBR>jF+3hv}q4(imazC$o=$J>uWSEJk{7)nPZ@y>hZSJ>ZZ|An;}3 zT`v_XS8Jy$hJ<^+ZF)R3(0=ILL)?@^ZqL59D)7)Z4@06IZsSFSWLzB?398mJAs3#! zBHkx~F@_&y{};);A(H3@eQ%)ZT1R8V$= z@)@xxX$UqElY{|OTl5a+b3psD-K6;%%CT0!h)E)+3LV@2TI$p8Dy0_)&g6IW5^8>9 zZ#e^dxOrRW>|Ap*5A8{~(+WBo@(L4o-R)`zKpW>nNz3O3IQs6kh zI<#Imqaota!nlIpyLiI=cfx~0t*dGl{*2r{cJx&KVY)iK@A)$a1TCqP&1+$15We;` znyf~g)PVM;$)gdg&pUb4iI>krNl04~{X^Nfj)Rb>D7IiS9qTBxl=aOU1fqVtN(a(p zUv`_$8l|YWRAp=;1%`J|r)T@hBN>k};k!HyZWU$=jX(C-sZhZoZe*7Ro+MXq3DnuO1 zvrnEIM~$D1tffmOGuQI(oF2p}&`}bhTRH8lBKlhy{oD`UX6M>aK{Wt${t{a2=vh{n z{8~6#WJI15+Z(ONw}QH=;1JnZTvmfhV{K4aA2sqtL`8Q?UkyCM6aSv8+bhVtDWEbr z`Z#OSX_{(4Fpc5*lC3ry3f(^?5*0-k(ur3D5u6vaBstJL?@uC_Ufhul6Yl2&HAJ=E zTs{-hwsL9O=eN_RuZOr~kJ1>~#EU+jq0Ki+)qQwfF~i;yVMpq>y|P@F-u}_p7?NQ(3j&&$ zR^~!QZVtFl%BB1cjR74Cygo%TbE&czGKge|!_DaND-HG}5IJVI*={$B!AqOx{TUBs zcFd-p0l~Sty(<(fB^|3pK==`CWpo%aX<2iElzFUF?pEWW+*Xcv)nH$@$Ik2i9AwN} z{-zVEOiwd_`HQz;7qeBT=DzkyO^^BUxpNzvnDh>L`$eW5WwPo5VGR)#4i`+;J}>`l z`FA*^UGC?qX%GWlu-xzU@UP_<#qutk7`DPe39zWMh-HKOO%sN=#H3BD=N+2O;yPL_ z9rO{D(|;CCS~EY=~`%oBcYW!-iwYE7Xjn|(Kh;2Dso;4!m|uNuL=7&u{YKf4JX zdc_x!*c}#?fY9!`jFyX6Sa)__pX6ZaxrdSO7}rgJgL%vV;f6JfRh|KE#2hJ_V$K(8 zrgOyZ-seIm*8IFuLD#DXhBoo|M@NXg@gJq7hX@dmcDnQod{M|+zoSx8BDZ(X_8g>T3MPBTX%yp5nnYoO6qlZQ#P zt*S}}rd9DoU+kVteAYj`uH~gUA0J2Eb_feYdl)R$Ma$IFhob(b9b7@K>Azm!3f51A zJq;I&7i77J>3&X$jdeLd*0RoB|2sUqT7HV^^Y?9j_eogkPiKD%P3~yS{=*=xc5}g) zL!2HT#QF^6a?v9NcQy)8%%d`8n)9nv+vXXZ^ME$8@^u9j6ru)}E0qt0P16 zihD|{R67J|U%7PEj;&6P1f4@94|cDxN@BLD&3gjt?Hc2F!g3~~t)b#(5eQpSl-2JM zT2n!V0MYdq8#Y9_5c*WM{V4x}SEgOG3A`bt!q*|07QUfRrO?k;!jO&aPb6g={&==euwzc@*i*8+k4! ztGnwpX87DtU+ZZ?f(0dvLh&Vq;n77(XA-b!5_@KN@!h|!Zs}IFpEIaQdq$j{^o}0X z=$n(tp$tz^T^0S62e@o~7%b_hNjrBUX}e67)ntn?v0vCFcF7~wr5Ar_wB)nB^)+&K z_pZJ0mpy`7wBSkF@ z6>vNhC4}rzUlRdnD(eMRY55x2`#V00GL>N4DO>-t0&~?GgZrbPeVRJT7(yOqrKfooLoEavINo_`BH1G zD?MYONw<)ghKlMhJlM>Q?pH0}8Quqsp0WjawtS9Pr^I;+GwrhLYxnG@IBm7e;aRBQGz_#etHfsYU%pX!H(kI-7KJ*wU1bZe{0Xm6vn_9SpI&kxAy)>NzJ!LA?WtsfEB z@55i)wRbb1P$<~zpGsZdiCL$6k{tTrv99XUR*;dBb&Ks`>NcC+E1PdiI>L#q6l9)6 zwnl26ou5hR3h(RCA{M)QG-+P48`<<4al0oYEXsaWN+_ zOdHT{ieKPW8){D4QKY2Bs&7mopvX8VPobXX^LJRvMNo-EhEQNqMIcrM9RwH7XtU_` zp3j(Gle6qMGii>vW?Cui>lmqrT52wWQT5~F(>9sed=dtD7SI&)w4%$pJv~-vthYhs zsOZh=5TdL`7{Cz~XX>ZjWZUTiXtfFPqC&~)b##-BwPYQa58dsE?oN=DO{>Yla=!$B z|L(P1E6q=J+wS`@B;Eb`aNnShr)Zk&s;!#&oHs8%L~d{(Q39dPq6~t>oBz_F2yKUp z{PokfMFBbz1xIXaZ=v|nR!oR`GuJgY(@%Ks&lGQO?LV5&&H+iwWr}H#BzmKJSSaN1 zzyvdlE29iqQbHwY+CY~VFG~7uZt88MvvZlw@*jdXpYuVdp<_EO7%Ghz z`8x@WYcA#tiqdBf88}#u)XbV;K4(cO8J`_DSXU^z^-j&%=?TUUTWDn>k`qe|XUI(~ z>A!^vS<*Z8BaH3F{3Fuazv#yQ2rd&5=^m>GTkk`#D5(Y;T1G_qe?##zDA{Zpppe&R z?=+)+d@n^$(y_Kz0@I8V-~BS44si3{I>Fy+MS{H#-u1?n(9q&eb`zidaWEzsC6dx_=6reqNS(Nq@cD6YqHTk}}3c7$=Fm1gBc^=z#HodV4Is zl5Djo+Q1w1;1BRFQTg+$08j>h3lRdXs-$?BnGr%q3xjyP$LF<&9Rr*8Y7gVf6SiKp zoNp!+@S#rIHeT8DfzSczE4T3qVks?+KZMq}bV`OVIp}qr4=?Q&bSrppTt|bGuSC>8 zl=4*?BgK$vN5?jy>??df~ z%CNbxGFCL~p{KUuhq8dF}V?)P;owK)ZsRL$MI0jHEy;shDdh?tN}W zMz-w`6YSlwn)`*LoCKYQ)BR9OKrAn$MXuwN)2jZrEG8o}a49lS7%tH_*!L}M9rom5 zV+O!5H#R#9w!<`LmUf+U*;DamMu1H1S45v1$5`(S(a&2&nHg1zr*q+|7uB$w4N*s30!n<#5Zqc6qO`8u{3X5_D?1Tprg5Db$yc3E}P%d_CV zdS;n{aOd3-(D-iU-ILocG&u`GGbD+Bc(lfnZ&4SGN{(sdu|Y?UsNwT#(~>9{2IY&+ zIWPr*wAX8masXdP4jJy9#*L+R$rqJsm?~~xGjx6Crie3(hrJ3vVTyvI%;I^ScG`NX zdMe$X7{%n891}<%_3>Q)ZKc7&Fp3Ci%IXb`zq-WRnq*sFtI2AA!cLy$pc7^U1&(uMd z9xZyFARK{rDoumb7%j3r|CYBb=SBktVNO28&9_Ac+e(%In~|ouc^l--UPzb>7fV^J za>0$(>?BBEhEZX;@h8!qIX{aQD=`p3j9A{>*bSqnTt*{EUHKl~-%~H1*z@x3mHHU$ zrg-{(-HbxlfIZD(|aE$l#nFMCY<}W+9)QcfOpy?Q$z->9gF`#C|ulZfv>JqMm^ud=PA6@c|yI!+|A>#Cnj z38w%MMQ?UG?XnSJ&OuJo=(L zhKtPw|4HN5Y^+{+eAtu4r}(EE&t%|{K*v}W5Ni6~T2t=6T{~FfTn`k-K*x-0(kB0I z*B9$bmgg0%R3yRfj=Le`ty?wi%;koo^c;Zj!JBC}s_zQcv$liCOOz=0Vm7HBUtgZi zvZ#}XQglOqlWEI+4!dEjqhq_s2+<9zi^ zd#;Xf=OX*|@QmnS3*GK-306&|Ueue7^{^DbGz!z$>?{Tz%4hDcSLWq#8#qS3) zLU6`h_i&ADMrI<_wwhK2gm^#$(&nE-8$6eL?k`<=K6~}ozzM4Rq@+B0TpDP$Z8%=` zGF;zE$BhdphX30Mleyk8$_TLTm5&>ds$tOpWKQk!THF{7d?S4bQDEb2V9}XVU){8g zFNN|;path15tFcsRcuK_&57pKjFNT1hzHdodicv|tsy?Vp*h+TrvM2p+boD=ET!-6r{)l_So| znw{mKMr2a`x1Y@MqXXChLvJ`H5L1nfyKQRM&mh4fV=rM$8?TA4`Y+$f z=Fb5Q0)#;m0-R9(o*qRc3$JPxq%T69CTl%q&O;yDiffVLF?GT9no&~d)z(HJf20?G z593I~NW~8>LEscG6UHNNY9d6Gz?mFE+H%>KU5gh~QyXLcO_ZX)@|fwaoA}0aA;}Zp zLR1i4C|O~mNi5E(c8Bot+CL$IP8o}CrTY|>ZP5y!u0IE5aWD{D>UB5Wpn8yCH$17J z-eJ*{NCue60;hp9BeL?IiVFUoK2op`{5XY@g>66KN+mJL{^_H4nodnWb_BNcgl&zu zFnC7&ggE>7`DXaAuAry0{^>P|?N+m9By>AbZ~2?LIr7;{)@~c8Ij8haI2@j?c(#r4 z1q8yC^dc{AZ2$;cT6;YiaA$4Wb!yT!Ufb3ZMK-VPx4{8jt72a20pptGUyj0?9~(ZJ z9J84UI;kC)KM_8eq##d-2?jPjfbp?z!DXH0F<2+W0k>j5?!V-{28|enz6y>7tLa3U zn0nVf(smo*&NPmmjAf+|A8+w%i}1%Wu56A3k+XDjd{wCq8lknyhm9QW2;BNIEl*Sb zaZ(*A7a5zkeDKXu2$pNwN?X4rV93nmEgEg(FmwSH%CTggjhOyyYyNYJk$6ZI{jw{& z!}W!hL0B8_AE%UN?xh=vzPc~?G4r;+f`jj2xz-D*h0)yHk1gI{M(Gh#(xU^T!a?TV zaLJ(TJJB?UM*emw(52d5Fr?rmH!7#e9UT>zYSfnTWGFyOA1{cNw$31jc|Te9Po>j= zwkdxzwI24hT*62maH5Jl6I;TQ!u&seOy6=b)+z0_Xi5jlN}>n2wn)WdU+i(3>TB71 zKq^|kLD!$eJl)`RsIS?otnm?IR$eUlFu&4QM%0*mWS??y;vo-a>s(nC-@cds5|?&J zo6n3@+lYF`e@zu!9bregaowGZF~lGgoRzb#H%39ch^LW2Cy#}T%H)gu(yrSJUYGHs zYn3m;4);`~5aGc+X533p50*0NINed31Zr_B(ZRA!w}dU}<=_<7v4?KS)zh`-yHOJ& z6~CKV@BiTVL_g1Dmq170eKQmD$k2!`-vU1myh(efY(ekdYjl!l`r=!F=@})2jaMHh zUylnfjcxrpBq^z!*AOPryT@x@|NHMv4%#MNe{vULv_p%Feqh5q;D|-b>l?SSx5*?C z7*Hr3i}OMyRcWnZYXBLXC%V2)HPDjUpE6v{g;xbby=TBZ$r(?)V(!NK$DC!wpGz`3m~$aBi4JgouzuvsUYdZNUlmn?E=Z2H{&Iq?IbgJpzCz=bly zZfU8&qHOJ;;;ixblxQbf9p?N?ffB9fqIyhmyvwRK*8g%C!*ncKN z8ftnt4OF|`smd?-o5TRd>-h9-wHYc(xrgTOdxZ%j$A3LSj)A1wFTi^@1d_mCGyUZ= z6x;U|$BN4fkxHu)z$BI8GSxfcsGGI>%3?xFcIQ$MVXNpRs_0cI(pIBHmJwA1(Qwf{ z5yE-SYUt8reQ+Tik*nJuX#YwvI9_Xae7T3WH~AIezzniAO^rK41_%rW$uGk(Y5ocr zc1iqR<_%z40)H}RNufx$_Qy~;iP*%c+OM~w?>$)4^BBKGV|ckM@uRxuXt1hE` zvIkewI6tSN?;FUN4h=`HoyJghB>BNDPTTAqz1rVmA(z-kdjH2)e) zOn`3`t&?2uNdxtb2KEM%vzD3$VQrDRr{WBs`sesJO;(3VLq2X0tc)l#2It-d>YTz3 z4 zHQ5X*D<8CLUm5gCS02bQgp;siQsNQ|CPtgLTsMT-%IV3@ZQMEX#yqRe+MYe$a-{6i zIAcAP#)6sNj@~ICq0O;gr-y+}yd1yfD~+c(z>G3$lD|XXdktt#>G0>!>-CDZ~->o&CEn7O+FUT_A5Vo%^)!d zG^=8^D3cSLW2}Bx1&Dm3`4!{e7zapX4dV64) zZNy2*2M5dWxrNKL^s?bdhl@+2-vU(Pd64pe70J8)pb)KjDRhZ}DyP+O#-2iN@ro29 z>3Jxp@`#08J3J73K0=KfN1cB{NlTyImZBMcmP0XAhx4JFG>Qy<3I`EdrX5=xJYnl>@l=MTZz~ZuEv4 zEp^Rij^U6j{kp%mIc&c;I(6l=WOD>xMw)>aP8O(2vNltAUq{Hra^YYWx@;wONq3t- z$}W2C7L6}W&CW{_i<-b&haPl1@cN@EDFL};Y?62lzbq18cO8&_vM%UqUtzX?hgeJ_ zs(*4};%+eIFxStFEpCdf=%Lm=Y(FpH^etQEGN9X0q&4X7;K$M9OSMhr!selvy>>>| zH`1^nL#J!k?a(u{&V^G9(&t?>>$B z%dqm10|y4@L6eNiX;tE`f-1l$b;E^xkwJ1f2)prTR)e%fI(dX>@W%>Idq(4%@$E}J zbhf=pJ+CKmwXX6@8M)j)A;Xti_3LaJ$(q8ZtA)>SG#$&!%u|CdkXlcrbe%cBIvt0X zy4alar@C)`KMw!Ww_(oEi}retK%woYZ*ulhD9 zBCU>R-9=|}Cf$DbeK{1(G4$lZN$2GNjEw!WDmCL;d7H}#q?14vEBA+TkvXH1AICF* zyFT-Ejb0Z*ey1Hsa_@lQ^63|>VSXr^05Utx;S$CTR(Y&Z6oU=>q?}fB&eT^SJ$+^N z;BpbHsptPIo|9^Lm-F4?pZNMjXPD_boFOy*q2%r-srx2ZVp;&^zCE=E)RS`E@}bn? z+sVgj@;Rjzl!O+RJvNCz%_fnUPtv%gcemzrs05RcZ+})%O2_XWH4xpmAiCs42K13P zYM$NPg-xho5h1)BQ)WLVkhP;qR2h@ssv%~m!=gEGSnU>Tne-K#)bhKn_(65BL~Q{luk!izP|}iKC`HO| zLO=&1RBiTq%U2so>R1iyCUzLK zPeQ-(bdNnUKV*b<*@VJHFilMhe|F5)i{U2Jzo?joh3Ue|x{JAK`rvu$i);R=a4g=nr>Oln`~GIfrw;?7>>kUi#g8Vb%#$7wdKvq47Y#RFdfH9x)It z^TYJVu^t%HaXdrufDw9Dd|&j6>)>|>&&b7E=lVZN4Pd1By z*XW>+QVijbrLD$vgI6j1sObHFU84al2o&T&bRq%66D45Vio?;M|$Cj`RfzXWqTGeGL%TC?Lby9`Qv7?-&(KJr8~dw13=`cpQ9KuRi^NHHiR0oF6# z_=l6|kFIjsdi`l%wk#CbQ9cW0oVFpIo3R1%8onVbd^WUy%J_{9HLq(gaZ)+zW=JmM zU*#q{ZeXN`zMZe(bo&G$m8G+!w)GOe`#!z4&^XjGfQky-K3y z)T^S8fP^%!X%m=zUMs^SaLF`(*>m<#sv(k@Nhl>BdmY164f=XNW>;7E1~wqyeyh8` z2M6Gl(flvUH86G(>Uoq^%SsQI>gsB|=p`5W??C@o+tNlCAr6%&abH49e^x1>uAH}^b8by zi29t#()2B77|nV60RP=hY6xA_#pPzQ5dyX;q7xeZ4r)aYXEd+=vN_KC!2?V?Z>8B# z&sufoFQ@qZ5Vv-Oa!0AdkRjD5Zyybf*3TkrFHreh6Dpq4+N*P?3>M{uI*|TNk*24Y zOUzXrId$`KPc0koig{==&xf89X`9>v2NROnHT!Oxd+bx`d5SB^^HT39$DR6r7Z=y*TxD=aR4M|@{ zxs%`SD3L#9c<4Y??m;mQ<{3e%Fv=Q?{x@Slr~lTMMyj25(da`4o{N8=%-67#_+gdl zTY0&=kap)sR4Gt#ZdM|NsAKAba$*Nw3*Zb{l+K=sP1L%dopdH__PZu>0cO)4OUm z*z#{b=EW-i4+WgbKbel6_2g?51JI3ZqtDq`qO$;;-<&Ap!66n+=~7$`SKd9Qafe&M zUxQbKCohDjx!FZt?bia`$|kiL0qc+8wHVVM(%q8{GsuPsq>(;pMcN?f1S>p=L`jKB zU=npk<_NB&78e#dV9r2{aW#t(S7|{+2jNSc|LpWztv)8nTq5Im5nZfFYsT? zM=_{`Jqy;AG?E8H%ZqOJm>T~JgYBJNW>Ni(l$$NPMowL@jmc*6#f*Dcb9|nzgsLXs zC#xbSQTx;|d3Osoyr@!uUkrgltqQHY0@PsCG4pmz`#F#jeT74t<8`_^=d{HEvwArB z>Ul;KFphYFkF&_f#T%7ZOrnTI8HMiw6Pj)R0`jln-_&0SSuy|42Aueh;rQ#we6u;tkIpM^d2R;^k%G!nAGqLy{ZpctI8v6(YO1{f>^ zQ09gFr%){~Qj8flq0@*>&TQRW6(4DCgihW#e9zbJ;mTiz#2wz3$joVmZbtMPx8 z(y~Zo{i8`uMB2z;Q@^hNq}g#rhb)mDY(MC}v)V#IhJ4aVTh|dz5>xEygH}VWDcK7# z7>n1CI-$b;0otR@D?b*gj+Pi-4)T>Br&ib(fg$jw_Sj8eFB@~t;n~Ei@!OZdRPSJX zFzwZ98u%91=2|Rc25};(N_+c*PLUr2+F|MZu4i zzG6C6tyn7|2u^xbYv;b>cN@cl39*0UCru%b#ytVk5HiCTA^xF;>pPQ*G={Ob!e4#^ zyJ%o+8=pWoh{+%!pSWh=xan8Bj08Pa!-sCVev-%Zk^)A$-e9!>I)5Wy#tY@?7lv~m zC*j!kZvmm;e)ea-qLA+FY~4OaGNtyXNA}*#d31Csot)8;Nqp679HC^`+-#|{^{yXF z$wR-{tldpf3I+?Qi!KR^d%I5lE#djgHCzpQ_3q)vok$|00?l}Qb>iZU4G}u}>yw@k zG#Sza&IaH7+nH1?P&h+cdKMV}v{83)y}C|nnpfXn#1sXFy{QSuskisq410tLd>hV# zRsZA1l8R@eb%1~0@VQILK*ERKxZK!f+L!&9@)#+&yP;XT(+0a22KyHz1uY1mO3h)+ zME_s=Iqo=hy#`-1DNc9(MW-xCBJ(Hz9dX~X`fC+y%klM7y{Lh~i?G6W@B0bDNtL^~ z42SpfZH&Kr!nY(tt#4pTvk~7OZ+?BWATLiz`55G+SYoGbtFWfjvZUL2X&G(|HA?J9 zqh=DS@~wg!4+1WP*;xQ;@7w_h1df6X8PaOjzmqRg3efjCavC8kvq98sPP>To+jUDR z)x@zENX6w!7w)FjFV?H^^~3c|7ne?FQsjo2ufph6WJn=0G~n*soo$$F#S@|-)~I^| z^!sN4t;?im1)rc-3&yiw1xueC#1nKko5M%Kmn{ec{M}zU!wr~zq{1o zH@oyyAB^_>zULu@?J%6JLZ;dw})p538kgOmuA9_5}pxmaznV;S&)g2jw6# z@umORr>9dQ-gvz4YE1y=Q6@!w@?)XYx$JJs^o=CCushli7r>j^hu`Y8_H+r=4nL@` zT0y$u9ga)Nqt;=wE^n5Ss}^Xpf+XW_Ypf@n<4c+n(^%*ikFKH`Pt z^?rSUf+0q2vYH(Go}HMTg1JLxCo>XPq1t`NfABQ|G5W>C7xGw!`58IN6vKNjkV-ew9P9S0XHny?K=R;$-lvbm4?ae`;U>G>lzw}t z*;2lNhKbT+U{z?A5_Dq{YF<2zdsMtHhXS7P{?Vwrb zxWcN?NddOQenb%m;_E84ys<=$JV826usRqcE<<+I-kw}rir)90|gJG6E9E6+$Mf@GdDN6_z5^^~xf z+Gf4XJDHEFc_;hsEoRt@sKIeQF>Tu|X)owl0`@jvFzNI}no$Zn#IZG}TtnmQmzT5r z>jFH*6MX80g=%HC=KOiP{#CFu`w#i)kC>tc62E(vX2n`E#|d7^uj|cEX8)6hI+_@I zi68!1SF>IXGoev_d?}K=n2+|Et41M#X5rX)+y@tD;wtJ`6uHo<{=>;wzrb732M@Ap z(7hf$_*tEAR*P?{{yDtl7~hls%+FY}uH2$SsNyz?Wq=dgu$AfMEH@!3+vd52(Mv12 zZ(cg4-B63d^@#+QrkOx{>DsNC>CBPg#$)IQ-#yt7zBpIDZxwv6mlE>arwdn}>8H1Q zzep{&t2G!MF;FchzUH+Mh>uF%+Ksj;|~!|EHPzll@PKUYX}K5 z*^M=N-lNa&JnMPR`Tp^o^W4Wd-MHWTb-k|F^}24;uuCxNsd8#pvZNTIkC2#?Gw~cx zk({Vy(J@6S)P!6POi_qM_8W|(G0gV9_xxh~MaTXRbCmX5M<%%h+Vfw&_saj_8 zt;|A^98-hy-fP@)bX$e6y6RiDOsvatZ{?Mo_$xL#5p*?~Wh%u5@9mt6I{NUahl}$b zf4Dv~6XKEWVtjrnr4~Vx$qt=aCu2H3bMiGsM0n$#TsB*~>sVu}NA54^yJcl`6w>4R zTk-@j!AqyjM~9hu{d2NspvRFUmtRQ-sbm*-LeZiE9=}ng@6q3&t5&~I?lntrjFLoS zGwn5P#*Bq7B!EAd6k`+;JXa+QMTMeFcpMpHDya7yV#Q)E&Da1_+FP%B7tYiY=RXxL^gew0WGs27 zp}>nNraN&UtD>s(_8ZQ9=5Btg0-Nis)l4)vc(=#(>oVpQMy5jc*C(bs>)_p{Iq)ox z@%yY%vPrg6FTC5GS=Nz)W{-z0B5a&hO?6wH$g+xvfuxC@hdCx&DF3>m;wN&AeFVkE zzVVxU-Rt?;B4wG-PhQn-+0e7XY)K_or<^UV;msCSqFht`C??P~d* z<2Dp0QJmE{1X6{4;f+5FftbLgrsgq=8Bx#XB<8e5W4?Z&1hIx73S5dzvL+3LnWU=f zTL%rS;c~lBW`svS@RboW*eh>`QnHx^IkF9|rEpqjT-YL98eK+O@i%0w!Dv_SgcKz? zTzA(Rd0$G8xP_dLMpW%T>a|EC*viVn`ZA|5EsomSEu4$oBF++~$>g+l{x?zKi_oRB zuLX^LeCl|c)+Aw`lvj2#h{}WMNBz##zc@58M(g%*$*Aw!(XvoO6#v!2l008>L%|s) zy|0PHWGPAD4zujZ6K)(3ND19hD_~?7+nhB`mwy@cGLYxh4DH_X>2(r+oHY6(k@#Gi zq^D6;BRA-aWM!+a5X3bx^6lGKd08lXNxf~x1a(8L1xl8TXM%605I z9Bb8~rGw%f+SiXBe_ZG@FtBN*^HgCyjbY7^0Rmw@d$>G_0vl&UTgGdn z@iQ0CrT{DY`XZ>tRLI@M%>(W|TX#0L%f%USH^=*%l-2cGS*e#p2JcEUjxE3t=S{bQ z_Pal_a0@(@ux|dCqPaLzW6ZsUcBDCqciu1Rr*^AKz~biFn6F$z z`m_$SuDD$}y;&V5^8y~FQ^qK=fC;?$mom8@e_}78ogciTaYa@_Ecf**s0aU>n~Q8o zpKs_jGpfk532CWXb%(g$|oI72t-8UHwxgJ;PmGRm|G{l zA#ZJ#5Y(6#+*&BvlVZ%jl#X9b=cZxH5_a*z6fYPWATqyo}H9vRC80Wy)Ur2{S;(7$Y7v*s}su zY|o0xi9 z%t?rH?NurzaK6$Mj~h-f9F}5Y7j*F#DVuo}VN?;XvKiD7q1_2>fJ6CR=Mxk)l98G& zs&k|X6GcSP!it*tD|MIF-70tTwC$nwZWs)s^)xac zNpdug5Q{8#)RBzoMpuWRO}S>KO+a+Fy6luPRbrQhdGr{AiUVX4Qo?$;BEhc}^pw2h zOL@3F+iIMVUX+GP$LN6Y`7-NXnXg+SNn|Z1%M&B!b2{gV@R|&4PH;19C)R#SgewhjU2--P|8BDiHQJ$I#t#$-yE+C`v_Kuq?2joYV=c3#A2Yv`i*lJTD|NHdR#Ne(3UM3j^! zw|K3w!5tK(ERDVsiCGUNls}w-Kn7n+GOkQKn#&+i1}!Ns#miRC}F`BW#M^zJ>aId-#m5smHHz4}sRs=Z66l?18frrb&7hY?(soUK)Kh zUoT>5QyO;rViOFPOjOp9zoJ~q<=(`;uKF_^EwD80p+a*fd#d^KCf=SnTJTLBvdJ3P&ZUT)a~t<^E+TR=NHU_)jiLrI#_!bNOu;Yij1O43;rZ&Z8iB?-eKH&>A|*YV_K9i9$@EAGxW z-CTm5%;ZiQMTzEEQ$4uZUa#v(^!j@w*5JBkf>4JM?**9|?|U?jc6N3zGUt>IW|TI* zT#ZX#t!QXa4E}zM+srq{>`}o^=!?mwshjMioLQlc1RECZe4|p4<7W5GFDboZes8Qz z5Xm$kHJ`4RFz=s=eA8D^o={P)g6ElJFb0`UM%i`uT|$tZQL;z7vom7vcVp)3t>&4B zI}MZ5uI0wHeK!z`009yg2uHoyKNOR^xeji z&^2S+4q$#f9mZd85zoJ?65$*alyV?Wo_9mlKp^*HT=!_iP93eE750!>Twe~^s-?8t zZm3uoI`eGKwDfCdr|@pNkxNK!LRVLpde%C7<0|_u(R!aKuIR`T;pphtm8)GM(faUW zTq|E^8A&9$Ll}3;4I#~8BEd%d^~)Wr^O4PorR5VsBy%o<^57Eu^>orjea*he+iw+! zg2y|JEzAZU-PD6F_*m=&Sy2M&iL=>yz(=zgp=w=>Wp|r#_Q#3OZ{`$sB%|gY z(bT=ayt2i{eALQrMy3BjS;d&*cS$-`2BqIC7=-9W+TzAW>!zkAHgj()p`L| zJL<^dd<6T0cwU9=?@tRAlDg*(L1oI*%o6WzzSsvkgrq-X)>voDEOkRj`F443{o_1r z0kBYf%0GFEK*qg7js^T)ZKW)Fgj(Xr0;i6~L`To7a_`3TE+{NN zVvu_Yq6`F?XR#%jge`{OCccaoSs2C{tLUsZYXWK`a zdk(cTn@k4V(E|Wgf5aXTPRNJwvgrT}@mP3M#~G)qv85K{ux5wj01M3&+@!uM47Iq& zxK9j@j1-SOsMUKn?rSNoa||@=IeO>v@dP^I(2Bd4Rh`VZ!NSXJ=RblF2Z1DTjK$lr zVnp6P!&z=IRsQS-4azMBf6;>6B+O*(sWLjX5YV{_;c%pPn}QNM40|&5tEn;W5hk$V z6OT@trsjy6COwa>y1LgRe>UeWC3Uc8-lr3;*$Bb*?vVXBv6XiEf9fVT61w-3)>}ya7hFH z^Cm*J9TIdkR@rSvdZGZRGPfk;5q;Q`au*rIC239sGP0$VE_|`SBX>fKzJ6yapdnU& zwpzx`%^BgvgaM3cTj}dGSh$HWO3H}Ar{bY!K)3WbAj$ZSqGGD{Bif9#c^@axi3S_aVWK}69%_b1vQ z5H?FKwL8WDqs7my03#^5+|97@b&7S79ov3eWwe6lQ46{z$ps_F1moxG&6zUav11Bf zlvx~2H}9hWRU%VP466#oNl^;W(IQ+@GCVq#dt$P(DMKIIQm0><*ZUTEHwPR!wm4l$ z?*caNn4QC-oK)$Ok1l81mm7CHIw=ycB4=0Jn}e$2oXqC{99?aQ0OCsz1xvF#8vHu4 z^5%xbcU7fIsICtlowhaajq&!QHOcvdQIFJpXuFX*kJKe0jUrmNM_N^hMzE zHo3EUdlyorz9vfaoOjy}_|W_XuUqeU6RS3CpAfq@+ zE334${vndAv-a~8$V5(4Q`Io0>64Go#xO~yu1iXIE_rUzjEPC~unw*0T0LtyL(=t? z@PX$T4SjKkp$S!dtjPEGn&N?}%POYyR}_mK5xv|7^5aUy(wOnp@R;-KXhM*6rI2dd z^L!viu*=#-^)>>%uj-umdFoJN&?1)tV^awZBi^bi7gK(&!HTvgEpr$QF<54^=5_$^ zF-gl+wBFsMx`dnOQN~+(dOyb_%1)LXUo8iD@r^}}_$T*6RIB41DlMoq`{wH!;J796 zAkk4Fnu@Vm46XuQA>1_^K7yRz0UB}?`}Lnf%K^RR`BCge#-*ZU|L?X=ED3guOGhDL zXAZTZ546REOpu<4znwj0ol84sekyw4B{ zl#eHTBGaXTkG0XMVz;0Rult8739({^Q1NigZ$2weoH6Y4*#WfAMcb#6pm{bDd_EJv z0v^LwBWwPW=d*6I)>eHLx&7*>Ry~T-8+woVizuPeq`QSW^wu5g$LRH$sze-I zy!CChP}C^lEd19%BnHxYcB4yQF1%?cwpd7hy@YK?x0Aa6sjBvHc}(6!X(7LirV1uX zqUHRGe(f}zjqAL@_)-d`K9E5xx)1<5f>^wuCW`luJ21wIaB8|j2M0gCbm_e0_CoRY zsl3&Z%2>T062nY=I6YPw9rxsOXDkV}-gi|Mc=WzrLfKyt9g_2|jR>kK*l%@Ws+LDC z`ugk?d`-Xstb=4zewl!-A)XcNzhb}|!b_f5Wp=p)b(v_$zcCv~I>bGooiD)EZRC3L zgKg@)?2Hr^W=jhewKO)?RXJ%b-?DYpj0XA`8D7hC9!_#$!R$4x<);cAx5&a*SM>Md zI?cm{Cibjj3CzwI10uW^O+GG0e}fG_km8vb9btjs0>g`V{hI^d1T_wQfi%wnne?Ig zbnJ+>q4W&3`4g_im%v_rYLJE>Wt0_twEW93wh(#y%zD6OU?fd2vHk?&7`{ zikEHdI!L=9W&(#N2J6e91)*Y4x1wORCaPUhAEjjQMJ)08Kj1_JIo&3r^joE>bCrr0 z$qA`~#Ghqz$7!Dl@#W&*#1gCAN+rX!5&GPzy&e6MdKZKEY;E}L@G5vdrTBK3sc5$e zDhQXi3fD`f)YpH?}6`1I~pRjdW5N&~eDU1O)0bv%TE&!5!2PC>9P2#fBQ|-PS z$a#w(7|wL{vZL!6bHDX?Ol z-tzi9V(qoh9G2*SY@_Ol-_ziC<&KBj1EDVpYyjZ+)giAoRjF$|pP<1=U8@8G6(S%# zkXF2XN&z zAv5g1i(-iwcttD7bW0LlCbTeo`IJOP*$qZXVg_xQK{)lf7ymwoT#9rdeYxftVy7#+yzm8qe3e=UU;bd$0kY&mNwcfts3 zzz4bSe32N_s^#ItUAYTDXB3NuW|?{`(03;qml{hBAjwgktj$fj zjiB=$_}81%G5BuMtCn$n6=3VSp=i?g1#Z}S?M#%0s+SNnqT`E6i2 zf!djFk_=?=uoveJ_pDqk$z>iyU5Zn#ti08LKkkjKWIVY>j9wLCs5;`Wc8v2Daw)b0 zAMKrWk-GR3hbIdDa{59n`1(dDQ>g|5SqGj@UFA|*F__JL*iAAOw!5HTq|ntaAHi)H z(Fsog{~lf+)(#m2^0fQ*X2N{vI;8{r|h;=*y>`m@q$v9>0?vP(MKu^Ssn8zaG8Ru=hZ#uWWy`Y+)#L zf5;Mkeo(gQCU0jPx3XXj4y&}AtCYZ7YTanXy+5x3etT2tu9)Jq%1XXyuvXyK5}J1$ z=uPd8Z>d=-r?)2jI#=MYTfsBzlkvRui%5@t(SvDG{QD~jJ8(@}!ke~yM_(LT9v6C`x$MRXj*GvVn;29fBO2`pES zd@n30D7e(pdb@6lK05Bhh2A&d+z+q+?7T)wMIg=^AQ|l$o7j}n>cb&wE#UMJxU_CR zoQ1o#w$_sRotf1ZVaUbVSu^2c%-zGENK@0qo6o-Ck(Mv8M@Hk$)7a@Qwh<<_wyXo9tEC`j%T3x=)T$yP_kzc~WWnal=ZusL<{<0Y%=J6%E}9 z(xC}UlFnLP{pbIxm@caybLQ^k=zgtw)Y-3-gE2+)sFTKrGwDy~9ubxs_I^B?sEkz# z*qAquLo7Wyl1=RL3rfsWeRg(Xe>t=flt;I}{qOQ<3J&sqMKVy{?(FI83`uWY`-W5L z?2Si0umW+xs5md#DnQG6Z*&Ocfyu$Rre_Qtl|uXE#se%}ld~-vQofriqe#GEC=z^EjubkJ@*C&hqSre^x>gE~;YAsu?3|CMaK`^cxy93RwkQf6it_2qMd6w3t5 z>Wr5CTjjv%Ij}^?%E_TxIsL##@CI+yPQri&VMKwHi<#ISt_SgVJrWUN%w~5Qmfc)k zwJqO0K%#CyJU>3#=+)rUdiTw%!zB((7l9(e67ra>{B2<5qha~3n}>6hbNhFWM?^hO zyqgTERrB9ZH{PK2$mGo(yEekS*ub==cVS4=x&)?FS*&-EjLk+`}zON zesUZUEq~HN)n!5{rlF|z*sRuA&VH&sP0i`Ee*d^eX*X@Cgz+6!)-e`f(lX+IWxgR( zVWyjq*fDF8&kr$+MY9~5nCV{=ivhIPq-S*o^f-j@%LpOA3(-L!5Mp$yjEsx{D<4l% z>X(D{sk_s#hAvB|#9e~G-S5#{N^;0pufccv9RCpR#@8bLGafPu6K_@f`9gVSoBl)U z1)9ebYf10YH+O(eJstp{Wq5;Jzo6&Qmo~9c(62bz-=91AXMfVP20q9TvT`hVIdHPy za&mXDz?vJvRy{gnC(V(>+j%IwG1l_<`T9!j?*jGNJ0?!>R_3i|G(EUf8GbQq8n#w` zH#2YfZn?#oAC{9Q{p=GfC$#z=A3JH`6%3IcagB|;_Mi}Jg2fPBor1a|Efa6M50_t& zged2BNh#tFE$O=o@$0RZJ6C$7O;&96ZXK#lp4bjE&*?a00krEwKFA<}$i=Oi6TiE+ zCn+yqQc+PMey=?%WYAacbN;n`4nhCNOh!VNC_isEntO!q zjm_>IxWoZ*`y_XeLBdArP@Q9S2QaPLsR(vn7aJjYa{#Igq1&qAO*Y7pN{4ElV_|$Q zKmYpEf49|gmFd;@_y^;dni`qlCGV{Mjg5^@r9nVA4BhezEmR2Hyq`Azy}Oaa5UiQ7k-1e@m=hxj$wsHj9PaDVzx!%O zQ|hWnTlJFNqs}m4lY%^H@Dl;JieiZ5Ns?g@w}a1zs@rpd}jdg(d|E>$@``| z^kChhF=U?>b`~P&Rx>kmnkv-;NcpiKsa)dRa_Xp-k_D+&)W(G3I9w5{B ze<72C|6~4DdHsZYBk|YMI!P0O%fy*DCERl8pS^(9W@d-ff)giCqNoa(X~?W872;uH zkdi3zn7C`dhAUL5z)mrThOpq9B!c zQLk*h-ltS%NxlBF$n^h!Nq8CtPK7i7kdEbOHB`*^ThemyUy|eZ6?Wp^@Xjg9n!dxBcPH(B9q-vte~N@*ir?c#2-=qT1Ar9T^{8a-(eAs@z1!nn9m1rgB@+R3AN}ob*<~RVEXG)%P~ngM zpAG$4=~zV|a%v|4f&OqWd*ecX&xl*ie0h0w_pQ>Nc>ON1%tW5cR5oJ}sGpQ6hS2vp zY``AM@rShYey9JBml_f`Z)Q`AJcEE}g%%qNEP&S3+F!NXkO+XrB2{v8WWZv%+dSzgg|j|wT{3YMZ1)VYjSxx7+b zjntn{amq9-HZ~GT?uY6tnzZ!CVRfdcAB{Qf;V z1pKfAg?N-%2xuA!Xuk%)SKO@D*}zdl&$NPfAhQh5(d_=XrgTHA-Cpd>6sddvTX%+6 z47Dz@yRP)dTdVJy;_7KY0z}=reCK<3>dA=Z1m1DK$>3dpbt8$(&m7G4WT+Fhcy^et zxkb!Kk2Z!jK4{o+1l0%Dh@7RR2f;fX)`Ni)>s(rW8GaMRFJHcd#_lan9w2AEWa#PW z=4<9MsbxNgF}!a70GKy5JR?H)DG7O?I=E29gfEa$T9l%*>m8-))pUZHL}2ew-4~#1C?XPjyVPW63^QC`l*rI z|NGBPuYVi6=q~EGw@#QOghDX2*tNh#nliKusPyH zSFJ8RE}&+@I@OQO&lXmQ&ls zvZUSj>({R+AQ(f!HV^SBOme1IgB!6?(c5bX7@vN^Mq1-e+L^3SvgM?Jiy{XX7njw> zc0fH2hZ6xXTNb?N28f2Opr=0EZwQB|6e34=ugb^i4T3H`*$(s464ubQYoTw7i;Kk- zHxI+GgtT;#^>(|W{^8Xk{5jauA9*f(k#+$vuO!|Dto1%-0+_w4kB{-4bJzdYvKaoXFi=^zX2QX& z^?`bu{aPY9^V?buD98a0*ta%?SPn?e+o2zy*y-|fwY9a)nQ+PqHlZHVQr-GJ2QgNy zz6!h6`BrK;(E4k|s#Ld$iH#+Zuw_Xz(tvCN%DFq%xNoV|=hv{#xj0e3S2cqpkOUMq zhwZ3`Lj~LiI0z%bD2#e{{DW_uV0B0lRDdQ>sBwS=~EpiZ-?0O0(pX}g%MQ`<7br& z^{m6HUtFS!a@6bpAapRl1aBv>M({M<^Z`t1U2!r~A@EX{-x~$0%r3%lcr1UL4%`1% qH=`H-2X^%TfaOrf`_F`!fj=j^n=|IesxhyrjkVP8so_-aKlyLe11l#0 literal 614623 zcmd3O1y>wvvo#Xjg9UeYNYH^f1Pc-fkl@Y)g1cLg5S-v4XmEFT2`-c1?lAZ;*x-Ca z&VAqe{e;_Vu^L!(_tRCo>Z!e}hVXZ)3V7I**hol7c#3aiHIR_dJdu!4%`njszgbNF zd4f2hn<*>EBF>28TBbDsal&$Zqw9i%MDz9O3psWh_7Mq*9!XJF`n_lR;j&efA;b@I z4LwWW&TO4B0y!13JH>G;8!-=r)2NLsTbX|<{Uj!oPir!PzohqKk>-^NeQi@?c*!np%G_#~b%H!2ZFF|F%VfDa7YPtycEWk4AdhmF; z)i`r=@PAEC=h$1Z<~A;g{*I*Y(QBOn-igemmI=77v~(Uv+y?uk(jB^TKTO?6@}K7t zHRn+|qd{lN^;0O!rO@iKe2$xgJH3|lZiyk(5K`{g34FTDNga&u8ldQ&>i z^E(Q&{JmW%3t#nrf27%^qNcukb|NCDe%A=e^f|7NG?bB*tq|(%O(xf>9N3}~I45$Q z;_t_BHlUm=iLJGRm?|YW)E-ck_r?4VK4-sD)yM_A9P=WXkG7IabU3Vb;D#Yjyj%uk zxZvaC+au2)e^|rm%F?wlrl6+wU0?BmF27{!mXifpQF@n*M%wtNRNbu-@!UT9b6TJ1 z<5z-e$Jrk096&d{5=-Rj`=jaB_~hE5Lg(;*jO^^!#GrjbX>O0=I38_aF{f^lAYd`= zI^Jvv&J;oRS@navk;*x>x*U&m^>z(J3EVc?EndQwTRk4{oFP}|g}`N^0oXFRgs`BX z^Wv{x&&DiV#GPbiK|hzyK(gYO6hX)OST-y@E2tssUmkVIQ`G<+*p;prSVW7NS6Hod zO-pcwe?1q+)uN8m_6kBhvoIpbNQEj2$)QMJkbJ^v*D7j2r`y3y0Nu0h0QYgfelfU+ zexto%vj=o#OCv$Qe8$3jbk*o^uH-}M?*9J9n8<3nG&K=Xs%7Z-9dA(P*s~8Z0}&z| zynm7hZYYu-n3yl5*o7y{*l_)7$15t;cpNSMT|;xs(l<1twG^(>wvi|-^6Wqk$<34mgd_rY7U3xx$q}nIz5kNS6a<+ZcHJkaUWs^!E{2G!4#JbyK!4IH# zhsbCk4yU6uEN^XMcbW_>CwPjIB-@)!iNUrD++)uS$;Igah2(~x4l{4=k*k(PwNZcr z+dNg4#dMNm9qYE$Oc8^4){Y9mx#;v0^piIHpku-`2*_)6jo6%|PhE=~!_(HMWlkNj zz?~61-8D2Lf9U2|im3f(Ql$Rxx!OL}0C2$it(2U4h_(+e5s@xqgOpy+S&C`U=WMO4 zd652E(zvPe@^3wS>L0OTjM{d)#e!qu1aash>q*`M{%rlvOk5&f{(Qr>tLGNqopC#V z_M4$jLsgb+__rpl;=O9rvdUaDC0~)ry`t@0=oAw;C4oF`FzrB0z>8SF^GxXSAx$Wf zMjrt@sV5hjj{-#@NK*dQn~M#0rR^SfCq9yHkl`~I&!@Gi#Fb0y^q*PG53_wTu@$j; zm$qoa=mOe>8O7f?5V@_qCcIeF1`{UAcjb$&u1yFY$@A%c#Itgt2#&oF5<7J%-yx0j z5&_s=Uth7oN*mQnD(H8`<$qrbUy7`NoR7ZA`@Isr!t`R=fP36AJ$}3mms8?lWE6Bb zG{exrZS{akq1|EXbRYozk({DtyFK=Cz`y{k*NW8zzh8E7l0vKyCEoh;5Nec2tC<#I zp3bJZi{vtcGPr?F{LS4e!+vvu^VgrkC)nGWn^#GSch)NJtchmE)u}}PDTgAdL&Sq( zB=&SJ1jd0_ER2EyMh5q$wgcs|H1~-qovOty=Rq61pRpFFZU{7v#KMhRqZ>R#7BhEs zLVh9A?%94Q_elrD2B`xYuqi?y3ZX?AmsdaEqBa zYoY1H%WOxAFSD7OAV%#Wiiizfr6isZL*hgXeAA%2_}+WH;Qf+WtWs3ZBp%%3Hmmg- z1k@})DUkOssub5mIL-2=ROZf)gf^pfJQcjH=l!S0?w4IT=s#s z2YiaUK3-c}J5VRNUqwRfSX^sda_YXmvlWKqZM(FTl$--jcQVPqLx}9>IvM29l%0&B&-M=X*xdT}^O3>iFN>ndD7bd zy&fzHUn(U}6A}bHt=xZK)^cPIAFef-{7H<@O>!EP}Yv09hb)mZLn6T=_QNJ_@R?pBtzsa`-1t_IWrV3;2F~MR&J`fwwlkdR4brYYEOm2@p@2 zN5o$^m*^YH%G_5CF%`oHB;PsT>Y4^HtrJ z?mxm$Vm=cScPbZ~L5{}R;^&&W6e2SDg!TbVZ>5w2Jx1I5k>D-$$g); z+3DdxUIJg-4EVQbbL!2YU7fskLb|@HuRTk2LL*`9lC&W5She<>u~@Im(gAXbn|Q~F zSmGIB;%UzJDeX>wHhBLbY$(2<%bRkE(14el9PhY1R=0FnIUqrv8~o5{GSQkeq>gVT zF^?C+XJtgW$mNM}k5c9&8Eq5siF@y;8>Ki+tB9zJxc(N<2IUA?_}B>3B>+Y_?vCa}6#?nk}i>OUliUU3E7c$*thE z*?tZ;d-~lDH)M?!olB|sZyz_eQq?lY+6L1sGMdWp`>p_a#Bo&19qYKFa>@0B1>%+f zpdEZYUI056PgPEKt0jjF4}lsNQ_m)(6*PQjL`7J4kkM2XG9-xmwjd_91K#uAzqm?- z*u9K#S5!QA%((6dIpaTY!L!@UmTpb#|0?F6GwRQCimaf0{Aei z`v-o^AA#)Kam`T$sm7gZnqfW?En!Pv-zeK4@|veVDI!ou9OZCQqMV?cOmi6yRX$i` zRGIS+_;CML^nbcpgJyJpC|Gr7x6MiH7PV7+d-RSD3PUfh1j_)iD24d4KHeH{oR1myF8O zB^ievZY@D@q2=Q)`BI~KguH-UVi~L5aty>$Cm~)ea)|L1z^bM?VNct)K<;_7s>o=v72!lYAKxbN z#{p-oH-A0vN9?!;@KH02Z&tU0Cx&Y;`K$hF6BR4u?*5_H@WP@B(IDg02b>cWnawL2 z=3Iva@<+Z_Ivq;Bl}0H&#-ck~b#O-FEh2@E8#<}Be)#ETAv83X7Sgv`bI{EN)>YA` z-_!8#m>E`rz*`W0c<>%iQ&aiejkV!P<-^>6G(M32=lc6vNH#hS|44 zEKY8cCp&jvo@jPlWOXFXKL~Pc|IUY2N@+$)Wku=Dcv<21{sg(b_iV5^&YEJGtIKG# zH)3zcO%C<$OvFs#Fp9Sq;tH@FOrXvXv7|TRjA96ESdum@j?*t4C$rcQJCp7mh1GaVye;ISk|J>?+)mG(a;o$tFEe(E9 z|0_9>%_8Wg6PO@C7Q8s{>6M5Xm!!^O-~}T++X&@WG*P-l1Z)KazRaKWfTyS&nlj>4 ze4qBU#Kt!7_C)ONAa869cIOCNcR3?{&+mZkFGmZWYKRvA1^z2?c$a=<>FvNl7Q| zfz<}dOD|^-+gSu_a*130N&nc1`nx67xZ)C!ccsGp4J9^QgQ_GIi7Imk@n+I=qIC7stmBT9#um37~8L%-e-IGlg~r6?@vB2S-E`yY>_E(b?kdV>1R4 z%r@LiS>8!n+b4j+e*s+eFMzjNeqXM#W{AsA7|5&_z*)Tmf4=f2wsKA{6FphM-Z@w$ z?`%%CI5XDRzj>LTsrK&JP%*z!rMOWxzcOXNMuOcfAfoEd$9Ek@5R71-*Da0ET8d)7 zgd7;}P_zEvz+KNN$8t1nzAtH0FIc4oZ=q^57YFtor%ENiDP;^O`-?Bkd zQtB@~2UPvt`h~qEvOB}0bg&hv<8jUObcX6zbgoK9OLX3fLY(m(STbWj&uL$VIof3_ z`hMxvX+KKglSCQ^FOmy>!l*&9YrCvQtfK`^dRs=7;OjYo2&_~1fDZ21kbV5ntF19{ zIX_?%WZo^nKUK(>u0g6Ls>?sB%7$1SF5NR83HPoS9bUG?EOq-~fm6XgcbOsH>SYs1 zyy}_hwM&kBtm0&)^2aC=*i6|82t7k=ludOZHXd;--|F$>@^6b{+gy9gq@LtM-JrVX z8w;?HYpq>YCMI#UPixD^mE8U#4mWLcB=&bBHJbYZVSAAlPcD`B@w82P<^{tDL7a#Y z2+A~MHjo)G?g-s{jB*CEg?Wz?x@Bdo)QRT!-)K=1^O+}fT7WNi02u8xKY?76QCepA zQc*mc*r=@CdFFZZSeH^7#)o0rb_gNFkxN)7(D?Y7W{-&@FN+#0d%VzEXCI;wGT+n<_|G%nq?i(;S2?`hTH zoYtilm!bhSwy(--lWp&yS1WYwQfgS0hnc2XF(351R^R9t#b8NPC-V2dBe65kD*O}^ z-z<|TZe%`tUoJNyu8r>T{KMjor~KOY z)@(bz+lP|9Qcg#WVdBuPQrY*(?V*rDx)RFfMF6cf_XYbu^szY8PIAI3k{r|xDj;P%>;%+7wKhNJRIF^DU%}G1o=iy8jM+Me z{MV@H-^eLQ&_o^F+4w)4+Ts*=i)?r3bHGKdM`C_A-@LNsTFo}^a~f`!`cp$d{2}0L z%!e$0gs44Dv82vm!er`(mh|AkYrJ&-D*wU!YFkj+izX|;x$gr+=jC|S#h>RgYGbF+ zw@**nNz?kZ?S*LgMiM7!MFr zw7IjWqGKqgI=)_yorllQS8iSCglijyMk9FpUJ_f|z{W`l)72Pq&yDrkR+$iYU3qLW z=3N>+^=GK7xi!PlP>Qp9bP|U@C6e3Bh!Wc*91B*A}+*BKhm|pKjL6Rl<6%Rz>TDNOE z9$r%|H6t$_C5mgC_8LV&%xyzm$TVVKre(iDaF6id#F&oJiuF{|14Q*A3mPx}n})(h zJ7-`CTQICstuJ-zX_sw)J(g$@vxx7*5qqT8JTafHC7!KgN40(d-+y##s2(~KWM!A; z$h-#Wn+)<7j7s(@2blIXWOzK6?|kbuuYlExDvpl-yLyMky3{D&{9@$&Y}@PRfcPpO zA?a;Ny<^+HB;0r|}o;Vfl|m-~WB` zQ>ONXRj23CY)-g=5=ya@%n5&yj0fU4DUw^2%96!?V+mC<4oh9{w?*>L|B@fVQIQk< zk3h`GK65#5&L;~y%F`w#1ie0FU$e+E&CKtX?-`#?v@XOP8M7MWaD0D5@Yn!X>zla#=h)PnWiVy1tB&y=^m^uqO;d=oWhb={YQh&+#c>G@Oc+(HLlRsvE?B<6t8JxN0R4}E` z_<#xgN-%`~)TOa2w7=*i-o$wKd5CzzHr@5~CEU*j2)_Gc%I^a2hXC!$)Q4LH@>yBG zdm)QOvA$_nU63aDH^jNIpA0YX0-f%bu;2H)9u)nKy)RjTLyn1rwXlmVTB1txfPcvP zR0|j^5ip+l6`$AL5#cWLR=V%cSor`aUm-w}^x!gS0o`7A1JKe(<}u#E{cq8RANq^# zIfp#yIy^9VG~eyhc6+Dr*7PSx*0IgRMRh7tc8&@W7_>OvG(YOF%suKSiraVdRd$uC z5Xx`r@cnd5ZlfXkH8Qy2mIo6&zQ1%X@W0wfua^m9W!lC4oZ8rh#BYkykirULKw>G|nk8gJbi1HFdM?EUIO%J1=$7LGzlvz%8C$S)}?R8TAj3rgx(?YkGS zS1`S53;r_;AdPxl=u#6eQi~>i%-=t1?H{L$4LMP9n*Q7C25DKW#;VAIghOeeeQ;>@ z`^A4g&g;{$9~-Is_DXh!O6MF$Gj|Ul_7>gjJj5%I+|5+y{u$*8AI9;lPjQJg4E$2tNwQ#jVxM zmv_18V%eBEi`3%nJ_QA6ya>@_$KhHRZaWpvS*SiMbuDBld%GqFVOOwmd?@fzMX1Qi zLnn^X9JgqDQ;WYgZ4lb0k<*jvXx08T&kJd!)Ft^Nl*#Q9aLyDzBQ;SAEbzLK)5~?3 zr1kmQKQ~f0elO{D4}Af=Y4m;MN-=82Imw4!uFvCLCD~h#tYl~fm>yWVUUs}KuE9TW zu-Gj=DS=we&K9z72%K{Ns<-s-czML2yx%onI{vqisC|Lc0=31^L&!m=QHmE&a&P*REkqZs4?q}~|?8FX>n09X=8o`?8!M?7c+IcFbpZZAMh zNIlA!maPc?cTwdteDD<+@8AYaFAG)JfR;L~=*uIoBMS^KF6LM@m)UTa!{&YFrxgq0 zAOd&r_5J{t$9hV?MH_KTwU}^vPpRu!B%Wl~G|(Q=L!OPX?C=h=lh9{2cT@R@=Vc=~ zliX#K1*J<1&8v_m%F0^?C^cV+wRU)43;&}Hu7<~L-8pm$)}}T$rq|3yuZo?alp9NZ zO*IJD=Cg~Ul}8-k8wdvfZKX= zmm1`GUkdNxTuI&T5Q_dCa%w()62<7c<>;~35k0-=c4c-ODFfBgI5x^JBt%vTJeF5? z&eS@>MAVV(`_?u0#Qla~2;owyJ=}ehMDCPqbrQYk$y$+$h&oYMl=Y&j|J|tL#WCM? zLAD*Y|IdeY#@W`NG={&*o-N{4$eLoK{MhQF$jMAVXt)z;>O&WHv-e^r$JfT=j$q?N zb~Ulf2d7(Yx#dbeiyp2t@jc=hKvc8iw9U$4{<0ow1%^$jD0xtI6E~9^&xA#A7cg3Y zB|WV*Yr7L6SNlG|AXxn_5gTL?^1%FID_ggC`cDaBv9>izt<%aY`;0c!1$WNTuPlWivoz5MFRYQ_8X0Q2u>dTfcjd_BYDl4OMA3{1&YQ8Y7ITi1zqC1%F z3VosdIZ|;FYXdgvyH|bh3)=cji_`gsVnAx+tTO<#&Idey{|3;20_J7n8#V!y(QPh? zTJ1gadCd0z?M00m>2FlI!>7L9V&U2NI-?ML};u(_t z!$d=QftR?!doU+9Uiik-{{v8_7id0PrFAykLrL3kXErUr0UNu@7{LnSS4PP>tRRh|!jMOW5EjE&Ycb!OW~0g*REaPTKs@Jpc&#$`a_9Ihx6=tT5D% zQAKlc#^rJUF|j4a?C4{Ms$HNH$9MI+7^w-DEE7-9Rq9F$55WEOWLN&}zl1i(dW22>Bka2U(V@G6r<+3KcYM%jYHeUg*8GM_m@=W#FVCdf=2Ga zYue@|>r^kBq5+Ge?Gods+zy=qcbRlIQ!5Lft4l?#Zcc&-okgt{aixSKxa7YtIDhlY z8rykX*=i94*n0Wj*$-!U5NfrACwB&t-|fFuK%LVIRSU*uvr5;OF_U^Nd|Z3+r_&tD z)M-8c?np{jV>N`_FRD!^kQ|>-OVzNW38#)~8FYklK-0zbvB?&XuNq(UB~auEeT>Q( zPlZ91l=l@TUxnA#DQLErHSu;JD*3YXU5bH3-m8{W_k(A=%it9ZFrpPfq>G7Y`1U~! zg6Ur5tnAf02uQ2e!;!phwIXI%X;+-W1{w;VXuQ~!dRxV-t955v6aI&5L<>Qwsf^w( z10g!4%{QE}xC33Sr5xUp4bCVi%Qnbs;IYpqR_AngRQ{U$;j@~s+yo*`o&7p+J-xDp z)hwaF^1`d~W49O3us+WDBeC?9%*Si0?`spQo!HddE4gdf>5D3dtljs^x20oGd9TBv zUEas+mFWO!3guwY~za=&5(;M+}1mW6xZX@b>1tjX1=uBY?G6(zJfUiS@~h z+I<7T>JVCNe+Adnc^n5od1i6lyFrLv&{obQb(-%u@7;lcsq)7994np zvXZ!aystCe>g~-R0&2RYYq_9(ax7~}S(*VXiGE(smi$4Kg$t<&_iz9WL&*`_uGH%3 zPyainI{DS?_kPKS@KU;ZMEN*~_XeS!gVvAynAa<4fEE{5%tk(_747lvet4N#8!Gba zD7K4iY6F7_vu`FEuYV05-~s_D_h@n;ju+m;OqqJ25V^UDsVIog`U~O1-b#o5_pMkG zh$gRFNY%lT+yF*CS@)?27=F>9eb{qHCnX}9%%KTdB|4vReBR`4Ck(*zRIQUM4LTLk1 zn%Zgg%RW~Haru@h4n=UNI6Xg<&h&ANhmc8Bg^fmr_Y!S)U_u6Tpc_>AhN>Lae}suV z8;ReKycdrpGRSGVJ?L2wQ?anxNgEuG310LVie3O&QQXuD>Rz^9N$O+faGK7l|Lu3{ z)}lK2@)-SAwjRMfz7p_h5NU=Ch1Fd#`@?Cjk`nbExyUbr37)sDOeTu;PR-YQKG4)X z_2Ec^POhSaeCie3V3?c9LI7J}6~hk!T2^YNeQ~)=kNE!fEi2M80?{%>CzTC_-1k1J zMtSAx()NIeH^A9Q1Q?dG&x5lORQ#+R4~!gI9veE<_L&y=M-|Q}5>_;87TxV#tN+v3 zGbV_)ttvvBHsG=PVh^(gY({l|RsU+~_UHp(WE97vrV1g01EEc6*|%Cj5&7nQ2P6h< z{7`Myy|T9Mtv$TF9pO`0$j?H56@Wj6(JbqolG4Db2cJ(nibDP)CeO@9JWo?uh_Mv^ z<`UUgn0Z^#FB~qW?Ca8C*6wB!2zjCpf%*nND{mY*G8$px^C*wKq-HOojnhOE&gQKb zPmH>`;>mT&Ul4f{E1SHCQ6!#JcXX&K+9m}tE0E8?8hHXxc;|1^e-^yYF`)$ z;q~GxH~xxe&srqwTIP8c-?wbc>6n;8m-nWT@V&^lse`NTw^!Tt;y1>rmu|}U18)kX zu&vmxL2O3QQIR{>adl8KdoA5nn*$qrjbem527gSSG|8q=v@qe9J`fRWsmpCDYpX@N z(i2svK)oFPDJbatedhUbV&-)I3*T!;yK&#$ z;4h2W-Vxlixis)~YzO(mD7NId2DZUkO}bCFe@A3RNd9I@kkj@hvz(xkV27FfnZbG- z>wVM9{zO*G_f8UF!v1woK#R zkc8SFbfF2Nox~CNQSGj3i^}08mR&CgYab{o>)x_yeH*Z>zWgN@o3igPfrwi7($|75 zHxZDL$>*Au!U=DwnZ{Pkpmd|Y7cEcp*+ArPxX1O?!{eSUjWBCEOp1DpkVdwlh3?EF zBp{-_EeX&u&)|e^z31TMa}v9gG;Lb$$l)u7nKKw~?fTsImVd-6DIJ0X!(xw02D_;@ z`+hIkl8MzW(;cwZ`#EaR%Y_aK{FV{>)WRWShPZ0+=9=8V;f;J`{G;X%bBEvQMFMWt zIJ%|YaVE*0wx5eN8{?ug=pEX-T?pEQj}Y@&PcuCetTu>q&DEtXu!r1T%amHBB`#Z7 zddlfJ@qLMLao_I}k3X3PZ>ZH1~5tuD0M9G9(E-o&H4W#>}?bYiE z&du?P{H1`yyMj+C-&Eb6$F8a@c~T8syoNurz8^O%QVB=3vYobVHpu7P<`cLoS3n0t5cz+ z$@LTd>P|Y_5nTllRd0R>e7kBuO{7*qA1Ps#;giOAkLJ@r*KF3`9`Mf6^Xh9`oj;AP z4bQl7C{&KbQ*^c;KP7uzW_v!Cm9O2%2b(s5X#B>ITy?|f+y(N7UE z(|YI!x{*}*ipZzM$;IawTSvkk8DSK(nSJOEf=u-bV(~XG!DyZdp$yF=C!jH>rA-`Jc%$32oz&AL(P zhl5~9#g6+ncEeFyLiQy`j*)Wy<1m+5K(UiM&yj$P)+1J`cEQ92)?G}Z^`p80wK~v; zt)o4>$}%zA*|ZvHa~WUHNz(TAyATX!2bJbPEy9+j;aUIi!wYAPMD7l$@fQD zL=y%CRDDVyBd$xRy@DaLsvdWmF}NSVl|7jd<9Q9K*%LCw2QSp^4pHy5IJo!T^9^lT zFGNq~MK_9OKK9mWp^c2%A=m;QVjLea%QwruT|sQZXRB`_@Nz&aUgbRD?cvvR=r!LW z*r7+HDVJ{3FTU!=l;7jsKDRop=fQmgE=e+=_*C`TNi`S-m zp0fdN6&|bT;fSFcM0zCp*H?xmPa5-|4!q7bGT?r9D)!77$0=?EnXXM#n`qE0hpOAa zp{LIoIKQRv=FBsyTxVg(*BjRPW3sBf!8N{k^9if-6mivfC+^gnV5FJ+=JhW*$sdSH|=#Ki17YX@c=`@Sj^b8As^`p3ul*2$fgiraDc>Z`-` zsPD)qYk+V{TV3}=*Yqv<;zbcd$~VfzL3D1-BOjHcj(CBCT>o?;uX3)6K4<1sGVqes zs;BuA&sp42z0i>nQ_fIR9>^%?lo`IKa(if6?nd{|20aS>KXLIq6<}ihu9e#|zj3Lr zqMi4~k4c}!A~hW}qB?q%2@HWn?dq{ojYLEbbiXcMO%1Fxw^(~%dHvkWY>q5*N@NU3 zg>*XIBu7KtMPR|4#S4KEGfW#&dAv+t`d55}iDZ~mdQr$`0^~)poLht=t>^sum$UUSlCj0Vhx5k7H)5W_U zmbr}6hi*Y{5ApWK1Bh6TgNOZO*O55CVHP7tyHM(bNS1Hc&*@-u34ABH4()w!D-+#N z&d%MHF6y6M^m1W)Afp(r-NcxV>7DM`eJ1d9Nb*~?Lar2qwKonbGv&Qkm;TJaHmLu? z>!Ia7Nq}KDB-=G2d!mQ&ch$!1VmM0S%N!d2!!uahVkq&U?+t$9LSRWuPLPx=Jt4{G z9?I|$$hQn3z|Fa7(8}snTJ4h+8Ub`C-Ye1HuZ+TfEvZd4W=yOp_Uz-%xAh(w=M<*M-i=9c zL-*X>^EdLn)26>Wzg}e}_K|5R2g)l1L~z?NMUil3vJ{Q|41Si^$H1=F4$5%bDhIOI z5LBJ*kzY;hWeJHjPqD%FCGS9dC4kSEE;;-AH$`8kz2f`5^JvW=@8<(nR?Ti7IFVyk zes?5OmvxJweg4IB4ZSuCnKZ#mZC%P-?+z%E@xioGX6I3_z=2VVXI?TSc`T--C$`-j zLVu1(gq0lpsmJ+sfr^=AxX_=6>{Uodx~#ae$7VN``u$Gi@~y=P_52wCd}ePWwRX`) z6?bmSR#L`NHvU}8QmamFPDmzGn?1FD%VjZ8#Qun05-JQEOxO$?zTaJ-3AxZr^yV|T zF+AyhO!l_OwhS%PlrdTg`i4vZ)IULS@D~rhBx88>~SW3b%TF7nlbnC-VSj zZNMqq{Ml0Y7JkqV`eTXksAMCRw`-UOg-C%WTvo8cyS}Ntv{{q^d3u?|rfMXG5?mfs zY^vpS*wPLw8%t%h1ClVqH2trSN`hVU)N!nU^C?_CI))NDT|-Y!1kbA_$ta%-oK(JC zFjV&0#We2P$GV4R+}NuDl3@m~k$b4?5z0^l|e3Nu$tLaOiD~-_22TcoIQo z-?l7y2}c!eL^U*2_B*kYZSs|?AVI7D8C=&)Y~FwXz2@OSqs}P;Y;AH^pJ&gsN_X)I zlf8-56aJ~>%_g0o??soMQt0}UQH0>5L1(-nIr_TPm56O=%nh+J|53BkTkH%VQa>Os!k(ASlXgCHM zf*+iR(9;wmic ztjS)*#e(=t$)8l5)VW)sOTX#Pey%6XjM}oM&4(6U5>H$I6Xu4$u2JI}qmU`~bqxPx z!()6UabQWTp4Rf1WuJzPoq!P(`rEJy#m3B*r>lV6vHUIIbc@QLVfkP%;VAKD_BlMW z$-e29f#}6X2h(?3oaHV2wPzyjkQPLigWMmVAy-1mlZ8Hl2y=s}O7}HPF*wYzhefR{w%&~%E{Exyxl@?q7mW3<2VA&T zw1=rl#~5-bIp>E1(!K8s%3AE-2h|(e8a4mCw=?Qrz#-{={^5rz!o6-fwHe&B^!V18 z6!cYd-1)6m$%f!elodlKoon552(XO42FOsd8p!=c_M$iPQr3O>JIEb1^_c1C7vPEj zXza8Yn&XKvP5F~*B^tN&+4MHN7IOlJYEaxzc0xKC9c%xf_?+e+?c`$d*w*tHnrtt3 zt|mKB?b|hSF3Qa#`wbtKiPT%HeS}2i9)3Af+fTC9lKytN}2BZpW@jo*l{l%+!dI>dKn5D zdsS4V@=mg@#Lulb8rucLKiC^EIfApA*)ErR{)FdMT@kIM4X;Kao0r*y`~5c%EL`?4 zLJ#PX2pq>g7!vUA%kG-u0la4)4_W3s5t@! zpEHLf(F`8KuP5GD?KaR|^fX!y9Z}`ge%&=#Qt_~G9hzO*B3c$&RNO(G==UMvWh$}F zo(S+n1e*+*!pAFq>1KkUXu#F|Pts_ZyRZO7{8Xoks3V_~-M5<`sP97wh%U0&O29<3 z7U39bnp6Ri6I~QJfp&w#1E6|+I2d+edJ;zM?3h5;e9Qh}xdv%`qc5Qi;ZmWbc z4S?41jvV?hBvHRWr+2*QzK$vp%D?ftfA>*bg=D%5YevSYwW}}VcBL{ z4!qZT-&kJ4-cn2`ufFvHffdVim3_QJCSJ4cXZ&Xtpli6$mu)il`ypeAsMlNI99sd( z_#BY22udGzAbYQX$yshgV8$|Z$8Ol6veJ@Z=(955j0_P(nPql@Totm7^LDO%BL-Xs z`>R+HS}AS4W-uhT+!190pDi3u181}Hv}YWcBZPC9-oA9bYWex?ea!G0R@GvcEy(tP zxzfo*F1G#=#ssJv=y46@n2BpVmP@iAlE&cRnL+=qN~4`P)LOVF9FrYETy&G7ZW_T` znKP`X*R#M8;hNY}hQVv#k+P`kacGpY@7_~W(6EE#o_auE0{*EC{ax*`VbDKbOmPV{ zG6c_rw@0*j5lGu{{plid-Mgu@)5*Yyk*WSNe$Jhc0KC_k?#+@Lm&YVxoQ}S~=h`3v zMgP9)*!f#lVLp{r)_tRzu8TUg$Dx-ZiZt5k|9tURYC@ntHR+iKk}UBWJBWZLd5JEL*@vEMhKO-rZ( zJRLuf25bkKlPtHs^CnJ6z|f|yYrh&ZN5EV+YDv0cCMTY@E(ZBatyq70ubbuIsHwQ+ zN@2^6eNa_Ows`w~ZDdJp-We6<7Qbk<@4EmhpF=1{m4%RWbL?rC{zzymod%~uN@3hy zO10VdjDTD8aaPR9f4*2=!)8+)kQ3A@WzM6BWrresT2Vu$DI#_VrR~yU7ihUc5bi*H2^%T3gBd^axax82UAI>oXAUw^5s=9@nj0hra~r9 zE-0KI&|U&Vrss*rW#a@~J%+S}JW}ha{8|2IsWG(4nUct$^OcsR|BG4bfe0yA(U@bZ z{_s~OmQ@XMe^S@u3S-P`8j}y(k8=?J5kO5x|Nfwa6N0%QBPxG_44p1hOTj$3q<9# zmzVw@TVEB`*4jl26bYW-UfkUq++B-11qu|WxJz({QfTp@#S65!dnsC6f){u9n{&>8 z$G8voB_m_xC3|OO%{kYcdwp{k5iXtMbQ&Bwt}OxhzE;SlU*!imfT*_oj8$;j#+({o zc}LXjpF`Z~x&PG)y&YsqN_Y4sA~`Nf_Xv~R^^Q@NTk+csKd%vhZ+RKj5f+6{qZRnI zOk7QHRRij6K^7ZzZQTYh{5Ia8*OuECecL?H4!SvZl9JFe@q`iSMLu%34YSzDjnJK- zu@Q5PFZ1NC>rNvhX?f1NE-JIkQG}m5zONZZsd;EhWt1-~VnQ-HVJPY*7Z(}Tw=YvD z)P#B@)7bHo!$_Z?B;oM5oho1 z1kg7Hz$QVToGKA@b2Mj8(a`?ZHb>?$S?U0FFwYh1OW@5HheVqTs-jBTk&4di9_=($ zmmXIun!mP&%R(wsnS=CMZp}to7B4go$a1rBDyxNW&p6cJKbEKq=wIb?R{VDG_U7yc zq?Pk`VRU-7tYww%hWps#g zSYF^U=KRHdUs8kN)TR;Y+GLw(v8{FPf#%zaW@;Qnb`=%t%;o#O)|WgZZ#}Q$4d!ZF zbMM2jIn_q@$@u!A@3{xX%5iu;@j6oWkehE5tytKI#=E*kU$+!O7!#cB6VvVH%YVD4 ziW9(AByZmu23w0N+MEhYo8lPi^5d2$=UzbzBXaXC$Fa2kS~SwUe+wT9Oq}6B@0lmN zM360iZoP@&E^PY?%wCeteFFU$d*5+BZGWz=I1V1E7DFE}m2^wpc`b4CJ+BvyQD7zo z)IdgCiX~m(6loJQX@8*|0RT~NC7_Hd?-*zywt94q4J!?I;wEfvfU>LTWTU^MOwhRQ zWMXOr0xvl;GqJv`QcPR~jdct=S`vj0@d!$OzYiMFLsrcCQyT2wvjWE&>-T&MzXfG~ zp!V%889!oW)kSiNc8MF6Nf69JpmBqJEt*tAN8hVWa7YI^-mAWTwa#yg{c{c zNf~thEN5_;QL>$TYt7-fX$sBNf9xU8^-d%BOI_ZYSotHoc1jXu4p*5z%)^V$cava> z;{0UL*!o=!Z7X+VoI(g$55ypaPM6$a4F4wk-KstpTJq+#*jQ;z8pA=##lbgO%06C1 zh7cvTEqksKY{tM>wyW(n?W*SY$0%#GNl%^5 zPz+w>&uuiBjkC19G|>IKiXKyhZ>P6n{*Fi+}-eOvBQUu7n#^YSqE{>7PG<467ERrK4hQ1`Oi}T3JBc z1gWwRKsiCq+$9ZFlUa3fYp?V1$l&n*!|KWZA6D=Ge_4HtOO*Naw^Lu}9&hm}PL4KGm3FpFcF6LJ|Ra4?qVk9EtytPOJh(Tsc!80`6O1Or$(Ut4mD?;D8a zkNb`t+H~FQbGUZ|Tvahx35(SZ%e_Eui2(eT7AY=Lg^?jW0ZQ6yC36zlwK0=1HFOo# zKYP>~<{VctR(2*O^GMredp3zOop=p_KMw5@NL}^l5*GOS1Jg!n`>LwKOuxmn(AYLj+s5sR3WivWm(uO&0jb@rBvaBT-^OpR*j=%Ru8RMWlH%8ujG-{w? zJQRs**%*laL1xP?ZSs)Ic4gM(bwIEH{$-UFI)X`yF*iW;hA~Ocz(w9Q(mWZ+e7a?{@eYvTN78^4W-^ z9f+&TpFQ*@oi5WF2f1_N;JVFjpyqym#dwWPV+Cb>C;DEnaFbK&PE@M-$Hd23*+ah?HKndX{mS(50xo z52~g4L$LhaM;CeDr_Z`4Ff3E|PdrTw&)2*;;4%zWgYuJ7X$#jl`@Cc37Z~__DiCLpp+_# zv`4%0l?a3hAPQFD;XaZTYXdD*qksF?E&v&ix@}6M{Z#`MKE`T=H* z+PqIF(2gUBu9}n2-47J#Jw?kU4y8ZSn9X7KSId{G79zTk z!~O^&N^{Ec7AORWF3mypQ@*{Fxn)T!=Z6XDsj@$!?fc=Zr?O`U|4U9L7c7fWiij(z zE$yJkszT9olMp)s`hKFRJIkI$lpbY#8jN#kD=?uWSTgt1FgS{(uJLYo)1?D6rPoA$YXW@L5etrJL%x340n#rPM zap6{#2xyT;C-FO33M#>SoGhw7k3??P;Qr^D2bKn*rNXzyn%n;HK??l70re!bk61V? zo6)x!!5lPP&U8*FZ9ICXmV|2;NqJT~(^ON(FG?xTXdZCe&@hS~>5ElP{}(|{iP+Jg z2({qR4Ci(Jv#f0u+m<%*#Q_^01N8I#GjRRHM?KWID6OB-AkKbNOx<+GB<>sxoCE?Q zlXyIvLr=fiMd>sFL1;J+^!|-b1F!|d&_WHu#=kpqf*n)H?@r+F#rJJN+2b6E?c4n0 zFWlLJ5%2sFt4|6r`Mu&*7d!>~wdtM)-=8U?KYZ6ZpD-70d8~gXjx8Ip_Ojthx7KfN z?a391c+I?X6Dj>g-6^TL+5!~j;uTvalL9}$y{Rc$miykwt=*$5 zQ`Y6HP41Q|3jN8^Mc(FH!+|r7-n2(bDE*AkP#wa5^%bP$eI&^*Y{s5-^ey7j?Y%E zysDGdY`Op8p~sGVkgX6AcXKi8MoefFR-2dec3Xu*(Zn?j8kor65?339_)UKn4LmNb z3Rax-!{1h~K38?&lDNC4p?X*s9quR9af%n+E3Ftzwqr+qOW)){l+~HQf^(X-GPq8j zcU|^h{e1nWe%>1aJ_hC=_~r?(&yOQ|sZ$Q}3#PixuZ41@d*{jv{6kG$iVG^@Y}RVu197(XS=_KOI6<-V06vLbH$h1{2*DW zFhuo9(_n7E4bMIa^mX&_K|L<`5qa#S`$VBn3U^k%Y2IQALfj$HxKX}7tTa+xI;KUS zb}la=buyD)j1wxu-cnPKlr{JLT9h?vHpaj9lG-L7pRAUop7bp|iZh3&TbfbnGrFT7 zi33%koKCMvX;`BlV=*WSw)4FuiFl$AT=<>$ayn7hy|q=el|Dt;myZ1j4Q4L5+`iA_ zH}+?L7PMfr?;un+&^CBC20q_41M?Mr($`e`B818U7C66Kop?>vw(mXWL|2`Fe{Et= zcVn9yuSJljkP(GmF_e)Ws_ZTB$1 zBvOFG#0TdKE=X?kkHCv5Nw zkOirz<=&6q9(Rny^-r1Wo_LKkqcCN`1zs)iRQx4@F-g>)WPvoFuRY>^sQVxark6VA zNOP1uFfSuCUvqzbURaZZcJr3%z7=E*4rBg6EyMn1>ZjY;8zZJMQ_r-4f6SsbhCBHD zUUdB7(BM}4iRhFF-=A=0AzI5l7j{(P>GDI}2W@*%SnFyz)~^c**)ozG9QNa2*wza{ z|Ax+V_vZe*_$7C~V*W;Vq{`8ok7q_EgUW7m+xNisGMe*|ZJsZ1QE^wpM+QH!>SmQ>Q2Ml`IQ z=qP}E(#>9$-fYXCet<#SIW8J-X$~OU6}`f?{D5%nxK4ctKXSL6s;Ff1cU0lFE#GHo zaELI8)WBUVkyeX#@eV!Xf;!A-XHe zHrrx<{dD-x;6v19<{ytNcPCJgLzVbiS0@hw*Lp=fk+JvD{=q-W1AC`LzZgYn)4^d1 z%oW)9ed}lAI}Bk2_0x&^1CSTFA-^PW8pzsXV})FTYZ#-|5?9`yG{phLbGV<(amsx6 zij>@E7p>-O!}~KXynB9(jLLfx|KWP-J=@H>Oa?2X=Q4EHgL@A_%JX_kT#zlz8+Di* zhqSM(OkPtWShrIvs%s0oNRb+^$&qf@F`5NdJni31Tvm@^^#M7wsk~>72s6$A{Ycf;YITgM-b<)3nPShvI>-)9P-RK|vT)OZgaZ+fkQpI#<^u$0Uzroh6n z$n937At(`(nGG0*lxgZ|W6!9%ei5$&4<2;~1Fd8zJ-Ar+fhM3zagd1@{cLufq|r7qX8S>r#;#ce{X^+Dc#9(` z{q6Qsl7NhfUv4o;i4EY7XXuYh$q^hb)r4SzT2RGwl@RrCm~Z=Uw$ukA z&w#ZXIj=r@8Y<3KS;hj_DU@x?nqp$B19QXA*}>neMiFW~y8PKhZ{NwD!hLNh2lDoh zvRcUqH50+as(8e3fOhsM^#Mbf34Iw?gFd&t&H{#~JbXBl*kI*|xvS&&FpA z-!hDI^+n=i`eU(3PN9@-BIlBrz3T42Anom~6>(J~CI_M+1P>*=tJexaMZq;@LeMPJgRQ-%B!u14VZwrm(184JBa-pf8Lr{qs-=^A)r=j5>EP z@YE=}i@KL03$tfmPG_RW8c`ZhrZhE6UBZ`?uU;J)QYdt}+$Ak=UU48E_0}7hF^7jt zSXY_HO(8ifZv72K$5oDWhu-qZDeRnojAg1BF@6Q|1-HOGIo}H-Cr7TFQqz&zotB()zfCp5&RmU+HAid&|x} zt0maJuV8TI?clyQYP(D^@4VFfwYAlGzW?=O=9u95IW@R!v#$%SSD9tXK6VS?7vORvXg-sWiNp`pWae|{(ylHtS?eVy**!F6^7xW=pOD z0(OXqJdG1Ig_+j+e~`qoDt%`#7cd8c-7r;FjXd95yv z^5!{-VXnN{q%s1VqFRDz(%Pm@E4bUiKPWHXgGtLKJKqm4%@(=*HK`%=*__TvGR3ib z2poo-m8dy{L^+Oslf#r$vGvh`M=hKxfLt+o9ZMbe-Jf{&FAVG)f!O9(ZC?@A&m+jG z+4$j!0#)2(IbCZi=+O5%x#^E?d#r)$KA}Rsm(6Kg^-H^`4ge2IS<(!1**FBPz}J+y z4ErMW(2j8>D-x^sJpIiKY~{p3LD6W8Nx%f0f+K{`#397Jmdt>!za~)NHv8!&p~S0y z3&&R|2Tp0>EM;l1aQsU9@D9nc0mqO`exG$0E8-P?YL;M$f2cxgm5Sic>o=vl-{Ebiws6TXmDH(HiRHo5Sd#a>J7`E z(3LwY=*zJ7o~+d+AOB`Hyuk@`MJ64yXgQGjlx(S(|L**{9n%S8nVA(Vr}IW>o-f>u zhN1q;^&`rfc}C!JsrSxWP&b`~sFZkX0ySA!e&nZ%`D0&eeJ$BU?QyB>DdwMeI;EhE z{nL>xkgq?MmI>sP-qJOtfpTpM8#&$Ho${SXIDrz1=?AXMIl?!BxQMQDiZH6)FCTYL z>7_YdPkrGZ_(E_%bRu_8v)3Y=WmylsR=ceOawZ07T7z(^$t zY+NVH2VV+u#+T)QC)DK$KbQ@C2?Sr6l>hal-O1!J=0i&orEDOB2WKvdD zfHFt+%+`g2;?52TnjrAGOfb$Ze11&0l6&pP5m|{H-PRS`2@~9Puc|7l&q_`YQnn;z z;_o5M6s58MeV+cU(E0!;2)Jv0Fle6l%4|JUX>?){{4Kh>Jx--Fgv~x|KgP!YIB3Yz zDt3TO^}q=t?y+)2Omq+KO)&BQuRih+{s%%)Q9-qlCXeF9&rW5Ucyk)~`s5k+YDUN< zx^zxS&_d9qNrFOdKtx`^!E~wU`D&07M?awh`q6uje&Eb_p>maj(5GY}{~==smZ@Y9 z`L|@?ST^ZeQj`(2D4{=A8aOic_~rZ|(|`EK0&pxH)&@Y1`K~;IDhjfKlC`-pZ~LFtIpN`K-M z6PZbh5@uAbbg~xr=9XD&8peZeYfncvoMLsn*(5qQ3mBaWp_~JjQnYUS)(R{_5POQ& z`jstaOM(`JW&OABCs8TQ=4T{B#yz^mi?04{A@;wuB{9SMUHi_d!WGn9G#9=>mkC4x zR=TZRd(RgaZPkC#>pA8dWG6}GBHR1K6u6w|tYd3YnjBulf66rhljVt`iWIFpJg%Txc@(0U|vJ)DbU*wScd*1o_Vi&xE ze3eoHeuwT}TBhQ#isDEbWEJjSTC{}cb8eX6XR#$C&R$xc@0R!fCbCMPcEPo(vsxSD zXgz`_Fce^IuPI$*uM+Y%iIf{P>(oLAik@WmP-8u?{M|W3O3mQonc|xIh%V3dMZb6Z zNTq2Agn_wZhw}U33K>Uq*`*9Ef0h)D2Zv+5eb&!rIFNnnMNi+Vc{JfLN~5C8&|9v0 z6Wxgf@WRg(`0tE><$uB+qT{XN|J6@kAnvmaECwdM+?qNVafVadp(aw29W^-P|~HH}lk9yTFQ zyVAC10`xKI^EhoYhFOsJOf%Qf6dfuIpe|GQBucKqS1;IU(IeJPKxT+6+rxQ!o7xg) z@iu4QdI-!?(#LYWQx4q90j+@%PY8*mb4I$yUZp8;Tiqn1c^8ByL^maQIB!b(qgJ@i zhT2(-f;>D@^WFVumQg%B$tqxQ=FIb+gVR`{F~{lxz(g6b*In#cNw~X9Y3-ApAGAO9 ztm4KWZfW*@tgw}jU|5(Do9L*u?URQyiSh5X!-cFxMjQ93I)`z4)sC3! zekb7cEVBR;SR4i$q&R|mdI3*567c1GU4^w!(AvgCYjBG}CpiR# zP=MPIn>2gC)3$aPRp(nAp^c+nt8-V@IBv?$8{ZpxuESh2_g00x0|w$?Y0dWV{8YS~ zPG4Qj2;9~iqf#ZVYPs_9w6TH1?EAYqnv*E_guVV?!8~=84XHIp-b<*(E7slVxIBN3 z0I~m(d*vEv_`vM3x*p*|6*5@!02G?(%MIeK$9{m3d6>uzA=q*>AD$1r{bS%8Hnw*o z0r>vUq-Fd@jG*L>F~6X$2@dbo*)Ozh^d$rTPCOm^c;!>F!6?c2(J`NRKo&`#sdx}z6DmE zaUhGh)PDSc$~G6O_WcekKLp8|ykLiAZk%ILkA{5|U$zszQ;^P3tXn`s~G zoHZ0rty~&&Fv5j_Mh)U0b7Ig?mwKk+3iQ5PJ1IUX5q;dUR0T`BTcXtoCb5^ViQ8e* zxI-KZM)-Dj-LI^3KM#q}Sq~dS?))o!f9cUH15%coW5 zZx|7(CxrD;?Kx3q##kuGXsQ8dltZ=QJUon5oXBl(iy~Pl%1s`AZwU-Z9R#TxjLXW+ zz0us9a(!I)l-*b`lw}lB*12nQtbdM#ufAj) zn7JT~2Nw?a-2|!c!~H71FzZhNL7UPORagTMO=c5uHJA^6R(tQosyHAi*gfw6J|wwY z-Wj^`f-Va4R?=<~Nn0Dxxa~1l&&%gJ)0;G=#E8|@oe=#%zrT2`$IRp~WW%*6XJF~Z zD1}OFmo0qKYx7SPJ*;fLGK0T4sXzMu8i%eQtV000t`g;seHhcCefrB%OrJ-3)cD>Z{%L(Pv zF%BmM4wzCodpOi(VV`zuBRr`KKT!`A81@ngoO^}F;sFUAL3pzz=!aO0q>OBf%(fDG zNi%H4#rD>21VzPQpOCf*fT>C2B9Ww`3`s5%=4yXKPz$bAm#DcKxa^kTx_Q5{gD{l`rH>~|G) zOV=(P7uZpDwak6aS=SV4B3!A@%}?bZKQ?38f0H!&%b=c$tr5|s+O8As8zQ{KMFqK% z;8(tq2S7=}rJKS;v(S8r>|wSqjUsH|zoWcwSnh6`B_!v(fPGfTS80&;6VrFvuhIUx zV_?Mmii7SSk3l^Z4k7?@F407@6*o1# zP*FtA^THSRD}sCYb%l*p_AYYlj_m;0Y_+yg!M|(uw5=Ib8#5{eBw=D|vtWptxuyFQ z=d*=A)x!eW{EMVPB8VuvoT>ZYgj>BY zs=s;X(gAn7199~oOYIUQ^`zu36j$UFqd4Jl4h3#|>_4Va08Z*HQ7L;O2g0%UYgl78 z^H5w2nb~r!BFdsn2G$?0demFnm_sE3a7mZd$g?1#GgsNc%2x&f5Qnq`VB1;l3debz zPmM#k0fr|2Y3se$RQi|1Khj4p-ZM#{-yvWxzen>YF8BGIfeeeht`o!M+UXN$W}fm(7?xHaSEc~Pvr6KHK_my3gHu@t`OGmw-Hb+U`bsqoPCt>I=Xv9dBON% zb;HEh+Vc;FwUPh*ul?|iT^D%dxW@cy+3vgYooXJnoAbs*y(TSi0DwTU=&WERp4S%c zD*FQyzL+5XYNPmH)-(hHu?=X4z7z~a;AS68ET;LZ#=@p|*h6L%Jkmr4S=2`;5b^$t zRf82IC47mpRJ1}xN_>;=z-&GY23oAOWi&C}8j9G4U3ck8))dcuNgQWO%uwT9+r_=X zZET29Ajva=9P?c^@+NUU=8 zp>6zQBzIQlU+-04pTt*55(wv9_S0 zpDzcXrd4{f=8XLJWrDc{jV+iNYMS#}v19bogo^96J^{QjqwMxK3;}1n;{jK4&Rz_S z^)>5HW{#kc8jA}zNGw9!DaHwS{t)jL^a)c_Af^!{O|*tRX5aRQ;9K8&-GE#gS#r|| zXe*jE8YLHP7{wnr5949nvlx}jmiYOt_S~tkE}sJ%*7gwW{hA1nv91=PKz$0xS*xc| zc*Wpz^X<|dcy2cNalV*mY{!nrE|mI=`hnyEwd947_XFM~S-y19N8v`r+-mAKQA(?8 zA_ooX!nd~oK?*rp^xStqPUB?1r;!(3oqSfx9vHWw1I2s*;lZmHUCWZlQ^u0a)PTP+ zpTsLfZ~AdvGU>4Vfc~9HWi2_b8#CBBWij$QJe;(=t@(Utk6VFN>ucuI!_zLHM3|h8 zMsj-n{GlKuTY8U-b)R}0N!qlCnU$bIgt-o+3YOx?+3JQzZ!Zf*gQGqx7i)jhxu5ui z=<-d6H68UjSPY4*jGE#Rk@F!7XFP}$-4L+~)?DfPwn*kfoWAZU^sx7V3N24(E9W$) zTsnryF_-q>>id)UK%z7_LJ6shHbwa0v^wt8r2vY#>zcY%j=7B%^wOE2MbS=d_(e`= zAf+Yo6UL6@Emm5i>o7&0?1LQ?wt^@QtT{I-eDj8@dTi90@CHs@pSHuw@8f$ih6ZX_ceaO;rGZVZK+WR$T z0rgAtu(pC4Q@}yWb!1Yw+@z$A*V-Y;!aMvL0UzL2{vz@jUR_uKv)*hqnrLw%@AM)0)jgXQisY&HL zra$W)S_|$k{wm?~`%DP8X0^T*UVqBPQ(Us7r|_zrQdqIg&6HnSSu35GAu~f&bl~y* zwUfvh29~+05(>77supg2X4n?mXh(J#arMU52<&!WvrBZuquaeLx)z&llcn-&r*Cr02Zxb@q|dQ9QP>%m@KYZieZ!q%;SK$xJ*e zDxGQ{|GsuTxHRGL^PHNYoIgIuyqxh|Z4H-FKMBtKb3$dmF~b@5i{!@{am~G{XakZA zn=mphZkJMp%$Ro`LiFWIVW#i30+BA(l$dxSOa&B*| zJ-v1Ly?M@Q=nX03v1?rA%viSWe`#>DZ1R!RqqltYNuet4WQroUnnZnEfM5?S> z3m~Ln2yKff;!9MDblvNuEtCR?3IFZf33Riw8d-T6MW5Pe{5C8CDvylid)!WQ3|EH5 zSGL9~5=HNqYOenfb8F`Z>S!v}u9xn04|=6MJd_FAPWXAaR-$XS-T0W74%ycBbILk= z9+&Wof8cCuD)>of?0H+eh4LmQG4dNu4s3Z~il&#r!v^xfpG>&9Sm00s9A17bIU`_c zh4ZJKn8j9ehR`D*XBKk`6rVe|V0iu;`n55QeVgJjZYFE$_5VunK@YVJ-WD&lu7*-V z-iXIcG=iqzYiU8SmJ|&`^fLpJf|-SW%#otWLTN6wV_Gb(UubUes~)7H+t+5QB*a50@t z)_`zUl!OCvf1S?gdTIM=vbY7L)P9$2sU-)&R2V66DW2y%HDrBWhXRs+Tr@%Da~m|lqodJqi1@cLwr z+I*mMG2@$v6=rbc#BU(p8+&0o_eh<+dy(aT=7OXOy9!l)75_>I`Shy+Z6J%aL1%BE(-lRr`S6h4RaC7JYwFgcq; z?4t8I(HCi>jm0+sJ}E_4EX!n(l-&1Fn1Y>Rwvum}Om?ZOG{I)6B6s^B$j} zH#fvn21bDc{6An)@^^NV5^%FS!Q*BcP- z5tP?0rC@;=jzCb}pwGYh>?7Zs6@WeC%O+)$mZPpZM^9cPtU7m#UA`vA2{$UkU;(dJ zDh0y$IFRFnENx!IeMFMjJ}ZGYg*w*B*5zTZ2fq@UaPQXKby+6h!e{m;3{+zNYEAy5 zoIhw_UKFM=fDisXBWqvJVP3iGcn(+%ke2duOL2yC3}Ch&<9##Ne{u zD0~7QKF{8+rMQLN28At}$`>9E3Xli&K;1=#?aIi~HZH#Ck3z2H&T|+$vpS@YJdW>)xT%ll%4L8-9D!OnxM=)oO^S-lP=)$Fv7|{ zwt0V%3A@CW$-YSisJf8L*-{bacE+$h&JZQF$)%bVDGugdR z$m<)Sak5Vb=a5gRwQOuPU?8n zjP*@k$64wVRZaTdrk=3Gyz^z_QpOptj@2VyyMc1wSE|e#nMnM%%3vr z?)}NU7iV=$W3Nf(FsC;SeSVkVAr9?KoiGnd57=JS+*`d^;mTQ@2$E3L?(-pB-yGHb z^(wk+BfU3nF@N<+?Rt7WTlreXuRUXEiF76ICbe31{ldGK&765^4DhtA@vi4Z`^d6AA#kpuoppqLJ1j!0g*U7;JrwO&~t8sDwSr9Lo%+vr`TJlZ5= zX*IOJvu+#Y>iAV!nsVAi;zzTDmtN^9EvC_{G-}e59lS!0g}R(w6O_Bh|Ju?(kJ7-O zzNXy@%)&XxdHq?~?R85_(*Eh2jg)#HFo4|Ts?Ch0jFYC2O52I<;5cAsM;w>O7kkcp zU1+TGBl{`R_zLpXQuJ3B(DW=WB%QNNHJh4bF?8bLw!IT(lDC?$KY(g33|ZaKO*30> z?;Wc8q39Q69FJuso4N6aOjF?i3*_h+Y_v@+9gjUR$;wN?X!v!0%IIG)z`CMi6Bqf9 zDkyt%I#0mbhHXk&D0qhPE+;4#;cQnf0JQ#8MgQUT=|xuO1n$t=C?ks70ec;rFxVsK zhd<=4sx6JR_*p+g{{$Q)2@(kg5r|zyK+uvXN_TPp{PF*$P*Wqq04w6zo1!u~p^xx? zibb?vn(U@uANtE_Lu{#TU|29T$nzefsk`CEl`5lByN^${@B$)JW4VYjpiZ~V4pV4? zh9(jNt%c!BdN9WR8oc`HYSEf$hIourW>ds@+B~>zdplNeg@$j$qau(?Yv;@fnUSIF zIi`@~!EJw?Mud_`a>DU__s?vcs5i4|u!2ww@kZbb4&X$pd;Nt~xGmJuAx|6}d;M0? zTmf&5Ik`Hj^7mPlgo#<_&1s&;!pl`KbV&bfp=xWG!0-TPqT7N4Ctitqlm@I~^`J31 zGnQ|Ax?E-9ogZeeLnjqAm~0t4jblqc)|c5J1L;N=#>((ax-B7aXgzp2}kK9lIDq#|LTYWFx!l0gJLRxh~=oUAhr?B^Ndl zSX(&2%1Js$4?73RdXN*q+?ANAUD zx~z95wW6NSaU8)>>!Y6jTS0$oN;D_2%HN~5=L#vJkhauGPRz^KHemYr_AWV32XI;X zXJYwL8S~ij)T4Klg{Xx%7}q&MCNS0+08v69Gl)-f3_f>HC=fuCQVw5`0i5BeD*hj| z>Hd#JSv@)TEMiG*9!B=8H12)>xYaMV3*)>Fe)7zHTKrpa8Kf2a$e`)~A-kE%S(0(u z7(I~iV%lSe#qLFBj>SQ$i03*?_b4?a+~cEi?X*1vV%jmqJR*H=7xVjRYAdXYRt)Qo zNn;;%J8se&L!_2~Gy{@sVRjc{NxD`j3twc*c$^7NgAQ;r?v^JpcLP3qctU)9P#NnT94jVHkRD&ZGsL)Zr9M_Oy=}I*B0Jk7XaoGjx>}>x)+Oxfkv&=bvj7 zb}g_9w!a2?%+axqAg?g8AZkbiJm%y{oCd0GPSfX!fn)J_2rN96tuqWdnobL_ln9E} z+yg*8j1Q(NeHOQ$4{FGF1Mhd43)(TO-lwF4WHsgP5Du^>220_&o_UntLgthW0!^Xa zZih?(&oM%^MfePG;iuw;+5Lo#w7p6_DRIc_Gbj4rD79}U z-ComR6EucBkaPxL(yuPcTKI`+1k#W~+Kw=rU}mw4%-RBy)5%cW{t@d1(jp@%{f{2o z0+3%ZtQ@l0)#UsGo5Vv?A}LQ!j*M&Rjxss}G5f@;D|4q5zVIoFGlOvswNc?wY**nX zp=xL7UXr#4`PLJ{oY-i*ztz)&=i9PP>* z)`i6Dp%+T^TUpYYOcm{FSf<*j9(W=J4ZVFSrl|eNxsIOqh;Astr1oG}A6_Y;LQ+Gk z8|l5_F1N|eslnZD)pSUh`{t4?z_&Jzl#tZwxKifXCOe$x#$-`>8HsO5_ot zww0ZibG@QlWcN$KFkwl#IOLg+0H;%Vt28pXmRF>ssoqSt%R)h8eCF+LTpLfTajwKS z`BymBKN6=96_uD}Okl`D;VRfGDsRevMr$}rp{41_kJ0py-{@g)QdM^#1btp!?(Hu8 zD0-DnE-{ecvyVyS6W$x8an@@*0DKSWQUDrctA((5aFp^kD;sp)KhIXs60$0weU?mc zKOJWy56Tl3i~Z!O^AyYKwO1DZnyHoG-F%`qpX^nHL#`0%RlbNito_x9{nRW^!!u7( z(Aj|gD+B-^1g9MG<0HAzj_y+AUH-m_knn9$%VS$mt-Bj)3QT@BS8m3f9E)r@1Rfw9 zrdHuP|DDI)o9F!8kq6GcShK3IV*)lKv%^-J5=M0$#vK58`C(K|2pWj{`a4_B|CZcq z8k;U-ID2}{Qtt_=E(kj=;UYvyPRA*q>xwSOs%^?T5HUY?Jsl3FH#X{d-?*OYW?h?h zd9*H2()E>y(U;O{wTQT*=cD;!zv*#@ZeH-a%pOOXDdgAd5W4&%Ps~1IuCwX&QQU`r z3R;8=A6qF>>8?T`H$#g1oDWCT3$fKl1f1QcM@CJ1O11NlY}!Ydk$lI~@l>jEI&*U< z>vY8Ps*Q5Qr#0wh_T*`pf$Dx>#Cio``RI0%?*w8YRh)no6k>nUJ#QZXR)WD|#`$75 zL}!fhtb=eltKM42S+8Ua^<5y-t`#T*$MJv3e`jS)&3uo7EB-U=FRi_MtZnCKkU(( zRo}EWH#r)JJPO~?6H$?5B^6nH8J)@GkVF$EGGvU1EB@Z+eqCs8dFQ!CD0=(_^@9(W z6n|K?8Eu)lv}v^@Cz#H9#qzX%?K*vQ*e7q8)-Uy&?ZU1WzjSmvXGf3h=t09Ux+pfg63_f3SMeV24gm`yrlCDTxb@ z2W^X@C1!U+tcoz<=ac3dK0M4RWhbw(&-M|piLfjVp8Y;s7(9`uA{~e!+2$}~WNwt4 zXsvZd;5uS3`3d3pAQ@ZVOK#Mtuy(K3CKg3zL}R=V%JDb3|8IG+%0I(e)gLbX5VJg?c0s80YOi#R-;`Jd&bO1&0UulTRhk$YGIHj%t30fqJLdc3xN9 zpdE$|Mx7ozx2-C{-BOH5AT0@2=sAnrB<09o9t)$NtKjeHL^aba1LfA z=x3M?e*u^zUXo*OcLweXi2F~6kMg`{@`aykS%*a{WgnLu*VLd4BkkTMt}RJO9Gi4u zB1zkr+FTpw?W5>f6$2Qsy=5j`w~5rjagYh1IlH4ZI{JGNe8*em8mj#M)bR-Z)c{)n z2NL&ADVtj=ogea|_h4|)&3D@)o;nXS)^dD3r0gt22?&!Us;yf;7I;N%afIN^?a;M)=H|5rRFAoTx;# z88F3%Cr+uLDMIK6LO3tbfvN3$f83Isx|d+~v%{C5s~x|W8|cXT@gSfA<;QS_q#0$~ zfl(H%u_T&jwnyeWTt>9!?(VoTy)LHRzbNrb>LjWcUHE5QFT;guxxqdNgBMeqh7ZUp zKnCO`Gg{Le=&pv?6Ue{MfGpt4CrlK&VNam?Ox3rAWCv?#Rzz+a` zlu_HLYd`CgU1C(n9jfKiyk|v{@g&QKC$X|yt6ge^54CbAgv-JtWBP+`Kck}mf(EQE zSj7$hKF5uS2IA+I{3kL`RQ@M2V}4<)8@8<)eJlD-=D^Vddko4I`s3ZPj5zx`@~UE^ z%j#{uxZIEL#2~@RA1YKa2Ql#AY5Uv$#u&&SE9iy|7@?Y@Y3i8 zf95@UFUtf*)mykZL|l+C@Lq-e(hJ_kxvC5)@4SO&U-^*-VI@(q`kYQOn5>`K$43hh zhkUUnIWRK#(H-j3b75*`^&&)!{`t5F&Bs8CswOAsf&_GnczHoew~D8B1*>icwI0%J8X>wQ4+u!+cjOPBO@+SsaV0BNHtRlL82 zLZR=yc$CWpr*u>N{_TKxR1~fVj_B;G!qW(8WQVq-B08#e#rZbKNJVC4SmResQY@q_ zxHe!C41`LRQH{}C8z%UqA=Zo{WvAal{ogL~wH@ur1mcK3j~7*SyNad-V|4?S^(AK{ zr_U4vh^#PU>P>-!9!Eqv4gZIyuW)F(?f#Y?J#uu625ITsXrw_ANlEGM?k?#LNrRN` zE-6u((Iqju9tyC$f~9*5@;tK7b}HFFyu95|v-1hgO*yz4 z=0dm#2=NwM^FR7A{c9)()y%T#M}yECCrsvfKNF2s5j=6h@GX3eQ-8oV;LBIxHln4R z!F^K559xP~1he`#ugaR*BMsz#G9v~_C01enmOSr&RrIJDQrm%=44GSH{9^p*qfeAn zrS`cvG1=2r{N`g?Ij#4$zZ5C0LAN>f#nMvcfGviDZLOYka$J(7)&AG<21+uV_4g*v z&mZYTiM+nND`J(6_{8bcG>)*o3@C*kVa9V&vBu?m9zq|?qP^UXW|5)WIzy95y1FLb z_%CI4-a1=m`MH1cdgzUq&0={|_s4MliDjxdvRZb}i;%>Kj{h1u%XCx)+_SI~lqs|V zDC|y6w^bWzv_v&Et<@viMiMG&l{Joj6|7<$GVHqr6*zvBz$o}OWqhj|GH4`g|A{qU zMuBCjfjFf@WwsUjfvdfPdJAV^6c{c6X%I=2O=8A9OFNqXo#JK$xuHE@k?Qo|SzYTS zR`N}!qhfM!aM=4`Z-+N?FC1QA@jMs!r1qlmIq=ogEtk$KdHaXK``?$?#4DJr4hoSX zvx8ns-OTPfrM@diF;A$lCJi4vHk>_TCYg4Fa`g+0?p6z(*|I%62kf7ruI4FHjoGxGql1CObtjf%# zcn+Nmu-lzA*OIK>x#GVXfChJTrWJmBhSv2w}2veT=urEi_&(AsKfQ{6} z3)}hEp6fr44g=kitwe4oo~{Brf`4?$3%h_$rX7BBB%?0?dRR_vI@3Z4t@l^n_N44jBnZjoVQ^r zI5Lpg7z!bp1F!zzeM-6bN%Qd6*1SUm0GGDl2&E|%O7c`tdV3wxK}nA3)*co|Uy(11 z&S^mK;hdR^nn53TT2p08*z-*@HXU+k-1~P@ajJt;J>DgP{h!O7u8itRhI$yvz5h;vaIqt@L1Q-uR)L1;(2`y*>Lg14s_BPr2s) zP*;Dmu{($BiG;Ydi=D_w+<`8xZ_@D+6W(ezFi*stdu)F@wpS2h0eo?t`)=BPcG`S_ z%$zeABU2~#d!jSQ&Rju5G zn)-UYqHg2n8Cv_|{77U|<%u|R-sinzgC@?QwA|^f5mZ5M5<|OVNhEf&jN$hnjYo2o zK!D-O{2y{b?BNvam?vA}w3(dF)$hWcliFf1I*MZ=R7c<&QeOGbVKS)XAAZxTUje_c zx4K}_^L?5Xw8PTlV6ZXrSHbJeH8~?&dF^Qu+wtIckL=_(e*bh;lyP&2$XUPa+j%C) zSdoXVL?on3Z+X$K#rG_bAllF4UvI4(ZLP!=pHtM{6R11Z#=riT(v;??Jr#pA{xMha zbA&@xxst9dohiu_FB?t0kuw;#2E6m{QVw*#r%6&3NciCMc_n?^OF=B%rICWXa*6qP z!?`hNy|FHY`PESN6`Co&<)uSzB^@H>c7^`I9wT^h3}rR&+PpdrPVaEcO!dCiq1m^z zaJH1s=Tl)L#<&DR-+H{X3^?}03BM7s+(s;{BCR`~$905b%kgzZw%ZQ3vg{p7Ty-7? ze#WD*CalP-D1_KQEJrlW{K13RhZhP)@KVPd=P<_J51;>aZyIS+t-GRTeRkIQsM$~e zW(Os+OTW?8(3ID3BudOzz;y?%*C*%(ILJtct-hO@+w{?58D)95cv;;jqJxwd3jCiQee1OiA_@&CpG(8-jMT zg{8s-<~kj!qDFlA=NWtN>-%MfG<~M91{MGcYPCI*)9|75NUT>wBcIL@8ZmwK4CQsl znL~=Wb&~s}wtc=qU?_**0sgM@0{K}$ft&MCM58m{F1Bgi2itt0lGDFQ{1WYTk1g?$ z_C>nq=FBBWNryHghcC}#EZO~_z+;wU`cAFCPIg5SaZ{$y)X_Yl_+~m?ly%T3Vh}XDi)Vu(V_^=G?2DV#rgy+!D z%%GlNCP)O632CdWR0`!zxqvwEL%IWqm4Zi7jKB`Pu@ z*#nMRngpj#-(zZBlo@4a&XbMW%Er=W!usSE0>;JsObY#j$$;@on~^`sHV%8(*Z20} zvm+3v%k7QhuSes=;cX+EQ;)wy)pWDo528llUTHPnF|b|-C;z%xO2oP&i1Atz#usWo zK1Ta-x)nyXStm*9C>+(uUb&#&*zcE$?l6WQsHYBaXN9;_dd*Y^yZlVIeg>)Xt{Pk` zhW4+f=Xz%z=*2I%ow=aGkqsPxE@^QWWM?^6`;-p`E>uT?b9^7ZUYsc1eP{deZJPAH zA|bE+!{lzefby@_@OD=!-F7E$*6kgyd;YVxD3!;M8~w_-ycUP>AN<)*8}^y`OtyI@ zZyh?u$LjG5Hj5allRRG;A|k?@R5o7LOTX{Jh7m+sH8raP~c~ zfO0daX4O?Kf*JAa1O?(Bu8jJ0~B8 zf`;yZl# z3NEyUcC*BtZ}}|bn_2oMq=?)tFviV72r4YSt`aQ3sRDP8^H2GEo8h%{veQ@bozRku z9H*#{KL_@9hz*LVBMjxaP&Yr~YB>U0@n=5e0rnv~0D88&`nB@YQ*q@k=#+4Dl1Q(0 zKSfa%KfSWDn?GiT5by2$8>DK5DdW7JCEkz<=)8ffE(}@$abnl)u{aO=phmD7tQ0Yl zPATrPg(2?aO;?amHucB3cX9O1hYR_zc%JdvP)-s>I2tEl=+=x^eHdqYH(R@t zG+w7FaFXZ`s0!^>aWxi027 zA6AZ~8$D55i-~F%3prBNC~@675*lum+M1Tq_ubo+_rFihxINSi{lV7QjP_?K;oA>q zy||Zu=S=|z6kc$i1#*LceEk{P(?BV|^N6p8Y{r>jDsqNz-w3~*E}W{Z8xoOM*e2|M zun#Y8B3hyxh|h1aN!nccVI+rU_D*CX9vQ~*IIasrhHb2+&>wkFr-lD6&PDsP+Y-rL zy*_10YyDrfH4fri<3v1?iN$cgmcAZ_=QuVOtTyZnSbW^laE)sVyM|%Di)6^iUGo!1 z^8;Gv{gC5RamG{foZ}&Lf)${cbe`gUHEkW{K4!7^*#@bdWc)UA1#u-fjMAXd!J4p= zUdR2zj4h#aj0VLPAQFEJ-+c70-uV66$vXU;UrdHVg`uBk7^|23nEpPTkMhc2I-a+(N zk*_@ofciAb1LJ5C~$<0zrNr; zMmsjID(Kk`_UnzOw;&?7fAS}k#pDAKuZ0hD^}4Lh#& zZv)?#&--P?bTk)@U3A?KaZo*+2Qq5EogQt;%PbB*+I92ZWV0l|1e(}4j zR?XN!Cw>`oH;>PP>Tj2kI(#)vX~AR)=mKG>bXZ{rKpo_7G@GeUi4Q%=x;Hs8`ohjN zN|IENFZ6)w8Pa}quS$rauK~d*7)3_&_X!uOyS@50tJ-iB%zFjhlih1ebNq88&hlFS zO5ywazRH)}|I%GI?f(*-{yGxw6$3pi&%th&$ZAK-aq>wJYPMGtN{h+`CLvIB=0JyhzYL7gj1( z>wzEWbRiztCz2bb$;XJ(G^_A-52qB$wV!!=)HZ$T8})Z7YADf=-;Qevo#Lx({F_Bf zz;gA6`|G5!kQzzPJm2Q*{n;Atpz-vZi9UEfPEr85?P6w0%Re-scu(cSA!{!lZio;D za_C3|uXo4^1M3Hj;0t%+4$+R%XWSI@3T0(+9l1BUK%1%aKTCidHEl*Q`mC)Y1$7?x zt#Y`IN+2*_%r2wsoke3B6>LYhU6}FJRRs2f z?Ux+h!!}#Y#v6n|8Xwt+3!9Nz_e{(!v?&`QlglP!kx*vxYLabraV9EgiNBvMczJ$1 zDt$Pjy{#L1EFTIPUA)1vqrGUFDFw%$Gu;aw#;xzq2n?N*txcw{o=nm24*5>ZyqxK z%L~8xzl5A%#iDmOMrR<7yA9!tEuDE(4f^`Mu*f98G3PVz!Qyl+8B5d-?`bNL$Y2b7 zG}tv(dQ2aKE?KR;MqX&FvFI z+w_x~j(-Fl8H(Oe*3octh=%W}WY0y0^S(IW-?cMAW@>qk*s+Mysub@lzYOq#Ec*duZWfOM!nWB_6&1hJEf1i zb>G;bF}Db*X_N+bEzK>J5(rWGq?jrU?{jsFEk1Iku&DO;O8%(YCR3by+fba3+KtD;l>ve z73x*;?%to?F6^~}rXRAvG7FY`()*1PIf{5!(j4v-OOV&euycOE*1HA$e@qMI(m<*I z>xcg1TFFBU|vd zae{SD*#>fbeJe@K%6pafOwmaWsA2B{f%~Qmth=*?yAsNVq*6a4l zFqb66LgB?2Y;2Gf2aOBqvVV z-<8DE+ls8f-DzOwz7D3+^*op2b!{6ab6p6I#(}2VZOGxK$|J$VPW+(@efR2MO4SFw zq4mCJpOrO+POoT{aWfA7d54s5VsQIaj9#oO zswC8#nLt>>O~SyJCfQ!4CkBjByU&4FupdfQ2a@Lxc}HAXdRYyiFB;O$Jy0!TrUuzW ziTvoK-Ebc|!lZ2gE3>KPayn;n`pW0%ef6MQI$t}Dp}2v zxp6L`g6^S8m^*oQFJZ`WhFf3 z7E2)CA@%;EviV3YG#@%uFS}33*P^vI*2XX&fW59hoq30rirLX1v1OaWN(rRMok+GJ zXL2#G3sp>{S4r@^-{XugJg82_aa^s?QdBV8U~Il0zC)-P7Z>*2r9#sRybT?D5iEIL zW=lJUa4KW}p@R!9Sh(lChGP|qRP*=EHocNIjFm+>QNNvtl}phh0wGe&ZMD%$|88vk zeAgRpt;2Wm?Y2O$Rrz0-Tb#Vc*nE5F*kI66^&Gq2!0tgajMZiDXwsAwToY0#x+M6K z@+BM+)*(BME;Az~&8%-Y)|y{*uv@dh*YT$#6?)2$ha}Ze>8lRzR{PT&PcA@$uIdNl-e(hFxiid9 zFChlyr@!?TQRN&FizJAC$GbZ`zd6~5TR)7b0~41MqiiaggHu&?mgV=b6%E1Wbg1Fn z2lG#?m2MXU3cRq>nXK6EX#4BOi{y|=$$(F$n@FOe&oqW`1|3{&Y&z;+9PLEHOMQ*H zYiB_wR_vc|=%siBDQ}>M4L|dvFjovomsDC+80N3`=?9SQ&v4 z(EP>z^kwYU!Vr%p3jIs1P6%rxo3-f%Zt?rbV#>`rrPTKd0_o0qd8l(iH9^0z0gsKo z8g#>LVgL(J5s25?>L5%m7|13WQ`<3f^TfyO{lZXoDMnTI^J-Sl&TBdsH}P1t#>vZx zITkj-tHPgnvN+D)E-TqU8q4Z$RhJ&;ThPz!LY6eEGn9QPRbfqU| zmGreXStk*E*ZS}YFL^@R=WIePBvx~Tf?1CN9;a#Psc`}J^TuQ1ggK^fU>g>0;+lVt zjlW{x98KIfaD5ottUjgfy6E{h zahc^f8{J3D7c;eg1-|>E7y+uKq_rxvy17U$IC~%%OJ?{4@|b|ks}`{U+3CrDp=$CO zdx&`15qve;ZISA%4gc=rA6?w0m<2B;`yKV``$mTn=}5hGZU=VOoaX6{l>j_(wM z{j!5=R)Ot~R)!)fGVsUu=6f}_Pnwj0X+L3+lXr-yXEHuZlc`_86;pXV5zl?p3g_}*`GeBaCz$h0p zA()_bK`ZvVLW5^=iw7_;YE&X~@TK$725GGVS!9xZJ6szlJM*q=<_|~)fCPyn>_xMq zIVqK|VAji&tN64=+!&={O1#Z^bG1_}&8qqL=)5nDw@ND9RRogHLcS;&sN(~nTYEj} zYCL|sJZId8$2>B3EUTI)4U-mhsg;Zw|J!`{M#qJ~`dBXed+CXsXX>|ld=V!#v_bv^OCZ854FcP1r^@Mt4?+M=X zzoV9m;tj(><-qJrv?3YdyR6mT`_Z&N=m3{v!qq_5C2kH#nfgf(#2uMj7yJsFPU3Wr zaNSgP);~#nDN#vih*_47m{e?UeUaKG-gR$~UGAF8E_2HRt34}|Nx5vQA^fX+j9u(S z|Gv0n4iv=SjUlu4XF&M3{$7Q@D$7};BEAyF3N`bJL~jubpV0}v>Bh?5-7DEbcYJbj zBQif{md5$e!%63pQ=$3!`_0#y3c9=d+`=?>UsEc27(*M{cu&Uw-WIT?-=ul@gO-?$ zT|^IZ21z#oDu7)jdD1FB$q?ADa-a)w@&gY4#SX6OoGjXh3!K@>jf`uQv~GRPf6gg= zIiVK^`#(og1>l`u_pZba-mlUj&sV3oI&gzCx@~RH@jADW0Pge_i+dP>xQ|v+L+3{7 z;VE~tCr8yrY#sr-iAI}ir@UC?o6Q++TbK5RmM>s7cO0G?SHz9yuEocRjtA|I5mLgE zk?zk7u)qit?a@wo*@I?NilQaiOKX58R5vPCqXY^Tg(II{9!!dqoG>BAOZgnCj(r zPq^Rg9^5X02nx$7Nbn@M@W=zqOz3%V>mn=lkePbe+Y>P*do9Q(JT-o3DudtAhcR!P zfa(ke7Zxq~S=vH?=X>jONxCD>9$W*FNYhNtUjFN-GA_tPlIS0}FGuE@0B_)owuvd8 zmLqdUMN|*O7DIqq>Gm<+zoE?fZIt$wQ3{}H;LmC=nxIi71!I{%0kYS2o=R{&1Nma21K`sgWyWW1*v5CiJ@HP;Imez6ih?JYl)<%x*A++k&2kD+g^{x3 zy3w!b6n)Y>75^+Ctb3d1-71Df*mX4_ewXvDD``6z9;acq9L;f%xd^_Trezq2r$x!D zDjI*umA#diuX595+Xk&*cZOLw@mcbF5Zi2*fwU%`1BGhWze;ypzA% zeX!<{Jjtcr`%O|aU#}&n0o>i-sQ~!ooe$$$(njT?@(*56qr34p-!_UViBy`?;A0?D z_p&v9)U@mDFMJm5yheH2S#1~6+z%K(gV?h z=};8_{<3#Osntb&qQ!<7Z$&QR(R+V%<5aOHWcz@YBoaA@?vC_|@7O=opn$UX6zNsF z`$3u`WuaOxyaTD0z9XSQa{KV^01Hf8_Zo-cJRxswy;_hzprjelSsSz$*Di{k0_=Ek z7Q|IM_C@`lqPEwjX|=ww%g(=^3n13k@5lMx$@$h>31+W_SLP(4D&j?++%@d^>=(Q$ z^&UfaE%qj6j^dlj=ig*yvu$6Rv4QRMH%PMW020%PMZ z@-TK{dOI+nTSlGh1IFtLw@;>KWNgd?%4>F*0n~H8n6$9MS)cIo=bJ&-ca9*8#9812 zuc$vT?{C!X`9a3yUYm_-|0mZv)2IE!yT&8(4dvI!}2Sdf(?;N*X zGe5l0r$MofC-7PiAhzMgFvFWusCwvAKQ)I7~pV${{aGAZU%bmJ?)4r5to zP+XfL^4wY5&XIF(p?cQjx#RPNedH>C;+6N--fI1nVV|Yv2MmNb_gIMg%eOrPD4Kdp zxIO4?aKWE{KWa(twg0y0P{X{Q@uxAYe0-2gM7f%a5DT@p3{dN&hni3)I>P}7ky*`Tj!ImMUHMjk;sns6Q;l@{cpD@Bw{EVm>p?TDl``vhbmT3Dx05z_tyK)<0W8($}iXxu1e-f--v0r~$o~P! zX3<0{f!rwK$!E;~Rn3|W-La;pKUfHPw_x91xW+TZ>N&S(lxVF zRdj>Y5S~$K+pdC;2bJ8K9&Bw!VS-0gBG`heTuU|yki(#7Cal-o@(D)guqxsvyD{)A zz5HOmaR=BRR%dIa#aMvGoSo&CutmSOH&__*kaym8IQ#R-`~n2HY>a)&h>{uy#HnX1 za7ehHQD^S{T}D*zZgE`ko3v7QYJc>~>=k+V>E0b#WMZSV_qVgP0fF)zOvV|OVNzD-a-Z3M*1A0U zclk~Z$0adHqYq2Zkvt3Mr{sFY#{_X3ZcK2HUb*BG!f}vQKMGFRSg+ni?TSmfru*|s z)S9rASJ6oclI0=n<`-NS7_!^xb5LhY)TONH#(FMdIlV|)s=isGZ z$yw%2s(;*X<%uGcB*%rky7LARsi#=RDUu34VT0I&jU|86&P2<s&7a`5OCLkG9((NHs)A1&Z`=ltIKnP^3^)K$!JPrQ* zKj2xlHC4;TSPx{@<74}*<#IAYczP-D&7d$~{gpcH?WLGt;LPoT!mh{(JfFqbSpLd$ z)5gsF2y+U(oh1xMztOov_w)*N)%&6$hxA>6lxgto>p&qTAeC?@5rbzCZLnjB({k`0 z!y(^n-sXZ-TZ`JY?&f->IssDw_=_sfk4*(HeEJsxyR(XHRzb_E&9@7ncSa}8_xnH~)wfppRm-lKWL&%IXjQ%F|n=9x>beN2dSlCI} zZ}r1DMN~?EUrtM{MgGP)1(*`t>>QH0F`Rs`dqN}LngjLw^rHnMFT|WAc_QTpOqTUE zD|oaRGQFf)qw0h0Uc!E=@|h^I(w)FXyAZcj9ub*_zhj&-9Oe}Lgo<0W+F$Np;)c4O ztk|x-lrWy=p88&XAYvV+wsvVdvE2baq-e$yv~>^xjlx)Ayr z-vvQfZuX%%1uwerjT4IcDOv>+RQ~QY$7_ht04=C(%WJo(0U~2XW?fsxxqMK64lF?ct?HxjJq>`@2>FiF-9QaEE+uSbe zAL7V@7_U2Qp)2YZZ+j81iIv**A;1glc#462Txx?%viXTDBBC41smtndxV|=> z5%66vmL$gx%l3JAeX~00yZWYe5i~9-oepEZMz%ww%RBU@ysOojT3i%wh*p}VKGO}S zC=%Gq5YjKA?qqM?INUR&ZdPB=(Uo&dw>BRbT7{5;{6Y#}LXp52`1Db)nGMkB!Coi; z*%#Uzrz^ZM$5dsF02giZGd*fJpSbQ$x4d2mnIPa4=FnU^3Ym45>qS&+gq%V0hTV=E zGS}4lDz-QR-gsgGoo&Pa9x?4qNdiJ8Ah4sghIKPi_r7sa2(Qa5!LSGg&YnIoelUv< zo`=k4fxmOof8v7Zn`_KNQ06FyEF;=o2_20$vJ})h%WsWht!PBXc4-1}t0^u>uNaC+ z3h=H06`Dt%vixawmew*XsQ@f(#^wa(hmy0*yn6rce87w!!M zlyxB#y!O|Ct>Ha;XTu!4(k6g~h7!n_u^B?G2d4|^-<*aw8upf*yc<+>Mc2l@I9GDW z=(c?n-!?(}v)~<$a4PAH3*`PP74K}2I(xe2GI+Jg>jdiGBo<4C^X^?}D^4>}Jky_^ z*y_8)br8+>08lzWORC-9mpvw@&$jYx2kdG&L z2fnhQf*6UNnC-p%OBPel6vA$HSd$$e`QzjCOZemzF^QY&Ue6kCyDF>}4y*e~5A6RZ zmID9BzK}3zY_)y_V({1Dr~Ho221VIa8-S-k&IWC%Ts8`ks$^gEf;@Nq0p1&*V|K~xqTL~Ta^5J2>M!} z>!wefXXqss^2;4sL7yd)MgI+DRW(@?zml|??YJ|X&tEyBCKO=&0Hp3xZ1ltb!W%j! zPk?aMt;6cy&9ThKZgqI4`ea$rw5Y~sa0%pO8w+_g(UL#u^V9?MSZvx8ni&rx3QsMl zX$eatBYCLLE-_iHAnCFrqdTVn&L7o94TH1Sbm!QqdB!e#^Q%V5(Rcd{*+Bo?=GqLe zi{8|h#Z->tXxe6ddd1Jd@H4=V!2iJLWL*v0)VDJP%DB-UDN7Tj&qF%mV?R^rQ z65tDZb5^F>B0O)9HheAb1Ao)HvQQa^mR(>Ba3ZCIWgl;c?zh9_J-*-dBuK>V;iW}m zD8v_IpGo%zw~gZaVb;Qw?}+^qIGgs;6-=^6_Q67Mwc-tmFUy$gktO3KjYRgPSO{}p z<##E^agW$gpL*l%1Zedc+8cON?<)S#uas(&WV{@iD~<{I@*S!LZ)}oSvNgUg9+H`&{gg9^{JW8}uw9c}CHP|Ml?A+>>h67kW8gNe z!99F2wc&F6l~*X`tJ+(&LlcY-W*Cu)PeJ9h{*Heb{v~`}1Ukr$a&iMl znhAXQ?a3zOM z9i)bQ7$;3p!`+WJhNJs{SWiA9$H)-3QW;9dJg?o2_H-ARA6e5~w3*WUL=1lA1ct^C za<|G}f$=A}VV~-5ZBK7LGyT&!zKa#b0$qN-(>oE@ENX7Dd%o;4b^S-Bdsjs+4nx|*0`d~9Q?CV17=qoRNq($I8@G2m-4=P z+{KNRzV_t?tH}!2;_C-ceb3o52M}NuFM5Y@oxn6a^ydSJgp~}sq*)_bS6DC?0!AG` zGtB|%uCTm*t2TfQ@oy^j(>?hFt+tcS?rnP&&-}=QuEEemev@;nh($6)Nix`BoInW1NDgC0 zF0YNj&+gfvsz^+(miODqnOtog2r*_`*K+dyvs6@ET)gpC>tE5NDZfvhE#}YTMT!LD zDX|tVgG#&iIgK*3C3yI)u-gMpt#73Z<3l#@en9c!imbqEQRnP~m4LsC=76wtu-bX?L3NM^bpXgeA#gYi16R&F z$q*eg#c2lNc#WJ1b!&7~s;VdgcoZXBeHGH>co1+IH->2`WT#UdAI`8nns%no+ZYD9owT0=P>VR_9o~*nru{{l#WK3XL#u(4ol@Tk!?e={#CYgOY(GY32al;$B5$R3m7nohY?K#IKg} zgGb@-7j}sz6VthVNXh2&FMBQrS%2LVJZc){+MkYmmRbu}aO7IO_VGI~u0FXn8 zR4O?}WbPE@fs1A~N1UsKT!|xX48U2QOF0!Y5|GV}i`2j!+-a?L0AH1=J9x+xd`(x@ zLOb)Gk-D?hS(;42A5ortg8;td-z@#^MxL)51$f&j0-f9_SfoN6#x-dJ(eex)FH38F zd$uRMgQ=d}0}c!Uf!9UE@$g7VV;3jZc=|6u%q0E1^ zpymcDub0gFzQC151&J(rn{S1i^?Pp8N%cYxEjhLV@4hrk7DyJ~PFG<>e|nUS*K=zPQ&KZX z9(%vThM4Yn7!X%Hw=p%R?{_~9a@doH<0q(Xs!0TWnIz1?O#Kzbiop$#9NZ>vG}u}@ zb**4A-g4}m?>v@VRqYZ}(8-aTtK-N`o`Zgo77f61gqQz>@2u7025cRJzl;}d5R`nO zFYWbJ07w{g_4TlbOD$L{nMQvRNV;N(;!&zWYh*bw2T%zX$W5WXoS z!`0jl>EwOwblyEF>2zG(-jc$Y54JI&BiAe8H{N)l?zcqByTFu_4fK!DPz#<9k1;Ez z$EIINMOF@PO7ZFn!Dkz_P1{yVYeWVWsJL3;Fr%X7xjc z5RZ&l^4Vurf!EeP&2@*gZU|I*cfj0(J{e#W9B9(=Z+!yzLrXDpGa#4}Qm+8PUU)InM z9(`xfn;uTkUamNX2b!%e;Gll5SuHykg9YLhl5_*^B@q}kV0hqTH)^G_{cJn7XNseH zEaw6`S&uUQ)X{P>^_CcB$k|OtiHv7bLSAbb z!oB}=TJPnH@JGgX^K-YQoUR1%FJ_(hqUl2($(R?y<5-eg!!twC1Y*toFKL9YUJPP&z>yMlPNtHUPyj}^CkaK@Va2qjrAiIe^ufNC{z$k?UFf3oHI}$9 z(=;zKfpg;zhxqkSlO$|UD!@98L7{U9+muumpJ)yk4XuxGV#>q=>seIP{;&_jmqa3G za7XvZT;ogatQ}5cJ>uu#4Jo3ev;+DbkLW??Lv@H48_<=>5ImXBY|8q4VQ3P3!TLo| zSrj4C?1<7Jy$$xzr|3e8V(g23L;Q8#l(=%xNe-99FiYmpK89ll=o3@n9H5b^me|>h4E)`7A8w>Cb)?h^49&TlulBL7+LS>@I0h)$Z zG=!Be7>tJ2-i_>rN)7!(+}TSb0kF;aA%qS*wQKZsY|uzGtU9>&LwR*`G? z*2y$!%TjRS`!Jx46WQInCfVzJx%}sjug~eCH#cO=E5VB5nm(IIB@|2?Ldt!48xVW2jE(C3^avH@cTY!=eD|w75UeY(wUTy$1WpT=k1{7BS$Rmb zCLmGVOtMKuS&jdi8F4*a#8SpPZKPuMY4KtV;A~&C(vPyW_0_GfL$-5reG^pGDtxxk zwG1&uz-pMx^0MTvI$=@+ZiEd&9YhRfQy}~1izjhf(DX|3{-@AECJZqIN~v}ln(*;G zCqh$$Zw_<-1i~Da+s8DT$q0`6D!4}G1L+-%_UZ5->todYGQ1OW&K4nKM9|rgq)n_) z5;Iwu12QuJpFC@l4~coikq|*=1irw@r!|QQ^st|-_9343I(6t5g?0=nB=hr<_ZRz8 zaF>&LbygwMBwwd(@?W!mK;@$jDFOFfF)(Z8d&0@5gYU;XD9^jXE+x_53c%zNkz?|J zBKaD0dMXIX-YfcfJp)Ugh9G57jdz3_&2BlXP?>WXdIm)cJv!>rV4GwQTgWpTt+Bz5 z{+tMDWP#R2Ix~+AF*o6hk`Y#nufJK+9iAwtQ6me|-P%5X^c3O=uM$XtZaOJ;U}PK( zZrjRQ7nqwT2gkDdgakaB?NtPSxlX3X$O4zrhmk>oIkZB~r3=irhy`t>dDS}C72i8L z2#PWkZd&+>${t@c9HP(JCbKj!S$8TqozM61Ob9~ahzx>>h$?ObKAnOzQ=9piSBEig ziA!^O*kUzM!hWEZpUTPB%*?+;@yxEcbiHUhD?W{R`eo=P6go?Unk5~TG)-PSDBc-Y z8|tqp6uFYQQ?$u_epv9Zk=`aP1bQ`ba}L7LQ5&yV#SMF zad(H}Qi{9F%f0WtujFTcKh4KYz8bOs}m< zHjlCLeC#)t7CxYJk#rMWGLwMbQb*SvzeGwyV(1mVv+XY>Ma~lAxd7#Q79FvgUR?5eLYDF- zX}W&2a$94SMgi?T%n>9ct>Gc>A=ICfyP7l7LKt?|0q%vbH&t`T*QR zbPCu>>)j=T!JKo{pz{ z-!q|2EFzgEzY!z)o9~O5pvix<`aA3&{}q#*vEjPs+rjo(`?}dmx|5Ey4~O7(skows zv|t|`-|dzaoj<~Ds*>`g&(tnER$*LG{Y%JqX6snNW;KtRW2^KPGlPe)R+oGgwF+UZD=#XGSkaHVn$lw8v>rx)-yG#L-}tBXK^5Mu0$|y{$djK zXDK;402QzsKb&!qG<(9LBH!m6e0@H%VuCbmf@Jv1RSv@?21m*a%_nbL(bL^MuIUhO z=P-cwPef+VelHzeLTcSRJ(-0)^zv(Ej#Yr&r{oX#ftb=XicxmI+>qitg-S?I?SWt; zYP?}YeS{F40iJ(@80LjiYgA}zE|213m%G}5H*OIC;@H-j)ge^{qd#&Mw9La&5$;@3 z?D)es4$v(U(Z@lyv)zivIC6q4TT)|FsR0VZ;dS%6n7Mun=-^5BP4r|vwkLeH7R5VG zw|0JLU``$~SU;T0IQ8!m15|NX>vXg?-Tl;37MSxL} zQ9U&=w=YvQRw@^S51m_dr)01+%14My-nA~|wF6op(TcV-sFC#pKVoKM7dm7PPi!O1 z%QB0pnS|B}GR6aR&B#f(jp+H*Yk~!&TgX4uYdW6DeE!HMB=90$)T{0tRJ zcUMQ)PU=FXcxhwUYaEm3;YQPSDn9_1?IT8r)^hs!Oc`7F&`!tA(de+Q#AO9HCl|2S z_XY&{lCpPQI-16NOgmu{e*tpg(+eaYidehQF4maEJyZO#&7A|eB>}3{ef$&U3W?-< zyQlcLs}ZH}yY>2?c`4Hm2`25BQ|%NsXiW*+J7@20j=i(|GkGNBnx}d*eDMD8+1qY9 z4OC?|xf2r2ai8<_J_TAUb4yc7+`7(j%p*-DaG4m+WxRZ9cb{Hx=giF1q%t?OlT%BY-N%ehdl&?8t66wEUpg7YtfwzX*B(}srz=aqT#755IG=2$I&)wxv1w6Wjg2iQLgc;Pw&TW zB-x7gefez9ohSZWR$UO4ND{JibF@?d_8sS`gUNqTac9{7vaA*625R~CX|Ux7FOinc z>?SvamClH^&NQWs=eYtX6F{Ae{$}!I|-~?B@VoK2Du;h`)v;oY(B~o4k859Q2g1wj~iZ29rJP z5}mVoYcQE+(^nI?+)j&`779FLfRU9}@5KE8EUDx z!Ta~u;&gaR7V}3#-+e2FIl#WjBS+YcZ;ud%rEf&Lk(Deu$G+`|(xCFJKvLUw!up82 zLiK%Lu4#>;B<0kc+(pZw5{wbCpR}ptKX&`CHRtai0EHRk2kGbuc{Pi`~WKOvbTjZYmk zYW*ZRh%a0oaQ-akc#1p`xI0#U^7z}5TS84drt+ABj_2BI2 zlS(t8;eGA|9(D7;8OP^vRv8*xQG>4o2219b9|&d9EgP{s)a0CxL6QZR^cQ} z^j!v$G>t%n*q{S_+x#`dpAPU(qdo$vt~<)Yy-D@KcuKOz?JOE-WXiAk6n;0DMcfrJh@Y!V0MgOk{Sc~ z41IQ!>NjX2sUPY3j}s1%$wUVuR1}LCQ@1>Q`^73Za~{mKf3bX%`6i>?L3V@t&J<_f zBtSKw-9ha1^;HyLI?y7lAEAMB_ydryOiH-u8Ez32UfeKU31%5j2I-A8}N zHn=lJvBeROoLtnLsD!q2j-YgUw?iU$_dxS)8$}Ecly(9@JM%$JqZyj?GDJB3OkRH@ zlhuzp)=c73!!l)nKtPfHjHj)*icL1MbC-hO`ei4*EKzDv|F$Aq z`co11X~mq8({kiCS?AI3X(&?UgIod`uk~nFHYsMCWq*9LAoN|FfYYzFAl-0hoaAg& zcANV%)5DdA%!KAf+0O)l=;BCGK1MG&zJ_$rROOS*Z0Tpt$XS&_dv>d!+kER(*r4$K zuSoVEq1lEJGDgyHA%5?q?SUR|M!z4|rgT`Yqt!;j?vgfjPqe-{v+?V1MSGdDF4hkX zZZ%SGPYOf)@c5?~ux^3fV5J5;r94lx&yJupU07IcBjl=1)^sBZjd_7HS?%?wal?UF z=mwA{CF^0hL<`8Z^d^_nEIO`;AjNi^;177MLF2%OuUc%j?L8tS=D)sgsJ2WNp%Z4m{u4LR>Rxno%UEK{&dZQP!nYH zdyg}fzfgsX#2uoPjpGL|$}b4UXd+^meuRqlWoql((>H{Kaf!;D-r+8BQy1TO!_dNk)m*tPH~l1?-Bpcn-_4CSy94xRJ~_(2UY?Ev{AaI~1bVrH zSW>Nm91wKbcA*Un15C-N0K93OtHbPmCX`KEf?^T#(KOy)wFYDjdW+}pE6Wh#Wt!5X zDg^JY;~7?&9~=tGdX<7p5EOYN!K6P=3dzZbPcD5q@aJ*yj!hh2pJ7&S5T4GYWLt7U z0%!N3q~qqbg4|;jhW!~VL)NA}k%hjE{n4)vTdh3S!H0F3zEA2CV3mljNK*xCqT$&v z)ZW94)deP=vRZ`_zDj$7!06?jdv=+lhMsHdb7v&7MiP(U$GFS9L^JXkzCHEz8s))_ zvwEB7oOF(MMJ+&T4$|z7sbY?LRv7#+pt!Oy_~TP!JqrRP8%brG0X35b+?1LmR_~t% zA|Jspr7bHMq0k;0)%X|S=8IJU3fvqByT9|VP3PJudSExsn&c39`$V|X=W?hS2{^3BF z%4zYws=57h$|61kByM!r5E=cDAl9KX>X0$7GTg-itqyv?EWAX2z*cm93EN{W;Av`; z%b3_6;qbm>f}I(6T=%jvb)2z&6h#fjZJB3uC$-bUzbED~L9SO5cDqkB=4ui6J?r0O zflFX>`&$g>h#XHbx_D?CEpJuc9hr7v^4cNGF@c#5Z0{KI#om`xPj>hOs6D~mnpN#C zCqI|VK=YI(RQDW=>_xG*6wvSBBM(*XEv-lmLMdq zmaRAN8-$AK#tdKKeZ<2bJ01JdC=ckSThVfyQdG&ZVdElQ7N|y!;vt~M4R(h6hV#VM z&^vaQ-|C*6Ln53;&BER%_*1OZMwACoVA4IUn>^gG`Q!9PulCH_Xj29U>fiKnC7uBP zP@Qk2T7mp=RVO{zHy06NE*FcXr{_lmK7y6Ugc+VO-jdairw5ntAr!63cPl!HxQ}Rx zGeK|$GJE;+B6B1hjS6?IDg1o@?mQz-kg+>(xhqzBEFB#?y!*q_6RU(?yR3UX zxfNcX*mQP#sj*a_Ta=9g5sp>TfMb@xqnXE*# zYcAzOe(K2J1CgT-ZcoQ|^9QIq%^*t=HF%DAy$V^H9x^3UB01j?PDKdq_??N_#KVBa z&y1AbYewrC^D{-QP3nzNRn%CvAmcluIq)Ji_en?DgL6!c!cWx2w#c<^?W>$Dd77Db zs&h`;Wp-s@{nw^+C7lYr2ppA@w#J*~o!VVAj-B7uAL3966%-GYXMda+5ut}-&03q# zwcIA-Cc1B7x@b9yL*|9$KTMR4C7SGC2I5MO<<;pzTMUwU6H==2S_7&h!iyTB8ny2V zC=7wokm++9560wyy7td!)U`%{zhuw_dr=GIZFG|b z=lHl3{{%<^gN=gO4TlDP_!)c^Q&{qgar|&G#7SeZZ$GG!(pyIul*C13;1vEaYUkPD zr@V6OR2B9>?p+$IsYCqCvz55;H2fn#2mm2!FwZxg~HvTD48}Z z#zB)@B?<}p*C&E8*PujNNz;QfQTVPW7an*RbXUZ9fP!zVaQa)BQhp&!CtKE#HeIru zk=CUL2iYX%w=Qqf6V`k9xVmCgS7YR-LG4V)>|1HhJgqo3Cf9rJBGLF!VOf$=|2iN; z3QPm%d-Q()8}0Qs*`P6S%+kcRI@RvyiMEEq=Oy%z>(5>Y0$o*Kb|R>!2K=#Bq#+nPQjPeZOqhDs5+KN$i#MLw;(J z>l`25o#%&BFW;koUg7PyhV$OK{M2#r8@!GkhM0rj>w0io1c@-Tl@WT=;v+70z4g8W z!l?gwMkQPCoRYAQ-)dwo-1n9)ln&Ru?r7({HlIJa2vdg~yqz=q6eg@>d9q7|vYnpe z=UB-Sm1#S=wv^;FI^U!ZvOqsMS@z0NM|UQma(NkZQbCdC1aF@NlCJloDH7uM((|D2 zFXPB&kOIWzVii@MSxt%nbMkYaLKsgFWr4y>nnd4M)m?c(^T|uXjEBG+#ukI&E^26& zRP2M3v}fW(`j)Cyn{);Z z2oJ$LN<#p>VQ2+hrS#auitmV>LX2ERyI1&e-K|ZU5MW#|D?De0u6LzTOYfU>^@(pS zTICRs6Sfx3ha!EGT-u|DY7?N7d7DtCb4&{(3y=Fn{E}#O;e_zQ9xjq_s;-5TS>1!w z%VVQ)13Y(e>oc@slu;22*NH3nm4L^hiGD`3e4Y0$c9z_m4D<}ZbDXyu3L@fW zjc3uYrVk|2ja^?%8mj_E+4$~x*lf-ix+pufBm6UK4SLzCPUh&Pke2n zD3qJpqH#&?Zi<~dZl9L>R891k=>?LKZ}(i=P31-&-N0d6KuCgQxGTWDfO2aDae*f9 zi^E+qXvYtp0aNw}?0npfTftcI>3OVtf{}bWDICtOz7e0kF-U#bQl>g@iT%RTXWE>)>?p(6ZwWT~UXh=8p<20Rt)Z#EQqu*eN_Z{DD97b%VDTog?ik zw`0i)1o!(7@1!jr12nn2(B(?L2!7J-2g!o;)t4O1$jbq|k_YT3x>7Q4VYK*8^idGk z8GWBH(wj6D0j>o6)jJ2r8mI;8EHaF|1-0@|94F21U218$ihE=GfMDpXpkATkadKLD z){ris_@Ai=M!TetZAuXu`X24{9)hQ8zC8tGMh1wk8#{Wt-TMs0+I@H~f?Hh-UY+p1 zG4o0P(G}fu5<1U-T3o*SQIq@gdHDJ&#iz3F}#t1^m8P1`#WKNXAiMrJNA7NzCO z??YlOp0ZB^;Z>oOG?jRRv&ac8kBB_ zz_}j>Dw`vK2i56i@;iq2{?TEXKJ})J+dD(Qhtf>90|K;>JEhLh0QZzULsCK($v}dF zDtYRipFlE=n`Oe)8I7k6#@9ceL=x?&z;;;9Y$qKF^jf)MXmd!&KCmRQpl>c0C#s0q zr50MP}j5Taoy&_VO7}NcOuImDU7oDh}9t9^~(_EPmwzsPO`PjE|&b!sOyoV!7(jsHK*8}OO@H92ww!_pyr;BD}ql< z$33z(0J$4~Ki$kB8B5BD1iJvd7+;np3FGLs(y=pAm28NbMaaOejvxlxr!&WxX9IKc zVENzC(Z^DshWENSh0`WuM=#|PT8ewg@3pB zI&x0#yA2~Ntody((=;P>K40%1_H)O~yWM`8^wP7ZtDIfXhQEW%q3Kno%yQ7y<@n{r z{REW};{DZRmEsR(HReFrV<63q-Vw$m zHyEmNW7h2k)B}zw3ntiA4wm@W-Ib}Hktsn6sv(-7QQG(~sB=BP42!h?4lz0tN#TcN z1oKC5bs}y8F()H({*d0en`NViyP#dqv3N{co{!~9}M?!84)gb^f z0s0W9Rnk-%lHDL0#HCo$a=!r3GARI_-gOFXUVQ=U-~03Kx7A)Hl#=;xOD_SPk6aZ+ z&%%Ang!%Ra#|VGUrauWV)7r<*wJ%D=S#0a3ccMPX6jjLS%ri2eu*8UK={};LAV7`Z zaB^Sc=jiVDFWc5CzTu#~dtm;_?N;KS>3$93Hjy!d{HQdK9>9C+%86nUn#M%Mq!t|M z?nXxQLe)hr#e`6*3}uC-Lr?(*8&PEA2NaeMf)cNF4%aE#QG@0x!pg#Rmx>@odI#g* zdJ1RHbY~nGZU5Q#U7k z(4`!l*lQM|!FJ?4tL9vBX zIS%g^M)WtR;8oeaS^la#zOCMoSa(V25ItyZ+UfLBaw&Hp`dRiy@`!eep?0aFoV3W5 zkJwa!gGO47#8gA-@Y6U-Mw>!^BO8OIR1aR-%c$+oz3y#S3hs zd*b{%mh}hRPJN=GEIg_9@Z)3Jg#nH0E~+3lgt9WVC5n27mWR*mnN#&~O?x1g5h zn(cXOQ*vo;8Rkz4CM_cS8CTx;`Cpi4`fwWtAgh0Jyf21^T z*d}ww7pad9ZwkvsZA0N~x4f$Bl4_St#~>^p(L81eI$NAU~1nEhtSD zim@1+ts%hTj|OXynaidBg!J>7yWk#i;wkBL4S72H^}~#Rp*$J3Tj9>i>mg8f=d=3l z-C|~^LUSoCB@-cm6>p|Z8J_-tV1^{vRwZsgHNcMkpe#w)~7wPX+25{57IPVU>prW1uZ z6Z$AEKl2QlHmITH1yObv_?`A$!jOQ*gTju3Q1_hcxQ1<{igGx+27DmVA%=4kt{ywy z0!fM@O@(t{!d z5I$oADuY*3rSj&4#hnXMiJ;ZK4rg+hh7~gW3hx7GbUC-ri{Tak@gTpE(~(=8sVeOA z1~ti};U7}=(Oj7&WypA+@J7(cupPv#wD8)%i^5$3ZoB-Br+ivOusux`Z=mr6T-Nwt zDnp;L+21qU1MX09ZI=322WgMBy9)NMiVrQ*7{)JSq*d7(A?9g?CIY*WD|xX$fUoG8 z)Q7XntWQx)#SGjoc3wwtKdp5)TNVw8%ko43_gcOuFx?>%|$HTo3r|9jI6WTk7oVZ3%~ z)ioVNL!)l&ahWG-IJ7-Jy^P_Mw}f@Jvfk^{_uAHM%?JE$3=WA@c-;|sC(*0>=-Ojh zLJzVM$EfsRS(`s^h&$~JzqIk&nN~j?^4j_(tDs2e!(v2`BCP>Eroky{SD7=F`glvb z!Y0y?hYpBcqYiwz!j?-Q}o;yU_O>{P2%NQp^ z-8s8yK>lYpV8z2fR6V5Ak(W)%xG7d>zEA(YJDH*m{Glo&iROh9Q4{&Y4AMN94LH)9 zaxpmBgb%DC*Ucfpx~)v=N-y4ah?e-Z=D$Bb<*E?f{pZK1fAvR~7$mJ+Q+L|(^7Igy zFsObezTd-hd{s3%C+N6FGsHYLeVpYNKmUHI(<3x{A!U#G_oRs(U&`fKysS&NO(^0j z8a6jdB|#z`mHw{;J1a$!ZRbrFFvSRjngom`1xpJ3#iO2mE4%0B1Hk&G{@DO~z9YsKxP1kAlokCU%h~Gt> zenBXgb8zvP5+fax4hg^2P9Fyzi-HYhvUPQ8-IOEC;x{&;G9qSiXhlG;sJUous}Wq( zmdHQwG!k?@x5+=AzUx?w!&Q}Lkn|7NzQYeUWVCCKf)r^+P_V5^Z<7h{>!3x(DZwxV*7nfK01FSId7KUAl;spx ziHL8zg&|K*p9rESUVm)YFlD-#L+XVT1)A&f_(|bUn;ZQF4R*kUHH;GfE#g0e0U8OYL~5%Z4+Gx;zq zGm+|0sdu+e(-|YE?8_gWTtAg{ z65_1#Iy!IJa-nLarBLL#M(#;Z+d~hdCz=LSRJCy?K8at|@ zC@#q+Hk~=v<(%!5T}%;trQBzu@j`X)K_h8b{{?U(O+!p2ru*UHb+g3s23(=z_1*^d zbOqz{2dwVN;zux~8v{om_v90S-f1@8*T*Y#M(9)Ft;zwJX0Uu*`^TaEs%Lcv(VHDV zBcTbi=n0xPeFwpgO9N4tf!i@|DE}+er2Zc_JfIg+ymemO8E34$kvLOx+Q-L{)*t|M zpG75U`+a+3cKJvv0BeIx`=&WwjN`heLY6yVb#c(mV6(Sp9Tf55N8`q?kAKHHCFU42 z6#Qu4PXBBKl5Sc3z6&47HKALKyQ=CKx@7Xny$a;KADMVgALy_GOyR9+T1_#G%}@S& zV=l_gd{iAnF=~D({P9)y&l%su^JaOYu?tM1+xZvp!o0hlZnk*b<#0g6{=O4eV$iLp z4vQl=X|F@EO$=FTt-;F(6uwK}*%Oa4r`Ith6APptq}-?0j;O z`jC$J-9s8n){V^gkyf}}iIAaEFm8;x%Zt(Ps+4#lXs_5u+|OsDM%ybq5LsIyR3Jcu zQRcghJdtPAK9A#EJ~`zkLGk!m7+bNQ_dkg`A6Mi?zc*0nZ7m`?a4A0%&xjorUm4b zQU`vX#`z9n=?7*(NRy!Au=YZax2!srNEIGNv;cF&)7u_1>e_N&A8g}RXd4F>4CHD; z_wl}t8!@cv2l`w&Z0)vW6b?L!h6Ay0*Sz+v z4+FAoP(bRC!N8zcC347Tx~c#f)OGqrxj=o8w z)?ViQU_nxe)-LDC*F{_bY4GV@HIkMS z%7UuGXq;y~Vi2JSl`0MFVvTEYrCV6n_62l95{7#Iph@r!NF;3}XB#s~RD@C{@cU6Q zP?N-!=Mx~ecR9UaQ^CqC=~Q{_*7 z5R}n%NpiPmGmiAv??{KrSb)V-dXM$t8Dae;sqp#kL9J6mb~a6~dJgrY&1MVQ^*BIn zyPV}|17D{8nSOK-yydlVJtctaFsYT*1}%MQO?BHSR}DYj8b(>e30Yx5kSh70@AGXh zQsiznI#;7Y72F$Oeh*BDyj*#Ip@jULfwU+UQFaLizxe8{@c$P+g#ODM0gV zF%36A6-@)~%_vHMMCY+W(|R>z^ajr7bbkqr#uRr<=rxvFJbVX%CoT|^#G0i>7w`y_ z7p>4)914Fk|ATqM8V{jc3}t8d>kmC~K1RoN>R3W<2d6cM2zaN3hV7X}4dLU$gTG`^ z!^(La#Usax&-y+({^kEMb9ym46h5G4-2E4gWu9cnx-nNHhFcO#5W2BJaE4+VlGb$S zq*SdMUXIs}K&087fEeRQa@wNvjC98n;flk1K7Z|f(hkVLd!}3y?81=@oyKl6?$#w| z5S4lxFTMJeK3#~rOBO~>`oar2mProZ~_61$N|&Iz79bBO24+Q>jFAz$>W zpXL7KsgDvE1=B3QHD4~zdMIh(@mD<(g(j2z_%St<45N?2FPU5tW<4Y*3kQBdW;?q_ zEc>+So=};iP=@btIG7djVlZbt^_gx*sLN!F^YyPs;MD}`-8izMN6Mg-QR$RjB4Nsg zXjg_#1a!mQRm^I-=N`WeggQ7)E9WLrBIGHScgU7?wOm0!EM*bLV&n&m!mB#SiW&F9 z^F{2X?{^IUrE?!lPmqT-tRo3Rjj5a2SP7Lwj?j86eK7WHk67Beiz@;<81?RXJ-H&_ z$KW2XeA8QS(Qu~zar8}YWaf07lv)1o)Dm>(ZyB-E8n|Dms4RtV91p?pmlWUBLIqjT zAZ$AE1btWrImg+oN3=_w_6b!hgV{L$04Xp8!)dMCCT-tBBKJwcv`ZipL`hf937Vtt zE(-0&CI-y$ML2XumUcqFy^B_BY%D2haf*GR9v-(V@Xb+w`X__jnZ&A=@`cb2S^T|@ zTr^3dkqj}K;&L>kcw%i@lRhv9g87kBn7u@R974xE<|WP7TZZf}*+hGXxvSA4nMJ8L z4m8f$(AJrk!LpWjy0fn?T)%8pBMjX6$AZ5uTY?-dH-frBTP6}5IHD_zUD@m6c6=H| zU|MP}C>~BiBWWE_%H)I`8!G3E`4hNpFnj2-w%p5>4-I4z_xX$ii1{!XTE?q3kE#{pls%rE%ncgR;q zSARSvK3!iif1J55ScD-7$JJur{Mi_JTus9muatV{nM-A_E` zlyhsCdbdAL=YrytEo>{&`;-Mso-Aj2s(~kBAPxlkm-ZUE`=jdBu2|u0< zRcTc6`#ueIZhxLwd-VRhdUj;9zT;8M;qK(Y8TnZu7IlrOL_h2VZn5+J!-=|D_*1$3W;Y}^JX5&&5K)rzq_-)3k#<6<9QEgz&&%%_QsFz z@Ip0&ps^b~$T+#`8-I9&iJ&}4yKX%~1=+}&Xer>G!8E~luzb{7yVq^@irwH}1!9?%OQkFjJng18H? zM=13OTjhEFjKuhK3Q)xZ7&v1SV(M_cu6&}C@ELX$3WhE3$P``2dTxGaZD#;m_S&J5LC;@_q{!Z`(0ywVMrsU4FJL&7*)0Nn*f~$2PUR}7hzG)OCT%u? zD$lXuuAn6$h>y7?Y-h>WY=NfgXoIW!!ZV%_Ufgm0$o6#edQJI$l+IUTN&)`<@R-gz zVqn+>+5iLdFC&AFPF8c9*p*p5pDv*XV*Bz+!soap%L2#GJZ1+D5V4{QT)c8D(_flp z^ibgvRN6gRkqACA#r8O!K2_?u7A<7Go;~zCasj@nw-yVTTychIs-HiPG?jY=GNAAF zfJmGHP(df^@RtX0Lucq|XJS3)#UGo~j=0^zSM@l@CRiahH_AUC)|5F4MF6xcX=FtZ z{X_m&DVU;)Q-bgF2b_;cQu11ceY*t6JOw4mwQZX*R51X4&@$?j>L1(V`D+gEcQU#S z+}FB)ToViK@uvN;bB+}RTS`20_*OuSpC+q#_3uI4)IG$ViRIRi6^TB+wJ8KbdO4ev zG7*L5d)6*f*+JY?l@qz5%aS7cVSIK(?xLmx(grMKp(Gu#E0Lcp1!IGc(X`a?9-TQ7 zU%}{Md6Nb$r2=x@{Vxy8p9TyE;VvbjoqumEFi$!1hgVTFRvp9xwPr+pv+j2ARGi35 z$pP>gsXa^a8q{5TGShqgC+rwyjB6DI+s5zTAThmcz4>_CsUS$hSaB-#8HaTR9vJ0` zYNPGyWONRYr|BVbqJFlB`}=Z6a@rA^(s?@Ydc{|;E*WDvSi0cskWUCCWJ4>#P@M}v z%62(U38}iT=(J5RaFJ_=#=nx5q+k_|=-KECE~AA843exX%&ZMx90acFManstV~Z zsN)^6&p((C&=w zfW<6BI3P!xtVdJK4k+bpYjYY*8N~Rqj(-bE}IvK$qE%S zY2egNl~#*fc#izPbRZZ^JlHf0Y8}1|ajgU19jjx*`+h^9R95gh)5`9s{@B+J5x%QU z8Y8LcfofFb*Al@92y=o;u+FprIr*swT2Aq00b#7Uh=hXz(%vNx*kxWfvd#-{`#vs!U7~I7dtw z3peceJf(PTd7i?9>0LWi6Dxd@Uy;Rq)CisUn{3oh`*Jra z9g8~sv#tFcW-EzVK+gD|84X(Y$2Y&Lc#YA8dTM3nW+(Q`OMp{wL*m z`A)GS?wkL=dHVy`iIP`I7^r@{qFl zx^~gfzk^Rd*kAJwI@)X(vO6G`pUttZ-$k@yA_g7}_2|5WfSf{Lzwo%VB_xXlP4#FY zf6VYp8EZZwKA|cxi~F(uJ@mP!@MD56(Ys&FB!`inpOd!e7p{HiC}ood54Jiv+R*3< z_1>uQVy*8=JV%Q}|9B~Qx*_>{K3A36fKyl}z+)h@P#IPWB$}G9z8BZ2$t@F{V0yVK z&MM`_5(&`wp?G?rdBE%9=&+7lc#`dYO}e{*Kc2fuk-O#(wq%pf7UYR2ti3XkMp(Y z)cE_qV=)*jZxp1`(~a;*wh{DehX5?Fyv0YBXog_jJZ4ynG`ea$ z8n9-JNsa8M$b_Ioh?l~^QHnNdaZ;)^CVd*`@gN%^DZ2Vt9MpowAQWe*YeS#KlDPHhO9)ayBJItCzz?jm4Z;9sK=xOarOY2fw6Sne* z=Wl8#%`aIdW&V+`P)R~}M*sX15m9od&QSgP{lgLmq@dJY$4EM!F@`~aT_*r)mFZeX zsmX3(1QtSbOIRF*oM5qvj9&c-4)26y1tyVp4hXK(kBa((3zbHSG^E`Cb7C@a%CK*( zio3OST(c-+o-rF3)+?XdzJgkkh8kf*8`-L!{PH9VoQ#3k#zx@x+S2-Td~{kfGQH*( zNBvtIpzg3t5ZKVpcDgq4R)f(=)0`(B$ct|bOMC6wmNS@cYroN?N;Qhv5Z za4=ST0Q+z>gVKL|zW6L5G-3L5v*Lr>nd0)grX9FhgGV6vqXcJz@wQMGaSl`fs{}#O z-MldbMNpoSU!SN6o5de`F%m;0o84(z#PZrNuDot$U~M5O_AX=}6Sb4rXQV7v1Cx|{ z;z^STdeEwm;ikcyaub@g{3YW7Y(cj=?YpJRz^Bm}7q;5~ar0)N8_8S!>J4Kn+A|LL zXFC&DC>MCte{m(=7EsUNE2E8l-`{!Hy4%8jw_7-VQm{|?+ndA8x%%z1|F6u6t$rG? zTGH8MW^Su<+T*&-%aLivAjzwhq?aAVoH#7|rS-Jce;FhC9 zNv|^kRoWcIFDUn$X;VrjtTYJYVuXHqw}^Jm$Biyml~5%iXChg1jui_%5tHFvrJH+n z%a7w@hmfbpDp5B<$4&YdIcOyYRK5-9SPmblc?O61*Fr4_K7Ojsp=Di>VN=)Hh&trk z4w$q@2Qc$W+K>!+qB&6`Q);!3(y4z1pFSs7z0POzV5Xf-F3fAVAz9$|1;;-Fgg-K` zBh>lKq_zkaWHxa3^ypYesK0jZ-CPKzon@N_0f%U0ZEo`~!heb(v?de*6bS+q9p3|- zQA}@>uXy?{3{u~HaZ{zN(n?6e>mgq@Dj;)|t}QZ)F_*TLwxQ@_r}Y)jL0M(87Xol0wt^mPC=7jgoIK|Cd*LbV zAc)19Y2ll233giFC5o(YEBt6wmo#L8rp;V?qiAF-{RgCjpGPyFn*~+9s31ei@68k#-xc{eW`PYtQ&=z3Q9|`=KEw@TIu*O zdTm+kp~rwc0RWv0e;cmmH^*I3*ja!3k6FN=wNLZQ3`Jqmc(J2mSU{3 z2j06~Ab1VUVy0jBmH<_OyN}jUsNZ`c&Bvr!iTPT*n5<&+Itbc=DI)PBWQUXNo8KEj zrKTxN+gJQpmdEB^z0MuKofuOl!xqfez?=T96NhLnZ#;t@eb_&oz!|d-ME&n!RKP=3 zi#io0)U13*WKktFE&J_u+V|0RqwznzZpc#vQR8~QShpmWgIcs9zMnW>URYO>RL&ntdA=WJh7%U z#Tuv7-MT5(H>?$F*M*ACO8DJ+60-Bpqz!}@yP4}3FQrbZHJ+bXDM2~`i{p47dICa= zonN0_ehYS7;U~RB*YUYCxwXY)F~}O|#NYAJ^~Es@jWu?7CLTUEts963b#Lr!TPY6= z`5O|B2qtB*e)69Bh;nWMz$STO5H5aAUCPG`_Gju?Q|E0$aWchmvyU4l z$&&?`zCMQa(~y5-W+#cFRSvs!eYiCee5uQxddMTVh+X_K8lD{-#v2XU)2sdwZjn@Q zn@P$NbB>exGi8?`o*Zw5D$vP}@ymt)_b!%VvPCqhJp{oPFeV@@qV9J^SeMor#*TS0 zZ)zMPpwX{=WzrStG3bcyMpd8>!TU~GMru}vz+0cYe}fl=2IrQF1 zY2)1nOwt)iZjg@HmnCzf?~2RX98TsRkp7rSn7`3X+_(ljX1j29C3h?C zf0%j;uPFa7YE+OIdKkKED3KUa8ioO8=n_N(loknTX@miWk|C55X{3>m4nabY?(UTC zhCAQ)ci(mIfAFklt#dwmpR>>2tq-R6rjTwMZhp=!TWur-4OOV19}sUw~H5U(;xkjso(}@DGXUuXu4qJL&v9faZ=s znPkI|0l{(r5S|6Tqz1X^*bAuDFEqVSv*yw7ul@UM2-jmiuQSGEbEnw@TBWpty?AvY zb~xdLi6ud#8G4XuSFluaR`)1?R^G{7Tr$hJ;0+qP8ln#w3I^nH(sP2q1LAvI3jJ@QPiU?2Ckj?m&vs7zAixopoH<#wb>NyUS8VcW`*d{ zKgWTQ4NX&nl|}Zl-Bng{3LJ1D5w2&P?k!&ZFJcNIe{Bz~qwF_H=QJRa()mB{8F?9P zUy2#=_XKXmPC_<-UGy*S(l+NBjLBTW;~WU`F5os{Ss@cj4m+T>(wr(7VL-q>?+_Qr zUs)rfM1-qB(_#$y?~x{t?lf-41W-W;Y5#dE9Ps^U)Wc8TG7=xYL0g)LesOZRDh(X0 z%t#VYJ!_I4`z_rQNtPWcEhHPt>p6TiofJV2DqO^W6d<;9#C3e!#r&hYKg^ z4=#V4Zv_aPM4Mjm(z_+pNRsPFHZ@6`wX^cZSb>D%HRED zEcSsP*P7xz?h6(2fAoB@sq_9oEdEk_)BCc54XIT91>5+6H}dlpG76f!Cz<{3{h2>c zRRnLB)gABnfZVNyU8;Wlvn&_<9uvN)iD-S=PAfkJ3h=yi5}l%1&=C=}%x;hwtop0>kAiwzApMp{5O? zGG=dq*27Yb>q;m8;j<&f<8xi9cY~1W28JnB`^w@`f z780F_7H3o|=j8v6!vAo6CwKf)DJFOKbH#PxnUxd8ym4E#v%o%{^IMyX%t*O*G%c>< zrut|c(s=^7&8C(U#gx4R6)l)sVtCEe!Uxp6 zX!kTEf1ZFz!Ls`vai%QrX@9LM@bF>Y72LPS*5S^W?w{C{*YlEQL6pFq-#CzAOnU)O7RMEB2sbRsxgj@SSwvb}Rt`~eA9iqiQ%l1M`pq@L*uF+;?IdWS=!GXF z2`Gg2$+dWCv((l5-s}m}EN?<@Cl^0^)k(f2%{soLyqo27{VOFc`ulO+aKQbxrT_g& z5?;BN4R|^0z{=gM=EwM93`Nvu*Z=co6LmnBOnC>o&)<}o#)wyqVqD!0apwdHOcsdiOSXx1Lj8WLLW54?dwg{r?e+iTPP!ng@SAT>3 zeaXNx;&v6N1`pl6q>rli7Xn4boCVmAnE32#-A0oM5oGA{xrjZoGE}3G7j#pdo1n_H zDE)c%OBD!*SZc`}N2pv`1H3E{d;M+XgmPkOxmf@Mm6`-hB zxS#X*l)vIER&hQ`@c5x_HP4h7fn%hgx8& zf^e_@$XmGPj!psJN+2<=6pLk~2u`hr8o%Bj1{5L2BoDMWFS`nyytJZ!W9USAaeK!D=#`5!v*?q z62>om=iuAnaPzExi^~1OH-s}wtc)j(oJr1* z;}3d<$64QwuH+k#aS0vSG4c42`NlscG(AQE$XB0w0i#s^j<75OB#b?28Kf^?3VBED zjlF&Xrn7j#*(ord86zNoxvpVPt`Tv+c>eE55&oBF@6TVFWWz!9^T&SFJ-5R7(`?FQ z$Dp?PkHa@;8PHeueI}SYRD&o%8(qRC?*BUYkIVwtkh7%{yx8tZs3G1q?l$RM&|Th` zR~`Z+KZ5%NEq_r?!D8bd_Mpn}z2)TEu-4nkG6*m6KN**E@_EF+1Aj-hiS2awsE8** zPN^s-lR}Cnnkn9$RV@UwL?uWdXdw^{A>}F_QV#iprsRNV0jJsX;L);rQ%Zo2_xQR2 zYBz`G@kvIDmLY#r_83{MTA}pSSGyF8tFsUP9AS)eNI}0+GFSQDnxRs4XN;Y?%EY|V z&Uf03JD&b$%Qoa`Z{ybr)(-cv-_|IN$y>S(s@Kx+pK!(Rvl)AtEfqNR_b{4%EdK(} zQX&i|Xu&ER&iTa=c*=;-GNXnBN?!J5R1dq5>NbbXBme7WlJVYL?<;TKN- z;o|Qy27zPhSbE_9t>@p94|~a~%nx!^AKCI&kqzN{8Ekyz)K2ixV5F#lXzd5{(G&?UrHRsHv>#a%T-W0k&%B?YlV0`9+?aDxZ4+wdxK;grg!CYWAEwXHP zGW`iw6Qob*5~1E!j&O`<9A$VCH#KpO$J;K<>?gG5aeJf5&!ch9>}w+Er7F(F z8v_njeptA-mUA>rhF6S7r2UluwR)D4Lj$r7p+V}1pM0KN!bp+8InIZ)c`PB*=gt>* zySVf(fu|W^#+cIWZ$(hrjh3`iYg%M&uoO)&S6mUc;TKbRiiG?IiCsTNto0#RH42U#f0c-G^f%1JN=@peL_l;AVIbN%xW)$*R4g>G^ zB|fZGKq^j(;T1|Y+;v9**@VxG&|EihP}Zj$6$-_?ykcNmHWf5)_B&{Te(W_;D$_ND z7E4+w4a;RBKjoC?r1agu86~AtwZ@^NyA8bbxx2m2##4(K)B_m81ncq|Aogk3QRO z-JTPcyT3#2nRi@~pKVE6orAE{iAjcIFV_anr785$uHcu6&A2WSPf~m2f=v|N@v~st z;>K}o_90XEtZn71+}$58i{&i?eg2uow0zrm(wOOq%V>!!YzjY2B^u$tCe(;*y#{ijH!OW54NHKoR;; zFD(qDe6Vucz|6n;Jypja7{}YoM1evp8#Nm%kPbM^x5 z;Tm~AD75}lebjXPqfy}F9#V_3*tp%Zo$fs*wo6}D|3{{@mFI60)5%R^OPE}Kwi$8O zth4@%c#F#~q9@rEyPrSvF0RWrEp z7+C5m2Kt~Xrd-3~A#KtI;)=6Lkag`tO_tAr$u@#%+H+%gCT4-pRSO!C)kv1IgcXQJ z1sTlpDZ2Bj0>rAKH+G32GZj_P1dv=tKifgC2=Ge%YmQHr0{lhUWv8K6{&qZ5g1y-z z>oNM)vZB&TZX9U4ze47vAIbz@Y%UTyI0%mCOmrVEQcRhpk@`uOWk8fhxp+A#XnNLs z(8OW#XD*#ztcOp}&ka7vfQqznN6{J$=Z6fa7C0yn#w~Fe2gH}0P-^zs;dQ@HupZd+ z@ZS3Su;)5(CMdv8k0AWK@GY}UQ{i3AC=vhbHVlsrhNm9+Y4Se6I zni4~nRV9|#8=)3QuNwK-Kt_8bz-Qk2iYR~p-kwz+KsoFGXm(FmKt5!k-2ZT_E6Zpg z&Z{Qnl@zI({+{74GOWvMXI$ohTXs3EqtykE(AuL|Bu{f&@enxIFJ{{|B#f_1*dbzer)h*K?D3NE z&TKKEJ66B3d89cs_$=HK!fzOsRQ;P6*z5tjO6!}Lbr4rt!9FhxANqWSD;#1?^torw zyKu2Ccj}Q?ZWvH;#MB_%}MfWCVpGxj&Y7Rm7E5)m%&cBO||`tb#MHuG!NC+6MNDpLddR4z2ECe!R@y?_CV zE>#6Ja4~sRNH5;WaW+m_McuMgKuu&C@t-KBAGSKZIiRV9o$i@+%iB{9aA5M5tDf-I z1ZfZdvVOa!8+5##8gDf96}eZANG_!gm$36yFy#iR`u0BP4r`<&wp{Kb(hZ#t_8`mv z6>jUrD7*nqa>BI=X18w7?hc9tvZRnT?=WZVUd2J2%8Jm~MEgW~Ha&|fi1E8LTGzzC zY$YSe)DCLnO^~voNHgux=e%LU*{{E!X#z4W6lf^W5&*s=?E{?)R`-t+P3VDL;3;7t z^U^$$=j0|;CvHU9{T*_Mt80~~KHOpZY3a{Rr)QE>`KGxgLRWx1qr$N6x1}#9^S$n_2N32)m)*>yqY@RAuysAR?eYPWCm0I!gHQ8(T1;q6aO5LdTH=~ zAN!0{Zns-!H;Bxqvr-g?<--PQd40aO8H)aXo*KasCeYwCIDc578cAwX3>!vefGk1> zf^rd}JpKd_`i{)XFrmKy%}@1D5A%6MSE=Cs?pQUlj}Wj;z(6n`kj_$C>8X-KO$ahj z!_EgY^r~1t6H#m40ory0NnclD&oX3J|DKcuQ)eQZWN}?CdzX~SgB7cTm)?7T$4EdN z@rn+v!L=t2|F(h@Nx%WO^W?)sMRi}H zxx~qQ?Wq-z1&i+*uy46{<~m4rvu?w{4_O!Bcmu<@SyuDGG@o(21hg{D^~^hdPP^#S z$)8o*%Z-Bpk*~_>>>9DUWTbDTew%n6UB9JSe35!B)9Par<$Ry6dUn#Y$wLu6vxb_gpFQKUlMT}$PX&E$-jAeS*Hxnk z?tFiD5sTwj6*IL`(?vs`+QZf2L!j?KKIbCp%TtKJY(ZGq-(}>?$VS7-96|ZCrv9-J zU)E-S0jgj3QNwZ16#pTV1h^G#U(2jG-})L%_zjDk-JN;CR(Y%FKHORTBz`eA?L6XMCV%c*;|V*Z z`X+6FH?mCEZZDlgi@ju~v`&l)MNh5uIzAPJu601TfAL^bm}q{Q6{ zw`>Eg2^>X|`QvQ5-}5VN)DqBw{Pmh6^_ha@!+AXsWlAD)!GEy8BCON|K&XlaCoD>> z&opU-H9wWX$s=Z1j9qLsn*YovXM zj}-eRIR7g;8yL^d71Q#H>uiXnoTE}}AtS4@Rna2czD`^ahccF(+@HOd`RxXlEX_PN zOmNGBA2?x#@r!A+kf^K??8H$9nab+%ezUE z2XJ;Wxc2WvZZ8A$9;ntdCG#MFvweu(Gn7*j<(;X!899B_fK0$Ywil^W*tps3YB-xn zp(CEW!+@~%H@&=8%p4pRIt90sLT%c_Z8*^jv3fMLrcL1!Ei}&cIAfjyBhf~>op?=3 z*v@=?tl}rAh+0K} z^~$$JW1l4}Kq_pbtoEk`uc)kr2y5!Q38pXcH{GIcYL-oJEl!Io%-{N(Dtp9b>zbM(g$t)k*Ygj1A$Y$<~o9BGL zBqCTNsUqtg5u9-gZS-;KQZgSP_YhRR0CdirNKkp1D{X$te$a=y$!j+M&3Ra+vf)MR zfj{_$3Jk9-Nsc$PrO$kRBR-7gLQP;Ki39Vm%8T(TGWe9RVfme*B;KHV+o_f ziG6F%9`?ASPEnSWM_A7l?7!;9xkC&X=c_BUuJlD&Y-4j}^i|pJU}w0RotW_=k0uo7 ztWtb#Bn?hN-l;`02b3uPydZGc)T&$oEU>iIe8IdF2=@&q+nWMu&}o>1t5D>SodtZWk3(4bGy1vc7z!U;Tu)uc@>*4Om+_=fwR}g~r9O+}C8dDOI z&&Ms|2HxrX?Z_W=;Nvi0h7yhHcAS#OKQ_n!L+xgG2w<&H9ohI#c2tNKWnxp9HhD8i z4SNjH>$DQ0HeLC;3CjCBSctwMYuXH~p-)mdC`Ojc6Nj`@D%@=E3ccVLgZp3Q3TO{E)L4xAr` zCbjen6XwwCi>M%JfR4z7ao~oQOQW8Y4>Wx0V_IWlyT&eZ`}ruf)&yK07Jdme@!jcY7XPJt0=&(ycaO`CRjW@ z)cj?&xlv9OycN`=26c!_<RDd`+Y(y&Cw8)hPf*BJL`ORIsF7q zHapYEzR!geGG@O-S$+RAW3<_w!*2Zm9%ix>mm}T&oI&)sji$Adto)P6YZkrpp^0og z_UqJvELrA&U1E6r-p7 z|BcEP<*Bg`{?Ehy;o-xU{FzR_R6M7~#l%dLdCm)rDG;gU_nQx=J@FV>2Qj?h^6^|j zOaU2ibe{4rcR9n=qoL>^&^tKEjqm8%b ztyR9WWcl6!fcg1Y0k4WHlO(U^u(w9Z)119N@(J&0-w^h}TX4!)bQKIbdUZ&?0DWmq zoijCJ(1+u~S>#{6ajQHx=yh8cWP{4TgCA<^Tt05ykbnLoRW5xRKRQ;Wq;tSsCE~72NIO$=Q70$2%aeS zr`y;3RNg!ICBYmaNf6$V7PWRY^L7eL;(Sc{=DjDbBRiX&!4sEN=9c}HA~C3JZptlS z?PfCZiR`r%qUvJDO8p{>C`v#ebI01DvCUI3ZVsnhD)Sn!G%wfQht6}mw5QVb?_EF1 z#Lk+|p>qPZ3)Fa|=3)y=!W=}UPS(=Zk#_K$-(ZRMGFjE1JBYLL%sq08{UIB`Fa?xV zoTR1aV-T+^c)NZTLeO6lgW^~j&gT-tTRi<}t>-X!;Y=T7;AWdi0L}x93qM)R-*uvB z;m`cRnezK(JPKWU0WZfGSAkgHrSZ8{C)cgmz@De!9cZk8P_&E(FIpW*fPp&fz? zgvvh=g>5xl8beRt>d?>D@=&1`$WLi!wcsGVIfZDpI6&J+mYhB0lA!{QiBlYSDVzyD zJ2fB)eSzm3vG8hr;j~ITzG2H1S@cx%Cv1QOXQ5zYVD0Z@6s)b@I-5~34U|U_L7bop zOM=x#k2vb}!Vw!5j(Vtq`roO<$x%BxY%eku++hX`2**Ua$M1Lw(dV+&WKVM<{kNcT z9-gQ^)BG<{9jv)lhoj68QJ(QGPcw>GJh(_M-9gA!xu5LJiD(r;)QwcQ zPQPPt9pj(j2{z|s%E&BbVqD%PjNM+;xRH5O7mmS$@st#bL0-W=AhPL=UNN0h@Jpp= zRR*j#0%qc-3Oe-F>1_7jIKRcCv)j3pMR?w-NWvK$2j4`pAqd(&0|m2=2?8qBEV^V? z8}heW)qeUWzhKFop&NNwkLc*PuJ(NWH23z{2swx`h|b=JrFa-jK&kKqnbU0i_K*14 z1y9lgJvPEJ$TB{-{35$()o|F)p>xnFS-7?}6M4z;Of~v@L7j)RP&mY?^hjv9xLfcz1891&Drp8FpUX#d}gJs9_ZMUUTd zM%MlO{0jlS2y9y5%#JQO>I-K4mKff0yGSt1u`FP5f(2$l|HRK^gee!OLz^`UD4lLW zZSsG+2XB+*8+&|ZPrG-fez2@O>FLXPtoJo*RlGOdUo~xj<@#8vsEN#3e7F;2=&FOx z&RGqUx$U0aDsTEFaer1SjsEKJdJkm0Ip*lBJ<*J0U{d<-5BDLT++DU|nA)ie(|ay7 zC-=#QY1D}g$9>}W6uI2TBT4pRoIt_RD!w&!K?^CjthxsGM-s;QDKaD*jY1}QSm3!3 zhVuA}LG~{F*nKS0%oA&iQyp9QoMH;gnR(<(yP^8(N(k^Rv9YJoQ zD8^9{Jz#4!v&>c8r@~1XGw}VrG3}Uf%#$Y_(hd-CXMEvfYf!^aO+A75Y zg?2ud+E?hY^YFOZV?QK5h`^|P0vXhP3O1-nN!)Bu=)g7ep_S8 zWkYdxoy(9Jsabs0-c9aH=zHa$$BVvA9P0bg99Gf?Nq6HJUE=lva%S%mn>Y;AXWK*RhabuQ_ zs%^b)m^^WtZKQKe1iZ*mPc=lBEn*;?<5)R?GOfE}({-&%$0cp-!r zlPV^VLvP-1u#8M!%H@l(k9}E60z&Lo$@2)FJGU^s)Ddp_N>$3bj(Udvp#0h}MbTZ- zt5h{qn8c|8$0ba-kq5g@oY#p>g{>r018gp*+Ux<4U_v+0kl#?dz;qRf!PsQ}(%s+g zo=M3^@{(mbewu>NYySc^I>t9L7|pl+Hun!5FZa)QST#3e$a%j4TRYK~QN~0orIc}a z#xz|_SU8Fpw`BU;` zB=96y;Qn~gLC|^O74W&_=21;!#9k9dxn2Ly)-|Z0z*C(3j@4;sGp;-URRGcx$TGls zLm^=j?v~N{^V@fa45*-^`6L@Nc~Pl!0e5QAJzZ&)V2eb^e|vgN&!Hmv==I+Wk-8~x z4WjziZN9S5Wck7NKr`*BB&Jw)ypvuH7=AA*_A&oK&MF=Z`Wan9kVZHdEepo#Ayp(N z4UF$p&WRXu{w0{~EW-ZCHFt;`pKdvYts_p=46M%41M-TT8n$zn@F^`T(zpG)K7Y%` zU?=pLOBXRR? z{)wPyN{(Fmx!LM7Dp+&-q@35$r}{?ihaIfaJKEl_FtS=5dy?q=Y}Ew(R&D9Ujh@Xx z=&hWf-y836qXwe*!?lyx%qmsLZ;FDazh`6<({nVyHIijgVV)N&U&60WPn?HtdL^#@ zmCq~B1)q}7D}81O@DncwrYX@Lk0#Yz?N}Q9>7}@wKIGc`Sx!Vf$*i2Vnym3=ask|_ z4)#jyz#Adeofe;-p|j9mDe^7Mm#mSCb)J0hk-HE`zSGjydP5`fx=Vl1NK{(H>O-{7 zy!sLc?av6Su(kZ$DbR20#2GI6ORnj&CiBJ$&1V02pG{hi=8xYTIP#TwG2_|H=x%;S zUh1uT-?=u0g7DlW#e7urvM05E)>u%@HWqL-*=9ynBdSKfy4gR%9WuwkpanQ#i`qkG z93+fLpx!bhe|)tZI=HydcE9GvjXq|sG;6avv1TWr63!McXp-KCLan7cJQdzsR<6(& z3RRDDAtS?W)hJq$`qz67GC3&-RV9IF*30;?1t3>0`Ie zRSuj-G-TDh<7aRj;V-kQ-D9fUp|e3eWIb{5oL6u-My=2x&cnz?2{UIOL6a@87b1Bu zi85h6dcb0iXN9jn(52UcTNOsW+IZ(uym}tAqos!6S7ZI4l-)_$IAs=M;)D?$N0X?P zxyP)5gQOw^>@`v>CkJv8IYcE9-ibx5^oh?R`)l9dxIr5y)+ZW5>Xb>rFlCL35Mlkw zHZcP;uYuXjh>>7QHAZc+~}M_M{H=A8J54oGCkq1>NB} zi~)!fip2l{jxZ#Sy2IZAYyL2xFB+N5SU{*t7oVBV!W}B9JD?kF#$_(wolOtD1R%!i zI(3hSOPwq^nTIJXp84sAC_b{L59YJWk7X@ks|>9mf%HY3Nfp%+tu)f&khI7}J3nQY zjdw;BlJNpxhBz}=_=Q+pn*M9nudYg9eaDlQRmoWKkrszfvJ$72wkq(xwe55f! K zHA)1C%`ScxR;@byj%Pl;J%}=F9uk$G#@KX(1vOSgfrsm^6fIDwS8a*XNS}1zrd^dL z{4qQFtKuy6bAy6T`ZL!(YQ=-_BS(ZQ2|Kgzaw`{$9CDlN58h^vO0AZ_)jHb#0Nhn3 zl5+Opc>2L__D1(bOK?SCMB2lJ9bNF?pNX=CvCJbTy`kz8&AIGfg0k@~A;*IA;xy;Y ze%Ff&sv`chv+|MR3}`C&FWWgr4*8zN$tb zB05819JlRUu?x^0p*8%p=9i+w1#0{0`p1VPhtvBb@|Q1#{MlIqR_|1MYCp+oOa!9* zs+GN$@}v#vve-K}l(S#>ooGW0l3i)i&D*Wi(N)za!E4%axAHcUl2K+YS_gmL@oaz6 zC5AfSjU`ZA7GJZ9GpT=&?x!Pgz9)x#vD7w{cd_Y-CKKNN&;XUyj=Z}xPPz3z=~@}hg* za<~6G`JKqED~ah`tTXAze)QF8>lI(iy-A?7(U88m52I-SqqG!4QZAhL!l?%+P>0O8 zD7@J3T=V}Iv7Rt^dxbCe9=}KDFe|O1DI^lhp-W^CeBZrko^R@uJ$6TYoYXL~S@2?4 z1|^PBTYq!7djbpk64cF4Xkpo?XDfoYt{4mWOCjnTjyg{2_^0{r)bxJId=4nsa$e6T zbNBfn%I5*>a!v$yNlZ3zb^8c3iB7e^leUsyd#MNZF|kGL1^FEDoGl2V*B8078ejFb z;TV;!?qJ+DAk6~^YX1zN&o3$s7mf@apk0_~RrrxpWIrMF?4-ai%!|~Vf2cBizcr@E znd%)CfvSQPRBdr>@j>vR+qAx1c`;;P*P1V& zn~m<2%odjqj-dF*f&4+DIMi7eMTUw@#EIs(B+v2(C8LABs$vC3%BacV=k zEmkzoZMT7s8mH|EvkEB|y}_5#9e$?w<5HQ-aRV&NEzF&VSL#+>2nRtrDYaGHAHz)B zNXu5z*r-p5uZagAeE8p! z=_TRbHJAjrQ$dYCS8`}+RxYcDVA(OT0b2FohT}zSIjJL_wTolAn|VqqWj?e!i6Y&q z>8ho3vW*cR9I}t*b#*^QzVsq>t@^cQn*UA>TM^(Atw-%hfsh$e)&<86OsgYaMC+WB zoG!8RTGiRWG;AwEh;*2^!f@Njggk?LHtUMW&H!=>OO7_&sJ~bd4__SK1A9a*29>0O zim7I_oEqel7ovYIsP~mEsB5rr?n=m<^kvM0q7K7bW!Fas1Eapm^{=pyTd<09>w1S& zQe|WF+Q$nXY9fwrsy#w3Ywc(jEj`ap)A7@h-&-zSqP=NFnqr@3yuNX6#?X^_*y|UJ znFE3J5;quy$RA(n*5?*Cus*2Pl>ZG?x<(WxFkSj=&TxdZ8cw)#9Sg&Rk%&%1y zaA>_^x^~%Jz^bg*(Af9DBK98;alB`pUV3)dpM>7Mvji3OP&;`C`z1b>{j-=|>vBDR zW9saeV(&v}G_UP&&^;MG?Yk;>+_uBRii}^Pf8`(dA|vEf^Zu~*fru;H=>IgNMI%G( zga7(b$>kBbyRrreTOy_-S^kL#fHuYaIV4SQJlormm+!DWxSi4V_XccZI#6r^2^Xb_ z^8)*)nUYSM*@7dnjahDT@2?y5GFQ793&TA(r$)lmQ6$2I!88sfg<8AurgEqP)ry>i z(@FET(}P`m@ete)%pb3nhf*_Jx%(0NZ4NLz1~ORe())Ch@g8H$g+&VlHZKEj=4So= z)h7+dLz=Gv%|~f*4P9>eO`MHkh6Sau8a zF&RE7Zl%6lJ21-@bi2-Rj4YM!rY7k){HS?)-EG#9E~(JCm2j&B7f1ep!Pnj;*>O?V z&arn8TOxr}Pn>dmXmk(Oq@|QZWNNX^`@=r>1uE+~`gDiY`kbxq*W zMuoC+V<1^sKpN1zmCxm3JMj5RAGMUe)qx9CWfr;J+LL zXt{G9DK>PeYpwc+RhTlu+zDz8c#3uDwlp}@bfj2U2(T>pOs_7mK_s1r^zHOzFT2kh zw&liRX@HVi9m0I3Gk?gMuiz$Dov=RXVYH zVj}tloP204I=+bI*HvW9+8}>uia(1g7q;+?ADdMIv(8wCQ5XcOHTVdw@Ku|wjK*w) z5VVOgpv$sPOgLV3h{Ej={aAV>!{-}2FT|$V`}XmxB9gJZr$=_VK%Dt?c!kMzv?snd)ZuwOV77!h78DpR0+hM5r@ZYiI{3kIB=W`cw z6jns7fw<0tY}D1W3a9i#+@fJa%*xxGy`o@QRb2QRB$@vZRyOeKUYWC$fNIQX^Z$L57`%`Gv`Xcv%=ov&Uv?-n>ar-+dZwd_z(KrYrP<7!Ze&06)WA{Y;G}k zKEckhK*@{cxjc*nPR!5$%k*RUhTB;$_&q`8QHWItL38PaGmN6VQJHh$F$3Aq{vKAK z$Ea}YaigW)4baI2Ia2mr%}ozEcGYe}@wJKA7=u>5wP z?*6V^d%vy@A;Ru9v%ZvHx!uTCBdhlU^gG6HDx>kzzr~yVjbBuq{^hxx#sO_McXa|j zf|t9Cv57O!6mH9-x~SaS*+`OtgwYe7M8z&A$$q;pS}C4ecrLaf^klK!rg7~6K!3g8 zwYYdKAo2ikGRx`pwL3+`uc3}nF?IE<{djHFsCkt$LAD}xAI$#fY(FHoP8mR0s-^F6 zPwptf`txFy#xn9OYOVPl(g#-ZiQ=pBmq>>X%#SunoHhgeL^uq-V^pnGP48(|%WnLu z|7Gp=*EiLt42u%Y_%P{d5nI>F>tj=)MR#0Z8*_KbwMZd*^E)!7H_MD*Ud-e2>c@1D zbXX4;q||)SW{;yNTv!sJJ=D;3S~EPINGt5utYSYIu?IM%XQv;XZ6j`UJ>Q~)C=~r2 zU@D}T0c(6y@CdddCe)>mLk=*aF=x5e74P=*8tMoryvWB*X>ya_=quTMK@}mi*y0y~ z7q?gd{;bI4Wo$|wY3@+0ET%V6U%*(QKzE`-)#kvjQ}iHGk*4J-&nPhq2E<$Z`9-!a zEJ(E7rs|(C8uZTOlRS2*;v+7QDfnr(se7^D(H?cy_FQl89Q|U}v80PQ_tQzlJ#jvd=Ov_y0(98!QW zKd)A%B4EdvTtDtoITT>8=0PT|ZwV=`+xGOu3-C0Fl@g`2 zEyzszAR$C`o%G^HQ}<#TrH0GNomnNW_c?v%rDi^joK~|JdF_SKtNrv%lLJxx zo1{P*lj{F9nr}ZqAN~h(Vq+I?&Qb^cjzBe1$|BRbjEi@VjbtT}7Zl)EOWFb3cdG<) zPqk$O0VZErch)iBX70@APRaJ@-sQ)|;iCVx$Ni|S7uIr*Eem^I_Ya0{$n+!Lb4U>$ zh?m&zODuYh07SC40xFVd<*P8E(-yBoU8vf9{~jbw)C4LA&d8=-O@vc_nD>hs+J7?| zep~0ORs~izDT)$6AH_@j%0}EVcYy>Agye2+$PEBG@p)nTeA5tTWOvI)_sdT+?jk;SGcZsK6jCzyg?D1ofz8ph!X(WRKgfywcR3fFd|soDfV+9_Aq~+uPt?;-CLce^dOD>sl*cWKy)x9PtPu!( z{{>ik_$FhErtbQY@jko(TrCT39M7Xe5R*1_y2^7r75TUR8^wrIV|*FQR*7eUVn*WT zqligcBS-kp?zx<5rXvO`IL(|sE#T{>VueF0DGVSXMsZ#9A5wo7 z9M12mZAk0rHRE9Urb;ED{l_XtWNehVEgf9V&HCKVt=Mh!C*An@Dyw|xKz^nAdBB3rb{4DNEz$1iT-0`Qfc>ZQ|}3r2R0ET)tS7@?wQ7;KbwGy^-_i z`99~r-#kCV?1*sel92AW&$@}sFPBstJ_n#&-$y&VKl-#HpgTOk%+-|vO(E`~PS6V-@ciT3eN?7uK{WHXnn=BIOV?CdAKUb zSDkS%Lo=?^nnPdJkEY*z?p+M7Ad$A{Sjz4UI99S}2EO^DW17F`h z3S!!`*ia=WZ9Z^Kh;8YgNg=AcwY5w7ou+obNM?Tamd3JuWtMSe0)KnNa7||G2Bg#BzC5vtsOPcb zU+g?kl;#*ON2*G#=nC@piCwGIk7w+e2xKj)JCaxLWY_#-_Yc_HI4Wa=zX8_57p%N1 zI{GIM&j?cQ!gab5;2q~ed#9|t0kJyGDm#l-N53xQ#;jn+{5`&UO-KIy{CEZl_`^O^ zVti3y4tt=NzIq@FQJ4{Mo?a@XrbyvjsCg3M7;#AgqJbB7SaIu2Gn%}za5-a}+oheR zpR!PkmTU*{ULXR9R%jQW!C8x92j02YXjl2-Oo|r>E7fv<=^V_KpD5eAxo7Ew{#yZS zV9&QVej6NGx0Cp?)$p(rfc$`}b6_2)L*#J zrO#t|UxL{Xp9*EkmOj@5@Dq6y-CheHCXuV*LsZk()pzRe8of6lBcWWQuKyU_>r)2Mt@3{A#diYAy@=kyhEL=jyZPK}wSF2Xi))t} zU{MfZgZYk*!ib^v6L0_J04IlP6gt1!Hz=NFCMsI*6ztLmt4xDN-}R^(EqEwjhzquF zD5qYbbk;A|af~t)Y`Mv+++;m*VC7l9SSyASk{h>YVRqKsyW#%vFrE3U?}ougERPe;RONxv=zUG<~Y* zpJ?BZJ9XQpdGSRiW2HYxoT1GT*9^nxKNW87Sk-f!2p-Q+G&0_O$*WtFH=rofzOPVR zcotm4(HkRiK$|oq@M-yWb;~chmuAWjAM!0s^HoCJmvD$O8AMFY=CWan3u}y2oDBXI zjYpvuK(=>}w!xIBWDae>`~7ANY}!;kib@*Z_}sWq!@~J{QrN%=cyo{O2zu9*i9cm= z?X(sLe>99dydOO5d>w1!O8Bf})O^_QY%)N``}hC2&BOn1vT}U3II~(9cp_2vb|5~S zr`z@qUadITC)}q-t12+kGaz*;Ac;l}!+;9Blu-COHZcx|6X9kO3>pRB!*#D8Y0LS7 ztT7;KL8;>GA5-Vfnybb}MTmOcG-eKFcR@qRz_M%jyek9-CcRt}+4ywz|L}BHVQn_h zmc}8tyA!M}6sJJ(0KvVuOGB~Z?i7N%yA&xDFHp32aV_o|+}$SsnK^Tnn_T64zMU;= zy$c!uj~-{u^P>A3aTq+y-y&pwVdvAW^_L+sY;7|Ecti|M13SWh5TCk7SOpV?fS`6O;*l)*rXDQW=`_$V9;v^857u z+)`maIjx_X*jA*(Snrcy-@7>kxV(w-6f``ul0PcY?f0&%j$K>e%Y%Z4g(v?R#8dPu zos}c7N0=oM6P2VhyjkC6mgi7!VZ4vD#9_al$mro9hN<`-Gg9!eg%e7Gq?L@q8Wnrf zCH}Ehv{>8g0Zs8`|EL>&&tH&CL~U$s4&j1Z=)>sIqD&mtv#4h<>`Og$whkqt(SEoQ zqtUsNQ4sQTV-Eh3h;pP0-IauRLsI`lo6j@4=fEYf-gyho=myyZ)E!Eiqu@~jLtQ3q zbihA3fPHM;9QgcQK`oHy=qr`>QM|cIrFU-);st^4p&3O4_j<)!&sP?+4&(#+a{Z9z zZ)2-li#~V-6oEsN3&l}G7^uD35WN8}}aQmria5a$>0%N1y;80)% zuQ^tS9@7J}0v>dq_{|lxfVly6khSh<+~(QJ+%LI!wuXH}3hiI|MAb%?4}rjhbs%{- z%rDa$i(Au`@QQ6mC1J;fj*=T07a3E_*6}Oh`+kuxU$Y8T#7maUVlRA~hyp??SzAwI z(qJnuMQ^QybcSdUfs7D#CgxZ-5BAy5jADPmNC$Uet0*zjOK1k!B@5zG?dgPvQ4OQy z+(x;Mn2|HOaFO5r*?TVd+Rg*7sNn>6o93@~YeR?XSypHtqSkBJv;aAJqV$NP{@#AI zC7(|xSR<}z8^!yls#do~Q}*4BfP8zb&<*RdP2I73^5@yw`?9EDuRMA6raSUQ>b`9k z_%&};PisGQ8!yNW<{Qagx^kbu0VzR;iUop9bIi|J?jJnTFL|5yNk_-T8T%d8u^S^s z5l#i^U1NQ7loMWX{R48oxF>RnD_^Z`+2*yo{2E?kCUc@z%*7YKqxjf@#EdE&pvo(z zUPjy!k$r2BhN4kmkJqSZ`qD96{ok1=G6264W%Ylyvhq85&;E4BKA_JMf`(Z2R6@4m zX1>Qw##@7X*pptsWknHjkU(=7#~wa1L5L$-b03SD+1E=Qf?HAJyQd7gXVXN}4nsy^ z|1!HykN8SdEBPt8OD=a;x}E<0rO0x2OTMU(y9XxN5lUAIRK08$khzbL^1SdExpLZ3 zdIK}8zlNfoJgwh$lZ~AZk5_FpSPx1Q_e3SzE;W6Omkad{h^h%lc6U$J?etypvp-~= zu%+v|VJX&fWghXMWCHNcqy73;5xZ%vIMP2n=Ml=XQFA^P%MHL`|DZAA>DbnDeCF0= zz9-3Nz~+LG;$?1-|4S|f&}6t-LD=LQyzrap3hx=UrOiuRE_B8&P)Op4aIaq6{9Hx>-SFFJ5`PLV4t&7|(NT(WHUQ$w^O1vQ1b;wOM3 zHM^Z9AS4Ro!jc(QX{v9$<|Y({T1+T4qI+e*KUu9M>hS+v8%_t-X(~C!8el^?Pczda z;?0#m9TD{i52kps0IU6^KG_n+mtEB4!%o5a(%=AQ-Fo zD`5IDK&>xyofm>fkea3RnNH3e&mJouPUrbKC!q~lklBlPm)yKVd!5^(P2%N*r=fM% zM-D{z;nF3Dj$tfmp%y~Dbm|s^U`1vUBR__39)Ti>x!+rlP!N_5aAFUQS7R}A4&O&t zB7oAGFZtmz2YrgPGi3OIx)P{kwB5gn^6!e^qDztQ(#V_*`trF}PjLtnatcqb*8d!T z9CF!kiKEesiRdbxa7*1dH>DC6qz*eeqS^VgjRTD0u{8>n>%d`OC(XKJiBE|jR`SS6 zl~77eR?x`7cSBuS>Pk;-Tm<%sDbs*Yh=-~O?D3RwzDT-SEWVM;khl=-pxF+~2cjg~Ku4ay3Yt-Rz75amb)4KVIuo1pm2uR%am!N%# zELo78qXH%snKKqM&GR$Rsr;SgH14!UGgmU=ZM&7+YwdMcg3PXszM!f}GI2A$^owrf ziTjBAKUBarLee)s=%b`?NXP)UQEtf#{DQ6EcIw@zfFZi4-PbXWdh8~K3QY3P&<+>b z>MXiYR}`yv3#M6Fc)mds=)dCFOHN;uqX%woZAuP-g3>ANRkg!oJR;9=HaEnz z&w7h1OgULbdaf)Dh?^f1t3GJ7fh%7erS1ew9*>c0)v!Dff#%S{Rw9*Dx^fcFNuiQT+1S zL+1bI^~_NE@XOZ$$ZcFfJ7rN?8NTJNWg*|_L@C5ylmvLjnchzqi0kp;gOde)r6FQkT=zgcPd{UX1|_DMuW^h zN(=bX7%4$m@EDviYmtv&zO0rV;gwA&sW4|c6HzF|M}=gQ-KF2sJFIN=OwgShK4@V3nd3f`?1KSZ$_DgAaW}DqEV`QQ-3C7U(iM+mS_ELt-In^T? zS2GR9OT`h>_bL`APmh#X4sLC-B*xV)thO?r52o~LFYl6c&^gs!{W6WSPM!iPwq zk<>_SJsUn1!h2ua{KN|kz}MaesAt;xc z3Jzk7FM}R00mXxTp|d3}pezb-*uS88Z>`FR)sRy+oMTtk8?1GHND|E6Q>~UrCWm1r zavi%DAyMs0YvkkbI4(V^0=C#~QFsid<|y}SuaBgGHANJx{m;1XMK8PXe< z`O1S`auo9;f>d+@H~bu##9Or^Q@JHA9TM(Xn>~PipnK>9vQzOkRAnnm0#XX~PWWXJ zpwPaYztB#0#U9-rcC0*U9SieIRG82=hY_r>7n*$bKPNEviS>R!i$XVPq5bE)v10Fc zpx6nW#q?603ADtqSzvT<9XO;3987E8A&_eG`n|3wl^L%=c7C@1UNkJ>m4f$$ym8H3 z?tLPO;98aj7oQL7#Bg4Lx<#(9`v`m^V{%WQ&@@UYngVv-^Z8sB7ff4p@toCbB#GgV zkjS*~lSD|kT74l%s&k{;KZwzW(1!_Qd~ho;Gy+J(TKS z>tS60JNM>8s6#3gMdL5@+!$_YI=&)?zg@z854ZlC&T}0%RAgSeaK6T4??%;?5Dxx3 zSSk#DPWtwGXQ1P748O^N2rriF-e$C~PVD!@FJiYzm(ng?4SEX>yaR>0yNh?_z*;Pp zD7p-R%v_C<3z+sUg(4vA+r+}6%xkt;6Uo7!C(AR|IS3m#4$k7LVxPG`80vo2GHPf3 zgh=>$a#?+@n%ge5Aco;gtXawM;BC>s)zwxv6>*+W-gFpDzMZ<38X1a;CkGD-P`?f7}*rmLOvIQ$M_7gumjbA0Rhw%Fw) zbIVdxwb7r1v7x@FxH)drA-N@U61V6{ycW*Xwf4rlenJp7h;N`DO!AGnB%8=6ArHW_ zr&@e2f{KS(B42FWs{mWK+`W_aq$n+hO=Mid*h?WsQH4h z&kkfeXg5;2#QzaMNBAE904X8Vb{6o@%yre&yELKz4$riG0B0eG5u?fn#q(gq<9JMX z;Z08cMv+NqV3mn8h?VDtPbmgUlM;l59_hXc3Y79fK#w9XEy6?u4WscLBRH`VReC8J z+C}t4Do@M!OAcQpXrE|0un_>--7eB2(_th~r?q+;JbwO9xQaI63CXG>gya2$;R|aS z#?HenJ6>+dsz|88F5hD~o8UMTu&^n#1Rg^n9}Hz35^acLe)B5kn#)l|xJ`OcN zRHXD7dAj(eFMSMypY?wEcgJF^LK>{~FCWK!xQx^mvo>A7P{)y9|_PAywx>B7_@}Cgvy}yyCm)A?%U*#gJ77UC# zAJ50Uc@Nr!R%saH4r1rBKuCbH_=Id(JD556fdA>@-LPEMxwvQmRN?iF``tA0Ruu7J zlcN&LqNZIgLq%i6#Ru?aUcfzNz#Ub9YxrBE@Cbf}hV?mzM~`p#o5Biu(YdQ_#~rXL z^dI}Nr^m+P284%V1^XfUg?e9wv>#nyoBq8h-iK=p3z74?z!sk-4Guc0UG=)j^FyLvF7S=j?^e&iiI0Jf?qvs>z6Xl%aB=^e zcL?7DqO2dt*`EkYD!CTOQv}(J7&Mp)F{*%JA@ThUE+O-C45%7hAotq#_S=+YCL}ky zfnaCPN9?`-*KEc9Z?gq}3v5_aUfz1!;+ci#uEKs9v}nnp2(Czw-EQqkAGFiL)r0Q? zHbJXnU7qI;R@3>iNsO5h?+?$)wORXdfbBcF@&3PJu!K#Y~sHNemfUlT>CBLWfv zjBAcm9Hmu#9urFRzGguu`15-I&2OE`cT;b8@LpJj1R=d{7rN9M!}TOx##yS1H(mtZ zZhzY_ig++ca4|_NH3!d22yQ++X$)Ie1XJ`f-R)4D#6U7p*+2AGi}$~OUCK^IV%JJH z&thF|gR2uk@EXmV>wBZ@>lWYlB9E5UBrE%lC49opjOV(lnd%2(EeC_~pR4NChPE-9 zo~<=dmC~6=>duk!#OKMMOgQ>l%hDM>_pkFl)8xZbk8P*74xfK3uIRUV=bYqzt9MEO zz0tsDE+Oa%1&x9+xlpjl!)$^I8CW0i>DL~85d65P<$h8qNFu?I(V-+$q%_hPkRC#y zg0#l6^?Z+B%jf^(=LQ}k!>v!38#C`j+ehP_xuQ?;%HF%3miwHPyQE#VjlX(D+k(3Y z7m3)KA`UyfRN=W!iG59nkZI}9R4`|i)c_nFV@VqZv?wVCi~{4nR=GdfGMOffhpz}* z59CYvcJ!$Gg(x%XnGuYQ%zX!0kjF8xvPx#&N0jgKp9qc>iJ!@w`oszjc(L%}57Ho? z)&7!B8nX)E25#fUdm^voS@=|2)TI*hZTf_ktnJJ=|a?khl$+rCpj4=8CG2fI8wnQXD(o; zXe!SJ+8_cp5L0x-&JzD4bv4>{iP zEKSXx7Vq$9xmcL?gMnxcbp zt3YUgan3n!r>wBdm5SCr{I3?ER&|Qjp8w5KpHs%V9JyPF-5Ft_XcQG5y;#p_PC<@8 zW*_;|YaEb3>X*uw#>qtnVOP$pe0iJL&{Kxlcs$A4+Elo-S(MlO+2V#r|yv5|4G$^g>SS9SFeRdR})#dbfoWbd)D=#x&W6ywxb=ua@=(efK)M zeuqbV>p!Ek_9RML5z-N}Nb^2L3{{<2eWF&)c_kB{ES{bo_s7=s#TNXaDAue}`*6IJ z=TZ~8FYvIw#aEwba{rtye*Y7T%TfI;6CZN2n1bW=WBl5K-xyM(U_r}^lg5WwaSUJ* zeVfu`r|WpOMc}R}4&7e*4cd6XOX;D=dLpx8>bv_|JhnXSCiwSa72 zx6`UV14;&yg+3fSUBhP#c2Qa04CJ*3gSAqZJw6lhoZ5f4;%-P6j|_Y|wb!^W>%@)A za#^=~^3={AG43&{QqF+EI3m50WC{W7Fd={3iF^$F-#I zeJ3xG1RWlqpVi7i7OU_ z(B6pr`5*}_@D0JU=wUc*Uo~ZVMBKV(Q3weBmR&O6;u^KSG8SuoEK4>V z-y-#+Q7l-&(fqM`?O%*hm9gdQ$A~X{J3~>9dDE(X`vky`Qkbdpm4#Or7G&wIQ2fa~ zd>5b|d`||VFs{g`IN(aP1xZw*&8(v- zth+!%wvw)>r9)Yu1CMjaKlc=h0LqkSEG zshaTnHX@)4ctmnk;ul831nJ#9eguKl1MpY^#a^_QB0on4&Q4L5ALUNI!QfChO(M~h z5&@MHu;&(MRN`ME+d1M+CISmBBWryi`imiQ5pGhO!CeO)t6~9!QexiU?tc}5grb}b z(%brQ`tv_Rf*=@J(vfrwDszBi+?M99Ggw7a^M2+6rGmYwLfoC$V(i!Fm7YX4+&IUm zNMwz_6O;H$ZsvvqBtmaez;F+Tvy^U+>Hw%y_ERjEtvDzn(%LuD-6>Rwxa7*{lcQ@m zsuzAVBH@ftZKg($Ie+jJd6oq6^?d#>)5iomd@dP@a*!P%!GxD`>M=oPwsoCM2?ruI(75@?CSNLO~IemUXU$3V4K68=v7*v zc*h~!NN2cRKW@y>@VRVYsmV_@!BkR?7zfTcX@NQ**kGA!8Bs^JUUnDq?0@AhlUYK& z-sUSD57uKIU+bLT)`=VhQ zrPI}@u}vH^=S1E1NX}QHE8=&0b(VvPcyAShJO<$%GY=y+{ zbDqCM#Ww{0TPB8^CeOF6w<{<9PtUKvRmFsnzhsdeVIG>>lqQV7yb!;gv&o9`J@#O^ zeHIC*fNoBk2!?xw1^7G-inoq^&htZFa+r>e({wQGc$Jm#yCm+&OxTnG-Kwik0Uw6X~3+GFr2tq@7NT`M_`16@-kK>WJI~I)TcR}n_J2*Yw-kT8>(9@(E5aMkI zzQ7n1mv40bRw-Dbm&_Jsm$U``t9A|-)kqY#g8>z_2)NV$r)Wxzkl%&g0e?EZL1|CTh{OlUTKVc;r z=huNn{G!!<-QB;nw570q{v>B9vL?G12&`<4eZx1H#+JjRAi?(ri+n9H+B93X0MOOO z(T^PJjq+u}inKzxK>&OZ8eqTJWB~;6DOiObAcpcMpIG>iaC;^rWyz_b?a0Co-m%*} zdB&MH%N?BJAMpE5GogHIq+s6$*AxRMWP^2p>hP)Ydp^IBx={*&KB_*KsH&KLezasE zG7aACTUT{vd1PjmOYqmb{@~f|5J7-)o^=3?7Gh(2Q8N}CsJM%e!1=%;V9SPDKBz8C z;1}U8+K?F{4;}*B1ZtqNH%dqW6M=z$%}`{J2a}}e*7O^`fUzYE;wj~k%|C>F(UGr> z6TnfRE}&#|iftr0_&Osl8^u_7Nn^)7@cxa42yl|=zKG44ayG&(!sIA_zVD|JjOJAg zupioXUGc}v#;it{%F#)WCExJ3k|A|bZg%kPyC3pLC9<2o*Q7ynRvb=uOp7+ME?9C#b+?lQ*_fS)Td3v20y|q<(d$Wx+&$|Nf*^AoWxLaZ%92obBGCcyq4KT>*tCJ zOL=p7r?CB)az|f6i^keMexN%y$dgyzNMdy3&{$H(1eVVBw_Dq9Uu?G zj`LRQsgFpKkKHSeo?EbtC&nEB#aR8R-zARWKz(_LyXFwGWECCEAQQu9q2-rSn&+3{ zd-eQUT(xTQviBCYiVP77eh4`Ue%arnEL2jvp|pBs91l2ov9h?81^-*CQLA8k22y^l z(3vP9vRe`!fCh~)l@NV7%ql&rM?CKGkn9L?qGReuJvI+KR>m_*wS{Q9J2c*!UI=j- zGtESDEinqBvzM|>!ne|p36NV{-w=;2GMFu{rHMk^ffXKgHZyscLCbyk8Q7FhK^O2(pf@5cW5a0m<2(Da&W=DC&?2xXZDim zR?s62yEdiY?isp22{Jh}rVB@+k}E`NTBLQc%RT5ou496#-8y2g9J#`;r{~jvTE9Sd zC*s2~`YGwD5L3m)a(W&aCxAhRq(KVx(JOWLnVO}51c(&p+B1^VA^%c_tTZ02jrpwq zhW}~POzY)?7XiWk#y5gY!AEmk|asylWY*6{OcCKHdn%%#^9(>Ib^WZBRUpx zNKFve$gh5bu@PupE9>M*`=e53pPSuodRE%dn4yqM*!4aa!9nE*L&=y;nt2{fr=^=e zr68t2H>*IadYBG3XdXpF)AYejF!p`sOcj%Tum;gHzR9&D;i9*aJL)-sWe|)enm4_Y znTi^p)PT*XC<@9$xpWB;K_qE?XCTVjNK4}q+{E|D0F6e%5?vJQRer1;b!NL(V`O5DGzf zwV=EtgAJKr8yJ(7;d=50laC^59^Jm?mH<<^gWie8M)15mPQg#Bgu78v16OQ=jMTzb zE0uu$*v(}j%WYY6+ORuZXS%inOY-2?ay!w#E5vOeNI%5~!#`AM^% zUAXY(D;P#W$S6-)S;3{JlS-ONq5)%C-O57!i(8OO|H6Ot74+NxzAA}kF-(NgvVN63 zE3a#xscD7uLFa1}jqg2q-@&zk@j%~!*kT8m|E73fz{ZI*OI6+Ln?ORNaBQ_LbSYF| zV$2bLkS#op$2AN35$epeyw0NVGf`i$lC0M|#qKN3oOtu$tMVSskhfj!Ag>AS+0&l2 zNBgOczspc{-Yc#;*~m3h@;-lE%86stIp|BY7ThLuOgRKvTzK;z&(Gp$rBiSQZXnid zXpP22BWaq{i#?1hcW+V~`Lo=qr1U$G-Eq^~;7CjL>p*j~&qTIwpdB1`Cfi=$4 zv9?9Etm$iCQdw>x`7=T#t3j(orH2hGtGo@^&WBB#ULJoZNas%*X07T)9Mid(wpYxQ9%AK?LcG9kuFF= zJK}d(q!yA-Onlex^W9KvCq5*rb7^}zCN#!5x?>#QO18t3pVZE0Tm<85ZENPt?G@pE2`9?VR-7Mynk(jCu#(@Bon{WO4&syg zAt<5AynOigqP@=Ad8X)(7oquMkhNw9tufvG5s%oQnrQEE5PEWucZ?6B1TyG!X zjSEVtYSASD3)cJ*p}?B-A?oO=G_XBJz!p~6Z%Lx?3rcR@s&KVb?4DEuQ8Sd=_xM<8 z@|XkYbbmqe??5_UlU&%EXEe|5-o+{moHQyReh5tFD>#@|5m19ed=Zb4Fqb%Rb>r$T zRhAV~VLpGm5Rw__mO)FT2?XUvl26jmSKj2qI*ODmLW`(#)P8N^BEhmJ0;wG~!PDa@ z%zsdFma-aEN<$WEO!@nor78zm4CE(YESo%fTn}zNW%@#`Z?&3Q{t_3N#Bi6pj0Al0Ysn&*>yBuEZZdV;Cp$W*Bp`!>Q& z^2`>@0sCm`qRb;D!p$ZsBlU4)W6bG$KQVwgmRxu*ub+q%Q|Y!SS~ObWj7=W;o}QoH z2PA$<2(rXmdL&ppA*#@pYFB}H=YL7YkkUDDd#@z$JqHblGtq)nH1PvuFqcC{4E_*& zE2V#>&Mmz|0tS|c4({eNzL-MU-$vW3vn1lMG!Q|R!?Y*_JkmAPT84L&Yh^Ov?mI&8 zmb#bJAt|gtBdi*G(NLmQdQyqiP6(}E0;DCA{tv$+D2@I1v)7DKo$+ae3S_#_bQ7P5j^H#>xUR3{cuQkT9!OYXv z`;PAFXJeH!K(};ggsHbCu!4)da0VOj_D+AYmlx_ylyq8f6*yC67Rlb^o6X)rI939@ z)7htcX0ZPiso@8Af7eF6bZL3SjCsA!5XtAKg1NR%_y;sjI?(7CKm9Stw4?hj7OBne zFrh4BgJQJY+iaw@#01VGuRFQb3vnY#wkhv7Xz+<8hoiUeWwcKD?HXZ)62T&db}n}h z%H2oWo8_IG_a^u9WG(+SY?gzn6#qMMjzDRHY4b0iA6IpX&cs5J)w~djZMnzM?Pv~> zJv{c8V;t#TKe(TvEUaVf%to!~9{F(nE*+^(NNLr2FiZZRqGDhi@8~KN6@c~eawppQ zBC2_{D1B9+%7aGm4)MfD%EjOsqn<x?FW_gYe_3Md-aD^sn$j; zQz-?rdDiZ2E{X@V3JYVR<#B~G(9+K0 z@>}EWwP#fix`=i_XFhP>EP6cdg-5YDW z<_q>8Vj_)^_ydeu;UErU#1?L=1y6W*-`xQVyglR@58$NjS9Wz+GVwa$`>rzItx?q} zS@tmEAr@K7`n6p`tQqRuMCWs!YTYOUUV3-)2-Lh7yo}xY5W6AdfBonS7m~}sM_cS1 zlOl2><{7H=8M%f&Vsv`UMI`1-AX-{$6`QJUr9C`da9hH=Y=%)d|Cm}`x`ACnA-|6V zz1JJZCT`*r)pnK3`J!E%nlf~hvR~WC*b2>Fw@Yw`!(yc=Tb1p}oT}6FTYS~UuYeC5 z<(Jh{?_!yBUfw36rHtjz2+!IRSV2c(H|yCmDkgTa$v=nsDJ$TV? zAi{GEw1c+YcjS*W4MqUwDFm$VOkTaFD+yCjFb16oMx?%7hk7CAs`GsPXrh zmcNp%RTgQ*m`>b`MI}1b%Kzw4-ZGJ?(4L_5!E4%Ymk%@xa0oJ>bF}LmUv$lx^`rjn zm6|d-=cw1FDWm)f8wLdc#0~mXRSCYRw%#$kdG>M5OiB@#6ZFPp_zCfERdW()Y!={S z;YFc_W#B%birDZqa6Y(~#C*EP( tH=6v@*=6?3N#gI*gcS4{f0E?m8=J=9Vv3U6 zm#qS=+5Ep?9?!L{q`jJAZKw}jknbeWE%S`|vMQbI)BXMf6b*{S zyId3X$aqZ>6hGVQNP`~%0{CBu+yAk0w~IX&t@*UhAFf#L1Xs3!(^rZ=;U!sTmcCkz z>i9o~i{Ds?yLE);yuNeqE38LuXb)qPDT`xp!aDv^Ih}8yM$wC&&>Z_UX?~fW1O6{3 zwL@FT=Aq^)l7+Ss`L(4@_AV?Bu8?eRlPuJSl<5`uXB_x;N_a`aNhSg@94~SPc#v8lSU<`D%lP^ zyCm8EHm~8gKP8-#onDl3tl+rZ%ELM=s^Js19f#k;+@KNajS57jFuP<0Y|RCwc{O?F zWgwgL_B659c1@j#1=q6v{X}Yc`?ZvvjBU1x=k}2NmSF0tNoZqLVn6zju-Vr)vP90B zq$UI?$14%$WRf$CVZUDm6I4HOZozZWIfmNbv z(~IR_03u1hqlF%z+$*r2wQRCDIT6|D$dBUZsMaoItEH)3Z*U@=UhNJxxfHyPI(lHq zBv!%SmqlHMLxFMrFQ!m}DQo&*_FXpqKkq*= z02FMJRe;|U)+bsR8WS4ZBekpZd+<3><2RNN6=Eiw0~6Wh-3N85^0kA%Nvq*MnmXH^ z2I^MFpTUT`R`kb6q*clUCR*Z7l+{1CpJg#c<71_Tf;bC7B8w zCI7ukQ1?bhB-5|A#j_0kg~Qy9ej#wc%iivzJ_QYN-$fe(qR6N3bWFRk$t^r7{MSR1U*n0YZ}MWJW(Jfx$ZrlDoyi zb_T@7P|n9W-RPt$-4Tp(?i^j*V&eGi$pg$nPPj`$sWF@<+`J?rXSj)iz%R(1GKXC) z3*)f&jr0=of!7|G6Rj0W+HDSoG;)2HtVdt;u+rOMHuozT78rBPze1WD=JJvjF)5(A z1=r^P#g&jr{*~_-jM-3+ZH3u@b9axBM5VNf|0BA&2DoHeCx5$YhqPcCIkuPHR+u3H z2}8-VxD-W^v{nA%13RKQS5~`Q;P&M&$$7!IF$sMcmpntsTX9q0F$ND|OkPIF_gtR- zK0eH=zW0G23r#C0HVFDh2L^{LEJMzhGn)dyqS3$*P(K9i!1f0p-8jnr0?n2WUKdS`Tv+O3% z#NIoIhT{a&Vwr$qY#XfjRNZPRQl}F_1)aM-jAGTu)Z|!cZW&@AJas4j2g<(4wj8JH z<4^rU%-#}^);DXZPjI%{6Gx#3^n&f+wjJjWuZcjzH1*G?M^0lu!B1E*JZ!%Q(gJ-< zN2raA`&3>CBgph7iqyhn1ub_2tDmNkB~H}8-7?4!tG;$20E8OxcYo}zJr~ghoO}JZ zM|;JIHUAq=KbX;nQe$tR5Bq4Z1R$vo)S1G)+o>N(=gb-E9uwpM-oIR=*B7-Bu& zN-XV6zD!wrT4%h2ZxT0G5C=~Q-^y7C4C*+#TTC!W7x$*SRy2#yY%Gb*Hq5W`gW-gI zu2Vvq34r|`s8gl>QJ#xZi;fvr2N(Rp$E`2+__!uDjN0*F;222wa(9XT@%p$If^W)l-g^Zw+lLs zi};6PQq3@m=HDT|g=%kLnv}A`(el4qfaH(qRK3Cj+aa8r67Lo*VNz~UfH(Ow)NlfK z!?vDE8B^^L{w*dqL_;_Vdb*r_PnB*iwYnyE2-3Z3G5?(&a?HpMMN5_~M`aIE4+2{0 zfQsoGEB@5j7S0i@*#%*PN7aEb_@BT|L_5|We2BCItH%m6!=cA|YC^|X!Ue0SH1I$Y zdkL7NE~0Q#cBrx`wliVpUxqh@BN)fyiPxg;6YF(GnR>Eif)kw58esweRH&}T--w?* z*7?KO-@_r=A((e1Q86{3;1RV0Ooeu-Gu-YTdWy&r7g-%0`Ic}Foe7o!YmUY_Wj**V z0H}C{91TIjOp;ME8q9NNW|O`MRzUoT7Z}LXO723y>v`PBmcd1}xn~PKG2{7msLUq7 znb_^tz4U>6?4UH<(-9YQwOU1Eo|2HxIB6lfy*=r8tW(9>#?h zt=-@T)Uw*EO9AXa=+uMMa3VT@BXSb4Lo*xj+vmu6ZecyeC!geVjFr!2&IDVeGh2ud zIXjo-b+jpm$uP7fpLzqXGk6p-LATUK>6%~#<3OM_$zLZ&?K{`=C~@thKJ z>H%whJ+DI#sWhVkdzNB%l1PJZYML|YC38EiI&YhH&L5sR%rhA--t3(aUS18q`l9ku zGF}$xKG$3J5T(B$K98OjjQr$Ar0nbZp2u_#4`R1*^(N75!)}S|B`S8!7LKjC2_#1v zL!J_$UUREbNdNuca*q*u0@!|?* z-ZYMY^$Fo{dN)=>;{=>4g2SYSq&S54zUu=gGy(V`z8E4k$xCDi@LQ*>9i7ByIkve~ zkk#>0^#^;=g)<(wl(^uhQ%Gl53Q_Mzu&9ICtDCtaRksZgE!sxd>_baW+X~9Y8)2WM z)!oCk+h0dywi6=-u%YVJVS`8OVUD2`gWF?{WCgaLTz$0nrg{3JN-p?!Blm#UTQbM@ z{;jNUnct8Zz9KaU`Flst!(jur+<~8}TASlCCrCrRB`yHn*weT3mECjCJtb!z@x|a; zi6`#l6;Lat#U9%lTakC-qXpTUMFGTh} zp_YV&1`W6n{kGgdVhQ=OeMbGlB3qhlhHxx71sgj(w@y60G`hFiWZ$Rq5F@{wA|~rG z527Bek3(+x2jjTzNxZX~cCFqLsmnuuNWAq)|0N{Bedtg$*N4i6S}3nf2M~=`gjF&$ zQe=kpVov5E+D?|7jEPny{4@kMk|zN2E~p>G_RuvQ`t8BricNT~G9nYYftx8?atsROhlmDj-cQ_(_;VPwX!2U5*KL?jEt zzCZ;yATgLiaFfzv()^K|)L-yEhCplBBvtVdrM@#%6|O~_E}oDshsxGu41|Lq)ahst z33RBCJc^DMile-n#4YZGHQo0riy_6Ph}PPJ=`mGP!|EtbYpD+Bh)~Hy=Kh|?EU~_? zan@p>HY=x}=#77N5F2E^BiE%U-ocH-A`+J9wK&3802`DaF3rf$+(NB9_gu9h&B_(u zIokpTG{s-GlZ$3i5||abnn@4n`3{Mhi0TqR=|kF~UA7$tFIRsO@fF}3ESnHu1=P-r zr8Gi=sV+{z#rN5$_sO8cyBVCmVA~L7bg(J83@J76^~%$lw(mVH9Qoz5dJ1p z_Wa=~PjZ7S3JMJxz%S892{YO+qhta3+p$xPOaX6fNQh@(MX5K`CDSw z1!olqdT2>N)StKdlPcr*q}4$$vm8s=BBl2?H^?a0_$*7HEdJtGVAaPk&RZ_D#eD#f zuXjc+GaEvz|FNVEu$xtUpdKtNA1!Klj+{(GlnJQwM0uYDv{7?LUCB!@%+IJyTx2L% zrPs1ZHtzYU+ju+mxO&2oIFdl#0w%fa3YE*_V;HF&Y^(ozi->U? z4|`jfcRrHFyVxEse*#qasovb?GOai)7^m!c@T8sWcR!%N{PG$+{?ZllXW*|lYmEbP z%0IT1z_4vaL*L})W=WsTcX>juKk;6k|GPvezFb=VPlsr>yDT&+sOn3cBC0YQfRIJi zK()1;jR78s7`4%GNo^D!((DWf+D~zO*0s`C6w9_eLF=H{I;fYpMV*TiJfB9d?Q^P| zBm?Yd?+Xab zw=n+6SHe{jozc!sCd=$ZHGQ0k0vP|ChOX1texq!VM?Gi#y&9L7Zy|VYqLIrbh6g^7 zi!>|Mdbj4UDt!Vl5N?rF(k5s)4IkG!h2G0oF0s(9c&#Crd%s2gu>QcS;Y)LlH#z&5 zE#mQv{M$dcNsUrV?0ctE(7mVZk0!_y#wKaEKtF}PW*{Gi#RQ37^>8)`%6H<6Tqo_$ zd0-!hYh&!7``4nmHPS2%=Q5j!T1U0m)j#vR4GV$Z+2{K%0sWgp74y?{eWZTLV<+jq z9~};Ira`YC1lImsS0`SZeE<0=L4c{zO0eoDPL+;DxM0GHoMc$9tUB#N*HuYbvHT{u zAXF=#x5gnjly;M*0LeAvU zU8clU_vNJ*Y#_lBIy4IFPgd_dqko9n%8|1#-s`xqV+5%9I3Wc;i{CVM4DYMEN;)00 zKm|D!0TXZyS-WlJU7kJ`F!_K9iL(=28UBYj435$%oCIRG*r#*o*%WlcJPYSRtVrw=2uzdnGSvk`nZ8q7Nf>m^ zr^2W=za{e>SB4wt;geEsAS9tmwFeh-x9G~8vja~;B7mJLOW8n_kX~Aa7~whJy52@a zvPpgX@4{WW+H57JU%*C_y4VzIOyyGLh?5}>WX+qHEo~Pnz#fJbF+WB_OO$?Xvu{OXum2TbL_P|aq$ zZ^u0y>-Y7T|M0~vVC!9f4?y$? zZCM%C(vOx4>j9x~Sp4?fQ>T&#BA(AlJ6y*{I1rmv2fy5~~dkxx?jajOp}2RN4UjD$+cU50`NHe}Pnok<}}xbbt3L&Bk$JSkP2I z;*b6wqMN0|*zQE>X7Vq|myV+H1DLD< zTVq6!kmCjq62cB~c8VvSdY*55`A58;%pXsD+ws!^|$=tufF~h-?nq-G~YY8MwDf6IZ-7HHQH?OhhOL-F!tGJpMBoP>sj{Y7OER#qx86sz(|>$U<#!KW^4C%Tdmu%$=ob z_l0SJA=p!V>giRW*NqFS2F)%uzseu{^1F{m7R55!jnKvYe-XNgPMCa($(NW+#=QND z@+Zz&hKH93y1)d%#Tt8Hhi5(?3H2(fT)`G|Y`=4ih1mg=TE&#pyOzz}eA#S@s@6y{ zCIw42*rM4zfGq}?*`L$aGmQNivg|V2Ex@!C#S~TCJr7XYjCyj4cK#Al{T<1EjM%K- z_B-wD0Ag0q#q8#`rKCI8Y}-SkLy{)>Wm4Lr=qUiQvZgM2WW~?%?7k0(Ns4Kl1KXtRn!a)y+Av=Ok1xIo5Mt}5y^4I4NShkW;3*HsJ70Mq>oAl zh{!E&$(A`gL_Wq5eXH@^IBKK0cXv0)35 z!!(n_xZcKH%aip`VQE#nN z#1iFNv`Y{TB7GFL6Y?}f*9q9M0oU)gLFn>NB^j(+5K;i2qf|=@j@USqTm!QW1dK}f z`5Jz48f{n6c8pdLVWww)-N_cw@bpSE5mRjXhHlWVMNP-cf_XN8nB`FIOID=}VBE=QL&zR7X@ zU6g(^tF<9CLsA@I)C>f)@|dzliKX#}JBMW0A^1~Wu~xf+(hD>~sK>J!JC=rI$^RIh(SA$7$F&SCfI&WYzri` z_@<9UgDnE67O-d=P?^Cm45`j(u8-0A7D_ir$VkDX@-a43C^JLC8XAu>8Is1RP%%@n(QofryIJ4WINrMHk^i9x!I+O$JT*9fApAxE2d+vm>D zR!H#x$_~(~!pChyH%K^(DW(`VMNEZBEvZ>0hB49}0jWu;M{E+=;07t~prwbX7L_ib z?FQP`NX*fAbZ--7=a}XKk{-ei7Wl@cjq!ci2hSA$7J*BM=q2gY#oB<@DQrauG^e8hyl|w?k-nqr~qi7|B_!56}^eUhE zjUUqXo+B<+G4?}<4wOqb?s|@VlTh_pY(GdleFg1*z#spczvp%J9Z_x?gO~dg`1y!r3FxXlLeX>a;mL|h%-Yw`5Tbm=gF`2(cu75tcnw>+S6sA0g)(LF{TfBxUX5B(W(<3ep5!(i%wgE>w9g+7u zrkF$a2Clc!#m_V=NLV6VWGI?f^&NX-?R#iO|X1}4A1_^MDB_-dKZJWmouJv&ZTPEqdu7M`ok zBgA?_Y#Z|8W7zRovT}l5zP)S~fTXD=2MF0C$Znq}zgqJ3FMj9lPyAEniKkAW^T)_W zXXvlKbTfA2?WnA!Tz?%^EQ5ZSK;*OkZhBw$jej4`ozr~x@M*qn9>SEXrD`L!3{b!@MVLJ4=cEANLC70Dlfc$}2?>E|_M1z;d}Vww@FEuMJl0w#NoKm5YI z#oMR~tjAHVio8^8lY5$!ay1zA)}su zg!#@#an>N=DB7-)js6MBY>}i!u|dzjLU!Z|pZ&^<*riwQa6m6F&uAtSLi^@P-gel+>U*u|uKo)jCQoqg;)aPH?%E*!GDLlsd-z z>IqUDfHoMjinb4QnSQE9*;kR&klLIkoWR%$ZEKVcBrzmk(CmBwZARE^2V0!KxtO6u z5S0USa$_sP7DFxtPtCiXmAzMRaidMcW zxn@K=OJ6Rw+^(B2i=W2}nk2SGwF%wTsQR(D@EYEGUaMGZLF*{ayTg zfcTSWQ=yCJfJ>yhhxaR}#X+=bNkmWvzk34I61KcbtXDv9Vl#zyb1;3BQ5e&Li>P8o z^nF6TLh>s}8X!s0OwMESi-@`bL{!uyeH20;r5qv|epZmYCPf7@qMchLYGT_U(Nc^q zbY*(=kam6;l>}uiu1J`&MJa{O<|sQSwHdMT#8yaQ4e^ms6`)4gyy!lQN+39*T0?w^ zBtdbItha&dzk*%bAWjB^W`#nXA?qKbU96!NFHpvB(|8N%EYgf|OPB5_T}xuAHxA*e zHR9|QeE2@!x`zw#%>_s3;qWxE=4Y<&-uGgCs=9N{JJDPDOXeF2Uq43hUD>5Bwz0|+ zn^j^gSzI|u2t$0eMP0pssAqBV2*w>ml`9za2-+S;8$sM9>g^8#Ey6NIZ$Y(l2$c?C z^8z<|9a2GT2P8Rwp~BdmmH?JdytleuX&iGv^G0LodH;^i(rj zrJk56zmM8JmK|cMp7x?&v@9!IF_St8j_aplP?>_syANkz5jJACkw|ts>{l#|_ci%+Y zz+n9)kVlZkLwxf0UguB#yFCH$$9yxi&rg?ce0hVEHcEUA)ED zYb0+;;UeYoOM3-PLRQ`&6suif3-2tL=Ds3@hUAyu+r}nMyL*Pzd=QCou7Mmx8H33Mr4^F;U=#R55RQ-}$NPZITWp@8^e);7B4hA5 zzFj3RcWA-^>Yc+p@znprH@^JVqQWzuk39LcYoxHtY?;4_+;QV1yQ=s#= z+{wnyiET}4mM~@qW4bSSq3S~#q8)%jsBiwPdk=>A=?9V4V#@%g#F)Hut@Sm+34D8$ z6u*naHJrPE%34Bw2p zVjM*U6r-}74k~E|oWh180C|#q}b(Gsg$(-nO{QLo8 z%Ftj(wXmJk9lCTs+eh{zU7&9S`=Og0C#O10IA;p8>_;R{I0 zXzK%{I3|UhBnCtg+8uNmN#2m+1YHEm;S0nxAo&A(rB!Pvoltv!*O(S9o0vQY?-2hQ z+0tK;+Et<)M7aYv+hVdg?II(!1GE{U%m^zPiLRV5r9RqRp_!J1<{<_LUm`REvR*`j zfT=O1!?;9Bbf3ZC(0YL$OmGrtc2;Pn1C$+Mtm<5%Mvzb<(SXXj*ce>{+qCtZVsr`D z-@xkuvb~C=%k);>KyRJFsEn%lF3$c1?PQJ*$8gnyXs6JoLMb7J9DGKM8!Q%o2kAM! z^~G0jyH0%W_YG_xrIw#Zn+MR@3QA{P`C#1TJ15Tb#P56${10zC_udVq=O}3ib%DLh zG>D%MXl6&d1=uzq0!ocwdJx2t#GvAoMr{!Ha)NuKnp8gl)?tcmvi>$ghR$kS|0=01 z2-R`?JnIDcdV;n&#ua3vjc&11&yi$Feu>n~(B)qKZ~|BCQtzI?_0~X7CJ)mR?7v5+D8< zdinsL{mS$F!Ixg*y^P-H;~%it9iw;@(g$&)*T@HV7c0YbNHsZ4m|i0-e#9Sr=_Nk* zjH8K_h{=SJu(aifu$iRTM4KCA!{xSZW-#o}4@y;36?}f*o za`fAON-=&E+wYODzIew%O~)15*%-a^AZk|d>|q0Z`^UB&`|PvNK0hY*1>Sx3c|YP$ z9>EreFumQI4!!$9GLO+MD*IXTW|b$tw%L^`ePw~@O_0lc{a3F21h;&jpA>)arI+~B z(*_$qg2|5(=7&jchE@&AB!q74Q9=)qU5FPb$1kDmTQ{n@*adeCrI+6I^|l}0V)*hd zv?sgV#``EYOHt8IA0VkyppKyJ_vs&g?)Er?BGM~r()b3==0UW5jCN-D>~FRF!LMGw z<2j?$IBDQ*nX~sxBqym&GHO2k+aE^fA0iu`BJW?JxBS|j) z9l{i;%b8RP!6M3l0^<~^DG{YmqS3`Aj9DOZgJR{?ZbY|UMpq+Z(B#b-v>s6JIzIRN zf59`Kza_g-zf7E+!sibo<<4!3QCaUQVZKg1JwV@GK=#IG-C&FM>SL&kGc*XsbajKo_J~prS)W5tYnjK)Wcgy#}2{QtTs5No*rg zneN=j0GlOrHpS$1C#`aY02!bb2T-b|oeyzE3q*8wy_+y#1!pR>nPKeq zTN%Q6Ak+h*A0TZ8CemylCpH#s6$*_q9+fp9LXwP>`bZq2t?f#|j1Z$kt3(p%T*%6x zO(f}t5QJjvyTUpf5cUeflttNHQWQcxB1Mm~n-t6E(Pl=y`3p!?NZKOnZ9|?>&ksZN zNSGr!piG5U6Z8>=5g-gpd>F>qvG4ZI;RUSHQG1+eeYGig6=M zZ%&qPfe~V}L})a{y8t~y>4p?60vIz! zlt#HuIEH|TM#P|$0a}dOCdmS24Zd1NXH#4`BQBO`7pp|Mz~IpHOs+qI4K04LLAm;; z`0Z0v>KGdVVe*RkoQaM;36igpqS!D8*t?V z#CDAsmM}J82Ul;4OVDnL#C0T&k$B@~#&bJg#W;hu4N)@u;t-~M=gfY$w1M}1Tyvb! z>@5HO3xCJsPyHr`cRqnO{}^AKLh%5$SVG%`vD+kh8LckTJNj*Gwn++>cIN<1_89Ga zz>{C?^Yt%&m-`iYpZR>`lfVA$ZgC_%NN64*OsAM)8)ceXT%Sv%#^4vPlbRoqEq@o3 z@C+@#_;3FbpDIG{0a7?gn}g0KXzfYSkbDmbeL{1ccJ_U|evMxLA7P!P7`$^N$gO1A zg4o+6HHTq(8hGKpAN%9yAU0p2=-<8kH2R!s_XyIibmwU-C`C$vjXUPURgTpV}GvWjw8|>n;=b4#@k>vTM-`e`AOHe+HA1<16`gxXXQ|o+kDmLFJaV|CvHd-Pt3|sxp`H-h zW856oyiZcFq`D8)4pA**`;SlzPf@J@2$xrP6w!1WmMB)PpeSf||0`ns0L|nwZnV{j zweK$C<&xnA_ugh_XINt%e}^Ue4wm3D=pPz09Bg z!tJ8+xpRrn{l>F|onyrIqqLhJLa|KkkD!g9?JC8nLFF24YLFaBEegR{K|+D323ro$ zE|A&-#CEplob8T*>1}pV3e^F$xq#9`kcX%y6X0)dZp$BjVTUK5ei3E^=K3uCwYMe? z@}&*J{2Hv*1QOu8%;0)NY$1gUWaVq*<2x5cvcVQ%x{ezjrJW5a zvSr-hJbp1mOG1)CnSdgY#G+Lf8{z8%NLoXg0uhIC2JxMECO|@WZr9KwrB2}0c8+*W zQe_u9oVs`lQ5K0qQq0hz@L`?U3?Qu|0!GhaaD`&^3Wh68Z>;0WQ+)jj**cOk{47UH!jwCtkfCgbQi{|LP{sizN-31` zpd%P?8JH{}DH0n+Xi7+eHCc6FBttTidE5@LqcbBIDBB0kV8R*+bs^)@PbVwqWYe*30 zeY7fw8Z;SN2XrQwyxkKnTeMQdI4Ajl7E7voBsoGeZRq9Q`NlT`LbF6M+@`nmD&kKP z>W6vebC*HA0=&X6{@a{EUQkDkHVxnSas>$I&S@53eGFSagk3sBZVn)P46To$^%A&k zw7w2%2iM=g4tDlpTULmROzx_?V`LAEP^MV5u;UewiD#1NQ99+lbo`hm? z<1D`V2s%4SR%%RdhSoI_#t0LXo)Q-Y?LuLLrS%We?9^zM(ZwP1l^+6$X8RbS>JdUD z&)yQ~6oM%?iSal})(|zpnA@+e-C|G!6y1W5nII@aGe(;puGk{YOHvx$b>HsT*X2vZ z>O=VXaX$C^Kj4|qFMvPdQ~p{oBDGmKbQqw*UVHzm(6-{u(AHc+fx#x&cgjkB~$ z`1J42_~sYxer#{q6Q2CqVPZUm&ezeVPmDQn=4hseG1(d}+a;zwH?_ZruWpdk5BcL? zxx}p(n?L?H4ai^dncuuXHGdG3twBD(YDW~0glkBgh7g?pcgEg6%6(tbvR4*S`yNWj^g}2|GJUs1DFsvS0vbg-JBHy1(I*sFDf)MhUAUEeut}URW1Ek2=ExC3vB208g&|8G?-#`85Pq>t zI~$V?-r3!AEA{pvLVE)(*O2r_{hf&I+ zOu|-sHGNF}tHp^F{NV7VLF z_dk!QOISC<=mbh5F%pzTr8S5Kh3bmWfY7WX(V|^Vk`s^4H6AL z=Ge(WP;*k~9Nx;c5HmtGMCk_O=A>2=6=dZuqFzJPHl&EQEvi@`vn5d1h*A^gE2MUY z$}Gm6M&bfr9e~(^Z0#+4%n0og(+g`T=fT$4e23m}ryKF_U&GDU2{J~R9y(iKqQy5g z+K)R|mv0~zpaQ|Kl46(HtZj>N6`%-pj;Mgm7NmMeQi>3ZE+W9piC%Z_K?K5>n8v84 zhc<>7x=hnV=SE9Gqm)O;~8i*nUQAP9w?D&NpzmCpH-(LZVM< z4I&NJfzCY=6LC?XoI>|nY%fA8p&1f`N18FF*alH_mSqRYU|frGk+A3!=LOomigt5! zbdYU#rEOmD({+@syYeksztXx9d33h-3&F!D3uRiaCU`E8+7w zFKvh(y}yfKXdhyhOHg`DW~iDJoQc5vdQ5S|Kcs zpiGIfrYn47hArl#=qqi&BxW7$SZZZTl3% z8+T5Lh(;u$^c+`S#+p|k-z4}?fb+pS z0DGbxw=Zgb;?FGKe&uJ$dS}VYqv+ly;+Ifrh{+Sh=sU!;ca5kHuB z2k2WCRsghXh;;_iO|itL2dY5Z7BLpBQ+JN_=3O&wwGG{bMYh6hwgPmHRy}`^Y)O-k z|CVy?!aF_(>_r%?yoAKFNHyn)r!L(y3gf^3-GZ(7I3_zqo*%)u0+r48a^1I(dKF&_ z^(3IuHB$N>DPH4Szq)hpm^3B)pIiVg@VVa@(4-L|b&G@AEcnQU?VCBTpZUrNtp~W` zy}MV9YrtxhUE%aQ{tvF8QC(yma=)sk@3PSG#JDB4ZQ(|?P@$-j)cmG@4XPy9=Dqm5F@nH@}CP6=?k+ z(Z4~lw0+m#t!E=79HxmYD0v=JoW~B|y%;Uk1B4Y!{vl8gkPlzzN^(?$RABrO(qf1- zBkGNR4(3nE%6E?~@v{+Ou}WQkpTakom2OP^eo8$6l&i7!BHo`Q%5kd6h0d81=JeeNCf!Gwp_!_PM zOG5oBHvAmAUqJCHT0D`Q7>Ag=K*<`~WF%$KDiVD|ThC~Bw#bI(DMq_A+Yg|0LZ^r} zLNS=oY|V)NAgFDGIl&*qq#V=;Z89V^DD4pvT6fO7t|ZP7T21sVsu^{TWi(+fol_DA zI3(r->GF9~t4ThQQrTstss>kVpxhi4dMH+DrpHKafwnpM@{6SA0Z5+IID`YlX1Ocj zOBv!bFb111u>D0BXA&&YL`(rTt5C_HY((ohp&5|E37V}jrrZRTkZ6!JBuXGFb_kPI zLc2}g-$7g7Ih##|l^H4BK+*(!PVlx{2#}i8dK7Pv_iiA0hAsC#Ba4a{d*BBcTVveb z`7h;&I8wBv(8W2RyL?Sapmii_A+$LJ@Ut1Qb|{lbp}*%+U65jexPuv7xIMozLDs)P zwb;Zr>!`eoVbHoo(um}jXbD|ZLCQNJ+bxJmlUhOaHY+FpZxVu!8Ghq5&HNC4F^AcC z^7a2jv-Mvh;VeRqb}h~HD6oz;8Cmc8%@~WD2x!xydo8hbC^JHNM{HKndqRqzt%4ch z`aAG;$0v$rzKWy?I^RL1UY8M@Uqvq_C^bSmiz~Nq#V&1F!Z7NhX!hLn)z(o&0%(J_ z2CWUY^x)_+VPh9nVN5`&8QRRcJY?rP7dr!*$wQ=+aOHJ^A0lKow=v7N2yRTP*I`fK zy#rAUzVk-E&S3Hp+4yzP9%I0`U1D#6pDq)ix?73 zN<&a3u@PeQUFjfWwC#}%3bY#|vBmYakTiyPn08VDIVMXqixrAnrU!I6Cv8Tw0j(V| z_Rwl_o551jURJksQ!ptwl(L;0RY24n1R@U0yNhimpJQA>KJNoJc=BsQtU8G<#*{1H z@3sp?U5y|X-SmMD7*k{N>COA=tXvR>HxPdqKY0Ln?vB^`wkFl z?!SK@C=TN6n4*90=#zV4i!H2Or^ybmFYxZO&pz)X_66R3_IW>Iqy-(iQQCJ{RZb}C z?bDrecTd~}P+~e~%G(5DjZm&#KsZ564?)@n{`x1mEBpKyId?Ac>EFIVY--|ch#pMd z{Xw@7t7UxJ<(^+g@4~nKr&oFMYmR0s5d9fKa{yOvkQGyu4ftS*O`m4+DEMQb-oO>F zk*~eGdp9&g+Swy$cNCLbwArRyfAzLlhmx+OWPanH;+sBl^t6v)4RHOV?84Jl9XPj z7Ht(nvbJxA3L}TDiP)d z$w&Mm!)ASiF(F#e4%@3y)+0%yv_%<>QY|izgm$%a`%N8gcp05dNbRs2akgU&2k+Rn z&Jpx|{_CgCQ|_E0+JlH1p)!Yrlc48V(_qXDoi|8wL|F!@(B?%_R-xMi5Kho8N~i`X zm7z_B&J3=c<9hQhTU8VyniMl6ED=La2xYgRg(1duOzF`1l(ao1#hlsZ1N7HlM45or zo~(BnKe>#ghY?w$o(&*0#E@Z&E`OA0fF2@>XFAS(`{vVbcln7l$H0cNl4Xo=9Q(oBblyoT0*nC`Fu2|@R+k=j-Kd4r8+fW(JH=(iIw}1gOx7ZQCoPHer-Sl{G3kBq&727*mphBQzy3Dy*%EQK7{m zTF@eh(`X}z4Dho8l{TOlBK0X^4aSv7Z1IZ~P_YxWbO}_bY=h)Ap*n~$Lv+?7wSBaS zT?Vu5p>0H2+vOk!g?25rtci7>7&R)mu2@TQ#1D`d@%4bD6QUdRUYVaU4Ix-WMY7&S zTzQS;_lo62L)}B?30M3GVHH##V>i(%(weJ8zYYkt(47M{7D&p8t?k_4PEZz1CMX~@ zkWxTt15O%1W|0(0QIaK3Ejwv6V!6?pFwbx$j{on6-hZMDxs4Ap@*anq;NF^sjSWSjT;KMX{WHIC4l{cgKi+11@-M+c*FK7|JJ|9D#=O<6vo^rwCLC@N+lcGt zn6kwbD~joLQh$Qn{0M(`4rE^A>vc?7VT>iGND50*n3E0;=!zZc@eg1|W2}8`IA+FS z|4(rPQVDuM>Ze#&A-2R-D~jSe{r(QAkCbZ%*kVOkP3h;`q+p2c1Tn9`w;Q}JV5LQq zkmCd~5F3XOz=oqmAQ?~%ucy4#pqK&KAiG%Ck%M65ZVOct;K|+{?Kh_|vn_megPaZB z@+@>GQJG+<5HV!Ugxn#E1-5t%Kf8>PNHMxbKmQ0ep5{9Z-p~BHr8)RGZuVjPBs1N4 z`j(^o8Ks=;;;RD|*B(aXQBE%!z8z`wS5Yz%awe+9--QAG_6KN*edfi7{+e(CqUaYF zqK*yTV;$>nq;w8xN$4gNcPN|)hnwVh0zZ+X5%Qs= zAZq}*{fSbw{W7WhC~Oge*OqX&SUF2Vj}$^ zXj{N{aC_Z!&U*cSv482uNc{%Q)gL2`zD6;2nx~r08jHJWSubh{!X1>52Xw?)BDs>~kefe}H2BA?l46-iE3$TB9bSvOtr?mMe;B zOMfsU^&61aadm-6L`}rFI3(Tr5~Raek7kdUNIAYix7s2HNUI*>7Kq(KQi=KicEo2t z@A<;xZ%-2apN}5^f5sC}7@j^cqV0YJTm3Lqy^fn+9maaCVv;B4a!73D4r{Jril@o( z5w!b0G>mY*!jG3I8F4kH7}B(e9`RQr;#y)qA+Kusw#AtSOo^*^$#F(8?dTUJse#pN z7qC?ezCB8aHRR9|4o)EMI=L?~Rf(ADC{b7r?`>Bj#zPX&Oze~t6jj9rBSQhObv-$Latskp+I9)I(F%{h{>bo*zB?St6z z6mE8ibpdigiWX}+uq|#hw33JAgt$0MTvD@JDgkwxq!xq7M^Z#0bvVQCTx8`iXGY(Xm7~D;fju8+LGhOP;^Elv?YQ- z%BTiJddL|K6R_Z_9y$+sGN4s_Kzc(CW6&PSEkc3jqw`@CAUZJpP_Tz0zBxixKxjsg z7T9t@Y|aejYCa-7Jf~#&l;bL7x>p6Xz(l58D%AqjL6#Iz3fchWB7bASNz?J)GvBF9s z2TxAh#P$Mtah-Ox36+8_2yMpnlcVEG4P=bnC3GW3RALHy)F{Mb_OeCr)Kt;R*Br=Df;>PINU2l&a)zs$dV{0gY> zlfS${ktT>Ohy-j}^21kd-X$_&1+hU>4cyy#gl3CIN0ilHD=*gQbu{b~mmApf8tq~O zFMu_9&*ONn8M)z4r%m!C{SscVm_V-bFA5gMa56AM04h zI@b42#|H1Qj`cTDl3pXO_UI2zVl)$0>u4B}vcsA<@D}TR!fFfaX86gQ2Caq}8W-5| z_9Ns5rKqox$J;D#o&k;x-oGvS*FXCTpZwy>m|}-!_bm0+tMB|r*|56y5dG>E9MAHl ze{}CAX(zd4% zE9CDO-+Nu^BD4GdfWw#m9$^8gE$Esv7*h{rRWYDxNeb8K=a(_|Yy8&Fy~Yzy82+>< z5sCL)OakC1FVo(9jdnT5+&BxBlKbUR(bNO@(JmsE*sY=3q3I=x@i%eRKK8~oPe?|zifRS!|l&f@F+TXKlARv6nO<^at?YG+u##7`IG{nMl{!i@te5uqH? zT%kZTAg+U$iNTVS$M}YFv`fF-CiNZ8HYgTwSQ5cn_BI@F`wQ_Wp2+;?zp{hLf1h&g z!}yKMl;f+2d+WZJqbIc+r0@VaRpiv+Y)|F_nf>9h6+t53#O;6wopveMf9x#e_X_o`6&YJm_V@a+4g+;wMvFoxlvQBNE8bK`e$*Vrzym z;wXgHlVVRXUV!V6Y=)$(8n7$e=qf}N#PRMudr#q%CG(G zEkzpC)x45dlfWJS^sk+rw9T`0)bkB6;^dh!N$hX!P{+ZG8DM z#zpG&=UDC+^oI|CD{RQINy_-nto9aDQ^PBT8~JNEf3M8q+ZF%VC*nv z?Q2EzQ0l3s$JhpPCiNaycVNMo9y0Vg!v~eZ-a)BNm*H!wAt4sZFFX!8yh7Hg$MSTgs`2z7^o77E- z?HW0)A+mvSC8j!nn9*EOR2C@&kp^^yt*&DH98H!S#zeP)(+R2lF&vjL@)yMJV;FlM z#o7`93oR()gdMkn~IzQ zsaxU3dk9L3BSKptGQ!tAraC%C88CB)>KdvZQz+G_CA6o=-3emUfw^6j{Mos4eCmt; zFKzKGzxQ7*`OL2u^oxfO^Zkg~fpkE>sz~hyhAFr>;1X_kysuZpW=f8eH&9!tugU$C zVmw3ab4<_t`&(eVuA+Ux^2P^9oJXv|R^PqUq~${WU_`$IZ9y82q@E;$nlC_Hp6u?B?&K zjJrf?t`e#bBK=}OV(589TT-_T(qruMZPQUXdt#WN`aI=qm|$>|>zKuZh@9lz7`&f& zspn76MV!7ngEl3Pq=?A9r-zSqtYaPPyQufQ!F%!I#gB-{f3K>a0vZwd-;FW<({KLf zZ{FS<`yOuZSnpF6Coa?6Tp<2GC3O#C3WG?6SW6~B>}mHGq~#`VyoN1@vF{X0Qa1(j zJZ^Mn)~A~uz~L#ZY<@4E(y`tviIP6|XS6a!w5u`}H0#iO4(Bp(%2-9};( z_10HF?xVeN30Yo38G&)A_J}#e>2E35+@y;?F_^p4>r`vEX51MUFq3PTZcJRAA;+GaK%B>9PmYG1ON>s) zrbKZJP9cy%qtY)ba)^j*p)vuK)O&I+jt~P0S6pX&_N(O8gg(@m;K)8=xJ+LD1gU)( zk%TEA^~GS?>cA`&>o7KwG}F%uVhheG<@9hUI5Qy04h`_R!Id}ZcDF!Jf+)WJ7Ph<$ zNyz#PVy~m!8g6u!*m?FB>-@uC{bPRfXWx3B;N%*y+rs&e;;MkLV=z6|^n*d8W?-zQ z4J1$KS5s_t`Dmm&B2wU`p=obmRv#mUYgpgmCQB-wrrjSwcLt*YVglQtdQ`f%I7Ms= zQa6~gw(jUxU}ZvYM`+B58RNXcx`xaMQN`2|O#!9B`9oqqCUrIKp23zK#w~_aEf7~F zmO#H+N7SKlj5R|6pp*%36PhEjGpGp0dsO_;7nlodrP#hCW&@%a;}J24tWfm>Zr8)t^kN4%eYuSE5E#5@NWVz?ijpTt#rgykmP{)6O{v2KeLz7O(g za<@%$)e*x6nlsj~$gv_e6HN4jiE9B8xgXFXnhatUDHTH|wtyu3RMF0%#$j}NR0=jH z=NUOTv|7An3PF*RAMy0wLNk=yrN-c^742dNYrsu9OnGFm8HNQ9u|SQo z!dBN1*HRcI>K3Uts9$312hebm6h1`Vewm;8KMv4b<{s$0eN&N_=BXt{yZXF>D z?ksWrz_jkG0f)3WjV)^=_lPWS+tDCJPqF z-gu2{q?lf%Uu}}xv-F2OS%Yg<%H%9M}UiZh`hhT5x+fe-p8hoGf8+28lN? z)n)9_!g5is=v@E>LHfRdmD#&`EKfX<`Q(%P5Hl&1Z$X3Ijq-9z&WmB<@>s_@*0KKP z=zVSQe(F=7`ma>=*K*GP^_eqg{>k3nUfH(o|B_OA_2R{gAA0Jkr|#*C@V(yNvEIjO z5AMSj>%a)(b|^+K;KtW5q@%){ir8+V=>jo%`ok?;bx2s9A*a{y#Vf-Y?;Rw*2jjt( z-)+};wwD$_#g7vuWsX-9f*I3b!SQW45lh^RmQqVzbp~$Nc|N>`Mo!|N=m+$$t z{dWDt|2D?t4U7$U_*^g1+>>%aF}g~&&A`p=*WWOIV>6^~CHLeU@zn~k0Z}1G3z{(B zX0Y30g=tO_(>jJLsAM!L8u$6qKf1$MqYCYlAH|P9f~g{I`p(bamP^Rj$nntu-C$oK zb>|4tP?Q#x3E~z5v(k2mD6t(8=QC2U13tkRQXCOO!s~M2&vqqQuMLGyC}uRr@#Y0t zUak}U4MM*`&S$W_!;LcDE{8;6N~nS+NUcRRVVsgxXs&NjZQg`34k=G{XjZ!GTV%Gt z)EN6B)z&lg`)BBv2IM}hJ&lTCh;5$0l^>>S&U5i!{2@Bsr*0Pmsv%m`o<;3xiqUgJ zEB&{n>{6`AX+&zLgmfCwM8DX>SI^_rV|)UWFno)!+@QJsF!k0m*fLN~mZbDNxjBQH zeX@ea8V&uxkPc&FYw(rgCJAvCqXHOG<4|L9Mi2pPf=-T1qJbxy`w<}-%z%6*wU*Fk zMq6`?@5s$Mu8PFYLLQTKj1>^?D5ndsgUO=7L$7$Bv1LF^Ns1m-C8rKs_Jr0A@4sb8 z7`uiPrwAU*iX?)QVINmI!~x17ibmwIXygVmu2PARgeRH*i4AiJaCEu^7|^(-~0kqX}>yeUpX+Xt<2{U0gY#+P+L& zo?YO~n8H9bgs#AQ zi!FLo5=Ig@OA6-*onoX%WP)akkp)E|#CVF>MO?X{813Hj*qpN`cQwXd@gnp#-%aRlvDi%SI3T*l2u~kD!P9sEQNa^YVlN~v1 zBDNUtH_J7Mhm6ku5&hv7VR1jfK8njXP*#-wI=GAux<~mMa+cItl{cm0ZC0ngG2KeTk0DOWMhWKKX9bz zP<4nFXg@-V1(<}jEy*@G?}_axa@Sz0F-3I)B$3M{ao(ZQA=Py>mGu1tVui>Yar>lx znAEWK3)GVpee;9lbRJv&N6Px0@lhFJ_1Xss^Ep`-JoYQwaOx0V-6G3mSX8@%x%=%Z zMmOnG0qoE(l$^d!wehCW=kprYPq3qV>pOtKgw}3{V!VzUEA{5nZ!?KYK>LE#{y9SV z{cv-{XTJD<-x8bl2R|OU_~bM6%@%U-0bKRsa5R_pQNu@#4jgsOql)U-<3c{_S5n z`f&h${^G@pKMwriXFvPd3t#xc7w(D$^WVpox+yiz}l;1J%xsV zRQpF`oE>g_h~`alxSz0mkl1t>_dM3Tj2+z@WA#|?M#5vCE4Wr>{`3xGb>A&L_4o~b z;(z^jq<8~AzCc`VVf=kGyTve~96EB^AK0(@96$8Rb9c#kyE}dQ7kB6n{tm_ZIqGr; zbrIEsbrW2%PEI@IaGK`k3HsHmjJIEWm9h5r21SoUqy=hkcS<15tjL z>%g01(oZ~L_|wyyH2o${Dw1R@Eymv9xBuw@f0fkk>l$l?x&mWw|CoJsjqY$k?oPwV zVC)JrS`k}C@`xk~5^?2{)OVmBIV$)`p+Gfbqyu@=xK;GXyjvs73qvtle-mr&e0=$- zUpR@Iox`{i=L5cK;4XGOQ!kN~MO=w-uaVOIq-LFvN=%i2F}CPIl@tnMS3=o3Fij9M~)1pqP37L z{YuceMU3Ffo)|%mfozd1#F()TY>_c4q_)8L4qOkxlbisT)ubh#V~vSx5%)4XOo343sN!@Pj)q9mYV;EgB;RrR!^A(~&h%RP$l~ zx7{1;`z%*kuD%~*cQ8I--0py&k%AOzly!_TP-Z9&F=mg{IT%MunbXZZDUH#_p~hnc z$b&(Z8yQ%^SrI?*XJhBcS;#RC&kL~j25$CE+VnwEJWF?QgJRSWyYpz?LDZ7^l9Uhd zqlR*Jofy_hhv$$~Af_OP)2Q?GixQ0!Okr^KGQ~Op)068K z7pM5q$4Kok8ZX5RS1iFLkVtAPj5V0L!T5mnJ>mj6RA^os65rY)rC^*QZb|Ml&K9I_ zc3Aw#W9<0y8@%J!mD_bxdyEr|@9@GYinqgfzaI(7|Ywye}^}Z)d zgOe7O88NKCVeneJI$Ch3F~yOYsh$+d0YxDOQUN%S5||D+A3CyY5m_CvPG+R;0)G6` z8{ZphBDsqLJK9#5GNCab@;vx!M+xx~jl=VE<$~3oL*+VN64g<0YhFzTJdPS{+2e}^ zxJ0*inr`o-q%=mjLg_z@@(}690@3dWreJQ3F}dw)HM+s_<|Al0#OU?6EM96Z5#zD$ z_D1|Il@0C9hcV?g)^_;W%ePGcOVKmld7k#>X+$5P+iiI4bANEl;;^TF?nOTJOIvhO zAo;_T8!z0piy?)Q!|M+cR~_BrS?bGwes9m=J=Vo1AHWnpf}1@|QGSh-UM4I*PV7EF zj3caV5w}8OM+ocWZil`(13Zm2FH+3D^|tdvlP%=PAAI87adb!X7g?EyvDL>2-3@$s zckgr9-zF|0Hm~s1&s~F`dslaVtYaPPSl75#tGBdx>)T zDt`0%yE4ggdz3Wf-p0<}&-%m@j`4l>Ve@H(O~e(PRptCIH)O8k;;a0t$FK5Un0j}k z-}~>Dz#sFO&p$&KPN8`Nv?NQ$riSRQqpPd@?mz4Jt@rSLKlzJi(fU83nmvpkEpTgp zi7$HE{dO2dO(!tr9wJjzPm`KcEU!$Mp8X3lib;-ahKuRcm4xL68egK?y1S&SaRK!n z>kBS^ejBxak80x)?4+Sy|JvJ9Exy>Ly#eS)k(_A{3aYiQy`#yiZ8fBz=U;*Dm0R9s zFwJiq{rfJ<+N?KXej1a$0oS&nZzM0bLmYWzi12W=-a&~~R zdt@RS*U(geaTwoYCe6U{6o=-Tat8gU7E*S2pN5jF0z&HuE5Vi-u?Z1F!lG2z z0&I1I95#m1F&hS>4ITcU9I3Cc&S7kWt062)avaRW96WvFC|w41UVxk=Z480$wE|7FC?QL`Wi3%w<&L&E}Lft&18 zj;;}dBgbvxqQa~XMS!RrF`5%O4GBqGC5%gmPgpaQ9jazf#Y#YYB*o!%4!LIRcPU0M z(H?w&aPU!bV-PUdTu`Jrfry_B7$?O* zg4}O|SB&+zdLLs0f)EyKh+Tp$QK>P09cxOYw3L(A&}3-y6z2yHGdNT&+K)(~#JG%R zkLDgD5wQV3x=tXXNyyZYCuBX1E<9APV2=uya#s=C87ZxB^^nlZY9U`EhB4Nh$NC11 zQ;Zvu`8iHdYE(;%Poxmh>^*qhjDg)jH9^#KN0 z-S{B=;!%1T6IcIk;5M_t{Dpr2^q9he_yG>l2AZhEb_ZL`z*&sDLb-8m7!SxJV)sFE z`5?raI6q0Mcd_-|Aq>Pvw5!PZI$ilz<6LY+C6l7P&0pVI?fywbcQDRCv5UQnBCl%m z3Ncssa6i#J41Dbtw^GW@FVh_+n$!~G1>9l6j`px+gCr%!lGLq{x(T7ZOj>@EdiEdR zb;0XzMolxY@&ux4(rO%^|9ifD>pI;ySb-j?py zz8Dr9^J3ufzYBfopErE+m%dI|O=)jDjBGvojwp_Pu|{)vj@bSs=+p1|zK?aRV;$?e zr1z!4``BZToe+_~3-rJ9JHPY7cmCcO^Y2v^RsA2i3q$t3-rlj^$FlMWMSUK>e&dZI zB|p7_jr;V6=LmTN+A-oBsGw<&XgpH1SmwkC)m@s4pT~#F8?i$fQ$ipXkPeBf*ZAotUuM*QnJ@kPof9_irM~p|0dQ}n8}C++eQt)Z!F*6QdIf*S_!q@<2C@pl8 z!#5U3Qs-G+e}ryviNv$FOw)X?G;bp_KPI&g(C&Iv3`Tk=mbhXM>pP5#6z&;9I1T+K zkSIp`Lpm=B5D>c>E}HF_zEN^F!&o7kh|xL5FWz$W&|Hw>1Pxcn>Ikdnk??iC^v~~{ za~(l?9g`~!c;OOG3@{VAOJ~3K~xi5+2Kd~!^NcJi63o})ROxV;yuPBggxvGY$1plAxGt5 zLPV1#h8n?QEVx23UQrTYF`mI_PY_m$=`(1DF%A)nh>&GMXu#Tt(vllHjQN*yTycXO zDpEJYmTCBo(}*osLs^(w92r$XlSL$hWQuRt0Wkur8o*gmM@jx3LCQgK3m@ zH7139T>T2w#;dHZe1zQZl0!jm1DS*^7MRf`>aC}7)&78nNY!9GFYbr39;OcLj1W%X z{UNDp2#X2*!3L!?8P#`LRoQrG}i7&CBiP2q4OgBVZ0JPUw`!xRzZ=(}=4hhTiC^ZlUKCIM;j)zjdtY}^$ejl5Mf<)t1 zV0CUuxiONuI}VFzO?G!#$8#&<DL?vOwFf=7i-Ax!EK|$Ljh=8J+sG!90HlbvI8yTA^`?O>aGK)7T9AMy%g)JFhDz zA*@a!b_-K4(cEF&-JL(w(^t^N_aQvMXMX+bw?rWgKl?JD{Kad;<)f729ayd-c7n+Q z^o0H!e`mnjOmVTPKIQ#TF^j2FM&3!k{ZGHiPd_;u$^`XMiqU22&FgRLwaDt|_D<6u zZV;Bw5W6q))Z@q3PscjevHr&CeQEGI=YAAGRsV!rB<&6l4|k{2DZm4t{`9A}e&aWO zo~J4{sn)*wmfsoIP;EYk21|$&L_C^>>_%w&0Zg??!je{7 zXkQ@jbW)oQX?c>=JcG-BH9qM3K|lQ~4^rw6;np9QZOO)&Pupjng=&zFUY0BvZis|yU=I-mN#`_!`7N8Dsk7E28 zl#^uk-zY7Vy_*zkmoO&N&hIBQSIMhq_&1+@4nF%=eBs`WRQTy%d;n`7A(@A;`Y=g1 zF-FlG(YB#$<~Vy5nZ1IWy+kN2bDx)Lg)`1m_OBlB# z#|g1fTs0?!5h+bbT_&}O)ICjFy^M(EXFmU8p@#!-X=g1 z(ZNs%32}$GYJV7e?+T>Yp>NJoRnHvVXMr%mdOIX#-5#LiP-E+BFfTz$Vmro`{g7c(kQw~;bnN4pf`OQ;H|Jq7s`VlCo8)PTwO@roSR z$>}~~`}eWcE_Sp)yD_iJ`YB=6O7CBT}CwFON*Nfqy1tNISM%$ zh#p(DSgROo=npJl2_fMtgK=(HY{;|01PO`MMGz0B0246&HBvCdW(^@D_5iUFF@_ur zsS}KPkbo%$K5`5tu3V6T6a<}5$<=_;NPR_)f{}8-9>he57S$H91x99cdjh7xO}+uy z(k@;`eNQ!Wj8A>_jpUeh0bC%Yg3JhOOUl^`gl>gIA@-ZtBGF4hY71;pAXdTYkZ{I= zRiwy}gb>Rk^VOhQqN->=9X=gbG+d+9!1*j7^AJLAN6JQ}XBru53x&Chfih z+oI_;tiObFV^TOtzjz1@a}q9j7ad`3}BzJ4REBE#||yUK%6a8acj104%nnt14dXgYwIHlU z8@m0|#FZzetC;o*)&9NL#m(Y2 zp*|q?=PY*0IPiG^%?C>p+d3fa5q`PB;`S-v!8oMf{jIB@*Z8H+ zZ!lVI6LO1DLvkHD?6JJP!&jdB!MU_QZ2j85o{-!HeB0tTU%uxyzp=~k<{h%EBWx(M4=~gqR*l zh8P=k7!kuQ`o%XOe6>9PCpyuIPW0oa52eAYsy{Cxh{&t=yZtL)`AU4|nP)nn2_ZZI z{Kp^V_D=L6)%?aL_2wF;nj_}^tL@?wmA^Ns*@ZDVOi^mEW8!j#uik%Ed9m{dq2D9M z*ZB9J{?Q~qe^T|Se|M3(|3|p>r>NK7pqkzS>E1SX0ZgV@-=kXJr{B4VF+T;j6VE*V z$NbKxmk-|fPniCj=hv|Q61EvrZN7@8is9%<#5{uWiF)!Hn2e?!Vs{qFYh;~*RQS=7 zoGudk3HkP`K!-ROZN5medHn}5W215>eh@8WZwy8JnH_G_7R{Lj(^G9)`26Vuh z>!_T598;~Q6Ut#JM5(J}T zG7e0!hK*!Y9Vt{e+mVwcM?)Hl0dA{1)SK5Bj%H+I(X>rA>*VM#t`wtbtk9}1Tmf?j zF$qm)QLU&Z*XRyQ=?i&4u_y%_0Q2Lfc_GFXvRaacjPWgbrD$-7QCz!2HF+IX2>k}+ zGiBV`MU3t5mL&<`+v6l|H$$|=*xLl-FlkC!JOw&Ks(>BcF2-OsB^@mpOvIR>7@n&c zIeEwy+uX)AhP0eguiYRVZD4$lHQ%A{H}Ug^)Q!lMr$4L_t;;C98Z>#*069R)t`uc? z!HhCkSFj!?A=ZIXj2{(@D4Gm@9La;A(UOA@LPeaH5%A(}XHo#k23cBydm7hXLEJ4; zf4mr@09845aw2sGjN+OeW0bI%B5{eqwA+9cvL&Vx~Nf=fYm$Q zRxXrS_Yg!zoHI#23)>d_qE?tnsAZJ>D#ObaCN`&6@i`lBgI zg|UkD0dS~GXo8#+l@3ggMajVGv*SPrZcLh#&2dhSy153;7HwZGkz&RfVORL&dNh1GlNicUbH{N?x5IyFuOBbA75VWT#&h5r znjGEkMbhek%mOpqX2mAnKTJLQ2<_HOxMQq>ri!p?3Hy&?ncyb7v}@PU;24${(0GB> z!IFBdlvk^bhi}HW7 z3;g(o3A*YORkeYwcHhg;m585QWmryO_X6#yH_GS3v8?ZjE<>J@LXGu38Z7hMkCPWS zi0Ng1_n#lV`&a(p66OEm+x*(|SLqHOC+j(UJp)@IQOWrZIqs3T!ot4HSN_Eb+U`Ur zI?<1T0 zT(ETemw^kIDpF0~PFIMH`0*9;u#2WOavqlxWqTIV3Y8tg@+S4{2TPaaBn*cS5xZ;T zILFN{V5-CWwx_?P)&42M;p3Rg{3-EN^@Q4Ua&!m}*H?Eg|1wSe%C8GTo7*TK@)q z`c3*p2l@!%@G*wP3hNq_23PMR63My3)e&`}7}&v)snEV5rA$t9vfe=Z*ATrywe}Fk zpT+n9T5telsO*56^Tg&U(&|y3`P>U6CeU9**96IHNSGnxn^@l$oPbWrX@eAIq_~E4 zyPz|$9x?mmaE=sPY<&~gE{kDfcgcDd+XSpP5T)>2lPxHTRRah)mV}p@BV4nGn24Vq z(Cu!M(-P~)s1N?L7WszYqOSH|HbASR<~D4D>F(NKz| zH~_wmQN_q5G6x`+kZOUeuNOa94LP1gm|^`2TP;gU`JI1u6j!YXu_?UPxDH`NmZlg> z-eb)kX*B^8uzrOxWrVu?Jfwc|Hc!(X;G5eR+mgd6!u)43Zi=ht*wHTG_8HLgh}$46 z8k{$X6H*FTiC9Nr2Ajli*c>Ah3KD{#COTyPQkmrv6$fB7t&h9L> zRzxg0WO4$;5W3UEZVg-SW9%y>d8wWh8;~AbEzoEQs}@X)^L zp41UP*}0p3PlLlr2JR5LEfA2Qr4Dm2@|<8A^e6b5w6`_v|q z`!VDxVj47Nl&+vETqGxhNJ$}Mz?fCR4ETg5hXxO$d3jA~17dNmEAJIzikYnq8aA-g zm(jEUtxF1AE!r>8WXRe>ENoY?8>F;GT6k0zRMXSIfU6B?4{~^nsPM#L0}&5ckiHnQ z){>INRddjQB0ygrhpwZkhtQ*L1?~`7qQf?@i)&sd#c3&VtZtBp(`8}H&cP&dXh920 zT8V{ct5T*2RI$|(*UX8XWmpK}9-^LJ!`403+7)^u#DzuUCOKUo*DXjOE+^#S5Ugc! z`(woAn9v_mPgnTy3fCOqoB6R>`PM;ci2aPzT_mOvd9R_KeHTAkVf?r0_YY95IY_^N zt8P(MOLEF+PRJ4@2`RQ{wqP^H8wwsKqq;zhl=OMd#8piSN*W@rjkxi5aP{rG+~fyI zY^2$Ilf}-%FmEB;K-0sx(MRx;*YBFU<-P+_l|okAIO=0|y7-8nUSrt5MDCX8=oaF; zlE$VZr!BnPp*i(kx?2|+76&*zB6VY61|sG2L6?xjq;a$uWF3#wsA$FOuTk*x;JUrp;p6Em;`bp4-)!?0|D#kpZxwfbSP}Sx~wLK9z z2mG)1yKOq1VvIR4c>g}jR%J|5dr8%_Q{7Y(6xWcSAP9u75d$J?m5F>oc$Sm^#pG8vtWM;KZ1G|OVLPLBc<~&oMzRZ z#*bg7YVS=!x2N79#h)SdkMOB~_acAr>-TRt{z+3MQ?L_UbB)+f$@w&<8c|PQyKg?l zR7+r4GLOj;hZUro2um2QG8|-_c@%AWY=+Q{A$yEw#P^u?fU14_M1Qq+kvLaE{{~X; zVcZDk9(=JC7Y*s)Be?oJ)y556Q?_NRL#pYkggH3(3M%FQWfZ9E$iiG;I5@|zeqo(2 z{l@(-N`4T1@(*2UdH#SG2fzB6M=_Hpu#;`Zr@wvIz^-Pm5)Lb5?>v6}=G~FsYI=h> zKSdbMGaRjG*1kn;zRlv=jL>}#O-I=FbYY~*ZW-sb9b)(K)toFAQHusowwbh;&@ZOw z{D>5nG;1G4>J}pt(r^aj8pt&{*Q8iDsU|P6{ujvUHl{mGPUpx$aeg4i33{|nJ-!3> z09W5eogoctXgWjawn{mXTacqA^@^qe>jup6G(!#!WI+51-$b(7V`I^hfvHb<*lGW;!7D zHj5iSLt6bbu0B)JrdnKLR{^#H@yABLQq3}^f;>2kfw&ZGmB1J@3&vO=mFFr8sk5LN zTMJ^8(0j;Wy&yJYTg6r7b1uxm62Ru;$I}v5f~jYW*8YeT*U&sAr%PBbB$G+K66P(* zIdU8@a)5D%gno>s4Z^Y^XO9~lQjJ$=wnP&#?iOzRuJ__xFcPM|h1hk5Lx<59K_(=d z?z>njVn||e??jA5UoZ%93*#LkdnJv^IgF`sO(5qQ=^QB@n`Jg(?0}e9rgpL|D2f3v z30Eb;>MX_>ghMhBOfEd?WKf$Sp2pTKIYz{RadX5NRBFgIi5{cNlK7RL94g{`hHZN^ z8JzFQK_EDWqfLgzSv0N}BR?gS9@ivvxj|ZuNx8)(3%vEdYcb9tNMYB88e>X9o?y!2 zp1Q(H5D-(5^#XZ_C@YK{K{~*=S8;3Kg3yvy+YF2Iq}4aUy@;g8G3z6&eh;FM(|Me~ zhOMvCADu-*i&hh?b;rC~N;)?)Iqvj9k7F)Q*&y&U%i2gB@nw%oJA7Ok#Y{1nEj6W*Hb*hMoI6D*udA7v)HmP@{ z?g%Tl@y#C1`g^)Z9w@PisySe_xI{?XRHF_z3N-6`@40VWu3_vL*Bq6_khdPH3AkpD zq#N|BGc^8nOg%>r=j3>foIQR#XQ%^}+``XZA`N3S*X4EN0^ZH3Ci_S=phHE9BXrp1 zAMW1a58kyOU;Oj|zx;2%Mm}hO#~6hx*y){nr?}qQH-xy_A`LUd>@eDV@ja!l?^X&m z-QETAa0+39>Nciwj9f>`r!g1z_~p-ikGy>8@3afxuPYlo7JJ_!K_~^Css!(U$b2l(+*)Z+^f z1*s2cw{Kwl+b7mJdg5ve6D|{%HCDfa9lhgZ&}3YDi(#>i-#Z8V_#3>Re9p0X?i4Ch zverZs@o_~mx9(ylg4R?xFoD#cMeI7R*}XSO^Z`{P+^>){FO~liT&%m{id%m9%>fU$0?D`?%2(Qy(GWGHLiI%NuK? z=t#@g2>mXj*-dQoO;k^#X@(tL$JKW*#}}UF?gIx64s9*A+n`HZ_|X_DgQPQvcleP# zE^Y%gf-pv@3N$wli2Ye~Y>@;y0JB4^6EwE;iz&7_Lemzex{7P>kh*np-iG8z{dM~N z3#9H^DN3>xIW`#Aq0wQ);Hn<{fHe`5O&QHTTqnmBu6YyR-lAWeCiQ_#Ly)afY-A!< zJXsw%*7SESkcTs*#Q}!1nCTd`Sh^h9`09j|Gen#wC;_3qUbUgBH$f-w@qPZn3 zK7r^AP3KAZ9M<__sOuK02_0r+onic(xEP_ymn5tRz%ifKmXbLv&Iz`Lc5;nw=M16C zXt3C6!q%%|UJVU7lp-rPIwXY_)#kX|YL2ZAuyu!x5KRTFkzz-%#BPfe6juq|-a08+a#B(> zco%V_K#ommW7mX=l2VOGE+gs55m%dJj;PvO3`g4-Sz=N^G?MeZ_HIsc%+ihJXS)ru zme)gGokn#7Q?*!nvYq3{cZgjD`7|jArY__3fX1I8R+DoY*= zm`%d`F~Sn6v67aTF=EIlIh-Lb*H8vbRZ}(dl1|Nlv1nK$bYM*}lG9)bohN6-*nu2t za;}K0^Q9Qdj8OGxR*;BsffUz>$&^&3R^(8V`T?pAVTF+jj3EyrVz*Y3)p-oZ7#B2y z?T<_G4Auus1u1S3x-+CM&ka~%4EUvl(ZXGe^n6=v= z(va0-(g9++@*36^t~~;ol;@n1?nYNw-=WE%nxQ$u43;!lQa>Q84a~ur!k9KSDIAjH z8{}a{>Jr!;G~B}3osys)UxNB)uu{_WH@BP+Qe&v7v4-7{i!@2(g>puo&2H_E4NkronM~|cC6t#?()$|6bo1u$~bjw8JzJaaYF;-R`V(Lq% z-}q5Kj2}q9_J3A%`;VjMQFOG4a2}htFf?dVT%7al^H(ANr_xqVDtyQ=13SXjw?FXj zQZ;DGRBJm7-D7mArfxkr18#>~=U7}$84iCIagSl_0U@2n7|1yx^v4BTxACL*Ud;N<+nD8f#GU7pUugKsZ+y_j z5aHR+eFVMnQQYi2mD>bYp*CQ$B_&1E9%+7qXFvNg_1d@IlR#--K8qQj!PZCk`n^*j z_10Cytq2DTLf&FnT!Oq^7ISn$?5D(jAJQ$9hj3oX2y6XKNP8$#a({*pJ_7PqDVbn0 z2b{g#we-ZCM?e(b%Vx)a){zI0%945qWie^LZZz1*teAANS6U1Q+n;MeZw+(PUu=vj97n-G+ZSey#i(n8?N{JDvB`uWLjmdFbiec;l)pVa6$7F3# zosn`wURGE;!&D7yTR*@JwBp;IFfZAKg;1JnB8c*&gzyMTOi4ZpFch}1(z0E*XRv~gCa~92OBs!8F zVWS~0PSZ(69@bG)Be_Q7te`W@`BE~*J~58bJS`JAZc&oh5|R1<%MH@d94GoJj7Ty4 z0#F&^vcWYOjgW&LA3r^Iv>=61IX0FvQqDLx$F~7HNf10a&rq{P3e%GM6+_Au&h^K~ zaU`o^LLix3#?oaEgF}NMMj^*_$PEUAlFBjez#73AhZXFvZ-RFiW|*1yXB&Ho&`d=xhV zaT$@qSSDv;KOp^rFh8VPze!c?usC{xVg3qv^*w&~w+ce86hmF**?)5zKbhku6>0Sk zNqUL6sL^<;n2b%2XfA2>t|4?2(x9Ypoof9RhWTZPQ*vGsR;MxMO~ikHj6lww(61k3 zZ34C$X`9h`T`DE^xEg5D|v`G3stmPhUlRBK4l-!9z&(vxLKj zU;V-#|4=52eq>5V#JC`@rZ9Q`lJRP?i);7FV$o^_p(e#A3Ec~L|7E`P>oL&sO~MXK&$$N9hmOsb?Sj z1TN0U=-|m=hcAEnRrmmj$`jqIPyMSI-aUz%Tx5La+mK+eu9S*AK(;w7Wk%f@!eSp= zefREdzUk?-2CaV-yZ_hHuRXs;JopDx*Du_G{+Jd992MqGZJ6mF10 zOTSoPQd1hpA24?J9vRS~#xg(^Kp&%S7BpMmrCR$v(&7v`o+?I|RAdzEZXs&PVFYy4 z+~CW{rGO?cATAmDQ*Zz&G-ND7M29JPxQO^i@%0GbE)jqHcTAQTH(<3+7!2%R!1#mr zu#|nhN2u3m7u&#<@BjEzc;+*oz}8RUCztV~UHtTwdrVp#D*E{ref2al-(Y#L#&16V zCw%79Zy#emp2Ae?*!IB}Le<(XzFx9_;};;OQZ^PGSrSGBKdNw*Aq^K%nUsYGt-xF( zhlbo)urrLAqPE5&1WlkzQaC_dKr4^U+h`1!ro*>Oj2{?|P9bKBpUw-97fK2+O|Vr} z{05WADI@iq)J+(Uwy4%_Q?Kn1=gY#l?PsWhT_sXp61xo~c2NH%(@TF!xA#%dHFC5_ zuE=2xQbP)DNhYR%Y65yPigEiG+Y>`W){L=xH0y5=hbgI>pm`t-n+)@BVAKE0VE3u) zN2u1XlUA>j`b+f74XXA|F%E_?x!Xo%16za05m`M9o*XIZ$DGmZOMw=OrX%1sBI{_{ zK%+x_e{AAv8Bw2>SMhHn5<`pTPB86i9f)(O=#6>~obf+=a+M$z6NKv+Or znZg@x6^xLgkn^!ms*n{(lm!DoMM{d)m9N6z1r;q^&!mXvm(-9Y)eVe z22>7lb|1AX;&2{Y&#}WA!97e^f@?lb_8p?9(70C2P}?HTf+@*sV-#FV&KnrNAoZKX zuEsT54E*Xgnk!-)BRNB@YLZ$sr=PN9YkF_-c-3M;;*cz9dksp_#pb)FVt> zBg8dwS|^7of-gm7DsUh$uCOkZ5-O_)F`H&GtAGByHWs_l9IEd znijS&fLKr>Ie4-L#IBIwQ60!>OkQ5bx`3HX3YV2RIh@A&h-yXdJvogqCX%8@Y#?}U&P6N~+a@Ks4lV=L@vLEYXeY1?Rtp)QH?FlBjr4I_aZh14Cg z)Nk?bb3!#MvM$khnrdT@9B&fm7s-naLN~^k35rEXq#TgEhnhQ#w_m|e?+}j8p=JZk z9XT%;mKX5TYX#5o0Hsh97Uz+;gNi8yak;{9e`DC#M2xAho9`nbW3Z??gaPr1X7(ce zk%#^wY1l;58@T&S`J_7dqz661$D92AZ_oLafAct@+oIdO#OTaR@4-3fP!SFuNBjlq zwO#zEOnHUH8FD_v`CEuD$AYZ`)6=ig?Vf}9Bw?@PlYjXC@s)r4lfmV!PTeFd_K1s% zUOnah2jn*Kzd{HRbUSi!FxvS%#xcQc&Vz9b50?e}ECWxEhmkkE!<=jxM9J zjjIl6w_e6r#n*=Y{0!$K2>(Txcp}UN-hVeJC9{l)-Ztpbc6hB?wHFp7W+(cwt z(z7m={plOfIexNBING4!*{0su!;V&1?MSPh&`(KmjhqiKd5_2yrW@bHNThDAF?5@y z1kB#Th!Em>F#xSWt)+ln4X8zpCuL9S*C6lVnmZ_ptq!r(Az{9Wrtgx>@9;an`N#a~ zXMY|)J%^j@V46Q7F8^!f;2e@1*a5h6ypp_PQDmo(UOBla>N*eiXtvxq7sJ|%@f=xA-SaXT)QME zL+U&^8)9eB91#ll)K5BWtEfy6F(u`f#>dC1B2wKg$#TwMn}TD|Sj;-1q;fHWIKil3 z2GFjUw8oeB!yJyO9I2#pjVOqcQAiYXDGSCH1W3*q<4SrnCyS~2vgo1_p(Cv-;{Iuz z1H>Xm$hjgfF0ou%aJO*&2IwKiHslluagMJewi*$wC5KapUBlIpdVGVNC!}GM6bxbw z#wn&+A`(kEp9C~}(1fdAL;WGCzl3Z4D{?#_rwi!dFrxv9n~0xcUBOI{9XZv6a2ey4 z_{lED_AHm@$w`RAm=qn>4J4Ukv@fZ5iAbe{PRMGIU{F^QyS5(4p(QNFq&Oz7YE&K0 zC46JZG9fM2(X@^6HDU~|_NaQi0NEf?Gt3>n?lC^0TBCYM$_C^bw!KZBHwc-WTM&z_ zBJ>VZW&D^@fM-0Y1z(cgskcCbCP!Ra#6UI4h!b?Ei2cTK@fZbd5*?;ea;nLxB!(fE zL4!OliDm5&SAp@w;1PR{ytqwT)r8?DDS`28M7Mx6##M9l`18~?OC%o>s|Eeu0&Ih| zu`G@(3t9nj$7ORFjiX~6N{6_JA6r6q8mc4O$pJYAh~FdS6*)~2T_-H3VAhZ%h^-J0 zwz>nxVcS@kt|pQhNS#GAV@8KKJ21cbFvu8N?=xDzPQTp3vM2_q0W|>=X82J#XG{H0 zv$V={&;7|=On_w#Q_l$9HdQ;tG)go13;NYr2x|;!8*vda0X2$idSY59b=&meO~_X; z&6|vt4RG&gV3Y$~1kKx6x1wFYhHhOW9Bz>oTf|TnLZV4{zras+sK>kIoFL2b`$>gEyL;tx_F&7|5@PyPd-64yNqkW zHaocXCWeSnOW#e1{TWhyfnl+Qgm3WW|M`Unenz*Oud+J4gmoLN_AWBszBk;nIJ|(m zQ`F4=tF7ne)-E^zWK~E&-^j)k3RLOPo4UM zKlp=pj~G9B{QZCXyV%}|KE(R{-@d`KpMQz4n6SL{1nt&K*!sbVh!7p&_D2}{BU1kg zU;dY`pNua30Q%Iwno;{rf@^W|3z&Qvs)9)o7f~t6GRboH!t&h`6jvP(^MoO+Q{Ods z8!|mP{P?kspL@>mm-34=qbIQ~FjUy;0)BP}KQ0%7Z#`N?@#7tgTVa|rbT>B8^bon9 zl6^6NzaN8FvM?NOL%M}&0y$nRNu>7bJ;@eJz>e?F#t z_!!T8<{H1l8^^!bmR5Fie->akoUpk0k1*~DOw(e-lHv?owb=0j$reqgN!|HUHfEN% zdPVBb7M`Rxke5Kljc(lK@V<36;DP?&QOt0vFj%VzZuADZYsqPZ^@^?D!J|%X!D*<& z`hZwLs!~9NR`~I4G#ZBF8RiDtSBO}$#$zL>LUk8$a~#X#mV1zT59zL>VL}L(v2{dj z1bW0KkP1`v=-`WK$&}&naa3<&+?{(G-G{|GVVIDIzha!e#xu{4z(0bm*BNhq9b=S( zHz&m56XeAf)~+LR25}R}6}W(rgh~%;F=9xa#Yn_WE#e*IhPZkJ^mUr?oaNyhvsfoo z8)#~a8CM0w7|@JnM_L`HZt0H~Qz~L_(NvC+GNK%@wJFbwD(E_{>2VEYP2{B^#Y9-N z$3<2NV+1+I%4k?eV}t50#x{shOjjqU2&psRN(|RXCaWO^d;Iz-9+_>y9N$B9l2V6ieq zQzEVgM2<*lLQZG#Bf+&b;s)y2673eEQ?P>K$b*u?9!ib19$Oc47l|07#4sW!$Q)3$ zSIgql{spYRfY?)LQj$*5zQH#Yrpm<(1;|PYo*dT6af*ga7!20#qHcoNhT&kFW^D%z zH91tMT5zQmzM8H`gJC#o$T?tLq+85Lp+Pl53P8eFiX9n{F=25Ta*gpbT-{+u8IdKZ zCv_e$s6bp8QnI9EF(#73Swyd5ssl9DsE!buAt+*FG4*ZXX^RviR#C z3gtklhZF>hb_oHMmmsmqHWWC49ty}q-t-sb!KkVfExTkJL6C?DT$U^oQ36{o198Q& z{^cv(Ii;Gv^$6nESHPWhdKEdp?4|#l8FNvsbhCUhA9R_IjJB?nwCi$<*c!3DkK8Xv!wrW1DwSr?3YQmTtJkHSugYvp?zWHs-Vcam zhS(sc)`>J!duDNC~vmKx9w=y2031b@IV-zN3@XnmbHzpJ1yq9=?T zp^j9WGh#a>wRug6T_aoO^=B zwFk)Q0lsqAKDUj#Bn_^=jGOM_noM)*CCE^_=ZNhNY1l%BQTg;XEhUuxC0ygc^~AnK zOtF0D6w&mh>$I0Te6xj75D6lQYU?^|UBBb;<#n$=Pgk2*o+IK(-6>r4Et2bSeM`)n zNOysxj!Z_(fel!mlk>BL(Tm^?k)s;3eJo8GR#U3#POhw*9MUc3r0zWEewhfk{aOZ( znIN)y-NGyslQ&e)(hWNdi>u`P655VY@nnlge3_xU1@bJj`@YHen->Y;A5w4LPqX_1 z^&J?2fu?DJn{=~n)IUHs4}AF3k>C5+({H%19{XtGL%;eZdf$M)2lD+iJI~`nhh>K? z&y@G^^jU21#D0&sY9PKuwfQu`^>6sRqum!s-9vQ8_kcb11H+-6`KQ>Lou?1)N8|#{ zsXGnq*UGYEapOK>yN@lO;jlbZ9&xHdpxlU%2)Iv!70M zq7$9y??FG32Jc$G0({kb|1W|6>9zl0X1@&FY?|iN-^1rS(T}QTFZ=}?=NsSwEWej( z{4&+{ereght*i{QT~>>Wq~%SzTTc^*KRdaId58MB-?%`|58&Kc+-MWE3GiOB?$GZK zq{BnP^hMMiVYwxCW8CP?OGG5BpxrAep|yr~8|!C?f9H(a@0PxF<{oN)nJ`x3(BSI} zuQPYwNUmPs^DU|kVOVZr={#vTD*NWG3RVoHV{fEp`xdc1 zhxhNt>^|zjnTTsPhUjL(;eMIPTO=xJz?u6mMY?MkK8wqXQ)3# z)`-RPq)r){>x9iYLwkz2ct1{ds3rk7-Y?jQ)F7!L*A1zL)PIvNUH)@4x73@La2qeb zt+{(!qKRhbSyF#5?dn0W7YY+McEueVs#3u88m*3Y=?;Gq!jIvaDb40p`b9}BR+}&5 zTxRG_VNAfzV#z=iGl!}tGeyJmi1nmyhhe#(-n{b4^QUe?>NiNM^W;1Mn_w8R-NIr? zvANN=ZC{*_tzIX#6eA<5h*DCLF;GvBuwhC{EtZTJN2r0x0%43XBBd7LG%~fpz3=f}ix{ZbV!JQg39oN|_d!dbH0lUr>#Y8S;qk@I9o}H%Z-p z;NSns5f6XrUHE1PZT3Mk85o~`g4Oj*9)66W{|9&<@D9{35#M7eAjxA1Vy{?g3FD5i zk$^;8PD=4wcZq?k=;#Wm6S|`bu4)R#A+k=td2#}*L#q|OYDs-VY%6R4$%Y?i#LpP! z(}H900;Z+JDOWg`5x|c#wk*bLTy^C+!%-P;?zc&+9n6-f&M}!_!YgaYhT%RwC8pT1iW%e?Y(7CwGTb%@HQW+!l>kw*`I!bqh$l#MKTtMRN7(^4kOyx zB}}dmsuyU^UZ+30o77)mHJc#1iTVoh39Z_a(q>dp7=t6N0>ffTP8C@?LVW{5i>qqF z#4y_>u4+U_NHxOO#rRa_pb^W8Nrf0xO+pohmBV6T&bp9@aZEji*zF*)Sq`dtT$lq1 z>&`$UCEac1pAlJ`v1_x}iqVS`=#12YRVA@6WD-@`w;GXNuxzB1(3A+{RbdZni>nv7 zDkF>$T~Tk{M06WVoAisHz?T0zIX%h1eFVVpfFliSqcWET*en&X-9i|njl-lOgu-PsTj;J>eiHq~Jx8B9DI*nF|Q1=Y&>GIhZ2W&V^+m0{^U`uj-fzj?$i0>Jl zdzy6RA%^aLKJ=^4@%Se?9D$toP`6FJ^YvE@PP4M;>ea8$It$on|%IPZoT1SAOFOR5B;xyOgiew-96;?AzWCMMKrrWt)93% zhYby8M`U}Jdi$wTn*1hGBI-SE^d27mHwU;+-$3-3-}!G2zOS$Uzxc71M}GZFERJh( zet?uGj5l9E-r#zvn{BU?yZfw;p5h{3dBf*?aukt(p!&~!zPjsRao#=-|Ke7hz=RWtjuYLH#AO3GeLZUlGHToQ4}o0~`I*mt_OsXiF3)$O6WyUNelqgNXa0oM z@iB2VBlhR%XBQDSpw&Q%9*Z?8&4`Pq$>|wh*1zP-pLp}6?}>^Y`HW+E^b=_H08Mil z7X&xGg$_NjRJtSX966q$o9!Ta48sxqa<`0Hzlll;%5nQMHN`+b-=ppRit+K)@5ZKm zM{@3N{Nw`W4srP+u1;@h?%q~4IwW=5q94)v<8uu3M`wv`LgI^5 zXP;wOybnv;pnpZZ`JJ^vd$&aW`_QIka_T?g$H(-mW70g(E$#z5O&DdI>%bymbVTY; z6T3Z@S9kE00}ZHOmNXgLXma?<)2|+4u^$=dzagc&G2Vsv8|ZkBhHa2doE0vyk#W6{ z+W!<;K1f{6a7-5f03ZNKL_t)nP=AYN>pL4G^rHn@_h@s9fA$+Y{9hc9-61(0lUAFA z$!k)m%MJRY_mPJW5Sk6b^d>_)Mr?$RI@Cujdx(IwTp_P=^Uxepbh$MMu6p#$QA8M+I3yHrE26XkQEC< zDah&rZqnf^MSUc=o;cqm#dFATKwLOh2S1J9_J6Xs1C!HFaQxyH zmhK~Cq-4-WEEh}z(wEXE)xveulA09t_)*5JqVG;Nl-Td3bJW^i7!!zCeZLh47PxJ6u?BgZMep5ekK zDoX6e2tB4PaTUlz8T%J8%p$4p$bF4Phvh98?IStkLLzl%N%KAGy#q|=sPp*I0Xc_~ zY!oGTW75)-@MzUC-gydN@6#QgM{7kwB+elZ0T+a7vm@sK*@#_@s}wZh{7@LDy+a)M zO6j9OUO`%w1%}mylK9n{*zcg4aDG9zI8{visxL{<9AKT&j=d3Ag{N%7tEp`ja+Cu8 zNP^%|1*u@@6pIYiP6=p8GoUYesjxf_{^n22?gkC|YO48@b&hr;dJRXej%i zdCeck7L zaaFYejd!--Tvbcr*2|>+E>iy<;7{K4v5)`PGd}o-U&K9g zFM0lMWV8!%56+KCzQ>j`gqKl&oo4%G+?#Yh45=b7E)csjkT2uiyI^~Q=nZUG@R8p* zfM0utTQ5Ju7eD#^$>x3RmkxO3(LbUaR-||r!|FktkN9c!#e@S|Vmb02^H zd$8qyzxv=OJ*Q7M%(oN{Iv)Q-_x3+@q7$9yM@c`j1~0(x{oe2W_Rs(P&;NPO`IBwi z{##%U{GTFn*XKX~`8ON~IXXJhx~|7S?Y-B(>+_xHL~p3lqWqUU^5}DPH{VSk<QR zy+dUgv%h$VpghIBl`H8u6-3k-`X z!~6niI3Om6cDWQqZM^Zm&VwW09*;(tUB8dmUP1eB^4Pz8bC&A&LJvPW#y5MoDx)?c z>n48u&WzSFhCQbmMS6RJ?&v*qwns|uC5*4(d{?-|*^}a!)HbAafXEfB3*MjrM34)eEKUrrQPSF%xnMSig;`L)-y= z<7Iq({S|JXWx@IexQk)+<17x&6UG(DfQjO(i^aej-$L}DFb}hkQ$=bw5$;E91o0&- z{yDrwDc>6@)JMdy$!NI)9ANELa=lOAJ;<1nu3gT~u)Ow95WN@G?e)!XA}*$gAK~?o zxY{HSQ!rnSfg5o3n6w;``ar4|s4mHQinXU9*W}?c-K`N=06)TqM5t%PaElnvAkO2O z0pAFzYY>@`V~ffh@oNMC4nYw~WNdBVm2(9xW@3t)h}DQw!p2ca`>ww}#a75LhkQ^< ztZbL#7ycQO(_evn2CW+Ejjv&2hfN)d|rNLtC&5#re8z zbUL(hq}AzSe!9OV^)C|Teq@O2BDGYo8pD`8cx*nxk`Oy7XogW&jOtvl1!gsBD>NJ- z(iOvVuu|?;^=P$#LYp8ka|x!d8P1OhYpby zzNxX?kd_y*ID7w^0vaRE)Y5H=sR-N+}yeDal>eLiQMCgkL3-9Jp~0 zT}6%)hK0igfU20H8u39fy@ci!X|)S}pD?;fik^P9M~-8>D~p376>(USJ44!puMaSK zY}iBn0ymyfZ4?u`pPwhi0dYO98W7Exxiu60{g7W`n4M)O*&+MLQ2Bs7vA`Lwj19f=o7@g zhJ_M`%T&{w`0?^}MsI3IEN?u-uzCPrRfOi4YUep}zeA37;jX(CmOQx|GxYC8(k9Cr zHO;B76V{U`v2ExM?qirgi{vNy?88ew{J(siet8PJ@iWwu7Q#ls4Y>{@;wr_wl9vtb z(U0R+*O9Qmh67x9xfF=5HS`tTt;-B6M}PD*jlIa1&p*J=eX2rIL~KdY>wNCxo2~e@9k#m2629#ygbh@e7fN`KlbL6JEaWmZ@`}7Bfq^%-+h4bW`)+r zgvk}E`uGma#A}sSTMY9D5IKzx8zd@1?w}f|ypV_U48yxg?FFWn?&9Zu<4^hQFTV5X zsPCp8`}h%$eC7}6X73`-FXO@~Bs8EN8Ro?9MUZPm_f3BP<8K^}_~a**uZ;g8_2!S` zCsRhJzV?bSetR#fM5^fl)%1YX;c3LZm+r{%bD!$?>@WWS-jAE`@UPv4$|Z7rnm#qu zQDL*^!@qur^k{O<4m?q37{mUk5_~%6*C?9;=ky=ug9^zANBCBJxDcqkZO9G(e4wkV;lRbr=vq+ zg&~B=9Bo><{kyPyNNQIY8<3QG^X8pAUaL#GkHpcH2g!>Saq%?G_LJY;jeG}+VS?8& z;+Ld$7piT38Z z8Tu`vVEBwYGmNzgq$;%Ax2Gw|k8y(>5!`GMPsiE4Up)H9l`_B+}k@nUH*f{;AQZ)4r zrBw%RizEe2@c1WYJp60lLc8m`wWkYfQatJp z@ztypb=fgF-%XC5+)uGk7t>8g#i))W(y)UD!AXm_TZI#9U;|j{L9E~ss+N=+kO_GJ zCkb3!$2ciBs3~K)B1Z>Agh(95sPo_jUoCOphMmY(96(`7weeNr>J+)(V{!GvRFjBn zURtw+BcYnn%}(P&PnH98G+URw1#+*YM9F$m+(MY6?we&i->#G5DOj^69Rr#ZmfW4^ zS-*pKHPz@Cts20gGDGSXK_QMwX;RwX)qnvc1B&k(^57V{OOh9gnU^EvXrx}qT}f|R1j(5`8$M8;M+HSO z6H@E3ETm|d6{fEb>-Xu@P%SLsb}mQ;r$%b^HOy}h8HSaSQ-^bjJnYeT5p|w0-6zMp z8Pa_v1?|(?2<;i~Tmu);#f02$5xX722qa`Ad9t0x`fbDp&;)oy1Y<(#cd>2&jfFAn z4d)^`D(VLOsD*k(&K}81N`X9VLS5n1mvm|8G1)-92lF@v(sCg6$GBiv-XgVINNCAW zP#CSCuER}dC15SbLFw3~mk#Y?d5$S;r@eo%IhLSAT5J4MK zcNz&9OM#RX2}eL6uBNCL)ElxkcSRkF5nCbkTCgGkoWIR(F6Tk&3X%g;tS!cV?jw0L zK;_X8OL{YxMS+}eA<~obfb<@6jft%b*bE=8BUMEF2yqiW{2M<_2zTL|Ep%ek;YHBX zWvW3kwy5xZl?53MkQ|mf(0={CyGk|v7M7J1M`SyL zSk3a*9=Q|HDL!mNWxxvOGMcx^)dI9Hn2~-;n0Rt(aii;Kn9&}*3$rCYMEvAs{AgZ^ zzRWRndjxBV(UZEf2+4CU7Y4pBsl5_wP9?K`%Qc;SWZ}qq<)Xo zpQGKsi!gZ}tzRbRp?n9s8S|g<@YtQ0hUGms+4#m)v^gMlj@X_iF788XMRkR^B{_NG z>=KsG;u|AOu2F5=BT^sDobjqYZTUVM`7@D%Cz0^RI+;_?YT z_wjE4Fu!(ZpDE|oTGi&uw6k*zhYvBd4}u(_a)jSn(9O0>^0VC} z_8X+-KK*LI{D@rrW2!Jk#RG{HTdY0g!~gareEkH{R*2ii)kkk>u)dKL!&is&%{H_7 zDd5Go{~l&=bbjsW(gLAPqIE z8+W5_o9Yv?xFI9Cv zuGwUI=83mGcBAbZIG+%?kJZ7P5B}l*vo_s-F!Z5c-NgO+PvXay@uRai*I>TIOD0<& zMNevPvUK8tsH(_p4dN4Ts+OL`%`}Nqj%m^eXkWPfOCi$P)3>GsksYa5%+CUQ(SGNxJkAMF;egR8NPb!8!&Y}T3to#JfsI$ z&0eD3eCZu;mt{fB(AV71c1K^w=L>Z0hjEiBuHL}cJsM`DZWj#+=X+e}N!FCYF{#1N z@nQLD!InV1F{7J}$jRZI$3zipan&KIuW;V*qg643;&x%d&c~SD!eoOS*Cjr#L+LOK zkrA;^`0-&$&sl=3s77MntP_8+c)t=r74VXTnq5ECoY z2%@_^4Tj}?xZyO}4w3W{S+BxGg#Wdh=^Z74y4wR!UR=NsR>ojuzZNvUf}R+9{_GL^jA>31sku9$xkl6 zpSo%A%`LE7467aD^?T5;!27=7W>P3cu=*l>e_WW~I*|I!g2#xFlx~riSIK#UW-B1| z{rF~vR>uWRgd_H2a`ynSyF^}|W2kOWg$5t~c15_(tx&1y(&#(-IT;-j`z5(O1-v=J zD-AXM@%^a3h#w!|s`4FBz6Xy>_rvYi5<13bpQgQW8LL4^Pp*|4%vsm0F z#Tm{WA+NdN7KD^GNpXt0L#&!$(xZNP$M3o0_iDtsIXOqj=Sck}a=i9RWBS`_XvnD} z=2verzwkdUc=*v3q65xP5G|wTzP?3ltu)z-5&H&_5jrYH0!gqOkW|6#7H$Doxsu3B z1K`206TLa08u4rMFb-pehB(9Ab;9%phB4m$OBV4WDIVhWt5~i|p;rHHsyYxhZemPO zHwEtyX-n!@l$;$dLGEkf@-D`wo~Ihku;C)hqkDlTsKRw(e~Z|@i=3{$?qZ+%mZ;}=@SjJC#Y)GXXQdg+#lr&=M7#0hX zfeuEDiuSG;yfK$DEdzWhSW6v<7N)6hGA=uEg19R92G?Y?FRWfmwl?_+_5mljdcakQ zoC5u<0ui)kZHQY*dq!RNd`TItm%@C_J>(>a&$tF$5K`Ze`d#|j4sPcP8Zx4ely<4y z9FY~N-K4*@Q;wnL!u_?3#Kf?uG4ZIc2+b`pCH13XN_T-g42T}##+jVYf*)f!BU!-; zv-cI{K<@T1+r&B0`nd3p7t@kn#-q*wec6ZJf$kKHPd_6}2SgwZWkDcEp_&RlG^M=E z9GZ;0^q5rZ&zOuzIR@$m$VSSJQaskI5h05BOsM80Zv_|aGZq7>3*@NeSdsH4I?Rc~ zA=UUC)x;xeC>d}_NTm51wNVQaz)wmzJ`ZaXKi^LC8qQn6Y*a?SP*_5W&@W4ZSZML# z7_F8-#)lQW#+b5H6Kzb6(^5PrGsMnPZSFAM`7&z6f@)D6mTAc-HfIx~hao9kr4O~4VOCm2a zsVxS0)D@{0tUm^wLA8Z)3AL*Xi-H|d*W;Q@v-Q#&Pg_aC`#C93$$1+;Hqvk!H)t`Q zmltu(gfPAdY+$h=>jd=!5;|1l8pk6bMzgz5H`^etYSM6)6oqX2Ff4ieLw7ESKK$q@ z_VZ8CY@SDSf;L-((G9eI2Fsq@P8n8vB%6Sp!+AmKL`oAZc`Q4WV+aa5#uP}g82UP3 z`67@p-7LJ}=8#bnNcl2pHRa()U%>1duDTDuahmCc|Ch9EXb&$zzf1OGq6JI^%MsFb zRA!9!W(>!d8IA+}av;fbh`UPFACtmE)Z2Fv#skgH*IqZJ1yHk3=0%2s)3pA*q+2sA zPBHo072#$r$#)I<_78rIe6vqCJ0Q0AF!a?c2#ar5T3*D3)3~sB)!ZFwY;fRaRHN5l zpQtmMbI+5nJwWI#^W}?o1JC}gw(a{>(faQ1ygUFA!G;tn_{We)iwZPm;VRI?<1=P7K}?o#;eA z{OaW_Q@=yC@#ZC1p}9p~-XNzFRzs~71OyQF@Kixax; z5^?#LB>MwK+ke65Kl=U6z-o`?gwX=}yHV|k$s<|r#7?}4avDKyG0Bj7^3d}5e|LjN zKJy2(N0&*9^O!q@_Z2Atxg*;R$S?DsTzQc%pZ^KsaGvh)6!q?%1u1W)Qc$%)4%ZMj zr@eKS+-{R%T@qyJ6!zq;>-s6DBti&7@e9b3^QEES~{Qg@>3GQXmfvd0687 zA%2=keNS3W(dipEWZ#xHE%~S|Mx0NW?UA}|)ZctZ+ZEew;_5kaeD(1UKY9ucJ5E~xj{Q;(5W4!S+w+Hp5?var2($h>|!cQ0U^IasLM#2u*4x-z* zdWoxUygnTV#CDS*pF`69v_~td(UbJ6OGvr^Ji)*J*bzVbsh1i0OZ3NkG~2K4`%uq_ zu_JZkQt+02DMiauR4Q^#n3VmShd{R&l_a7#T$O=T3|$X7JABmRd+pC4w3Zpdk@khqVyB;+tH!*GWn;v<6FBwtL9LUJK8-$t49X6}~aZ4Y6xz4{E~3 zJ~>V>%fzmSv;j1v)D#nWs0*^f0^395+!1Qm+6bu9Va}a1QSbS1YcGq+W2L@OM+YLF^p;;s_TjQa2{I z6~3zRO-5=(@T8$a&5_8Ez*S2uI7}xJX5?u=;Tt02$q9=~a*hnkyKzpz8R9x}8;W^9T9Uhs zVggz$sp-&TwpSL4!xD=%M8Qv=f=;mGNX|RBEvTnAUyt_rHaWi}^=C=L9;!wVhwc>s z03ZNKL_t&-W#}`*@*+AqrW#)%#ruir98O1Mvx0!>0!a*&h@q&Kbyp37_K@=&Updo(bHe0F ztgq;nn~(!yN_L51Il=lJ;_@COOz|6+i2azLiM-sPm-E!q4VulHj847qru$qm>fP(O zW(D?s;$lKvM%cabhL6t(DLTR%P5q<_iv!w!XC+sUeKhfrPd!8GFVG*{ML2iob!j$F z9?oEHlY0E@?H5mbbOuZNRO2^J{whYb`2uNmA4B&pe%K6N^?OLYNf;k~$3N?v7AnJs zJwEiS_4j=_nb?m}Z{1toqkC^Z73N0^Mte__^ZjtJ;1@pq@>{Z@PjsRa{m|;f;62fa zPV~bs)xsYQ^;?^-)z&v@4;#{O8JS(gO|PM&7LmG)!c#7y{AfXcv`cqhGqVnRcQRx4l(s^ose9qrrSCc*Ia zF~fXCCgA)O>o;NYs%e;pnlx-;u_BA2HsI<7vG)uylhQRF|HP{!_+t0qCl?6QXQ(%3 zEN+}$mj)e^0Zoo!kr8Q2npBpc0|J-^$W<``b4B@{PbJ-_k=U)UA<->&$;qhQoGeSs zA}+M}=`n&3=eyWy1=5it#8r!Tk)f|4Z=luM@UaCJdvfj(tx&&T5?n(WzmBU3mRodf zmRcXkDUen>NzUi+j1gB8Ci`V9{?!9P>?Y)X zjKvKEkGUO)KTUUh5p+hFoa2$tjCky$D>OgBaC8>c_u%UlepJqXaR*p8!%~B#goc`& zra(nDE9Rvo1cT_X&VyH~%}DAzL_u6cT#J|oZ>0Hz)bHVGQ5hqCj7u{_l$0wpD-y0j z{ai6CO<>3+6|6l~mB`YpP54P!(CEkX#|?rICWd$+brqrm5_hrCLYm=#Y(3(8)GG?) z7MK+CPC|xMSgWa(GNN9UG+2)1Xq3XV!rn!UG-R9~kXljoq}DMkm1-pG2Nq009!l=G z8b?wFmY&oKIZNT^t{$7ouS!6?$C4){O6ofmF0@iTx(XDM$*v zVIx7+un7G^iLD}jjbyObU>Rg>_^y-K2v{u1a#2N`;p+7Nv-f7dwq@6O-*1dL=UTfv z-EFIGbu3nqqB_eG?KrU!#0G*SghV=eC_N{!owDP=5BUrD2Lv|YB#vZBb{_I1B91KC zP@ESli2+BJZAm7@Mr^9sRCAki?`d|k)|zvUk%zgf$Wj$qMMk11>U_WfF7Ca1pS{;! z-Tvk`ejir7ZDTs>I3snfSj@_CD|fWi2m@xgD|K99_49=E8dly<@SfTH)i)1qcSAf; zx3_UwfFB@T-%ulcp>Ajk7d^PQBHB2=Z0zxpu$;WdZH~9kJRjoX z2xY(UctEQ{94O=C#O@f0PE3Ybz%4XhurSrmi*|H+4awmuH7Tm`;4b|u>2pQF@S#lt=H6UV%9ytJd(F7 z?DSc*yN>qPk=Z~>Tk`mjI6>4SrvClN_zX_QIpvLchi|30xTHILi8}lT^O5Ryq7{@!s5alWBjoKlB7Jlef=zrNCyuJk8HR|fBu zu5_jEbGg7T_9*Yp!<8_R4nM|C^oIizz9qtP(AF(^X)=Yr$ggO+a&F|sEr}(EY zeg^*Pcj3MK0V(lCZ1)uCpTg~uyuS9H4BoQ6ft5?>?o-#-D8px1=2zeTAI14~-Z?q& z%m3vWAN%>=BX15E?|u+{{Bv(MAb&G$&u?L6PTp1e)pZC*bc;1c&ft%LBsfYmjOy+9B){I#1x-N8=3fBcz+!__V*hgzKHu9iB3s@9?rj z)2uNsfifOYH&0Ui1XvOzwNfGq<1|s%C*+HNi|RLw3%bDHzfI~^%nrYZxuf9$As&%d zFN5AEUp_+}@1Su5VPqN-{W7hZ+&V-$+#;e=`#T>SUwSbxZH^Eb>5smQbOn|2 z*0fnHcnC;+1us2i3WQWaPO-e)^JxpBmw4P_14vy^r#afaPMi-+7ti8s3A#qZmOQjk zRvQ*r8F2eFt6TpYVb4HRgkg6GEnXu z#s}2}NH8LWGF563>eNQOF{n`x9IuXaMmMjxC~ks^p`oB5Q>ny7$Y|<8BklzsyS;fk zLb$-DIT8q%WR)Xfx6JJ{U{SSJh%-Q-PJlw(e`T=R_K~&^HXMtfh=BG6J211x^UVv!?KNhd0X=xv z&NI@cLCS7L9Y?HOL+lvKx6pn>=WT&(THhvb{}gUNNL<{;tS#0xvp#}Vq$}7iw8BOd zLco22b{T?FcL}$azOAD$p7&S=O-F>;4viy(KwoOLJoo1-3Abu089cqPIV!J*d zs5lZoAysOM_zx*y&ZRk&4WjdS4(~{lkW6VzOvwiRkb$lGvlZLJcplYZE zYAxj5CDZyLA)QjENEuK0$1h&I>+kyg`ui_tUii6x#rAxL=?Ag#1N6sV-Je&UU>bjj zz;lf~zq>)1meg^ijIXjd+EGgQmKdnetx;XDVa2z@xcy#NsBe^jerF}qq78ACJ9+p2 z{6{IjO}D&8x4i#t7XZ!8@d!u8)}+MZH(ckp~j8SYTFOLp}b^?_;o5^k@N^)vk1U%wB3{k!yD{($t4UVM#@ z{`^a%bb`wdqXSHr*XRx&e$&U4?KR3c!|Y{z+hLpgFn#emN{9T!E1%*w@2pS}{M8>O z9e$1O`1Eb5o0@xu(wiHplqvka7J4KD-YJ zH_*@%yEW)U#3|DfyL>;TY%w`P>ihV}-#z8m{?^C~e`iKHSs-x)HI!B!XEs9V=$5CH z-4Yva6Ju(e;Jl*jdg5YixB^LF376Ocn03NRY9&LnE#t)sw~3HO!hCqg_o*=>9Af%u zboMGXK1Z&Pqv42JD()5WC1ENcTb$QgI@cG3;WkLd>Oj|D&>cU(Yftq9>U4>@QJw&ro;Qh|5Dvw+xrpz#1`w85)eDoKd#J4XI$|0p3VR?mZ{ygKofXrTp$1XpXNW29D&r`|?)UGV_mKp(%Xm0#GM~S_* z;x;oh8F3DE3mz&)Gpa&tX<@|BSkd5nui1E`{n1_8`xYSRKAmXv$74%$PD?xlyaY@u zED9l3tOm4qQt(!QXWEjF<3yM{PHL>(sKD(OQ}r!ndmUs)m|vo`P=-n!7ML#y3WN%| zqwHp=CgQSU*;@)XDHO-50|gC+)re)mE5u%?VyKOz`2(DYgo;Q@x$4w#F;cD(`#{aH z8Q~+8|*I_#dU8d&9bkWMi zBnmOM*QxuptunUXE#8PO?k&t(tAJVCPgQUONQh)aDqadiK~d^>NXd~Hwzw4BBQ`8s zdN*rZB$?l!Y(9#WK!|I$r$5;et$q`cXGjYWJwwVdq5<~^^n`AQx+0P&gHxv&A%Mk* zxS=}0%&9w}Uck6(`+y|q4gw)4M#Q~R#yM^sA#_OBwnLr3N{^%qLVUFy_hARMZ&MRu zXxO3&YCdLSP6!j#oxFJpPXiK~3GP&a4(;)>UvPHC>XMp|QO{^tQ1mt~4jPfJ$HzeS zKu8UPBB3^u-1dBC4_FZJC8hLC!$8pcgo9HoU&A&}0}puNmjcuIpF!lq)ZGKtuRgXf z(G0i`#Po=;y3c(1DR!IpfBo9l?;1Z{!v1(vaM>r}7r3o(R;0dCi|_ZJ+5SE6(6}XZ zg3J4;W5+}U2DICuX^ZN0^7;;CbZqk^q-)4*h4?kP?idXb30tf@!FYB^NoS0yt*_fSss93H`~bU0KT00%A@UfpKPIjok`@mT`Fb#* zjy-iaBCnsPtUt)O{Uknp9^T`9!td87UOeX`|Lgz7?tF(2A7I-4Cx}0R#3jgeTo-5^ zfi1`*G>&v#q}cP=tB3SAfA7u7leekMkOCog-_dLPUR8}lLxYVy%-_8u;rGj}rC$!TZIc-o(o}DzDJ3zQVAZWA%fWMqu4WbKisVW`*lRL^tH^Az^wB8=prPJ@fgi==`;2 zs!g+&>fO$9`!wC`l&*Jj{?n9)J%90^|39ohndl{=-|)p!`-hb61!cHSN?&7s@+$fA zX{?-(=0eRQb(rBIbn|I%hAA!qn|iG75I=1Uzel&g&wvRG^5#0C1=WhgvGH|l#L5Bq zmK0y1yZIZG?H8CXpTo9KgC0`1@5AyA5BJeHBK{EM9QPASzKzKV+Fd7vW5#vg=J0)k z)z|Rw8e#bo(|7~hJqh(TmLCJVfdpwO%x(b^2=OV9Lo6IohijC}Q^G7G2|}otcW9pw z&45v-L-J5zSkcXIH7=wskhmoE6Jh=rsFxqc!YQ#orRyJJ-34{JhSyv8_za#qtjv(K zB-au1gheCt5er>2?Ix&Gf&?KL=FRM~IuSfGU96x6;;iveqi(V3oSHMk_5iUHP!MhB z8;v72o=__ID0ou(MThj6uFpv4`;qpDRYSIp2Eja1&RbcYNE^v-OzMEyQA<&pHhhLw zN{*9uq>bz%z@lkxHYvzaU4Mf)rwn%oDrS1(feG`e%FizZ)Iw3 zduFRf)Rnp~Z)+uVfsh>S49nV*yV8=uEuqRD579LufPn9lqS6=oO^)qm5W^{q^ z(Tv$)ORWov2@q_L8i=^h@a%-eB^pPxf7Ht6rW;sMYQ98tL7JUn4sm%w9Ua?TgA_?x z$tB^WhBMiOuXBW58izO--NMkQ_}Ge=BneOuNz~j?r;fPD;DHbv>2`Q1WQmw9z$2j# z?RQT?*(f@P)@apE(_WptYYPIQE6k3!xJSIssDol<&q|GjkPMN8=Ov{W<^x_j@-!p# zO4l_mu2ms7yl^W%YyE4~y#Kr18|Pag^u9+1Kr;#}W=_qa?IR(hY><3^zhJP$O9#@D z?s54LE9aEm3X@~B2knmtafT65TKc(8jcpsd1rjpkhJL(*4qH$u7qJz{tIV`$tGmp6#BGeUSkp5_Fdkq+{Y`KYWP2i}|XH*O1HGl-&s$o}oWDXMX+VcT7z-&h;tPRxlk9 z=osg-_jHf&&9wW%mcwUHDZ39N(-VaG6GW3{j+?ZP-A6*VW`6QA%(vu=o0RRdY)(6t z$Dbm;(>?8!kUCJy_m{aBhKH2-1)h(1H+*b~{@Mqwbfqi3tF8>*D_!YI-~ST7OPwB(H#gt;X84;uE>%27YW*x){sq7C zH(vh3dy(H|ed1>(UihVd!S3M>pAFdh&r~#_auVe)@m>D&zP(rjHS4D@0qW&g)%DIi;_!Kzx?C_;I=$zljacQm1FB z+v|kI+Y>rrb_!at@d(rBD8u_H(_P~H0bV=u`W7`WTQj%4j_7N22VZJm(8qh&xPkH{ zE@#BqyC+1fDCP17)Vsv~lro$$tZ&fkR>&=Mp3AZ)m*Gb1;qd&Ndim<)^ z0mvUDec^($gWqK?nQC9ZKF<~pSZ^OtrXyr`Oq`z)7q4OY1f#Un*Vx9thus_riiv}* zn&ILX2xpY-0hl6k1hH%E)8P)G8z^otv~nUs1u$X#`7XNaRhz@-cb4qB}ZA!vM^Xen&ss zlI2&Qs-1NYIe#FLf`NU4WHoPpq;gn!fd$h8Wx0t zkPwY{XOPd4Z~$897MIj=fJ-w}TalaX@gEfpElC?TqLL&T-A zqy?;2s0qZ0{YV`X8bP(iOXFgo41uzX>>l+95!;vBiFz}8Ys3d7PlgVOv^;>_Gr&1k z3&uUDYpfnKoIghF&e6p~;_@NW!y9eLfI<5qahJg3BjrE2+|C}S;TFQn~`^| z01DOpgh2w1P)+Duk=i49K)673hDO7xVq@EvQfiA2F}2igh;7&LaZmHGwh$1z(9COh zj7rU+VFbKVcLkYML^s6MJ?gXs@6eeN`%A1Q^4h3*h02^bJEe934XxlW>H;^#YAekX zYf0NUUsJ;dud$`MMO)I+BXJ(ES`Zn@!vfCrH0Yew5JNB+R81=$3-vlTaK=IYbENn~gw+ob`+>ClO*CD$YrQ>z#2&Yak`LK!o}}}; z%uZfoir1Mg?=YT?gvBQjed}}Ay2A2Z%E9Hk-otz^=!IWu_?%Dt%zhQ=o2a{am+j># zY4aqrqgRpl*x>-MJAa%yJRtJUg?ishveM2&T7^CF>)U&L)pj3U-W^SMr7Kiy!$Yzdy=v_MD!YO#I=ka$d@RayIB4b^OH~gfsK9r@Bi&7aLQl& z=aK%4hfsSo6;`_)zx+4XeB>Y8WY&L})Q|MbGfMmd<3%D5Pf*Jp!u<3t1ziB4JEL}A zLCOv4c!K+yIJ-yL3eJf88E{Toyh2)h{R`l%|0>(fYn1#fFel#p*7ylW7%v{DPHWHy zP2_=Q-*67s>fYkn(;`{AEgBlcpZGfM~^V zx`QuKF6NZsX#irs#<+`W0tP~wQOgp`%YD&M0ykU-V!GFg)arz$1A-h|IGf;2LNhU%+U}KMM0}_R_sDzp9nFkfLLpefMbG#U(OvLUIi4&;M)xy>j5)xrP z5t8C1QHx^P(Y}EVYz)+*Xf!0fQE=5bxN){a(m<#IO#p-+l18)(Fh)}3|2E%H@I&p`J6z&Y6cUWIS|pX#!)oP zsZ#`;xX9QLsgqzO&@Eeun|nhK2+|B$6iR7dTD>745XGvZ2~-^0kBxWB3Cqf~>p+y) zJE8MEH@9G;?-TBUnpceHbI=t*deBF7{RZhj%ecLT)elg{C2_jdw2akR001BWNkl zh!F`hL^653B<~tOTLOgcA!&J^G96O3*D!++6j-&=zqqFCn)zutl4fhBP2bqdu>t|F z9d1h0gl8r70q?XeDooz+5A(L~fKiL#B@+6E9k8~?0ekvH0ZB%v`y{ih&@iJG!Mu_$ zw}kWn^$lKbq0ABH;HQXfN&O>ac19UjSe`>oRMGZYeNLSgSQ!cF9MOZu(4}BSv9ZTv zduqA&yXPS;7EU=s~OQbkq#nIRbB-@p%mCw0#2;Kf#(T;(->#12#y(BH}ljawc z@d%Aa?R*Fh2a{rTVHSO4nEAkN1> zuKb(n&(IzJ2(f#XZoWlk6JfE2ydtJZnL30+;_M!=-;uZ1$s0km&>fuO`3}?WCc~r0 zm|y>TNn2Sj_pI{={GTtr_dvsUTOaucH;MTvEZtlD=|erDpQfBy57=_Z>;&$77vd+%DP2QaOVsKXJ{_IJp;Fa2S?r9U{OE$#z(T+q$F zy|l3*I6XCw%uc>{i{!3!r7QiR>B``}(v`0C{jX2_%*0RrgU?VlM{L9MEaK;g?{d?W z$2r@F&r#N2!2C0ByV3ksdf}G>!}-&gKTiC`>#*v{TjyiHFd+2-L0{n?{nc0AY_|TR zmGdiqwTJZ8&kf)IHRf2vBPnqZ$>DO;Fg#qXXtURJ@zJScG zF+O^N{^S+(9mX|noH1V9rsOTos{}Eo?G92$SdJ`jzWkowdy0i_^$O(28P1*t?o#R* z)7f?UYxmxK-^@SHG&p(vG&SF#?yjNZqNSWLU<)KDTWa_!6i6|k-OJQ^Ot%+K>QN^L}4)^5r2lM{4bfDYOxA4Vbn4(+dy{jXSCxqM;Rqh1n&d z#&o$t1VU6o*brxmm3bSrpPC63LQ7h{0m>DTW>A%YNI|;^mx5R;pfYRe%g{B1g?m9J z8sFEpSgC}0#+EBQAAlG^M=X@at(|6V1U>eM^myGjJE41&!UQt4aqN<)!wq!)Rm3vr zB_!e;o^@dOJ9{12#px%;; z)d5dvXs<6VGa9vxz-MV@w}2N%N8Wr~oOg)c#L9K*^f(${qK$wgoM3CJ%zf9dT}G?(ou4?G~03?juG3ZXyhj&M5$fLfkMED^$h#6zS;N{-+W2oT;_oB)k`GvA~kRo*x@Ofif& zYiWK$tCkqPN?JKST5KY&~(3@yL}{uN5s`Vbk?qkv*Wwui$90N|CrhG zl(4!^T$~cRdq|p)tW=v(w$~}clT5pteDvpUaQX4y;$J`ZEZypPy8Zy24G^A3b%zhP zaApl@6IQq_srhMC&j@|S%PppLMrRw+;eG1(6nVH!&#NeoAq*E!lee#+_R_b!KL1ha zzxanU9*jRu2v5_k9z)|15(^@R)ef%%WjN*NrLXZ*|MRE#Cx7FWw>A z9GP!e_WzE&xlNgF;`tDyqq<^N5kDowM}&h5>iLJMcEq$@ad_vm-;~K$h8e@d=c&`@ znYN!s?L!3pLF#l!oIOO-h?^r8u`;J*#mWOxe4QF9W#-79qZ$uzHo*(Q7!)xVeGZ0!@YPV2uVEhjR*8-J+f1`6k6O=5s={ zmR74ytqW{AW@4mUoFiciG9c1}9iVC08-E=(&7m5WA|YgApGk`?c{s+V6`>!glY*U) z`bRLI@XdAd5J~YKVZOo386se|gK2?P0WTP(6?<)Gm}@gVBp^X?ZD@g3bXE7IP>QM( zG*jO&Ekj2w2WWSWnG-@AgSM)89ufM+ajlMsw*tA&soD0MU%?dZGRV}*0zF`=tq2Ry zG$CO^7%)?Wgr*&$m(&`tb=St#ZNR38DM+m#55Q;bwbUL#k)WuQmW0=es$r&Bsm+)l z1doC&A0T;!`wFuiJ}wBv*3;xrvl6B~Q#5%SHP4E>w{dH9k3Oh`Fd|`S*azuQ(-x_e z5xm1{YDHu<&|bU{7Y#RoTQkOHa@ z67)L8d~a|%+Bqt%B#r>#jWJsL2GdOTY1gXMQIOEiK5nwdY#4PnCl3)}i$p+$5Dks8 zCmp=O_qB!%sou~rS*W##7?7?uQ#K@m3!&Q=)Om-q0sm7Ds`HuD@RD)w za1Q8}O1C;C&37PrkyV!W;n?!e5_WMj|Nam<`Sy6``?F*6&&oEtd`=UDs4`^4L zv0eg5#%l$YW+W+%7u!P}17={;J_TRf2Pz06j*GV9yfI*9iQ5ME4iTZ7U6QxAsZ+x0 z0bZ3lMAGaMxS(urqM8Z)elc0vF<+)b%CJJ@4Am8Ih{~FA`vhfNAbs11d{Qh*Go4LQ zaR@ER?*#HTP(~pp-xvJ}mK*CDSs)S#aZ5V5N1R_UK71a_eZwbkjXHElJVWDzmavQe~ zusdBcTfW4wc^=hg2(yVae;EVHqi1j`-Qhh#woaCY_gcb#yP|HEu-q`-e}+1al+Bm<@ZGQQ%ir<} z{oU6`{_Y`X_U|*l_CZ2-NO$rI-Qwl9ASTLs$+*2u*?ySm@`#W9z2D(i{+rLe`R~8} z@xn)c?l&1PBiIM<`Xi*(YsBRvLd-x*vD$XUba|UR&dFz=qsnjdPyXgtzw7V$JJI?> zchN82C5}%r?G|+Ju~5#cFs`2<53dvIcQ|!;r7Kx(w@*=q1wJgOyFv+G{1P(!7Qg)BH#a*! z_KOF6@S{IL=srxh`a$~TpC!%y0^Qgl56 z@N3_m(fh%lyhEHnOTWBC`)|Mnh!grXX>pgB?xEoxs`m)9uaQx9{|D>owf%VfHGJwQQ%lL!v0&2~_+4Un zNU@&0-B8azgYl5Gc#7r8-M5auz9Cao-q{wF;FQe`%5=$?KMhHNLoELYb-0DsrzzVX z$I1r?-F0GbE!8V!&#YXdVL;<+NVt#2Ovx3mJ@|8Im(a9k+8pA3%Ixq}v>Va5rA~sE z8Ff5v=3+S{=%ZGUGYyonvO9a0x^aewpT+BM@T-6A>|H;*|NQS9V~d}rJNOVGy$({T z;~|cc*EhgU2(#KoddD6pV1`rE#M(~@&F5Yx~eCjqZVh;~3C z?~d>~Z-%8y+g`IRX)%KItq@5RO@>I1)de*#kZvUP8)`YA7DeKQI^7_|$n54zI6W$c zcK4~RC(V=FI_aD`f8$hHk_? zv=nfO)Etl)kN{;Tn1yDJdPNx97&aLVg*x@9Hd9uE5#oeX5OeA5a7I*0mUnMQiAx_vhqh<*466qcxAz{;XV7@PC>ss=+ zG+YB-$-9V^R=C!R;9@i9u)QH<2(^7?%QG}hlwr0<7sP#eTPxUVC0|NwH*;d#2Vsk} zw5SN;dx$KB&<&u58-y5uh}Vc{p-QtAMcR)|Sh?9;l< zfcH0ey!#&qRAq0zHD^?icFfUE1{xrAPFy$|0uq|}7=j`ygqgK%RD?3_?eA{aFcDp# zTS4fhJ&`vCu1L2}0XiBYp=-zluiDB1&D)|yQDPFT8k&qaAKSJUsFPx*xJgSK$9;lV ze2+4yAQP4iD_d&Wp?Z(BxJ1GX$avl1(|~h9TAUKQLYPHDY)F>w_zVdZ2^Eb6?MCnw zHZ0yO3w*;ox0=w~s?*nz z5M@psZ?*KlDslb@w-HSb5CpHsE!k^I83UpJ8rmsv%y{uMsXrognY6kG;*1wJDVq*) zqd#ufOlbSJ{(tPfS*$GCb*A^NJ;aciXS(N3b?>c_#o{QbEj1*g8*Wf`gN9lnoBcpP z+K{_j@&Ndu;eIba8|VkYb~h=}g!SMD5r!;5f*`d4(h}UVpaB|8N+fF{tLo0vIVXpR z*n6+#hn-coE^DNkqDbldKmo*I=FP~+$jFF)|LgnP6D;r6MUp%|gmS9KqA(qO0C);V zBJTVNarX8moNE$sb_X4wfOHmXF48Ur&`ZFF@LfmK->AO{t}adpb;1CP;5v{NaRD8B zN?Fq^Zc%bWnH;7|#`~9;)-c}rD)75}_P_t$D%bk-FE!-+_h|QjkhrJpJ@Petv#!U` z8|0cqGrxxm17c5;bHfM!@iAZhl?T&fU;DMh5B>0~Y`P4oP{sq&<}uRIBaFuv7*8%U z9(7FXdz6!};?f`REC1{p-|e^jm)1-FoAUIh6w7-^d6GE4Qx}BaMVmWU$ZOAZ@)!K- zfAweZKYSN9_Ee`j)%TiC6TGK7)v5j#NsM3NEBxgDa*dFlVd^dt!XD1GwQ;-LClB8M zyUNDC#{cyW>Do{J`~~oTmwxYkxMl(EhM*B?71w+n$3U5OD4Qoqn;pihyDZLson~=F zvp8b7`v{JYGKliY&;2&P`cJ<<$}qlFpZfehH>V%PFCHg07YW@G-z+h=K{QaNv*`3V zdA!1O{|f))fBO9`clSL?MqsK&C2zftBo6pkZvsn>lxd%|+(R^y(iODa!tT9{u7%vq zaq%%kVjVG0BdOe{tY1L=*Z8%ceQq03`%^yox!d^74>FYx67qR`oRMwBg@iL0hD2Vy z0_i2DNB@|3^~b<@R0^J&1|QaE(ajl3S>U9p=8O!uc#LoE;QI&1m|4aou<()PBY(m4 z%8!DcriH@J`qokyg{q{9dxnGi>;8_~Q&NqaD5thq~U z21+@=L)M50Mr|S#|_y%B&7D1A)%+#@9QVlWDNSYwF8zhV{c`CQTup&5TNM!r^6Fd*uYa2wxti0t8`qvQ={ zhm<^nk}!#gw4}+9foJlpJx5;GzPReQTC+R?Su|hPZ?ad0R{su;_@LGL=k8hIWb;Fa}i55xZU5y{nY5 zuSQs?H`xwjnmm(uX0vKgXsJbOubI7SAUI6eq9wE%cL~f1QUqxrc1oCS(CjHW*77Ua zp4+4%5~QsRS(zM^2DC#}AXT$l6vTWjrNiO8)HJ!wtMP9Q6rqd}k&O6)=Dsejr~_$n zVU2SMvw_e}APe%az*2)nd42r9j?Pz8QK$xA+60>Z7~hT+=V6MDdzf9_vZ)ssQyq<; z9g&vZTNfl%j~_puumWnH@wls{dD-Ei128 z!1*<$>>#D7Wq2;(TeoHVy6qxTJr->e@)U5i7zGz5u*#%0x4`)gBFAL2En|C0+&!U8 z1xg@qgtCD!9@g)5bAoV!wD&M=Astet9xIMK^teW9f?X|F| zE#aIz?P4H_M_A$dOrBaSH|P}VRERoqcJ*h|rJ6Qwj_G!8zW%X}%Uv|>;$lnKIijp* zWNi=%F5bcPK6wx{pU1@w_!XuXDft{ltHW=_>OzE5Vt*4Y3d|8pNW2H7C9lt5ZbW24 z!a%dQLppv6m2;>KG&_F+Kw2JVEjbvlH-#eCqShGcK>-PR`@X9Bn;pMu_+D`8x6N;KI? z;P?3Svkl|r14ila#RxiZ)!pL{e>T12^FQ_ZeUkhb^PP{-?A>MO%u5{I`a$sLh}|uG zSCa-|wni6`@%|-bjI3@L^UGhwVRUD%)nXTYKk42*;NSh#zwz&|>VEhK%nm+4oKMWo zeEBVJ+RGT(oLs@hA7r}M@#$aqZNBi&-Wyje+Ff!wW*QEN+_T>dI)c!VJP zwe4OM$t$wEMccm2FaMX-u>Qo)Pe4Wq=%y#Fp2Tq${4PFp)N-zxD%7&3tnS0+7R}BJ z_&6dG5X;POWv2Ke_=Ur{5zQ}P(=+()JU*Ua<&c~L;)IeK1Wz89SXrX9_~}W+C&V3~ z_6X%@hHJ0mf~VP8;hW93?mT?I#%zZ&yhNH-_;8Jqb}8#CXh1niMc!Hn&UteM+ zU*StX{q`x|Pkd$u{=@k8fVlGZUlp(Ssbuv$h;3`-6*>bU$E8GRsXs-yo>OQT#!^|P5k$4K7Hv8-rWplJcw z6X#mHvZ#e*1u6qnjXVlos0qpn60fQ6Vn}gVCs^91q@5ss#6=~o8>V}%@?u8_eS^`E z*1P0s2k#@X6(raL{^+hQ`R`6g7u&eVvpkfsKTLqaXba?VkN zZHiiJKo&4j%HTkhrZsXFu$CB=oQ*s<;j3^#cMW}|kpG`IODTUCN%3`h? zi^zyssp)Pnuj4$-={Bj}0QIx9tk4v2Rtw&&KCNs5W$=()T}YxL9@2UT(Tv#!q~N0x z+gF(#K2KVnVO$+zt0j5ZATHq4Bce!AEVq?wTsD<`8#}yn5JsFdgcf9BXmaGyKpm;b z>)TY(3=kHDQb4@JJ4X?#C=*>GY!H{W=3q@diz}EjY?v_~cZjcv-n>fqR&YLJrLX(O z6edTRW{81=6TGGc3_sjO2~k8XCr&uDgz>U4%>P)Er- z_;^IO^JUVsgO)wiI>-fOLZ-vY>W;6}M75uRcSzVEr3ZDGi}>zlEf`E^$m>g}T&3Gv zRpT`d_~s1GU84*=WwS>b0+wE;ncb(@`^c%mw+~NJi7zzm5oy?C+8olx2Q1iWJd3z9 zI9mbreDz)w+YQc780R5nT--<8eO_}rO?yf@gUAe_sf%lVqAqgx)e!gl6ggnL{{hUd zQ8u?pCx1xxzsKU>C4TLnfA5)@pZ=wu>EwNcc%IqepS|bj#lGj%JP`CTM# zkZd54X|+cg&yt3F4Wyl|(TU;dWzVeQxV!ll*+85c#Mc^5RTLVAwb6 zCADu^j#hGHJboIr%ZNWDG%H*@FMNrb2N<+H;$M=id z|KMX^41DbW*0;QY(7eRbn(5?moE_l8C45kPcekPiUV}5JqvQo?xPYZShNBtX{5ebm zrJQi@;#2tfGqke<^R@Z-sH2eU8NEt35w3w5cqSFZ%j|ncJ*(yH5t!hHZ?qi%q z%XLEc|KLKv@&S2uk)#3f8MB&xR9z$bBB8%dvwxNG&PPf43ch;+*Pq9^XSVNhMDsoT z>KdQ;`QPK@>EGl_KmO*zLa*m>;Sld1jMxH5GZA+0;Fc$d+b6WEYDR5kxdjP`_uvL( za@DZN4i`?a?9jYZ%_TFSCCo;qlUe1n`W9&bM@?x1G_{;cr^qxJzN^Jo?hUUiqLuYY zK-?114vh)Y8B{Y)?l7%(2wjiqE}=8R{0JWl)4hGlw7?_~TQFBhM*&S8p=~Np)~eA_ zlcXJ)6Y3oNA0acRLgX-`)qG(4nJDO)YL@KrNud8S)hIsRvg-FEyj&o-$Ak{4ip< z$5{*F2#Z4KUBxq4#1e=a=C%pQY-G!zsg{k2!zhTW_=mCyq%q>0P^PK!lSfY;ASqA+ zJ~|}CZ4&w-&WBnIh!O)JLKh0-I$#;ny86pPEx!_>q9(Mas9k_GO5Ii|xF%O7ai}Rt z(^)mSx7Y^MZPV0XIaIUKDXw#rP5ppf5M*iSRLJ5hYnW}yc6+K3lrn~z8rA|X;hH6? zhFGE$gQ=oEKr1iTK&`EqV?54f%rdG4H-Sp13uH=7iFxq7C(h~{)f_8I0FUcBN}6Lt zBpg=;ZfYMoS4asACx?}LEG^=D#BEvD#bZ>1)rm(!UAV9u@QXt172kNKrN^{>uPgGw zp{bSxiqxb%8p)f0Z{f89FEgp<3Rh#$-Y3Kvmg5$Wkx9o5C3%#}>y@BeBuI}*#y17= zp1cWIa+u4sXI{a_C1coy{7vGS=g{>IMLVQKikv~bVfGl}_P`oKKjE4+ngc2=L25Z* z=%GyDveZI8lkK*Hn&2D5v~HVO!eoPR4Y2w_RHxDVk7vyWg1K5Q_AQf7!Mn$9A! z1AfH&`wy-e0PzDs?xE8_9(ED)Pz>RZQGSEp_}Oot{r{Ri_QgQ+a*IEk`Hi2LzT@f7 z;pif9en?o{ewb7I9@Lz<#a9s8SMA@$$xzt9d4uD1*VgI ze7IN_wJc%zm@@u3Wqgrt_2>M`zu3I{KYXfFo$7D0P7U5uo$6HYA`?FG*+)_LI6j=i zhaT{xm{>f%R3{#wS&KK1!0 zxS2jgyLSoSodi$BaCpoTE*ueNuh8zkiY4Lr+CKz&AE7@_jDl-lrRkqXouT85 zJRIV>AEgXem_PR1e}mDgr3IQg4r_oj zu-bU<+YxDwkxi_)25TvkVPy^$@jg`Ub*#@5l*v1SgJw@Clc8h7bOI7H?e2pcO+mmX zgj%!|x)U6`Xz4J$NuDC1@9P-zw8nc+XqT0NIX@z;>xaC#+he8M8q^L6HQ8sTSf<8V zbB`%Z#{m}_;w)^N+6$_N(~LALBqS^kC>5Jv1(Z@v-dN47&<~_dBySo_1t*TYnPJ)? zUP!6Kl95XRYiSmS<_>&|a~q&?I4i?*%Ud;v6~!!7^Q5$t@c=sl2~#!AZH133#5JVj z%cSuDk$@FJT!(C#kya`S!#jtc8DTb3)(b>3-TW1FbBk$xp1ghxi5ZcxCPlMDOF_IK zc?UC30lK{teCyG&ZJJkyIjtk63mM6fQ>qvlO9(5> z&u~p;4;#K73${R0kGO)M_467#HYev1S>PIvM5AFv8Y>3CoI@JHd4b$)KW9)39Uba* zYv9VZj1DLQ#n2outr(OrOH9WdmYZ5&B>_ZmK@k(IRDN<9{Wf7Ml{qU8@s)8b-Uv;? zdBwzIIbyO~Q^tcsr;b6_l-oTL?-Tp0_%4xFXVLKn-Pxg$4z2&D)Phi?_v0|GHSh9ke_Ay=c_lEdN#v4rCu=mu=UaMfhYJGd}I!a#6C zX=~z}VX5+9mB6?!#1JZhj*%y!B;?eJ{Pyef)Veu5m1>~iD z#3KHB;gg@;z~^p&zRa)w^KUNz@wUh1OMU!2qa z{NE+a9w&B>AZ@_=bscNYv2ykFgz@-Ee)8v^WA*Bvd|xMe--|R}rs*#bcfa}i--<2p zu0y_c^!*KzX2Q+ZsP3WbJ%$oUtM}2&E}-pM#C?K~ICu)z9Mdjd#?R|T+Z@ixRm00M{y@bhQpwFO3OW=3@y?51XI=ObA#iMgVImC1AO=<3L zm0-lhO{`pIJb6E9d5z#F!r}tHzm1f@c>5oMdj{XO#Q6<;7%+LTf5jZ7w779b$}>cM zkn#97?ZqzF!7I@dGsTK)3%QI?ch~ z!S&0U3~g65ySM1Zmd)J{5_#zG?Ff?5*VBzPok-&ets!j&5ljr{5@~R?h{_8B#8yZq z)kK>FR~$sPqtb$NxteIjgDZ$jU}qrh;+uOQM@VrvH{jxc^Vd=HI6o)M3u(E7m59VZ zXcaS$s^DzI#SsYsXpmkgQ^eMNO(>Q`$sU^=mNV99#USKZ4Uyz3k2YkykJVr@hoFSn zNa#m~TMLFE5F}z#SIe0qxR`Lx5q(0*wDnHXT6G)(JD(u2?BpnChP9n4k-zLh? zR<3QA2;B;iWi9El4y>u;?g$WSL0wZz%B*!r7Q}^Wf~D;@SRq71ML@PG-fe%eCaETA z0!{V%-knK^^OWpy0lZgS>nNFRsyvYgzXC);GTr|5#PE&tLj-xwLU zB~Jq)$B0XWu7FLX@hqAS(K6yg#)UodaE8#{0!`%2E_s}R?BgSZzAogX74@9v0->#( zS#e4kr52cN>LQ30BdsH8beI~>3p&)t)SW(o*6lA-0135aejO{TR%QG2eMwd%{b33e+V zJw|IlRQFGZA%+RZ@aW+}JGgt52v%ja02 zq`$zNLu6Pm5M+aHlxD9!s1D@ZMh@|;vajmvO`Xv>HM2b#g;HjSSYS^TWVVfUi~Ngk z;Yb3E@HmePNXavgz+GXcvmUuw4dt562fFa$U%eqJlpu@fJ!;U)W_$F?8v(a>sD-m^%STnpp&58;_HVFI zonVlo(DbM%G&%V@o*Cw(`ZYN<)8saAO;~eW9AE6hNJEA@^vrvCiR-xst{5hSEv>D) zYS~z@U3TNMQhE^EK~|d ziuPQ43`i_s_(puYD6g>n5crg5=Th+8|Ma>14pRiIkoGRTwF%YBJbmi>*Yuusp~OLD zf^2MNcaC}MIx!lO zaS|nQwYQDtWLVCyTAckny+~#X{#!|Tdcs{dB&GUU+s$P*=!)X+s^a9eCBoqKZ14sl z5ncD}qq&#Y`%ZAn;M5w37!vq20>T&?cma(OeQtAKq85O7cHdm@O`LAfm+#fyXC?%@ z?cPt*NN7rJSHGO-RpnPew9M%Vb?8(;CJbu`=$Qi?cr60z%G)9S7X%|?60Pz@6{OEOf) z=|{Sw_ZYNYK_#tZQel}XdfIMT7aA}7Byb~>G^RySwugkmR0 z;|BVC83UCRT1ccFyVb!;eqC|pUy=#W#R-kt)Co~m)#S>-$A!WW+(t;Hw16rNh*ISy z#yH{aD50p4UQ)d!s)IN%ZOjgI2~451!NNV?+KE4U6#Uu}*iL;95YbhG4Y7!MAYq!F zI`4`rLcxVQ#1yW$>heEV?H9T}#-N7FZ;G+wHeX&!1)@j}rpYD)a_B*=a649%cle|< zik3!;`a5+V|BP>Gbo0K`(NEq+*XjFVm+jj4uMO5&%F!YYMp)*$OQiRNL1`hLd^-V; zyT=wjTf0*#T3xoGd6>J1VP6*(x_U#G32UZ#tXh>O{Azex*UEHoJ}ZUb7u=*~m~(dg z0+9OJ3~iuCi!Gm97!jeW-qniMELmHdk5n5j@vD?3AJ>c3p{BV}mEDk=qP_3TgZlwV z4nj%bu_5(roCWN4SeTmARMO$lWSiBfE;Gxt^XeCYPYNj=gAUYX3^0kY;^74z-qQVG zZ(n0wU6DxDTh2%g*0hpiU?s}I7Kg*Ive4tINU*>ZEnrTEu9Ew=l@4h#LNg`)mdT}q zt2#@*j?P~~c|r4y_v+p&U;Kvv6%Hf}X&?woEyp2>oK|KA=NhdFV@S@@NEmND=~?A# zs2zIJpoKpQ;xMPgjGwC(sW=iJ(qCkAF`~y}=ayD-^1p=vQE{dfWF#||f&ymF9XdkM z0e`w()C|LSO9prbik;nl{GG?Oe<$)W<)uv?XVpl*ku7e$v?QYKjro;18I_=C8)o_; z+ymUB*IHGGNz~R@Zx++=PJdFYbxoQy9d?TxLS`+Os%?T;h@THwLLN<&eT2<<;~zWy zO1toRkwt}_tA(=`E+RKz>by;nfFuWrql`~9zwb4;b(NaB3#eLzrHJ;Y@gMJmiwBJ2 z{gqP0OsI~e)&%Qn+x?`o|3=fF$V&CX;JY)Klruf=Hib@3KNCd~R+o_>W_M?Jm#p_y z#5$$z$749)HSaW|IjVkFNQjc`#J4er7Q}lzi?!?~+9LdRq2Q?8rmh1W(?^%#G!i`3 zn-!@-TzCQGr+3WO;5ZL6K-Oze#~(Wp#h-$5r*<LbhwvZLo4o;DIqAnOo*;B_q zPxOEYzI7HW3=e*VYHA0s2;P<}w7gMpSYKbIgTtA`;jPd6=C;!2i7&R@g|qkZ%}cU3 z$o2q;bMP5;AW^paJb7@75F`+2F4r0>V_ z8i6cwld`cj_m|SiD_-ztU|tN_HCDSb(#}pX&8TwK6+s|%KC!a?6Kb)t0!7_t{Wr5= zHN|PG{1C4BjLD{e@gU0Wq{Fuei*<^*bZ*4yV>Vc{ecfg}b_Y zES0hD<{|to6nF7S!EuQ;eQ?K88RD3v6ujb#`&diBV>pw|fDnvmT$`GJyJizps zGr19>=8@%IW!w?(ayM``I@_fBZI3tJJB$w%%=4M|VEUOla63!cJ@+U?U9Z$@XoK$D2 zlrMY9-Z03-e>!g)qJuaNjzyf=Ur;RShsq=tBkpH2-9mgK6HWL`xHOtzFguej=l6Vx zpYY7Qijd=Xy-k#rL4u>KQH{KYUB)=*Ct%8BuD(dzpPb(Wt|tS^fo*DqeD0TzE>xUW z;V9-yID-(OGAVgGd8*MvGm$i$fvo;m*TY|G)cC1EHCISY{@HG7UPOoYk(^Gual^v4 z-b!u!lrHjz7!~-k|Jrew#xcyx1o!Tff!?8qot$2v#%GP!_s8d^Zwn!K;0Ei+VD@p~ zmhkr1j>b>JKJ&R;newo~T&dwzuTKo84lk7fL0Ulw7vzOLn8=qIz$wJg>=6g*P`b2%)I{feN*yz>Uh8Zp}TRSgCk>wazT>37H$*u z=PYA)WfL%L+U4ey5>MLm5I0$Qbpx0o0XVkm^eHp=Qa?_hc=g}V%ikE1e56RhFZgD>X96mzt9CFi%;x$z;mX-F7!^o*suaPn^bSk!bzNbd; zXdCz<5hKZvllp!+n84g$7#hJ(H2rcon>&Ld9V6Knk`qmHfMAX#uf)4d#H2Nt_DM_| zAQ@4K9mqthq`2BjD?2)bthnKwY zSUIan!5;EQX7XI#us|OT8At3ww${uq$Z8N`nArh#Cu4$;)eEVnN+>{luk1)qbM{_} zG+>M%=OGrETX0JZC64_$JQHAyaQKTm?=OBnxY3hhFgyiitbI-*v$;jjm(h;!(KIl7ZIy?gQL2;J@;jQ9ai?D8zw3*09| zUdIJIYW-IT!p$!6TKtIzC&uZ|0BY?5hcybsE}#<&%zkxTaK_w*jf=mrCe*n1untmX{cJ-Y(CX%c&_3{lnMO)M>wON*BiD@9S zFa$)vh>W0K^5pGR|V;P(IEKS{vp~FUg-q%ipB^utQKGHk!jpD_A3+ml3Ev zC&jH3;)@Gi(4CwbRWJ3Rp*Qk}Gm+IdYPE)`8WU}vGSGE;IJL?k=Bg4Y^40ABqTkD; zt^xANn*&o+gJlW2g2b&k43#sm2JhMye%ti5xM{I2%$$|sIl=M#q+rU*UHOn0#{046 zp30e??b7BWGt7s_ynR@`I_n@EWPv8m?WKXsQN@(U$8Q~27ss?M_}#~5G}L_*2V~(N zf$ms>Q;Qu!`*#4JhLUY|Cttr7z7W~MI2qY_OI~OA3R0v+Y*lwY=)mXC3GC-cV7beD zo;`zot&^-ww8-?6H)r+axTzB((Kb6CxJ8p{TC87W(Gg~?20gY7KU2w0gj**#>~>AZ ztn7xcPcZCOo!WZNEJc}18!AA?DAn^bpvp6F&mc^F}*$1o(zYaqtNkb&HJJknD+QtQl~7z@X3Q;W68o zTQzw>%qS9Cz8@*mA|*#|Icy`Paj9*#&(Ca4>8ewSi)=*ih8+{vXOi9VLj_MwO>wex zF8=M801+>@ZeeCMA*>b(G``blp3#bc{{*7WoC#hv?4J`B;NA4#E*;KrJ`39t>A|&M zPZ|rqM>v}9z)ijK`nYD1afs_dP@aUFI%cW7cPs3+_7+eM?{5-UB5%mM^Hna$Kg}B) zIbZz85Yc!;p8B69V(5Ju;eS2-V%~z-`=_$?1+&3{922K9cE*JJ$YR}!{pLk3A_B@n zD6Y)?@Pe5n&qOH&$8JcEI0jWVO$Lp5TOz^^c0XjJIJoy?V!h>UI8pDwyMt{%-KxBF zdBIVnsKwo?hJIR|N+2f4@B<+YC})4P?tb*?oz&DQov zRnp?`B(!!7;ys-%Mc}5rZj{mMkS`ChCMicl=R|oepD4IUQG61+oACwzsTknO_z$b?T)IA zfs2X3v5o$ZA=^sn5=hak5n<$vx(H=p%*|%tgS5rm)FI`2xtDZ5$4-jA9bkKbuZdFZ z9@c_69j{}zwZ~0>Lx#>sQ9AZZDJg%8_}B^aM%)+ca*X`ODMWZJXBmBFkA%%R^gd7+ z^X0RKyqC-3F3`&v^LNmKi4~avgGnUr(Ot!17Wr;I;?$g9mc#^60!sw~v}UgbR_E@M zuxlUbglQ4!Zhd#+L}*l_$xDP#5&~}Jw*~Yk4t)<2grfpeVAw+KA;fP}P|0ZWgEcmonXzS(>l~UI$afZ2u$8pYx~T4( zHp(A_&Nft4p;KYyW2&o9wrBWKX)q@#YWefqhW=w}=6K{o4Sa_&L1-)wMGTebm(Ly* zBB!sh?&qxom6=)4*gC7q)aOccF?;Nf1iFM5`UPI<8|iRz*K!dMI2h~xr9Y7+s>3FO zrKHNUrG&O?F*&f`TZ{%7TYKMbVbh z3sY|60)5|-h1ABh89U_e{g#TVpFtHbOx?$Xs!$ct!706L^{>z~Pu_K&f6>+{&N>9S zW{{J1|8dX2r~1oLhBTJeO87ax-pKp;aSrDEq|6!=Xt;Y z{rNu83%3)18=?JLg~S)Li-d-+v@0Lhk%Dl^2^H zNsYz5h689H!l?^9SItr{cG))p{BI$RcoZ+aeb~V2ejCdH-*ftgm_4Dg-;uO|=WMyU z`84uv{)!5@!hUVRzFnDdXX#g`^jS}RJWm`d8(ySGT)yeHAK25^L#XMwbDK^)>$#&@ zrmJ4*SudvxkK1M|-YW)gtKR!h6Y(e_iia`qZ}h1Zlfsl5xw}I9XaBUDRJ^ZZYsMF^ z{c1=;|B(uY3FrxBkE)20`h1YK&(~Ktms{x(wk|Z!*Xz$mRj&6xM2Gdjl)U_c{n_Cv z-}H(`zxIT=lYc{UH*fbT{FjQDjO7r>;!hLiwD8?Fj_*&{i`!wgY7xnuKG!~n zFVNEApZGBEgNJ9$n)rdYizo8jaZk`Q2*by%kZ^L8867m}^e*>#`E*%#hE!;~V8UZQ z9`v?O-|pJ);8;!v0WZid;~1!lWsr^5pr-?1*vrhU5tP9H>TLy`RwBT-xfBkYg4& z>z3y%A4<6GpLomBs%g|L`4TzrZFwH2Cl`ln#&4ImX8Y@0Rm{lvdY`_Y?YcLb7=L;2 z9nBs^%v9V1ZZ6Kf#i*UX&_I~RnS-l5&PhxWlLAT;hi)tu!oqmpO`)}jOO5?p1{4Lp zW*{ta95@6-Nizk=7IxgAW!+bM5KaXoc~=Sj%s?|e96!?3EnDp=?{O4=0nDP|Zg&r9 zCGhvOE}97kT2^*H;>QZSNLK%Wt0eaolL zmAD90z8W20wd&J$ie-V)Cv52axx9gWskd7nIej1UH6OxqT|yEb(N1m&gAS#dM-us; zG!vjYkzZ+qDHYMSK0yH(29#wQXw>2y9m{qkcRB!_`Msd)V;5chsJB8geMqFbBMyMD z*sowj0)%G)++7!&r_pelsDHG?q)41bwD^jEz1UqS>0jVFF6ji55~`nC>M2 zQV<|GvkQVT1RQ%tW4|uWDgx&cV{K(2WqRR{$I_YNYAg_8D54ARg%>OA%MuqC8*3Eq znYlsLXJ+o0hYw%HEiUA25V3%2T|x9RUXUYDCN@ql^-PhRNu=rwMVM-W0$dENId4cg z*(5rEd_*JSNgN9C!QEGMZZiXk>yE-vg9GW0@A|tRSS+q#5xAfyVrzut3(Sy6_eY_LSFhLFsirM zrd(p=EOn|JTa#}c6Xt`v&9SO);(d+Ma~*$m%)6pfBle5o5a zNnDI-{5Je-tepbyDG&Y#RcDU%5(B$o{w`c0(Fo4^wZ6K|)2#b?yYZ&=+Vpp{1LXn! zDB0%&+SA!jX5WW5O$jWF$v9(jpud|-%lkz(OGEbcB}jp(T znI~dNyU&{Lp%vrR5oWfz{+2%O#!QU{o-2>PXs85vQz&Rn6euiJsZ|b$zdIL#^RqT5 zeDA>@+QX;eNK|KGN2ulkB=e7Z-xW+j=LO>E$oNGPr!70UUgvZ-2rKE=2PEyAq#uUd z4a(a7ceSLn!3&3plormzt}kT|70_n99;1Uo$8K3As|Q*i`_LjEAxp8Q@4owpl=Zi* zx|sKwv$CE{OfPNOSc#r=9A9I8wS+$B2tUE5 zAl?oh)1~r8Uh5!>ul9m!jBsbr{yRRUEbF6-hMtmZHe-o&~?VV zKufhmL%|?7JY1uwK@u|_vwLw}ffVg5Zm%I~D8}=r&Rk?eHfA(G}5|F3pv z9xAF>k~9?ikw%$ST5*qPt6UG!ZG4L-rdCG4h@^IATr-!3UfeS>&^bCPdLi&BfFEZP zJ&JM=TNB0JE=YZpA#xEfa~hqZ@w+8G2&lD-2cfb^R3eq5S=!_9fpAt~?>K6Ee$OUF z)t+UmG{0DlV!N*BP(}`yZzxa)+Zx&UCzvEk_1{(>AR|YtSgOQWw7l9~ig=`%VbXDX zSZ67@uKJdrZy7ug!f9O4W7WHpTG#lySzE2hzs@6~8C!6}T&~f{e?qzb~Y~4Wyb4o0C z()3wcA40Q)N*p2v;zy<66f3IIitAu%nS~m#8m#pdrZvhd6b}S^hcd{!!T?YbkQ&|K zQ-9#$Ft9rZ&c+3YyR54Z8HGkSPH5ZqDJ3Sjp9C^v28(BC#pZmo5>RlW`V)~Y2A_l7 zJ7@9WgN@VXrZO9gYQPWDm4Q17mxKuHA%h2BdS<&x$2 z!*-KU(Z&Bhv~OAD+oy zohehbSVdWm<%}KYPCZ@rpLGe4rFA>){7#hozmvgHW)pr%ne3#-T--Zt2=3`u&)+$P zAwLjimO8q&$NFD1`y2~L4`OksnnqgNmd5S8byb;mI`5x~Cf~WHjS2;xVFPvwjGH|^ zpDFBK=EL9b>3SDfcG)Ep7xmtD$an8Le|mHG++ ziRqUIljvVFoa}jgXhf zLJlL|JZqGaPe_(P{+>G`@o5IQB(jK9rW?}Fcg7bW7lPJsk?`b$=G?>ABp?Zvd0iE8 z>$=~jJ&M`*)p!;AzS;NSm2Ph0exhqVO`LegymKT8Agmzyfd8j!;90VI_ArdL|JPFy z7o4KvoA$M19QR`xZsYvF!u;&2x&V9epNTn%dN~IJ@P?YyK4&X!#9{xQe^mL3%0qwS zeSA++nibH8UxB}$7@W%O`DuV9WWH~b-!#Gzak}I>hep}#sF!a0{w4>E8f#XldHCOU!-PiBw)O1cM4Qo~Oar+F(b^q$`q%#K}Xt%?{6f4~!H zzTSQ`PKfLaOc89HT|jQ!a3wtxF=$0%gfH{m>tl<3qV@7%?Z1h-$E~7w7`y4Lbfd194x_Dl8aLSb z<5ymxKcxQza6%29kPM0{9)OODvDKXkMS&V#2j)#o_DekyD#uwGmFI>}#x{zLt^u0a zr_$cYVKBqdQeN04!Z{WS;0JCe-Qhi%K~Wt+`?C z!veYT5p=BzgRP9W3S9>dblFvvCfidOb_yuI_U2?GXNJ?jvClDwS@gFC5c_$D)_rg< zanBybI=qu}g6^5;{G4=WIW%#jmMjG{Puv23z>cHwLEKM~tBoL>;*?rZ=~L8%1b$+8 zY9g8%1_U;nQm`EfXJQy^knAAB@F%LTekfSSu)$!3CHQc$X9zfRGvX1@Tm}ZpBEe{V zf(EbSeHzK`4?s#2t?kI<=oqt2_-01ywxmk-7BGLKx3MjQ(Y})8XiJzO$;?+l5=t47 zp|s@IAm?L8;S*;^3+S_-sxR1#3iLmWtaw@WN5T455+TT$C+Ujw($e{N7UG{pZf;Pi zyd*Pq*%_ZsNx|4io5*h-M)N>?+c3nb1FAseQ&)dRrLtOBaxQ9&{h zEI=yV6fKNbQ~kvQ%WKD+bvqZW1z!V#Vjn#VOA&q9kxPT~p;0-IRl_+3Rwrmv#mLWqht;aQBQ-_H7(_GW!bvXJUS_4E!6GDG&}w3ymY-I6@*jmBvoxlmfs+{6_( zwpA8%kfi}FN{3wd!B71+{6Vds2^*#b9BcjLu$j)79nWB0nXU<+tqHM5PUfs!UGIoG zezTWtXHKDP-AE^y3H!j$%5E@E$rC*ZLM{)wO`#7__Tv{k_f|0;I(^9chVO}JrttMl z!`cfNRkbvV2 z$?IG#8@iu9O|JDt+ybQ-rO7+mSL`^*hFp;DCb2CHIBxQ8G|4v=EhhAcFLj(giG!Pj z(kJbF6GTR=>2X}Q(2PLU0P{INM4k~WBX}BE_0^adtqYp+vagpvBU>1KyeH}MRkn5Q z&mKMKxNoCU#^E2roBZ;=m(!G~oIcBV_uoIin7h8DUvZbqNO$*E-V5jdpr2kx+`a}a z+JsfdI}X(G_0GE%r-hB}f6y%W_K7?N{Ux<*`OkQ)cd{Y&f4)Vcr_E3ofls4>DpYdY zL)U>yxbRI-jPV`qVT^F3H2gDOFeT1E4Kj`E08aiJyyWoT9;+eH_E5R_aSjOE0}Q<=w8^ZS3pjo(i8dty3$J+55NCgTz6W zk00CXAybe1TpJK=GCIQ2+m;K@K8hKrC3}&JLgL-7G$PiHfrT>pL7Pdan=SaQ%&!F> z_YhwnfK3peU0A)b6L$9AVuvKxB5z3-wS{Khq1IXK27-^Gu$2k2dK-mh9f40PZa-DQ zs6%ftYW+gp_DlUExRm3eNs;#Z8A7L#_xHKuy(6V=B$Iz>Vj5FhPKplyC{_B z7?y~fwwDBI1{l+c%|EuNbMz#z>#lB1K{~1Mw^GG3f@^yvVS>L01{QzjKW z-*bNbtzS^5T&@S1WBouYgk-11!p~oU{ z_7c*Z;P<*p@Q@)=G<=csSX90J#(uHIHQZ5V#OWj@J4k`b7MLU0hFJTe{glpvR-pDctMZ#Py(>$9=biJ|}a4Fw9RROor)I zCeK9gK^(phja{cPAYwgOpnuv446@IT-V3vNAW8r?wLSNrcLIZbFIn(a%osbwhz*Pd zUpOCyrmJFZV+NC4=}m&3Ll?k+H{E0g{TR9(rJsqR8x%~-;_6%*N{N!$R&+_L>pXL- zu{Z@1$#M#zZ}{A<@CS_Kkb^e@PnB_H@7Z`S;ofDE~o~zeTL7!9| z;mAPlHHu{%PW4pua;7QU*!wK9Z^weFg~1!Dde!c?u$~V-2i~{-1S0|lJ;z0KX%?<_ z@9*v}P(#i<1(03Bin3rlTQAmG3D?V5_m{lbjxMh|oSWAG!}f@utpUeX5*nj!?Nllo z`Qsa=Wd_?>@7-{^j2-_qmwgAEZ7%gcZH;H|cw~L)VHY{@LzmW=mA=sp$K)7kPygG> zl#1`UGv_N?G-Z-T|mS7G14)+QA^?%qCpk4wn$^Z13z*ad_=SpL@V!PIyT zd2=thn7~yDZS{r+cndk5wzteCM6_+mj2L}+@gyM3UikSV`0C#RUXjN8Nyi7y$J@lm zW)Sv<%EM`(%nYBVaBqd_Vat1Z{;>V!_x_n^_O|djxWGA%0M5nIhTXkVT@YT_+hc&{#bCnZ^J~sP>U7jW%^KdZhxV4g4Fm z{%55a8Kb?NY&sf!qn)lCXx#qJk3Z^jG=8#->DyN8fmd}YS(tv;OKR-nUakl?hiL`6 zlqY!4jX@Qz{?!z~l)@rzFt;(TS+)~})E8*wJ;#q)j*s}pC_{xay{-Fsc%u7}!eT%j z++EicbfAkC9oA(Zqs>(|9G$aNuXawAdPxxB_Y!f#M!`mFM8N%h)&*XLLb3u1XZXUBw1)T#6_k}gymHru<7NlJ+cTrU)? zUq|@iICT_QLPXJHF-m@1{fmxPTw?=DMJ_7M*5oQP)U9yN43rmqCfTE$icIF#O_ z>JvA4EP^ae$;L*@9kVT^(48)WMT?o*O1k3o*Ev*{oUMY^%1aVV=fm2w((MM32g1pR z;A?g=6u5&pwAgIqQq zq3T3HM#GgR!PM6Y;J=NQlzj;4E>RJ;QX6F!^61Nc zE!^1VIXgka>=Bw$I806?rv^K}o2Z7gA`p=m`CkBk1B2eoYDybKO6FN_^PGSBGIJSGHdXBe`AAd2nU9yfR{sPlOYGZTd{zXWRCTV9{W7OLfRK- zOJo~EW0d45c+dPkMIclfn*mgGfhSp3_wTMMyX1^)4^h13#`auZ(UqoJb6MZty%-V%2h8rh!9wVHprVLyO9ZJSntm^o|6X+VM zTd4X`x*~I4N~EGlEmt~?z~?XB|3cU8MmMo#Tb0TYlQ+mFV(1S@_e5U0D?h7O`-g0M z-+3M*d@qQP_mQbyWqRLY&|esiHM_u#8LDTHQJA|>$rq}pd2UTIL|@G}Fu*wt4r2&8 z8v!f?M#VI(RHa;Ak-B7zL{UPTwwv=B3KY%|N7B>B3i*30peqt)5d`bY)}EkKP&1Rq z4*ATD*5Hh!NG4g&dFF`OWgNdRXc+p+YSHTtUPe_T6;p@!;R;n8nc{>kWsS^a;w4uL zU=DuGPW8Ync}?0uvKwbg5|3MUkPdJblnRxI`Sy;j&5#Q^N661ptgLdcHg$jT?gz?Z z>iJ5ap{}qQMO=8QT(AAUnM6=@7ULnx?Rh$X{`kXin=)BGbd_A((O=?CtW1j5W1|@C zVuvncDO{v0ysh=#kD7?`zg1#ezSWOd zujmm!6@eg7-)z(@9ipU?q9*20l767qZCTxc6H`96e;T@s%33Etq)@;uSU2DH{Oa|Z zYjaGb=HTVYVDNQsLa1<*)I_l3jl}i5MQ$f?>ro*Rm*bQBpe&-i&Up12P$QIYZL4}9 z)Iy86RlZ?djR^IX4^s;-aar=@*hAa;D2pA6L+mI6U zzyE~kQ)t+};+Yi`43co#`*@hd~_1lJw z+mSkPKXmp!twZ0wYudpR9vyi{VXCUmq7r6PNkWfl^ESS`JIr*r>KOaaM~ z?2rH%wxVTL51QfZPn%>qcB0m(maA@88oj*K^ePTvtIYQ}&U(~M(H{@r9F2(ZZKD)% z2Lflwox9m2Elljn!^){|wxKQl>~VjZu7Op3Kh*cV)dg+n%j{u&!p7`s@kJCTy{TY~ zz^iiPlfy$*4@nbgIOxW=N)78digGz^46_~U8vIxF zpIEw_UT+9H9s{Y)+N&6*WY=#%l>qX{EK3Wq(wJE?Xxi6UdHe|pLUz6GpfLH zQ81bVf-~g6*<}|-@6OL}1K+f+pSixpe1p8b7z|hz6CAvB2ow( zo}B->)7;#n32hKHPa%VmiElNN_q3UyL<(-oZRZG{LyP_&O4u-P#t%OCYrmUQ`hsr)R}(>z4d*w!fdBm~z5)Ou^F7Ga#yj>HywmfkiSl@JW! z3*21!rO05>6@xO_m$!9em)A@~#@V&{`6~iKD|c`QA2#`R>1RlBT%(GQTcYLGyjZ|X zvbVJ1PZ!j7aH+B0dP3(Jo83L#4dN&>e5aYu6GZ(xRd>uJvR3}iK&XhZCqea-aD$!^ zmC-3+DmKW7`qvM7D;a9aW5$Q;mqs$JVN+7CeBF%C<$f{iX;v%`!?26gp;JiK4@!$D zkbkz2^Lf=0I@2s)EbTxI7={g&lm6&Yx7c*JHUH=5Z4S(yc8sKhM!n8grRgcV2027B z6~z4+t)Vw199jn98-)+rPx52dm&=&|DuxB9yG?D}lUT&pz8vf?j8s{G8pElSxlp@t(^y5etTE}$cLaog70 zO`GfN_1^v$&=mma<~t=)p&WbXy&u74s5#8rdizEy$w#VjW9+qzffLd2*7`4KhPC= zTl?Lc@8bvU*O{Wj7?IwphjGvS^OobpiK=r@Ym-7oOxHuISau9qdhR_Cp-RwAl0Y9V zM4{~K$^mK5iJl-8)p?GdIVpmJ(u(;0Th>MBM1FNT)cKe@xaRTM!mYD1-6kiZ5 zi=UGLxk0;hq)5O6?nTk}_6SbfAE&-yN5|g_XC(H|C?!d@Cm{D>=r3Z6G-B=SudXU@ z{`qc&NVpPtJvnc0$-pD#bll;v(=uoJ5S>Ce0gOZ!?Ep$+$x4l5zgHykjCV|vR&_St z(UdksQS>{PLI53R)>mdeSIOx<;eA`5$yi@`SNmb7Cjh9$yj2b!A3)grcF>rmI(~uC zob_9O!-#LK_iZ8cX#tM~xmJbM&KFeY(QR~rgqi}cI3wEB+|}9^Ri4F`(IMCVjQjfR zd*f5e_s}81^I@k=P<(@2R;CsOs*Wn)w~`E+HLG!IB*u6~c_7&irqUWsBP=e1_qG^2 zr%|P(NN}%(`>v}_#|wN)SI%PBXNtQM29^X?Y=1A&fV@CQHW#T!?G?or*pGlEn!v*$}j|ekpAwl1W*^R9U4s;7*CP5Dd$9m!3 z;xSl>gJTkyxKQnYv$h&VF#4!+5&tw}>)n>^kx6vXQ=LP>4o&>Tu?tqB zbbA_6xB?jYQzLuO{@Xp-DB~eo?pJ37cFm@O(m(er{ z!oTF7Sw6v2`4Q+?`LSzZ)D%5}+La8IhBKj=3MB$h+u+Is(^8N8oUjO01>;U`62?VP z)liAE!eJ(*p+vh^n;}wSg6##HKK50=nTu7}%rQ!VRezD@x8>*^+bHrA8?nxR>M|C9 z{nnLss@@8q>j<4}z$DoBpMd^SR5)U0S~VW5jD*6sM4OOB0wqdjEyibDH$gIkRq;ar z@J~O(Ftm0T0^P)+rX593F#o#vF3s{9)xyebYs>^O0irucHXFtMYz4P<4=ZJ=E}xzr z1snP^Fl64MY+`$CwS=4x@ex=)NKOn`DJ9~(=PuGY^?GCh{8!aU;3cbDu#;_loPY)BmjBqz88z*0__ z`0{iC;Mm*3&HTu^77&mJ;nUId(Un)gSQ_ zxl)pW!-+8CCW{~)I_?Bq9rxD>>xmo-D9!GPY*w$huU?MBX}SUf#gl<*^7^p{h67VE zEzox@m>4pOW+KB*^%P7UVr#42riVD=+xsiKV|s%VlUrM!`^=)p;CxRz_RMvuSB=P| zl&lxek4TpH3>Vb@OAOlPa_< zK0vKcczeeX=OLR9@DbkPi{fji$y4A?kpOq=2C7h z2|Qvfm$+yPz)a55zRz1!@C@e;`X8=&BN>y*m54$$r84f#SRz5^r! z;^WX0BtsEtEYV?ps8%u>S2^rSb-Di7(r%ZP;!Y~q!P>(zpOv%&%?TjZ%dPGC@*j`W>T$}YZ-N6!aO5KuQV9p;9x;cK)aN(c;QevBeLrhsApJV_~A>8<#d#ZnY#d zys?z!Lb63Te~219YLb=o0|o#pf9P^{w{aayczr(UKZ?JQ0i(*hFYr; zromB^;=k^z-ga+fH!d`I(YxxSRrWawtAS<H{ds?&kJPGU6-Kch6cKZ`Mc(DS2QYZ196a(x=Yxv_6c8^W+_$)j#8;PaG_e2#o-<#lah#L zO$_L{bz8)h>^P3~pWK$VETq;M_ElMwW+uRB4k|MZoW(uEWem*Ld z{gJG67o(Qrv?=yDog!wMMWO$R3yh$ncPDdqA<9bqjOz$zrVkd;98*}FqA-#SNVP?a zX`;?*+TsrV1|-7yx;=yVTE#HpNR8eQn}PFTc2_J}x1BB1yC}|3xqL9HS`O&cg|hAN2j?j- zV>t*FP>_4ME=nnL1Hv_#mS|k5V}iFzsr+Fs ze7kCPn*I5&gacYLSNBC+s2hZQ8+~)Vn(PT?wmM@yv(!hm<8TOOq|s$8A}i_;WgsY% z+92^sX^Gh=Hx8jG$XyNlqf=pafrB;KDD=0YUry$oOor2Dk(IbVe3k%c>u3MvLApS^ ziD%T|9|ch>=!|)D6OAASPu|KD#cHooeyWpslP}7H|A%TyiTtJ}(^qZCddL$tDWxOX z`X#JcVjt7t#D4qWn&>j(%@0oqB8 zG+9rjf}rri{yT&JLGs_J;vEG4yS-Lmy^Z;=Ls_USI*x;!Hr%Yv8}_ei5)Qh&9~mE? z!G3orp3fCdLPd{hs(h;qw}XAJHy(M+=uQGYD|jl7fKb4!kdK>8;Ik5HqrQ=(|F6d) zc^u4ij)lH{WL9T=tr5)8#8$NT1*XQh=Y|B+Dey@uID8vAeSJ2$*ARI4lCwMb>N4s5 z#o33YwK-RRou0ks&NA4fjsDs@86pNS_D>Y`59xoqjDVaNPE5W%v&d0OTwD3FQ~5s- z{uY#K!vKGp=%B!IqG98y^Nq5N50%n`fEN%NSQCXevw=r-u6Sew7?aZ*4-cWRU49-a zRYZMLx=s>z>Xbz%8w$jrtpT)uXPtPvT)ZgnVJ=XYy#-rAzeHhaI z&H3f8j}rZ&8&anXG){ec7}k)md0Jljry}ZCDy>1~foIrr|EmH6_qU|@=Q)zM8>`N) z_}7^?wQ~{R6Bd(8EEL1c-e3~ah7#~wor)j<4cHn#pggcf1~#-Uw$bK|wPs9Ni_$~* zk-FZ1s;w=E*C1Gzm8-IQx~iL;1xLq@Vv}#*Y6<^01RCSk@9Y|14-ys!;eFylr^Ouj z`BNChCoN$rNdZfD2$lcdM2sfSlmk_=$F3J`uC64-!Q7z6Smu5(fc#fU4E2|?S&`bG z$coO!jk`u<#{-sYMBF6hB0#S&_gtnAQRQ5FsZ%kbVvHoemq+TbHEf0t_&W@{Q}l4* zWtIB4@)77dz$eNIi!k!uI9GQcJOvnGDJdAer?ql;b7W+7}4T!6v!4JPrvm$Mn%gcjz~TwE?)>*|CnykB89}y=Tg+b+|u~ z{g_gpo33rK5{`Fh7hZDH)Ja)OVnn1Xbn8R)I3u&-XUAr#`oN68s>wo$m-la}l1tb8 z4{iROj5-O+dvXH?x>ed^v~Tc7?AS-8St4_% zxZ6pBLDuF`0|9;J4ClD1&~OCZm}Zg&&+(vYwxh8Cin4KoF?|~b>UV_1hQTy@k!x@u)E|!HlPI+(xlu6B-#y|B$q{lu3+aqDs!)A!bb`UCl_6jOmfiZ}-ht z?NEKXzPFf`M)=b{rHWn_+=NZG$pbBqgRy2*ZOKu(~9)PF&sL~P}Co>bVW8$JZ`15glR1}vJ ztS}NCfAW&y^7qqDwVocvHpz9wRSd6>AXH(zZOl2aKlK5nOYfiU@@8SP*lD!!^v1FE ze(hs;4_ifVj9hL!O!hq1j2cHxODW{OFFSYhUzZ1z><{2WOj_*7wAMhr77#DIEL_~! zgtPoL9vS8$@`Uin=I_6-CmCl$9-qfI(PXl#C;R$^EOm6nKae>;H8$gxA(oc~GbnWp z#1MQ$|Iut~O-Y=~=m7~tr`${-cIcMrm^ zpLA&ewIcp3rVsFrR&y@7|2nYT!<0m{qY3O+e~YdCK0GUUVXF`Pc>bCrp&H1?`G}Rf zrjOk|OWfcF#4S8KYMj=XPdPHZpuER8+Qi=w)GeC8+A$)tm8WUI@0(9)mH&zl@Rr50 zT>;hp;=a{ocepmh&->>y2i06jC~E-Tc;Ye}@3kuBjSYglRmjEF18;g>CB|YyB{-(t zgP`=d9Lqv~C=*UUSoLLa;#ImcQsAa}i-8>F6QUHYQFK1NOU)c{)?kB$eS3XACAt$~ zWFb8e$MS63$rJ*lpa0%A(|a{rnM6NYrL(8zp3AV#WQ)G>)i{j)jaqwb;&RF6Ob)vd@8nhybKy5)p%-zgSKkdkt`%eLNDh?WlG}d;0xr^QNU3Cd z*|x8g7v1~(JhI#$X5d0mb48jb{cg#{YZ$LpvWX-ke3kR&^)8=8AFPmjwJM#&Zp$^% zWHWIr1Jx9fT>_%xuXf%pe%g4+33`KEUx|}aOZAkemg0nSxX;DgtlTL%-%2D#sP2Nj z-e%$st#&U&-zH{_zc*|_W8Vb*mQ>|cBivoW{*#X6)OYh~=Jym;XBR}Q(8!)~r?qI> zUH+e^f;!6&sRVDuyj~$&1C8aUz0=;5aVti1W!cdi#;phb-dL-QrjK0?ycjnp9PP)C zenke^jvh+m6jDvK-|^c1oPTrzN^^LtA^5FO|YL&s57Pv{|Sof{}=pf1-9KxWyH zBH0-3zVm&!G=AmWfR0gJU!T9>pTF$6xdxE_rDFJdZgFE?$ic}h6RKI-P+zCos4FgG zzci58+gD$)Sly4i)|#q@K{74L!hp7WNjL`+snfTY+U)l$lV`Ra{u0xIzYh(Rf$iCM zi|`Wb2H$%!HB&R1o4 zIixc8j{CZg^3dTAjN&b#I4iBOyBKQ@te`gqra@jXfe>kf2GvMGBMr0^hfZ}3VYj_e zr4#fc*gD2*!q2fk3CiMb!UrUq$|ZC27vZJwUWO5_U-$a$S1fb+da)E52=nmd!xDzDSyl^&z6cg)JDL8EQU+ zms&-#affZuMpvajogN-lh>}RbOg5!7EU(+KH1FZ{9`9XAnM-bB6FOlBp`x`kbH5#R zaS6;irLolNP%YYWXCyXF76aJedmSV$6MiI+MSt}G5*}dBdsjD;T_16|n1bB#zUuNU zi3u?LtN-?kO7040$J24LDG+~Swb(QG?-Oz|kC|U<#+bi+RGPJ|{lZiO_XMOZmTY$o zVB;dSB)eL}xRWW&gbVK{mFl@_L$05=K?BH#x^fo8jn`|>&vD~??>5X5szEdjoyVz> zUJ<-sSf@G|r46aMq;^ND!4LXzN^v1ghcYCztv_P8-9eTx!?|><;KU zp|D*tnJFf_2m;}Jvp)R!iu=v6Y!5Op9Q)&SZ-Ht?JdF=aFjCKn)0iCRi7jt3WetbV zI3)Wr4{LR=z4PgL)d+bH(SF_}qN_@Vvj# zKIt2HD=yR-_L~G;u`qd+eL(LKm*N`#cB_AA+Wdbq&vy3zmHWnklg?M(bDyh}cUV8E zOC@!9b;e-4);`CFoOj>X^(Jk|g;nxJQ@|dgp?_%|n-&fcD7)hm`kH1sx?`kN86gtQ|96pSjD_gAR-HMQHtwHDC1XoM6XA3=+a)@7O5<-PK z-i-W_y0}u4_i3w|4pO=y&da0#%B6bX9UA0q*vk9tm}}D%1Q&9TpA*Ub>6D?B^04(| zg8ORBfr)qQ``i5Y521rO`|X|2bx8Cq#${e_+Zblv=m&%#dQE(H`9!S{@rZ~H`il=) z=hR0DjxRnRT5l12+=BNAe41!ZKjFT*iU?!Rq>|!kJMqL~aTOjke?M{=agS727Za2# zUz3@E>D`{SSKkO9QZ@5Bm$H32QG{Xu$UoWAV??fsjl{n#<7`0Z5CW+sg4jE;jymG2 zo#fQL9>`6bzkgvCJV3ZU(w{m*Vp@@YeTbxghhsK&1YYmZHaV_(JMWn6p*dT7NSA9O z3TyPiF@rFoTuzwx0&gsz+x&6JN*Da4KF7;B_|2NI2?+?yS9-RP=*tpgVu;RE<@Aq~ zOlu(9ME=G2^zjhS#-)E0>CdrB8BEEKuKFG}+5rlw6j1~jyqDiUmL4D<$TCfpx(%I~ zkjD=z+7J>VN3BTZ-wDK?@%(tz8$mtpE=UInf9xu0qKeZfdZWS@V+u&K>b@XA_-O^@MUkcItJEET%AGQ6pp?goM4R2l&aw9OYZd+_ z;Xh9xo+^iO3xRv%n$wRKIH)gwFhw}JP+gHmrn#TQ|MoqKsC~k0oRkH#*a`)oN*;26 zPdIw~Kxq-RuF1J1th@zZ0Cag7O&Mmy)8 zd|bRa<`lNwZO}JwwdN)>0opAn-!vuAs_+&OqXi)2j;NzSkJxQ)9q1&UrxIeVe#ijS zHJ|OKRCXro0ne4?$zBDp@ddl<^4pIoFqYzj*(mI^Ngaoz>~htjzD(gJ9AYV5^vtaI zvS8BxTdK8qTkaOad4d@o)V0a!T$JMPx1)29YttMI`a+ATrj+qZWnIVv&BAfOj?c!j zjjVM8dxn>mtAO-AP~9auh<~itBxU*{lktzneHQXdtH!}Hr;zOB_8j@tfRXjlsSxju zTXi41mNbCw|#p8v2_+~$13egtvue22RU;3>Oi z8-rH`lCnE{*z?Q$FpDTWLyZ+d=3z(c)A5ePMolUnIL>A{+& zC6yu-_yW5qy6N%0$H+wbi+Dv?WKp7))GdV+N7o$`k7=q3;A^Z@Blg!gt>WA{UW@MZ`uYhAKj9Kaac$CBUsnZSZej_)&*Y zcJ^#;SPMkq5k~8J|0hovkXX_HlL!*GS=i#ZuJaH|1rmH) z9*X%cLLh=EywfmfZ*5~}spLq;vG(+%EBJ$jXxDxLKNlxfGr~Z0)QSxHL5d$-)?31q z7*-cj;M3QD_n1oLoqlC_pPz!iaxYzEW`2ou^zLho+^Hu%DF#5TIxj|Oz*+%EZgZEP z21W3wN zr3**qFSra5?i`8`zaTRG*8>gM166fKSNp;(uSxdc(wpH!9GPbQ*ROYZ#0?1ub|jzHQ$#mbyt4*9RQV&esY3gah9MDfnIg`1ARTFcp&HrDua>{J2p$-!ZS zmEJJMX-J0=1t%6(vZog@N1V?+$16_fi(Ab_##@0gnQr`sssSXj-VZre9xRjV@>RK` zP|d4u_)=J7pP)RV@bA~QJi9<}|JHR(iP$?k$16|2L)e9l9dw2}z$czl+eS5vCECMz%2*79B4pLa5~(4B*rsi+PV#f=~Wzo7VBa|T|7%j;IIWBL)9I&b4X z^_CdGzE(y@LB;grzfTz2M#3L=p9%k5`h#5C|2kZ2Xi)x*)-RdQipK>QF(C2R5rk+ zub8&n9;~vv29vjvza~>9z4M-A89jb6gV8|Bj-b?C&i!kH&NJ?v&eB+4!$Q;Y7^u~w zjAhefp-gPFSh5EH1Mf-@Z{}b)3r?v6;k%@sUupt>%2A#+EBpt)*A+nt8GL$lJkmy8A%+D4 z@8W%wABmaq&gu&8A;L+SdF9hX2&*hy4P*A%u&tbC>6}Es4W$XCvWGd;5^|}XIL>cC zq*SX)$GYDp>qG0cBsgohd=09g$}t_Zbc`NMGNN?kweR`=zT09Hwj)iUj#}@3Eow&^ z7?t_(&VWpv`&n|A{#yG2IWtmoM!fii?X|w%t;FnZz=#{mP54CPiNL}Wq6Si2XMQ}N?wt8SV6M&H#pPq8&(H zHS1JA(UwwTxrXiMqxpn##eMU)_A7EbVB0!ydg5z->4dK1Pd#RkdmL-^RGAes7T_jV zWM-S*0CuJd*`qjBM$ef4W=K?EXC9iWIR}n5`HNf^q|0(WY9iji6zgLhW*c;eIh~ck z$~$p8dTtfka3-LE|M$1NO}(4vVFO7{%a7bhK}Vp6f5L_KCm_+K*mJW!Y6G4i4@B~i z8z`BC9ns2B!*j);sg+A+T@+xw z+NsZMY7v`Ir+P|!%8`gFMNLmnC{fs}>)i`hI`GCvCMGitqs+{ZMVAe& zuj#YC>2)tQs9Cfi6=9!88~vLJy8w!*92i54273LuEPdh>O~bdcxN(V>eR-dvF1ujy z+I`%GMfZrQjo!x#Tr7DAxlB~WX2LMunbDj!q*o7dOc7;goXs@#7=Va{o4kLuSs_ih zBpNAZaVqTDR+kjNF$H0|&$_;|%G9Zs;OI(~{ z>v6)arcI^p_rMYO2v!l#5x5S4VNw1JU^jww#A(GLuA2em7yd;&(hJTj1a=ZpX*x%btz_wn4_L%9V(&M8`fa3D98r5j zk#d>zTg0sSa>V25{5g%N>#e^n@DeZOdxUian4MD}Pz^RVdlH=)e;114t9W zoucMn&R++e^7_s#>b!Yu6hM21X5WilsO<`O`8=SOR}aKCzFGIrS8t;cudfH4mv0ZF z_17YsWQ*i1xJ&|uIE9BommuFeSZSstJE}ETEvxuj1mY$1e@E^DH zb(c0hl#g>>Gxr@kgbQO>{O(Kaa{~6WzN@}5+`N9on+&J==6{jh*HHKU(7Z;mg*mR8 zL*WE%zXDGb66y>Scrr{u_1fhIx!1u$-gCIq$H3T36B7)HL3o4Msr)r7j`u{xRL1VtjOhZe>2b4pX3eGX2k zGMWcfCUIiS`9R%&Le-9aa~fzKlLU8S8t=j|kv1|l8`=w;dg1a|3|oT3m|Px@;=0r} z@Rx#p7#0y3nhDa6{K@exT-t1#7GdF2u-Dw%+_dN*K{GzB0|H@TfFWQS;MF|ty?BZ@ zhonIU)nnO>)$~MV^dZ8odM`(zQ*8R5a`bl?vV!N^xsFe-L({?o5~BLvMf3|`&xBOD zB^18F{q`S0_<1UrXG5E@Jgg2cbZVduli4o!#9UH5@(I+Sko|8fea;9+I@KS*tJlHVf|b zQ$ck&-JuUAv5n`UZnQ!!SoDk8BgjO>#dN`SCt-`^u1XX(X*hLFXZy62TR-Q@RFO(* z8I~nYS?3)7O!1AkU&f#Y?BrJ>oSwmVMG0M z*YY=|rga)k@CzYO%FF6~9a-T@5z$Zw#DKO^e*5BASb@n+eI&Jf#Tyvx#Th^txYg81 zS7(4L_YfW0>a&n7wtEu^%*ZASonIz2T~nLm;w0vRA>&nTT`bEL$)>`3WfOhcd16rmE_u z>^c#vX#MVF5l#$*{El2FAmzv7g2K1accNNLHg|-`PtoY|br@@8 zgO1okIREU7ddp?TRZ0Oywdqh>tasXL5|8Sk7&D_>fU$W7U4m%J|H{ z*=v_KzpQ~n=%Uj20kCT`-rJ1x;dpyULP^0%%E%4@R&-QiS-!x>)f9zBP*K8qb zDSn~CjeK*plD%`BST59e_cn1R8@SyqgCilrvbLDsN14MdXx@&`{h0)02yRm5%G9lV z#iew#?%`j9j?=&L0Fk?7kaZTx_3Jbj_PI@wri74$5@ac5e!0KSP*3m7gl zCtt1h%`Z&Qq{@MP)+%R*wjNg+59~JJp`G&GaCwo!jsn|!j5g1)sEJQkl$kk=`nbYH zW+LxSOI_?2GSA*SSsVH$6RE?jJUgh?@o1PW28G#fRsH@w%t-3MHUiLk=kYM65bZOC zCJj5a=Q#LJCU&fX{lAqvak{CXD6NBmv(*UMMBb1X>K3kaHfjLBd9|%Fv5wwvZ3>jU zJX(V~v3cpD3?J<*Qxai^ENLT`3O1h-sC&!5jH>Oz0OgDr2nr)Gu-7?$ocPOBN2B?& z2)2{WHTnG^c+r_e^M2q8l!nPd0XQx7h+H60h-W~caYQRs^9_Y9BKANZ(@_>h;$?P` zl2$Ghn=zmK#okgCxQJD6y(n^ZRvHIZYyKjMgW^ztM-MNR{HfT#n-cr;GWUmaAaNnR zs$uZ*6h6Z)m#K(Tb2cZ9)b!#yEepoitp@tUcK5ket;WdKHKOwV#02)KT4~zCJq0~U zVuo!_yofpn18ZB^XhE6`;rpu9nvI?0{j@slIP1c*r&MoGsZd;KqN7jL9ZxMhretmm z49DPuczb5*fj5dZ=@C|kjKV$BEbcQhO2nzx32ygPfLwAgM^`EopQ-t0{#IqiiHyCG z$sdiVAX%9j$;qi#*#wM)F72wrBnk`P4_tN4i7QAo|IkkAX+HB>0!I`|V|CbJ5>YtB zQt_B$*C^9`A@a@kc*NHjy!10$^cKD&Ednmn7g214^0ILV*w?n?epfJ11vU4N z1WnP}4D8ce%V;4I*{kH@<>f3-_SKg4D)kfxP5R2U6DWzp>zg#tywZ@K&rtZmanMqY*Pi3}V zUAhA^d&e-_KFE|NKbAJh?I_;ud>MG1^)`OyVW{pe_Aq=WwTU3-ThNW1!~jo8H{;5DUY*{{Bdt_v*I_ ziTZi1NZ?P`OQGW%uGh5tx3@`d-;Kb|E<%a-JVKH&Io@jso?RGB5pgeIh1Opma? zm=sjS#$r-U+S*>?3qSj73Yy;w5nJ8WDVRJ7^Ou~ck+$Mdci*ghVN?^*O zw+PAIKze%D@*=u`twW}SND)v_1QAo`d?K4^P5+}7Ys3OvVwSNeyb0NU+ICLf8=s^# z9k5V5WF|$&fbo+TeL$bTk4?jcA(Y9m}F+{LWl+X!2DUp2RJI?9<_uCY05fC6#6#<5y*;JTohasKr;McBE zqtkl&blJABR$ZHw zV#%7tRAHpg0}b?H0u|)+yswk(a~j8j9z)^Fbi({$bt_#EWSECA-5o1cdGH6tk`U#S zafzj3AtEq{biBny);oqur%YPF`UGtqk7Z?I|Dd}8wQOnC9-6@u&2r|S>)<|)&r28u=GHL2aht$8$#-w#R6 zr4HnVA7{2h7FAQm^IhiXH7p&Hk$ul98ElI4QToKb^gy^G3~@Y^9EV_C9I_BCk*m;^ zAzSpI7Y1NuR~+Q1_G{#*qyhgo>Ff_|g&jpsU%D&ITSyjlff^T$Urmc-bK4mzZ~81` zMY{iN??YwUL$JDFYs9Xrdp+&!9dpf9=O5;h$O4L-rq&&&=chi})X;wZ9NYXe(PTWv z!jVavP0~7?u8Iz$sK`OSyI6ZDd`hro7+jIcAb+og{U5m#&m!{Q#=ci{mT1r4Uk2^$ z2i(AY>k86Rbee9q2bJ~;r8h|TM{#FOSB*Z_WCk{&N(25dgt@443mTkUv@Z&Z;;gU@-}5xx~sznGKAT> z=L^*Upww?^B&z1U44}1FlapLIh;W_Twwa81cZp1f)#t&1c29*9E0Xzl6uoyJG#-(D zo8l4-l_Ui%B2@MbPz-7?S+gEe0`{skL`-PhSG42OB3`_rxyaWX1fyE*B$B0{LNP4d zfEj8DwYKNuwrDiEIKVW;UXo=8>?7xMY@t=B5K#oep*HCecLfS3%@M9KrE!=nv>_(b z_pi+W`ZaZWZf=;#Ma~1(5TLNQc|pwGoNqu0E0Z@{BkO2UwK^ok(BB`YunkCD3`9T8 z=~XVa%!?7os99H7f-lTUykK?Ypk5<#9tA@p6PR6m;`S*25J<6Tf6t|LciUN+guxI2 zZs69FqLt^l)EuxJ`w@+{{5T(1%l@t~IaMXDs=Ma!eBWH4!U9>BG;iwk#2#5Ry#|+r z5Otq0ovHJX`zJYiOe>ij2y7F2uf(tIB~wA{wLTw7m2qGj#{w_mjCUX12Y3ywu=*DN#|7vX)wHCIS8G{K z6qw|35=VLFqIxkH_=f6gX)d~kxKmihfx=oP5y5o`UQ%kiy_a*MCX(VO}CcmUt%1oz}dTX_4u zz)yghN=D+JS+$iQhg``ySz}4%DqMudyrSLmLQBTl*2Lz+PU^Ri4(nkI$CWADA9J`& zU-&$O+Z;)?%0rmD!s_ZS6Fe=2d1$%0i`^y!PTLps&R2-+4mKx*90r8r@lpFDuVZ~5 z?(QoOUc{jHGwjCH>>HVn^Z#-*W>noFX?MHrU4cHM@m^2WPIdQvV(|BXa5AhZBR}(X z-3zg-aTq#!R-;^z=d&b%k0f5+$wT950sIFiHv%;YPe{XUSg@FX(014CNhAI-_EKg# z@l~dT##lNiKf4TAX!Z{#99@TG1KWGTB@`!`9W+h>{bij$4=?#9rjlfn*WCxt` zJ!MJ6q=awBZ1O9(AqliSa+jOVJYS`K6???9_95Ci&FSKBRNZM(O2j$ffAB_%6zTDk zO|N_I7DbUSJ3|#)% zJMssF)-+t%W3AG7R{(Ps@qpaIdT~!ktuhrq3}h$QWB+9~ zf#EOl5z+1%j?Ze+H4fl(*4@~$US(Mn^nh3Uv0FmgTKnNeWa9_-xlZede4q+64*HuPsP;FD;uN5gD_@EW@c{9#f_((rqFj}Qneo%MxH>2P!=GtB2Dk$L z(so=au}#{x{z2EZM|Kz(=PokVyh?(s586jYS7(YB^EoVe^OClTnFvM}-bxwfx+ZeQ zgC?-OK+~$mn_{i)LHg?%xbLzdN*@UL1eQU2w3XbOgE*?JTGN`s2Mq5Rp(NIUCa zO{Tyx*;C@2)GulKI`@u%Db1N6ZL#69P1$H!{-BygWwM%uldm%(D={&zDLv2BSr zkLz^7h${n+G!Nf5@{|R3n};$_q5H9zAvd{KK7J{`7?NDFNPdk&bz=UceWAK{v>fei zg^&hs2V>O$H-RsEG0nWt8e|!~X=|GZq=YTkCL#?nMeY4gMNw~t1d)H1w9k`~b~9NW zi*!?;s6GqCgv{YMQ$3-?f3QsyB{wtgp0w>ww>#5#7Jsq%8O4$nX+eGhM07Cxj2>s% z(k-HKOYAY`X8^UOX~G+AECYu}O_vVlc^BZz4$Wr#zDhI(F=wd`X3Z~DprH2_vb%?c z{dKGSs>-(zE@J3K_dX(aD@d!BM5kmDVm=lz z73m2snenq(hY@r{@8(HsML&Gu(}HOTN|z_4k5|6(2H#TzboBSfRQ#>D{evil0>3~z z9`qnq`7dveyqdu8n&RE(<|v+1e{U-ZfHVN`6rVE8QzRG~@6A_Yj`**s8c?uyP z6*t*0_JknW9>5f_-f`k{RcOH!QR;BqaX{QpP4ZG725L5m0)wmB3-&1I zC?%U?mF=z;r}%r6+3Z`&e%dT0H0rza@)}ETdxVH%oq^uNP0trJl{l8E{*B~ z(+4Sie{l@O^N4tZES*OG_3+gA`qS2Nc0RD~*rSQ|+_`T~$LCx<)YN=j?a}Je@_WE@ z3t?BhR$!tX-OC&dG}Tjh+goUdHdPuwXUCbplz`f?rp9=#m5t?Az107PS<;Q^q->G8 zKC|})g4ZH1k^>*89-h_x169`Lsq7BVJJ39cT!i5+@t*dHXRbH&>#j2<+>3eq&xLus zC*L1ldZ%mrXT2KG2^0{9FzI#zgJVqFQA0_h5c`F+sJ2~gYr_3w7 zpFNX^&Z)9pxuc7y6%jMjT~l5bNApw>{ClV(dLYP}0uX6}Nv81WHY;Xrzy>rpnf!D%*q zq6I}+*hFX_{B|x3DZF$D3K_lS!BGSS=*lq&N^Ci(YjK>T;3AtOV+}*i8)Om=DM*-& zvLX>qXFL|1>g*p8v3N`MKdj)-DqZi55{LspNZ4kSb?#s=m1PMDVQrp3{u5FhMP53% z^k}Q6F&nC_2gbH6NhiFrl%o_=ZR_1)DsvTeLD-0ExGAj%?Gf61U)3J&{Rz+s^a{OUggR_TCmvl13WI z^5!7%1n76FR1roD3+F>@7Yv0;xlT^{SQay)AeT|zGq%q(6=r6FOzGaN#Ij+(f3&t& z58=>2Am)#G^C^^^BV3bE2j6Yb`6-b2Duk=RMSOn%d^04r`i1+qI5R8}<8jQA& zJ=q>tXeiN581$13#nn4jB$7+5BG-c&2L%xRfO+}{DZW|gvJFW~lh|Fgb0;C7V2&%T z`xuDO#PF4pAxxuwCPXR@r$4^No-#`(uKQw8k?u2t>daS!5nkRK1>g>`Ve>(U>mWf{ z3wLiV1!FWxk85P$)@VDw;=9THI}Wh6ky=0Hx=EbEOkow2Nbs5>j{{e8W$k2}^IDMW z(FxbHPDzymbSoK#%%i=ONV^}AU6@_9u;gSzymcPOHII9eWXHbjH^wQ?+~Jxw0C^gn1+P^h*`!YHKmeBrg%U+ESlgCz!_}Evn##LCozeX<;ESksyecEsLNyG1y&WQ^ z-S5YKAaBL06F{5HYjv=NP@*PBIz0=w6Hc%yx|p!|cYla^7#foIg4q$w^FQYC{t^-a zIE#PQ2V!9EpX<`gXUT*9X6~EvuIGu{-EFp%R(Usk-E6W6{HEg{CK$~v+_GBqelfxV z3RdDYxii(ls_M4)hh7qTOI$nQNxzDz-qL?|^r*0#Uv1=yMC5#NYs)sYTB%`rG@{wu z#?(<`RIGJ{SFl)!hqb>P@9X4~8jPzK=?XspUvqGYyz$*U<5kwQV=c+m!w*0PW&k_K zXWy&LDTLSKuW@!TwSU3+CpG(pFSQ(8ivJxBU(hi4wXFt=%W+BM)gHgNuKrVG*=@38 zdT(6WHnoY#Cj1AI!ee_EQOC)f-O|;D!PXT86nlIdr{{HOyNB%F=7FE9^BLo(IP4SU zDqd4?(|z{V7w#?EoPhLc3E<%w%>Y~iCQiiHo{=aQ|E!#ss`Jfl5;O;k>``erC%c)Gp%usin*c z`V8gr!UCL4ldpC)ailpChpZS53Kr<)Ne4bVSUYxK zz!*S)tNI%JMGQ+LKWCcon+688{CM3VJ?R`>`%(!dwO9D#>jEfiOeZJj^d%=?$S0=k ztu>+}9VqV^xot?;^?L|4(}d1TBPY2rgA-heRXT;(xHq)AjI}Z^<9w%erQt7Bb(4}W zu>o;C>_O(R`Y%Ty{t((6#Rs-U+=UphrHRbqpn|v4?xlGBY7bw-hw&yB2;2#W8KL#Q z7culhQ3b=ArxXchFao=rAO25!uE6AahvJYbq+t<8GWn(IcQUWHPwc~Gn{klv)fwex z!fzJvT;hE*1m^gpxtt;D>mCas93rp@W+h4Oe4wy<(5IHK z`6r#0sJNs%#jHf`30tA`u`Z!GTnjLP%apcfx#fFM=QI(?ME*nLZ_O_VEF=eHuDF@+ zPLKx4x)XT?gO#*SnZm$RJkH*qb9Dg@3ZW)TJ62(WRLtSh#e4`>LA|cgpPacq$CU}# ze+Dk>z$;}eIFM;K|6{S^F{Vx^*uz4COG~WPkq=d-@ZcCW|Lpd4?9;p?zF+#t z7{hh*w{9V@VYjNCVz97l4MHXByR=o40gpLEx$iNRdyKKAIFxUpzR@<+t%I-zkRlpE z(?@9L1zcy^SvI#&Q-h%ZjFo4QS`r`~+b;ge^!&VDVMpyiZKc7NN&k0Iw{$2d38*@#D)&caMa6FFy#K5T1*#GF%L{vj z$OW%uK&xK|J@URVHkfckO{2<2apN1^o-KJMPFtR9IFvB2n>SxC)_w!ISMdW{Uv7iD z4f83GTu%8YOK*f9xkq7jV{WEr2cbc#q8CB-IVcWe#(O&@h@I)PXLZeF$(4H;H-EDr zcB>&93z<}iS-v2MX;E!L%qV5*n}M@nN$KDF!~Xo9?UN<-!6&DqT(OdduBx=M=IXiK znmJIuV6TAOGD7|R5O6wAnr>4h!U=7HV&|F5vM%6Pv&^4+RvRzy9?>*w9=rCg!t%d` zu-Z>02*uJptmfP%)fFp&&VnnwUJdZ-W2!MTw<+V-Pz98qn6u)_mbBkQRBAp=<=Qk=6%n0vE$9rREoft#1PO*2r8CcyZ`%1 zg`vZ#=P~4ny!7z@0mMK%zX}YCed2^V^tD5`S| zshQA#sZ)qu839kFoy$A#MK>!W@?)!~xQZ;GgHmpS#Lqk0~Gd^qR7J zh!TI0+}vD&C{vpxA5)Am>veZ0KNlr2&cIA4$gfYw|th_!Bu>$*JE#XRV?iai3>En zkI-Eq9)9z?f(mb@*}*mP=7S78z)%0dU-7wrKD_yjz7P8FCoTXF;^+J9U-^Hll&UB_TeV=q@@Sf>Rf1mWz|7xFR^L}!34^4LnZbpq5B{Xg5 zk8ktypZpdV^$Gs>kA3evtAA(;aW;+Ty|bbQA9gr8d^3C7xBcwLwqXB(pZ?T6#N7jk zTp~0bW&zbq&THU0F@B5R`yX!c&pvTC@gy?%N64k6J9_>#(GFUKvO3T1^j^~H33U7- zkN@hM8^V8ADr(fglt6s>=i3+Fg}8eN-HhhkcV5NjExQY7x{H#|BcTx5jE@n-=nihy z8h7f5a9Quu+r#AD{ZNc%kqGl0?ebgnw?D7VkC9)A#-t?=y$K8#R2d3O(K_y9&k=So;^YN3<^M*NQXw=beO0q;h`ypCd~1BR1! z;&JXFp++x765=8*8on*nm@pydY5+PZG>g8b0illet|FS0k{r3;N3$n%J47;OMjj$% zs0LS9jFs&>EO4#iT$q@ZmFG)#q+UTL?yk@Hrt(Si(A2ShPc>+zfNL;aK-d5sX6wp~ z^$D{Rtjs_zRn$X!0xS_Rkb;kEEP?hj%z3nTh=HF^N!Zo&oF^041uAzmPqEZcCgadB zv?iwoG&{^1Oa>UCxFtSR^CGl8W1LK!?bL!|O&^8g}-SxvGcf{T&ht5N7@hM?%g zXVnyNjWBEisx?Urj*`H;`u@1dSe6~W3ZT1DK%ES8EvF{`K#Evc5)8E9$I z?QANq%4psYx)Yj%Z$X(eZV%An88?>^7ch&pyb8g2!N>5@{fk8MaGr5jixr)Z6-AIL zmS9{rMC}RlV{j4YJ-z{_fwG$+euiuMdc2v|^IDLa;LRRYM;Tjm1e_5L*0@lReviDX zoZ2`qw6h**(}aDg#(3T}ND+!QSP2ZP`|vRk;(!Y;;KDvyI-K7jX6VqNc1nAA{9UD^ z;*9RxbubvV85=f~yd{rc!Sat0<{cq!CqD237aW+crmGYnBa-lOsM|nWoCL}!wd9R~ z5LV=VffjI~M^gvP@y^$BHfw5%Sq5+e&iAQ9~i&iUVyx>Y`ShPg5gpMzcf4x}X{710@%*8#o(rix&vZ z4%ryP<_7=vfBciz{9OLmuedh&=x=?UA?$%X!o2x9J`Qw?FOzq7k+$~|noT`FHXB0d z(IGJGnu?K#2^lgXBSxX*iM7>QVt9^-$h!FGT**c!)!})?gfTh zS8(nFjCVZ$>{I`qfA=r{{$9_Y=}c$(N2xP|_e^Jct9tZzXY6hr5nM-#hPR%yxy9#x zTm>TiwG?|F3__=YHd@F=+p> zNW^2#A-qF#TNGSvg??wZH$KO+z_UF1SOa&@SP!*?WOsYZ=O23+`|$5Rc8&l3lmCg$ z?L?a%qU=6EoZlq0E6iu41jc@eZqAcRBAt8<(=YPbUwz}zv%hOO>ya`{1tRq#cU&J( z@?C^>pXS`tuP$CI9a=7uhkFnY(FwB)H_)*qrx|J3BhHTN{ksB-4MEok522Ru;d$DF zYhd7K8`98YWd!s1MaG81DM5HjoPQnl=g?d*X7qQC7*8(HufIZQzwo+vhF9vdAG^t; zAO9oz)d{8l@6h}RdGn))U*nn{(L~WLqR-;|3p5v>C9N)!wg(J%zJaAoICzBS@JUKP zBaIg+VYQDe4XnLI2pv&xmsfw=!ooU0~G-nDFReg{9_ z;JYPo02E4p0KUijCy9qMW*5FeS)U{IhvZxv;1NYr!G(P+j^trR+8&U%`(O(!tkC`x z2`v;4{^jGT4EsoU65|&1i>dHOa4wU^B_>9mTafgj9-V4oaK>UKMOf1!4I8}=3RJ_!IF0dTYqKNbe2C~BWbu}i^eV88!ja2roP1$fss2OSWlO>|~ znc#y|19PG$h$t>dHBUShSK+ND9W7VBu$DUF91O{VX+-=E{D9?xu~(CdHV2!L#}*ed zNTygJNQdb~a!NHV=^f$)5r>;?aNgl#MhC4U{b6Frj=bdOxEFhqYOuPZbe+-eE90u- z9f;t9;Txso34LIKh*WGw1VJYBLsQd$u`+C3G;Hhe(NTIsLq!C*rY3ir`76xvR6S;i zKs%YdJ(Iy)4e7EoOxjx3#|}%7^N5XB6Q4l{djlcX2)Q&vElO(y7ebA@T21t2aTTHA z0wvYw*d9&=Z6#w$HEI>aPp`!UXAbd^lJ^L6ru3o_X0>e0X+}dJGzLCW#(--QE*O}{ zxeoW~fDu*P?n$o=JsPa9wgapEK5JG|6wMDnDxs0%z(47+4UNdnm@#Sk${ zakwJbFqO>(MLc!8xysYcMJQXt=_%5Vb%A4;VYY|&w^4Z(lM}|xd7>Xcdn|YO#(;w6 zgjzLcB^n{t1u*9m&cW`^q8@|jloa=b4v>N-$YY?42gG^8%{->nWEa;~b6sYr6)bPT zpW=#8cFnX%nIR`kH$Y1n4zT1%+Xh_3cb@Lx6xWVe4w%chFhXf7I;2d+Y0_e)p$v0; zcTBVQ6nXauIxHyt0a_ATQZ09DU;7%=w_8|WMEN?wUnA%Lh$CY5Akv(mJ=j>yM+=H; zMtn0=)QGF+UQu5Q4RcG{%+R9vFp@@x=39u{;=GcE^El?vxhd(Jr*p~xl28R{K(`6$ z9ky^&@)mUHC_}}0XqmIQ{sFuNBzDvJ8Z>W-%_-yd0sPFJ@g73nf*cmU*bJr3XG`r!9I_Sby)lNG_k_4gB&&(qGn z!?=9{85i^?@1og%8q~;}eXtciUWdkGc^`Fxk3BYCfN~X)uhKQo)8D?Ep}&l-ZZMvF z6WpKig&#}6o#{{2&wO%&_vZ=C@V3p}x;+D0TwO@+-dUP8v}>YUuuF?~bo+NAD;J63 z06+iE+rF*L@8C`zVssaICkF3_KlLEKeJ^o#k;SF2BH{J%ZVqF1{<&J>DDOdUZFux| z|KGnC(|2b&)0zIO(V4+}rZc^ne*QNvk=-Mt?PcP@KHfD1K+6Dn%7;IB6-{6Np1}G) z_nQx5?N87iyob1do#oM!FRfC)7?Nyx0UH~(w;w>lhZt|SJo@oJ`a9*Z{%uRf0qe#( z8u<=nQzNtt-@7UQ_+uk5zTxlu>tB7AM?bM*ck2a0cOQ9s4smxQKA>8#vPH+I7>DO* zny>P2e)-n-{Jj3=b(5qgG@Em*9w?)bU znwMawFF`N!?i|5&`27{BKZoj!GPL-1K#QlHy+D#3B^@DUN$hTdEitRT$Ke?|qbF(Cjhxz~jAOJ~3K~%$qZxiDR(@WqM#8xY_ zby#9#kU7c;IgkUsXVpzv& ziuzK`3KZ7}ZIkIw1gDFd6b%NUkW;{l!$+&dLc!P9r=C(=E!ioKlnu-XQ7az8jkPSP z)b~j&Obb{YCdZUA5Tha*C^jIa;#?~8Ga}BDH+$%=!8N7g4w{IMhIj=PeDtVR)2|x! z)kp&YUtfc&AZ|dUn%m+F=BgRso#KErI`Y^;xf?69YJ}?=W&LYzJc(PNju>lN=@lHPjz;uf;Bf`6MRQSF=WHFg)kd&LuA}l7OHE( z&kRe|(09%dSGl%k0vOJD;;bf?^G?V+2#cEB4Q<8y+E`P?LN#$6)d2RPrf737NH)S` zu<9t3p}~p*iQKO#+Tuf90Ptahxe@0x-e*cbgS@1uU>54gwmKvU>S~daWx==LeJ#Ly z@h$PG8k3<_a;k=RDWIjnTOo8?a+@hRV3rY-(ocwkrjU2_{SfmLd!h%)mC5Teu3e+K z;Nnf%c_#NqVt}wROSCD=rM^i~F@35x*tO zw3_9~Fbz}bT< z=sDd`bVthbR}N6UWfhxYx~|9}dG)p0Tf7&PeT0Yk=x@8qHqJZFH!tvi{L)+Rop{SK z!F_5`3oYNkTzx%tvmL!}QM$K%+n;<7?;F}Y~4VVa0mgMEVX#W6tb1y&h z$*=JH|MF{ZDIxeu{oL;yV)`Sr=ibBo=-b5Q>Fe^Pq#0S<^-bjF{g^(&_IS%{Es(vF`Z${LEfokALh0_!EBmUwwtVyNoMK%v#V92^-4r z0>A%19m6|FWBgqx57*J{O=9;!y0IZU1v@}|i(jl?1%b2Tk+P(u14L%{?idML6IPP( zK2dT*Ngck~qQe5yITC9Lj7TE&Gt%}D@ptNYW4mIuxJo+RQO0wO+Xa}C#|If#7eLmO z;XG-$h2sUJSrYhBzIf#!KKhAgIcdJh=ilNNk;i}OdEo!zBcHw-9o~(&OL$u%Ia2z9 z<^daTWBFND%{Te{FAjY8(;be2#6p~Zhm_0(SKS16c5!hlfrQgVy8}RLQNJda$A$C25#n|@ zmvCZOY8X!*LdQd-UE*6KbX%ImG2?M%vNnslcw^%QO8*1wZZBw-Pcv>VL%v4SeuqND z##pP*oMCo^F~i9Qnc$yl0aK~TT_E?K+{fu>H4SQjriesvQp?9)9J|L)304)FrQl4+ zqp#f6evVBD3GWJKfwFZk6@>Y=x{`*gt?6#81w=KZIi=BLVi4B``rwsPfCAENYav=b zm`s9G%5WYlcjMa^jYLJlh${{k3MDyAl(@VDdVnFQg7+CL5LzW|g`9;Lt2t>HF4#+C zXeU;UMlnoMZJQUxtny*uLVf?lBhu6qudj)>+&9=z7aochWa()lgr?N=t^wlfqQj~I zBEh?9G+l63F$#HGnZPbgiCL@Lt0sEu*uRgEc7Q?*f>^*cN?8R2hi{b_AgfRkDWhc}%I>Q933Z!t}^CabzR-wx(7~5zOicznPNzY5(Cl#5RO@ zcV%1qgt#8(>K%$vF$iOe_?p~wO$}As2xTjj;>lxu;2}cOfV7T^K~=;-=>us1$ha8L z>?oM`uV7deA>jIe60MaZC-|!r{b7dk5V70GcLzvpDAzC4 z&*ssg={9-zCM7+I4S#{O1BD)^H<0igpZT>nF6QbsH^|K~rT+mO&%6p(P;x|f7tnNw zX$uxX68Ig;j?&N3;VJxbhLwXif32^SI{M?w>`vZKDMy55gO3kksub%v>5uv7Z$C$q zzBXO=et-1na_Cl{vgY;S*Zuf6*sE?y?KTiWI3&2R7u#iQpayLVAe@8!(k zJ=2-a^tY%pgZE5ldQF+|=b!#D;_L@$7x&QaedjfcvJ_!BzJP1qhpuLPrM#Kl`r zFOS>3nhr`g`RuR$?YG;0?q8h3n~doHL(u1bW8h~#`AtfHfz{22nIAq2(&6InH5)!m zXB3uLsiVp++^*@7w2u$9F}#ki*E~ud=k@;l2z0pE!>%K5E}_Gcl6w4XhLZ!D`5lJtTz!kL+NR&V%7;HO zr%wmWbQ(=>cKq-&cV1p(`6b}f1IqMeLV8s1V8aQ!;UV&Hfo5?NKRd0B+kPtMK`?2s z+@eL%!C`hvoL$AaNGf~e-9^NYh&N28w)UJbx5N2U)J_q1iVN4Va)oRwbU2S@Xx-N+ zDU&x3;M@Y|A|fLa3KEWSlF?y-YQXzDG|Q`qQ}Vc=wM&s9a&C* z7l1v`hj5_-Uutsd6$>fb)m>=m@NGxC95EjdTTsS%H7F$De8h!>#Vx77L>`uu@dy|G z9NZno-4$$mKghcfcOETm9S_gvD7Hr#Zi5T>;1R3o)wB)Oh(hp9#R34NokN_%xAonh z?6B&xO^jnLr53A&2m#90)#RqHX*|u4`ihS6T^*yBiV&@3I11w3WKh=Rt`8L|s|CvF zretFe4s2{GuEn%q8i50(EjR)yf=E{(u?EgirczncZ5={$TjFn(U#`mrBgsv<;jWd}D|M^VR&T&0P~(!fN_hsOGDh$5d;=*MW16 zqBW)Ly<#S$O+aFOOfN95g>fad2rNzvU^ndslgZ!jpxKAGtiPkVZbD7svd4;I;_*#j+{MZ~R}fc=sBjLUZa1Vybilg-0a7|&i>_F} zTnU;HOX%o8B0dPEC_2C}M9ezO1S<~JYMhFz$5dDp!ZO!H^`4S@hpE+4HmUoV6{!h! zXOMQ3VFNTZao-el>@Xs*N9nJib_7J(F7v86MU6>H0?qg+9+2DR8v^CKhyFgeDnC>V;MDqf2 zVAuw9*h6Api}B(bWsBt#WFQasQ}UcJe}QqgM#ech)bqHAP{x)rb`|eZ1n+jl#ZAhc zD@X~LM1&3I1RZ7=3&>aMYmhJDnkCT}N*NhXm-tyi!iKcINO1{lMm>8}3{M{S>LQKw zI4$_5#{bPcngceTGQ0E*%x8v^`@pW@;#KgTlHOjz^E1D)0{#L_c=Uf;ke4m+dxY5m zy80*d>-P}ab2w{I87bp!qCJ5u{)Er`;XCegzxuqwKEmGr#?p{QfV0e@xBCf1&V^-|A6|m8<BohcP?6fpsg?y^rxu=EI*j=CdDrYn-S@Kfa`ddx*0c&ABJu^7E^Y`wZ4U#MVE+ zKSYD~vBw<$`Tp~SvP4otA<@X1aQ@aC@~T9>$+)_Lmj{Wnufl5}0sjB8w7YJTty{y+BK>}QhfJoEd#?>Q$Ta{0EZ>e_nW*dklf%xD`}V;Gz&Ph|VD2 zN0JyKKWgWuJFwTYy*~NSa=kXj5f!J zKZfK<;~Xmq(F6SadNm*b*PLKQ8P|^^Wk6|gO~QpA#f2F;p3t9cF!@oC3xvhjC^2I3 ze4TZdLf*{DtBa`Lquq{(Yfx(`III{Y3&jFCFRSS+LTCpp?NIsyOgxq!L)~L&=uqcs zQJMpY#KExGlE!mj)$}dJAPFTk;Dyi`+Emu6B||+pZ@5M%xu!d<+-_tCA%R=fBBi*5 z5^BiaM~rb|JT{OaXT!UMP!q%z0YN=~Ci~bGq*QjU574h1Wi0rvCTG<p?Bny`40;h-VMmXZRn#8`p^Qtv1OkP~@yARg~CnS@1$A^1Qw-=bkExM27OhH*+T z7M)D1T3RMPVK$Q2p|VP~Lz@%G4WgN1iL_c`@d)i+#&2ImHWwg1LP~F8$ybA2M$GP; zOokq_0o0>D;u}MpLo`+H>ZbA#q_I0ZT@CfMDiNfmdd~(zK|>v_ucoWq!8llvX&6HA z^MWN9Hwp<46hge~qQysP0n@qs6{XYMEFtreO@ z0e2Q`5A|!bJtnS%9OrO+25kn4EpV+XYnfcyxM`EZ zC}y~zi1(;x+62&=E*`18YcWy|)B06HHzHx6j16Y>^+mLri)EL*apc&N)3TcF?FO>o z0^mn9WX9vIni5tc4}lb?Yc7BpD%GIR5nNkS;yEeK7tBw5)KpFTw#~4}3`Y%lgq%~o zFSa942)07nTZ}g^W14ZZ0n!EH>I`vp9_MC+`Tn$#vP)PTO=W4@l(<0KhNiR1wJxQm zZ@XNv8V-^d7=;`cjO)Ps+>MHEDJ`a3H1ivXjHtc^^O2GQX|s)Q3zj`)95L^Zv_WQX zq1_J1Swv6B!{wULEdkejbJIbZLP-Vj_1bGNRHrEb9kES78wHw**GPhBkLymbv6hSF zu`WRI9XXxFhBYo9GHwLt!L_fG#s!udnr4ApD9z#)-MMcNk3LF&be23W>iwaeP16n; zH&3Xn$%CWxj@%15MXVg++au8HIB}3ilCAN!LmB5#E)#Sh^)T8_EnO>G_k4?CrJ>lG zEHiBNFqU`f?~&#T=ms12$T1;yh17fcH!V^E+om)PuIW)9 z!S+G6>hCuN!?4nlR)Vx^tke++X|Et=)!Z%DQGE@XBjjL<#(zjosu*A31h^eIH~=bM z=u^Mh!0StR;Dj&y++UBum)Czq<}zb=8g15e%Om>3C&>8$d^nGW6I^?3;>H_>Fdkea zmKnkA;udF#u|c9SuD59B9~^jyW5f8?$B^&<(ln@@Q209C?hEfg7AW6fy?>tJ_-V>v z$FrYPe&^G7|GnmeB(hIQC*+eY^e%%iceLK7lw(Bh9ix}ajFuM1heJ<@>rvlwGIo4B zq}!=Zb*fYSHR;sgJ=LkcRiFB$8SC~by7|M*?)wV4_?(|UHEuGzyix9)jY+Rs*qS6p+BPyN!2FZ|*?w&DIQ>F1vhT+t`^&pb-k?%~}6 z!$ZlWv}QQE#ZUeEtCaY{dluMz@#kOTr$755vDspM^+{$IzK*<)xd>A;@vX-hSBIqG zMgF&c`8xco@8%hPXL{!OIj)??bc>`7jz}ykmg(yE3~%`hKb!f~FMWw#g?9dru<$gc zDdc#7hOvIBh@O(?#PwxLY@uXC1lU8gtkftC~3*@)%yvZ;zNUVN2nhW6e$VEH`j*x#O%y#h@JpXQEn!F>er6H zBYyU6J@fn=I0wv8E-;=vfoGd;dxf9vqutk0zRuX+&v^78md=pVl5yQ3vPSee$PIEn z#!>_R7%TTxj;tRE^Vb;N7HQaJI9}psJ-+KfHuXC8=P2n0BFcDthP2+pHeV;%AEM=T zhW#HPEFZ+|qlnvv)Yg>TurqNw*C1=!#VwHOnhB)MHf7vFU4!}sIVRL@;@f?Ew@*o1 z=*f>z`p3!9lak<@eJm~**Jo)K2lV?7fSe`F9O?ri3X5*` zM;fIXKU(mb2{Xfzq5<+Kqzr8{nZS8U(`u~lr>m*TGA^NHR956MQpO|HeS`MwUooCs z#JMGIc?C|k$Qw_J1>YDhG)PcNO5ZeM{Wuw0ime?f8dJ2iCVI7|o{b4-QLGxY0fd_9 zye%`ZG=V0ZkP|qyiN{$gUOCmrNt^l_xE8vt`uH|+eqq8lXhkr%9)?3moBBTR?eufC zr?2v5r^B`cVOCKnZZQ%(HhkE|SLjIlLCb)f{u7 z8qOuxKR(<>X;?*c^nz>ariGiC??ZDo3-) zb(B0pvIcP(5hIQ@DIU6>FzZP}hb61`!*CTOAhJeVVL1A}sjzQNn7>KvFHT5`mbhuj z;~Wj1ZmT|*x>8f@=$*BRGo013K3N5<6ymC(+3Fz$O5TES3+rbT^UQX>#`@+0OOGS- zb!DJ8ClwpyZdDdx{34PEhbo{Mbs6wA`cEJam3`|%!MTyq zq;j@5XQ*$R2Imu&P6(YLP0zSeY&<~xrfv?5tC|3o3BMtSh+L#_!1~RnDdm1tLVfMW zLQGoI+&ZiGKsS=hgr(^t5{{>hzKH1&vz>j`H||4L3!>jb*dlIf^4l~qt{ZadYH4A} zxR7zOrq}>EFsv`4-eK-0687=khT+HqFXNYgK)ZbN&c?k9#|(#p>4S)DQOb;BSL+6i z%VjeMl4UgH%(S z_n!XzApPFIIpUe;|CIiq#q3GO+q zg*Pdyn+(mxyPkA}nrt4{7fJoAM0fA0Y1c}fBncl54^gKW6g~Qo#m=WX)u~Q(?>aSj zPj#wq)v)srv&AL+@e$v__$u$ZCUhclS8Jpn%>ot*>U z`v&jN{KhWV#vh|wK8))w;G0$b;-H&aA~c?(j1QBB2N(|?;FG`dC4TE)e+v^$jPaNM zgpxbpNpg9bX6p*G?OX5okg5oqlQRta50KY4=#Rg~eD6>0nUVZArGNhK7If>6qV2Y0yv|Ji@~@R|_F?sfU)=EV=l&yHUZdeM+I$4`BI3f7Q@rcPFwt3^@yB$#`{cOB>>+4sYH+;yGy3bV)$ksbV=KF#36tgF$y8npQpy(M8z_0=JZ7|;%%f2;g`lO5m4^v|Fgj>THG7I-qazR1Om?jRT5$%W2bU;g zq%ac4g!7PwGnGZ_>wCehnv!`_$o&Ah;Jcxp8f&UmRlouz*48oaDlT~7sN?orx0%J}4=0PdZp~+7s;(+*(lB=od!blk#;;NaFz#z`!yw+r|B}9~x0(s2X zkdU@8^mDBA7zvX(E-0=kxTdD&#c3_8BQr`tB2X#=TF}ZlX1aC>%2<&Taa5ek_;85$ z731oD(il;ioNUO$j5N$q zr?iWV(SlXxc9P6R1$(zb#O);XON6G@q0;eaGG{c1f*IXr!_mMVp zGz?JIh~L22Q1TAqI?zM1ZQ&asv%%K$x~7^$iGC_?G<-<4SkMZJV#QA_dIcBK#9G$r zdd*2=rrX{p4GqW|zdQ!Xwi%0}I425$kX| zFgyrTGN!$VOQQR!=0ofm{5;fgi7(QP*FV%}fWHlW@fW_q zPkv62gBfoAajbigu=OTseLv9@sOvlro?_Q9dzr<~*YI1%H0=@aQP#I$-1Hn@dz7v7 zFTbbBJ070F>^!dR@XaxCy+z?=x~(hkdu(BG!aTlCDGw8GeU$q7>O1?~{ljm*&X*o~ z9XDKLI66zaeRqY# zAERt8GweSMVT1FvJ)Ahh;xWVO>zKXB>LdS!fB#QE+;#qL>nDHpB0Byk?U{#h?JmB% ziM0E8m&nOc@`AK}65Tw&cyf`S{M?`NyPtmbZ~L&Pl=_!==JR{ls91Ycxt*rqM$`?^ zKZX;9GNafANQfp(3SJM$V~dqFl6!L6B8@wU1Y9@Ze8jSY+~fQ$G`zFqLUb~6y4vua zj`u)TwJaxgP*%{rcQHmdbDh|q$2Sk~6Q4W7fB5vhr)-}6+!G-GfOhX8d^;yBZc$G5 z$lVNW4scDv%9EsFMvgZLTW_EmkzzEL`NIFy!!N=g{L=*X63>42b;R95 znz4=_yZ7MInUdq2Je&nx5auf=fs$sFGRN|WrArJ)2M`_NXZYDGaQrxVc#vWL4BeTV zSV@el`w&|oVfqqI0`$&^d+J-vBIZ^Vbm}&k%>ctJAihD8Qe<<-+@=04-W@W=9Vo{n z*+%1>B!YSvj{-WOKGH_{@ z!B^wnGQ>d$9W+}7=j-Fu*M1^B$n9dXX%BGX?D`ob0~^8ZOa`#kr7Hb5*P?S{eZSir zEZ{mQqod?GmUk;F*M}O^luRBoISZ;v=nCYTa&;~EZZeg9J$KQG13YLAHa13zg%ay) zXgSxiM61PSIaV_rQA(-ixbuaSK?jd(GDxPxnmQI$TxW!KtjBAmn&%YqSQw5xr8vfw zQpok*>X#L<;p>{q(Wca>;M!V}=BBhR&WLMI+!*+b_jOXk)F4M}C|F6T7LaS?O(w@T z$?*v3u7Ru|Cu}@qG@#!h%+_>^tEBM}EP2p1#z;wX9Et?N)FDMND-|1oqE1NBqj^Ng zn1DzD0vZyOmULXp64SUq%6Zb*kVe5D-lE%moz>OL_3!#6etw8bB8y^qh6YH(5}^eO zB*jXZ%Hy;a#%aM%H~?f06!3yFE&P8JJPzK!WEt>hj z`sR~Z*~QO=uy_Rt*RXL$$vgEuhoWv#neB9l>+!y)r~+-h&f>@^k;e{bureaH!FZJZ zrZ94=Zrbl$r(ifDZttta^+(b2F-krFyMm~Pv?NPFYzeZ34%?u8eZw*!dJESoRx)XG zoo4pO%+7u7&b>cv_K0I-SRYenPxH*@F5zs65fR-duV3Jc|N7N>llUIf?|%AKp8a31 zBiWJ16?Sp~kvTXw;cNy>%!=h}nDg zqP$)G{so4k`&b`;nVv6ysOKGZAD50uhr2Yhdka?NA&_z)4+s3#-;oK+Q=RHme=l@u z@Sf^a@75L~EpiX(-P;Q73F>F{!{%Mc%2*rAKdeDp0#+h9e{Zv$KK}cj^z|QOw*3)i zJ4du<-+%YtwUK4}G8=yp_z3;=o=^Pp@4pMz@R?tTz?XRD`Ip%oK8omhM3=QqJSME1 zfLz7-m-(&#;qG?(|2FiAe|sOD`>*iZ-$$73Ge7%h@8Am8yYvBN?+wQN^9=rJIBa<4 zdAnzR>~~9IFOkLI&_5XZy?{Yq2t;O zcK8G(y#%XkSQ!a>Z((Uk?$0o|b9Ee5V@;VB1!?fjn(S{P9OJrM)f|iGiR(EwUI%~m z9f|c%{Lc+#F+keK#+q~s^I^J_LP-lsnN^dc-vir5gJE$Cp%&&)8pOA#zeGu%5(BE| zFbaNNxxf~wiR$gQ;Aj^i*OBBb6asZs=gPj#31maNeU;p|q}7bDh?vc=G((+WzCfD2 zJx9t+LF?wGl^4MT)l))gW3)qoXWXk3{L zIETp=Y283hi29n!?J}+%Fm+%Kzf9O5#K99co-|e#?`_7gO{V2-;}C-~3L*h38Ifb$ ztl}I*49Z9x9Zm(&HRJJ;QWT-k1i@M%h@h=iwrSSNLM=wg;Iz~cZH(d~Dh_c%UJG$; zG+Wa>#uPs{(gw;1L#pLoc?i`|FNtxrOWHVc^k7X*ZFY+Kh|7Ue8pJzP7UX!LM!9Vb z;Z2+kQ0iEE9{p4ZmQk(97UyzBFk~Z-O+~AOdhJvV35LYV22PR5*q%@uStir0kOx7{ z=(b~xz807UZ#ZL79|J70@@bBVj%Se>lHtXxJEG1S{0?Wq)fD8A!?CcEtnIk5gpq4 z|4I?aep=6!A1%(1Q3(s1khrcswk%*6NHQSnj$#f9Xp?YFuElEotn#2;LDUgfbJ8#w zk}Z@}nZ~+Md{A&!u@I99O_1reG2&Y4d)Evf2CRY9JJL#_t7Om`pYKUIqCTQpuzZ5q zUm)Q%w)b8k#TlgpR1LqE=npTGH$O%Rm$7gI&1*8-SQ$Yak_%ZP!D%h;1dxJ>QcUXg zNj9-B;5NZWCPYZiT`ihy6RMuV0&;_;EjGt3{qYgLIU$WFke6UBY4BicjEq=Eu?Z1T zjA8@oH;8LUz1Gq)xP!(35Y2THD;px9B6rqj8Eb0a`2=H2Ds_|3hoK^2w17rjtBk81 zhU5D{E>TLs&jU`jXy)K%4%LK&6=m3^l!nkK(rqZQCH5Y{2#XQPQVYpShghN{p}BOO z&GiRxZ3{Xh%x>Xk$AsB+%!K|>;YiVDAF~bO>g0zD4cZz_h@dOR<0X2 zr2)GiD|5_lGM?-aS8p*p{{qg(I|=Gkj-VXjW-lQ=P2&t5CGC)g$0_A3&YwWoAPlJX zh#KSZQ{=qD?|!pbHT4~vTThY>W_bS~+HKMJ9fXD~1<4z5H~6W~y-5DG|C|5)U*21R z@^@R$Jbw=I4-vLI+U2T#=GBoRrQTmLoP(6-l>7vngE{Ti3%D@Q?OZ1xUxxNo@{OD1 z?FY!?V>mH%c7$pqT0>fY3_rggD_;S-M%ey3zWw$k#p(>dxQX_AxN;u&&M{cO^)IgR zGr#^P3`G!og1FhEUA~21e$$jL2x+|}9bRVam3Z)%#QewnUmtSnZSAhotZtFJ@2A)m zRPH_`G8|nX4>!sF!xieC>QtvX)rZun!F#Gxy`JwqO6=V6(oTwqJT2Sv*X;NFQkKzFo83x5#NmGdxT_d=mKc z_x!oa_qxGz&nf@!Vn$dr#MQ`;uQojQ-1lDU;_dp>Z#Kk(A0cc%Mc96e+3x#~kmAh`3CbU*3jF%IJ6^-F@A;-=3f4mAnSTs0yKyg>d!z1K=fi`PYprp#PMcQdZ zdkYCOGzel!N>k#`MtpdUcJHq!X--T}Q1YYT!D&yseMrB*M?5)0PUkRbz~yRBNt`&Z zwb9CK1UC@cW7PFn$>c$a>lU;_U0d7H5sVf~Eu}beKWFUcC^yjdI=;!M9^>p~(t4M; zdIF;+GW~LZsHEZ;kF;(IWT_+YlhJzHuq(NmkIix*^&S+$d|k(=hmIJ9Ja~{1Z6d@9 zWUXXRkK4*f2XS@V*tI87 zBaEAXP^v+wfEJtw)GD($RkI+E1&h@n3tI_@nD&qbv6j;NDLop&4Ony_hETBBA!;~R zUzeH=uO=!|eMYGPYV`JA>)XtKan<}TxthEs*Mx2C>%Y~R;=6)|h`3TWUJVJs&_R-THIv(SF>4seZEwY)HP+S6AD43PB$la50lD&=U0V=2B=Z8z;*V&T*Xlj1J9x*v*EI(m zzV;ZDC26${=CJIsyai=WDJ_WL+(4lgm#K@j$a6}j3QDQ8lA2`BzP=BpO%QeYj=`!b z6sNf0aN!&!ohSCkh#XaOcbsFn zt>=1sqs3>$B{ZFbUuEz2wD}@t&8J~O z#NQPkt)gqai|huDU~##&m-en9qJ9mwD+sa}3arBD6y3QdqsW{xYb^U^c)XESXU|WX zKV9JPwZn8t*HaEsTfXy<3rPuhg_sMyzWQ9x8+4pT;L}`oT1NCN}sH z;ZBjC!Z7^8>RpMN>xAB6wKTWyif(@uz1e|Yl+)7oIl6t*cz@{j2sIaOH~J?mc7O)P zCFjFIT#NhKr=$^i%lgOe?73FH(SPpn%ba-3^{rChGGD$^Ba-swCm4ch4>`8iyx=O-iT{wg2Uv$Jz{d}P;Hdy=@v9FjvW9%*VIG(96ZaaKO?mi6fOA;-4Gac zWFme#jvR#8aM;WO)lws-W>j(ppu8jv4G6 z)vik0h);(xgbp!7@9+rGS+l=9un4K3zF%Rs;Dyp z%DRxDWFa^6keBiDvY_8fn24qQ%#wt~9}Nxw-z@8!*rkvtg6t7zvDSE;MjP?U7Rz<3 z$rR~==SX?8Oa<+ivBtFK(|(SW8y1npQdIXw?wA2#ekcf%5Ah{k99q*sFy#}>askwW zF-ESDyNHbbO0ByV&yP{s3RJI=G1#C&VJRK4`?U`nlN|=+MeV7o%CZ(~0igZNwm2)) zn7bM4$o!!BKy8#v(AQx{5t8KTA%8TaYt#k8N6v4{y9?=*?TT>k9;}s^S+kd9euc?6 zg|w9KINSEHa#%=l7TM*{Aelupl0_!japYID>P9KsL_tLdK3kQVq^M?&lx{Z=baquu z>&@AQNbJKMs&hln;h>TQ5I%}w%5!8t+_fG}KLKe7F=HzV_ydtO< zoz-D15sec}fDw(Z*tlv$u0+B`u_{v11XzD}%oiJ&M(LxG`JX23XKC?Dr~@CLCj zU3i2`#v1LDaJR`t*WtM)(rQO27qd)!3D7d<`~2J@Rg*o%k7c#q100&?2ey$A6Jg`L1+MsS8QZd|BK853VN`;D--BBIZ@E3=Vf z!B=S$cwVrhfB|4k=keSXdV90le|?3r`3JsuYsqj%w?Qa7>A0UE?#+sNeHozpnmi0( zeEn^Z9LLH=#5easbp2%H;`ht#5%$*E|AR%}e;YQ(GjwMg@}w)zWBq!lbf4m+HC2CYfzM)L+@)%i%A*k8@(J;NS!y%lgByR-OWT`TSM<(NIMzGk(5 zIiZ|f!UoM!J9SKW{4HF2=R1K_*U7P_*;}ppTzWC1jRRg(^qLo*M(`2OW4~Xp3 z_}Z4sMKq?8=nCr7cUn9p{g9hOe{;lI@UT?vy>s?|Wn5=C14V0VgQJw(E-j^@cMn{| z{jZz_r~NKWZ(cC6bsjl{-Q6i4HHdb1#EU;nnYvt*hM%~r@}AyE3O2-7C;8WWFU6{3 zE$YRxIf}E*10O#@kM4>zM~9?+p-7zGm${05!mjO_k*B0oH)MHF`>fqa2G<)4|BWHV zKWHc1g@YP?**y|+-kKox9Y}@mBmG7-G^cw;E_dI9pPYiQ^p)409L=@`C(I_$z9Q`z zp}A6~=}ho}0gg&34I+2eA%j9AIXq_+mXWvC39BdiQQ<8Lb&%|a9ikQneg1Mcr=3{! z!%PM1zQ&*Ce(&M6`|-PnxS8j^v1WeCKaxZg%|)bFbFH=RVk-2;{;^jQ89C) z`UE3E6&wS)hgz}2z7h#V97(eEJ_oomMgXE^=WKo4WA=RK%oOP%XyuAwgm>`67Fo=s z^Aj)DNy)uZ@kpre*K~Pv>II(T2O9c{%W^-dcdv2!Pr^q$qs>mgY!N|7eM6GZj_SlC ztI-+DaJ18##K`6V_lll*`!KF4Pi~1#CV=6}7`#&|C^6EuFqv&6xfs^2Z|iE6l0Ux_ zH_(4Q%}Vs(p+u7l+f?%sU&A5R0lOs_`;22#(%6M0NYV`xoPxE&^Bz;M*sP5>C(LKY z1}BT))W!zPTv3iXj1lq)dbEHOY~t(c<1{6uZp_N^c#3We++Zl?HV+q<2f2 z0?Kj!E;)LZeenvf`Qb!drj$*e_=k*+ZHLNO0sD2^s!l5-0nx~bFF8R$COfSK;AG>t zxu}2xtX-pgMda%TnL6pK-6*K*EErTSQ%=pWk$uDSF8C>ZVxAAICtmzs%Hq1WMbR=; zs`3JeN=#M-6_z^r4*gApi+{mc{#nU0CId9k`y=(|8}MG9p;YbETHTJ40E@~QZq`sf zBLsd|rIUnrX`%Vc276)k-;CH&Cw1xqJ2DROPUS+4+x3v2QFf3r)zjud*l?{kPIBce z*6;0Epzl>u9aX1m#Tf~h#Uw+G>iTsYf8o0%n6Ur8AP^f_jQpz;QN&~^)`xTY`7fkx zw4}NkhH($F8nchWNHSs=B4;Ov4An{1NmXJj_r8*xQiV=sq#*0~kG$07Ik2vG3^3&n zV_Q?`hC|D$yoAE*8#A;eaX^~Wgk&T^6(q5mt2ug2VA@S?9(p6AP%0ZQi$938@T{kimzOam^{G1=4#`sDRPfU#-)iz zWrNmgmftsJy)^b_&TN}5J>ozJo*rlzMSP)MgM8kX$~ zm7zDU1zr1*x&*h7m?-LBIy%)(I=BE1`4b?VD#NOtv=`F(nZzw%i}AlG)hXkz-$IJ# z%U{~tA~Sj?gt~c#&axzj^P08<9CM{$HoPtBL`pq9zm#dV`8!J&CP+yq4G z8ua^ON5iwNfE$K#{~)*4Nt;8Iwp#=5%R#R4jn1g-+}Z`wo`C`{_U^`2wi{irWfoA+8{-1`rw4#}lRYdY?%Ww(F9I8^4Ma(B1BOxp&Of z!6KLT8jQ5p%O@*Sn$&qqGNDbaLVHzDmD(WEv#nt97O;0{s~QD{A~G^P9kwFq&U=RSiK;|?~O7}p6u?6HtaC1 z3xvw`=ze@CTfJ+Cb<|jCBsUvSdIH4W6XY*4xf*L#!bV$)!b#&>7q(Hw6I#=6cg~3( z5`h@g`sA(tY8mv)JZXJS_P=k0Wo>Kc_q1f1Fp3xUbIV4cw;|+`QH~IFwkt`h-9dm=(& z-{jtAK`g5vQH*1ss;loy5iAl2cI98;-i{ndJ$~8UZVj<&|Uk z8}6&Roln>BFvK{skVbLzeqe1!Y)Q;PfU}tQ6-NICaf9x_95v?#*;%J7f4ii=I^J;n z&iZ#BEiy+WssspT@LSQUUM!EVE!$1LFFc4%%>sGOJ}dMVo8{++8|tRe{7IB0G;#LS z4e?sfT|gjozaQXx+nvAmJ`F$K^9EYJphPKM=i#KRKfr18&-OTE5c0{n&%DZYs8+*c)c4~&(D?Gxr2k9R<;rc|sqP`)viD4DuXnX`r^-fev<=s?W42Pl zax;81oZNNkeNw}Q=&SjPz_nq{j{w&9voO+UT>l9ExC{Y>Z%A#ymmQ|7uHettFTVgs z-#(8wBD8j4zwN{4wIo^)71?#~ie626E3#dfeYdoU{uZqh*RXN8LQcyg!zmZfGC9{L zN%!Ao%^Y96Zl9bMj@p)|jc8I+Z~QlMqYnl!^q?*N$M|~gT%KOf5KzG0;RH0X)MV)7=!IX6XKu7Z6i%ON z*^C}X=Y5bLIiMtWCFr9xK~c%{~`7!zr}4LG#G zNoWqtNy#aNwr9#n)y6rjER8kQP}3yZeG6M)t+Nqh*g8SO~Dt8B?l?kWoq1_W$w4qyFkh%xX~;T~s)<0d~IYC~FT` zX?f(xYxuKn5Rx}wM+b5=7^fN#)=VcRVVP7S0a`@|cEqGyW#hm%s>eJYu>f}Ja9=e? z!pEpD75KWJK)4tYh%HEtT6G28NZ6AZRg}Gkd*u}N>tu!OEd3!$ZMN)X>z*lst^39i zyO{YpvjA~*$1a=&0Q8*1;6qq^XwsVtbHei6><<;Pi6JlI8Gx>so&xXTJ@(LFw3vY- zdR>UJmL`Z_J!AbgmjV+WZoz{k-Cme&3{w|LO|->}pXc$J~Jh81z$ z>SbDNXwCULN-cy4>vDQ;B@6aHPFqipDtMXz1sEzy7FE^879VG^BRf@DxfOkhK6m&X zF0QVLx4!yH`*(E9S9+Uzi8Oz)_E!O~S4^_38@9GuhX6#jKK^$B-Qg$Id$7;qGs)SxeJa@b9&P z$ZcdR*>ts-wJeZ(kGO@ohDwuunPf5J7{F=4gGa&I+BFYs5rB$LDagj?h#K^Y+lKPY zAy>tvwpm+pW^J(qku9D6AuX6tS?hgmcimy5&WGFR)xb>{tO)s0j(qrLm1}mC{u_9& ziyX#j4o)%_{HXW5xAaM#fnX9)Jgc6)5RN|z&JLKoVCHfc4h!)gDeHEKc&jH#MgG7_ zwg=>$Uwg;TGgEA&9&R>mH~R76orhi;7SbuzDU=z0#x188cI8I3OWeJNmiHcC4?}wP(A1m&PX5E_K7y2%5i@NV1Pin|N*fGo_ocEH*h{-=-TR(9O zUI=sk@GrMcxmI%@t#1o1w{HGs{a8p-^1tAnGs|=Jh%$Y>cURCl=ql>IrLue%_|`q8 z`EmMaX7=&$=@R{Ufz*vOOn7r8>huZl;gF{WJj$lH&7_Q;rpxprxp04GOn<(;G6uh9 zr|bim@#QY3^RZFjTbq7<3aVy9p$a=fZnbw{<=jf3_`Enf3!tg(p{Jsjg}LjDxbs}t zU8Uyv+5b*#o{N7oj!qtqoTN4jH+ZkkE15HF3U@}`lJEHXEdu#zg+7*Q?vLmsrg$H7 z{M816r*2_?&`(@zd;m9~IocYw!=CO~hZD}gRKPb$zz6-9RrR+D&;-oENIYPvB})$n z3kCl6`m}Kyz0=T(@pg&$zmU(ph^OpOk8P$&kD=X+9LCXIz_$7BDEx~iecgu=bHaaWTFAUF7PlsUFxt=aP~5esfC=euF0LQA;^2p2TVrUA3!&L;|LgQ6{; z?1|gj4;~s)Be(0dfoDLjH$JiUOLEyvjspSwxN3K81+A@nZBoyqhnI%vP|(ag?dXs;I9>M7Ff9o!&_WmP<| zRw=Lqe`+|3;;uEzLN(fXjs{$rVZF=PrW|sIlgJgSP0ATUG)B~{*lS>x7xCZO#T0`u z!yAHjW`0KCb&XoUOE^EO{mP>JO^LhRfRh)Q7iNZStfp+hTmGJzy>39x8CM9=7`3jF zTd<-qVFXhHA9P@a=|MDy$6^+=cz)U;s{RuhX!-~B(A$t8GTq86?Pmrf&nQ-YN@>_M za-Z>lH1$2eF^)}5*V;p6s4@7A?4SCDv@1`j_7!z%WCgFGxR^M34HN~;GDPWpchM4t zpboetaSGqA)}aO5fR?!-0ErS61hc8lwK;LtP5SPx?jW2G$lpgHX>3_PU{#gNiHiu% zVmsm73@t*R(Ph+J>gg)j^9z@R)5=#X(E9}B{AnfRzgT`p2+K$q&eF9ljkP>!!QE;k zVq1E0i?-pPkfal{z$~c&8!=6I$yDg97O@@|x_3{l0dQP;Q9k$U^xD1;W*UhA%kiHR zgqyKiU&Fpf+X`Cl_i9dr4v5&CSThq?9t5%xUSUSbN?3d8lJk}J@rW;6*_I!>WmZ=X zivEtKiJ~nFz^)pu$0MhsUXCqgxaS#z;$u>NB=;e@`luvO?9U54ji~2%JHs*kXQU+< zv`1aWZN|jv)0uXtVCAk~Vq>SqrDSxBo@|dE)IWgPww4;3ZN9g;TbM_JJ&V9B3UCrm zt%CvUR9fN-jFn^y6j`2rnKcbspgU*anXXGl_qVDgsQGwDC7VdCnNFaL4}aLlb@|E6 zl2YaR#60wo&Yr!)^hXXiX`-iJH$3f&kTA4`{*}T5Tsp9zXGitV2vi*YZPc;?5KT!Y zkT9AsCm=+Zbpa3Y4|&xFe~@&pP20p|88C z_0`kuH?t7>fax^1vk;IsMMnp0n}7rol4LgcBhUGO$Z;By8*@rVo_CJ7)PH zy#yy#^fz`~hy3#X`6lGO?OVTDZ+cl>)#pWb%;^oq;s%gP++rTpmTTn#ZfeOIE`=iZ z#2_7y4lv676O-((2!-z-9+wJy1ONKQIKE>|-aG}$Eh%>21|Kb7_8ExHaf96giu4jGb9V+KZ#%a>e#r*tLX=dHi5eVE1P zMbPdMoW$E-xwO1slB3dC6T2|5C7)06wOLHz%=VY3B+Tvq9issl%T5h4UNPEw| z74w^R!TkdouKqek4HoQ;_uHx1a%FjaU@s6HXmkfG%TtSg*AOV*DzHgQaa7bdAfLJ< z6HIX0usib`5s%}@YadC;>W0?vL>M7$T$w6k~*e-nuX)#Bz!`A5s(xVor@Nk0Yy zV@oDsE3T?Aty`FVF$GVtvQ(!0o)uLn3pwNpMVhs$>puwz#y-(9rxrYj@=;s&OD+%g z*Hf6Pk7PAdY{D64WXy5|5s09~W>kJ`A54Z{;#S(~(wb_WvFc{Er?czocM-9B4OFp< zGv6IUs?c$`n6)|{ywt6q7O{MOe>EDv=W#{Vp>ti6 z4Z{kAJxpQq-=GTG;CH_uHgBcrv0Wv@tjDaNY4}rP>5#2@U=?9HaNoGj(tL#e%|v?j zD5Y&WUmK#^SnQORN#mVt$o$B zu@^)<&S!~O0s7QSobi#P7B3ikc1u-kKu+>EP9%p|i&j;Qgv}O#$1XF)15Mj_1&jlw zWqZ}^Nx7ntznlJTBN*xBxO5*K3d4yNy1OP>v(%`D~HK!t){R{O5V_;0nJazSGL^4;$>+oTjV^H7vj%La7rjvEhn}D=&gFWT} zqo8TT?jd}O#1WRO7sJbkdUb$mPgaleyWo}%xbQa@I|o1N9(u_|&c0QAjyRbA444Vrzk00}tMML1if;Q^AlWi-zpOxaD z>kK;O>V1ON-9sLgb<>B!)>&uoo!`=zFR#`<5eyEu8Z^qvnA)+ksyk|it^E^o+!3PV zTFxTQZL*0=CSUu50~2fIlBiW~Oq)uDH51Xx^+ zJx95z_5nM&tH@&JtSs&wY4rx0TZy^@SkGHdRX*Jg`3NLkE%Zb8tDQzxm@j_-8)qh_ z;pf*RS$33lSxI&87mu0KZg?8g`TLeR`~g=28=z@RcYQ`KrgGC3XBb@nfCl}W@nhuJy>8OqhotCirqV z(VLEHsUBY~$|uCG=UbQnhcLwHd$wY1a~`317ocplFJko+HLSB+LK3zJe}kjTp3Fo# z;D}3S?-`LRqAhMho>@Moh_dX(KA^dNC0^*$!1QwyaJdL8A^zwATC#R|bK0WGjVQRiZ0l%!O) z=@uULO6W_jH}$T%=;ED1}0AJMiDa-ph zhTAzOAjYIfPLi}Z`@Bt85Z$v40~wM=X6G!W27UvWv+^k0(-6mn8;{cq*DF1kcs)t5 z%+PjDHz)!(z+0(K$z_|ioi4&`qlQgIM?*r}%5GYu3`?h@6agb&VT3`fEk>MP7WXVy zi3gDwWmLPK;9P18r#zTJsWD#agPv2t6P9Jyh@c0-2fs-OU4Uh9P6z~3@zG&5*t>KV zi=8tI=$BkrqYm)7q#{z)s>IbE4ODdH<(KT3eVx#=$}6XMkBBM6=*Mi_d#IMggv()b zt=a#<-lJ$$HaGn#jxD4I+Zx-^Zw#?t4QDRcjHp`dk2ZBUfWb+ze=(Q{Z#U#KkXrM! zp{q)Db8DiZcK-}cQ*QpybGg#Sob#M=&_;I;+vhFuSZGYmJpF=5rHR>kVn_}g5Gbjr zf+fE0z52Rr((E3ASY1)B$LyL|6eG0>eTBM=C#10;egVa&1?=}VEleK`8E>VTp>DBS z!mK?zZ3Tso@&A&&RMvbP6r@~-mMgg6)?E2lD#k(*22%}#OQZp909`~A+37*Vd|cUY zCg}KK>%%kHI?$4txTG1m-UH0$>+?GePPs;WA0f(rffNP1v zeSNDrnjQ0qrglT`>5Pv{kaNnYUjSi3syrt(GfTHJQoJ!kXn6smGytPxcZ zG^8D%B~YuT0A8b$;+z~69~hEzogIyv`M}6>vwY}|w1?!QOM_J>fwg7RnZ4)^Nf>>X zY^Zc_3M@w1`;~;1BR-6CH6SULTcbt5&Gzd&!8`m|ukVe_+ zPav`6U}U*6pHDEG3W$WmDx7K)92?LuaVe8f><2lw=c9!U5llbe#|eemSI@C%Jgulm zN-0!`f?{T>uj~0+t`J$y3Z9E%nhRR*bgsS(_8C2xS z>8W%FrqeUpx& z?3WwDSpx(lHU(lD^NPDq0XQKJGR>b$p>FnNZl7+QaW*6e9NlcM{Lh`o|26hoKM44^ zJ*Hi&o7IS$pE3PFng&Khd9a`8c1_VmPG?coMjoMGHw-Vjrf~f)Sl-qjPgZOhbUkXJ zdmbAH5)8Zp7liv9-XPXX?7M9HZej_*!)Er)o_1W47Td>HS(>*2T(GYfk%B+N9xxeU zVDkLk#an{p&oiDI2~)urHaj=EY_wgy!ZrWY1JLqj%Abhw#1;YS*yUq;!8oM6V$B^j zgOAR~7<33WYXGh_=mzD6<$H{jC-WkMU+f6|(ci4FuAU?7`X|EfJFNIb0Q=_S-8=q> z>1uAk1UTNz>X%;Utu#f=^sBH!ym-B!_%y~HF-=I`rg`KHnpEnDU!yFqC!F?rCuZgi zVa_L-$A7HVIlC|GyUPnYUAV?a-%%n#XAFpgTO?M0#Cnz?* zyM~RSRCi8|_?lX?(f~I1O)CrPOZi-Z?Xd;a_NV}Gqmwaf1%H5(m!D**{iwFb1+7>b zK}#`$e`w;X4eAJ(k!=sJ7jLccR6F=HP^-r#HUv{%vv|_f#|qEcj0-G@yl=T!3t3?- zo)rIAD}7;}K#YNmt2nY25P~$2E5hFnH3u>IURoG z*SI%pY=xt4U7B;sEGqK1j}U!J8Te%I2tzImW0DNDa7+3JDaLa3FA;Ezm~FxC)C@EHOzf z(oN0|Qb{f}{Vv@)3C-w{l<@zv0M}s|g)4zPV(H+rff9JmVEm(3$*!$p1Z7!P;hDGU z+xJ^&Ftievdm)5fanMq2PGcxQrjLmXG;92iP8pe4*YreIq>2UWnMDmY-dUJJcEVMb<=It0vzJU;lJZ8EYDidDj9^)jy+{R zx~4aQus6|S`RDdX%9tDDK-PjoTcAlFR4UPEjRBG#K!YY6*zKqZsPsX}k=lW;UD-UBl!0^`37k{94`o+w{47JO$bI_iW; zYvb7o+HTL1A#O$;kKAq^ct%JO8O3#|9vl?ImyWtfe1-B+17IO;mtiQ9ek+69JBP4Z zwmfLgB=!k+tdBpYzHuuvrreZd;cs#m5`#wrxKBm!(^_OLL#jN6y^>XtJw9B)3?)nV zl9)T+<#BYsd0&1A;1M9QztYtap|j`?$%VHfd9(%%z&eAur^@Y7aR5?jA7v1TW)c)J zIT@ck3{$o&{LT9c@*ITcekn8_XvaV27T+QClc<`qU-S7s4~!0*XzvmLcSkb8jhM$D zLR<5tv~v$UYdZZ6`Lo5jVotzu(Ew zBdz8tB@Dw_DsA3uUT!p*Tg2hF`wl%&bf7$d{`W!nb@IRAfWqg2xg8`o<6#3)QgJ^v zrS0J8w)K}9Wu&LUai(gY zU1EP>F3OKY^ObcUf^nt#*GbCl>CM|$M;4PmTc)7Rl!$Y<=v#&FoL8QzJ`IchAUE+E zKkdI+SJn2OnzZf5nH>TQRENp&kQL`A55ik0&vohnfY_(rL+r_LW9*BxQwtK$cAacitIWrnnzv|@NZQb@at z_wq^_?Nz_&+?7hlMrg-iBcU7WM~owXjqO5qJI^B)XJ!i#oo_f6IhrjnK4{}`I!eg` zaS5je-Sk(Wqc$DB*j>Ul3uT;;3^H*bXK&1EmI z_45fMH$;vpOI!w80U&j^gGMKvXecY=X(TqVgb8y|*KLuC{d+U4xNr?tgB8&#>n=B1)4R6M!j@ZT1MnZichGUCHAMv zZ@WYAsay@6ER)mB>S z&nyS@@_h#a5_mk1Rn?NEfWe(`q>RRtk0i^4PXvn7Yp_&mPJ+?{gc+1*NXBK`C@Li} zWD6KTTxiCia*%knMJf?nEU;vat^l7dSlpIxdgLh+Jv}N_0-xBlB`P&w1p+-EVpLWL zvvPy_ak=L)1A5FWue~DP#5jUVOc#K5h}9{!Al`sAW);tr?%)PkBnGRjXuf2!GtD$e zT=Sfu)~ojjIJ(E8={V>Op4g=Gn77dE>}t(V;()b|ZfTai1Nkf4E{gc?3)VcPa7Rg692NI(fJlAZ!${4Xj1Yt*+e zo|%iS0Vd@zU#FvL_EQ zdS;NK10d%%rC<;UbEpU?w^Z3ewa_ox{aXsUR$JAnRmF|*`91ubraUxCBi?cnl~zhu z9xrQ1COhhbE|~o95Re(;q4d3h3eDN+!f}n)GlSBo@*g9jDv?IhZz-}_rVHb#h;d4% zZX{_jg2$P!psZ$Tpp9ZRB(4K9CqM1HuQVJNnY~EZk4|}VnAP0$qucr7>u+-p75DY; z#KpY%$H6?T1Dzu_pD^-hGcNec)yHnX`q^X=_7-90UIH3U&`C2J&1LX@2)vcq*a`M~ za1{x!7TNGg80UV_q1Hn%Gmf1=%T0f6+u9dJ+_)$YgZxM!eQ!)pC@PD%c2fBM3QJ9X z41FF?cM)Kxd7Q>NXU{@$0a;z=&IxYtG|=vylM8hqzpy2x_wd`dHzz;p#Dosz`>=Svl>*3ru(6vIrlcC{9Xp-jw_do1=jZz)Y8x6)**Kbik6Swdf&jlr)!_$br4;e;% ze};h{i5w|yyb_q&;1$UFF?|_pS?L+w{Hu^-GlR*U94uT;ldd$N&{Lq4Tc>zHWDzCE zD=wD5Z7VKG6}ezd;3B25$J#^m^d_zgus<|pO>7Qj*B=OtaNgy7!3W2Zd4g{=IX9Y* zS-QmhAB{Mhkt=ZPCY$N~NcYhXiK9KEFT2F)oht?UJEyyu5L{xdFSl`*BfChfC*NW^ zc*Me`dr11TI5_P`vWexnw7B5m8lvqrgD+h%gtu2UN3HUS`-pVkgK}TJi;fmO9M2J0 z!p5L|g1Zf{@@d`7e>0D)LEUgAaKmLZxVp6S4o=r#wwqPC~pp_jOa z*on_KvFVfV^QZWTZ4%8)Odpx)XzXIphIUxaOwNbUJ^hfXGVBzuloX48!WobH+CNvB z6+ph5!!Fz*w+j2x9=~l46-Uj^`ZS1BXUFa$Fy+a_$m{xK7v8Z!iB^|p1`KtpArbdH zs@-$c1ACPw`rged*$3!nGb&8TbrK*7L|uPO)w;xI(zvwIFQdQxIu zrbHE5#LQq_;fqzrd=tRHKD0W~`f9_JT4EE%DWYM+{?+C#12+c;+>YtpiO{PqUqc8F zjJ4P@-bF?VvnML}k@&)!M93Mt5MilM8zo}FZb=OGr|eEoS!w?b7f__U_snBQ%px+2 z!VF#l2pkC{a{pq7b*cOp^}-p{5(I|g(|pKYj_#_f2fmLvu`n%{)Z$YCH&+98L8{$^6)pB35ikRgV;Q;J(Ey8Z|YOX$qcYmtUGlkOKQXA3y3lOXC0}=1IR~ zRcX1p$?`YJsJ*TiGDMA_Rf_;tk_io-LaRS745?0c)nGlL#NcG@j#|g zY70z#r(vHu!a-oR&1HAu>%Me?L(5xqb`Dv3#QSp;mA8uA^=DqwQ)Bz7ddKZtH}*zyZm>?N1e{ zDqJW#s0s^DvX=&SXdAYchXz2h@8F5)ln$C0Ov_>CC>(nvwrQ1U{OD(=Mae{;PVnF= zS=E`?8stVb62)J8`C$(`ACzGaH)fWDc%i#*oE|{kPt^E(7Fq!aoZ$SL=p{OF#c`3H zI7O&HTWmYtvLfCR9nT>)a!*vtTY#AW774kgV9^hvuWV=YKULFCr5F+7bBDWNHj!Pd zSC`}mQ&v^ak#zZ*54E($#LH#vU%Ny)Tj(3QL6O?|E?>4~%+}?fIDRFcL}1Y)s}YjO z>AVekak$|D)oicvp9$P7wq~YGa72!2yPl2l$X`< zJ2a1bie!CjmU+dpGnW)wFdQ=XcYAJqGg;&Mxwg5&15Nzv-7@ zc4@h~k$(JOzxA03smMNg6JqAs$1?@+E-eLJ=;x<|**)8WE=J!a%2UHST175u5|)pEhw1Hge9a z^Ut8MySVl6b(SR%G!+=Q3u$PRaZ3kS&pHLfY0_uI@DT;dwRm73qF(i4;z05JQk`>^ zPpwq2MQ9Eh%|eg2WYO=QzO#KaFxB#fboUC~dOFJsT0a=}->1(S48ow86$~3;!H|+D z-HVf~)Rnvyk8vjrdPcfO!UUgB$f|^6s%n%$q{*+8H!t)Ho<2+=7t~HaKtloX<_o-^ z433#*ZowV;NiW1=&Z_w72ebMJ_3~Q3)J+2P==|7&^poQXKDyZ`26yr{ zuGBCedu~2@cNhHmZ&%!{+Y+@U&swwf zEpj*E-hbzrP1Bh7umhN??((+7=y5Wq;qajvF&a`oqcUuJ9?(%|$nM~7(7$T@rmHd| z!f2yJ#-xbcu_%m_Dxs9=`x!ooP$9dfbYuiqPKgrQVYO+TN6E63%|aU8Nbtwc8$@Q1xkPat);ICyo0KS^fRlQQlQURH zTyUl5_%R3tUMHy$&C;$yUbgw)Ev@3x(77wAxgx~QDC_JkR92Z5a?29tl}Aq#J1|hu zU@ITGK6Be5SMkM)&!Fb%2KHc$&R`JY>tJQ8TD}EykC#*%1KU>wTxH5KB%u$HDM6u; z1s?R7Zg>D|fUQU3Dtp&Ev;k}rN&ZcSSLW#U|+A0Zx+W zpc7P8*4g8fRXfi)0b&d4FHn|Oa}DU(*IFT4GaCozlAW4ilD2)*9gr&;rM?5xIjn08fLz* zB8FP_1iHC-Y-*JAyp$CdVrH>DBc=yPw=7U>4(J?d(;90l%zW|0ntAd+ z&9sA1NlW^EMI3xl!CFL1;&CclLiKb!0Z9>TCU#KHPfu?JzLk7xbv1tw(F` z#-#}<1X{GE%x_>$XIR>i`%A25O!_Ai9YKW9DXj`30?BDSZF1~|gk?buFAq$EGt0#c z3C#^*4PI;)u934Rq9u*Ud;r$uzS3X2kBOr53acy33KEMaMh)iqHd>Fc7X8}LjYzj= z-d3jV^Q7aqi1|GjtqF8KCH5V0@c*yV8MFh0J3iI%cEu5xsCA&rh1j3d4o{=y_Pd|^ zZJH0KsD22vBd-65Adc9;&ip%rynDgF+dP8uD&zSJj1NCTOcSBMgrIcm`%Eno@_jaM zOOwxsQx11OLb>=um_N55=)QkB*Y?|v+s}dg>Oi15cA}n;x*qAiI|i5wP51W4DA6)V`KmEVM(;vYuKT2L*C!{q&5^5F7 z8`Sv$$XB^|`ISF+spY@nZztbepe}`tQu}mGwiy`A3+;ExuQS_n-aZ zBmTo*euFwa&-mz!;o3XPIpXS^aCj5yhO&Ex{^<25$cZ|x5%~hKzXRH6^NU!!fwt$- zu}9lI`jfYi{sNnhG1)*opydV7MjXCC%zwl%+@;lsb%hwqyD%RT8X+H$e9EW(@e#kl zB_aJJ$d8bROjz9|^us`hRgMOKO?8b4(5V+3vxjpeKuQDAzP8>c& zck+n5Ij0_;MLDL`rwE}q7kDX{Y1$J+sdbOVH%Y5Y%J^eG<~-~Od5^XA!bnZ;*e)~i zF>~7!1e&VzgX7xAX;)iv*ASTz`QV$F)L$@;FQM%|arFSJE2KLm>Q|Wde~DKAJ%&v~ zR+aALKJbw7?uO7s@@n#=S{#T;sgn`xfb@3=aSz-er07ZBHeqv*$cB)V)IY*B&`P3A zga2QIAeTtoV^O?WUN5MJj+B)C_!6zXmp!HCrn$(1Bn?<7ZkAdggp8#HJy0U1gB!!r z+z`$QG-4Qa2AP~`3xK*Qs@3P=!@?Jp-jM9aKbTSHD+$LuJyJLga`(s3X;XxXb zCjtSJ=*(47#AeEvX>-F^Q4h)gZX>}etu<8dBjFAqY=~*?jrE;jdP?dmb?h+{Vm}ds zU@1AyV8II%Q|FupfuzK4f>4)Yyb*|4g=z1_bJI4XB@q(KYc*4+M5~dm$A8|O$BP(Z zaY^sK4QZgxezoF%kJb-dYza9RtL2GTa@w4aX9Lwm|Fe{ zp?i*y3wd*nmPQ>DZR*GyCFVlRh15mrp{LH`=RFD4KpJ-Fte}CiKPIG(*!wuVR*2#} z@R0plGF+F0IBhVvH$hU@MB@QEx;pa94bn1~B&;lX_eRPJT4Z@z#3wdS4Ri*Ac zMXjbxW5Uci-5wARs4hmT2{Bfx_Gs(ia0{KiAW=dirUF69>5^uZRt9G%-+z%3Dj@?5 z#4giH^92%fTn#CZ5SlKD|E2U#BBT}c$L0$$hQHS;v?(CWeyvRdadk z)eecCjJG_Y(=q4?Iu}B|AmxeDdURe9S6g&k5#=FpmvX4=?|hQ=&CfkCai4tt?mLLd z!v}y)*J!23c#UrL<#&C?=G7(RxJM5+v3~1|&LN}nZbU$79kFyi&2{)NI(-D;HM-T^ z<@?AfA+HIK7Yg1(`P%E0>5Q@$=Iv92c|%@3@b4HG&tvl`@G7nJ%#WU7-knf(U&PAq z@>~Dt-3#=8QnIO_#08O)F{#LRd#@4}mUyd#^#k<2Vs`p{ZAY6G_c z-Z`p24Eowa?}-ie!=z1nnSS#e-Rj|YG zrb2h)6S<>CBy_h4=@jByNPL6g=Bwo0hmrIW-N}|VU#FGpl#659 z>J|E`}fCE}0nNvwy&RaU01WqaO~)w;?wAYhU5;U?9}z7|&lMZ{DI! zy^n~?U5;=4K4m;*zI+j_Ph(agUNRiLNgNKCL=7mn{-;RkuOfDbkhkRU6>7PTv=g+hFp`HWt%7br3fi`48;R*nu#cm9LRfA*BurQ< zn5}`r<6-f-@6lF>d7{+;W9wx+xj`#vIl|k*E|C1e=$JMhK^vV#+bSVdM4rOhgxSlG zK1wY!x&H=jTKP?FydcCCrT{BqKT`D=kw|L?asudR^O0|Ny5nZp{1lQCicseDlB7BI zfU<}ur=u+}#)RY<-Csw-nh?$R&kC)s5e>L)rD#pGxq|>A0g(ifh*4|2B&!Kt+hFLg*@1gpk5g{$_+&u+A3|>JC~-)JX|dmL%_vy6%h#Pm>5eaiQFC#XbH4y1KKCok4wK{z-z!ngBrhn-f(E=clJu{{|m zFqRHOOuY!s@G`a7dm=M+2CnTed$=2scRUQ);1NCqlZ5v{q9Z?gm z5myyqLp~DFkyaZt@j_Ew(vmTd`#`IpQ%9Q`l5CkuFw$y-WTaHE&=!0}Bjx+Zm?50d znj*6^d%Mj!5;KyZoF92fn}~m%Tgnn2FA6R-e3r^v_OU~~FS^l=Zua(ir+XouTq38=kE*MH&N>0KbEnV25Y<_+`q+QRXF+v$fe zZl9vgx4=#)bO?#E{Wzw#5qX1t?ayoKpB?c%3T+=tFuK*7w7F;AUy>xCc8f{|HW^OFzojifLGTdXw3 z%QJ@K@7}1L_s3`n)ZKml=|B7nP*i=TD_!YI-+TwiSI+J8HS69B~+&h8M`7wq{sR+OLl<#T@P z=iZ+m_Cqf*{uvK}Z~Nl*b3c%+`Qyuk_$$woyIb@(KKDI6v^zRy-aboSCl2)w&`AhK zA0n?K{q_HYm@kQa0S%mAe3aZ*;&4G6E@<-!W&Z*;21+XA;r+h|*XCci1vmXBC2ba;QV4~!UfUJL%yn2S1 z?~~G!h_8WpcN3k?fDIx&kT3zv%V6gDrzmd?96$5FA{GduaI5iae}BuT{+m~5`hZqW zsM|tZJs=NnVcdl>__#6xwWHP(LU;v5AN4E-9}QY(B$OAf%EdLmX_tbf3sQd<+ucBNMY_9CHndtW z>z&nTjWA5aymO{4;s}b`FQfY_K(Nvg(?n~Dwi~D=P(~qHq?G}Y=wrr2Hvv-v5o!s9 zFcI<|v7V3*&JLYU0U>pTG(1FGhiN28^wD*3!&hfFau7Ext$7hu3X0?{A)Z4TY15iE zUuV+6Q=+kX5?KN?H?U1J5)W9#7Y55&&Weh zJo4mb?8F(Nb3to|Hla5E4%dn^RrAU@z*>@XFtcMKtR>$c00NSYm>QJm=4$}rdNDwZ zX&+eS>cwz%;XfuNcFuAIuogTGIWN9C&Cdk_xv!X3Pv=UHwn)rEQ=?A4zac@4?a5-E zh!L|F@%j7I4OO#%&VU+Wtw^rKsE9-m+>ohDstfAApiLU=D`RIbCMvOD9 zC4{x}nfuu{P**e!r1eM{2WlhL3 z2z3QAgLMn~BS6TGB$^a-=aUuDsE2F52*f4TUgh6@?SJ84KYgEh{2S=>5$eT@gl>mL z|1hZg+ekPejo*N9pKkSeV*mcguzYeCySql`rzq_vNFl_LHm})lKSt1-#CV29|Ey}$ zDew%akvAKr%O6LkZ#qKbyGXrCe;|a6>kC2XDVmSK>-Re6RxaI!9>iR2CUqT-~xF!c=FdFH*M8kT>^<`9M=$IJ0}&SebWU zr5^UA^agW1pz}4%t~p=#XiI-O5xYX_AMuz!TgMSC!N()HMofETsMK~uoj%UE{Tjo` zo%h{7(-l%CWVk^%eva5*01<5+V#fIBIfUzkc!YEZa(79H^;=kld4TW?ZTibxygm|a zq?L)E_`<}m0VZ!D;cJ*~D2EMo{xBgOkmRh@){e1q9W6br9T2PZ`5xnRje0nE0xUpU z73z43q zx%Hq2j63AvEylM#Nu8gAxFK&=q<(^^g!Ir^jl~mWqDR=YW)|r4Yan+P^Kv4kk*c>S z(~QI&D$fx$p=FKf+S8)ITsx4C+!b1R6V-tfgpewRCnG~p=y!gj-vS7SDMHJD7w#gHF{pf}@P!Aai9kDYsd9qe(MslE)r3A2^feLMoPz}n${0za%-?FD4 zjpC{Adg2CND%5=k^vxS(TIxEw8BCdrHd@SA4@>~hbj5<1kz)eOa<~<^#mxsj(;ehAX zI!qHrBPLHb+PG5kmCTuUitbn3Pbbh)#}_h$&F_5QoQ!;3fU5 z&OfypeDT113_W&2>m#86?HE<{WT7-K5qr!BUGLy>Po0i1O{Cx~YB!4$+5#4Zv~p8d z#EYxy>|9=LpcAyBq`{W(`(uu=cyc#1=aEWqhH(YK%L>&3W)U+eheS$&w1$2IRfSlT zHW{(2h*V5>gu_5B9l04{Gn2at@c@hj5@tdxU~@;CGOXf-58Y5|)fS{w70JyH(9VFDuiUdxETs<_M*Mq8F5&m^q~+!^Y>_&(Ru`NPn#n# zXX5G}VO~+@j#k$MF+yj=0cZh>fs`7xikDL^rFOCcjY6FQT8xmCyxAjipiU8!0HG4Z zu$Gvoh_wr1|A-I@s<$yaAxQ5y7mW}Mi2;+&6YXt>m2;#VQCmi*d!YA7^&X%8bOrti z|M4&XHs$=^WehJ9yAMIUMre^bUlLNLtuE=;uMoq#OM&067%Tma*El?YHh!F@K4EgW z_;FHvhS)<~-9cNR?9Uw2Q@=qzy`(#OKpCHR)KOg#O+dp3E@lVX5-$&Qr?1oTI@HYh z1!((!soVdN|KY#*vt`g)eU-YsLtXtCsXu^@)M-uZcOU$?mZSb zwzrrsuOs$xHaGtW>E7R@R!eN({1|onh`N7;-~5NK!9V(w-oq92hh>2HvR`{eE|ayQXxq}CN}yg?nGNA|b)k3Rc{JU!$4Hx- zbZ1|Br@6Z*<}GRm>V8dHnxSLAMeBxoxWT-Cn&ISq8j|-Xs9qxS$KaBQ@tb^}pZV4M z*u77XR?kp&*N}7#=XMsfY_aKIQ0K2P$FET}vfB1k`*+C0inN&+&USR`2TxkKpFDk< zpiCfvG;FEUF@Xt-8LLmxwlBc@7`(0Y)Do%nKExBcc!M~618W`S@|4*!*ll7T$(y@= zQ~3Dz0)$FRTS9(~>Eaf3_oGbP*T63M^r!u2Lmm;kN9^9p}Ik)HQJuSbmNAJ zxQP@}z<7xvP0VAWUQ9?jV6r9;TL>FZ?~-V>^P4()(y%AkC6doc!zFcpi&{V|(dtMD zcSx&ySZGjBXl=m4K(J+mni;r2;v+)bGhRG{+J{J+Kny!f?vPiNv^ocCw0WQ&j%l@F zaYX7(RF5gsOn3AOaee80VHIq?0~WFHO>}=ronFE?#@Z1v2Fh-Y*5FO%N3S9il*5N; zhhtK|EM(Qr`HmDX-H9V1^qxYzI9tpXA*PwI7B2&u6R9r?3${9DphdI*R>+$NNN8^O zX+r0YIxn!-xkFPXw3(I)Hg~8*VvK}jSXE4YoY+Ub55#Uun>%V*F<(AI?Ds$lbgr1fSS=pL5RfFx6HYinwx{7pb2Y2YtZ%`q4U)9JR4yWVrP)t zU=daj12Hupfo>7ifb^9%M%3IXi*aEG5Is@8FgMKtgrxwi3R(i~!VA|RlZHw?MCN@Y z7Een9o(x`s|6W42#7C==``OvHVZzcSp=+dJ>-^ zYeZ*g3f3#B^ThCiI(6n}gYhys)}lgpB3MXRQ*3X9&Kb8RhO|m+>ip>7CVS3;L~|sB zB_v&QjDua0`yJI_+GXk+&k{tax*{bZWI;&i^168;n7@u~Nlo`hcoMP%2#_;O#rL}n zjgXBcPV6CK@e-rfoTD4#OjTJJtPPz5?cfaJnA}v?8Yl)580A`(6XV7w@^ER zvO@I;!Zk>qEY|9njFJzL5Z$yN8nG+R$)(5Ynl|=O3l<8kCc5=Z-n@=TW1dbijtDUUCxl?+&07oWc!h*U z8xtV~VmFi5drz;|o}@~x0c(jc7@-C%xCz~+@Wh~v$q3og+O-VCPCeZ!?=%S|__dA} z7PFOw;~XecBxc-Pw>6Z7oo$AMg0vn94XLLA2DI^V?F z3(WO8u|6bpZ@r7-|5zbU45zPi{@TZByVptGDY1Q=nnyI9GwxnM>;QT}DR&6@8Yy+8 z)gwaQQFhM~<{m8_ZCVpo+xPrE;AvG%vGA=0{drttb`R|JKXY>z;5UC^&(HqSmzZ{E zC?6va7u4zc2OGQ^7#}`|)w`tOb&$qTG`>mR>l|VV|EiVy=NLt0ENbLO;^GIAz zv}uiWJ&d>cTmR*T(-%iB9zGt0=@^O97Q4bWGE z>g-0bSs=xRgvm?LEO5C0B4#fWS3POC2l~((*kg1<-5{kcao90$uTiIrba$4U>(;Z z8fGM~kS5fcX*(}*GXWKaB-Bv|(P+AqZuMT&r8Z&Xh7bor3Z%_BR_6uf;JzeQ(0%8o zMeJw=;;@XMmx%R&kcHIy&xdHl?CCgbE09JUc8Ch1CrB1-0<+nT%G%Mb#s6IvF0tBT zw2E^VOIk|6j5w&LS>wX7C7>Fx*-Mq&ar2~Ln*=LD(H??h1|$R|3iTjpb)15jmrcoI zyhN|i1j~*^(6Ps4CJ%2xTOp5=XQ5y=VcJ1x#6E&5HVfmlBCZxzu-O9WG&ePy__kVe zv($un7wI+~CJkhdqyu6DgcZShFLojkLPe{m6H^aDA$E^Yt7thxR*k%wm}-a3o`Q}` zviLFgR{}6b>Lh4`Ho5859eZ-NP4KuG%GAwrs}Qm$B17*fMj^VH7Q6thyB5)Mgw8$l zcH`stdW6)TDuI|HAq(AFA>-+1b<{wNj*w}ypel@8p+7NV@TC&Tp0M_9Szha;1Tj<> zrmzKq0HQQoQk)d1d)lPbF%bGh?5lIns}U4}&4g~mTJUqO0j)#AyqH~qc|Wi^ngFHl z*U)+(Be@YIKrh68!CZ)85|)y}h-tr=e@ZiRcJ=HSj?_1*9eZlJYmhGpO`SiSJ+<8| z(3&`&BLP(@W5Dc)HYGy#B&Z}q_Zz^|+7fmPMk5l$n8qHgBOf^z7HL>7UQCw?+6Jse zLXL=R78J?6B-sNJMy#!fDbuZ9e_{mcyvFQ|7~Bx=j_*GC8fzA4`+*pR&^s5|Q6bUa z!@Zw>hLJi&hC(d?O@7R2o*-9Cbp1q|{aO@z|MYn;iTT1tSU#V8dLlxLgjIu(5sk~j z%knIb4R3KizSxGCi{mwd(#Gs~1U1SOu_Ym_%6OzZ~KR_I|Mu3saB_nw|R zJa`$=FQe_tj*OTuK_1bn`1>etk<9~+jvjIT`d_8YFVN%D8+r3Ot$y4WAg5a&Xz-Rg zAWZK_7SG$;=zO10{`88}zLS3X^*`Wup1(#69aH@)^pctP*L^|%+oaRLbpAB8TwwA# z*KWSX#hsUE`+I2p3cvAlkNEVb{~!M4^B1&wP8pviZhzF@yRpzT5j0ZgbIRo(BK6Dk z*FMKD{RIgEUFk|!y3+TdD}(n+KLD9vpZzIT*Z&fEeaY(9`=kvQrPYr0%|8G>!Bi)H z=9lmN`DV{9l6FkYYr^{dzrekp!g`DJYlP!Jg)O_%_oH2$SZl#r_kBIy4n5Q!lF@aq zzwgf-o!;f)?mmoX)crAONq>h>JegBYXr=cS=eKD-f0}xj8OPW7XTSKtWo+NI#Q1gY z10QUdK30&A5yLU*jWdz-!EzwFwq^Dh;QCG(t9^7Vi?GE{U3Rw>KU%jj#5!QSj9onbz%9UZxFcyxlbM(7}Q6(nJ0 z?U>aj#Hchz$~MxPkk)SAN~pv<5pqM%*W^`XIBpOEbriG+8Xuce=EX;Cg;gK>mlmiZ zP~Cj9qD&VnH~+Ib#y||Q;z$I%giM=|2FRTt**qbvVKMv`v+wb1Z9o-LXIL~(;F_RS z2;%J37#-{BivbW`x-qvH9IXbZMk^g<{|qrF)UFYtQV$(g4T~coc&S{7D_VKmyqgl-4c zu=$vB=*TN0_L0_<;7-4f<}Zw4i8zzjASJ_wM#zzn6fuELY1O<4NwX)pMV)Iq2w^#% zW^TATc4A>s=eQWl8$>*Li+2VtsawIKVa1KL*tr>I4L3XoD5Fo|Pxn13EWr=gV?WO< zU{#6H&ue5sNYDU9OsR)No58A|vpNgNf-q@~&=>M>z-oupg0`}h1xC;V61NaeNb5a% z*w9+0T=c|YM(cn`a=x)d2tw|R)Pq(-i*K)lB?%AI5}kQmBW+He5{{WPMCuq(t;D!p ziWVzaB}8)+f+=NMQ|E|{Cs=KMzKPPNGpshM1*~OA5eRO+T8Ea3nNJ;*LncJ%!i=OD zD?Jk3+(l-@3fLO68IcJjp-m}cL*hN+s-XJ~T5llz=bv0#)iPGjuvMh%@4e%?+om;; zkj_X$bp~q_EQn8;nE3U&IzE9I$w86e(tr0pEz8+`-43|8;{ZU26^R0V)OPo_0Tbmuklyzev^Ot?nnKZ=+mE8 ze)=>2g6GdGae%S?4f6OBx<4gu?k+f(j7=MANoc!E2(NQ|^Q&CG`BCaFFkXI{^|jAG zc|Q1kKJ_bKr#|{H#hym;5fZb%cf%Ir4y*b#e(e{oO4Y7(r7Qg#uPcN1N`LnHJO9=5 z#N!VU`^0elyGyUezOcIf3R+&m=DYm8|L(Ov_o#*#dPI!yeWZ~CMo>lC@Tah4SNeW* za&pe$;(~tuC}HU-|6Ww*Q?yL5idgslqjie0I{7Noqc!9C3-n^-$0-jfK0d4+@U7>V z9z4Z#af{vbMMAsa?|kO3VXP5}RGEmJ6Z<#$_5b89e_HAIpZ(0*(}^jP-lK*3?F#vT zlnb;(>fwYqYzZPj2h9tnLbT<02-Bn6)XS%s_kTz%zr*MM{>;mtoq0S{E_s45BXU9A zUndU_eYEpEq?#$)=dtM;w0m^N6HFgP1|Z);T4mZM=F4lu_><)I10+04>SogV?)&bO zmLA=n66|ZFlP|&i6V&}c%CDgNGp6|(As&`9C(!CKZ9GHdu$Uwx*eStw#Pu0OX!|4d z(G7wQ4A;N$t|F5GiJiPT_GVN zv69k6$TQY9pgk=B*2wFLa_Bt4Sfh{7i(tA(Oc^h4I&)MEEnCuB3AxY?{emG4#3ax+ zPgMeG>7UBlm@Xq)gs_?k$*32eNDI=CSkcm0iX zo(3PCM;&*dD0OyfXzG=i6;)_c@=CDGf>hY+?Z^NCAOJ~3K~xZG$)0RBoRgZ2whQQ9 zJh?qU2#!i<1+*k2dJ50P$E(d&)CppYpd(TRlM!2x6><0m*a;!tB4tmRYBky{v{5h` zIT?AVzTI;o4-?Mkt*8dtp{LE)NUL+_g4TxCEw!z&SrBcc<8x%}-0*9V8L-g&+$k%p zZhSwxxD-AKu~sxy$6JI*hyf9y&OR0|VKH4h^VDL?)G;8v5(a1QvhblpmM5r(R{ZZV zt21-=(B_^rh!>PN)|bjh&bI#s-tu^NyBM);|H=BK7leXKIbH2AK?91BgcJ}_}H-6thKnO|j z&?7!k#}%=gkYR){Vq;I86;1ch;T{yltb;V;WnkJY=;MsgqaDqRwb0CNaWokcguKy~ zoK>)34c$=#Rjeh*T|ga8jVY@EuoZ5G%26xO+nexK_hu~x)ecvo6bOAG4v&x<0~Cdz zdo)QvqqM{+P&a`xrIuP&r|x1a@(W4~J4AN4XEaR|Ye{!a53v#{(*mm_G2KHH%Jv$y z8oT3#K|>PZ5U|B3sh(Hs(Uj2T&G-G$P#=N~DfU@Y{sp`_#F|=TmG3wKKl>nG^Ud zeCBhB+xah(t~|qV;YGUjTMN_{mtTSDUoxG*XipP*T1i|glxa&`+`+A5KDo+#_7t{# zzon2wMd>IU4e?l0x)SZ*JR~mlfN0Pkx z^ayfB(mlGxmg(dIOYkJ=(CwW@;~60zwj!$dqD1fk$(LFbn^^w zio_Fyd$^sD7I$zl&TfC4GTvu<_5vaNE57(^Z(g+i(%(7-z7BkyPd*npc%tL<*_p5W z)t|ic;#}uC*SY=^*15rZuJ>OhJw>;?#&F>WubVXQMRBnqE_dYZH8z_oz<;V+I2<%6drv)`H$#%j*TN&89 zsarq9@^_fegz3R`>a->;j%nk<&A75E)N#dhdY!tBOb;HCbVa)Gi)gseVqgS%%b1FV00T(Z$vt?G{_}{JSWvj>%iW%K>$_2D)bQ5XmLd z<-oi>B#&2HGWqNql=?^f?mnGtzQHWFvax(fTx`&8!t$DImytN3X(Yr9IA)=x`^IZ{ z8E|I0gBOWMe@4CaVJP=Gc=Av1aiCjV=j_4LOq(mz`g3%b&gc%l_m0-MO#|Bp*YWx) zY4sB83%5Cb@d{oqFi%2QeH$rj>f958Quir2O_|v5&~S%oa{&q8AsrOLaG4mLJZzXw zKFW0Z45cnPxcof!LqE#P?f{V$gpN9|KyTyRVmSN(WqKMx8=Lb-HeA$+!-iTEv-{Lt zWIDMt4b<`JV(kdl2#L;tB44iDxSTK3IkXnrn)Z-TN0uY*BKST z6eK8C5;nCmtJt@aE7Z`WM1wL1Y!pcK4V-~!TkumSbWWYMmHtUUdnf9&PqBfJ2&qu3 zBc*GJ(I9AQhzWl*c0FnRErM-&TJ(2r_dR#qSLnhB-2yx^Zxr*1nlBIr2U}ns@twB! zO43-ydo!(11??(U2I{OW73+`=q@|f>A@tr-P};bE=pau4F$Zl)U#o&mj);>M zt>6uds48&)NydB@iJhTAsI#|}vsQE{)G4;KbXDSzUuVrqGP<=iR&r>?QZ*}eZi#uz z?R$k;Ma^)TKm^-$jajYjx(u>!-GPY4 zo-#*7BidzrY70bRIZ<|silO~XNG-MO*`e>+bukHb=gd3BatErd%+8F^3!Ve>PFpG8 zazx|-Xl@7{qx}u52^&}B0(BhP;!V0n9oG<+s0^@MAp1ma%@+|_00rYR(e#wVc7WasI>NKB}2IE|^X(lh=F`iVJ;1il zQ+6+6^Vj+EkKqTm@22O*?zzr&u5-PO&JEsky}$b8a}kYK2x0iC&;)U~k9HS`<5m8m z8@v)WP)71@Nxy#ghS?w0JTIY)P~JTWd9L?YdHy!K`vCd$XXzIYTk7e@$&~5dhsYELP7;+VL4 zg*0q&HR|aF#^aAry1z_{i7?#*`3?t9f1h9Z!fPVHlH~F_s_9ht+*oZk8r zb-sd~Ey#MGusnTJ`AW?_WpjbLTS2)+OfNHBx&^?&)j#CH_x}pM9otC!@G`pCq8cb? z&k%uHc|6{zD+S;=te{?lG1BX?=zl4%9oJtF@%dPO_}%OsBYwNWXf(>=%HK;{DV;YgYx$CA_RXjsn zIxZEPBGZGOxNyR1B7}mKWS<)@7ftHZ)4x z4<%4?Aa(7Bsw#WuqnTgsp`|5*6M93GhL?zs_`3sQm1@$8>V#%ihL)o9Dj<6%Y^0T} zRmHub7AS3t)Dylp`%%LcY>e$1Tnjot*+D%6@7Qxfo2l#j1hntVz(Qzz%)ISk(?d71 zI0Wr(;^i_nU%=>SsN0b;Wy;ur&uE{acFd=VIyqi5x)7X%TcQ})W=%|jXoRpuYDUzN z7(h;`yT;t~nyAwSFr(Ih#snisIBSJ=(-I8}EH>t;s1uU_i8@6}ZtUEayp@Ju2w+*s zJLr~`@`zxn0L?JdD7bqomlBw_dya8x2m-6p77Z5d|LQELdP_f>(Izh@+$-V(ana25 zY7o#jPz-yVfl5nc_v^~|0E#KDdzS4KASc4mINM$VHMjBkh8bz;W=%a_+IF>6Xl;uH z7Fx1;3Y6WdEdWTs(;0C&wtY1YU>zoVTt;4?V<3mnjL6W6-n>8=TN#{tOO&UfBEgZ6 z+qD@K8(XSa;vNv|kjK;#_(W=5kkYj03CEU_x6le1`w3CQY76qGUJ3naob5Vwhb|d+;r{s=02^vQxgb)>t?KhKUD3fA!-BR!vVYx!2!>lL9JqRTn z5&L86RM0qKHK9~&Uf|&f)vhg+n1arz3ht?$PcqY1Xktqo;{>fZwH?3L%M>#j64lFo zVI-2z4oT}1q~DRpj=JrryVS}WRS}e$JIYQQ%iAiJ4UtM7dt6$HrG-{R2L%%WhK5%0 zv`>W_PD1E$?Wk>Sim*Cq#t*p&mUT*u1*^q4iQhs_nQ$4Gb)$v5ebzRPfM%y|4bWjrLxCEzZU z>xAwKY5DDU|1JUyS8kE6eu?d^k6^o}$m2&^`Z#|V=VgqS=(k_vxBvSW0Ql7BpTQ*& z`?p6YJ(A)quD!t7t)8-b61#hY{_r(~g6bLCeFvLPDZ9(ja826ete zIr|e_|Aa69t#`(Iy%+tpzjsJKd>H$}HR|OxrioB1Uwv}NFMsYe0=GErUgInO_4GRC z;>Svy-{E_}_xSX0JxT8J2YHH-eI)(FOsahMxBqb`4aGx8NBu6GG34I?!Bi(&-L$Lzx&^v z@Tt#zgYk3;=~?>Ae@^!XTKe*t4e>@La`mWPqFy2#?npS{jeRfLnWFl)aPDexcEnW@z-`g^}f;S=x___hbhxhp%wYK$m^9g>B3#Y$v3d^3Cg&o`Vu|6)JBrknmJMn?x#?`L$`jJ{?Z%C8eQLK zSl(uH|2{U}B&@HY@oBKDz*7(hYPj7PgYzxI#g|zg?VrE;HuH&6=j&jH#Ni>+<^$Nt zlG^p;Q>PNJ=`rG69{WsO9i&4PJ=_k=W9Xbs&WunhFwc;*(IP zg6lCc8EKi(qirjW+Frq?1E!OQtggO5ov$I$Q0-fqH*KkN5A~30NBB5V$Aa@cVt>}m z85@~*2M9|-e1N9=gzh0_GAtiZ$4kw$!AZ*@$h+b+41XDtviNgd$Y7W#A zTB(->O5Mi`#ULo6LcfO5opRPn*y4I_hH2fKOqQ@YVpS=*8LMegk)SPUCk|E67EWu- z$^A(q-X=B#CqyUbKpi_&6mvyfiP2htoV7BnN=p#OA*0>2FD>e*0#f3ht=dMm)mn1j zO%V3mCm>zh0I!aDJC_lo=C*CML@4dC8e36Y&I$!W*HXT!Fwa6Vh!NtNy-C;%M=$K( zSN7LQPo6582PVIS%a#~lqm~0aU&6~#+pjX?wn6(5&%*d{fhvT4A|6&?Ko!!UEsZ-y zauz}~#1nD(5M)5YSv#+E19eNC6J?huyP3G&kcJVni036W-HoFtTsbbLyLP?ZKaQ)wqiBa{e5J{N>9*42@x}}Y}gEw2gKXsY=&AJ zQ#KAQJugb*M6*wDi%>&J-(z3M8$*~oCnfbw}y8t+n z4AEA8*RTi$6L12;NC-~JEArIJ9wjxVb1m%{`{=kBUK;M+lQvX6N3_ zVE5RH1QCJ?F?qY@9jI%p8lF=tdy5ruN8%2wG@S)oRPP(Dr5ULKq`O7Bk!GloE&=I~ zZfO|Ok?!u2Zj|mAKPj+*ePZ~n(|@= zc2#Obmf%!Xg9P(`)lcn|M;wwnuo)knP_!U2v^m2!hQ>m7X8bo*>L)L2j|pRsfAwDc zY$OZee2=$tO86n#o!txVk@i#r#l=pqD{No6Z6H6mKlLgnDlUj38y4KM=(OXwC#=tU z+`(N=92h$@vVD~AdYQzP!>Q=1s_?H63!0TYhKZQ3-nwDWyAH;!t!upOdR<;LayIwv z$#ssM_YLfx^L>!}_w<}y$1rGHFrw6PD_(aJZFIlX7Anl&_?(XB2+pgLDrFx^GZ1_NIm_vV=0zUA8&DbLs zM~5Vy(K5t9hZ@5MRJh$z(x@tKXVxBxAZI7fn}2-*OzFb3OSUM@J+LU)r?5QMdZM1x z0HACRb=l%%qMA#U+v3jWD-Xz{%NVo2=Z4w8F^L!C{1mlpMDJy~`s11-!p+3GmZ+k* zbf4u}*!XR%X)*A)b2XlgaG_DB8)wCvH*)t}Y;mP`kj*9ymkOw@&t0Xb?yu-^#dj%} z+8M4f1x8;BIm_Wmu}$>ON_@tcW;Zhl-2dFNH6H%0 zY+VsLNeC+&0eg{%TMaMD`RMhKz?oFFD``ZsI=tja?K8cFVCOU@BqZ#<3EX~|32a<; zpUSmkYh6l@U|SBfK3YR|aNU6HjWj^7T4 zpyLiPh35R5oHSQYN%RUEBdTqKK?0$14sjD5BsG5SwFFfIW|O~E8|GrYqM#D`V$y$J zdGiJ9m_WJDq*Q7Z1V6ks{INrN_P*F`Z>9ceE~Kcujv~^-NGHQ)V5>-iHNy)BtgKb( z>VAph+I$XbIA>LR?b1X}ql~JYnylyjo=oTsGf*vs`8lA3Gt^&ue>UZGeJ_YRCZD$G zQsPp5{C{>#aR@S6jQN&Yz7QmGn+-4aB6 z!oAAMs%T6^w(TA7JGhWq15qmFDigZ{m(mSM)MhxIMdN`!k-o~B>(zUvLAL98uK$|R6$EhUNZ3fu>%UNbt=+tRxS zUCVZ^s%=*h$TpBE_o8c}mmc3$<&o@H6Lsy0rRjAemH7m_y6uy}w^}TYJ~`GfxGFf4MJ>KhX>`m)PTbuN!XNj;H&w+vaDwctIn$(S?AR zM~}NuxBFb$gXMS=2cXQvX!!yt{6zTCKjpXO%jH#mD4CIGpopQl%TBL;({rUjQauT% z!~ch~55@YM2?THxlbuIg%M;-zL_x^LMo_k>f{w@t9FF0b+S;{!jF!s}EmM!g;n$|j^&EkIw7{W%n{N4) zhaT9}hQeS%jp`aBC7!z8ty6Sf1OWh&2M--MG-}d zIxpQonM*jf+L{>g?$Br3D1kk3xhzLcsx?`52WL{f#dc>7+hE;e_!Ty7*spfQGI5ni4trCY!rrLmB57(N~P8Iq?A0>q7H`ruVWqWO7|IN$+cp7e8>-T^>Mo7!Wwd_f14}3#}2G1vC2e$GO&r7K6X_1$XR#Lom^6J zN;}O(iwx`jCT0BRc{#!2ySius+<<;mhj7_vnH>R%VZi-SN+HB%@KUO~(9t~9@MC*( zkp;s^qJzOA9Mno8ig-mk8NR!msG9^~oYw(3TL@u<;#idx%v2eqXv*`~)w!@^1-VxO zhmWjs=4a62DX1o^-tk&=d2VR zX76~2U5G%geOLdvv|B!1TXyy`y{xnR*+W}Dbfzai`r<_}V%gLd+@ErnHP&&be~j~- zxwq+TqwoCff<$Ys8t-#szAz+yG(_x!Q7!SSndEOtNS8Nc3TIj;nAbk7=xm$nPI5Os^>wfL-X;ymH58bL`HR_}?oOGUM~)0Ow; zZq9gX+i!3Y1<)h1k59c^bIQ1LV$Tp(_I<1t^DV<~Fl?FwqKJs8M5J{6uCCrC2ZtpZ zno#=eUkY}#2M>nsDzXI%xiwWfiOTdJG|7!Wuts-;jKpxout+b?H;#z%*(&%f)=)7} zrg~sh7*bbKbp%S)keq2CqkoF4k-kFzO?iV$VfDfu;>l(;Z)}d`9S8KV0@kgW;g*yL zX6u=!V%uXbTKC7JqI@y+TEa80JT8s8?r=yE6$(L2&9#AUvYl?nwrJ@UeLj7Y=NKHT z#O9^5lSg7=>VsOse=%(yqhj?O*ZVUKQCPY zX)hKKXP^b=_m(qNQgz?oIH!#l%wu}tzqcGbEG41I9ydg11_mqY3elo9eYCzqf@`K7 zLd%c3#xPT_?ZTE?cxjBVjhFuO_9jReB;OX0g08Lyl_?YCq*_~ z@uSmk!BBA&cj~~VCJFx<2395kg70(;E0?faO5m~J-4p!-AI!hVN6t{N6KbYfLzOWrH24MW7jBy%w9?rS|7h! zTx~k%xY}i@zJ5E%#*3| zo-;RT=5$OGo8H3a9&Up$GKK1j`b}0 zn7jfT`u(@XyfT^<0P5Gns@gxV=z1Y4+*tUepse$SUmBUmjZ{JzjOKCr4dIiXDe+C| zvtm|c2_4^dKg><8u1nDdkTlf9U4{rZ=LeMp}>w& zq*Ds_Vax9O4lHU)w8ePgMi`d z@#uR_g-xDp=`1F_&k+m91K^$oZBjHAnLYRKyh6Y|pA9B-UuMKxA|HNV_PX7-H|y;% zOzr)K-tsyN(y*^Ru2?o*erP?%)T+i0RwQ;%AHW_;DV?$hy#S|mxvounIoGxiM|s{I zHQjY}`RCjWpS%`e|97T875rC#bmccxdPbOi!j)rSKGAu;nqwosp#Qr2BcuI{Zz32r z^|Vp_%U7r5wLmW0LdZAX!80HP5Vnd3ox*i{w|^^IbhgAA^St?pF=_P1cWX{Ao#dX) zX}$VMg`(3{yz?M+k;uRDy!_=BhlEO6&@*_XGKcBfBeg*V*nmv3M=p*e;zX6s zW}ct5Ig5XA!Sw=YxT5;eR=Frp2yjvKQ?I~xB4+)WI9V9V4 zwQ1~Urv3t@>JBu~2UE^t^*r0!3$$YIEGo^bZtFcpm(!X+Cz7LnV31oF<8j4-dp_2y zx)j=Ztl<0pwr%V17%-2$gJ;rQq}Lf)wvNtEX}m@2<{deqn?Jvg_!tUAT68R(qFLI& zscCx@ee;^hIjIU@It}ndBj=s^Ccx?VF6r$S4fAP<}1`5(l}^ZyZ~A#NBRQU z59&&&oGH)EvcNnfGMKGtop$pMSWl2HRJSIPy4&NsFQ~DDaO5MM(ls6_2slc$?)D8C zc_T^-=uylfUaLXr63li(EgHL!#&N{^1F4n?*cp}La9kLrswh{s|C@it(t zicjE9ahFLGlH+$=+ol2>2-micx_p5D24?JO%f-%Qvu`f0PU*$Y*;*A^O+HtDwZ3^! z-^F1GKTfM$jn8D)!R3|s8?Fs>G3}J2goa~GGxPF0EHecup)Y#2(|jsNk1(z$`5SXf zhxb9OQFDuq-SKUum3{V)zJF0U8PHfvl>k!O;E&)PFfcZ(hV_p{luKIQR`q(3B9r#} zjTW}ZKW3f_@x=sO-U_$<2*8rkb@Bo({i#9liWplK@(C6$1KqTZXIQ}xeD5yZeL~0W zd@H%t!fojyaIC~Ule7I>O@M@*1u3paADh|=vmQ#~yzpw{6ZKbm{_h4YPtK)1B5(Qm zD?sxu7AsVesd_!FhrLU+-}5CL#t1M*O6TUIl}<)B%Tf^a6#D*J>EAP=G3K4+2O zyB%sF@FDHlv_Bk=i0yAiqjHw1S)wdm1-G@+`7D`iP-NbbKI@VGgwBK&)Nc-Ga%?Zj z#nOj!vv(DcB*UVd5Ueo1nm0U+TT#bMuEYd>{SfNKdg=n)Wq0LLv*vTwnc;q5l9;Zy z7ST3yMMypK#1q8`6XVe*$jus{3N}^pBp`BippfLQd@tq%=Grek^&#m)D!i63b-88A zperkR*E3?4YTzaDRR6A<%P`vcG3vew?j%V+={#o^CX5nvSW@**Z(E7pYNMuA&=8|u z%QZX_N2|%4{!q0=OTP@s+~MJuLy8rS4yfWRw_Q4vKqsJ6mm%@i5YNp}z{Q{cHQTq#biP27JhvG_mA@^%iQjm9 zv1}Rj8bP5f4T|aYVSzo-B#jX4ThS*!o{QhU@m+O>JUo-FMZjv;9to#<2WZ5k#1B`e zd=#1H{*tA7tnJ`88f`bOZFA+^FPp{s$GP50s=p8miAdk`(2~8y9YTYheX{fTzPE{* z^ITbdywK=)@p0QyN54%{=n}H3F}gB(d_E=`o@wcICKZOY~C2rJMlsUcg>A zR)2_IaId)U)MlD1&q4W`smqRK=@rTIy?EP5%H46y+TZ`_;BHkpvi}pp{jHzaHx}e# zFH_KmHUtVW)|8g!vh$BARrhahnZy~;F2!Yi&*nay%DpPzjl_*qQj?$0;lL!mH(HkR zUVw~~e_LK1kN*Wjve{lWat>#|l(02l9GoY3_0fM7t0dK#Vz43{t{*Rvf+7(xF#64( zjVON9{o{w6kn zXL1B~oY>R^J4cE+GaSJ7*D-n)9(cT3yGL!HFuuc#?+DwX;!GM)Yk#wPlrhC~jC1-} z6P2gz4`%BeOL-YxDmE6o^TG9c{?am|QNb?y($f5Jztc~`{>-)U6sbZ*7co((OcU@| z#*UOUf-k%6?=po-webbcE(he~YWe`D+-P~{@DlmLQaA3i_X(uA%{M6C5q`zkd7okH z{=_lC=9k?B$*gNq(p{5m!9+F0GdJXN%qhxbk8}8>4#g&`h#l1rMSW|y!Bvq|=6m{H*4g5?(SYx=$LnXglA?7qfs?H_$-`?DTf={Eo);PTv9X5_ zE8wlORnrd<=CY$J1?#1a*|%MkX>G|Uzby{ak6IhN7cIxe3mb5r+}Q2x=c?*c0Hue* zf8H$&gbmeo*+64Uw<3km;1KC-96*+-fehdH6@gSY$OqGAwbBI1&-=G205p)0CvN%JKYiL13*P)(>kzQa9FHOY zpSrJYOw*GB8nrnS|C%oC_f1>S-{%7cn18O<6a4z|hL9XEQb7>1&juj11=Yg|@d4jX6q+7y!##u*89ZA?43NN;*q=fRsz zE2;a|s}Y&SHDZ}~fv|L{jE!EC#)_s0ecOIrZoqf%3;k@+WhIQ0{xn~6?S$AF^dVGv zQwg>mx3`^4aT5cQHtqYVIil>fQxCsml=Cf0yzK7@;*ZgcC)cYQnN%ispMfMyG58J98b?3R~7F2mV8iyc3qagIu1 zyE6@>@1mG#BK!MUqJ%+xmQH`W;Y}?WRrN*kd=9Qpu>-}J!dKPIdBLb{wzd!xK0b}5 zrPM~`({xh5JR6>1=W~8uWeW_I_yMSu6yr4OG%m<)d0Z>rqUi%^E?FeE{KE^%pvBIe zxx~9Df}vnQqMJ$@qFb4#TT7~5S{x1-0VjBB9B9rW7^O{0YoFPHl!M4nPyPPYDO?O> zTw(~pQUplxl*>c{KhQ3!r&O_7w9x64c#|!xa(T+6;*kfsc;=8b1+yh0LDl+`Pf$`S9ELJ!n_X4!71BhYojosi&(X@l(yBPJW&v$~T;1 z3lJ-oUI0*CmY3JP7O90L2S0Tzv2HFNMoN~PmmcemV9vrzE@hbr8063W)k}Ule(QnG z=5^G{-BFWD>L{cYn;hE6eBAU~`0=|Ct7E=34zs0ZC(I^h-&t6Dddcgea_+azUN1w( zQQ81FE9=hAsxXj2UBRx@>gzyXusV_sH70ib*7e}DdQ->icz@octUMZ?^5zrCHHs3x zA&Xo1JR=RePC!ABPe=;ZHpu`$av>|X62kd*#l}>aVY_njNCnv7Db`h<|3pa9F#YX& zys%~@5CxKJN6QuDL+eg9`7O$;%Z_SHw4~vLkKWFv)JiG1h^H$BExDeZ;`N( z+RW{f!HLN>EDtYv1Dj(CHw4AcJ4!b%$%Jvgr0e6}KeiSr6CpiymmWJA!<%u8bZ+Nb z`-e_>1+^W@A(Eys_~J*c>jg2L+?Cgrz*i7~P)y<@*5rAHVlfh)))8cmglrBt3LdJ) z4Nm)&73UoxQ4~2xmZ8H4nwJnnL(XG6?K6=wr0PbOq^Ugqd4_$FNli@i$|ozZy(zf} z61tiGbV^$M>XMx(J@~s z=8Lhf#8SYvNx(-1oMOI!fD}Fc?h#|Xs-!9ba|%K!if0@uuX!KV+;EOYu=a{9%sx<+K|2fhl+GXCw9wz_d|r*GT3b zYa4efYf3XesqgPlF5P^{!YDq~L*I*;z~|XZp%AD8)u)u*-iP)*K@T|9}BW9hLvFbbb9V z%U^b7#O@YEw|P$QO(1R|uIG#&V%AEAGkmur_Ha{pXf+~J!gWCd27{|AumiT6G-$2 z!gysTMNQrbclstRGFh*GqVHd9%pK21(GaTAX8yo2rr$f@o4-8-@>I+@CTO z`RMIm3uVVh1Z#mcg!u1%g1ssBKdKuRiQoTxMG;VyolrSBUy}y1ZLHw@YI}{>5Ff|z z6JXiXf2ocs!vSZdWy6|BREVLa?jGP zy9UIxnBpVM82`M=8q}NLtu^Hc@=u_Cexoc5zK|83b~Z0RvET_+5Y&P z)DKd%-!lEE>Z(Wilg)q z{Vy3_-c52?Ryvn1U3cMJq{1r7)%6X9N%|>Ualo3k!*}X4xKyCwH3)YeVGvLa~?^ z6~}SaPlFB5#xb8ij5w~DWL)q!yl^sM2mS&?vBD=Y_jN*taE~y*zKPKh>$bH+57G%l zv#05~m9+OGEP0iEX&Q*^$d@2ev>0DCD?b$r0;dULCXPs~qFYxJu(xP0 z3m3Q}e_lqKH@)#%dIFmV+W;+kIB}hL-+PIgp(Y0FIc2=xP^3(OIefZI`z6}-$|AkS zjq;3|#(Tf?lj6fx(DA`LTeJ!Bs5@`RNILyH)xI_ zPXlCB>qqPO0Z5u5WRV#yB-=oQq$cpk7@AsI9}=I+q5M7tBNRwzD~(kxI)lfok4IF! z;XXfqsE%$8E3A_#6eNyQ(VZaGgM9CHq4>C(~n>?<_uMH3miE%&d(`juCs zEUYqt7EqLuym-C1oYyVA9^c%1wkCZZwv)K)vDGt0qF@bE|E`L)rFZlH-N#=Mi^BVZ z|Jo;S6h?Ra?{98}J3YhJ^vtI0;e9a@=@XnA`JRTx=8Q3ps-s4E$1mcK`~Du!PphXa z?IB`kX$*hXsE_?){hFfOZNECpVp_5h@U!g}S^l}08s1Sk-NM#YIT-N$i ze1SNBIust{{oURf_=4zpMp&yiP9bW=9C6aKjI#KrWV!EAh@S$l1Xr|5hu(4&T%rQR%(j;T7m8o8f9s zCkukB-FqdbRjpAh3VIO?@Bha0WztDD%Te8yepeLkq)e#-Y?Yo8u8vzY8vh#+DdY5l zu9d6o;hf<_cAD-aThW2?Od9c%BQE;2VtZe%hIH2z#n;D^{Q{wzy^9E}35V;w)hF)) zTGP#IL4jyJ=LqqXz!tQoN_QDk5A-L7AWowWswjng6-Alfw>jd0!lB&J`(KttlBH#p{$hLn-v`K{aNTmsFAF~KI3Lu&Bg+Y}^#~Rf> z3&f+XF4G!!hseqj{|+ZQl(sB)oa$@Q)So)=g^cW4c61Xliumq z)}9seef+va;36|jK12{j*V4$aNx*JPAH_*&dSR&Wr?fu5Q5k^Iy=`&ck1)6!(|;@F zY6?{^T*u8)_SQCsq8>17A!M&()3s0ja_%L0DjZ$xJ1 zMRmfm7v@8+N}5-K(){blwIghbfhEQ@$+iu&4g7{LK4S-752>>t_LkNsO?nZYw~(Ye zPh6Xz=YCK;2(U5hmA5M$AztkM!q~kno%N%QfcG#V`j~bF8iT+bEhov)AYBI9mC#(0 z{vMbV^rztnh!Ph-P)uk&hoKFlz_0m~>aGvWV`yJ8?LGF@J0U%m*ZoY7?#7a+=FU9R z&sN#Y6&$y8tC*yVL`s8NVtT9ZJYIt(7R=cDLz_? zrVpUp=;#GS-B}8Gx;M(Y*```~hZ@BU>Bo6nLB9;SQUcR+BrCo<1r+snRe9m?pkP=T z0ft)iSe4KL=uKEwJIGw`YnAk%k%@mGvm3}YSp6;b;(D1VZiaO-SsyjFCB|Lo4e;W5 zvM4uj9BA`sJNnX6+2tyh84CHzpBzEMrvve&=04Me3>79)pchp0C8=hF$vh$a76o|Y z<0Ebw=hFA|E>uuw-w}4<*TJmfS2uJWQ@UhI%u5mG01Dz`m1%=j)YdpNplk9>v+!|% z_eiJu#M~%rtUvt?QtEMP${SUQEe54w>xHE{09HsGcnr7QmC};Y}Tu6~^tm#Q=UJiqFNo#96u_WJEs zudiodhWOu(J|OW0Z_I_BAzpqgni?Ds_vm+gd=h3G|L1tNA?^dc68L+8*Ybv6V>438 zGhyxVZ=A~0iApV&_+Qg2?KqR%C7-P6SBO2gfymOJP%mk_Po|1r`y4wGXE@KMs$B}v>ZJCS8t(xAN@ZV`cvXXx%gcO1kO_p zx1{-Mh8-v3blXIZ3kx& zdXm)Mu2v4eh}qzft}Q>^uM*hs=&=e08y|Fxlw{CY_aAgY{DKzVUQFxmZb%=Fl2rL7 zoK-3e@(0xSq|JDk93w{FckSW9^jvRA|sJ@m{up=g)K}JPjqyfCFa!gj5%aE=mf0wXYRE1b(552D({z;nvTq5!0ohIJQW{W>;}@z2Zp|- z;PLJsHl-vZi#au+Zz{<(3M5>Q{z*~oXUIo_|KKJZY@@~cF>fyK5V-)PxkM@Ey>K8r zXM|h8I)L|lGq3zGK6K{@J(!KOlg-REWvQ`VxAZAV(WARs25PD;56f~FBhxPXRg0dJ z^~T_9lnv%)<{uuY3E^J;;2Z~#@~Pai7bDLw9cs_dY9egpfHc{k_H>3r#n|!7$pNOnH1T3>-BRQx6MODDsn_M zW=Nh**hNhbE_7wA2B1kOmY^iM-4ffEM@KoLs)Kombuq0aU3}!zUM83cO46byVt{`s zMU9gC7lpqxv&}V9$*E{5Nd-5rOoLAc*P{C_W#vb21ycerNS||sw(xAMcSB+`S7xx( z7+gnSIKC?kgFNmer-xRl017BOeh6mzgacBCX-lTwdM-a@?F=-TB+T)pEm<=cJ-;?; zznG650AQi_pEI=5q_ia5*;qmheIaTj1coEU)3jg5;~jMc*XE5*Mt}LZMTtqNL{Ok+nsZHGr@c=lz|~R)Addi^*dau$;7et4T^pSwi|IMGegn=Q zLdMsQTzFF!S5^4$^YION_-~*xlCOoJhW0YwrarMST+WLvIgq|82pwXKMyoauz2jdH zxW^yMIbs1c>1b!!5)+RoLI~yJv~*Tr|Jxh4xyh#0dAJ1vL+nLDs8=gt0$?Jv-Sk(k zT{T@iu&1hI)Fa0%!kw<3)tL-v>SIKng02hLN%`>pd@fE-k(f^K$eK+~0 z#3java($t63WtZAr@;Sa^ua;GOtRHQ-G8`SV=jPSRA7Kjl&*Qz7-5)5T!KvnzXhLO z3p7u?-sWpmq;pi`9gt#Hp=IZ?{8sD%>PiF?RNux_2m2tIw?)q#RGb)^33|u*ofs+{ z(lreVH(dM&SWK$AWXk^!=e6eldX2?8mHDRt+0Vh=+@SnDSZ=<#Cf2Et{CIief>C{$ z^^V9t%J0r(vg23V724WtBzd&*r*f>6N1!2(_}Q59U6TL#c@BK`d0~9Ozlh0ITwY&P zN`IBn9V5u@EH*Cvb!<=pWjmP^;o zP(nFTxw?%pu#S*yL>m)!2KRL>7XTRe{7FbvV|jIR`WX@#UfF3N)K-a}pglMVsEr!4 z*a{+&w&O*mL=;7})IRu}U<8^eEs9EMv5nKtl8ACRVZXD3eM95>S!z71#^}yCAROff z;Ato%m@5o-XG0+XIeb#&gG#Tg!VU;)@iUQXuqVJeLNp`{9HAui4il=J)mX*AOo!CJ z2NGHMHi)_O5k?|ugW9I4V1{SVrFu(x`(~Kd1E0siEclXpHA~(B-({^TsaV> zKq7Khe%X%7u8TQQN!neusy{j2&RicWh5B^?nL)H1XjP=QYaKWh%dCNnV=pX2Wlz?M zD5P!Of6Pw}4{azY`fPua;ZYLswh)mw5wv89^_kQ0e%pg~`Vgv$ZqR>DIhs#l*+JVv zYlj;Ku4p1Ll;Rm}4R2PpTr!~+0_{Z2EeBR70VzBXob8hPr3^lnAxhJ~Da^wK*+fWa z29`*F@H#BX@P$Hstb@(N1o49dtfWv|*X_7iJfpC_)PeXT>ZjH8yVyRrbE-a(2 z^0Zgmod0<7B{%i=?hg}jp2{R4fvVrgHSomR+L|~EY#+5{mIjAIV?bctF1nPgnN29u zNb1P&k_0+HA>12Q`ERxc#jCxG1EfcNM9`C8Sg0Y}7OOBfs^t#BckIwJbWT%IzXKg?DkFkg5q`V&e|kghfu;F%@C(!fFn^B9=q zSGUd8i-wu^RoX%>`Y86V36BvPZJXZNeHf9}mI;XyTxPhojSYeVqRf_x5AYd<^BmaU zZ+(WQLGCtNN!x@nH3V!2+vIfWa@g|pem;LNe`2j4?&4aqu2(3v1d%&8Y%;4BgLT^? z`>{-L(WgiX+z)3VP#4K@o}wc@yXcZ+idB?h%jrt478hiNcSQLX=t-%^2~|0??1_8G zXW8*7(SJ zuHV|8v?%2*0egZzR!}~5`d{IU-JKZ;7B=E=Wr;>@SpxP#fn%=&L4Iq#+MCJ}ho$!5 zlZ1g{QuPp?;P2J-%%NqWqtMHM=ToJY-o`&fOq(MQBr5p29pQ4l_OegHCjBhA0{?=;vdP? zQ30mJmZo-U&d6DIAkW1d9Wg#9F$(D9rW-h_Q%a^OaqJs>CR?c1ZixAz%e`Cd<-OXc zf|5snrYOxXB=@(vh6O(AJPHvV02+5{?xkEKUeP7Y*0p`%8?$P&A&Gma<}UrNcMq43 zgq@EXjwdl=FViTeXst#k)5YSIkJlqokO_@Jjiy67YrjK2=NE!ys1VWig`SlM2uuV{ z{ZklnN9bISOtIMHdxN@bp)K?~WQ&cDDvr<*7#wT|#NXZnLKrVm${#jI+~hxY57B^P zyd(-u#qfHVov%5gS&=mOLNcQ8xl2fIH^=N zp*vay?m|bw5r~4aUGgD`p|6c|wlu{MJ5yG$l6>Q+!QvEad5p4bvzq#(8PQ z%nB`6S#TTE0mRZa4dF8dZm07tcRX(o7Q1&w>Y?4rIy| za#>i%$AgWl0%>uE{w{)gAtbSy+QLV!3>~$y1Y_DK8sX*V#hFR#!MVS4=-z3u z;Qu^RV5?LANW?8XErUE^5iEbc@@y~b3zn@#=i|NL<=rm#EOe9Id|=AK~W&zFB$u&A|d< zh=J#^cQA?RiX}@yVEFf^=|zA$2uP~n?K+$&6db(0tLwVV?_-7O3DigW5IBxO zLtplV?Ccs_F#x3^1)S<(eGbl=iXKZxFDGE&n3x~7jkBmz$R!zWS=yNeVz+EuPhb!Tb{Jb4sg7Bu}0gyWOvG4INWx3%~-b$#Ob=lGoPMN`-2% z5-&BQPdXV|xPIz|Jolan#ESR^z94)%q^!{!Ml4S9wSdI3M~n59#o#IcFY2flW#h=`9%;1_07v!c16FJh{Ba* zDU;JNyRXnh!*~L3e07|@S&($?b#W;vjQfbIIwqEMLg!O?>-&%{lt01g7(L1Ja=Zkf zbaaEhA+aL463Hx__rFz{?D-rXct{m^o?PZmS=9i#r}3Qr7&D=BH*o_?uH}vaZD=PQ zvWr3b-MFv56~6(zEpX3A#4Pl}OLyT+$`_vVV&5`@l|0w+Sha%UuU}8kej*yOWvS|m zN(j&y?@OuW{W+YtWpDS0e$f}-l)8KU0d@EGok?xJuqVZhZvmyoMdTJ)lVz;02w{zq zg4Te1qsLpzfmY;8rHVFazsH`=IARO6r}yDSe~wRD+mmD_6~ob$v@c2HQeqqR)ycP< zVgeSnbvE>ZW|N2jA)KZtD;rVfpsAxZ+lZ>0Y}2;kP{V&ydYTXXKiswNSGRfwo@K`1 zJ$F$rUVd|b=Uz36F+=fKVVW%E)}7=3j^`KLz5f*K*9$dO5pQRX>m&Ws8Oz#?&Vy9n z7D`?QdT&C-oWC=voA{IUR^!l`vDfAIf+yJieaf0!C-bP&BPxG;*#7}1LD;_d-^4kl zZnkts_Ylg?Fj!F615>FeGoian_8G4k;Sy9)zQ(+{i?dG||0Y65nSKD_Hlk;^wHtn~ zic?X2jga0XF5XW5ICC?kinhe$pCSF-pZ_c9@GLT%5Qc%}$=7jF=Jopti%9C;gg8=tq4?oR2tRT!_6B>ee*2)Wg8JUYvcGNN4#-yU6y#Q zgs!5g@jz>Jd{!hxGzr>`K)~vNWTnndNOgxaXl5+x^m~otnrCU$fZWoh-G0TYg7wIb zP3ykdp#}B`sTrua;!=rbxHopLxMD@H+<3pZKqqML=Xcon#nL!U(v=5V`54pmwLM=2?acIY9 z2+fGLDs3`bni&y7D=sT@q|AY2j%k~U2x;4rp=ANqwWRm7#HyodMg!nN&EVFxV-iJ! zp%FB-61!%2?v9h%jNV!WvsN@V?1NRKYy_J@yMm-<*oFk%LBq>PXOvCZm83NU0tOZY z$C{rVjFg7g1{_xEEp)YVPrB)H$@bh_NM!b2GNnP?6MA%L=!~ z{gqGTk{VjEjcJ#}7DHaDqx zO|35z{0{Tk30|MX{rYb7{-l|I+^1$C^un-um9V_TEjIiKsWD&Uigb91ZeeBsX37`| zeJjTU!)!vcwB+*QK4q+UUE}qbG9DA+C1H5?vRtZ!0tCg+OVAEEZ$O(6Zl z&m`{X9;WpjWnu3oE0N1r|2kfN7>!2_Q(`NeC)&cHPgJ{TD4zayD=;mu=k^XXbOn%FLtqYupXOOfeCa}7X$bHJVhubM}`FIaN?AQ3g zDRzFFe3{AR1#atG-_>_;rSGgy{Ze2&dm543@VRTm*irZ3Cq6U5=iY+tYrOKGy}^@T zd#B>b{9qK|on)y3FcS_)O=J{*8Nwh0n=}Ldjb!G5g=}K4n50DG;*_-WQXZh}< zVc$+^?;i8!h&bNlFN(qYd+&dmu6vfaSh9cfE7-S5JeT!N;FypU_d_gRqOr9T=LgrR zWrf6jQanP!3>*^Uo8U&Bo}!E^YWWIb`qyOtPjDNE{a+^bH!u^*bWF5Eq|2BWteoLd zhvi3lC2;A-;VjU|LOD{+xw|ccjWgkk+H*vnVa*PpMO5T<8^;h z`s8Ong!CV!KYR-9eh1NKn8%yAB~(t(e*Fk<5DxE?w=3rL4dyc=9K8etYF?7M7tsAJ z$aCDeyvyf*>MMNwXTAY`i&<}=wrECI95H)946pO~fAn~|u7H^>a2Kkob zP!P(vPabbW85(;uyh>gVFdgBdbc15!0tt;zYL=h~utvK}Y>L#pMAV7hx*6nUiE6^k z9vMznBO(zH+;%BiS6m#41##IKXw5lvYbj9;j`kIuBQ-nPTWhU1Z3u#Sg;ybj0ZHKX z5Mpj6P&;&j(2B7n%*5dWjYgS{D4V_39xjQnlr~~4LhJ-d4@<@rF$h7?-r!+5;oJ;I z`d0i^XBaP~F=azD9UtO%DT;Q2mp~aSp$|Ls${MqXPd!!G!Ec?=8-g^3@(flLpJGd` zno~!`O2DjP7OXTB#eC6*Ix*D7Zk!aNL5R4mP_=f<)(+fpl#r(8V_CwAwv^+-(60Tt zS6VXoAyR^tFphh5XB59(9a*aqN@`4ES0pw=)T<-av7I65p1_(JSj7o(L+A=ZLE^MS zQAE^d%vlMlEktNCRG?&`ilV8d zdkKQLp=lz-nR&Bp?>zw477$uex~1OZ(pqBL1+&oDuREhT_E313V?a}F*aAT6 zMCiAKB$VwD702q9kT%rq3d@ay8{!(mx*Zknl+bN)dNc^VOw8Lv9cSuBnEiS~ zbPP)LK%Nz^0m~JLEkP}B7n+tSpdk=aGq#b^Qoh!>-;~tO(c(LdP$4YNAqPsG8~0Zx zkb5A_h|!P{8FBBZ(+#9Ew0i=kTMXmxumy|{Q8&k=z1QjY&KrK``sL&s-{ z!$cfjBrU!HF;gzV$}LP+Xn%pt?fiEOq;yJLe4Sce$Hz5h3yK`!+(Cn4c7YBH`n@}7 zGV*u>Yyj_Y+XEKuT$qnR0%tL6p)mylCSgEMLip}q5-rOL` zNF2@{8yquBSPkGaWqS=PMxO5@T#_%|;8&l&2hYFveZ?m}vm`F|cx61I^Xquk_kQuW zX#L863iU_P;e@n0r%NN%A||JF-Cc5CP&ZFfryEr3NW)$7_&%&Yf%M0OenG$5V78C2 zLZnASM7-j4h3AZ?4I*H+M;RA{bVeAq6!Ma3b3(Ve_t?HI>to7hh06nCYUdkeK^c#*G857nad>>u8z3#u z7|)->rmqwI<^Nyr3;!bN@BQpernA37T)mIf-#}vzGPlKQ>&SIO2_sMb?t*@~%8O~` z001BWNklZf|s5+M=amL&Og9UeD2@zD}Vnl9y0GrSNe}kR|fBuu5_jE zT_5`wJ)v7dnE$=fyARcMYes#7sjj~Bi~pmLVC6Z6gZD8UeC^wsyLEFy89#{j3*ZeT z>_J(Q*C*6@NzFG1dPvy6B=&D%bq|||W=btyBCii|Ti}c9O!dDe?alP72bld4F;+Zn zsM`}tS+w$(bV=wP_apMzbCh%kD;+gM&94#{FA=ENe1hd$!0DsU^vQqzB>1z;aFfOZPyWdo_@DZx3pVj1413R#4*!reOiX7zW&0y&cmf>?v3s1$ zmlkV4$(P5Br#th=UL!2NK|Fjwoj=N`&++kJ|2n_%V-w#^0o-&kpKzloqUi*Qgrx<3^_8a&!Al*#&41?|7ZG@yQ8N^fyO{w?x{IZ z&t7KQTqRGh(Vg5Oq#Yql*b5;Fk{X)a8CZ+BvAKecP@@ZDFo%Xk^Kt#jh02B?RUOSD!v4{U zwAdk%5h|rDh`CYQ(glkeAyn$#krEzQrKq_Bv=Cf?jm%;RXm3b?ItsbE5U!oj_Lr!m zK|u;L-4 zW<9Z=urX05bzyjDs0g8J$i~;LBzL=)u|lgza8_~%Pk7n>-aEfqQL3-=rdTUTx5K0$ zK?tciQX)?v8=4+N-4mB@K@F6{OSEzwVdcF1GC?>c#An#2=Mec7#s`ln%jc=P>l_|` z#9i1XO#2&95>_{abV{4Hv=l%pp}WVlA2=LtVdXa114O>b>e?C8_6BW#&g$BiIjt+| zX2SA)LjRCvHwiI#0H-F(cmlHWpu;#}BhC)iNXY|Md#oiS7V1H$=RK)KsuiqdunkZN z%X4D4ppJo7&cGg1r-;acHZ72NaL&6y%~vUhf=G{X56cgcDKMUY9qcRbJGPDqpZUV` zn0^=y*B~retp;-WQDlFW^E*a&{3eT&x8L`^ZTsfS*xueY;=0!a@XH{2_VjsQEcmkAV(JE84gq zE+0S>%3)18TmrVl#eGuWQKuWU=_;)n!fW)G-se8ADbx9FgmbEHk+6C1V=n4XU4Q=< zUc}<3Sgc~WVaUYw_PajLT43D1gobBG%hxF764YC?cAYl)xM3Prq~iyKnYR}7d=53bLfM)N z6aANA>e0)83k`2GT=|bo+m4Ww;nHiAX<*u1!SV^M4nRlCM_7Iis|ihCqw0NzlRNas zk16{VYcG(jTtV zzIhw?#`oR`6Mpv#KSkgD2=U}OhUFV9uKdZ5&3Y5x;vunH`wATwjHSUA&f|uHXoP-4j`eDj{k&Ku3?8~5V=d}|0`;Nn*Tf4 z-}La6G!ispK-n*8wxyN(1lCCZHpUjFn^;(4CbV2B+7sJ`Y8larn|Ot0?h+GEGp9u>H)M@v~y=_Yk`oAuxNzf>YQenNY_L#NEbuD|At22Xv1}BB;%14x#MD9qR%idEnSIx?47MGo}F-mR<11 zu^{3iuSQ=#Rt;6CrrtoPZAL&e(%R5rA_M{UXv;1hYCsqvO|+>;RcLh11(t@jHKn;; zBw9SM?BEXXS|E>(>kxPZw2;`AT1#-T==bfA#a@XOE9dsu-!5U0a-(tIxNro z-)BdJl%7&M7yF;!nBx=NNpshGUAO?+-1TW9v@sESBMj~&rA2CXHgq!qHAF{u09%KK z1LOo%Xsr@r2C1m^ppIW?CQn2D^3I7K6b~8zq&PdBg&{mk8z_6lZ2>nFAjO2ZwOdV7u z^nOlhcbHkr5n}K)Vb%9j6E#O~mQ=;8M{V!ht>XOWxTof0^5%!2UIX!#^SP)`n%N9Vf7MqT7z_sxnP4$Z($Y*T_f}g z44{{p_Mc+>lTT8tVP#Dh)=-aW^#tF3nu% z;2+mAf^M+-m^M}N<|b))h=v}*+t{R7xlNS`iI1^EPm_*XE@5>7yNkrHA?=K|Uo-9g zgi`;IU;USN0dLIy-iNO-O*aT!LB8;#$Z(CcI7Z{-=U`h9!iTByES7Jue|VqW!#gZ* z{WksLo%^@4UomceoV0j`{?Z%oI9OW($%wdH)tWcf!Xf}8t-gS@k+=ZLWkeTfnw|gZ zEP887=tp9Iz~qFsA1M24MD7#24b$N&jF+J;kZwa5R>ae_~N#T`O z-fhGPWr6eqK@Z+o83UpVLRKK-)?e)WHOpy+9DPfk0-@yg@n_5#nVKE#L(%A<-u553B zUPZe-VX;AN&v=@UdWrGy0{!Zo{`fWIJHj+HZpgbAsnZhp9ls$D&l0-{?S$#n7`z2Psp1Wk#w2EX35|CwcniGv475zo5b!4i<2)u{ZU9GA>2pwG3b_Z zSOF(K?k^3o;38l;hr|^Hh z?{bMun-_@b7Q>Z0gl_L6!txyBX=QQg53ts#(VvIO3ri&6MU^60Nxo)zU2Tj)JVs*`dtN(OHCx(V&OCbr-MoE;7+9zqXEs2*{%`$M!%k05v|fnPpc6t zLe0+oY|R}A4n=ul`ESh>~UEM#Kf_C<_HZ%xLydfc>5Mn~qktz)cVdh^`h+U>kD@Ru>wlk)oxog*i z>0DgMoLR(@JWMVOF7Qf(Iyt}B-L>=Qh2Rd_<__4<1Mg*sgZIrcReMl>9p(JmP@ zgZ9Mbg0dfdZRP<<=gh%^5u`WLO2CwCl`?v>Ds~4#S17x+yTr32We}hRtgR?} zXf+TzXX%G-Lyc={i4eE6{mZnrAjE{GJ)yfx9hZk17W&BXUzDrv}wCR!9?H4g<0OFvtr$LdJg zZn2X5nh}Nm=nlI_9Zcsm8jXU`Vf8f{Z;{fgv>GUfj&c7rro)%{={Np>Uj;t>v(NHd z;YUb%i{<(Twf+b(-9(0fw9#=U(U5*5D!>GGLTks=@fmh^ZZTc`PaIwKdx!nw=Lz8! zamb|Q+0!Lm+Y+-aA-EW9)?qR@Ke?TdhDdFS(U9a}l5uFr90F%evC=!HBxbtBp0V`Q zJfPtc)e@%VAl(aZG5ufM;zi!g0pM7vwW#Y5ua zF>-#D>FhS+gLBd|Uw+SX()tYJRa$u&dwj~@dF39z@yhpRPyc1rXFk6q%5xOEL6a37 ziN*o!oE+X{b^Q7dR5K1wB0>5uNP+1^CuDrP(s*$ z*1qQB-EmDidX+W})bWUM^9(8#9rg@I4+x7L)7cGZBgPWx-$wUaT0LTX@VCj29}t(P zK7MZoWC%Ob@R)RbhSk7ywq!a8u*BpiiTxwOazoueLhLH791+464O`Ok4cc@GsS~BX zPS{7XJWm^+qghALb7KD(Yb)}2LYXe31ZRtyQg?IedQ@7RnY!v21Z$C&BW=hsZwRMBXJp<LcesRd>GDA|8y@Rd+`N5_&XLG!}R4?v+x6AVLjF zh(@tKXo)FRJTz^}K=Iwm`#@aI z_S@OL8e=02m8k|W7pcq6MQu@#-UT9q8Aaeu*p!T^2^5>%t$qwz!v!=df;&dTkSV*~ zU8wyA;|OajXUrfV-(U7NA=Evd^I5?oV(qEho)(Ts*Z0M6-ssSe-_Z2Wp4o0$&nr*A-3|yTJ~+N zCBoj(9Z#AI)*@xIMxqkBK+7vs6wwK*1zPW;=Is)qg7#-1&BMA-unBAPx=B0Q)M2@& zmK|aXR1LHucvzrej$arWQiQUJ#8n|KMpP8~M&8RiO^!AP%2@}sVpct1aX1I5-YfuG zSz+xG8diki4K$q6rX?*4=$0^?yGTwWE#9PC-)E;kO0~vkzi^LV=GS@U72}_L{uYw1 z5f}csFIwr3AH1(v1d#X^im-Y6Bed}&$gpBaXLQG3qZ`f`&u>!lOSJkSH25`&*2ue? zlxZLgkD#uA4iIeg>z{~sK}-UY1S!5|LSJZY0U=>IQEZ^>AJXzfo32A!B4qMDQ+5I& zW9=RiPf@+gVPDA8ee(1sefdbDaFwu1NdFMA9`a2z zT}JyeLVSzRe}kHLs9r_&lyP&Nvj1VqqsXrPCX17AA>{_rEeQSgskm#k$BH7H6Vi76 zaY8LI9WEi^n6#MS5N4c+qfX{G&l}R(Fl~-+sQwS$nJW_9CL#?6aVoA`(S=9GW`Ki%PTzZlp({3OH4ZNhMc;qo87FF&`Ij>E%e z$>+~db}L{-BsH3A<)}sI6-1Sb%&@2v)UABR*+%KBHf}MZn8f= z;j>@-tqZs4pQ8Tr?dKSl&k~lI;nMy0eN5e50=tGLc^a79ojf2PUcwI7$(y^R^?Sc} zb?Z~|W=mc^$I-LT1F!yN?#lbtE3YWO`@#<+?PG-XO=7(4_bS?GO=)F==7;3-yL|d* zzx33Ve^D28QBUha;Jv7ex~M;2iM&aj9+A(l6Z-G}Caocq{dK1Os}yzoWj2co z_T57}8bY%pG(=*5c=~e*1#Ay#M@V;QHPc#$Sn!c#4yZgL^an(PJK5?9VIh9gZl~_7 zm>L>71bI(F@{Pu*U{rkBXvCU*B|wWAvXVR1^gK7Hqp zN}x{Hfd!`biAVQohnFbwJXWq^#_No`j}xYw9v0JWiR%Z>T|F#-B{2mTph`lb zxVz61A`#U>=te{kgK=d+JZjyH$RtF?@m?;oveu6#?DU;Tt znn8;@KJyBRv(wTdEzg2iP+E3osq`|7x;)7k!_);wBhXq!f}l}-6hDDXJQd0`wi$U~ zF5Za?$RYZE6jP)bqUVH8sQD^&N+^?$RJR_hJp^|Ks|p&Mw`^h|_QgfuDNql=krLJM z4KzWL(XA(@O`N&6;K5;$T7}k>v@}9gbW&O|G&x2m^?nGWK^YBW4sydtK}ai}9+c2e zj&CTC_lcl$7}pcGEgq1j7$z>j~IB zFF;+;Dkw@^8DU}cD|2no=8beMlsbYH58yLqey}(vI4qmHd}sG^orJPiBp50_K@o=N zj?q-n1wKLZWaaJ9R_BqSE$(R-G-Y!G-+R z0-6_uw4)TGO##d?1LHQ#*I36eIC-H_Cm1h_cr+zwrOu&n5}S*< z&ByqiZM;Rhg33r3Mp|9Y&5zC-n!%uRH+fo4m_@8q%GhD`GSo`m1}sNxq3)cO3CeX$ zn>S3QxNErd&;r=O1!j6=#wv&enhwA~D+fdmG;7rA=TkXUM10K#$<){ot(duBR!avW z#4e+K0aNG8=88#mPIp-_oxecnA3zO=R@!(AjY}jRfI>MO(+VJGbjt_+sovt;{3^ud zIZOExW}jex_$>eE=ad!)vivA1f1L6DJ?hyW32kQo3Mv_mJG6UD?B5CHnzkn#9{m(? z`4Vy168krZi&G>W=+o;=uV0^;u!$}JYFJy+az~wx{BKyK*mw)-zwy9ZK$b1Y0qEi~QJb@%wG*X}>X68wWN zeu6OkEzCWn7~5Su{Ple>owp3d7M?$z?^&*RaVD+0)t+H{Ng1kuH zZm_nX$`LV4PdWJ3B6Yfi*#lyK|LJ{}Ri)%(XiKzL!my!Emub@Z`C^SSHsT^8>K6x5 z$Mm?!D-aqX7HSo&4%De(B~Wuk1P>0N=&{D8=g|PevaiNet!E|5vyT9-s`M>_z!}tCB7xnH^ZX=6p^y_bZ z*T=N!5+X;0?%VDI8+6M@jN>t7d*yuvUVwCbkGy@3the~97I=T=^NHUHKTQ}uM!$HL zu((e;{^O@e2}Kw;M~vGW)c!V{U*aG9{2%Zi{`q?y$rp7|KfrY%@LtqKUDThiy7DU1 z=|{=kCy3jJ#MR+@KM5Eg+@$O$EWiE(b7Qw)hm`|uI{B`TsWqX*nZfo>wW74TM0H1Z z>1!?)OGnGe8K$@>YbG2{U!?AjQ61h9IpZEU0Xv-;nrpN_f1fSzig8bt0c^pOK(%qa zO3wEn91+&f(qH}R_ZD}bY(K{0$~S2B!{psZ`Cj2=A6R2Wmkp733*{kX%nDZDE!?`B z>YY(85y7OnGwR#7aqMYrqL%N0kT9Wr@fIOmW;p)+r#~+H71Q=AZ8~N;{{-FQIi@KP zigOf0zjgFLi-i6LA>BcAqU>%V9bg*}!&*<9POxc2TcPHF#61>d<_ZsB6aD3{Gd=TgqKiscyhh$!TS|yb001BWNklm@`_AQ@&mth`OF$Fw$(^9dF=XghMW zL3)eOt%%(ou|}PaF;>K3LgImZx~3M;X0uqOvlGf(FbGOM>u44TNeIh|N+oZdU)yXV z^&707VBr#sOGFeaF7VYNs5~{n&YaQ#YZ2IEQZQ>SCW;~jXP*`!gz5~~7|_sMtdl3O zg@%6%chSt|OfQKBHoG>T?8+QyCFWwR|3A!pS@94y5g&OkJ(^At8L1hddJxx)87Ng~ z)(GKAIGuQqmWey#72+ZzsY|vs6Z6Rxqamm^LT`*4!HT&ku1W=3R%$VK zXRD(qf+KLDRaBe1KKrzv0d(z-V*|>$ z3%nuCLU2`A(QRgw3O>qTDpm!FN{R||a2E}<_qEe1)S}d>QM0?upC0#|s8e&{TMD5& z5PC&}Qm5(;a}`2zVKBtVwDsc=dgm9H4Ay#F^&T-x?m*7Cz~f5*4Tj+FO{>0{lD-Y8 z;Y-#<`Pn5uU#KZLVMU;NqYGb?0wH_z$Raf_(WHpj+!~FDR5YxeSu7*8dr(ej zrJK3QU@fD=h)#*JPmB*QJEy`FiJ7#T(Ac0QT4|8Q1Ga{VW(%wyQMN0}ID&P=Hb9C9 z6Kz_9JVta!T0VGs&rqiUvys@lU?=earli$RBeJ3!?jr3bqR(UDMH-6LhU^Z+#VLd> zmiOdwL*2ejw|EWfcAVY&6fwO-TJ7i-2Ws(k(V8-z-kif7XUt46Bm{yPZCcRE0xOM@ z9aIuKBlL=(XmGAL*}x)UcD$Qbl#KRJwNZA*Sl!cIzDpZB#`CXne)22)#y|b;pjyM3LOS=x;^8XBJ%mSCy~=dB zMz=VHdWG@)!%W*kiXEyGtsYYk&rqgg%r>Y*th&2AqzxjCn%8LP3Ed7*$LokdEROq` zKstPVh4NJ7)mW)H0u@saU~Bb2-YYF9c!(u~RzDdg;V%$7O&s|!Dyjh{$0!<&HTUA2$ zh|s%h)EjV>(CK(Vu*YVf7|fUZy@e1N-$0ChzxAO|GDEO;~^T`}f*b zC<`>b+u<9#Et6bv*7kR)xZEIViE!yJY@5HYGU2T+`~=?Zl_OKAAgU(uHS_0?y;ctokUp|D7eV zfDE+a0@1WNq3ja%^qjO_qG@_pP{a3Aaj`{1L&J6c-v4&Y|M|cE+0Y!HI2-xx;{#nD zQ1{XKFSH?Y4pynTe_v-UKplIq3F5ogY1v<S@MCW350-r^_jyvZxC zG(Pi%1z~Z7%|TI)8EHmsBW>DI$5TT82+@SFLgPr?KLg<^ zU4KUG-=IuO|1wOVmMx_{q?r(9A}t;hQ%@ORVA{Tnrr)Q#`c=x~9r^KXbaf5sM#}y& zu4uZ@ChZ}@>%cLZ)<|k}%X=7xhAq?CG4*hc4qJ-(D0)~IT6vLfwM4qtJORBHy5V(7 z6Y}Oo+H@N$Co}@e0gD?CdnyxczhYRQ5w86bntP_*PeS`4+Wrz2mhL>4VIG~<Uzh z$Z&ZNu|n&qIiV8B+s?74T_Y|lLM87bZ65ItF(R=;lfpO2yZP)QB5I*)=E>dekn_46KSND63$yfx~igdS(gs-R{R3;w#Hxft0&KJ{_;)E5^J zZRWxr1heK&(Q34iNiiWPHH!;0rkI8Ke3F@M%Zv#&taC?h&JeO9Em8xt$-%Q41c}bV z4V}AhMctkGM2q%-H(V?NEH&C*XjLd-7Qq&kZlUv7KPc|J*1i#`9{5&WRGX@z$(+9| zf+>hh#N{a=O|%-RyCX*`m_IJ9_lS;IT_BbyNl=Z%Wg{#`LN_@-vl^!Ebe0;i9B5J~ zqF50^k28^jI`?}P|1qOgKX$6{gpec5d|U$}Nb|-*80Jod_~R?k#(-+)4UVBe&5n9# zhV2s)Tr6uNE(D7_6?b)MNPrNOhCqwt?Shn?Q!CO45khCw7HLy(r+HIqHl`YV--1!9 zV&xdC1&sw-Aa4~!i9;hsK!tp8=&)lA=5c(TU9H6(?rk;_m+A_yqZ(#*bTA}lcUCtS zzAXvT`AEA5Fe+sVonkViCZX|0E*j%pA^ zrhsY0EciK5df&!)z>0GwB#cPfLFo`yAg7oekPr}QE=-p|umK}dM@1tZJ}3rJsX0=n zCwLr?7zvt@GcLYNE`Wv|7CTy9qjHxf z8LL;Qb__D0F+h2VGIcb)LX@i*JFrX$3DXll2IW8*&rv-f5`k6K|pC^8fDS3ndP{RfyUq$Wt= zJthX}3#2eNx8J`MLqkRx_}njszcSA4-~ZwZ$m(VE=nCswe~9P@|J`nVipbZYeU!t4 zJ)imfTl~tucz+lAMP1YnSX~Ib7j;n=_2;jD_s_n#`^5$0!%q<9CA#CQh#sT%EGjn`#}hQa!RGvw z)W5;{`dt?)n*$}y;=3>kb$5c*Eyna+MOiNogwe3aYX4NAHBk>e38(H}%}0!zn{>x- z&ivAMS8D;NNG#Z4VE_0-OlMz1!?%$1WAy!>O`MJW8F8^i^e?KNI`rST6xe~((X z?9WcujI35QF#vSvVS?2h$?9B zVpN?HhsS|#y&*0)v=*q-iabSPY{Y)+4Bni4ESXAw=@C|%i#8k+7K+y9LgWGR4$5SZ zJeW&67a40-D2@bZ)yK=V3T^LVql7}Y+LHH+*;Qt^1es>kfH5ZvxsSIo3wMT9Lpv9E z6c+`b9B;wEii^i3VkMz!SSyI^G0CVl;4+q$9wKFeN`L@m479BCNUv+kvBO=66%R=( zF2;(?&N%ZhCsQHDio}YBFpDfe28|PlFdcdq9)n>87bj=b17?KePG4ho(-wbhGtd$o zC!j!3(%>Su;}e*zw{BpR*<@L=5Q?FLk~%|zKs2VkyP1*bM>tkTO=QDDMNjJTM^{TUNXhz;qX?H$KpP05GI{*llfyHHM0rcLs|H1kkEjk7B_ zi~FF>&!JKTvQmRGnIlY6FQ|ekv;){=-UOM$1s!eS$OnIY1rZM?d=do59EO=HqKasv zR)tc~?wmLnqD!obV?s(t*&iVxxf8i66!DO(-dWimv=E5?A6*|PhXvE-3T5Aj={35O*Bmu8zR0wBnb2J) zE+5X=5d{w9%?Vbo61oNb(c2u(Rrs<-&ZU-&d7R39hcJ|W%Z|M(x@eNXw1{_P18ZV@_t zikq@BPr`^{wj_3$5X4bmng2sJOT)>HTQ7*Bs6^tbutfAu$)z5926;R@r! zpQgX`Q9^%-w7f@Ly+*UwDVra|_y`(KsPzghA9MER5sS-TC-m7vDe6esuLz3`Wn7VW zD_VJo1f>=AvCj9Yn%6Yj5{C!m^A*+{2VuFVkqKcWuI_Mnyg>8w?(SLMBYcOO)%k-H zZCtWBzs|J%24(ybU6{}iY2o`ralEG_Ba-M_|7UytlkJOlX!~!GPhTJ%pFV9S#MM3W zg9kMG==+Www%k(>1C7_|7LRENJP1#+iTRg+K zxx!Tcgrt>LJIdpivE+d*{gqeW5z8e5m1csbJzck_?2oZ>(^IgQXO0p`18H&ZJGQZN zVL6USsD$B^vfVHqZje@$&>d*wRmyl1+8xL%By_~pJ;Gu`c=9?eF^%Sj%Xz%R5UA_>4eISypWXF5 zZv>=B6hj~gASF&5DGEiI=Cw(R-DF?Sbzj%2syXKvc^I`%cdOZClWZ~ywFj`#!0B46 zZnJ7tee)Z?FU9){w|9lr9c0>v;9*q~eugq$V|wv9qMTu5(ptzn#hJ7*BRLUJrc)sf z>a5*)hgN2~9d82`RR|+(F4*kqV91Sb_lS^Ogl(Jagl=$$WNain(94=oM-WjgLfTo{ z)YYAYrh-HRjnvTte$0&2oRO%q}>aIa1-lSZpkuPWFcGgoH@{n(&n>b@W7aNsuB$pM${)9ysr&( zM5?-IrRpMN+XU*^A)OFoL0TdX3ogKP3H$Q>44o2(1#zM-F2}5kD09wa4U6m9#mcmV zbZI-6!ewD%yD%KGBOl;=84rOB8FUodHX;^~yu8qwx!AqfV4aQXuFe>t6!nm}l07`H z1x!}R+@ljN{4cHSl9W~&YCT$d54cK35O=f2ys&<^YY52$z3}sA;)1Rgp`G|K4G+jL zE*^$pE^II7MdEUv#1RcD5PPtN$t&mG7^e6o8UV$T5t|2h8B1uJ#YKo8AqaIeB)IE5 zue}XkRS;z{3OZ0k0(86XC~g+0qoXI zAoj$SV=96yc4q;K&@_2ioA@=an$c|p)Ln|(giZstyCx4age^=fkQHW$lq#g=6J<17 zT{+@m+o5Vi`URl^ZR)(kWm?g#!J3gfNUMfKV=BGhQ-Cr>pV-Qg81URN4Ybzzq)c#S zUt^)p4XqP-CGO}??)uldOsJ*vAtTJE2h=Gb@fQ95ZO(50G**6$Gz`RH>2Bx`h$%xz zv@xRP5y~czy4&R8K8SZ%O=lxJ{d08dd$drQ?*APK*GT=dXxT+{3v4Ol6(sG`@7{)} zg#L^+DP^-mzoVr71l3*4y2X7ONS*ViwSXlEF;LDDttolcK$4GDQ(A#}lR>wHt|4>g4LdnyDZ|Ika|QNU*#je@f`h?&+_sA!k=mCbe-Yg5_$EyAX&}s_HG%G0U=UO zue+1`To#i?h!de@G_J5|mp1P$;%)}HN=Un;oC#rzghBuoLFv9|dsrK(+Z7Tv#O{cc z?_#Z|O^;&|F{@}D2wkDg39J#)79o&#XQulZohlL%1R?e6D2t>t@w))p5Sy`Zx|jh@ z3G&)AS>5b09laN|YlL{2JTwTiV~D~Igd4TUfxD#DYr=8fxkueRM%`Sc zfAl57;d{ZZaZ%r)#E{=OBUEu$|62&5k|vB9Z5<)JZd{ZO(@K!H?y_u?2R>>!`=KE- zKK<|&Qg@xann=56-_>Wk%daq={S>J^0`}Pp%*RDt)P=x%Q5SVl?^Jn^#lPd{|K$rv zc^|XfBzAk=j&6pQJ5;?#C@;`o`Rap*?;pkb=)c%yYJZR6`cKlWHmt6H_6@FEu`g3c zf$%8(>X@>d6Bf?VBLOnZ%^b&3R zWyJNVTWq$s*zKKlNh_Akj)kDt#-Ugg@ zoSaUKc|1>&xrU-hyuBrp}R+S z=^oSZ0TSO2#~p3lp`JWR+J!t+hAVG$3&;M9x*e(0nr?Lh25C5U9H-kFaiZF zJ3zzA70h-BvV-LrQzPU^vt7h410C1^+IdU%kRTk;4I`nCNIayCmk`;5oQbQ6w0lU| zbm$yWl$@_o$4g8{9cgu$*qtrBK1UolMzj!eB(EEUKsgH3Ij}w2bKGZY#4urPhNOfp zk&?HTyUVLXLcfRyGp!_dy{(K`+{M^N!TN@t21k`7b8*mcv8(zCmjTiMK@@C+G^3NV zP1Cw8R!fkei#Y2o@j)gmL>CDS+7!LzcER+z9;pxtFab(P9Xe&|9I+zK3=N4O=1|t4 zgcb>TqK*kILQFLT2V`Z2e2S1DLFGU@?dGtfiG zi^FIM(QBtmaM&ig`*Ybh39ejf%hai(l}MYF)U$~15uJrr+-)ckOczIG?0q|G@2&79 z(x#pmGlHUR;Rpv;u}wUbQsN?f>w>xn>W|SDu^`0GMNY*#W3>rHcc+H9h?&d&+hPC@ z**{Z@>#7ck#zpT?u7N|KcwKS=IGyN1SA_#DrHs|mUH$UJ)e*W z(F5c3wPWROd}msV0!?TQbk5G#UQbBwou_?ziO>T%62-Z{c_pNw`9w=yCJX{-4YUz* zzf3~VoKai^2_upQS_>}j_gicpKsw^i7A-r}(vkC-Pp|+qWuBa=kp?9W5uw8@qMPL! zkeND1(jcUr;%@iaVd9SPkd(R=rUsIgl!es$v1?{O z0McJ)b9lf<|HYs1@n1V7#1#?-B;9@R^Ju}j>#go1c7u>0hUm`mxFC00z-ptlK;5or zaZN~xzyVDzk@__n4LC)TlCokU(54%t;TW@lI`zbSkB~+}KE&F=ay>TMcF%DovL&Vy zj2+5$4H8J}M$D%R!ly%Z!J23VHuSq&kdg5uQ^!Z3eU*OCoy+6lWy;ZyA>E_o^-Q<- z!W)EI(<`7SgkDj7oHh>}-|pGH`7G1n1|2u3U8c>C`FTe%7OaMz68K9(zC}4Yg7g^Z znzFrtgncAT%Q`YTcZi%~+;bE}_gedWNQID(a>fXA0ZW=FR<&Y3>zvVLTbV<8P$oquu zyYJ|f(1`s8RtJ3aH~0BCN8kBF7j;oT+VxHdyubSneb0BQkNgnUaZwj_QQvQU@>lNz z|2H4{qr~;E?bBLM2})UQ`P4r=;SK$SKawR(@n;wg-cQ;+VRh}zEoXV>4%oBQZ9}<1 z=pr^hM4KZ=uT;AJHg$rwy-sVpr0`|By|0nd#P;3}_4Hx#`s5901wp8rn^?O??q8zI z6Slg=JY8dZ|4~n}w{JcOzleEaeeeb5lY6x35lY=7$O^(hoAxPZ8_M{3YW+9p_Dg)~ zpWlXm0{XG<92EEMO3RwoZe!v5Wn~#bKh1o!i=>}GUwIF!YtIu>rCZ-)vai7w>gEx2 z-owI{wmG29_lfyMy7gVAJ5Mkl{fESGK+2zHb?{ja)_IHS_YaBrD=6<{+cK>lFvLiI zB7_JGK!PQd zDMW9Z4vVmA(H88%!r5)=EsAQ?;sQd5hM9+`Ntlp)jL^v1n8t}Z?ve)~WDmR2B7|K1 zoS7%Bhv@s#X2&T+d^s04mbA6tZRn<%R`Y^CXw*4E?XcwEZ(@jo*2ugaKqI0;*t2tl ztGg<_>)W}dow{>&k_T?yRwGix+kAo6yj|W!LUgn{E>JHQp9l|O>h@D}{aS^e19wh=c2zg~phc0l6 zu!N9htOOTY7sjpN4r^DHp=ltk6v^(ku2Uo?LlzgMD9Z#xk5P!(NFCJbjyjq`ax{j8 zh?=-5ADs1V%^lh?ocn6s8P5_BWiIB01V~s!%r#&cV0ObUB{Oy-?f*N&*{`Pdr#z z5D*PYqMRjkK5#K?e82vyf^)aQ3&>NM%jzj)?g->#O$>QMPwG+O)z@Vhxz?Lkom$ zhv^pKE-{{xR!6kWyD8frBcw~@-6KMO=I2{|JrWd&BWXSRWK6w+=~XC3l|o*hVn@5Q zagV%n+r92RK0i>H5x>w=SZBB1L1@b|*C&ovg`PWS1 z!fklnBzsB|)4WFYCb{1ym5n|pL~KnTzuY;hKG#_9|GUBFEwwd)h|+*9(j5}h3E|`>8gG$B{eD~DOGtZYFw_)b0@)I}Bl7B= zbL#&}Nt}_Gu+l9ZCh_U7T*k3BK7oP>e;ij^25VrXc&;#6Nc{%+ZMw_l#VG3a3Szs)J44$!t5^q z`8zIkQ5SVlf3wPG;qQIN<$l!aBftGHv3s1j&aAF|^(~jnySIpGL)koz%~uKC0cCm@ z?S4l~UqIy>)UgobK6(F;)&5-wN;$bjt9#7%Z_;0V*7Q-Tm)>Y<%MB4S)Y*{{}rOP~L;ie~Y-fPwbB%oPaz}-Db2s zhPB6>VJyca+u`(;pQ3KALR9kZXXvkd26>B6w%1o$osoB6rEWin=`I?+^=G7;Ny9zb z!ZqynJix^)62gQ^#M&Bd9c^;9qFJV#t+BQx<}Fs9pz1Lp-XZp943|9MMCX}NFB5c+ z(5_=zvF0tNEjM)TK<{F9o3#G|=o*uV#0`WkIzEQP>!j`&qoHNTTUEmuk}^m~knAsC zi-T4q5W1d_5?UhD(R)b4NX&&Wye8JmqF}|^XW_boxxZxIaz1IJc(~-d9Qyy7DSM^iXlpmSv35 zBA6p2Xp6JC`Ou1;^LUZP{VG1L{Xwb zh*3~AQglp)2b=l&T6F{m2$I}E9iuv*b}MLdvEK@mV}pGm^)AjD!LOk$uc1HcGLb-_ zs$k6pOzLZn=GhFJBi&x3-&Lv#oGGm&B#L7%)US&Wj3DCKW=&udiP;Z|wgToT7OkC+ z!xRYOwr8!uC$&sHz^*xC-j9JSu21K?Y7umsk*G*i$BO6#)-9ImOdFkrYCs;A35KfF z0_|vxsWX8?a)xee2D4FTL9G$Gy9}3KM#n2)m)(6lCNu;Wm}4WYjF_BFOGp&a1()K& zu}~dzkqfCmp}+L%;=~^u-!Z+4)g3HU&@HtTI=M^8#ls9`mpVR7D^H^Bx^uoY)5bN{ z655m)D=h8V5|q3;gi~nc3UC7`#MSch|8dFug*Ina~MoeUF&$FdsgI$qu=P z)()uiK5g0~Qn#tHox%O&lAV%Ot0%+dwofqiTP!B zsGmJdcj*Pj!-tW21pO<>@MgrCB5dzIMcv$^>DT$>uiodwzuwcu8`S<1p>-Yv%ZAk7 z=70L7BmUYLc9yxDI?K0T?xHU0Z$`aS&f7l%{?B*lyZ&nZ9PsylK;L~)7j;n=_0E>s zV|2q!y2~$q+mu^?Fr3m4U!-n#&~}NuIiymV?tK9b&(hVI()|Rndk@Die~@l<3)y)E zJ=|sHarD*q5W6j{tuVcd@Dl0ZD-UoGWwVPCnU0?&??BzIi9_=A^Ve@PMqV9Lwrko( zD4YL-#Q(+b{+BPmX`4U&bm5VYeUmml&Uk#2)zuf@_OY1`AEwT?!M^#9CG-BTPapB& zU;o$4?H2X)DP-pgFW1?97h6jj@ip3+~gef$vHG=#Y>+7cs5B=$PJXf!;-7Bv! zAN^h0n1OrL)9d8jTW?w}U%E{#18u&}bab6=_ZZa`*nN=eXu0Ox(yR+cphH_Bghq%R zS|20W8BKQ(dj-Ns>R+bWs~(6m?Lx|kL=SEf8$PzVJ#41Ar4#W z(JmrAB0XpV9X#~SDp|Kw3S)v=Q4uhA=Qf3u8g+EMLNj;M3WkJxh>W!DWlWX4XT-JtKM#!ZP%()@ z96UHHCY+99f}J``^#N*Pb{1(&5VFylGg!}sM_HViAw%dKoe)!WBmqJknlnM8b93j# zJsxvI3)B&kwFQ$9peey-tT<*O@8RK7$y|_x^X7I!Xs9R>+;Q5f3$h_Ej^A?!H@;6n zWMOtf?1fehQ}NG^k#46z5Zb)BYrz=+>Ie>-&BNaokvb$H3=8uXf$DB#k%bJbpjFR_ z6tp=xKiSR&PbGxO#|!~k@Bk7)+-QoiIGmT4O>b$o*Er0tp@RuwVUR8&uL98^bcR*& z@iG`zBWgyBjbs(YzeJ0-Yl{k|!Nn$V6vugpU>hCy8?zHXT62D@1m{k>sPDKAt-^E+ zx`MQl#Zl^GeS8WJj4!lsW}y!5!P`2#8u= zI|gD6RbAF+q}3_VD4T&ctw?<)@A~A5As+tL0&R9NOo9+|^ejJNVQaf{RH9EL#8@yZ zSgmL>qE$jz(X7V`)U8jbgy;@los~KlLhukdt%-7ihXJw#mu4le3pyWA%Rb=yU#CFX ziU)YD-OX;u1ED`gcl5?&E&>nUv0?ayF>fD2)I}AmjgX!rm=W^(J;-wdNQdMe3A>Ih zDOWMBV5P%?2Xbj^=+sf?i57NfRf%!Lg1QjbZAo%O8Fwh#EAGmk0%aD;=|f1|BXk9c zhlD(n`}-~s3hrWEogvkz<25wwFS8cK!wqF-9wRylrb@I*n=;n<%_t-#@42Hq=1Sdq zS3w+Hh}N>>_z&~n1=S?$lSp+vN?Je zY1I?+D`-2Q%^k6S^JJ|~`G_We*R5ni+=9~#{0Xpx##r7zYW}?*-NYs-t5dRb4R)VZpv9@I(?QPU*;pf z^-~0S3Rz!fm3RG_dYSbr(;*-ES1(h-7x~mb{{GKf&=G4BIu5YFbD|2uQD=A$djrI%ZGmFS=MI{5##`@Csf74mNeYvXI{F)GtVeL`?1VZmzDYUBZMp0iT&AkCGLX1 z^n^5F-H7Tf+H?igM_hogRp?%+#se2rnh^RGF{`sesfcVqBGMF-p3oc6xlqzqOb4K#^NP^D_E0U-yv-bl zsnF)DgqVp#LPAG~GlUMqU))0|Xc=AjGk0pKMnYQ=yGTv$W}*g&ii9&~MTJAyEmOi*kz2&i^$P>I9}6Ojnd`X4rQTF(h5Y=K#S7sgR^n#*Q)vf*9TYBEl{~ zTgc+xtsamULcXcL&opb%xx)+e*KOPsWL2@Vp0Y_mCB_CV`*;u+Ud;qCp>F)1mR5?0QPsufkkwtH zu`Tkwi@afR^OmzkxWxr%#K%m`MqaCPZD(0TY#)eXmDVC{YmP>U?&L31M~DGQ%4*k` zPK8>OS{%^;i=({hvR#H2TUgJHZcwHZ^T1CFNZk0ujjj<9g5XZ#B1jL0(V7vHJ~&@7 zxd1NblMCAUR_lV%$S3Y-mq^|*${11g>o)fzlPX>I<7m}!4}xH}Mv@1oX$zFi!wcpm z0lwc=1+5K@g3N*i!@{yH;v9Qt-WRAn+7?E)onf^{%K^gdJnDtnY&N6wmQtRiwGFXz zWI-P8(55xh={_MO=+rS7oiU&7pw->!5;CeC)(lg_T6A}GXoL{3)|kT~=rMI$N!?5w zZqx0(!hG`6Xua&%sBZH6mn27-dbAlK24`fW z5GsN!D2a%*&L?$L(8(OFkqT@D^b*tIL#z(o$i9E=+Ny-zJEZvwWm&dnC)TSN&zxa$A_5D z9*1^<@G|H`7;X{A8-&eupTO&nvGB%c8f}qs`Us|X2*Zn{e4qLBDs6m>IGm8zN7SRI zsPpBw6wU#*M^~^}DC3JX{*ttPiaK6mb>(Hoqa$>F99H*vedffsEDhdi6whe&3T1u? zo!@LIxTtSkpLlx1M?dx_%=e7c-plFyLDK3rxj!Tf6V`;*2C8joCjp&r(J=b`85{Sh zOlUN8>alW(d3%|BdrdJt0R9_(}2!vy6uyc@d@Ji5I_H0 zpW~Ci@+H2Xa5IZ{(Q=HAH-YaieA_hArZsF2`S^c*3{QX0oACp$pZ}Ma3F*C9zDZ8| zNXS^H$TaZ}e&ZfdzR62`jZZ)Q)~L9Px~R8N@07s%gA`zYbc2{K>Y^^{N2@;k>m4Bt zNO-NC?YpXOB!r%(s|Ww|iKjQfpYh@U@yq1<4`cZyv&?-2$Db>Y8!O?0%d%G(0!5L|Fv&Fll;H@ zqceWtzyCDlxJUH;s1Ec8uM!sbot7RXyW1!h#-poDr`MQ{o<-xI@$vuW2Q$n5$GCkWjcA#MqAc^KA? zkON^lrEI@Kiw*4k?gUZ~bP20tB+TT!4KYtHx&b;5o)oVWWqd!TFF`ycbSLQaI8r~) zPrvjh{KZ30py>!<2D?JH^MyBH6B5FBKM?f&K<~~dBDEN82B@S#kys&2h%|S&>I!V- zj7%xSB<|3ao{$0UsM~$I*9*LH2^MP!lL27`NlB|2G!Z%v7()t@MyzIM{MCxqMzy%S zNGETp7HEt3m{LW;W)zrT8Y6Hwre1`xfCl#Yf22>HjW5oGZ3O;o>)UUgEksS zC9X##R`SjXrHtOXUstqt39$_E*jw6lMKnQ3i!F9>!cov-h@830GY2Ges2b%g5wp9N zL)xHIMoR;Ik=RYZ9wGx;iz5b_E#DE5?mRdQ7iOzGa8<@mm`)bMSY5)?7PqB{3t9Ed zU6PVLKrZ#}YPC^Ng}jDx3N{5S>TAJp*$9hkR#J7rD7iSUO{inQ%t$MtZop=voC#@d zh&G57lMY&-Dd?PN#lRY}3WUzrb1qEAq|WXSYpp_y?tVTGIIE|@UA!4mubA7>sY%eu+p$eSQiYbNMaaof0%C%Z5D9MgtpQb8 zBu)UMY~3v@;?7q|PUSckJ3L@5JaF*}c+gaVe%H4P!*L!O(CXtwM7>?R4&HLV1+z?l z*>M=mE;7dEXoWg0Xp&x$>@MbD>f$*K)3oq{eZO4fhqQvS!NnwiAmqU(4q^sm37T6L z@ndVSQuJP?Y=v&E9vUX@L{=57WMXoGxQvEYXxqi1-5Q~5L@nr?u+~Vq5WC29vchaj z=tkm7$-^NM5;jL96o`>f0wTfBr&b@=m^-X>31RD+JB1B*Q`bzJ6IK-&dP1m}Z7>?4 z+rhRS^EQ(Eg?pVF5=PSQL>HhVxQ<-lzyOAy#Tr$4<}_WhV6ZXtjehlRM|E*ONvqYv3v}e3{%IQ8v3wr;h-8eU5}^G(w3J=QnAP6$ffK8nsAIv!xM z@d>-G#gnSI)3iIis3_G{q5)+lzhs>vskWU{)s#4EZoV@ZOOe3LxlsvqE$PqU0 zqtlbj$5-jE+$I>bc^_>DNL(XnB*v{XfTg%FRwHNz&iTO`M;7fs+ODXZJ?8l`de#%V zik7?Rd<%9yL)-i`F|7#Uy{vD18q;T)j(!50{aMOJ=YTt&B@sx@d$xV7t)Sh*>;)vf z1gcC&PdFli9YDzWjeJ#`SCq4NQMVrGDk0L?!{iWayUeE#)9<{5^fPQ$&cyu|xT?C{ zVLo|+a{OfiU*r+J$IJB&X?zTk7l`3iv>q^>-lX4u>06$Y^O{dCqa0%`QKm-Qe3L)? zr!T=j`?epusMpoUpFZM4zw^JdcXWrcd5XNg%9Ngf&^f9~XRID!n@>^ejKo0EE0nVX zg7p5-pfj}|GjG2^DqqF+UWV;|K;Hie^4`Si%IDtH@#l=&mzhtmFx~qANXO6rEBGW| z{GRvaXJ5X@Up#)BGQEp>xc+>y;)IU51-rV~3p=T27cU@zymx**Xh*1gs#Ht%H zgL1ZJKD&yl(l$zL_ZKVPNK8j`d(TtHCulWcTOK&t(1>Iq^&VnnDrhqgZL=kmYY}ks zu!xEV)pZh>9Djg$QoRdoie!7C6IemSNNb3lGkVPgYvST(4M=a08^OYYWoYh{CLvNG zIL9?~Ga(6OE7UQO2k7<{t428u?pUsl&k%E5MH*bp&V5_VvG$;4UKi)DsyQEcaqqSf zEG^86AXp3w5@V5dmryiqg3gOaw|4h3H1_fx`CC3&J|K2_Knbw)XhG1v*!+2UAVuv=p8x;vhX=9 z5IaSpQ8(_Qg+SQ`=F@;RC1!JuuqoDg2S*({+LS?7%Oo+#)G#$-ex7dstB@OLq8+`P zvaJ}8QO_R2G$V0E$j)?E&9pg_IwS2&bUUY5SyQXKi!B&|MocrIE0j%IIuC@H)>P?% zOp;R%C(}TT$LKs#w^zU(ayG0&D?5ZvF{!BfIMLcMpC)(tatvt$+GAB9m>bg{Xix#KVIlP^8S5vY9N;wHq!!GRIg?1Iy$DmzX zLir5gly2`u#=9MDe3N65=>0jvqBAhBTDBN~uYsN0^_0wJG~c2B6Sr*3u$sS%Q+ ziQ=#&^fTq8L%JD>r?mV{M4bz6dJ3!ODdSage?l5wq_QTon^-(RLi9;jsSAs}62cM4 z38wd<^(cWdH&VMpvnSDdfXL(`@zkT^RqFI8wHy$vVT^>dkJd+^-Um*pr6ab$q>mE_ zgb=Zs-XQP>gw6dMOeYUBpMDM9ewIJ@Rd-qb;>W&98*gxS=iT(zzD~Wrhs;kh9Y0LJ z`|7uJ@@nasw+HCi38cb!^l3ha-2ea}07*naR4i@%nZLH;++XMN=lVbABftF>HpdSW zOGlM)^{29)j*E_SawV5X|pT z^kk_b4YtIWmoeliw zANv!2_vs5)`$c_E^-c)9Z+Oe>2L7Sj~y}J|4%6S?-Ey!k<#U* z?RoMyxLU9f+!-4P6D*mj=~ZXj_*9=?LeJ?eOaR&FxgyGhugV-!=|qj?LQ*#y^ntX1=a_5Il1+2bS`Y~K1A#baX2C7iIh(mH-%O^Y?cQt z*r{`OUWnrxA2-EYV%ua3m)92So?t=Iny{MCc8PLY38^mJUJrx`L6NXVEG}+b7df5#aq;+$Fv7?fmYHhl(S3JQ9)iNhI@qAqivt*=v`R5fogHcpn@}4tc!SN2c^HzIR zZ|{}l4q+{UsU%FDwVU@GS{rjgEfH5fM%o+~q=O)|#Zla$x}Y-_-ZCxG=0K}H7Hc!4 z1r$XngfzM{n=NhX5jv1apeL^isu`U+=B=PjeXJ&&8{G>*6>A=f6we*QJTTFE>w5qd z0};AaV{Dnw@L(eg)Y+Z5X$M%s03{kUV;Y4v`FTMUZ3f#!T)U92wzzBK;(lKox@D9F zm7w&O{MfhIXy$H8Z<`;{W|#)%vBO&M0Isf4r$pUAlYkji2_$v=f)znhLbpp>d)y+y zpZ6thkQ6c7Mb*5OR~D?Zn@VAs46)9|@7E3%5AY!8M^V5Xv@3Hrye^?)0%At&i%%Aa zj|VMFs2DS$KPBWqS_^44Qcij_6hsxB83HxkHwdUE@~Zv6?7dm6EooNg z_gmi(5qs}5-TCIt%*x7|r~=A#HQ_)F2F4U<#U{fXbS(DyN%yhcoRV;`_d}{IKHQq^PW}%p{=- zI7|2Fo_o*PC*nkW5qt0d`mf)gx*cibf)Fw>3$=DoXBYc?olA2emIgWo#42%N4i#>h zT7;;lgmg~qI;?i6DZvtT*P-WE5Xnf0&X_jwki!QK^y2P!iMUfbDWM9I8?kfuoW#s@ z9+=Jo-TrK~uYq#fV-nn1oo0r02ck!`df*&aPS~B?pu2Pjs|!p=B>C$EWgzB`klpR6 z7O2Ib8ZtDxy;DaHXd;B{!J=uTmWXOZw0Y2(MNA`gY=p3Zsy7xnx$1W=Ai`Aak| zQ>Jf4(jKksV{H$a88KD|(Ci_FR*^U&G_dTbkHtA{=j$eQ(bs7aYIVmvs{Xu{3nFJ2 z9b&^AmU>Ko^d;J~n44QSXs7RH0%UPc?p~kxY@5qW=N}@3$BE0x^2$q)ew}jiewzLr zqFe-VD5by7~GVZ<~L;?~E&Tf)XP-`lq2{gJfX62?s3ESSz3 z!_n(a^_U8}95LPV=GS-`4Lu=WM*2_E=PxmpYm|DGQXZjg-iO_k6l z2(J<1Ijyar1#Enhn4ib&9&LOHoeW{Ybn-ZA`-ojZ!h)1ClGcQ-B6MhqhA@JUK2Y$xas>^=?+qkHUx~Pl#v(+=t zC?ETYyI8A~aY63i8j8TC1GJnW;m_98`uBe32(9lz|Hqw_ZPqq{{>*OiszG7f8#O*xmbP zRIg*UbTJ^cIU#-on~2>RW(%hKm#E`Sru)CfN`8a3xP~nsC9f*O(H#`#7l^0q&QEFe z5@okXkVAq6YS~f8H7)g|#Yo6K#Ehhky8{12wJ}q-9U*Nn23pO8lw6>h6hcH|@O1IC za2IH3pgr1pYF$E!G|kAU)N_N_TnK0hrFTwi>2b^Hj+i2ijOor@Qquyhg^2h_>5v#bfj-|S4`+lArg%7^a@rl5mG`@04gyz$8e+v4rB50+S#Ehp^{f4A_F#MmIp5| zNl!cdAcRQBBeJ^(>4TI(5t;VKd`(f+$7^ z=7QGF*RNF|x@e{o&1?80ch zxBw;qMhIYIbl$HjL6n(7;KE@5nt)Wmpqq$ka^B_E@fxiput-ydP4L|dGqNH$azS8r zRVr`T%gq%iJ%pz4K)vZ>krJ`gc<7Es1vl2+4;Me9dP|Dl>6`4f`@aN&5YOT z!F9Wq39}2=6;^dOd{(ds7O}RRn=ApWU@d!7q|{1KEBiW(iiPNy3n|cy*l&rUP|IS*jwIB?8!J=tpvF!h>^UAq16HCp1;l7A zgGNFcT^!y;%FdmY55oO|I6L`uSE(_b22|Z$n>vLM&;*o@x_y+`ofGpZRTgMXwDFLT zBa$Mdq|!0kb@c$ zG7vEmF^$BoLwE0?ZTHwrcWArE39-`c-$yz_$4tzb&=ss*ConjEBzH(s%6jFleDV7x zL1#9uJMY_+3AsWNT1&Jk5yFUcg2Zc#XCEf6UIhCB-Q_<?@Fl~xD3 zg%Xmxwu1)5MnXSfbva{QHl#SwN`SoDPU-h4h7#&&m6*`1Avynz6EiwsO96aBjJlLKzG7>_uAL64yO~qV1fv zdy!@@vTJ|L|NX;X9)R^r&s3gz_BR>N64(c5?Y*SMZN#=%y^3xQX>to|3tCyx#*Vgo zfinIf=&yqOEKg&kFu%IKsEfL&i~1{2t-V64 zcNouZlKVIRq9Y(o+w17=)AaT2VGMrg+02XKUm=7i$%i+QeB|!5&yBFP^6L4E zJoW4!{Fz18kNw!=Nd8xe%kMzCGY`%Q3($r!LOVunOSd{9tv<(eei^e5GTv=G{gdZ> z;s?IUXyso({o+45<)c6KDmFby-RzT=Z~je2J-9B2cNDo;LQhS}D78girE`A3ouuKlOfw z{tD^f%hyXjxHXn|@}vsxxInpzV^2 zE2RpmulkfBVEa&~{`*F@hD(;eb)LfM=%o%NLU(u3B-Ly9B@*!B-uwa6S? z_CS1lAdJphC7vCJMAy8v05dZejL;cUTu?Q0kyye57q_4yRK>@m)p@4c!hCF9*zFeX z+Ep{^WRA&b!MV+5lu^h_XW_cA7J!1pN=V(D<{qg8>Q+3EuX}*c7%`dLI^2|Ls%tV;aLj+62zM&Or z6$}GOL<{aR?+odTfKqG3G!SXz#e~Jicy~#gJS-1{kd@R|LRXMbU4Ya8WK0*1H)+k! zOAaoC*2wGr=7-9>kIg?=ZjFcHt~-p5D3l8!w_{sv>|;Yppdu3#Tjk8 zAT&Q-#u6Ky!E0?Iq#0`wR!Gd433a;czY7D_Ja|n>9^%zk>TaNIU4d8iwWtQV3DoMY zb)5s<@}6QP0Y>f%^}PZ~UAFA6ESkX+Ug(Fd;j zKr8e5YaLn@P|~9L{d^aY>}xZ1f`l1=k&PG~|A6{6Dp81irkvlRoIQ?eBp-c=#kDU` z);ADc68rlev{o9^_Az4aNW-~5Z%iFp5@}^fth6bj^ZP$70kh?Y?AvMqr zbUI)=9Z5@YY^te}J}r08x` zEMesc^qkm_=-7F1*evkUJVY<{=cqpFYan!l{9DA~Hrsn2#JELBS3sA9)F7Q8`HYa1 zx?5u7WgOe>L2iDGDiXaZg6>hLzeC;q4wT#OnCCiibxeqne(&y=y~7TnK!1#AM#sxo z2;}@KQ@sp!hT13L{8yP|ptDBR9j)A%gKXcpe*T9yJpHr3!tVH#wtJGgxJgb+NCj;> z%sR9c+VmoA^9q`N2j!Fe;tyOfM!!+&yZ>=dOxK8Am_wi6TE*o(;`#{PT<6a+b^Y-l zE(Kl_J#;!ibJF6yGbMtvg)y#EgH6!3B2cexOIFY2N$>MwJ> zEMMT(>H7%zZ!!)3rsB=i`JKnn-3hgPfuDODPv%em_X}Ru?__cKe)94vdGRu7_2T@3 z->aUl*sZUl!{1^$IpnF2@ALCN^apQwTL9nrvkUa(d+Ar-Nx!OuyiZ6Igy!Obc8G4N zn;W#BFg83+#v?GWIbp2<)a(fv4-YL7fpl*(^ut!Kep}E~tYV)9@B1q^6 z(*xJAhX;iSaZibPq^xyzc4lJN9^gCOi8*>uo3(w=CCCz!fEGiGi>yAP+XxsII)CF= z8=@1IftqO3zKfSNV`5Wg3@gxr50Fb)R5|SPMh~V_Sj=l6C}3gs9}jh(u^* zK`S1F6C$)-k5ybWO$w#VZ|GlTT0e&A)&;$EN*$MIy@HhmF*dMA8BZ@^EfLev{4-x8 z^dljk%*UL0)Weg8T?sfhdZ8n(X zr4jo|Eij$Bm>LHUMicP{M0_A-R?sPt2bo=gfwC6z5JGnm#8CO-fsP)i^r|92f{!|tqG zS_6WM!i3jF?zVG|vCYn6nvwTq7K7c2P;G=hIF>?;kb^trlg=*8NQ%C1b&9mHBSfX!SEhAvVer8%D3*ta z1v8t^wR1OGi8GhiuMM+|SfEuwRUtfBH!g%b%EE;>$49smyp8UBj==~ELq$oo!Pf#u1T%)rA5ykDD#HqPdJQ*TI^iieq%0Mdv< zKx#KkEg9fLwEJvtrGYV?VU)_HiR8deP* zgEs*NcPgje5DKPSBo>fJ-Sud>Pl$KW7HQK!-Q6H{m5>r)xPzn<+HQsF3T%&NS4imq z4TmThWF)Y|!T{(%7r$Lnyw7ypGo8N1;@am3-3iz|ET7OHyhy8wSoWFDp7e9D_XVsS z(61&`Ct@^0f9|1@WeH3kkQ6%1_MjagVeO)8%Yd`yRU09VlyR7?><3tT4dG>UdkwR9 z6Y|oT(p@Cw4H61v-MbUok}s7gNUxyc$cudH&o|RR*<5j6_2!s3d>RQoW);{Y@4fg{ zUtNVzkp3==S1>KU{&a~q((dqo{Lot)cmKrqO~7yR^iO`4?fK(O^#*iHqC{*rQR6X^ zUSaXl^D`#wqP~%;-(MhULC9|}06C=%QEpIV`DfmjXP!~u4|wX?&m*Ugv6UNOOOkdF zwqR$J_zIgBUYrqi-=JlJ{lraTe?)ii`)}!?0V6M7qYjs-(+xb>?xHU0P1idu@E!sm z!Lbc50Kd%}g#vwf{hz=D{0{Ic@ZSRe-8;Nr7j;n=by0sY^;hU#0CH=c7My{p-r({b%3*5`XWPBK5Q6McF>$3ww zRFaIuUa6ZEW#d6?5{?Kl5{8V#nJpU3gTFLmrJ}4U+m5u@5r+!4#dJf+TdWS$7Lnre zxM@MBfa#jB$OOxzUS%fNj9m2u9i5B*=i#Bf7quW(u_B9fMIo=>b&PJ<7h#=XBS!YpR$b(WRu-3VI zFB(K8CJ%_x-GE3yL`hw67Hx4J?9>9zf)KEts@>caXjp)5e?v$US~G1|5J@1>e-AQZ zJr)H~Zzi~IDpZNYPFxV{+)-QB&>D2z>v)0CLz)F(%(KO`&G>*3tjw%d+zsgFRS#v; z3f5!}zkB_-n?dT$!|meh*IZ0sK?vCq7*+(!9U<2_I8Y(nN2utBJZK^Zlydt*4MA5P!#E2BZrbNptw5o@zb%KenyOi9StWD68+*upc1@s0G=6h(Q zP0mq{op~TvG|aL$?qcxFK8;|CGB;?lQO5_@B{WUYpw*>|z-ELTAw=5t2};~2#tN*k z3gy&8_;e@4KGIrGOe3j>avsq#V^U^kx+em9*P~6r8w!I_gCVVwmpY5VnIv`tv`0L2guiH+DdL^{g2KqtR`tOhI=Vz(xC6>UA) z)@XN1NcXT7+1~#qZ?x^*1sg$jSUp190JcI~B=wm%jD&ECwo9}rLJLS*W41tegjNnY zzrCbi+-BMy)3#U0`{%?k5xcw8&6AkzQKu^eImF@~A>GFG2(3LxqfUO{;$b1g;^Cid z57E6@SlwYI(aHkhj1VTq^J}!J!{iL?HAoXyo*;G$y8SP)z57jwtjGt)=oBcMB@zen z%80{=1b?1V?Vv4@e2R!d-eESO+7Z%CLi{i(zks$!XyelHAODi3wTPk(%IP;#w`)v~ z$;)r2t^-ef{L`4e#Q8UUerC76x$>zjpUistXCn~##P=8eJOk9e>z^mCeR{=?%s}C_ z53JwPK3^2HDUrJw`2RLaVnaj3;yb4+IaBx3^T%Jl`mX=Er%(HYSl*O*|BdTc=D)9VA@E+*MP1Zi=6dFt!0&7_ zmIiibdnkVwiG?^E6Lx1HXN3GJd3l? zu|>lO{TV}EGp!#%r^l$L6LR^FNXu7=yUUdIP2lSpxMrMx@RP`=JCc2Xrr$!pyiMG@ zOAw)L9;J>qAsk}mGIcznPIs`nA$O1Qv1k7_)A)yk_{;20PUyyWakjZmJbjFSKnj%Y z;{;tm=uo@QuqSj!Yhpe{n+JupdVp^C*i9>jOD_`gNcdCZT(vom5IbQb#{g)qXvj>P zYuL1hwa&#g3)J;3LOvo6#<2I|!>|go`>b#8VQn=tXay{TUMKeF5E7EM4^!L$>TW@d zirPrkLOouh#T|;FOQfMA21T)lBiv~T=};L@J@ChBrWM0fNef30gqVRHp+6&Ru2AdX zD2XYdoRbyqtL&_W#;0s~S zW(+ggpm?}f-}_1Ndj|jjAOJ~3K~%q^FAG`?E+S1nhOg>l^{F$+<{aB7q^``O&x6O0 zSa3Xt7I#I;G|8hynXD7k8rzMm$bK9oLf&&BR(x`va*m(J*~J@XYy|?$&f}*3y&^X2 zbt{l@y0j#dpB6|;b7!ouVM zgJf@T#5(We%7tff_Op#H#y=1?O@+3D#>`uuoxEl3Hqg|2z+Z^|HKKD=6{uq-RDlqP zX$0$`W^_6tE+=FTU7U8#W$yO8VWiC&w-Q#g(XlQu8hLL*OF=h1aS){UeR#mT_Ef-t zRio{kT}^8AgI^ypS7PUY*N_6L4QV|R9ZeBCUtdzp$Lo7;s21rNE@i@i&K51!gE-A6Ev@sIeOhBpu zofep0cKk(5j!qz=EnxICi!QcKE3{-^Q{woCHu>(x#O&VnP_NW05*k_}RKO})TqsX5 z6Bn5--lc37*k%9{!{slb?K;*DK@X52Kr1x84Ynf1h)PH7CU=jAM2H!o_*1a)TwyH6btDaTjI2hPk7aYL(DF};c8j?~}vd%&2{wny7!u$_xprANY!m^U<2FJ6| zz~*L68@$=F9iW|(26MzoyF%kWA&nsCj3-y=_l}AE?kisT)Uu#%uAt*(T0H_Piz~;J z=@Gi2q2VfZ_bP|ay@&7p@!#g>e&p4!{c`G0)hE8M0dG$*d^7d${@6`Id=mN0b!vZ^ zWLFtka_|}DsgEC{JWtZkQ1xYg;U8Xvwg1bijuWx3nB=$l_ztmxyoG@1zv`8y9$jWE zzP-GY&=sO(f8M;Pi+XGIP6@ov0zdFozV;sgPXqrJ@M~Y?dkS+x#yhk>7j;n=by0uu z_0+RVeoH<=w|qBo@hHj`>TL-7==cbYCq3I{bH;dQL*qH>_yyYLOGx}AKllA_J#6B; z{&}K4_W@SJEtc2*@N0^@4=OFzbo;NNrUS4&j16!f><{VoUnO;?)bX2;?JabDld%0HKmVh@#&`epbHs8B%2g!p zfqak{H^k+J<)z92I^+OMi1d>B{QBM z(b^t)Z$qLYQav1M4m}E~FdUpw)_aWGMj#Wi|J*kAXtW1rFc;W*7csg$g>~}mthUC& zhO!f43Y2mg+99L~tV3Hs$ApBQyr-n@3~QA-u0S63aqN%~jI=6fJ*Ji`7?n0n7$?O1 z0%16%wS8=S1nepj4G9I)eIy(@+fxFBo=Baot{vm~dqIu}u_Eb3tZfPD9)t?v0Q3rK zd+yBD2pOzJ$6N@+t^plp$&Us8hDU-C@<`hztYsuA-QGP~G3xqJw)d27KN3QtKbQ#N zobmjCI%Rhs2O-WFfskg$t|Mw;rUBFuzqibnf9jaW*oBm!6kk#{2}D613F<*q8Bk~C zhS=QAsZh5bVimeZwCIf6VD7*z9$IBBfC{nGIcl&ui(7@b2(VDV!tBNopU4j&eK`3h&%iK&bgDkY_O^wt+-qTsE z1SY|nkp~YM^YsBCWkLttqENQUdC=RXHw8vB`>xR44pd4 zHd7}Yk%G$|aNzA)>AG07L_cMpK(z?S>GTc5{;_<3VfU z*o%LF>bI2Lfphk?_=hcsy^F=QBxoaPbu!~HR+tVm4yX}_k+KUU zjYw8Zm*~_XvcX~`SP$CMYyl)vXIQP&apx}R=_*DAS%WO-_BO;^(6&PB9!7(-A<9I| zYg9Y5Tqfoh$*YNJ{Rk32kCvAhH;-d>oqYHjDmzREq+LRWSCKGJ;e->Wlj|NZ7$}dgyem4w{=to zMxFL)o2#fE02^X{h2iKWLjN*ha|HH9y5;BDoZh1Kf0OZakEfpf&98ef?t^;zXFGN$ ze~UOgNg5spo&a4D@`7HLm@KeX(Rv@NFA&P_@Udqa(A1!IDg{}1tqU;Go3y{-Q6NB z??7-T$~Qte^XvEc=udrtcKSC+a)r16>t7|rPZ8vtw7O5X|LVhI9QYEOdpAk(TbZQe z@BI_ypMC#}fCq|B%|eJTDb|+5plTVSc}jY4)3r#y@HNc z*qw|FmtKCje^wH^vv*O(&(qo`dHTopcq#sEQvU=Y9e@P#-pkZ+!F2jKrniXQ61YS; zzeG7b!DLI!k<|AXGh25$1?iD+K-+a_9ndLJ^587Yu)!G7nmj$63^Ea7Cddd;F^!0L zYJ14@h<~(sDVYZ@<>0oq_#mk044X1qBW+p{n-Uf~XgwhddDzh(JWm5`Tw(1A%6N-B ztdZz$B+*C+OLE_kw1>4FVR)YY=yTY#z&1xP7;`I;5Czi}W<6>F${{*_ z1XU&WCrJMs67G{0cc_~K#*+^~y@v8AAud4OaZ1C3)j}X;NZqjNrkI8MP0hb|n3h%;G>3zjj?hdSFs zvDXB;1(eOqmu<|!WCj`732o-ik^)+yhi0LKr3d<03y_7pFya7pSAptm-)N-e)mFtj?hz>SWA5C}0*Np=_OND%sDM1SJj)G|fU^q;4}YZ%O%_xVVqio))e{*dq<&%z{pHgJs4?%qRfo zv<9pO6r~nl6X4^uF;-+@>|&x-p_NF?>sg`QVpdRHIY&93qe-!{r0l{Rh*pVI%65U6 z^FG=HCIRUT8-*yo#?wLx*Y^(%W+r4snhWYQ#R#rx#(>OC5^XM`BEC>!Z+`r6oX6Az z$=-OdKWNRYDA;p`eza^=J)qL&^vuDBhro-RvY+LoLzTKujCrYQome z!Hr!dSftf}?p#oH{DPV6{!FUN`CLpAgkkkli!mJ@8&ce0@S{Jk_(aMt8?~rs($j5|MXuG4i z$u%N!Ku8g>NSg(H(H$XGa(AC;yNA|+xH!S;b;?#q%jZe)3T3)PzWiy*X=6Npgt~bW z2|e2K>=^%&2h^2;IxaAKfu<9~&^fZBDpCSE?I99S0&ORhaU|<0nx7-uZSvv;me*e3 z?D=n_O|KBA$FO>tbnuctj|FofxhZXz(bgeZ=nl@E1wCf0Cfc+FEtn|6hOm6qpAU1T zZZeX0h$&U}K+oupUZaj1rqeZydrV~j7MKpy>5xEA=w7AUyGO`E8+)u?M%xjs-XsmT zSzdjf>HG+@4>3;qbq~e`_^zKB*q;7tEDk?__1BRyxC1;!y5(!c;T*{YHY;}14Q%rW zb-aQNH`twC=6nCg-{9x|)4N~S-{BjvKJk5xkN(tcT03QZe3iw~n-9Phvl4 z?bmZ6?;ESW{pB71-(y=x0KE;-T*f8nmKxuB=-oyAwXb(V;Qi4*@%bI#1AKXc;`6{i z2mUMIe*^aLK(&jysEfL&cd$(O?qBTK-WGo8KN|mBnO}dpo_QwlTk;VXclV9u)jxU| zwiQxkapgteMHhxzM$IT^*O^ZDi6@Wp^e3L<6TH3P32T=~-9E8<>z(tED$?yp@raN9 z^dA3|*B!*)^V9Dk%D0n`-cPqWB`-fs>|l3#KumY2@eJZ4#GJ8mlWE(N zt(o2Nd#K}`t#jJ;dxWa=%a7J8`~dIiR&m)Jf2~ zV=5bBt|*u3_O}?PXl>5CmCo6tEmLRT2>OKiKabu=i}}SM%(TYdmpQxN~Ej@ z%BDx+(npy!P^MXg4MH5$)mwp-8?6Z`I2$!2W7>)bJdF^CS=`mvk$NE+F$?3F|DV8& z=Wx-N>fc|BAePapgaG5o;8@xgkb z&>5*~E=+2sEySJK8mQyUQ;m>%0d9fL3x@sOiAx6E|@xd~(eIzp=gYl8rFa=|@z5C)^13u2AD ztTO_lx|4HE9zxlVzRt!KW|&BLAeIpE{+i zGi^743Te1M1B7?Z?=*LXh6JruY*GwGa-)p_GvDXfLBBAxcz~N#n5xh%+yY*A?nt$2 zgrxM#f(4mfhyg7E?Hnzc*8GF2SXAdrXKx;uHV7^bnh2&nx=pk(QH!r%2_q7>NZQda zl)AZ!Y4U}OpR-o7A38I%MkFd$168}3JuTGgL1Q)=mRyK8gIa|k>MmP=5LC^SV@U2J&DY&?f3@N}ZwHCO@Z&H)6Et9GvES*EA?132B+u<{-D= zNEJh7VSHYP$WhkRKw3xtd8TTer=`R2*Jbg%C1D(!Dm&0u$b0vQdK<0pMXV8$U~NEkz-mHE zPucAVc7nFg5z`;jUHJmtV#9d%I<1_d^>eh{9yYuO!t;KuVg%XHRH=1=$m`r(NsXKu zRy!maB4^%s;*^-r9LH2UzbCUJE{_T05n4RxD)np9wKX~p)M<&fj(YwWK?eH6SLl{& zN;#tLZaCUPuArL1I(F+0*&GwQ+r-1y=-T_(9pB^IpZzs%bXQ1v2-;)TKsHFY!$188 zZ+!yoEtLuTAO8-9!w->{gR{i@tHc5Nqc3>y;e%S9u^iS+=_YYh>iMHAFL!M32A=-0 zU*;1(`c=#be~qh+pG7zCrgk5sEKkVYn|n{Pz55tC-GladKJj0?y^yxQiuKGh%Ez94 zmDbj5j*l3Q-Wn#Pj8`brZKAz-vHhYhYOZ%e;QhD%#OMDpUoP(c$$AC&H-Y~K_|Jg< z;T_oaMP1ZI{dKNq{(tt~E!ei~yzBdY-xy=gb=!NNYoES%>(-L2SOgnPDydYNq6*5! zk|9;f6IBU>ARDUk#tRRr&P%FNc}oRC>jEV3WLL_7F)nz(3OFXn7LsMlmLX@rGsS`)?6Z$pV)i?cSa%LE=b|j$C?}1>TQ-@*2VqkuJQS#f5LsA2#M|-d@33qTLq87PJ!= zCwS~(f0;TMAxy}3NaB^aL#7qubQ#YhlKx<|%S{iym+s(E!r_j^#pXo#4&aHq)gRu0 zyvB#e+2!{GpTqqT;_4c@y7?^^$uSMo?ggrCsQD?>3#DE~yHi5g15Q|;VcQMU{vt6? zO#7=$+r)JC1!DLo7#}7bJViXLEHD2iXRmyKy!`;uUnQ<5x;_*7W8$g_{$2-U{}3T{ zl$~O}!g8WJe3iI(5e+AdyGQUL;I_eKYJpDEk~%c}fLUuvMl!V&;xe=#s1R}Mk(iNS zv*<9NYnCt~e%_g*j&>t9b=1-!^rXehcrlz)>afOTOI$sR>M6Cp7hz3Y+(Nr!WLQ!2 zhBAz#)y0hI;#fTjwP(X=HTjL0}4 zbeJ9Du z%}E=ALCnO}9>RvQ>lsgvP!E_dn^V@6l$b>{nHj9-pGOg$g}K>@EPz?#r8?T2r5>p{ zVAlA>PQ;2+#*0l1qRg+;%}5S|Iw2I&YKQBBD(7K*J)#lmg|KX5dD%KP3e}p6kX!sj%yY^#4lT4@WV5z0?uA>-uLk)~)vokifYlL=OHIcU= zIhtG7YMP-T?LZB)d)i%Q7rM9On}9~6i~*ZEv=<185R6iq8$b5vd=^EiNGR0OVb%m? zi4{>N?;^E2;XLFpG^R8%0TVv;7;Eqzt4rJprLJh%gc;Wf(F{?*H`H3NvO+?^#zWi} zs0!#7bRu-GAR#sGt{&rdObd)#qj8IbfU}~GmqG7fmN6SxUHWx~S06*enzVYUWjNLj zD;+f}8rS&P1i1oPcSRA}WiaT_< z1=r^8wi`%1r0Z|B{ZkiA!$m9~lDe~oxd=THE>m|syZtHo?1;SkP0sp%#li3%YB8p> z%k)R>8Y3)Du;~oTM~HTyuMyTfV5LKKN9fKlTZ75n$uRFPgF_}go^^$_`V!+86PEG~Tc z-_uC{@Ko-eB@XYV?4Cd`+@1l|W;^ZUQ} zE{}Cu7~5X}{s!#1R9M^;p?c#~W;q zU*IFpe4d~G_^aQuvDEl(y5%D*4!`!D#oafk#l`0+(}!?Ds)5@=gA7pj-UY3`>Nw>N4n#U$o11^QoZ?S2h3`^2-OK_nKml*d0eR{SD zP~TP}6_rHYU%~406#f9MOuL|;ew@@jNPqa&#PB{)o7E z_H~b49zDnQ&I&iB=64}}M)i)$A#EoLmM>vLqUv?ZU`)Hu66!B7=s?#!L0BAe@W5}h z7pxsbZB8v!%o24dkEFWZKd|q5@%s+7DxSMC>IF*f}9_w9`^(}n3v?x5UZ)V3gdF-U`im=NN37$BIFV54#dUG;jMzYP-f>H0&(fY zt_huWqKzp#=0cu1oN?(a*KMp`nwY8X6r|-v&2CqBA+C%rd3zlk8w6PhB8q#Qu@0?O zyowW7%r0GHx~=2^LTtDKH;_S_3%Se!g$7jGc8$_8mcBjCTFbU7%){n&(}~>sUlL{$ z{5gK&E)%zjzj+6FoUseDu~$7*nH`t`#*>DuI1g+)=kfX+V5fRd@D=4Q)3;k2 zb&|#>o?XIdb6l1Qv?S=-F|QG7Su$;%w8Bu+l{z+6I4x&HO7K}6YofK*7NAF>%>@XWZXj`sjTaHQh{X23Ru#57x6d~o zv6Y4&a0{&qqsuLD3}&{lg6FpHRGT1}Lvy9Z=JXYxNXzD&hcm+Bqy_d(9f~3~keJ$- zbE51kZV~M(DjhBh#1yLo)1X*cB06BMNazSE4HYBOTF`UAre?2o^|{Tz3Ai=4eHfCr zP_%juyf)&(sfD>+zAbcU5Yn;<`8G4CiRZgon&ZAVQ1DVI?u1BlJJ)vX1+f-NxR?)A zYuidv6QAAZQ$)Ni=*)a=%MR5cFf+D6l0%MI28pRHJ}oBdq?9qXMImpteIrl?M-_7o3$MW#2i0m+a9l{EYN}kqO4X}FvlMYaP z++f=kB^?l#JKUDk{RKq#q}4dv{W~OFrA$wf&pO7FuVVQHKJ(w-;m1DtMauADrqlU$ zhdlrQAOJ~3K~xpk!*rYL#B_@F*O|_akeCoJ;DXnLl@75z(w{*nP!z{ND|K8^ha+lv z7ERAl%ENfrff#vrM0fD5LTi~epexXhx_80? z(DV{v{nds9lM_U>g`V!;OISXR?H=UMf8vY$++X?X8%)~&ZYB0U(&|yVgBO_YydPN{ z(O>w+@4E``IB{_uJ9`M}Zqr+bkB{-=fA>ND>0f;L9Ub5IL;bPmeuK|Hyk-RK0f-ApZvtLXn&d0>+fRuz}LU) zJ!07|*xh+I`Sf>@`r9)??ta~`w^Z+h!23nuN9Nz(?8*HPfIkiVk2FW`{kmWG>wdlY z`ZLcwgr*;1v3ZKPF%}0epsTaHUz}EjI;@#aA0l?wC?^|!@^Ahc|LCth_l}R{qn|!N z!xdtx#LZh{C51i{`ZH30k?oDEv`&Y&R_X)dBe^b!y!942s|&c z>LbsrC}~X?J8D0%*_^%ZV)V~^`XbU@LYBKX-I{i{)R1(Hb{E*+yh;cwU`g7%-l1$b zi^H!n+!4Gu#jhb#z;!o^NQzaXo+*Uqslykk!)N(NKlN2U^2`sB4j!UAdI7`P-+2(# zE6C~=nr@SiKS-5p#NG+9g=L8@2)>|B0kc5pPjIVftc?xUZ6QoZJjT33)o?N)8XWB5 zbVbcODkpd@GYhgoW}S-Th^IMl$f3r&s+TDEq+kJw&F$;vjpeHobvmHtMRP>fh)P9x z1ox{gKnS3Q)N^bmK5zzjgkR0NS@hI9Hh9M&dK2%%v^yovMbU~W)nOu`&w zXx@Z+YMtlS0CO0a&d3aAp|D$zq`IdLP)4QXt_hN)h7PbJN{)?{8y%ryp6&|SsI(9{ z1r4uJ*NZmiYV;;XHL~$6vd1>gye6L|2$8rr4<6j&-s4sY-H1pNAmd_BRilhYxJbi! z#7wP*7e~qxSOGm|hEi}D5ebMWDu&NtgkF?7&KtpJ*Y_-xstR%0IM!tkoQg?!e7NX|boWE9^Q0%#(kk4@-5`=D9$a|p-p{;4m4*9s8cZ!1zZ2_Qh zc7ai-v%oCk)GCBH?@yV9d7xXhghi2x=!k|nq_Q$EYAM<~^~{*If@-t!H(qTZsuR@j za;Hma`?p)Q>3!OFvztG+1qTnfC7g!sFpG_h;6_*+la|MbmgY8<1s+ChtXSQ`5Wxhu zf{l@y7i~V*4I%1`aL|Tt&=G8lb}tdaKp8(kNCOlD#^&Hw@EXZ>5iOhcerjz2S!Kpb zv|vxHql`zybk@#CQCtBnr2YodLEb&e@hjKx@|*4ZYjKUTy@bjYYIlXM&!n_%&iB5M z_mMhXVcM+_DrxaN{o(WI0_xC#-=XA!O%nQdu-e=ogI>V1ep-uA^lln&wE>C63NtI!?e4GPuGdd&*5J1{22Hi>KS=^nYcbd z^_~OM@&c2Km~9X)w}3cp-&0mP+#Cs8y7dhNp^S&PFB!L2h|7f4N?iU1tE<2Jy6c+$ z@I~tBWp?oc}UQB`OXf@r{v*L_Q#K-{TU%lSPc|g!6fAIC8psyR{2-? z*^j?5n(cnU#q@8vbGkwNL+tE*tS@~7`40CO!|6rF<0r@`-^8Xb@lSsGg@4bZy(+^Q^!k_=&ckg(>LM$76xkkdua)K8uJ}7>`3j%;`0%zq^PBr{!jFIYae_UG#;YvS3KoG>D|xr&BcFH` z{OkPuUwN(xSiVBi0tq`5A?FSsBhnXAnrP8F-#I8Dr7crff~;r|QC_!FZ)Hh2eTZ5w zf-G@~_-L39NUV4nk+{Rko^kpe;`on=@>~4ePqmBZ8Xh3V1G?qcu_~0ahtPPL#rg)* z*)?2^5L%mY>RY0f)#l=KakP44TjqlCFgA6Bum4z)amgiw9}&~Rs>bSy#wi+_@f2woar=TH4_&?SY#~s zcwQi>qlq9DiHdZB3)mPigAf~*z(wiStto!m3Z=|#`v4W{uDS8t1e1aWBd%s4O9w13 zkfd~-l22PJcb6)3Q*%L!0~MjQR!eYJ7ZSrs6D1ub21gf8xAB>?YlsxwmAp&jQComd zXu@N+F*F&Wn;mxCIfk;bVDtIwc#Qo45LtM zbL{p9Mp~Zp1kP{7&#kY{bP{GKZEP%I6GQ?&M0^swZ|+mCuLb0If@c^0V0 z*?c*Re4%qfSHYRF0)?9A1qy59G-*OxbKF8)Lbq-xi*s~?XcKvZJ9#JPrfF)P1!$R_ zz30r{d7NnD;Ua}PD0NC$ov=~RsIcfz6|^sHJb6KtaF?HKuB{tlMhvtQX&w{I0%bd| zN^B?yo0@x-YVEj9Gc?$ZG6|k1Gzqbr-LA1qQuUdAR7ZJ*2W@ATq$*& zU8D#$c1?H+&Ap18Z-a&@sX5^qiJjbaJfpK1Y57}3C;$;_?d5fW%Z#_sNLnbhM!W{7 zj!Hp7A@!$(<(B=c4+jfXu6UX|id_wC#X3!!G6UAF-Y;(8Mx}tKA(8Pb2Ch+&G%>G!hx&d0iS&-XE8VG5ij7ucKv|kXy5uxh{ zU8Rl@ybxBPazGgazK?Wvi?Ar<-9_AQkrvn6Y*FrD^$I0VP)}i@ZPBk>B*a4`oTBLz zw6=~7fUw7EA!rA&g3XJ1CN1PMN3e>d%SPY+7id-@x((bQeJ?u`{&ax9?92*tBVLYhGh@LCwd6@I@9EzC@i4*uIh}`zyr$ zgnYKZ@}D3sucQ5r7`Bvp0p$U7u_C0IM`H;um3kg_nHJQ1i4czo>6q#Ch}a#q@I2X3 z^W`_Z{?XvX3peo~2p`7shCly*8gcoxcLb3(a75}hXgVQ}hXh@d7H>Y#OTr#xiF5_J zL%d%7mcRelKR85o4-$L>+7ZZ9+cJgg{IkD&n|Ioc{0F(({`vxc=2PD!ZQsk7uF-V| zxTdzKKfOS$uM+J!Ru{kVJxAdEVXwdU7Y9D}H-ClwR7myR)aeI^ixaxV2~NgL!4~BG zCF*2MCtszEU*Z>j>PztP|InkqUw=sJoe+5cF7Ut5R@DClEl};v)~^D`w4gTqE|0xm z_v?P$FJY%2X0>@Q{o#!_6nDS9(uG&)5ZvF(c=HbUXLu(A*vxu@M700?SykVuBos6( z8++|-)E{}}Rld+&r%sPk&em_qMLgVon7X@;`#mSE`Vks8bp60^`xG%oq{&(T*-}1xPzrxSItxx>j>m$#s zkWc?HLjNS)@?pFeV%(ydF-ur}lsdhSx_g|T{M0w-e&*N6HzQIHv9qfjPY+;qNsvUz z8@#0EWa`dvIVH$I*xaVuyh=z-gmg8OfQLwB1gltyXe<~)K7Nc^9!2Cb64q#(2vlUT zn*~zE%Y>I>+y}`2FF*HF|N8D{`q5vC#BZ#ySXvO$evNQ|Mq@g=f^(H_eTKvr2yumk z5i5(vC7e1$g|Id-rw%Lf6re0%IHLH3S4CJ-i<8zjiQR2Ho|1fv*E?8#p}8J;lF-#z zz!I3l#LSemp0FZ!+e_;xAR}=xAqjl91~v zxLNG|?%d>@ht#$0mDUsHk>QSFd&QCvvr}_(6Z<6O8mU>DDC|Ijy4P9kl*ShoYw7kr z^HU|98?u{*XkZATYkX4QDrsE-MO09nIzX;73s`2iq_^;|^9rE}(Rt@+S6U0dw(Gn` z5EdlST!1qG*a^hWn;IPy;O)S+N8hgyvjbU5Br-=JVoj^_S&_dNjLi-#;3jz$QO5R2C-Ws-m zjK!ojiY~x+-A~M{CEIqS}Ju ztT<)Y5Ewz4x^LCnvS4%~Mp{+M5b*58!;u;y({T$@3*CHtS$6}zWoR-s17N( zWF$0kI&`6ZhX{m}(Ij}zZ9&kLx?iG68a6~usT7=oOC+Xsn>QjYkjqCj8jDM(2u?n` zK$#XyXV+kRk=Q*??5`6RCoQ;bxJ}(XNzG?C-z0QB`Q%B0EI>*V?zGNx$^twg@r*iL z00vU$I7Xc|)RKw8$kPI3L=g6;k0J3Es)og5s0VoHAztR-p)a&x!11s#pFL5g1wJo+ zEHC{wWq-i5T`+Acl8&+ZEUEtzadDe)aFa9#rr{k{-omg& z>k8TbE`yy??R`iNSlB@p7*`Y`UY1xr!#yK@gBX5`e)Ahhn8>FOkjHs~2p?z*HDyH^ zK1lQvN<78)AB5!vBpA{S?fbA4tQ17f(0-4GCGHW+9aeX34opf~?U+svv2vBVKSJY( zr?Gt>jn%G|;ZT=Mb>@jwypXJYe@&!!4^6wPFCZR(}gl?iv2|S_c&39Uf z7Q_GXaCw6Kl)4ml7p|jNIrWnaeav#4iQOIYU{!r zzk!V}Bl2Z_;iq4LpZ*U$>ihMFxZVkYcMJSkny~x#fTw`*Ej<4#^K1N1z#Fd&-mm+0 zzwQ^nM?d)hv44!zCAy5`u|!Z(|^h0!c&B0Pk-S=etbmuXq-9_^0h8jEYr%-)> z5C=l+@#%=zukmh)nYGqY-68P?HlDEC-bNQUh|62}c%?Z6tB~R`o=+*;L*mhGrkhWJ zKZ?XF#Kmn=e;r=aS{O~BF39^U*zh3ajK*(sH-94%UD3I(JdZ0}d(3)7HiQlq2iKX# zhj3pK!oIa9TVtkbXT-hP}nc3MSG}@j4NCm z%n>bwZX%=%euULSVt<<|N7#OWn4qa(dBn?ty3drjMML9u;svIPi0?dIxqv}?j|%?;B)=a?Jas&QCFTPRbRnh@>Y zM0f3*gHIEUYnh#Bt&Lx!%_5#)HO_+OM7J*3)H9tlW@nj%ybW{vd~4x$&Y7!V`zB(g zr4zdX0yQ_TfJQ^p)Y`5sf;ura$8W6-rJ@n)(8lM|L{ckGH3+G-LTlGVP3nw|prIi& z%_I*8&kFsF8ZCcEAcG6`~80oN_X&XweDL>DRMRi&ICXmd0>}hCLN&udDM<&NBvM7Irb( zLWXfyh<1|}geDEuVOh~mTPt}~JUg{SkdCm(jSaf*3BAqxy@`D>frKVdx(h1a+Um>f zR$jNY0dXfRjc(nx8GFS_Vt0FqO#$tsv4ZCyHe^I(kDH<*&AHz>7L0rXHdM5Wv$ItY z6t7T=Aw}>y2XVRJmWk_AyhPjdeD5j4T<Vb)RiN0={3(eZQfH?`whBsVdu&KwSG7BU41O~gN65I}5p z(L?Nn?2gUo6^S&5;(T12G#;^y&B=~IS~#K07>;;sPG&PaH>bXpf@R3NfFx_~cWdWQ znWQFPlyZhHi1Q`jz^zu=mSA(6Ln+6&1SXP(9~8h$`~5nhk#Wf?|M>yh;~lRhSfyq zD=Fm%v+gaw?0VT{qe>z?p`GILR=nErdwE9Q0oz8dOxmLu+me8Of9R%ESGiwHr$9}UQt`% zEXaF9RY(UrY`hAV2=W-i*?)m>3c77`kuTA3h{$2jcvCMhBS1oJ3u7ka?PDy{AE7^Z7qMTF`d5jom+;U! zkLrGn4UaR8kC0EV@e_ah*ZC(u_0_jL_CM_PnWqbI=es`ce*OOSnLk;8FY}|n^t)X8 z%2jrHf#f~z6%Tv*dV`<+_zC|pxyA3-_pIItffwL2z#jpAj^@h!uYlJxdcJeL0DKVm zJn+anG^YD?zwX!fTbVX|KsEsd0*_RpY^ZcWyn-k(s|NRwiA3*d8v_Ga>o}%5YIbj*BB+5|9+Xtz`yV%_btS*0z z#l`2yyDeqBMAqV$f zQMMDdJ;4Zgxkcaq2z9!{v~zTOjm{3hABFxB{pPvXwIf3l3ei0&J+j2hqfO-^WTnq(TMvs%Dx8% ztZpbn1kOkYJL+_UIt6@}iLv5d5t&2Utc6&G#mKmyS-Lf~(7>*sK`E0qAw10CN#4Yk zyc4?hj0>2BMh%V?z1xzGHW#aSOQh1^lnHDE^FZMMkQ3q zv}+DfjZj4xkA-*uHacdFOB@zXS~iz!2#po0g|@xs_YLXXMXHo+M3MIX28a-nQEL+= z&qJiBg3pdWd=_~#_Z*EhQPoNlTSLqDBk!5FQG0Jc4>?=RAzBX8R_WG>5Frl%54DN= zbEcLRK}EeQ$ zS0yaKM6eRbLr3Tu$Ji09%+Aw!J}Y~eW*2c-HvzZ|P{$_lmEgqQ8fP{`K3*a!q~#Wq zfK3UBxiKwmcD;uQ!5e-7M|(#~LW2={xw}6~RwT5?Wl`e75N&~9b?njL)G;=}-{uno z@al+8XsWm`DZ@3Y4$av+tsoABZr{XA*CsxBbDWo~L~&~A8?r(sG>r{YpaU)|R9cW% z&CDF-Ce((7L7B+AzB$4bMD(sZTotM?19<8^?yxCoF)yafA&TU7yjq;C!R|33!KeLccaDdKWcf=e+58fde3}oW;hNPK26UXqy>r zPR)wtCOkK6$Q*bndqTHwg7bcjNk!rrOp!VzVj4(Gr;PU&QV10<*0#O0^PL=U)8-bI zNZo5=MQfx#EKH|?I*xem5xqh!0m?*JACpI+3?0~j_OH+%WNN;Mk85OFA?bt=ZlDJ{ z@-E>efhj~IpB|C7N0i|O%J>a@*MqK+^#N)9GShAec|$pUfY3*3J)9R88doKDdyomd zvEyBpx{p}Bj_B_p`a6WhtEAO!s85o4lCY_CNB@ReAEWFZ2MeUtt26#buxW#aj@aFr zUHV-MMguIDhOT;p>kY$+-A&^1BKwmKIeiv8R;oWm-XF0%d;vW;V|j1`AAXs~Vr#Xk6}^AP2}w zt%8I|=q5Dm+jXKA>a@hBBis$7qCg#bybFZk415JzW~9qSG%l#a2QXWLbVz)j{^09( z7p`3<%o8C!PmB}!_(Np5!cTnaQL28C&3@0%{w8k~l)wuuZ0&u(e-8W&;8%c8y@luB0R9u;bMN4o?$`agU*A_{jk}iC z-+aqY?KdY3cisbb@$J303xDo!Kf&k2`;paE7U_r(I?R3vH2Z^`V_T% zkhHu_T;C$49j->L2^$W`y9X$HsK;NzrqA=we!5*`eC+9t{q&#F9X^CEBa2HfqUl>* zq#84Eu_N>Y)A5I}>D}yaMpjq8On>+)-LPdl1ZsT&k+zTAR}?$I^1q@u5{kJ103ZNK zL_t*Oukee1;pFXY|2I+N$-C$do+PaU{naQTbG&84dm@*rri}_qCc$7b(@cn(MkPnw0jiuYm9deC}$5MWa4V`9mU<(mhD$~ z-J{_QSsh^he%f~{t!?;k-^O!ce|8BU*VOSc(jCz)cSvYAR-rqgY!9)#BBVe}d-9a9 znvevd6^#=rkvc|v?65APF(9V6md3!8p6Z*%*F>rLDP*`#T%M56u2O0f3DTlqoisc` zogK5Z6nP05r&9NkG6-dexMyrM2si1Ez5?bDt`Yk{T)spdt|7icV-uMqBs3-RuEQ;0 z*$E3llOZH(K0v+445(ZB^{eDi5ba5;W9q1w9W~s7w6L_{_C?$iAvxL`eSqS zrOYkUUBfGsu?1I^ads=x#5}8T9(S*?u{`V4M9f%>e17@}#Mr`p#KEW3gkhKFa#knw zmC!ZLse3>wEeOpZZ4CE>Pl}C!I?f_}FKDp4?c|zf*KcU!kC~Z`3^g|qukPC*Vswbj z9V!&8#=9;3Hqw~5p^1wWr<^t3YDn#GX$c|PYh0Gi;TZbnIP}6@rmu(Q!kpn8%?X>p zrLlxW_ey_U_}k@@fFJwY`OP1HLNyE_+NZFbUz4#XU<-dt*vQqAV$ z+=)r41@cat8+dliYVUX5nr`>RxAP#ab38-ixHbXWea6tndvlM3CcKKsES%3y+}w_F zFHpvaL?2Nbwo>66OU0Ls+1lP>KK%jdFKE z!OaL>8de}S4s^{6+!d=6Aqj39>bPi3!eFRPxCnLTZrA;~?Gu^Mv~LR)&^Ay_SV>KD zEtSw`5XW<*?E7|_+RQPwhFMU&9p_QncCOMMbG~>yi<}x1U7QEYmBzxYimIa_6IKHu zrRL_H8pb84HVjF$hLISpInYBPt&B1xyb3|v`6}VwF?4?XgA*1_%vEg*PHA@jn+e^z zg}V(~0Sd7{qvnK4m_y)h6XGpgdulnN4hM+dq+btA`^a>5sf8fMCcaOn2|}6YR!38z zPO;6EU2}MsU2`l8Bi&|))dgu&v8hsvQnm|nuB4O+^%SVMM?xG}UVI+yChB;T>GTP# zZV@JQafZYJ?Ke!@2iV@2Q8IO;3^#FomYP4$ZkLGhLzpIzYpB9BEwOxn4hc<)OGZO! z=P?IZEm#hy+`uhT%ZT$OE(iw|OFHy!b#?wk2AH?#LNIFGB29}_X z*^)Y~+FY%#q6@)g%W!r`w|V2hskl5P96vyn1Ag%@eT|Pj^Dh~$Z>gI{sner`(;m^W z&09H>r#nbopy64%gD=q?e6t~+N@Us}G2VHCvi(7vO1Jtv?#{`JKaA)K_xao|j^{eF z+|Lj;s0e9sO57Y%CnfI}!1>%4QR{H4NLn@LKHgZ#bx(*VgvCu%MywtYIc(b{tSHk} z;>^XBG$L%U>4>NUVX?yNpMu>V#r!`dub<;5K6Mk7=Q(-z*Jjb{JJ-nzH`zRKn|^wN zxV%j%JIa1dw|*1OQd2Cial4LG$Me{(IgWbz!{_Oae*f$3#f4W}&>!zbPX<2r%+33u zV)yHQ-LF5$^-c=B&@RD$dH#7jbq9FwJ2#yBb-(V{_f<(=LJX+<{uhI9SZXL0536=V z_!c<4Km1JM(NFvkte>J^KSt;VbbW^|wF|@HW}?e)Ej6&-}>({{QjOzjKRmxW!^~ zjePnD`?vrCSf-XUth`9sK2Pj_lYjW%-3y%CJ@SJr*6$;(cBIv_SXsPI;61Ojyp4w6 zWW4zU)O?Mz7gxmPtK{hb8=6xlj~57@sD7KW{Z+>C3;g28pMBGx=Pj2D^^@;mxq6uO z(eE^u*LSJBzl7HZiJO7YRqU*%YDD8hq}A8%io179l+YzYzSvws;W<`Uud{vSJygFz zT>T!KhkuKFc8K~18Mc?v{llcyZ+%yBw+t($JJ%@n1(sL7!gTyFW%q7=?4SM)zxd-L zKl*eerv$}8mAPfV+Y_ZUkW?5VnIywf>n5AVBHKCFFjRg=}tZiZCstM3lV1NAwFm z(p(e*t_bZV8n<91OU%tcSb$E;0|rV=%xqbH8lLD1qT zhuC4#7FJwfHT&&Tih+`u5i9{9~;y$d|#FZL`}{)P?G} zt4wQQS_dQ*XO;#iHM-_EAJ1k+JcQ`S`rNI3zTPfa&#vGBAH8SJ4P+Lb&$lZYVGftG z;$nH~0P}w@8k{p5{XCn_5_NV8rbe)H2W9nRAcOB;o3qSv2XW!&UC8DkVKdr7WEQ>E zkEsM__^NnQD|PPNwwe*#g+v5H85@<;DKVL zUl0;VQbejZO$;c$UkIIxaBW4hPzDc(bNoU<%K|GItBR!!HbiU^%tp$vM#d$fg)9LJ z?v}5^4%2o?orEMxYG9+NGFja1O#@9fB$$Ud zR;7-C+=wGICO>DVF(IO)9VIU!vKbjiL8l3`j_%+HBQhLyl)eRNNmuR@%MseQ6y3#2 zMC;6>mWZ(M23M?*W<*O3=;*&Ef>hG(Dbj3_ArN8@3qy}PtX5($n!VfP)f?W_K8kH0?Jb-Y0|()9<0-k7&k@Q9lOKV)bbx-*EPB zJ)?w%P&?A@hO|7JuelE0?o!4=E&B+!u)3yMz48O%Zk=`@W+Zw8I<=?7^cRdLN7VWR zQg^9#h}A1#EhP+?tqGhV&0QjaGAz)31@9p65+^LqlQ-ZZq5X3{{p-&%t#1qAPYQSW10L(=Lv&@b1rgQ-1~Ql$7r4 z-qZUo>#{EEvi@JFAGg5!<-EL) zp87*XD*gJ9>GWfS%~iI~?DOfD{)o?iZu8zZP>g@d*ZB{A@4IZjdy^EeV=R5vS_X=q z!1N}+_g}sBwteu~m-eag1@iJ5-N6@``cF{%XGrbI+gP!WR&0)F4!^?Y_TTVr8}~6e z1lbVC)N)E4?o#@DP+sQu|G`gotnC+m>k4^sh=c*@9xqIPfAuF%6XP`&S6-v`*Qnuu z7$PAaqvIZ_dvGxlEeLhI3bIRHZJ0Iz?K`SK%%_wTF!i9kymLak{Fjt(d>E~XY5g?F zNV9YA0(WsX7v66a1b3?vxxqN$x%X@jBAnAg|t} zsv#^u4$xtrsVJfawT2J_Xu%|UXiAO5`Ae;7ax?+a?P9tncE=zUrWK+_*)B1i5Lsaw z-J!}ZF$6-|fZRpL1*BPo)aYs8VT3qh#i&E2Oc5O>Qu8{>hS0rCIT_IKbJXSr-OhcG z2E)_f)AES2Szu+KacpK+X2BFxpvi?zF`@K<*o;V=NLi6&3@5FR-1iW2x^NN4)@GJqaF%Cv z7p;!Y2@XMKW}&+h>*(TYTo`FlolVJ{R)()%orE$8=IBzyX5sYQ8C+d-)d{rm*L?pY zo9P_7HoJ!nr(IZbA@)2VEktL@%KYatxU;erp$xu8akPi7*x-&_V|K{Khxmq&UFh^t z`dPm_7eb@QFX=QpFr^YwAvW9j++4aV_mQBs5M~&}`KbOpjem}9U^aP93iVv%KgXBG zhob+v3)7vy&NP-VkNBHXMzEQ?v7qj!yKS$M9 z-fVW3huK}OhD2vUrv~E6kO^;gG&Z9yD(4AtRwELFGmo|Nb0jKK!PLX`Od}GVA)r%2 zG9-0wvrN8BGehcpQLrT@Lvn*z167gWpY zA>|H{4%I!vG(mA(N9muy^azpr5FXI(-f{tZSfZr^i%58u5MQG^`~pQ6^mlJEoxaZB zeDe+d^!Xcf`8%Y=6+%+9j-;lbX@%7lS_1X7qnlhA_Mb`B-$F(#VjU3d&cp^kf? zO3V+j0TM^rE8nK{uT$#*HXc$ZA;=EITco%l=CwZ~)E$t{$N%4VdM^BkoKztHFibi**XbilX$(>J%SqbR+3RZhk+ar7uk-veR{WPC`>3`w#pZnT7 z?mr7(r{6wHOdltNEwyYJ&aTny-Fe6N)a?#+*rV$AFdHbFB~!1gu4THbZ@lCC9#_jN zx0%*Y5t^HP>bE=o;FsUqbLM4T)@5DRUvm981m6D`_#b}^%Xzo@Pk{g7C;S#mXaQFX|eDMj9gUD`Q`$7To24Z(a*&Jeah_pK_7Pl~M8P`|*VYdY8u%g+y&vO4Zw!FvY-UlK46{gdP zPrdjr=8%K;Ucdhz4#3Ove~+nu`1Zf{@(j)D3DVA+e0`!oEV-v})C|BT=Hd+(i5{C;XN?GaZyG~Ihk zYPPnOs6e%`wWtlm>ddfHB-if*)d`-llQ($doX`l=}riz2ZZoE zMN2|Q!AhmBBU%~{Yf@Z<2|d!B zxVQ_1APmD3h*XFQ!H7|bT_iReLTU-Y*}cA#74AVQY?dJBy2#AiJ>6fh_L~= zL)rW*l!015Mq>vc88NvyE+GuWJW=dipftv{hx{3pAc|?EPDb0Dk=u&u3X6#l-I=M| zYwl`nE1|gudWAA}&S9+CgTOQ(l1MvXRY3*Q)OG}0>+GlfDq^Hh1ku)ofK+KFgOk+q)NLpG8}dErwuVW zbR;!KYJ!JNMRg}2Gvl+kaB9h&!e&b8BLs-qhyj><1iqBUw zvoU*koAiPt|ID%x($dGfId{6w3TcS$*j0s4AId4`fl_f6XG69N2XKf=twt?jZsNf( zhoFVVHApcRi$iv&uFbcSkZ>kx#hsG9ACtJS+2w0kQ4sOqr`49&RN^?fqkC8p7ecde zReL-Oq{Yl6J+JeCH)b(^=K5j}u@vJ5+C6`b%)wnD%z<&|x6t|jeoi^&!lWQMc@rg7 zcY=ypfu~?ll%0#KpQ0o?{4`x)rh;g4Xjvx!#oi_;W{fo;X zwrgVRP4jWP1l?e|V0iF6b^SH8ewi0vR6hHgBa`lfU3CmXc%A9=84th{u(D*V3DFT` zB<54{@&TdSLRnCHVd^_l_x9&0>pReZaSZZ9TCWnb z5c0!ktq>+iTdeF-&R#&ubyOO<)hnd#m@=-=@dlC(3Gq7E6;k__!?o%et(;r226PyuZy)Uf})5TnfCGby=4M zcPZmBWpjnpy+1(xxIIA2F~U2Kfqj2{>bE;oKT29Y!{Yi^A1m%&RD=*mf)w)NHMHJf zTpMXjl<_(tUL(W}bvgju5}NPP=3~aQ1GYyWCO3bEG^2~fPksx1{bQ(pn7tcs0bhS# z+xq_N8ZqtD?Yes+?|zT|Xa}?Bu+y95#anN4p4M?eSwDlsYvgvI)D_copWOZ_i`}oj zEoiQ6S8VTp2vSS`z<~dY&;IHTL;EDkA%SjowDoAZN0{Dxf3D@Z#@MisL{0A;F862+ z{SF~4NZCKjBobqXaF@LJ9Yh-(Xkf^ms>V2A>R|vA}xPx&A z78|G;vqZCc4IweCAE4DoDT}zv$E*Q8XjKo-}Q>OF_XQ+-!=>-X$f00YoLyw}8CKo$&T2Qu! z1c`H!{e(y*t%Q)zh&d4&cg+b0v^)11gJ6@nqj2{vbj;K#VVse>dst2Mr#Gqnk~(~a zIQ~3Da4F^)4`nWQ49~ zT0enAWpVgzw06{uqSFCdc0B+{6RCl;@{#qBDpm!p4fL+4mBtmJFwpOTUxkY7|zA2+3l?MP?)SG%xH#6o5cf> z);ttC>#Uz#ICEX-4b3cMmKoiUp%_R|a_a~Okr7J;vFHw8bA*7OTXK$3o*AhA_p)(e zHs*(}-HQM(&v0@!?rhkz-gR>=&JMOmbQV*iZh_H-P@Pc+W;VN%6S=8KBZ#?5EhMNB zg1fL)9ix!C8PlLpCKt9b-_If90>6r1C&7hc20y;FW_0F?*jilRNR_fpXmRImOw3pp zcM99oIzEJf*lscFP?9$$lFlq-K_?IEN-e~+LE;vXiD`Z4qFRG9fG1%%i=>4iNgOKy zNN{I(YG+|}ThTr+ZH;*46@s2ox7Xa6+l<607%d_b#4Xe%V!=7eDUx>v%9MSYMl)ER z1%5*(!-}98*K@}}R2RandJv_Vx%h>GL_*_@11c`~#tk9%#B?7JreWb>jv^FoE&|-@ zD1g%LjiiMKK2Ck47H@8tg!vpa#*>Vhk)pG~Q^P!I^RNM?jbUf*2B(TQQnqecH>vJg zPJ7t)lZ<*B8@_~%4Phm0^(Gv*FkXSQAT@8I_AP`K*a4}e?w&8JE=a9lwX$Uo?GuuY zXu2Z^J*6acvjh2rcK>za{Cih6XNZD?fwcHGb=XJZDKUMSBpa$d!?1aZRE69P93_nv1s{w8(&BERqE_g-qI4jpNE zOx)QpZJ(!(TbiA>n*-MOpJN(c#Vk{Y z-+h~$EV9A$23mKhx=+gAA+C<7!vR{J0uD*7V$%e6m-dPq5UeKpqh}y|+}-Y>;j_Q` z$BzwR``v%=I-h>&Pq1+!?)(*|vjenVr*5v2R(FZbfYz2e?qN2d+q-C;(2^OBzKq!G zgzn==k{?+1d#YwjiY+BCezf;r)@5DRW&PyqUq?j4WnI=~UDo4Ex35#;9j3)k(H`7> zcfs|M3J7KW6lMJ_T6yg~f8N1(leqT;d3oQ{hTo<7lV^x|muCN~1R_)4qvMKUd>RQm z5GJCXVb&AkF{!;r2oqMXqw8z*$4|4k{@rigQ}C=-E7sPVnMK{Fvn$S$g? zHXg*41|$~Z;x$4&q7GN6^$OHo2m_(^m{l}FOe50XCN?8FPC!Q;9n&CEF;f>HM97PQ zcK@4*R?5_1BE&qt<9TH@qT_W!Gtw+xyAXE)^6Cvvk3UF^TST^~F0s0x_IY+PEtrN5 z(%?c=6!Ox?>>V}`+?8jBm0hOONI5IS>|kC~1=)aAV&h)~tvI_hwrdwlrvx0(ERGp9 zW8(zvL`VT0_NZmWSgsJn2sEfXKw^*aEHXZgF_D^6B=wZdf@)7-wu6-&UQcLjwxb+f=9kE$MFj77tG+WC0X%7r4OJXaeRYj5!tvb(G zf(JfnBKgERHoLI|A$88r)hWzl=0+V{+}2rq)Dfid5&swzi|Wf{M9YLoM{Gu7p8QMb zlxA_U5*H(SlEJFGKTASOrcOo-vn#fExEkhy+6*v2tGYw8RWPAW9?nJJoD<*zt@%2J zH&{ zK^Y-T?tHDIpj#LD^1|kA0NI%ZN1GkHzIx8}?R?$gp$w=103ZNKL_t(8u9xue&uX|c zwTDhDgW?V*{}vv`Wb4g11D1_;xTtdE^#kk7`lklh29^&}reXt>f$< zPCyh=3P|>2s=ze5%epG0reY;RR?>3x^RxOnAu+gf+T4X3Jd@9N;d#TsaKWUC2WJ+4 z%{eBdnqYH)mjqn24Zd%vky0W$P6!Z_HwIu9^uzqwvj{H9KocgB)G2}BIFBLBd{_^S zleQpXO2aTLGC!jly(5cjY_i;x_%s zRb(2ebsy{nLen6jQm25JJKQBe9UD|VurNf(onNz9Kr`q76k+Q9Jhv=Ff{(GIgfM~? zw8oj2tAxh6-&4Opr|TfMi907qO<0*A_J|0937dl-B~?~49zLUN?mWYE zwqn|x5@LsLe}?tlju1EAgj_rz$A_-oxVTN-Ji)ZNPSd=CV2FxO_}FW7%g6I~$FoCf zuhj7k{@4HF0sQBWeHp*>(uauMC!zUwXjX~#%D2d?;|n3xFk-VstoP@Xwj;F)wL|Mo zwr7WkZ7y(8b-YPP3+m81zN9&Wb?c4BRLILUCO+vS_!J8wj%8?!&E1BwU67Wig#3WK zxI@`|)SDpR=N6V!V7tJk6;kdY-F?Iw%JejZtK{wk30q1ZePX0I8cAYddF6Ycfm&|R z>EC3idwljc|C|A^I?{2>?|kkAFh2jeSNQBp$83+@qG?~D4nND7pCq3>howwlK-i#l zK)H_%uQHsyhR7F*x&wFv{70!Mbhn@C<1^5RuO>XEQ{n+V$M-D4bHplq)Z_#QfMPs{_M zJ0UHvviI^0;QjH!-cO|(3E#)1i+Nyi@KvVuG1KrgQ{6+RJ*Wq$K1oQK$Xlpx5ZNJg zuU!bl-(Rtvi0zxO_;+dajJ$J~w75;`UL~XfBG}k5Z4M~Qk27tqv--w0KK;_4yk|GB zu%S#nWjcJ1udn`PUT0-U8;pX?V|x+|38#d7O5JX#=HWguo*|)N+XXht{Dt;0l}CgT zom%qF8Z8~#FG)N1$juKA&#Pq*vmHX)v$*m0ro?dX8B*$?`z|`JvHF~kCtHI~3#J}Y zcCj#b#T}2CIlQa*nDwxrh>vrZu|+~=>?0`zlr_Nycl32W7HmO~;3KFaBN7L9O3A(l zGEEcZOfeka5b_FNeAO73&W@AwGiPY4LV`4$O}m zrMPewf}tfMW(buafr3y6nFH87?Omk00F;BTJI+`Q)EVgzsC7XI8Hqg-j-aftx`)+B zYH$b*N@$JPsykt+XoaRz7m5UIiXN0C!JV&BNo_!)^J$4j zXq0XVNT3XYBo}VaS-TPinazu;%tG;eP0T^^v6+89;|HwHPFS5Cw!uTdY;@P_)VpRE zTbOh4sJ**T>wGyZt{Q1+=vEo0;LVCTj4d|4JgqLCo(s>M8ypRAq^9NT+06YV8`MDA_~(iF?n$jX!l%=DU1pSo zAyC}b93$jKab(5_)5gdPy$JM@B&Z2C2FmCRXH);P>g+In)EPX%cqD{Qy|EJfxT!)& zemvsL6X!!s@p8Pmd_PYm%)ieCBa^xw%>}JjnAXp+xb_81l<9N> z_2fy)=1-|7 z{|BpUzd*BhN?yLjFdQP=4}mlY_nFSFQO=$srjEE2Bo~5=ght5=W7;+h_ihrFTf~l$ zbV{>4dixWY0Cj_w%(S_I)*i9D#KwisZub@?KcH-%hI$39D=OI=W?LACuQP4_kn8O~ zxq0rA;i-C^v07Mf3zmm_)SjJ$fhU|g&vBoBjw z{kf^tVkIJR^5%d*NIM7%v^;@?(}zE20r=EPY&V3o^?OPQ7?}_&W)TRKVF2k_Ebo9+ z1jQtKS4|#yZWJiviaIWd`75-$_o%~7Bpi~Gi@U(=lGlrd65v5;c}+x4wi9e+<)esNF#{EEvi_3m$0hLo5}qjelMs3T?VrFh zFYB@{>)+Jsu6&8@$rU7h0`$Wyu6>7)9`E!tjcjgzfU-%HX960i29!B#84dV|$0Rc!wZ}pv1go8h02sJEY}9u32nONX;%~bA_Lz zz&p~@RB^HMk&#IWm1g%<;^Ix}cmTsLWBC{{ED6mWqdaX7oK4%e z?gFL~iy>jcEK$Z5v?4RIBA|p&30d9gIUk&mss{mTAc%Qr)EojQZKVzo1yVFpQ>Y~( z7^Qaxt^}D;0DgpH>w>Hevzapc2i59aU6~OHrn3+zauGIW1xT~b+4$?G?d%?I+-b@j zm{zv#1a5Z?9fY9-toXmj)-_uRvy*prO4rGe6Jd6kULY1Q)WmT)=C}l_j;;u;Q8r3y z3(^QSHGV!#pk~A(lpztbT{t~Mm>I)z-td7k3e%QZ6rUZtTNkiHav?e`eZOmU0o$rU zRAPRF4UElW`W}>J+9Dc=$)HrKM#hbJ;Mjcqg<0%2{PTy=9l`}KBt*r}VV$LZ|vlE)^fn+5R8$+_8YShuF!z}PN z6^Z9fBO{ulHEarwdLBOYT0^ z@n=c-1~G03X=1$p0a6oFXCKntCcz|5HX)Rl7v}ArvMeK;FBB zP8qGwQu_~5w@>r0ZobTC|J#4b@BBAg{^_rr0;j+?FP8P$e-gQKUEVeU_WAk$4acaS zQp(fsxNqOBblOA9ExPXLM|%Hxed@);^}P(wj6bnlip#pJ%liLq{kR0)|LMoHoXfhb z%et(0uiyE5BY*oJ|1omckSiZX`%e&8cgV{l;?f--+PBor0sZzmb$gF~^JR4SQ~uy{ z?=$r_Bs7EzSM58MPA!B?$UoTCmy`&+yVw}g=!wW<3y%oAsTnZ2ic0tX*86={mIv5$ zN`JCXckti`e(vencc{$~p+1S^d$hCja@y`-wtZXr^!L~H-V2!ABrQ)7S!2@<@OUS6 zkV5y=H_$geg4B<2`|LjOXFqfs{@}&JXMW>*Ov5wuM^Cf7_R71yzS%itJX&M@4N5kU zJ75tlE6za%57((u&Cvz|mjQGPMe+lz`WRtqH=bUv8N>xrOp!DY zn+X#fp~4YQli#U`7>ky1VZDy1yZF%zPax;KgK1}&ZMpXNXI6?8Zt$_Da8tr;nTkq|X%jffRuS%dT-8&W>T!V_Q%tajLxsZ&QG zS7IL(1t8O*3fsN3jHVU0e9Z;+xOS)BuHkP^hs+@&=O&o-kRB)q*v zAA@2obJ!K`)XcNvd+VIpbLZ*B8_zKi=JW`3~D;Y^;~ zhfxU)l#Ng=xbr@pi-o>iUzb^6jnJ&X2#jYBgL46S=dhjNs7dWjEWF}SV$=n7Dm}L=$NR*hz$sV*a)euE;O0% zCpl*roAZ)$TaoC@=UO9mG^j1YRr;g9is{7EKSv@HTcg>%MQk>xc1&kC=95=w_C`V; z(J@fA4P~30GcC@Lj;&j-O&z;Z`$`=^@sA2MwkY|NBuNV?1N;0~p~ z>H(HIp;JZj%-?P%LNj>Sn&Q|WDd^ZZ!`q3bE}#!_#;FKIzqbIG7$GWR8AC7d9CeJ2 z>_`*XfW(2+oso9$I2K}j0-Yv|V+xVHIHBEn3!82t$JbH&IhuCqjNk5rkVc@RhD=_L z#Qa6Z%?--ghtMI@PdhX_rzDAWV^1HwCP=x&c`s_c6Y)_sh@eCmq;&S~MKJmj65hNzGv}nzKPePG)OTbz6 z778JpIiEUZq9jx+WgDPZENw9x=Vn*HLWa`d-MK=oTi}Gn^?!kt5go5EP5b_=1Hd(6 zbBbA_)(F#xjy-vGhrBvwI(r_%l8_%hQ=Y9)J$JjN4~{YlIw0%`H;5WgH42 zO>_ru0Wh4cn9iPqyicLyQ@{O({J}5(;4}TZ38tU_+=gHLjoZ}nLB_Lv59@o6H4O{O z*ip+p{^wu55C7SFek%aK^xFp%eIDTnM0Y4DqNjyV|Hd)cTVUV$k@6=l>#{EEvc9h_ z8N8QuS(o*1Uj4KGbj{y?=?{7H^e$o_B8}IWPM?9`Y*Tc^#uLi;Rdo9@I(_A_sk}e5 z>fqr>kGM}BS84$vxG*%Ash~v63ib{oc^9u0AB~+PkRSbb6@n;H-uXfQ$5ExcNho&- z+b13?@Sa!D1__A}9`4JszJlrly1mE0uWWWH(+$#MNw@n3WI9$PiXE^hYwKo1|j_g zX>o?yhncqfq{Sl+jW{7?^7T$o`xRCPB%LC$GOc&eqQrPaZnqb2TkPHDcypIHevmrs z6Ks!hyC49_W-@97jxDP%&%4oL$H)taxrvL}-wx#8xQkK|@X$Xc$j-iHqbex4d;6LtP;jW|XQs zyd?D4w1;FPb@xfjHwbY8J8&+g1hg(uJE7^k36Pe@j3=)X!xIoj>UQPqKb_SpSrEeH zU%EA-nq6dNM604>LoJomjD!?C6}`Dd-FBXG9s;pZnzVJU=XN=drn+<1%!O-K)Sa@iqGm8=Y6!%j{-sv~HVP7rm`^-ht1PLd#ijwt zrO?-rUKCf9o_wxIweFGj^G?|AhTFt9(FhfbV)E5;AXKBQ_>Q~Gw1u3Fk>2UVJ`3x zwIBhi0q1CdS)2{$+}?h6p{hGg&z-<#LTswDX$L>gw73v@?ouWdbe)*i(KS(+MSK5@ zm=sf?s)ygj?16cuk4Rk8EVt;iBE$)c5H(PzI{#NecOY~ z+c`UHW!6J=4t&!JOdiB%<17{jkYuE;%zWYVATRK6wF+4g^XKyrAT|D;nakn(s?1_; z=c4(v<$@6$^1R=QF`WnG2`*@gFdiqmJ#*JJ;!fnr*U#o(b2bWA6jdQ)nXmOnj+Z!# zTC1Y{ET&0BH6YcY%y}u)9%oe}v*;GE>H%(ke$5B2`i>6Ch@H-86c7^97&ZWM9UXrT zXdy`u{s0l~uE$)H>g5184x@Kpwy6L6ob3ahh33 zyMwnFjy{R<0$u(PF`p3Q7At#b70Tq81BrKuVWQc&MU07l{VbG98H9HK2&|`sfsi(6 z*+n-8l)gc0rA}9zF&<6`i`&$(Cv~@w+!4YqGG3#O2b|ry!s5m^E`Cqi*s%G(*?Y4e z%d+dt@3;1F&bdQOIaFpv!!!4;L8x1tze({4} z_y_n&K(|GTYT14y4O^ln7)GOI!*&lE=%z@Dr7E&mS(QWO5E&V9hjY%}do4fg6Iqq4 zkuoU?RxBWZMC6To&pq?`*0+A&{bw0Z665i02>XP^(}Z@1W^tP^dkyo6wA!b*gaik2 z6+!}Yh(96?OGE?Xw#B)YFe?=?Ad1Mgc2}H3heC)Wy##!B0pl*NIeY~3)$Be%^=k~P zLVt3Ju(?X;RtV?OB%Iteyn8_I&r|x_q}6L!-lBOxhb7{6sOX#l?O&pl8?^ixpZ~@8 zMuYYD4+x7r%%zGpiSKPjD7M7;iaHVT&}c<~z=T%t2zi=R+91|p(ju-!g2O17_DHyn zXp0f+^ox&J-UH`w9AfeYE*ukg?vwimjH?Ueu|sohD8^hcStDVKc}MIT#IL5t@C>4>?4Yu32o9GZ84D)2T9Fk7SJ3*`B8>}_7) zL)wk4lLg4+2 zjX9jvS)J8c{eP)Cw*MtQ{^^$~$B#4k^EkRHids*xLuTdXJ0|gd&=Ply<`wyPmuBxh zZqP$Q#*C&2liAivi?OMrt?wgfN(qQ=ABK&bRw^B(j1>DeoyXs1{my@JgHL?sb$ol7 z?a@VM`|mtPSgeldp5A(<^E-J#vRODTc0dJN4sXm`IQz|Sd-KVo$_$NL@B z8bUasV3=jnb`AHBX%Ai}&JM}jUA9Nhl2+Gnd5sHKC#A?So;*V-SBd@^%rEh?+xXB| zr*R3CZb2|#!;i*&Bt$G5*%niHlp$1A#l!)pH6C3DEEk%2B98$b7nlY@^MEj4 zQ_>8IQgX!{FpXH67$tXj?{L1ap9Ns9a%fes)ZyehK2(uDHY1_Ah2?ADuVDHBD=Rc# zKwLy*q~r!G4Z+I99aRiZ-Uw-%2<=E*9um5dlIGah;vA^kg0^7QB`ZN|0M=%S>M*ed zC&5=d&X0AuU0*R8E`h8u6PzD_0h6wdu)C4Cv!3+MmXaFE@jhjUa%o5Ig#1VRm8 zDzUnyeZukxwgg*E??0k>0INZ5$eP$Gaj^xTv2hn@DU*sA<_^_J>Z>Es)FTmMV@NXz zXs(K?nyOe!k;jfX_AqqJ7o(XgX;TB|EKiKy z(?F=Vpp4C`;G4Hq01neBP(~i|U0q;u$kxxQLaO5C85bK$cGbDtgKG-D$%tk~6LbVs zU&YVU5HST!zOG9>b;v^6dXSM2JyzzJ1yEnzs2;qpJZ|r`st5aISww_D;F(GakH=#C|L#Ej|CXG)rjyo)- zMyzv&rl*MANAc~PFduPVaXw>Zj(0m1A)~vD>quyiaB)qyJY-zYDcOyD zZPENJxsN#aMbhRKG+oDYlk8{Y;VNMm>v>&PXt|CYy~G#(>-TR=17z0`e1o}MPZsj- zj91l&V~*hwO-N|)0&k*Jq=L?^Vi$w~)rj{SH0@UhwJKT!?^n3?9x5IgTT)pdrf6<} z0da?frhbReZLsnH?O!CzF(OM~pLX{ealWO$|17CLB8Z|EQT)S<0#jjJox|*e(A>vQ zjq9>K4{jgVZV)T=JNQ)ZslA61$*ChBKaFt_Y?o%fK*DA4k7ITj;Tk=+`S@pErrZD8 z`yvni=U=|bPk;97NZjS*)(b2jd+EKAUg|ix{VaJikhkBUJ^1Da+NJ=X{9HqS^tXtM z7l`vqH1iw8-Pa!ZJR2LvlZ%YAA4FEW{LE)|`O#1PA)orxch}f)R%dlq|G%y8gur|E zJMbQ7byjC}R)4XoBTZkQe!jycZ;<*!hSjC_6nJB^X1m^^<2g>z1G(P^yAP64dkoES zdhmFAm2r-7NZUT@=45k#=3A)$E}v(`@c+Y?7$58-@&mY_w99*Mf1OJR8NoAdcF7OU zW8*az7r*(gVRxzD!-Y9vey6&uq#*9yA70*C?edWEiX{0K>f}(R>fry-vIP7T2#j)_0#Ft$vg^-y$v%f`g`xHHsEW+Nb2F*xr4LZP{XS zo3wce`ma)&o}d@denCH6#L6>BIEOU%tMk@v5SOcbl(dJI2Im~O`)F=(w#BHTt@Da+3)TnJGNNmcM9B-hdWgq$ljX`Owo=tl9~0(| zN!ti&bzPeyEH?-P<7$uLK|>h|q4oGSVK`t)ke51gFRIneDvIhs4MwsVzBxo>K+8U9 zQy;9G-CMY(pyLi|bz5liz<_THK?42#049XF(d`Wt6QB_$;BOu42k}N)Q`JDNLR<(5 zQ>~8S3FmN%9kJpn4;XnfV9QM$*CM2?tBl#1kv2jaCdX^2uFg~1faW)yyFA1%Kzo{N z<9HZ|rbFFr6o*K~0hBRfu7;l3#Oh@to& z9^ZhD3x?GUA>v)3AoW-YD=aS&(MjV?nDj@>{ ztMKl+jKfXhB52NtDhZDcLJ1Yqpg7QqhN;0{f}1O@QAjoHkI4aChH5UhDbP-Nyrp9N=GegI^Z(-*=8EU2 zI>Y-tLR6$#$e9%r6xh7AqQuL!fZtkfxe@1c8nl*+#E2G-XW zvBUd}kH?f0(Y%We4qU>86>$cvJ!O+oos-91%KCXsB5E_l&2eo{=x*SeTTpfoy@P9R z5cXb!)qV1C1u1(F7T9=+Jbsk^wK>lJJBI#Mux|iYnD0GJ+`G$c|5hb`SYSMQ4l9=s zdkzT~@QuB>FS^RlFZ~AZJbrPA?96wJ zzx@99Z#D7^9Hp|O)xYa&t>48(NA45i13qr+u`$d*BEpCSsnY;rsI^#*XbCtuspuaQ zEH4;VPZ4H^h$)61EJR^Ax?#?DCgL4=z*6(~O%-6k$BtCvUDWZca$Ye@?sihaa4oYr^K&ew=Riqr~}w#icL5qmV4kNPGS|eti#h z&y#Wd$(1eer4RNwXLVL*byknoyI(;5^6RhCviw!79|3+ASOfp<_jp;(>a6~H)Qi8h z;)Q>64&Pk`jL^LcA1fxt2j_r`mB(2ca({s84Z{37q$MIte6xPrpXOmf+O%NTX%B9_ z`TOmH{=r4YqnG)K>tEpy{@}Yivh}s!-15RtZNMU=XGr5Njse%aMU?i1adj2#KU9;- z%jA_R@?to@-Na^<}xyLoH5#C{4d}~>#pO;s+c=30>@xHJB#AiQ5 zh(AViaFzDrpAy0wm=2WU(A3qRY~YSSZvuw!5@UH4$q!I{0>%X<84ODQI63X0+Xdrh zPTnr6yDJ{zd?7u!KpyVkn%CI9_>G4Ww<4@>K2I5b81Js&+b^?o>3>7z0nk?+R3-+2gk31}-^*)*})(|CE z$2K+BFs(YS=>e-EUJ`QeC}XuhsDcjKpt=vX0gkY8h>Ig3-UsVZ4IqLJg)){ZZirRKU71gq0!33- zMJg8%SI5&ePi8_>$QfiyoUbr4WvI9V7ZR>LCd@{RfNrap?n1%&9qY0CWc{O+7c1En$8{Xe#f!j2`ihW^PE(>Pkkci|^DvJBhSd z6NK+w>Zq+&jKQQURe?TFF4+v04YevHS`kW_T)xvt|EaL5V|DFnnVh+y za(_iC(m*GMt1$_~GW}hi_9?2^R@2j42xpY?JtT ziV$cQBe+6;6iMqDnr4uz(^q>>9t5#mClpFk8N#k|FMM>hkq~N#oS$$Jr=fA;rtji; zh|-k+?V=FoxsKEO`bD*XsS=hOTpJLR+OSYp4YILJ%B@3!Lwt4QRcEwdxxPoqZz39e z%R|RCyeY0Tkq;iar6V;wP83nax6mvE?L8tD`5~by<0V&7H_o(vFRPYl?*$1}m@NtN z+99UIogR?}GhfdW#XZ_QNEOrI{Dig$lcVoM=pP7H4DSV1M;_h7&&L_AQ8Wo@E9fBT z*ieI|GGTTH=X)@Z*+7_OG|zD^P6Bv6PQ*E!IDD(PY59%%zLx5!xfM4_$qpSlhSde& z0zSCv^lvjECNwqG*{w5ln4#kw?*%^_tB95zYSj@Py2|qwiI@bu=b@n+vC4nW)kPfU zhM!6O`&5rJs3JUa8n!3Zy=@@kvC>d-M@a!B5t^Qq_oj31eUQ7P)jp~JA)J4<@~7M5 z`h6yW*U~;B3nZ>9`_*SkULvJnQiB~!YALp^=k$`0#&Au&M*60uo!`UFddOEO`G7Ji z&c8;S-NQF)V)q8(?}Hu@{dLM1DBJ4{!~exq{l7WMbK0FBq*-(^aGAu`K{2kF=^?lZo>cad8AZpcId-*4XSS&CUng8ZW1UX0 z+Gw=WfmXDSXh9`mgQ4Ya4P49A=BqpaO<3tDx`!l>;c#K62DI6NbbOV(QM`YZ`T19A z=J!dPKt6t)Je-5P!Z+8*X^ECgP-c|zB6;gjJ;sMSvL#rcAZBYATk&1+`o{+zgXhmx+~_~F0*I$!<# zyWcb4{2k@-zj>X|D3**|E|N~3X4veLHv5e0i;Ra)GOioO)$63~pOH8Jn&0~Q4?a!z z_bxv}-1!@{I~Un`{A-ARZ(-OqnYKM5rRUI5`2PRs7GL|l4~CsMtFt<*vwBPYRbue& z0iOdt2E2b_`dOXTS)J8)y~IfW{L8qLbKt%oarse~=2Q+N?%bfP4j2xfs=|X_r&+#= z4@V6BMN0pDc)x?|-ZHYMj`6``ptp#NHytFJ1KY#rNSjwF{>8tvQLW$pyJ6#+1r#;l8{q#c}o6F2){3B-=di7f8@ zkq#gGm0i65VZzS$vDklUV%y#(m223r2c^SGK$;A`$Dcofzyn!Q_(1n~`TT&?16dl?_UNLYReY1VYj z%jEPrW!y*d9Mgz6rCnykJ%;HD`d#2UzyF^d^YPEzBKHUQ<~+^b%ZSg_tuyr1dDXu8 zuu<|HAtKF|vgyc!$1%qTPnZ`7BhF-U=1$|~r6$|c_{l9}c$ca$*bLLCW}$4Ccn=ac z$gl`j#sfrOquF^0iEn?niOXY};};q3K1p^5z=vvb zJ(a;*bf<>am>|9<&IhFRNbtmNMaeU?>|i>_at)Ai@td^~qvt8PMf@>&eE>MbDL%$3 zh~ROtj;R`vGA^o^LPSEDLX2`nScHrYEu}cx*$Nl#laF_*z&7g<$>ec|;X#8B8Hs_S z0ZAD@ON8K2FO;&FM7|za;Cv))XBB(l4Bzfl#gkYCeMEye-YIJ~nr@d2v1Bv@Hs8^{HM`idNB>$to) zqa>jurjV>R1!}0@rDX8Fx@hB$LJ=$n%-OU}Ga`O!q6~z_;iRR%+mW--1glQdi5)C{ zlI-2YV|CN-)F{P)p1M$_E{g-~O~IK{I)v0^5?2u%s?%{IM(7OPsmKOX!AGqql68YQ zC9OT;CxKB74d6D^W2V%mhi^0D8icw`ryStiD|v8re@wME!nTUvWfaneiGK^B6?|uu z(Ve1k>wAsHaH?;Z;F^kCFfq&$W$eLDz!_q@m}XoYst!7I)v@cC&<@pIpErU5UJMC! z9hXv@Um~uq}8fZo7xDA?GXT*doDjrJtHfE#B9JK`lavV5Kex zzTnNsC7`3PO(vf)DQFJ3#MJSq`*6wC8EzR6#dn5ZXcaHB$MT#s6f}>7W|!e)3GNUt zho~#qu%zS*c)1MifQv`Kic-$uG)&|3u3G4!hKq{e+f3dbKt3k;HC`5^%@ZKU#GMth zgIj2cjK@!d7#2D_5se-%8Ii}y+mErj{x|W>HPmW@#kqkHkC>hRBCv}-2z=%j5})|h ztHky)-R>*YvyzYxKZMwm_}0@buQB#ZY&?&R4HDL%yOo`tTn%GvKx9@~_|1dbq!Gwz z2HK)Ip@T!@02k)q&l93z?SiP!)+W~K9!IOY%n!ann7S*fpBn>$E#} zr*q`#MQtF=*0kKFKe~X#kC09RfA2RkzxfZo`L_2wt?dupC%k+N>M*w^50_}nBkvT< zS|UyyA#TXIFdXj_W&s}x#d<_Hh~L&NDprNXAZXeluRF@xK zTZFwMc+0J<<+fhyBr&e{5j`aC-o>1ud5Pu)A#{Z94q<-pZ8Ku7*&<^Vu|s>p{u5v0 z=*C9~<$1={UEr(ly1tXd{u|&^ z-`hnwtF!uRSfBco<0~HxeE#Pq`_EYc@cTc1hmZf-pO8)*=<}3ekGOY}*gT3Em@V&d z|3OAh_VDqLcJCUotD|{g`E}B&W!PLLnkUSTu;S^Do}i=$ICt}5xS91GCx_3Yo7?2= zpYwaabnP$w`+o9s|AN*1XUIGb%>|sbRphV{O)ESfVs_y%B)&IlK`}%I#2k737;uEN zw}}3OAsWgMv66ATziT2cUck+-((K-PB)Ba45$5mA6ehxK?;4t)WP5aj-~EUGC!hT6 z0h{3x&X$Oqp?ME+GrWyt8==|3>C^abK^qh!)Q4d$_1~L)hT|W{#eJOnBF)~H@y+qW zj|{OV#J4{zgq<6p-_JU|z{fxRD!={98(g@94r3j6oSqLtw<4|A&F&iQ_esM7tfA=& zp-tcmx$i3CAXnal!yX2C`A}WOm}5NgsCmLn=;n&Fm1|(8h(L5&#lkEMwCGC6eL;f5 z4S_PDAIikm6{$#uIG>oZrT*jr@O8aW^Peyjsc_}mCblZv<76d=Jy8HWe`FvMd#$j3c!8Ks4 zuIuU!SIPjWrwmm94sEFnW1YlqX|U9wI#7}WTEeVukIIYnFe|v$u;NLnio4DmVWtm7 zZz)n66$+B-iB`+Ro86gEMd~TT43{fvA@1Bo59gF|7awLgG2A?X%UBxG641@O{!MUI zxJ?tD!L)|kh3bxuVMXe9NSmgL@u7s6Da}!>0bR{*Ke1~c zhjNk7&1)mc-Nm~jEN*dmkK$fHhi57IdBoqtha-%EJgzDB0Amj;FY&S8YvGLxNaG$} zx6|q(YL{`LqglQNP6^GHJ_XDk18H%1%2w)iAh-B#Ri`9?lB<DE3DL{c-&6k096ZrmoU3Sn4h3IlKP7j^E9(Hp?hcdvvUQ9Cy#+Juji%Ng|B1zqYUXo{O7;%4gUN0 zK{WtV0N!DJ2D&)sSQ}bNoLt{dBJj6gf9KZ^@bmL^l4$w<(`s+6X8#^(b(PRw;*+0i z`25d(u>0k+I;*oftEs-*0;pdjwnwe z{U!c*KF4o7Kl7_2*&iY07PDKg^7&7l-CX?kKYWEx{Q5w-zd?tqr0qxXVXQaUR40Os z9g<#WSRYq`gFSHP=$7}1-F-~|oU&HN^%JDk<7o1j-ooTM-T7BZLXoK z7dg54MZOnedu5DOuJI>7=XvmIN4slu2Lqq~xx&wW<{HE1-$M?c$M#>QnY|0T0uaAJ z`x&DHFV_e=*FW%br0pJ>P7rzL0Jqcn$$uHJm#+|;Jr?_~eCu<>dsee|$aZ~&)?ea3 z`5)(e{_oeD{r9R*{Q3@g^ARjRh1=}o{VrnFiIV#nA}569>$K5ncGW<8qr5ze7yPCAarlteEf)`?(I zRO=`{6wGDBz_$Yu6XV@|${>V!PZ>Ry72idQIn?e@@(M2w8$dfxTpXc!O@IGc@@7lP zN4U5mA2*m@L*$#xcK;YZfA{sMc}5;*==MphzlPZrVmBu)?&G@+5>gHPvISNalzyKy z>_FepuWI5{4LVw()*&3@+t=_-wft$yPz>k&G@6}(h)PD9OkgB!A|+KsgYzC-Pv|y; z#W9)!Wo$_6bCh(KIQtq__DCoDz+*^^G;;xIDLDMBI^jg1q=Iioa4luqph=NZw*%(U z(bqBiNjPJ5=THrnQCp{@^!AVe%Had{WLA5ra4g>l3NCC2zQFD-Vlztb(cVE*-P36bZ8NK3aqR^N!6vi? z=xFF@xGC(;ia}=Dhx_`6Izx~eo>!_}Lva7_&>E5ju622c|y9zqq0eX9^0W#b@CoZil+K(s>Xp$rva-~tq> zwqP5LVh$HZ1gZ--huQ>ajnGZ3-YhWG5WkX*vhlb8v4OCtO@fDxV0aVLfe~jyw(6YD zBN(o_bPZ@)Vi)nd$8fTwjG2-GVb+5+lbanVq2f2vULGU!4lDEE! zt-}Hxs=)3dv^zsRA5o9JGDvk#s;{VuEW~zX7+Qpcgci#u#KkRCgxp`i(jCmYI)xGq zwgVy?Tyuz)8I(X8&Vxl@Sx5N^)A)R;{Beuq;UZRMSjjZIH;}w$7$3vZ9^x8YT;cp3 zT)0`?yk#Fk#D^0`A8Wv7lPWW~IYQG8q^D8W;`|S@ck@r^o3@o&2f^@at*Ps66Ozx%}B8ZneAYDk$ij~f8k&9TR(HU zjz*wPXMOUYzQ}m;b+r2_%zczRw8X`2#AOgc$CmNr5|%q8-X+d%(Hz{yR7jgkSSkc3 z#P&`_%e+IWdXz$F-V|YDGZN>A%!bFfck(#!$_Ls`z%(iM7Y}8AYt705A?^}qA8bE) zYdH+p93x@3W_fV?JF-k?byjEf*QUOk0f|7Vj3waq3Ir@zs}$N>{s}`Uwrj@y~yA5`ob^V;A6jX56RDA$5(LO0l2oh zyYdL-2rD;n;TmQ2J~}TW001BWNklynaWyMK zUhPuSd2+uYux0+(OUSzf&3$WW5%T6TjIZMJdymH*eC-@8_i)YeBdCNAT1|6Ap6$^d zpMTqTeW&`uFPs2>$R|JdWzxM%B)`BDF&$BPf!PJ2-FxK&i@T>a+qp~LdH$Z(E9XXv<~WFy%(kla(o2Iv2bxOkb_?yK*- z9n{gUFQLO@V7p*@l<^`e&l2V@;Np_*+&A#?;vVOHt`LZUHqcK$XY@_5{(CqNB zx`lBtp3LiZ)`~)Lal$D;{ribOEECgLi%`ZYEH=yPTGfFG?=j&;!{p?Bi;L7vVQM-# zd23Kxbxf9_>Mfq?ki{#pRn$C|_euRR#4TO}W}dums>3(DB~0%0b4)3vE{8g;*nvi9 zmKhf}xR@}VBa(2gCl5=+s`z1mIH6r;28-mL%AgH`ez z$eY;B+2bUDrpj}+ObFG5Dv}Xv$VKqMa66@9EJjCRqJx}jd{P~*f>oSv!^03YdFb?H z;v<6;!$-xYYzrTZaR3{%y4{^2B9+0dmD$ZHQ+Wz{%L#O-NQuy$I-V=SKvNaYwFv3B zL)BM@d@MA9X`KosS8%gE)0kWd4y6@MI=}RDqfNxi*)UJYfurh}Ba>VRSWINqJ3*&WH_ zJ;d$fTts3Z(37pJL7f(>uurTEYk_g|80Z||+@YDj!Eo|CxDglXy^3XzNT$vaKwMNo zPLCgFdvDER{-m_=Im(`0j{ie!}wHEz0Ic zD4Q#M@^k<0p(FESzjBe#ei#=o6Lxn&S~OoEkGu6+lq(ZgQcK*u^Dv}WTn{#&ubSm0JoyuyH99)#&v^^4O)8Oh<1O;{NiuoU1A(|$*HX&k1o-6 zE8KkbFhuwCk|%ec!g5Et^EB|~w_WN#|BqJu*r)$@F83SAKMlhR==LJwx7CHuh>f=> znrRoubO*Pw>yLs0;x;)M>z57UW0HA5_GsGvrF7WLLyx+;P*di_> z;q9o84^l1;NQk84?mN6>XLVL*_1B=jn*#5D%A28JU!ih!|9vV?=3D($&jDNDp8=o# zE-uqqoz+=={L^!`$A61<_XWcI9C2}jc6sfQz@~9Kr>q_$wLgkm?ebHf-QjnC@$ENp zf8FZ$|M5WOdVQ5o{?jFW=okheE_%N3^Q*TAvj2vk{qJ63J3NiK%Y-n)xq$IHy7{-@ zM$B!=MKIqYT9cQ3xkG#YMtuPI2j9Ez`Tk{m;TQiOe)?BejOB!GxH7q#^ORiekg;Rb z4YvI%!}>NZF3H1;IQwAZYU6{;)*Iz)U(6cKW7(B2@OwDn1DmZ2LX`* zawMk#%aM}qQOY%f{3~)!$nYd#wjnOwS=`k;V|eg9A{TM)Ja{8-_Axcu#S$N9_*T*W ziOP76hago9bAFBUYiRGHC8Oyw!WwK#h$G|hcoo1TquG%*=Rx*q7WZ*+#5X6Dd`wz_ z*#cyRZ|>k`J-#`nU$v0tFg}5n0c(yi8gK)ly^iLNoUh`WJz}#V54%(~4sst5t75O= zk?3%f`c`s5k|K;d@L9dAFc98TSe4^=;X=->nHrc*1#KKCg26lu{zd$o*Ew!Cz(>f6_4tC zawum)Cus3#Hk_OWj#Wg$c;aA$&{PpnoFc(P2q;P!s?(KRnaEf|^9;*X#46e#QthOv z4^?1ojX=dhSTIs=lTbCexRp227gXed4%PMSr{7^z3u>v&wKV#PKU;r7fHXKtPeqCN zQn3Z*kZkxy2`Flg(oa@a5qv91oE+9>xR`ORA*STf&_0k-B%09fr5b2i6lWFBgG)8o z(OM`GX^c2GN^+D8r0V##x45L$1RZ?E!#X1_pcF-<(C+kDQpTa7Z1&NztjE38<8VA$ z66cDIinb2WdRZ311qJdHeC99t5kg5?F zE+b9M#8yyIAWhrY6EI5OlDCnPp*&~_Z9SIKwl-HnYqUEF@R&Ms_7#aC4nZr2IZmNx zeqsiz((XLKwFPuVE{-zJ2;BxzLvn-p9rAFF(7uS4jxy9KO7Vr*++|3Y!3Ip%N1@qp6rkje$dqw{q8 zcTwHL@?#kHhlF?^J;^CTg*`1T0zda_(YHvvmayj>ul6!NB}8D|hinr2JRGjupl=^rBxjxaac z<(jy-k4T~$zlyRF-&m#8~#QAreDGLVoB}ZoB?nXKmkF-T52*zwEtf ztS#Ajp7(s;T2-~DbM8HN?75p_HwV*{CChdYBsQD?HX?}>69W_wAOmDmYRd|oe}R8C zko?S#z=4MrCDBok{2&3#5-BkX0z?rKCy_$aaeAQGB)ds=_r168bcQ`t)mrQO@?+J# zJ(At*!IaeZ1sWH<&pG?luBu(9_PgHidCU6a^e6X%Klz4Xgs(iBfPc=v_qi9b^}mai zkD^^eY_Af#8-)E0{mqZ!`60^5ImWz?lq1Zp5aT|o=kf1h@8*7iSSG|nLig&>fV4eI zhu9BxKlg4b7pHSaP2T*Q-#}I>bSS??cJCeC(H-5oCk>Y=sip^jOjv@@Y_PGKqGdG9 zt(Jnt0n1BFR>U^dv@11jD7uAk2aCAY`mK> zU7+b^#O?-`&r~zQC+q+fH4xgGxHE@l?vzw0Ljf1u8mt(m4MCmafbVcuyyi0Tw9jr_ z)69=i-O;R$ZkOnZL%V;Sw0Vit?_*v~Z>Pm`HAafd#5XW3JD5~X?kJQo(6j<5B^Ab# zeIyE2nu=GzkvNV0hf*;UgW@J=sAyR)i@nYUCtn5SnGK z?+=Suj!3L8!g2>b5;S36_m_J;UU4Sm0a7nCYQ(t_nq1BFd5iT*+9+{lQ-RsU-Zdo; zg4)CYE=Cy}-0O~rbHj3jm5#IxQ_@-J=83cwf*880d|N9*GXq3Pecjm=+u_6j*46!+ zw@#eb#B~|DeT=B+UnEW@uQQ=RjAmXD66RjtH^ESKw99A_O5dQ366S`6sbI{TTM=Jk zTd)*}-9WN{mIjGNE{-xFA>kg-(Bmc76sjri4Pi2!Lo}>tEn6F@Bj+<@f@NN?Y@nt6>jICG!VBSUdz2Sx- zHDKd>3+0G@Jj1x#Bg78N1?z!XL2#_hN$WG1Z6K^GcfFjci)i7S^*Wf(;a>3UxNdK! zV{fTh-J}esDdPcfjonKh#PbgAUO?U1T>B{I_Yk`S!u*(~y9&)nz$rv*w5 zzKzC9EDkO)-1rH;*DaKoRMv04PPc#M^hBH^46efA2Mq8}MHNzdQZhofWG)x}*0(dv`xf zyZ1rby@vV4uiqYv`(A1nC#XEb#vVa$Zu#Wnm-*tO-0542z9lOH1WN8f9R1ZJ ztelW{`^5Q%{>J^#o}DBIs?sFhgp#h1(-AiOKBc{ZxMR7e$Tf<;$d^C!!mAc4KKQGDKt67P4`aih zg;_>q4`rW{KY+&bXh`Itr)=)R@=I8{Lc7d$WLpC&Yg~=II|Fh63=`f&a0fR-ypYF^ z5L(QZjN2BmV}t+}$Z3f%BhFrZohsuVloOin3K9*^r%3s(niP}++-I0I)sV1)gbtBN z8BPJ|xMWiTsT;e1`$aRF;DyKj~ zA$X5=hjh!M+qelqFq>7jZfDR}_G|2J;%SC!EtjFRlq~odaoHgZHErk}J}N^W$pd&2 z3WB+ysz_&8u0~THq#B?O(i%;uDN@l|MmCO$jliwawOeb0x3CM{gbawc&gW|GmMMj5 zR*hYa)eH{Z#NkyCsv%c)Qdz5Bkm6`J!Y)kQOzJ{JaI82%66jPg2Vi@pwn7Ni z6vYE&>D8 zX=N)yYzncvWpZjQfGa}j6){J(9(zwU=`31^txdQ9#Zyh}#-*W^w17w^G_}~Uh|v9nsrHkrjfF*(x7oKjj;GbPo1RJ{Hh=Uz+33`!{o8B2nTqM@3r&8+e|-LzsF zijuYwL8kYdQ{Re0>IE6Ba+YOE<_^km9B7tSu?Pf6;r4U7EAEb(A#kao`Q z=4sN&y|@Zq_CXtBcLSO=V_$&%nmTURSdLg$ycp8-#QB!-9h(hfo?&@O9%gtsppU=pu2R@Liza-H)YnXq=Oz zCG8F%WJ1$}ENJ%O&Qck*~2lxXk9-hd>^poi}u+zFnE#lPcXgq02+! z-XX)`Im-4yhV5ybZ&UO^;_Pm^)AzDC_w@I~IJK)Q#N~DJ(Kp#0{U~|dGORBX{W9(& z**-?M|JR7C6{=gx*urK>^F1bqB-X{KVck)NJv=>)OTovFQ8vevY3c_I*(G*2w}fz= z_T2Nd?Hk|oYv`eEnKd)`J_awqZ~vW};J?pL{gaEdySqq!Q1KI8!{8rdad3fd@f_{x z7nzpmPf93+)sNaD%*igna0f?Z*l82MDpj*+=6`xJ8=zhIG89*eOE0 zqgnP?TF~DJKt~y3=ItazfdjCR-t=>xln8+WZC=h7Dig*2v|g zpMUkX|JL6+;jjPt@6%q7l=)vKE*?gkeY`PL_CcD;h|MjQa^-{CK4vr8`Ad{xAC-*t zu^OLoPG~`!H4=6eb6WylTD(L`j+EhE+@B@RujARt;{jf_ps|)2P0Hi4jwj0%;@%}x z6L~vO`WDNF6HE1g&|SkyA(R(sW-n0uJ}fVZ@rW=Vu(VGI4W-nP_s}}3R$bjv$Ek0t zC>oqJH0W44rsHvkrGUf;oSq8ssu8a`RZNwGIs)#MBWh(irQia{J`nK)r9XwnjKo9K zlyTR9IU2TTQ>LVCMOIjDD7%^{G>o)0w1r}ht5*XpE1oJJG&H3euP(RQuT~-0B$>{IaR;kZ^eL_%(TX7x?J%hgOYX)8Y{C@Rcfh_ zR^F`x?z#GDG1r8yJE9XNLzJ9iO(+Z1#7|aV7YZJ6YZ2{n9}sEkHc5TIR935E-b^O3 zBASWwOwpD+LNdWi0M0-$zfI4o)`Gf{!Arr5)s*n8VEwEn1}!7*Xy|H6G6^}=Wr5II zT{LJ4&CJPtB&{o|AhwQ6!3Leiz^h4~2PdU!kV|xA=14Ro2T}r0)S^6hTtRLns(msH zn}W0sS#dYaVr3>LfmqOBNEoJr+mKVJiv*?A~-@TnnPpH7G06l!`F1mJp5DI?t8iTx!W(J;_4l`W8Xd2px67 z!-Tw5v~kb|(NfFq6x0i-;bgodJV(l?P#mWw!{e-AbaL8C9q_a@q`kdWVU*mvc78toN`TSlSZiLfT%SjAvn7l2gPjU^$cQ6y|L$;7dz1>7F*YlkPu%{ z(tAO0^= z{{(~s%Jyyu`)JoPKmB~2l=+^2(4pJE#wbQf_b{vr=nb0Hea!csc~eHeN~S&iU9`Ch z@)64QXK0RpmvuDa-dQwmn9Z+KSdse%-*)6-Ro~O0^1PnLsZkVcj;3>zbdKTbS<3QR=BHoZgAmkF7j~wv-}pC@IKRVX;MLFdiGTDJ z^3lJ^jt@~vAfCR2$_Z{68)oP<$x?=n@#F$|OlWz5_QE&b&^YyR9~(}H{x00ly-j`L zm)89BXD<`R2gthv+SOa%mj32l zehxnSG2;9_;_84-mPm+Tg2gdb4e{_h*O;oEXlk9U%LF=ALz#R%YS#^6aV-! zC4JyFGw-b`t@kOT;QkU{{*QhzL-vhiT%n{JjO+71u)&+w=dpAH@j7<+g-4J0nP2W1 zHuqA}k0J2_#63jk)ey|*DGYVQwA)cOBi7Ggp0IIF=*G#&3D{HIV<@ZHR~2zDzZo*-$8A}b|1u>CE8s^yr-FOh-Z$l zJY#$HgLt0Pgcr$P$m3bG6M}4!rpMEMO^1;y+fqPcLFYTl@d7iaqAu5wonf2)wG?k0~!F2q#+Q+35`;+B5s74VPl6jw=e@r>X-Nynr@DH zi%2!ABy2$n64I2w4y1lT?0PgVNb6-aKxZ4;`DQ9eTGesqc0^Th4@3%TJz^QBrCBDh zj*G_@#8r_u5`3bZ4hN04JA z1Uy$3rs@`AM7_bR7M+bP|j%0NXO z0GQ+AXy=4Dm7+~3i#RWs8x&()$J>iy8Ykm1)kG-iQHiGl{L=S1`g&rxuMJVs;y%heprW5i1$4{IbnPqX?e9!_v+5dAi3 zJ44%#5mwjf_Ft^mzx5eLy9=|HkRikbILd}JEJ)iGHq6O+OOQw&E}*O!*QbH5B54{W zE)#jz;$;p&z&i2(d3TOxyot7lKt@BZ=oo8A+l$~&18{ux&!KXTI6Fs-N96G$7Vd&} zh3kqi8_~AZB16rH7t~hd?H*;!lwBY$U!c2qh19PYj_$$JMa&1pH(2iLe(tUiwXev9 z-FZ%~-dlSD7r{7bZZf!V>gzrVTr z3blJT7*9R`dX}cSkFYo*uHHJMKQ4RX?mFFYkwG3nyeGybO?w&F5!FL9JV5F129{W8 zkO{yo?pP_fZ*dNh_#!^8P~AgUCp7cRbf>Np=ZCNQonk`2K26#@Kv_5JN??BW)%UcX zJMinYi+48p^rc^Xfxq$i|H1a=fS3Eo+mEB#;wlVwKcTxpNee7RJYPrj3bTvP5Slmm z9Aiq}UBJgbqVKlVJq+Gj{f-T z)1PnIY|apJ>sL8x zjoG5R?O7-*yxfhp4YA!IF`&(^rVwopY)(7dk=7$Q z&ymepkb-s@w*j>d)dtH}*>qIH&r1WAaSCaBfb~~EE+g%hbmJ*({2Fid-292h8=CkR zX_g;icIr81XD%_U&oXWov|*sx`)+08-o{p(kd7}>`~qWtH&*7TL^KxS@(PybcnNrR z3f(Iv4hq7CxPM3*AHq_{?&v`zTty{O$~nmA*o6aBk11shcQRyh_(~OncD94G+tVy+tMvhrV;7| z?p-yVi;zc$GNO$WbtI=)O|w#yu&AOjqcP%MNB1d;gemoj)bycgEdm=M51?I5GTSZ2 zt{QHW<2wbCk;*Ez5s1*t>KJ?-eXpEV8zXL3)4Wq0;ZXGaSoXp#)N+@Jhom}>z%+OFNnkG)OC^eSrQxsVpna2A4 zR&iHrgaq0YJPBT$xYr}`q?Q)?ju>ZTGu#}FiwQwdpEvgbb7Qt|7*=`HUXQ)B1&R^| z2MRVCd8kI97pf6Ct88Cw{kB2uC?OO?YEhWRYJ}!3gl39es}r=4`UW2ynO7_TD#V#r zo^OFXR0KxT85)L~yq6trBcA7Y>F_iIGdww-B4HtfMJB|FK`j~c3?;(oNHY?}N!B1) zDPtx!b+H7i>FP29!D>+(U_%6V;;fpxK8+!O{%B;Re`UXn9%4l?iL})U345*GTCuxIKgWVQ&L1JPfYpXQDC|kpGAg#}0 z)*-Y=P(m;WrM}<07AtizNVHy$MQSmYNKHmMRAbs}Ivc5p%%(F+DtI3m*OAce5DA3X zB4G~pQPK(cMsfQZar<>rd%$q*VHOv@&iw2P`1l+-U&O|Qmq=z_6U)|7N{5Yx`&GKV z?_fEAU!dQdL8T*hEn!(u?|_QVD8qE#c|fAz78%zwM9!kK=LyS^yxk{lR+QZ}#OE}t z4#FO>8A)T0jVHu>9*GM=+(%@Ax?>hlPn2v3nKCZXHsZ-9!`wiOW`2xFU_2c0GE&MO zZlxAYuIFTClyRTjpTlg8HqW!V>!~_@)t|=F3dys&5Gr1A=G;pNTPU#Rn3x^$C80eb zgzwF#eXT6b(6~X^N5chX-8bLz&)!nIca3uLK}Z(~vL?5yE*~jnMu-AcB=W;ME4NtJCe@r0I6-Zr)4j3*&nI@_n5tIWNim z5kLOIG5_obenqd=KYjEPpZ@&+!shV9sNPT8o<(#Y=^jLPJAz)rIU&w3u{?X_4U1d> zw%6|`Z;#0NIllaN-}!?5JJFZ_=2brV%TJPyXKeTgOZn#ay&oKo&(a^>&v5t*q_5t= z;oZ^uUjP2*PUGW;8BgwE-tEJzMMWrOWPd#7Q@?x}^RM&8N8fs==Y690(%^j@cpms0 zz?PR2x(56#@P7mUA@z^S=~w*R|Nb6s(;eN>`&S{)KwC7tJua2#fW($Q&ftA!@P6{~ z2hip((4KmPZvXjLB}D5-@l0q2+SLuxdLIvuFrGAg>Yo6=R(Qc5rfbd+W-Fk@#qiuy zh9f@l*-O~?HGb!pez<9accP#A>?5?x57W-?VsZCVQzGb%)vS&Q%cqb_A3^M4`rQqF z@8_O|pL=K9@P*IZ;8Tx(jm(m5_%O3`&(L(Wpu%@cQdv>FMcfHGqG4cr<37^nDye^l zFa6?+?`#7m{oZ4lPyK_Z7}lpDKElkNp?j0Gj`uVAYY$Vl-^6&DFFZE#sk=Xp?)^EM z)xET{|C8~$W94IbT0n?MGfeI9J1oyDvY}DRP{)D8OqtEEV>o3eU=5+kIF6-E835mb z?+`t~eGhy@?60Lv)_7SK3?_YemlV&jJ>!^LT4z2K}c-=ET_9jZ%8Ry-*l3#BCRFcozvWvD4p ziPSQuiNK3EUZi3MT##wu!l90NyJC4pDIhV^%rn;Xl%b)>jN$kkZbyXqma;uyikeb} zIYNVmbuCLfngMf~xm3QVN0froRO2cIEJMj7K?_JErv<7~Q^>_JSFnf$$GV!ly*0M& zMJI#K5v`?Ib-W+E)DeG6wGb@MrVrdcVHb>wcMvCSYD%@6Rc5hy<+}>tS*Pa!rA&rH zSCfpv32iM}s}Nry^_5o}x>8e~7$J@n3%J!(XvHfON)E^X-Q4ivI2nySZUHIPtQG|; zd)2s=04|JuB)H*ieXd%kyO4%xggv zK`C3Mlv)Otwhgjt07EB}+Hrzrd|n7`!g83V4SZrHSHshiqMC{GLK&v9_+S;);BhJ` z3^jR+SENW1Od>7{k_qjIl@@?F8;Q+`6``a#t}O{6c57U6MZ8#?&?X7PXuE>#X3#an zE>0$N#Jxd6UytL?u^dn}G;C?-k)d>yF_FhL_;do6h)1uf;b4^1)KjumL|gF6$Tb%v zk+*4ls=GlOF~0MpyW82wo}qM7@k35qD(Aji4o6XS}1^$ z>vcq~`;N&4f7Oa6NLTH!_f2Ku|DS0_KL?N^e z9xHD{43UgmgQ^l^LS`rUXgFJ%6`$h#jQHv5G6NHhONU08Ijiw;`wSNx)3w~yR>0E-7i`37P3JZ}3~J~u67zCIDF zT~D|F68+Id#^WR6>KMi84=-Zl4O}k0YZ!qq{Ok_+m;BYoo?!ppvt)akHvJ{y>Oq>l zZ!tSlS+;NV+YDiM?LP8m$#D3uaCz!o*`{}@FaG>feB$w%vL}ClxY!Zq*XX(pNWqBM zW{+Weo}6Gj{2G$J$`>E~&JTOj@92(xh!Q^e_=j=-vovS!BepAKc0|-3aJ;NY^0RJg9-Fx~qU;EGA!)>~wJ9_`h z#p;Jtd3(*W?urCs7T#YM&`&-^`~4j-bO-=LeX zh_hotdrWKx%oH0J`DRe}Nt5YhV>3h&((ABGcRxJV~;6+c;O-9XCFakX9=hFXtb#xynVp> z2f;5gTfMwU>qJ7(+OGZv($HhxQ??HxO&~ODAmHW9#GtjBwvB6=#m&kH%*)z{o-QC3 ziR}w`^G8H|8f_A(`zB@h5RyLx8gLE7_9{BNLTEQ=*dWb@yt#{U{b8)!4Q}KiRXl-o z=*$RBO{3M6Y^>Z|Ybv8tD$~(5&#Xpo$pT&zuqD@WDMz4a<=|>hT&$6p z2s#pG5i4^n2X;4BH1n3wPS^w;(6*LWX&eYH^d~cL#R}x@lDzFIch_6QB1pz^LEBW3 z5JRZuNeINC#lyNLI2{ZfTrVQ_i*Wb90*2_kmeW z6DOxct^T+;Dt5cjO~}<$Q-xGVnl(D!OC8fM(5Cnm)u3SF9){O&nuUij~_1bpXwxrgNn+bTS~JK(Gm8 zpr|)fw^zbOfm{t%QJ*jdPAQ6IC74u0+^7B0C}}sN42lLrLqVG0V$MS1EIwoaz6vAsV*Y*Iwcz*)--+IL%T#fy9wIW$pDK$#kEj?=#_6g6$b|K zY2Q@rhANa%$s?3OX_pBXYJ!^xaJdJWOn7OI zP&I@Vlx|A@XLMLFZY~hZ4&*wqdlBmrWv7r{z{>%~y)?6{&;@MVC$yK}u*+L;Xfk3# z-k!nBWxCT(kT(xd<`0pE4X($S3&Z9f+Qp?;6|B8Zrlj>f;9Jx$gC1aEN%25<8)f3| z8Hp{;s1t*HVhx8qn)ciEeGV1pBX4j=Gx6VpRs6|p^%6F$DZ^>f<}SKZ-+h%Kd`oe0 zLfD-q@6M8s_n}D)>x+!LC+UV~e&E;fX8OIyGH?aB!cRVa4c-1YNgu&WM|bM-tIBy5 zVS9L*w7G{o42(CPWQc#j?|$ZwN%~csFY*8Ksb9OnczB(%yBq1wGU$D^AhcwPZz$u7 zr0olo@FZV;^y&|LtMBNJeyIAH&wd=8|0uJ)hX{KwGCTFn*LACOMM#@7^zDbpM+^SN z{T{g>k)8?;kYI}^|<3_ent7MKVcV)_vrq!^V9|I%Zsd4NFh(#b5X%_<2Ep zeIY2+6| z)8X2Xw+EEIP||e>H$Vew$FEvw_^Y3r@s;r}Fk5~QUED{U^~B~FZIA1S^>CK4djyvS zyQ7a1(sPyfD5V-Ux5oBgDFu8+85d~SBhE6-atpe~%Mxsc*#dCf&XU&E3=;Oi56C;m z%1yL6rsR7m-r#bRs74vjVeO1KcV_cLEFRI!t{`z!M@zSNWBCJw<}A%zFf;OYK^a;! zCgukNW?oI9SjWc212o`C5vwUmf>j2sKpCf0pH}v6<$hMPCU%ZS!Ay_@X#gosw@iet zx_CJSghIPFkd7KWHXu&uj?k{)shv!~g0^ifEAzhcbK6YG0VUKltcA(AD>RFOw1!H? zgAtp7H0&W-xw0;W7$J{}4T7~)exubCp*CK>&%krR3z*lDa~q|OPRByoc}1d9C`EuN znVWnn+cIpJIJLQsc2{$kX*A!^ptS@mcpa6_TF+M(G+O1&b|A$l$;mr`J|JNRE;Wrf z5ePynG<%S@&T!PyEOJc*TRr()3L*n;bEI*!IK_8+O<{ zqN??GxM1#-onpC6N!ZH2#;S3wQ{kSAVq-P(Z}FTdAW;zQ>MccHsXF-2urQ}3lg+8ZcZuSaQ77ZpmV$J;KUvlj{NG5O{_lwF5NgB8Vw zNGcyCE*nI8M87{?VK_NY&OIVmnVr7Ljt<>kK>7_5E&+xOchTnuX_wasasBEvuPG(3 z$eVMdpw*Y8663wT)~u>l{!Pk<|Y=`%0=i3@n#(H*_psuQyxquqOm+4(=ZJ&E~RX>hvz z>$J-g*76bTIP-~re8iXj_Pf96^e0g7wZRMU{S1+B0sqCnhHbc`J9_*2$-lo(oP8MN z9*|RXr`r0VYc?{T9P`P~USgE5^Sht<;Wf6t73ty{X*?os&%@~-Y=yX_+WwyN_4gGQ z>eHVO3^(s3#$y*1<2iZC8NOf%b~Ab8s1LxAs*Z0+_BiSwsf^HInnpZG`D-<5*4 zPyfmpvR@?09_}46jAYx;wlDD;|MAs#?M5=mMVjt3?dk^~DZ9m$+sQ@V6@wRe>@nf1 z>obH^PiT$^`aC9($EKFv#2xMOkmll5azAHSKa8wr{NcR~ub(z~XZo$re1o6))hm>n zA7<$8CazBtng%z4vLOx6;rSAC{Ug5c=n*^$Km896h>N=kZ9O+`scDZ%{ThwCYOKsg z2qW%_YN-w2C6c$N$R*>xhTNgEHA;tjCbR=?1@Rl0Y$>}{WwUZbckVLcb9OiHrr-T# z{`#*!!~guBY$`vx{9DXUeU!Ld&>lQPJAchJU<0weg4zLI?j!9QaHUzCkbI=%)$8sy z`kNOh=>Q20qJ1?~#L%z<8S%2bJvD$kmd=wl5TjvbOG*vm197SNW=>FO57wk%56uCU z86g%53)Jo*^*!z4InwSP;5_(UH7VDw5$`Cw1$ZXTPtfidv6eEPrkO{^b*7Ba&3n9? zQ-)?r=>|euZ$8s*UJXd`S^}190#n7&u9l0bJ9$@Gnk1h_{Uea@zQr*D3A#XRoX}2& zQW0`LugR>CtMS|3#Kt`|w8Xh%*%2*x$w[!5a(VUrsr&%hIw7PTckL}&~`rdcL* z45W2S+9;v3i9sn9BT-?Ywdl!i6>NDmtb-BTTur-N$M(y_z6A))NNkK!I!eC9I55n; zveZNopHiOzEH@}M?OIZOega^QG#Rqe?pLEZ^-V2T^H7=7ZK@2@c_N4+s+1(;zM9JI zO3)C{NNs5cO4>t1MMMl6g)$IgMKjz=VCNCa2<@^=a$LaN7o&lw^@C7K<#^TshwU6X5&P^D=iA@Yl?9i(%Jt0G@TsT52LBtzV|4YI3Y zNo7*RRqnddQk+M!Y`Mx+q>bwt5D^U5RlI~!$T&o+P;6F@Gb>ea7b+?PEtP8<17ejq zENY}_i^?cHnLt|CjTsF_m`$k5RlJ`7o$zfYO${Xp)3Wkq)g{asn!xpIEG0x&NE=C; zP#Mq`u^2(en(t?OZgrJ>lGVoR_aIDg7;`+kPoPhdn;4^ZEb z;{m0dqgan?t|NAi6qXeo;^G_PtRYOt7l^BXl^58#^c(=q{E+GBF@%iSKxkf}#3d!) zM?5`4bmvi>sMc&pNew9*=`^5D(Y8bMQLH}z^@x28ix(NEBfMV4@4Sfm_ZFR}%?yhp zr7Q{l79|=q7t!WAt{Exm5_#Msr#a4TK;rgzb3h$lkR-a=DDgW=3;oE%ui$575o4-hpZ(ELlz)rr$0_jkz}>Xm{x?)t@2wxtaPHN;`Vs0^ ze)~~E`wQso0YbBfcE@-(Qh=2AV7fwy593d-@XKF$`lq_Z{*%w$9blg8^%J{V5v&q}4uK5nsJMIe2TR=gi78xJU4z`2&N^D>*$o+<6(d3SfK3nR1r z=h(!Dnd~M{J@un=&Hmw6&d}tK5&ac>cYybE#CgmTP1=y+A;0patCZjSHXFC!dH?_* z07*naRA2wKyX5|w@P*$!V3W?`q$4@QyJ`!V3+x*ZmY*cEe8cjQ( z{)Ccx(sGxQ&Z7=QHWUURLy3sKh&D$w_q|p_V^UyTeUi{WMp^$NpZ=5olh>a86z%*` z{JdfJ^3(4aZer%p;_$O$q*;=0JwRM$ygp5_C7RDyL160)O5>a|oI`Y%*}0b(Zq3QV zKGFBI^9^W5Wl6~abcu#T&|S1#stnRBCp_@JWJCjn%VbA(^*PB zQ!je`hPcjP`T1r~rnvCg$NFAfPi0KBC9n?{UwoO1?qUh_w?A9GETak0i zR!x&9Z>SE~SrwW}G;)Gc3~eXWAWue~JZ0?2Id2_`LTKw4N{JLx(y0fg8kQ4opiU|` z*A&Io5UWwEuoSlK>e&&O;5#8}vkeAoYO1_<;A)T>QgIT#&!7Wo*^oB1RbQM|L5YlN z!itb1l$`Opt^!~vwRM;VqX$#ekC+vV3@%dA0`=7iYqAx)-8LkS0>NmBG10NdYq>%oEdT!p1_RC9YdsyTW%$jx9t_ zERK>gHfEZ=4dRVF&g=6uhEc`laVhvVp^f4i#fVI~RJN_u;I_*fVT$b9p zuhqevPCNvufoG)Zip~O)VnJ}e;s-&ZK4= z*wv6@p(LXW1GzZb#R<)PNiG*4D`{0lGVKy!W`wQ~H?cB~eMLOvB3ov!N1Z2|QKlUX zPaZEKlCj)TVvl%-l}t``-Wb)bYQ;cWEGJYl#r&4ps;J+iWT~46M7$|wXh>}DM}sy- zNrL#Px^}Ivx}OOpf?1(VZ%|U|@m8vvJ0~R#f{-h_HdmLsx=e`$QN@LVbQ#PMH(kZ3 zq=>fD)|v0n#^aJF8;GB(a6TxiQXSme@rbU#BC1D#RB^LdFpv26M>} zHxPQ;iotBrE`U(yU5$11(qn6ZR*5N_{Yiqb%m{8kp(j>P1Ra*Dtyo+x=e!ju?M>TiBI?7m3 zzr=N!GG3wNT{edcW@o1(q9}o9NGchj_Xo#rr6$1EM=qn`~S`iz7#)k>j_}rJT zAYXZ$(w(DQ>|)a5ff64>(nXdpt?>RBF(X-5`Ky``M*QrUc5(QQp(I)|gobV8()+%~ zFMQRpx^bT1&y!tW9aT2qI7a$YY3x3A0y;n?0uCW3*lu2&wrFW+7ANGi zWIEofY{ay#&e(EG zgJLL&M6N+@exl#M#^%;VHm4`_J2yz_9BF-)uv6#sbaD=4pLTJKb|c!ZNaKt$2xHm9 z_Zw6Lutxo@Dptp~GDL?KEYTd?pxL=biVbQRwxV8PhVzP_-$a@XaW!G-0%>>zkpr~J zIEkdPs9{dlBVO>c6&edA`P%aC3Zjub6@2TN)~5+PHHIp1XOPJwnzpwQTo}nNP-3Jz_abTe0BJ2CYgFfqxhIEIU9;vf z8?myalw+*SX}TT!Oekd!O3S!1unrAEn2q=@fjC?PaRuXvN1G~0BNf}BM&(Rvi#EPK zR{`6EitQ*B$B`%5GG+m#x>TnIi%tCwl}bgyc&M&iB~1z{j5$I!#QWN%U`-XP2dT$~ z$yPW9)tcy665D)%g6|9yL6Rckcw^Bopk1}*ra?$4BDlKI;45=EWx>h_G?mXB1Kxv# zskZj}fDQriOPo%`p(AcOTozp3#TY4hL}jAb8i|65fGjKbH(ti<7-;dKN81T6#Z2XC}=R8FQ~ju9jx=tqM&Ms&-gY0o)Qbj5-T@B57F*rN^D81#}Ii4?Usb_JxY3n zY22sZz8#@G!RFt>PR@dU3giOQdX2bP5e>i;8$@bQT#mEKMRpP2j$0H)Ln&=Nb4x-b zA$TlRRx#w-T_PKD=^^eiEu7t6 zJQhs$p{bijCAGxq3T1j2@eNITLRj2F+YMqu91m!O%EvAl%QNEYK`b0#vO#Q3$%kZK zVK#r3FuTQge2&fShieCg6w-JF)f-I1F@E(3-QHDvdx~#QQG^&9GIo-AE9}p13~4H2xX(zzh8TXGcEw z+t(@6nsIf&Y<`4<4L-a?$t}b=g#g82novJcbi~$AP^JU4Jwd}rE&+=T;wD5h&E9i_ z%{8X+JUQ)==^#0ZMQRA?oRI$w|KiENp1Avkf3{%VT*2k@Sec{hF$Ri1M)Fmpfc(KPFKj6_fj-M`6Ve#%-NXOzT0`pPkX_>aH;6Cd-P z)F1rD3;fFOULd-j_4Oy{&U_#L9^BON_63HM%Z$h0MdX`&`HMeayXCPc!L)|++({Ls zX!983N#H;Fa^hcn@!d~`pZ*ii;&=ZZe)cF~_8`(En(hcfb(p5pky6fr{RhNl z1~=f_Wp!ssrsRq7WJNlRmXDLt0IlNiRvji-=gbYLgu&M@$qwPu`;3kUlP(q#`O+< z)*!)H-~I(u7L*dU4$=gyaPAP-9svU-?*deZWP89A&k*G`up5;45P8}K8KKOple9_X zVF%0ImffgW@#OJ5m`B_Z>UylqC~*g@U{(+x>GrQyp>=988S!DDOdV2Mv>hq2!BUI5 z5ld^*coFb8yGBM)9jgOPdqOwiyA8=gb%iB|6SqyOue{VSGx7)}JvFC;+BBn8bSRA$S8m$ zVp7VKD`(Vsnt3Ko4kI9)5QJ@ebRloMhjR%y$)vFkgsppRRIl! zoI~XYyGS0kOk2MVnhH>&)gUQXz*V;~RynmPfe+QGS>jfV_0_@r=I5}%w*%@tzKf(` zU|NHBr(`{Yx{h`yBTmSpV5udIEjc$^hqF+o8o*>IX$>V?4HKiTVODXo_>gctkd0DW zaNB!%5LXZ}Qmp)87mT>6?oC%3n)yf?GkL6IcRqt|YFoI}khjn;iKdLVK1Ggx+k)@N zxvBGgnt>Q4XNVILCj8v+UeU&Z1WKM!CWlReG#)=smFw(B($JCzPZ|SO9Fla)kKVF? z3pHQML+BGFI;M5Nq62DZSlMDY#46%K#Gn{CJ4&pZP|nwYywGF{g|s$`#k!>M`grT)(VyDEV@198l!{)=C_>WB@77-0pDjz-r1rT zGUjTan&GN5ygNn1v<=9cgDvVfTt<0 z1bQE#UE^o9D^gv8sl}puLqu#A*O=BXBjo`^J!$hO>#?Wb|4WcUjyvQuM`K^l7m`pH z>!ylJb(1ezB@=Adw^HrsAo51=U7BX7w$Ky%hV0eO^e-rg#O(32|KT1C-;-5`xu5}`tZ`*&dI~g zN10a7W79Klv>tn&C&yc)P0w)aKH7t;2!^^5+*r>OKU6mCkk|9-9&T@7(>r%fNx!1) zSKHSmkR0k0pmvCVpv_UDD2sy4O(=ArVukhL5 z`-{8pFuPax>R$a!m-q!A_`Xo2gK65 z@^|0i0A!tAcb`Gxt#gQ+3Knw*>SVx9RO1q-Z7+6(ukEvn%cLGUomE9f57URkVd8C zkC2DUgnmiWA3;bpm}wNIlS?2S|zGU8?QY zx&RY=x8<8pjuJPNe2VD_;?@`u=lj~~ZyU-K5SPf~3@bMY?UHGlV-m<|2R6H)9Y{bs zC735?LBbYi;4<2rVz$3+1*Zx|NIh;PfVH?VRo54DbS9Ojy2+%Way_%J0+NO5CKOG$ zzMx*K`^z9ls(5G+WoL>j(4L^&dl|v@0+)`2U2x+Qf+<4TX0YtZ9c;a!L(YQnM5Ng}M zTchC=*G$BfQqpbG2+d-ORoPjeis)8a#@jM!O0Q*-c@p1kn;`} z&mq36pN-TpH+>ENiUY(=hGK-dRAEV}7>xKjdsWaNk@COFvo_RM{?OAr@iWUWWWW73w!l` zxeB;Q6~;;lwRLR`#*4N?%7cv?Q-{p!qDO4fd40VR@fKInK6GzTkozcCJ z#yT#|1w%+h-(bQvU8wN36__<)QvaPE~TDFdGAv%wD;F?I8GL}o-Y?*e+xuZ-y!Z9X7DHqVDCoG<%&&-_mRei&5eDeU1Bj|>4=gW3Vo4SS4-1HNC8hh5^jXLI9YFgp7CzhBeci!xmQMc||Ob{F5@LcAjHOlc7%6OUFU&Z-IL70xtGM+p@I(?b4{x1LO zH{LXk6j|c^53q8MxY;31kI>C-<7TICppc{pap;NbgNg#;c|>NA1+#!!M8dmogewNy zE~q%1gXtx(AMHBeGr!d`p8UHsJCEV}ON8z^e(~%Z1`|aXPIigI1wwy`c)G{0{Pv&U z3rxCK_v+`UBsUbV@7V2mXAck&-|*B^`hkXK{|wd#De(Sxz<>KgKL_9%_1l}h$KM`! z0N&VdynEIDUfrt?vWCO^35&~weoa{1IWFXN28@lDf2tWU|N55_*#FCCf9E!7eT%Sj z5#3zINe4L0BBd-To0mw_%Z&Q1pZWl~A6LcA5wL{CgOn_9am2n;xm1HwWP1TsrWwX9 zTyuxKW67WnSMF|Y9(Nw4o8M2fIDB7m_w@?xhS~1Rti}hKj-CMi&D(DCrj7>-J7V7&@b`x8sKHoGu+I~&b>%lpCeC) zg!j2<7!E&)-}xlX?tOH-Zx@+iAaDZd6>_==JizL9q4(dSxxp#XDLTKS+uv#Z z!>^nnq{ngL5)w4mD`Q6*I>cV0>5qt;bKoAu zw+*4chHGD{pXFQbq@-b=I6jCB11T!9Igf@{&~Ct;?vrV8{V_qFW;#uz;p15O5dL-1 zOBL52)c_)!;oB3e8LGq9Ml6r`{q|HNweunsh zxD6Tx)D28_7MVQg4E!9kS;bi91`9?R=2gflVA+GGzFwDV2p67`+?J#25NB8^r6sN| zBeKCt+6Id?P-ajvqA%jY3SrK4e1R|NEi|h5Z4X3 z_Bw(GHB2n#Bs*(#ENY{b!sDcIB zl(HA8>Ulj&o7E)#Vl2}YBtPLec@ZycDFNO z#i7A!s(eG`vMQ(()GK8a;zlU|_dp20lTjv!mOv)*;Bk$jZNfE)wDFV?#2KMA)R!vk zltesjD5WOgi9lLcab0n^S)p4brs&9P+2R>MomMo16+kJ)SCQLS)PrcL>}3p6AWoH~ z+ssD7e1o=;wAy7_x8y0*0I_WsOiA@UP6Ddny%6>d*HmHDaut?S6l^jOqn%ftv4o6o zAg(0wdEGf4u#!P1L~3{$r3Jr)v`5J= z(CxlViU&BigKH1BqU?65HjlG}v~N9!OS49%Q+!hh?G0A9FGG9<-#w3Xi8vZk{w`Mb z5V!5KDHUOZRhO<%h2B!$Bw`qaY1N=E(#%gNrK6+>?htVs1i|8Yj5)4R#2sVes>nOG zNI%vMFgGBB;(SBN4JGc@@USg1CT$#u!?zySoWs%%IS-WRDA}XUrlJvwfCO^fsSIZC z5Z6PQQ>4M`i0v#WW6dj=mX}cXBH|NyvybX^T=xw9*%h0spI}_y&-&;QiuSZSFC*y; zIX;eW&eQEa51Prtf|5MB+{E{<5&G(Wm2k{(^kL!v`SwM8mnd}DbcNw?iLzoE_bJnc zwE7-4KFuFK`7QX3H$Tr@ZsXH++)hTgN*XURm8geT42NBCA~4i?LpuS~jITZUrWpFUFU=757|rYfW`Bie3)NjVZ9a^43q)VT^bCfl*}KJXc%C#2H2u*V zNzAUswdIn~C-9NZjm`Prk`TdGTGx__;4VNa&xS*}ae1`KNL3=7!#i5Po(W z+!9XjXS^}c-TxQw7!3Bd(y|S%3h#g0J1G-B|K(3$?%yNqK8n{p2pd#3m>UpAEL|Z@ z4--$X@VPI2i?2QT46mdcT;dQnJcv!rJEA5OqdRz&X8$^?!|zf~AFd9Ya)DA>QamA8 z2D?CLPYFBEy(JO*Z9|Qo;pj3YU&ZabLX0z5U%`blsJ~8lx3y3Ues&uPE3!R692|L= zliUTQzg-!CR=;QX?ueq_Cf<$!LqGJRdze%duGN|-1Fa_I>n)S$2ANmJ#JI+wO6fGWm;Z9}e3* z__m9Tn`J_0hSJ1*@XZ2~?N2nCx;MHKjI04-U_$q@%dvVj#O2-;)_+^_%uAOJ~3K~z$C zvB{t*6|11J!kra-mvHWw6g_dhKxKw@136`!pK#4Frfb9wz-83=EuLZ<{#O*&0WK5Q zjxq#N6iO*r&XwuiomMVx62xsC;oHp%(XozWq1)rQio4lyn)sY+54tzEL##2Y~ zHINRg;x2%dD$-k)Z5Wy0yRmYz#~G4i#Xa}|D;=bU7^_>}hf@T_Bw*5FxyLlswLy(C zdd8FOra+hI7q>~1$6~=GLTNx3pst3tiANek@E}T#hVQ4!o-Qqh-{M!wn~w4#_1sk; zji4&zu}6wh#E|5$6v0+lF3_A}Cl^6xNO@L%vm~WUwEHjM zj|(YJ^arm15aV~q%lnb(0M{f+^c5chm0>KohCKe2>*;v1TagrDgYP$tCpAb<~ z@-k)@k|D_Oeuv@ z0=O2@2r{c(R5??kN1VgCf(rw2)q(ES4c>4JmY7a<$mtQrV?oy&U_$i<&7ua-^$+|x z_G)H2+GBk)F>M~f@pr&JjI>7F8~_n)jc}8uzlQ7I+?4j+Df6?>usM8)xZaS~H-LT8 zu#e>hlxw8?DiW`v?KAwrlegf>x4yrx>2>ZOZsMjZ%+G$8hSx~TbELG3O&unR*oZVM z-29N;{hJhwRqRNshY;5xyYFvcN$VZPlLv_Fr?LDUew6Ehe|zb}G|i*<*$z7wpWdDa z-@Us1L!8e@{0M7SeCF%de>yuN@72A!S089;S&>+ihWTA}U;^0CVK$=g(-Hb}QXe#; z;a>rd0e=eY0$+u9B$xixS_9wN-v5~ocue=|UfpG-9VE;Mv%6m;e%2T;fBc&xwH^KW zzuAHOHS3FC3cNJ-sCTFv`PvsZHMHmhsA+f>POmWfN0DYpcn`L#xk#}MYJuVQWlFw@ z=!*3#pFm_E@iVZ1iDD^H#tqZaDPi_1-R@P&xC41f(mQvK|M|HCq&b7O%R36q`{7ay z-yY)nJz_Zzyzs{FBwxYrUBvY_-%(_oR+mxE;=(D~+`>j7P6vec7NK1ehdFWBr3r6} z;_4&)?se9;FEd>~M=Ads(td|8|Jpl?wog5!d{ZCC&mX4SeeT`GT>vc@2{lK%bBkem zfVerw=Jt8oGw;tcj>&^2MDFZD;?sZPIe#N?<2!L1p!U}K?3ex_GW#W(`G*P3i1rb& zzK)TBQkD>I;b#f2A7vO8{G;Dp@Q1(t4SxA6&yuH$tZqKc?DDg3yC{F-B}-!G%xiGw zHF9iOUi$^o&F7f8|BLnNQ;2+yDV`-bdB=b`fIQCG99@KX6AdrXEpD*B_HoRgM_GeT zHBcvcObi#&+fM5)jL6O_l=2{P$fyKtx`H7%ee=D^#}^=E)F0QtrV@b@T>AtEmk~+r&3gN z{khalXfv%C0}(_LY3(p8_(rhoh_R~}OfyP!h`Rg7&T4#W2=T$@PKtfs6>!9q*g>`}7fR1sU@+EeoK9BMPP zH6YYB=};YoIV;)-;wrz>O6~OKhZx``Lz$pifaSn+Dl{&kUI-y!$zdho!XYI`(%6$n z!_tPF3L(8p_TR=c;Ac-$HV4=+Bc(a<^kJ}xQf`9Gu(XenY5Ldb7BAxZ6=m!gjwj-< zOB!}S=9H-e=ds3M1hY(uT3x1%gAx#LXoy&JbuOaYK&sb+tm^9vZX~AaYEEm9qzhQO z2-+g*s^CL|iPaFcEX2tWly)YFf4w!iipUi;@MhQ;RJLn8QOHKmSI zRMdoSMY@~~Gox)4-wGmS8%9;1!ir+Kax`5t0Bg$n0Cf%*966PW90_Y&dkV6P^EWBB z3v|^To)+6VUYV^F#W_RTL>d$5fU&^LZRcaba*4QWetLz@6f9-A#z1BzN@mbNvhG3N$Rk85LfVTM8o ziilulxGrH?wk}zSrHaI|&FQr5N9Edz{fI;5REs963EUyujn)t*N@*!EFs?i#MVdmf zUDU5pi-1tX)pcVsauV`b5XrbO((E7NyEV4jrDy>4Rj2nEBVhRqB{g6k*HQ(p9LXuu zG*flv#s#L;b)4fCZ($8+p`==(1h$On56LJ(pjv=+g%>(t=kP9e#wbQ~{aG8=4HU>m{jFF~Wlc!xu_H{!roRG%_ zIi8^m4VG4zB$}ot%w9#+$Y}?gJfa?TBdT+7;M|%tHi!cs4A)MWW%5|@HfRAGsvtd^ zU@DaIAbGXM+&7q?fBwz$E=GU$yIA=+>G}=g(GJ$FasE{(4OVU-Zi8!1@$JbwI?*I- z2=klU;LF!2e4jFYh;j3$`0{@fuiSo5QS%#5MlOE&S>omb!|{jMz5MiBI-5$}p7^1k zx9ekj#PV)@ukO{o`q?Te*Rbi3VSSFpT?)MG<$3aS2;t`sC;1Nb z<6!VU2mD>Cd+|U0F+Bdgx>rBm%F=*3MBgLD^}R|WPFue4RrlUWu=nb(RZl&o{Mp3^ z$@&1-{CV28tHOp&{PORd;C|;dhV}>ium9N{;t}4Z{`fa<@T*^bhH;o9Hy=aKKZ^_R z{KZ=_)SWP$&XBUfiec?VT)2P)2;C_XPPY!hK%N$q;XG;BXEmN7ry;&uuB^N7pD(pIr#Jd3yyk(NBo zfj0%-+BoQ;Tw+Ymkn;=t(XT)Mu4DV={pZp4Qf+;I8OKbUlYH*?J_YtH<9dzLUDh`rX7}9l@99EL zL&xUEW8~#ulI)xOFaO04_{{%w109|~{4T@QE4Xez+ZC#*wlbH7G%m0RCA~^$Ut)g# z1%~A=W@phRkf({fJj3ehM?pXhiblL&<7X#zhkbj!reVOfN0htdBE^hvULY)9slhOShEsGJNT(B4c8Ku+oKoUhN(z*GKpE$xX+ncX?IWPO zcrS>KxHcnN$g!u4T3gLku&7l_&JD)d%C$7bBx6yl!>%aBjFN=3QB11)7OBonqu|1X zss}Ng^OQl!aUjJ=zk7%@4$TR5ui%?Qp7_-Y*oNt++4)GJNH*8WMBBG@AE@Tip zMw$udl{8$SYz&#LaBhi)6*fJ95kN~7;UD3oofnBe~5EyhCHvVUm2N}f{qJfn4>ZQ3GyjSc{w5XeKq2Rg=kr|d-oL0`R_mLcHJG+S@ zDxlynl7YH{xI*4YbuH%|#-lmX7epFdlku~W&>LBWQf;K$Z% zt~zT;j<{CRM0K-v6G#o+E9P*ND)>qj6RR6G)3lIdVO)C3SfBBS-Eis(dGr)ZRpcBd z(sYJA?cqXgiBG3{hzP1{a=w6XBB&x5uDMN#|37Y5~uPZs#uaqN@Z8YQpK_= zN+fY6L4Y6#5a>pCqlbIXwD(?Xy>C9O(+xDc0W?U#v8DTb!En#8&p!L?bM~{I_xDVI zDY$}2K_u6G6%6SWW2E%5UC9Sq<{Y|Kv}pZqcL7U+ z4j$ju;zO%w>YB86hHI&5;sUlckIlg4+79P^yWQSM9zC*ZD1A-h+IBHS)S+XG_5le- zoGyu7Pw8iMzlI8jH)gMj3^?QGA;Tvn3)+EE`rHf^wOZGr>91@;rGwjg2Qsn|;!7 z14%m^tuNyy-($M>om&gS&HQEh>-!XajJ!NRQ^WY0V+53f$pGWal(N5F1eg-19$XJ@ zsAwsx#(7B{N{sy6AeGbC}&<7jOIqDO!;X%nSBqu<;1-uMpw{b<>J%5W_Wxl;Iqj_wcb0+C$VD^5!H6gl2^c9@U1r3g~NM zyIt4sjjT8UCo^0)VjLAqi4gkR)-`>b2;Eh}e21K%nLQ8w5NNpDu@ZZP(b zq5XY?**8hDz_`FadHUiz-x>JWg>Um`=g%BISR9 z^-D~5(eZKQ>Mjxz<_*bK7n?HVe35+kH4o$f9h^Un%})`!1@bn9&UaJ0 zf04BQFrxQyZ15iISbzC-#=Xkr#YJ@TljP-$Fu%9!_mW!j`XrnFD@+gX?fm+({42;OC}|{@7U?I5%`j20gy(avm$+ql}rfz5wOBq*OrfLwt*GzeDaDoZF#^ z7hr4d-W;8`(&8s8%IL{O(9MaObkr72ik5*e$u#XAMNcC@UN6BK+_cBXLhf6}#f&g< zTV7-xVK)aZHiT|b`M%?XtdXJ)8v>B=6JHG$UoiwiHN(VbR5N8TLL4b6f;XHbERCdf zAWnu_Fm&^cc0HkeiL{zg(hjCgWp@ToNsh9aQF6g_09_&Bb+8^vhlL5o3`Do)bRZAA zl?m#bnqa1EQ|oo~-A@wY6Qy@(35W^!I`VB+kp{j2F{?#a0y@@xls29)X+UdH96G0r zJ#KOx4i6}*2phbQsGh;{619C)m6DFo{0b%fpZwl0e6_xI6F%{oDJGAR^De$u)I1^z zE)k|n;_Q&HbB*<-{fc6k>{RY;97yZ77F0?tv&(%`@dMs*vkR_r6Y8;3dXJeyoFh)E zlWppe03m`|pbQ=xCg2O+uc65x4H6vI*U~xZYckbEUrWeZtqW6I-l$Sjq%H`)Rm8c9 zH~@ldjn#HQdx7GSSdBRmPY6aHgfbd=BlxMSrBqZCvW$`3dweLAp`+vm9~|NpY{WFy z(z56gw~H~NIpDh`NU66I!^u!t!IEkMy$Bi{(tE^WUHI5?YF%)+uHXWg_=>F{RCBw8 ziij;91AMGV6;r{rMw}~UaJXnSb#B#|bWzED=7m%F80{y!4ggs=iRp8Zy&J<(0~ek$YV_*hpz6k-V_@_EhBnFm@Qy^fVl+r5YBo^|8PYf zFejB6FwpD>B6GwZr7S#5p2QFz0^A0Z0m}hojt?W_ z;LzMshPsI2Lc!-i*)Eci3+fJOr!S%TVVo(Ld&6kX;9P?X>pESLJ*soWE3UnY;IXWA zq0D?O9m6AR)4X{AyN<+HXu5%MGhsM7N!-0X!Rn_skQ+aSb4HxKOkQ`}g~7L%r3pD5 zU^>G%joF-Gagw6o?E?StXI|l-{KBhuOrd@9*LToy7YPk+%zW9_ES~sNc3*IyPqHc> zAg4!BgUm@q7iF}#g z`Nbd1+yxl)B-6Ny?+)MfeSN5lj;=ewfB3)5e(12qv5s}Df3x-JrxQ%aNI!cKq5^j;afP4#y! zN#lL^<%e;iwFm^iZD2n{$8V8`jBn1P_A$EVg=*%FGeWySn%lp8 zTu6lZl5ukm;Tk2pI|i({0na1-=s zxAuU$5!@QzUW9T;Vn$w0DdYW!ujA2XrCxx$0de;c{omk|pIP!xfAQ`{uhR)89nxO8 z``;1Xpp-(GUS>GFN*dme#3NiB@M9#WhURSy-f_7{?mP1ED(&eDxBcy>o{6jv|6SVo z6ExHF#Mwof{a0=)2`NJ)txhp6KE|+o2#pWY>^^^+xqDN2eL#QsByfh%MtlUAqorWF zLh=rjDQ+@Sy1;O#lwpDI&oCZ7f{Q(xPEgW$oZqX)OXzV;gY2Hf(%(SKRW2_sBDSW; zH)xtY!t@Gdn4=O2#gI4`Fc+a(UWiDfX-=%7?1*H5=N5^q>?m8Or72u}jcuPFtW9+n(zj&S!>2 zTY1NGz3uDNMMYEHS1v*`H_}GXUJ*c3shjBp9|b>CbO6(Wn-x}hnXpoQ* zmn!O_2x$o9KH_ttBrxgg#Ek7?c2JR`;rb{@KViCfOGQ30{Hk4tH zylD}4fQtiYMi?>I<6WX?i;jYp1|-rM#2Uy6VwK9SBTxqXajjXU|1jAY4G|U-1c`@Tp{5I9Upj0c781-J%^SPi0mM) z!NnFcf$<#956Fqf$wTJveD)>unLp!mzqH}0Pd^CmF`Pez-`E2PD%#OYTNi%j?#+MP!T?MP_8iaD#43Ix+Y*=;H7 zee&X?JoQY==bnD{&ewAf^3q|_;+h}qI$zuXpvO;OBA@*1KDG+j zr4Q2Wevdf2cGr^g;b@=X@ZlP&8otHne(76x@XG&+>b*30Z%Tl_27D6u5#SSV`kMm( z_4eQYd$^04aI9nf>(&p$fJv*9l>Ry@?>rKItaq=co|%$|k1;**A)37#OisRf*W0;} z=}uh0Pp>1dKfwU}+^3iP_AmVq#dl(Sj<56c|L`To8;`KDa|FMGivh9BSOz>(But3$ z6rnp;$4>JHz%LPw7}hJKKO#&nQrs=Qc_25c>5g zbaOx{^O`tOgV5u{A-=szY+fOaGt%lll!D?Aed|77M8P)+K^Yd?jhpCC@}Bh0VjE~O$?F+}3DFA#7;o&L<7FhifR&GIi+2j%N%aCZQ;-fdh@B(%5iN!$hdQ^tcr{Sl z8ViEih~|u10;>s7O@bSR96e%&mVlNmgOthyEn`Id2G<^;*@5h$Lt!+JYX*e2vSn=- zk;(z~u}4CxM&9VLGT{6`+}(i2Y9Q*Qnj1Ef`xPQrkmd>^5md1>!^f*=ZYx_CM;OP}XL;&1OLb5e@w;!+i5cmjXr1FcWwk>)}lzz^*Jit<^smrK{-vfC7 z8y-gDMZ)|FU3-PJ*(JE2B5a-@4}XijBf-T3%-&BK_Zin`5SQ^%Q3ol3^&6bKgyj*s z1=ysbLpD1#SuC#JUaCV-oHTeJf;;cD9isxP&`)!N+PkyJ*|R2q+0i4j-unfk0@CGHAsL=oTws z4kyC6fV8QT6LtDRkWld5P#@bXstNIeYOMTMCy4W??hxD(VqIKBS~OdI-?wg=INh?Q z1+i_a+&SFDU?Zp)riGF`jg?!d6|S)?6rfr6cuWVif z#7X77I_hG;&Fzw!*d!Dr)TF6z6mwEjsA0QUp!Ef)ZmZy|mMO+i$m>uSJ95P-xM+1c zz&u(E=L(t~rdpHJqTmLcUt?ji-7)L>9)|0bWRx_aKk9HkBcY-h@=&?Cw*Ag>QE|Q= zUm_V7dVI4ctrO#F&idLpLc0QvAU%fJ6wQXw6S~WU>2=~wJc=v0`5}IKL|UH0mQOHV zJwupY!dZ*iHf8GuT$2zVAV;tkwLtLKYf&Op@zwwUAOJ~3K~&lJCk(5I+kA-F+y`-r z^F7YjLNZl@>`*@=w1%H92-AyO48=ag@z(iBntDu&6x0mSL{SgE#k(nH4QUwgj1)CI z*C1x% zCkTm1W+NS%_ZgRG(Y%9bL;+96bPp*j#0=kFC-pZNmIrhPuWiq_o;=L3aY`QU$Jj$c zOEbMj(_UsA=2$W^0}{W%?8I{ni~AVk0sZm>Pkp*4&d(CM2k_o-lOvo<)i}>5NW-J( z@DRiDW&E`-GEO^scN%O;LnOHkP7VqB8XtY_0)Kx0X}teyvy)W^JUr$)-(eXjT;(mO4 z1Mm9!^CY#DagRI{^7;kJ_$}JquiSYJ`Bzu(wZXdw{Bz)AZ}Im6{CB{A0{j8+9{|Tq znaBG7S8ry(49gkqJu+adG~~@`%Ia%4{lQ0?j`h}3_d%vRj}mv5Z)xtnS+U)~spqir zBxAqisb^l{b5E~-sBQSIUtI!U1-{D9{6S#nxrUC&$WIdI_Y=BDfgSwbbF`BlEeF6p zlzk|t&^$5@=W*^3b0b==BYcl$@7t82A@%o>j~)W<&EWn0pB?z*XD^}o5jKYhO!n># zX81EmspUsL-XOY+juXb^ z8HUv^&Q0)>Z!$gmb>#alAat^#u^yXVptMggEF+)ztnmfDdE4zo2FR%z%-XP8J%Z;X z_+7m3>R7S~3TUqA2jdFjU&D$cZ$6Bs4bt41-gOAeYv)MIv*g8hVEihd|CQU1sgHm5 ze%$mi-0XnanQy!;ms|4$7iRdlfp`+JIb}1y)t+6}Cm9z{V*V^~+SBY_zvZ;KCSkgV zlQlVdob4jv5v+K+-HS*$&F1>c1pf-sp2N?lbSM9q*eq_lfRestI68;rlSsQm=RZIx z1EvC-hd3JFN7KAWbK-k6lfyS{dtiO#BiQ(p_}P8<*@9;BDxtd!Vx)e=$}Uc-0cYA! zHXb*5oq&-06O79Z-VF?k1JdTpNd7ba+0$9>C`Jq=}txO)-T zY(N8fJiu%Z7YvgPWz|tiEsY9sKyzOWM|a4$oKX5HzV*1KqBz7G!Sa@?TX8#99BPg< z+LmA2f(b@XGd*Hh&oJw@#*3%)4MfFDMy+g(?m!*_qDGt~C^Z2Y+J>|dtnZNI@ZEyY zC88g3ae5pHp|dz$@MuB!jE1nCHtcko@p+*s4bQ87J%nV_(a)!rZp(RjWpzH^j@fm@A^C7MjIL zkMlwv9nhi0g9gO8x?iY|=bCEtd7p7jK{ZzYCkt5}nBvvJmNQ#a$!gff6WU2l=7&HC zYsxT!HsroT$C@Hmtpq3$>H>tXD2|dE;xyyuJ!ObkAGfH2z815&26Td!gcHHVR3C(z z-o0Z=#RN^2H|nES_Gn{BGEyH1Qn3+DOqk?qiUGI;E+OhF<2KjNk6<*@Tnh?sGM+`N zAt}{ZR<*h<=5PjCsi{g7OOBKs*>B&&(K9Ss!q$WK9hkO=#_eMoNKMiT;Kj&EYTy_M z&TvgdOeC!fIj%8$6f9Rmy_S3xVoaC>%IFXky8VRaK*@slLQ$}>ROG9VbwweC`u(Mj zGIpfV=;k-iA7T66aI4VJNe0U}H)72O-`uE(i#$R6h^gT{#Q7!4=78LTYo;_IQ$~T_ z)$^oKO=wf{=x`KN6_pNMH#LRo3uWDrHv>LSaNdDza5jJ~3GGOnuI@M_#2L-*70Nh6 zoA;sP8QlCbxG7=s9VYwFR%A}

WOJC}`E(bi@iSnRU3tI`-Bk7O5-*2^ zqF=v$rp#D+M@6(&ezXpx#+gcQFllW z+Y!lk$2$H568nm@uN-GOgUoR^?L%beKvs^b&49GB>Fg&B-ODRh`M-`1mjs6TQ6`R% zKT)@5F#MxY5abZlI!nMYRCza?*f90ixbdU^cvS@YmJ(`Z6o8h8r(ijK$ z4Ldpbt^;?s_$7DyiwyLoB01xT*Ew89`Uyq5&jvAI)D`MEru&9oVV(S&ZqWK&%2$~g z;d6O;I8VI+TkxVslgn&lC?ZBil$#kd<|G5b;ZW||?3M-@dSuUs3{e^9zo;8VQskkIs$>IG>%jE_uwYfnPX^8|*(Uy>&cUzk z`_08LJPvyeq`=;5vpD(j#ew-dX8Giv*`AZ7w6Z->lOyfAb#s2EBqltz=ihtgrU{jV z3U2vMyc(rlUnm@Y54>pSXWt~1B=5M|P#G1unQT+paq97Zj1>vS44uXi7HEKb?X`x0 zZZrd`^^JdVSxy)u3}`f5&1RJXX_ARXW7FzRF$oD}$Ln^1jYF=Gu_N?ZFNqNi4MM+I z-ILoZMfIh8+NtR?GJFGY$36(<2Vihf7#xs%0*mVxhm?I}q{W;nKlsll$tNPzQ&phG@KF@1Osq&UGiM{EG%un52kvzn`9Tlx2S~ok0YoC1e zzrGnK2(a~{AXseq^x>qL4tpAi*ePCJH|C>T3J$uK$6Pg1IO)3!>JAJ$*4N$fc~NmEJU5r%-!iO>3Frnbi&Zf;rnpHt8G!>fZ*x42f$nF^=C9vj~Ap#Ji>p1G&qQ% z%*8)RZmoL*N!$Z?Y(rqDXFPBCzS7Y?IMN&iNn@L$ip}z8;D#~H{!%`JXM0F)197z} z@4K~z4_PHZ%o%f=so4+*P||_g1IS(Hb!!G}q?K6eLwxQPAV^a`ut8;1@se7T1o3=h z2JKGRYuuuh)?zU~e(qy`5`Zh@S!*jn9;b*i{m9?s$=&z;STxKe<4Rq57C|ukhelQz z$_|8|h|o+1$hzfO2sf_IjEiLPcrRqz#tRH<9l-^I2G1(jQ^)nW9ORA|ooeJ&YeguM zi}?^3@fU&#mPo-3Lh2*|)|=lx<<4=JQ1JNdMW|jS#WvsVdRWB{)?#DQ3o0DsB(LEa@CN@YILd4 z>RsWxN=4N^{eU;q84}2XKX81PgrCF;h3HpTp2A*AIjq!}F*1|8($zGb#LOyT7jkwx zDGK42q2cbomaO{cii?&13qg^K{m`iijnJzEmRdSI-43w!pS{76oLZJ=evYPv+0pXx zniU2G9n}gSk8vwjOA!4Q&Z8PB^zHPy+gk~w$5er{I*E_eO;)q`e)ygT9AL=GZC_KO zQaGE*$c={*t;*y(1_GWiPJvBPS`L5bn*txUyPeC-1BxvCySq5jZjiHm#Twg{4Nn3e z`8;c;A^c$=m5Us1{1_z+-(++ z=9*oCmko=p2Tua&DK}Dog-;(_O?h-V8}18+TrW8(A3d-QcKf?3meb^#o(KL59J?<# zQDa#!`_bAn*Y@G9P@o(6S-nrOD|2bVo|XMxa9X#kD@_a$&##f-nY@D)lyBD|L&zMn}uYd3Ensc&_S|=;o>&9xT z{Py?R(3!!|plLd9EK3a3N+n*dZWc(P345{)9?52XDhXC?X*P2Mj{!CfaL3=Y4XTHY zp)$dV z!=tn0GK*t;`D|#U#aHsyapJk*^F!1;XshIP-=PX z&w~~rj!qr z!jI}4SHGsf!QR5WnZ)|B#KHYzhWvV34(56lO`Q#=@)s01laNE&_z6jCq#8WI`cZ7d zL(iBI;`!)yZgD7*T`UkP7`LOIM3xBow$o^y#U|HnKy!-(_PioKsU#E70p@?F*(uxv z+AyqwpLy+3<^eEOQ7)hrNVehAThwmMm_`RfL9r9PYgV!cT-#p#4ijNHQX8bFp9c0SQw%(^v>$IfBI# zLAmeI*3P1bxc1CDoFgYrqbZy~kPColxs+1Hggqhq!TEWNOHUcRgfdVb<5}&bd(GB$ z8Fb1(fObB}`-{0mJn15Ied3B3AD!-}MT~o>iR1FM4*+kxG!=^w5QYOJgAC1-mwl#N z?!IB4xuW`Hb+w1!jvqb@W{b>9JpqHOukSpLmmW^Cn=$eMs$#uae&Nbld6Z7m z*7E(v@z=a&aEvku&zqHQ9}ejaE_NEI{(!-(-1~b$+kqx$BlR!nUMQ?dHCoFNsr*-{m?vV{Msk!&IAO8C z`K%)NU2%R%$1B5+;rd$t2VIBAy6-=45C7_T4(LO`?Jb(q>oE`&OtKruZfUwZK;p7d zjSL%Ec96qJK7ibcl2;DZ&vxoofoHGvLQ8qg?7%}3m* zI1*hn(ww0IihF#4>+X!J9PnaB&d|QwLjZwZyX@-^R`QEtV(T@M1W!m`M#QxU0q_jh z(=d__37(rN&#`*F@o?wIaI~4O2!a|B)!Ey&#qdvtLUyi%VG*v zu9y?K3vwGy(bTI9S28kbcnUj@0~0^CqtV(o9oh@shuYjlDV4ao0Q!5A?RDAf?Fd5T~~D zsoXpPQihuL|9wh@DgdifR&fFneiu_$Np?-nS<`b9C} z(hXlYpC)?ag>pJD&F3dKa`y9W>csn94_**rf0_`=HIn{{LT}rC?}a@$;7qjHLcx`* z6E7aw^A5XgdV)Nh5Uz);;~|w)))dKLra*JcG0}TJ4abbz4VcbM1Zn;p$!eQX=40Og zZfSh`n;cJPFLN{sD6FyRqJoJrCz&Z#J*%;nRFWAZJ0SPqo(M6 z6S%=&RHzBru!b)P>>Cv(88|>$n;+*V?A3TF{k76DKQ2SrlSbcmjAGLMoqiv%vc*oS zv-|0Z|4UO=^i8{Eo4}6`&6=hn@uL)bD}U8R^0ZzUCj$4EnFxY1h)Kd%ntyDu~&* zTAtf3Xn)%ULVdyR19slFlIZqRLi*}&9Z+Z8k7nH)-N9X!dlnI@KPTfsF@T-Mts>!r zdzD|)ysmN_8DWXIUDl?OkF7i^D#dy{fA_N*9z!jH0M7aA9ls{HUoIDM8ZSNMFvEyR zt>w{iCHLE`fa5H!xbeLU<#Z!cU2N5!uUq%ZLvU}sf*M@aqUg8yNSi}?LYI-a(SaRz zpU8Q}lSaf!g4~Nf_OBg2psmgVzLHG%XErJVqPYB@9&60KtvzjXj8>y9T)tP+B$haF zZI*h@vvf+Mis2GbPwLc(*!W7QF0x<7ue|IxulF(?}K2`j9iU@M`!98B6$EB zu9A3F^Ib3^v6KJ9jF8iOjtXtE6HazxVQt|qjn@&mHK9FMOaX)WNv7#?DfHvLTQATT zK!j*H7<6J8^GW9zdnDphmpIf48p^Xtff3yH{haPdcpEpx=lIkUF-w* zvh=6h+vkHX{wR7jl>zRSrAorwj^-rcoD~UW6%-H`oY?+$AUqtF+bIwb`RSf$-d|Fx zS`_6k*>UCVakZBQ$4Yu5e0TL(qSAXIlPvx$y1-tuoO^pOG0tm*Y`agg@+ae<->7MN zU`bh##-adX1HUfnUscQO`CZayN6B_UYE8AJTvD#rW5u*i2Hvc0f>eC3*?tzT2I{!i zP6>+fb}ywaFXeW8+?sud8$dDo%L!8oM{aYGnYeluvhaT!Tc7{e*mjA~noR>^2*XA> z$Qep?J>rvb0E?=MWOo?{QYhr##H!EL)ITL4g7RU8ZoIYK>1M6NM#q?P(ehvKnPbk& zrOW{UUw*wZy(3=w#O1)6`h7!Ib;gmx@jrr=`NlwAgN1)^>BME>u9ojNUL zqj(t(o=@Y-%N8;(VFY_kpIY&KVF@{2I#)>FawU#484tO=fQ|ziC;xVePn^NvTp-Nd z*3_fP;pVpfm8HSmaME@+VQ5)lV6xbw2H$=eCG^`B>vJT(wJF|HooPR`P-RukR5w9V z9*7C%(;;M*f}Us7AtU9s?;F&`zBnCxT2`B$DXB==)%CC6Y#sRpBl+ zVGtC@PCPt}aJ$7lDst`trF9IBS5HM1$ad5=kkV^0`LAQ>{q~{vjgw#N^wXIKzp+p( zT`8Mf5My_jJg%q&nfaho)QV(wvWto!Ru_*G4N!?Ig7wtFm*$A6wpG{MY4v)zS?>PM zqI6YBvQ-Go&9-|Z+&Co_HhQ}b!%$**ygwr~%W?(u2-A}a33A)fIQUq;t+9{#ulTYl z(9s;Hg)R%k;6Beh1q#;bPhLA2fQKh z<+JloYiPyyekM71?oKj6KCZ8X^G_xQJJvQXV^CTf>ow(FOi}dD6!NfD3nW2|49$Xr z)U4gdm7t+OvEHbN`fPx{za~I_@w7 z4A1Rh`TeQTrQXNI&pdjoIeO0D1LEXXgQOkRW~iEk`jdZ%h1Vvw6JH~f@OI;^r^q!q zm2vGsvVr~pdr+!moHdYo{m@5hVzq* zM`D4Jd4Hjgi%5*vj68hr5!Y~~k!4?M?-T*&5-mt&xqIwlRet+P8nleG6aTU`Kj|d_ zL0{)*!R*bS1UCKjx#Pn>pddyEy`q1+Y+vD`+vdpAGoz+0vTTj|15XiIW)dbf_JWy; z^%Ym6QXg8czKbXSwIJ@ibDJ~e-ItC_ov^y}Ktcsh z&3xs|VyZ}D-{bR?g+>1gh816sW}jYHiVIjplJOm5u%qfn5}POj!$c~3O`SNPX^7{5 zq@3#2bHsJq+Ff;NzgyWSu9Q`*?m1rU#U(7H0LSai=qwQWohD7B(69OXoJJLE(va`9 zKt4O87O9TgCtDO2-lgi=o=~8kkhC<1LizwZhe~YL(htCvSTB)OFXIX*m zA=kk>&cB6^4U%TS9sCR8S-hNY-!x0)X3seJ0|tJ!zt59(t?c`I8~UFQI~emN6K2!B z(ab`5Jld%eaSH3j@ky{k0a$y!NnSI2M6Zvn`E8kZeAA&%e%tLVuV++%pG|>r#HYHd3!sD3YrQY`R~910=PpSc3+h;8gXjNB$u3c z-MI7^4&@86NAbcQPjB#^!tBqD^)&7FZf**2bIl-$HY~R(BF&vpX!|XTU)gn_PV%7R zLm%5`*9*BIsycb7Vwk4!ul>!3axpneu#b%yO&TMMsjrbg-EnWq_iIu#cim!e+wAFp z0+1<~<>K#clGoTLgB3-zslu9WMWjNX$l^t|Q%Ofu$`Y^pACdUJ;51UE~j-hIi9~k&w6wT4*r}H&-EABI94K3BH!M{0onR_ z)NVaYz|OwPy^cm^nV|4v_N0-H@^13VM9)pCIoM*y$0;8UcQZyR84~h-GHBJ4U6P;V zTYJx)02ADhyWT8RKwn#7;Yw!#;W(L)78&kQHRa)`nQ46lL5B}Gza9=aRN4g22L)nr zy!Nj5h9o+1XCzYy%D*)Pn7$4^>W!f59?U(7JImO1RD-kf0M3EOv2gc)Yfm1WblzTA z##Ydeqj{BxX4n7@iTVRLFpp;W8LzkJ&X_-X=JG}rW_>Se? zdyGvNEp!H{r+16d!AM9T>N6MQUx}4C#GTW{#tMI8X%a`+!Z{sSex7CVYbB#k@8O0@ z?~%w8#Ykf)%kvW+eHwt#2{POElEJ+0} zMheL)H-FDrd7=@%CtDno{4x6~1q<3P&TAE{*8|VLH>?j^CkW$}GT$P@6pb0T`KH&h z=W_&OAU;u_$+(9;^O+7_{|3@yKD@W7cq7J~XtFr5)3dmabA;AO zbvF{Qe%$+^fv2hVm82AXQm^;YR2QQzz6I$!Iw+4 zS6duQ0c}I1se7@usZ@uw8fLzzZro{rmR*7iU;4pgGE@wtlA0S0?vrw)@LOjo@xU5o zd$RoyZ4Pteo=d24Ws69!aC&ptAho8z;St&Lus{gJUbu4I?Fe<%3%q-5WTdG;{Io(k znS3X`3Ms$l&A?f#?aKscsL$5%L}$=x#n#E>7gj<|<(w)(R8`SmZaV5D(^j{s>67vl zHC7Jlc*}|_o*S1(xz)3Jn(>(^m1I87w||#TcgnT`_dxTUPkZY(Ynm8d@CRoNHQZ}Q z0>Ogr&^ibIwE=h?uwL257>AZT3h}1oV2(4+!rF)s1s|TDJvEU%i)?2SpB?=8&zox^RyNb0Gy4eSe z^>`Iw-*3SwD?&DyBC$)_vPM@?;8#2eodjrm{+On>=T;*nT7nA6!1L1(Et+G!Fh^G_ z@@r!7G*IWnOwx=viwO?BIs!bo6uPpT;F6bp1=f9MdgHgBDF9bWk}_^tVSVN1uHJEC!8z0g-*lqP(W=2YECEU0Vl$Jz9OP%B!*tJ%_%_;k79dZJl^ntN;c!DfyJJ(Kb;49QN z8Ehp<+w(Z zIFcZ-ckL}Bj2RS)kG7765G`vf;=@~)mM7H2@aK$PCqMcjef}@Pi;MtKzvn};8BI@e z#y^x1#UCl(yF2}bS#+JT#KqWQJ-RIFem8o_&|5=m3h$Dw^3uuYfLAjd$j{Yg#kRBr z$8sfzW(^eD$MU^!=0X5mMhoX3+O&yQbA8na@mHvfGOu{g=P^HCMq~r1&ISeD#G~@#{NJ2){*^5LoN!01tcrzEJa$MM&kT4(5 zG@f&m`3y)#fv^fl597A}t6YFds**Sofz;E2+K@m-JsK4Tu#BBW-@HAfi*K11bZYOF zwB>)z5gT?kkkM3$wM48Cj4Ai>>5(A+^*0YW>jur02ALOHTwL~yHK15S-9`x`@mOZo zxaSDtj&o{_sK(U=ah%`Y+&_l0s=%o)KSM+NZsO@lyiq#Mlyub7{@^e}-wawb?#?Fq z7hXPe=wF-q#nq^-+3j~4zD@CE6W;yn=sTy0j#ik;*Xa2t?FbAiPkI?^)ZeO*bao6T z)kS+vF)aSxmc?9>%J@w+@5`f}j`D{H8n8Ds2(NN=(c>J4< z^?1nkx$hB)8}{j_eSEi4NxaJoKd8&W*H_2A$!nC2CB>p8%c+44eSA(4`7c_$8LtP@ z)V4jBzl8`6br)jG1i~VF7yAfP_W&Ue_g$cGfHt`+W#2HcT#v-H=K`d{U>d(1Zt+Vc zzAJpyLqvVZm!K^C(Cu=#0?qYx_NvF3)tpD;+34WS%iKp}B+iZ@5RfGv`;L!?7~rap zm<$CpZ&mpB6~y84gJQX46`v!ABS5h&3jh$E+B1Y_bqT!fFmo!xMx%FB8Ze^d6XJ6@ z_g_2ia-0ow|yTS~36aU~(l|tRV zOULUdn&zu6s=0qAD{^^nNt`HbXx zoOUGZm%VPSB5?VfDGJ96RtD2=>&;xBntHrYF)1C}o z#uNYc$Fw%SAY{~3EZa)ik^Xcx{ zN%VWu<8^mx52S^VuhQHFF94nLVLx+6`)TtF-}oTbpI=gs;J%`{w>_T^=Nz4*~Z*{ zy5F_zJ`**pY^KMahDPZOA`#_I>?foBPiRJl0)z`IQvV_ z3Yt%QPlY!1R4h45_)EK95dvyiO#QR4nW^W7p<~+5?VZ zcpT}B4;8d#zde*-)kR8^>0H&k(CRabNGw<+UQ%OISDBxim+|+Uw6M*`LG8_0@@H|A zZvLRm_y*P|5@)g#b3NMzG?giTtOpDFC?Y>nk9tJFaO4Up(2Y53?L3O?+HxQ!!H%+? z4R};UdJqj3sGo{JgA*#WP0+5CgFitxN1e0j7H)&1{8l@i<0~x@H8*F(j-BFo6|P-T ziEh1vV%@7F#eXvS5i3qQHG7I-(=dhJ4=(h*U37CI+^>BDN6AwP-H8fU@Z?P;nGuBA|`&giOuw<)1wb)Sndw?l<8)1dA9WmV8_6r zFcQlQ=%#LfO`}X2(XvQ|4>yN`@HBrS4PjpwK0UKv&NmoqF&3S!!?8;3(iP|kGR-2Z zj`5X&HSEe2k0%~}cX#rXr-~ArxTVE;){E)v$0q{NQS0LV&3+vd3rLC!@<$Y6^97BT zZfe73#b)Mg?O&7$GNo|!&KI%SR&wInCuuF*Ie56Fp)a6C+RoUuQwF{r3 zN6Yn=gtXKP$@YcQzg03t(&>&yWLUbo za%tfSb_aVa-%Egs>GGZTW-pWLe%rB-n`~5h&9Z|yAKf&zi)1)eieFsH#s@WO(mw9)`q@l6}onD*DG5?)KK?cuBst>*b=}mu^j5*Qlf1Hq z%CSv)_=0BgjzTJ=ZVfrYo%=rF&rRavYUi_uf8nIli>Jd0o6S7aV9^Xdp8C?0&Mfj4 zo zo|(jQufAP}!Ti_PWJkF~aw}J<5_`qlF3U{Xf(0R-8U~H5FUDOy+2G95{~WMOY1Aq9^S$eb_{99&>8Y`L**cpUFe0<}hc_F8x?5BaTYYj7X~AYn zx!->sJ<=@TC&pd$Tp$4F^iLO(mO^r+ac9cky7bBm-Z8dGNd&TA`lFQNlwTzi;_mR-}Z0x$KtuGPQ5Z(59|9EU{i!v$lTv!E@Ol2_OUc~!-6T64p#`H zMPM*ImsD+-fI%`8XBE(x3zDoLikPp~qA?Rk_JoSr30@Z*hV4s9O#b#79Fm%s-8*f%b+rq?s;-Khf{)=Dgs3>B+k>d>Oj3*k9!H zo1+QRX}C*|hT6UoMWiWrm^=ZseEQJyoUiCx5aZurL%M#efT$I6|2QJ;%_s);s3+-Z z(6A+~O}|Fn7J?Rt$hnTiFX-7>{5MC6g?ERnD#XN@p9Cq%Ji>m}D05-akt%#+VDVpG zv$SY;)G=a&4zjzuM!m%Qo24P{(dpCC;aDq|q#ZTY(Z4!(Al6WWY^xGX%}ic4yNH)z z{HS6}Tt|~Et8L&KlH31rT!0LXqePcyTaZx3mPz)-*4G@o=kLvzwLXi>rNwuWgAaDd z6^HT>qNaZeo%^b2AR@T!(*=!T^re5BKJNV+M&3F_nam9=(1Ys0Cd{Q%i)FmQ3%G#5`(e*;=n@(DE0sb{EcMKqIM2- z#l#~QoA&wt3c5^$rUZE>e(;~@@Jwsf!3T{4EuegQHg@H6$b19u}s zwcF|Kzj=^KTma+CZuF4W2a|`r(Bby_7HiYVz>L=eL*+^?Ds}+7a^z)eo3^KO$DUHy z%=o|>OV-m{8}aBpaOJ*z*~{}#f}5io<=~(ej`)|#RT~1P`BZ(7-1BsYWl+ES69!D` zptcB>`O4l1N-nX3H8y_Vh@MTHwd;GRK%B~J_WGLD&ND__m2NC8Q3>&{-!1k;NWQIg z%s4+8qoPNpl`=0H%9`lfnVZa7ajasn`MC4@&4|4x&IcgnkoY$MAP#{-ixM3jMoU4A zipms4fPzUdTG?5~;Ysfx)y%R_&|oLNzUUb$8hMXSTO1>xXdPmWMnH)*bdnXMdDig5aW09@Hlpe(a zzDI8w$7zNE6F#R!*=LjYc09#rB^GPTA9C|<4k-J&t9mCW3tcw4^~U#x>RfD!DXp9s zoBvvfC#h_3dVC6od379Cs`1qenMte4ne}pDM3&|5{I-7^RKEY6DOQ_dpG~SD?QW(tVwW(dYb=>f(IX--3*1Jj5cjNmE`X! zxjzg#Qm5qOK!nQOQ4YmB{7d&kyr~Spn=!eG52g$WV?QDg6F?Lm#S0EOJG2?a&}8!R zsVWc(ZQocttx}`lwt|MyHcp8hC1q&3Fc6acjfwDhIl|eWfFgDRM zeaU=R$yco1Si)G2S0b=Pn`6eK`Gc5g?#U$+2dh-y&|^&2av?>$C#u8VR!Vr7?vGJl zIA(li{8~mO&0r=vO&CM*xfKA508<>L2uGEz-OE~+mZM>Rxrr08bfiIOVMZzpr}(ab zrB%z9JhO5l=Dy&9y*WMXSvMN^a`Yb~N}rbceBNZ1I3Gt9ic)c{cPNpoiCE5bw4Fik zCWC{{yYQR#&GK_hCl((C`@b;}Z7*-i7ye$qF6S7_x5dFL8uI=8r4af?UKJokR4SJs zzkp9g&5^y3SpHpivYj&L8^x2Q$PZKu$iC=pdSDO){v(4B8P0ACrBzy;QFMC6l-|OBCi>SM-B;SS?7x3h z5JGJp^D$P-RR1e_4jJKtW<`LBR-#31%vJrLgz7c9H8iVDl5nFC2gg#g_tIj*_tPa! zC0&;u^NdVridgRLUi~a#+K8lzIM)uKh8WLC*YV;J6{7!4h_=9|Av>^N7qg%CNyMQ1 z?ra+LDS)%5VvHUt;{LPv`2i;mk+|>>ImHF*VF4`v|MR-YmF1b_o(%rI z4jbY`{=KA{U8Qj?O8)aN0kH( zA49z%ExsR!tSy7UJbJGKzHcB$;?28@koAJTV9Bn<^9g33@H=Rx`U=ka^J2&F#iPG! z;q$@%<0QP_actz2%Cl}C>AQOrr!3`vl$YgBKHYAe+t#d*TN?Qk2em_n$) z)sx6){Dfcb!S2hwT;S)k_W|^%TAKEqt;-iXSGH&N#>GsRH{}OXa|oC_Td6ZcTv8My z#5HzRiz$dF(A!Ab*YxByZ*;40 zO8c1D7wpYxl$pE2g3YMHL$4sKNQEvX*nJeqyk23yT|`Zj zwn_xG>7v5_Ffe5HDy8svziYVxGb=4+?!5sq2pzpkQh!x|BCAYOP<{Jv?rqDVgaQ!q z<7bPco=^ukC@UQx*|Ty!Hen)v`5SvoEN*sr|IwBs4$~OP(66t?IW0gc4avl=qnXiK zVJWV8scl%D+OrO3gny?0&Dj211jP5J+a+{r>osTrVziUzd0YQ$@4Zt_CMFY>SC+r{ zg+Hu_Az74M{<}iAx^G~Qi8rY9Px^ml+RFsfM;*xH@xEec{y$h(bg%E9Vn`wn-lbgJ zH$9E#3s(6!ZImfY8_`#F@eCa3dNYloW%FyQTvw(uypzh{E?C0xGZ$j05EshMa1AT=D z;-sa5Lotz~N;55M%C6_fI$EOs?HiYEUx}Y9>br*f%{rI!CcF2MU=tSQ^G3ogTSLGT~DmxZ=Eb{tj7uMeHjm83M|c)Z>lo-)Sq zO^!II$$Oi+qf3fE*z@daP4Dx<>oOlt>^FRkil4{g|JFkE+M9lFNZ!%T`WBwg`Aw{b z7v2!qNwX&CysPYmDJMqO2R!Z$#Ukb|Mez?FB~?7y7DJ5L@8&w#!Bk-pWU&NM%u0OS zt;wN54kA6d+8?D9+H@R6n1Ws*#jw)PpKE&gp9sb>>3H5!NQzdAnAmql@ILfa9?V83 zMm$z+d5fJTF(Z0zL(cydXe{E zikP?;6}b6Ui+cN?it^upQu=PKocWk^&rMiL6>af)=gP%H5_ke#G911kW)xso8|2@- zeKmjXd}w5TB<*PK@frgWrEv9>u%gbcNs9I@fclqU7FmFVR$BM^UmDX5e<|L@G1Ji$ zTO$pX4CI+OF}kWKv<>dP068wARa%}$eD%ry#kOXoqDh99_Gu@tacrl2$NN;7!KRBQ z$V+HLJ0BTVNv2nyh!5;T9+-CHD1shz{GpiybGToso`0ee@$8|_&6zYaWH8)FFaYuGgLRj<2j)=hHD zcNp~a5mD_dZN{hMrhq9QS7p>3odt?xayrFKAS&E*jb1j12 z(<2pu^m}evtv*HqpKc3pp;rgZ6ib(pK?a>)Fz)Dc-^RvoG<#hNa&zJMy>By)Kv-_u zOFe0>@B@{UuJT9>zPhNBk=KlfDwWBwshF{Z&a5=0{#H(fi!dkUPh*@c_pZ8sCaL8A zA?hF$IP68oCqgePrip=pFeOO0z1gZpR24d-x4%!HQTEMJmK{y_#P)?@c+B&0Ip|I8 zKzz(!ba**K=L!pNiO%xIiZUz>vu0n3Ko;nb$RKNX7qSnIqsbW(K(EbXNr=u7nc-v=EtUW8>O-e70q!jUSAnB}{>ss_Gyp`H0_tQFW zV`m4@@{Y4s{o^3uhS4$J$vqSe7>_S7?+Sbp{#R-?N<(T>p0sV-(f#&rYe;CmPQh-` zH+9N`BPWMwcyZ9H#?UTuEeQECS8sTKb5z_Lz~p6I(B7u-d%_rC1!Bq;H5E}SqSC1d zpH9o$mcJU~Hg1fkTKh$|6~-2H@R;N%Lc#SwpZ-n2Mrei7Vxt7LlG}Kb@s{QNrJ`y) z2_1U1pQ)SJNi!!s_HcIjz$7Xw>mRg*4q%^XyA^Vl+5L`qtE)Qykm}`oD@<@p| z1ZUVmAzPqtLRszlkB0>kimZJ<4>=d|VxQ768wzv3k{^``~WdwN6~?zgRk&)FQ>K$R6t@%5wLCddf#KoF9Hy zR*@MAb4$uhgxPO9drXUZhQWA^18;lBzhK()Ic0aUC5Q?S#Eh|5M{Y`A7 zHsFOry>Fq)XR>32>66C5L|#^a*1DjE5}BtYJ@43#ePNdpp_wbX`MSVmG}FYJao!?5 z26?3{5_R6%8XxkWa_&Pc%!ppJo|F5<_ zdwm;We~@!=iElfHp#kW{*Gfc6eO(0a%B_9bpXwuMc1MdcI^s_R#a;Rz)ndEBxZ4ts zU7K9p@q1=(@#`-5>;#exJvwq){cHO(yoaY*=2TB6`Z#plMwh&C|P&>Pm* z^Ri6}$3LS7QmA})!E>Nc#hy>u$w-QwI39J6pJnmL{XK2;VNAbs-BQi9Q_FRv_fEZ7 zwlpnuD`#U%3Bx0?$x@ZxLv(6&jj^#DFh1gw>{*hCKT8xNM5!V zW?PraAf(K?gb!ZtqA|iJPMV1w3|^SU!=vnNN5z)ZMKhqMaNC}RMMeZJ`CN8xt`6i| zNSb+SQ+$LZzwD5W{O(@4u$J)l>aaPBT4Bi~X(eFJa$~yNY1@UTkv?8QN~Cg9+m`T; zh6lwQ*2BKX5O7rCkf9_QjX9$$$FtTQOVF#};j>!{Su<`M4Tm3{zUks^@bhaAS`=;J28ydjg2$gR!42tw{JQTyLh@m& zy~vO%Rk?B|Ej;!{s#YhCU#G_MHKUpGY+vZSAWPPA{d9p1Y=s{daTBJ46V>&$XN6s7 zPnVh`6wysVgW24woBjNsI6o6eEYj34sIn4aXU}J>sJ9dtfa|GIkGVnLh{%Y`NIZsD z9lp@tjt;XF697)ht1Do>%0NQ{?5{*$QUhrdkl0$4aIORYlk{xL(TK#8rOLSWHX4J1 zy8UQk?$2Q>8@U_@SJ4qiY8j=pj3jA#)=Rqql|6(docr!{p|^5VXhKISu0#qmvf zBT%4%Uysn%h3R{WYaGxXGt^ho=2{7BP+d(@xW_I>8N__*ojgrnrF|oYFR=IRy>JBWV|$SG~;l>&%D*Nj)3SA3aUiq zS~SOKpRDZnB_@?Pj3L(fx-cGeWvnl~>y$dYLB)rJ-m2g5KBkw-#8={I0n(Ib^`*1k z0^Al)sZ?2H6K@HAB=35fpF8o*;4{t#SEpO5hLX_nnO+_}D63QZS^Y?)VnXXfUtZJ1 zkYFV&uHa1^>a_p3dlVa^6I(><7*z@ZE4RZ-OiKBGAHE}F=&%RKYe{LQ$!KD{?n=9i zmSg{-pKG7I{WNO*5vf?!fbj{(rDbi4-x5;1R@}~q3gqyt2r%Xl;>>$N+}c*fMA>tu z##{a%Ti?ZnJ@P!>$(@N_>z}G|_(!+yG+45y=}761{XLWpg@NEmx{}JRA|5N0=rec8$i|XPylsqIQSZ}R_oMXgiY0vCC+SPlW5y9?pOb4`422Dc zkwp9{=kRk|&Ey2F57(W`3X;LzXZ~ygVJ!(lzr63K>yO{bCfW{3g}L4{N4%sK=#hs% zBO?+V&`=CpQp@XE#M4MA*qrVCh9OE59sv6u4L|oq!Pk=nhI1kvRvZ2vFC6Gxy;HPY zt!^AIx53*(SGL9BPhXwF{y7T~AhDKu$@(>XQxCDun?hb)Uf1AJesVOY%S=J-cQQKU z`pTWlXa^J2Z6CK4A%|4-Y`lEbsE^tdOxXq^d22o$E&0@sf%*#W0A@ro*Qsc8O&NEh z>6>86?56}79{B6|AbXz_9c)^L57^ix5&UXG6C{GD=#(@#YyYdDFJ=e-S3ol<#~zs9 zetMvu^0@AdxE~nKU+$}4^LsoH--@1MomC@f{z^g9*Zmm{(BlvV z!(`9@+IOsJg~h7Z0T?|RtW3FHY+nx0N8&H=TKL?s z4k&c%9>IIeC3XV1w9 zc>PJpBvV1ed#sf60HeMp+b9nXR_nbEXTbyj(s=?7|0ZjFq33yk?WU z>&@r6AbQQ9GezR}mB-Z>1w_YJ)i5Z)_%Q8LPaX8}uP<~qWXz5078&9&gou5)odXw0 z<$CeYii_o)ONIH4hU7&eM{J=KJKu~Pk4#G)Qe*;JY5Dk7fhn#>N#%gY9T}^slCq-OSve|=bO%sPR;;%`qNBou&Yo3$ zv?)(cgkY=+%7kKuq!@h2%XcZN#j0<0h?4tr`a5V$g)XTb{}cCy68U9}ZM_`GH)xnCGZG5MmeV*Avk`kr8m}xpDy@d5D`GmpA+A<1a z12Az6V2ZC3-*=|iFBTNpz#$?k+P}9ifkR*CqGawS<~_&o)qXw1l{IGvG>zWfzU;-# z?kRY-UAS@S2@PIA+z5RGoFeJS3K9drbj)#C%fI~APmlH^85trrB;pn!$|2$B(0`nA z)&>rbN3O9xl;MbmzBCEQU6c{0usJ6kR`XNraG!I%e6yZ4WK^}i)~;o>PFy%&w?erz zTV0=(P6Z9sGg4iH7S!zYN7z;e*~O-V+WSCQ+V4mWpzhh}L~T{+%8lrD4f{c`&ycOg zAZ--FNBzV&={+KP68c~itRgiAlnkeF_08{PZeq)|WyfenU_LH2*$dNE_a75cBJEw0bOO;hNA~*Fl8rk z%mb0+V>}OtV}je5ZxIMZtG#$X(!v+gmR@_~B-Kr5B&)Bc;ntT*;EcccSc?ODckFZxbq z_k;R6gkDpam%jX>XT57pV?j!fNm26u>ms6*Yb!4w+x~fGNoDu`4{>|WoBcmp z#o+*i0!ey^2$VQ`Y~S1;I|{-5UfvSwCGyqtF{xsrWhM##ej@z&9dAA`P zDenGxkIc@BATHdVPhe7*=4(fL_k=Qvr)UhA(CE$9s&gcnLJaBH&ADosAd6U-N)_hd z{m}fnBmCcXm-)?6th|^rNu<)7KFim%62{$D=n<>iVfAaXez>0kfZv3O(oL%=yF1;z zivK++i6qI|n2pUz5)iK&JN_swWW#wPPd5VS0SSn{X34*yzgr6<` z5EcW|4`r^Y&|h>Ufsx=_74HDLj9Her2zy&Q%x6xqe`u8B#`Hw|l=kNF@a7W|Lc!wU zX%FmKnzOwhacd|?p*ycR9Nt6e!j0iRNAszvm=FFI=X7hzIop1 zz*t{YKlnAGzNIw)j3KEn*^&zNRy#xssL9sN`A~O4g@MTZ)B*fy_Rn!QCc=j7MIXQ!dg4E0*oHT9+AtEO(7 zi&VuhH}jCeY@Qq~zcT$5u>HSbrJ;`utVuDrHn7Rb%gyJ)eavuS75)%Nf8vosr=u!yk}*| zgW9gQ75We`>Q^d~Cf@$s%wcoslh=D&i%VTh%}tN=|mHpZa(`oYGX=-GEQ zs0!sgs>J{jQ7?0}+|ELl_w(o7)a9}*1eHFMOAIA{fQMa7^kWE6_}=2FAjc;qmYVc% z>paa;uQOY#m9F$Z1|C4Xx=U(0;44piq&m|$LkZiP>xWn~KG3mJ6v3tW zh8>>#S(aCXoH;F$vp$$GUDG&BuHb=@iFt$01$cXIqh!YrH}U%5!S<<`Kb^?z4M;j` zM^884$g8hVT}GC$oce*Ccx%$V`j{|Rz zILmWthrcE9>zy$ADB2MFFV-Kq7TsHpItWEhRCwiFQl>@x3^KB{<%qCm3QOsS61>db zVspMd>25!-bA(m|4KE;%dv`l<$sng?Qhw)1tN|bi<>wt%ubEwa{b_K ztN8gGJrr9StX@s-e;7QQM@5XXaGHyb6`^VT6;eec0lu$@9#7`Lhg~~&tG!7bR-!9! zX1fe3(Ht;DrZll>eql?B7WiBezkU<(;mW_A>w+HSG**3kBHQ-=rs0?Q|Et;Tvmgs8 z6T%s!XOc-cB6RZ3YO(>k#@6?)8tx2(xL`;M^FAJo9cf)?&|ZEky3urAlUPQ~`SS`x zKg|`aPz`0auK2Q0(Ne*Y9FFkI0vJr1PyWI*!ewT7WDx!h7kOB5<8$+@5{Eyw+Tw)N z`h%NkTNiOMbUuYZ@v1!D{lzk0ac^5{R^BMh_0nbYhe;sDJp-qzO-?I(7Z84y=46jO z7RRi7a$Dsj>K+ne>=ebe-7KCTE}Yl@M07JT_4V9u5M=A7^@O(nD-<+cYvS+@ykRL2 z&SL!iMnQrr@bpLFxWEEzuszA}koCeN+`xsETbF2VaIQw*Rx0qCkzWppqw%rx=Hpl2 zNK9r6n`$@-X5)4XIWqUn5RM-vd zgZ9|yn*@q7RST{2C=WQjmiU&J8g@$YI!2gDz=tU?tfSai?!F$vaZ-xNbNIj??@w9z)KKJ||1II|F+$ATgZhbEdncFS zTQXrSr4lAC5{Zi;Xj}Ck-M;&#W$(yo6p)(dPj@?rOn^81?a9+{NJSRFS%RRNB1y}l zI4ziGVZB9CRW|-zikaJUG~{NBKXls&0fNcuw_>R47=S+$=Qz{a90{>U?ngGM)G>K%(dA zA~uK~sO!}PF2}LovHkHS{fAHAlDba86g|60?x&kF*Mz*)(Fw}PD=_RTcV2XP^>YuL zs$$#gBazMa#s(%6Q3>U_6EFYQ`1tJjaJp|%N+V2uUuTWRrW>q%BiR*~Yb&@xL1vp= z@73d)BWqziqC)hHRC?1H>BJ#Is8Pr9ECN9`CAr7O|eD!tr$ zMJKxVDKhBCKWpGs@>VG%_;WAQh+WE`lIICzqqy<2Vg-}Eua4(+y+JOwp!i#Nx=#t6 zzX@hMyC_+GhC2VSV7amHH#2q}EAtT8FtiWacQ6LVDeL|F%cAlDAgH1{s?54sGvpYl zBl5~qr{R$EXRx(E1Sf6zKxLIK4vqEM<<=lIwc^NP8bD7k>+O?!?}KO0dr@Ad6XIx4 zhTRvd!J*N)RdiZrJVZkPCaaDv{j=2dTB%CIVfV%3>b{0#i2QbbLIRy~I>{WT>YQ#% zWfe^Q(0nwRcQ*D-P2*pTbf=21-I0=!qz$}3(d^vj?Yjzf$_}dL<%tNSQ4;9m$iqSr z--FLlus$mj&0y#Ju_nav4n?M32;uf|^`lLR4guBh^Ey)`(>{ECGT0{xWmM&sX26@ms)98Cb~p z{JG^$P-ispexO4K{7Q*Q=o^+rYnLrZO|jC_#DsLKBG22XDnYa8EK_5um0Y)VK6sdK zK=|qQG&jj?LO9Z8b{nf-5p$Fn3#Tky)wAvHd@Rw6MNwkCCOqbk(1r_f>4&j+^XZq-r9p!KM5` z%BEDa*>g>}VN2u!?Ls4|Gkd}|ptAT4QZ5JMn-W8cK=C+6VBo2Bs`D=>`xDSYt!++)~J0NV6uBbjj5&uq5mP{8rLxdtJF(ipcgV1Qfs(+`xZ!@m)j}Nib$l z_~%V>%gfaGiz#htVJ|@Fw-VM==zl8SJ|e^I(u3YMK*Ys&RzH`Z zHI$b8wFCl*8F{dcg#ZHIarJ8|j+kOcSqPIaGQ(DNuOh{#w7>by`MEgd#UgbGRJzWG zRq#MPAAa^fy6NN^KnGW6HJFKHL|cP>I5Mj$lNE>bo{_W`l^brn{ zS^~ctx-uG9XoEkijHM*V=48&_0N#UgisI`IUJEK4vj;iMMIMtoO>7#TJHic5Db+=- zi|K{1;hY)^8Sm3`=<9QtbIs;-uTsgWP}AnIYf$ZGz4_-?#NP2yQ@lIGdgaZuG|e}O z#XeO`78U4Ag# zypfJcl{ITCX23{uH7K<_@rniyLc_$Np=T4lecc!D;v)xUdH=;Kmzj3?zgb;xVOq%Y zCl(4ppMCBB;{ph;1vlckQvRs1%oL3ZP}lAA!N_em7yf#2cbLJvb==Vof52aGJpd<- zPcIj3{7xWIBX0%h5Sb;pspgg|I=z-9BEWQt?_(@s5>}AZZ0pX(4K)%X4_ynnFP9-6 z%}9o#EkieJD)x`lFUio0W+T56_NU+LgUG4om0~hc0X#TEN`n)Jdb}jp6cpqZ-IA4F zpNj*)7khN$~tJ|s`hUc-FVp9^J z4@=>YLD%{I1JV_L=kg*Ytq%4nSx*Mc%<)Ces|`zM{;!`JNt8ajx}Yrkmz(Iw1G&O| zq{o3AVQa_l!1L=@a2ROUcB;1ejQ%W;cz?w+g{BN#9>7AVqe@>ojLz~iv6fY5IK6aW zH)|T%$7RU_%Z>ZJ^Lc5Vs@7(WRA5-CEz|~Z%-T2(Z#bO;nAKErr6=rEXF!D_=dH+9 zvpr14NkEj7JcV(VVJLVi$f!QZb!nbVJi={IdBA8+r}Q4{WG9+Pwt>8Yg;iud3!v!T z?j}Oc?#?L{9^H0KJ#v^xmPS(0alYBmFvMb&e<@5mMm(nClrX4fIW=Hb0K=pZ!!%Ki>!xvC`eUj!AkrXEp^p*r*9yqb@u;M^)Q!`(T@ z=k^v`>jc0haI1RDSMTDwcDsnCj!ZI;7dMV^Y#Ok$M9G&6MVy6B zZ7PFo_O$uGGrtfHzu+G>VVL+EsQnVK^JGC91aHqz$Ll;lKUKW}CYeRIWEl|e(+2Ph zTAOVMfpcEur(R~AKbbt=)!^a_Nic_`5hRN9v0ybP>Hh?`d{ABMatu0QfvW{n6gWk1 zs(xyp`2`Ml!&wtr56zHzfJi+M=s%0C-{Gy_K?81gg?h#;o_8#YuPP!$!k5zm%GjTH z+4vIIpI1QeYifM1chHH{@;9i`YD3nUUT(QwH{37UJb%81pMRwoQ)?M^h4sGNeAI9z z!T)egQdy)tNgO>Psx=q1Tj5kta^1@yhRTqs?&I^yUOvEm=~*)006;6q zbAHCyFj-eN{^%fUD+yG4dhR$QPam^^6l8T5J5wd8^jqiiu= z1h!kr(RVT35Ph?_q=~+6mGY^>&!F3J``O}x+Un_au@It!jDA+{)NCzmxZ*p>{Vg&R!@g&fP z0q&BBkZSW~Edg+?<$LiTR~m~YH_c{jM``SHzelFXBaTwA;xdjB?!4fh$-0StOZ6_- zu53!ujK#8ifYc=UFu<2aDhku}=%0aC^vQ~h7yLs*6X5_U)7d5*Krs`lmaXzSji#RK zkz=RaIH|8Hps-AO^Q?$@c1c^FITk$t?%e#4C{B^g1`A-ZGD>C+J(0Xy`(8MsCFz9Ntd<<%&y>M<1ZkFT-4k-_?{;sLXzYiyL~l;LyvPpB8G7iv zUz?PFEwSd$!fKctd(kp47DX!Tz`Y-T?iiCE)$Btc@tzpar|Tcm?j1)_L6St-_{3E5 zNzs&C8-p-*`t}MB7Ld+PbCu=vzODYZ;a6%8S`SYiP$>mspM?Ic+o@BCn5A4yZ&0Kk zNFPu9puEMjAgwK;a{xxl^mbY#)3a2SM}s=?^9&YbjJzyvYAC#?%#A-mcUvXc=$+_> z=|OWGfi@~&AfPq=m4+>d^OLA);ty3eqW61^3dz9v zV8L&<70A^gr-U#GxM{8D?N*tyO)Ht|1iY9zj-eG_I~#VIXk0|hAgNO{53`O2*VQf& zL1|@eA;LQ$r`1lmsfB14uU*DBLl*|n9XwQsqmzKEwbn+Mib1m0{3(Zf?N%irX4j4! z5R%$SA8RW0LBG}6OPY*J)KkJ~9!{JjoqN&`$Hkw+ZL8z~MZG>bMN_=!v( zABMmb?fcFP%aNsSn{|F$V;;&fs^qwto&{Q0W&Lfa`o)nBG?X7sdwLppXgdM-;LS=^3lWnA&-(?>9pYKSoaIfW+)g@_o92ZH zzBkKK=4EthrhJe2J3 zoraB;wwnGq-aV%T&zg91lA8&xdK{)!r2%p_Lt#2btzq1D2Fr~A#J zLx7O$crZ82dc3<{c-==>fNFAtt}uX<=F&XjA z_cX=X~R#goD~YNJ9alYNz6>-LmYFS;5>< zCQ9In&xKB6)}BYINfi!YtVD6SgG>_}1fZUVk;+0D`cq>4xi57d?K)*EDWJSF(C7Dg z`_5cOtoU3(Rp!|-L?W+>o-dJWEhLz}kRWPEcn8r!cKO$~l3pOUmUg_#z9gn(|G{a# z%5O!gCbrSxceXmJj(`3)33_K}uCm{*7pV33~YgetG0-|-6s5BAY^ zMpp=rf@Dh{v;*1^Rxr=iuEFVWcWTKK$)^Q8MfbV}Ru})0< zZGHnXLz?MVwhF2lZYOFmx3O>bRL-K2W~hdwC2vHrLxeT7of-4-*KzJ>?16Zbu%%K z@RWfmWr8wO25MDH_p!n)i(@B10wt(zz*47ir+v@xIC<}vz%*itAaFGHElYN3E~5y8 z6Pl`;C>f6;_c&1}`{FlF%(`Z|pH%)J+&qUu@-->Q#w?RRN^~qwi>6IG_XZVK-hcLy=t8Vyi=^%n zL+9T*IcJk%V_IxFB^G%Cl zr{_~io{Q3K3=CZ28klxaZF4`g8VK}ul6>aOTx`di$#18cZ4bnS!pR^iO(Kww+MLMnx3RgTu#s{+^~7x(kI^p#gL^r)Q2>=z zKkhzrQZ5CU*9{%~WHWoMuWA)Y8QE?6H-wfYFEzwx@^Yr|@IF!u>2baErD(|%KY5`F zF}Npn{oX=VCD$<+-o_Hwe1%^gs8qRmTfDnYa}wyOqc?}=#_%&NPB8b5g;5}p5)q|5 z)PCR?G9s+T{q0T81~r@UuxNveu|7O-b2$4j>D+RLfD89s-%|6oFRUojnn&fFurcVD zxK_HOMsB}%ZUegsR275&mfQ194SuIz=Z@<-;>l;^d<_w$(I$UyUjB>f)U;7NME|V7 z9+4L2mY;@mIe@TdxmO&L8*=(*xLQ(^0D*1)7(2%{`JmQ9mP?YzrV;<=#Nkv}odHd$` zdDfjW&jg9hR$xan3*XL;iC%tOn2S?;*0Rzz6W`r1!$3S0oO8?6$0kcR0Nu*PJ|o6p zC)+krXv(-*Y$_DQ2gP}+EvliMU#;T-RmKB>9OopWb>usSNF3}$xwOPNq#gPXsiPrs zBccM)2z#IfYPm4VN{5UoL6u^~rz*wb=274h(NeTt^wywc{z;Y_PN;v(_g(-6k326f zeXw8$NFLk$H!-v&tNS#_-l^4xf43=Mfe!7J3Jm>rqg_nahBNF`1eTe9A{wL0@?j*L zxMXpGP)=T1CjM+~!Iq>j1^MPUm@SaAiu^>8DHVQ&ocU=40o6B#LUolXG?E?Jp&KfE zBv28l4N9Vt-jB%!fT9u)j2PwGja~VAwvb#Xr~=JxK__HNI3aUcoBYcgprrpT9fo8L zx$eE9i|v1uDD+WyKza61k2n_)kFAj`EaadW>Bnf_X3ZlwNUlO8XC0jDuY8T9QGI4S@W?L*WIIz6f^Hb^qPn1ze4#Wm_LAgZz_ssn*3%UG)s>}WKEHmI-TKSH9)AEU7BGC?r9#BC~jj4I6_7wG~1YAxROgE zd49NQl#C(tb()^$My1HoiaSM{+l!$D;De3B79{l`1A9n>)dq5IHGH@CZ%EwDUR7MghXqK}P8(LOcO0X~e@ z@!@k)lm=rwFTFsb#+;s-S`h%Ml>n?jCnG8m&fd*GGNv*8iyX!-rY-JK{RAyvC$IB? z!=%-aRw1Zum3z&p4J*;)*mp6RF$4+o)=_98nY&Alt&2jaw2LPuwx0)N&&y2EUf>Jv zh=?bI&GX%{fSWQ!ijRE|w9k}>BE=J`BgkoI6BMwj=P%#@BPFEfTIUXZo=c*kP%e_)~mK{aUO-Rp3OV5*= zdC<<=?(3#sPb>dx4eiLsb(l~?&#yBn5y?WSYS3$mj`uj~dLn@E^2+U9yIIe3o3lTX znk}RhLLWqFYS_#>MU*fX1+v=HRt&=*Q(RWBl8MP#zaX>xf^|!T^HYRPL}X+BI$pb_ zT>&}T_=#<{T}^n@&=wp-6uT_scNrJgX75J%DW02kK;12J3BaIs5qRj}ef)dx&VAwB z#?TLQEUc}I0i8j*=O<1>wFizI`6m)3rZdf){~t|f9o2;6y?;fJkPhkIfFa%89Ric? z4(T2tyy!c`2Dfi%n7{d>;E+Yi|(iiDt|9tq%Bs|pXRrcO@A-K z#|iGJRGn1ge8JbY6v7xk%Ni}`qUaT2P0yz3;*N=FnrOp(EIr$v>)Bnsn(J(%zMwnm z%SU_h_1j7ozzWdXBvIQ@eL}~B`PNACfXuS5b2CSkyF0=amR*=sv=R24w0#%ev6~Yp zQN>E&twv_Xc z9#z&T!>q28MyvJq@7@7OSuqeIVsH3OTe+n{nfp-Ifnd+C;o&e;nZZx-0(610SWxhK zu`g^sDi#GW04?pa&MBew)~FRUkB~GQF`lhD6}_m6mvs@96&qI#TmW{kW6I+&aAPqf z_?EbQ>KYjDlFXSNqfYCD&R^4H!_ihFI!1~Q=|8(O-UXz#zA~>5#B_BINxQ~B)^j!S zTvYmm^A*2O0loI#R`}9E|2*LbHS&)b@3@wNEx0Ow7|_g1TjmP6{uWE=somYH1=-re zFDLI$sxXC}eeBheR$=LFwd-$kar@XF1vmSXc_*|55zZEPo0v_^DgsH)X4Js&;GF7d zH5U3B)M25y`|cMLzlvSOSFGn43UnIf*3K|SEul0G(y|mxrr_}urvlL2MZ5_HY#61e z3S~pP^wrb9diQw*{adH0!B0BGv6Z8Z(g!634kNbBHRDMu@!^VR~wfpJWSv-3=jgKU%G)5n5Wne}M`27+% zJmjoD>JG6}Cyz|}pll~?3cBQSs1-JKVeOH@?=F&eN2qRxX46rxvMUWhD&i4gHTe`7*gGK?@cQ`^Uij z#!-0`(7QtU0s-cS=!Ny}LQ@}{)^6u|$qIWD*&m|wuQ=B#EZ;Nyey#7%s%xq9_~ZW#2mh-8 z)NNEiv^)jue|1cwx<-`rW17P;=5E=0ccyG@Dst_rAgOkDwidl8%tNS{lj|%G%5oh^ zIn_yvRuv2#M$h-}LJYh==%wd>f>a&-+dX?38czj`S3(C?y|}w}`1&pChK%ft`eX4m zM@=ap5ILrZkLAIr?>o_XY;OI|l=SehTrpAF;pm`fEC_yqn3I&(OYV}KL#A$weNiO{=iA1a5X|p9cMRE#}4IM=(0ZSEun(n zaI=_#^L_SLqxIe!pP1b;Qz_ia3Dm$C8idn~gk9WjA6e7hf(^Bm4vSjThN-+bE|rMf zHjwkC6JpvyyLw6A&03W%BugckRBj(p(oNDlL48-py=u1Pv~?F6c5~6I9ZIr!A-%Pg zWwjCiIbOe>V zS?diG?v4=G>rB^qYQ6Vw17CxL{m-bKGOGfM8Ul{jVg8#_?_r6r#1$W zK4R*3#WUR9-Q)SMW6uV9cT15!y`F5Q8dH8V0McpU^2`(atWa3wueK-sc|tIb2eQ`ehy60d z@zklQ<<1cl6DY#~3a9w`AwJoE$%{YdqoKrkZm?p5MU};p{i!a1n6ASe4Pm?~H~=!> z<~r7v?fZ&0>Q%fbQyBlG^ke+ff73ze7P3U!9n>2?S3rfSS5*-dNr|g{CC%Q7a4S@sXG1%D4{Vc@F?POu_D9%AGf}t*P<`fmmT5=DCn2GI z{vl3{1=Fk1sldq^z>En#tocJP|K6NDnnNx7nJ4xiH;cJ$HC<{2^=KSx+KM9_4q{L| z`B^Z4Bnk*g^Lgnmq|DO%+J{@5&h?jw0Dr;p3)XcWOxyp|9i2(l*9q*FnIZAsU2Irs zdg;}PLe$CkNaYIs9rZ{8Xlva3HB^uPYb^6bF-?lzP;1pV>8g~26_g4WH4#k=?CzkKB%~ypJ7Eu;MO~NjMkFPt8^o5LKCxJo<_SV@Xtu5eGr49BHp6*(3tA;Pj5M z$)*K(g;dovIBpCLyfP-CqKM>>R&T@>)X=b{`;NDW#nVC2f2~HHPAWBYS4>3!&nnFB z8;`S{_=t-n%!@&kp_A^w0Nc6rrvL8+z$>yY&|ayZxo9kJw3aINa)M^jC@+W?AX}Px zZB~b{|27VFr!HSd>$kddVNSL0#%Hb0hUl!LY0%81;wx?Szs4)n#!0o4hi=9i3G{Q1 zAwZxE;nH(=0c6ke{BKHgv$qL)%^+HjR0`2ttOO{X7O6_^GTJ#1NR(LW%!#wk<)oK&{a(qAg;kpRzFROi88Pe*Q%YGEh$Yj0lOETcswDFjyTE*TqWB{xlB# z4FQ)od#jZ`a>P*8pq;1AYOJnc)(JB0GA!-=yfkoICO~7e$WENmo3Fk8Xxz4dOo1ty z`FUeD8lCojL*VUZw=0&4`mZV{w_rSnpe^SwpZ_f_@rtdehTT^lnY+c=&Ii7gXwysf z(Nx-WCsIOVRkS~ximNem2;I~3Mu+VZAX~Yf>uO{;=xj8E%zTy>7r*P4M#?@6 z$(}cOemJU;HRH*BxL|y`U}u=hdHh!)lOC#>xl_}7hkpFbJ5$$a;$)T5ft^E`GbWd# z?7ZyrFJ}WqnLm$g1EsTc zUQuCfwVQNGiXvP(Qi_)tJS#SF!{9Fzr8#sRG?`Mfb$U&$dKEc}ds=GN$i4Sd%KUt^ z-1m>eQ1GB?0uBZk)D@U_IZ`MNYhl(aAeIQbU%@pgTV569Y#Pt#(zJB+^j&PQ76Rp^ z#quu>J7?24TI7IlZ!~gP%=J@|%5L%5XsZ)90=q9-=nWM`;+>eCc;-*3TQ64_QN}_e z{N=YB27N8?#1>%J>a6K*?B0yQjqXZ)6yEmbK3*WBFb=e|hsCU9cwaEKZ{|avS0Gow zcFdVFmna!Wjhg10-0SKokRD8-|eRV1Y+c1B%IvvxhcoDsiuSE*56&YDmX zWj==u?~h96MM~OdR{xacK2)#iw>Kbg*_>Y}F(6By<15~+$lWham-P5XkmNEt(zyK? zJhSm+J!DiOyX}?P;^O+gZd}#wlE+5Ez`2solm!bt)`}aK-qgPtE2fR*y0>>cYv39x z*X`M}->_+L%AjpiUlyV77ab?=dV&m@i08GN$T;d#jpky7Ah748s?$>ESA9QJ1zopQ zw`FAc@U@7Z{MR8S!SOv8^a~jj$vCYPGU4x63B-$dD^T>`CTo@+;^}gZ`RQ)3f9M~Y z(59UvP|&SMRql=0`Wvex@q0L(&F)uKc}oP0^D^p+kX2f*k}Ags9KeJhqaR-kT!8)y z8|(SyQqMAp`6({4J6;Y)J6SOp{8y}5qV8o4fJdd`hja0E4`DVlB+egt0K63V$}r@W zYH8+J(QRWON%&fk#uL0lo?m`I+k^D(e#es07?|y^ zIB)Ecj4#h&8Ke*MlqNDFNbA^CR#<;IxoBg-bj5H46p*m}0g;>=1vSwek#(C2=-UxoIt8%MUtmoB7@5xF%6{c7Y_R9P5tlIGSZ+ zo*=&zn4qzvBOSmjyAg9bER@djo3=O9UEc<>%%`lfWU}a>!z;iRq~XOt{CUb(BunvcNG~s_ zkUZXcj0ZzrGMdlz zaLVSKxp5I0G%@P!fixL&tRKnzLD$h-oeYjmLvKge|!>IJozKkN?A5|Bj$Rlh) zbdhIpE*e%<#($p1G@I#8ztWMo@0pzc`Yx_yY5gcx)yWlcBi)(msH050>mtFVahY}* z9O(ZP&EM-v$D;&@)E|Nax_=5yL7d1+>v_{&+aZ3DatmY6)6q)&LbMbm-)<7R<7dQM z7tykfoQE&nZ%JIs4)X0z&Cs_Tvedn8&6|Pz`I~(-RTBQ;DAK~f$1*^w$C)QiSW212 zggh~1X-q2oij4@J%e*XR*;DwW0EH0TR5FGS9RDR)6(To@(4|ix(do~95UU(zCUMTP z%_OhMDxK)>ureu96U$4LU9`K^Go4m}Jl6vztH?^jgonp(7*8Omt{uO#Ws$ID{A3bV zVAf9<7gr#Wk0sj6OPqQPJBR{ayA*^ufBS%TC54PrMP*+)4U9g1++p8b3(Njrz38AH z<=?+bi%T6Gn^->yknJ1`+08M!EgZZntbRPc+b`YMdEC*FxiChrBA7pj6GG7csBK5F z;)GsC2`pG4DbsNIaLCnTC9)fYczRq&)(iXk@2cc%0Onw%YAQMUnK@(8Lxh1FSi?N$Rm_d_3g+X>L8KTcNd3iIE2dPGn+NgL~#Obof(OHjAG2Nrtp z#iH$CEIHGodcLjy&^91e(@9fL#?DyZ!-nqADqt-|5nOyWj$+p{=Sb-J8GVL_}l`Y!R|pR%DR zeZC℞x@FY%5XbpMjR~9=Gnz1|E`kBt6gW;6c{>`In`qDWyGBmN^7Zm!s5mJ-hAL z==d#4kFBq}w(1~odC6o@=$4^E32=uU_;hJF0pgRHf%KvIdLax=0p|MYd#sJ$1flaq zBe(xhY5d?$X@~WU!trwdZ^36PqRtAoj<{D*ruQk zf%*HM!oAWbC;x)U8Y*xSWYp_4=I~m%_@g2mrC#X#N3HYaKT7LZYN==Q1(Z=`pS1*@ zGXpBgx3*@xtOnP+tJlB%a&7j6M8A=$NGHZneQ3}k zu#TL~SwE4ZxCIueC#~QFHA(}Ty|}vAZ0bb@_%XAV<)x|`GwOEl2HR0n(Bw-k)zK@z z)@=o5lwT!|;(3S=Adb~QpIVE`>YG8Vze4jcU&RuGoX+`{ZTwQTf^Z@}`{+Znwl{5P zZRvu~pXVElb1|A$JZA=S^4^|TIT*TH;_}t_5IWk88l-tMP|`U~shOmSmuSn(4!wG= zP}UwSBHpB_>?y?);Fi39F*zqW@tIRkxvbVVl~4frp>JoEUE$f1aoMzInCpnF1 zVlhTyflOPvaZ3y~CDMokE>0}JDq`?f*6>I?K_&|?HAQ6AQ5W*}@>+TuLbA1f(`)c+ z*LY7-+X>;go{6$F1JPErMG^r?1F5nllQR6ROA%}s;Dr#Zc?s@32|S*ntw5ldlextuOGvLlOTvuywuiM`TMzn zy%r}HUb26bYU_1p$Lj*B;Kg2O*|Q8%whA-Xv@aFY2MyLsSfk>{J=_-Aa#fC zHwLy#rhhwB(>dA|UcZY*WPtb@1kuM=jxf6g<5q@}M!GB0`R@1&AEw7dx{T30NA2JA z{w831ENxF_mC!_I;CK@05gk}*xIs9!T$N3}>U{u$DO%}G%Af&jB0x@w%>Yq9XYZ;h zx9L0bh>Q-2ySJyIpyLdIGzj}7QNxpj)(2u(CA6n?ChD8}D<`aov1peVXJxv?%`t0d|37d$w3W4o4 zU@OKYq>)gvY}|QG^yWgb*{^0WyB+f#Jw+Vt@TZDV@QYV_reOXW&Czll@&OP$kfG@o zffCj+gc{I(PD++~7oB%NGjsoD=#a|?<~_{eqGj(1_6g?O4MugjNt_lRlWarliX9ey zA8Xe-*-iI|YH55!b^)e@juy}DTXtOo@gA>?bZ&E&bZ!R+gGOs}`>rP$3XvD*edVvj z@0=%Xq>W{VaWdj>66$|I^qPee4!a!fNF~|D=%0?i@w2lkT8Q9O#>SvKnO0uQPTU#} zA$Nb)pEoIhygSz+L!HKlxXt)YB~PY7ZC5 zB&+`;Sshk=+QUCdqds5cH`6KADf*;FCs(qkwcIqZ{3rWA zr4B1;N~43rnh&nq)h2e5`(ejV|GJ$j=iGHBC{6y|u>jU}fX3v`sk)5ef# zi)dqaxR{OSW#9G|#IQw_6LRZE(6M1#$??ZSR>Mm(bb|3ZroRT7A zlnSg7op&yytL^jWR}u0Q(z;AKaqf-tPXjbpm}&xwGq5^yPA>P4ydU!|Xx#s``u%rl zn>52)O-o5a08I%`hn5P~*cC`8RL%vCY_ktHpQWBb)p3uSTCV4co!$M(idCi+&(mr7 z3eG8rTA@NoIW3t*j$bfeK4!UEN=kp67W8p;U75h#x-O#iCgo5d zb53iNHBk!XxRfqPzFcagt8)}d)YK8$+#Kq);_c)2Wvyx52E5;?Y`^_}-n&DX#LyUL%^1xKNzz6>DDQ=9i~&uPUPN$L?n`dY`#_f_(3|Zi$|& zhLEk&hoZ{4&n(=$clyF4oJ9VHUVnS++Fsg_+G=_%rDrs^^ntab!+&tUgZVhg@3Rsm z?^!1w?t+Hb|B+yHnKtGbWn4|U%80q5ySd{~M5xF<0v>(%`L9-O7jdMXfi4Agl7AF} z69LtP9>spv(1!}r7zh>Ht>D|G6S&}1WM&e_E*kCA2UR(HNN91-mRVcML)9ObvM7j} zxaYm!l93Nytjd7kar>{v@qP)|$xz!_Sg)OU>G=e=%~Z9wxktSN_P5x76=Zlbib`zk z6qDqbib2oBDalYOtj7*z6CD0!XM3z+rQz++;a(gi`#CWZY@;YD619rgF~F2Xs^oym0L!|3sF7z0mB8Nin|Z=5KRJUr#!1PihK0 zGR!o}Skbdnm|)3K3i_cyz#`ErMt&9DM0kt{D?ptTgBiOw6;tW#9L^j$M5eB0i$A`J z1F6XMWAkY_jMtkBpbsI-Tcogrnu|*J2Rb)67XjGP)JizXI7nHV^7lwBvZD`H3KxH_ zlz#W)?oY`>!xyW#-VkRx%@)@!Ob7MuY zNa#X~2R_lj84Llz1D3qwt-^dlt-l%g2bofRGj{kGy>uET@*hk%bWD`kv^e{-wfQ%Lj8Gff4Qob<{N!B z2k$8k46WEsiIxftLq^UX=5Rw$jL72O=yLE+4OgVNQZRy;)uA!paAn{0cmJ@{h3|~q z>k+=L`lX|pe{a`1p3%|KJ^bjnL7xcThn~1vg={%Fm^^+t85|yzZsiK8`XMqZ^)BFl z-&V5iHn&>^FX9!+*^7|tL`lJKX4PZ^eohb6X6a=d9Esbn{8=9A zw-%&F!vC$xj7oU)j9>K0X{a|YY&MxX;*dBOyrjZFEH+fOKi_B7@uW#rJvXx|xY#R} zH@N#%vhBHH;@x7*KPj_ee(Id*EfDEhDr(N_%aW#13J$ceV{3ikH6@EtQtEo+ zuCpOaJY5P!wXe1lSDH~q{df=xBE`Wylu6{-p~RS%$~md!!&k@8GwUaPCNr>kx|MCt zW{B{Hon(N)$YAFV_qapJU*@{b!WBQO)iTrpxEU-wFVQNov2e0R3X$SPCh3gAj$3y9 zy<7uDmH@jEuH7{3$dog0fjWuJYd|FG=99115k5CO=1Z?mpquHCHY>5l{0clyXmtEr zaMXovc7aq@^yZ2OO=vU%2qxt_SRw{(D}HRpD3z)@GC8ALD#VwkTb;I0IhHQ=cb0Cm z{SbWpQ$lAgph%0NJMs5FFP=qXm6)C4F;N`QU7D4Q{MAJN!GraObhRDR3=q8WrZ2Et z5eL4qI^RV*;PaG@K9*NAKFixN`-2cBB2=0AF~twRtKB^t2uwIJTQ$7c89jgWSGAMl z)bXP~Arzn{sd+HJ|L9`j)sLLb|5EE(Ge&)XkF}pfP>d@i0!Gt0Gi=O1o?ur7Yo%v? z@q6w{dUOI^aQx;u*gqBw+YLS8|6soq+%|cRxVrTj85w2KN3?NXz+uK`=tJ0Anq$`B zeUrc!&ohiz%{!xT@^Qo{^gw##)UkM!K(vmw89%%65|h)MDwHJV$9Tj9FcyB5Lp}>& z!62ar3mo{rQ*e?0{XCx(!QRujZ!Kd>FQsE7IdFlXu7*Wk71GD$Fqsa`;lneF4t0OWi015I5`O!)rFZgGLq>&-xoJ@8tzFi!kB#e9tuC-Wm@rp*Z z=Mt35?PeVoN{+$~CZvB$FwFz$HSi{-Sl?x4vmj`S+U+q_D$CCB-+}ImJDt!MQzN67 zIz%(-3Yo;eM`7mfvo&MOYICXPIDbZkiB6B-y)H;=4CLxa?MT|mXREN0I@9`+$1U(O zdrk~{7P|mF?*Q~spN>YbrEsZP!OofS&dlEd6*(mM;59Ngop^L1(;&dqvZb2g zRkRlmoKMKUJ1uKuVT%k$9%=Yd_}`7EYDGaAo))xfta2j5{b<%Jsg?QxN?lY zT*4%sU!MTqGN&tXAfKPeyQ@}Uu*2u_5c$C=(5TTD2q7|sU-a$0U%xz`Aei@_KxblF z%4f>zI{fe>w3CMEjs6=#d$$F6UlFoV>@R6PURY>))5aev(nKQ3qc?8{AMZ26gCX#V z$M`yn9WStNkConmjXeACct?9^*Fn$|no%?YwF>e7qd3%5V##gdCh><2sC^}sVThyi zcJr5G)M~i*n@0Yj+xE4E8Ezk$0VG&mwc!CqbK2`f`?&RwKu08lw-^BuIPh>CJ{#-r zC*@{Ys(b0O_;BMHGT;-QCJ8Ir4IV@7rZ%35=PlZUv)l=$9LIY6lV8Ooi%&VsH1dZ! z$X>cdGPL-%>j?L+_(X#x!SULzRj(>*)I6Rq>+;MYjUJ11h;T@{iKjG@2G0clN_n^RK4p+`<3sD# znv)sQ3UaCC%LfuKS#Sp8Fwas=Ei9fN*iNY$xgrN7_lzv5zeY!$s0}!^klSyo-rV(@ zoGj1dq0v?sR+hOkQX7{#XHDW8D#)I}v`%Sc*yGK9&M*y{Mc9n7aV0tE_mDr5%o|z8 zQDUO+XA##fYhBHI;rjxGK3k(L;W9OB7>l++7OLlEXQE8th3r}3WKH|Ewckx)>mK3n z?;pI~Ol3xAq~Z52qzKdduh6)uDa~gLQn!!S%i;Ye>%D<#AvdS}=P+W7sol#shAGX5 zwXOL94(90ajO0gPSWC-^)YC~Z))_b5sODQD6|LHRvki-QE)nSsiKAtqRR8&_x=ymM zX&2sOUnu)2-$f?*&+PU%tp(LHnkv?rdy8?46&uNB3J(v8oSsqi`B(4IrV(JHGB2w2 z1^C10N}!OQUtO!l0c_uUmF(yW-KSj9AFfAd9G?5hb~!t|cu^ z_yjUe9_CLKUp)5DLn=RMlLD;$0~* zDtQfi@cG{hsmNUQPe4gHen8*}nw8qG8289~1c<3<_*5Y;{( z?NXjuVfwM*ZmT=327$k2vtWWMrI|`))lEhk7W-&WM=`eWXW=!T^kKQlFdEay*9n_1 z97?=f+T+PVFn%70M*wt6S#e%6K%RM**w(<2dNd&KBQdyDwlLEH>!@l(EbW>hqJqaq zAyfY@zTZ2honA@RBHP$V;rQ)T^JR!7hsyfYEsWlBpD%Zb2b*at~8Prkn5=;wb`lf;kwA&9)CECWP_${#BJK`!WMLN~BWCoC^7qL4^m>Wqf>?Ja?3wNNOhGYp zrsY>sAyr+pxmc8X-zDJsc$g7WIf`oEx_IJH&iEoq=Ffl&4WxQZ6-xGz+Vj?{u~RXl z1^h|His};xOqKFP_q7%eZV?5`mR-~f$K+#$Z^jBk?lgppw~JW<(zieFTW0)m>RAr0p|SZ8$`9n z?D~k3Th(ePK0E^nCyg@cTbt+ps6-!VBa}kPaq?M}sZwWEL-EHyZ;4>OQC3YPc_0pp znTuM8o|IQt`rT<)#;&aCO7PoNU!1k3paF{vdmCPj?}Y!A2EDKFkzX>Q5v)LrG3Akq zE3u}8!*MV`MVUK%G3!lHmTDYn;va>~`;GMkyVa5{z zG1Z-hD?pR{ECgYPYU*qh%)!xtUDhjYsU2(0wn8w1|L+A@Iq4d-%2Y^u6>-xfVy_H` zckMVX&99Eb`@%$c;)qS$19mtC7#*g^c40H%WUV7QAgR;TOAu-Cg0bGSZ9%*HYV>^4 z@BMR~R~N$;hj`1Ud-wg(xx&S2#g4O7vStGODrD_P=J&u6&fa*0F8y93vY=P(Ow)J5@zh( zf%fM|U&K@c>avd;oLW5YcIS)cgCcxWEza%`s;)A|VgsR`snAfz@Qt(~b@_0q-H8QK zdNYP%c1Yt!w}p|Tb!30FOdp}1j=BszM~UXBWB0{^+I73bYGHW_*-6K94fJagsd zS3=~{yHDpW0C(MQx2rS2f2TK6_kif=mX(o`-H!YGhCp=ed=WQ+g=c=wIb3_HYclTq z>x{fc!j!?3C6<9mU#?ZJk&oCMZy({AmB_?vv8f|H4e{o@_UL+;2y*-yjxCj@Wb^sSO%QNb8fN zdQPb={-8CX^GZopD7qktOG=`)x3^;| zj?f5|jTh&C);ynnbl&_=d>ia)mboNKgNEMBfhmAZNa{f}&iY-ZWQg#UH)04TympMDQA#eJEjlIvB_uPubHh!DC(^>V)?k!oZ z4c*cryMICaPP+re9<}bNzpO`*#AQVmo)p-VE~+&ezQ}L>c`c_9N9c+{`aLU=Yo^;o zRI5&29s=XnI!gX$>MK^aT>9hl;vCgb)%6UH<$qeQP8omvl%+b&&;T+_Zw`iD$GhaT zEYmz*yck2kjn1T+y1G^P0Sc4Mn*V=Xj_(8(uJv$c$Lp=%l%77ViQw{Vvp~w08Pg;1RFr>8P z4Hz8h?uoAG+OCJeQ&l${ITr)eeZPX8Ogb{wMI!a{@>pXkUME&{`i-)-k^Nh>L*07+oNT2gRs2UluuDIu34vQZn=a zg9WQhzk)dZD@o{EPoRN;?mqQehmDhzKgEQ=2x=EPUq>)_{{}ns1Q}h7Yc5Q0!iJ8B zRnOsXeAy}nkZ(=w^2OZnm1>}(%&bcptCpV-=O`EUZ|qWsBDy}E6#3D1v9~Z27!u8r zBvj7v5rU)Iu$Td*jJrf&u_&1Y*Zdno#G=fxT)d&v_7U@O06!$Qg>Mi|VK zxv?mxd>3b#$`DX>C{QovXM7bJ|3ci6Ktrx^^8hk^Av9z%HDRy8!`0p)HE`f8p;NiJ z$-!0IT>rfL;yG zjG28+Rh&absV+eBEQ2XmcRz6li#3PYK9)G6tT6-l```B(YxB{GFu1+Gz%Te{JYbee z$CC^2a>T7ljd@E!6^fyp&ngf^8-NW~iw&)fe^!yXlV8;$oxiw4BeDoK+HFu)Q$PkE zyk#Sc=2>2jm`-QerY2ugz{zY>S0|}tc>O$@sZtKlh3P)#iz<#hF{v{tr?^Bp#<#G&*q%%9j{HYSDyGcG_$&*-(BS^e(gN{ z9?i)^E9N4hcc;g`xJ~`)5$S@+UQ5;y0?+kgYxChSt-q9sL+bE)#CV$m?m2%(&3-3) z%%*eZ6>=ojbD;ARrw_}}GGM2fuiMqs{zqSs)7P!~uRULet_YfrT?RrW*p?!Z%8!3S zL?716xdq-po@&vfuzF?w0`(n{v+zKOF6_O;pDAV>-*x5152}r%5+-9dkfsvK6R{)}0f!*6yRrzVtqJ)hW7B*E<> z=AtF)7JXip4^du$;nS*=Zx9}Rb3e@C9C{L)`)&NB{?y&Rx%H6 zuOC5+*XK-T!XY1?Hk3d7JtL95!Q}n2H2CQM_%QQS^ndsBX^#hKCHHQ<>zAV8SEA&~ z@S~=|N6*&&Tu4Ih?a1&@5`EfhDA!bX`0u+0TNyxOXRJSJId8nP^cb=qIu#D-*+trn ztoJ7_P4S8(1V7bo6*782!ngF$Kl$`mA1A+SkiIJuEZIezH&jbc1q{`;Jp6SHalZDv zdQGn%{o*-G%+9IQnCei7ju9Q=eX1i6!9H;BH?<;>ev}-74J_<%So4!wo(Vc}9zJyM z-L2)9>RYu%L4Pw{ln(T%eln7>G(`rr0%nBESrp$y5Jh1un~r%7NbI#sPn0EanQB5Y zXEpl{Kf{J+!<5e>0DuNub1Pn|UEK)`#uu!+-7wKThzNs%3{_&I0ntAqUDF*?0Zt)h zXi0;iti`qY2?l!zuOapsJ9?=9a$>8->L) zMxes`U=#_a;o8eHZcf4E)=a(ptbK7&Nliv90l0iVKO;d!^>v0tDL)Cd#FXXmr`zRWVFs)*4mThz~o2ggxTWoeVwZ+&6^tZbeD&d4N zRgNZ>y98;b;=8Y)^Wjr$%;hQp#kBF0H;odPedrQUja|MkHH1LtXc_7{PaxD0Q0cMx zNOV(<_tJ@gu>)E=?!nEEscd-}nAV6>WvA~rJ;rJ&Eo~hxA$6vJA5h6hUI7nb!Fz)z zRYH!r%3do_-HyJy07&n?pR(64+m)NF$UEl;J2tyW)W7P=`>vswU2!n`>rPJZ=%vpG zu#-~$)%w-z;-CS)zyqFm7ia81{^yZ(28wnp#n?4IwLMiZ65dbafI>8!5yzkP<)mJU ze*!TVLNPm?MKP&Y4PX$pAP4v~`m#s7uo(0L&ahKVIQ>Ouy$ zQ8ZRdyiE(@>g91x$FN^8RI|NgRhUu3jCuTR9#^@1aR*NHK{_WZOWYYi{!)5Yq@u6FH{vIOUFD&C#easl`&bfn<4P)tB(B^dWW(7R z3SWEFl$g;d-9|+=)!$g04xxt}`;YMK#*^_F2XlvM6=C)e;UG3@qdQ*!{9ifAQelbz zl2?myRxSlfAmfPHx)w&VkA*aeJW}W1j8WOA0b3SK%n& z4~&7mU2ox=e=|IAjni%-3~|`N#QhGvuHzK;;E%GEbov^fS}}m@UdKqL3MSl`bZQez zW-wrxp~jbZ=-!lYPynof@v+SjymRnFkbBd}DxaU8{^IFFAbeyI<|Y|nVf}+PvTi*4 z;{xksFUU!{=bD?v|HUiYtt70}Gre1-A z3KKH}gO$X`X>aTWLht-0yMCB^$`JO^ech=~+cDL1Bt-`t@Q4>>AE&@5YgcQ{Gi)mk zrLJoJnwLGGm=H+94ggC8nADXXM0l^AP^u0FE@f^&+=-4nAUCU-WljaJ5RHq+GiaGA zGR^a!yHk+g0NfN?$&hxYKAlTgs;`49;8iMHV&|#0c=z>(Ftn!OcJlHliBaKHMzTuF z`^jrMYXKTq;vrr{9=dxPUDvQ%T*vLuSTp#a&iQ;Hz0%Bj&_ooYt=^@w_q?&l2V2iS z<#jkm3U;hq9j_U;b*9N~fnkqmx$)`uR%~D+-A!86%&&yAIASlpnmm1RqSD3;N@My;a^v6n#xC#ECGXwQ;G-)Nf+F}I&ZWG4@8O|H_AX2I`6kiRInki5 z*XrQtAIoi^&8@O>Z%NAymDDDD z9npxV-wVC6UZqZ=^c%O5Xiu23?XP!G;H!sWii43 z-aNkr*r`X9!Q4j=_YlQYCs>L(7cGxyY#m#M6E!I$bVYy%cylAv#GoTq5T+0|ql)EF zg+>V~H`4qW!+N@Fcp=m0ad+P1u<}yE^_&6k`Bj2_6Hm}Jx?CaYAMb_;&Mf})g{;q` zJ6aH^Y!4#Cpr6+839+@%kH%6?%+~LOuJ{-RpB`&-$XNW)l#h?X`a%?Wm7z81R9iZ- z*nG%J^j5}d9nJr8HudHMfqW)HjG&xL!x~=gzq%~jqn3QDx&p=E%flU=RE#+M&8`IH zT*Q;{x?W;+x?peg&;R}K=W!ex9EHDBjGdR1rh2>Lrs~PpZ6dA9f85Nbw8_rZ`KXt6 zf^};x8)z0A#_|{_0TMIg1_>v@0?!SY)!*7mV}8;LO4X;KluB2FmmR0}ASSYHQPMI0 z3A`!g)&h`f;F_XL&S=1iKV|cDjS7aGKMM{?QgVarvnK@+6fukT!8jMD?Im?7K(=%r z-QK|d-IJ_HPBJNzK)1i~7v~HAeu=*#*@S-lg9%^{v8zw34|o|NuMR(~--9$+uLYUX zHX}($qhewP3=$Wy0txqL#x1B47q##tN|%M{WmGc1XB((%L>x#KAZ>=^YE=wu`y4ut zOx%o|(E&pE5q18GPzoyA<@`?zQUV;e70J$9v!-SKZvnG`(*A5>`TAPb6h_*r#gac8 z)-^p}$N)$)3s1a0BuZ6h(k#U=a~=N)#$%cnS62_kksB6ML~E%SG8$#XpSSiH+*}M| zr^f&=)>y(wb_elcCDPKLc_3qOvs&f%AZ>iy3@(8w`P^Fohi{DcmIwz3Jp;`}xDPtDN)dg^=vO={s^xUPFNs#tR8EXQP&F>uFcD`lRoQ$Rg~*@d_8;ag;T{1^ z4beJ6elp$ZH^@KxzuijU441H@xbVLE=Ry`!vr3r`g2|e6$ymb?98)=fk80rWR5NXp>7+$OO4Q4B zKnG}xFtNK5P{-0)Wf$ix7%ZTM@nJ)6ffcM3C9I~==APu8niq@6+hJw>^jv4DT?i@A zB>4OBAuA+1;`fP!@6j1I39jzn#kaN)HCrz+6t=AuhuVv7S~47FY})AT@3y*prN;|e zzlw~~Bz05!z2Uw4X0_(;oHDb5)GZu;I6QuBZZojUgJ#}wpx2fc6+8XkQoVDs7IA4w z%?k5URMkM|GR-F;wm8|G1T#$-0@^iXmal)y6NDl1D}tW~EzsyVkttW{ar*ETqnoYK zso3(BB=uQzpL-BGR2#=G8d;m!7c8NC#7x&JTtZ|dts~0KO5r8r4HGPV|0!-U7~Qq* zO{~I13%#0~Ox(C>vy9ycf~ve^baR3JSR27}tyU^wyOtA8(qemXn{VME`?0F8rsmtL zuM_V8`^XFOxv>DDk~An(Th^~G^w958LHMi!(bae896VFVj&cHQ4`P-pWP$iP8G1`q zSXzE|n6JAncf2NhIq3gP-gc=iIooMHjZa4RF+LFy&bv&zk0!xl>yKoPweJ-&o-jd5 z%uPc{(G#X{s=pv{7QZsXvC+@sG{~;k@I|PSlhl5139Q{_mMiGqZ#effS=@@o&|)W7 zCroJMSE%{<@2ai$DCg;q_>ZgoxFT*rXKoXR@VtK)*6lycW_sGGXwR8t%$yPSJGrOO zhtZ>}QscE?r&#NGO75?rB-#0+FZmvU@O#qIw6hZTpztlR720r~?t7OVQ}jlkdLyLO zK6uvW>0@x?i%a0k2L=q58gzZ20gcqIk4{P=f&&bT304&~I@Rlf9G=@Me>_rf=M)V;AzwS0`cSFfI zl9EyVw&#D65{PCz#R@>Fd+$nkXbm0Xv?=@iE0$?LEYkk$8W)u{y+XD>$#^zXvd)lO zb$5;hGV%tUO?nBrhucp6$B6D5py7{RLo1`c^K2>(DjB;tpyhxFi>8T61-7IC1C3pr zR8?_PWtDd$$gXXKrS@Z_LdCQ&x}t{q!M2L4&BBh}#PxX#>|LWs3K8?aA3Hw06X2xC z0?V>;hFzt6cza{|cAr4UDE9QzR!{q82p({GZT>jT6}`in$)?PQx@jve_3q;!smL@g zzeZl$i812x^EJd`US@x6f6hEi^e~vl9J6U}c=b2}4MYUk2f?c@+8muF{b#c>to$0H+LY@-$YR5246#T$VgM@F`l(&wjs&p*_BM`jwvQ8Dvr$BFf85&h+Lko z;MMr}LA}pwpymw#Y*?bUZ`2h@nj?5o=!2>gSB#;WEMP{Hg8y}I^7n! zyIpYc`#*`?=rvU9qIEJnD9!izKEC@Je-p5#)IZ{PzxtlX>>s92IAhTWoYzCN@*g+{ z@$m-zgCvtjT8MHiDn^fqayL+;`paK=@rBFWs^U;l7HB&;D=IlrEAx}*xr=f@?uj`t zl4S|^K^hVU--XSJZtO8UL?rtdX24TB&dIW*EBFfcSoH-bYA4Wo2QONWFY z-1Ga~d;f!(IcJ_{@3q%n-}cz$PKG>Hq?L{BPbHkbiISsjBQaIqkDCsVas23C2d6IeIfnNDp3xXM#C8+^>KE zor98STrbU*sv(aoUMAf@@mRHeG%0EU?P_xFHWoY-Cdn`%xIrE`y%`&fs(2xaCmn^K zm`jM1Ye>OmosDOcacp^2#A?=%!xfjAu_;>kr1LT+0b+c2K&-teq1oEWmWI`=iEIwKI1`ZXGH7@;2AmBk#tm>q#MEa1dI;#i)+;;Xu>#?L0)Q2v8pb!Y{vyVan^- ziz%_sF#s^B5TN7hjC}dhcjw%GJl-f7u}bat@7f#1Im6zO#mVw+=6drDTIuP&V{bOnR0|>$Ujt!d7?VXiJIR_v@&ji4addWG>yT%w_tSScenyqntG0C5Coqnw?IlI z^R?ClWFoP>{;R=Oup$2#q4vu)E2fzcRo2Pr_<`~#oBq?K_LPOt1uYn793|q_PG-hl zx2^aj8~?COgz!O!ws>0Xoo9;62~}v(g;F`lBF_a111C95<-FF2rwk8sih3UDTXRmh zuJ%8)(ojKU*tC~VfbRF5?N()3Ad_#?2obsMB_p}}>@anjx4+%H_v`@c4s`EH30N)d z-5ROe>dPbmLBIu%3b8%c_QuY*m8i$v2iMm)+~s9G0mp9+9)*f=SYU+#t9+X`*nU%> z9)>;5`jD@6Km^y^Z3?uv8z947yY3Fi(zcobq3z z4+dU0=c-8_GY};WTBf*6qfIh6_?NTz797Z)Qkv2}*L;p?yxN3@;{#;HINI>B|Mf|i zcM&glPr$wQreG@e%gfWj)4|KrmxqaOPZOlvieJ88*o8g+x)I+zPu@68em<=ZibT&5 zB9%~#`a{5OomSa&Y6cp16AJZQ$r>w4wt8y2_L%vXl+Ri)Fy}r%_wt)wb39AX*@HR>HtX9_ck}Sdu19@!tPSbBi{YavSPM=E-nfj| zkLQ*au5W*RVFyOp(eRHlJtD!=V_-htwlh%+L&1NsY=RMwJ0Hfsr;#hU+yp_BonYXW zPrv*fOnyqW*l@Jw;$kDbwh@@GI_}WdHfX=rh1UIfx}Ch*F5>Gl1fwncom3rFS!XR4 zdYVCZeubj6w-sw%*1P0R1i&oW#q!H%I0+LXjh?CYSF=^F^QFnMq82ct#(|u{V?Rp} z1zy2tJA=kqMW_NSj+I%aTWs$dL7cOSI-!+3jBiJw9VpG9C{FTCuiMCUb=yLln-Rt@ zeR#^4VpFpTjCGD-6Zqc1Hj1ttPr;6?{I*=b6aT1=6>ZR3v}8Fr zv}ZUYGie~W+*mkWQvv*3t8L&MqZx}J@LyG2;d=6vSCZWgc z;4M2lWk+&0_#)t=uhKdG?s1CU1pLCznwB`7kIBvEw9YuBmlh}D?8<5%! zu6z-;g`E1?-$tSG*LjHxioK4c=w6wZwb%sl@^2vJ=dR^rM`bomE|?!uCk3loUHCZ# zg^g`*TKiBTyn6hCgmQ}K83@!Lp^(ribaJX#-?ckMYL9wRs3k=Y65=pZgrHh8_-Vfkexw zN8Zf&xhDpr3cQ9L=m4TQrkV~_1e3KWNL95aEU(BeR{Gs9JjVz8^y_y$;uy8o|9ZqQ zG2TDBRc|`6SF^pZ=yT^CpB<%5+bFgk>^Y4U9w=f{fD7mtr!CB>IobRg!#SoESv_Q_ z4^KaIgQ#P|ZCGv33 zD;oT)&{Ae59CtFsIy=F;pG%(eJ6<#TY?yKPPPrE(X^#MQTw5AKpC?BBq7M z#I23<4;K8ig~;uDBR~LHD%-t)t0?-4cyVD~E#+_!rWR5`6om;GTqhxnlOHhJ(7>7O;W<;Un~FKxUASLx8sZU4;iu{u)cq7i)sZ_TUXOrPO(SjLwP^31-Hs%?kg?gM}C zzkkZSPDjtdsdRfO1TF&8x8`4sY1#OvPGJO<6Dd6Vez>$=_rt4JlrLcKZuia_WSQ{e z$K+HTU#)ENhW1DYhvf(yfk%UYuA_%PFq|(|V!%bx$BT)1^fyeDi_LHNsQe%SRFm3{;jZYVCubd0DS!S=o1OQ)>Slw|gO@|f*u*Vp&xJ4& zdIsv&$|S2{kh;(itWP2%%3H}x>Q^ag>Z<@i>VMwr<{eYU_+vmWxPOHy=3y5(?20=X ztFKHW0=ZVnF16y$g-ayPQu<9lQr6GT&zZ%^vI?*UgEbe0m>JYnS3->#h`Dx^Y|~cI zO0nla-vjbi5O;dAh9BqQbl22#e{Y^cQcG0E=)0*b3t-xteg}+mW(IoK9 zbo2>Ffd+wtadP7T9)g)uOjX&7bG zB3hs(rnvXjMr=lEu(Y?EyR=e$J;FIvJ<9v{ImOd~nH68JBKz`b>v zoh8flOvJxv#S+;%&4OC>6f+5IvRFJSy&Ka5$>#A{q`B9|cAsN3>|56nl(#xx z;A*ePzTxL9S3ub{s_>L2?-h)hrEiU0s%*rJj4yBZZt`rRCzk?hxgH^hRdK_gi4VTf zlUUrgDlBgzjMWnKGiut5kC$jM#?71RIPYde!hE|O*Oi~cu83K-H@4Hrwnxf#{5{@Z zhm^wG%3MRUR|{v{W?&9+@lj=CV#PSiwENZmO>OQg#oa-Lni?$8{&;@|cNAnNhQ`+c zS#ts_diZfHIE{~TuaL`i-=Lh+cO=@#?o*0747~&I{(c*^Gh0T|eBR^>IJPfp12s(5 zDI)`3-e)eN`MR^u=xT84^3RM62M*03>j4+6bW5hcBNaaZQN*rvPQ??ofc0t=k%68= zIc-GEiugs|(tM<6au{8kU+a<24mh|Z;Ab{o zkTO-mZ7%#C{EieLA-xYur^6Ast&G7x1ZlVd+05@_o8LUYZe*L|b*%I3~ zmworna)bI(yhdW?f8EVQ$js5m``!O_r}_z*eP`6EhEl0Hlr%gP&#*pn5zDVGPc?+{ zyzDrc^Ez!QoD-dX`d7E`T5{zs-8C_aYdGv}s8fj3Md(P4s*Ov39_z8i$}Y!v)R!sW zSO;ZriGwNnJ9HN$Ugb{XpJD!n+fP{RC>_sZ&27uGyyMKrEBxHwHve{Z6q6yf$X(Wd zit@UFEkN6TNS$a|_1#=nw-P$nH@tcFN@;QW6U`sIdHBDc!P4LxsC7Rmm7_n`3<`ORgRj0%Ffp!CiN=h{R2XRHMrfgpN4iAFBs2&davNVt3K?hX?C^=vRo zdYuR#=uNZ3@IJ5}g0@Qm}pAUi=*>cB$gmGW{H!t5;Jt#kgmHF_5Q++ zx&oT8QCGVo4|_eRzw@G5ou7LRNm$%X@pdN=>xLg+$U|x9kqR1YqLD9eN%=}CvLHX( zN~|s`mQVL-NL<^=f}c6)0Gv(8Wux!ht|hDdN={Em>VuJ;f@_h=KIhxlH}PA%{2z6I zH11-x82-A|lpZWx<$N0eF8xU(Dbc4;#YZQh&)}}Zv3$D&>I6ZwW z5~^30W^Y{6h~3<9I=g6hWd0KJQ;ge)D`|Y*KV>?$@h;k8f-N0emzm8s&+=_?al9S{ zEz%$}qhMxaud0KQDdxf|MZHiIcD+6a;QcGZS2f_BwGxo^05SsEP*=$6u1f^`Eo~ED z;j_P1NLN@SbwT3r=lmq5PW>=(R+6zJiBWDyJrhS7dph7Ay-X?gC=@lT+isC z=@@ZBSr(77O7D-$LH?|Li4v(VU!@fZz{jp*`lDt`+hh+Id&lCD%Czi$=t8&OGiS0N z)D5qu^w*LfjR7@Ucicz-$`=8Sc_%?i$6h~4>z|$vC+@=_5h%}a#Sj@g7LZAtt=tw8c5JHG|4*!d{^{)`(GdL)^vjgP)1Z;Vm z65<1|bKT0A=mZ=*_{ZaF`qKdD%e0KR?bt>8aM#C-sQ!&K^1)=ttDVyCA?bwuOa2y3 zDFN#?_Dom_0a5{Om(vUtYO(F{{(VXu$%fi*@?ISvP4^T9vEhHWj)pxJo3qnQNuZT& z`wDHrsBIc}^R0S8V>`8pwLi0KYD`MR%na>K>EpAMdASK42mcmcx9vr?m^Zlg_d1i| z`Av&tr_$7SebC`&$MK@YVFQ;5;v}uKRExipnDAX~u0mR|H>L%A#J*`h%_6c`a=6$N z;BH{Mh*I7;O}-?{dez?t;R_>@oDp$+4UY_c9DJD*!Jg9DNI6KDvJWf9mq>Z#CF~;foNLm1TMP(doZebEe!491DFm;sxH) zn@sNSymw8%8eus`gCQta|%4Ieu0UQ36!?jjJ09k{XffA=WOH&wLGA4X-&uP#Hh z0jbMN+bZ|@hTgl~`TjIBf12yQHppph?Qv&P^W$hzETV1a-|8k0jwmJcSbM#_j$&eJ z@%~%SMkf0u_IZ0-q}93=a3buj7d&oK$m1iWI3XuaUK~9nYS!oCOufzN>r(|RI?<*3 z5JW6x%d6iYBCanyl(%siJMcRmXAp!or4H0KwUq-1fhKn&TOTaF01{UyLgOUO{dIVub*Au)W+X#;FZ->cZ3u^I#Gs4EZE#+e0!dI ze%qw`8Zy>a(sU*!;AddcZm}DDzL&)4hHOG#%$He?PnDvvZvb1au`!2UwO6`q%YWNo zmRU?pPWHcDofvo~D{3^GUcmhm#dNl~r_k$I`WK(ql%EqPN!+{ZXe@a*DJd~zZ`q$O zIfx@UZT2QTseUI6;oEh(Oz9khw`t;g>mg3&v?jr3EKF=wSRZmQa`Wjvh5-S5CEqc=zY6fJB>wpg)+muA-hF8PVeb~y{Bk7*h5Wt^&qIdCeNt2 zjICj6oy=zaW-Wj3>A@@^{XymokUnwOBxqP0Jk7k@R<3qyY{gZV67M|UJ?z3`L#TPb z+elR%638L+s|nTI2E#E2(Q4>Ki%HhRbM=C#yRvsVzFK;TksZu<$M>AV-YDw zh_QRZMiZuBdl3IfvmVt&^e54Kc5&x%n8rnw-F&&QPM1mOZGznps@l{ND z@K8UCaViEiL-19dVqS(c1x7X*b^rNSVS03(#04xx7B)dLf?Q2>>~X?#YVlY%==J-5 zH`6gSeoxP>+eTlk_-RPUm=XKY8Q8u>!m>o7t9FmE=Eh-@xyiAr>GLh7@(79XHJzK{ zhRa>$?GHm}1>ZY46yjA|12|k>)07hX#Jx`R$+=5T0I`+P;dh}CvlfUhZ_$^;Fm>mb zAT40wQ_t^WqzjLl_GnBhPw+mb#6iib+_YnNt;^W~Xu@JkUH7?QaB|jw zBA0PFxr{9*wmBa4i7Xa~_oiR^_TZvp2 z`2Zaa9ff?ZXp3}tWeVusH5O%-yl_UN+E}fkS>z9@sL6U9y+LDJ&Z`vLp=HMy%}i-f zg)EymEpO~B7Eujh3SMds8moppqwE!S{!9TH5;lgeSp0b*t*R{f?78MMaeGT7f&-+X zk9VGT35i3tM(lIc9>2E4Q3`~c+o@mnI#L1j)UWy-zG;=Hi?;J6nD>~w-exArXur~2$uzBHNwo9*8s zS_dD-lK;IbSZFft5M6`>-JLbMeckEfZpTp0&{h23h?(gkMGi!_g;f{#47>tG)_&j=t7FE4;D|cl*!29H?S)__z~Q^*>STy305dPP{X3?lsVt0!U0%RHol(^tMKx z(`o~TADZ`IW;3Ym-`)drifO^mbHM17s@j3?0GDWVc5A|P*O{| z4V5gSMi2ZK&;n)j9ndPEH#q#fUN#wC?ymSc17FmDz+1Y`!IZK;{=l-c|&$)R|MWi1@?=K3^BLUJ4C*~GKQ}~pHi-u7x6T=Sc z;||F`KlpLBU-Is3n@rU_*RYFSwK@G;)S;k#vj~GlU|G~jkZ5gLU!9D7xlW&;e(5}W z|By$RV3N%qc}JfdeVhK6LyRWULXldU{cZ%)z zWp4Z74T;=gdA<`mSWT(bf(_m;->8GrAG#gZzn?bd!IC_Z($ir+Bf%Zj-UlT^v9SoH zspNn3(F}I|*_@Obh)NWIyX|K!JBgDI1K5^`{4-6}{7tGbYi0D#m#61^ftYH-2iqqDQlIM2;%I`GbT<6T!vvTBw;g~dlk_qps`ynACi zy4=kzO4X4{e*Sb)kN`QLAQ3Z_C05NQjpNDzQPK=73>oP%Lwk!`&b-msy9ujcdXnNxsIATy< zFnrcm+GK^+XE|rufNgVe@W1orWr5Mi4ORunFdQr7-ol_wiwDUWQGm+(w0)WX6S@si zaTDi(w=y6&Dw-VzOjB`-K_K#Dw0`4Qi+N$@dtXR3=rx8fBc?<{5D~|vgp&G}THX!_ zRhqaKv$^PDqcRXydC>oj#hE4S)BINr5fvtagchlck1Z*}>i8s^(8aXQhJ+(=E{2Jp zv6tSm4^vkqWG>K{HNZw*dpt{t$$o zT0#uUtr#~Jqn)!ia$mpf-DA*W2?xZj6v6bc%9EVyj~+kj;wq*JDiq=^S(N7E6IT5A z6JRtHKU$eUS1KE9o+fE7_MG|>`sLfhC4njXv`DeEV=o9K8~$^)YoJGjrOSA*F@#YSt+WOsOEHF0K0z~KbCP@>Kdt*lCLQMqQZBg{A+L}QnxN};m|46X4 zu>r~A;EvL#WY7p=AItwZBfX31njZ$ol2} zlfiZiUiRzuIZbJuk9+8KlH<2(ej~}QlTKq$)GXEc-kM}C^v}E}gk@%4tj~or)sVWc z@u>b_V==dD0X>Xpb88OvT(xWcz6LhatTu z>2r*J!y8otw^GU)>P6!z#SI1W;6g4pWpXNQ{t@j;GG(z;9ze)?d#eo0yF{+qBq(U6 z+-N2x{+0y4m`$!Xx(ScFoTi~q#HN06zfTIfwLE@_{Si0&aBsf7S?g*HtZ|hpTK9hl z*ShssY0t)|LWf1c&z^rN>~0%O^`<=CoZSh`raB>qKRQf6lwiP+`WrF*jYa6oe*8qh zMB^8K$xh8o=gnADvqfsFVt}tbaFMyXLp$lWL)ljx`>B*S3i`(-^Af~1G-Se~b-OPL zirgUU1tt(OI>Q*WQkK`3F2k9homN$1KhYzO1}~d$!vcT$bb3p6s@K($3VnJ3Ar20R zQ;kxhfk5~Dv-g;!pmGv%h~1fI?YQA@o!+yBx2SrMqew~`Sp zA0r9#S-u+1L^Y)xsbE1Q;=g(CW$;H6Yw+c{1sPX4mdvKV&7Yg_BA0MPMGR;^08R?F z6~860)8rK7Zr>RIZJiGDJZ&r*4h6|CAeXxUZD^5T~nmGS#q~=fTtAUk2 zlSTxzwiKUI367~4-@qLP-K&-jaYA(YshvkEtSt;Ug>hgEutP!Vo{b-pmVd~Y*|j({ zT*5{|-?x~6Zk-gsYM4}A8fjW8*)5;w-VP8^r2g_fzAW9ol~aQ<>B-wLnRp(Csvsol z{uq;{&+PylqlcPkGIFxv%T~qlf#yKA^o>GhEMR<<#VX-72MIWGNt5*!!#20rrxV(+a z<^??Y>Dv~mcqxP4V>)|FY=jw%qA~ICxf76WJo_m*GDB#TAHJl#`~lUJYtywPvTnS6 z+EEl6)4yrWUSluzWO+R72`z2q3HTI&7j3F@y%VC-E&2`t*d;JvM#yvg)x)gC&}utfs{PNOr!&h6K%ZrUdBYmb*h!PXfWQl=6oFoK-=oBoCFJ zOKu8oW}sr2?kM0 zzV5Maqbg~^K9wXvm9T6eN{6XU4=8$`M0WTXj}f~CDnB7}WTbeDN&ng3rD)RP;CqAq+#9=Y84mh6G%TU-pwS!+l}8>Y zDordF2&UI&Aw5c6J69fdArHqwz)bt4J7}o@mw?Y(+N#Oe2#^7&u!(aU6K9nx2Fc(@ z2A|S}Ao2o}LfRxaFTN2|M=bml_`1@r9{`gO!97&ivRzKA9Y{{tKAZm0 zATc`Jz1@LFsvj_&S)rscNcV>7o$nb9u2`^I$i>9X7DvFl-T=vemPd=3CUCEQ_h3P< zhf9x_l?pdc5w#l*L=aC#M$|u>l*0rdyz`TD)0HnNL4BfJJuA@LgMTiz5UCe|A?nS$T0B2<3Kj>B zq)Um+n@48@GhASs_RCl6`k+@G0_}D&K--c^O6^2BQ#B@9EY)KjDE-xg^K(65phFnU z_*>s!uERB^)*Cg{)m(_QHU%0b@}K|E;{L}EsEBvBU1GI>GCt7@_}GfulU1$DnLE5F zJT?6`zKQkIu%5V=98uFlV-Mw^RxVA=GUNAnL6!K6VWC|PU8kbU8cI5SN$gUULRwxx zT2XxQs=yS872*{9abiR_C}j_7YhqGtE_AcNXsfa=xF%7ov!4ju_kyppLZ+=Eyun#f z?bi~!AL$mg)47pI86W{$mt}hs6R{o$q(sLcjYSv_IE^7PXk%~qA>EedX{l83d-^`8CZV#VnzHyY8fwIW%jvHNw zKYvOx2_~@7(D!wh-)Z@7Y*yf~9gGI>C*GVHsE=#_EnACDF_Nj-+z4e6Io*eQ z&@4t@()9L6(a!BzD3yGPApfNi!To(!+Y01D)dgR=EP%f`>)ihC{MO7q@9bL+NjyOa`W=uw3~o<90dVHbA_cz>E7 z(j@RmN<-oy8cZ5mThDvz!XZZxpyth0yu4o`EUP!8I(VfR6tLPnzY z{`?M+HF_FNaDJ{?x`Fz8+NMt>(VuHoFngw$(`3=CM!HmkA7Z?a>-NvjZ4{I{*Ec zWSjFowP(`-%Wg!Lh_D)`KXtZaC>YfdLq2+(u=u|LWULGNsg%)pZ>hnK_(R_>gV17b?Iw0efWS;3z9*}iYZHYl5nTse?r#3`}D6bW&8m=aWFjf^Z%xJiFMCQ)YD6!;b$M)Hbw4VjJ%PM)ISm~$bsn#L0 z@e7Wj<*p>&#lfcE6-MP?9gnY6JJw>k*0ng zZ4vR>r-8V|XWC#L=ykq|Ox{nyh%4P@N~nAUwH8^<868-vStI&;@OkU#7;)8F2bhk(k5ePXlT{3vWT!>7YRom$6BpX``vk8)Kl)JEvVsf zyOAUx@N7tz+!br{JnF|hWKeoTiCIAk?;6{)pnKL?!c2vD_P>XZX1+U1_x7b4zMWpo z2J%^BmE`g$%L#!u18;iK+!EFHLa=BrxY6FkEd0)T`m)YG;^t}PsahV~`QH}{G3hY| zynUQc+1Oaa2ik^(4U&kELTJB%_;Ukj?Lk5gJ6^5aO&Zbwk> zaO2k2CPMClV~8&H9%2y=uwGorWq|FLtEuOMJ*>8yf!~g`BbV+d;w?(FYsS7?I*)@AJxPU4R zYPS6r6O=-@zgwNZfml&}hzgH2(#>-*b$h7_PpHht#*Vn&CfMQ888sjVa{NPW ze2w1%JiY7Zq7Tmxcn9WnW+)kI53b4M*HhVjk+xpE`?hcH4`vPkSu8YXJRV>&ksX{Q zPrgm1#(L#Njoay`)9gugyfXi&-}#op{D> z*8yItk-2+lvAN;XuXTm*x?3$RIBkFIm9HTM3JmF5JJofpMu~e3Y48!S607l@)DAx{ zHHmvIiJQf*EcW5(7P0SHyGjW-`C4wv{=}we!pROl(L%qqph1*|NVt%cm9FF&9Te3L zyJO$xj4aom&UFovDIA6-nnT72DI5$SZDoE)q^Lz&=Xm2M(x_c;oSzGK_P|L zwQv7!AkaR|V)>`c@lM-#>HR8%p?_V_TO<|_)#ttdU&zgN3clmfkM+w)8b+(;TG>6B z@+B&U9y^Px(Ei7RG0g(grYN6Rq#w8|d5w~2(MuerjX~2bPHv}e;~TwZgOlaWRtg`F z`b_K(B>#!cvgKtM0By|v^V-!}eV-FllAjBH({ga-p9cA8^&x!Nwnf=KUqS_t46Fwq z7IXu)u-|bgCeeE|V=&CM(GgA<)Pxz)pbt@%PQ+MVNE~SW_w4deh=kg7jz7ro zS&o>I9|9LhAVDqViTBX5n(`qlAD3mpbx$2hjDjo5eSCp$fC5&1c1^;+#gC~5o&34@ z53|`AT3a&(RyJkqJ5-}JpO^I4zSGmrDMeN7bW7DHHN^9YWJpB8Y+lhKy)_NnG9?|| zg@eIsFCHu5Dm&B@&{2sgl@69Lf48T^HPp^FE57-3ErBh0>qRr^i2nWI5x(HTF!L6x z{i+<4!nCk}<=LLGEn9}y{=vy5Z(nO|Py|LpaAksH4^H6fz6^~|5XGqPShP;@#+xwF#A*jJ%sFSD}p7|*@ zk70JrWOe^-iL|jwc+qDW+C5e4In+>mk3A_ z{ZRSB3!jd><$2t5NT`51*TW3nU5Iy9tXNJ@jzgfg#Go04} zqDHpsLRHyc%Ym`xYT`Rf!cq<-ojr{q0AN750v76&`}vAer0-5!S<=M+>8#a3e0-B5 zoMC7z$H~l7i8)Htxx3s$bWf0Ogu39pKpK3t?d0itgGJ@|+pIiSd?y(51hHv>gg#h+(d#_0-UTz6ca)3FsNAKOclsvF!-tia< zp+%QBXed`W9)kW3eiWe66a+momgisf@XWY)4Bg^XL|;0`;3RImcFk*Y7Z*R6%ln>B zBZCb;D|Y*RiBiEIPjD_NZCd!pJxC(!Kfbx$5Hzyh>#UGioR61e#~0I3+e$LRhoB<7 zf;BWkyZ3>TV&?dUcjT_96GU)jAp_!+viHf_iA15cu9g|b=QFKul zUhR2M#5?43Y8(u0VG9+bkB+t)diNh!3#fgjy@p?NX&bqCgsx(IDL}Uqu6I~%m$Y=h z-yM!En#yeMV#9yceEgl06lpr7Oaz5BeR}&Sz572Zd57Jd550USUh(d726T;>CvFa$ zgYsYFbHgcPWRD8X(fE;qN3GG#H@Ev_*|f4yEnUkeGya-kp&gUeZ3Ro)0xoQ?j$UI< z^6_E`KltFyecO3Hrf8n4YsLM^L%yRor$p6wMrJg@b~kU62iG*Ja6M}#64!i*-LkQK z{3Pqox4!kW7;hKY(5$AN-KOO%n7iZF-ISra06BZVj0@dUT(!TrKuH>Kw` z1_YTL)*)!3-A}c`rM|?!7%{xxJ;_O^2;V4|WgnbA;~h!=^byQ1Bvuo3*sh4oCt5zG zrz}7tr>Q_Ho8AC#&euzUVy|bDF>kQ8J$Czan$8srv@S?W(p+zKj&u%$nd@W;?Q~l& zG6$SPCeXMTc^X1-T@x^)3_A8V;9%YKAOGZKEgU7S&U8#hIyt`1^#(S&|BdWZ=j9M| zDBkWxH{}kJs@M9q-MPSSx2e)BKj)8pTz_C+)+xkZl^8W`R&150-yHG}O*$BBZ<*3S zUz{;wqPgfgV5XZoIM}?jyo6G+;53DlGWVVGuBpW+E*mspy6Q*chMys6AvbpIf_44g zbn%G9pzo@W=)c=nEruCDdF^?YDrP49bX$wsrCzi3zd_&C8Q^;ZGG9hP>{z9=oH@s_ z3)o~GyOlw=vnCveVZKc}56wH*Eh$QG*^aTY6xndpuTOfnZ4bIa=)}zKsH{ex{YNfu zY8;k>{dM_i-5l+v<~o5`DruC3it?;%pL&zF3wVo=0lnC&D8<`OKT$|WF-J#ZS8)vf ziIzH0%kj*#hxtmUIcYrS;N5-mQ}z5X@JaQS@xwwMVP~4vBt<8m}t;ep+K^>)C%=JL49Vlsr&M)!y-;kxUC#t!Q5B8A=zrIp8 zDj{aD;yAcaK{rRz%v8OIrfepEY_K)7BJ_>Cwfv}64by-?uKD6>Q@Nd3Mc-F=IJeTYt*qkW0 z6n0NSjlNPE%XYEBXPlHpuAZ}2#18h86zlo^%JiV~zdZFlo$9o=ZFD)}O0{*X#ibr( zy+cWaJF!`q9D0U(bf|Rry0dsJ1b+7V5UG{Gq!eoT3ZxEIJt2}V`JtP?gP;|tgCIpD z{tu<;25NlubU0DVb}y@P4vAmstO+gy^aTkeFH#4CAYCtQ53AJA{UZLb$Wq)OY4?563}cJ1YN6lfwK}+Q{m0p7(`txz&_gnousMah(YX;L=-K? zcLl?YJ!bU-Mf$lss7SHuIPpd+GQ+(Qd7H{L<^~h!5rQC=>H3HMXZfgkuTq4=b4n=U zWjImg0PTPHDlj4-XPm#@X)#2+VET^32wR%4P>GoQc-lJJ$&^)(14svp3X)ZHBs`a* zu`GJShr#rU7D?|vzLGMSiW*%BzY-BvJ=2DStxmAh+(T|%F`qwIKc%;)hCgut{oFvS zDj(Bm3&rQbq`ee6-)^&1X@^}{M?`M zlKgK`AmWwORN{@Z<&v|9=ovw*QK}|w_~|}$XmVxY_6)lf48LUeV(JD!sTYxY2W0d| zE2;v)lG5x84)*#S`O|X}aj)4&LRN7(+mRdMJ2-Wb8{`g`dI3SaVP$~k&b4(~Ke+=>`JE@jf53}8ZE*BhP>Bohy>AH{R!0M)D(L zoSc)r_gZVtd3hMLh-dW#r@ITV`t)&CHJEf_H&qNrO;k zpp{mZN{@6FE*K8EfV0+pG{eT0HX&m;ffiPv5I3K7KCqevc~a>&if}yYp3r`~-z8?V zm7OL}CYX!N7?{j{i^KF*vteLdsWYsIcwMCQVj?TNIlkphYJ~rC@x3wQ_dm|IR4b*O z^~ZOwP549V9F@k~%B|jAOf@{wqC$R^lh^u(q8QMo3r}W`w}{_8Aw-w)^A&Iy9k6PnWFs3 z*n?b$>lbMDnsXz@F!g6iGwPoOql1}x?Q z|Epl30vTDx>weg?BQne(jA=e^fGecQPGpVJqJjU=V6ZLyY$^8flLJ{{d_O-ry2i)H zZH4RL;O@ck)ZN~R$2Vey*gqQ$hjzKhU%3eNoV{5(P-PpikFlx4Jfn|tmSUfI#v%ru z7cEP7tb#`eWw`2Nb;q2H?9ps#&$oO6+ABI(#bDp&hr)t&zW=Ofts_!)7Lh7*cL`4I zcinvt#DG4La^+!lQ8RdZm}ibPvg|!{+i>gVGy0yKWl2{WDx2m;CGGZV(NlW)qE(Z& zAJD~Ng-quO)>WP|-DG|qsgIpu_ahTE)xZjH?}W<6x1R4vMiT!tah+d# zE`jJMCT#55#=mW)Lut))MQpU^EFzk7`#FCq<5f4cyBNiYM~QnY5lp(NLGH z3<-#6wxXQylQkum24z}j-C9RDT!U`>S8uMW1%us>uRv~|(U+W0wnsTuO=yQA=VtcZ zA-=ROt)XHzd%h!`;4~&3rMd+>!X~yRzDQMafo!#j^c{z_X2@Cc+*VzwcVsbCf>{Lu z5q69erDq6rC9h)h$u2orVQI(kR`n#Zm5<>1zykS+%bR(> zQ3+2Q_G_my0bSHw8}P$g#&0jG56)bw_`31BwBg2|rZzV4u@KLI<$ymUE+N8hzG8;E zW#;v}Xt$MV()Y8GSax=J zGUE#u`At>35rXFv3+1NFK>{zf7D6*;qg(vT839vXBPajG-f}kO^u@F85I|zH>FOB# zikrLnKP^Cb5`a&a)lPgwcYJ&46{M7MYxx>&P^T!^D7@sTKIUb1}0pGsWqtMd+0Ph5vnb;rPcZee(QUV(8 zgIj#%K7lpYOP9OR!1>NUT115nK}IW1^ZF-7eJ8w6&DlwAVgiKC#O(oZ7Sls{o2nEe zh-!^qu+CkF5@_$QH z?e7y0e4~QjF`&h|dAD6Ot|sDGZ5@XyE>|8H6O@z%s%X$3-GLRmE12g6-!uJ9e&7f5 zJ1Jy)25}LgJ&a!SL|*mBnCz%q!OhegbqO2Jt-#nD%uSck*PbL;SlD=M zM5CFy1tTS6Q@+;=Gg0q5!G7_{R9M&V^f^kNnamh1|BJ}H`rupLxAFN+6uPEk_#l5i zmxYVYRid|(-rh@Aj`^IuQIiQh`NA)#VnO}Z?BU}p3z3RQa);|8kRdcudF%*iKm037 zTr$%1P*-g?f%@h9NOrIg2_Zi^y-4qJU}^RR?FrRxsqo)`8_>idK8*kFxuesVI6SAE z35&m&@P^xhkC(K@BxdnA?7W7X$kg-I=$`z(vS^>3Pv24J>O(-+S@sOwFosX_nLbZS z_Pzpt6QRgBi*@nK*TtHBP6Oa*IFT{P!Oc>OYU*db+c*Y}#i)u}^*9S~J`u7~2(T^cY zhLZ*sP6Kdi$H7=Xg@LBl$dcmEDTjo@+f_*<9^CkPoNhl`ct7~Uj}ZP!~AiY zZ_aM7N`VNFYax18>Qouu$U%NbI)t2d0UQG4<2dQoifaJgN+tTPzPt)9YB#rNGsm^- zT+kR~XAl=3Sv!*JW8AH6FR(s^h}d3XBZQo8J<|C$l)Ei$)@Hx2zdAU}D4SVb7M-x9GA z5mBsZ8PqHpI(a=>qQ&G1BVP`8yOPbmmPi0IHjf+*Fhea$ZI>P}V61s9=Jol`KQM6Z zr*LxB(lhLE)1~tI&8h0F4)!+SrksSbTg(G{+GHe&k6Y zQIs1ad}nkUU0181!2rw;kF6P5iBqUl`4r1ScfaCTn}qQvf=)=;QbuDRJ_$vTQmBVC z8gpD@Fc%RsMt(TGKM!9}UmX)uaiM{`B|8n1Q#vkApNa8@j$jl623P4N>Y@ijP6-Am z*$25NIy-ss-DxdZF=7#pT_=(_j&u+KuvLo-f}a-t30Bq?QX;+`gasC|vuLw5LtT(% z;25Mhj#Q0fXt*$%>+JJ{4HBd%d}g0rf^!5T`0bkD#=>0=_=aQ20N4=fDL25LN|{zJ z9j+xrk27eH9xZTsMb<-{tRNF^=NCTNFw>vECCaAsnMl5j*l)FXED2{t?>CWr!#1um zw&uZeCtH=_kXx_0SPqaJXTl0J;`5BGTS#eX)^(hn zKpi!oZ}=B?!%V>VOs_3OWM-vEx6v(o%1@;n$nWs+`(8?B?L(>0KSW&LZA#t*oY_my z*#(W^3Qe85ewcERS5kc8{deZ4R8^&kV33vMUM0M*>7oH_de5W$) z0I8}iJt8u$pk*!qjKFml$Ho$!%tv&$w^Mn9Uk!?2)6F8NrSi(@frNi^j%fJg&}!_& zWTqY@D)Q^O%n)HBPBw1%ymk5mLzZXg!R>(I6^jwKjUDZY=^?@#rLKMrI&Y1f+S8*{ zALCC%n9#sjKPxZszMWcWdLroAEvEQSYWzS0w~9Qi>7j~6d?&^@m#ic0V)@N*t?F2P z&K_jWdrU`LRxZa8&&g~BZwng^Q}gFC}+dz9Oi`T<<=o9^dR{j^) ze*(dYKJr(tL^X*%zf|JHBbpn)a|XbC_gkWa&}5MrzUXABzY3f z=tka7(`Mk2Oe24}Tncyv4PDN_!#6%1UA*!0eJ_t(VK^KO7|uE*4!Z3bi(-CTDk2HG z{Jiaw>dpbS3)ov3i7j2p?tXokelsJ*pNcdR8jY~pfCJ+ZTU?!K<9e%&Y6)Q?Y9Pjn zpz^K`UXQF~)<^u1*R+U!!eH=tlbgy(4UOWr`EUQhw`F}q>o=zS$LAt;lVsWeC|+(o z3eK+PyUny1s8S%XWUTOobq1ewBa^-)JRR2`8&-<}61?k$ST!|aGIr6w^O#JJXDr`f zt<^PP2etDiTafX`xubsZuqniI;^dSyFuN7Z5zJ7)O&P;jO!W?h#?^0G%)35*J9q#l z%~_L1erbpqQpyi%N+@s6SLT(rQQ;BuC0NH5G2Gsj-YlJisYKCNs6q9ld&KszcInopkPgzKWEG|dbR z=K+s&R7)WRSAYt7CKYT?bSK4Ui1eLa*6M|+&W{L3?JB_Jz%z%ymms+RKqpJ=@Ydni`SB{cWtjhNJ*Ep;-;n@FxKyz@Q z&0TrIoNz$uJ2}f_RxdlD^?}Ycyz02#wrMF5fl`ynA$=L+YVn$jR0}ptE7aAvRAwMM z1cOSSj`5>xrXme8J86VYHAn?H932BDO&$11$53$zblqNw6w?0alg0AHj@u+v5x)=d zrht)&`Nbp4HyUxeHJja#6+Pf9jY*}Q0`^zf|B?rh;OGf+ zhve1DMIl})N+Xps9nl0&&(pEM!Ui1Up3mn|^TZ_JYr!eUq5T5YDW=S*CkeZEHRc<> zH!U`NcswxK3@+g2Q|Pc}l*QkuP}O_)F5NU%sSVC4QmMz7Qgr2@3hvNb6h2u@1x9t> zkPeu@&Fy+95e<8g3r8EQ#iRrHV@=Bd;grLY=IKB;C-5Z?p<5qAdF=!#%4~#~juED| zpxGpxsoO7~t-8wD+$;P>SGp}12mp*OKlCIDH*6WAN^QJKIRh5HMqR8LBUP0OW*vQW zIyISz=WoW(xatbwU$=l*F)Iui%;wQlR8;-q>KakG2(x`sST%jS6Ds<&6c$ab2s$Hd zAH;+t%9LG9IoX)hSd15(Y6{rW8X4PI<5EUdWoHsKb{!FVTq{<7Lg`J4h(kMA#HoTI z!8v(wrZ8T;Ef=BYLF49pxq^fGt_aXT7nav#TZ3X0niSc!GO-Md?P^-`S1VuyOb(jR z<*-0bcP<~QcUE2`eU>3SvEG>UMGG18GyFtxC67 z0RP}Rdg8vYZ#LDn+NW7@^ExT6X3}ULHKpNkUEojSVs*_djzNna>aP<_?c(upg{9N` ze@aS{6*Rii-x^@q*#|M^TL;>HjdM>dR49i%ZK-oH`bG`zgq4lEv$flg`v#l3zgO89 z1ip)TYaE_MCaRvAVD{)54wc@L;;)5xeITYKJ73;Kt2OODzy#5pVt%;hxGVN9D3ma zi36&Iu%XWAJ}=(?*6wrdy_~&=py_{Mg{qr3z`P=ic+;fr6^mV%Tj|)>MbFj$U5&nH zbDQ*ym7|&`twYZgvOWF3O-X`V0C0s5&F>GC+=KomlgU`p%)q8;xbUpH0#!|^VH&wA zn?0?1YFo-XCL$&7u~qO6>~xfmG4evo1xZ)}B9&*cK=$7HYHd&(@`Kt)*lsR6yJ65@ z3$ukiXRYJ_QZNEpk#iUSQh>{DUIdaj-rwEO_m_ETv_dDJmLq*0QneY4-42|ISRrOw z@&4lH=U&17x2^Dfi))?6dhUs^QcUzjxhIV0>3QBHu%S><^SLDxp}Hazx8BszqR?ANi)`0$@{6C8{0`w**M56 zqra*;R<)TK1phtf7><@mb{q-Na_i&GZ+LN@rg;+!CjiJf%0cfX)^mfsyU!g56q{lS z^$=4d(D4S^A$5Nb|6j<```$?Arl_-TUqXP3dl6s{6EeQ_Sh-d+bz07j55bZd?VLRJ z8{+V>u+?mfR*Ee=?xdT->5DzU=(;Qb<{=-@eFx? z^geuj8%e~U8c68P{XtK=2v^8vu(}-^N1D$c>QpHq=T<-E4u9h+2rj7}Ku}h$j%QWO3@gKW)B-WtBPVu|_W)pnlN0NA(0)D#N zVuwC;c9jt*vQT@7b4C-0RRMY!OM*;rhf}OshVqBBbF%R0Oyrai6tOIUhmKf+>{IES z=kO(fHW*5h4i_OUwxq+O!nOk`U?0KaYy0Vh_9wbxT_Q_Tr!~x$uhi6|3N+Q>(tx(L z;Sz}xBYe24(BNZ5DHbs3{@R~$d<1tv;$TMw8o7BLpg)N zmEy<=D|M$pm4h|zAmfY@#8DluU}GgksVy`5H8o}M<&*$T5{C#MyLm6)avlZNIc1J& z{e!+@gNToRy%!vz2vn4OF#v3~)??R%D#)~hTU6+o6=3Xg@LM;Ffw=H65!#0A+9%$p zc;=<^kIJ&~!Nl>o9J<^PbqVXuwAKO$3jNVbh=^Iss!OvXCv_fMXomvas9b2WmHr^4 zj7m>>UNAJe^QiI1J}g+agfR8Dsj;o-$B^hk0v52IMU$0&B1@$VKZY#LHx0=&Mc*;7 z|1XTA+jy!*yL_3Lfki|nwlj=W#GC#A2p zG(~>W`32xGAZ6$riHxC%QkFIUt~g#n!_2WZ0|_B$hulLk;?5aTRHVwKPWCo1fhw{S zNp(+Azt|BbDSc5L1zcxz+S6{1oerW?+F{orp7;a_C(g+~K0xGG%{##(_%cvHK3Bd; z10_fg1P{pXT4NOG>7?rZqyne&QAl*H>Qv)}>hpas(p16zrI@KpGl^b%O85Bl%HUBq zMa^Eo!j-ovg9Euw3s2oDt#2j?EIV~TMz|DGjgh;1W}2fcBKN08z9)h%Zf1eP;4}&0 z?7dfVap9d~?XlyUT?gByw-0_P*7$G>P$H)aY@5#oSN0OLR zloPt;LR`NC{;F@ksKzX(qxi?pk~TkC=P^&Tt6d9$<=>7KiN5qN3kc;YOtWah(h_py zJ0ZLkZJJ(fPh*m==W7PZr|d6OShhMoKTiY6zCA17AEbe7cU$j@b-^m1fH{{w!evj! z2l+{yQcnSJm-}aTOOrQ>jI*@b%%q~qh6L-=<6oZGmf>8)+1jM^X{_Xs{Hv|119tMW zIpP_h8Tyh3{) z{jd4;HkT@|n+Zlvw4wfl#$!;PuxdtnI(l_2J7wK!-NpSnJ=9B{F;|^YHmONw-uWf_ zozgBUo%vDv=Gz;=;c3XGnDKSMIdB8PAwl|OSgdJ3U3e}!f0$n3KiZ^=e?z<|g7K|* zy8kCwV0^VQ0)I$H0)~hnUDof;#i7dciilKq_P!6bB=nHixr2wbG?kAZ~e=pXcgCywa5lycxq*7qwPRm-i$6Hc+= zOICe_u4&7+wM)~hh3Cw_^1EeC)EZW)_Kclfy8m^PnRJ2bMJu+qdB@D)cE@3?+r7=} zo1?#l%BF7#wW08qbtGzR%`d@h=9hA7Eyo92T$6z`ZNUAq%oBlYE=@UlO0BAfv1iy1x_N4A06jMN+klNvJaOya(@N$HSZJ|#uT=O}SErO< z{yXpT*Cb3@IN)L@&PBxn`=n|!pB3C6x27+<98k#n7544PA%(pfsJ-5na?ybJJ~M8# zg^WtwCY#ctuGq<-?GCX%xocE$+U5tPW^B(mVFs1Dl-gBnudO_cHs{r}a zni~7l^esF&caWEN62?{ko7KMz6>5V4rdC{T_EPEW*F7j9>%7(LzTI&`C zVf^tvzRp`pOm5l2a+Kr2EolX&9mI39>(O2AsZrR-s@aj~jaf>oPbSvmD%92*2eQtt zix_p3^SM1W*Ixz6}oG|(&;S`#Np|FJ5)kHFhk5I$`MC-9UJ;^y0o+*?BG79=R ziiGTlCdtL0^%J!SA&fol6tZqkRV2n*M=bl;k8~MF2_Hw=bh3@i>&&Swq^ z<%%_(z`S3TNlry7Rs(a-{gL0AOW32@Z@G>P!~caculU7M*#@XyT~1e0w5QjXpA#Ws z7G#{Mx#>t#;cv`h8>yAkF0@tfLQ#?|`j4mP8XIA!Lgnm`yngk@vyibS(*L7ijAg-O zH5H2~i3&SE@4>og5MI;7HKs7Kh1(T>$7cZ_MlrjL!b}?#@d+8k$T{kMYZmIwJLNQ1 zbDU$7(kt>?n=k=sj|w2GKt(LZ0@JH3P*tgtc@MqOjCgt{TR1jWIkQ#M(&Q_m2JnC~ zp;`@ZB}-9GA;l75`mKKXMh?p)o8oDD+leS;a(loKF+l;nNED0sV~==fRfMXNFMHxG5EAXKkmHho zm13oC_xEFoEfpJRYDYQh?;ZWwY2hA|94^Qg1ju~@AXg*4L}=$KR(2#C^YG*(K1r#e z&sE>Xah;Z$&m^r~1bqJ$Vg+dI25*#rj43|Q@8XDJ{X}GhE3NtFK%y$lL0X9=Kn}N5 zVggT05Wj0zvP`_sz$lQ7pU>*uW;j;%)t8oIksFNd8Xz$7e08u_Ke3Uu^fHz7x1)&t z>8@_KCc+Q_`=E3wW>Ppx^X+knn|hpbfX0rFqhoJt3-F^GHX#q*E(MD2cy1);YgVfVL_QfayU zRNEz($~e>+c4E15zlcWrD2Z~cq=Q^~1lsrO_J9+eju-Dn!>~u-mD*um@NMjn|Mg;> zi#*Ba5OOLDtDUkIm2|*sYEt)ne*A&usWu5Mh=W=O5m1S!C{H;EH@#<4MR(XYAFnjh z$hVT~Ha*<6Y%c*?6W8vw`vdX3-Y`dzm(tR zf4H|9-Ht3|o~n+`^;j5Wt-Qv>fc@~+021!4j~d_&iWLtO`HOF90-&BC?(+iecz)=d zvUq`GQxW^~w~0V(;P84G`7p!F4?w4QU1H0t5cFdAoF{i{N+JT!fAr6$bb6N`)*UM5!B6FmFVcl z#uCw>x&l@MI@uinNm0!M7zS=@QTd(B*5pAr42*74ocDj^npVCD)?mMH7hKvj{Oijd zydAL~JEps8m5*XzyrOh%KIvN3YY!ZdZ5d?diu2YuAO^EL8$6t_K0cmwTs2yyBdX*AW)#_@Zck??ySJT^?fKhiwOH?qUr$dfaKZ2Uy!5pL$0g<7R@k4BP+V-qY#gcM zRON3%JZ%9mskKoS%M$UYC01L0TN&OR7d~4pUCwUqsXJH|-M*p!<|_^-OQhci8ATlYkY%M^oI&eay-Lr+zrYWCJr-p6f(TXcT@# zA!9`uEi4n89=hLMOzsw|g`e6|#i%Szxl&944J8>JTG7F>NUFFmB_cJvTG)5G4C;Wr zbaxY!gBO-)mCVVd`IwFseWc(@@v{D_JYx_hTx5Hv&XyWh&Dk%=1bYL1v_PM`pxl=(>j zr|`33U}{Xp4NB39PvehVvAP}YErG8#x4=>0r@16~>bj(!b-g|#nL^|5f6BI*=KrS! z@Z1TLACnOC3PfB7!Q<5<6r1GyV59{MMZQVjN$u^L#MG-5u#nSu=wyMZ80GM3 zf7u_sL%S#_ayWIG=)!E+xrZ{GWTntQ6nAJdDg7TQ-MU3-B^AtisYZDu^YWKJMmDtg!?$LI3X2s_y{kZA}82JS7x@OmKX zw3O|Mjozgk1|$=sdO<@FVP5=G7er4?%wE8d{%vnUY~&ISw%ZuX2N4s$Rj~m%48vYm zsG{E-l7RvS{Fol?y8$aLAozgI+cSe}Q}MpR7lmpp07&zA`Kj}Vt}DkxWc7VY|39~h z`Pz^y<*}M2L`^D>!IFCxqN$20(6N&DS%UHjmnG)g^$KrXZ=(0*xwWS7_4jA?TYm>X zAftY_V2Fqnr9d^`sJ_de!`#`ZdELY4i$BOdnaPsa;@44;)8R6?;>E8?-O~{%WtqDG zkB)CoAzl!*0dU)WzoX?hHRh}{NEPj~O*tD#jPtH#H8UA6U=euly&-S~XnG1K0OF{kmF zaBzG1>d9`nbm*msb|hEb_nl|=iCm5Ubx3jOR(p&7@zI34-d(W4#zSP=&g6P!g@~L8 zne}<3l<@ueA50<1y~q!TE~7)$v1GU30d186xtSt@M6Qph)i;J4CpzzM9*$R#6P;k1 za@gSafaycs#NLVMzX+Sg>m_Ady+6471U9cm76t}Ari#SghAh|m8|4vrz1fBO^xA&W z??0M+aPl8BBYnK+6HEOU=+S=me4phg*6SuBRM^TNv_Yf8)q#GgrUZ;hpQBo#Xft0I zh^4eZRP7m65(E2M6EoQ)U5Rpx=dPfs;e&i31_?V} z>b&gOEwL;A33wDRA~4FfRWkqxlR*ZC$(scN z&v+=|JDxDvk8$3GV6^q3jJlWMIN%5E8Qk)uycv-lI@Y*7x`;Oo^csAB*@;cl9)?f0 zQBg|yDu+a15n3!wmz~Iz+g^N3@^QnJgS1(^$klDJ8G=G#Nyhn**?xK)^~8irCwrh& zLAeuVdzC}jP!m}91OBC;`l7nll`T9CS^W<|smaENzsU&#^|%-=Up~|`oW=`Lga8>h zBtx_6nJIY{d;R||A{A*n`5PG%%m_qUm0FHT#{z^GoYTk8@5QVZ^~#SsR`v=Km|iH| z1BEZU3_5a@K}u~C2jdmq%N!UHqmlVS7u^FKGM{N>IW4Y&yyvT$#Il~AzeO_`2y=9z zRzaUqF*4B*5u2~T2mL*Z^uNtTC%tq{YUr+QAQ7!I+VFausP$}NOeAc)HJ z@da_h_sb%yCty=Baus2Ix~Vy|DHy97^BjD(7>XPDs=+Mg_w-^{haLkFj1Czz;0w&j z>5{^dwQ1nv)5wdLjMdl;{b?7171}1v#F*148uDwFEP+X`yqE=fbr5?ad4}9eLQzp$ zkt(w9GNoA_BL!Io86`085sDRCOrV5?E06yc^~T+_tI`d_T9|5Ryi6@H)Le?EEZZpqPMQH1|7o$`i=Y`zc^Hm&Tw2f zNvTsRtR#&XDMG2qp5<$+pByMhloWBPB~{V((zc>tIE%JqDn_}_MD`9Cbo73TuuJ=o zRXB@qb)<-FSk`bqDP?3P(_4>}QY?P@=S^QbghT#CN~aUR%xnC&1je0$azDvgi#494 z9>(FWsdiX?wIvo?hC)v-FJ@BW>nnBl)ufM%j)T^K1bmi-0#ldCoj}{)zP~e_FwZ^@ zV5D}gX(4f9^_KIj#G@*11y~$$hjlXBv`JjTDO#!M=pRr`9Uyh{`F?67CL=v`^{s~mGA$5V6|{9w#I`M0X)ECj9NG)^3xqA-&J@{F za+6S6Q{-X?O5oZklNB}zBFAHt#-?o~ke~L`0%<}H`7`H7XxVjBqU7h7ei?r_gsEL@ zK87=4xHu=wr38jSd6XI*Fn>2CPSWKx@gfhil;O!axHxrUW$~ZZ4n+-#k0Qs%+s=}7 z0(`uO5gr{Nf<*3pXHR<;W~Z>1khRK=l<&rFEl2Jt-xvhM^7n67wlEAh$n53rv)N~s z@;7;a#dalcQDN(@ySM{LW-zA6DEl4jo#Y4+ZRiqieA)qx96N07@25^&tbcAxOsmC} zRi8rzMlbUQ?gE)4Gn2a8F)}-^HmWFz2;4(z7Y8xn`t^dQc49x4Z?3+Ta6S9EbpygS z6dpcr|6v(OHeaeMB|wE-FIcL}4RIg=62?z;UI#sGifPG3^IZj)pYJuBh6en4i?a_E zBg2%wP7w>5r00X+(v(Uv=$^)pM{h^8r*0$mxA!yO@6y7^dC#9_s;w=bH;yw8rv?zv zAFFDe>?RHAwxJQtpVdv?GgmXQXhrp>vIr6!hqoBvgMkM;jo=^Y0w z+`5@xE4p5y_SEwOskOpmns3`HsA}v6kRRM{gK=3y_7l59sj^EqcLCG>-iN`n*==lA zlViWd;Bv?0YT?J0V1|<@a~U!SW;JeK9lpe!-u*sQS;~CiQ(_c_Q6J1 zy5wXoy`DyOFY6xXR0jK)FQxhu8@~7K4r1c{4$nnux&dDM0=Y>~BfWEu@IN5kg4zK6V>Nk(Ozvr)mo}czp=dDZPXxRpy{mg`6GZS_p zCcydX1vl^6YRKq*TD>AV|UQ#7TuN= z5Qf|&+_ig#05QBZYQU!5!{r^QVpJ~-5pP%_c3d<*7-emTDb>)-f~(Xe4QF){cljPs zB+k|;%=|o2fHZYJkfmz~=8xI!B`oE4am`4Lm#2Mvu!SY*7bXVeAK=ZyN0+V(ea{qh zbmsQn!`(Dg+taVVbLHhrm*n3n%#8N{@lwH_b5Y~-aqkOn7=QV2VD6wjJO?zE#kbkz zOf=(X8CKjPN##u2>T?$VB>qrad?2{u+%ft!F_x~Nz!$jHZI3bVHmp%I+^blE5u&D+ zsp2qQ@VvB{w@gU5PdB?V@Gp#$@=KS!=+bFs18@q%nu9gZR;!FA9j(~hC{tq^DkI}2 z#(nPKtvk*)*P2nFssxvh!-Zw26J#72Q@{^mP*%j35NOj@U=53LFk%_f`}P4nr%csL zpLa;U`1=-`cJtQ!6jafN;usA8I`f;tI_5Jx>j&{Ikh~N{gd_u$5_U`(4W>-V9}H44 z8LRlZypX|%P6C1?rFb0%FB_X^kbm4qL376F=p)6sIbp^1FD`srB@#!_mP+x9*$y%& zDGjPJ#VHjuFNs_5Mw_85F~2FoWf*!bq2==*F=$uqCN*|eL(*iE(wAQEz@or9HHR9) z?(F7w;bP$aXf_F`o6{o70fWQ_`MJ+y7t5t?k@4re__HQT_v8$%rZ7#0Jk7YUM-(xM zyOlPaXQAVIS(&1lLn*Fhs1&S<^->m;(0mTZQo1fm?#Q$fHm2?%3842!tt2J2-7Ae0 zwg~U+bP%tc*46kG3!2a(_^4* z-=01qp?+}vSPz%QL;UfOpZ_i}N*URBQOi04BSa(VCK5`F8YQ&yoW(#=q6q}A5sL!e zJeTzy?JkMx$e>Tzq?wpTud|O?Agt@_l=|_~6xVdcQ=r zdZAX;ksLY#>ue!)$aPTITdpkoX>T7D^6;m0f*1z&*&Ua~aI9NXq(}2WRpK}qnk|!6 zw|c}bV(z(@jo+7aPACQ|RSH24?oN@LNp8JpKjy2aW!D$K7w)ksZL1d#Ra+QfMnNvR zx~VEDxVV-H6h7eX=u+SNlfMmm$3LQ~Qx>Y8{)%3s$@T+P?xDSogx3=l|7f z1+4x6&CUPsJUP!|WBQhhiR%VhKC2((9djSLKcGc^-d{|tBfzg^Z=&Fs+2-)laRsHP zbN$AJ?fLj0COUeMfdAdI=hX2eXTzkp-9?aQQj(X=gZ7l#MdJI)vVpja6k=ipocug_ zRQXe--G<=X#dGdV1Bu$jum09wl`J^qqC#8zMVJzO7b z6`q4bLyT(9d}m3-Z+mm`kou<`i|NwRiQ1AMBVJ|;a22IS4lshM9_@B3wXo=DN)S=9 zRtHl0xJ2z!`vM@_f!$~B?kpok-apbC?PDQH!ZqdfncL8_kjtaXC+(iV?Vc|}erw#^ zsX#S(dwd+LHpBONHWpO)I17a8ZIj7S=mRn{Um z*#nCqb%D}(x9Z3tt9YWtr`&7loKBdaN{)E#NiyEz%Cb6OVb`@z2Mu4$n^v>swtzt` zbMM*OsS)cZQhaG(r9goB`Fn-EK|}6WKKN49-|^7|`ApgA0K?6AX&yC%z`#PEvp@wZ z3e_yWxE1YEdXz00`zxbt{<*8bXlZE4Y}=27C>*K?rLx)Go}sZ1R2ZPRENFUbHkRMH z->5vztK@jqKc?R?K=YqQ3S&>tfnwEDffi8k`#IeE(#&OGZTg{|fmv5%gG@srM($EH$|g( zGG}*tgJK>HM&yD@C@U2Ybqw9J3zak<8*LnCKO-zKt$tY&9M zGA>u2Oh$`U!Ba*M-7o`?=O|PpqxU!iV0R}m3d+87=1sIvY(-N_dPJA_jJ(nMP{nMy zYiQF&v_gLd8Jo@f!XM5m-MsZ8iPZ@j>x+q>#!${aW&-gyQ>Uop>jf7mCHW{QUS`2U z+h#igD`-aSdHKk+m5}sw23V;q6!6ZF#4)Qnb9i#wMy?d>$?sWzNhd1IDB|{~pJpmz z_;&S8g~l=E752pIv)cu-$ckK+;g>O-`yLgHR}Ip@Qcf(u$<#AREE)l3AoiN*skkjm zZUS=_5#fZo7FVnIoUd~>Xt4#(=iYwbHOm)Zn=(ey`$fzZD8F{^A*uO?ds0XZhtMfT zkt!Wk5m&2a1T+g$$SkhXq_H^z1@0VrvWPT& zvg}(TCdFsrHzhVUY{umEuODG1ycDVl0;IfPMiH6nvy)&EEM>0}H${36y5nfvN*aDY z^Jh*XA41E)hN9y{P*E|4OE6L>(u8hNA;o-VYSr#Vu;%jS9|rxT|14apPthc;tN_&| zjlz%;SE~S*^1&{Et8-QZX)3_T)wYm-spmYJuX0eE=<|n(Z*4&%M+t)|=|6wSpoB_E z+AE>a@z&z#b!3isN%&3GeD^5W<>x1*Z_OUJtPE)%>-untrHU~QM*D?dZrN*z8%AlwSRxg<}dkc{>`41N>0p2U>E#$x3H^r zk_t>cMp~>D3h^t^h^d3cR?IyWdV1C^daaBc%d{GQ8853+s5N7YxYUhMn+Mn=zX zDAskvc%S-yGe1Ng{^#W|^*$x;d%5csP(V@R{^Az4zG7^?L36IHq&`LDU#{>BZt%&;%h0Ft-sebmw#D;<1- zf*g-t)Vf?K#Xb6{*dhq|ZEKV`o2-8yQRf%x%ZvUOP06# zszrzvY?) zY`d-&n#HVUjw6L=vrB|Y3V5I(2!*U9lwWcda%G9L6On%<%I6Tar%3hx)7vb{H2NY0 z)=4Z9`}2Bp4NKb|;M#lO2auL)-BU;D&>m3=rb)WmJ4jS<&>v!TWp!^;%*7_(ZRX+X z(4Vw!_ZeUJC`u6{(=nu6`s_OGsnf2H>JPre#{fXDnHinvbczfDO;@k^Rx&@lupPT~ z_s4)KHF6c-!>6^eYIIC;ECb|E_@r6HYSq);xj$(1Iz*?OqUEyr{eUj8pvRjKz#Vnx8n6Q^Gvq2P`z|;6? z7{x#{ucUbtI~baTh6~N*8;=D-NJ?1QQdCIQN`^MTPcN@HnTq)MlofR;47g3%T&Wl_!rYuJx;vURx&OsFW5>(^zWNCIZ`);Y}nfX8ScnQnrf zo*5AAfS=;f6MbAbN`L}DbxOh`F>A?%ryhfmiyg+kj?WLbA$f=wHZOML|#25m*;9*hXN+GAGE*H~x7S?<=;*IHlzRT@q$WG+Rz^hyQ zSZIu1C*<{dHh#b8-rYt}FIdH!8^S=lJ4aCZQKPAj-|&}Ru4dL(oUQ`nEk!BJh(SMe zWBIGZBNi!mnMup@NA$`kjOB-r=^LVI)Cz_e#aT~jn|Wk40-@B8c8r5ow^}Dg?bc3F zNjnS9;=WPT?rg0{MGBnOw`$7p_+J6Wc8nBq0!9}BNC(KicRre$!o$ZNd>Cfvgb3sY zQOGJDig41^efKFQSm2eSS%nl^Sy|yw0`)8boEWl_7NHiZK^v}cM;8A(z9ttebe0?}%_&16 z6+4W9$%#=eiJh#0ZNX>yJ-lRAIPuFcLBd=l0X3Vf!Rh~_=`6#Ve&4YFqmgb#cMJh3 z$k4x|>UJrhcmPU5yv*;ICT#0vwW@DL`qoJ+2Ca4Uas8UqK_hN3IFgG;L zK*8%LP!C2>yt9+Bcl#5&mV#?BvyGrx%dYSXDM*5MiH_!9C1qB1ruH-~|Jz%W#pucV z`s9)5(X%!y5F@#w0#M$SLOW%6{ZAjd24o;e^-|RR*VVLl%U|QKnbW(h9f`IhR`Z*- zy+PMbntId;TaMBG9-vp;Eq`{dKE6)4N4y~MG@m7ab|7F~V@T(v_E}V_y^Pd5Qkazq zD)f>*y_zq03x*sPqRpUa*{wPQ)OJ(WpgLPNcCN)hSb5#cwP5j`+a*fd(}%HO;Uoi4 z*H}PJ>23DG=j}r~!nb^R3kX`Wy_La&NE)UeOAm}Sdp4aG=A7+oXI8ccwH%OSv^A;m zxVBh-Vvp&$6M{838CN0|2YT4Qfc_J$)`!)DnEMg$l!~i&S%i@-sAb|FdjLA9>rh*p)caL$~29p-86iS-v@+Er%Kv0UnU? z9VgY}s!x9+F^*C1B{gF@e>B_RdeL4xdm4ZWZmhmJeS&W-5$_!vPehz8rX)zHvF*k{I8wrM5Qa8M-H0GiF0E|-qU?A zKp7l_czS|`w>(Um98CD%NcRMYSihk@>@xLiKj(dedSJ6ND;iI=6UZn#s#!Tc$@-m> zQU93vS3y5)D|^K)`1nw*O5xo96^>{+Oj)>r1|2myEohKJxkKo6AZ%WMjm9Ki?0x_a8c-6ZQ{Z(c+#H3gmlSQ|QxjgD3CEI`z89k}} zxPUzM`4Fp7h`OF+EM-nbb0&ygsRc3_i&E7)?;>E9tNOHE>VvyNY#i%wB~l98sS-!k zI+Q44P^#v(p~X)vd&JYrRlZ@N*L2LDHCJk*iHD)N2(yuogZ6=|s)ii^gDG@UU$~R{ zRtmShnlw0KQv~xO@bDx!hX&h(jI!`k6p{>-GYHcw+bVyo~sOJwNOtA=3Zi^t#>3<#PvzCcrG_*>4#(b>HV1A z#5}%gnM#;K5`EI^biIZH`eQo(U$GaYd)1653`*+j0YOl5j~DWB2_I2gYWr%ySznAb zLV67+oX0PS|L+Be(5Bbo@UP6BnSR3!Jk#|GB}d8LfS-+C*tM_6h>h}a_g&B-#KqC@ zIYvDv8^KukX(;nuI5+uNtcwvECh-`XGP80x?=Eh+E~Ge_a}Z?9T&7SBia9n-WG5jb z5Z%E2#SQ!~euNtXwQTHJXaB9WUayI?XnqSAUe)WW#*$~EmusrkBHYtLh6(lfqKq4_ zo-j)c>cZRCrgmUWTxg;p3r`qpmUGf{mtHf`@eqIWLlVLD-W;dwIBh8Ew$1V<4h$){ zB>_M;FhyTbTI zW7tPY$5xqw^TN1PQC+#1b)}W%OD?;56gqHc`iE~CrUI?YXln59-1`kv6Bsh`7oi?0X#dfS`T7pO4}XFG_re+Jl}b(OBTeO8g+sGpz}tig5rkLF`iB`J_Zzjp;qelQrnPnTN*slstF7S|_n7mC_* z4P+yX3PXna{%j!{|!EYA?N6dt1!Uy#%d$jHw z`$17g4(N<#C30}sxj6w895~dqCC%w8`jR}Jq%T^@o^J&WF3OCFWcd3@fUISF+1sP;G|kyTlaB*q)xN10(2JqW4*e;d>k5q#EGF4^ zr(TXrM&bM-GGmJgNjTI2fh`>1k=PNe3zG}4oxo1_gV^x(&_H?PZT5gC(0Qxb{5!qN z-$gu(OblM%+yXtn6$4OT)g|3uvyCOgiJzQwB?S{VxqS*NL3Ls;NfL^F7kytj?ULmD zkC8$1W)}s9Nj5H>1nlYS14TX|FuA1mS0kS zI2qv+jUws5K4x-1RnRhi+&)HSm5=;(L7kG8A9fFZmjAIZeMnz2T;QtFmTkxUG%J)9 z84??K<*xfSFWi>;*Yhy$pX-b_u=Kp}=Vj6Sdav6wm_YUG5d@4&#kzakEqYxlMlT{A z$w$4?I~Q{P7i?!r`C~|K{>s1>I^#X-L zgO~}_PRaRHZ$Ti6eUpt!Qr!t+ZX=>*U|QwlPO@>&*Oz;L_U9?BY|^nV^;q-QKW7;U zvRLYqisf3*?BL+HA9CU|xe(BD_L&tAb-P8#Cz ziVGY?G6}|7xN-?8;ZOX!<`6{^82If^`&Aorj<(7mCK!^02UqSJD1Cu)Ec^-DsKEa% zk_iBz8l-_fW$Y|>-l);i6t8OrQ5iDB{-roMmh?}E6g~NbT;`@P^Y11EF5yCgwUil* zf(#rBwaO$pKi&BqTw*`s>!kSbH>*P-kp2{4j$Y8#S-xI>(0)*?oB$6PV5Z|J@9I>S zMb{dsf)|fZlKyEH4`}*y(qy+HqAKypM$;^9zDNPqsX_&-62P>-up0>JCn=sS)P=SQ zkLNFh@B%~?cOUMUKUe3fna%q1K>}cKdQE(*23*YYD92*JTdb|1;fsrx-JNUa&rg3P zg0}<{v7R9=P)!dd$dr=?y+wt?3bLCfjLsHEmEg(-z|0B1 zMZ!d?XCdGJrMLkC1jM*W+e};n{C<8H8vgW6Sa0%#>HKz#64_n4Dz0H%>JzPtmu1mF zByp)qFGQ$eu(79&!ns7gNKOQG0UD--Y4B72+0#$iH!IFeIZf?YCI0-F$?P1(3vC~a z0Ua}0$+4ey&luqy08y^;!6acFtiG!;q?~J?0wV;r7cR=WD`y1*dPH%(m>@C|HR|tCD-3U6+cuSMWD!!hUr11q%>+3EA(P%n>sL8F-;$# z_i;9N+#uVDfIVtFoA$q2gS9f9*%ecIv06-(3sNM1ldDun)S4ni3^VJ870dg=_G@;Q zIIB-pJpU>=$^~mg&14n~?H&@~`!Q5{a{Hu^#WMEOX)Ke@jP83j3+X~fN0H(R)yuqcV*y+baw z%2ZYh63jKa(eB!oG-CBgIda1-O!1BxHN^9OK3t7Zt8TW*z6$gbv6AGOgLt(Zx0%{w z_?>?8nW{0^?{9QMc=B5J0yMWcUcFlUEtT!@Y|($!zU8wc;50GA7`n=&vmYWI9NNf& z*>=DdyhTd9ixu=0$`EC#O6dn^Tk`Yh`s1-Hs5~{)&J`G&8sX#5HcG66P|A6np1(WXR+8VaUaYYoh!-fsoQPy@r-Tl+`pV z9%-oJy3r>eorHeEhT7j6iKP_0$~P06LZEr(%s78Ibz> zn=#1-hkTrS$^SKF=2h$E2V&sa@dvfVcWw)W3xUU)9I5|ahThtsP&q+URoe7P>V@mg z&9!?@1^Db(=%tpX+{cpyCk)!CLB6HTG_pA?pnKQOS`A2QWwjcV%Tk}$E!kQpic!tz zLUEY-GUb{;4=*>=q9wt|`P=b^3j^sZs)y}0rsMPf1Rz3R)ZcvA4ifg*X#dn z#y;RySJE zZP=JgB2WY{H|22Z)}nJ+0oD08yls$uI^I%$8I8X#dQm=l@ag{kbuFP0txTJ2ae#iv zr4(FGt1xr=l2-fJZt47?eJQ6y`1r(d{f`_S-x=v|el=Mh zk(2eji-WO*WtVSX-wJ;@xBVCJM(^)vmio-x(~7&$G&Nn_JC$CK1~D*hpVz!DP4=6e z=b5!UUt>SBw?XU2TK+3>qj96ys{di)ie?(X$BAuIY5zrreH%GFRg`u#VpWD>%A+qS zf&Dw?j~9sVS;LhSBW$aFZ?Dd;=XmT5LL~WbAYKp!2^2>eALV6_SXzv2i4*RYCqG zt#GgIVVaqAL#o;W?;5(Xt@K>t=OR{OY)$5+1?ma`gZ?1gIq-0TXpZ9;csXRIM zsJt9`$fLo?sB1F#@aMP!Y*-rZ6k zkR+E5#PT#}m5wDQy8s5%b%5rHH({~}nvj#W6)QUlfalbGq=I{adDKL!x<|8gu zzo$!4_6&rDWUN?+JA>gnKV}IkiEnPwX}pS6*R!FpTh^n4&Zu77VV73lz=z7lIk16y z7NX}gR3nr>W&iMi;Pa0x64McY&r%W^RJHW#d3@xo6rF`@7l#2R8(D-+nT zaqgqlx#qQR(#~RgIG0^Ss33P<$jqx3k-2;I*fEYv5WU)0>G{6;WGNR>Ys+rXj zZtqZl(+;LQk9Z zZO&hWxoE8&;h5l8w{IWpJS1DJj@{88zNG+}O$(68tY6C64TLdqopcH|TWd>U^S=qZ z9C@xWbc%m-df$QZ>R9TZVwwL7YA=1w7oVpFJ6!@BcmwzP^3Pw&yXfOmn1o`eoVTtm ztOUIa`@DwN>5g=LjQ;Cm{vDtqWkPjTL?zR3!jK4QxO=Ip$Jadgxl>17<@lgTuM{87 za{+#4%fKgn=3{xYZphNlg_^Ga3$mG4Noe|0!+*P(p1#QMf<*T}8tpM91et(EV+;AI z+@oxdg4FFT|6O_Op?qx9+y8#IO7jC5YZJvta5atpCxVbzTV?b5cVu!oPj&2tfiwgL zeFkYl?iO}P44!WDSf}Ev`Ru8>{T4C&%3&Dl& z-DyCbVc)jL9EWOxE`sjAC*PG@7*pmZ`VRc;-b1ZfKVLr4e)MM&$v1$t(utdec5sx^ zRUVcQ!w{uyQWB}eg@M-DVz~g}ERW`JVE<6NS`vK>H1d_vD?-hT=t0aIz=2Ba zg|T>|!W7^j4zGTXs3dEa2UcdMLWAOp^$`Rr&#EA2NLfaQt@K?5rSDb*GaUHfW8xk1F# zx%ZkSN5XV>2g*Y`O?CuOq~a(3M0dh+0v4~YC;J7g5KZIEBqDHDV3!G3734fJkp}ex z-+3$y5R-$L);Le^VzqPccY%RZwD%Od-dR9 zK~#)q$$0-~3Lh4jl|ov<#*1(mZ82U2v#J_%G}u{L*HD$U94dlh(PliflPl4Erc=uX zmC|v;`;||D7G2CiMFR)hjD`lPh9UC)Qz<%R<05hNwqRCUu456h2nE`Beh5!(9FU}Q z)XJGa$ykR1dk2G5e*OohENLyufaXe(RjICk+r0WQO~Qa*S(zeXP9;hxZvWzhh*3!) zDqaNJ$^$d!Ix;R0(8UWG=SgHm*Qv0f$hADL$tZ?%_!`k^Y>erlNwVw4XnT}Q;Q&o=#LqiJ)C%UNM^&)JNULahNQ zIenRs&Dzmgihwp5qipC4*xUwG1LcFUyZ|z32wn+4i9<6wlyzF=l+CYUzXa2oI|I!2 zB@Z?qs3v8Dvv$DB76xxHbvId)o&!lHsHEeEgswC(SiW}Ok0GVo=nEP;vK(c}V@=GC zm5&glFiF=jLV8`ubC-O@6w{$hm|BbE`I4w47QNL7ob40RCl&J}q0}_icZx4!Rlqec z)-NNCQgGHRCRecgJ#2xYraFsP!s1V1IvkUo8`ayP`kibG!%&~x6y2HT2)8Lj`T(vc zQ^$gN*SDFj;HYT0K);NJ1ohn=CjBHl8!gouD}9L42=6|jsZg~fsPrYmzZjb}a6a_t ztyGfQ>23ZjVnV&N>K3fDFDzw(o7?p1i^!JvoPqDIILhrP1p_*cM<9<1JOg{ zYFK%}6`$n*A&{lbTJ?P5q=h*+&4IA-%r?50jt5^nZoN z$V+Gbn<1cu;hU*#)|r_U%#t_eXXkB98~)PVt+L{|i^Se+u$vfPhueVDl0f|~;Blyn zOX!ksj9A1(FN08*y|rtryG4VaUia+_j$Q|4+d;v$GyFy|;bT{gAMb+?sZGv3%)GlN z_0{WJ+0xbP%{rJ&CsCRA-A3<2b}q826AWDgzA@ac`cAYa8$=HVt8*|AFTk#9Hs%fv z-zn>GaHMM4gFrdO>>zWzCuHi(X9ESLiKd16Md#<66~3ai6raZJ$BQdE=kEGUVrbq` z_>5>ucI(4Ufj&>E@sEV}Af({ke@&l&C_1z*#KU)ePBL2YJI>I>J=Bv_{QWjbpI^hI z4=W@RwxmmR>LshQPJc6(eqwi@vVcJHeZ3D~?cSnB{3eij|AhASu;`mz#-E`1i~3ib zRqb|f-ttm&L_=|*XP|-oHp!&J7^vUn!^Yu&s;A$1#PKURfs8wpFfhHuB(68BzoFjnAd)yV_jmZHj{0v;ou6h zLO*#QOleY5-a8^Yy#@iE9aLk?_!T~4YEM)Db|a#Y>DhbrE2Sfw5eEp1Ta=-|rrQLZx(HFOT97@c=sU6j`1evz^$vi$8AKc!er2$L_Nd zPbkS2wT;xQl)mE$4dVNZ_?sSAWGI~z^y0=)KEiFX~5ymkK4T*=#;V_EW)sJxAH z{}bP+87?gD!%M`7FNa@o~OWhudY1v#J%#&hYhb` z+0eDKzfZlLqVQy{s-Y(PZ;}>DZSC%c;REYsBuVC;Qvb<& zeB%W~4t@UfGGA_}=6jO0b_dP-SO`r9)(?d3SP8!b0eL!An~@$B7C$D0EK%=I*VbRV zvW3&?mBGQ$`3qVg%AmC?RyctSr@)|2gnfXU`jdlSa8jejg|F*>(Q~#tmhs|aMxZQ1 zDm633iV7d+?)IA}wHh7i9U;l_DeWW%wA0ivGj74Zw}A%H%$9Iw?1|7GwqYd?a@yFW zlnpjCO<{L#jrccsP>a{EzBmndS<2!2DcYFrfEA4ap{EXGBdKpxfdQVfBnmgv6%(xt z4wZ56+=`m)5Q`NP7?063&oL3(ec#!EZz8@0@b>3aq5ZWqq;~^z9H5m;WAdV1aO*!9 zJiqRx2F}-AqiJ{_G^()53jMKirqeKVm2_ z`ixo0Hitp1Ow#GIm{@T%@o4s3l4y-Q4u4^C=#-u;Di)M@Q9?>zPFYH#loFJfajBhZpWkFQo(GfpQvAW90fuaZy6THcQMY$BVC0e3D3&*6h2I z^Xs~C?Z4AwPgt-mfcdEs+I2$Ct9Asi%iNmU_}G+<70(9&U&bS1cU}-Ls4s_F0Kj-y z(Q`|ic6`I^sBJR{BdJtIu@D-naGM;rN=1=eidrTl07uz~B`uYxF07NvD0JCXZ=X05 za{=t}aXffcC02h~w4~bvJML`%_?UZl*NEp(cec4NF@V9;YJ%UBF=T^`&Tg6m8K?p~ zn3jB?P23HAg6&Gi3lz~Mg5LrwZ2Onx666R77Iu!#bGU0WI7=F2Sh+?7logsLbFumK z?0Qx(RqcGEljGfFH!9<0I3lj){I+>Wb2`hv5bklQ!+vxwUuv*0M9OlUjdG--*?5(G zA{FY$A*V$$tYl735<$nb+4;_Z9VZI7fL?4JmPVr=NklQ8gIHMw{Xo_X_P+@LuMHbr zksD^*Vil$2AFy^F<05y~_)5nI<{(rT!++^Nt{ zd2$C)h?KFbK@t)@;rN8=qrtIGi#Hhr)?iwn`N0|gUDFGf-dTE}-0(Ua_10 z7H0u*z7L}S@HyaVI#LKGb+(_@!AvM`k-oh! zFVgxPr>dX6ks$(3W&(y_tZrmKWTopuc-RF3XB-p@^&~_i%1F3@MhhsQ9Jiw}JlD|G zT$>`WC`I;p*SnI30#$}c(EpK7LdI?hLLY4UB5fbLCeaj7d+!ze7+J{NKs2-Ka@fPI z*^GN#aBJzGuGs1m3V)WcA{EqLhCkmc|IF`)fzMuJ8)WI;zv+2p@AnU{ek@urh_b^M zK&CT|G7U&x78AV)6GayC%8Q7|j9**ZGeNrZMCATB6LpAn-0Gv+z1ykqdr>iB(Jt#; z`e6c>n89|`UMleR=Is0`>SvGg5d3CnFSu245Ii>2x#@Fer;`iI{qFE4G2z~9#{Qa3 z@c8CTip!{|(rcO;OY?RwXCVYtVq70bK7U0&`CtE1928=55%AM%UhDS}&UCkZ+QLOg zrY+VV@=-`}L&rKohV@fGV)y^eK6U zDyChe)17rfqO*M^EvnyvkUhq7FeTHG6$;Qrl(RoEG*k-F~|Mxs(m@-;qOpk z_39u(Y9i6~(-#fD+}!v;7V#XKwLpPq@PN&VF()!0#QgN;r6zP-l|Kv1{-T$dX-dsz zj*6viJ@qDg*7mC46$EuUQk>H*7E?qR8#WZN#`bkRJ;?@v_EzWldR79=f4fMSO&R^5 zj63{y0FF(iQJYQ949zw&jKG;l2z34BM}Cp>Nd%AMB60J|%|;5dj9(&$x`uYhhCJ^P z&3aOB_GO2soo}BL!C@t-J0Dp2%>@~z%Q7PqrLMzx|AykH9}zioH~T0}7*HE99OZO!YVk)H)hG(-4t(U`ECjr^+FhvT6bq*-hxCGp!Uu~~-oVJ#*uZTE|HAro@AjaSNo zGCws3%k~G4%wRjG8xA~W?6 zsVMmZA=(-|3N7PjFNh9p#x)dhwjdy55y{bF)vZwCR)k^A`LzkKa-l_mcNX4;{xMQbF!T3M9H$Wrf!3&ZGN zsUde>6MZjjzM#b>=K?r?|4L6&h0bxd>kI-nStT%EmD5X7lPl|5RRt0qKu|R_({QS4 zuKvVqqRIk+{Dwk^0MCyhr%!=X^B1?svHTCRf_F-WSu_1&Bs;uwd^*JO> zUQ?EGe2s<91>vE|N8+RZp-v`MBq05rMyJ^hrcF2g<%Y_F;`0_?w81fuhy*ZW!9ISp zN{Ba(Bf*IE8)P)gkB`!&9EM{akN|ZYq8;rH#Tfcj{w^g|eeocTgZjEU+{bT_qhlFC zavWCj*kr~3D2E%+l>f0F<`dz}gH^ZN;6eRF>Lr%*KE$Sn&tD}}-xs2+#a$yQ`rWeQ zgH~_YUR*xLX%or55O>O8l`AJi8mlE0R`5$0FNm+$e-(INryr8!Q2zO)@`(j*VI|jV z&4kI<7G#V5bbhs5?8=6cPHe{T*P@s~m8YbwADAs0wRGQ7ZvB`i%rrLCqf8jH=Z*1Y zn>h*6BXa;c5@)`Br=oAaTmJ@~gA+{mi?DD2lQ9`T(CecOblkdG0cr$CFBP$`#wrlZ zAzevvc+r*4jWzLY^YSHnF)_3cn`l1jBK+-2VdWk^QKW4?{n46t*EKhXj?hJo_U`M- zL|D5HF4{uSLPqiJ_{q`8WO7m1VvwxuwTFIx$+S_dEQf{$5IV5)Fn9Z$adLso!f^{~ zGg>S=xp;|{UvCcz-uWi2;p3<7&M^ULtmAMMYck*VFK|1(6cv;$>BXxRf42>G_)V0( zeY2_DH-90~6Z+Z%(5>WO!0SxpI?|fAMiOxH52MJOsm+y5pJWn=XDyQ0zVToaLmlyc zJ<0Saxb8Z$sH-P5`R2xZv9exL<{J%)*`3OSZ_WTplng0tEb`CWaCvRbhWY9m<8FwE zg^Vd5C=(?9EVk7B^(u>Rk8Nd5aQQz4o&ATTki@NkcVQ_(FHgH`zg|)E05QWZd^Dw2 z&nDjT0bejag#Y&u-qa>VKrZ{i;56O;31^_}N%lwjeC`M`rGuJWtRwv5ClfV7AM$W~ z*E^uh=*a&hrMUcLf($Wks>AH*rVI304&k~W+}QEks7nS(zL!6}v<{^)UD%bPz5}bT zbr_wkP=_l7$aJioo!UT+zv$5`$B9vw9Bs6%5QC&&-flc#%3I89HCy+NOuqThOcKT8 zxvUy`et8+p^wrNNX?}3v;%_o@5}XXR=HL73z0X7F*5}+9e@6IkKj>|zojFy~g6Deb z@9V&KYf?qtjnjL->JXrgrL(zjo0TUu-Ho!0buXt+crT1X1+|(DVh?rllZq!rD5$kNme6rp;h}~l+b3(XD!8a%;hn!5IjdKl{xVxW zx{bM%54$@ltFANL*rAulG_h;viPSiS>q`6!!i&QiYWM%{wkD~4?^RB zWbvaLx_CBtHG5dZR2(SiBSiVcnh8W7&fb?aeJ&irFJA+_iUCSdRaccUF{BAno7j3j z+R1umv$jy>^E;1k1^)cU2=7y!Dn(G%yNPwSPn6c|adDW?g^aQ-FAOw|@n1FZo=&sG zi}()xJW3gh2HBa?fvIUy*_K($t+@^5s-tviq4d!jB8d@wo<5-_$6`3kUhIRT#S90_ z($zvzJYxLkPwA9*-Fr?_8=Z*JQg8J!=uIvRxFvYJ@ueFu}bF1HJX{k3Xvql%~gA7*1W;Z8Ufqa7r^j|Iew8E8{FVt=%v9hxzhB6u(fet$PX^Z15AlpS%qHFBZW$~2CvMB z4hUHO9LZ6HBfy^B`NAeQJz;^7S4+iVT?Y;-@n3}MO^IP2fLWTLe6ip zm?>11*{#gj9m6kyiH)4zSrRhu@9IY0sq@kI`2`@qG*0qli(#!fyPzA$dpDI8@d|1E z{_VCzv{YWzEDwqNE;ssn2~Kts@PH$)l%~K8z4nBl8B&&sz)HA?aH0$-OiJLB*UU_9 zsTdGj#(M81&6+ZBp6zZizM2M&cbXOZ@JBr!L{)hODsFka_#Ibrg$~pYH@jqzh(?|y zf-Ql}Yyp?uM8V9FDnm&vTpM~+Tv@yfwMq8&Oi@>`B69A!&rE<6*Wn;{NIQw z9d78?ZwrhZEI|_v(vge#b#zrt!^3bKe=Bfa$cI^Ow75f_ReD}a+S|inHl@MhK1I1% z3M^YgoeV4!n&!Er1?WF%e(}FeW&S{^V)Z@&Ow?H1I_)quBusf@$aSLoEBq>=sVP6MQl^iFpp2019A(s4rV8eg4QpyKoVq z6vWVet?bBQ`~x;dMKI7?arN-^eUB9aiMu8QR%6!9uw7GpF@9f&Fz!o7V2NVG<9>iU z{l+=TI@~rkO+)QV-;d>2x8;F?3@w^jIez`iG~3(fOvMjet`vaLlZ>9NNZeGcP9ua$~Z6vox)hX0cGurFYJ9e_0TcZ>A zDHHQc$i3%9q|MUleRceM(E3@?y7SHEz)p{v==jRr#7|2PlvJ>VDiQ*_hq$B3zXE~Z zj50z3B(_}6{7*f+EG4&z;D@JiDWn16vVak+CB5}p0apB#0dnQ(=M z7@ra|tOmc744!#aZ=#T7T4%p5p>?9LJzp)|LiDiFxjXRuk>Rji2P9G3 zCh60;>;sZ~{P#iNSFa@jGlj3^zT^ms$=9dh+5I%2a2NDOCxfFh!|Eru`QbY^A|fQH z!)=gTnORUXe87Q4nURk}o9PEgQqq#CaPpMp&;Rm9^l`PuWBMsV{JQpVQs0E>ojD0@DKF|aOe7U~>veorlo3WGVj$YNV*4emEO5If1pV8)b5yytA~vTsNk*k%rL;tLRXi)!q~I?HAg;9ewSD2?v@&e6Lm zJcVDL>nK1eB`P4Br2~UnbF7nJ{F4090U=J$25}Ij{LKLJg?9$8oo?YeZb)X=21) zTYOIp&>#OIkaRz{$Q`}3XRlx$B`5IRo}ML)x!+XUOSze->`^9G*|dC4xh^LKI$52v)B6!* zm-I>|LnJK5DI^o81W5*>l150FT$!;8^A{hs=jiu7*o_Xykztck#QK#mfqFW0A9guY zjl3_T{8!~r;_2AdqM>1laz@rOi(kt-L~!p%j@*6JX6uSg6Y2>5x-NvB%!3JWRKjCtI-3r^ zE^`QvO2bm1i-Q7!b+BHj#X-AINR=WONLWiU&r`~f0 z+7MhS&N#`G7Jwcpa)w_ZIm~#4V?*=XR-4N#4$S^+AtF>%LLVy_zR7AF4#y)4$_u=# zr-M>`GxyPz9k~<9*j@{2!NG`p8n;%oJt^}wSK;&}qF)r;i3y&xG@LkVI#o&YIh))QY9Ua82M)Sc(XQkU)gnhs^|yooN=->bW4x2(lS z`-yX@Gdtfeq(HI=dt~jTdR26e3tCr0Lem=c!h8{*(!H(>8}pp4 zN6?~~^HM65%U<~5sMi(wUVGu8(WxrpetIk4Ub*^!5N&F`16j zvG5%r{*D6hsLb4enn)#$bB6{8e-e84^tCAe*|;Emu>+bnjd0V$(;H8bzxC(S zG;65A$K!j?gdD>nij$}}t@BAy@mUg){pT)J&%9Ssgu z+37Md;IN~5;a2kN*zk$yKRA0Ru5;g5k%-&B=gPaAvk2tlAj+WtkKXlp(hP6P14XJ& z@+d;$-{xT?tC{jn$R_xaC^R?_lt&cF7@s@1-d@^S`0odXrmHY1PHGJf<7(~25Hn3n z6!gT)ji!uE&c9nsK;`gJMqWK2tD)~@*UyiKB6i>Xas_gr5~FTiC#0OmQkiA~z3$4n zyAF7J0zx+st{bIX$VIgyq;X#=qdH{Ob1~4(k3q)ft(Vf-b4cjM&(-j-o4$me52oj0 z+Hcn-V(j7DvR(fD8^^a(Q{g0R_7jq^7zy+miy=FuGN>D^v~WQ{8&8{2=>17S&0h0% zi~4Tj-Tr1ob?pxID;543s?IRu@5w12WLdD1#!ioK|FopxH8^sqLJ>KEN4wcs#x6cA z_|}5B+k#1liPjzYhNvrMXfwgWrlO5>ecS49E4a-f&+)$)6jpuZT8Jyq%eY3n_vd3c zoqOP3&BoaQ>QJ6I;VHLx{cC=rse(W~;l;8? z3j?4s-Y-IdS|pqj0qyyGlB)k}6O+ER_pbt@^)KMU{-taGa-)JfV2kVZaAt~o= z;uUP@BLpRSd9b3=*o~Q!z<+6RQJJweB|!Fy&$|47ay}=L+4Hg36qdK3**c12);rN%nBJ4*+Y3Ks73H7Z9`*KgqOKV$7uz)zbkZFI;dK{qNMk;b zv0!@XI`dxUPx=TfZkkoOKQ(iW8dsL38QW2ZwYY>sNyU(om)d(6h^WeH*|ag<>3vCk zka`1Mi5$2hmxuVRd9~MwO9@3-LkRROWU)CxW@3$WWXPR}KI}3PKq(MU5rtTxBukAW zbv=Lj%2HBZCh%fMD0j=o@XX0xf`IDF%XUv~Hr|Py+>+1S3Fu%IF|3$0aa3M+U_PSk z2TPff`p`C;phOupCYqXqRUB>+9WzW7QV6#JvpS%7WC;sKgu7NQQC^LYYutX_kVf*J(sJ^HmS^5UW5guAHwJxv% zfXN~P_4j7`*afQkVh&w21YCA-1EA~eRm79nbVgB8+)eD=3rN(71)Hy}A?V14DcaHm+O6K; zA6I+?I%7gx8TLn{t=Eg2r2)#x+JRaZ@ zVYvNQzSNm>gRv&y&9>8s7Dqx(2{!?H+ZtvW2`D>jkV4{;iVZX%jt-Rie!_Vu5;3|G z{n5>f+d8|`V&1sIQI14VQVhM_vt%|Nz=on`F$Z@33o#qj|n<2*5%{F%lYW@x3a~zB)0H3XmFv0I575?WqPv#;QbSA94 zfx%S9nQumWW3VESwC#sQ&#s~y!^KUo%3^}`0Qo|TjX9;k4*ChhecvATN!9kQz}ZsQ zgi*|^h1QYj^~8>p4w(jrs>MHUvJ!r)qa9qiBV6mhchi-X#Uu;e`FgtCZpYG3xGF-6 z+T9uMSv*^(SAdwEVKSvak*UXCJA&WG?biR~M9Cbm5a^)+x?cONmFtbI4^FO96ZPh; zoxinEkqP@vh~M1uu9#dc|MRAeq*6~l;b2XmDqa2iF3_oV(btEkHSfm4vrEgXX?FF`DGW9HT9+@FR{x(*B=Fn+Of5^znRCiflf(xI z<&0JOupKr!>T1~Y)fP@Z2o<3lY9oAU>h86XMxx`?*Od9NUA*~-L&zj+{t!#)&*B{r z_HT`LPb|c15mgsrxr8U*elISV!Iz|DE9zOch}xU0+IXCBe_z9hJyU_8$Bw;7+<&t;M>JS!b>ce`+}Gbwcc;P);>lezG1 z?(@Ma$5)(*3dG_4WW8pxovsMH|16;Vc0DX8<(4fwF41soz2~dK7(-77@|!c4U3Bpe z57{J&a?5=5F5jf|J*(n?X8!j$`5|3Gi8BZ{9;)}=t3RUu*q_vAn{K49U_4GV@u?!BD%(;~ZhWwY1SB!W zqP5;yWSumat(*f#hki^X+|N~@3}A}u#$fONfon#h=a;I+!kmN>XP=v{IW=Jg2N`P@ zn*QfN5UY@v9IN2Ms9pSO3AD@6@e330JeV#=cUfnZV$7+^h@1zF-78@!Rw@KRRm6s7 zfUf+h#$<|TS1_`VfQB^ArzbS_qVQS*?0j4!l~kjVbF?9Tvvyx{uu(?q&Yxsj>{&; zJ$o9Y7(Qq_G_|(XX;06`gXLs4?@Rb~MLONrWw_VNS4rxf^%fsYg9@*B`MfZSR0;TH zd|YmNnhh(B`NTbQ6oK*)? z+-t!6_dMx~iVYN}(uL2#IH76RJN%!(4FS3WliC<*7sq%KVN7)w21}grcLuUx3HykG zhWV&I?@rTFZo0fA4Wox#iJkHHBuZd{%JIdjF#2@y#zVCf~Ts;Rk~${R#Q zLs3U2(Ab5lE$!4%;WM5-B}Kv*o-9WM0<##J!HShyWF=2AsC2fQxJiz%X-0*UNad;4 zRqB$K5?n0hBhuNTNNN=S9PAX1%)H8pDWU+7Xlb)vqr+8Qp+?gK%W+K>oT!;i@DoJ zLk#aniya7#Wn%w>zrX&b7kOb3kS`1kzfd>u%!W%8*Cug4C#g#yVFya`Qv!%Z!f+xg zX=(T6tt42Y@wOznnWbX*+96FfDd z9j``&TwAjSgF^OAu>fw(tq%sj`1Sf97x@hZq`rP(#On;0A<{)ZVwGyl+Fc*zL)P1U zC{r=I__lcAT>Y^1di6EDy!O~uxWN|b={tX8yJfevj%Tg-M-mR9taRz;>ugb#>HBBw zdw!?fTb1x^N=&K}JZ874qrR%Drk&i*uT(jI zc=$}VyfxKG1GZ4fS)nepu7SJtDC}X$A=uNF4|NBt!#n;p{5ajgB~ zOyFhkrDP{pc*8FAZpP!>ds)$+%Ht;$LPCDJ> z-VSdCKn0$(>1en5Is*7@l!fSVf?QKNWhx&lu@fJi&%N6-S8k}tfmub|@SN3wZs?yx zo|5NHNv}k**WLTX@*uZ?#_EHEe53hKSv9Qw8^pF1QON&a3!qIu^RYFu7 z{)}_>Z)IueZM4cqKTglQ4M@!6_iqBo=FXGiq$|p#j3ATU#h7ccxu%%KE(9d?NrMgV zX)o|`_d%b_z`$-A4li|g+x{!?s&8Ypa<0DD|1HB$4w1Thg?i~dv!}WkfhihHI4y^R zyvHl3Z_J6XGM@8lXNgB|ipT`LG#+T6>dK8@UJ>ftJ zwa)$3;mg&`CYxn$eV=NNJ^3-o?A6zRk1)iFZ{vG(XlL}=joYiL6zC6Qpm!5C$ z6UwLHnm`V5Kc0re@tEdjI0HbG6pl~SWNT}zfM<`MeBd;mzEP(cl-EatB#3lV22M?4$O{Kp(QB*U-bU`%hUV91nmlUeSeyx-1!D#m=vMZJiDesf)~6 zrG;^|ql97WVSt);)|yKZLg1Cw#K>~yES5+wYc^rbs3Ystx<^Tdy+pBmEpE(heZE68-L8Mwe9<#gb6+4|!Bj@=2NW1`A|qk~NyCe)V2a%~WHX#gd1X zAq3l+PA;A&h$p$&;bbrM`c3a&a1*IvfuA#;hqrl$jHwo4Q}9cz831}J|3nFkdiVY9g4Z<`_}I? zC{Z!V(-`51?BNXBm8p{Tj676)ffy20<*F}lq4OYjs$#XqLR zmkLD?T8qAK^k@po(Md>fk0%SGSD#RbTAOMpgu;UmAS3DQeQ?>w{TN-NwPOBoI2ZaK zlvP`j@7YKTQ!5wOk`zu z56t7M(?yO>RFa>xjabAlV2)4DYD!_Hu~SHDe6w}ED>ji5!)P?;m9kaR3Ms}AVjjKv zxyO<5LE}S*3_4aeCAzT8aFURiUeY~1T!lGsoU69)95N$^(&X{pXL!M>ZeAy2o^Cg# zA0-mZ4+5=^dRV9e>&e-(X~BOLO9%j=*YE_zCFGD?oeV{J zjspS2HTYGZvBncYZt5tPUtNCvgi@&*EL(r%mvet|*J@oI6Z``cOKLg>0^CjqXR#A< z$~NF?@FYQpqDEJJhbz=`riiT({#m#6=5Y(n5YRC6IZz+k#!7JcSRdD7>DfWQAo+iv zhmsvIYiSOJ3cG<=m#cLj+e-$j?u&5QPG#6 zsf+c;!RupDTA!ury?@wCwHs`tjNZrQ)1L3<`Km09`Qu7(iluP(4zl>O&c!pJdwKA4 z+Vf>J5KlyC)1|!?wZil!Awh3opm}FQxN(y-ZWhQ@iJ1*>Tl=@%G9PGM0+|rnvpaih z!do}wUw@&@WeZW9+~GHZpPa7z7hZ`1B0X2@`X2>#djf0QvozPm&%AxMFeBXuf}SrP zn*B}%o}4NnyuJP{yD8Z}p~4eRd|timP=@nO>Kd5Yb5uk?s@)68!Ubg1OB-UMzTis_%M`& zbV^w^Ww(hySCLI+#h$7Y0JyC;$^X1+AD$**f{c#flV9%A8e-~(e3=xp>*RsX0wU84bNDR2%jHKfeV7++(_fabCMYZ!Am8}1*t ziAfjmQ`X{5R$Wo0dNNh)>Vh?_Ve`T3DjVynY2)C5;$(Z(Gjp`Y{ukg_!{6Ex2^v&M zFwR4cGbaQjL@aoQA%;E*rta~WM7hTcxZM)xsbf%@Bl5K!hR+vF|ucCPd`o3(cdaiJf9g<1{S~(jX-Ub|N^mUT7h%>OHFaMz98w_ELqLGR-1e5b@l066YGPdAJr~>`;yd;K@WzQ(!?eTqX0Mbzr-h z+g$S(LH9(Q7?=h7u`glp92mgY7_ne-%N&!!~WQw}}&gfcCR zZdLc&w8)du%OZ)N#^2^&JC?9oA*EOK246cJRWsx zuwQKK_P4em6HiGRpRYtQ4E%uOtIbsQ{1Ims7IH*72W&yyU=$Khy5T8N>L@5rCu()Un)bY_3h_`8a(=?n?UF zBpkS7JE^qh!Piuol+EBn-{CDfz*F+5*vd-&+eULTOZBhzi;9vO9LYJ&WOfSFXiN+F zxy~>ux|Dbn|L`qSVcY=8@_yAj7?^vCq?m<=vk*Ke8w1m8~lj11Q4T;UR3ov(N5G!b7L_$E6^J8EVd zitKFr9i_oZ@06F^enr4w&fTl3N9Z5fAV(LpCiI6CHh*J&6TH41(wj>e+SoglLui=M zv%57Ps!)yK*Rh%<`%<2+-zo{&xlmC8=|G<~)N$tqZoFf>?7*V=5Meovk=O*{Wnp`Q z1*Nad7>=&)jbza*0VL_mVJtY~IA$#Sv9eqKdpn*phj7I8QYR6UD5r=Ix{w*!q8 zZSL6*dYV1zZJ+!5mgnmueCfYi9&fT5@Nc*BQ_QZ+MC`_38L{YQ(8W|nnt(5KbBR3k zJ|1=me-b59)1D8vW3bOS9@dPF>&ZDl3LP!m)`hEbQ%eAN$G)+2edW0a>)4;JfLAC{ z6-Rmpb(gym4V&rAS>LlN!p4>&9)5l5qYBr?SJ(z)JZal+lsv=r0!Zm6#`X%r=bfVn zVYi2SC$HzdQ?tP5yJg~cs0hcpL%w&XPL53Z8U95MLA?xKX22&acZ3Jex6KTmp!;lc zvCd2B*{l55XUN*iAagoCIa`~TNEuHI4xaYEXq6hc5a;pr<^9Bxb7Qu*Z}Y(m{*($# zSIutB5##w8&hM?^Se7z(!hiiqP~H4(krjS(*iauZ?&;k*r{27n4TsunM>6Vl7NIIw#Pp3dW-uGiPRj1DqaEAUONACfo*jDidCTsR4C>-Omd&O< zR(;INc`STD8<=RuTsEwb#MyY6$4eE6_C;a0>(z=BK(i;ktl|H#l=W|fw<>WkqkToM zrlA13Qx%dMUi#R7A#T*dx5%#|<~OEp*S_Klw>s+TlZ{G6k&qs-j)I_yQ;MWf$pr7? z-?7o6MK9q2#s{VkdTn;<0$CZBP>f(Fq=x!Bqhs?wq}j#?{uu|ILv#Zs=JiLIm#HeZ zP`ub2?-EWAk8&A1&{7b08OBf@eP^n`;Wjq+-eki`3Pp6<5z{1L-)+`;G1C4+#l4&l z1IASFD=OeH3my@qrm++X8fk_5a4@$1i#DNHjR#L<$V;Yx!!p`on0!jHxlE-dk>iA# zi}1y>T5?=e|3zG1p-k~2nBZNh%_YP<6GDL@8Lmo5E7?y2l?f9K3tLbfYZxa7(N&_` zWFh{MsLSw*qF6n@YeR~+Iqb+HV2_gOP?KeV+c!rPG3a^ZOmB)Y3=QWOGbg70>!ZWV z0P;S>eIf+POV?bo6piB0^%Fp`ZN|mNDGX#o<@4ClD+y?E^QjNos}8=JaHYzUD_sTa zLd*B!y5qg6ZGLBEi}CzM;+Z9l1|<_9&>}?lZx2Ys<>g>3U^Y6+QlK0>e9paV%KCDEmrnSklD zuZ)ro`3^fw4j9D9L#_w+aE2ivT{!X$k{Xn{i;?D@fDf_dbq+ESCk{UUO1|)MjGab} zCFe*lyFqY@tHx&WQ$CK=B}H-3G!7l|u*A?(B9r_&Z=Gb7Q5Ps+i{Jy}MUm7knZyA~ z61Z_Bi@76!=6~WPNR*)D>xUN#cnTz%$d*6_R7CKPvrkn*)Oy>z#QK#VPR zC^iNsC9r^@@-eQodL=P!)+R1F&NGpVAA1@ewu8!(8(~qlik*;AaBOMs$kE-{6e|P= zN6PCQOAwfuX_3p4s{-l6%`<#i9Z2BS!Ezw+Y-1bNvAfF_%bg44$xv`lV*k*HHPZ zqHZY_f}zg(N6guu1XUP*Cz{pDg2^(Gw*`SPZY=a8%D9=Po$dx{erZm{Puyq}AaUy( ztz#IhV?D2JC-m`8yYRd#_)Fwh}$xq+m*Y?|tnv%rm+-IDa?u(J=`~xEBE# z%*e!)JAQhUe{nDdGJKRaI}B_GkS?isl#jzwNHuSWlnc`l5|P2F2tf>-hNoKTq|o zp((hy@r0ym|4fqCj-lHlNu<(p`2O5;T~}*!>?Z_-hbijFoN3UO(si)~mdP zBAr0tachSx@Uu1#4z~N`?8*1GB>r#jCwXAYrHI)1V;_0r#hZYk{RiqtZ~lMNn|=-U z4~MHi_ouG{U|m*DH>dw^Oy>)q)X{@KM8Eyh)_(po=PGJLl}jjgB=^xztS5$o<6zWc zBCyWi6OMWdHWuA5DYqsc7~832GszA#)#~%L^bX{`YEbEV`!;1x`LQ)k*WvLJYBnu^ zxU4@Bi2krQA*F{|miW2lW$<$SV*BNhn$IH-3;Sp@?8`$sgEY5&Z`aQK$0MtJhsWD+ z3o}M>?>0%cK3gp?uSjREcc72da@tOV@I)HCzMfc`HwaM7@BtZmyyUJA64OLLREpT_v9W}uqzp(H6V-Tc~#UbqEXl9r?n zu=-cVESzU>^Za)|!NV_$gf ztdFxu@RvJrCI2hTm?ErP;u4B&PVoNLG)+EbZsv2b*pU(5m)tlIZW>l(``a_3n8a4n z+w|rK!v#Q!6Iqs?9)#A!hh*Z6X8ua|vO%JtOd2(~RsNUkoVDu1^Hs{%KNCWgT8(E& zsXPlF1D-eLYRM~&rTwdUTuLkiYN?t&?(bQ6-rJ}pd@uO|H%T{gNuOwEuggXf``{S)GgisO>3+#9RN_anMjEK_%B1^q z@B$g`yDbG>!p2o>Bx;iP-o#)~Lr$xP(#Zep20GOB8mLg!cM%X5;R*API-F^!YGPi7 zn7S-?>gG#(AdNI$S+*{fsFShTxlfc(uOW?enlR{!fP)7{Z1?~BwLPx_z2SW{7v-w#MVi6Pm^tKA zYbFD<&6=3v8ENK%^(Yl%5s5gL;21p&8jM@zpD_E<_cy@D>XiBNsjdw7G}`tkW6 z@V)t+hLT(=9;aph)}ifK1K3-DyWNxo9URS_HuxiI@*o7dn&YGHChYBxvEowDTJP+y zX_#<(;}cy#Ft#I>b5(1eOj*C$i$-HUEO@Kgl4>g`PO_UrY;U)k2NK zte#1OZWR8-g zzO1SB^H`e8Ih57hdM@3kr2eILwqY#UROs^DDPwU<2+49@+BqD;rpo)z8N9TuBBGn$ z6E#XH=DCUb@2e`0{()bZWEa7A#%JSwd_s3J+Mp{c-z-vMIQlsXS3)|!V{@`jwfWyy=PWK*C zp}g|i{e8WBZ%^KOuZ!vKp!TrSiH;zVg98;#A79b5nt1V3#@QhLzYgZk(UDVLt?RbX zsP-+NtL=t6S*II^oBy}=5h#Ws->Lt6)L2oEx4abADRR20l@l5e4|JXQIK!bn_k4KD z_?K&=x7TrX6%7CODedsol?)HU{H`{W#{6Hb7YFB?|Hfw;#p+fuMGT1D4v#FQ*>tJa zSqD-+%)k|iPN(g?d4$8cA4imd?U2reEMfsfKk-ZKL?<0j!WROp65M2;%J#-8p~r~i zs?8F;%7!%Zo=~ZOjE^{(Jm+BF5bcLhG8a6S!U!#)S5<|Ij8z`6+=hq-!5+bBg_ z$zx-OSpUk+@F8iHZhsUvxP-Q9<*>6l`i^Xh_G}+~q$sN-O-w7{zSmSs`~JV=qKL-Z zxRDuw6)VJd50%&|QJ2X<&+apCcMr+e@fOq=u|5Nnel`CnK@4flJEy6GRtknTAyZ8o z80qYs-(&OUACIxM)@EJm0mr#}{TMyDJRem>dPvGy6*y==M{U%$czR$6M~RT)isO5Lr!&;8Eb$ZYVq6x za)n;%6*o1S&o>9vF@9@u8eEBN}yD&0WIFV8`w zqvBE*Y$1b3noYRN^=pBU1QV6X3cKp|nbo99QOa4gPbsQpz&d`G@wA(ip1AW$w9~}2 z35;Y;n*<=VM|j7~Dvc30kcLAqAD=|H7HAqwQ!0N!zr@VS(cWk@w?G)!$=sR8b*83- z+0NvjC$E%=BMWj3=dR#+7h2JweI)mv@F~tJp|WcogNgG}5g&=6x+X`;hd~{!<5&c3 zMIpul-ViO;-!-Ks%C7pZCpH#CMN5iA5D*5`tghTSIAuh5MlgP;j-Q_uAlYQDD+r~F zN!EK6*#Iy|9wOQaN`0mwPaYTz!(o)MaF;P`<#Pc1(dQsCFfy7GmMf5C3spfAcn`u@ zfkgK(GX^Kcv}CI!uK7-+f z4}ug#WUlE?Id{j15sL%)D>@t1PcpHV}@o%V;Kc|SCsLqffjve zkB-)-0OK=3fhI$Ew63}u`BVj4)x(+tlWgQ{Ox%vOP$&j*Ix)$%O|yv>A;H(HkS#>C zXk4&-nDuPCC4Z$eEqD7*LCB079&&l*0C!1rk--ObA{<$JR8-BmlRm;m!vlE-+ftpv za7lR?XCh*{+4$nEvC@(wvSNaTQd5|Xbb=z(0`;9b|6Wy6#js2)rC$}Czd5mCF|@`^ z39!gXTzB*$QBP@>jFnEHPpz=gi+3aCFx!q$!J@{G$JLOfQ?3z#+SlsvuV<(Op=jUyGvR zeqOJryMRdN+2idsW~-Db@B5k-uHKuIZU*Fuirh9Yn46d8^|apg^)*%XzP98$Nyzru z$yc$xJw$5JwF_&9huKj0eG&8dIu?9e6x9#~151JK z9a(G%;+BlDuUNCCG$iYyqj91*2{5|9nU&gM)cB-c)Ngc z8yd!#y45(z8WF)2-DTsSzbvO2^cu z-aA#>{!d0B=tE^3Cr`ntK&02GG$fLPB`%PH+sWlTm8Pq zg-hw!eW}Kn@p85)TxHUVFIo!LTWg45MT)0**bI7V;4Ac^yiYi=@7nhzfCu~hO|w|z z(NVcUuoI=VR|V~w(HtmR@Ns`OO?Y3KL;(|mM+6EvMF|x&@3|Lxw>SH{wueGg&5QW< z7^27Oa~2tcS=gRT7zR0n&{)k-18Ka2vyuM-!IBbaQzxpW7xdMrN&L|3 zJ@)7tI1zh+OtaNooN`hP2-F~a3SQvHi+bu-mbd@AF4dM ze584vG>fKAc^~NRP?QeER-Hb0oJG2Ma2VuSQe5DL<{s{xrYCE!xW_Oyiz^wHBS8Ww zpS{IU46m^LqiLDW47!tt8oAh{qFU<~HqkyQj)MDyrfXZfJX+W-Prf9Z{u(A|C&Tmm zDBs5`BeR5=+|Q*HNO#cB+RTaP#wiuQha&`SmRzLt41=H)FKFmB9HW>nO&(w#`$TeQ zjUglJ|3XH!?qKv>$A?-72*rgJ;#Cnl0POy9Pz{(anJNyMGmFH?q-K^z(|%9lCrgYX zHADYkggUywKzj*CeA5jkOQdSbG10pzG+Jsx*BJX$imd_VhB9eI#i3(Ehs<~&S)dEc ze9Cv4;V_L78tigfZ`Ugnsz!#v|417|7>eR$OBC;je7=1$)BF9H3;_YsEBQk`ZlLrf z-_xRH$(qtwG6nyB#;H3;VIt319$Gl>5E7F6l+B@M{+V89F&U_nlx0Eh%$ed}kSP#W zfT}LdG&2!PUTiadJNiDEg@h1bv2Aeu$P~wb5}O?UPjn>;kJN$;^mnY?ULT3v#q+=N zF-ck(I;0N`mT*$|7L2U8sI|QA#)9rZ0?K$TyKN?^s%8`KMeo?aem2?GY749{Az0UM zewO|E#3JZu)n=2}53U-BP>e!bwYD|0b}#0Rf%?d)YsQJu74?^aC#H1TUe(Ppvb@f! zj7k{RT8gS=T@y$$R}%8B*=D#B*vY3Fc$@U-?8a7CagTk5;No!mZ|>WKb0$qxR+KQg z7SBOh8cT?p{m9@Udo!3_|K+cqC05sE$R5jxsmDseED+ZKFpAU$yIaQyX6HX1!U$tI zEJW@Ym|0WGv(mP>SX~=DWq7>xnW>#Y3A{RtY7F0=ZXkKN=m@@9ySE2eiTT!VKi&@R zXXtohdLo*=(}P@|(5MXrXIe4?=Qids}gFpS_@xmjeeZJ+T?0 z2cZVM^@U92?pw`Js?@?qcviiag?Tjd;Y)i?f+yqjhrz4Z}l-z61&2t-n=?W-G5}Ffmj-8 zvoHen@20OG#;@R(v(?3)#$BD3i!a-yD^53&YSWOA=Cz-G*h|K-lHT5Pg-v~{(XnfJ za7(9|v`wf|^u${5F5^i=G>OPPaaOB})w-eA^YuK(l9h6ghWu=E#DtQi(}2LMkxH#E zym&p@rzr4T!}KY=^h48PlIUIzeR^6&HWLi)^Mda0lJBQ`HVpSGDh2Zt#M)ahIQ=XC ztj>6>>9eM`HQM)Th6t2SJ`T_Mc{{yMm19>Ud8Ki9rJZUAE$bqda~)AFQvOdi@%Frr zIrAyA6gN(+5X%xI9}QGycZ5?S!E#n^X@=p+$*t#;kJ&-K!-`f~*Q8~~jN(yQL!Vqb z^eCv7c&C|V59$@Teh+p-v%>b@UhGw<`I*_KsvK6z8BRn`h6J9+ILOBd3P*wu-s~2b&x*O3Ky2kC62QmPdV;X(+>^IpRr+ca5Cf> z*guX5%XI}<4Zl6LQeU@<2Pxo?(K2Z3Pr#U`iTsokKEuR>SuwN^tTT@$Lp??s_%(peJj)k!+`2Q1FrN2^hU_+zIu9k1X}6Z=&%u4H^5!5G zg-VilyVeSl^d+XSCHP?Tuh#~Ss|*yw!;B<%cW6qYQdF)nW=v+=HTb6DvZsi8dyyGBa&d7=QiYLdb^{njq~)m&5NB#zS5)JohKgu^ zPj<{d=(OeOqR#C}uxRkJ8I`j~A<_a(ITUUyM$l;h;e3SNDs}kEWS=!_IP{3btSvNF zKUtF*M4gP7mKr>oj(fR%!HdZaw@TL^MnlSKe|%f+CGz}H!J|JyY_?v&H&HWpuVL)36%UX%`C zn}eMsfsi^H7e{^%bIZSe4pNz7!YF1(l_nt7)wI@=#MnC=l`MoXgn@Iu{O*5Z7z#+$ z>qFf;433*bgKH}=CGFx+*)2H`qKPtn3>8OvzM|qS=^1=tRH{H?MUA4ARKYEkEsx0% z8S^t;C*LOhMaG5{9JhoSCXs*$wEevb=_p_I5Z{*`0*c8Avk--+D)&4g<}Jyt}PhQrf*-2d8X9@zK^=!f?>Iq17_S zJeFtC#Tl2}gjEtY!cr_BJ|_+1f_*jkIS#x~&2w$6F8`=t&wI| z7mXtVQ8+)bb*0N!&yFYKSTxAVKOX`sggzsonX&!Cbff@|^I3q>FajbcqchM!-zmUI z?pM-`eG8LiyroK0O-t-v3AteQR(j#|dMzTT5usy`YRtGS9#%E)t-uoDyle`{{Dj_j zMm+u9C|*8~kgROFst}#b7iS$aR8m$Bo8uVAwd9MN`0H{8mwTSn{fBSE zCLgj5M`HKo8Ptv@FBiD%*jhsWoTem0o_9<~ZaFWWD-JBPqlyLF+lcH&9t$Q;{rIE95t1;pDQ4wTT+g&b!zZWYr%!fXs;WcAhSP zp?q~M7rAcORBJKav;N)qj=Uvz*L%seRXNX@w5sy?z7mU z?Bz$87(=NY96S4Heb5#pVzK1!XX|$sJRKaekyb9MM|TU%DQ_;!Kk2zbv; z#6DF0>aw5}vn1a!`&OLH&vIn^#ZoKi?chPH)i>7>69%USCTfHxK6fV8Je; zVE1b3T!5zK@(kcBoJMe9*N!$d#YV!}GyZ&qo54BWa0WHs=gkh`a4sJGJ9e*NtIonW zC#<|ZKR$4%mTOy`h)7;TM%rpeok0JBfEJDgf^Jr!aT)PM!puyo!3iD;3jP_D9#$#8 z@A;K=NqIbH`SR4y5n^&ly~5P3%fQQyBT)gvg3avFW=#s^lRyNF^So)nP3fM`n-{MG zLy39|PZ{36?fcdAQ=Rvvz4!kRI=JiHqD!9!oEhiG>kP2%BY)m5T{;B@Oh;M;ue$f! zf8I87=zVQ(FV=PXBntNw!{lJ?0{hllZl1gi9wAEk50Cozrjr;{%2-Ub35#FDA%gmz zBJHF;*FWIzWtNq;zLo&^S~F5Dd2t$g6_NA}1768x`U|CqL0I@22)V7d={fc5llP0c7rg z>5Fxb$#Z0UBFtuA9s1dpU~sEHT_X$=6WyBamMVw~0D{Y_|2T8+@09Ms6fe$>AlwX& zpEh~~YYFz(^);-)hjKIH07YH|0!DcYRr3a<1AG(lt7`LY89H+SJ&;NjwY=P3!4FP} zgD94vv}Jg^&bxI}wg>TTJG;oY9?O>#G%cJ`q$5%13NH=to0E(Tl7L9N?}HEVA1z!! z%KOV7f?>=(P2~#QTLvofU;;XZnf;;nlD1m1&LY*Bnspj^>==$Dqychv?|-9$2;8&D zQ7M6FaWyuUZeb8X;+V`##MHDC-$*yy1C-()qmo78#oR-DlnLjXrX>&Ow zveL?^7I1^6+g`C+LovP71u6kZJ4z%uhraj*eS;2ZkOUVwZs|l;<@{YX+`%&(HMxy&(LcfL%)TSx12+WiIsj-ch<+b-Bfz6g`XQu_8bjSon4$`Dyg~}LGraon>D;dv3*_W$0gjMK{(#l`HO*}pHBnWml4gNv1t026o4F+ zr2f-SZjv-iT0v$U*OlSPAZCasiMP>5fmTo7B_Lp`@fdTh&!i$oNq`4QO|b5#uP|N7 zH)k;W#E;GoxDKqYF@s!1#p@~r)&~E9d4Nm{VuTc_OcrSUI5>M53h$QI@^bL?fXMl% z0DKSwJ=io+eUy}*kvsj~{~{shJ)mB$9o4gjfb|m@4LD3fcfy{BK1@O!s|`mUv~9-z zkgtOU{+2k6gAVZE&UcgE3)6@>L}|c5)qREaZ?J^h?mrOmN}+#M#>E|R?Kyt=&BW+B zy(K&v%+9Xxqo0KcuW+s>QwgV=IlSI}7&RP1B(3mbt2sB5n-~Xoa$?5kabZzLNVp>^ zm1OFMgPeQTabJFZ@|7e&EtxovL!(FBpmCN(h{0KP3w+<_=a{BVK}k8kz(1lCu9)C> z^0ogndKKXVB#ImCPZNgA;VNVmF8MndPqUQufj~S-o%?qyR0=g7&#$E>*}apWA-vl5 ziMx(cbY9O*q_N3NIyTG;E1!3RtB5h_-McaSIUcxX$Z=gI&Afx#x&80a6F7r0o0e9i zFGmL+Wl4)xlJrqV67|86J^KsL`Mk~orswF0mrcu?sQ?qpy&TdUX}H#@HZrM{Jt$|- zVTap{_;fn=ZMB(j!WP`k_RNcfM5{9Svy8>^+?sF0xOn0y1+w`8R#?nw_sUj#7~0u+ z%-WOyIac-Y34g_t2#M)D#o-?zneAtOdJdVN_;0@syR&kUaWHnm<|9GuPp*b;%~TmW zeJ?Uox6_wnx+9I`iP*sz-wt@W6*NRmmN8F4Wc8=C*=B6#)w6@W&ZdOK9`2BDzE%s* z#P0ueofP@3+vQu^LHkSZF}JCbKjxl_{o(KkvpW9yZ})pAShPqd=kWx0+VvDm--+^6 zcW6xOxpGfxSbz(!^Slo-F=@{3QG<nxMoLEEA$^T)4v7@yt zt9m>>yw$r~8M*y(HwZE#yBhiH%Ufs7z$|X^bg9)Eq91y%)%pb0n7&wsZ^qW#ENl!&7Z?6nj!1rITw4j1CPU(*C^_+UU7?K zztZadGR0aX0E#7(ixX@6I(ia+Bk+1?D65Y-**as!N*+LJFc=lZ)Uf6ee#*+Y!~%j~ zOuMyUv1i%6E}}R!T60XxeI{;$VPHUa@uPlqB z8LK;2&%En5w)0_J52JpG%KxDj5|}(DS;u{YrG*uOPo_$*d;r72-%R=Y?gs;&nrqw@ z%uSCXDY<2fvuJ#DbqK)>#RTD-nRs(Lii&ELSPKs0+`YBMSRhiqd|&D^%N>XRI|#Q92QB~~_u-dWs`QVv)1-{0Jse>K&5}W|jY76dRcmle#B3cQ*u13aqnlv}* z0pjmZRRRe~lx_qFp&IWDXsr4&FnOdUM+ij zQ=wcXoceY7f2~RWuOBCW41wAGi>c|jcg(2^0f|LJrxg0PmH{_?m zJ{+TBTnUBFiQzvi9j5ET^@W+Wc(Bb`i?{KFrfA;LXjL$W;&Vz85mhClSNi&#Kb`_Bo}6jpb#S}1TaMvJjhreQZlFT zbzj42ETpE$lDJ9Uq7NDLrNAY<*}BWG2Q8EqGoXi7*;P59!f*$^$AVcHKC{LF6juol zp>A>IZ13Zimno+7_KHjA2puSGd}navitq)g(Obo=nkHGyaDj+_2y6OR(DOy<(i*;$ zD{+&C7nE~cKN9?pogtC%qZ`n(elw{*&@{)FDtVl>Kb0_L$uQLa<(_s&9bk!Jwl`79 z0n#5%IvoBvRSFMlWx$ZsE{3Q|LYrucg_=_bl7+*dk_04vb zw8W&8GfI!#da~S;(30dFn6qYTAXhVtK$-W0;b}e7m|a1i_hDUsyn9_(!?w&Vef;^h z-yTmH*S@$2=3iX2W$dJa*_jBBPZK5#6X(J!?KKn%9eoq8>3Mu?=>-*W&nZ)$wiZSG zZt}{(I+Z#OA9#Hl4wjupes&ytyU}`I7w=ERCNM82K=3=cQ^HI?u`wpd#3-=&_W+fE5MQBIU(B*ukYI-VdcH+u8wDPzUXV9c@;*^WKOVmPYBxxXt0(x6yhKM4KApjPg#fH6n=8k1$JIrpk7PbFLd_-q7$w>{1q!av#smb<)p z8puUbHuu*968#ko;90CLGgl9L)Ru@lT0vbM{N7F;mEw}3-DeM1zUEQM*nvOSUrLp+ znd3OG_L~oGhSH2hvfk*s=Ja1k;^BjBOVEoTatc8RR2T5+D5R*G4|(Ke2r_WwrL+TN zt0Ka9JuW@Uz}pMs-{e=wKcyx}_%q+H!Z5FM8ydT3;~%-gW2#_tB;NRtfxCA#N1kc}N!=(q#H}Ft*>Ao(`0T+oN@oNPgd(7NLyu{w7#V+!H zSzH&kUv(5Vh$_aw?P(3x1@`5JMpSCftSb4m1vqC|lf|mV0KJJccZd_m+xDjS`&B7? zYnDV678RK|KtM^w1uZu|z@PmsDIZLLnhGBqUrM><+HDxl!_^Rfx&SiA`(|8r+p(iu zvdZ^wf5dAarb_m5<(StE6(3F>F&+1hB6$874s#R6KIU{(v~?NLO)ob!Z}*KMWGVtP z9mJhiMztySrD50VvD3@++MV)RX;7bm+u?0bwql#i1!ObK-Id)g=8-&hHdp0QynA|P z8lPeUmU#lKiepqxn&O{U@VXKSdC$Cc7Z2xao^g`psO*bYWwDczzfO4kSSr{v;>VNhR+Vei2pdYBCmPkDMOoJ%@4^3wk)n?mu>$k<-i)(OqDel1v1h*D8_`1eXHEDGtGk zYm4lBV~_ucBLvunMmSQyo2_37Rak z<;^~nMaNaGgtNc8WA#-JR$9!e_8O*$()JC?y_lJ~C>5C^)VPR^-RQ!q zRmMSCVq7~ml_0=x#<$z&>pm$SyiJnx^}jg;ZQz+)#B8UjN$6bKFgTI^`8_P-@n$9A zpBHS+&zQMBx3_1ZHWb)aob$htn^y6sjDc?==Ukz4qmPm4Y_y?1u{0lep*rN_*4z*>plC9jFb4v~zj?DVskNSlDyZYxT#!mfw^Fx4~Z7F={AJ+2n ze~&^5C-Z503-_?Q=r6`#KobXq%T+Rg0{gf(SDvZ@(XR{4GsL~aCkF0%J zL8@C0LPA*KzWIS-!4 z5+55eXJj#9?==!-9K>tpQc|Np?7fUX*b2R#yuPr}enj4A|2%#)o7Bw1i*U_MxX?u@ zs5`#qjvX;C=28x6#C327E3txQgLoP^ky6&>a1Q@WHj-vmN~=ppy(2}} zw&g;@AIXp^jE#4Ltm2l-1V`FcsAePz^6X`g!@qR(i9%A7aLbX6mPBdefy8gY$j0ox zQeF+0Nv?Z=CuH&i`$Fjz%rOe*gsX-i{3tRl4@N9mC3^MbI(8X^Wk(uckvS%A6qikD z)psEHeaU7WI^WKSjE=#2F8fA@pZZeJqCrWL=sjRfYOX`!9RjaKt|*ZEL+Xgilzb&L zm+JieNb9wsmCW%#WimOKevjW+`hh?M+Ym+1hG0!$nn3yvU!HGRFFK9Fifc9#Bq=M$ zmY7z`n28k|(XL5lkT-BANKnow77%Fg&{i8=HvA>oc7;i+#5oRqrkT6;mQ9U0J@5AY ztnn0?ujqM$UNj*g?b}8BMqPzU@>twTMhyWj)FMi4212Va3Bocul-Sm+D#(yI{%V@e z0E#Ad%G8I8JbXYZ6`+ns~y6;K^*dk}}hr@`YMW}mts7z$p^qPdm z1)~fhgAAD&HM+-ofjXEnehhF(B_YJLa2j7=h75ze*%X-J5lj%!xz1%u!T3;w?mirv6JMD# zAeRmG`w+h6`|l_w3PQMf9yhmT#+VopBfAhMiK!I#A4qF~0zepf&ouVOPeEHa1nSbz z%6(*M?hT-i;QJ}eWo%NHrN&4|i}trpe(1-rytJsLFj={M-aA_UIo;(d=Ao!*lvJ=K zW3&t~Hv=I^KEpavjsLj4#SZ_LY>v+arj}DVach%wJvDR_7k>V+MQ3L;Z9|7;Bs%|} z7=b)928OJ;u1u+{@pxK?3t9K)>`y;0Ewjl(c3`(De{wJw z-(6$i<)b=mxo@N?@&UVM)Ai)7#O1E2X4vJ}cVDKN(B1uPAMM<@zt!$-eh!{ewnKtX z2yLL@ny8p9B=PY-D2Vv~oxov3Ia5dJ3l`lk3?d4BUdXF<(HzHg5o{WBnhUDu7TCwB zui2T|6x-f>{Vniv(8RIP>vzFb)1SDXqFLl(BsS(6#YfzfPBlgZ+h4pcmJ-{Y?%tW6 z47Pg~EH-Q{v+JOAsHXiuEk%fs=B4NEPlaA-ioBy7$Hh2Xt)zo$erNA&u!8QJcCpa)|GJ%Fi`gFH~^EcR>Z{-s1#^t{Y-5>LrfdxV-0Twes1^O{v zXhLoU&C(qh9h89LBDcpK&j~m@Wj}4=uslkoTr|)4jQf~2_@|*O+B)x*H~CTHcEtoX z25$3;)qnil3c~W%;wgyq?&rY+J%qK9eUa4UOc#T8uJ=Nrw0y`>dva;1S+h@9#~}uv zSJhKu7SUJeN$4!Gh(UACn9YuXG~CnSrt{AlzMhhKIcYwOk#+5@WgVljfelp%zwK)O%dDK^&ADgMa-1)$EX+D*J z7Ace8-sRF2Cd#fWDX*6hw(e!fw5Ohyz=&*Yu=lgfW?<`QBQGyW%+yw{K5r!@VRTcT zj=7}ca8BCzj&*LvnkM>?jJj0FDoaCd;57CZ0`yG?7OS*B} zB2VLIc~b!K*yH$(sFexzWJ;bH>QxxSs-3k~WN1Gd$%@Cr)DgWO14>|au8Rsdv|%TQdBJ>~#5Qn0ic7T3)BiEdOr7h)=x5Ar`zP*9xI?F8WS8!;++% zELWYRDI|uazMw@u>i8qb-~zb{I*EcLhQf~lW>T>x7NSRPL?BNutmt4#gajr|rt0N4 z>MEtpLx&zxjenGF)AL$e+eI@_@6yIEQ{t=G<6<;KuD0wWW+5jFIAZTNx}*_a@ATM$ z-=`N=Y<}MK#1auz+KY&;$dC<*jT!k(gGi+S2EC#R3tB8O;YTJH0kJ<~J!0}y6WKq_ zzkXSJzJE)o0Qzo+{*^b&%Y-dyoJp||z%$mj7d86Ula-ktql~N7k-SYU4rxws(pGuz zXVYp*(x4%Bs1Cr#1bZQ5)o?FtF~LEc9v(;d&eObQ!@O2;L7=KCd?mlh9Xn>8jh!yDHU!5KpV z-9(E?-3d09)Dj1#s^u>HrL0udxWPR~jp^l0#oxP`z#dxx^J{9RpNI9(FhW9-Nk99UsQIVZ-8c(;{yg#tqHL4TgJZLhQ{Jr5= zbdL*0nifB!nK}v_p?k}c9+O^ac>&(d-LIDtwKT4OJejF8x=2416xqtB7#Wfn& zwUU4vm&@%gQU9y%?SaZf(@Dy$z=#2<6U}Ds=aIc0!_;68G521T0;=YkHc7PPt(z}_ z3k{adduOHl{u>u7D+zcrO3lVs+2_>HP~FSg(1I5P_rCT^XzT8n<6x6b?uY7RrMM)l z_}gi)*Ish;L7}+N=eyy4TGMZS)vupV7PkbHez^)C6(e0jg}T-_0y@}55U zl5e%N`bbkk0Mo-mm4w3U^1tVrMXDCm(yz2b_uQ?S#1f*;%PH|q`P3~C!(7qarA;;z zclsTGJs_arRMZ*e{($5b4{8=N*=^`+>P&Q>Vh1fy4Oqj2h9*N0YHfHKiyH1n=K$|@ zchMHA4}YrX5P&wNL}~y2m!p*wyyDt`-jo6t6pa3+{n0|j=u~jKFO@SA`~-~HXnizU zY*@!mDucrs;g3olf%jpt?=LN>za_B);)sca;g)HiTmgLd67SiW>$5Ov1*DKWXJ^UA zpnQhq_Acc4#Vv@pRhs#u3o;qxJ=f?HuKDqlm))J$;ci&=F@Z1phLy1PwwH03uNoPR z=%4ZFm95CtlcOT$o7$2Sq}O{=>fw*q7vC*sS!rlRf}flR_qn|~A6z@kT!@>7%~w2I zSLOo3=P1gsoIndYt!p@rMD1Q$7M|u^D}=jmKv={03YXfY zXASCGAWN9|uHRr-@f*MU=&p`I%kU<7A%>W0+`dK#iY$i(d~5TOS%*l@!7>TkO1YMG z(q22g{4Lj56@!he)|*^ODMf{28OJE$Q-@M}dLZ9?|-<&2C99>%_=UAP86Fvi}lm?dRn`GB8L6+{lU+gj~ zDmhySqB7zOM}p$`@ViI;Ax0%8X!NJwtQm%Zz_-PmtnrH|V}$od(k-;%cvg5T^3w=~ zpEtEJvvImg(vAEuyUIC$aLjs6Yk@?HIjb5xt6alpT8H|#B=+!IDk;^Z)82H$nJl~P zrg9P?nN!-v?g%cEi9(WhzMe{etsXCi$BZrshP)g_Bt#(CS{o7D|Din*2$ZLg6l`o69FU_BkUd4G~i-8O!y=s6!ja@hFT|DP_h01sYCnLEE z1)3*~mTEebUPo9^f!kC+m;W;VRX;C-VeoJqZb*ESl$1gy8F~m_*@ZHv00vO>Ddq1H z?s17)gu8w0#u1S}y%mkrdNvI`BX9KP(>x!a?C9^}(2*LMm6bo#J^k4^8E7eB5W~WM zcOz048zhKs!K(7H?{})SF8a8S$Qz(MCiR#fx=pSym!4kcNikQ9o3xVLH0%p;E^&xT z$;hN=5{!3}H}?b3z7>#Uq$7$zRp9WGHsu}Cqa?W0uB3PrACi2Ve0iTCmZP_+;)y|@ z1|lhQyJ1A0M^K2roP42G23_QIT7ryF&sR`V6B~&!A^1r{tSUb8`YvoJ*L7Q*|H;bO z#%7Y+m_Wkt8S;N;@n&7~D2+0du}A(CVnc=`eIb@Zuo2@9>2Va8(mG`^Wn=|mK z>5=6!ex^QX)m1|b34dl-6Q)mfZUdJewOD=ifn|_g@YW=XmI~ck*0T5Y|*r8OG=96_U)%0;);6 zdPCo$#1keL9UX(LRra%DPXEAAm~WR&(@~Bl#~IBpswu}|kD*6}Y2Lr1zkbwjAX)EcLHR&x^W^O_HPmhArGs_onW3!d%6(2Rhyd~HoyjKq zt#7Kqe{UeEA5Z;z6R=gZv7+A|5cSJii55`QwWa74dz9`R&B?k;m;{`0RtPG&YvBP2 zI>ZV3{Ql&6=GGYb=6>pZ+$C$^5puy+b>MCb|AU{Sn~*aeGjyKMJV{n+@mm#c25tg2xa`NZL<>?4ZH6@3?Rmsoy*Qbsnh=0n7^oZp;;SX86Y5l`-BOH!L zK~o&;2hJA2MtD8{w2QQ@f6|qB|FRwZblVB9P*Y!{sPzd5dnKsv;m<=1_*ZQAj`U`_ zzQRO_5&Q1v`u51@KL#99!Hbu$IWNzjSBM?cc@*Qq3o_n zlTB8|cnoe4fQa~r=gA6gQ5p#=PA4;4c`6Y&( zh$M>+EVAq!6N(h$a9UVBk(ik@osRjUv2k_B>Rt0#7%PUgg`hm~p>2Zd7X?PKpPh2K z35Rdvnt8sk#q_PxwowHJCwz1KG|EWw{=OaG9igjf#hKIJvycZ8mZ%0!jf4s3EC*OXLZedk>Ow(vA$oYK^Tz3fw&?m zqXq5jfqAQ5{Ix&l%9w5nznoB4Rr2X?`BAUkl87ZSVuKbK$;ViegzasO=^A|Qfks@71SX>dpT0&)CQ_dGb<tZc7_k2 zM+UJ<=+UDXb0k%MOWOFltH&3Z%&3<&t1LrUI1_a_xuSzQwk$G0dyI!lnrSrE zLP}1|sm(_l$4-@w^?k$%HgWJ`o-fr7@l#k95U;R?r$=!e- z=mx?XaRQ_(YHU`T)yRKl6{ulK?Ef#t8%-4u#c;COvH1;xtaNC3aezLE=`BC)^SFkH zLSPsOOL}94-J;iSm=a>6R{c!I|6a$x2RTaBxgk3K!nhj>xYj?B%OxP-y`wmL? zMq47N;<1z@b01PyP`+6VSm)9H0%u|UN;;i$`~on^Don_<`V^1fV&?qDaTx(d7GI>6LQ9YeKaR7qPK=FV3t}nX*hR=&+Ly@ zoZBaFCCkHh%fvpvEYkDYbT+8h?qR=PraMXXvZ}9Tt9H-<wKpq}h-+hX}+R3{9kNSLL1F5Q?pI2Ty-a* zVTzB=$}tw^hYV%n57XyQ6^mxMHMzeIg1V4GfoQeQ2yuQr%S7Ej)$6sh_LI5nXs$4P zCSfL}L;*=jqG$fsyY3qW`%;!)j@n$>8R+X}o`>|KZ0VV>l{&*GGOU zqo}(Pq)yXx_ctVQO7#1>_%vK7ahUq$-$`+R)p+oQ_tbpLzAY_y{5xJ=S>?gaw?Pcp zza>C%O<++i_4#_?MWpzr4?Qn@4j`wOYjL>u0t7K7A2Fabji?hw075tVPA^roDu2Lq zJ^AVbo-Hp+P0yE@qPAKR+0Lbmud3~WK1y;zx_|lAr#rx2uNE;P6A>GPvPA1Ii;%K- zi@U~)v+`b~9zp+4$K}0t{GIOe>*t3v!#O__Hqp;tbs(wnaLS3>MgC5O@!VC`cZJ)Q z+OPTlU2Jv=?T?LLZCp#H8~1HNLeqtNY?^2vjwroWjj#M&9%1qmwPh)f&`^t{$DHMx zl-{=WqJcm9+HDDVMXrYGloV!kKe!pCLVFRvC7S2m?Q4Af7v{!&X@Zwhs8&$?3-J+R zdaRwG`cQ!gbfD~5T_r_5Kd(+Lx__ETFG|%9NApW@z#BPFHt{N2H@!OBr^qKB_GT26 z!r17c=s0RKJqNVe#;5?kFV3>3)(|H;$Tva?RN2NH?E-qQ+dKb);+0LZ_fCB|@%+v7 zR5$dyzy_Ko-0zCF1<#5;ET@Ejo#67DABotPK|D9DEY@y8B03+8qAX-3b{-{qp{gsvg+} z!WN@B4t3a3n=@n3M9gYx?p!C~9*1$H= z!U+X~{X^_)JfEP0YJ@tD^5giPaxeu#NVVBeY}j>tf1q*6ax(cg-8&I1&tZEc16u4Rn5_GbqkDtp^RggI)3)Q>ADrcM~?l_`DVRnAB6Z` zCT+}O!g5Zl4addCJ_(tA`K%Rtkx0kf(wMo$mJ$gZrs6*#up-L?1oE7F8EFQcqTF5d$UoRgo) z8{v(&Jcvg+5gbAuplf78L-Syq7(7o814$ZP0HXWc9|Y*Jio%h7@jU3@TK*jw=XK~g zsyb~XV#=Nutn`@GLJFpwi7x4_a$^Z(3{q;|dVIJkr^pP5V5!S#$WME=IC}`|JinM7 zKL!1M$uITvD0criPH@#0Wdd8RyYEm=0v!_{wfdPQQ@W*`UW7_5d5wI+WOJrzy^!j{ zWLKW6Fuk4O|CSH4dVISAT=dAN?aH>PlK?rjU}7>bFEX=Bat(9T7P+6(Xz9g@Mx-DP zhW6aQetOH=OS-Wi(HBz)9WI}_H;?d<*<&0mWQiJ#HeZ@49=5t>Lyfg9Q zZ?^(o$6{^+XI#4uJ_mod7GUhG*zW(6R|Vk=>OX2`uF!QiSc_XsxK`9_-{fn|(KM_R zcK#Q>bU=MQe4$Buv%WWqmMvLdq}khQJ{5xyu^PX+GV7T+=vzIC)ff15e_pvhTX@mR zx!->Vdmj8AQKR|oe<$l6+DHe8tCqait4DC&r!1s& z)z=#`UOf>WyCjJix*rlDjE#&%v>Qbp615$o$Gw8__j*;ywyTTlr=J-9zHlMzDM2io z-1{#ep4+wK^8SB_pOQL$KxDHMxOLh8X#s9u{Qmw6qRg?sjAk4vuJVk*c1?visNHVsDR6Quh@2BUG#B|lyt!%it_SL@o4H0%7j z0~{ZYA^{&NNM3I)6A(4-BxEK%|2O?C8Un-Dve!Zcd)~D=qY{!Z#+6tTZ7s7Ro3n@l zUTjT%(NP5DiMWOIbezZqg7rI$Q%_y?iawOU1<+XBPLSa zm!Ek35LTXU8sHq=TpR<@6L4*Day7xBY#9%UR$}dP*5;r)lJ>Vtj$gt=&(30lb&$g) zABb}5v6LPfzVAB|yb5}ykoQ~FH?0k7jkKE;GCUy|^KzR5eWdj7bq{6V}QpPvYc zRGGSu<_=*>Ihpz~)k@a*$2tK5F;*+Uvd4fYCK&J>I4e)=-j3FG3^2hw!#)T4!!CC+ z5Q!a>a(rJwd;!&+j6NI*((`LkiQn7eHPcHPCzx zw(rwT)Isdvv3^30!NAcUUjb)tJjk-~jPR`(Fg~eAw<@Y>d?>=_97QchMj-HSs4#m~ zLS^WFx+qT*;zprh*rN!e;-iAz{|Dd)Ez*%NU|A!SP<)2aUsUYzT9~8dL-sdsiNMO|PMX|Ch`i;SZ827tpYgA>GBKU+z%v@dLg|beI_X<$i9j zN-Jn!GPA+rtEXtDnX6tQx&{~oHm8JWVMlTwjT;WPGLcJ*4k>ts{(F1=Ht_?W3y$l? z)!~&G2?i6_v+Im6FzJmPVH%1mMTWX}%P=mI(Ys!Lpt?(1MaIW??9gT=A-0LPVO6TC z^RYue&wn=NaB{2U`N~}(xss_`{&klad=qVf%xXl~zma`aZPABT`BODLLpBa%xQ}bP zGO)u1#AhQG)PbDD_m`+|_g9#ML9*;eur2;k{@!YvfzhyW3^-cFd%j}9Zv{DR3A-v& zJ-Z1GUmv&F-6f2!(?wDqPtaI`{a3CN9j)1k2c^;pbE?9(M%n0gXbNOiZvmE;?3d!z<=#zhx$ zj(g&YF<9qf)7e2C;v-r9a?#;4Fo2S^V8l$yX-q<{$M4b@*;Gr#B7fx}>tarpnEnbR zi}+F}#9{_B@q_!*c1YH10cl(W+B9*?Yt&LxOww=jPB_bcyExn6r?pJW*8KL+k(VETUEamlC1n!(4_AB7<$i^L^&}dsnbrn%DOBr#ht~ zF0>+7t}(@7I{(g8?2o+X{35#v#x{fz`K<#?hG!E#@TFov3y92$ELfD&n`s)J7f|ls zj`=h{!V&s+@o33)?dyHirbe;LZMRT&=td*OUH7Z^RHXPdPkDrNK=1C@xB&B5PZJ8g zVv*^2hmgU#q{FA2JvsCgI~+4B|NB`?CST6?*UE0+G2ZxT8lM$4x2LtQzP}UZ&Bu6o z%5?q-JxX-(xz&U~^Zbs_LaZSB&Qtgc#Pg}C(w z^{MbNEe4JBti{Z4ro}w;PK*rANEu#LPk=}A5gn)4$6w8vO6c0EO@Sjk^-A1+;&py0 z#oeJdr9ZatO3>Pvhcae8M-UI4yXC-o@uv_4y&BQuLf~GYI9+_Vm-c0ufh`Y><>$YP zYKcK}YEf#vhl>J9paG5t_cQHR8l~MoOO`SsLX=&3VNk?R)36DW@YB(R}>{ z0@WMW&NeI9MYGu%D=gWPB1j4y;fF`l%%4apO}^1E@>kvX}fig>L* zJ)(>jsW0%jLtv@fckTl54L@{*g)GdH!7hjB7@HSTH1$zSQ0s;X`U__CJa^?Rf1Buc z5maJVIAWl`H=BEAP(Dbun9GSMuyRgWR!nkbIUhS3%K zsD6Uz6t@LTNImFxo0Ti8WG^tm{c;(8d(x1vGF^MFE?Uk1%ixE%klTN62s+H#MyPq zyyevixaiB{^a;60xS1`nAL#h_)5{qE-`M59mqdkU(tXu^Xx{)RD3y+m0YGwb3S8J` zYhNI6`#Ihd$08my;paF~lKPE2-dRF8l|?c6Q=A4loW$6XJzcv(4mh=iL=Xuik8x5s z)HY`DlK+62+s~SOTXdq4I#y&b!+usUx!gHv3p8DcddG{{Ww7h2>{GC%RfuK8l1RCg zGRAR$?bUHsKrMt!BS()_8}M!5rmQspsBFcHOQ1(be_Jb=uIR-Qe6sng0% z(l{__Hcen8J;uZ*`PxZX-2OP_s9A9ctkDM;QqRy(X|dD-k$B7D9#UPZfR&Zo)=D(C z|64M)sD)=Xp)jlrSS#|AMzCW#b3Q~E_x~xwWMVWppo zLuHxFK#f%MCNltt)kdz=IX&$-OqpyJy?MGFs$H4`F~+nU-_w29SK78b3yvNTn@JzT_1LmQ-}m-o+3#}zgFS3JFFMS9Gzx9B+;s#;tTz5p9z zBZr~|P>-;nKhOr~p(P|bwqg)KTJTQEVTG(1nA6?t9i=KlJtj-Kc0B6Ycd*oE)%$M?}sr@j^ zuy!4%<1aP$)CS2WKnpO7#j|0%n&)qd=e&vVt@^2MKH@ihabll7;9BGRzJ5E&LOn7c zW@6IX#l=TMxGoprNan1GE&s?}3AJPb|8^bRP=dOCVtmSEtDHyvRu#TZWXVW0W<($- zlc1mkPLDQM>BxbovftcL=`8GJ#BUJXFF|$57R|w`^ihmZ-Gpt0_nge=&6E2x(OZk9 zN(ZCbWk*#BA*W)9f~6!7RRl*26+VzUqm9ic!w#uGdwjlRvN~~;Y+S{CdMwPNUFWL3 zS^SDb10=At6$z|s1yJUHxB7DQ-jwqYP(Gf4!%iSSH7q{dZ%I?<4?oMeRvdA4@Aj(R zX=%a$#UVM;XTnAYLq6Yt5&YoGvD+WE>dwQHy|f+V${Qh#yLT#TDe=UBY2~C@L(*kD zTrl9P4(v}^Rx3X=BcPKap!u}J)M%LXn}*^V>qqHY83+BcFLms`&p)QcD4@7h(|2I-2->Cg+-}iX)TJ+J%s&`6efvC^Ooh7CGG>BqP0Lu|5ggaXquf6kTvrM*S3esM)l54e|Uuw_)=FV*{K77^be*F!K z*wh#M+{HYV9at%C(-SDXJv(=O=s1$^*PEMr8z5>f0G3m;bMJ|Ao zXT&=+O-Rqf9Qg6OW9LkJMBJ~>l+}a3HuMC7NQd(Fxi6Rg>>f0{FMM_3IDvOYT>@o< zPNL2DN^Az!-Nz$0x3B?XRRu(6MhqnFs5$S8e9rqCn@^WA1)v%BlYWEcou&wfz)-%Q zw3;hoN~fuxb(_=SXPgv24~yDeiRw41!5Ydrs%1uWz z)^c%(H{Rzens^9S@)?6F!G$;osbCI@sw!yQIrdGdP3hbEV*4|iC5g@rcL|rA)?z!1 zfu4D`rq-RysotD{jYdlOJ4oBK?Yd-Dz_tE}lO44%W9Z#>g0cwUgr5IF9r&G+ zh#m!TfledTj|$5NJnDcLkfEGohK^iXrKNuH)fzL+wTG*yI?BW;bs)|* zBzB{X-rtv73Ub6XBa`@oM%KZLbi^)8U#hZP&PPE2_2F~s{#5Id(OE*z3eqoe3g={= z!-4{%;JqOJooIrRB(}<~!9(v9TphONv*i^Wz6TJME;%Sk0huAgMNO>q=bJVOsvPNCsqU=M=aBTF0(yv>RQ{MHq!H&u`4yLJ{TZGstVj=wF(xh}Jc znYBPbs@%~p)AEQ_Kxz(a$ zW9!eTC~i&AgzLKescPv>STp(?NEW`3zH0NRE`OvezBV0G6j>a1DFp633(_Ptbq#DJ zM#zU%WLHk0HDrZWNoi1}2V@83xczI)n1S?YW{_%^RIP~u`KZuXw(ciF)N~{z=7@Vy zlO&~L%vL$IxDK*GdKT*szcOWYahAjBq0D`f4zRJy+jM&1xef-)5&2E2GwoIKf=QDqJ(tdmpsQNfwaM|peHVsZlnccoUzO( zQ{Ee?3u}Rk&6oWI+8!x*I}w#n%qOXpE5s_wtb?1}i1)oJCL&%VKjffObN@vjK}ql1 zPtTW3MG*22Jttf*oLd3;eD438Wy=}oRe7Co?5?v~=k^LZ){`WROp4qjRI*?VMqJSAt@&yZM1*{) z>xmJQjZT662NA}ipp&C2%Zk@t_5wN0(H*R@JC=#?laQU@+0bp952`tWJ}LnqrK(~x zo{o)_nxK(!W|}&Np1AG}d#Lb3&;D;^6IWWz+}0wKM8qGpbq*YHy)mRcLYtU2ib};^ z>`6`Xe#AU$JqR&L;V5YoqG9UUlrM~sBpLiubo8nU=ApRhH?lvzsZT=4 zCYC$m>dcuS-229IGMp-EV1bp?y_(8oW&@%5Rzxn?0wV5Num#HBkzs2CTbcZsqaInR zPu*lRGKN4D($q5YWqX#^2{$V-Kovj@M_j_jjyjntkr{CyOI?ixt~53}Pw0d7+k9kW zeX_!UW-7ziU80l<94SRPV3KxPqdA|8Xt795#(|?a#7k})s3J3NuZ=yPK{pbHm&Zyh zX5IRa;PaN~zl@28hWnwd2|jHmCJL*8yRGJmDvHvm0sw)0l8v@Z;%MlpO7Bx4%aREy zsWxE^Z3hjV8csz?;nZv_?CYTf8xw-~;$wbS5to**p?KBGQvTxvPIS;R6fm}k^?}q2 zO5ChEhn+yo6{0$nkF??7)O|Z!{W#(|{zW2P_B5K`fce2U4bzmLpgEzxi+s+qYV`F@4RlHjd&4 z`fM3iqn7fNfJ%U>CDt;qFcz%|g&TbBFRYv=hu@k@s>GrE<%i!P)<#~t3=^{!iv+KJ zGmp}~V7q)UpMJC&*OeWqB_<6cqwFN13w9}^NPlP=@ zXI_u6XmObC&Q~@DXPNl96Hym-?iAy+tWO{BIF?kQh{wQt~*<2UdLDY1P($VE6Dg)S~Cqzk+VD<_2do|+BEpQ8HVal}HM!q_1L zU15+S)j&^9jQs)uv1~;K4`4D@Wm+t>Sv#Ie8JX!1CzQ=5G{69J?=RQrOo|_F5yBU$ z8s$8S)*PLI3i$#5v7xAt@J(N4c;&__z|0%t>N!GNf1B z&>@k2dIvCeU`~F49O*g0nq4l1FE_5=&{9c4XDN1Feu=S-Q;7TW!X&=cmZ zE*728$9VYV+_key$1~c2feL-Qp`a<}IK0_KZOO;<)2CNiY0~kaKdGNgwyX|%Ccv-% z;HgKEGDk5%Uy1fCieVWmw^_HY<{xFN1>80zOWVRnPTsKHKfyv2uW{gOx3*i~GI?6B z@4;B*g`I!bpR4@BrwlOzSL+>WUvv>=z7^oBepqF*s$qYuZY4;Xn3bqF0X4ev-VNye z-QwO0m(oc1cT9K!<|dVDs!evf1}}Y9yO7^kx`? zW!_R^BxA*7<~EQHHu`#ZkU_hV-)op=bVmL8{sr=ScZNf|sj#~0E>RuOOYPCH@&2l%DnkY4AdJBYnwv&%v> zlj>A6ET^cN$BwSPNGC4VAGt)2v8~9+b(rH2xs~QPpy3hzog|*4b|{Iq@PoupN0oCT zgM2VdVYa!kTcSu^TdCor+oOB#a*eO4OW<9DlB!~v&>75J8Nc=Y&-tplu`4+VR2dXGCx2Uy%@SKl zpwPW58*K(w&Ne1#9gy*2HZK|$&-!y8kkl7U8J?q_ZL-6RDGeqesvJruMW0prD?g77dto3@&mAC8mG-7H@fsx5BHq1;UV zo=wX3k(DNnI_gW`H`C?m=*NFOy+KJ!bN(xX%k-vLYqUURb~l!wSk+NbE#Zh&NRA0~ z(SCS2h?`5a`s=XDoed`|^iztF<#<&pA~D1sOF*pTBk_96Z(`O{V-*L|*a4ze`g9P3 znnubSdcZVGDP>29cLFfn2d_ybR)X^(Ns7N+CLtx~$4J<_+~42EKyM(w@xKdy;q^|Z zprm{JkCi!7>1|}w1D%I`O(Ak&LqE8wq7tiq zAzG`wrRr^otX}kW4KB=aH;*iYk75vAjwn9X>y|fyC!s&zYk(!0Aa@bH!l%M70^dwb zHVRs0(#)2q9{NCxgz^V!3u6bLhs@^${RuT(hG~5#U{hvzGZ4p+;ZwGUC9r)VMlRFFG!@!@vnN0HJ8`*LaR1-P ze`DFZnSrI0zzP6$!a;`a5U~m$h$9+;FUK5}q-0)LHva~?V++4%$PgHg#0mJN#|f?Q z!ysiyOKdQ1X-8!B&HWO>K)@hms*uD*Gx&jJ4(yK+F%(|R!&<`fm|Be)11stYxRp3m z<|K?Js;oB1C5exP0!I4iM&?8eVremZXx#`F-FO2_qu(1Q0p3R|IzIVYb@qqj>(6x8#NNM&I_i z)biEr+@<#)a#QBGA>Q{{^3X~a^=gU}>a~84mQ~9x5bxhdK}U6R&r&CL{Wb7$g`JPx zfYcytHtNxIOP;wxAue}9V{cfW6};Rv89TEm(cbhhz!0qCu+iPKb!+040_S?Yj)b53 z2Ayh;l>VDskLLIB?zu$E3f!|Vo%LTu+%*_SnzHoz0G%^Re=;HJT?tb^=hlEuGD;&q z(S$Le?EC5bEV{U+lf-Vnb1#n{?ZOc}UL*7vH+zpz_n>hW5y0zJjh%31Xth*o4nKn3 z?|u%vK7ie*p?{*^dT|)(Nza$8Ip5 zccoC?A}$e$&ZE~Y)brQPs^*=qcqtf{2wuW>QWBo$5N)G z1xnRmJt^l>*c5wmoMzTg|Qk>n7OIUuxzsL)aYf*3QE2BQSlIwYBC zC8=w};4`>^^GVC5ADK6$KxNZAZu*iy9bM#lJK(&)D8;;O%Kl7=oei``laoWJh{#ZD zpwh&c7{>u0G7y1Cr4>q)ed>~=t}BdI#9R`6$dW8pNs-Lz)&a2vt(G0#SImN}z;b90G=sww{$#s%TMF!*sSFSjT?YVSUEO zlJLwav)wy_bBF7I!6leZT7(#Ra%J(^vKa+Fz;5(x#zYkQ5DwAE!-9v-kYhqhRTf7M z6k4(>5nS;p6F8AlK&A+!a_%J$B9N6QD~-%0*hL+M*iC2gEMi0Xw| z4hcd44LK;XY4NdO8py?d9(+P#nKqhQD=x3rkjliw-bQ};WQQ@DlwiH@+1t!~QZs2Q zPRLxZEvq3Ca=Bi(m=HN*lq9BrCo!2!32sEoOpfJTsZ8N#yAUA+Y6(dS5SoxuN$MZI zqt=S(MzZ(htjdB-GbJYl0z^(^sSu_{M0-Hsj3^~RX>w9TRru0vBN?sdwQA#hPPii=9A}Zo^3q6O>A7oDn z;ARl~>bc`5Z_fGnsxRr{kVqMBwj;|S(bg5Giz(OrNXnK3@P(uIdHWZ5JD9c6oFc0PA8gMCv%Ka)LIggFI13&&)3$Pn4zj_ zx>|8ERn%35kE!5DC{qsom^o`T_s_mO%_p+>QUToOHK{c}yfW~5SrkjiI&-*o$^;q2P+Y}s@F z{Hvd>ze>Fe4AyaaUjNn@x0iaUm-_dqmxzX!da3WR7MmzUMG2tz>ygn0Fi< zX>tPV3!6_#XiUXoQgJ?S>8)$vUYN*L|dSp0-jL zt+;n_M6=n`RhH|WYJDSpHurBk1I}?q_D$kCv;=z}{MH zSn-2QHhcY z9Fce*nKy9MH6$Uqy4hfaXT8nbUCg<=sCe~a%14)5VjTFz!y9h)j_sIHO0XFtQH-Qm zvVu87l8UCAAhIvWgKZ+^o-jcAHfSv7B#{cgw&7|ImJ>?7aXbIkz+~gfxt;w_Kle!O@~kh4klbWU=LC} z4BA&S@Zg;%?yhON8AsCzv*`pSGmF+Bg&>5HX(O>w;bUfuAdu{>r>o0irjp`CQm!@% ze=uG`D4!)#LMubgfva_o(u%4wAOyR;<%1{JtoEKH1&g+!X=_{>5V`z-!CJhtWCTq! zDa_Se%9W-~G0m-Y*dS<{hAIoJkBB_Q+dV==bUQ(chUg8^_1G}-?&BM3C8>-;iA2Z| zZCaENxN%_cLrFjCQVv+moI{%WkX}UyglIrzhz^n?g&k5%i3vo?7*o8{ib$g~NhSs> zxLWUNyAB~MOx5!LvG*p;vSfFa?{9bBMR(fUXD6jPBFEj z6g~ooIZXHjj4%eT9>#lSMIl~63DHQI%J#OWi4zeX-t4vp+waLtvQtVaK&rDQw>=Ra z5$@scYp?%rU0tpjti^^%N*X==U{M+ffz}Ne7a%!~n}*9SuqqWV95;OT-Il!*Sjx;M zO}V(rC_#!N!9`pWxfCuXTyPLQF@`+quM4EANrL>k$QXx+o;c0Z$uBf)w^?6Zk13Y-u0 zqo-*qw2-v@2on0BaQlG3lLR!S#+Zya&>~S9!3)PV-C((p79SNO0+~Q6jrE>=zo#%2 zQb1kE%*&2B{jSLiWYfxx1os3RBk1q~!ilUNDR?VZ)Ag=#i=T<##<3lsu08Fzr!x6E(|gasykgmW*i4f` z@JnC0&3AtA3=utDungWJML}>886Ay^*qA7!VqQoNYR#RK)%E#!UK^S|Z||~s&jd9l zBWU(Eaa#!P9A^ANUDtT;>H3jz=vg&}Mg2izHpE0CQh!SL;gwr61{;u~<&7s-ESi$z zMSWe)3i%puy&JgLj@(%azIa;m>g@&V{fH2P%BYXMCox)}MaEzJDDpmHl=!i`MtU1K z-&uCO!}@?xil$T?&kL46M90KWLa*GK^Y(eqKRoJLH2E6*(-dOjttVR^9B96Dclnvw z$v3*ujs9cl$1-@o{q1jmUfc1p`uo5C`=9u8+~`I>%PO?wEBBVXdS}7KZp>*z7ci4K zcf2UBd73xMqqt*8pNI|kI0+$g3;WbU`?Jom@*#4*wFD$T|Kc%1#F`j^kN%JzjizFy~<2aNPC0{cOj6-=dY_ z%Xe2Ss)GICcyL&7@38*Jg#anx>n|PhPw&9ImVEX8@*`6f|Lyx1{I}P)JZlr%F>=07 zn8I_p8<^K7AKL zU=J}7Vxm^hksJ~412HA;9IiMxT(RD)x!5{-*Ws)~7?(@S1VErP6=gHWhlHGNh+_zO zq*zF*vY^`~dg~ahXRZ|?W^;GbCzOVm^Xk;iqDq^b-Gjl{964V_3NA-Q|9;N52%_Wi&?om=SfffyWBSt6Sn+j(3N zD6I*uC&fsR6)EV<+_VEp1iB$n7{RQpSd@w<>n)_7VHi=%nxmQI!O0x!A{X0%r|TV$ zw}QpI;$T*SQuGcsdrwS~VQRe&F&no4ZWxJvK;w{7BJGGE5kzFOgCrz23X(F25Rg-P z%X)#3(C+sfG#ViUgNLdt=?6>iK`4bX4H1O%@j;cbV zWu7>AZhV$T5>vuCh$)gnVpd6_knG2ihKfzU=h?-Yr?Rc4JP{Rn?5r zX!J0!X*-AsGbPVslxQhv`vFPJDRWa`JByEj5G7?{P_ibtoN6?M#*L9hQ=wEtyWQ~J zcg`uaWORanrw^W(Ao+1(!0IVg7>P+tY27fTMl=66DUd=DMF~D5L!#{fP|*^VLI+8q z6Sr0sWu{PDD97phei$sKD5>Ww#=gf7o?UO*^_r7eiDJN~ zi1jeW5k!fWf#f;{`#fbW5ke4s!r7h_hWt86kP=dA5Rv3v{&ExwrE5~mpCQCt?iFlK zDyNXkzi?Dl2`OO5Z0zg8(Dj~&XXhvcQVWs`c;_b6hyfXqQ@Afq=5C5PC9HInu@H7o ziVmT20bKIa{$4CH8~A-^m(1LaImH_TxP6`os1NhvjO`(~yl5Zu>0v|LJNytpDxT~-ZLrL$f~^iTrK0UdE?NsH#sHToPj8ee zu-ayuaLS#-=0jy}tES{&(QtPbxpPqQZ09&@J?~yxE~TR>B4v`?o@-XKlA~pV zdEd4C!MtR%x16;j*68ccvC@)3<$}f#KbT-mDe-K*mE`b;NO*qj z!IDL-SynAicb>-=8#D^g;GogmIjDH~*6agj ztQ3Mptx2iJ`tX_WVfKT|7@8lk+o89C?>*Y_&SlT0PwZ{X6Adyi1g|Q|-GhQJ-&=fY z&&21gLQ8)B#TC6>^V->l5EmRRK0dHwUEl{#Hyk$bgQmMr*bQf~u$8)@E`%E%+SAN~Iy z9M928Qk?C%=sK=CPY97^kxjkPM|PcKQ6|20qWJ2)Ij?>UzTrpJ{iB*&PdnaN_Z-eE z3N0~GeCnv)Rol~!o|o=4%&Ye<*sHSOpjH?WIp6F#S8|D7i|zvk@8AOe-#?%8e}DIax3>vnY94Pq9sz4h zsU1x*uxbj9mlbLn0n>I$Li%sI~PD$*DP(-($NNUgG|5EC&4gsw@+A&_kMd%D3gHxVNOGDeD`rYc6p zQJ{?^22ff-h&UhdF&k!D21=vhx?fyiv|t^Lsa2pp|CVZg73ad{(pXJ6u#~li9c=HEmNyB6#of2`2&3F?0bz zAjUlQK1P8ZEkfk;QB{`sF(9SH`@E2VEGKiFoElAG8viFhCz)BBi_|1Z*I;*<11s|L zMkh(qqfCi5C8f6%g~3`&3M1C}X~e$5=o*AZX_3nVW5jufA2M4Qp;4x$9doMfY@5yZ zWo6j)J;8fMYmr*#mo5}hnCX3Rq>2BXBAIcU639e!J*aGABBn(I$j5w089_lnsv`f{ zkEAdXQMj?gg+vOT%|4e%c72D98e?hTQk)9P#Dp2p`8=0geveH9 zzRyK&rXdYiQzx99{?AmzmNHf&o9;R7O7dVTRJkCzo+vCzMwEm=M8<$j37sT_z_K>% z`-o72!?MDMfRY8L%L=_8dAb?c_8wD6e25dDSkT+Z!*x%k1$G=LL}Jq=w&MmZa?xE; zlvIUgjE?23MvI=^kW;kv`_M&NN)(<*KFH}CXAu{fuEEB)@#REcsv+cn5AGefmB%W*>HWj>b_wWP8?)yKl(}M~h zd)95wa#m3mx#+V{j&XFrdyD5GCLWz{C}r*qxP9=!h4qvY&#pS|EprLzv&~35IING? zfsvDC`GLjY2geO}4r(5)d!AnPw1a2xkz$%g@;-9eTAsCu%WlMIMQJkT#QMN%=RGHv zmbcD(e*VSf2heAy%aX%-K|vyuEg81~{CT z94yU*JxO$<oU9rab&;_; zDbd@&v#Xv*7dwt;a8Opbl@_0u315+Q>AAxk!e5ZpbSanMvezHE82>3MXqC;0&tB?pDzg{9>7vf|adbC&hz zKjL;YFL?2|V%_g~>)DRm2Qz9TINN)UKE_K9A@byMOKBvx59Zg!ODQD=>v{Rs0;3~O zHzSWP0>`W7x|RC*|7*v@K@vP!w}=$Ew^F=%Xa1qo&cD`jW_b4^@oYO>|J~Kz^8dfF z=D+^)C9hpX?%p~?3Pl$T%4kZZ7_8%B@96BvU;}rK8VarG$3Q^wv=w}z(HzWc5MY#I zvmfb)p0h4x)}oJ$eqcWuuJ$AAUB{j2hDb`uXf31lw8IV`VbhKD-N4Kwx>4{?KX}M| zK4b3$qFA8Pm`8S{A_z@PmN?}vQ_9#F=Q2h?Y0c1gx&6H?P$JL{`T5>@vgUH@(1j$% z%xMf`k5l>BQ`!(wbT;BbM0Pzx3iuf5$H>q}2$m3XDVl&Op(XNoee{gp6GEVnikEK9 zS?>nkJny+0fF!WaaXbqsQ{Yozo&ufCW+zJJap$5U8;TJb$7~h_MNt~&MsatkDHY7i zd=vb|yDPr`Xw7L8`S}-@ynWu%cA07U%U`?0`KIUXi-CWBvSWw^4o#t9a0%xf3JJmD zyho9FuLwg(l8^#X7=##s%;!xhAW?|uQM$l8kMowjOK6c3dXAiIwl30-j)119YGQCm zk&x4Zm{jq))Cw|ZK@}BB&v5-F8;`m|77c_zj4mfET_8z`5Yu-^%4t(%AiB&FG87bA zAj8N(DVUWKV-%5`S{sLs?Pzm3Rw^JST+HlNBg8Z&4nicP%-)wIV6CUn5d!Ro0cQuc zuEWWihy>LjiUL#B*M)l`CPW%>!E?FY<73E(04XS|ihdXwt)=T7K4}68lBCdOPKPQ< zS>z%v?;Viv&LV7&P$h`OZcL;x;8KB2M-_3cn8dNq)W1o#j-H=R~i|t$>ddIR97%j0j;X^{35?RcU)gqgR*N;d(ew86O zi}MyGC0Yr(AbI1g=gEcRcsXO<)D%}cuKJNkflm@$&X7tIgCk6)ZcLMDtqKTcDkKZw zhKx5T<{%6yynid?RBo2HO(2L90wK!8<>h7mw-~cOj3kI`SUXV=yB#5VP<8$cq{>ND zF`a+>NJ<`&(!`Pmg3bG^)TFQ{`II@lDT3{QK7XN+%kJ`rlgI~NJ{Cl=%mrssp{fO1 z1`3sNB`JBd&g0&-xt_HS#v!xBW5z2TA1-ibYxet=swyZ={>~RtP-w6|?@f6`-sKejglIt- zWbpXtzz2{bo94C7|c3 z>*=i~gp}FrK4X@g548QraM`jf6DJ2HuiROFDCF?;szXGNcY?ovYs02bbS@G?-qmX( zS=C}fzAdit3R($XJZX6GBpb#l(pY2&T2In)kyeSo@%ZkHU@e@P2y?D|vD>VlXH$1*_ ztamLB&m$p1ZyoL463M8Fp(wfR6Myx+HOp$tt)=1BJBuGGsePjx-RM7_K39YHJKy;Z zfBeUP%wPWHUy@SdSAX?a`K{mjEx!KsuYc;#{@Sno8s2;U{LlaV^W3Bx-RMR)`b=6i zxe#Z!w?9<-@B1~nfFy9xl%Hzc{VYf;ffU&Q{Vb}>oR~ab)+`FaFT7Ckm3vF7Lb0mN zXKw!etop`_E84;F&rdCnE_Tcc#aZj=5x-3||k7c*%EVY5tRmHud zl1gnk-#eaMY>-k-W21ugcF(r6Y&yri1H;MO@Wngx4{?V-O4%@x3^rkX9y|Z`>udhu zv85d)N2`jvhjY#*16WIm){?m{nbjqmea~6zfsT8}bI!N90A{&p2nc%T0l|~aK-cFR zZ7m~XlFUoPx_5M~<=h3%uLO^0Cf}&Iz~!!Iv$uF*kh(-ChfxBQ!HI&M9q2YaMoJ_K zQ#ADAsM}c5J^%ZMWQqcjpJZZqogLpL_Y@3w~<8&b{x?% zq46XiP*P;WO_Ix^d>|$$i;7taBo%3lq%?}AEOEipb$j}uqmhACAvl@AOGgEFj~b2^ z6;1j4JxH7`3ch-8#9#YI{&EN(^QDxji%4P{M zk)lTeO1yw!E<#0lgOhtOm%oUm8+Z1yk(X~5LJ5cD1EdYp05vCArlo?ea5Lh3k zDFI2;+29al^Vx@hP7a}KT$+qztx-aeTsDwx0^d6-lMB9FG!qIUo8g6hdgE%J6z2B_Wt8?Hcp1QGo3M4Z7fZ zRAR#Gq~Ia$^R~xaycJ>~#1WNTF0)GlRW%4zkubPnmszkPKV~AQ$B`2^R)EkYG1_bx zdzoLqD3HRF&?G0p+x-3}m$yxriNl++x%qrEqnq9@jS!09GH(>A5lT{&B`HQy8X;!1 z&YhDL=erRR5-KH>NEi{=cLP2|>O%A2bV2VDSAC+X4gHwe7nL>)b^z}&wc_Y-g^Pl# zz0Kymjtry6dy9-7(T^x)GJ2@*X}253qQ(>zKIZ2ldD?Nr+a3`trYLCU1^ti_20$*q zlNpbbk_Tr=QsvKD$oy}V6RySKgC!-!Zr{_nL~A9z7YxbahmOvHC=Eu7%(X6xoUE=T z-q~D`=Oc;0=n`j}o^Ej1kf^H)HwIdlQ~m#1Mk^MTp)#JzD74P!KyMwjj5q-<n$hKBvvrY$h2!On-Z~y{E!EQv zU%G#Iy=^|kTt2AYyPiCp6)Z~0=<>A7RXebnS08)LnJ1ML<))_{0~h-q5d$iDo?UK< zA#i(D@Z#~D(|N(kLVw6Ke!A{>bhbka%SJW4yK^k+l3876zI8~nqvO#N%j0d!lPk+F zyuA8I162sY<8{w>-`(-Xh3Dnl2h2ZaF~87?`^R%)fdBXVm%RR@<;(ZxoGkKr@1Qo^ zTSk8Hc+0Ez4?Z=N(e{pa&vv|es`={uS_^PC*aI9+}`%0LLHD?^Cz zU;k;%s#e@PZqQPpM4~d1gJ#d2rQs|07o03U{61M9*!DiNjAdpLuNw22?k$cN1^?wM zr##&ZoNY(8o#lIvTb@{t)RMZWxOdbruM8~S>9PY9&apPy*uYicNahNeIhry z(a)Mb7lYS1$8Y}TZ}O*q`llcG`@j3UzvElq`WCfAJT8fphNj z*qj^P=tejCk#%cXan!WDelc+Cp#Ds#hG(0OqnY~O28B@@MN>$+wZBefd=?GXbFwhJ zd^-Qc=H$;?g_iu%tH)@y=FJPs`F0?B$2-p)_fHqO#K?QP!QToAKu>Z z#)U%&#k?ueG2a+zC37I^eWfEpz`?xcs_j{Ko|Ja%ZA3`Tywq4XuxST&HgVBpA3wO-oxk;oi+3^5<t4jX5zI3?e0Bin!oV1d`GeML}6;X0tiF{T`zf?O@rqJy(51mvfp~Mt*dD z4bo87nHlJ;r5$^Qwj%_vrlu(xiaMuJTn8?V`SB}5Q^}khjDfqy^XoFXe|$GGlb%(b zU+;rsjgJ8>CDx5RJm1i}1gb>q5>;frAtdHa!Ln&4531l`(QvhCiN53XK+`mu!$rY^ zllNldgY#$w%f|3*V>xRB%O>wzkdYsUi;{U&P_+ZWTIxoTVn7p6!D2*07KWs1lBoy? z++<++WFbaWQRh)&>(El;ZA(g?rk>MOB?v+ABiNoXIAT%+A3!OL%G})12V(G<(Hy}? zPl|RzHRRFuByt)~kn)uJbEW6p)tqDvA&~5d&}A+MOF8u{h$tyhS~A*!T1T`L>})1x zUv7JPmq@atu2N3H1y8Uex>^u?9*d9OA(g_1T&O0A)F#oChS7$M3kjaG5G+)vTf85sYl9CS?>tc`qL3shh(Th6z$k?ekw{*sd2V;h zl+(^Z$K#GEX?6G6Uk!U%YjS&Pozw^mw^oCGVgB%mOMC3$e zu%!6Bq|88a7*zoZsRiqOVC)B+btqXNB!rZ3k*S1`2{|B$LLjEQh?Ejiltecm{D3Sf zs-j|uBT6f_?VjExyr?imUR=s0YS{!7+GJli6ol*zUnIfD$L}oON5f7{tEn%f8OtqvOp*eLKB>2V3>^gNJ@fjKl1o; z$KXNh5|SqPfkh!Woe3VDUoo#MOren?5n`kt^P-&+lHfa{%RJe`!xb?JtP6yY7<>X@ zkTNm$J;n%Jlx({Jg<#o~C=uyg&u+6vYYj@!cO%JJ8ex$-AazM0C02_ZxK0VF-rFB@ zkDNrHkS3?pMRd7@%n8Eiu_2*E;BxO-tquq}wML<()_r~ZkL(&=Htqjk@w@3?b%Oi>uRuA|TbfnlK|g-RbV zx&{~M$AJ_A(kQG?Yi%Jo#W|xk8>H<;&p}41z){0{|Ga+BE-a# ztB!ws+%dYqfBw?ZhfH1{BLDno#~T-p7jLaT*4zcS*!DcS>{xf6&8T?ma^Q`%=L@GZ zW~JnymUQDlr4|43&YJtjGY;nEhpw;UgqJwqbi8tG`1<|DXHM_FdA8@_m8Y!APc(M{ zE_OX{JZ(8^176l_Jlvt+mAgwqOzcO;hcw$^*-@w5+WlTZm%l-#~W+PqM)obzxcwEgOA_?KRAZVUC(Q0dq%gz zg~-=mSbe7V>P9!Z(Z8=g7lZek-~1+j_Gf=~{pVY^ZZQl4+wB&BKl-CTLI}Yh{^1{f z9#40p8{OzeKQS#E!-J!eXKlx$%RTpxe>etg+dDSxz|Y?;uSdxKy|n8mZmN$IS~9Q9 z4}CHCS!A^2WLZ#MICj1J$TsMYQ%Z2z4!m?AIsF6fYKOqFS4jOJ9Gz>N; zFMLS3P{l@Gxz&6kbNA!qLd0mruJ^pV9_gImcvbOaYp?Mbg%T86(ED(m27Rv8tmb^v z;dPu=9aU6i6x2+EA7` zg(mSVD#Jj@c8D}cq?9aY+5EWL>`2z3b!K$xkT7N%Deguj5v8E6N)Bg+rf#UJg3)=5 zmS`!txZ1EC1+r+E&GOi{51zWzI3F0Tqbv-jEU?aD$>xd_0zWbRj8=G`Z}LS-95k9+ zs~YD#m-`W=_S`#~^K8?z*^P)=aMeSI66-aEj<^_!&M~?XrDjxhL0uLYB@t4xcL^8e zdjZ2qLMdupGsei%tDYAR%WQ^BH>2m9frDCbx-7V8Et?^6didUAQA&yJJ|_`ZbIA}j z^CmM`*V`RI2#V$qFBH~CoJ$0mZxADprd5KFB6zZ4Wol9k$e2@bRhjqCE=FKPiimR# zAG1YMl#-AV*Y2U1;=9q4PzWif-%A2|LV@a>T$~cWnU$JflSC1mCwV&=z5*!(BIU#^ zLJ*vzHkoy(lw{KlC}rqfPWR3li6GIA9_bu0j1WttNbiltPif!~BXuch3WJoAHdyR9 z5L}=ziqqASZm>w1aT_uEyp>9!l*Gj}rE0+_gEj^~SO)7R6EuLzMxs%Y7{p{c3WU@s zA*aWUL?Yge*`SpMZznFX%Eqo#3K8`bkgP9y?V;`9`AT(BPdM4OeB!G>@UPTg^^I8 zpD$j7yuYTLN>w5wOac*flR50B%!P3Z2%fg{gdi!Z0#7dX+YgS>I{Lw)wWc2})l|rJ zG_OcWbGc~=!BUk4A$lq!kXAA{|B>g|lf|6l`Ih#=ve~vQ<}+df>my#Iyzn5x2hx*Q zyFGRoNJ&7`Aap@Y63|GkQCefY&kH+)MG8Sx<>YaxHA75T7f3Ou`45*3SM84TZO^hX z?0QF~6DNxTDFoknd&3*gIv#BT8)vBNny0%xk;FTfJ@dL`+dCGu<<7zELz7T0y zIchKho^5+V(5U2?>&V5n<>BR?WnEy5B*naNx*t3THGJ((!8cx5e&*udloAgwd(PU( zmmhrc&tB~UIjrlhXVch6OH~R7PV;j7G^EZEUZSMa5@Bco(^E3}1V&_;=y&hM0Km@s|Jj{b#JZK=0zj#}zb%;CN9pbgQS-TuTfK1H@b>x0 z!^=I3+WgST)XQB@Arp5Gik}E;_lJ>Ik}us|@bc-5ufH(z-@kvs>(6@5_JPBu;QmP- z!;@05>m3Tqo6oLLLU4PjcyLs4_wb|B+Et;cCgb<5XPTp#M*j%d4(oCW)UqylwjDTY z1Gf)n*I38riQT%;luEGe?KNK+^V7kTG$$5rn^1s1c0l%~Z7 zQy}TlMkA%b+ZLfJ^prC5-sk6%l7zrT*zH=J9dUh2iUGAa&WHplv)LCV-cP6kvau}$ zSf7Z=QyWcDHjI5o9D7X2r~y+Lq8pG(BE^)T_W8I}Qe;z4f_H%sJW?oxkCS%qqGr<2$lh97#$MLqO>=V?D@=5<;jPpiS41gz51G&N&2$-~yvh*eFRtX9KUU z==+Ws)r3C~`E?62-&~N99#RoVq@**lAb2nf5CTayN+TjCd6k(=?0|@Q#}ZK@ltGAq zh>n!T{9JAz2ni@K75EV$p0jdcTCvElN6E|t_G5mL!I4sg=yExrgp@236N3~jA}E9` zNY8ijMBX-1X3qku$O|h(PZ;<4ds=FuDv*HI8O>9dibTSC%h)+^d**5+#u@9qrSpnz z90;Q)1sJl`LX3<lCN6gat67OQhHkV3 zER~8Z=Z5`g8D03uX(J)wi}#P&+owF+?elRdB}rwBN=N}yN(!BeKzHrHcDpC~0a+|4 zOpTBwy&V``M3}s=6H(;O$H;zk8JkfSNGS>4LCJMN+N?61uM?a7NQjAt7cCD~n!5)T z-+uj)?>`+l>k@Yk8(z6r@yppV#4w0|w`LqJD%S1D z!}Bf65{??3&y%Cah5S7=s|BwdXzm_XeDThlS^3$mk6rIscb=lq*XQ()k_(A9o@{us z4V)Y-m{obfbXFTW7ua@=llLWu4`wBWQT*W16^|~SFslt$gW&G*jN65oCc=6K7YQ-H zp4Xms{LACS+1B!3zH-DDZ-06SWpIJl9=E)+_I&xl$D6zFS7kIW-(K?O(<=^|9d{1u z51>MBbfX*n$JLK*@czl4{0RWR`@6r(@BQBIec<0-e)(no>aYHaU;gD^=5PP@Z~66K z|8*WeeoR&U?=il7qZ{4mr&CR-`Nc0B64E8FpLcw}v)nzJF{{i6%u(xd3+>gu=amz~ z*B|`bq6VI8*L%ME+EbqG;kCyDMr&qefl`7oM4nzc-ni&_v~j#}*z*go9Q;uH*B?f= zS0yhTm#q7ax6XFFbo%LBvvzPizHIr*t%5J#U3@Ov@#Dy7$^E06|LZ?J;0I4zUVGZ| zY!|rNbv#{nAfWFZ!4248`MDPkxqn=LBsc9}%ep{JiQ9|G==qE&i?i*(VIw(R7CgBc z*o}!7Zk6=bVI*{RVjfDtYG!Ed$hI3fn0;{DDx(>F9)Fig;C#S3k3^$XWKl|1^ZYXx z+m_ZQZXX>|6o&17V6;#ggH}vT$=sOz(rIoJf3zNG2S@M$<19&NE-%-JY)v7x0+-WF zMrxGKEV;r!S(NO@5~=ddvyUKUqU#3=mB;%x`#ob6+&f)T7KXh|jLs9h=Ww1`j+c8+ z?;J4%tOH$CpbX>K6Jm#G8cGF5MJlDyDzSLq&0bR$yz!(%q@0+3`LyQ2=^Q;3Y$+)z zwP3I#MEfbF}d&Zcks+wW6Y&UCKn>nsR*SJVVA|Qyt z;r+zgjfo@;Qe|FkOp2U<^h6;EF0$V2(T(6atJMdEj}rWVcL731321Gm!l;ZeP?9M% z>+@)~u0ZMhoGBn=F47ahA(KZ5b!`k6N+5KB8$3!zgphcVkCCD@1UnMEV=v&X3(LB* zxTrCOWbgwQv>25^3MpFZB6Depf)IL?DUs4JCP@-G6)TlyKMo9Ij~_acvjk=EQPK{U z!TRg5SSeCYJQE>_Y_bZ050O53lt@G&(ZLd;M^35NY;q-(mPD6Ju7ox@H42mIHyN$E zK$r^Lh=_tHH6emXmO@5!S+XDVXQ-4!X@$3uI4z8%m~jp<#+)vdCL6860)bqtR%aK> zPvvT=09`{ENl_5c49??rJt3uOB)%m09`FeLoSB?c!9ok9%x(8EW$tnY^9qoH6u^0( zDG3Qg(%FPXpp`;Puu&kD0TO;H?n)_tmt~}zBt&st97jqBDe}m_clm+!^t#ey$U=T? zDPt&-01-0lSt?2JOgIH{vR8=d$1`&m^#5Y-O@3^<@_NtD>UMXUXm8x!O}Zr$lc2`~ zlF}=rDi9P5P)W>?5Kp#KC$+9eSe=XC=Y)8jt=9?s~mg>U6UBQ zQjfyANMf!@!EBHc>hXOFY%>Y-6i4T@5F~ecj0l}$+deUal}V_EC=fDYv`T@mMiB#S zgk$i6Fa*5sd5}p?+vK>FDnsN2MtPL+NIwvT!i7NefyH`*k}5g$jV5@3R4R$BF=Y)% zRWZ0g^gS}Pl*%*FlFJkLsb?y#Y-Q|D70>KX7+m0=-dXYOw@=9Og1uQsA06wuVXy%k z^f=Fx3)?%F9Q?im?~+5>OZ*TCegI)mI!9%S zkkOJGiP45qCoy%=dYo@5votsBP0jXRMUiVJMdFOFnt`p!hyQlZtDF~a?8BWC99+{;*@&OJ2M_KSv|n+$-@uMIF9=zozRv$kdr|=#vX~ zZ>wOr=~!)AGCVI{&iLB%bN8!{3KyDO;Q9pFu+76!HLiG_r zZVyj3oNNMBS)Pl#kXXW9Z|R@FfUYu)QkvK9te6y@|MkbOa=2)?b$`ui6Q!`GUKPf&0fRI+sFrq)4rMF?e<-l2@+fynJo;g)DxSjpZNSIOg32 zM@3g**@h!&{~#^mJVsgqJL$raPGWYetanp4^#();wD+GvbXWSL5AS0R{A zCggcV*S7>ekf{`?CZp$MRpXsUia=!qZErCIbXjn)RUxFLwH=FkV1KIFo)%~+5TfO@ zNv_{aYm62cnN#~V2^T&i5ZIb3uFOhYgiYI$nJEaty7P1nf_J>Io3lSJan^HxwdQoy z(OS>_(}s23&~=Vra#WBE9;QlCW`@CgI-A=2qxUT9KD7l0kI^Evf+JZEk%y~+|8aYT zh$p;wkTEZ3ynEPif8}`d(6a0Tc@t8bdQM2zY0aLZYNdg0TRtHO{+HLmCcKCwh(W}=2JlHt zj_E-mBvEQ2F^T5{WR!^m8zU|xTt^Twjg8bP%+IImA^Hw+?rcji5n?2sokM5S#)MFy zCq$Xt!0EFaL{=sAMD%HL_%@k>^9f372q|PN+6M0if*-Fx5_>oua}vvgPmJTHak!X- zT1g-ZL6nA73%g}bJp`hNc-u2Lj}r!+8-la=@w3~`T-(ZcYS-}Jer%teNj5szJ^ylh!QraIn2LiP&8>$mAx6COct50T zUNoI!bX;%$^&8tsV{>9P=44{KvDMhNZL6_u+iGmPag)YLpZVR-`oEfYGi%m4=i29s zy+0=)VRXHz4zE&VObE=N3keVkXtf|^R94JEPzJ1Z=wyj>smkL4MUv<1M?djc1BWCx>4{Se}T8d(hY5txDE+p4Cok-*gw&(JPR9<42t7Oh1-QBoB{b9-zC zd^UC(xE@VY8bt+X3qTWQRq~>+D{65K7c~V&s79jOE&31kRH3Kh>4{i^E{27y{T;_C zJ8F+lr-|5Ygnu{Lv_!UhC4@ThymX=H0^Xj3#o%gpgddGC?suL%;rEs}xOf3?)tTdt z{S)1L6HVIpC63K*k1OeSU7_2~97@BqnSLuQx)>pMDE}=L71lcTw3^fQ>yB=9bmIe5 z>yP%I1ja}6m8bKKr+>SUu9=C&dG~u3Ns7Gai`KI6%uVnz?(Z6&vcNC0o+CYGW4VW6 znt89R99$CZeU=^G3r@`hXS#@yONgbRM~3>h(sSbUQ{a(iE2#5(JZV~FZ>H1zgP zQpav!iqVTN$DRvz33Y$sQ-q`Vc*;6nd`3yO95=^aPSZcS%L?HcVVjsPqHghr`S&gs z9O?Cb_W4J$zYUz8&R3eui|6&cPGA^yE#?d@<`ud6li%tuiGR>m%rV~f{$0`ea(Q#} zum6;Yi0JmenY+)&g~2HrI8y4RXZKC8)}ZfGYxng(rTN5K-cyNIEk)0$GIhZ#ZwvHAjlnhX`_yx-iDah%lQJVM-IOCxyV)B@iARr-#A#e?y6qBpcQgg z!PD>Vo#~4IVJApHHxrcaYjop0@a;(Ak2A|<@`dUS@3rUsH9*`hBQ=ur1gd$ZvE19d zWAAthzfR!q1g~Yj`!??nk3-R!;8e4D*bVZ$LzGIndOexiqoDM<+4f!?_V)gN_dS%o z_YQkkaI$tJ)965l=HFU2u)lZqlTSpbRJaAXTffa2C_oVWyB$T?^;9bB&!&YRCas?3 z+wrPDs30DR)+{@sR8)w1(&O&B&nen)?zCIq&H7oqH|Z9L22i_dMxB;%-YKnx20>r& z!3;eFTeOAbsyLd_7Vu}8|7;4Rn3}Wy^SZOpm66mE+ZMiY=zsOJ^XBzpCX@=kdz&ObnD=~ILZ(#^SHO#B(Q+NOlZ)-DV+br4jMPGQSg z@^vckl4OQbx0|(JdTON!!LuZY#32AAizwmEpfk4Ct!Etzv~cgsrdwrG^$!!`G)0<>Y}239*7FhtYxwzu>kBlFZ5xSj02=rTw$8 zn2{4mZ_djYj^etCrKgxk*m}xQyi~oC7XZhpVt&+N=QLdb`znG6gh;2TQG@2Sa6H5 zjMPV$&#Om}iM88E#9(N;|H8gS)8REo;pI-4{fkE|_zYv+<12doYu!*u7x z9PsWZmwGo1f&@AZ%6zk9<_SVNg)~4Wv3`3BwWJt;d4IM|CK`;}JAy-HBc_z04QJ=$ z^)G=)t;ax+#2A|+AZvHos#P>0#`5MiUQEaNpj$ErpwHO#yMyw}94UDdu|QeFAJhR0 z6XcATgF!rv0j_CW7!Jc^qHqon`#K?u5!FPex(Dxk+*ELX@W)I2VnajA-%PmtB!s@0 zRVEP6w5ZNXGv9VH%i_XcA)MlP%bMEcZHwn4$+26@r=C628o}e$cJF?(6up}>+nXMe}JlnMsmCP}d8lD`>P8_k59C&HD`PIegciTsQ_EXve zxnX#}vEKcb!ck_At>5E_VDHCukT=6vOV2u*O}y~h$xXRr&0~i=du3?+SKp7f3Ihv<-(p5w~CMwf{97x*NT- zyyDU253|h3BqB6i^#1yNrN=^lZLIxoeC&|hgms(Fb#AlI%jkQ}$*F|j_BlWh)}xjjk{3XnftLk<++ChZ)%6DnoS zp@qMYtjZI|UE8emsJGGLBq?yV)8CK~+|NNG&(PqV!+q<1ye8lEmXqG`Rlx1=+j%r!TfGg3eA8xSz=qSQ=y|&b!n}K@ zqC1$>9gOOSqTT71&&IwdAQd*{fq#LiR9A#S)Jlcnt$QiI{M*07rOj}Zws^RhiT2i~ z`7382JXmgy!3u1b|DY>uC)7s#!mj#$iR7)jZRK&XA0~cS5gzA?s|+ba+3sZ6-kEzp zxO|SKhxtz#hqC)|C+cRW%JXT7tp*X*jh!Y`I<6Oe-6bvAbu5^lIC}Kn$=dT}BPQR^ z4;P^5dFru8YTz%z0kTnQHmOG;UqFSjwdXB2?3>2_Z~9v&)D!xIWJg@5#Vx4mMoW6o z5sJr4pf}6i_Kiiq!sF2}2t9~h_1SQa4NbF^mWYYpF2B{rX{LA_5OZ*%ADKpcG3(rU zouE>vhRE_~#p(1m!;^eH%luLihx@UN4sZrL)p>3StC&|7*eGNO1)`ED@t4iq$eL?L z`*H)BEiZqg%71_T$SkhaUg-w7t~5`pjH94I%bfe>;G*dVUTY6+&!8pg@tDMznAD4mlSINU{7h`3 z8toW1x_ReL-KOhRXsfjfPWs0s+qoEYiq;g6s7j{+387q^5aTzL$BlRmJna9fAECZxn;xG)Gr=pan77 zSK{!hw5K?(Zufsa&v-&cRG)Ci3C#adZ z*=bV3W)15>tNRU7i6Hd~9OZP@uq$f+OFZVU0uAhBD++}+8_*837>8%N z>b%OuBAtov{h7{O65LLX(g#c;*zE!&3PhPOT1kSVF7}BTAMu@lmfK-)sGc3uDQgG(wvqc1ud5v!6PJ6fSeF?a*IMnI`;+C z*ea2khBHL*(R(HHJ`~XDFzRQW?B`2SQ}7~pOj1gp6773>_8gWMweQG}7M{}6fMz9Pt&VVJY5Eo=UbY8x9en$B$QZ!?q)U{ z6UkikEoI-!-ysxNgGJ=Pd@r{*UDjNjFlm~OE*~c;13rm(SAoR&P1HCmnu_Ugm>o0g z^@qd-yd8c$4gfTZaMRyJ2|NIdqQUxpuD4{joBL3_@Ujq_%S?C#l0aDJGgV13mCr73 z_y45e>UA3OX0`Y6o4)T~b$$O2*WN#=b?^PB(twwH;eXD42ef`ceJ|O4`z-(du)MSw zytT}|KH|D6UVP;t_S~i#uff2 zC?kD0%vC-MyVKrx0+vT`77o5jRWD}HV7~^SqcQdVU>V=LX%0?$PwZ%0k}!@dG2PC` zWqeQzfxO+*YP1VY!Sl)TzN1wa005uegBCjo8f5S5fn@>NVq z7)=Y_rbbh3a~|C1Hg95E|6zY$zjSallun`J;D7$SlZ(Y(18@pOr)c@@ns%jfdl0x^ zQl*;L$IkDb1m?#iS6ggrpQwk;Ct0w};lBsLO`)v7w9-;%YHV%QegBqbW^!(0xUJ33 zD0a5SvR5IiBbBUtgg-WglI`~MDbZ}(Nmu6&<1e?Nj7ky*A$A5%Wd<)aisJ0HCKKxo z*b8#_;9u;KW#E<=3WZlCBf=2u2q(A5K&O?Xo5<_e8kp>z`;#6RWrFJaOGPUFP;>3n zEkV~M4a^jCAwUG1u_-1__Ur;m+z-=N06;t``Di%le$DL2Zs9Jgx;gGG`E5==vq_N6=WP}P?f1CI zY)LLtS``dPF#@83BSZ$3zj;(!5gAI<6rP+F(`VPX2g*MACntT=Jo9!?@4b%%$Eq<#ij03HxO8wrNh zY85VdKxRI;EI1tSz4lUjjeSnvUZyhXDhX(RBg$Js77|mZCW#!NOdX^P#5H`2Gp|3Q zrziquO5G-ri>SeB1tmpMta92ykzTWfi9swO>(+Y0$R(HwPls`4Dnh4t3&Db{Vlj-4KprXwbAw()s0K(T zD$o*YeMK|*qBEF>;u}pUfgHg+uo;3DfuZ}~Yh|==q1Iq%2msy-B*Fu9fVL;+;rw|&JpNZX1}`dr@KhHW z(@4#@C9fq^_8aUlwV(5WmQDz^2ZqGHr;Q(O6wjuUXNK^d!^El#hypo(=m=HLeivm2{OcEa-|Nfz(@ZUzolzLx9BZu(n2`0Mra_G}eWOKue@2k_{k5SpY#7e!W4TvitNhxQxGu8DI0 zdXRb^erBm(&{gGR^I?ij3@v6!mNC_m5^r{jn2mH~XJ?!^L8=F+nIm^&NBjWx-9cO6 zcjjf_fFPJXg5-0*sFW%R&DY$mV2XhzhY@eY#7MC?+%@o=t|sR3O4gu zT+4n=;xr|PEyh|bK6O7?&UA%a$E`4hCJMOLPqlntkN@rQwpQR*%(H5xonYI6Q%p(X zX*OvCQk6sE@aJ4O79lNGg>0e&6J!&#=QoL@4a=5%UQST?RfE%_N^cU$Pe&C-b5+}& zb)B`_Z03adR`;aU&7%(az*SP&df9LG#Y26mQ{mk9Qy&o0xDc7m^$dG!RNSH2*xjh3 z|2g1K%(m9R#E)PkjAY^7SuX8HZkBfUZi(0a!d45s=P`P2#0$Y(`D`HgbiTM!?R}~* zn|97y&lhxLOvTpSpkJ?Kw-fx1Ny>?xd6X#|$y-xm2qrd8Hf$l=$P0DHp;3FL&GYY^ zs9mV|%XB=}w3{+mAcP72X2R?kd@LIr|2G72_kVJK+3fq!Y5p8wd6{IWyX6P3guSol zf8h7s5%vwZ_r2-m`MsXcdBOjGby2(8dnx0N?{MWT=3L7~;AXLq4#Bf_2rM$2vPL8P zZ6v;}rSJLU%^6&qORh=_-p<)Gc|4gK&WpR1!fZN=mEz2LZwxqFtUKDT-&%gVJtpMH z>k@dG4tkUpthasoPA3_;k{UC#HUD}Ncnf{1jI+)$!{Ch3^npjN<3d;_66TB zPpIB0eEX*c-NP(CFPJw(cJ1y9jh2GJirh0=%$ZjN%=WlG=YLYH?w&otP0V_3D^HJ3 zBcV>6)-qle-A$nI_xp0Yj-K;Ob?NpRcJTt33-djfhJC{W&*Kr|NjZlFx@%@8dlv2H z*)zR<%V14jQ5kI({Eee&oh8D_Mq>>Pmk$vWK@agf>Qg1*?>YAXcYVv0FY*1MA`#N{ zbP{f^`c}*{PLpyCmh5)IJf&8AuvB6e4J+X0QxpHpde)KdIsjwGQC|Bu#E+^>H3AH3 zT36YU0#50sA6FBBYq?gvuMb>vx>@x$jAIvhE2SC_7yK|#wzNkRy$he8FLm(^65nG1r}VHhE#MI0QS%&*b3}2*Fnl@ z2C>ZtJfrpP+4gR|21{54E)oEhqCuYH+ii+uRa{k<)uQ#{!xGBn&CI;^;uEnb-BN}W zsjEzRI*sHDi8WlYGRWf2sDJb5;lXlTg(n4U%?Kcks8u9VL8xF7B-Ht3hUjmM4Ic{3 zpl{K5%Ufe9$e;|Bj@dAb?4|7tw3IRC=A$nGW=7Ge!mBSyZ!Pr(vhL{(vVsukBp3Oc zBW%*c@_jxDz>2nroOjcG_}!j!&{`nPQV=f~d*T?koAhOngxgIKIuS#SXebX^g)VB4 ziaQHuj}asC$xTj;P-nSk=r0OQ{{BZ4IxLVhX_4bad~PNeEFi%{u=!jNA(Yvd3uDLy zMp;NsHES-YxfW~Wzg(DBBw<&vjPr*F_OOfxOr=xSqhaGIy>3;Ef=q$cnm^;VPoPH(f|!PBYkXt@l(8z?U#rLLEOXf9+Vp-4D#bu7_%$KJ4Bw)<$$~_s#E=nS zk{AM!h+EHDWeS@Z3pBC~Qpodv)c1Ea?9U_OHGiIVp54Br-myE9r zKo{cOFCia3I+``bTpWkq;csmm|96-C1{M+L1*aRAk!{L>>d7_0H*JYQr=Jz89-B_A z>H@B=@fOTC34jT2S`;C~N;!;QWUBoXkduYQVtc;1Tdx$9=ITzaI+CT3%MtAfGC^t& zu_k9fWT5H+%OU8U3%<<##8@#Md#JE;i>T^ zsqs?KWYXimT`59=Ys$IkqE^0AS;ZvLev;nZDEdhxPr;L2UEDOF0k2I}k8arA86 z&$ef|Go`Cp&!rA`Jc8=Kx~W2^(56}NxY=*N7RGjA5dRGePr3EF5}f6(ppw2BD@#gz zmW@BMoD2B(a6)dsZE!c8ir*n*((Ap^de}uc=I{Br5bl7a4QLK{g!0+Lqm-OVt5I?A zta?TV4um_*lJB~FEjHBf^=WItge<{U0mN~~iF4H7L*IF75U`t_SHBHsgvna~I7CzR z;Jg1NE$f6FxJ+Eq^d%X|)j}`m7|3VDkKX(5K3g2@G50U=6 zBfVd)^Gkj4`2;~6Dt;a*K3Da$-EOJ&RVZDp(>hU#XDA%_aq;BC!S8yY& zOLnYf`+Q6Ri39{9uzMB>%C8$qG;223JbQ{^kv{)>GgL>?|K@XhhMy(B8am z7OlonaV4Fzn6;ojbR*gpo|%LyUMLgdYwgfF1csA~#;*s#9xOWBh}p%5#2xLg*?B53 zDK{Nl4CQ}fjz=4>A$5{5$)WEn4Yyqdcx-vJU7;)1DzSG^a&p{UxRNwEa{fNF(!z_Q zCok19qVPSgxrHWRo_Z(XLj)>mAr2)WCcqP`diQM=8&(6h`^qoD+Q>f;8zyZ}xt(&+kdr;b0>dUP{sRzUEMrr`6bcP6M5iE_Ik;*lqp{{6 zVP}d+N!u9;0eIGMniUp0q;3k_hsN#G;7j23jo$*|p~%fGiz0xfZhk7ieriZ>eN~}| z*1}}6O(&|LM-PI89%9`y3uC3O)x#7+n+;s4a5Z4%D}uweau%bL&YQ#Q3;QM5kHwLv@$ocYl4X}gyeL1ocb|NN-oYi7bA$alO!<}Ufbd8>N32Q z2;uESEdH&bH^@b*1pW9cpLWPi$tOJJ|Lp?YvcY7EC!XjdlK`5Imi{u&;FAFsZlWnE zUa^1S2vV>l0!0%%^4lAMc1nChro74Qu{t-M53ijRh@)J7f!GLOFl`bgcL-;=I;@%Wd3%OG)p8#f5bm&>hpakd&aL+iAWGRGB-I(Mpf(nKx zj+U-~y&3G843sMDtr&NbkH8=n%uEu{wz+ltN$Qm)FP1nkw4s`nH7#|L90V7gcer6v zJ+iH(oGFz_SgJ0b=-$i~itwXG6qJJvt#$xSiM~SXB8Lpx$Dp+52_y26T&31%Z9+AU zj`*qwU68mQ!l;<@qlltI+LF4ae=}{5F}ri$5Rt_o{}0F9O+oQqW-vk1#7Qqry68&2XKf-(H}C#5oGEHK>^Tf$ z@g2@`Y&ZuG8b|I%?+xdhHoj-oS}eab56B(;(HM13NBnHlL~ZYVn>VU!TBzq&?)Kf9 zf6Qb5_iX#Q0t5E=IrH=KCqKdLXjODHXkza09OtS|F&Oai_{&W#ogWSz+AZSkyKJ*g zQF%AvU(7nTUkXv(qB65`PfYN)gQJPKi0mm?w%^bGxzN}{0Q3RdYmoenRCq?Mta zw4+<{^H>SU+rM=A_sq@xTT30|d#%(}#T%HI%p<+OtE7~eE2-Ewycn`}rKSb~g*~{o z6M$$<@qHx1NPwKQuuimUf+tYm(}I9SmnE zk3+wiV6(YO6f0Ze*hMmj{WAM!J>&gYn+T7$vwJl!ucUEd4AT6p_<@4t<(kFzE4U+s1l z*eg8f2#&X7U|^qdUTyWbzYbDa)!hd4Dsf%V0>CEH!z)!gVx87gob)z-p5KNR8VMFz zZPzTUKW1i+ns{}+FB~BzNt43fhhzYONLHfn!m_R__YU)mo)fbV$xIqgCooTn)}Mc8 zF)NTalanYGyB+;!lyiX_eAz(G5IGx_Ad@T|5=O7us}v$8*d4(ip4We3esn8m9c9J$ zU*x8v-==2kRBxE{zB$CAC-!)DN>{PZAg#KCSL6zM`hZcxf45w|S`(+YqTEZ5^mTX1 z#3+!jRU#H@NK1i!y2=zk?KEt7Owu>>wCT!eM!Bb(DT|Mr?KOqee0_7>EXaqw@5T5>qS8T!++ z#5^mWzx`(RD=jo|q;iRz@1+x)F$m$wJIrQfz}pWlgDM$0XH2uL)p~}DXwaOCkw64d zWK;mF|*N;?oM0Js=*SF_N`PAgI9X z;zSO_Gr^n~PJy40wj!R`w!|D$qbg&jkyYrz>u-rTOY{p~P)F?g6K6^I2S^DFUC64b zFj5i!6}Te%+wSKT+=~2MWs90w94eW2h6TR->6xCB^+vCJ!H)u3=t#A!v*cTP@z?20 zOuCBOSqMx3mBI-q)dB8VsDWe7$JC7sF?dxq zYoQqZ%$`U{M0DO(_|o#c>oZwGsYp9`%{XxuXCCv)VOHcTxXD$Y_LChH6H=hy5r`hq zAW4@D%*6|xqQerVetr_`kLZ7H`{r`aM!x77MJ2&N?wCd%*#LwR1ckd312u=&#Lki< zeoO9ieJ3MJ{bDa-dhX>QZ!B^_q(-y~g+Pox6d{Q2HJ!#GLI$fCLxPdirHh`}xfi3# zSS0>K6dnaT)5JLx>37Xm)%`6g;K3E@>NzXY3K8JiNMb&C?tdU3$t2swe(0V9p z^_SL00uzpLlGk%e2`ypc?*n&WDERb-fTBVp@yt%*l6mqDtb7)Wr%!C?K#{o1GZ%e9 zNXZ)k*L2QDpw~)vn!JT%Z|6aCuoe0mx}}#Sc0xtq01BrV=x>xO?szQKA{1A|JR5Pt z3Wk{+s#qr1bI#NFI8W&zjZzdFh%Oda`fTP8^2GpOWIc+I>6vlZTxNYXL*^rcD@tg{n)#C#HsK=1=R5o@gF z&iFg5N5;XnY;RIw(yLr&Y`#;BtlfX(zUPf_k}-5x&H7nagA44toBl<=`Hr@3K@KI#GO@lq8 z6(<>!sBY5Xr`4p6AGab{W@{)kA)Jza_i-$CGz0hTDcy>xt&Kw*K0`CC3^hraTH^Z> z$LD!0Tuv?j=Ud;_abIu!+;I)n5+PHjX4p{Yp~v#C^?*=-jAf6CL}vy%4k zx=Uqsc$94U*E^xzFA`;s)C*J>Ra{g#Wy7+imTw)abmw*8SuApPO>$srSi~X!Ji-@eMro z*{caBysjf+(cAr4i|!w^L~UibuG8M}YTvJ#v-$T%a(1fL7Ef@^%qU|)3b@WwPqnVS zlJ~S%zJKlcCj4|D$x`dDgio!?0OlJ?Y>a^=NJ50eW5-8nQ45^@D<|7ZKlI5sW|O*O+cQMqax|ql^K!k7^Q4hB zi|AHL4aXnfWuJ@2m>tMp2GwKBO7GNoM^wv~*YPl4;GtNVD?O?x)Put4aZ$eSvgqvD zWt>C^&WzsTztVcZw30#GU#mpcY(oaux4Y=g5TUIVh)c+S zsgx_uVLAuzj~UbAKZ(y+3GRu>XUa~ZTXx0C7kzgpArz?+sfG5kc}imsZL*{acBUh4 z0Pg{uSG*m)N3(@4%hu4v5N&$xk`AHoxhyB(!k}HG=~~rZ&Yb6z;ghkj!=~pHP0co) z^V8>YiF|FL9OB5e{1w)Mv$d98&}YAsmMM{}Qmcnfw2EOGPna&1Jxk3dCLP5{Z4Z@V z7I=>m48?G?=0tS6l5{NHoCYe4lXI3f>p#l+uo`etaX4LKNA@*tjAQxHJP7yQ*6TMo z8^uT^ZJIV|xm2{x$i|s#`1`^WD!u^yqf{N{AF|N~y&P;OY)Nw~2T&qaIUbpo@;87e z6QTIPdDwXo^i(fz7?EuGGfGgNe-s3tUPDBYLP;>dIV%?CEMF+*kvn6t{s-2zF z+$+9^IlWKnx)HxT$1n!PvWi2!PJwihw>{Qc%w2#3LeKsCRV25R>6G6ON7VFk-OY+ zyrv1Wlgfsb!3e|tnu#wz-s-Sogl$QQCCe+p4yYeRe+c=>KFDK0Il)Wh<3p-Su4a7c z_iOty{NeuJ+>SS=R@(NeU*A<+7mcsFhPvh(tjI5cBx5DhZdw4aAQ1?4et3)J_{!r$ zU?~wAv~WlX4@tJ+-tpl1!M6FM-rl_H_m5zjr4Ka*$Y7wR#*WDi?|-lnKyD(5cG56i z_H}vpUW1uDZAF4OGXNVP8ufB78?5Bz?VA{%T7l3Dl^-9=j;}?XBneL!jfo2bJUY;* z!Qle@o=?GpVLrNf$&Sw}l%$CO0+ZU#u*oUtc|!o~ujC`|J_%$X#bGj4;1MBlkZxhe zIisNLQERQ`~O^!IEwy$u9$;gJ`OdZn{&-cXGn8>ZHNuja3A zi#VDuC4-eVlc6WZi>H#IKe3&h-ac|{C?n+qVbf8wvZq7(IXQs?s7U0fpU?=duVQ%& zP;;jFG<*x46RYG6ZJ}=RTAwpKT>DOT^jmV^GlY!W%naxxP;;|}ijIy}~#Jx>ph zs=@BPt(>rNG+{v5uY{7;SjqP2#U3v6e{AfubOwTl*tBIfX8&s$K+d!dyV|S#5&rB7 ze;?{P)t%;k|1*R%>qa=*GC()62W7kI$bRuM@c~Bs2DJ$_{mUpnWS~J zt5h}1K@VM|O@8a6{+kIS{Cd*r6Uq&dVz2Z3j;t?7 zk9XsN0g3&y>DPAmU-iPZeJ;7%>!CYdBzm>$0$26Sx0=Gd@AEy;JxhH(y=@&^|8x2W zqk(%kMUwA-()*6UpDgch;o;%`YuojK11Dd)`ri0IAMO988GQap-(4N^tN>H7cAIK7 z(zFOOmyWL@*14DW6=v#Am`G$VkY(kye|CnPLUZEjbh24=$DCV>Q@mF6`Uxmn#inaQtIQz{JCbq{JAvU z8YbzlmsfZjJ5DXX%VnSd3ut-jHd*#F_*kE6K7=Cxkf3K#e4I(d19083olO<4JS% zKTDJK+~7>Q3>AVIpw5o}#-W&hcDz zHi2|8-6})6?G*rFP5?f1F9Fkf%UeDOHk*BS{p+6mSw?Kn^@1+m3Q4w8C0pzuX{t zHe~@-TqQ8R*kF}MW?)rI(MizhiEw7tTy^@ylt`dUv)#0T6q>4jq#!NWsWc1(vr}1- zjryRbt2$hz8RjH6Qy|=CUs)B}&+%)i@yZiVT<&C$J!BY9mshSAr=5AdX3Qgv2E--? zP)SJ3F_Tt-KDXtpSmT=AzJVkyE zizPRH%K|6XWaZXZS+BxB&@hPT6GnPu6zeX$iHya?j2Aw(!%~P}D}IC|JR7a~85HNF z;38yV&U+anOLctP6_@bj#lR4kSL?`_{!8cspCVyYOaUF=0%1?g4D(x}L712gIjig{ zY=nlz!&m(@FdhN-o((ZlXLw}~l8b7R{kgWQ!_Np13vzMIcaP zYRjOGp|FKak8HQwVp#9lsouW^8o8@Eb#o6|PI#AD&wh=&;8MczoIH9~+FQuFXfau# z0X8ze^%u5uJ6gmf-Jbc3BzI7PzVlC%gOfRj_V89uoKNECqcm?WRno(Db99ye-Q$n( z<2)GzSSJ8)E$lDnyJUCz>bu?9Pg953W}KBEgOLP!bg_yKulU1+y-jX9&esQyD~H=i zFO7ZoTvi&diS_K8l=0!rIp(%|Fy63ccG`_C9x5uBO2|re4`2;__s*BI+E8)qi4|M_ z84QO#w486NpYrOVGQJH558c^re3uoCK)Av38Kqp2`Lim_f09mWU=$uVM05Yw7yk0f zie9Fc3!+58etX-y;N`1o;RWLx3bNR=D$lM`Fw#Tt@u!h!1yyPqi=1<>tNV0e%k!_h z>n40ORGLd6xYGno-8yfD+7*E<)%S;;GjApnzY3NMAGX`NIW8Y|v~FMj@ZT^Qah1Qi zo$bTpXEl5rtnk&HKeM@?{NwWu^#`JX@cQ0YE%xd;kN5fia`OMZpT2{U^fPXt zOE5mS#LrR&LE%);B05g|;>;EV5VggAI}Cbz`xZUh(0ZqX1HC1MD601(P0V!pcxfL^ z>uN94RzT*$Uhg|uh@rb$8NdKi zdB)44B7kZzfb;y|%oi0&874wUi8jlK zQY9+#@3wPI-fx})n$m$l3iX}(j8=RQ0>GnM2P;FYOxi2=E#r*vjWM9O^tb|v%{wQqT`!~^5(pIKN@J?`G+!o z$kJw_8^mzVmq+%#;woKKC{GLk1OS?E_3*$=u8ZD^Kny<8wQoQ|937{D^|_#oR6B{a zd@vvuv`ZV}vtLd9 zdKEK>irTs?QKmp{Qqa6|ja{5t6wl|D3cy}#ro&1z1xX%lOQB2-4V|Cd2!bn^^3ecD zPgpo_TZC(uL%O)WIZH5f&%v0M+Mov*1<{o?M`IaJo9le;rzFpG(rjmRBs8G&?&cvByxcQQd_|0ggldiZD*{oH2@i zT+4GlhrK6noG6taRiFYgW5&?nW`UjVLiyPf1reFwn1Bj`{Gyz8I&x*`^!|OLrWo8l zMlDQJNNHyqx}3D85=>ITzs$HCCr;HKibAkmzL;kR@EqOt=;82-Oq>g2S5WKjd4nB#q9>Rwv{#e|=ONgv^?IJ&@y2larF-)$ z=itD1HfzDDmHl5!uHwqr(;9+rlrO1x^+mZD7b|zyBU%6M+ce+n{9pDM@;DoR1MgOa zd5hUONc2#AeAvO=*>+C4)V2a*ffe33+9=IOna|5uUz)%7`O$VXzf?lJy(zCRIT>}P zN}Z8=PryQPp9+`hacudAI+?39j+n?m{@NVCmd#RYr z3oBbP9F&1z&a4)B0RcufKH>LkMf&R-T7_Ef_t)y75r)s-xxu@UfcG~gE2@&T&4=-5 z*N(H7GoOiLJ}{RSyoF0S0QL-Z7e;rOUfArP>F!-Oo4@984#y(aWW*rO&T8h&yzN@? z1LXvIe|WmgJGI=spPYwPDKv1*w&_#8&ZWvbd?^RR0`(zCylPBJ##uneQJvL#8r?n=wRme->Ibu+!N zmx?gR??YY8ve)2%>!QD2v95`|H@sOXa+3U;wlUrRF0(A0CDm~3(2<4Mhu{CyKvR)A z_oySP$vCkxCt%trAL@;*SD&^V>Y+YjlS{0Kk(8jV3!4;ELQ1Fz;6gQ#Xdf-Y=EaD9qa&}s-<*JP@B^MfMi%LcHG@Q&7vAI7HfMZyC z)?8e~`nBeKY+qK#(p^5FNg)>zX8>G9F9_WgJA}YpeVDF{%(x{RME^TmIbEhvgp8)TJ?y~i7m$5 zXYfRCnX{lMn^N+^O$cg(nz4%$+DKtA1J#l;IDycn)6Hji2V6J|ZB@{DRIZd=)dk{|IZG993&YtIlZyRwvBQTvt&>H4a&WnA@ z*+N+dwPs&Jo>3ffA3c()(UWQllO+Kef}P2;_jS1J@*9Ft` zLUbWm6lPO$_%fggrvqURiY!U7n9wsxx*OTLsuy<>xUE7 z+X0*Xb0|Ya$f1I9$tcJ}Y(4c9%q6+Wx`#Ta`BpZ5`b@9pP1~50LdrBQlK=%^$*bdfC(G11?{j$j4Y2#vv*KoN}j&fzfB z%LGSKgb?=4o2+iV77S`!K1;QNp6z&0_G4)>+{st_;`taTMjSvt933!nQy<^R@Rk)nh`em z`j|lpQN@+!rvctablzensSl&m7LUV{*RvKZo3WoW&T355q)4u1^aaabhQj&5$PLU6({=pGfD91_jev8nG152oS z{f*KqI(7D!`e|$d=nfWOa7D>+F zcuiE%&5k?6G$4+};>FT}@{vh<_x**dzUC>hBuZ32VXzVf}ZnJ{qP(%~d%cP}bvMqd6cb?fu2SZ0)8by+nuhCw zklk3+J+LE_z4d~fxSAr?crE?2X$5(*j;{Oe*=x-*Org;ybe`S+wyA4HUAct5qTVr* zo3-BZULjp}>o~^BDjJFWE_S4*YlHWkE$e==xnLKpC9udjMnQiC9k2KYJqooPOnpG+ zG$Eux2xQYRw2`{Yu&}W~dY?-MG>JLa0FqU8>$KEZCyo?C;6ASC4pnqfk>mjOM2)22 z-$eyNkK_pd`#F%E$R+j+bIU!d(C-eS$%+*flP1R*^fV8%8SE)6Wi;YYUrPFxF4?zz z(86sg?8B!q=9h?ROwzgLUCX{oQ zc_t5C36ZgvX?SF&y~8i1Q_Kj{h1^?cn45e_4-SP3V>)Usuxe;jiF6yQ`Wdm5bN<4n z;*ihKEr)j^&hXp+wF1n?Jn%rJm6O0rQD#Ga%xTWDG4&E1MWekT#agzdZU@irdM=YU zzwtNJ@mX-!`0CgTkP>P&;;Q@)J|#-!dfTWLVhojJuUDdZ>W%P0Ly8qO>=N8<7!>3| zMHM6dF+ELb`}SU%Oz0@4`D?73DLzPij3_7S%u1{HpfRLhg2|2 zUt3F)j)Gq$?9Y8JFd9x<{^DEo1pWE;L+>f;4`r*7Pf4V!tx~oa@k^xFE5dgg6tcRy z0#XCHKhjI-Vt|MPKmW%3smBSnj@yk7w~=uC46i&LdY?RcrHsNsa{CM<>DP)4t;O~~_3<@IyN7A<51)gI%B9U?WPuEfM)Mwf2}vnfB3u}NA(!Cq)t z8bI$a0Z(t4>hD)yP8Iz`Q2_9##}Y>7*wXgA6uHicwhW86e+x?-w;%UEWq)cP=ZanF z`lzEGdBfO1&JQMwoC(n>E1QqpTm~SX+xoki^fAZT5z3(37a1rTp)hHq=2rQhGU9T)xS;eAoOxGKUCK z=;=e*5DRT~=dRP`-FnC6yJ2!+`>KvWOcE4{b7mM8*8C$xYW`lsog8OBYV|y|pEB`9 zNuHuidrWjtfd>&w^KX>YVY$iy5@d7Xw-vYU(mINXxbKZ+Koo?&w31uteIGczF<#<` z!2~4;;F(-H9Ng+dvSCY~{Io3RpFs*Tt{jFAwS@@zct9xXV_T9 z>MffkyY_jFi6YETdcTNmSmt)@ZO@LUpMm7s`kmnu0j zm}fWAuj6_)NPFnvPm0_*EWvzd!6;s~^1>{YyHhjnJ3@a)R!lW+z zR?1>VoDgk7zjb&}(+U-nti``k99S~DUScR<(HBu1LM7=klD3B^pGU7?pADCY7N@g_ z0zG@Y9YZDSFRcbPhmUblX@$X|$#-4)O?0MAcQt9BrMk}UqH%1)udL6t#*Ys|hK;DG zCB!Nk^n)F1jwTQDr`q>Dka^+n4$hp+mI6{83C7itbm1Y-8$yW@B`)C6h05Ip|0x^W zk{&a8$+__w*Vn6&6@3i(JZX0x(rK_n@2bwTA@@9014{1NTzBc^`U>HPS5v6gFcE$% z5RZ6Y1b^{kHq5hoYY!DsoiV-5^JBCX#*)T)Xz}K;3>~Z48XQn74Oc1%_*5zW*bKI`239in-HIrATGyOXJ3@y%|+wPG12 zn(8^5+GH3>dZK=5hx4*$kVlCr~L{-=NvK`uqLG z@a{1${cos}nmA*385`EqcJ;u#@JB z_dH7nJDhoD+}pQvqlQAGtB~2z4jVyT%dS_s(D&;*w?SLU{AEch+=UjCP6{dfpA0hqqy7OgZFDjOr%J1~5GYZ@;ecDKnkM#a{TV zL^V;aL=&yYUlY^;9Le?2ew(PWFGSM}2-PSSyfas8Av5Q}F0EpfR!@M{rE}Xh8N(JS zl>i_slg*YeuFU*noVfHa+vS5j9|(wR)hdIs4C*JNconnmWi<}oha0%5VTzzHeK@~5 z*D5CO81+fyi(;mm_N>N({^i)dUa`Cqifs0yZFAqN$V-8Qn6hz#1$wbz`Ny@@WpcI`(aVN-x{mbSG9; z*A#6Gn7u=3#BB|q4@g+KtEY-MJ7XxMla!L1)4bR4c#P^*fMYUU{in@Z!JMIF4SJB0lSXQ!&LY_Zt1vPzbL;fj7pyZ!V`Hoe zm{S4ImG1hMNsK`Hl*wcv_f-^rf~eeeD$nDSm;j{XIZUO>Ys`y%3j!M%Cp2`)_{l{x z_}G}IH`88LKk*f8fHwu><;;RWO&a)h-PIgQg1Gvn?d9TRZ1nC-*YM$woNKAz9H=^* zn*YrDH?gU2*p@5YRn&2j_dcpdX->aTg9g>#)mYVmMnzTyq|OSq8IB$NF8$iL6g>Q` z_D~_BcTeHp9<)S$?T0nwCFJW%p&Y(};NHO`+U8UZ1N~H2d{exVaEPeQs#+R8o$}L%z55TQc+Rd zmgf6{U;ZfRGY0Y`@{`f{YpjU7UuZzYVT+?XgRu1LM(OT!?y0!p_cI>0lDX8sJOz*V z?8hSfh#lyrBqsISm3R9xDRdiUDBoAMCZ=B<&gVwn&aLzYi0(D?>?7r;;&VIAh*0F~ zrVET{@Bp)ek$yZ}3>SQ>uOOPF`jOJ74r0uH4#O6c-h3wJf7hF{@wW zp8+qQubdE{Ux%P4ecNpL#TGq>vhEzjr|1hc88~7u>oWbm`qPcP^q>2*>dMF%?(30x zgLWbtT9M>^GyFD?_Dx)M%uIV95$k)p2AT2BbnUn4f1j$mdkijfJkN!KSA4q9?r(nP z7wJ!aoDPF9k)0sCecG!}P@pUIe+M`J1X3#Q0bOQ9ZrDNow@WzkESRDDNx>W5UG)8S z7wDS`)YqSN8-qf&t}{cIJo^;>YiOg9Jq38lL#QV|F8BHej{F?Dx;#v(e(*t?nDn1A zqRJJ$Y3Mq!J#bk^E#w%Vk#!$c2|{hP`yT_MN!_CL_4^@PzJa$>uP$fRyTh$JR+eKO zwvd03bRt1tzC8V1PjbADG%aBIWjRrCltroyMjV#gpK=7=)@<%j#&sPIy6qp=vzM>a zWc?A}1?JrrUn6|JAabK5rm&!xkara@n-9+35Eix|h&+%4(^Pbg=5t?Z@-L~8vhF2BgDOMjsy^86Es5-qiAM+X7V?c(8N9-842?Ft zeN;7suLG98EOVR4;4JgEcqw;-{Nr+P+uO)_4&lQ#H*#~%Do6}~ zG*Nd+Kd}L=UN9EnPXN(&s^x|{7;03vzgr-tM)6%D=H@XIz}4Kr(4K#R2(YL0nY8wl@d6GH&bmpIS=w1s?aHJQmK8#^2%f_JzxNwJj3>>V01(BGM4 zf^^FQ(?P7NLywWW7%Z4@IE_9FmYH)@x|3090_Hhx^PGh_EO>6=5bnZ{{YS;cPoJen zzQhxeQO=B;ncPX{gm%+eeHto$`XLangkAqQtbx2oqk~md>kr0++nyGSPfMt4y}@gk z&y+=W(gC(j2d06i&cB3tEFrVD>14B7+UOHpabJrHveH)>Yn+x<+mszuy~o^SFclv@ zXA}EXBO^D{m+VXzG0;a`{8~ay{4l=~1o$9S$p}ShXW;b2HKzRmqZ}*MzF}S~(fZ8n zeKWb4y$LWOiXpf;u_v>1&Yb(XBKOxxMl~fuSDuD>@H}zq0sVJ)P@Ma(4cMY0?H9?H zFqCy^fwPNOkVS*FW~7Q=_(Yt;&oon|UmToQQU$Mrsp{muWoBje804`Qh1T&PXDiAC zlP23(vFCiY7Q6VlPDH3NC&v2Vq%5pvie>Bw^YU}N*}b6zt~Y#JyFcqIZX@5}vYVD_ z7cK4Lg8r=aZX}$;J{}PQT`jGAUnrytyS{`N;cQ>4!>EUKXUv>|IF*AwTSSrOSmho z7ndi>hq!bWGd^u!|A;;O88$W$$`Jlwtw1*}_ovpz_9b^N?dUA3+ng&5>I4+39?8w< z-OdlO;Kmh<&s(tG{JI-ciBX*l zmJ`%JOw^~y%9vgy>pJA|HR7kULL+SYs#~iV83I7t6ey8HjM4j(Eli9I%@X zHwNKucS#)C9Nt}oY2JX3=E9A^2lpA|3jQ5bC#k!%NK`V ztJYg6_0yAXMA%4C5!Qeloe2@_Qe(joODC(@S_t#?XNk(=;K%t39Fno#;M;Ivkc7=5TB{zBz2|%Dv2n*&XJ?`U^zIM!uN~d_3GVXa zYnhVT$Tstg!@}Af`p?!M^Ns$S`x-mZ!Vu*3##iwE%-)BElRq2d%_}XLAbhqwIBaVu zGpsT>Sld8r4alyo2)S5^#u%H^Ol_1N9~+akd5*PHGH~Y4Nwjy>bpFGEvFfC0N$ucn z-QWCN;?raE)wyu|4a2Id`U#o$l$$#~0Yv(;;x0-9(U@McayWbvOTiDKxSv^i zfg+Ca)v`t&2fuw6VH)HbMYR2B5_RIk6I#3}4%E#K3vzJAA6?ka&H;<}8W%afB#f6q zfq~6ua&$R*swv=kndBidRn2drce9Isi5obSKX{kR+14bQ#g471@sAv4;IAzc z0vRWf`*Jf&Q(}&v1`&SK<$;nV)Zgy!Tx~BrW_;v!4LqE5arY?AOG$r2UEC7>ta!JZ z_y(`r8nTvggNb-@NVC=S>(H}g&%^&fSzMC3f6TV^iv=nt{FxOoZ~y%*3a5v|s20UH zT0`wn)&dEPWF;2-;qJ6MfW`*9i%OZ@Soqzp*Y=OC^f@BuL3J{k6Xq%D0o5@6Dh{-g zr5+ISkAnq>p4kkNfq637defyxM5>_ko~KK!g(s%gS3m!Rw0puq?VO#+JLHEf~)43o$?c6-pr!h|;;D6wY!V~&5; zt578B=NVM;SMx1p^O>^NyngAXBb76=Gmm^ugx{9&ieyrv4ZY_~r))B`jdIyP7Lc($WEL2D~h^> z^p!ooh1UA?1m5e!gm*o-0Kg;*x?gOGYVdeqiDu%F2^lf``a24A=&jMa7fb^Urat0FHAl^a@VTI z+R@KFQ3<~eXTFPO2^NCf6sfEQj~*{1(knr4*zANgfX%bW4fuNuc#Z2Efp#ERz7^QpRw6yi#oKfq&6a2KzWzn z+Mq(6V5`L7zn`M|U*-C6HOstF2bafmsaHpq9JVE-I4w2_UPaPk-dO#RQR*{!N` z@ingh-7h5l@7Hu>xPtU?pPs#Q0YMI;zy2LEhPN&M5rS}nqz7m z_ZmZ*?ku+3dMafZ5j2pD1?-Cg;vPO zP$=|C48@6^%N3!@DSsk^ddF~~l)OD*0zFp#ZI)Bf48sM1HE(6Ew$~MFv1S4_-iBu} zGKy=v%>1w$ZSAt(#xj!;d`rB6714%hlv@|Sk9 zZ!1In4TC42%o$VIzq%_oxhCgOtDc-Gu5jLv{n>=^H*!ptp41Yf5__p=vVGsZ1W}t@ z^kKI5`*;o?Q(=RYc?jiuJ1@LiF&(2L<6!h23RzkDW(i@a>x{Z1-}A-gQJ2Y!2`BWE zGVXz=yh?ym7ER?Gog>Cat>In1BBI~#YZ@N5+G%8;?|z8sIuIma7Ntv3; zcVBJNXce~qev^+v`qoO8>9t=;qx1%qK!Px6y?CmTjF+hK<_0h*Xx>Zsq^nh7&qrw9 zUrvZ~tRlo*v05_krv5>L9RK8{5U0Zw2O@x+rzTT=-s=mL8|V|_;#Eqh5YahT(l1o? zp#7CC230xy=48-FAkRPYM+b^alfB5q>50j0aaLASmL1Y`xI;Wm^(ufz$&~yK;IuKgKgD6@Olhf7 zoAk5M&h`yAjn6tH!Q`Yx>DzFncqI>W%t6{t@9=`?+v}z8UcDO^cVV-mVKcKHsw#Hj zcAG2IqlIvN$RY^_FvNw8T4G(4k3h$h;#z$ExVRcl&1M=9IkRi>CHq82USLdT@@9$y zQdnXJTR)-LwBi)sO$N=z>=W}$eLQYIiqEL@VQ@fN`A`Ar-9#n z>X{tt!PWQlhrfPJB*3daL~-l+WwM1jiHw>0Ofg72;XRpKK6^dfR+w|9?yJJ&oMKztnq=}NqX+UN|-U8A$Gto^so=}1A132^HNpg`*HN( zSQRxnsZUyV>wuzyc_Q)TDu}L!aAKN}GO5i3!f}#~GtvwoD2yt@TNlMTmtS8@i)7ah z&>Bl;sPg`5tHEf)zQ|IkSRKP$24%)l@+&6CG6ugs!sft2_7G<%qV(rvcrR&8(!&D0 zIO$^&N6o0fVFdrD1(;TPqdn<*Kd`i;dQ4C3;A3K@A+@HbsY8f?NLfdGWbcgJ4Jqg0 zT0rIUPi(#w<{t`n%FhScAxhxEHSxCbt#aGQN4)QLLWVD#X^U=)g*xfIs)>f2R1zj~ z_U{uxh!EpFKN|(fvC=Cfui+SV!qxtz8yFFXF6cS!Le=NZBFOh$P;pGzKtU%t8<*f*7~q zqB9Jy91r1mrR*oUVUQKqqIchV4cq^i4$dg&3Lf7bn->1;sCL~}swbt+YRZ4d{>AFm?UBw3i+N!Ruo?HJ z4(^_M_wgTfXHTR2!be@a?u7-SVpMNAxGh?BBtbM1gp_-K!E{~~{%4_BT>@h+pqMGF zZGRmNRUo*G_q|4L%H_aK|7y(n+DG?elu7%oXN0UVt_I% zM!l#oW3)VzdXqHMTg`8D#PG6fo7luPk1ptaMeyUCimL)ElfAnI9j>XiQ^Pef{lT6$ zuT)VTq`Dr`o03;sI2@ik|E}1KggC;cAo4D&NNw{v%{h%Z>spl6zd=tn>%U_Rw@X z^F>dw8rhSd4$Zi!#k{i+;+h zHjI3~3c**;!L&`Mcc-LnR2!Y_JqW?IU}QJr_#Q7|!N~jk5A|WR3(9uUbgrcoGSQ#E zH8cC{RF`zynCkY(@W@iG5}=XF5180HE9XyV{VYW{SXS#7K;+U;TNjnC_GL@Z9k9fe zj&=?94ab;OEw&0YVOl#&qo#54S9O!?(~;fzYvO9k@36T&4NMtR&*h_}R?GGeU?Hxs zr{~+2F&5*k%MZ4v9={W&Mipgd=CyEdO-Cs4{9`Q1=+?CJ^csio=bQAPzX^@gQ? zcapnPES41(ZY zX}rQ~`l2eiGksQlSH!x{Y+q)EnVNP(*Ff!CROHC=rWIds|0$kjC6Be7$jDFH1@KhOaplx9$?!YnVz2d5MoUnEA&S{PFd8 z5Mqm$(rJ(Zc8&zRIYA}W;e%yWha4xrbb1yYghe=hY5sqltnv1EWi_j0P6hxh6@`pF zR2j=VetTnQ@^zL^>>BK51YrPjRv&4Dde>o;x+L7oG)On`U;l8o3LM`sbqih9M||M9 znS)0YkVsBpv{hgF zf2-ma=;c@X9o;#6{xjwEC1-Yd(z!lsFg`uHdqnXaA&5s?pEaH|gB{hbBb3|&IxiB@;^*$SRviqx_kOZSe0q&)JszL_ z$PxVXXMIvLVS4sEu26h>9_X+gTGT{OTi_R9or6cpW_8fckC!(GRU;=#PlvpG%WP1n zT02c@To<XWt-&gcV~zpnXN|5g6GLZ*{YXj7IJ|%4tiNTZ zdI=|Zt`Mr@ICE*6mNboN*#i+lGjolyeJ@Q&?7v<7ZTVqm`jYeaYU}Vs`0-D*xs5=_ z5|jqXYSnM$Iy%j)Qo}_5stAW}dM*#6_*}X%l_vh?$Y<dOj=7&^vH7pJy^ zdf29w(Wd#T#cwmlbM^46KZ6lruK&6o)me5!NwVD>{^!N#>y7IcpIx9_dSqX7UJ~HNMG_zFi%UG~(#$3%SYbtJ7Uk(At%gll%Aom(zI>QXOW&$qCUDrtn8DGGf)o%u<`;%l!ZaL zSr!IVm;+cKn*A&MEjNiwa^J6h7=`EVkdPwLU>J=@RR@0CKAx5>u6znze#=HP|1ytq ze2$OWf5?1BIgR!vs9O#8iXJIQN|%H;F{&z016=qxMuLGvHss z3bJxrjGZ2Q_!>q-X=oP!5l}@v5*0nu0?aeQev|4T)A7Cb)x zauzkKNU5&HKdQE|qVj8hhYF0i=(QC>Ez9g80>jBBr>{9^Ui>PBB-EBKq*7?Ba>a;H zPHyT%dB8B#n{fIqlqRX(*G*>Xm`r7=+^~=riTYt@!KJ^H;LBl{rb-GYBv%`_^T#)6 zDwL?t5m9qRqsE>Zy-CgFxg2wjO@^WH{6GmJGy5N{%j?vrT^xZI98Gx#iB`r?{ zu&q=yuA4_8_N5YTeXBJqi8z59p!b+Mn8~#+YX4G0Mp9df zZBKAL4i_Z!H6~$KnX#eqYf`JxD4nC0$6xjG4vVt~UG@O;b6|25k3;-z3(#cls&BS? zaOoslMr*jp=)aEZ(?Jyf_e~1xeT1x6SBI$4q=V5r zW=bG78CzZMq4zJz*g7T3!2N?lZK1TxPZ{tQS7Vhl_|maIG@Tv{w3A{B!T|xE4zAeB zD?wTIe~&rhjqk(v-m`M%^k~?jrLzmMgBdzSGDWKn<^!ZY)Xr1qb;|#hB>aeP&r`|I zHU5@$l1=F712BpMUo~C&b22D%h(ido`%JRgg5qlQ(T5Fm;Kd!KJ_5JQZFymksRC=J z56_L72$K)2&KudPwP6QnAnfCVCt7DAl=t$qd?4~XkpipE>>5{XT8&`mSUAm>cl7CC zkeAdGUt~HyAM<3)JZ%Z5K6+lOz)%d{yglUuS4T6P3^+#rKu(dfkqshXrLCfXO#Br2 z!*2Q>s*wx8!mUOfk@k681rY5>Ggb+q7&SvOQnJJt=HwIvvC4T%+=+TjM&&M`g}FA- z;FVaWe^$*DK`n#LAXvuba|gf zp&szP7SeKNW`I+9UBbTNP$~3lszWUd zTtP1BiN`{gS;*i1A;VOb6Hk<}mJk*L$#)v(rR0ewGxZWT6ZU*vdg^qlA-G$vR{4RS z++CuagW6z^11i4zTd|g+^&4LXFSS~HM&hxW%|+XnaWlPWM}3)XY2e-8Ei46@oxRLb zAl47S)qNJ;p{zXm*Yt3HZf~?m8eP7EUfK>~O&!acQH;Pl+w<9SIL=Evr3SXIPfbfsV;j z3f?!4Mw#iY8?>@@5TlLAVLF_)-fQc`jHvx|jOD`Xr>$W+Ke8v!fydpKuMgp5IX#i? zOkEKx)@=u~-Fu=Ah(DA5S{xW8D{-L(`S$Ks;;+x$9G)(sBgZ8$c0x~+)7p0|-)x#1 zSum;dlVop)glTX~A=T6V%F3}Dw_Xt>_7vU7_RSu_JtMGM$L4ZwU{k{rjqN5SdSE@O zE}Ev|_kH|ZC^+w2UhX1{X^j|$Pz}RmE918-9BszCnTu`)MJGcK&4L=`r@f~tyy>zkl_?GYA zjYJuo{E5V~WF8*MSR5$36)4;F>!VgIEF+z)kBFN!kTMD9Bx&dQ=;-BUi-6hO7mVjW zeR><$t-vxZz?}^0d*}lB!-d9g+lu9jLn#t&h zi0ynVM>(s4#gSaP0Z)H=kw=nb8K68eXLLqYCh{ND$G?lO3$h zc4yjg;*>~Qsy@@=T+E+E53)$iW~t!igyBQkIZ&G`frT11RJbp&oQi}1Iq}7fC+iIC z%uN+lBfW$v00G9+N>;oS!&HgaU<)$YEi=ZqjCn?J)y9?ZXq2&tkO|{`1jcrGiuJeY zyl-)fxw-^BcFHce#nucnX1<|Ig3!p;uDz>juEv(JwVuJ@YdI8sS!@vqtu3wn@!(~L z>cO0cSJ||h*g{$3G_*-Y2-g`nR@Y}J)unS|F>x#m}N7xwxy4zkMwQ7dq5TAM6>FYF z7|Iwk0$-1%8WWPCi01gd&t5kwip;2^-S?qR47u9foTi})V>(8HzKs734jP~T> z>2DQ|yhFvrQjTvw?g|@Ah0_yJ+N%+9)5~L_EUQ$a;1Eu^OV?;rPl~B^Fs%OIRO^7%>hd~+ z%$dz*ZGi;FPz-wYz}!q+o;^*l*1-@>B-LA6J6~lgYk<^_@KL!WPBf^pIqr zp+2iWUitVvE;zlsym;p0udA0k>tNhp<%-~$`3{vbA!)^gPVd1!l__#je_~na2Phhh zOs9-HDXC&+T;GUd+qIs_DUEOH8!ZdFkX_kUEOZx0Ff7R~O;u}I%1;F05WkLkjNP+Q zAp5z1?a+>RR191$oFyW8jie`z&)Lxe<8OEO?G6(Jop;$=c?AFS8#`u8uHJ25lDZWy z4(8C`(?F6HbQKqNrCM@(i4)w9O#5~4|By3a!|6uoy9S-E;G^|E22BlfQH9;k;QS8N z#bGZj*0XQsA=O4}^!!%rsWD7T<);a__Fc8~@g?e3F6a z&Ox-MFIWby5Fs~;YeLm=CAiUgyKo>m=g z1X699MqC!A&DFtCcT2$2(o#0bXE_}sA6K!8T0-!VgYkxW@p)+RDGo|;b>92sZu;U& zE*Wwo64?VA-ffbP2du>Ixf!7bu|Yrn;{)Vo>ydz;gzY&fe8lOdYi@*Jg;9RX+mqw& zsE8(c4tF5j=(+zy!a=nXez&jipRITS(dz2z^Y{4cI;(F3Q?wSll4HEM|Drtilb7ok zf2Z+P;r51i?*&@Dy$s74ZwR5gU`dup{_YiO^q%$IZL+aM$af*@_oeem)rN011^;z} zH4*3i&3o--O1XtXo1=#VrFpt*V1_NN1|&8YT0C-T<;Gwd-7dFP%B+r_}Tlhs=zj}EeD?u&X%6O-yZyDs|Ijss)sX(r{k&W)_(>&xKpdaln zmDQPEZYan5VpNe&Peo;ZgH}J>MNY&|d^v@At7xV$lo99=qr$Z|(%3zv+BodXG=bjq zvK9WD3bg>1p0IG_jQRA}NXL!Hf4ji}>+c$>7~~$O`23CLmnq~iA9iW#Rb6y6_^$yq zD!1gV{0C~|^X>JYCf4h0^l9n|df)fh>1&2Rp_%6f~7qj zR0c9_Eje>Km_?ucCkrRk_)qi{nxp*Hf{~U1?pZy4wyHH@>iC<;nO>EI-?VUi_+p#w z?z&Eze;(GqUE|50<)sXAJuALj($Gn_?+EF?TN*N$YE<_>_dkV2h;#Vl#~nA|lnHX; z#M!I4hj(T`UREh(2%E9k|78_=9aTH6@xs1zK1hbSYHomv951CdCl61CUJ0*Fh`J$t zs8r+5d$0y3s91+2FINP{-jJQJCJaiZ@)F9Lpn&Ze3Xqf8NpQbwEmk4FXxe9K;(rr> zm#dF352z_Sw3wX&yz){+Lcto}lvQOS0jo+A&8BJ18UD5^VUc6fF%V+v1aj0AX0CD& z-${o@zO!3w{hdb4gtcO(R&%V3Zlq^OkiQNkN8dKD_m4^UdKsmvsp>~s zb(JeC#70mFXXntDzIKztpN1YYEZ{(z?tLJ{m7tOI*b^%4irnBiP4Gjb`b06cuihs^ zTAgxuYTe>AQUp4?(+tFlYBiAo6k)Ny&l%92uD3p>A5nD2)TgB9vkqNub4ZX{eock7 z^+VBz(a26^e!l^^t(i!LPsI$J!BiWQLEZLVaPYvHd zkDLsMUHO6uD+S1%9M$b1Nmz_PCSv>#DNcbQuJSu3d*WT1l#^(6?+OIpf(__wchY;^{a}UBY&|raJ9WIT?H{^Z7FX zm(j}X^&7FQg|Lg?aq{d*EkCMtr^7vTvwT1-(f*Q16AII(_sA6Uw&=sBqlwV8;wF>& zH75FMUP<#AQc#npx*Anp9_D0oEmq3p$F!5#m#V`vf~8dli1nm8K?^sTzxxS)64j|> zW>-YSlXdayYCdH!$!=c#7M@?mGBOqZMBs-H*BMFBIH5RQX+1X&XgIVpefPdxu%xf1 zK8gNMAoUfa+A~Nw(AXsovDkBwH}GtKvQoYq@pOdE=XrM#va8ONTF47NYDaveuU|_m zKKb{7JkqnsCH$lvYp3`6TVW<@ia-Bzn|fT$p0c%*dSyE8o2{;$h`=tUS7SWK<|i%= zJevzb{&)4NC;0jDvY1KUCHRVFDdI&eC3$kvO!?lC{ophwr9yi*oj0}GvBUl@U<^)F&rv^X5cMp2FUJzB5Yo2)BesGYf4jHu zDN7-gUe?G+(3L%>0XaOC?YZ;#L^L<^Zuh7EoWFqSEWx&H$5gO8 z>G}ji{s`w-!b2cvzl~*VWHfwtq21AnF3LC8G!l|g@P7RMi{nTWFVQu)xt<6G(o(b8 zzvFJ;;_%AkX0Q!^(ZJS&^?zCb<7)Oq7&W@;H?QMR(8cniV1Yb32gfW&AxX?+?W-{6 zKWvZN?~*d8_eXXa@_tTtojw+O)8|s4n=Y>tv6Bai6Pz_@c@Q(q@*ISWy@Qh5t2&1_ zl4g><*}#zKNmX%BU=y=NiFKJ(RVEaXY;SO;ygOAihh2=l#8)>>DcndB1E5g;i`~kj zzb$*!4EO&Hr*2%UTtvARaBCwbTO9i8gM6B+PR+X)SYkooSRgrj)9l{qhmLGWu;KFN zZ=CIje&mthMUlvGz3>xL#B;8CPNe>99t!@Xf>Wr#XfiY47_muGwi>^sQKT_<9-S2s zv$r)!h77{0NewS+MZL{_Vo%+k%}JV55i)*ppZnrZ-;K6&82lt|MrnU=qchsFU_y2O zAC8P#iNsl&C$GlneSjh^qOlc9Tx`M6ER?oZfT^Y$Ms6BI%52BjkOm?%m?+$#*89#J z`!)4V-uTe_Cfr)mugsN`HUiB4k$)1qnd7H{vCKcl_4 z*_52&dLz`-VFBwiI+cqscAoI3dgQi4F-8L`A@HJPXXE;=4(TuvX>a9}(MYNR!o_t9 zu&nbbpBsaca9(SW3da|ZzF-yMt0FHYU+9sIqs3;21zTp?Y05w~Fcxqs7cr@XTUhPK zMJ9f)5i!DN`8VsUg)cECSUq1r@l_xMIUoa0>MM_MQmU% zwKvI#H7O`6BZeIp5zPZIkfh#@(pG`pH+@W;C?H!P897Mji4Tuu7vgciafLCoENFl< z{@79`IJQQaf;ZbJ`Ovr1p!@>7Ria>(Hd1aQq3#bi>s34%%ITx+VXn9;K-f_F(1*c_ zWDYzMme-x>xW{3=ndv&e5Yr;q(VIbvQOv9=QOD^9rEj3AAVcqCpdeQZygzO5tYLWiMyK?f0P$mF=CFfVx_!syohG(3AE#SC8v(vjwCB! z5C{#b)QD&>cZi)ExdF>YXT#DQEO()9_~9iB=(RVdwo)jlNRCn_k;ijetfwM^fI-@K zf@NU99r*%8pORwsi;1JJS3Nm9$7OWQiF)^fFFlqf<=qSWX{4?@Lz*XcK#}tQ18_l) zzNs8cnw`y_F!s#LyhojBk*95bu7nWW-OM=OAD9(AE=2a@$lXnYl6m6cwOcja74uS}kX$Ck<>3+^a(+UJAY#wem178+lVzeRbQb{>p!9@(}7HVD+OXvP;>a_l{qyN=RmjF!~X zd&!@3o>Ka{>h`MSe|_hae|o&<(ap%NAGx^h^PWsXrQyL!ad%bo-TO;!eMN*pZv+4N zC)d1tVNt5!yRXkbg-u&FnPa@~EI)gG!1^O8!Ee3xtM|j7RaXbgKRs=Eyo-G2!J6W8 z5EPqviBT(lcD|!BJ5HA+b)op#^Opbi(Vq9OJlbeZ*K^*i3Y5w&{oV$y4+B4XHn8mv zY+K9kf8*qrPW`>APu6RUn%)d2Q2vWdy?4Re8kh>;3yzhmrjd*v!7X zfLcuGjGNZ;KfOKY>R|ctd56)GlldnF+uQ4g>%+(g*CUs^j^(W4*|y_!o*k{fC{(AE zc>nxBts`&Vs(Eno&lh-io#Ugcf&B>Id*k!hUO#$94|mpcj-%)D=(yYaPkja{<%eJw z63#__@_5IyL*&k_1;z+^7pMxw-A%=|4_xee?%$drg`hN=`==`&U*1sM^ektYqx$CA zjQ1}(o*x3<^=JalB_5q$@zKQ=>m!Rr!O5~}XwPMXqNxZJ)!QUOQd!-bWXVz^B8>cl|NxEL+LLTj_ip-B5ttZ96>8eIaN8BZW{m7GDU^xqH z8pEAs#kQ4*&~wpR=5>t`nm0~LN+Y@0JDwjrhdv->#+*F885u)%4HsH5g#127=Mko+ zE@zpOO3Ii6r3l%hOC;aJLkO8Qn*u_XdEDBMMBy?UR;Vln#z-Qd#B}};pbb7aAVAV! zTeJvhqd+&4_$ZJ~B zB9Vk9i3kDuuFt}~0HH)4=Z`MCcvT6LX!=6b&HUE<&k?c^nL?J4q(rF9Xa+KlfCSDH zT?eW})VIL5B-esHkfP1b$w|0*i7W`7RL+TJPVYyGN$5Jlj>VmmT)9orj^DGbR)V#=fN!cVU)r!oDa!j2h7p@^8t z2|Wg~lXr4>DrqueBHFzEaFckQ@hm7qHgpEiRX*Q-fY4_rvMDC4MRuSoQxJkj#u1XB zC=Ei*2qCiVE%ONFOw%+qubr(~EaqJ82c91jy3WXiYP;QIy`|LI$sd(r2#((7-wi_|VyK$}AthBYxjbVQcSoOHsY*@_5~VS+ zpsp&SEV$TPe)??7veHNrySB%NfEE)bBRWzTGb378xp5SHz6_MfOx!dzH^KlZA+GhM0&kp=55F#*lk7WufUtnO(pVW|bnygtG(3Q4^9xDS;|9 zS}H078i|XEn`2-K~OY(5jMQ#o|-a1j4pZxRtnbvZ#@A<(8TQ0Vm!6hE; z9EZ+vvaC5<)@)`4Z=BXddVcbB$8Lx;O+oY@i@N~(w#SY=N@Ekpd}rn^5$_#FE3|~FkgOYnF&gzXXFePU&(oV87ZQa6g=1y}rIDO13Lc!)oUJO> zjrm***e|P;65oG+%fl;2S(dzh`m6W2%ckHv57xZ5+>(3F~{oO63s^YDC+3j<2t z@cZ95dFAB&f^}zA@xT9e$$RGq9$$Ct`o!ohDG6!=4;GrUWx+RZ&%bhm;onMGANhy( zw>*5_^JE{G*CnfYO*uW^TNl{3mZwiguG#~a2g~ohwWcm!xq0E&B=G*=2S0cz?!NQR zJN(ff{Sn{){`dLYzx`W`G5pD&{0RWt?Uuj!o4>(Y%X+=$FaPo{`NKc_!(YcCy{cFB zZ>oOY1u8{$f&NP{jVhxtQZTCuj1KItEQfAn@kLyt37%aaSl5DY+@4W=!HvYPs@?|v zpLZ|$>4n24#hZ85pUIEy$H1e@J==a{a98}N@7(&@Sg>EVu8(78Q}qrP5?W<8)#;+- zY*~Dz4&z`Wql3l#tDV!b(2U+w7{z~lYeh&q-n%@q?=1H>vzMDbB?WKaTVlfvKYe!K z!{>Xxai`(Iy5a6G2=<5G^7!h=%p|^bTJU>sZN8RG+_UY#WgED?ZoW`w+Q&pliAqb> zrexO-Tp!1m0q2E0`Lbl)NbapO+F^f?v`%n$vtT)|hzYteFuKg^Z9C6k z17i@VKD(FtVPLS1Qs;5h?P26{S5Xy378IiA*bgk5ibHQX4i*}f95{|1Atjqd z#b~FX9Y2EdC@D}Wu&5h0s~OfghS76pU2}bO?2awHt+;6|odvhISR^uxynnuBInOEC zhhxWXNX%z*=5>{IRUxR2;)o+A$E?W6fa4fQQPWz>)5~iLnK(U}Aw(Vrj+1`&`0B`i zdH0-S*Rz<{Eao%hZp-8I%jtnua_mL|0-FR{8x$c6P6D(v7;RWLB~?|TMPwKUY(LUk zK^r26c7UScJ8x9%`h=QZClz#s%6Nn%Whkh!tPlo++Zjw2x`5CSO;frJ}J zgs70ZUEkenrnj2y^*#L0~) z0N(?Gs7j>L69+ZnZAY{{s7Nw~)uN%ODz?Xg&gEu;i23!SEKp^EOo`-uW^gKj8wX+> z7)HnCjmP^yQy4DVj&T^Mji9MZigLz-|DU}#>DBB^%XF`|Tg}()Hu`XzPNy>k1=Nx1 zAkN|Z13+TJ6e-0IKw?HR0tqo=f|xNw445FLgoMO|5hIGMqOvpTNQdocckJeCR`d2} z@U9)U(@tlEGfhDe&oE+td#~s|MAKi7S2#k>!kcaA4#8%7%lb|{F9RQMY)Mnv>v z?+{8Orjc@CN_brMj+}c$3`KO4HCkkZNJv$&SnkvJBmFQ^)rP7P5YM3Y0RK%8+a>L;_h&*CHrg<8lqDgi$5m zk%WP|1zCdv;m6_#y(B>e4|zup0->tORXZU%!hj%_W@SAiWK9SjO}VdWDKV@Qe{fj&&17lrK|gD1%%-V31skC3v2@CikZk(eVnM}#!RAs!<_O0xoH1Dzsvx zD%gz}Cl`E%5J)1BQz8~_NO2|$g@_6vXQ+0DO$oY@S!Iy2JX?oh;9|R@w}Dw*Vd{!m zJ7aHdxcBg!(f6g9W;41W~Zdfc=oUb>W z?gDmnYzEJmEmF6H91C7UHATz~1rH%aMoLr8onUINxXHa<1YB7!ifdlirGb(H$qhJF zl}kGdNG0?uXNeF+AkGvqw^E{j-6; zD`jyLCO|30;5(G3TUlR+R!#2 zwJ9-%LZC!qZ`Br|b8G3n;JwFN-n@0d;3MCC?}CTxfwL}h@6_{P9hkR=HdyZ%tS{d=1w)9; z+lpD+vK_;hP%snQG4i;JtQw6Mg*)E+$m1?>vW<9W`O>lZnV30zNPPePmUo}@ob0AB-zdnr zU``gbK*>Z?d9Ln#GDP`#>fy!6y>o{Rk~gnEn|Y=(idU{JdGBP+(Sse=_M2yRZFV;B zc@#UNQeE)+p?wnYzton~( z9$wUESEU7Cy|v=4n=976WAurbGPRMcTJtk)YW#bxm@@zU&N<(GFtBryH*Ou!{Kz!% ztkUeyYh0Ld_tA##UTIYR@|(w>8YKRot-cTi;xGQ<>CN!hfBo0_ z<~P5|pa1!vGoP0cjt~Oh_{KN*lRx!NeT5qM$QxkO00a9lMU zE)2JiW?v|Vpsf^pt>T^2k(aKl7{i)-7dui;?9G3C*El73e7?au%d2~epMPcVr{0#7 zGynXzv;Nfy7zqV-iG@ZjW= zfY+}sa6a*Hm!WNr-}t&Ak73M^h<6D8lbB7|4S2%|Hln z>)L{p6+4@8KH|K?dQICFceb-5Qp`9$JYcjAA7R(^c;`8oNvRu$Wc6bY;$VaQyS#E!Il3 zGHh%aQSZi#z~j8p6&W9V9G@NbtTmeRHIXg25*`+ar}A#effyR>PPLdrGh#BAt!&y)BposX^#6KCcCn zqC(DsoFX#V;sVq&KoGs1xTgxCD^P}*;)DS(q@3w`SAMTF*s&w0x&*?hg83*at)QGp zAWtDuP&lHhSi~>y!D14dij#K|GA_lX=zyUFPfa+CoP4=~=gG~A0sQp$4MK}4+^+De zMFycHWJ7Y1s04yUv^}|6KvD%EkrQALRXGoWA}D831jZBwVw5C3C&ox!Yuqrh?E?YH ze7QuK25T*=DkFu#M!|0MWM4oOwTehARa-MSg^!Ye#(RkifrLUELvpq_x)a3(Oafhz zkw9q(kUco}VSt@fKwAXTJh_;oCC7oH1eahisWqlQ1LlN?Ys*zehJ&!SpNiYYg3R0mMp#vg%q!b8gin}x=aw<)c5~4Rb zA~)%P)$)9avpLl{9hq8(lj47EmCQ22x$;PAw@#w!nM{~ z5~4!HBDe}NDF()IplzC`&%ziJK6q*^XpCe&O^&z_kQrK4FbGf918?3sWYucUI*+x1 z(Rte1;OxlQ4@J05;dB9&RJ&LnaGD7bpS#7IHv5h$(ZrvvfA|@|91nvy@c$oCz_qn%6W&vK|8|C1$gllVM~# zINILPjR}1jlqUFCc83&{#uR*vktq;d*`ITp7gA$!oM$~>rYrH{+JqEFnS_zk}K&2RF9AN+t{_=R668=y6R{KtRHzx>O;@LRw2Tb!Mpu~>XEwedy0sQ;5I z<;-_K*znHdo(Jp5&SlzKW27Qc#?Zy-zi2lYzUZ7N6uZ!TN=O zTgP+mJdAvB;kkR#^QG(mHusYI=L7%a?wWT`JuXR(_hwvaN*Ga$iQYvnb|Viyu$*^^ z^PS`8UVe6VQl%wEl^5F2Q!ZvqBc67CkCqkx-&e2j?#Yh(X9E{~#D+l1U>d=f4{NUO zt*9%->Be!|c}5=zDPWA?=As10y>eyF_5J#1+L#+%EZh5xUFw3MC=2zyc@be;h@5OJ zKX}w}=fsjIj-Q+aAx73?L<&P)RcyP+ddz%qYH5w+rQ;H8@Yby*P1SL-^_+JDC+nWt zCmXe$#Q6`+Iu{Poz)Y6hu=4?B#>mAqjTiT`~#LTQ#Y>c8+Wz78KNe4O5nvBpj zU2hr3k&E+zRU^2vYN+&#i+<$VUc=x9&UT)Dv}C1FCbMW1N6R_ZIS4y!ETfnqMnV{H z-ZQT?&N~DNsVnA<;b=AEa5ZCp2^YJOwt~~MEuC}RIT`S_!;P9N3&Ak<^n*uLbG~-t zki7{qeC1$)4-c^}a@r@dFw{yHhNB;eL4a1I9FPgJQMlkSB9?L7rsjCRU;%m$=!Ps5 zA(WTalt-Ryds42Mw+-tqgDQ^YghZJNqZ>x+3P)8K_E!r+h`6XwssTR|lO^~OkrPTv zPPP#jJLZ+(WZkpr9l!XcW3KJbQSYAe;M5~jTgH|tW3at-6V4(MGQ_Sg%-5WWq5|(R zrbekE=33hop;&8DjHE1((UaVWik=ilk}zb|6tP8BlaQv0Xqbw4SbT^A0+Uc!gk2#7 zve_%s+A*{3cEtf+wP+EEA`^rG-4sXgbd9I12qF=DH-&OFOhr#_&kCp|F#M20btCV5yN> zLrh2{S}UR-$+4g_x?M-C6{MPB80l@Knawa&h4qCqE2O}NM4$qZ(JHfAwB+ascEFAU zE*Vf7scJ$0{8Iv!mN`0yC-1C3D}uV(z~gAwZm&0$2QYoYC#Nvt{bSEcEU@P zhKZb-ESX6JPi2iDfh;xX8K?@W%hW^izMxQ~#8fS$ft(UmtI-lvstG?+-EgOZw~gMCtpbao`GiWoD__86^M?$4>s5*wiiaD=yzcq( zjU~G=FnCJ1>$I2`qA-tB&QI5KdHVPLdCkraY?>Ort%f*6_1v3$tH|FDOe*rGSYu6SC z6}WR^*_&;jDR<#VE5yV<|L~meKD2lgUwvioV@M1rVTcD)_}IyM;Jc4HQeIOV_1q!R zF=d|Y9AgyEER=q<4i+`*uH$5FpIP7?Y*|mwXP7J=1%#5b)}?J-++?8F!xyyn1!cs{Ql;$Nwz# zg$TTFzx_4o@m#>()N=ddt}^u-$c>_-%3IuXhgbJh%1?8w($tTiRN4yet8M zZ(Lt;bjWKRQ%FwbDmt$y!|NPVtOj@3IVIRp_U~KB_`Mo5gTAO zGgP{qhcXWAanUeqXAlJ<6c%C1B}k-cW@xP##(`nCp>qK{cGK~x*lc@@QAnl`BazW5 zu`~`N0HrzIc62tCU@IT63VfdN{<+1DmR3cs9PV-JY71h`!?j~oDVCMO$UxId=B8m* zLzWdT3G6rmnR*gBVyg7-Ml0BB<)BjOQzbN|6vEMjMRK^y99G zHnSRJ&4ab4F&Uj<@Qx(bi0siKl7eS9M6#@LDU;A7QQ=i4)dCSAW63EXRmK=Y^cHU| zy546#n-QEP#EAE{1V+V3po}CdC5c{5?$wr zr9n4yjD!$8)_J57n3OTH2zElXq#_Rk5GMCtuss?3%%T$6Y z%?T|7!`R^wL}iM=O2HMF+Tfi<p+w;v_k(BF7;G5uE)tcYsv3+^NG0%- zup7}&F%(%yvNWJIs#=nzB1H>fC~nwlj#L#fxQQ9eEFpx_mn$ zN))cHA!0;eZUhIb1_?;vXjH|#m82xd(x7BQ$P$)FPDI}$qAr*QQLZIq5ywaYf`#mn zfUXwg7zo~hXvmkGT3JvHm!mZW15vIGHNjBora{RPIH)Alb&ZdS7|>P4-u{a1uE!^dOp&wohKQz?i7;*mc|wusnv@kJF=_G%kqZgD2-{^!D5o%q z_nIgaIVof=0eQ8qNJ%pgiV&!0R8?8mWHK&zWLDH#m!Pto+4Tb<25Kd#WJ0xy{i-Sk z0udNOVm%~Aml^GV9RgcnFjbAw4MJ-ct)UwOZuEp`$w(S)Sg+5BJ~Ern*;_4$At1$$ z!FhUL0wd8SsB+sCjDxq9vG*7O!8^=+#?__f`r(3z;L*iM=OsHYFh($6&KNu_T7!{^ z{YAyWs%6(ZPIn_Y26iJH?(K7We!)8@BXzZ*F`Bb}L}W!jj)ai7x;JMwt2telCP`g= zinyCnab#DL%4kALc%N7u&UkRW$46f{>p2Xk%dCj(Woa_SI z-lCNxhKzNNkYKJAV~A8*&{mS1GmqCJjmo@w+~AVrbUSio??czCK(L%ugwdg8;JbGo zvGtHujo*%J2Fu}0Dkg+h^4e9`3kv zV)0q=r5pP{@m=A{Udcar=cHp^U-19Fb>$O+@MqQNBi2J>^a-H7Xt5#dZv1CKJs`QI9#@v=eXvS6kOY%@p#vB z?{r{w`-Q;!GcV^HP+m4ZTDNcC2H-p2`OYUk^R=&i4S>J-o4@V=;jH?Iam^PD1jWfVxi?6d< z^XAQczJKqG`a#dpqI#xLuj6IS{!DTIp~J`YTzsL^&B&@2pULei;RrweceItFtrWX4 zaJqFoUXNsovv_Z2I9yaev+=9iNVFRgOd9rE^Hh8tonUkhpAFVUVo3NT&_?rU6KKuA{=7m@3-?uPShdATgH zLY0NMNGO>>7JNV@1X^pB%OyjKy5A2YR(6N}n1I>&M`Lu*Y; zkGbdugp?2+Ay-AXau$S~@FcK@n7AhAfFvFosO_5$8Ng!TzFw9GRO6jpOm! zm2kLjWDJ6~X&607Wk_hWG{g|8%0xb>@^YSHn7-?o7(2$HXEt9^&uVlG=0E*;`qx`9BQcx_iliN~h1Yu$(7U{NaTf7|z!4iBd&O4*1+66_H zRER0aC=kiMM@EY_hN@mNMv9YGBm^}g7ko#Io)i);D`d`e>n%AabUgS0<5^7T_$0hl~ z%~ihq4QUXu2(GG{5mPR%+)!Ap1vQXJQX!?D)MrVKzFdE@CJBiUlB{M$JWL6Jd+Hj! zoXf~=hX{^zi3E_7fTpVA6io@FDKT9sA>xFch!(;~wrh}rBsD@QvM?abBvd<6u-Kdm zra%?kM2rdV9Y$#)iJXfB9#bZ!NQjPZ%p_Tp84+P1Sxq35aw7YYOqo6qN<&WMP~HI$ znk*HvT9Dm9a$9n=$UHB>Ybj3RxGLu-C$diyjw4PYB4i;E-hxu(C`qDBMnp17NTk+f zt)K)Cc0GM`m)42*u@|%DqI2jM9ZU8?qn=U%0WR zE(YTmzz^lPUIyFc@RS!VRE5-n6cbq`yzP)F6z*(^>YJu=Br!$2PmFdb%{fzHs+zWG zXev#kB~2xH@I*6&o^|IiwqOqGszK>oFfLQzR7~I!3_;*?!J3qdJTQ1qSg#R6qE$sm zlBTK=N#SB5W{ZfH90NHO0aq!5kQGUp5;i9b&pCvG(nuMyZXk?ES$-Z>kpQVAm4JCI zF~G^#TAB-vZE94>2igT?t&&@XA)KM4z?P*Tv1lyE;G>HC2JIM`qD(xPHF zSY`$utV>qI7z92g(&(_(p`;+DL`tAd%NPT0JD_D|zS=8|?W$tVf_^;1Fk)1R*@+ABoChSocD=FNzRfg;;e(!tf3#rf}E%{&PRk4%qqjW z8;e`L63pi_9y~cCx^s>enm%cc_vWms$a#_+@6YK+hYbbI@M$&Lk=h7QidhvHr?9=z z2lnO_+tG5-mGHil`NW!S*=S~!Px)?dxSTTnBXIr z zvtPiL{l}^=MBshvt+)8rx4y-Cy=H&^Lk`B(t5*T|$AA3CPkg4X%LD5l{^1|Kh)sG? zFY3AL-Nzl@{$R`fbzn1Qmh+m%CAFUe)b9_kLJ#}2Yzvbs%{^@zQKU%L{oin(IpySSibN1&g$NLShE z%GDJgoNV~l2V1W1w?A=lX|&|^t8<=oTi$!T<@M`(&$J-5KJw_in6!qGH*MYO=8+bb*KG zQz?uVEE@SVsrB|z%gy~Q|MQ;XY&&weET4Pnc*(WZDfcf1F2+pA3LhX4 zkWw=a9VtRxw=1RRu$&-#A7W*!qBS1GTAu;yfEJEWI5$Y+-L%v1&D|kEs4#UsBbD z6euM*+?#Xn(J7rvG>riSA!Zhp{8R^jj*0co^5yF-w+~ys*IOR#EU%jhv+$9Z9v@1I zcSKMtg_4o|SrME4Xy|R39GJHXvLwXFtWtz&epo7n3z7KLRbBXyTGtHL;=L#38l@!1 z`zxdxSa)^`PAXztN|{6=BM?%Nkc=S_)8!jhCX&DbXoM_ozy3GZv{X=Y+Mz5>3 zxNt>8l94%;qTdW-j34$FE zxuR;zN25rHm>9=_7#uo|WT`RD3{MdnB$DWh12|oZqDYid=&HiGk(d=J!9psekp%0C zi&dC{PLKkj%J-H`5jb;U^``8}ITu_3#hIGP#4M)ptjT2@EGehrs11dUTc-Urph1aA zq*IU`_zpxFLC?{G-BDU0noLBdgus&~bcEDR2^pIN=$sP>g{bG?J-PUAix?Ki!f4H9 zj9iLbgi_m+FWtZD^0_q$%EEvu;cZ+Z4txjMm;GQu7bHo1`u7t8J6*+-NR=bHY7nMI zYsJcFeDuW$?n?xp7Ky9`%1CTXgyDJw?9FV3y!6$V8%8AeBT(jn<%LVx|KJ2XlNjbbVnnpPd(% zV^$TbSzQ{?l5!lB#K@W$!nB>CxVUpdnTjA3N>x;Kg>{zX9WXUOR9((#8p+u~3<%W} zw=%`WO$-x$A{Qaq_lOjTDdLhMNkv=NTs@rAXvwbY*yhM!JyL0|>@7H0&Cyz7w4yNz zDJ5rH&tg$%#@o>nW2UZV)V0D+xRIu~)* zB?3*;5=CU!d)B?=#?hR%HiVdp!#NdXNJyF85P7ogiEe=4&^gg*jgWzA@Ek8^^h1wS znpbZux#$AtJI7`Wbg6LEWtuqJ)Ad!Js|ueY+s+a5wlHE-qNxlkT>|}lR@9Z?V5XVZ zn(bietfHB<2%(r)1?f?80}{qatS$E^nW2fD$a($YgcB}TUd7v=RF~1=K7Ojn|B`XEJh^SdCRKGbT+XaEY?@-%?ua2 zk#3C4_X;w@{fJVyXe(Ns$-$A5p!o;_qB4q?ju#l!@!-s|>qmkT3`z3f$redw5R#2a z9L^M17B#P3n^S4Ym#)r9X~%=Ls8P-daALo;=>0@%{IE9TWw~rdMDCULFQ$8lN;<+Xy_uAr;f4Fm8 zJ9s|CLe8-59gXIvm{y&h86ROYzo-}WS=HN68d=rzHx-PQq%7#|i-y7%BJlp&ul*X| z`qsDjvp@SYe*gD>|LMQ4Rx6ZJ{N3OE9oE{XqAtMc=_yd$Sug5E{qL@SeX!+!zPIJ> zdEn~)f|sv<9GPH@p>12vwgY$WpB5ICkNjV+A3oRU&ATT(4>yq;2eaoAMsFQ1`0jgW z+DE@MAu*ZNuO!NtJ63(xVs;rkz)V|^s1j8=lS*6hz34wls? zI`+%9LTb&)E^yI14xRy||N6BBkJbahZ}D1E8e`hVEH+C zL^;FT54L>o{*DhWLP_FX)GV8VaBwkl+Bx1ocbx3jJX%|R>9qrzp90@-ZLi^IzT>^~ zfx~4BWFDOL+&k+y?-HHM)RjRiK?a zhU2|C8NtPFxq@xA+7{@43G?_WevmG!?p13^wAuV%9)RmL$qp z?)y(Rg&AndB14;^F||e zg;a@Q^b@1+!|$qf0T=UA#$XvS1SK*}Ww0@zrC?Smd@SSWQWiI2icnRCkimsGZFk1U zh!heZ0~Z^|=prE|a)#goW4|lGYPvzoB21p{@)ud16u^DuNqIn33J# zY{$h~5YZ%|FuHsfq8rHGGPn^TFT>OzhY^_rNy_5>6zZu%S3DKoG*EtrRJu4NbEeHcWb`65 zS=Y!*mT>YQ$Kr_1X+jFrWKoghjywc}GALC|fpLJ4NL4}3i0C0XlvbpW$;uR`s#FMc zaUN6f4EUZL?IgU0!furkQMJW=J%!t) zv?PCtS(uO(zKFdgTdv&CiMAZ;R5K;$mGhBvIfp_{F6tWM2)PK{Qq^Rs$kUjAEn%;1 z*_*dTAVhF>KGOR{Z!K*DsvF6I!owFZ(-Nr^DQXfm%x4VU4xd0)EjbhcGo%SXk_w^) zRTZRz(x42)fXu0k^5=vUl62`3cReu|wynw*Wg3)e(MFRJ%ojC#^U~N2-eD_45H;<* zMpyNOw)yb+bGhD{RTWw!I-AHc68%WZE#7BrjQC`UK#H<7L}GmE-bdFYKbENk6xK$p zh?Be5m(aNwNJ24R%xEi(PL^)4j5(FY5hZ9&D-cSMr6h(+LB3prG8evXCgDRQh7mW8 zh~$}R&;D|cC=?fcLH0a4>sU0J8;1*wR_JMcdA1ul+gf_-%W)2Yh0fGX!*Vv~VmEUC zbj$sdEmd2iRLwXyACqNxf3Did*)gx9rb9d{3{m;-%vy z{a`t2G_{ue;8Bl@iCYIPuU)A)SQxIXYK|85)BDZ5(Y$qYNo_h#cY(7_Pd`TPK74{v z3Z*l3U2!~bP^vVDODJ2$)RKd?C58y11m&iWXg5ZQ!nhvekDdDZv^v{5He=$-!SfqH zX0^se!IO>U&Z8aQ{$R_yrvYU&$NNiOX)2T|ft!PiobN2(c{p(1Mb_QG7%F1Q&uxa- z_KtZaSk0b&Ex$jnI9_VrzZmIlcrJ{^*=}IfNP?PBSf8Vu&s2(^`K0NK`aic)zT|&@ zzR#KCy|FX}}-b3I;LzVpG3yXS$|ZybE)1MzU# zQr$l0+wVR^J|0=N8@_V$*}BLvWlq+X&6s%Q`tywR- z^qg!*Mpv8zTEeQ;Tv^q;a%Ik}{_$P1myc&$-*0&DN$HSZ^qvrlo3X7mSC%!m4rd%K zf3_}JJ-Gy9{(OsykTOd0TquL*t?lUf_6J)&crtQ6K<^6IOiV1kMtkP9<#4XKeYE%l zJ9mFpKV?%ri@>Xu;J?0pNJyu=bK?2l2j|>4oU=DM8!yYp%#%&e>3ZO$eZ`wMW_mzp`Z|Uw0_z?N!H;+Fz zhxu?(ar>a*WY=^5bcauw4^DfYbeX+*%PU8p!t<&rrtqW5oJEH-0+fY;=L%dJc|j4i+CSvLg}XSO5SZ07*naRF9Sw zzxe7Nu;!oNJLk#HA(iIw`GAcAg2%g_lxlKTC?(L-IBv`om#8Q&1Ud)WN=%|qLQ+QK z4;E&63D#`7acbHmPB#|o0M2 zVT$WYXhO1dF0+u1f4zUo8#fQBji#Bcrq~FN_LY?n%=*aBifkXGm!#(XX&` znZ(?}Lj_VM=ZKPmlqI8e#Iz%&i1Pt0BGpXtm6xt@y4#`715MR3gmMp^O~SrW#bF#0 zD6OVrtI#?kXz{_5$f#yXc02sA2E9OPgNO<8ginwGEi*2IP-+r0i>N46Q}nYOKuV-B z2t*0Riir?xaWW!NvPMYoZlsGnYI02))8I12ZJi;eh?rcR2!#uYC?z5Zg3sjS$Yi9D zlkv2i;FK~#6t`$f6xXXL;e87&uvJYMMsny#|37>#oBAF{l52eKi4g*gf0nTWa^HNQR0Y52uTD)AhQf%W&~0* zg6q$Z7*jJMB(axKtVHTDOfB97qDhmZ@(ka1>AlAywMqO{*P$XP1VTvq7$7QeJu&nU zBFbb)VNlgHImLq?*|Z5a!8OA-3}6SMGU+&_f_#E7C5e+fx>-l>lZc52iWq~ylo|gy zg07IJOb+E3M`VBpJW7t|DX}O(2%HBU6W6!*5#dsksI>v%dImpWjKN01;2{miXCr%+ z5$)gzK_E~#MiIF9t{SLvkfc2{KQZjA#=+V{B4I0&-kyNhGE$sHz!FZ}GvWz%eCR zZ96ub4b2dVCc~3)RVU_kB(lFqQ#aerQY4@{#5CREyEWoHC?!&+&p&*Z z8UZQ;ks%A65Fjci$SgU>ePT~XHz1`*I207!5GbU?`A8HFf=TZ!al}DrlnC^FAb0_I zgzHjcYNSHAu15Nn+@y(@o%s$~p3^i9nt%%fAUQr?VXdPoG!jYOyPIZtsRGUo>`i03 z*U%spRasE?HOE(5y33x~Y{s-KDXSdUrip=Efwdid3@E9{ixQTXDgpx_vWffzkg z6FJo^Ei;1;kw>Rj>=c3*W>ccj zJlXWron>c|Q|76u-&;p%1V@t+r3Jkm@KN)fxAxheXP>(6D@OF&#dhFg+w)I9J;5l! z?ks0oC{DM57jIA5AJ>?Jk2oJtV{<0cl2zSvvgw(e)qLabM^kJ{NnA*d-&yf+^~vjFA;gG3(;UnT-Z@^=cF+zzr>mA*&*7+zF>`WeIZeL-Sq5L@=M8! z`KtbVR;DCMrhVNf?MpwaVg%x2B0h>XMN2 z^FT3tT2;BG)Pj@7excX5y<;br&mFq(tYux`@7})P?cmBR5 zYf_TSgG zwP>Zdw>PEfEzx;)@<3p~_G8MufJ_PI6Gge4v*`w|+5zuNluFmi)z)JG7kFwh8)M|b z$%Z>~$!oVJPc0sf76nzlMnuo13tZGKtIi|yIh72A;23O62qJ|^iAWF*S1$U`)bcWJ+O&fGBcaIjqQZ5)HzLd+;GLuN3*lQFoSUnS{qQ)}0v3S|*O5 zK;y=Mv!?6uZlJZPnUnS`g3hMrpL2oJb&ZmWc_BHu>bdMJ0+>%L7W4VY`h>3c2%RB= zN&L*{@wUZ9oi_Q1QFxUol|p0&o0ztW0K(vHa;sS%MrWkK`$#`{j#q0ob#ek~sqr$R zbw*>s2Zs`oez5dIz=R|yT^lu}Qe=6-(0NRzk;*gl4N+Kftx-CSKr0g|O-Aq<*9)Sz zD496ENJAikj0p{ZM2RFS6H=XrvLS5Gl$O{I{(;dg~I~ z*T=+4RT(Nv4NLF*h%4|njyCHAN+xN|U%R%2N*D~EM%nYC#M*RR(bXUg9*MUfg%}Ew z`;pM4vFm(>u9mnU5~^Z!zozpTiGCPQs2-63L$-#6qDp(p0FdBu0YYIV&~HU;<$T#X0=P>f=A3KA2-B_9 z0+N7r7L-Bik{AMRXo=o|)MJR71Om~ezfov|6go9SYz-nCPxC17MhP)Fa#7hw??sd& zE@Zsu2}HXc9latoD33^ z6!Ym67gD%V4C#71yQ-;2?yMJ*h-L^NP0nT8;zLVBP-YqZU|CKS)(O^a5{;VPV;f7~ z*LW0Br655e)8v65r0Y~<2vwl69N|;LK}ww*>xhxPEpMa;*?j z;sgwxM~(NqQG#4+Y!GP-estJc=g?A)&BTCKib-KO+9|o(G|Y0iT-O}0ESWZdXVayI z>m<_{NecPX*(c)0VoYpOYaK-<@j}vB$ywuh?|evsy|TeJJ>F_6<5@L5!UkkW;^(wX z>%v>RGxp{Mlb>1hPRgA9dB!w@4~{SBT?De5^~z=KIJsyC6 z!^PIIu3Ju5>3GjqTc*ZSE-PL?%z0^N!K#x8{O30Hg-FeCB_bk5?(UZSpTBaKsR)cg2)Z!yWKJIh7wx!CsXji!mz{1wk-=(w2{{(>Tq_mKw|4P{yK>g}D6 z;jDlggxK{uF*AxcZtw6f4=lEA`03*nyVLw;&+vKm_@ZH&MeZ&0kA)>aOWoV6c=@Vj z+qc|*yyn$gi_e_m?5yLR$1Cp41z)>8#rw#4v*F~b;r8ycqu!DsV&IMa;;VqQuj-j= zUKysjWIHyOo~P<=pvrhIm-~O$`f}6{;7|VKPx$TM{%!vFkN=pz`@6rp`S0KQ*0=bh zKl&s7{LlZK&1S>j{LSCo{J%f=gFoPRe&=_-tV8;$zN*iy({<1B#-oj9@eHoY>&lE` zZ(i_Z-SPfe!`JVAzKeHo$+7jk{HNg&!TqzIt4?rxIpM|0$1wrc zN6uCaKYY@%={9UT%Qs*7e7uAdg5nt$B>$1xnHUz8;=PL=>*I5sBOfB0cHpIjVLAEv zKjdSCAH09XkDqj`2gT9eoZb1y7d-+&nH$R7&~!OZuC|0&W0c@y+jd~v zu2BT08Qj?`xmfo+Tm@dgyL=`$t`P9@(VQQBaLN1UJue>&U)b3_$u+<9!h(N#XT^_J zmYqeJ8Vgd~uv$Zmw7ui1?h!F?Z>fq?+(sdl~=s{v{8u zJXNLH43a8?uI*?C$NAMJ1#@)@Z9Gz_8#i$r$BHr)8LYzyPbLH|!0D<5pzkgHFt9(% znU{)td)3WmXWKfC*A^cn(n$K&(zG=~XhIa^MG|9Y(+PFw@Nz&3#n3m{!DG7?;X7Wt zvqS(>9Z*<~FFWeaGV}uvJ{*`$3!--vT5>ecxVu{-W!hBxz?dMPlHcz?+Hf?}+!_M~ zGDogI)p>s?NO_3Kj8$0W4Y6=LA(qPIG4b=5E;SH2Rh)DDnZSv8% zn1nWw{s=@&BjX-JR24EWlW^E=@zQa=4Oj~?IEKL@v?h)ZfOV6cbAtoxCCG4N%nTvm zog2AgY2@By1$G!Hj6!RP^8+F}v`C`L(D#I($W2b37xXq8^TQN2M7&EuWVrO)ktBDw z)EVCP1e-*S;4SEkD4;AVOq59BvHgINN!UVaqE_RS{g6aYNTOa8W9Sxyn1n-|B}j=9 z8YRc^{d5w6}n$q}?LMNxLcLX7k!VtAc!COOzmjg!6l401TrfQ~>fb59reJ|$; zj4)ynU<~hy0Toh{Ck6|)N#d@qP(l-Y0G%Oqo&?lu?Ez3(L3EA~5|=oMqN4*@kV2Y* z9fjA?9XmSBkr~7R9|J;Zf>emn)g5F;(+wyhPD(5uJeW+AX+zUCw7noBr7vX5o@N_VLsxE}8VxmMKefH4`8fL5M%7`R+J zl<;F9rNNRomy;r=0j`>wZMPxQus_W(mEmHOKJOXB89f>p5TPGgyFOh*QW3pN%wU}p zg@zDQ@ZOLd-9qJwbxTYdyC_nVK}sSK7a}gEYeo_Ywnt|v?Y3nVK$ZFG4NS6dJ8@M+&EDMDcj-r_H z@Ty0voI8gfHIAhiD2oCs({VbVu4f@6Cd&zE%0khO9*Ep%j1~l=XeIPbx~Fz0={#-v znh!4;-n_G*&<$7pfHqGH3~ldt{nmtA^Nfqp1)C7;f{hBaGBPVPWtmeCkt%ilWsH%! zx4g7B;nt4f{)a0rnx2cb;9noOF*vnhvl-adg4Yf!UcWQr&a%V~16pamekV2BueO$T zXZhfC&C5HHuN{{Bk2m+Yy-Y0Kzy0wUKYY?)lw$u=&{o&gS;r6FJEbfnFCXoU0eb_P zmh{$P^k=Ubm$RJ9b<1Tvus6%c@LNZj70)H`rtqn}) z=c?}5c9ukki{x4|uQHa?g2g1`Yj=0~ho5w)F!1){E8e)h^SL7L>AGXJZFuW;&YRD$ zxBT}d1pMmDOI(1To^<@^;RQ!K6+2^o!*vbTbH3>~U$?xl)VzLJ@aCNv-QfA)qUDEA z8WxlMnT#_bMjjk*xHDJ0erx&*@232!p1lqiIm@cy_^N$wf%kOXv79K5cAh_=?Y~=n zX#y`K7Vq!<-tRrd)%~+S`!jy`cYl|^{L8=O$3OltufP5}zyJHc&v(A_oiFonzN)Y4 zbL({7bJ=+IX4B_>?cTiLCl4D=*F9gq`}yCmA7ftpqRZ zSI<4(@htW5yy2bGj;$5k-J7uIUrEJ6pBsLa=IGF zy@3z|g%<1-lD%ohi~AK1PHW0Mqp^`y+jF=#V?O!hkGZ{Bl>{Quak<)16lueYj~FQt zBHd4yo1WuU&2A}pbwA_RUS2-6Sbg`T=Ke{?MGJ4~GM{9WQge5&U|uG-RE!auucQ_v~eI&-f<+^4k3*1^xZd@*DoYf%+r()dkx88DoyUZTl!AV6w1dURG-e#Vg?vI#1v<89og9%qIzmX!u!OBu#3%@ZOB4hI zGB1(&45by$wka%2rXW2ZB>|5c8f2W{!$1s*h{hu@nIsAbLSjso5FFOxti=vVl=DFn zd>}}T$R=q|?^|+__?RL@g6$9(bXj6_M(+l+F{mtu-V)rklU1RmLaA#6_P}M6a1In?wdXRYuZxi;*~Citk?^qwIm)IuPM z(7`8%t558Og%YF9YW+NDN)f9yeEh>p6`T7Vw_7Z)Cgb)9AU(zsKf-0 zk~j+EVidxYYmzZc&5!3(Ae11u4s`T>DQigZ2)FA*}L zl%dQuSz)j)pfZ_oA=eCCpMvBVYGh~`>Oj6kjIi^Q9wGW=)uY>)529p8lZ&6vomJA8U&^w2y8_L2`8c7!g%@ERk=L4M$1QCeQ z(hmwDB%SkgHl)8>WNfXVU-euyj@p8^f;`X2irE-Qrjtt?kXk3Vyv#u9#FzHr=H7@S zE<@%DKQf_(km7Uu(QZ0dx*K>+1+ z2D+rN>DsG|PS4^9Ri-d*w!_@^GK=aZKTYyPc%+6y`vpc zGqlVVMoBhJkMxch0{0(pD9jT84yJ~Ug?Twas~&AM_YRgfLkOQ#+s>jzWM`6bZ-2_g zrsdT=#j8gX1|RSNUfG%Q@M6olgCLVY|Fb0oEGk0?o~8?Y`^5>b9pwDx8+#uUMZfv- z5+5$Of4XJc4cywTot`d z^3m)w-zVoI?><>E%>%C;RlKlYe(LvrRbSP=ufuu4orU3e-EzKZ)86SBYE=((-N2js zId_&{U2|WmzBueZFbgG#zi<%6RS8^NaMKS24o<;|-6l9ACS$!`qLyoYtTJeixZx zJ}J4ZJDyxNymUBc(>fNJ{12aRA(9)R%=ITa7`C0``|qy!@yWpMV#3krARese?Z+#g ztSp-$vNM_T*4~oA1>XK}MUa-oY(j0}s_kj4B2*I-`B(Lo$UfL^p>)!k+G|Aus_m8(c zTshvjyF@9$rX6T{&+blg)DlEwmr$7GWS-13 zUffIfWsC{`VHl`YkM)5MPF5Vw1Wj+*wml!5_nd4!U66!!pmqY2Y3jBoQ^^r%w8n=( z+jUsy=)IuG4Oz7y3XSaqHV872^Js0!@`2?LloCWX6jY23*-8-gLN!dr% zO-I+RQRNO&8OkzaupWs(Nl%Os=N(Eakdhz`&RT?Xh&+Y*BxI?ivle;7a8~K~Opfvb zqNi;;f^!5TF{B{4zE9)zLTH3YVN#6zP(KU_*B}Khg2<9M)Py7uX$gJra8VFLkI)q% z!5@e+1KEy92M?o63M5D7*ys;F&<{3oP@^Td#F-@~_peljJWD}BgA4T5W2C|(n3tNv zdCp`~@Xq5kT+}JpOlqR*lejiI>4ZpvX^b9*_S)8qlqyY4h-=(~Kq__fAd*62w8Tdc zP7qZRx5OgX2*M|EN2t-+=~K{E@Hjl!G;I)7LQ)7RZvK8T z3P?f-)X0J+B+hNPMn_x=yhx%SUvu5SjgGXWPl0qoixF?D5|h~3WCQ&hA~ zgA1NK%ZUQIA>C6#X2>j2zyltvOHAU#qyPd^6bVZ~7=O3F!2{ribe%~_v|B`sL@5xa z8pGp~&@K?f>wq;ep4_04!(3V-;B6A^l+qZjkxJsCpy^z4pDRU`=ShgFw|H+;QoYDP z*d+2thwT$G1PR(}q}B+jZv;+j6Q(8^&?C_&cCq&sKO{s;j3Hq)q~c?Z0`Eo?NI>WU zWDCwEoQrc9Cul`ZKrx%<%!(ZEdM>vOtxNA`-SqT>MQMXlik(VPl_h=G5~N@;pVIl3 zt(61;&U=solNUrZ)(v=@n5+>HCVf^*C6kCBJW(Eg!c+hNAOJ~3K~%`pz;WI6r!SD6 z0f`DCi9RvVjWGn3nhU|zpd4X{#LT4Ymum*Qg&3&ohUfxVtyyU*5mqyp7!2oAcnbe(EFKxiP>|fmV`k z5JX`Z>_CV;tyVano~PCYtQHg_ds#?{QmFx_WuU3IWXjOmfs1-zT4ab&bAGv{%4YOK zLDN}Qb;q5B;-!NMB?VQcDYRtOIW}!zr%aP1t6@ONoJsYxJJXKBz8%BTB5<{Cu!AKA zPbLKyE&TXN$L_3P(>m%tFj&Vl6GSX$t1bIGpS;E$Y~W(uu|JJuCh`w&ACpPXONUdw z{pylMmElH%NQ{vmKHBi^X~TN6<-z0Bh*Mf5#E|vOa>1*+ikA)we))x+r^fpa7dijq zjXlcjipML*{YRHbr6^2799bLatYc9K-aImFo5;g;;@ASrtBjpWb8*(w4VEJNbiq{! z&{AT3WblCxFKb>|D(>w)pP+NNDA=1westV&v?!m;4D!L*CS{pSORly(@1AUUvQE=wF1H=0mo0~L$!iBCzxvY7&ECj|g!qFbI6p7X_pAEFC?Vh*_vT!+ zmLDH)Q6hzVe^zbUfe+3$yuPRT#=ZGh*A`!r!2A90f1em5-}=_KzSP(JcP^zqdt~FQ z`f|!g@CdTc0d)VYa4N({W8=n*BYIJOE~BvrJoTywZ6QA*)M zG z!@SJ+>Dsb8m%MN|rR@fuTx}7eCO4Yh*{6@cNYW@|@4-j$x#_vR_58mNu6XaPO!DTP3G>o?V)G(}hh+F@`Ec9Pb_3#5HbDRg!E#bSv$^Zn?huS zKx;)=Xy%on9nz+HQRQ3>&%Yj1;9Q4`0vkb!ocpH(@+`1( z+;RJ0My4_@HfJ1_kxk$7gLhg?mZ43CR*8W+D>Duz1!a+(;_seWntscxM-wtFFfuaF zBNB;}l1#$a?=2{@ns-lI9(0av>rmR_T}Yvewx{bIDxXp)kMA3-%+N&z2zm!$uxJTY zmPX~Zh)6t2ON@lu`%_lio^|WlbjdM3c!@GuVloDY4+EJg$&|!7M;KYS!ud23FCcoG z#08;QRB~-*TH;+GcuzOD zPEFz!_dT7Lc&YH#5(kUbLmKS|AmF_tVm{i4 z4jmzSh|qhI`#3O$LP`3d=xst)U@g%meyfxt2T`dM=Hy-a(k2OD!H&+<045;_T;CDN zNpWd0@o0l3DuvOKA~W>XBW;%gt;X{Y0&OzZjinh<;Fwk^^()$jFtlk5T}Yyo5Yjn} zc%(82Z3w}Q+}rfDO3v0adLPDc&g7qs7#Rp6;)O)$5>w65BBFxDOU0)5gwUgOj*wYm zHOqhx={<@*g^eYgM+#n2S`$T1h@Kb(LMrePsTJB76310AbbVs(BF31tI0i2={(E}; zO_1Curl*NHkMN_@R|!Or;3G;#OkOa_K;Vc$GWg^=F7k|R(myMlF>^j!l55Gj>1YR^n5aaA?~qa9h9m}C z=Ws5$nLp}EUf&$o?&L6Xjuj{~o`vz;OPK_G;fStHZtCl`p@7x~N-dd{hIwU}PjZGH z>ZZjMnuw(C()2@_gEof32t@SML!c}S^Sq$U12$Nkb9n2(gH#$94d4(e!}lwQ39TSS z3h)dbLL87%K^O=_K&UDOE=Et3AY~v3LqMPfXr*unyj0_Pv*a3Fw@r8sDe=ys7>L&5 z8F|Js1ujNklW9XI97ahpEzwF+m3hJs$w*aXXsy_t&seQ%&Nh~Z7j0^WUk#|NNcW`> ziFy1ppZijcfr}xaw!!0(2In!uk}v+gp$71~Q|`RBCjUnPD->aKTfz4Nc$k z?jy_oPRY7Y_wmEy6_=Zqd2V=NFXzRB`AwkKyfT!z;3rSE>`ygczdK`5D*pc64Ne$@ zc$!(-*%a_)w50DHoppHIkx8hEoV}UGNWrb$ii72ZB5%0=_>y%quxTPME=z`C;A~a1 zznFaNoNPPGhvyrl@7dOhe>`t-!EtM@`Ecd=zkhha{><>&(Ug086+*zTy|T;8hf~g5 z!K3pHmzypHF-oZNjDuOmiw70EpCPgwEK2^5U&?uKTJzy$%SG+zU3z}$iC|VFzVE9? zQ@;Q11s`tPG&vz961aCTrS1dAm)pM4 z@bJ9hm7~u$!~k4wQ}Eop5X{S*cTYBaxbo~SX3VOm&yP}qS(P&@bI#Wd53gD}=h>|^ zFC0vH>0rV;$6HRWIyQYA8UB#ZHM^67JIj)n4kka>X~gAJv7C;a{19URgdoU(?wXwi^N{8CjH)uirMjdN{d36+S#~cyiTozUjI4Gi53TD&N{E zd11eLN|3wW&s_iWSDMH#Rvj!#zWMSlaK-)8Et}49Ygs*&1!H~M`>&gp*Y^#-^5UF3 z&%ud)p1OWM{Qu=z{89wo-}&f&rCGckwWdPQ^AopRO$4)>njaeBGwsEV97?#+1l z?jb@5PF5X*DA`+9>@641{IdWilPUk)dbVCMn?K#5|F5d$5*CXcmT#Z%{`tVWXElqe zWLg@u5;z~JJImF!V_7M_{^ma4es%Y`(rKTkPMZ}@7Th_gm`*3WZVI;cf)6eRb{83$ z`Rqm5Vm85x$ivH?*KW=E=GXQ(*nhtF>htPnpB)#Qp37lHG!<{WaCozC{orKHRWDGc zNByRgfh@DwsOY_9e|JGy6`Y)}IoVk594^tH@gX;x_7tUHGMh8|jPbX@1^(&Z zp77Ii*w})vzq(*nX7tu`*+gOo+xoCOw?t=gMt$qHl-+mSIj1GhQ=bl02+`F~IbX?#UWs5O^sw|LFaO4DD z^lW?2-mF9j!R~Sg=P!BhtfmqxJ&!J19-P$VQnB3I z#Y}qI?UvRCE?NOb(Y3IuyPL@XB??NTDY6tM*LNNEI`GTn?96kT-qHJz7A`TOw4oV# zhR`#gOt`bVM>{kOBG5TOZ3RK(6s1LKPvb07M-KOLGAYT7WLg&N%nI_%5Mz2~9dA20 z8_>e>zrVSU_nwO`aHw-8Mf%%+Wj6jnBL#29I}D|D8V6*Gj=7#Z1_l}w6^)2j`+4{V2s zH=5o?2J4BW&`L4h@M*+e5o`@Hpp_xWNbnAo<>>J{)tC@}eLHxel2b1Pq_Sxw9|5WW zw*l87b&kL%ky7d8t|f`6kWd*Ck+{<#f*+6~Lxi5Tb__O!#AOPUNWyO)B_Ce4D6JX# zmQqK8&T%rwMv0Dr5Pf38MwO0Nh8s3-aF&P%nZm;$AP{J+Q!2f8c+!TI;7QbukPsIt zL&%KWL}o?K5EKN@dRybIC-@d!F3?6YbRC0-pyW8lpTsaByTm^g65&Tj?5GXu=+oz} zP-FZR3GpBl0)G>L7eh!wXbT_Vp)gUpivEtDd38*mQiTTl`w5+XvV2#6>Mp(ixw$ZDCc^)QeaH!#egfVTs| zq=wuzUPVf78Z<>!F<;EE{ebhH-dS=LkvO7E?#rk$)OdYKAu$G`byz2mqCn}2Rc*0D z&ombZofCy3&oi3dp+(^SS)X{~L9(|H_{)|}+u>ZgAI!KmH^x#FIW`EozE7Bl5J6~! zObP6HULvKXvz7p<2}nrnasoxBxpTD7<)-1}@`7#a`5?emFR)%PDGO$KV&K<g_EgYymZsz6Fj+gVoI zE!Ga$0Gk007bW-hDqg!aMIcz!J*Q`ut?#M5;oGn8lWECv39kU(e*KW%c?KIWS|k_l zb8H6=_V)m22bU&eVk9?;N%4_s=HV)`I1`xsqd@BEbV?vYpKLf?H5~5Fm{w0431Wmm z&AJ|V<)GrN+a=%m<=bC~qx+k)Iek3l2M=pbHjY6Ry#4XdW7w-KWxlR|U2Q}Dw6lv!ywTX%f_opT;u44iBN)|(sGc0YI?ZX=KD zfzuYg`N}>ApTcjhs*0-ma`tpz)&IJ+TrN4>U#4s5a^TTLiyRxekV2lFa~#Yx-?+QW zufDv?3;RF+$=y#nuK%pE_KOA0!+#Q@;q&VF_;_4Qd{vMf80+==zxptKG0XSJFmQIc z{>+hp7x}7PdChqindMH zg!7U29$#>LWx022Nk8;Y{jA^}xfZPIfy<3#l7SY9Tle^)VNsrQ_cIz2o4Uc3ioUHA zBl1(~@83S-2M-(ipm^oh0vR1mlg@*+J$)bO`*eKEiDoSV@18VxXZht97Edke3mGYu zCpbr`Ay=OJA3Wh`IVb-VRD*Sn)tdk9cE)eKGGpj!hEG4$AFI=g73*!st)-=D zlGx&13O?F44Y|?G3&|#n^n>Hc#fF_p3b&|oO{F4Tw_&{z$oYgABI|9(ZUyr~Fv;Nb za?1xN8$LMSvQsHuI$VH&%0wpff~^zm%;0j{rIFvdrR`HFn3?2gm4q~ez?h87s}><( zx0)bjVsiTsshb8r3XBOftb*IR5GW8E+kF-u(IbOTjUQs#y$=g`>%5d+Q+1UDc^0g_jjE!m==6oGjj!6YH+$z@GM z5vAg2p*fi46uRWe+A^&Si;1DnrnG}43<2*tybBP!6kw%H5<9yZ5r!n}_=GeF(Ifnj zT+BKnHyJ(zd`OGXwi7t-X!ON zUBprlb%__6T%-+_6oSE8v?<6{Pw)daS{%uV?0p)a&x|3;i0fN?bV+;yqO8C*D6P`* z+a9UXIHk7(DmpAF1TM-95gquB;QBE%P9?$7Sz;FuS%nZ9baJQT-H0MkNif61;K^hp zDv6Pf&}|4&k!eMh8;U$*Z~_}(TMrEV0HG(tqGJkz6CyEtRX&4|J{!DGPHX}MVsc>{ zoyP9N$X1nt7{<~3=#eU5l*GFso%;Zpl5``}R;h?}16>bsL<|^X=z5FRk}xy`6j6x( zkG(hPvg^vObH81l;ucY(B)|%!GEG2GJ+QeTZ!LoNLz_ugyzSx>SgD+?u= z1popOfrz+qi&O0KHTDURND@qt&XnvF7RMk4inwv&p5pF3*IeI3RW0Njw&|1^mh*&G zNE4z!l_0U?34S1|3MmaK$Ks|I5~V<7Ny-#1F*n>BIiWQsWPs4+&!GHzgHlMLaBQU@ zgh_FMcEj{I}O-EnCM+-MQ=Y)_2+2MsiMaWhmwLwXRHHy9; z3Om-w2}6-F5{wXp2_JUUDWS<|vZxV8Vzfb8NrmMWG(K`ohyMbZY;m3})vJ8Xc zY_lfjf*W!^A+_b<$&%X#mOIOu`$wN62=R$7;RdKYw(DR+9OAP6z=v z4Aa^O{0+(DThz^pD|^^}L8qvs$KrW*e)Y5EHFg<@7)j8+r+C~~mW|=*xg(7utJRkE`g5a#0P35emeH|(vzPo=spA%o z+XiU9gs98D=ZB}4{LQDA{N&tISHw?FIQE4doAYIBJe(V@PII-3d84$^X=QWU$NPLRi%_TJ39m5(W6Jd zj>Gw_sJ9QZf1TQIBy*EFTkV+FzhXpas+1BT1kP3+^D497%C9|tE+o#^V;RRwnObKS z+H-d0=|{)?*I)<$R=bfW=Nld#Xg<1Kvsa6k-)q-91{Y{*Ub!FNw4{K3^lZD4x&7b6 z!Fy8*`1nr4d&icGjbk@Pd@P&s)+&yd)g<;s+_z*J|8iY!9nV(-S6lbe4S9M!@bra4 zO2e!w*T=1;r5_7B?df@8Q!QKj%1Bj8iEfNMyV$U&Js%t{IGic|?|*#8={j&dB*u_Q z3ABQl6&$pR`v(=@xxZxo225zB6xP~TuDy$O&vh4gZ%@9wCN4HTtIn5^{z+K+<5i)|@jLG<3nTE0 z#~=%H&if)xKw;7jPV@YFWLa%^@79+aBDWTX&n|YX$lP66l!4FAFDJ*1orKPkp5FD2 zVI0ZP^Keh_e|@ya(M*4%=kG8&oDYSwIgLvSRUf#x?$BCc zjis>~kq7#2&9)n8YD=RMCv(f~rDfX>{P^h=r&j~FmQ_iI&8oN&-N@Bg0of`&*zPOQ)3N18QCR|j!*m%ez-VJP2 zj5t=lR)bkF}cfOI_Hu(bI^L4Y7>=c0DOZtTtF(lPXP)4nd|L1HFr^`@phM z>@6CSn4HmFM(Pq!G`DbXU*qSB-fh{9o<(DkQcyPw5*^8pr0noMl48Wz2A7M|Eysuu zf~?Fm-fR%c6en*M^rIvA#4wD7Pw5gdcm@{>w>L+E8wnAFwxpC0f=Li8;U(UINA;73I2nY99I6kZ9mQCyFSo$HaNMKO>j zgoe~rK`2B=R^|LA*OjoeR7U@k>tJ%8s~i|ZAZCG*0*9Ew+5$tjL*_sw3_fJWlu=Tl z74*YE$_ZJ`k-9-DLw3FdT#15iD(u{$U|6c+s83$P7)jBQ{SH+x5VvUA}fi=iR5B=j)hE=s(cpcGD<&=j?3#Um9c1% zh{>_-dV=?$W)K4jO->n+Gh+^fq>xHtZ3$BoBusW3?Hy8C%X+(IwHZptldNbPg*AqL zh~!*E;;OQ=ZOt#9U9#GZsIyD%p6sJ^35beOme9^TB2`ryI;psjosTF1wb3LpqYt>D zL!?N^k#6*ChBcesV{DDm#clfV(6VSW)<}jJS?Cfb`tGe67wewKXKRuhxVvb1=eS0x zIX`%^=F=A)KYlWDvNXJNvZNaWS6#>EdH^BW4w231OJ63J=LbTej3f!gHYlzyI$9Mt zUMLQiEmu3wwhs&;V{FT6+jFt%`QXEvhx-+a#)ID>6srs zUtLwIWL$juY|Rfp+tL{LcMr-WS&Z>FzW+RI75i16Z(85~ z{`Y?!hx4}H*0-)(do_2Kis!0!t4II z*g4O`lR34~ynE74A2vN#o#V3?Bgiu1pK>O~L}LWYD)atf#r{n5KYp?0@s+2xmVy6L#sI{xsz18(ho^G@F2B0gs97kFm7KG6HjeD?ZlyR9^hf%6c{ zp0-T*Vx}O$ZYZH8U&WceSoiF=l3TA4fN$#O=N-@2k^4uBuN;fpYbzQP=?2fd{rp4j zU}kyz!g0PHxqsAReyJI9d#|Q7nd4ceGMf9xEz71bSG%F3OXUvHQqU*?!gIK<`O`-W z{^O&guRZ3ks>&#g68KQu*DuyXNsGSTaj{!-yquwwq_LV=Q_)rxt#Nd|gzqS=IZ%e2 z1M8t@v)Qt!6*5a6pIvda>Dl(4J9~!Xy@rHD;5nKLe)7!o_=O{2k)}eM8Yv{R$q747 zEZ`WKLOXhx@{GLqq!`e)Vw|QBQkGNbObK7nT4U?tX35c$NyIp^>m0Mn5R&4-tvRRH zJyI!7_6$e!z`74)8H%$uffb(pJ;|arBr1OVbj5%D`*YsCwPf^(XIGA&o?dgl_FQ#| zt(O?1Q4(qmB1e1(q$FuHxZrUvfrMcwAyrZq}A%sHflF*H+ zX)1%!nsw*!-XU;|Ng$;rq#7UkGQyrgYK596rjin5p~%9JV;QMeN|IAzyIn(`0=Hr) z;d_*Dt0WCtTXM)GsmQ89OrdJIzW@u9%1DkQIgXS5n85cSGD_Eln;8;Kt&zEi7%54l z7A$Ngj9b zr7C0Yq%W)i%6Pp%p-h7x*5nWnsyxG`5cJLwfmrI4oO20Zn-C?=U2WEKxa;k|#zaVC z6h1{lD2N0hEm=)rYSKb1*AdZ8?&G4Mi|mUt+WT@&awZFl(FUU_*NB#=4D&`aZ!Ah{ zhM38HLAvOW*lgBhO!@v1q%jDor+YLq`iP4ezuA_1Q(L5JNLi4A2c-xuLG+}YAc53P z5!S;<_8WvSCDcxtg8z^bAtPoAUJGeN#EdYu1bW(%DIox>GOa3Ni;t16ABcVcrBP;v zOd~1fg6f!p>rx1)2r5fvvz8o2l+1YNkV4?SLn@6Bij;dspSalcc;^v1BUA}!tBvC1 zU`fntE;a*NuDNxzPmF%z*;*_S7Y1?^xRfDVWFD!MEC|3D5YAESg!75f!;9;I%4%W) zYoRrglZE2J@$BUkUpIPAS3NO04qQQrBose9b8O=k?_bWSl%TPhy{ZJionCq*nf=*} z#s)4oJ!h*eV@Sv(AWMX@Xk*YyV(SKN$~8RtOmKluo_Abs6Lr%vgtCr{LZXym8v?&r zMUIz-gQY@@fxYFNKe)aBM*9{eCGX$a&|oBX8!cUIe+re$u~krd|g>B`NQ`Qc<;pV{CeQU zdZcral%dvwz1DDhzv1pdjhq57&BT))LZs!*C+%E_r9mgfP44J#Rpksnc)I4(bI^1fl3#_eJZFYPBTS270az1iBNM7{UT=t1SesscDvE6?gU|AKaRKrQ7oCdgO0^cEyubVAW^#=M}TI zbgpAg>;|&i;az4GMtsaDDX28u-;>&uJro6-GSdo9necKqOZM@(1z$w!AA ze!~=9HH9kp(#W;m3y#72lRx_^*l7Y_oR}j5g3w%oWlrX&de0M%_cD>`fw=3Q~ zc|9}Bg~W@EV~o6((R;Y4*k28tZg$MR5UM7nWL6tGAGzK+j+d_r-@Om)H=2jbhP`>k z)9aqLHte?*E=6Lz+1w>UOgwrpj#%M2Y9*C1XkAh1C67;E@NDDh!;X{VLlAQy%k_FkLSx&ONJmOLkg)`SXibPkC`2Wc$deeJXsiZl7<~zq(#ArH7_HI9(zz+v zDS4Du4ImwOL(stMqdOIA&|-z7hp~#QJw=r z-UNMFs@WW)YeLHGb~|DiQAr?0B#f?zM_OZ53Hj=LDl_R4WRg?R&4emIf)a?Fz*q=O zeAtAPRdEQC%GhyoNF_i9f)7OBv&w>YHpfr60ig`nP7-4vMUTh>S_*_VXrkCvCoLi+1RlHt5?%;4 zT}K*SITn=)-T_hA!n#I6Vc7}^Bytu=UCw{XnUp=CQFt_-O4anC1lFZgxY8*)VwOl# z8q_&BQbn(OAXcM`TV&DX5HLUR#>FVmCN8eIN>hU}PM6`f)?uSSnjh zyy$YRjXvPVkuVO(=uo!8wELuxOR!@CnM1)u+=TK?A$?LQV#q*g=1?C{TKWZRF{^8= z%(!vOc69-HK*aLZaV$T3jviwbwUuZN;oIkP3u)nx7*&4^j>503|&Kj#l0&iIB~pSSh4{>|5~Md1C5zxaz+e*ag0 z^;dZB`OClj%U{RKzpc0RFIx8w8oqPS6VjHSJiX%X(VWA1^OYa;o6hm$CztHinGa8D zzH@)^rGIn29{F!Sx#TA=Jp1#O?>%f@8A$>-oY&ktob&u@$K|@GF`19i)e467xdobPDtivRMR<8PFvTNz2E1pOG^*o}TuN6VV_e9bUA`Y|wj`Ss!^ zu>FLAFGcFcYP#Oj zy8wK?5%V$g)8}j6JJS5|2gg9>lhYmNo4{rqNulh>XI5}$Q8p{@9k+aVXZE${;_K@6 zehu-Gzy0|&$~I_iI6jzj=ScG8V#nDgVoeEP*>*kMCg9?lJBJIT%)E10@$sD*wGrIe zulP^z-R16n#oz-^E(ZSLsU%>zb70tO4M&TLXIC9RdA4OWNRE#8ShTM$q$>j_doxBK z2r(4qqLlQ*fDZ{oz|?JFG)|05r8H4OMq#Bw;HYa!j*j`Vc^RrVt6RL=aMeeau8iDi zEzw%B9TS}k%xlA9QI09b$mzv3Kl*ILXX{9%4fk#>FnQ#BJ@WL*(K*3lzR#Um%jMQ{ zxeFLmle5QX!+=EChHlsqTtrF-+9It%Dm^hy11YCT$PlO@VpXKI3MD*UKOh90ZW5A2 zuQNvXtSU!s6?gXyzql9~MhailkE9ULvl^o+Hrrj{LiYgx{R){2BCiJHn*ob0z`1e^6WF=&m(fk|O3W58O2HpJ+OF)|J%JSr+j$f>yOgdoLGt_@Y( z(LxoD@aQ8_35?ZLMv{CPo1ebf^q(rC7Xnc)Avi?xC<%xf1Vn#ToUKtHLMS0#NFs{h zJ=RnRTjRz^2o52^7(HoVP!d5R4QqlDsAdlr6A6XY8X+>nFp$VN0x%G#35uM2!C@#>M(%Sa z3rQBHxML>-1fsBdGll=0Cy`r-GPw|iVesIC$KWx#ocA0?q6nlQ3X?cEPzpQ)4( zM~X=-4 zDj`jkq93*>pi&k$;X_g?*cGBDCWlDHd50{V=PatyYY+;lD+>3ygd!FzrZ28yA&Cj1 z_hea+OIcP(Aq$(9DcH`;%4be2lLKNxN{gHkS|Wvn97v)(ORKser5#-uNm5SNWT{14 zb9vpd83HkuaKlBNkqUPGjvNzN1(MK3a88MwJRVipz#*Y^aUTaC2*HtsDz4ucfb#wb z5?Sg(k`Se_CuEt7Fj7)!$RXfEK$?aWT?u$ps@(f{$cgnX&=2Pm8YAGPLDvhac7alo zkOW#eLiA*zG1_93Bx%by42(X~j}B{VhHe8nkadL{OOwkOjh%R_S|(<-DI$4QnuVKw zBnC%}B`nSFM#iLZQNrZ#ln6u$7*RfVjY$|?9MdR8ilw0?b6{Ruyz80Qmis4r5kptxVx0>HJS(eiOa6z zcv+!LLzb3*KI<4nq^cT>`F#BnL%?}Qtst8Uhh`kNoL#Q)-Vvf;Gr+vAkVZ0QLnRDb zKl1EqpievQJX!O-2X7wSdoZtgcDd#3x_|j+XV)V?dA{L%?OBfr7s`UqYQeIW9M27p z?#^F}>3WkYuGxJuSA25r_@|#=a{F+`UTa>i#Y%wU$l0n#%fxqYTmI<%1NPe2uVpvA z=gDQy`Sr+l2qmOY!m`nvEGr%!&uHGz#PT{(SctN=zO}5#Qt|w9$E}05G;?25F=bA# zcGNm?XU`IxeHnK8+f--kk;^V}YyZ`Zjt~={o^JVU9a*#$d$TW3B<{}}*4r)XZsa=; zPuL8JA3a}_{J@_)+JEcZd|Pkp8`N(kWbM!Y{Lf!;^uDdP^?zD@=iah7Z1k4TRvphT zJN8;jRcQ+8bz}@Xa*90MSKMD3e*fJsC$Nqo@pnI4@yQF%ox=r(i`OrRloZ@Qno}Fa z`D({*jQocW4&FE&cG(ziEe%grJ*(Zw{#Sf}zOGI$JC0`sfAH%Xd-^}E{^#Rses<}Z z)fMl&CW%uBxVztQ-6t-)OjFgIu198PTRy(O|H{8rvM?M0c7x|)-I4r&P&MzK%->j0 zJX$tvde3Jo&(F_0KD_hhDWS)Unq^b2hZ}*{>cZ6Xp=^L&S1Dy~@6}W)bFt~@on?Pf zl@aDfd?ADaF1J0;E;{Zk1s~mR`1tOZ8y#OPA2J~pA?xev-CIkBG4W*Ocy_Vo z;V~Ctez~xp*A@HoiWk>g&d)p0n$BgCsyJLOFh-M*tjEOTi}HLOqa?|SkRpLrKsS(bMr(sL#hE2Y$jO=C_dR0_RI?VT zEeCs=nbmB@$Z}SLl+?Cn2${=Wz*d4R6_t{Vri2(NYZ!)+ei%SkMLZKAg+(briY4$( zN`sJj3Pm$0)sVd-=fp5NL>`JbYb%T`*PIdpl^s4rLQEwvFc@+4`o)ts0_5e|hyC`swj7Ayieql5sZ$bOobkP?I`fpu1svmj)NmeBV* z!nkAHBxds^rcz*LWIu8m5~Fh$O66p0#bT7(p2zbmX}Wy_edDAN>KByt)M zDHcBN#G#g=1ffX{Rw%|X5WFl%fGo-}iiDtWcW*>+aw7N$1acUVMCxWAr7LvkN)sYP zf(IhVoXFlIltGXSQbE-uHx$8GD`Ip+pXhUkm|U*YX(Ojh^o38IbDA0sHzO~F>nl}3 zOcX|NnMOdW;+U2K(#^9JlvWTESrojM(w3NZ$TX6(WOaQ>^ewBYA?$pl+wDNW-NPk_ z#Py~l`o5gk96&Tk3AV2B3S5$8S>sberGQqlgggpKmYOUKLP>HCNvL@NAtq*x$pGIGi{1 zz9UQvn4A;Nc|!CwrZmfjB$zcdn{mr}7fB+pzg)6sGRsDyiR{g5?i{rE1Xm-RZ6km8 z(+l=n&Ap>H8{^YTpp>F>85a`&<7Zd=!_O{xx{4$rSvD2ztl)Q?PyBr4_{CMv>Duw= zcFXTR*n7Fo{id~O41fCZ3Cm{1pmTv+%bK?O zeEk?x=0~4x`1Ewg={BhIZ5`>r#*|Xn{e0sX$(VaPu?kv9bUdJd&r3JG}aj>iq zV&wE{M^hP=t)(*hbL^qb_^T~coHwK=p)_iih<$Lcezw+ze*4uhpU#{PX!28#K{nx*V zmwj7r>tC#dfRFAjI9^ozyh=~83#ws(AJi_M=iIOEe{S1dsY5QC*mj1w>(*S z7PIDS#oaH~(XwGT2A-}wpFH34-EZ)r_R-x1XKT-==UZB-u?zg{S$ZAU`%JlxZKe0Rq8-dVo9AC(l`+N-`5!}wR# z4R0)piuZ3B)+4O@k>&KhHdgcCc!6DYYz9v^47iZ#edeaV_FbJg~CtS>gRDywg zj0|JKrw*AtNo1~8YqFZ0n4$us(NY$-Y%JnnmJN$pix7!!yCp_P76w(dq~syji0r3( zQ8H^=q&Bz!Rz{59alPq_<4|fUNkJe4kJggt5-}CT0eKSnkf7@#;ATjouvv4uCiwu; zqLc=$O5j<}D3NH|3V~+VZSb*(g;wbzN=XAkW36B}1hP^jnMq2}+QKM>viUa(HeM2v zB*lPL5nUOQ77SxVi$v2XVip97&^h)!5lISx(T$`Cwko)ZAPlIMkidr$0%?pY48>en zu_=NS5ve3ObYvw+qCrTBtQw@27^{&Y6H=Ox9+u?$;;43O4)zZ5KC$(#@ZJfeP@KJp zavhQ~Vq)Q@6cEBvM95eK#pp?)Bclr9BYTJeMW9|Bk(?(ETgV-Udoz48_^eBtXXwdp z2csa#3Z-@7sLCQ*Yg-q=F%@J+j8mXq48>h66;epDGzhHL4l5FswHTu@YK9v-a!M>~fs2vdZX`&JQY})MA`baD zJ)cXs8ab1FhmSpi;;6ona8vTQDS_67zdMDhmFH|EWQTV>QY)HviLr*kM~1$*VU@9H zk*Up;*cV(8Q%8rYEiq+s4n<5CnMGUE40Af=q1$%@f?R_h(ZFcL#R34_mx z+DP0O2svUz!YYGW2@SXa03ZNKL_t(EnbqKVesPTwFh&sCkdt5-97ZJe=NcCZk|rkD z^&>K8M(2pGrwbmX6h=U$1KMPK)iAn*(J-cxd$89Cju)2YFHJga2hXbWxFk8+pR?^f zy^nNbWZiq7uN|#b+&^x(e`3+obl$~vO$q~d_cf>2j@~(js~)W*`->WpBfX0SZ6Y&9 z3T9SPTg%QDesXjrm@NuJ5*4G59LzPfRpgM+FNIY>+cYfZlFv39{{H8e?9VIOubHCy zVrd}=G4tPkbjA;!uKC4PJQ{ckV5|bu-~@y{&(1^&1g*-`3muZK;E~<&Pg7GWf{l)=dfMuxK0J zxqZO0wRBxa*L7d|+7J_uUvyk{neRRPW>DO_hckZo_>w0V9lvw`t3#;1s!o;_zjJrS z7`FWQvlZ{%UNV0Tav|r;i_O64<(5Y$hTng<_r`AF-?lov9(l2g+}@kL78!6;Lcn_` zb5h>&Y~}duD)Z!G;CQinDe&f;*$$qI^}wzl`QXSfuMLl{JoDTNv*>kb!6n943l+UcGMG1HBS z!TBN%O@lx&W|B~7QxP4Alt_Mn(1Do|r6i>cDI+C>kO;xyr{Jr~WZJf63>oJ~Mi;qK z87ayb@S?V?U0}8DIjA$gcdugBEI=BbT@5^bv7?h2B@|X0X3Hg&F>KZ=hTt)xY_J6a z7Zj$d=?0JM`ofjX5Tan{I+E`pCsHm!NwhPD5Ey)-8%A=9#JnTsB15GV=tRU)#Ng3I z*1Mjh8+5fG1j;yhF5y!-m62{C1wQ2x@K!a1E|pPhQIH&2S+rRq;y_NG7z*20$xIfC zs%|m60hc2rF|nK_I!Uxr1P#dtlIzP!RSI+kQWe)=%q7fkT$ey0T>`pku(8vzcXE z@nT(?aku6T`-_V6>yhhj#K__THMKwlK~fphw)CUJM~TD~G(%2kUBc{A1|MT_u_{dp zk>vWq;MFy%stdY7ThJP5B(qAf*$p`F$vJbloDuy%@PQW>XN(>OABiZ0(uJprKqy09 zHzZ%s5yCEsF%iSAxTcvx`BESUPtFmAM92hEpoBrVfiP?#Cx&3y`9NFO_)xF}LRy?J z*HH8W!cSafAi551Yg9ESdI$a{C=MW&uUAzOM^zcGPd-BQNCMVWgrG_gSPEo{z{zxd zlZz}Wl`5|248iv#*HekWt)o4B_EcKq0w`1XwFk3?^XnahFGvchEFpw~v`7U{Ae1e* zfs{%kVB$M71;6DOK_*C%=p3T1k=BBryxGwe7qM;1wVT4kv)7ZGHx|4_9A9$oz1xy} zS9sq_vEFVlIJDI~xO)pFCA%^6|FicdJCbDgx##apX6EKYghym#MrKvk#2U!Hytk(N1N)n`<)`9?SS`hj+$);MOSXHbcr8_w8^ zV-ZzZV`fn$PWb`3$jk^23wLw#W8eAyzppAaDMc<88>Cls&e1O-LKh@yD7DH-=+q2o z5QPgqmnA98wX6&F(0%?gK6agsdP51PSMS zP6nr#AHzqKQglAi56@>QVRuqck4mZ#S#4W3UQm=3BV&j$;(Ids!+RPl&?+0AHztOv zED#CK7fWRFRAoVuicv8}nSw>*kVRtKEWtZ=D#7jDlHE^OzbrQmms^J_Mw~1C6VBkt2kOFqb72DSFaA|pXzTwq7yC3^~KCGA$ zo7Qo@?&&+r(kae1f!Ckg`&fDHhc&AUUVm|Uubiraf* z_NEnYKUxrCLsh6RoC>`?A8~Nnadg@6+MOMC$A>`t@c5AQd;{qT~m4+xR@ha)4H*OJ$6RJ?I}%K4_}ty9Nr z`o&nUl~MFQ&|3e*1~0&?hf^*$j_y&*j~*;|`EZBB`G{2)Il0_$=U{ft0*f(nw(9WS zb8DjT-tn*RH>|qEM?exVf5Cig6*iY#ax(5v$g- z?lM{Iu#ziV!Q;ie{yA|nNaeA@jXxTHWmLL>|GxgZHxoRAnE>IbDoi!d5>ihR1-i(_;L5_fw>amB zNOV!5s|uwgUDpt7M+i0>^P;D_>~qnYGJrsp6HHlRH%p`lC~ZhG*LDf1(7GTdPxOI4 zb||S(CKtK-V6(XzBO(PbniL`~_PFRF^oZyXp+nVUgesWIz_xAhN#{bNqR8oDArV3$ zW5D&z2dvtw_Ooa+ms*%?>_!Wr%bO4(lA_HQ9#dxCkL^L^gte5K;F^Kun+xcKPy^z? zL2$&_BZMA`=JFVO3^qRxLK6BdLPnx4D0cRMi~umo(*;4(1^NMnVuQ@ZhRGmB{=1SY zr_p0Trhu+SxYi>?o>GX2p+r$4ki-Nc^bj0M%ADp#4yjU^dDL)i_$F^jL}WRlP>N#s z+({yrlF4kxV{|T&lBA5&&}A;!^F1Wb?kJ~4$CV+(oZ>4>$y|HZouhLdxu7qnwo}Sn z=PSg8E^35Sc~e9P3Y~xUv@R)>X4^Sb%q&&!9f>?G()&oa^=$gcq{^Gd7{jz0GpS1^ zBSW-3)7qeop|z3Ab%*UbViJfLP7Y9i4;0I7YD|+ zr;v$JsVKE#+ceCl`kMJ#YKf5o=M!~R()E^;rRBMOe{BMPSk?!QF1DPldRATF-suLB zJmo38a8UE32bUb~j@X-iIaVF+};~MHYt|U$EP||f_G1r984rH9XvS;_~%Jgp(wP#x^O)~+Pgp>66NO< zmiv&W06tn+H?A~obA@jB+4KmW>6w1k^rafS-~ayi`TgJjecpfn{ZIXxKm5Z#1mJJ} z=5P4cx4!jd9?moU6cu9PM-NxLd)9Kc39P!57cYQOf=MCR8!MilmwfH{*?&q)>rY!v z@9BJ^{(`^DsMNHsr|ARxPq8^J1pK$VGj_+Cy+><~SApev!|_EP8PrNp8^QB4$(`Mr zuRb^Z#N@&+lUnQX5zJF0n1v8Lw>RQoR?>8q!?EY$?2*m zru2o26{HZvc-30}9}B83fkUr|76L0R%2ac)mWw_%*POd z<^2n=L7=s!EG55kYs?$BCrn0~b>lc*H2mN}$F@D^S6|uX=2IjR?(WyzyRbYs-}2g> z-6t|bdyO*0OP|`^XE!DC4A5ym5QN(XwORcp4k+?c8xQ1kVj7+RnJyB%CEyPHa1{Wl4(-CsC57Ao) zJt_qzA)y!}r6virmc=l# zp5s-|rq9XUBqUNwLKJyiPl!BjjR2X)cZCn&Z8k=wA_*~!efyk3t16N#iP5p%tZ{8e zL{iozr4j6n^4M#oq11}qvEkv_nt#2w;BYpjzWw5?S5vfNtw+{AEO0nIxteTFC)g~7Q4P&^{DCAYClgLh@ z2oWm+sS9YJC`ufPm>>#}$jskNabSa%Byvd7h0G33LK8y|)>2JMLIlTPxKPl!7SS~b zJ;FJM*H_6(jZ_*TJTiJJlUg zp+}{N4GLT!3I(#r{S6@z(FpSEFVGo1>p~t(4?S%oD2fu{bAebavuPY+j(tlZ7?lMl zi&e&ph&&?yaMANcR0_h?<4IzSIBOAVFw7(neJ=Qm zAe12_nV-`baW0^Xgd56&oOg7c#Sj12^%g0>H5QVmo=qu@L5jrAbcWW3i)~N9LQ9zh@} z4I)2AaF!54-ZW8~7yu#AGNB{{GJ+zy4kaQ*Q4H7gh!!GaBM2z60h?k!-~u$!wF7V< zk`XHsX+{}mV6jd%v6n-$hb23?w z&0z)-IU#>8ClSE6gxF=KvQRl4?;1qPMP-siwA<_JP|_t*k8&x=@lJ5~EV0lV)7y&xW@5gc$Ol z5Hqhi_7qsA^^|HbWt4!{1zbo>D?_0p-sf^kmvYaW78zyv@aUAu+BDE(&@dd47L{^O57FC!}0tX***#HKdQ5J&%fW_aBmTL^|>9Y&oBg~5Zu|T zdFN=wo2P-Vzxd>9_d+Y~?2mc4+Vb|%hP_w!p88mYfdBr&4(rbI!{dgS;LhIodL8>w zb2r4qJI6~jftUBo&+V!BdDZ^3U^Z&G-1gks&0PNwQ!c~&oDG0&YZ(=S>FDwMQ6=-F z$<-v=Gd|FmW%Z13o-5+>jZ@oIRVC#r(LgPeZW+FA&f?$CBOca zJ@zIWetfd!eBm}0&C55&*NKYV zNiNJfUNrphWD6|#&9Cl1agz7By_&mwEsM70{`s2c_dlQAyJ{?_%Pn6!EO_lHSj1nJ zS{pzJIu|)xw=B1w-UWQhH?dj@c52OhTre4F?w_nUUVF~lL{%7yvSubn7`?_Wa7rb6z~CIbH|uov!gAQfP%s*(4s7hUv%?8pmeaGO0(`H+|c-%gIxx z*|vR7%}PlKo?BB%ZQ$OeW6^6&RWY58alx{0tQ&%hehoB*cWN zG-n&f!3c)9?X`SJgp{y8aPQ=j2d681@K~4dcx()Gv7qQ8i`KE}6WSWlD&WjZK(Bpj0-W*R935Eu*qv z?F`aL_ELcqaMn0Hnl2VxZhPi!%bnet_fHnAMZjpo{=6n8SgabBjYVjU77|h*x-EE* ztj6r`-?&DijK?Dy>qsKtqClveeiR5o9LDp7LK}ms3@N5O{v8}rCeY@ZyO|5A!VoLX zX+w3DkkgtJ6d^(K+0gRVqKkrr%tp8tv>{=;7FA__Zc2iXGAkRYh%#Gq#+2;uk8u|( zLX;S-C=4tcPo^Miq&8>;K6EgYQwb7Mm!LGU+aO|29t)`nX*gJN>M3c1kOe6XWmrS0 z6;czU%cXT8V*5TD;I5@K5;H34Ln6dTZv*G6hUKOwdeBwILRjbV9-;)THNhuRiu7Fv z(Nk3u3If&*ru$%?UGW31*nlxal6OF6pdtlA4kmj_d2Bv{a}gVR`sj1g8i7(JLgjR} zkDeekA_i>l5YeGcPWmaShzJ@Nhyj`qaMtH^YRc*16g&iXU0jE6Nutapb_kMhK^T%U z_@oeFFgj(0&^6I@xpdAsl8;#FDT?|!1+J@EPAS_Kq8rl59m#Idra}sj?N|A+h0Nt* zBw~svl^MTAYmCk~m`H`|njuMQaO7*6RFXog9PW;aw%g*ZW4Ye2RXGt|m|W0y+F5K& z_>c`|89VSE&|pNq4uzg1O?DhpX4tB7_{<0%5nMK9r6R?QHo-fDut+6QMMY9MDJoLP z#bhZEVh3b1Kj|`Gj$A_uHe(w|pbSbtSxgb$phZvEE@_1zrGmmpj4?##NJ3B+MJ_Z< zkxEIl(zJHKUm%gXCWUsuL9~F$|1R?NRuG6jkmOKeXGYmzc99r!qFYOmkC8+YnnGVC zkMjnuAfN<{ihN#y#8iQCsq%z~h{z;qg~S5&sA7}y##N>B==rvHSQl^+Y={^G!juR< zU{Dgs=ukL>RtTRm7rPL{bwts0mU>j7Wn|L_tn-8qsRvBdr+MC0lPu z!cdP4V)*P^wvLpu)I7S}u9oME1N&nIv7}Imqotz{dDHvme8Qq}oUZ~d z4*dI&NJ65Fpp**dbLYZzR8STLm#ZxgmyVb=xc$hj`FJRX1EX@8Uy6~sDrj7wcadqO zI6hyqXdGIboS2W9ogITi>yknVlr{(fo2KPps@bg#(~)7*+Pr}rLCHjj9_Ja z=N>o2F4i4yK3Z|K3XDe;w+||gF1FML0)`iF&CwFp+lI5{mezJSA9>^X`Sn!MM>U^R z+&k_#Tlb%8@NRq0ruA4KF6t!VUy8xo_dUP;+rJIK@BGg1@aKR2=e+jXYe*?6iUQ}{_1B)9obc^$ zf17WA``i4{AN>*E`qsCYPJc0Xwtg8EV&c2+T=H*^I$E!Is6 z#ei{Oxt^~YzOrxl?{`1B0OY?~0yti@bY61VL>iY+O6J0`VRXATfyLI*^*uHUIxo4s zKg&0>E-@|*<0}7|Pb;jedG};Z`=}+R#IL+O=XlZcgNGffts|v8nklyy=L0*FQD*6N zp4E1PQJT%RAqIyrid9Oi8^@~kgaD;-Y*T<#kwk$pin=s7-?Cn&=em8 z$YhAR!li`5#k$A&L}5NCxVv1nynni8+lP$f8ISQka@hn@6s-Dy zYa-h|Af#eemRzWVug1BRDnmaUNyMBv z8IQShYs}Sc%Y7vXLZZu>swnV2vRW^Qt{>8lAG|So7wCG2)HQXX2tFr4iC~ODB$*49 zq{0V}5E5xLy#wJLL_<;~;B)a70U>g_@~Y=fifqUWL{7nILy8U?T%N&J8J$rXiR%R3 zCq(cRMM-58mC-a!9w8Uf(1*bJx}`7%B@J2zN-a3wv_zkfQX#a2=m<7rDTGMbn&p#~E*AZVK{>qoaGwBS8K2&5=5MS)TZK_v9~9F@9&sBmapvqqc9`Pw28IUODo#%QE8 z5L-t$AcZDFnq%fX zM*+fvsYz-?2#M>ws~FxBq`WBt5OX0SQh}Xi1^VjAT&|+;%u`;EEoJ z(#ROlMpITLr5@8Ym&DYe=-F-@_uB@g73<9!Lf#0{M6{NSstOwuAr1**ASr=VQ$%#c zK7qFo+I-zfIVCR!^ocN4Hriv%8zcDyBZiLTd!#CnQlVrTQp{r52roh8&sr%ZG5I`+ zkYG}2B!b&B!?Lj)ts>L9#P=GR0_B|+zM5nyEuDS=Ws@ocmKnNzdTI-+-sOG#CX zhzQJefnz|+j2sx11*`2A?>t)vNtG0ZLPY5LmfcEc^ny|dqX`k#ou!o7@Lq3Qnr+0} zmO@KrQFC}OM@hx1+fb-Lp`kV-DiaA&v9*D-ZDiSA=7|R(NFmaN$fz(#DcJUozV8Vk zPzp(36!_j#lqI%@(@mf>J;o?1Bak97A6HD4Th3a`&a9?s0}sxYNFi`3r<^ICCtf6R*qHCjnlt>bv*h$wc)PZsXkc8=CY4(Agt z8_)Ur$qim1U^XgPcY$T&zHp#ue_HaJZ`@#hf5rWap7U+P*&h zM!a@7dFp$NpC?U6hL;X%F1BmlJ6`ef;SN=yD79eIeoE&?N}OD7n3&un^U-p^P3PD% zxfJ)Ap6QvMOkc{o*;MUtEOjGN%p3norXTxDVf)bM@Ky_Og^*3 z;-flSw(L}bo3r94em@jia`#5f-5c4M{pRcYG=1RVWy8BC4YqHwq2QI%;zjnTQ2cMN?@{PW?p-?mY;}8oJzCb}IPV zEyMqKapy4}_D@UBN8USYI9XX%zG8n?abvf*PS66xm{@m~|2SIEbe6I#`TC1D&`M_B z7$2CSS1L@tR=)jc$=jzrvwFifUfQMixpZ;0vE1ApQ$~<>dSn z7X(6TCR)=+O;RAVB#DTNiA4_=eIP~&GO}7N@Xq0a08{32C;*X=Lgcg+5*Kp8mR1@U zE!tZ!C3{t3Z#?37+43(xddSV4F|(QDhwoqT)`K(FJ-mEt#^K(Cz3GT|j?UTGKBEc( zM4<>n+k4;l{ zO07uIQy54Xtak{J>7Kf%@<=#&R9dM-jFu1+!S{&Jqslt-LSsa@E^}j*MC&pivjD7v zlt@y5zItyDTvMrrQY$~a?jomS#Rwm^$P_SAVH8YEKqbNH*3nZ^79&CkbZrZMc+BAy z%V$u?Myv0U1d^(7ArPfTn=#4f61gbckOfSGp_f3UA+edyTa?O1D{0`g`phOy5u^s4 z&(k4-j&~6sLjGQn`NC0F#L(k?VBOl><29UDeM}fVLf0C)4j;SBBgP<%L8yY{o1BW) zB_gz$M=MP(K+_sN#GTE*PcCOdSA^)1N&<$214Qgw)xr7CEV@q(&Qs_kn~!ms9%2gSQ;)%(z%= zS#2_NJ7NgdX0tg7+%_|d%h4FE6`c#ixtp;IN$r7O6KpovL7|EfqyfPIc<=ME2!kLX zq9?d+;P@JplKHxEZTNq{#O-WEsW@P7VxEL@aC26&H*`5%Zr5y@j;bg~u4S=W5>-t- z8s&9Y0_PlEpBdmbBoGQEeCGZ%on==WY_o-*BE{WZQwZ*E4Hm3Gu%e~76?ZM}7J_SW zclRL0o#O7UZ7BuH$@}5_fUK3kTKCNC*>mkuqpxkkne~o--4pWhQc7na2ni(+r9{Z# zZ^Bjx2hqNAlTmHC2Z(*Dd^Fr=PR0L@BO6C0JKi4XO1Zo)V_QH8NcccjAg~Y8P>xnX z0ur@(DmMG*h3cf2^~35fE-&`XhMqE5aHwM40(!rf4l$BA7{~JZEC|ZKkAp^HO_Xb^ zswkBJ+KvUS&l6HwO1^NWPvFFl8PlTqby1^_+1c9~7c_L>^Nfe=AR#5jB~YcB>ZZ5j zcWfvag*7hPzGT5ALxq1#mOG8#L2pY97PRK!yu~TNF}bf$HQy) z`%QU~6a6pu5ubc4rytiE?U1|nSw5}n)7@fJDlbHimQIYFu;ebWV}|48O$RpMPp5t_ zIa=;(9MN+;{B_;v@{`L-HFC`Vb#ODL^Z5{55csPDhzRziqcz-;GPoIl!<2xc2$(G1 za*o|-wp3X2VMVvJJZ^L{<6`-l1Ue^`qrRkw%1 zzbVp>M2M`_>zlTqr?wgFtxx#+bdR@VX^K_-R9@2S8!z>aQ}#%aeYL!y@=i->cPt4pfr)5MF^u4aKgosy&r=>Ddw#%=k7m7Q*bTHO<^y^9ByH1 z&b2b-xI~pH;ZNJg)B1zrEQX|Nrt57EcZ>Yyob7HZZq_DuYnP`pLtKH%(XS6LsK;tt z@oN==;=e|S=pmSAAT67*A`@pGwz?7Yw%m#C2KAMG0?-YjnItcEajHjN>p%A=wzv-=*MFuG0b&wMjMfa37D$V#;W5Mzdzw6CIN~; zz0gmU^dbs?z;}KK4P06Mno-dji;amQUKtax+DgN#baoqd={fjl7VSogxDSa|4=E}! zU(olxc3%37yh6X4yjtAXx%HEag-NjlUJ`XrdyG6Ss~JxaIN{r5DKUOATzmZ{oRp`E`h%`^90g8x3ONWXEW!!^f^@1FUZ? zyEHJu4g+auzsqiy72}R_;3ewgkZS+K~8F7mXgM+lZve<8{YrVOB$ zo3XH-ylM|gQc16NTj5fNJ!XvLlt7jf^NZiTO`k)CTQ+gM4Fg6G;KLY7IOp`CcQPN^ zj?5{Yu?LRba71CdqGppDOK})cUy-z|$Q5FSlK4z~yz^Nqz;lG5GFc%O-**iCmo*l( zV607_j}V?>MC@oPVVlW+pG|w9SZI;Vygvu;QCU-XX%P67M}IZsP#8l4n|m-gQ=D3M z6hr;%i4re=ZOL4Rv-;WOC5*oK`(jsEEz1fnR2DNDhA&GWKFR+M1sd})+jV}~pqjhi z&>!|F#>DR5dL8NRA(WZx#<6S|Q^<_2;J~o~cpDcPi|o2nC$5mF4Wq zsHf_D=VQi)Flq@|641jK<#Boky{J?LWs_qs3%$a~!SZTs3Njx)$HJ1>_P;|2QIe-r z)j?r0agy(?wyqJ;He?^&-<+eoFZYkZmfjx8@zf>sZz`!&5bX$(=XuXyAww)U<`H(% z2(LzA!=^BUn>JP%86Am3zuYOr&ql#C81JE>P0V*kvA=+cFvCc2^bV&2N+qZeDrc4v zhrO>{G#pYSAIZ!hH_-VX-(F7{o1!}xqdeogw~KP1@vrA?uz7~m0hag&+9Di21)*%5 zqQ(MR_1f6RxVew%ip^EzwpoG-iZEUDRAgdhrG!KwasU_vfyFS#mFIpOO>ZcJ!y*7CJn0vSuUhkHoU+X}^v)m&qz6}pQAo0_t=4vo5N z8YN)qK5s+Y=m7?U>nU?er$w8;e+aO|q53L;xN0iRW8p@6dwi6Nf!RqCA<=!9d>5mP zg1#I!Z=GTuU+#9*9+%!u zFQuuS-|(7my<2u&K8lp?>O+^usnzhU~qh0NSh(^HAmM9pPax7Y=7%Jt(~zlIco|5xnm;g$ zZBX5xI1Cp}Tpcg%kRTzpA3n|Q8`A4Hse>%+P-Tp(T*r3)g>H8qA2ajkbTujF-lr>o zQ@$T!57>5;AD9Pb5Xnibz?@nyv&U>P)!l9v!g_dzIx~YcEyRg8Ei4hVKNh!GG{IRC zCm~&zZ!>hx9!c~eP4e}3`SI3iRu|gBjDLXmvCq?7%&`|q-FBw`sqK|~ifDP;@BF}I zu_~-B>1ZRz3$36!U1;4#*3I1%Z#vaTr`t!mIS3yc^V7l|3n;Z1M%7k>M-BkSKQA)> ztn|pKb?w;ISu-_a!?gk|zzZY{+`IkOLIt{AwcNTdGt|#U9Eoe2ZGFYi(7Lo7@Hr@F z+7W$G-eAS^A{7zgtwrnhq;R);b1y|cLgupo z+7r=lD^Y#l-UfG*LY@9{+#14Mk7i?LbQ~oH4-gh=Bl8#=3xO7+f}(ro=c-hCrT5rx z=#b=r9W8p+$Zg6wPBc&ZPy~hqA@UvlKqSU)Uke)j4tc%*jT89C6yDji{Py~)*sSHU zW8Tv#_zWJY$Xs$@1{Sh##?z3Dh@a+mcPFlWZ)|7fb5men zHN?yfa{#)O#@g--^Jvi;&;c9YT%5>h)V!{S!Q9X+&2Q+uyO2V>gkdWcI+V3ITL+7E z1J8Fp{7geKDBpG7A0r#u+xjNLp!3A&sW{te3yi6h(2mT|)9w-S;2TT^lm*S<%2WOj zdik))(zw}zfFl4~D%Djg8#)-;l*FgdZzIYh8Tpm91D(Ajjy#H&(M))U0JCzA{Q7<- zTTOx3VD48!>2>tjVe+3q96pV6=*weVtvvzIwF9FAyPSjJ@L;=LK#pdPym$`X@}mQF zz$AtDnPXRI-D}2I}?ygP8cu6qQewoNu}Xk*=gp}TY70w zjmCf#Hi_!b2!MG0;$s70nnBPB=N_2#87c|sjI0Ru>Fsy~S#uLrEX?_-djI2fiVBla z_B6#+fnXMR2+_(jqg84Nb%%p&bPuw zK0zh6{LrbOxb20qehBN*W?oOxkmbwGFTe^oO$)u|}zS$b7T6sw_4F`p2 zK*1l(&K9+lRCJW%I2wdpTHQq+-(zNb5BH~%r(|wQq5I#Al1FWsnU7}RSvRbw{c0_W zB`m@U6F6aabcw^xp#b4$_H20gWP!@z{#1cm{lsa&eLMdq88V_`;^K@uQ5s1`d2Lr> zF|!y#oyFz=_3c`IJwsNx$>XoFjGxX6M-_V7P$(*5H`3=6hx>Z_3Qm2qq|BbUg7VS? zK{`#L)l(<(muwje;u(iz-nhOPdm!Pc=#l7PEXf=fvh)sB(U=mVkkC z50Xm;0?DVvlNEN~G#L3f1zD;hN-rD6>bv;m(ZjKlYrRmem6(g!$bSI%d_i~B29fF^ ztq+U5cmHmoU6JC{(<6WIMd`>m($WqKzo}!MQ!4>fML!#*Gh=gEwK~QT*c)^8s^Hcz z2WTS7wGXCRwkTTT3tcp_LJAj~FC!K# zXH>jOPZBrO0j3R?9M2!Fy{bdnQd`LRb|H+!KUORXA|CL+>_qzADTb9>Qv@K^T}(N}fRP8hKwCLNQi)|BP56MW)lla;1w$ya8&G zL35LL9(xGtHuIiU!>|LFQp@fg>tAiAx~y@^>}SU0M;_73b|3 zgm5tBL-ziyW|32KZ?Jlw?v!nJ+%9MwGdj`Kba-vmJGhAkkc&y3K^hD(OeDzf?`m$# z(io%`xXXHBRoeP#)ywDSd@*I33j&`82qYKOWy$O(8gZy6^LzC@j3~vpMYh}3ji{fy zPliKz@wg~;6FDBHmo|ECSLGo-Bd+nu!DPSLueSc2CKkR{pY=3x;&DkDWX}kANmp;W zD6RWlaf@&HHy=@PPu0C%pL1U{Iv~9UZrtt|Ni0P@X_mmIHV;b?2gInamsc$JV^VBI z%%6fjM>b6XopJ>v+HKmKq>q04KW7A=f({&3U3=|1x~jY{ormlrUqv16@QKn5N6gl( zX#qTJB!2rL(?=o1_b@6xp15EvrSmwH0KZo@38wpDu| zx$PMI0~`xsV~{>~HTSu<8APO`$y1$BFJmNVGOW}vdmW)VNeJ(cj2O~#NdC{V=OAu{gkKc*(C?!Ui%y4>nRiG~JKzMtl zj8)*;2*spFaDQt!>tRMbVO>T)E+ZZdMdtjY14lMjBGNXs^};ox5sRnPStLG0U*=NM zE3q5Ifj}u4h#H(2JiY6pctJ6v8TjFa_^3sV#dv9QrVD7Lvij3z_WB;lp{p6{Ao_QNN+Po0*j?&xTv2XiH(rJCau%C&-H&NEyLP-a`_Hv2(YAJmx{gJtg zm5y8J4}Cdha;OWl>KS8~w^raRDWm`vazen14=KzeKdbz9DTLsd6H?MCGyuJoIcIGm zl3b)@M-zBTNbW0#SW-VDG3;5;x_C+vgaFuN87=inuA{g1Whc$=T(KYCKrq4d%qMO3 z#k{P>g%iHicr$MlwtUuVNB&M)a#m3~p|(D_UuJ!=j)D^?sVYZPzwr3RrfOG!tZu# zJyDK%x(UV1`4CG*&}u)Lxb}>5UD&iSe^wCYPj_@GoY=Lil(A?XOTk44v0CXU31Frh z{I;8!sv;3m>Q0ncF#t_&oUyt)cE?>A)DT1D;;8lKrS|R_`NUyZ(cB+{4`cA&lQ9#79&6^SxVt) zJ4KWBkeINYC@$fAOp2&LQ<>!Sl{XZU2JbLASF>Ro?PB|8f;Y}{W&vAts`jDR$U^e4 zG_Yt1l|dn$E71p;D5%0)qw%YyP7dHXu!blqX;AolMJo$Kxb{N!r@L_?PIAWGT15kF zKr=JOmY;!9|I1Q2#I@T|wDV+DESe!AVLC=c$Cyn&-+E`eMAE%%r zS&NPq^2`QF#n|9OtEp~XYNrCAFM^?Cw8`nc`!~nF3&#MXIx(mEtx&FYg9{>EG&fSk zkye2CE&x51#)`#am#m6=+ySI(UeZ4%#Yh)732V|NC7$$u^gDT7KsJYR4 z{dS`}`Vw=?@ee zh}lzr%i9}%$-a zr6(e&OZ>)}_vXj?b;iyyi zUpC2-PHpmRXz2E7KJvCiQ2M70NY|h9SP-LDNL3e!0!BC9g&=gfj>@D*>QH65RQkai z5*6K5#)+`$cbpJrGwj@jafe52D&OAt54OIFO4+B+pyXeBP?_yTKInA^%-`9Ne0+!= zVEvg9v$Y<4nz^65kavuy?3o=xC%NYQu<^ObBFMl^gr612CH%*U+n2V+;41f_dX3}v z!ve8g41`6I#0u%93Q^!2 z8br;=2SE@Eg{&=k$oZ*XzSj4WepK+mt=s!Tu6eYaUm*0~Wq_v?Xwm-;A4wDNC1TYS zRae8iPN^XS#lZGnLucPz{VqAoWcvGdQTY60zbWS$xOp@r`oYD*Gek)L05O?2&!Y&) ze#J+U<;Ita6#$WeQ>oacP!I!$ufyFc+AqJa|L+B05a}c)$ijacModo_{p_Vl9)h=W zLhqB4g3T2_QUHV%vI$I+RFBBiHT~%Cz7+^puv3aJ(T+xneh7gv-Od z1dz1LK#}ZaNx|Xtf8(S_6ZG=sJ5;7?DC;H?`LK#zh$XEabjMXK*FA<%0%k{jp}@D0 zBr{x20BX*|P%%1Rq$?_lQmUvCk67V(Uiq&m*DTaSVz6WGW&xwP+%0a{49Jc;*BoMP z;$F`rE5gJE8UGlK0MJK`3h)V)$Yk<~S49_Tt(ZTNenWuNi`MlI@^e2e`gx;cnBxE3 zuTbjnqj%j8P8vkJAAUag^!d5dMNVJ2nP;?+(N`T)R_Yv*D9vKhSsDHx?^{YtSMppw z{gsyKMCuEF4o&h|dD6CRuPf;2Pb<|o2#R>7Z5T8Pw|ZR5Z2SyeT&b$olObL&3nWr1MZ$F!mn!!yOtjW_r{*o=aO>?gunhi`!_>9`Flb)RZ}9UWcQIB=k@T>xc6xq)oX@R zu`0N&ddA<16J4s|l19oquMcv6HU)gW{J|c@GyBkIgmjH4e?i?HJJS``)&{pfj#=nj z+vp(spo4#75kUb0vrqhiPc7^-u6}tWHxwA%LC+&WlZb;R&feR2r<AcdyO- zm(~0?H}OvX`%NOn*1trxmE+zMOB-SU6^`e)@5=w}+e7REpV;+mwoII_H}0k}_>GaSci)a-qxp!bhipE~h-nHs;n9v;GA5)YEfcA-n+F8A;TAKAA(rt>-Nr z*Txa2)5HMduo6+s6;SUWTk#C8L!( znFXzYFcbFRY#8Ehh$nKa2e&2W@r306LumBTyz|8Q)T6)k)A5pneC^@cBa(^2(dYLV zGsf&v^qr*$N}s`&Z

+
+

Value

+ -

Depends on the value of format. The defaults are:

    -

  • For MCMC, a 3-D -draws_array object (iteration x chain x +

    Depends on the value of format. The defaults are:

    - -

    See also

    - -

    CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

    - -

    Examples

    - 
    # \dontrun{
-# logistic regression with intercept alpha and coefficients beta
-fit <- cmdstanr_example("logistic", method = "sample")
-
-# returned as 3-D array (see ?posterior::draws_array)
-draws <- fit$draws()
-dim(draws)
-
    #> [1] 1000    4  105
    str(draws)
-
    #>  'draws_array' num [1:1000, 1:4, 1:105] -66.8 -65.2 -65.1 -64.4 -64.9 ...
-#>  - attr(*, "dimnames")=List of 3
-#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   ..$ variable : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
    
-# can easily convert to other formats (data frame, matrix, list)
-# using the posterior package
-head(posterior::as_draws_matrix(draws))
-
    #> # A draws_matrix: 6 iterations, 1 chains, and 105 variables
-#>     variable
-#> draw lp__ alpha beta[1] beta[2] beta[3] log_lik[1] log_lik[2] log_lik[3]
-#>    1  -67  0.37   -1.25   -0.25    0.80      -0.44      -0.25      -0.31
-#>    2  -65  0.38   -0.56   -0.12    0.97      -0.52      -0.26      -0.34
-#>    3  -65  0.37   -0.95   -0.43    0.66      -0.49      -0.37      -0.52
-#>    4  -64  0.24   -0.48   -0.17    0.67      -0.58      -0.36      -0.48
-#>    5  -65  0.19   -0.50   -0.41    0.47      -0.61      -0.47      -0.77
-#>    6  -67  0.89   -0.80   -0.50    0.78      -0.34      -0.54      -0.41
-#> # ... with 97 more variables
    
-# or can specify 'format' argument to avoid manual conversion
-# matrix format combines all chains
-draws <- fit$draws(format = "matrix")
-head(draws)
-
    #> # A draws_matrix: 6 iterations, 1 chains, and 105 variables
-#>     variable
-#> draw lp__ alpha beta[1] beta[2] beta[3] log_lik[1] log_lik[2] log_lik[3]
-#>    1  -67  0.37   -1.25   -0.25    0.80      -0.44      -0.25      -0.31
-#>    2  -65  0.38   -0.56   -0.12    0.97      -0.52      -0.26      -0.34
-#>    3  -65  0.37   -0.95   -0.43    0.66      -0.49      -0.37      -0.52
-#>    4  -64  0.24   -0.48   -0.17    0.67      -0.58      -0.36      -0.48
-#>    5  -65  0.19   -0.50   -0.41    0.47      -0.61      -0.47      -0.77
-#>    6  -67  0.89   -0.80   -0.50    0.78      -0.34      -0.54      -0.41
-#> # ... with 97 more variables
    
-# can select specific parameters
-fit$draws("alpha")
-
    #> # A draws_array: 1000 iterations, 4 chains, and 1 variables
-#> , , variable = alpha
-#> 
-#>          chain
-#> iteration    1      2    3      4
-#>         1 0.37 -0.074 0.50  0.548
-#>         2 0.38 -0.074 0.46  0.098
-#>         3 0.37  0.016 0.57  0.363
-#>         4 0.24  0.016 0.66  0.981
-#>         5 0.19  0.542 0.22 -0.174
-#> 
-#> # ... with 995 more iterations
    fit$draws("beta")  # selects entire vector beta
-
    #> # A draws_array: 1000 iterations, 4 chains, and 3 variables
-#> , , variable = beta[1]
-#> 
-#>          chain
-#> iteration     1     2     3      4
-#>         1 -1.25 -0.59 -0.76 -0.503
-#>         2 -0.56 -0.59 -0.47 -0.794
-#>         3 -0.95 -0.61 -0.83 -0.589
-#>         4 -0.48 -0.61 -1.14 -1.460
-#>         5 -0.50 -0.81 -0.48  0.074
-#> 
-#> , , variable = beta[2]
-#> 
-#>          chain
-#> iteration     1     2     3        4
-#>         1 -0.25 -0.36  0.14 -0.26824
-#>         2 -0.12 -0.36 -0.02 -0.13334
-#>         3 -0.43 -0.33 -0.34 -0.41513
-#>         4 -0.17 -0.33 -0.14  0.00039
-#>         5 -0.41 -0.10 -0.78 -0.49269
-#> 
-#> , , variable = beta[3]
-#> 
-#>          chain
-#> iteration    1    2     3    4
-#>         1 0.80 0.70 0.067 1.24
-#>         2 0.97 0.70 0.367 0.28
-#>         3 0.66 0.83 0.329 0.99
-#>         4 0.67 0.83 0.288 0.46
-#>         5 0.47 0.53 0.447 0.83
-#> 
-#> # ... with 995 more iterations
    fit$draws(c("alpha", "beta[2]"))
-
    #> # A draws_array: 1000 iterations, 4 chains, and 2 variables
-#> , , variable = alpha
-#> 
-#>          chain
-#> iteration    1      2    3      4
-#>         1 0.37 -0.074 0.50  0.548
-#>         2 0.38 -0.074 0.46  0.098
-#>         3 0.37  0.016 0.57  0.363
-#>         4 0.24  0.016 0.66  0.981
-#>         5 0.19  0.542 0.22 -0.174
-#> 
-#> , , variable = beta[2]
-#> 
-#>          chain
-#> iteration     1     2     3        4
-#>         1 -0.25 -0.36  0.14 -0.26824
-#>         2 -0.12 -0.36 -0.02 -0.13334
-#>         3 -0.43 -0.33 -0.34 -0.41513
-#>         4 -0.17 -0.33 -0.14  0.00039
-#>         5 -0.41 -0.10 -0.78 -0.49269
-#> 
-#> # ... with 995 more iterations
    
-# can be passed directly to bayesplot plotting functions
-bayesplot::color_scheme_set("brightblue")
-bayesplot::mcmc_dens(fit$draws(c("alpha", "beta")))
-
    bayesplot::mcmc_scatter(fit$draws(c("beta[1]", "beta[2]")), alpha = 0.3)
-
    
-
-# example using variational inference
-fit <- cmdstanr_example("logistic", method = "variational")
-head(fit$draws("beta")) # a matrix by default
-
    #> # A draws_matrix: 6 iterations, 1 chains, and 3 variables
-#>     variable
-#> draw beta[1] beta[2] beta[3]
-#>    1   -0.48  -0.083    0.76
-#>    2   -0.42  -0.392    0.61
-#>    3   -0.69  -0.101    0.40
-#>    4   -0.95   0.079    0.11
-#>    5   -0.63   0.063    0.28
-#>    6   -0.50   0.473    0.14
    head(fit$draws("beta", format = "df"))
-
    #> # A draws_df: 6 iterations, 1 chains, and 3 variables
-#>   beta[1] beta[2] beta[3]
-#> 1   -0.48  -0.083    0.76
-#> 2   -0.42  -0.392    0.61
-#> 3   -0.69  -0.101    0.40
-#> 4   -0.95   0.079    0.11
-#> 5   -0.63   0.063    0.28
-#> 6   -0.50   0.473    0.14
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
    # }
+format. See $mle() to extract them as a numeric vector.
    • +
+ -
+
+

Examples

+ 
# \dontrun{
+# logistic regression with intercept alpha and coefficients beta
+fit <- cmdstanr_example("logistic", method = "sample")
+
+# returned as 3-D array (see ?posterior::draws_array)
+draws <- fit$draws()
+dim(draws)
+#> [1] 1000    4  105
+str(draws)
+#>  'draws_array' num [1:1000, 1:4, 1:105] -66.2 -64.2 -65.5 -66.1 -64.4 ...
+#>  - attr(*, "dimnames")=List of 3
+#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   ..$ variable : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+
+# can easily convert to other formats (data frame, matrix, list)
+# using the posterior package
+head(posterior::as_draws_matrix(draws))
+#> # A draws_matrix: 6 iterations, 1 chains, and 105 variables
+#>     variable
+#> draw lp__ alpha beta[1] beta[2] beta[3] log_lik[1] log_lik[2] log_lik[3]
+#>    1  -66 0.098   -0.85   -0.59    0.74      -0.65      -0.30      -0.84
+#>    2  -64 0.291   -0.57   -0.31    0.52      -0.55      -0.45      -0.58
+#>    3  -65 0.185   -0.32   -0.28    0.82      -0.65      -0.32      -0.62
+#>    4  -66 0.223   -0.38   -0.14    0.97      -0.62      -0.25      -0.44
+#>    5  -64 0.313   -0.65   -0.14    0.83      -0.53      -0.28      -0.37
+#>    6  -64 0.338   -0.74   -0.20    0.80      -0.52      -0.30      -0.39
+#> # ... with 97 more variables
+
+# or can specify 'format' argument to avoid manual conversion
+# matrix format combines all chains
+draws <- fit$draws(format = "matrix")
+head(draws)
+#> # A draws_matrix: 6 iterations, 1 chains, and 105 variables
+#>     variable
+#> draw lp__ alpha beta[1] beta[2] beta[3] log_lik[1] log_lik[2] log_lik[3]
+#>    1  -66 0.098   -0.85   -0.59    0.74      -0.65      -0.30      -0.84
+#>    2  -64 0.291   -0.57   -0.31    0.52      -0.55      -0.45      -0.58
+#>    3  -65 0.185   -0.32   -0.28    0.82      -0.65      -0.32      -0.62
+#>    4  -66 0.223   -0.38   -0.14    0.97      -0.62      -0.25      -0.44
+#>    5  -64 0.313   -0.65   -0.14    0.83      -0.53      -0.28      -0.37
+#>    6  -64 0.338   -0.74   -0.20    0.80      -0.52      -0.30      -0.39
+#> # ... with 97 more variables
+
+# can select specific parameters
+fit$draws("alpha")
+#> # A draws_array: 1000 iterations, 4 chains, and 1 variables
+#> , , variable = alpha
+#> 
+#>          chain
+#> iteration     1     2    3     4
+#>         1 0.098 0.043 0.30 0.119
+#>         2 0.291 0.133 0.26 0.609
+#>         3 0.185 0.612 0.11 0.483
+#>         4 0.223 0.561 0.43 0.842
+#>         5 0.313 0.463 0.21 0.044
+#> 
+#> # ... with 995 more iterations
+fit$draws("beta")  # selects entire vector beta
+#> # A draws_array: 1000 iterations, 4 chains, and 3 variables
+#> , , variable = beta[1]
+#> 
+#>          chain
+#> iteration     1     2     3     4
+#>         1 -0.85 -1.03 -1.02 -0.77
+#>         2 -0.57 -0.76 -0.92 -0.44
+#>         3 -0.32 -0.65 -0.94 -0.31
+#>         4 -0.38 -0.95 -0.54 -0.49
+#>         5 -0.65 -0.91 -0.28 -0.75
+#> 
+#> , , variable = beta[2]
+#> 
+#>          chain
+#> iteration     1      2      3      4
+#>         1 -0.59 -0.120 -0.206  0.053
+#>         2 -0.31 -0.066 -0.353 -0.424
+#>         3 -0.28 -0.552  0.039 -0.239
+#>         4 -0.14 -0.550 -0.715 -0.513
+#>         5 -0.14 -0.153  0.147 -0.444
+#> 
+#> , , variable = beta[3]
+#> 
+#>          chain
+#> iteration    1    2    3    4
+#>         1 0.74 0.61 0.61 0.58
+#>         2 0.52 0.57 0.52 0.83
+#>         3 0.82 0.72 0.68 1.21
+#>         4 0.97 0.59 0.74 1.03
+#>         5 0.83 0.57 0.46 0.52
+#> 
+#> # ... with 995 more iterations
+fit$draws(c("alpha", "beta[2]"))
+#> # A draws_array: 1000 iterations, 4 chains, and 2 variables
+#> , , variable = alpha
+#> 
+#>          chain
+#> iteration     1     2    3     4
+#>         1 0.098 0.043 0.30 0.119
+#>         2 0.291 0.133 0.26 0.609
+#>         3 0.185 0.612 0.11 0.483
+#>         4 0.223 0.561 0.43 0.842
+#>         5 0.313 0.463 0.21 0.044
+#> 
+#> , , variable = beta[2]
+#> 
+#>          chain
+#> iteration     1      2      3      4
+#>         1 -0.59 -0.120 -0.206  0.053
+#>         2 -0.31 -0.066 -0.353 -0.424
+#>         3 -0.28 -0.552  0.039 -0.239
+#>         4 -0.14 -0.550 -0.715 -0.513
+#>         5 -0.14 -0.153  0.147 -0.444
+#> 
+#> # ... with 995 more iterations
+
+# can be passed directly to bayesplot plotting functions
+bayesplot::color_scheme_set("brightblue")
+bayesplot::mcmc_dens(fit$draws(c("alpha", "beta")))
+
+bayesplot::mcmc_scatter(fit$draws(c("beta[1]", "beta[2]")), alpha = 0.3)
+
+
+
+# example using variational inference
+fit <- cmdstanr_example("logistic", method = "variational")
+head(fit$draws("beta")) # a matrix by default
+#> # A draws_matrix: 6 iterations, 1 chains, and 3 variables
+#>     variable
+#> draw beta[1] beta[2] beta[3]
+#>    1   -0.66  0.3552    1.03
+#>    2   -0.39 -0.0109    0.32
+#>    3   -0.65 -0.8123    0.59
+#>    4   -0.48  0.0005    0.86
+#>    5   -0.60 -0.4373    0.48
+#>    6   -1.04 -0.0197    1.13
+head(fit$draws("beta", format = "df"))
+#> # A draws_df: 6 iterations, 1 chains, and 3 variables
+#>   beta[1] beta[2] beta[3]
+#> 1   -0.66  0.3552    1.03
+#> 2   -0.39 -0.0109    0.32
+#> 3   -0.65 -0.8123    0.59
+#> 4   -0.48  0.0005    0.86
+#> 5   -0.60 -0.4373    0.48
+#> 6   -1.04 -0.0197    1.13
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+# }
+
+
+

+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-grad_log_prob.html b/docs/reference/fit-method-grad_log_prob.html new file mode 100644 index 000000000..9b8626521 --- /dev/null +++ b/docs/reference/fit-method-grad_log_prob.html @@ -0,0 +1,173 @@ + +Calculate the log-probability and the gradient w.r.t. each input for a +given vector of unconstrained parameters — fit-method-grad_log_prob • cmdstanr + + +
+
+ + + +
+
+
+

Calculate the log-probability and the gradient w.r.t. each input for a +given vector of unconstrained parameters

+ Source: R/fit.R + +
+ +
+

The $grad_log_prob() method provides access to the +Stan model's log_prob function and its derivative

+
+ +
+ 
grad_log_prob(unconstrained_variables, jacobian_adjustment = TRUE)
+
+ +
+

Arguments

+
unconstrained_variables
+

(numeric) A vector of unconstrained parameters +to be passed to grad_log_prob.

+ + +
jacobian_adjustment
+

(bool) Whether to include the log-density +adjustments from un/constraining variables.

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+fit_mcmc$grad_log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2))
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+# }
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-gradients.html b/docs/reference/fit-method-gradients.html index 06ad91dfa..e022df320 100644 --- a/docs/reference/fit-method-gradients.html +++ b/docs/reference/fit-method-gradients.html @@ -1,76 +1,13 @@ - - - - - - - -Extract gradients after diagnostic mode — fit-method-gradients • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract gradients after diagnostic mode — fit-method-gradients • cmdstanr - - + + - - -
-
- -
- -
+

Extract gradients after diagnostic mode

- Source: R/fit.R + Source: R/fit.R
@@ -181,54 +109,58 @@

Extract gradients after diagnostic mode

parameters.

- 
gradients()
- - -

Value

- -

A list of lists. See Examples.

-

See also

- -

CmdStanDiagnose

+
+ 
gradients()
+
-

Examples

- 
# \dontrun{
-test <- cmdstanr_example("logistic", method = "diagnose")
+    

+    
Value

+    
 
-# retrieve the gradients
-test$gradients()
-
#>   param_idx      value     model finite_diff        error
-#> 1         0  0.3291950   4.51253     4.51253 -3.28079e-08
-#> 2         1 -0.0351071 -14.28740   -14.28740  3.60592e-09
-#> 3         2  1.7559700 -31.76980   -31.76980 -3.22111e-08
-#> 4         3  1.7963400 -12.81930   -12.81930 -2.77577e-08
# }
+
A list of lists. See Examples.

+    

+    

+    
See also

+    
CmdStanDiagnose

+    

 
-
+
+

Examples

+ 
# \dontrun{
+test <- cmdstanr_example("logistic", method = "diagnose")
+
+# retrieve the gradients
+test$gradients()
+#>   param_idx     value     model finite_diff        error
+#> 1         0  0.788454  -2.90866    -2.90866  4.89447e-09
+#> 2         1  0.538042 -17.17950   -17.17950  2.76466e-08
+#> 3         2 -1.725840  16.99630    16.99630  8.87535e-10
+#> 4         3  1.854370  -7.63726    -7.63726 -8.95156e-09
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-hessian.html b/docs/reference/fit-method-hessian.html new file mode 100644 index 000000000..d8b35a7ca --- /dev/null +++ b/docs/reference/fit-method-hessian.html @@ -0,0 +1,172 @@ + +Calculate the log-probability , the gradient w.r.t. each input, and the hessian +for a given vector of unconstrained parameters — fit-method-hessian • cmdstanr + + +
+
+ + + +
+
+
+

Calculate the log-probability , the gradient w.r.t. each input, and the hessian +for a given vector of unconstrained parameters

+ Source: R/fit.R + +
+ +
+

The $hessian() method provides access to the +Stan model's log_prob, its derivative, and its hessian

+
+ +
+ 
hessian(unconstrained_variables, jacobian_adjustment = TRUE)
+
+ +
+

Arguments

+
unconstrained_variables
+

(numeric) A vector of unconstrained parameters +to be passed to hessian.

+ + +
jacobian_adjustment
+

(bool) Whether to include the log-density +adjustments from un/constraining variables.

+ +
+ + +
+

Examples

+ 
# \dontrun{
+# fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+# fit_mcmc$init_model_methods(hessian = TRUE)
+# fit_mcmc$hessian(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2))
+# }
+
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-init.html b/docs/reference/fit-method-init.html index 2183783ff..e66e48080 100644 --- a/docs/reference/fit-method-init.html +++ b/docs/reference/fit-method-init.html @@ -1,80 +1,17 @@ - - - - - - - -Extract user-specified initial values — fit-method-init • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract user-specified initial values — fit-method-init • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Extract user-specified initial values

- Source: R/fit.R + Source: R/fit.R
@@ -189,74 +117,80 @@

Extract user-specified initial values

this in the future.

- 
init()
- - -

Value

- -

A list of lists. See Examples.

-

See also

+
+ 
init()
+
-

CmdStanMCMC, CmdStanMLE, CmdStanVB

+
+

Value

+ -

Examples

- 
# \dontrun{
-init_fun <- function() list(alpha = rnorm(1), beta = rnorm(3))
-fit <- cmdstanr_example("logistic", init = init_fun, chains = 2)
-str(fit$init())
-
#> List of 2
-#>  $ :List of 2
-#>   ..$ alpha: num 0.0834
-#>   ..$ beta : num [1:3] 0.8592 0.0505 2.3089
-#>  $ :List of 2
-#>   ..$ alpha: num -0.859
-#>   ..$ beta : num [1:3] 1.54 -1.61 -1.08

-# partial inits (only specifying for a subset of parameters)
-init_list <- list(
-  list(mu = 10, tau = 2),
-  list(mu = -10, tau = 1)
-)
-fit <- cmdstanr_example("schools_ncp", init = init_list, chains = 2, adapt_delta = 0.9)
-
#> Init values were only set for a subset of parameters. 
-#> Missing init values for the following parameters:
-#>  - chain 1: theta_raw
-#>  - chain 2: theta_raw

-# only user-specified inits returned
-str(fit$init())
-
#> List of 2
-#>  $ :List of 2
-#>   ..$ mu : int 10
-#>   ..$ tau: int 2
-#>  $ :List of 2
-#>   ..$ mu : int -10
-#>   ..$ tau: int 1
# }
+
A list of lists. See Examples.

+    

+    

+    
See also

+    
CmdStanMCMC, CmdStanMLE, CmdStanVB

+    

 
-
+
+

Examples

+ 
# \dontrun{
+init_fun <- function() list(alpha = rnorm(1), beta = rnorm(3))
+fit <- cmdstanr_example("logistic", init = init_fun, chains = 2)
+str(fit$init())
+#> List of 2
+#>  $ :List of 2
+#>   ..$ alpha: num -1.25
+#>   ..$ beta : num [1:3] -2.2653 0.0495 0.8638
+#>  $ :List of 2
+#>   ..$ alpha: num -0.172
+#>   ..$ beta : num [1:3] 0.437 0.285 1.047
+
+# partial inits (only specifying for a subset of parameters)
+init_list <- list(
+  list(mu = 10, tau = 2),
+  list(mu = -10, tau = 1)
+)
+fit <- cmdstanr_example("schools_ncp", init = init_list, chains = 2, adapt_delta = 0.9)
+#> Init values were only set for a subset of parameters. 
+#> Missing init values for the following parameters:
+#>  - chain 1: theta_raw
+#>  - chain 2: theta_raw
+
+# only user-specified inits returned
+str(fit$init())
+#> List of 2
+#>  $ :List of 2
+#>   ..$ mu : int 10
+#>   ..$ tau: int 2
+#>  $ :List of 2
+#>   ..$ mu : int -10
+#>   ..$ tau: int 1
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-init_model_methods.html b/docs/reference/fit-method-init_model_methods.html new file mode 100644 index 000000000..437e01d11 --- /dev/null +++ b/docs/reference/fit-method-init_model_methods.html @@ -0,0 +1,180 @@ + +Compile additional methods for accessing the model log-probability function +and parameter constraining and unconstraining. — fit-method-init_model_methods • cmdstanr + + +
+
+ + + +
+
+
+

Compile additional methods for accessing the model log-probability function +and parameter constraining and unconstraining.

+ Source: R/fit.R + +
+ +
+

The $init_model_methods() method compiles and initializes the +log_prob, grad_log_prob, constrain_variables, unconstrain_variables +and unconstrain_draws functions. These are then available as methods of +the fitted model object. This requires the Rcpp package.

+

Note: there may be many compiler warnings emitted during compilation but +these can be ignored so long as they are warnings and not errors.

+
+ +
+ 
init_model_methods(seed = 0, verbose = FALSE, hessian = FALSE)
+
+ +
+

Arguments

+
seed
+

(integer) The random seed to use when initializing the model.

+ + +
verbose
+

(boolean) Whether to show verbose logging during compilation.

+ + +
hessian
+

(boolean) Whether to expose the (experimental) hessian method.

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+# }
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-inv_metric.html b/docs/reference/fit-method-inv_metric.html index 1b83c5025..cca5c0ce6 100644 --- a/docs/reference/fit-method-inv_metric.html +++ b/docs/reference/fit-method-inv_metric.html @@ -1,75 +1,12 @@ - - - - - - - -Extract inverse metric (mass matrix) after MCMC — fit-method-inv_metric • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract inverse metric (mass matrix) after MCMC — fit-method-inv_metric • cmdstanr - + + - - - -
-
- -
- -
+

Extract inverse metric (mass matrix) after MCMC

- Source: R/fit.R + Source: R/fit.R
@@ -179,128 +107,134 @@

Extract inverse metric (mass matrix) after MCMC

Extract the inverse metric (mass matrix) for each MCMC chain.

- 
inv_metric(matrix = TRUE)
+
+ 
inv_metric(matrix = TRUE)
+
-

Arguments

- - - - - - -
matrix

(logical) If a diagonal metric was used, setting matrix = FALSE returns a list containing just the diagonals of the matrices instead +

+

Arguments

+
matrix
+

(logical) If a diagonal metric was used, setting matrix = FALSE returns a list containing just the diagonals of the matrices instead of the full matrices. Setting matrix = FALSE has no effect for dense -metrics.

+metrics.

-

Value

+
+
+

Value

+ -

A list of length equal to the number of MCMC chains. See the matrix -argument for details.

-

See also

+

A list of length equal to the number of MCMC chains. See the matrix

-

CmdStanMCMC

-

Examples

- 
# \dontrun{
-fit <- cmdstanr_example("logistic")
-fit$inv_metric()
-
#> $`1`
-#>           [,1]    [,2]      [,3]      [,4]
-#> [1,] 0.0481393 0.00000 0.0000000 0.0000000
-#> [2,] 0.0000000 0.05628 0.0000000 0.0000000
-#> [3,] 0.0000000 0.00000 0.0407205 0.0000000
-#> [4,] 0.0000000 0.00000 0.0000000 0.0702359
-#> 
-#> $`2`
-#>           [,1]      [,2]      [,3]      [,4]
-#> [1,] 0.0462024 0.0000000 0.0000000 0.0000000
-#> [2,] 0.0000000 0.0708086 0.0000000 0.0000000
-#> [3,] 0.0000000 0.0000000 0.0509179 0.0000000
-#> [4,] 0.0000000 0.0000000 0.0000000 0.0663432
-#> 
-#> $`3`
-#>           [,1]      [,2]      [,3]      [,4]
-#> [1,] 0.0401421 0.0000000 0.0000000 0.0000000
-#> [2,] 0.0000000 0.0599614 0.0000000 0.0000000
-#> [3,] 0.0000000 0.0000000 0.0409305 0.0000000
-#> [4,] 0.0000000 0.0000000 0.0000000 0.0636495
-#> 
-#> $`4`
-#>          [,1]      [,2]     [,3]      [,4]
-#> [1,] 0.048948 0.0000000 0.000000 0.0000000
-#> [2,] 0.000000 0.0612541 0.000000 0.0000000
-#> [3,] 0.000000 0.0000000 0.053141 0.0000000
-#> [4,] 0.000000 0.0000000 0.000000 0.0843771
-#> 
fit$inv_metric(matrix=FALSE)
-
#> $`1`
-#> [1] 0.0481393 0.0562800 0.0407205 0.0702359
-#> 
-#> $`2`
-#> [1] 0.0462024 0.0708086 0.0509179 0.0663432
-#> 
-#> $`3`
-#> [1] 0.0401421 0.0599614 0.0409305 0.0636495
-#> 
-#> $`4`
-#> [1] 0.0489480 0.0612541 0.0531410 0.0843771
-#> 

-fit <- cmdstanr_example("logistic", metric = "dense_e")
-fit$inv_metric()
-
#> $`1`
-#>             [,1]         [,2]         [,3]        [,4]
-#> [1,]  0.04674080 -0.003735660  0.002672980  0.00238423
-#> [2,] -0.00373566  0.068142800 -0.000262148 -0.01289230
-#> [3,]  0.00267298 -0.000262148  0.046002800 -0.01442680
-#> [4,]  0.00238423 -0.012892300 -0.014426800  0.07403270
-#> 
-#> $`2`
-#>             [,1]        [,2]        [,3]        [,4]
-#> [1,]  0.04818520 -0.00541596  0.00262377  0.00791809
-#> [2,] -0.00541596  0.06098370 -0.00895976 -0.00360963
-#> [3,]  0.00262377 -0.00895976  0.06049940 -0.01483090
-#> [4,]  0.00791809 -0.00360963 -0.01483090  0.06889970
-#> 
-#> $`3`
-#>             [,1]         [,2]         [,3]        [,4]
-#> [1,]  0.04188870 -0.003366470  0.002556630  0.00143941
-#> [2,] -0.00336647  0.072588600 -0.000749238 -0.01014960
-#> [3,]  0.00255663 -0.000749238  0.062633300 -0.01869750
-#> [4,]  0.00143941 -0.010149600 -0.018697500  0.06842820
-#> 
-#> $`4`
-#>              [,1]        [,2]         [,3]        [,4]
-#> [1,]  0.050144300 -0.00512010 -0.000121339  0.00550570
-#> [2,] -0.005120100  0.05424960  0.005528330 -0.00520478
-#> [3,] -0.000121339  0.00552833  0.047886100 -0.01343470
-#> [4,]  0.005505700 -0.00520478 -0.013434700  0.07104140
-#> 
# }
+
argument for details.

+    

+    

+    
See also

+    
CmdStanMCMC

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("logistic")
+fit$inv_metric()
+#> $`1`
+#>           [,1]      [,2]      [,3]      [,4]
+#> [1,] 0.0431002 0.0000000 0.0000000 0.0000000
+#> [2,] 0.0000000 0.0583438 0.0000000 0.0000000
+#> [3,] 0.0000000 0.0000000 0.0531465 0.0000000
+#> [4,] 0.0000000 0.0000000 0.0000000 0.0786797
+#> 
+#> $`2`
+#>           [,1]      [,2]      [,3]      [,4]
+#> [1,] 0.0454991 0.0000000 0.0000000 0.0000000
+#> [2,] 0.0000000 0.0809962 0.0000000 0.0000000
+#> [3,] 0.0000000 0.0000000 0.0496686 0.0000000
+#> [4,] 0.0000000 0.0000000 0.0000000 0.0688616
+#> 
+#> $`3`
+#>           [,1]     [,2]      [,3]      [,4]
+#> [1,] 0.0406769 0.000000 0.0000000 0.0000000
+#> [2,] 0.0000000 0.054686 0.0000000 0.0000000
+#> [3,] 0.0000000 0.000000 0.0544764 0.0000000
+#> [4,] 0.0000000 0.000000 0.0000000 0.0675343
+#> 
+#> $`4`
+#>           [,1]      [,2]      [,3]      [,4]
+#> [1,] 0.0434081 0.0000000 0.0000000 0.0000000
+#> [2,] 0.0000000 0.0669723 0.0000000 0.0000000
+#> [3,] 0.0000000 0.0000000 0.0466104 0.0000000
+#> [4,] 0.0000000 0.0000000 0.0000000 0.0654703
+#> 
+fit$inv_metric(matrix=FALSE)
+#> $`1`
+#> [1] 0.0431002 0.0583438 0.0531465 0.0786797
+#> 
+#> $`2`
+#> [1] 0.0454991 0.0809962 0.0496686 0.0688616
+#> 
+#> $`3`
+#> [1] 0.0406769 0.0546860 0.0544764 0.0675343
+#> 
+#> $`4`
+#> [1] 0.0434081 0.0669723 0.0466104 0.0654703
+#> 
+
+fit <- cmdstanr_example("logistic", metric = "dense_e")
+fit$inv_metric()
+#> $`1`
+#>             [,1]        [,2]        [,3]         [,4]
+#> [1,] 4.31917e-02  0.00404353  0.00504066  9.28594e-06
+#> [2,] 4.04353e-03  0.05741780 -0.00262368 -9.31332e-03
+#> [3,] 5.04066e-03 -0.00262368  0.04816930 -1.40772e-02
+#> [4,] 9.28594e-06 -0.00931332 -0.01407720  7.29028e-02
+#> 
+#> $`2`
+#>             [,1]        [,2]        [,3]        [,4]
+#> [1,]  0.05586860 -0.00758105  0.00327655  0.00186063
+#> [2,] -0.00758105  0.06515060 -0.00821290 -0.00560473
+#> [3,]  0.00327655 -0.00821290  0.05351870 -0.01152510
+#> [4,]  0.00186063 -0.00560473 -0.01152510  0.06850340
+#> 
+#> $`3`
+#>             [,1]        [,2]        [,3]        [,4]
+#> [1,]  0.05023130  0.00119935  0.00439778 -0.00486186
+#> [2,]  0.00119935  0.06410140 -0.00267351 -0.01118110
+#> [3,]  0.00439778 -0.00267351  0.04945990 -0.01316990
+#> [4,] -0.00486186 -0.01118110 -0.01316990  0.07295030
+#> 
+#> $`4`
+#>             [,1]        [,2]        [,3]        [,4]
+#> [1,]  0.05218850 -0.00108995  0.00814572  0.00609612
+#> [2,] -0.00108995  0.06646310 -0.00283644 -0.01047740
+#> [3,]  0.00814572 -0.00283644  0.05347060 -0.01085170
+#> [4,]  0.00609612 -0.01047740 -0.01085170  0.07278000
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-log_prob.html b/docs/reference/fit-method-log_prob.html new file mode 100644 index 000000000..eb45ac2ab --- /dev/null +++ b/docs/reference/fit-method-log_prob.html @@ -0,0 +1,167 @@ + +Calculate the log-probability given a provided vector of unconstrained parameters. — fit-method-log_prob • cmdstanr + + +
+
+ + + +
+
+
+

Calculate the log-probability given a provided vector of unconstrained parameters.

+ Source: R/fit.R + +
+ +
+

The $log_prob() method provides access to the Stan model's log_prob function

+
+ +
+ 
log_prob(unconstrained_variables, jacobian_adjustment = TRUE)
+
+ +
+

Arguments

+
unconstrained_variables
+

(numeric) A vector of unconstrained parameters to be passed to log_prob

+ + +
jacobian_adjustment
+

(bool) Whether to include the log-density adjustments from +un/constraining variables

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+fit_mcmc$log_prob(unconstrained_variables = c(0.5, 1.2, 1.1, 2.2))
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+# }
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-loo.html b/docs/reference/fit-method-loo.html index 86fb957a4..30cfdd3db 100644 --- a/docs/reference/fit-method-loo.html +++ b/docs/reference/fit-method-loo.html @@ -1,79 +1,16 @@ - - - - - - - -Leave-one-out cross-validation (LOO-CV) — fit-method-loo • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Leave-one-out cross-validation (LOO-CV) — fit-method-loo • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Leave-one-out cross-validation (LOO-CV)

- Source: R/fit.R + Source: R/fit.R

The $loo() method computes approximate LOO-CV using the -loo package. This is a simple wrapper around loo::loo.array() +loo package. This is a simple wrapper around loo::loo.array() provided for convenience and requires computing the pointwise log-likelihood in your Stan program. See the loo package -vignettes for details.

+vignettes for details.

- 
loo(variables = "log_lik", r_eff = TRUE, ...)
+
+ 
loo(variables = "log_lik", r_eff = TRUE, moment_match = FALSE, ...)
+
-

Arguments

- - - - - - - - - - - - - - -
variables

(character vector) The name(s) of the variable(s) in the +

+

Arguments

+
variables
+

(character vector) The name(s) of the variable(s) in the Stan program containing the pointwise log-likelihood. The default is to look for "log_lik". This argument is passed to the -$draws() method.

r_eff

(multiple options) How to handle the r_eff argument for loo():

    -

  • TRUE (the default) will automatically call loo::relative_eff.array() -to compute the r_eff argument to pass to loo::loo.array().

    • +$draws() method.

    + + +
    r_eff
    +

    (multiple options) How to handle the r_eff argument for loo():

    • TRUE (the default) will automatically call loo::relative_eff.array() +to compute the r_eff argument to pass to loo::loo.array().

    • FALSE or NULL will avoid computing r_eff (which can sometimes be slow) but will result in a warning from the loo package.

    • If r_eff is anything else, that object will be passed as the r_eff -argument to loo::loo.array().

      • -
...

Other arguments (e.g., cores, save_psis, etc.) passed to -loo::loo.array().

+argument to loo::loo.array().

+ + -

Value

+
moment_match
+

(boolean) Whether to use a moment-matching correction for +for problematic observations.

-

The object returned by loo::loo.array().

-

See also

-

The loo package website with -documentation and -vignettes.

+
...
+

Other arguments (e.g., cores, save_psis, etc.) passed to +loo::loo.array() or loo::loo_moment_match.default() +(if moment_match = TRUE is set).

-

Examples

- 

-# \dontrun{
-# the "logistic" example model has "log_lik" in generated quantities
-fit <- cmdstanr_example("logistic")
-loo_result <- fit$loo(cores = 2)
-print(loo_result)
-
#> 
-#> Computed from 4000 by 100 log-likelihood matrix
-#> 
-#>          Estimate  SE
-#> elpd_loo    -63.6 4.1
-#> p_loo         3.9 0.5
-#> looic       127.2 8.3
-#> ------
-#> Monte Carlo SE of elpd_loo is 0.0.
-#> 
-#> All Pareto k estimates are good (k < 0.5).
-#> See help('pareto-k-diagnostic') for details.
# }
+

+    

+    
Value

+    
 
-
+

The object returned by loo::loo.array().

+
+
+

See also

+

The loo package website with +documentation and +vignettes.

+
+ +
+

Examples

+ 

+# \dontrun{
+# the "logistic" example model has "log_lik" in generated quantities
+fit <- cmdstanr_example("logistic")
+loo_result <- fit$loo(cores = 2)
+print(loo_result)
+#> 
+#> Computed from 4000 by 100 log-likelihood matrix
+#> 
+#>          Estimate  SE
+#> elpd_loo    -63.6 4.1
+#> p_loo         3.9 0.5
+#> looic       127.2 8.3
+#> ------
+#> Monte Carlo SE of elpd_loo is 0.0.
+#> 
+#> All Pareto k estimates are good (k < 0.5).
+#> See help('pareto-k-diagnostic') for details.
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+

- - + + diff --git a/docs/reference/fit-method-lp-1.png b/docs/reference/fit-method-lp-1.png index 6f1042311f3747822df57d6412754812fa258230..ca30aa56176b9b99fc0993043dce36dfb774ee2a 100644 GIT binary patch literal 86194 zcmeEt2Uk;D*Dk#yO}Z3OsRGhFh#)qKbO9-fh;)!nfKZewpi-qOT>^wAp@l>Na|G#z z2%!X!PDp485CWI?c;0*b?q9fHMl$xu&K}uo&OO&N=X&O|lJA?|Vr3FwqM)E)y>r{p zf`WoJjDmvtF$3M%8FbDE{Mmu-@x5DyXUDVGp+P6`?7--M+b)oTLMZV+AIglSZ8r)E zQHnc;`c`3un=_$#Pu$)Sw{h3ri%F-%ahN|Ev_mzB|FBkgALlh~ZDWhtcjIoZua;cC zexD^3p0dd#og^wIssDUPqvtihxWb!vYg-YgZNq^@Ri;HXnn=eP9hFEu)c!-q`Cm1{ ztIjjQdZKsUUjEbG1kpd zqGg^vle!c$E1N8l9<6LQeq-duD^N_2J4w_in-cugzGj+y$$#vsaN=+Le2Rim3?uv^ z%#zQ39Jph62>pG2tH9|xO;>&#>r)cNe~~uR6)y9SvoH0>*#3jSMp=~Lf38TAq;~q} z0s(Wnu7562E@cV3^82)hp^rM6UR0G*n9`e0JC9)oN6SN<8uJp~TPsxb`%lV)b0~3} zVd8j77YUbC-2qby2`U)X3B%Q*S%b{KFBwywVhGd08@YI4PT*L$6r2NoJgKF8hwM#R zc&?@LTDShO!L=HK*wCWonM-~TLTNJ%t`amHWkb=aG&MOadXk}z1HQvj^SjyLhYBSq zayexA-!fu2F&3C}E{xacrZb*R;w4L`h?A3=&U7I(PE=P5w&HXCzQ08sf>NGZn6;Lw zp4yb!Ce-ev8P1l1Msdt9(;9+?Bq+5g{3yDqx-NtXKGhr3bTz>6EN~?w;VVg;$L*}B zIf^hvykoyv&=AEk3;A|c3HNVDc{pz05cAUzZsSwuKhQAu=+w^*cA_gZjH2CA6zh$8 z;8sNB88)NPyDV)(q#u$Y9CI63^%a(cq8x8Nic4x^SZ4gi>Zzt}B3g3Sc! zYX>o)U;pi%{91QvAgwUtRr(Wp_4K^e%}4MM!{wca7;hI@e?g8h-9(1VFo##%a6M{g zn&Nz{D>J6|F}WRXkb)4oe{4*-9p6M+;U*m=dTVu3C7K?-x?(bk?Ui3|5O|^&vorqw z9bTzYRpH@`dPG9HWqdI; z)-`61lQtT;pTnyMrky=^eExx5^E6F^c<1TCS{wTYgUp$cOtbIa9M!*K^8gTe^s1(| zK70&}L5*~|`^!5}5#K^Jo1z7d9IG#>e1s6=;ebEgT>()VdmBX*(DY(|Pn`*VHmX^(J*ZF>V1E-Fi!gpJ!X`oW zGx>3O&D5n>7&j4s?ev{y@wWB`a$y`=Z0A&k)Hr2D#x9tW+=d-z@yE*)W^}W6ThhXW z8ti(4-BbS)E^EAd{s~H?`QcP7R^x@wpz=S{H|zVA8t3uLPiDRhXG%#F!Gyx1-Lwig zWqHys-9TM2lzB;=%=A;kFewkBf1o-l7#iqeM+k!~KDcQ`F;}cZsCs;g9?|_~4|^Y0 zx=gUglwhN;T#~@XbeElyVr_ejU&um4 z-945t6^oDxGB4qQi^OB2uYy7o#Ut#UhBlCVNGb83Z#RA%C(*MBpC}Z560e zgos8UoeU=XU1|?BMB0)F$F6aN_ss%yVC>FA(9e?@!zG{Y63Zsn2&y7vYbf`|e$BGb z@@-`Ay=KI+b$U4esFFI-1v9pAh)ynr_audI0VxUex_6)4Ub_IB*{=C6N9~lBr2*&V z8zQeE*>IU$eKbV=njm z@3UjD@Hwu`ULv#mc$-|O4rLeP&z!ugbpq2vy{i*ErE{wxg-9m!4|_4k&TO3=Gt^V` zW=Dy$-H9oN&5#>!5L1^l!6Mab`p%N6qvT7+=NMZ7{eQAmo!;0BR|uuNoG?Jnjcu~> z5^X*`kDgy8lnUbUTJzOtIcK0xJT?Zjtmhi|+Oc^cFd630oYq=Dv@5Kwns zbMGIcA1th*DxoCMMzd}*cBLK&oLW`f9h2uFs?pR-3TwSnH2#ie`(6cLu;+bIf3YT< zp7o&w`thU6TM(D9@THZTI;{b$`1A#pJnBuu$&`7nK3A!XI~R9s?bOav$fdu2XZ-Sc zbWC;i488h!^%O#PD>8tM&1Pd}3PKm@ zMp`tKd!3j?zJYe;K}=s}88R5*&L*3RKNEVYm8gJelN*+$9s0f{l8e5{@EI)!b=?pgHzV|bW`d2#{$+ksVKV_^qv=Q=tlSDP*%QWx(&g_f$l z95;_n7H;R|&2YB3Ffi>6V3sBcm4*H_1qNb;qKeHfhQQn^%icol&Cwg$^sKKyT*I>8 zd{-9Z4TFO^(&c5GGEuxld9j+7-QeP${}GfXv&=Cw$#;jeuW1W}UGRStaC#r(mG#-Z zd-IJU=!FD_b*F6Ub0TvK{a2GOC?32mri*TXRkH@6Sm3HD?he}E>MA6>KCr;Z_F)&3 zzKz?7=7Ot-%rvVVoN*Q|BJb`g?_(c#y;U zhKamjx75`E(%1rFY+=P@vNin*)uDn5vltxQc`mVW>yNvxbz>Ugg2|Y56T@Ka6KJvG zNGE;U+o$vMYw(<`i}nJ*n(*2+_3Rx#(s?5SlO^aPKFI2e4Z?zs*5 zP+@o;H#IkIHY&fF63X)7tC?STUMkyDq|-VldwFyJcxe)=^U-h4U#P(!PBsnM)0$QJ z0lCNbL@@rsEpYz|kxI~Qz>mWFZWsO6*uhv7^1`?-93m-G>yRbF^lKn0EjabhF_h6$ zz)D%hY$zny$-eScQE}ihNx*yEyaQ?SrDtr>A~zzly9EqYSq0drP5Za1itD7VI?-q| zADFj(r}1YW1nE1pm`&vvITdvWZa6I`9_V#n$6c1HQr@bZKHzpTD zZn)!(ol^#)(J+ciA85syz6!ks$*bl(-hT0NRFbT3;n`WA61~$CK`$Or^Hu#l`LK z0ohg_Jq9(AX}CA~2@?YzCc**GS6KY)R%E%413w z2n9*P5Lt|ki$gwboe6?|&50Zl5Wk-oe8B{vU8=X;0O?HlW-CSX@=`E;LD1W5^k({+ z%OOI?vRGI%e~gm^eF3C3$C8vH{&-6HCV^+wJFsAuva1yF{PgZ*DMkWbvBx$wL1orI zfzBZ`sHeS*Y7Nx(_l_qA3xbk8=PLCVf?i~L6sneEafTWnr#qB=u{#HAQT2gff6e`I zsKC$eHYmfv6}+0*U&Cnk3y5cme;BL^O+M%hlESyp4dAW=@TzPF()bMcS!tPquylw^ z-Xy4F^?h|??O`$dj~~CS?O6G`Dl}UUF0<82b1I9dv*nak$_cf5sz0cjWA3anBU24- zt@(@C*fh|HOmntO_)ETYFU9-a0KW-c*BHaujsH{B4Oaf(C62S^0yW$iLUoIryx!;}d=k<*i$nmd@islbtm;VM7OnQM z;|QZnN*ufQe-_$*FCCOG4^z$>?2Sd8H~$bW>#yw3HwqxS88);6gTYgMk z3DvVRHDSkZ`T_$rC(`MRoLTr*Dji{kP^4`be}8}3PI!XR<(=uD0O@= zhGK!dV~Oe0nHX?J_@X+G!W`uw=i9i#o)dn)gzu)Hfwp!hTMH|agt=_x(%RpKvZ5~~ zeXEYJemct%+1(c9+cEW9wl$~gW2^M$2%s*#A2j9O@(QGzek zC=~cY>P0rK;>P#gST-WRN<(&uUyh}`L3c(9?bXVox?hCpRjxSSd(j0O67EbMJwwm` zJNKJoaqQVCYE-sVK&lfF{0L?1?sbgC0`7ZYEP~6~dX6dVYJ&ovqEvtw7?c1!B!Al5 zO3u&GV`{5iZqJUIz;YW28@FVw@=n#~_3HAB3SS1}T$?EUV5sh!{6nfUw}Rh(`gQ^F z7{k?v?Mx{-OCNO?#ywt%Q~zIs$D&XM9$?K{LFU+uky^ja+|=NDx)AZ2=Nr;)xtteQ zK1C6r6za{0-IepC8OTnU@bMYU$z2wYyhrlkR6eITM0XPVdCK@zaC>_Ei|QWjj7N#% zpcGWcwDBGe+ZO#iBF?VLXRiKleYjO1Y_iAZ!{MxRDm97= z5yZTMo5H_keV)j8#umOZrrS7SltsP;uCg1k5elW!6BmEj*|GD`Q3dPU`^$DQ>!Q2L zW|wCSnN2arHr<2Iy4(a}4wal`XS#6<1uxBH;Rl+>7cQHgZ?QvpQZ9DCFy! znGS^US1}5P!QQhlWs#ElO~9$#Kt%xEsD0^8JB|Wd=lTjA42YRTSF{QGwi*)M!=Qqz z=%xyOpc9j*Qt2hubyjlq-x=CLwV=G(FSLCc4%zjQ-E2HIfd{}D`^Yk}Rc-^g!PP$i zW`=mIx8V0V9KOt&bE+oJ+dFtE4m_>8SC5AZAGnuRtyzKO<*TRlB4dJc6(2m)yuA@) z)mI=M;GopETbQn(fYBJ-Ypa-l4xMsfnaT|1#+#Hz2$g08rG}MMTqjwT;`XZlyc^^@ zrLX7-&oLD~3gZV~GQ7EY0(X!ysJ%snGEaE8`%E}#hEiJvO?2?U&1n{8x+2rmsP84T z)G6!U+7l2Cf<7V=im(c!_+U=~0kySnT67m~TV>Dv@IPI9P;dc3xO=po;+Krl(nTQ_pz%|Ng=;@B@*LqDB8)0 zM=WfY#TYcX)ep4-81R&(N5?+BI0UwD_pIn|IQb*7*Oj!E6KyAOwCUx|vNvXWA0sJt z=t@hlDmtxtTxInZpHLdZy?Ne#Gy38|1!?#YP3Pu!(<{6TXm%L&+BP>lHMx}qwi=i= zER*9X@&^Pwow#sT#?Mg?)2S9EM0wwKm;cfEkquRU=?H5h0;@W(E4u9&PH66c0-k|= z>vUm-)#tXjEh=O7GA0%F1c|B-eN}tDJ+dOVCe#LLww7A36M&(2K zc4k{;U4WQ@J}g)Aoew+Ly%mb)m+Je5*UK$BpKesWp|2%RiB9Lw)H54f-rPQy#nq*; zj(7sfmUs~QWj=b*>4^T|Fp>RUEciTxml|;%V#{IO;nA+p&!&IEd*zR(yGFc=k?Z&O zP%$sEuM2aeduV9I8l88vKC1yQi&5hlIA(5;#g8l;4z7{+cT!^^mUL0P!FF5FzTqp` zh&h(coEg31YHwqr-+CiXc}#<-IaH#c1NCtCJdRF63mEO=c63plzOSq&^dO%#ktw)G z#X#7GKgafovB<;7{OYHA-EF5Z_+%f}F-9mAeXH~jl0X@H##x29jDQd1X-?u9g4Jld z&rMIdO)8XXhUzbEdnJK$S2DnrETV|!#r_POmm}em8aA`Ih{D@;+ILF&QD~(E!v*D* z&HbNGjA~M5xY`2VqE`vR?Qw+l75%T$rLuonNtgOjMcYmf$vgt-B*LhUpZ^86gUuBv zH=_Nd^j@nA118Q@fFgwCQ{^lW% zd2*O1;n+NgaNfkZ8OIY{(Xh?h*BIH@26S`loQZ4k)!$K^l995d9G3N+8(9kb+T^x- z-S@S5-E1vrt7x^E>a7t?P>2v zT6o}Oz~6win0BH*evb&TwQY4^WyB2z%fp?}~No0$aue41XJkL3=VF?})MZb8>1!3hm> z2x7yfQ_tRQ?mvrrulN_h_|{vi6H|{y*qiKMPJ2x*1>vjcqnnA zM`|k@87g6e^!%awVOt2eYA-zp)TyU3gDr5{DO_k+I@*a^JI+)@cERbZ$A7f|#d?JV z;%J6v4pjb(!}~9lW0GNeKfzpnSXi$v%sV$fSlHNZJ6x?n9{XWp9}Y=&s)2CnM990f z#m+S#Qe$~3wMKOg1TWRRmJmE&Uph(SBAZWSG+n&<*rIdp++m0E$IH>oZeFoJ971Zi zKW_}&t~0WLlOFZ6pG1i94P%|OL!(>M?j_G@WdE_EfSzT7wCX%)OrH08BnFs7;Ttcv z5mj&i@CqRO?2!Vj=q54b7Mt!h%yd0i?(B=;3L?D4@`e%e+q~l#UPuU53z=E1pPL%k zC-;R%Jz^qlh0H1`n4?k2s*F89CuUfkde*~-pA^LVg0b%r4W+}ox6GcC!m~p@i5-Q* zU1N5e5JR-x^pOco3u^BITtR3f zCzw;EsE73tkk!q7?bYy;Zp4PVKAE1`<85;#NeJw%28#y8WKarn3g13HhtntF5Q-T7 z>OaR<#YJWr>`nIz6&kpML-;3Z4%5(R3vo?1 zC>@9zcHRx*`rsbmnVt|d&r*fpDfYt2fY?9%PQa^aqnS!w8n9)v5K0(;e`q0Ub{7z$ zf3E+rIPg?UWMFfEe>TL;Xg|z1v%0(N3^B{@zU7XdlmnauVe%G2n65%!== zgPI8U;nTG0uisJai45?3n?qH8F2jL@s%v4OmutfDrlk{J7nqL)*3tcM zw)7-wFW}1ivV3T4sOD&jU(1F9F@r$_*{F!^W9nyY`}Te0ODSlQ4WW~Dv_MT3AXB7| zbWe3#p)jL80X==^>LpyTxw^l9crU5=Gr6@!)^ zd1NSH>RTFzPL zCiNj&_(I`^S>)4Ce6~6bjHLIh7fFmx! z(yle!D3hLF&B?X%#&pbwxfE6VVbRWypMReI+M6g_a!EQ-&*0d}=LOv#)rMy-Rr>bh zU)f2+w7L|LoMC$S+J$Y_q?8FB1rPh@KrxKK7%bPq? zArTPsb(~~6@k{=S#X%Sw$9wA%*z2QjP#*3LV{!kPg6f>^vVq;qk4(j)Q$7!~9D6?P zv0Vir=44-CLyN=ZXVD&hR}{Xtazeij8uocbcaf(QF5I2e>OiTi2cK}U)(1`v9sVG4 zI#EZmo2dOUtvmHB67JG-a(9S)(Czi6M6o}@xl0<rA-q$qDAU7)J-aA)qpspKfK4$$gWnON zaBkBmYPB=&nf&$b4r5~DkF=nNyD~%xB^zp6r1q@*HK%)rAMQrzb^Q|Z>qwcar(u+c@E?jfUu zMo&be98Vr%8nckuQZ9mJO2@V7T|nQN-5KeC4XWv>J98EX(3+<2 zd^0j-yNwe55j|wn;NzzGU{LXZsB7F7a!l)fNA=D_FgpoI7im(}^bq2z|cE5d*D!L;{wzDC{H0B!cW9wl_C# zU}S^nz6M{~KjTghKi|w`qqTQTSHr;ys4!?r&|I!%!V>tsqbK0G_x7&K{}}iWomZ>2 zQ(`q@dgr`tv?2TrT%)EJt&Qs|m2Tf%rU04&-UqXAgamo*H(V*w@c~_Y#x$72m6^ET z6x!%rv+OhG9BLBpmQMq)RoZW5M^Twko^Yw>6Ye5ET0zeNjM=%%$zO0?(keoyZ#^_~ zaa_W;Xi$fNE5-rn-Q?~zq~~t5D|aUIK43r4bkP{&Vd&<|e4v9pcU9O7`R&$+wWt3g zYuVQA1zi5UWEU39JX|-3O&5_o+a~CS4%aREV6U4)>HKZtj_gGX_5$K!(R-FC5Hcf_ zG6NT&3&sVo96n$v?8D1=?!Ni5^hp_KqUEFkljqKub86S^QSbifyCQc4>Ec^n4ejz; ztG#VnhN20`EK;x6w2(<#i!iI$)hh@qsk$>R?5y2g_>7vWLCP)ltsl1@+x79mf0>e} zU*eNSf=ywx^n0EQe)(?jbVG+1yq2x}d7|B8Cn}Fi!6GBVZN-LGhusBD{UMb8Ix_|} z1u^l`16D9$AP>H0fiP>h@sixMVoO;fHl=@AkV}@mzTSHS{3G#+CI6x_8?EK-SwD-~ z$m8TyGZy^Prb@oI=;W_LTwZSoy->W(-LqB&(4?k80x;XdCp7|XTy6(EVSnOpo% zufY{2+gW0(J$GZTXh_fHj>v26%C^(Rg%c@5?kMI1hJ4?D>tba`Cu5oqhqj+^SEvSv zmUJa)Re*P4-us$6U}R;PoU%@F^_NX@r`K@mlBqj4wB4)!gr%rHXOjyn~zV}K? zV*YMwiI-$g&o`S?RUtk3`uQhu=*rAs6-K%h<+gf+JA|I`8oiJSY06YNEb1JklfmA= z;N0*cT>hvK@t*tpVdI{s^&3AmsK~~R*-w*!$lY4GlDX4!OSmbi$;Fb3;`hAG2@(@~ z|B_GTnhun66hoJeTu=0n|bi{OR0W_Jm5)zHAwWwZCR~Zo$|7YDgNb#)X8; ze5v$AE~}ff9&aw3K;dRm)R^6*5sgsq?dkbR)f%pfVCIgk>hMUGpT*A=tB1LqRfY}f zU-m5ySFq%TmO*j249WO8edHNRye|`)>47)WXL$_QAPh|foIN|+GjL$l7PBTXcB6f z!;8jlUqicIrRu(Os1+`^?U!0v@^9t7v`pn(1w_ZXR-bn4QTFZbCSgL<`Wrr?A24b+ zJ~)LnA;2=}Znwm=$>%vHDpxFNu@3^KVP!hn0|=&4+zKmvPLldpv&?`a{jtOn>I*MN ztOKNf{Co@c?jD~hckZz0(Z=$|AoZ4Q;mT~EtF!DDFlGph6Jec(Js+C+ifo~;M|$}S z7Rl0fq$y0rj~=%hG;h44f@U6h`%xBAu3R%ZY7hn1+J7!GiqJzr)xM@;m02t0FXwE*5rbAT;fA{!b)cf|ZR7ao5=A8J`j=93sdn9c!e-kXTuefwn z=89#4uKWk+>6$u6PVx1Pe}roZnSl- z3aJJQ-b)gdqXp^7)PGLSx@c{$`bjEG$?e)s?!trd%jV$^P8t{5hq;L>9M;v$AKAKc zMr)xxk2OL1yI7=kijBdsOzdtd8mY?R(Um{MxXLqT$De-#7MR4iwY;xRIDfM2BQCBr zw7ESU*+HMZ7?mEN=vw-mWZjE1XY#WAIq4;c7`ci8r(0*?Q0s|&!AMx9$3_6v_C+(5 z`}GbX!ApeHWK25>0MOn3V0v`lP^sr%e0}T30Tj7P=Y)1`zhzG6!OLM?$)4%-DXEn{ z95~&fkg=+CG_=`x{`LtVm#q7W;q})TwGn>lO3xmMMXLf|8ChnFD~Qg?f%kb3>{)Ru9}+)wsF zx_`?{s-O7`bZ0{vj21Ssi-O$S{%I1FaD9%l_41LsFKy)OruE=g2OUS)$sOChpDTnu zMG0Oatguealz+Mse<+U8Fnm2-vsrc1{UvUzFT~lw9<-YA85q%us;ne0vYa8Pb5ePqj&Se<41u%&H384~fb7nH0jAlvWs z73n#ze=9R4OvRjX7My*1CJ*H*a@a0c)BY6pe$%$-Yg(2_I*ZnQFu<|f{J7R;Uzl|> zQ^=r0m6xRQ~N%_|!|W_xQczI-*?x{cP8FPyw=d+tQAG%Ra0xTAlac%AojXhuNhNNTUmU_s+_F zqW)+5(skgyDYUjW|KJ}@q_|bs6qS<#Tr*G#DX*}X1~w#qg#WMM@mi0*l?>nn!P zT^|uM{;$X)+wu1yO=S-EbtZWzY%K`VTf7&XuM-I@^y8Da4#9?3_JSO9lXX|3_#32_ z0P4=Iww;OkaYvnNwXaM!oNoYvkM3eV21S?#IBpO+Qy6|&s+0S@p}_te!7`}Nw{I;# z_E9lul8B|oY|HR5@sY~*qg&lu)|1!_x0T(o&4-!i3G(I8IA^8d7~4x;1`AvZD9Y38 z%DN)G&8Z40f9ui8_j`DHdv60c>o^D$x)j=_pF0H2>D10zlN=v>!LQec;h9ZKH3J#E z=$c*A{QV>!&f1D?0StsSr?YUYjFIRG9T6u#1<}FARV;AJkV^<&QGO5Q9 zh$}Ue&JcxC$8U)p+NZFfsRYG5=%>%R6YShrup2&l2s$L;cWGyA;b#rn47EJ6xl)jV zmUyz#av9*%JaA^H`z1{CkJFwg0R_tITx*%N=D(#ZY4y46l;dI@^va-EXi`mYBFFs5 z(;%B~72pZj%O>m3cOQ#Y6FKKcI%7)A=^~h=WE{=@BD}lNBVj{O;5`U$o!0IzE)I6) z&^y!Ggq$*AP?O9(#|S2IpNW8ZPRk7hwAUMSc|pI|_iL zByM{treC&KynJioYM^TfS=iV((gro$s+nF%gfS;RMMb_kaO_U5Ru&0*+*x5batf6r zeV0fBh&O}=oC3~RNKJ>d&)pXWD}DcP;#QK*Xu6l2N3I z7OgpQ zJ25gFF>liQ!g!5u_civg_m^>a7|QP#G<>`@Vn0#{+@8+NUu%X}P{{}e$bvsKi+IgE zxMAg^#(0vS-r;qS+fN=yLHVBwqp(XsnjHg8mO3W^81x?7fJ72;ZZ3rc;RKRUw6$a- z$hEd_?N@14hw2C70fi{f4#*T=(5q45ued^?o(aq85P?JbYvCfZvXXkD5sc1 zb;L~9<7 z2ZrBA)t#>LVUhV&)WnUz3xQHF;b`Wi7apDFHcO^wa;#iSI^D%yRNv`clE!o%7RXWL ztmAAvxih~XM3(}SDF%Z^if%>TQv!N(bJWjRG)F>2alM--+u1AhSa&U($q>lWS3&Pn z_rBnd8wT3n1Ng{*OU@74VJbz%>d70(Cl_vs`F~$Rqj{6cy~lX2491_l8!MtY&=~)FMoF5O%Zf zj5z_l*%s>u8g}1p@%&4N;=rTzdtVSvi1xR(_376H!)%AtCe%_mT-@*z77;^E@|@m7 z5S*FKpj6L^>8kd<@w3XSEk2ebwNNreK&A`wgPaFJ{3jQONxQsH2#vC z)0w-)^)$+t0r94uB>(IbU+NL2YZUnakJLq7eOX*Wt7||?j7&j0&K$T-7GHlVhno#RMUc|4AgA*WUI_#r$N` zpvw+8;odG{dE&aeEj;jB1(MHrYIJ-`!W!nO&~?QepesG@tDDkT1MU;1&elg97oCu0 z$>3jAOQsbpgq@SJ_ykBijnNuG-lL=iw8keP9^s&x9BfOiX4VtVs>w5YqC$#}ikDd3 ze53X38np0${gHvD(1WYHB9mv`JjBFRW)8`%&n_gKsxV!l{^5}o=o27KVVxSSe6F%~ zIyx#~lSqHk*);8t;*|+QjA5yhj4z(zKg@kH8xX!jN>q_871O~bAMf=mdL_ZOLAbQk z=HTKYp&pZ#)A~w^Be7DboA0Im?s4c>rgab$@$5A8@fEypi<|-m`PXe{{xyO~{u6`# zNoz!su5IvdA@bmhd&#H3Q|s;BVYqcx>==jP>mRZ)#^eCKhu%2-79gPUrn@A76_0*g zQe~c?=M|elacVz(s@K2X7vMr)O7AE9B#rCCjujxXUTWykISwo967bfpe2|TaO~(BD zP5X(o_24_)B{H7Hf+3_>r%X2$p~J&-oLwux8o)damO3Bbz(hvvfgb` z3=X|~_Ll}d=NI?f+DdVsmu=Cjjr(G4ew4N>{>8}qkce8|ky+Q= zjee7N**9~+T@(D{u-?V*@cJW3SF;OcZ-~w_he&4tS3m)0>N}=dt*L|4BN2g&b`>t`@%bJcj)!E zKz2sW=Vpu4fw+RfF_1I6Zbm#OKSj+lK@aIv<54U6w8@Vn^D@?M+7#GaE)DHTaXr>> zTg|MA#a+3P&*RUq^vh0G4FB=&aFysTZgq}&U%Vr+O6rN0j#h5uo@d!esm>2i4$2>8 zk`nv<)b-hhO94*Mw)`gUuOC@p8Tj|}fipWjObwvVE)c94I}RI->_0Sg2a89j1M8cx zPOc&@#LW|?$7ArB)Di@%|5otIh?p+>S|}@uvraF>6;I$pQJYml?L*?KpP6uVonCcl zMUz4q_eJPxKXh|yr^!mq*oVup{UA!_Xmk&RI5}yb_AH!KWWfMc_X&?q?zllq-kZDs zC&_Rx3`W4u);bwGBDIb(!@nVxC_7wA1}z~2&NX0P6Ax4J$!C)IGM#6*!%4Aa;X|Jg zAg!Cp}Mcs&UL+m{%?T~*NaW95#RG8pxvbWu(yTg63#$aBMzGU$8>-#7l)KgM-mIyorU z^bME1nXJ;Uh(C>xvjnZrysmaK24<9l`@j`ZJF5^kL=I8Ohr>El>V;OZqB@%EfE7FM z$>fzj)CzUjesE;77%wT&pj&a!9&@lBJv9Vs-UpfjV)hF9$&CHqYS-gH`UvS#HqV+6 zr`u8QbG~Z8iEEoTH76Cm$FzBvO}rCcyWsb}+Xqgy5Q>S*U-28L4?ZDfIe{W4K}i$I zB^OT;aL7`ZvvTtecQtrz+rB&*v9nPCo9lR4X7Y6szL+#e0@iC(t^Y+Vm@>E=o-rRu z)xMp-f7<*&;*;qE^JXW_6%$ENl9{)^-4VFyzghrO-Vt5%*@?;EJ20Zwhri?KEpYIe zRD=9B+b9T7+syBFI?J=Y_V}o`L#2q`q45g-Gzr*Nr8&0Q0WAEyn0r>p8X*fLsne)q zZao>EVf*>=h#z~l{ium+P#lnO;IMelur2VCTUd5$b}Qh*sNnAPh{XGcPrPh-^ftuH zYv{uCq=&GSSw0s+Dnebd(|5xs4I2Pe7ey`x}&Cbi@%rT)opuVK-%`##w_80@uaW6+#XU zn9Xkq3IuY?C>1D2@JG2hHFokbqDT$8`j+=fgKdJtX zw=B{q6lE;$V6Io@3ENMQI?YEmd~H_>J3iYLIVL%krFdWC{DFLFK7P^g{aZ8DU}`t( z4H|46UHW?{Q@y8BDQDZAJ{Kl{1;wX?E#OLvWKVt98M%+8y@nX$cbcI%jy;;7yqC;a zyMAIi)!wE$SIQE!yL%Th9uQ1&0HF@cU43?}Pqy=;i1s0QT&4T#1)7$N{cS_a+n#O7jhmUY143u{?1?_+B0#ntCi zJ%w`|z^s)yZR1O90~W!{OYrm-;7y%RrfmhY*|mqkd{r!IV~vDP8|c@f-Z{^bt>$Uk zmQ$TjsZocpdtxzu=FzM$&7DAy@QP=@@IdCrW$2j@{vs-Esh3oW{uPlabVxWyoxXm? z<<>viKZ?pxyx?hWlVrx8NkiprY`0YGOkCGxTYrwFht`MRKh5Yn0p1UM$nL`(;RABO zXczl{kEZtxPrZ}wtPmuoEUj{(>YX|ngb5k;!#-T~oFCW;$pxQ>hfKQiK7kBht_iq! zVsf@esC5VEH`Z#|1`y$S&^_bm4@+UtuTQ*3A#9+adTB%gj0xJmfCDmVgM#FEv` z#E(908AbYNv*{D)Ri!7;m0_F03l1;LKGa5%E@O9?<~#6L)0mU}FiROGgpCDvOOYlM zcAy}%fN))6#xyFd#Kh`Sy7j~`_2*A;?riK{&G(1KR*Y2Im$bTflX&45e(<{avafqDNB|Sw6(go8P=&v7zAecNl2&e&d#uvLXmf2s8CkCQmGoz)+?vzmb+vp zhXSPo%+=AvZWO3yo>aO2?~cFn+13xo(z`U+C^wONFC@7PWAdaQ5i3}pj-lF1%%_1~ z0YFM~69$ePCw>0n3oKonVWRl)g~mC{R+|ULQSwOb!u<8#hwk@nbe;B2C$(_?S}8jj zKwk#N7f$83ZS#W;KEx^0-Yl17qAE=)Rp#KnuTQ&XQGUrW0_T{&C(f?V@Hg|Yl~ zaX~>ROm8;4ggB$oxl+iiaCw`oJFLC3wY9nVeRFd*uzR-%S0JR-yRT2bAgf(R6)RNS z@A#%EbgN+RDkdL1oSvlFA?>Vn(xuC4*sFW@e6;8{cOR?*Yka_M*KXg;vqDDNXW{l4}QuQ9lYA^P&k73{JU4ObCdj7ak|oA>Y- z7@09Ds!5-VuT^$rF@H3NG|~`J{n#nu9)E2CPuqWwOoYdUq}{SAvtC-d95wS(lD*=f zfFXGu?av!Gs~kZKOOy-; znSPgal5o%`olJ-bA#;y)hcJG2h@BIoDUp^$Uli(`cF(>$LH|M^=Jk+{`S>zQV25#d z|MJhIy+hBNj^g)Qjoadl3S#L$+Vn1;^ab}i2>aI$fgo~toBHEgi}s0%k6H0YgA8j? zmn9EBd7v{ZH13@8_}^#+&uj->xltP;>bZ1$@PW^UPA(XcUZxQl-d=&2<5<`*{}s-c zV%%1EbHhD|b0@v+9OIK#-;UYMyc;m)(q>Z2&e#^Udd=M)vJML`tDVTCc0 zNrCmQP8j*0mj}4B>i!Q~Z~YMEyLOKvA%iH)AP9)W(4~ZgG>mkOGzdt8v@}DFl!T;! zlz`F<-CaryEe%6TN)9y)&Ed1(?|1Kg-gEwe=ZE`>d#!b?>wXw2 zomTzUE|0)9^~|>(U2M7$OCEeXIIjXpxOrtx=c%DTKo?Ll5 zpUd76)IIaZUGLB*uqM!uYidoIR8snmkcMWlvd<*4`b&&Wo?G_gyEkvica7IXGG93t zR8q+-jGsItQ!GO`BNSJP{9NJ72;Xe{KW*V}U$yh}PJJuhRKw9#2p`40JL-jVk7>FP zeJOT797!mv$;Od8jvGTKKIZJngatI2hkir^oK;|r4mHs+D#l~VV!J1-SF>)h1{>e} zt1esH@@zecK(z7T2?WOrM}_GzpU*ynUT;o?S9<0x`%hwVdDydz(85H#TQNY8Fz)~( zVDX>A`lr5@;g6NAoq`QOs&dG9dQ-ZBWOL58q7;H(I$Os}293(<8Iky+<4ZJfwO*;% zqJ1aa7s_y{p!nyFf)L8=NO79uYaRHwF9>Rqb*tLX3>B77+_ot@8z5?y*+O5k&#C&! z9)C1*9#j0TcCSqR(ydCg{MTX`^knAE;MDSY>61jfJwc{R zzeVe77 z7e%@sZ^?A!)CFe}!Uo3EsRBichTPIH=Z=aWOLeXM2|DFwZvq9+lZ;I08}8H8H|OmX zK_p2vg{Y{D+abt4--XHuz#n7N$M;1V4e}tj645D-D_@@3BnaJ(uA2Z-8*C@hc${UE zy|vf^+bm9Ruiizmch60dBxPGZ#(r;9>(X9>UElQx97#ey+QYpY(W3yOEA(C3b#Q-X z@W%UBTffV`k^5YX_?l(986n)Ix2OVcctD7_CMG{WKAaQlSnte0)>SACsLcRoS4wO9 zDT~s7Nm@^i3NCb;lsYC(Ode?ujBC^anQwpIZ>i;5*|g{n8KzB4j)V4um+y?93+Fkq zh}?#r<7EZ0v$);P_5na|<(_V79?@!@J2zR7=K7<`mAy9*Qy_{}GBYS0zuk)NcYM*b3A*0r(C+>icQ8r1vxDU~^~&?DKtNo4gbDR`8UkOYZ3 zb`Ci(qB(rzWX-lhbV&$IYC$Re^`_DqHyL>&Pgp@JdQ=&8{wz@X@_(1L=|wgkk?zK%S8yO3s5?i`JTs!u#y9dg z2TyNixIW1QEXU~+yKn1Hv>CVQj38jtyIAbzNnK)wdmCBl-Qrt$z9UPacHm`sq$p3Y zKCu8(yF3GD-}u<+&DK1j5#3VxFCd0XQEGoDCfW;Ath~C=7*KRE3UJA9&r4*WjhJs% zU*2L}=hQg8cc2ndlJ>X1kobMREkFM!-5gZ9{$y*DL=&rt*D1{I^G;oQ4@0hBb=w~F z{2jgOY*wxGeq~I%zgD{kBi7{7PEt1Y?eD)#&5tS+?XsM2zn>I6e#SH2B4IqU_B!)j z;p{E1IfpBS`PPq1lLFO3bvbU~eHn68Z_~w=e%micE{;h;jrb+#Wxur=J%$%vEkIVC ztRLND3YSz<^3Ps%#36gHahP6ni)-nJ4wV>a1_4xQ?kYUR3WV?ce>x-sufGe`DeaY? zDWR<6*5t=b@8%32lOfLCwPAmp$~OnhhsxngMP$*&R7&H4-W_$1nRn#%tfh9tj@eZ+ ziZj!dHH(_CzgaQ2bkS=$oDa|{u!%BL;VvIYEL^n_PN=~eU)Q0CY>{a6!N9yS^~jg@ zQ1NUm)UiJSEw~0%0gMO6A?!Rrf-*?Nrc+1_7q8gIyM}4vW)(mU?jl7L{`%GNHYLo} z*e7Ho#As6gTUy4gX8Ptdrjkbtit2^lv(Ilg>Ao_$O}!`!wkw~BaW|sKQGT+Gaj*Kb zy*t0@^E=aM@`5kB?DTJf@vpg9fnzbcs3?loo;ydY$Bugm+qEP)5&J2;NWC-&v5sYP zF36IdeMUWfaWh5~$pvIwLUYZZG8#w~<@IMcEscz~jTY?%b`;0~TLKN9{1+A(r`us1 z*fx!7O&+Ajq*9O+ zex!^HEXp7>V@AD_eN-~TzDNH$RaLw_--(N%Uza7YE%+SvB>dt7ba32_caqtl&6yiA zFL-=C`25)--q+WF!*71ha=+IjM~1jA)pi{uF&W$6L*?yaPLrjequddenda}~8E(Y+gLLR&L($c{f}F#x60~UO?7%5(4U2RJ z;_i|3E6ZTojh|`c)xDNgFBc#Z(?U6A!# zR>Y($A;k)L_?4=~AlbJ7;;6zc>E-l^gwC{mg^8o$MWFanr?tTHPm;l!)xbl^x z(P(|3H1waE=6(Dh@Y8Y91xd*&V3QOD!573$6TVo})UxX(Z|{sagUZ&$Lr%8YL=;N8OF!M2m9Zse42~vaQ|hz zH~@rNEYLw$t1cDTeY4tPW<^gDsfH#SAKRbW`#jRfl#FDr9C5kkdsz8&h6K!f4YN1P zL`r~Z#i~cdjlWDz#!gnwHr_Pp+#u0frOzDM)F9dmnN-s|Fo=|M@2jUfy3ixr^qtf8G!q=Sr!{b{i9%1>yAoH1er6s8w z(d}cy+$QPTl}2e@+LBn+)A?_xzroU+25t77o4j5Q0{mtE2c0ZH!|*Vpw&^jOoB6HQ z0i~nv1qmX2$>(>O_y3}2DewM(N4Rpi{DdqcX1ep*=KCCR2J_{%h<~ z_}R{I#Eg)tOgMr9SX<6k`^n5Q{;gqQWM#s}y&gJ6E0Pjq&h)EcOaj@$$W-E`?bzI` zr%pyb16@aOmFL;SWVVx!1E-L#M1;u0L$eoE7K{1cV@?LkHIiH8x=etRe>V&3ejpRb ze$s2~5~!;*(5y1{FXbt1$nsEH zdfC>_mfLoOIE*NYrEqJk!^h3~SATp7{^vyhhq9^i?8>HgrAX)S>jjH0NHZyUen_Yi z&QXox)=NN-Y@N9{PBr3z8)`^+n6?`SK#=!15O(^|LYR!AA&W+P(y2f7;o@McPEyo7 z^ZHKk4By(#&!I8G6ch#$g0D@m`6Tf8(FaLKCIjknHW60?!gYcoyjug zoJMFDD4!@r*urmHBb(G=bD``!m==w2!u@>h%W9IG%o?<;V#hRBj3d5-e&y{_$I`t2 zaKu1ghhSn-KI&`ymS3wd8c(5_pXA50iXrUKE5zLIyJZy*JZyvS2chVjpXi1?@tLLM z_{|N!kyBzuh*#yxRfFHd^?_CqGeURRCGvRja_Ld=mk;d#R_YECI8IOlA+3B;v zFoIp0V0>45xy0BH{;!C93Ae5ttR~;Q$?SMzvewFL@eNftaOQSXi2l4)(_*)n{7G#mqL zBru2lrZI@w@2jA^>xb0+mYcJHPT~WKr`LW``FwAZ*)+qzcHOOGrx`^hXsgYZPx{~% zmCL&O3q>t2>Ls#=+~$7@`j(T-ccZ1rx~2+ zt*|SUFS5u=xC(aOziGDNJ@Vqu4$%7?c^n^+IK&|sr#i^fq(ZxyR zr|7xpB)GlF6(J+?tP|r68)#-UxA_j!?NCg6AuCwY)@~xL z-9DnpO1C77YpC-aO&`i0Hj-bWiP>sSslTJ|t)|@)Qs6?sYb|{I+SE9aXRP7cQH$N`gYYe)T(J0PgSEE6}YOdn&%` zUupLExDqmV=&3_B{69=-9M43)&1+lpW|iy6-E@1rNjvY&PDgk%&A1J1XS24bBni{_ zb_J0Y#z<#d7VzDR?M>^f)Lk{?H zMP?P$kYwVJ*<|bV>rK*4czlCBp1pN1_7|@49xk_%$M32p0lr#HY&hGoOXBlKPSW8f zI-5S_6WzS7JoUXPiV~o6PWj)rIKMarP^?CQ?8xIPI(Y}l2O&}T)Z@WpZO*2;_)MUm zUNEnlnx`sRtP4?vz8>b*xcsm~2B8FVIlcjF%#2^l2%2Dp2($H$8ukUTQK~7{Qr}zL`SV!lUX*5_Yyr7{Ol-u!WFabk15LD2ypq{ zZSr0?>ds2B@Arf!i0*!jS5vN0lW&ntC+JNJIlDTQQ!K@uVM>}MG(1|!=tn)%AfkJx zwfHr>eZ!fL#f%N|!k<>3nCW(cOLsCjJ#R;HrN0>;T-WMzD)Y3I-Y@NA+}i+?vai|% z>urF*dN}GME)#?bPC-n~evI z$$uYx#4Vqrc}n&)X|4WE568ZKhWRxIbKJME`>NVwr;h!R@Yu(inK>1gWn0pgyKDL& z-(J=Gx}_Xqz`q>-AA~@sAHNZj`UPSUm$%q-Ejl?+j#XZFiUzS?n#+#zt8G`s6!$t; zuCGPb8mCg4t);RlFUkA4CuwIy7K&ibuD6Z_2O}YP$VUnjQ4e&`9^;BJ5+6YcZa9<{ zvVW^=+kHOU9kA%oq7`5wmrO;a1J=_sAMMtc)wYh&TP-#Vl~)1p=e(3WL>U~j(&}Q} zVC6Zr3qh>Vt2<29&VcWd=|vcVlpuOV2_6zIP>mT1ERsn&5hp(rcZH{cx^>npicgq5 zp*9Lg1KMeZ(}&7Oi|fN;_n7D=$Itz}%FtN)f_6f%%%Jvyp`852$WqFLbL`Zb9`50D z<};JPxKW0Xe7#cir+GMiR0-RYhdP+?J+3i9%3lup5Be8dKI_d(Ju*LYpK61TQ5iw! zsyM8~(MP+Xs({^ysKYEAlia61QYIZGQo}ii{&+%-wU0LncE-a1Tb}s*Li+_*WevrZ zNa^LrOm~W&DXZ*V-}pbm-m_K^X@4zHCD^ud*GsP-e}E7`c{~ zpx>Y_jv7DxUy-#uuh-INN~*8nLR)twTr2^Uu3d*b{LWu%Ep4tgN^?_6vs$&u-&uU$ zBKl@U=hKIM9}+z)7#Ua>xc|9Dl2+{Du?AhRxNoNE@a2Bs36=sMcbw!O2=d<$k6l*= zpl-}sZu{y32^rOHf~msOKB7lYYM((`+${!%*%(8RNfev@im=9W@(^+{0F%phalUY9j+ob1KpX&uITxaaT~^2@GM2=6zs z^<&*f%9>Kq7~4Shx3?Q*E>?2%B{A_6OmPgFI5hR6dN3cVbcTZm%aoOsIt*j#ERQCS9G_r=!?kGCDgx{ZriS^!w9N;{M<8gUuV%rE)E`Q4Q#lZw6 zecqnVyhoz;?F)04LwzxNM)Lpb1yE~T2#<85o6b`H$diilwHs=bgjIRF>jo@{c$p(s z!p1jg1>U|ZT$y9SHkPVj&F>9c_I|A1r81i+ zDwJ#jo4;3D(RM+l0Yb-!V^LX>5(&dQ(yWB1ke=UxNZB!5WCM3aM6R^yTZaE#7VnV! z9?%QFkL>zk&HfqTW3_CYbu4`A8=yp`BsZXIoizFT>U)&=donYzgQD5R(m}T8(neCw z)$gu(iGzZA^8PG1WsyZs3v)pIMPpUxl-pWU>XMA!;(-mIa!CFKIFmWg^3Deqd2XIG zV`$$}788Dn^RL$c(B9?N+KC9RM*p^8%U%c@TWtyPyx?>>a!&n%l8jk7srzj~1k&}4 z1Ms5`D9_LqOi-lddNHFN8_G@yJ>K%~x7fuUY%&+H%5w^Y5tDjYk^E>mX(yHM*z)%$ z1<%nzdbSy4jc=4i+k(Yzcc_Z$$%T4GTw#gsQn)ZW57@0^IJBg0)6Q$S%?VDK{p1vO zd^jJ~a!sXGqdwG+u{>@!h9zkEl6dxJ<-R>!lCkK3+z$FDUU0zc-Aygry;1e>sju~M zI9}-tTm-J9bJlxPC(j^NzUJ6RYmK8x_X@OYh%NcI*=5zPfZ4c`_`Y@Y9x%Nh4&Ctu zMP)ZS_gEyXUvry2Sj!*-sB-d*Gz685NUaHJK>mN>}g z3)iLy>NtAn_^QYFxIk@~CWPB|>9agR!y*_K+XEz3h2-JSWAUzn1k+Qu+XxDdFc@a^ z*cB&o@US2GC9_d-K$$>N>z`Qti<|Q96Uz>n0&>_Z2yqn;T_hP8hi_;{@(UUH6yWGhleIXADG7;xfrJ19dh?64_p6 zDqm;oi`c{2$nbZKw|9K*?Ztj}8mvAK-kljOA~8{%%aqNEj5-Xn)!u)C%*9}qg>c)YwPP|vXcUpKre* zz`z&89I9WW zr6W)O`kc;((N7fFm<2M`d3D%sCug*+=Mzu1!p8 z8ew2$$)HM!1g0{yo3Yt!LI)=Lo%O~)3_4@c6xWsPR^>V-W#!o34k`B7bbH?H6RB{o z+La)>DKGD@HvG{cx%9+{3f-f(=uVSyjBr=a`80?)$nVNe^*rE3*6<|J_j=9?l|Oil zPXMrH!lAjsJ}9#XV}h!$=?!wv+`u7dRf=lV56|kxlF4CZnPV608HitHFV7Aq-l=!$ z>aj}Y9&ts^{675R>qCe;jw^F8EUi!X^6)ly%k!yCHC{E{za^Ozk9b4S)kpr$gKPI) ztHFFT{)%@A{FkQqJ)hK$N&_QQo8nBB(wRDggm7*EEfLo@RG}O=Y^HV-6vsi2Cwg5A zSUqsS);Ezrv0y%{ReZ|`%&9GW0nMnh1;gDpEuk0AbKn@|#3=Yh!Gj8c#fd+BO3j=u zUv(wdNhh}WkhUm$@E(Xvdph$RK83_!Wx{vntG*-Xj1_@}Qor9wS?DW{SrNu+d~i-v;YfpX7;8(JMNlSNo0e4wpa(XeLYctP& z5aRCLDYeu%#}+1oM*>vm3O<|%lxtZhS%Ht^lK&73ZW^nct3XVrApgd^)i3H4+IDmd zKsevxV{WOz540&Ro|>1xAv9)%dk-Cka`3?6b*v`z%aTzQGXSuxkC|-$JjLeoK%mjW z#I0^Bt%in>R<@7%j@6pi*eMnG+`L;PXF)JP-hfDi$e`(3ZX1uIHJaUC=At$7w^-xG zyoN!B1kuUU{78Iobd9?5L;B#e@oiYS&h`4?HRB1&n?#4lWO)^y9h#YWlsSMGWUL_X^MUY|fw?A|_`+yy;1ghaJ^BI?n z@`=JsDM`HL)ySESBq_0th+@Y8VlBE`a&YS8s;Br6TQ&I(gFi~0KIZk}XSar9y^{A9 zY_85%YrWcCTFsOYNfWsoj{&NXV}si0yCTUVmpKUd^jBJR5e|brG4QD%+vAb?8!Fp5 ze=h=O+k_duFC!9|O89TVZLsMLsCQAYH!1NV0JR@nG!$UuMiP#wmj3!l?+^V}!;8L5 z#aZDA3yoc`K~q_)rne}%xAENhmY@$9NTA zE!O0oTyYw+ZFQD*@k8(X>q}a&r#h9;0kH&mLHmWOOQ#U-x%eupp31V4q4{&v6P7O| z_8QQv#o@Z7nr(>U=r;!wvP1P@qE!iyeY6R9V5~i`@J*~-xKeHNU(d~$j*0%eELN9k z(a&WEo8Gn5_DVe+4PAL{mhDd5* zcuZlZ#}~dKIYt|fd;6r{^Fabf@i&k5`^U{x)X97EONWUR10sBWq^Fk}rNSi2nDJ%z z_QImuG<{PMsOyFY)#sV=Z{Zv>#oO5ybE_dWaH{Mtlag;juX(iEM0m~S3AM48&yXo4 zTANa_Eo^17;Q{__ncpvr3uI26^7($Ryt%W1xZW*@fz&tcw_XMn5#$pFEni1fke~8D z2fCs2@-=2ZYYA1t3D1rMIK_A^{=vroGvYLO%w;NTPRkMZxP@V$lQ@k~;s_tj82$x= zzbr26j|v6y7n=_hdpT=nekoI?%J_04DK#)ntg-$C&{qplgWXZ~p^Mo{dJR_9r1z~e z>#2__OSphX2TFSFd;sc_2UaX4i?1gFk8zgrH7X0*XI-=H9*lXtd%k-tuH;5cl=ZHQ zJR&yFzlr>YtxP>3KJP$yT+}-S8X(d-<)T=JW(iUBuVpOHx9&%$zc-HMbTt~64Ph@k zg16yR8|yy?s}bf}_%d$qaE?_=vrU;T2U$Q%v)=$iZ1*IkwNVL$=qzX&okF->>xTM_Ta6$;IM)L?Ij>^j$z89L zD3L0!xp1kh7vD?@UMny^*G?{)dD$+$ez>PIQ(eISAmsQ!pIW19^ZMFysvdpML~XMq zMuIr`Hz0%jm53wu^{&MRo5K^_mv^_yb=2<&rE|(F9%Bnh%f>Im6$I9_|GCk3%m|i6 z*L)Z8EmM!Kk%$%+V@VyNt9xj$=1Bs5cvT_WzbO-jA+(w#kI` zeE?aJG|(uk+mhT}+@A5tWo^&5w{>!^e0b2d6THMIb6+OBlZTP9yy-I6ZUsDhV_gSG zyN{z+&>m(QICN34L%-QCde2ufOF?mAHqe-lWgqA1bHcZI2%_*?2HEDYsr8=@w^9qR z^GlnInp+D8G6&e;6Ue1M?-NBez59Rhdt_Dd>pA5=Za!9N@5oG*HkDTZwq$|SrLFz* z2{+`w9^n7KVqi)VvTG|Ztu9NRSZyXV^-dZf9k=vE*_%UG?km+a3`QcS{lOP3TP1Z$ zo}cKObn)FbT`y7WK6qQ)O!Q&pv%=sPXsMD7Z$efh5zalasBLr?Z8(ablsCWVu|ZwH zTf2HThU1UgCO@@1+dIEHV40(3#jR_lT1IS4qAIdNKt70v#&pekJO%B&W+4;D05hX? zF{32b-wk&zr4Z!eL-1x&h*)>V=C$|8-FQ1(8uQAZOzm@M#Y^XVBzYchj%PtT z-MFdRMO9>f6=vsjcTz98c2eM}8|W~Ev{%~6`4W$Dvz}cT9r|ZxU|9*2F87(hZxggZ zh%D`A_K!=sqivn-c=yK|IAtzv#~s@5XVJeFbT_wU>hdzG4{N??k72Kv?(FGzn$ZuS ztoA-PL~ZVgNwbbVuzZG(`%?6ONrt@krgRPl0hCX)R!L+>FY#nK5X^ngkzaPa#_KM$ zI6<$N9nnEWe_72qRam+b!&5(Y9lfC+tga|TkC^AKFoysas5J0-BD{YrZuR?zRVkE_ z!Ql|?#O8vF_D?{_2v}vYz5{?@vAZaWLx6OAK~)jEuJ`081?=GN2x42R#WD^Q!oy~CzUgbUA(%#9EB^Ql#I@53ai2?(uRS*GKpQVP?8d;L zsHLWBJWY{|l-p9o$8wc(y^gkVd&Jv(xZ{N(QOgi7+>#}uE%8WQUJ(j3*BF@lB2Ts} zuoj^6gU5% zmM>mQUY+hNZmoBSucydVS>H{pySz8YXg4;na7YRwGkMPbKvtv7Vhb+{>A)8K8nBX3 zeI3lZw2upu@*a%EADjbG?@Z^qiB|f0zMN9{T(C3$nn~nJ(we|c;&k`SW_r%>@2fO= z*wUW-Y%jJuO;VR)pY{b>S~cR5JNryBx(Wh+kS)11pD`Z)MV)`^<;i4?MBrBag(&b$ z01$+I31K_RTWw!&FV?3zJ|!9y{jYfSenomUjiwMl2%p@$={2~)@LJ8iof*4vin@}( zVKP@QXnE-OV7+6QHr%K3!)8n7X|bz2H^jRnI?i>3v1@`byda;D`&v9o=~x)BRBNFJL|_+xq6ea2OGO;c6fo-$r&|-yiU+9eT}Q zm3@_&YhjQ2erEmkRzwnpQ`%jzX|3Khl|xPKFLhJ`tGd;7;Ad>NSwOxrBBB`?8dCtw zE(LOo_G+Pv_kI-Peqwse@HSshW+YmCWCx6`qU3MXV(?^>`mjkG7F|k&vdZe|B27fp~QBa zn&Y57qOa0lH~tC^f9qK%lr;R${eS#gsq?z}8qq@=6UR!&mS@p4cw_psm5;bzWkKf} z79dCKV|LT5+_sw%!;zZGbwjRVw%vf}c)!$*OvZa}L|1-bxhAIV`!(p?BPLsQe}V-X zdXLPK>T#0g1i_G1mQe=6u&)+h!0cCenQIHUi1ERnM}JX{f-#1iCeM@os)d1;)-;D3YRRGMQ(OI6Bf$&6~V zfx4x8LLYh)JE#KLUrmbDr}yBP(KA3upWCw`ejtP~(X9L6Z#+7#K!>g|lAhzT)OJ2-|6fGg-LvL(>`XwF{BA7Llxy&hOGzO0wfSjW7VO`3bS^j-B}J7lEJ@gz7VR<6?bfW60qJNTrM zOfv&;zC=S5S70$uxPnJ5rPpv;WOIJt-a_dKbu%O}=8MN}8O-=&uLfU+Oj?<4+>!V! zz6)G<-XT5OI2=B@flO`~VQS+!&L3fkGw3dj`%Vx$dgdw>wQY(Qm`eX2nwg^iRO9RwJ$9s}80auS>B&6Y$clPF`_aofa@O|3x%i^wqmJ-6e=WZSUMl z7gJW$Ckt~v67yw`t#YM6+<8}G(_c#jnS6iOxZ_e|UyS{NIVffz?I|I;(H+fv`Toj3 z@o_4Yp}n~!YnLT)vRwJT#B{be^!tX=; z_ce(Xsa_T<7n@&!CKngPVSdPFI74dU-|oLP5gASS=FVP4sK}L0^5g4}MZ= z(#SFEz{!)V2nrAFY*@w<$MbBiV#;hq9$+{8CGqDx%{D%y*npIR9@Aqjy6uEEzZLv@B*o6?*uXi{lpEP)Xdz0 z)YGq|fz-tR4#AaICFc>-txO0IcF*8#3*ovICKAJ294zhueo$#;bn|Ge@>ZbewModC zgZz=i=fe9kfOOnLJIrs20IBz`&!HDpNN3nSbb%8)2`eXcxXssp0c=_GAz-QMx_?& zK)KS1P*!A-4|}5)XPZU4eCki9g}=GQYI!8>pV-X1vLmB<=*JVicASBX9BT#gm@V7_ zCLfbM&{1oTh#s)AJ3`2H`fj(x**BFQ^!BcDCB0=S^YmqxsB zFBz~(eVxH+Co(VHO;^`&W%o))`h`-SccgKkn`UG~W1=x*OU`djE1nkyP^uRGeXOPv zNbEsovm=D!f3?2wFcmo(@x49{JGcPgXiF1AlX8qmP~C&`%1@EXTzGNG&bB_SI+~fD zD?Ss8Q+g=XnNv*hwPVrQj^jvV*Vq-;>WiL8^0<+)ww2T~gY`F<08)JG`}mz=!Dbw5 zS%wRgu}vqO=F08APs&}FUrPU5CcC8_BsNl&FwQD)i;TaYi#@`OjOOSJlKwX+@+qf7%;m*+QD8Re679`yS7Jl&tC>r;Gs=XOX^My7%TwE4j6?M#s%Yp2QHu3mWk+(55G83DH_%@TU8sm8?|NY z9QxXG!XFJ5G;+3v>(dXWmI)J_KWYlvNoeKTUZ81KXhd(VCMI{lGpZi=JUrzNNIT)< z7n;u2=oTjj7sV=MdO%Iv6kGWlsRiC-o(_2TkMy8<$3R^JqHMJ~mHu>3!Zi{W(X^F) zVMfMmwiUr%lkmzqb%lbs&? zn9>fM9r!>`&2e1P4^)Y`M_rv$T_uHJ_XVy;e3CY`{>Qog-`=bI=hnR|Z(CGzFdgQJ zNjVF%Haclhz|X|s+__oXThovz3EE5bY2|KW$go8Gut){#Na@xY*vW)Yp4?a^hpmj$ zPf-7@lxkqHV+*;))$=Ie^bUjFn@1l6!>iQ3-Iuyhr`bDdV9&rloXo6SD0;+Yu2O&d z(fc=oqQfz~cMnC?c`iGINd&Z#5-$o+7<_31WjnOK+LL;!C6&3DP3xvnmyi-?+EMi z25~NblDjSMJh1?E&bifMf&1?1QSl#g>&}!y#NsR4O6G*)dl-V!GG%%?=o0Z_>@RqY zp>rZ68t#|N8?BY5PCgmawoZB)1bd_zEgZnf<4;I$WBh87SL95U5)p{$r7}L`|Mdd2 zOJP)Q%0xJl*0V;zRE;kqdL$cOO+tzOPR}`vc(KRkBx_PH<%qujmam>>IN%OmwSAEr zi|y3Ll;d)NM;3)aWL|Y8?=DIZ&wmBYBi3C8Fs!v*?fm+JmckV%^*FZ*BPwtsX$?&IY$xmai#Y>o2gp^=C7Koa6ifg_sM*H35Z9C>6zx9BV*;qbWnbt>mMUS9)E!SE!GJsxO zjfL9Jvz6;t_Do%8p5faXeI=aGr4bE!?IEr+1TI@?Y^|13kA|RCM&eQN(E(Qb@v&RM z=`?|{R`wvFr0)~McbTg^93ziXVx5Jp(09|$9bQ-U9dF68rD`+TInHp>lfXox8Lz5~ zJtc$Aa~$(@7mj?@(=D~GST7z++^%GQJbqp5>%^X?s-@E5jwZv*AxI>;Q~3Qi@5BC~ za?A70E+m#zt64|Rhe#sWXTvUif;C6F=M)cV8D2hT-;qBHTS$y>SyH%sQ+IC*(zEU& zaEiI!M=zOkuA9`4er=nH41F1QnQx`0hzZC#dNne66KjUT&{B69T>7Z3^QFTDk*v8R=%WE9jU5f#JJbbHBbS(>q~t}ZZD#B2^n&WeVo;0 zzE%tWVxagj^<3LyaEFdr19xf{ddZEUt6AOiys0Hj_1XN^>Geb66D-iG`((Q;8fz+!%jvN!aaaYxj=ZqUNh*Y0udn#A3P@BB?0(lOnxM`$N5aiqmn2P6 zHo#t7`FPn@xq{Q>2&;|00o?Nvhg{Ky#Icj+c@}T(7MIa&7hEz=PT|q*H2dhp?{b-6 z*QuiFS>)ehpjoj@cYPI@JzZZ+!SR+Mcb_#eRJiHHFxcafT(|?2AzDFkpOevz-qOes z@JxQz*BxhL@|(pyRW*N7tFg=Dq73EvEeV?FLKXDAafih{l%V!PIX) zNhF_J$}{kC-4RpjhoVf4^n1my&7569@)XR!SJOIbf-f`|-?})&y6>_?sSYMD`nLB) zge&VsEYpMA3K7fm0H*t`?VHc7z?Ae9bq^J2_BdZgG-o1tuZRSdF%? zU+RO?jI>IfGQBR{f9_oiZ8yBI@;HIVMO3bWf>2zWv0w-4e=U_~bS~eXi-%r4cKU?= z-4MmXyYtVr^FOmRTSAtW7v--!@7=wXZ3pN5xwivBmVM@3+4^~cYBT)7Qh?c|*V^T3Gho(}C%eP=fkq7lV5{ZTiV0%x zX+&IsA=6^2@ecsq{@%__U3CVz1losOEIoM zL_Ki*6fY6aNkYY~dtKuNnq#(-X8b0dFVs16F<IM+1EM|WJMrcPet z!()5Jji+11g;vZO`j!lyG4c+lyO6fwb%pLb|;gO)A2z(yMT3{gA%* zJ0f#^>{NWF)%%eKq3>Tl4Jr0fXEB z0PkOQi9~Zn9Z=hEhiNnoX-A*!nfGqTi3aEY3wE0kM_Km&!86zknZ+u^Ap)ZK=W3&Y1guR|S}N+@E53tJ zrpELFQg)u}tdAVG8_Ltgk(T9Kk7W5!%Ek?%@|jx_VI9f3Q zE;oVbdu@q+#*7)&&M2(I5N1DLJSL3kG3eb;Gq1VX581Oz0IdG*f4IQP=rey*CZzSL zH^sChn$$R1Qa=WvKtqnCZ{-Q9KSTVU0c#+Mp@yjCgJ9dp2O7$xFx9^ znc*-l$T+Q0&H;UaOFqTp8xAulcoxLbKZhZo>iD$%%ij(~C%t-EL@5;EO3?!)N2b_+ zy3}fO&f8nm|Iq3Qc-L)&>>6{9=R0_*+DZqw;k3ppOjCq}49c@oO#Rcaw2ZY1UdX=tn4%H8%!NKmlG*Y05!0qP_0;DoC|bn! zRynb~y*;$>Rr|*{dm1L09PP<-(G~hxt@m+Em^N-8#S&9^+8EDf0F8!ZnNq5>7R%~p zd>aOTvl_sMrB!uF;rTJmasy?^)gmNg9x^nXC}PRV5p2BPOLOyY!UN|^=z8_?NNdBD zcWvixGmacTv6~hL1)Q7eDl2BCpZ-5weFa;TjTWsSNXL-UFf@X6Hw@j<-6bH>4T8kb zsi2f}H`3iA-90ptBi%E1zI)Dh?(^Kg@W$S2uf5jVRB;@-Azv1(HG@2B6e>9c`iW9c zf0*kvrzS&EH)dO;(k6zrT$i~SOpTQ;po=EZGS4SYK6UE2A7W>iHou-y4%cv1rvu1} zBN0opI!Wd|i%8e$z?2_v@7+3b8_xKU&4qBSchtV5M!+@=#> zR6eKljMAex@Wu#Z%GLpU;O=vyIJMTno2@p*%xnGjSb9KF)2nRb1|T%6dO{iIWHIpJ z`K7h|n7$E9)0~usozS1$DxWcy2#!0s`-GUssAZm%ul<=@b`&faChgL`6mAfD zQ06~SNiEA*Qf||mpqZ|~(PgOB8J1hkA=5pO1COyOeUno zB4CpoCi)#a1#)o>{eDFojaRN|d&UIEv%O0#_vxYs#2`RBM*czvI$669aI%EpPC|3l z|BSp_>9Jfa!G0M11X>g%k`VR6L+45(p_MDlpXb8F$y{S_WKV2Q`HD5|RE*E`+lnuZ zH7n~etBt8Iqh{rhv=nGdg}y~OE8$_iz{N-R;$>5x9Q{dH_P5c@D( zy*c_if<@&^hc17X5qC$loSXS*cJr%RG24_XU5499OQian_sNUa0H63w$F9B_&Zabf ze3Gf-5A2FMUExpS;n zpNhEseVhMBC!!Hg(c3cB>1r_TAUI_kDZJw}LJd2pc{`Ai>`N)iHx@05lw85qreTxU zhuhjP=XlwnzO}=SK?fa0jdGUZ#69Cr=JZEi>mdR%3l-^itu6}`yWe!&Ojx+veBihgU)N14s@kMlAsVaYZ)||ZCyi!U8Pf@;+dR_CaEFYgs!yc`R;e( zOoB3%xR~LAxy?$VdL_xb2;U0Pa88^J=OrA;OtE<-J2D;xf z*Ze3n)*+C|UUmb`Xm~mU|McK&O~sL@w5$diKO(-MlzD74$BU27_m2$lb%k~gh) zI4GLO@tT*R_B-<|W?Wog^38UnFY4I@@vx{v<6efM;rNkwu(=d+0V_bVH@m1=@@+^k~Qmw)$>0cSj~2%Jf}B zi2ajT3F?I^-{)62A${}qlrm>Q^LxgaTH23~MTc6WIkgpZUr|L~HJV1zE=0Coz~0(i zIB| z>?2#hM;PAG;{PE-jPOX)q_M$>0D}uTXX;(78p#`i(~v`~G4g*=R8J+i(}7|CgPbB( z_mv4oELBCUV-EqskDy0!-yNMCs_Td>Zy}M0PmFM1w-B5Isif%h1m98_=hTb*?YJtliDj-bmsM<42_1H9WLskXX|391{%DaY zu!7`ne!UjJmf&L{F^FmvUCUF9E{jRo=X>S!F8R4C-k!zP?C)VLNv2T!mpDA(R)FSYwB_QzvBpu2sc6kM`)i9S7z#Gn~ z*{hRs^>W0&4xJIDu#W3^-@v}>=!igp%;9Z%Dwt6>>^+TQX%RXM45ISKYDt`cd6{?`J@eNRupu*F3G_cP490B@L6k_Z`2| zewj1E8_oPHm9y8P*+Umk58j!+(%*f~xOq8LwKI|V&TWH}g-{ksuQH4ku6Y4k7=3M6 zWv%ka?`kdG2|YN#-*#5?uJ%toLkXm!XK>XxnEO)H+b2&q`rFR>$ZO!8)T94o4A0_^ zjf>5*G<(+3;M$e5RTjbzx`!@CsSDIA!{cJc2R#p$8Y8aD;*O1IuQNwi2^wO2;}@C; z6M>Tk3AiAys%JRvMV5K2uaID@M}xPBt5m8OwPc}csfNsuli6a3=5wIK$431mpYVMM z|IeLtK8ab0P8Fy661dVpk48NTbfTSDM7TuUD!g~%ND-{t>}guIv{_(h8)7_fHSn}U#A62HWr zu1B+t&rYQT)yjYs?4y2BTsmWaI?${i_`P~M!(x#j*P#8E?`|cY&O#&>N_lMCY$>DG zSaCm&i&fbXbMFnoTet)tL1qqDl}Z6X>(?ZJz}`BpamFb)=cwpvveDsl)T@qJh`hv< zq4`)&?<%pw`l8Cc2!0sswW`)q1b&Mtdg!0Y#VcGHp0{yp$JwkE5tp#hDv*e2=P<3x zaK38^mo35H+8fpJL>BR8X~kXbPqX;> zcg$(x?G-=tuxO@X120DyG(=C4wWOSKyS__k^mKG#4Qxy4GpvEWv7l-Vkc_lj1R=6< zt|ucZfkG7D3|fdW+*>&2kuKd!_r{OWhc(f3cVHBm&ZU`w55RZHm#iPBIN^4qt zhj@gQg6mH@BBn0)eEi&W)I1-#LE>L%`_~X*mO&e!O4I!erOiopx~dh6cMo|_;pbIW zOJL^I>=zGLxzB}KuJ<4d;1n5cwF<~EJlv6S((NkH*p#`4`WEf~d6A<$9P;OeGZa;S zG}?SHEQ3R;;R;0tcRV`D(7dlhzG7x>j!AcG{VUZYnan4)dtc<@7nFwD`@i{JO2Zz! zk6bA=)I&s1DL&zcD2wcYJk@zB-d|K*`AO8XcapW57QLsEwyv!~ru&1on5O-OdNBBr zN}*H4+dG_S;WEKx#KJHqYd3N?6P;zs z3GM`Sg+wQ{Jghx2zENPdt{!62!DPQ%Yy4E6Z(K!WK%!k*@hm(4Y>JK*CXC#g8zK|7 zc&fgn@(LNw@qagwz>CR-5*5OIg)R%U-pB6He=Y5SdIr(*nPKN?Up}vor0jT~^po_1 z$0AlnLb!@wyulAErwXLs3n?9cqDGj*_+214Bg*Yw`0%k@t(&F5hC&ww!VO~vp^6uRFFGCD2F_3N_7dz0nP?G5Ap4!d5A-yWO&atQhln!ZUYm~WDQe+(|S za!vG9N2?{4N!VvtVF?#9eL*LW8IGRyl5Xdy`drBB*I(Ziy}grR^y&P2YIUzA!gd<% zx=r2mcT5&P&%By&Qp;+zbJ_sNRiAmd&8EI5K~U-2FhK*`Zk8+1DlwhC(%m%<@LmsY zPYnRB=dy^RinJo=WF5-x(k<}(%vPV)P5SZOLP9b^RSRWi%3m-&-p}Y8H1pa(=v|TmBj1X`&&6zKh0P50?eibrTgU>Q;QUGZ@;t zuJfEwd(16-s?1J4@(Ngixs1Xy8=t8!#2h=HKvN|V52m*^w*41#Xz{b2f>yJLpp4@G zFWJ<>-%@>(b~Lm+kQ~gS)5+su%Kmqfm1A^qZyD~;AaMySj3anIZlzf&`4`h*v&Lnm zBVkx%iD$o!4~0JTMtQNbb5K~hGAA3>MWg+mL^0Pqy426Lgd4oD+7o#QU1WJ94(YAL z5s(|nV+A8qfDG~lUTJ@NGXp8d?5f)KKP?V2fmA8d3!5-2FE7k15!XtNTx5N@fv=nz zbDUCWcyS47^!WGP{dR=Vj?wEPB=;kkXjoGvj(qunIw?-H@@uS0Z8TOMdBNQ)P})L! z(aFeRwX+8_UCLaF_H>ErSPeW2(BXG5v24TR9nGfbYov0_3tZ4))+3&Nk;X;Gv*5XY zoQXJ!A^)a}x|#HeD57q8J|%$kkQ>ANp&Lo3=9B@P3GREO=w($l4^<7)H$O)2(eP8?=HQ4C{Sw_hM1oEtg2qt&)HOmHHlA#FUxn`l5>;73ZeaSi z*lR|$A}-Hm;-_p=f%g1GWM18%r~p;^#4RuBKJwI?EE~TtiSX57>xxM>ek0gAR5Nog zuChWi*A;Gh4TZOQ?vg@_zh-b-0Cu;Lqc= z#vmh+;CfLgjU$At%vINImEDVepe9F@fIfy9C|QXWZu5*RLqY)vxPy|xV(l=yaWP>W#A^%emmGuw2b!ekzbySo2g#&IhJ-a)1tT0K7-Pc6qm(|TUDu_vGZxx<>8qRq7m?p+GHHsGhCk^46H z;IRS+^8DJGW}dH*0`bz04ywo1qLS&0nCY~ZoxTfMBIQclWbF~0ZB?#xjYMS|^=w$0_zTTE-h*0Yn{>zO_xq&W6_5U zTTArUXrIDGJM*}m5^97u&=9RvE^V9HI_J0O?WRbSmd4*O{EZ{J4fUA=JN%o<14jax zCi94oD1^z6k2d8RyihA$_hwAWgE~)W8%nf*_78bx4& zDkJbrKv(@$4sriAmUVce#EZdSm%0m_M2)&#V@S;4gCKJKE~Zm4GdtdXa}< z%D-BGL}5Vv`O!(%o^fNqWtxTylG#w0WOf|>1ZsX93|FS@pm*Yf(xX>^y}o?46=h;v zM89dst|O(D!jI97_4h|zYyxDPvqYmHXunQsyJi>XLF=Z}41oJGN3vsiR@oz3;d>NbIXlFl_xA>@Y9wKVygK^^&V<&VC!^H30=9>6K`24nh?2kLSJ*@kQY)+8t`QLpXQX*9n z^p!RM)oDJ_H)*d0&kAKLkpEuq$H{Ve=#;yLQX)-}7}PR!et!@>w@oB8$r)8S5n~9QBnltlybHOR5LHbne}K47?wgW$hob%xw#Ohq`O2=G#6U~H zfB}?7BG%#|NXwFD2$D`>AYlAVo5JvUH;ebT^3r-4gytXz(rRj|Ysx>%JIfoeaZqMo zTya*_M%@13G5JghcxkQW&(3sKrT)zIx4)D|0QT;V0!tMxOVB2szA}7Uhi(vm)L#AL z*}p1SDhLE!RJ1uFAu@kM>+MC(7{r&diB~u)6LBl}NgS}HHF^C(F}(ilk%}*eMjTf( znacK_%)MjgHeuzhGvP`NpvIqS5E+U%?BSwz!@G}0#N2jVXcnF@IZcXsJ=3?%nM?;q z({kla(S(Gi2#SBoqOs*a5g`MKe@@<6Itgq>jxX#gvFh}b>GWZn`}Sr|iP`YCHTUkr zFXfMSoMU)*UhGyThA08Je;Rhjry{@As#BXD`0-Igw=!;T_=#dWtZOHJMmdO;B34d7R;@?EB*%3*OM3|!Gt{=DfW0L1r%7XShFcB+LfD1OWn)>Q16SNfJ4qjL(( zkk9CK6seq2lz>rqlzfeGt&OTZ=n}UV<+#fIhBZ6mL>FYS;O$9k_Jg!!^bdUTpi$qD zJ1u~+h4!gi^dlvc)cb)mynO!#c?2fY5U~+&vhgJ+q^*$!*2hMs4u>rVX5fE`!ki0H zn79-Jzt+7L9Mj97yLy=CO{LqBc6LuQ${&`=)$Te#cTdXE0##uc_?)zT zOM*#3A1CJdIxKr84gALMxaN;~JkPScl&T+* zxmZRMdP~eB_Ru}uL8{zn95tRx1fz}w$qq%25-^NGFhr8+ZY42t zcWlgfX_LPNAYP`3M>tPGe+Qhn(OeX76w_R;GJZPDFw zhnRQg)7$ToI1AtSB?>m^mTo^ks84?OQ{l{DtC{>2ZIjzCSEp#kam#RF5?|0mQ*=)K zk+KBp7TP@Cf+kM+P6mLvJB&0_{N#TAYKVkiyb!gOQWcpG>CX8kg0^YVlh=<1o;SbJG``+0G7a1CAN@WsY~OV=D~Am#7B=Lh)`16O zMl7?YGwEvQfW1o-_$rSrG+3t_GNXgo18ljjg^msP&8Zcgj<&voFF zM$~xk{U?!voM7{(x7K8eb7y8Y3cV1To3glg#KZ1P#OiQ~?Bc>Xm09o2J2U2RNuEt; z={ZJJGNG4Ihjpc8w}@Mu&{ZMqXbJ6eYv8pX>0GWGM!B0^X|uu*4{_#>qVCU^Epbl{ zjh|(|(I=yGtBp3=MR?=9@|GH4k2lMVqb%5QB+4NDu_394+SNGX2Z-NDX;0={$szoD z8gWt+V>^z_-8jxkN1c%kw%{iO_SpF)3`H;GIMfH@*Kw(T_2P*SF2jg#LeWk_|7GwI zhTVEb0CHG)iA8yjFH9qwY+W>bFTdNSV;qkr)6UmAha-HtK~1Rmo5AAIrUCNhW1FhxVSn^E-)n2}l=;01JyC%HC|aH>!gsFv zlP6fGP6=FQnLMT|-|_z4(C7A+b&j&cQrPP{eN}~=T|BCQ zS*k}$OSC!+^WDhfSIh}6>3k?lg8unCrZ`ut$;XeeYI$qjLKodAhuvP`F494c6Q}>w zz2?OxEBit#Gu~*~g^4QH-_bDjXQq;+hPSW&bH3lDMgF1duVC|e@RWFx182O*2`>FF z(o>RQF?m8zoG22&x=9w8lsK?a2UqJHkH8IqdknSlKb*;f))etoV(p-ZY+v773CAm% zb^Ms9DI@hC{jrgCenpn@;*gwyO^C#I)(RS{7*_8F#!oX#e7KWP@B9=jEI|2G>ad<^ z3A}tuc4|t|9_-EDn%EirpsFDcdKTrew7-E0@#f6lv>Y%bv-|9#9i_Ni}@P+|sjZ=Yaf~XIWrY^>YwF&b$+luwMe>M$e6eBR)#RuR&HcG94A{3K9Rp`IVF@9M8ojzZEuQe(Vj3Jhsg{&aJEcG)h}>QkibU?yIz2X0{wKibs1^7^<#8p-tr1konQGz0@>b1{>Gx; zhfNSW-Ae8se=Ze`zc@DFc5kB63;{|`i5tOYt#p?^Mj79Gh zje;u_jVAVqbP_NeJsB4EZ;KyI_S-S8{=h~$!@16t?--mMt{$&EPzcLP)YFtjStv?+ z%wH}wJq;J7wO+2aG}Jqov$Hf@u|j2VETNV}e99#MH0JN)&8n|<`HsHqQFpwg`S+3k zP^zmJv$)g*oU4JjgivEI-u&iLT{9H2zNfW$T^AF^ov9@`@aVT`H}A*I=dLsQ5KSO>Ztno+#HuCEuymcLL`6M&hL_v=o1#RMjSf6Z`c9`*b#3 z*0^}3?B=DpB?)OrA84F(M1!0UnZmLpOxx&BB^{vaXgUCoVEy$fd7xZS_>Cq*v<@YQ zC{na%e8dn9(CQc)jJID?s*%M{7Mq{L-MCZIqdvlQDwJ~bSM$J}>`~ZfZlX2+z94zm zwym>0iNPilDL{T0lsGBr*uf;yU*DYUvv~Sv*{Oi+4ZYUTN|#32t$zaQn44B7Ik2Uu zQ+wx7xBR|cK-Pt4h@Z7tV=hv=mIXeqy=PavF(WOCS`V~L?+f{1J!t5&B0b`2Yy z{TSoFX6L0pa8O=?edzH?esmw=f;$}*lnq-<|1tc zW=y_0LyE~Sc)WFVV61KF20e!~|LwQWxqu8QEw7Qbm=E7k)Cl?4<#v(|_$n-utK={e zrNU*1Lvv&z**N3DpWe{*mB(NZeo|vQQ_!EL8^K6%qy{aiCY8<$RRrRkW2MG)N{Vg7JfL($*i|aQ}ws*Vp%CDRFO3-#oDHZ?)kt=^d!C{k6xL zoUkTWCCcTrz8^FIS+MuuzQQdGb_nrrqqzh;<6a}bU${cP|4J;b6nHB631tX-s8Zsq zT%eI|Phf)laYx=r&@CD}@)Hw-P|1AXb++ZRr1QJ#NWWJVNg`y2L>U;uf(lmIxP+yW zG>Np+M+i&S6_tU_1(sqOoTmSvB55;U==&|>0OlU@E?FUBf?5w0U0z&&8gbodIsF}Z z!$ia^D4KZG9sSBWnzcsyY^Ocx@$Q@p%B+YimV}G3o=BKIb5kOnfPDx!5Tm4XNJ3UDC`(?wAh0h6 z_US)BbQbPd)FVp4R4%UUJ(FXpxQnpI<`}sfA6v-iZCr8IbD;ZUQVGD`^VLK?x7?H0 zeVy{z=`}#E^%d#by@>?MJ~Tw`FoS7ZNUxD0-j>TmNA#wb#r6O#CwYd@jg1k%U@E|V z4%TBc-Howjf6A%;TfFP~l&tX>Q`zR3jDh#qK#AN$j30!6P}3U7X&r9pa1w6Vz05jl zGCJs|+#NaLLBMCS+PvG4e3(fm#lBNYzabaijuX~GnO3=^tC#i50E=7C2HG6y9zwqE zi4QMffFN}rnE~K-G}!!dC1a`u%bg0LM*i55spZ21^dFvHlgpOmXF$qtvNsxd%OG8y zC)Nd6SysAav6qAXCM`|;sL_C8g_`U&sUa`_a?IF0Q}>>FA#WyHya6I3ws=kbXg=3v zT6X|JAx{-akPmn6>|Xtaw(6F(u!-!N$%DR>>&#hIdFVYFQl=K~kTxLIi)x1GvBPAq z8~gI54Z|NCW)uKO`-zG&Kk*7|OHy1_k_I(*(fapBYEJ}{K7IDUPi_?)44cld@OOH@ zYyoPFhor(|L7T-9rLv_PXX6DAcrUs^`4QIVE(`>yWqrS?6|YXdDi%7izYa>cyB^vo z%$*6$=w@m+Dztj^+BMHPxU64t`wljdG^X>$lYF5*oG&T)BN+WodC5;WD^S`U{UZUB zCWoOLOVydAL0@(zH`>Htz=I@M!cUT_l0`??(id{6?}weh*Z}oA<|5 zDxNmxlzS%oPq!UDWZjnr-o2H{Ux0<7JE)|Uu6|)~s)!W%Zo^?G!hp%=D)d~y*KF(H zGv^^hC#f}sITG;B%yU~|eJPk778sKSeF>@T*uuUhBfBnJ1F>{OZxfb&7w>`V0 zD7z>-KYWu5o*ySd&xO=d0u+`G^wbsVMhAF0DIyNy+GuD%wVcb`sA}=ugN`~D@d1GW zAWEUEAsZWv3(?(og$>M2j=mIxBTfePvx)e_}_{iG>HvHdp`}X zaAGA1S-{FJXD4jBBv|e)Sm+CfljYxISwDV}ADl)r zTA6VckOT4}zDRxgyK508<6B3CJ~&SCm7=wU4e5vi1fhal6Qc#rC@U@i(vFjTT`Bls7mxrU!dNJ5H#J$o9ucr z8%`&H_Yji&BZv0(q>|`zRrAJN`1IxBYZ1yrn^qc%Z*Ei>3TA$tpN@tv;)=G1Way+&P-B-5;QTGgW+V~?J#wvltSt_&E90g7s!9^w>|{7=lellp2F-g$7HYiy=+%SL7INiV;Z6W!i z^nG!{e0#)5>J; zuDWzKtjIhQ4PGB=cIbc!_T~$>17&DPk6@qO7f(Nq)o=N1L4QwQa4XuYK8aQm)=~UY z#uenWlFbu%U1IC(pIWA{c>&CMpnv;U1VNstL-H{4sf+HD7NwUyK2Ttd#uo&Sz->V2cbpfW%@4X|Zt(B* zcsQ9nJHye_=nMAflfo^J>h9?}IWHtAn}P^#a(uA1J1JtdWRrI-I4y_&+W4nm_Irnz z<*DNkhSwLv=^iWKt|qnBc?N|i3cCS`-e|;$g5?ITkDAc!ug#@7MYZU^fpzX~N8&2E zxFO86z#E0>dct2c@(@IJgx~DGp-K3vog3tBLQ;EdxYx~V7WZvPRM+cjcl{EmZgYz%$t^NVDJ%ZX`yAWog|l;vYCaw4q^6ffzx z_RGNZw+cR1Q4%2%3ccJMn5^}cz3bo0+UE+kIkE(=1s$ca&KLJq_JCR7fV8pa?MbTb)O{E}wZ&Av~K+^39D{ z0*9hNkI4GmTf1t9djiXiM+1ePl#%oMm?8TLqmd3nOLHQeFr!*5FREcOp?~O#f&LQe z64$`-hnH{OK|Tj_4T}Hy&zF}li`mEv`xY2r&y7?wrub`+OEiTyDcn1V^@l0p zO1;{k|3QR*fNp6)SM{yl8p%>nE5CQp(D3Oz7u>kaA4?lTcgYBBoQwh?`9%6bh66t^ zYg2aSMEM&WgQv(z_MUZgJ^)Vc1tK1Bn;QC_AeK}dZed;IOix1s68@(pgjBe_1RdjM zmgYm2`ED9MFGLb%H~yKlu;s2 zTzlId$9s-j+EeySY-|S1x6%W@lCLVePffrGH5+dENA1QZWzg$G96s(14c@QuvK#jE zZ}?r?UPYrMM$=Qfq9w>bsa|*9@^0EO^tF~$LF{_+fwAGs#B~l%#Fr9s`BeH+`E$CO znc)oVl(BakFv7Z*-@-v(%|1VYIVes=UlFds!a&XY+=3+cPm3wXP{%n@??EoANE2gy zAgs}t4$$a0SG<*hxoSi==huRDI(I2T(t;f8Zc1>7uZ+JX1`BE@VtO)5KtDRah9kvi zlR)V(NWb(ddfrU*1cRT4t!yL{-_IptK9d4JX~>7Kxg>$!pDV^JY9DLnc}JDTAG$b( z`#W>E>15ol#(A5@qjDlnHs+}(qbBVLX`}r^(CE_zOD)dc<=;wdN6a~YUs-w`5S<%# z-8b}KfTR9qO8u^Uy4)$!-{)tCT$MEu0fqnu!0IA{L`kMmgdoD_^}h5sTCK)gAVzPo zqvaz>Uu0%?#1hh2BKEzyvC+V+0+_a&db_(qtE+}KxWB465^T5{DR9K^p`c@jv70UV zJC;v6+?cAWSSyGVxuK|upOC4v3_K?d0bNjqWoK6Ih)|HM`-wJ%IJ$1rDee}_I~2PJ z%*^tpC61a=i{@wmlQ-+SuLSXXRptH+{~}R)41277@uS1~9K$LaKhar?f&GoB|BJ#M z>4>Fdwb7x{pZy9WL0mX?tUMTmG#k>D1Pf9?%WK#(>Aq8!*hh@7idqPRg`3Xg?~kCV zpu`jF3qGsIWX3C4%a0sbQwO?e)49mpsJ!UM`V`Xl%)M6A9p+JfA!95x$+Nj7N4s^j zB;spdK>q(yGc)3;rr(O??S2H8E^;Vxc5eTNnGvembSPIa2$gHqJD)sp(aG{-wZC5( z5!Fl1_-=Od_HQRiqYDbt*EQt=m%)xDJPP7hx-4q3^9|Pdv#ER!mIPL|U zM6k^#ckI9~L@CLdx<$Qb5H#n%T7XKpf?vp1QJF^DEwPo-1BvYhs9)_Aj=IRCibvUIZ5#_hf;V80dIN`@McGvp;yYMxa7>doB7Tt&d6TdvciWS3K zR3HlxVzO`!hU!{6D$9GB@@Z8d|1MmnRJ=1<+!SEvHCDf@m*fZS`LF_^lP=@Xv zrVnd&6{0?TXM4=fa_{F#TuuZyMetxNMDA9A%VyI@6i2TeIeVQ=PacA!p3on!xw$Xg zLHLW{Yq2-&Vsoxf92Q>*5cgJ_z95nc0CwQ>?#f*!ZU*MEgPU zh}{k*zv`L^O&t(S&(%OC0rPu5+VoeL%YCXQTQeg5nE&pUVUl5oPmu z_nsyPBZLO$coq1arokcuZY6k=_r;U;hG;S+2(h!iVbpM-NB}!wkTCiwr=B80lO9z_ zS41%n*D^lesD2bjJ`b4>;K~y!_Xe=P-p$>D@|!MwKx5*Qh`cjL%&-jGe9{9`z=R%T z-O-poB~c>jpv&zne#ej_J{4=N9Z>;RQ5-AcWZKDotfWHc6}n4NRTMpdK#=p-fl8?% zh0AN|;LL3&m)AgYQa=w3q=^B-3gHB6MES&4d}S>gzFo#ih7(A-REwI5OXD8<>K_<% z7<5IoBW&@&WE%zCsK8jh!R7^8^@QnMwmCEMt2$^&E6Cv!TW?4Y9QGP0v@v0 z0tJJJqk96r_;10)aH`oO3yr!Iiw_CtZ%93ko9UeTrE0Iv*`4M@~*x z4-gFP&4E-{n$6a6Lc^9EKCi1Z`AKiyj<7&{`mgY-#d$hq{v)=zG~JXgZY%OVA}FetW#vb=SDRrqBIJGwU0)< z$)a%jn+E9m2K^F7E#YJ@eInRypW$OjdYe~~cCM$Je0=g3ZpOF}V9D}`HP8i=2C`+k z1+laTeaZfVRw~pbdq=6@4=0WW(U5;tivp+CjVl;(b0^j*9_UO9=OAZAE;IAVR(&8~ z$TRzjcZ1KMRS+|b2KvpXwG6kN`(5zTS7Ods-1;7QL_Cd{e>Hk7m$1Okhfay;x_SS% z;)RrHFfT#8nXHI#qeth!*6syj%q@Ra`UCLIXJ8(QtT+i>GcndV?}p{JV8})gx9b z3Lcm(fP1euoBrN&gFsaAww>l&FKPe@bSX) zGN-*CZxhZn)ZyI+Vs=aqow@4gzIs#uor&BAu}r6M-vM5~Ju$Z+W-glL0^X`Og175q z8MD53vO#|?2fr2~!B7_lTpG_`UPcidFW6XCyp$|FCm-9XMTCA~K2Fn2SqtjlnY;{6 z@Ipi!pnJ~jWj2;lRnW>jV~zM_gT>m_e@L7g8s78_%-ts2513iPlrq#5kSS%K#cRG& z8tbkV$LydL?z=`tBdo-a_l;cZfS(?~8OuCxJU^@3AVua4M}vdZn07zdFGO)%AN+@d z#$T3g`L2dGZT9O$R!;waiH(3_Bp5i_z)bS_ zk~%M7YIjeVdj-tKy+xMoH9!)2KlX4rH73PJr|J5FB)S7AaB*5TBdHtt#`>q~a1T-% zY4RiCd!>{ni3}=$L~gH0RF@3OMQ5FLvo>BMYm?U@ZpDk+RQTgfw}HKU+=5~2_DwM0 zD4S^f@vZk0Ovan1O}N%67?U8h18v*o3EfKLf?%O6J!eBECyKiDSV&z1!3k01GhIBO zx&wIT92jkumbXW+?yA8wx=(v~v>^ZZ@+PAm?Rjp`3@j5e%S7N7ypkIsPM0K1^Ez856oP{d-} zC;}n-7$R6w8JAPEQRKB7>iG6O=w3t3AzXE45$6!;zoZaskTs)cQ7iumIis{yCbiyH zb8@t#6#A4y0+7=UocL~*uN4;nP`XQF>?s{Sp_>*{Q}wYSU9AcG!)p`E-5UH2` z-ldxMEMG?|SUHY_ITu1eKHdK-p#UIQU+Cl%K;-NkvI3^lk>c>nbbBOxovuT9fu zL-Y3jYf~xw-x=hzXVl=znRqVXCgQ_~kh0nwK`v=4b@sk%Ezn<8!qrGdaEs`xSZou? zf(ccF-8v?*Ris#%aRTdonM#WKj<{~AeAAHJ?;%w2R`-Egt9(X!1)eq7wvjI+`+?Sz zBMy`P1_Kwu8A8E=cI{fS1G;J~xBNhMW^npR7TNLoOF}wHTzJLDK$`gE5~pzQ0h!0& zV2wWlSMlI7CCJ8^jH!4+-P9xbM{4)r$sS;;HJy-jeqcieA)pU&)toTdLjphFFUlHY zIM_V@$;u$=UCym%h&|VWqbDe}-pEX$<2~q3b?dT3=W>WDjtb&Nd?bueck=x+^RCdF z*?|@R4^>~`5Oo){OG?LrbT6gLFwqN-o`T z*Y|zDd%yb^{LXLAnK{qQJkOjA%FD^k|2YwUH6(^qdHihCcTq>uQI15jXIICq5s*)m zheEu?IZ-}B9w@8e4qpB`e&rfXDVPw63368~R@aiXgSix}=#+)rE2*EvSuh{xPe%necD-Nq0@u4>lXFBsKxGYq$5Fzz=*x*}l9N-#s|WjMHnYy3F(rGAf89L{Nl- z6e4|SQ8sJ_?z>4{2$N_alY8H1co1u z-wlY?FiwV?Q|&$ESF?0)Ska#lAGG;Wx2BTH9!RCiR#-M@>}c%Cvior@p{6FAfb#Li9wtk?vr;vlkQ(r3ls;oJhTx>F{~k2^dNlh5=c_1fyXvI!FnnGKAMe2(bJ+yd0g!vw~g0R z?6SOO+0)WzPa1xH4yeo3G2EcW{2qVb)RcgdjmHJOAzX z<{6WZGX^SH?)D_f`?q5(&d~kGV$1TRAG7Ix|IxSqqmVT{Anw-44P(9X4Gs~-8}W7c zEfGkhf~3;=IWWo-nDbSyYT0IhbIpk_gwG&`j9%%_MT&n}u0C4b)yhHM~60A`woHMLI6kQFkUpQYBFp z8^-3a|3!5tgR^M{fvAfS9Yeecjv8kM(aBL?lb!Fru3SeGsTW*v(t+eObN}DvD?i|M z<(Lv(rH>iSsgdV_&CB#grGh*HQ;c#@K)d09%k=R^r@=OnoKJKe@z@^<1;m?3wq);| z9Ej@_jQOElgQIYU(3B@n@ACFSiiX6Q*ka&z8M69kbecuCJwyh%g@V#Y1Q*aW{HQuv z@+Uilm>Y1<$n;?Y8zxy$-DzU|9rh>x&+mdkGG|$2=ZHFMGn5DJ!YKxP^-@zB5?{qK z>M)7$WeC zt4aX5pT0{{Tqd7SSvDVOovdyu(WtLukb?Zb*zypa&TV??5G7sZQ~Ue=vq^7!kdc`% zO;4kmp{0a_dvcfcSXOX{8jCdc zNRWpO(gD>m*{nsGt#wi3Y8XcBG~3_Hq_*;{isD#L&wtQ9D5f7=i;GTn3$zUY}rPkoFb}Id?g;d zxLwOOOL5X4>KduWRpSNJqSq)9w=$-mSQbL)AcwH-XdoYDLv?(ErBR>Czz^iH(;dxb zok!fHClDi&TD4uc{K_pmKa=UIsY?_ekm!C}xXAp)U0BYjpo3oI(k*(ZwiEs9j$WMtt zIti$peb(xjuPasI6{S5}SSD>bTII3i)xmzkc8al*6KRA(^=R$s?C^v16 zA|llY(~`bttZ;k6_6U?6?|8C#|0fK>+R6zs$w9vtxPN-?8WR{Nof`Q3u^ZW&JUS`6 zw$JeZ|IGYw^lW-`U7t@zZ;l&q_a|w8fh8q~0vh-WslGuKAW&X!3aoLlGmH+mq+HOmuGY$(_RbO3o)E7wz;4Z&Xzg~Udz@>` z{okA(Oz-I7on(t~vAOxvHB2C!nB@b4dVS0A?`eKWHOGf&|b}%!-J*(&G2s2d{#|ziZV7>LA3=bM2U@4z zx*jnc3lS1>J8wfF?O};%4n9)9&7&jT!!7Y@#Y=!Aj)0*c@OZ3E2ff3FZq%Waom}L* z`hcg<9|5_GB@BR-a$q)|k356p=*T{Nyp6ABv2O~g1H!i;gq)VoQ(@53R_Mq)XA$sX#xn>1}H?=a@{ViX+ zkw?pwZOU2m;mcX%98mWtaYD(Tj*=BhP_V7~3zF#Zl6g|48c%H3m*pS7oeTsVCvyjo zL2KUoLC^%*8mmZnG9x=#yMkt6Rq$L*2RGf+ANjvCp?A^dSOw_#6_O@OMu3SE`Tvo79Mc^|>{|tIdCy2U<&uE?G}3 zLa>V;{>3c2T+@V!ym8*snB~`JOLBEwi1I7FaH|GIkpyopJOr0uHPm&&m>w&^^XF5LegOl1>opNVyw3q^77l8 z9!los1riza-iPjI${1R*p;T7bOJX<*{^$4~7?TK(ESM-8+OguN#pn-Tl}{7qU74(X zuQ|*NX|FArr;EI7BK9-CO1tguio+}sM|CPN>d6I})6|=*%?>S?t{jmp)ytO@qx?dH z;j}?Uxf#5Ar+F`?m4fX7Pe!r4RSZ}tD$#q!Z^yB|QYQMnOLmkOT#R#VqX8m)xO)?Z~0 ze1M-iX0x#lVL4>v!yXqBjmPZMTy2bC{QjkC$9(hdP7(Yw@Rgn?*@Ark7$yQ5`!F4P zD=Jnesg?_m;2uOU)wH1kOjV@7msSO%GQc^x6&L5%oiR+p`)u9J!tWX=aKe4LJj|Cw zKk5-N$t878q;*PDKd%FUiVP(-8Y_W&jT0h`TJw!&nXh5t_h_x-!!<#yPq?5pHgVXc zg7n5tX?#L9%Dq^D6T2FG*$PS+6c0^iOxQHU0=8KrrmTqpcD*V8RKyw5T_KgLtxE7y zQ%sZrZ>Xll2?+4dRLWWmFD{)xKZOzEG+_inU?|pjQg;O;#d9w~43zn)Ozk-vvj^=0fjny0M4@SQsWE z`C)7l#5P}x#}mWltTn{nUvmtxl$9|6t7yY=C&*nEL%?iyDO*x}zL{~7upKl&cypmz z4P`7Q3FV=$@8Ijyg~+cV*@>*Hq6q_w3@MppMw9Nz=moTxB|x9n@$|2WxHBpzVCtBZjox zizrhgirV`O)$N`mf%?jHWN3s+!j5Mmy*MdxKO7|$EQAdwIE(wT2#Hpb>a=HT{#&)~ z{7ZS%LGj~hg&h1=x3&7g3t)ltXaA3m#@4GyHsZFNyb#dp1IVV)iMYW@2-l{ZSvm-? zd824?kFg+@_6pd0d=aqw_$MfqHUI#%@Q3>H*x`ep&pKxPDij?W%e%LpW;q@*o2_|3 z6XY?!5d@F?^z4kpBqBphc;AwHGvE4)0#mDEak<1Xp+(ah zchONxbEa@&LJ467Lh5kunzE8k>psU|H03`h6=qi-{7bwH$eI?C4%sl@j7U>Tu|zwriKUyTgorXjXqrmh1!?&rkx1IAg%=p8e`y+29O^ z?5B9fB++24F@iHswvwv=j{L^<^KG8eq1u7K75P@1{q)2gJEQ)|{ef5QBm8BT40Es9 z{~#`e?Z1^F8_AGi_F1ZMi(KUdE;(zH|IXu)&G}JfTEfXN$YWbt0GT<$~ zFaSXFN@iey0H+zZxvRMMUT5$f$qLsjQE)$K~o275iQVmDRv z5N^b8;F8fF=@8~ZAi)sFif3{v#vR&m(()lLD87!J8HxgASGcLcMj7d_SxmH2KJ=>>-YGRy}bD8LZm535iew zs};#R&wL(ir~MiOVjadGuswpP8JWQ9-LmKRrxAl+^WNzJgY)JH{QSq@;rJu8f z=#7Z}Q=JBHhPzv0D&hGpcs_gmF917?i+ZCV^Tef=pMyWTNPdZK1 z!ylEJya(w-NHpUc0En^omn=iRqNL0BSB`#~)sr-mo!yAlvQ2BP6Q7QMmgzWxGN-W- zHQVW|cQlB;%}7l9^p<48S6VUc06SygNP%h3MMoyx+5GNNh*zWW zPIYq{olmObw|kE%8;qqj2rL^i6|i&mJAsV-e$mn)I4tcfO1Af-~^iO#0H zqIf0;Cixnum2b$iA=Md#uom_+^TFd&FdYrgCDv(|kT~yo7x9}p9|zjK<>f-WjQI8L zulg_DRk<$TTXKJ(^i=W>*D0cjXF~GlMNp(itJA9{Loh4aD!GF76T%IGzKZ=`W#rRS zw=rvTPMM#z+02=z-4EC8mJ}xFI}jiO8~9-ZhSzV38Rnuu^}boB-#8d+f8Py%LTI7> z*o1ivTrSZ@QYTj<9(Y8PADgWpma6_(|E|eg3EZF*{4f{~IYbJKmo^hP!-UJ9xBAN_ zyE4{>dr|@Kikj%y#>x`JC@6j(zpU5*Ei=@=p4DwO_wxFVFdyFE%i^bLkfBeAF88#f zb`V>!$G??`2H^}s2ULy_<@DcY+7Wf)G7*cQo}o9E{bo`g!-TRkWpbJHYOT@^R6~Xo zV!3o;WX!UYFjwhIB3*y3>Go2yPOUFeS-FhH!AsmJtvnfCMhKc!sKY7>|qqVpMcywY}C+Jn&545sg+lGb`2HGaBa91vMIFB#YOwcc-FprcJ zK!ka!7LJJ;weck_i~+9f<#94^SD~C*(1^ z>2i2a<-&Sucwz03cCzyvM{dfjE9x%1yE;+)mp?fu;?rP96(GaURICi^^idP6=l(u^ zeiUb~WM`X%d6C3TuEfi(&$n_-+jNPAEM@ z)m{Iy@ffAoNFpuSKk{)w{SV+=e#21xVVd&;NT?pqAlCAn9R*>LVoGj_L!F zXErm$@S8w^C%)u;WOeM&fkv%iz^Lm$yaS1{reX(O3n=4?mT2N$@R-bmeUqYwHuYC+ z)LeZWWl2&NAyX)SUYmd``^4B+mI3&eR-`*+9V9>HO>}{mkeH^lw+~lm)&nO-b^J8R!TK3Vrv>D>~ zmP0x!29!fN>iHG5|o|QEO=@R;jJuJHSm8 z!#jevkzZs2uRK|5IF1!Y#5Hq^tIa}Ph!&e1`#c+o`I#nt@xd*Igg8TknIP#Waj3V^ z=dChT=jf=zX}R65h8rsALY#q06W-A7{F!_Z>SuYGR{q%+{Nkc?P&fjpY5xUkFfJic zo;Rqd|1`aQa~EnY#03~~@^wdorELQ+N2b^-?N!WH2fY#pz%)^r^k)6(J*a$L^#<~9 z^6VMFBN>WrDhiF?C(%^fO?ro~21;MyMiV^(c)!2fe{C+tL3HDPde|&fBZ8|z7*aK( z7=|fc%m60{k`Bd~g~8s?h^;DycShbqx>qB{eV|{SnadA(a!$jdcwvw6C9!=jV*HGL zuHg1`?;OR|F*VlUkf0()6lX;W?ZR*1inA-_J`vyN7vUZ$bI^!czbgED)P>$?zRvX7 zFr?onOXRGkMlpy>{M{RI2c=*yMRqF0JcFr-zm`I{UdlhRv#gaJqRj3`E}1)Va6!Hf3|FW8^iNp@6A-POlkVRG!f$qSP8E9(hzs1JRy-5=+e@inoe zujA|ypjG%U{3y zUj@yn_K`3wJ-5_~GrW-1kuV5OxvCh?`w@C3pyu3-fQ8(( z|GeDFhy}_{qko_?G3dK|875t`t>lbE_1?hwRfKww z0qxs-CjqI5mvphrG<#^5=wo=f@GmWlQjrSwA0$&NVKH(jOfQ@5J~qvjT-gNxu2G4Q z?}=fYQU#@m=OW+hcnRmrqnCr`^ofXf>q{7DuKZ@~PGN)0sEM@3w#u_EQyC4lqgj_C_ zG*?{X65DeX81%zl%-2`5(@-FQ=>rzR!;0zt>5wbK>&Pw(>ac_GFK&5#l-hYP%y4RW zYKXQs77c}hQA@9m6;-k!n%{5J#^L&qo_4xeslE+cUf!&Z@%}i;CZof+;j^A~XLPM-?==C#M5&=~ABc~I!3cprcN2R-g4-w6plnjL-uncv{XEYtN9`rO zAE^A>+p5kFS^{NC3(zX8|Dv1D;jaU3UYSuBukkOc z5cJ?It)C&+Ka7HPSUYg;$xa4;g&g6T!tI@;sya~>Qmc`h!-X&yR;r!~b&!Iv0AZ2z zC4G8s593`Xx+8i(XWzsSz%t?@ z_y?t>B@QH{)J>fZ3}#J|#d`RX{CgpM)=1nr>LS@cq}|J&J0*RiEH=fMXuoTrHL63` zYV<@bHkXjMVhx-at7Zs6`^=C{V+ock7THLEoS{J{Iilj0@z2Cp^7{2>2Y#C_X(QAe zSa`X+iKLAz#SC{H)`O?~up%kxA*)mxDn_+)tQyxefhb@2A(q_EsBuD}pD}p&5&YB{ zkHJ97`I&xh`mU$G=hMK$gz?qiE?GGuDn}^BORKtD`Ziu=9-Q<3lj@yw+BJUA==0T% z@bzoxkd8|pZ%*7ldW^b|iSSJ(7o5d-h~27Tooh~z8}3WQTwWmLvTS@Kmw-9ZWui3# z_pMLb;(xIcCyJ=gZ_r!hKB?Aq`0@l%J{lS3kE!JyR8DkICfS^to<8`^u`hrSq}$@4 zzdh>-WD%9QzGbxek~#gu^M8}hbIuJy@!9Z$-*vhip)qCYiWiD!aZ@zYl$*Dl)7j*E zmzN1a{W?e|v7ElEPiXUBt0~E^eQbt555y04m_Fh#&az;w$BOcA;ph@9z>8U}!@h`* z(P0Gz>E@5Q54`zYd{*1NB$k12&n!BI$@amVsl>{#_+4iKHFF@(( zY?In%I`|=^d63k56>C`bToX0%b(!-AlwTjgJDg$e9~^^qkQA|p>12h5i$DEERc7v~nQT{b0nRhp1(xy1ZddZ`h~~QdJ;Q>ednh{G%Y~8p)17OqkxO#(yIoNVstf$tFbBN%mHH7Pu%d>W<>ZxMPMgb8p8V*$)-{ zS;SXwJ~vV?r(WH#+>_LEk@i#--ElEZSbb?2f|_{eBr=J`&j^h9+^A0;rt5KEhq1WG zHr5Lq9lqY@j6xIh0L~>NR1Z_sf0N}YRxt*)gVD5F2iDhx!~b_}7ct0AJ+Kk3#9z;v zHRv?2NovgL8f&>q7|f3JH(9>Yd_%I(yy2IxhHz9=!oO=F z;ZF3FqV$8s;|~)Im9)RH8iM0G{LAm8HfI7l?S42}2Bl>JE1 z!`ySz!_X6T?)c+2u>oFlx?*MAMhY$W5n6HZiQZcfbl5S3aWzu#uPv(XoYOkYy)Af_ z6)akQ^fs`9D<*4yCZWW!aoZam`sk!RE3tUDSb2Go}uKz|u@Qg`zkr@}Z{;U5QA@{8$&s9tY{3|M#AYZ&k==sP>G3EzwgM z%F%|k;pRS5tAyPro>sDjm1OTL1URSfTDD%gBF%lYfGUs8{s?=9*M zsj@;4^+<$0;}|&*?}}uEW3pQcqTkRy$g1!U4O&RAG4HJ!cE1aFL;(=q@s_nz|FqC0 z(5`64>vR(hB5cx_>Azi2b#6sya-Aul*DFQUyJ+4R@*pkHWGe)rY@Rcqh1pu80h(FxY#5IDV978X%u`%9ucZK(ssStQQ>6?+xaJz^hOB;v#e<06Y} zO5SjBEI-55D^+X~J)s)pJ(~XCK&)_4uAp znL-w)NA%n)H_ZG?u58Mw{_f>l#9rp|lAovvlgf`YiE_88=x?5b{;k)>ec(FHZVBx3 zozoE0=$sPWMi!HfyGnmfFthp`OHVF=VNZjqpYbk`&dhouhR(EizW>ss{TZWtolx(S zW=ne2azUjT(=q!R1tqG*kt|Sg%10$CjSkAD6wH$N5Vb#-(80#m5fsbxJyVY&#G~Zp z`@9@IIQ0^7?BCrC{u>b<5I*pCke94C`i*i$QZ@|-bRw|QNtIXzD7D8WDPy}#fFK?k zk1Y;6?!M&wwxT}-4swdd8_lW0Sel7LnhTjxk4L>ROUN2Wq#lw~q`m!M%2Kt7y) zqBxqd;KB<1wFFGM(@D9|>wtqV|1=!iql-h_mLD^qM|7wx{ z85W4BX?v}clI$-dKu0%1R$KV&9@$U8j~tNNkIK#noLbG_@m_EYS8C-L3)P#1nPc8k1U3I`Q<8n>T4~` zTj4XSINo}57X9FWZ}J4hBom2D!@-LO^|)H~(kUu_=VjinF_Oj8L)D!N&_3;-ifH%q z68;}vrSOpq8bQo!k(5@l*7(0nCD)c7;b4~#ueJZ^&X{+MtI(ioRcj*?US6~o0S=@~ zScjk&v+<@b1u><(rMJ0b!st7|6R0HFKBWsWZL8C1~YeECcb>QegeF6U5HON5nUBBaG2rKpHa)#GXUs%U^@X4ZGq8rHsm@n z9$ue{Kx60525U4u2y1`%kN+8iT7&j(17`>ic?{Q5JgbcF))C9RK;$?Sdn4dW(@2W9wm2*`BVoUO^l z&QkCL{>9v-1;_79WdbC(D~Y@9-Z9(saHAoPw4Z4tqT-}}WAC>@RK31(PSXr1M!O@8 zD0hn2A)>kl4I;JWMmVFvf45=*Y6~uL#u!R!A1r&KCshYFuifMO(y2FpS~gDX**zBY zQe#5yGmUiZF&_~{TNXWF=pp@M;=?t>PO9ocb|~ccS>o?A@j)^#WKXOuBd~}H@meJM z7RE;pP!E7)ac4mEl4WFHt8d(%Q00Ja-Hu&Xv|%fUbNV60A18v6)KFaIqX&MkP(TK= z3j*x8WL2&IfBW6pF9#EFK{k0$*W5Q~O$f|V{;?9$B@^B!Saz(bM<*?Sv0ZyZx2CU_ z*&zA~sXJQiscTROV<_Q>lC;zN!=R7?GGdTWBU26ac6+;La+@7p&CBED<0mDhy^07! zD%5XREnMX9?nLM={8yGfwz0wDMs=%b&%4d<5RxR zY9Dyz1bx4gGi9OWqvIj_EcntAyT&KE&W!=9@BIhO@nyjmQ<@NjD3-yLN+Tg}b%VQl zKHp{%U-4~DJM2;LaP*fBN+;X+?i;Dj@d2`1?WdrFWv?_Q&4rp1zt=Sf-|GsQ<(^@& znb5zWr!OgCh1}+0Mqx)qAxn&;dE>E0U3amT$UR3V8G!&pLxY(}2$^@7chFy7UiG`} zmE|n#@@Xv3%(y*Ol~-kyiWT3cALShF6)TGtSB*wE9=M^O&Yk%QL!^@mmDAT~OhDck zpRtVO_i`EsWT~x)xtHJh%U7$seh3@moYI6Vu_XDAa_-Vq8xANC#&lHpK{}=-0|Ms8 zCvJ&O59$M#UT=^CiQX>(7*p;J)sM%MpWyzP%nB-ssFy4%4OYXz(@8Qn*psa7a#Z<+Y{5C*CsA(GM-fb zpsY_SJmUQ(+@pAm&q>^oG)-fVd_@Sa^&-M5*7ax}M_W-A1MVZ6N#+J}V|nwq=-#qG z2P>QMI)u`QrC%>hlTrT4VfMZ(TWaDswpi?%Wew0`mWdu`G;$l|P?#?5gfg>uw5O6c zid4pS7c4~t#{+`2UsFD^t!ek5_YmX82hg6RRpC11HW^APhF{>FCSLTRPdd8 zW;---H;Kt38EXaED%9&k`r4<$btMp8>z%ett_9009KraGDzkLP`XVF@fP8o&Ru?4? z^5k7@A-LHqf-R}3IhB4*oB!m0ivk-M678GMtWfZ6K_QQC-96>qr zEBce`s>QWS3UE^AGivD^5uw+PcRwMQ)r@poLVDQ?Bcz+mI8|sy-;U8v)q8S!f;2KQ zpyAftY4Ryj^U17kC)B6?r?CYM-JU5+D+-To|8N1QY~Y>UE0;Z`n*|>A#Msuv7xmc< zDoNAq2)9VXTKQ%fs+|<&g?zH&`xYbM5~Nbo87o-RmPU@{|C-0QUcXVb7yvD&uzj2&j_d>kx0(7<&=H(=V#y#V6a8-Pm8*7^)*Y+llA@dCdEYT^QNZ$A+pI zUFT=LEJjq0KF_YP`&hK%>FHcV^l0XM?UUny5B}d=6y0@8<2n(qB{1G>ExAE?(fHQ{ z^W)&ur*k8TeAU@mO3ys_4`=4Gr4bLgcfVfz{M#GY)l2!S>$w{K*JlQDVw0x>4TlQ5 zSD&EjBp07YbJ4Tj8U#LBEIYCg7kJu(ZKDH~0`2{aP~f0F(>WH89Y8g7*`=_Osq_}% z+}y++VZ&(EOeE~(Gd_M80TICkgW}HB=wg@$Z>I?Ys7}PPmjjVU_uqP&F`o{fH9Oaj z$CW-1>y+im(J|1@iBAd1pBt>r@rhAJGNog-t0P<_2P@(DJqE)i##@5(b#zSi&W`;m%kw-70#<+L} zkidI>MHVAjdo`y+clo3!gyRMyLZ<=9o4HhnDuKm!@e}i99l&(l`KOcqDb~+GWgf=9 zONBe(>fS)I!oa)dgkj&YdwLs!pihdYXotqZ&Pb7j7oJN#jv~<=W7aIhJ1+@1fHI|5 z=IL?9lP!-khnW3{DfN{KOwpJjiuc2_Vkluj5HfCcZcxSv4(PeIg+YbG;*u z-}g5wbSMBHmVs$DDHAtoh9O6e>QA+bn5nUWj+&DNxc!8+0eP`ax{kZXwHDBnzD^lybRW}a9y*vRs~rl)Bh{Cp`^k+v>!-RNjUyjcvTh$Lg30-QX%EJCax%ShcTmj2 zosTQ&z1L}b;4~@VSs72=yzmnK6f0>M5OcMxgwVOShg`~n#*vPc9=eiet@&)lKmF#9 zWOfl5;jFt|cL$qVY7OE-hahroo}0N1Rm>K|4B#HffIIh_@pdsM;H z6LEO@fyZqs{;dUnL6E*G?s5$VU<0`B-JHm_haKIR)%kE_t?vtxncFDK&fFcn^>(7h1$+w-JSuS1N zhdThXq-xI!un|VfUTzriatQ3q`&FfECTaz;NZM^jRkRDq@3YM(V{2l0#P6P{_pk?x zPdDC}v>o1`-?0mIbdMJX8Y5PHWqY?I^vn2lTTmqaUFmHqvY4`1Qv2iG;!O_~t5uSc ze3FcLXfuHl2oz|OI#1(x!x0wrvj(WBakB1d z%(XP*VwSeo(p1kC1k2J0>mWHWS9@+-suQyV=kITLDKQZ7Em*Ez43lg&O0G_S6U&u; zym=2}ZqFq0(6NWjh)hv7=x22B^(Mn=!^@lvPqBU+fk4RRT+)rtK9Kt;tx<)1@}pXB zsZdt5x;Bn@dNM@y4`EKW@GDw+lDgxMm}632n}+n48)*eKPx%h&Dy@bI9~AE`#@%nO zYQjE6Tv+;P(Z36{s4ji_`S^(}iK%fI%{g}=<{s(t{bJ@vFU_@IZGrplFWX!#>)EZL z2Z`~cS{}_oSwAiaKZ9~p3S>kAHO&#Xudg-fvB>%>p*!`QnOB`zg2)ST=IXh*WaLXn zn^qvz-W4(Y+Cl6*RKfn)77i(W`srUlw}}Lhe>P9H-v8w4ZcbYN)OD zU4X20B7V+JFYz8cEYDZn-vzqaA2RorSD#QkzD(bMvTPR9TxKc;mTu2Q;g_jcd`=Cl z{dTx&c}{5;OP#%lcx2|M?-;@>KG{>O(VmVkYkltQ&Phvic}WcX*S`EIa!; z8Sj~nVph2Axt9v6vZ5m$bJ7T70N3*4NBpCpXUE=q%}y8(?+s;%#;=TRoj0$AHmeh9 z+Q~mN*{<_^Te7ZORN9ZtgWqm`Uc<|;Ak`~S$fz5()V6mEs@J#^j~8#qdG(E5erqYc zqiOvG_`6-^ z#3=I+mw&3^;Y$z~v6{G1`!`1Nk8Og{6S93y z1IKk{a}(aGy%{k1YmULCLAjMD*&J{C1_RkMn9Bn=Vh4Fj|JM#@Pv~@)+N4g{7CP)K zfIYjjS)Apfy_iZuJ_*D@+!3(tx{hV5WnPgeUkcCuC^Id71bpnGz;haxF>esIWA*FU zj{VQ>4;!sL&NfE0zcs%~YiGRx$4*lw4&36aU7BJi{?fq{Ody}hG`xNM&WoZsUinrm zUB5hunI+-L`Ikyeg`>rbQ>J|injd`0X^K}Oz{JbvYQ?uZ;DBGAab$BKmFLFa{#Yuf zg0=(Gau)*%dIdEN+lj|-;6eui>Gu``I$dY-Nqm%u<$~H9>^VG6U%w0J*0upZZ1lrc zuNI#*=O#I=mV7&8bWB(@7X0scU*tAIvA%X8UfdN-D{0ohP-{=GnB7F}iOy`$Pxa3) z6%O8Nr^|TfiFV>|{Pv8!be5h-A19>Tw%TDZW0MgkK(FjMFbJuu#UK+Nt*LaFcX`7&!r)xoUS9PgLKix1K&8?=?Sm0 zTah3!Uh>u?gy%8}k=v?sIlhfqGz9zgQ%numR|U34yOjD%^(evIIyIu3;7W&M&pegH zjhB7|Sg@Rkzuso;w?;R!I8BZbd-wG2nR!Z=(3$`GO?T{=@}{W3MSD9A4`L?~$!Nfz zuuYG2&-@FyV2|Kvja)hp{yJABlq#0TFNY_Rj>btni@)ZX=1M?8WgMli8*a9UP0GoawwAY}ERYTxrfBGuLw z$P+-QmTD@VUsVGtlHX(&F`AU<=2~a6PbyqbUwipB(O3_kKDWfu!^{n}C_UQOI@+2O2k?K%fqw@HSy7c5i+U(%Kq-6wv9)=V1Ptjs% zF@t)%aeU0YSKTuaK#7AYdtP30U489YzWbzB53@AJj8rnTJTes+Ct9Y{-L0>svM&q) ztc+Hq%@P0@zr&y9WM7`xBnJ`{Uc57psbw7}OonCayHf8FL_;$sbLroC9)LTlGIisx zb5M@Ms-!{kN$)-@{rp*JNytHPpK~HkT4~osa<~B_3%fz*h#s_GmQR{?j0fZjvp3lZ zeorr5UHK?GO~u7Bgf+q}9;kzRkLOL7c%1#G?hU<@Q^8?vsblI|oA5)IXLKOENL=Rl z9b+ZZ2!l~>@#MXzspDGP^kv0SBuGBPgc=e36px$W=$pjV9(4waifI$}xpYz$m8sq5 zrvwRzV6yP_{FU&*XBwL$ZEx^$BXweM>C;FjT}w18JjJm+#5L~4?I2>j>KV|)GrCDP z|59M{Oj$&FA6itm(JA$Vq;#5wiyrbg{427a4w8Y&HgFE(l>MK;{n?D6}nz`b>P-w2cuw?;FrGC=G}u;6?Xhsi*}pI7Y|&AG{Y6Lfftj{ zHgGidGG0PlT^0s&i=o7Cn~AHh9PV1(1DV5_^2E^p@OF}~EygL@3Z>)a8~%$Mm*&vDP%>hkab~;W zCSJU2zdemDs6)R!q@%E}vd2nk@#t;LdL?o^Ve+$hE2*`*=_60GB;iuFRF*1k13H~H z*{e6;EQcGO4|G1G$HLITwWeJ)>ZC%~w6&R@4Bcj2MDWUhSRqwE6GQTUK98Y;p+n+L zaw5@0RpTUoA|G}n0J9$7>UpQMhzIz*HwPJ%){p&jz`{c467)4pj7uqC8!5RB@Lqa^ zNo8ORBM#LQxKf~v>iMioCFj?uT`DKgNI#Mg{(stg?{KRB|8e-(NkkHv;n;i2Oj*Z# zo7qJ6CVQtsR*teqNra4KJDejTB-!(D63IM9_PU=(y+7aI_qy)yAHVy*`*+{2*Hza! z&)4hucsw6_@me9Frk=8|P4ay0eZ@rIglM0jU5!ErnxApgT!$cyhes8r9{@D8JGX;2 z%7g#n4)rKG4li{Ff>};o;+hY~1c74IvJ-4H?vf%hVwz2BpWvx>?w4|4r%01pf-kah z4cUpen4B*(uGrRB z+q#SD72FfUl+AbP-4)Ituz@-kh_#u+59jR$;vWKL9M<8lvz}R>%8I1SDVl7xVCf4w z(RO}Q;c?csg>2=p1Wey)7c-t9B%I{9*yCsjHM8(Y+QA^UW8PM#(maS_GXPof>m{#? z0cJ&6V31_UhsF!)JrUe85=jtyAFDff*ivJCQ90`~XKHfJnQT>XYF#!&s<_V4Ur`*U zpNyl&LWh#4)HNb%gnq62?Yws$0m| z*Pw4MQN~{nPLmthlGgialRH?z&|EUluErHGF2!g4UTr`LrcaH%k^k#`?!Y<&p6<8c zhj(E5J2+tw1>!vh-5NVEohO~FPt}dAako5@N=ZgcvAIt@3;m5v=}_k48=SiS?u}Bt zU=^8&k003_q2`M|))_I$^_TsWUIxp11s;SPFmeke#2%tNS|MLEEPV1sEEHxh(pDA) zszjv~U_*6NN%w3GtHmwL*S`-`o7MuB_9jpz{5L;(QRw}Ar9rJ&;;Tv3znP|drFw`e zGbGu==+@{1SYar z)VnX%tf{Kz%emXzR_8rNop#WsA9k$YVXqT_C`4O(5cj2Zye;IqetN8U3%WZdARubg zCOBXFc&Y?X_WHiT@^+tKCkKq(T?MgcA@QU&xv-UwmOceqqQx1==WucJUaPg4f5*gQ z1`sG2jiEbSk0<=5sx>bry9{F4o^b7aWo4N-VrR=d3bl3EedmGo&X{K+G>Ci+;LWNI z7v&q&yEAXJ`rI~DIZl#uu69ZFH;BYO%fGc1s0F;RF5=8W=1J>bQLByvH!K?Xauzx^ zm$6*61Uc7vl@9(6b|I+>s??BogJ@yCMOS6pMFxA4K_bpTiO)}B_de3MuLLpY*R>#F zv%cN4w2JWEou|K#4%jExiRE%VrjW5HqbyboSBLM-{J=BokdC$N6qet^k-EFmszzNs z^J0CocMyvF<^f0dG4`lW^OcS1JJPB!v*f_bP5A?rZZ7w)dBa0k(_W+$D%n1)zdsbI zC?;nro8>4LY#T8DWyC0_)+|@7txRc;Zj;`+#AB*{sO}E^@9S);10#{vqwWz*?XKIS zy`1jN$=^vCje29c6-nwn2mBSc2Bu^KZVDQWxm-fBVXkB7ZEwGhS+1KULDUq|acH~U znfIlxIb1DEXF9U@_BaCyTM%h8uNM1^gv#ZHv-DXnPoiO3w`vrsNH;t%-ycko2bOFq z+dRA*z3ww7ak|{(Zh8e}hSJ zIoWLqT_pLoeImaHywQ);1V7$0 zid>)Li=@4BmJIK7AhNwOK)-<$7aFM&_wsi+^5pNo0gLV=^o(m&m)no2a|)mfdbnq# zJ!$X9d07Z5vm=H>qzunDZtQ?zrOr{pUl(NH_V!!V6pmx=XO|8}$PE~#@9xeVEaJ}@ zV{ldmzS6b`Hs9^;vyHU_@w;wiFoqMi>orl*`)pQK28>qjJrD<7X( zW=_fx%N5tNYQTe00F7$RXX%~;>pEMR$n0F_+>}X6WAz8ijzCbU?z6k*6YuIwNpisP zBrw?>FD|d(-%aHrQ_u*_JhuB`ZuY|%z6bj1AP zKJ!7y@7tLCj*j%Rim6t*eg0hSYaBeQx2>uxiKhvVukFhp80AS|B;_*MV9`BwJ3k_8P`qKR3=I!WuQ5F*FZKw7k=0#f_j2FiW-X3msBKF zzzZeVzaXOb*{9>zbWH3Zc_oz7c#3sxQ5ciMx%UBPHf!p(znr3)npy`>^f0q9ya7)W zj?1u~+v!mDvnFo6gRpiSh0YW!s#}dUjEuO$pBzzp%WcEvz#qJ4T@VuVI2}JsR87 zw|;cTaKPX-P-+Cnq5LMsa~iW-(&4sYDspUWZ6O77@8{*cwIe`{7=&}C=w{T#{^W3n z)UoO+1m5)W>x1Si9hZaX2mUqnq@pRhs&eU$Tb-cR6aK!0u%X zV>c*{U?6+G&l#vG9!-~-ZTWQ9!G;Ob*g<%Ehh(r!VxnIIwYM@EoVkes)c^@e*GcxGM2< zsk@kI2?EC4BNDj%#Im;avn^#2*j4Bk?_6R5kZeYQxL{(mCXf496>(TU8u4Ljza1CPBMUL~crwHqaa+ zCK}Tq>ZXcbyVLQ=d9hG3K1wQ~I{)3(2@j^3vls*9*-BMD6gH^r0Lp@H>ftx!eGVuS zR|vkc4D3GQbeAvz5ZAJABl|P zMhelCUW5m6seWqH;TBD6sZ28wlJv*{2nismKsU%%L4Wk}5d@3pPn_mFu;@JJ^TRGK zbjNM0q}Mo z=Vjlm*r{Ep?$vtp#ZXR>tyW^U83z^s2jr8y-Gr5e1#TBCZg1-=?cs&Ipm|xC*R6^U z2f@lpYC5t+XrW46?+ig4`m!&KDRY~z06ey)9>&0bzTk>)^ z7bobnKCPk&N}M~2)Fm*dWVe>W-z5;_(ES~b$Pwn)`3zbaJLNK~0V>P&y|HT{Z`R0S ze~hbw_UTA%n0L&c5Mbv(CT7iGVxd}UFw}752Wm0Wbm376JOlGW+^P&8v~W3)glVqd zJYq9`b7vz~cjV)A^soQCa<-T_bD91?Z@GS!yFTp2al=kAN;9f^sm>Wy7P-ALi75W@ z7Uxqn&d1uql}2L0JvQ)4lkx< zdREx#b#+tmw1wgjA-ofG$)vPTVJKEqWX(eU2I%#SYHeJz(T_Tk5iZvynW2!IRLv)e zh4F{O!y_W1Hr3(8{ZE|3f?`fi6$+5xjW_sJbb+*0_6%#q__L)evdF%pOf{|5hL=MS zK)*r{9?`a=o|>-?S3K^LON*bloruEp#d-^OQkiypCxwP;UKEx#F-xaq5U$-tk!96b z(o0i~N2!m7H?9l-iGx`h-hUX-2g-vWoGaY%z4%lGE}pTjz2YakTK_fn6d*MG)7D-V37TYxg$)66L#f!gT$J6bb+k3YSh&U6cGz73|B`8`u&jIE zOv%6eFhH_5Q9UngeJ`%(^D11lhAZ+Y?h3_3U3M%z0wBPfE&+0Ak(#{64Jq3Cvn zQw~!=vFC>f;cc}ZVUjks@mrgpP z07^TgjgP;oZtCC*YJ^#jK5iJcK9?jZDlEG$!BnkNGn<#6%h_wet`W~7%80Z1YJwCE zaYBCG=k6K$pfLi4nZdz6zuD(tjh9Ty5_esnaRo%^Fg|jv0(vQKbq7@F_u!7PP1%~s zgFj!O5=+pDsEhIm@A=E{t??_^k&I!in=?nwYy*HVP}s>w;L|%r;ol+jx?lZW%{rj&B^Z&PbxY$^KaY&d6?j*bIjMg z0G;Ty(ps>=nY}5!O{!jc_qCDN5w!axUP4~nRbbAW6$bw}Tw?{)^xesD zQTaJ)VU?c=v0YKJ%i-2U(ia;@8ZM1fXINoQaCQ%!JIdfcF60^!Fl3B7c_FE@JVqEJ zSfIZ3dEved-%}R(cP?IV2^jC!){c)GvG@?6iC0^ z>sH3Hql%+ydoTX;2GnU_pE6ZK^F|}lQ>(rdVylh+ zm2Jj6pR*##G+a5`D6$ft%%B;&D|!n@Nf!|JKJJ59k>^B{&ct;RVt;{%LcyQ7jv&v- zLmJNOK)*q-)jvdpqCqvH6eB5BQD`XUMg9KoVlKEsD{dXZQPH&%EJk3pc<$yG#obQyUIUIm((&z9dA}i_&}P zEhjNGqS3OOORW?yfa;7d9bb3Y9f@!90c2A~Ll;q+kz){Z0Qh!w)}2DYq=r>n$E=U=~r>D3Z^~c$I{syr4h?q)rOwBNSeb zji6IA47Ks85s;6kg;ikirD?{f0j}AwA7Pnu4cLb&rp=%SR3a&6&0lp^a-UM*tVY(a zLy0y%^3tiHU@;j!NBGD}3$=-%lQQTgVojaHd}B^LU@IeJ2`b>;DyQ|ot~)W8v|3z+woS~_XLlVfe0V1_VBqM7_~g;rdDX1K@= zx(@2B>9xpOlnKxj=)H$`dD{n2?j7l9u}J&unlR=KbKCi$IFu2ZLI+;RDTv3+rK8|6 zp@a-5+_|Yjf_tV!ab&Qd&6j3#`m|v6mPXyp4em{oY5Y~)FIp6O*#YdBUexsP&nuG* zYkV;M`mn@(x2CHm%;V(JjQf%sDi{Ihke@~+CqpQD`9wp2@X5OPoQe0WF}aca`LN#1 zbSFw;zH*O#LpTU1ln#-@^Q*6+F!u5b(<-5=eUAH%#&uuVD@-P+1Nq&L9$as8`M7ld ziaD?n3e*UKFN9bOn5Ub*I#)fi9V^s6;d%y|*(gnocrAG0K)KdwP##B6Kf5aS^zsI6 z!OzgJk9%Dv^jc^ULIl1vQA2K^yaFso`HX&Jb?Y`Tr|!c4nKVk&#&;m-{r-y@rXOK; zai4F~G@QkYx}CZ)-o!^7`x@n9PgoXJFXQcVuSRcvq8cI%mo5>6!4H%FYZeY#(xjjg znlwxF4wjP>D&{4*0SO9vvJ&ZucemKRIGySp@njP!cb=nyf{XJko8k+;KfVr|;?jnz z)|Li*IQf4WBZ7^HM~6QAmM-Na_Tu~9SMLbqQV-n*M}CQ?Xc3&MASwijjhzSmLojW_ zyKiv~0k{l^Xa1e)ft41abkbBAml%iN04;`V z*nOg4o}xHB5iYm33olLsR><}eRH zS>S=Tp;UT`=K*tNq zgI%qdlWUZu-+7!ft?&N>urnKcsa-8e9hkMB{S(g`+Pb=$5pI>?5%TERQ}N+jYGcm} zDI}fyDPDZs)nE7Qu6f7>N zQ$6ZfRQzTvtjs22?o%*r)fRVszfQe<+|&oNRjg|X+L0^)ytIBp18!*9MV0>}y?UMz zpPd|!t9Y08J0m`C=OD^r&aYM}zHUb$=P(DI_!4A-86a5^+rEDEcFK)#<)uu_{*l#W zp|&LbD^T_MbZX3XR&6MO>pzxOOrX{}91n;(M7o{Qh;%g8e&T2E-H_B~oQR7ansQSl zYmnUuGN=dLXqX)+BK9Xpw8>_+@pTIiMt96gmBX@m=9AlpuN^UHtmW}Tn`S_58U{+> z;Foql<{`L2Ja>P`19dm*XvJf;JV#gpHOl3(F&cTtye3ykVSEnTLFfwD6p?`uK$&ov zPQBm|`p~=lXAYRABA^DyEiP?!ON;Kv-1fmdOXTaQ z+^U}yO<4DN_8?k(>%`g z^c?&tywzoweC=pTqE@!?*{@*SNFW%wt7g@Um(u(_DzrzCPvVe9lLNdVNUrfZ%?j?3 zq%pqmrqUZ)ep6#zaF5E1Ryd}uqAld&1|CLv4zta_C3@*j$AYbOnQH{T7wg6^?xQRf zEi-&CRa#qIoLM@jiYf1)@iNDAg>1)CA?^;H?p*=?y@%fQ;ohKLodDYv1Z|>xPL$tG zFF}#5-Ai-t57{@wQ~0fUYL8f}$f_v3{|Ua~rLaL*#kg^?UXu|D>hy z{lZ0{STt)eOusz|h$IdD@=21oWn&d2qtz3ncn!RDT|R`RO73V1ITk9s>C2EY6O*8V(Wv7M=+w%zxWMSiJKB2og0)A*sn0)H@p*F1U;yg^_nToNJJ^Q8$6s;N zs#7D4nSP3jdS+Mut{qOATtSriH(2I|#LAf}7&_eecVVosngAED;;zI#a94Km8}!=W z)j~l!&A2Cgv0HfTnk3?rzy|&y!^8Ipm@v`ubUsnazhCJ)c8WgS)4(jCW}~d}_V9F5 zbuZSjR0Sq%=E!u0|GS+e;`&=ZXT6Mmi~{Ffs)#Y69txb%1lJpx#Lv`g=KeQ~B*DoQ ziTsIY47~Vz!pJf#{3-uFemDiti;p6Rx(wk<#PWq!8y7L_h)`+rJ2!ccx;PmhPAs-}<{a3Fjls6_rPk2BO@>tAZ>M!w>^yu8T29dl1jmxC{*1GdIi3AF2D`9 zfXK-4f;8Q}kk<{7;JteoMN~A(4FkcK^O`VuVZ~a{5s=(C zvw<>((K8(Kh7tym4B!qh0qOe>NKC2>13m(9rE=5=qAFsX5)V&n#>EYJ5VfKm@_)om zB32;H4|s4)#-%A>`6J)pp(E;{IVVd=W9{JKEG~qrG_b)X+=Ie6 zTMRG!`DCzZ^d3@Y3P)5X+@um5DiNS&rdZq_-p};kw@`eulO4%}tJ|!;2v?}YMY!-n zkiecJ_88TNB1S2>?*>VU%MY2(6%@CId7L(Q*0=ZE4-{R`t)17)jNsx;#3;-j{s zQh`ZI4v-Q*CoB+FCu2N8VwWE#NsyvlnGJkPm18C(H;{yL624d#XVb0moQN{aWDg8+ zw8WSdK<_w+YzSfKZ74Rlt0+yLa!*vzi%bS=H!0-f%|H-zfdx}af33)%vJ_E6sx-UE zW$8=fkc?@EGZ;!0b+=n>>y(_1S8i|}S3x-kgrY)}K9i55U9m)I{FojqQieAWuZn&R zd(rlvI|w~lft;WS-`_RIY);E6oaIy;TLD| zpIjWDrBz$~D{zB^xP>Im75S(SQ8Zl*`Swv*Jghb{A|eNuvEr^$@QVU1-V7U=zqgrD zcR(h65=l$RLF)iXjiYK!jVw`Kn<0-Jqw1dSiVP?|2_zvAjQzitBJr3@5cmwOpK8d`g7X{ew@I7mmc8 zsDEg8kMadqeZ}+fD*>Ym0FpD0?MeR1g|l!5X07Ec?O&Tz*k(gM zGZTUzSh#xQ&taiP`J^*5U<}{nKkm4+{|fmZJmLS7(J_s#SL}OtbPzXK@E7v%7=Te4wt1ja?D8d8^b6L(%FtAD~8CF6f-!=6owxb>0;CvS+-&x^OnZ zl$RdEP{8HMV(>H<1XU-<3$iArNNTIrkMFS9$)S<|l zucrUy2b|PO`$4773JXDh&BP)pW|95yug1zdKDU>u7~LerrPf zl8k>J3Bx^dLL7K$Ja3mvYlqER%BefW)Q}*TX^ylYTs1QT zEp^;>(Vsy2gOkExH5XLnU^TH`asUSi6Y^0l3y(oP19ubXKURu2H4pubN_PafOhEvF zTlyc3&%fT_f79P<^TQDH)UepyQAoF}nk+iY_u zrlr+99V>x);B-I4c?6gpohY?fTh9)d{xM1?#~?D7nWQT9{2|dO!ph*a!JdD^gGF3tG~7>#IQx=^VobB!bw`pRSiA);0a3lksy z!9cf*p)?IiftH#^yG%CjHkUD-aJ%dAF;Pc76u0$v)?gN$Uj6`iqKrF;yt_iUlTmP4 z^{PE*iWS%4yHaVd+vH4n{3uX9r24i|J;A2gsV463hO|I=Pc#1;yfNu)|0ZVtFN(gt zM^U+!^e{ty;_03rv{lOWA31Wl$CdWv#IxBF+@S)l&*UJ(kukUot4Ex074n5wma+u0 z0ukbty-)(t#FHf$0&Bb|2UIY~gK2DT+Nd$|7q(RiRN*WrkqOgy>3@v>AL0E!iSr~N7f>)W|%4L z$rQc~c%T^;Gg#wuPz5dKY*gqDgGq8V@^lBm@>1>>og{y}C{=EuAlz*V^U(22_I8Bx zr6Sn6lMMD5zv-IqHS6m7F|E&U_rpU)y3JGHx!3o8yKCO)cImUeaG#tw1!3B|N@!;8 zhTmzV{uh5h-(0 ze>Etuw9@t~=RxO#Cere0Y-^qL=_H=}^8>`qZk^PC>xV9`s}BD~>csZJmSq&7_Iq-HTSP@af$3F^&^kM;<45@2qZskIi_`^axDyP4*At&FW`a z^IJ43UuN!1%Dj&a(P&0L8eI?Y)DGBLuUc7iR*~(`e~A=U|8J~KoM~~Opr;Hl4Le>> zIBmMq$dw;+SoJLHr*Gxg(>DQ}L3fUA9XBPq7yLeL#?#kKG$!?P2Hoasc{BiEd3@Te zOmM7)`L`0xB(=!x396l&p6c5cRxjq;WOnoXn=Zg6n`Z^i z_Wk=}#~1G(A6=t+Q*I$ie4i)7ftj8Jzqb=_$qMDsjj60ow=SQqx*;FEDbcs6-*{%Z z_~ah$EZ!1*3$MkjDy+;p6a!Atk5aOJZY`-y4>8n)*ij+aftgJA2eJ5F}1>7b1J;`!DE%cy| zXP5Sgb5=i-7_6e~N3T4qRELoNwXufA>5i|N&6VQ{0<~oA;(ZHcVpbmErLl_l^THk_ zNn!(p_$Px;A~RaxZY6f%FI-$nobLQ(RpPl8zF;3}2@G>qUgc3))q}NK9yiafG#Og! zMn4wSQ3yat_i0g%Qx8(U)K z-^5zx&R>T~Uh9fL?&!8_j2a6v5L92)V!xb75M}@7IQWNWLhOk$)s!jT?qfsi^l{b^ z>w#UEj{{YJS}B+Q$J=yV`KimaL0qT%d3FqQw6gJz(-zG3G0?^P>xq>#$^4JJeS!?A znwylSxM((!vdId!OV*CimZAY1p2)GVlHIUJlC>{-amwbtcurP15j5G@+ZaV?w^3$@ zSg;kbV#-zr+j{EuI!}V6heD;!j#186>%b7-bE))B$nJg32%j(d(h{y%`Pl#7b8p{F zG^UX;OEG&OuBIbc)AM>*yAF8t);>~v! zKjwY1_nJLqTW z4B=OiEo=%T4<`TMS(^O?dZ1T2V)@#Qr-yduxOuEH^;mlo+s;ta-y@N4q0`yFSj-s| zVHvJcszh6J@0D+3ekQb1XKS{^SQy%U2X@Lo0-e(3a!P&S{0b|!&fTM3=fq87zb6^% zxIoX9KGifUFFXTyid>b14V zP9#VNar`^wN`W3n@przCdZ!N2`zuekUXl0Xr2_E?(5kgIfLr9vZyPL8X*s}MeVr}g z5!zdOOA3X{)zB3Kc^mf?L@-I8HoRAnw@2GJY|Z@nS@HLqHqYzU17vKTWkuoY_g7HtgCI$boN)oPj&hczE@L_sAp}E5A!i+3h@kWf4*BI z;V%X(v&mbW`Klvx9c(dihxy&PnbbwQQ%SFXmcX{oL%nlVHQQ6+@()^BGG~Rx@`;k*{jwpXzeKr>79H7GS+l4Rx2$+b+ z;`tV>eC?H=+O5=Tz6EgCdRyjoPve;($H+@|e?&Wjh-B|EZ2kIxs^h07MH68G_e698 zuuVm|F`ZGUryoM#0fm!fi+4+mheIVstVi75UWR<>vNO>BeJVV1Ykl3&x;r%dozMO6 zHQvZ(nGt2|rxN2l#f+)U}0zo~k=I!Q7wa9jP=%WcQrsY)e5aIsWIg~D}mzhilUoi}lI%f8`=#S?U@T!ssN zgah6Y0dChXL(31i6g*+oYIV8u3=(ZqC1)f~u%w8W3cF8;mM`PB`1UYtFgHQY@!=t#_rtwsQ1g|#)-0Dd>&wvWRK|Q4 z_;G*yc9_7FM|<5jY6P~df3F*_YwS%keo#NrGBs`kVcnxQvQ$A^ggbX939(x6>=l0x zZHf0&?|HhjpGF{GNiS-KOqzZso!L=QWR{1n+xdMlBKyV#iL~##GYWy2TtZr`Qn-;& zXl^}fIzBZF8KdwqZ93D|&2cDJm99TM(1nG|K-bxK`tKGU=HLLLuI!{4T6aYV`|c&I z$q9qtcD3`cg&}@Vr+jwGY#41eJX`yzHP5)_sDgMC%RMJs_VlP66XsKDR5q)-b*q%7 zagi%%`B~lBTc^cmd3G&)fvGO=`t8}G>8Zf+i!-$818n z>)(5g?6#@PXPa(tP5O>=p2pAbUy6_(r&4i4#=OgwEC|o0M#NlOn;E=uTK9mfc_7`Pkpx^6GdGgxk4g0>ECB)u%%^4`W7#IMrPyV$+l-X>gyfn;KF}t{A z6yA!?3cHFssm*&iUNmV-{iWC?skW)jF9oNXMFVYSj)VS^Bc;Pi%$6NtrGdyC@ijzv zj*HRC*AnBg_54O^q;=1eF#57L6Ah10L>~FIB7y z^g$aNVng)$HznlPj#ZlY?25r3E6Sd)fHq%pDqdX~kndvFW0Wnqhcu@36LdEr!EI~; zo$4jT&U^}L6@#vG`F-gZ-TVHjCAElz+}UoAZM`cjDgWEhXc5%0c(Q+gUXP_v@+%=$ zyHB(a@jcbMHU9&XobZM&jv+0ZfCMYGw({kBh=GjXFn zgTJ2;FY?waj@K`aYyGUax*mqM%8wW~zM{<)SPz}Jr)(|6JGaL(Beb_3_Aypc!F~0$kHfjU`$Lix zMMs{_oM~+BKw~7NXK#!Z2b)X_FPt_d^YoBhUGO&kF{Af30U=f8uJK!> z@?8#aM6+Oo?8yA`4Mi8P3iJ3aB?8-ZjjePTbDoQ+pJu55TisQWew{6r5@Q+Nxguhb zh0#kSGz)6i7Oa!Jzc$Bjfj9sU@(&nOgg~Z(Z*<%H9bC_pJqT29qRw(mb}A0|g*A$o$D{zuLN;Z734YY2qy zGocU!@;e}0FC1J(dZ)*+l37|cX-6|2*Xc%PqhYN%^P8}6Gi1FZoW0A93wBzsq;)fV zT=b=!3(Ktj^`o-?!s`T`F;DAvuz$Fy1rHcm=uLHF9_VDJ&vC>S9H&OMiR{dV=tX$irDyv@WAK_XAS-2yJZHR_)9S4CS|e*i8FS zt{hsZdzQw#lHuDZk;<157<0ahZHNI}ife*nWb`7#N{05*I@PoYP+2#4<(n*&TK!vvYUlf*y<>FwOlwjOjx!JtcV?ERQz zED`P`CRfE^lF$P~c(Z=FOe;?}$)KAj8fu#EqJJ#r$B;YHCI!7I(4BH!Zc|3rB@C>k{LwBp(=hlvb@{%^;_KDX#^$C8YOb?sde)V8#>9z(u>Wf|h%i z(~rtzetrNzk%g}g&mEjE z8s|7`Z*X=u`ZXu>&dAn=W3LP31Jwtz@+tEtPG$c;^Yf1IjKl}R!}{G}9~e{!1g4^+ KTyfjx;r|8d1Hc3T literal 82170 zcmeFYhgVZuw?0hhMY@1=1w=Z6^ePxcq>1#7Akv$30s#~$f*?|*D$;umC6p*lL5P6V z5C|ZK7CNB>$S?2ly!V_tzCYo+jLq1|-Xm-8IoF!cY|lz~WS~t;%|T5>L`19e0Qi`Q zh%AJNh{TbSoNxt^`Ei$UB6rl)1`^JMztelrNWzKA_kpE95fOLXfBpd}t9#BwMAwOQ zfEp$td0WWf?5ED}kN14&?>q%)kg@>$HMPE%dyw`=4RySRdawre2ai0-kd30|d(B8v zR3w@WBYN@T#q}3LFRs6P31f(5kNpMZ=ZA2Y=zGTiC^|dN%L`NRx2W z7HpWlsQc&o%3%l}YLfJO&Nj{4zZLu@1yJ|*M_OSLyT1XL<1u;X-%u=5d?n=OpUZ9! zMhdoR7_0%u#1M*YjhOAKC}{L6m@RCr**5>rAjRy%QNdwbfKJvBZr>dVL^81pV34GT zG*fT8D;;a5{d#VIby&v||6 zSg9W}%W<}mxdN)9k4fl><<|Z1mau!#{I8vwqYk1|qwUv!Dd)6Z1{)r~`lC7hhLNZ; zKph|T36}49;a)aLII)Gsd$!3p(5O~@luFtQmLq)+j;0-L3>?>&_-BnMN>KQtLPBbf;WDvx@dDPX7i^7K#r?nfs^^Ah&}wri{L zikL;QLaHkv6fjm=(>6-W{!JiN9t`#uZNC zef`C_UDI1=bZ)|F)_5ZekP zbvIT+1wi6RwnJp`Sy^`5^4v!K!%B?^*q|ACudZ9G)12HW5XV)-7t&9Z{AYQ|u-uOt zWHRm8xhWDUW#f?K01;Na{aeWNA1irQB=7nUNovI!Li3H`V_{T z7`Seu>^xXV;c2J`!Ja#YGb?0Kf&glCTq&lVkL@ot&NX+GKCECYxOV0M+&&_xg0pzQT+I6!dxe6fJrq&uF4@Un6cYM(6nGL`XY+O3dA-KEcNG!7bft=vcd#R zGPxg*e~@&{rpR09bi0P9lpxz*BIF1^HuGmQiaC}U>Tc;ux|%&@xYAI)y9B~Q%9Y}&7X4ltRsjfF0`$XM= zcK~92!fB+6*o<*;J8yzxHt$P*+-rfI_G3smcn$<%N3&|W36QIIq6s6;$@Oi$j$_1d zcz-y*TB7wmGlXte3$+U5jH4yo{_~#hP0_8y@w&+gRwsYIsrN11bI{td5gB3|_8nbIg5qCB0e zh@;0kQ1Z6&wg@(7p;@0OgwHGQ9xvYRM+a>^wu)1lO0Cbc-N`#iZi6}`@BTr(4M!RW zc?W9;9^X5RR+^-lk~gfx<_jsW_y}sw---7<_Uqqgbf9p2-fTAl;IW@yC^#t%lyR2r2I6ebzGM%7U+>B#SUi#0?!!Me*0Dv^ z$rA$b<#k|X9ICxh=`24Cu?5WQn3Qaja^=-nliCoSUuh$od0B4O<+2)1&sS-kTj_qw zk(F^{RAjrXa+Y;uvV*xLS$feLc(|=Ee$88;GT+`nW#O1Xqg_CsV0(X_a|`t}tV*TE z>`N`C5n~uF8m5AMYQBfFloJt4nCa>|vewa=_W|JLJ)v9&4 ztJ=Mkoj(9PJFErN5@!a?dN+rhGgIYEleC&<+ERqm&6KoeInQ0&eR%I{l=*+;D}FPb zs#iUJL_T1)jAiOemapy#&BFzHw7g9r`$;j8KY)Z$MDIL`opD3rvFGl3>Ld$*?8yk` zKR&nOmVbQ0$c7cMCvB%PG?I4rM$wzQY0w0s+o5;;w!$r2sV1ycJtux$uFOm7Z$r{4 zi8#vw^{nAL z@&Q)bzoyXZ+Wjc}jW6?^Wm}Ku5LDIVmk&Wc$jKxx<=xgTto8bQPmro}RcwM@ftl*> z>&bWCuPng~VAD;`S*DOvgEBU+*aod$g1Zs@OX%Wx%3TIjD8WgoUdz4lhT+CaS;EE` zb5t2g3jd3_`BIo&hLh)z#VzMnbJjs0>BPpy)>>|$`eRdHbwx{RV~Qs_Sx1|%3y0$q z5WfLF{F=T|+HC4sZ0eHiKLT|gi;MmVigK)uNw5?Plw5aBqNiRJSu-&QBVjW`W6DuZ zgpfm<$mQy=Im;mVOjcu9FksgbbuFx}K0gcD&1^BT0>0x{On1^uW!ioq*Y3rGbFM znJ$J91}7X-+1S-HtkS*+y8X%`sR#ML3~E1#Xz}JH){g8vYHcJ(6Sf(>$T-=dU&A~# zS#7Opl456W2`2Nk>-FREb=~U-m$07F9ey1MRG4qRoEzGlBVHKZl|p62zWp0<)6lYg zi^$=Gq4}$q`Bzjb8!;0s6C-9&%v>J5T&_bSR5xcSO9VWACA?s=coRh!|4 zl{c>ge#md?k=f~JL8tD5P!bVb*t>`8Bt-&Yucwzo?#sV-hM^13*>AD7sMZ}N=AoRY zZ=-DsC}OKw`XyQQmya3u4HMgSMzLqCT9c%Rde@Oa_yny$B)j+d6y~K6;G6Yi`7{9Q zq#9x~m-W+mRs{iTy9fz5aRJR1xebX5k;RpHz2>=V0I+V%K0l1u0knLC&sx4_|I{;H zQYY=p-zKL&ubc1#qD2a}Cmj{qF1&{`KlR}g)Zz4jMpp@0t1EP*(8RkV>{{_VTY(oL z4W3fF*ezMG0c1q0ztOoL9?1Z>xZSArQYPl_P=_;-(Lvea0Kns~i&9NjeJ)8(Czcsm zX93Auz_kT&wwfl(7f7Z0w!3}*#q&8a>Ucs$(Kaw#7!=_GxMenlk1cOo$SI%8&#Hm4 zgfR3D|Kll(X=goAEkL2HN=fTJpv52)02#7na5!VD9r2*8^qHmXye&;9bBJD$evg-e z9f$C|t>NKmSM@NqD=8s%bh5!6HpuyN?Z{7p+)lIELE)MLbnwEPXnWX~5Iy~Ok-6PF zqC`EPyVvI2Y@VA;JV=y_4f)Qv?19@{7e5^CUFeANliH1TL@E_Y`bcW5Hbv>bz9>^x z;GRmgOX-I_X^td%bgaZp7yi!mTYdcuSR<%9{~GvTHsKYIExMwPW@dJ|NIyBd$c=I# zvHx&BcljQcHzajG3iK$phacugFu~_TmsN$Nc7XXWGeixAejs{8o}~>Ou%d7Yys!Qd zTc6aFKu=6$XBLXNG$1YT5fC)03X@)qBSTMD^WOjW16?7XK#w|DPvnfKK-vvk2SE6w zIY|$3<}<{iW=u#Z1C-zQi64Gk0=+30hM~`WF$IS3!rQ8z&3>>})a%cN&rnTG5o;5J zm_J5HYl4(xjZjrE?`8z!38^zExdkTSdF@I_$&IJ5zYurst*35egqt~_kpD|ZAdE$J zbMtI#+cs|pI|LDTw8)V?jD5SfLj2-ZuJDumgP(Gw$E3BTstPJ<&nqfEspoD@>SjzHe`?Z?ei<$5ksQi4HXMqU7`*nsv`g82A4HiIwp>e8 zU^k0>lK$CIk~yxXSXJPhX~Z<&c(T-C^=f|Ebi><(BMiocxHXh}ss>ZxaAl=(w}1?EvuQ+B3xkvux6l1a0_!*9xEy18gIo*h=E2 z*}2i+G-vKz)uJv9^%`0wcEgkW0=ObKg#BhM8z0QxX#SF|$aK&aWX+6(zkZb!5A zP)|>ug>_QQdeOO%Ga7BAwo_PorFc|_uygu7@25&Fl{kb+ zW*tQ`1z5-AOLy6@q8Wp@w4ft2$2qWlmT73SOI~X5Q4T4_bXKh@;Vu4<$}a0?#^271 zTZ;vK6|S6E%_5r6aXjQ-#p z=?TfM;c|FuG*wZ*d-*`8Yd6x;bL((mwlfx(>c*%cq zKqzMXM?2$>Kq|h{MqRT*1Foy!#O&%#W*8VqWMgYDKPo6Vb7qL*?ptiw1kDHEw!>D^X@n?c(g&840xesg{{)j%vY7@B ziop}22_yms{tI3;4!^6u!tQ#~I#rn&k^~81PZ^Of zp~DGwouLY?9gQOo324a(Gh7pgaXoeRHjZiwcsDu*&)%q90&5?&Y{`t)geF_3PeR1` zV1-)m&68ZjNEecqwnP3E_BWIGvg{o%@u4E4;}`(5lu|l#+Co(>btK)I?BzAr&z8Y0 zZIL&n0CSDYm-CL6vhh&LwwksJ!=iVG1Iw3=aD{vHDAdLxU$E zNADk17r#La$M?{nNuVZEUi4gA=nqGpd@&#kzevX0%E|+Bun}%7eDElpr87Ap7ioRBlF1<{b=0bF_}m*CE4IGHIUgIZH=fA&b;srB-CZWGMZ|9)nt8Vt=4_-2zoR}&AdHLIITi^+~$jr?c9LjS_sK@D$?HS+=kxM=sbqf z(;rU3lZM0t#^Mx11NQvAgRbZ>{YlWr2gHtKyQpTUPoyxW88ok*dv1!{rLbHIquC`z zSu?9_*S~22)~)QBEY%XJ86L~AqLWr;hNIv=>{gWvsAdKZ!ha&wV}k63G9$V51x5df z1~}{Dm?N1-$s~%eFj@l*7cj3e_g(26(M%Z5edo=D4Wz;61)+M1-4Y1|Eut;_T{5)x z*(^&2`VP`A580i)@M4aasFkhN7#Gtc>#3{`y~tAI9W06+9OwIjT)f)4_`uE3alhTn z&pPd2D^o^Aj=y^dtl+ESE4KRt<##yE=irZ0{b|^pLA+r26w%XVA(#4kv_y8e*2TBW zBhE{&nE;Umt?q`uI=};q6yBoVaLittE=R30J;?sv>1WtWO3*X(iC~WBzZT!U_U1q%B5ERn>WxuBU+;X4a{l!(yn;K zR7wiO%@P1?)hxbB7EyOoM9|zca}|JC5g80+asB|wru2n;>}vJ4 zdzCvUUo8f<`>}K;KOptc=f*RR&BkfRTLajS`r=o%i&RfKKc)BwP#fafr(`6qTM2|0 zJy)HiCo>~4XKOtOls5w3t|tmnAA5FVaho2tE{iVE?UX;(H2(Gv{K$mLkr-8x_Zf7?Sa@Z|AfjXQQD8qCMEdhBm{cIGRx{iw1Sd2n-Ez6 z?cReksXaWK(Kr_PDL@t>{?(hl=v`y@+eS`+^&;>meSmJ$+l_uEwihlaVMq8$sBZ^T(;3Pp^)pUD~0^*p*Dap;NuNGYD3z`|G>W;+F$L5 zJJaHh5NLVTQvdiLGU9axAI!S8H$rQxZ6;b^q8zoix+2wSJ++(p*Gsf_-TBf&#BvB7 zg#QV|(^l-`ds@widWt|lWUWpux$)Ou%+G&{?v1~FvXex+WyU-SJynEMajnEVBIy%$ zkF#i}-%46jeBWG>GChu@9OG>}YZK(Cr|W(2am~ez?s;xz%9YSXpZ{_Jg6HpB($33N zAYa5zL$_Oc8i$=c7;o%r#8?tq67s}5{DRcs&Jg6C_4!CEWB9)&PSJq>azQD%;mNE^ zLS6R``4&ldKK89xorpe5`_5=`ge2cv%*|gos~}wR)-`&v%iL{%M~k^TstaDFC z!`p1qT0{_%c<>2VplNroOG@CAy#t8nMV$z||GG$}G0RR^SXF%uNWYeEA?0-c2>2>! z?)S@|AcKa>@z|pryMNj6XG9LA+qAKPS>~Krc=MA&ZHBbRvSgtkHiRle+eZt?JIc7@ zTY|lFS=r&@i;{)&Uf>2VW67yN;LTlG+D+wGk!&Q>N9~19o+oD?hZ^lfOAd~|uQvCf z_q&JCMg}>M?m?&4MvVHyL*Lr&sU5z~)DZO~rYGK}?qhx-BTD+}|Kuj)DAmjp>&>va zd-c#LapkR6&XLKgLSnGysqAHF-&4<~#0Yv)Z?Bowi8g>jYvUHvvgpwTm&=cfFz))| zK}22w6sND3XfRv;Y|I&?0^*rtF-=FLQ z+1?@a{QjG6czHs|c?=Hgkl=TtdwR=0H72ULA~z#xzOOg#Ae-a7kpnESgxQOq{w=jF znj95H2M>mbwtGYnCbxb)|CwUcLD8Up;WN=us`E4R>1JhDfDdu4g!hgvjx=;r$XYGQ zY!uIr3Y`t=iv2p-K(3G(f^Q*o-twue88COPU@M8vl3#gx2Y=}OtAfh;(h^pc@NjZt zdR&rhK{w{{%Uuf;an8$z-1yz6#m^QMW#S@Evvm(qT{o;D1#1=7w9j&1EStjkmaDMp zy*qgKaxrv3M9LYy=L&s?0^qu@e1&LyR zpM@h>tFSs(=v~MKAL@~VBucrz(T6z$?S^MgSkxZ*S0K{-5y5^!;tXrfTuTIz)vH*8 z`Nsx>_~?qSSR6d6J!3h!8XG|y;ts*9$s$ZsQYn7kn&6>aw3|ggL*>g3x^iTZz{?d4 z)TYL)DQ)F1^N`-u*X4`Hd{PkJxbGBn#I<7EXjAC=={fmsmgi~CP=meI{HV2i@o_(W z!W0$FYh37N^3SYOfq^;plfbOO7N!%*Sm>Snf6eCCpyKN1U9=&hm@Z9ue&V{}*It>- zlX3j8uBbSsjMssa5L5W4E6 z>p-#?X6UMuMI{*5K1!HNdXvG8ScE@L0jy@mwB+6L1$%#p>dfxS*3Vrfxi^!r!33)F zaZ)cAh9};0zcKxA4_B;Mkyr0zI&Z*t(c~`OxC3O>=No!DBlz^T=P-sqWLXRQY9B zFiB5QH%C}$3g_Ud&1#Ad8RJGNx@i5Ha9>ZX>%m^8C%!b|egZvtGHHRXUQ*`So}G@R ze!&7o+*Nh9iQGCCL}d9OM)uCX-)cjvZ&8`h*CfQ5hRu8-mV5pbeDOmBO9CGG*uCj^ zM+vXom!lMR#ZiweS*dljh+H$oPie$_pb4A#ErY5BBJm#*}+Xz<>ki~7Wm&Z zZ~4INIkC1hTx92let452eQlm{^mK?*5bmzyb6DA9h~C%1BiIf*=XT0tyqYz9hMod! zVA4<~vSi#e_){b^xJhx9g2!;}3221t9|&vysPz5>geIA7L=(9f@kW`x>`04Ri8?01 z-^T%scx&@E&peS@eSOCH-F%BaT3aSbMtp5sWj)@N#Q9quP(a9okxiL`!IO|7Inot& z+|{~M#{P>=V)?G$Jw^UxM{i(D{g~mzl?X_jf1SWwQt@k~%G2$VoA-<}PBF0KXfz3) zOiV13fH@-X3%zjG9XN{y%wj!yyZ)cDZ87Z_rL+A{+i3lSZ)J}kw2+@L`%+6SSxJ6g z_y)rX-5Wr&q30y5A+eOfJ@)hV`pv;>H$(E090Dv;5VpGbh?nNlGatgZecj>$k+qWa zCdSq;d5I#cL&3N9!+9AcmFU1{)Jf_&er z!Xv3XjcpK-&ci(qUFwDl7SVyTX8#NS;vpqh2Jc7lvK^ZPct7U3dfmyn)v1(To_Vh3 zyxS6$4Xo5M!4v@51xp*=HdY=$VcML@>a(9}Uj(RpepVK7e=zjBM)LXjEmrjjkA2K? zAlR|Dn~exxZrL3>8WFQde%|%$6a!3*>4%>clZzJhKKm`x-CKzskU|$#}|_>wFF44!m## zA%hsAoz$IZce&@R9p``6W5$O(_9RfjvlkgK2_R9Zlk?(X@b0W^%5f^U9kjH8HN-IP zNchKvuAi#<&{X~-n_;mWYvuX-CIUUw3fUpN_I3wngnW_29AiM}%|yJJxb&geww@($ zJMco0x4=4}4`9wcXnL!!g+3{Y!{A?Q_{Na6Oz{udLa`d}ya%McZarkM%>F$&k_B3^ zidjkZ?Hhdp7RK(-PwNWs%OxKZs&^z(|Kg=!JTyGyb|Y>n=cS%BZ}(xx_YnScvPC7SjSf) zehz&+>$qSKxxwhIo=O=mIW!g7tT5md{~^?9wE+p>y4MUG+a**m{;m=i(@yFn(au~F zb*7#uoff16>?(6QQ?b{#OJMheMR}oQ>A@41vPT98EXV8GVF;dY3>U1C6*0y1#M%z2 z6c29CuWc{n>0UqeuB7XJ(=hoX5DkSsvqZKe--)XeoZSGGgh3JQ8){r9f<`s- zMG=i1onBwL)5W=-H@VUyjjx}-&hhdu6ocq8lgPCQeX~zApW1%P!Q^k@rk4 zA+B}Jokej@B@L7cooz@%!})NV2jOobFBPzuo8=db`3T1E>3)p6ux0G|^3g&{V-dCj zSe0}Uf4kE< z?X64T+*bZ&cN1Q^sJQg#2@%H^RTJ?&mirF-kCY-=!bxWmDK+39RmH3yvR5KC*yCHS zak~GWY{3hJ9TZs1*O|rDmW49aZ&_wHBOgvF6#q@ zcUe~v#UhxWPABlb9gQ@7SNbgD&9~^l8EsDbf8xv>gC=2XF0mVe(g;Fr$^d4RuCa!8 zPrjphNYxd)@TH|aDL=rI$Ed`b)@G5E3mB>T48$F5;IR13JKy&-YP8)k4_ybUrQ=(; z^NqH@qm1=9Sj?_R7AE%-Q4k_NO>8fT#S4!y{rU*B=#(@XnTHGbhMES3(budUT&wZK zIm0AqQvgNBUY9&dcc8Hy1?}H@Uwe%|eIdSlFG*i{>(iUKf1((Dh$YY?MxJDc#ExDi zlhRq;;9OriWsP?38#tXMOL8RB=i%Xy{*%dvXEZ+_zuKfOdc8OAi4<~1Nv(~^e6~P7 z`0TEZ0c>*&f`tK6z9$J|b)i%HwjlK&+@rhdZh3SptnXz45AW+{W=!3`ItKwlN&@I>kB(5|+8^EV6Q zzL(*}4qn{#x2@GC1g}Be-z6(sohqjK-5Y7tfF!hC2{GL}+sWj{-AccBdeBg6aZJev zgWn@~;ZcNaHJoSAYG`92Wz6n$j&CmETpS@MuAVe52t>q%_ms>k9-t$r#=sxhDsF03 zE~~z8%=o&o%n|e*(xdK#h~b)EX>azs{Xdx2yc8lfHiU94;=`Ju!uitY0$imrJ znslhqI_#~1IJ-*6N!Le6?@ABdtzEYXJ#wWc*{t=FVy02Z72MBpb+xK_%N?+$lBT)E z>5*B(|K-OG5953lflQeE1f zn&_2v0*i!PO~-`E0CM40m(heFM7-QkRc86bp)gwfL%!}3Q&OWv9Y9Z7KoU;bCT`vHu6)gF;|-2^uIlaL z{cC_r=u-L@RBXUrchsrw!i z9gXt9Q|DxUF;`wdPC_*y$YWRfX77@&9ar^ebjW=>(@t`Q<|j}M{RO!N54%ZUD8fSN z!)7h#6AY zoGu&ZaSZ10h@wwf;rcfabJ#Try&S2EjtiW@TBc9d+GiJUzSjrs9-aKBfeqSiCYmXV zI|#3DT(pGE_G1VXneR8O)J=M8n1xA_RQjwRkixpi{Z^=U&)%DO74g3sb~*G#dWhf- z!?su%m+|Fr4{M8ddkhuA(Q_?7PQT&>&Sb!DVt*M;2(!1C_?m0gu%f+5W_T!` zne&fZ;jp7%!WZ>o5Ffu@(AX62oS8NatQw!2#2( zt-N-g2V2XRZ?4p81_`FDyI1%Ed;O(@`{50td1n|eYV=L^In|r8$RovD{Z_tR3mwB&M6j){PK^7 z(5*Dwsb{OFsHfRmO8iMfVnsBojiby4Br0_0gIjHRQnu#!!>jc~@S5cyI=DYO#;Us!Z z4z&Q3M7GGLIMFY^j19p5Tvo7Vyzs3;8!)%Nxr+8;+zE_-cpl&t$FDC=c=dcmU>d$- z0EZsKO|y>U6=GBW>VC=O1qob=_b|;grd;qCZw2E7Y>e&RpMsV*(XGr+CQV_6(ZP?* zIDJ0Ps8tREa+&RWeR8-&-uwiko}=XZF$D6AT4h_Ipi=G-mh$vO^>%CT*|*Nx7Fo1~ zE~vzf$3YD+lZCw-C_V77$P~8o8khfN)rYMP9c^1M7+wtSFlWqYIUbtoN@xt;ZD}?P zgCsRNLcFD$ITKiU2~BCpd8leW_8ZA@@!>pyut9=g*|f6}v%U*z*fSdgk z8%~jWNSd*Z4G|3OjefAQ%kYljY+bvS zNtk^qvL`uvZHqo*#n!GJPom^P?E=VU>|yh%tX8gUt3?XofIV>yV6$qF`G}HRIbFh%>=M)KEY-Pt-u_2IRB+M9_P1}+YRoXAgV*(9*DK6N~o*k zy?0KVDfs2#=lK>OW!ACb$gSN<8b)p5p6-G;(Db9&T-;-@cx0$w^7|igt=k;qEBcc| z!VKHD2<7tf!wqbB$6WWb?uY1i*R`u)Ha61?X9b|!zxD!}+9%FVIU{?sdV0Qm_ z>=N*WIxN>_TL+jjZ^vVzvxuC z?wAVw{`gB|hF5RA`X^>E{ohz{ZmD=P}J&*xtqq zo*Z#Q=6a;w=@RU0$Vr{#gm+H#C7NmJIrTa|Rn_pc*cAD zGJOaleBE9*++Tk7MZ@Hm&}^^}V>6@?!zyEAxNzUe7INWApK_CzCi>XezCBl^YX_Xe)k z@@SlUwNULEnO0dDib4t5f4H=diN8OzIkr?1{WeEYtHpdKk*jf zZ>x0ijyFb1Lqx&^RelHL%ex#)l5W#EGJ%mIE4X&@M+;T zZZfcbc)WD(N&PQO-_E3Ue9#T`QOL#BUCrg(USu&V=Phh~jf@aEfhb-fg|GJJWWwKl zh+k6ce!}?|Z`5r?I*#(i_#(ADk-Y%%cY0#=kLOS);oWyS9mzQ3*mm^~7ow|dj8E3! zLZZAA*}8f}M`}Z;_1g}cLwm=iAVgHaC@-seWG?p|vRd;AHJwS>B-*jcl`AMU+7#TA zLvwM(E&f*>p$~E^OW|9MJNeD#b)`c;-LHX7N4i(2k`E5dvMyAi5u8pWa~<@8`v-Z{ zDp?CwmMU=}>N`4mZRYZ)Hf4hiZA%R$$GGRqFXQaaEXt@Zrx?8i!=yAVP4OK($EoC> zqs;H|U!>F<+aqljrqPVDo3#I7Nfy7+b;m)772?l&#Dj{isG5@)<`;XM2G5a{r(Hwv zEHlCi-RSGEpRSzRnflLX8DG};%Z`nI)Qag(5#R~tnYpMr3B7J)nv)rv`=F}!VUV~J z!oWne>RFG`(Cq+=fz~L|CUjA))|W6x?`(tUr>#cEDusq7?`t;>EKb{A{uaR+Ok2?U z4^Hr`7%abx+S`JbE-6|Pg(o+$jlbhM3b zHV%fQ@IbQ23EiH1D=Nz)9o8ubAFb1MTuD;Rx$Iw@7&#sSSkiXotfkI=qm(+X z=kKe&$F{$bRxw1B41&~BDTuag#BvRrL7g=OD#6|p9obtm`O-ojjA@XP#R~p)8{Z%z z=e5fLI)qNzUlV*3uT+~7v2WO~{@BvxFY1?<%H#TzC7#yBchV+j(u_alUMV*vydKIe zr|HmrRT?pR%obMaT;_sMzX1{peYFDVKT3HF?rn6l zA?|!b3A;|baaG~B{Q*8A=sZH%2K_X}yX)0NImEFo-?h%jydh^m9jEf!z3AN;Vavkh zhm&?S1mS@@FDYFoq{3g(BkuORpd8|{Q5^Mm#oKx_?x4@RI$p1( zAazA4+P8?qzbu%iMDjV*?KL6ZI^D&S~@)X?Csk1p}X}PFZ(+-iwsMC=oUHlMj?PVLn@UFytL?UintlS zg^FSdGE{ljNn*|~tmG3A+_phvUEl^^zXMi4`WNf`*kfzUSn^+RCXm`gT#wAq}9#BxF#PCY_w&#-tbj*z7}lN z3(;mvfhw`j0`K>vrHJ$X2sJD{*vxM=I=O*~a5*rdmFdam#zA3pKU<;YoWduzls03% z*i+E4Pex#{$+o@Lxf(KC-1?pgtX1gcWdxD#-3EZ)`fsukMLAab6XGl zD#-W8$7!*mY)^w$+7e+$Lhx`W@gx9aHHYV49FglYFRsMoGLBUz!;@Ika?D<#*)t(S+lN=}hVs?!eC>WerIuI8Sh#*X2XB8C~GY zXbyn|^QG)uG21$?N9&lVdoeYxd$lr-qS>B^eu{r&xyJM3f^IGrH-%4??3r37hrZ9Xo0ABs6h6thYb9X-EKiM zf>BSB|LC#)MV$7!YUlT!g@vL^(UW{0ZR3*}_z7C!IJYu2uJxdC`WXH*4gnXiVwO=g z%)eg;$Ey?_JqXL+8Mdg~*|SjR?{>4k6~XX4WTfK|+VPqaeDLzmvgS+r|ha65WWC!(YR>raN%%+JIO7iqeW=#FU z>fbU;s0Im9AnnBTkt7kFgtxtsB6_;%Vn^U?u9QKMEPvh4>|VDl5WWOw^b#;LffSn0 zcDofHbVij!R0wJK#+TuMeBCDW^U~v)8cua$JKFG)UODse1&{U#|MY?N^pC%1V=&Qh zGwdlHV{NIkZX^2CU;~7aAoAYu?UMhou(4~x_eekD1^#6N&$k(#uAW<8W`2SNY=`q< zX`;B4N8Oi&)ryy%%Ou$WTkSCUgi2Ji{Tb1UcTpXrS?Zt4ot#`Qv#RT&o?c$F`W~j! zU|WklipS_Y&(Nex5NZ@_+az?$Ij2y3F|%Qxm2%RzmZP>-;6?R_FiDHam~s~7yWC3S zLc6@33ZO%4fvaSyPq1`7h4and`bk|7u&C`$?X0Aax-X?OF3L0mGsO&WLMm?Zpbg`L zu1>DDPKo6vrPekXuX-Fc{8WXE;o%FR+OtMnJ~&IA2k}DE*;{v|1+@`R^$g_X$ekKL z|JJ@e-~e-YF3q0wjXlkOR53nJi|GL|6&0z_r#9%pg^{iO(P4i36B4VM4aG9}cK-VI z`uck)Xh(U{#|Q~;6^Agn`;BRqTnnRLb;^m-n(Iypxnroak<#Dy!!2GxRd6m+`u9au z1fGJ?<4)f~v9?UXsg(6QcgEF1C#oyj?^1ptlZ zKxh1r&{(b25YVYh(CKYV*s$Z+Z>`ljljortT}$!1R&tRm)DUObHq3wKX!I=z$e&E{ zgm`aihUE3ir#;v4&(p$Ky3wyaDMy!3(Pn$5M1kor3f8>VaT1F=NgA8>v|=_TZ&m`# zZBT(7;(Ku!+p#pA$twzkWAr<$1C}*|K9gqvpruhddIGuBTWoWOM9IZAUtbTa#>Z#e zNR7jjs=dalKX&#E);dalfJHJuJQ}50ya(U4x_!OFZAfRs9k}Hv3VyMqTj!L)C})=l(azc=016fj-7p+)&kw&mNpi)o#FPiEA0m3 zFYEg&Dv~xmzaK5*HgjF+y6^K0RL9M#xX!98;RhK~AEI}7@50|Hp0H<@@35=9&$)B3 z{GVmu7d`!-1wc5S*_eGP+ApB0GiLo#M1H5Q`-UB)HnlB|n3I?SAZaWX*01%+SyGrC zvGq%<5&R9z;vPJ(MZ?n}d-cUr z&~VuU;K7@Cg=1&VvvRpwpYlb1g-?(*KxhA#t-UvSf?H~V) z3M#;HJ^)s~MkiiR#dP4Ge^(DSa*))snWCUt%si~YSiJ@P3K7$=w<#d{O@$6NX^tRT ze$^%Q&ePZB7kDBR#BA~5X#tnZGZ>j&8G<0|Uz41?@;&?T%ND>}v&gKfIldlx zQ_$U;d(unu-Mxx8pNaJH!pL=iK_MX#8sGMUZb?401svW#e|`*8d7kn`MXxStpv>{N z@lV!Y_xTy$u``K7O`27c-G7!ZcD)t)y=Qus_X2pg4B@0p#gep{5VKubq9;@E03)=L zkS$d44O@A-xnWb`8cMz#!MWq3L3+yf`;+u@%l--7Phk1kfix{k|h+Y^6YJM z^W>u$fbo)zN~ffuKW@|b`p*PuD=c8fX6^_W*~C&#TsOixC#(dF(rNJw*Z@PD{b+oz zW$YYb8U*6CN)!*3LEk!+M!&zH=`?8VC$sx&nz;twOU=@pb}BP4Hg7B$>FrwbdqNdF zl{-FkOGx>UU-E3wYM(3D(bnP3o{PQt@kaN(^WDfw?J~%9;yo+a#29ug+qMs=_v^6~ z4s$zmGWEfVlu=}%lTovHeBhUUqs7S2p8U<7h70mXf}ZQRF4F<~0JVxXuEI!hk5_Hp zTrD`A=Sf8n-U%$_)X!Fn{`t>KXFrv};=RJjPcJzb9UT3-CHEiQU{OC(Q6DIZh+{cq z&ti+>)sDa`D)rLvxy9n-ET%6AqJiKR8C)aXk8T6#4eGxPqc?cnOSJN63rwA5m6 z_rP(@(91kE*bW0I(}iI;8N|3rZdm+V;s2uwqlTaz@;McjIX#Ucx6hgNk#`&7%BX|0 zo=}x8d@eFG_M>)C)cd8&oT9Ih*}P~NVMgf`^+vM?aUW53Bar_J< zjGWq!5OOz=cuKy?iZ=O@V&rT(g0XM4Y4^ZPFGgQ$@`|Nvk^WE>9@7{kf@1FNrd#|brRyd~j5{@>_ES zJJBPa%m?3usB1s6re^Vi)O*nn%0Y_8d@H47zcl7U9OR#Yyzp(-7W7wU#%Y7C)T%{n z2}5tp9LPN%*GOihy9J|`N=q|M{=B76h!$jBv-(g|Zb$pai7GvXsi}rAHtMTvQYKbK29qFM8&w9^aij zobBtx#UvZj^z|E0l8f&LyG@Iq9IIv&)8ZGY&BJ5Zd!&g2X(z~geL6V;&-oX>sKo|SeYc3SSZ`q?8p-*fQoM5PkCJGXXQ9n>`l{+o{ryAN3(B~I$W@`taOtq`%!0usYq*WU2Xxp zPG`K(1N9<#XZm9_`#X@HAkS^^p~jq9G=H%Xs}I+j-%It9n}=$sX@*%`itM~+NcWV( zXemR@*F`gb^dL2+f;UUl<6j&TD?0~CPLJJ1IjB7FSrzl*^XcE+&Wj_Sd_PEP`fT2R zmI3c?yXg>bNQUkUM&JSN=N~nDtP@yJ`DUeV8P*RNzm?I15oDWZt*5ROJpw4plXfKI zOha3<4Jg|w)s!<>;ZtkvjS0Oa*e<-DxIJ=H4|r`ZrUzGBV>x|}Ur$=ZLfoof{ahC5 zSIXeqaS*BUlF$nFjBF$pVL;`tJ>MhsFriD3sh@`$+IYFRZMn8l$MQrJ_7B*1s zJ`6J33mC0TS9xX;18@34o{9He!D^mPm%;33J2tgj`BDoV=&z-h-mM5eD4>OYwRCOW z7DbIW+<&o{FLpgHQe`=hZ?H3CDxiE&8Pu!e(aabicXKDET`K@D^lMZqs2rTrP`cwt4?OMN zT1FhF%g;J0@EM7A2R;H(N+Tv-zgpSHK2^HOkH0%2!8&*rQjK`{SVd!gR8dZ+P4DM{ z_|B-r?oa@x!~8u%Bi4PXiCMcfc*&d51RQhw+C>rbG=QPgmVTx5gxaJ}gOgdQY+7~; z>c21YRh&<6U7Qo%XoFIlxn*l$y-fYVYX~SvL*-{?++KtyA{4NCP)>iwX$3mB(FJ<2 z78DY>G=jq_NyUURf(FUr_l19Yb6O{-vZhwfOpX5~He^TdWM-((a`$EAS}NSdO1BS+ zy_wqihGb|!x^t(~X<*?Mb=4H3KF!8@Fp7L5GIGg6nkDtibuQOey$68mv;3h!81SfB z)w|^j*V&S4`kjHQ{ywj2-M*EoNL|qO8^WPw)@74w$Y}y`#Jv(OSG{9qESr`uV{vFx zr!(5G5Kt-bISAMG1DDA#3M9Zo-P?Ugql)Vs3(NO!**_o`Bxiu$li=7W6vd7n7Eas#@4 zOi6s9N2dO^aUh{CNPCaMEQ@2Mgl9dN`O-lD{<7xQdqZw2xQ$*`Dd&-KvGU@`CqYly zs}p*x5XB{Yv^rmN_PWIWcf6%-Dap8k+cAw~mYaGSHd0bvs0hd-f@LZCASVM1qt74t=RwE4aR{TdMcU8fjG&fY?0eI>+6 zZguBW63)?+4Lq0E@@v5pT2NHP4w)(*AI2p46AQuQmGwn+5AKxb zJ;zJMr=kOG8BtgmAwfxfI9Y7yna%?T{=$@C?gmIeQbRT49rvFEGyT6nheaaE7JXh) zG?+f(@DTYADJ%XuW_L<&^)2*ExET`Lk_E)9`HVcI4Y;#RA9h#GWaHS4thRHRTzoRg z@L=doWkxnT&rMt(o5cmCQK1 zmFyf{(+2dEu7`?gB3=hd_vJH}J4}_#;8SQ$J-uj-3X-bYuh}xBdeq3Z9sIsz$=0&D zCX|WKxt19Nh-3)pAkEuQo5EZEeuumIw_a}go{6QZzB^AhLSYj&<{wwEmS~Zty&+B==y@a@sCz4n{Ik-})5jbk)2a$iP(XT9UF7kW>LJUF zPftnoG)sQg6!DM&sOsM2KMPK(fhszpGW0Z)>f6D$FtICOu{X6RNP40VLPEofu+=`9D6|k zir%$QQdYolX~31p8DYU$`n*t92L~VMYsQ#Vt)*$|*sp$x_-G6MiHOSYhi_9HN2rLd z;eAs;fTmt8`ghbFFn<=YLaV`LU{%LHdHWDxt^#%I1q0^MSe&fl0DLS4PzH0*f*zMW zHN;UKcc9XTz&dBF<#U+WAKbeb$nW0GsYJGi8qj%jn96Zm{u0ZI!z8sEP3NDjr+$g$ z_~%RQ-w=13eXViHRfXZ0gY%dspqS#qGQ3VuITz{2mSFus@5#pZxQS5Df0V3k4QlN6 zZrrZ--wzM5)m6@@yJXBeuwNr_M}nm-8F>m25Zp=;Ls)M)=wRDQCJPKQXFMyhVZ_4R z0@hZgKjy(F;S8#_hzV^I9+#$;-aR|vCs$eb<|gM3;jlkB0(iT9>^GHlyrKl~kFw<% zyLqFe1P)oL@y-yu{LpCzP@CbEaBRcQU!b2RO!?4 z1>s5H_M9$XB4k(~nV`c?)M|XVl+#1(j(%>zNM(^g336}U}bdQgyrEN)OQex zVe%vt(<4a@@-c+-UTiM(tEL`K?BU_I#=b{(om#6GDyfc#O}Q_bPZgst3rf}+((8b2 zr0WE~$MMjEKL$eXu|3*u2wKi%ohOq_XfZj<5zXb%zNZr6p_Vm)Z}u7coTLSgv>dKk z)Zed#gZTTuyzOAUzsyjQ@kL|##N;J_O!yifA_}~@6Yc=+gDA4%-SvM|KfMHV7{Lbt zhZz(*Xx0!ta+DSI+NqfWz~WK%0z>?QM;YB@k|#(O%L$u0el%0QYl*ZI zI$U|#y&<~dwDK;k`RBPX<}vZ}8u9on*>I1DjD#_%Q+Xg%Ih4H~G{Z)&3g82ddIfxs5twVrjY zZ}ceLUdgQeA|fw&lGN);CSu7L=wZu0gB9M{YEaHD6wEaHM5;Qd)!$S-Opu|FsEs#v zS-yLZzfKfK9{vZT1VF zeU%ewr7KQSQhD}@E-xKX?xrI1l3IxCuh&BNXe5pJ(E8j?(8m?2H+bOw- z^ADM`m%#|V(bR-aMlTlqzcSk^WUM==lGo;xyy{-Wb&X?Z431f=xNwQSebXPjkdbtc zDgkN=hy%&pF6B7H1!qTmDtZe#yI`<8A`bhzr+aAeR%U2X_2ErswG$&8jrWJYY;Dy; zDa%iI^~w2b&7LuvNOo?m-i!?Dl`@@1%>G`1R!Q5n9p9i-oJV9YTV@G$5r>|)z7uHGmr=C%6vK*oqzH(pl?P3!OR^GXkuT-iO-IBWjCy2d(hZh z*>WQl`rn>2q;9x2vY0==kCQQQKg56L(EI(r{WhB}A>3wdQXQl6(yOMgzlPP6D{i!G zygz;VV>~|(ut8Bkb&SJjGDC-lo*_nGgLM(%P~^G?*|;XO&7O^wB7ccgSlHBvW+aBt zIp0ubba=&w5ii5Y$0k(Av~acd{xk3Wv}Df~>i&i8T>FYC2nh-JsUx#02KFA;%{Mhe=`)>Up(uDqo*aR*S@!abm$Q6z*$c$(-%vmpwYfhD%uvtnn z25encY4`pmS8b>EQ{UBs#Q5EFT_1epwbg^Em>~t@?$RK<5oNq_*328#_qaYjo6XI&=C}!iIU-CQW#^tCaXrQi-9j*>=*DO<^Wj{9H{rX1VP6?LYFe8b)%lmdWl*mfIb-*0@vGuz1XC_lv|^UJX6wtT zlF-;!!OzrPb5gM0V{KZlxtet7ksUu#0;Oi9?3Pt8tIg#xh5gn3BPBUEk8Dl@itH=; zem+V($4wd|l>`}}t`+)1-#8~Dr?`~WL>($Tn2cYNi}VzBU1Bb1kj2SI$VoIa1(F)r z=;BnK(UmY``WHzaAr#~qWoMr;P&vux%K&x&ayq-dlTZiQnJJv;PdX2_^B7;o6iR<_9CM~c<4qrp4gxb(1DrC8<=Om}dO-mU zAF7o%CHPfTPHK#3=Pe3urSMIoPvUO_Xu4*3zcC+=eie!gnAT||1w}c)R8T6b{%23j z`JDX)dWI0M-W$&4(t2VIkzNHNIr3pwWGxS=0vX^NW)FjPK3z-;@$3sT7rDWI&3ymk zMh=<;Mnh2dZ*P`iG%2<*i2n@n^F3q5&f$f3j!Vsv!L{-wS{Hr)f-VjffeRt?S8P*O z^&LSa(6NZ+;lpHqC&!C$l7@M)x)#Z~X&SYcMNw~w;{CDYdvS}5I}>9X_W@SuAXN8f zfnVX(UBWyGAi(=tPlG|lPgH=b7s1gpDO&^11rFgz0Wh_3F_t>px^s7dd0xfOJY~}i z!nI;ZsJrp{g0gal#p4P%VGjfQ=6n?4S`f2^4g8x3(xsVez@)8+>T75K=Uls?XdK;| zq_l(RG`DWstT?D%Y#VFg1EZi4pZLo(2nj(=f4HJWaK$lc?Nqk=($a zM8FjX9$(1?HYNb>pM<`2P9Ox*qh(8aqG(S-bVgYSZ|<18bYp}&dQ ztlfcZ^fmP1q;6Gio)!09(NBLz)wq$7`!&f13eJ2Hd3YK4^PXN`dA{NIS66wB2aMR+ zz_a40Z5P#a&s0Y=m#3$`|BEeuc+G9EXFDCa8d`lrN{+dqjUfzCY+QkMo0SJ7&}t4U zZ1jv8M-2BTT=!8;@_vPmZMoFr1tR9UFCa+J%CMuUR6`%xV6?GLD1gF&w2K+{QL%B8 zVdQOhNpf{1rVrtk zlek)RYsEog@74@>`g5KKrZ0GB>KGoTp_+vwO(@d&;V`Sp%x4eDw@uYIaukb-`BnAp z__Y2aT-;d=(zromfx35EZDBf+rX?-hfypLRZ1-7lfv4tzSiLCjL_RJ>0Sa|dw{ncQdlFr$5yn4LbdTDiYSLG8W!W$-nA*r!XycT&)c zQPLFHZqUx%)0--_Qv^x1MJMDa91IwT4y@H7!&)0-15Md>lr}x&Hx#0$L&@P~)3^87 zh8!Mt`Qjvh3y``)fwQuIxh<=JMzo)i)9;+u!f1o(-i6iCp>%#fV0#hFzk#LfignEB z86=wAryH_d75mF71m{M`b>?%NN_w_Bopju5f6>3k$`2TUt3@cw`cC7Tbqx!3pArkc zhh=~GU(GmV2de8PM%})rluTEq_7BYt`iVg-U!-c_E%qtqwbg8>d)NT7wRPweh`vzD zAZqxCakrEBJ{Z;VA!J%PUGJ3w+T=Kd@)8N0f)MoEmju*SS*(MH?;q`Fk>yPq%VZbY zzOpOH3^ds~q(Bu|kEh+qWztr{cSTUkUt?LzSM-uddJN+M9&ww_P zBVG-bd-j^3q}pUfK3-;J$guui2SZ?gt#+MeTaJHX)y!z*`+ke>m;O|R@0c^~Vy5XR zhtKd~(B2AFw)1lKbzs86vc>zo(G+}s2TSH)`&nrvSTpzYk99%&~! zUapHm_Gh`W6o9Ic&H{Sv?-}hb`bF_!7hr5BUEdudyptIm_{c3D`KqLU-66L|`VAq+ zT-nV2F5YbYzn|=*2>u6c=KST=_^?A>Eo^#f8EzN$Pju(Q-6N z4LMY1q_TG>H=mw~GRH%-1@rc@V35%9bm4f@}$V&`&_e|-vlZn2SDmT}}EL2kDvfq1)y znOcfzzO*)pwtxXWEkuq9jeK|41 z11yIp5xlv_Jy(S6R4TTS)}YsLG=#Dl;6Fm5*O0bQAGApmNAh$@^+`V9OAw+~2)JD6 z1HO1cOfFe<;^%iMTc52usE^8RGi4>jGN=e43f3OhG>^`utr~ni`5q~#L|(|xnAAZ^ ze*9_!d;_ZC_`jjID|5xggBHc)XIcO&zBr~nA=~$V7=ST zInyi65{SRsawGA{lQK?YN31)=P@r7SLCimL6_biKWbYbC-c@Q{AppIXx4o(t;RMlV z@P7v;5QKI-OOp{}4Xfk1ahzOh*|=4#zsXt~z1)qtt!7!qsk_5PVXp?D28QYq6$=8F zA8&pG|1Nk>6dcF8^p}+RKJWPx{^uMS%B*2<_Usp|o~51@Oao}DRS z(MM<;RZ4rA44i&bqim}4oM=A3nQw7y>&t*@^=~+AAYkxl(ARsmB%BB0p zZmN4NXys#x8^3}K{|soUY95r@aNMq@f4$O9H>?>_K^@PJlbf`YsDwM$sNO5W3}6&1 z2wW7>J{H!mnRE%zrbTCL#4NOZ4`E_VItk^w3hKEuwH~tQE;z6pmR@4@1VEw8IescFe zSDGae6K#JN{3rggR{VXWmcrfmklLM+T5QOXTGDOKLDD~^-*{BiUl$OF9OP2l^^rF{w4te9}mt!HsPpz0sc+7WZbCDS(MmC%p z3onk+l~_p_SksTYU0JGk^A<|G4lH!LV0}&DW!)6kxw;|iL#hl}dQ=NA zj(whg^@Oc?Aj=Trx^_Q%m##Z;c= z&nMULpyfnmke-33OmZs~MEC0PY}E50tEEvo@~#lIpRYl`&#}IJJDs!*_GzNFTc?9T zntQ{V3jw;V@}5M?qyCHSn9ZDtHPQ(tShb`Me489$7{q!}+|WW5419#FkhDIPJrY~X z!)%t?^KC0ytPY_TQ<#m=oG7-yxkU9U1A43@0xl z+1GpN^Nqp)gSsRc(8Ga3;e51DSwdq_GLz8^9cB=H@;5C6hwaEy z1%CXc|H5BuMGJL*qZ~qN(B{`~-D5JtdT!_gZ&87ufyYo{RSZr|6b`2QRYf zDPtLiI9#&|INYt+I&Zlh=lnawT1fYJRIE4D*f-muz$c(WxkXQC^*MA?jBM1D%C0^n zE~L*G|Ac7TJ8-KSU&NxepN80S`GqX&EG0z01>D-sLBEUIsAQ=#o3jY#qdQ|#Rpi|d zwn2)lII^K$;>}VC*P6F^8U(a300u>n%k?nxOh0JQ2(*YQ;hZ@JRwmV(2L^1!v5-{1fcvJrbM0oLBZ4IY?FY0w zp|oL-p%qMPSK-uv>E!M|%A4V2J%j`{lgrI~wVyB(ED$rn3hS?gKD^Q~A)QaXnDzel zzv3owu&wIxDdv!!^T+U2c8~xpnLGCg*M1)h#)bu1>DDDH`~4xfElrUogUhwDx86OX zI2G?I&>t-mIat~_@wwr?4t+zvDHBdRuaCqR0|SjwBu9_msc#EDtb&<`)9fI12Py{+ z?Zk1t+O)~EC+s1!FG%ObYDyNmtnR40{sd{FqD~+1NuARU@dogK4$26w*w@U^o&l9^ zu{P%I9#)FWh6aDW5FUTj1d<``D6Kte*fUW?##jM^$uPXDanIhGd?zKUHzF@s(R6RAP)rBd~+&OYjr;0>g>|@)I9G0sV z;@QF3qvpG>tS1yIzI}G|*1RGXcJHZ%lG->2vAZz_+vdM1mN4N}=hzl-JFK6agC@Y? z@xO?{kj0u;XwB*s21!YHpZ%VSs^c-Tz!)DqI3uQb-neSqCZymkhBSPJXGO(e1Dqcf zN{CMP=_f9tZ+ltYS*WBM@x z-v0%2Y3Xa+mlzVi%=A2{E%#YL{Q5o)z$NaNP?sjSbQdst7Z0BbO7c)fs%xhB>&}L$ zNgzhHyVWeRV)wnqM&_;Zf=EbvTHj6aSQB|ldvb7Sc`)==pOXf&_kF}~xhQ*Q&z7=v zbrkd_Eif~$bqZNFd+|c{BJ)?j$`>u{u%uVThHR2A0fWkYYWyVYBGeLTsEtTrAFG+0 z_w0TMws4QVd6^&5T7@bFI6W%H>VjuoH9`xr&;|oGtZrL&qF~rcMfAI!hr3* zHQyx02TvIcm0TqKhj-2OiuG@DnW4husHbM5AioYG*>vWZ0aniJR?;+Ad||2qw+}%F z*so#;yDbf|ftCd>d=eJQ=AJnu@TExtc`84H^wy4%6GMzdoiFGzY4P4#&>#sH(U_Dk zPq6!``ZyE`l|GCz z;oi%boRBI9ciY7VEdC`!an1xahnYVDphSO$ROxgO*7}W!Tc#G2$IPHD755b~w^}hq z;hB%cGbySjNl{BhKSCI3>-FYILXEwx^*zjyT5QxB+Qewl{P_-b1R&2Rb~bG1?peW6EjkxhJbv~{1@raGZZ(ceg#|&H{mUd&;ZyFvkB+Xh)9#)=Mq5+} zPp1Q;Ul%gCZc1of{A~sQSIy2IvBzQySP^1-apq6y5Ro^olZ8X|=JX8Rsr~kw&kH(0 z8EB=(7H!a9CUq9lMh2KdkXCCGg^onQn~RnA{kX$pA=wA9FbjGnDLj+w-0SuXLThd> za&D^rG60&i99EuNAobU3W}j;;bfVXW@p+iJ-#`md_xo5$T|nrt&LKW#Z(=zM1SnK@ zA(8p8cNh;rDHwp&+E3%p2)cAruia{V6gYkzXYi+p^tEv$o5_fuE3HTJs;;#>0>lMF{j?IVgD7f2<6ARcRyRRRa*An4 zlBBuTh^u3B13Rsc1(IV01*gD>cYt|P)M zyCF+B3Q?}>;l%eXi2MkXv%*3?33`fJ^Kq>&m|4M%|A7CM#|dd}T|MUE{CyP@C&9?Y zPacG({03&Rnk$)t?~%JEQX9V7|HXmguf|wciMnj*^(?xXQ1!PCl%l9J!o5~ZOwW9j ze2Pxf4q2jHh)B9_r2?Sz`C1uH3~yB?FQnBhwDlI5t%b-}FRkm`R4+~Vqx_j61mEyl za)Yv+yT=;xwfKS_D;LOH#5dG)*gOXZucJWLW~o7^?Gb#*MyYL2L9J9!(!Sr|A@}uq z$GxwycVWqwy;V`r>Zp*$IcmVlp%U`jpyGIYUSUA<%}ypsAMHk{kRWJ8>@CgKbxaUN z_wMi7V1E1WTOkK*d9H3k@pFQ74J5oenU$YbX3!PS2tQ8g(EP|IB%AwRVaAHOCxiIiI|Wr#@52`B z?m0?`gp^=0tCqYwqV{PUN1u%Nf1ttxiF9| zUguWPz_<-4!2qBVbN+(&)!mpLFu>deB1?!j2U4wZ=sKND5577w0N`#PMVJbhWqNFLmS; z3aT3kHjgOwS}IUB8k1U*6fN?nOD11GnGcQ6-ug43B34Z{Ufuy}e@XC8N`6tcumnte zuerEmTUkH4W`Rs5n>+-xW|cB$*NdNDSPs&!J9h8!$rUx^z%YLRDQ2$Foy=}q5O!pP@rIoS-;b1{4ek>w_5P=d5UM1{ zNrIkYp-49cwihq(R!vfiHZSyUpWJX5^fqIII?WIPt+@%awl`!HL}FI+^_iDXi35+X zNqE36HXRG<11e2bWDVgIBkb-M*wn^f@B>+m2Gz$GNC?+IFkL$ME?gAU_HE`9Zni%N zGZ*6Ajae^sq$Y+(OJ4F`eYSwwQGW3a!)N<=A+*=7inntrJ@cVcFk-hPap&$4;P|ZV zBw)7PA)wNtUF>*vF7sU%C4_B(O)@MbmnP1~wY$}{UdxqK7a5^*;lFzJ5EdXy2Jlt^ zU&M2)1`~c2YN876Jbq~1O!kD~{7$Dl$tyh}+3U1+`j@b4@V8HQeOAj4F#>K+dA5rn zCbTyvd&JHWByU_)-K1U5H6H!CSie-B*nZwKkm)v$az{%kBsc-PiTb+?f)jj^j4AQo zl*AT{x(_jj$VJ>x&oU0!2R%ToUblR9ujKrob%Dt{{U_)nm4+vzyO&Atmz83-&XYbo z(X6!ViBRdi6ZoxIer+lTF!xX`61t|Q9`Gl)>Yu43OEubmbf5xqPVNNB*_cKu|DvG4 z5<8w#gXsi^1;|n#UmA3|S&E()rU*vR=Lf!;yvXYor-B<;>Ze{>A>%)BUJ+O<%8N=q z?w3W0a#-4Bu1E-%<*_QRYZHwGYHC1aNM>c7p9p26#l{=@0DvGWjlBfue03_9TRnr5 zB69Q+J!P6dk_&5^eLCdCa~Gy$u74*mRLvoUCTIp`^d0243AIE{3xy9_XAm^%FuL6k z{)wzOF`FH`XuIa#8D?mwHka1H9yy*0n*?{<+r z6#dvh`WkqL-O~5d(caE*FSPWx?Z#@SvQ=w!8qnp**{=~-#XFBqDfY8#*70>Z?*Nlz$Yww$J$cJhKaym97&QB&$RDqEyVDw|<4DWJU8Gh#{ z^a(-CZ~_OOki0m4m-|O27yEI1>Z4@A(dS2_8;l~hfi3Y?wGp?392z7&px2F^X_(8C zffPGARA=v4XG|%4lN1YR{1FJpy0e+N{Pw-y%k?- zCJ~}KChQbFhR{VDlj5Z5s zK&U?JlVD?g=U1^R9pUCR7KMF-Y5qMGz@7H^MX*gO+E0wF;B^>;t35a5w8lK^SL&`e z^Y3#;3Nl@bWqDHYvQa{lsB__k3b8lS`fdly>(NrrVCpnmHy;ROnF8Q}W=Wl*^R8}Z z4qKLye68n5vHVq#b-AZ}9K7>Hsr?Nw`i8`m2=Ga}vucwI>Cpse$iI}E@aDC4SpBbD zrzFqT2Dy^|4hpi-207@!rq1QG8u;geQ@U?AK<{YSO|64`?+$3%q4bPQ`lg*CtLb;rucZ^5eWltRbYW zq>lPVju~FEk5vK4Sw4Y1^K%Ti1!@EgM%)4by8}FZT^&=L~j!7liRKN)Ydj>V7$4U*KDR93_r0;$IiOj5^9T!rBz{B&0UC5iD| zI$5%#Qa{%J@PurbQ=;38XI2T!0z%N$-Q@NB z@r%+%so{=aA5d4C8Y+X$r#1W05R6z!*8avi)#NblFy$iWp+cphcdm3w$NA`sJO5%J zsLJ_%sx|w`Mt*1V^XoDw1Ga%CCj*vfeZ#g<0=of*Y!a1Thl@)%Vw$k*%7t@z5@&K^VZi?{+;{N73y2O9H00}*G=j|Bn zdP+=n@_1f1htSyfmhA_{_sGEb<^Ul~omHE0KXOkp$T!NrV3Q$Na8DK}nalJ1Gal6OM;Wjais#KF3{I? z;iT>YM@+~1$phC)YwR>WIPvaVzxU) zVzyPHV_JnRoR-)QA&1%8XR2s1F2o4pu!~xw&h2qu2p_%2xKcz`X^84i!O5xiGpPjy zziGlVxju5ijCY)=9om=ynti`FP zMw5C^ckMuf;flp-K*E@@dD8hO&=L(Lv2J)DWUAxEytU#W?NGcG=3_Yx8@FR=;>q9! z05M2+v>YT?V3QUE1g(XF+uIMp*b9@k=m5T7r|tSgqj_NFCP7`Nwpw+^;(;zicWaWK z;GCLjRLca}ZUS$2NNyw{@@0JV@5u4P3(8lV-8b%}*$KuEO57rqlRIAgtV$eCySjWj z)$`+51b*6y<7nstPB3j9#c#E)P2~|Z3o_i0;-iyqeFkyKDf1F#*7D`Idkd!Zi8$H= zFU&BI_V6k&jW*e+#4;zb zxyQS*_4itCv>9)d`sFeU@UtXFNvgEHi$_%3NXo*hb@Z6frP zqC`;br65YK&S83-!yY|$V5i~Rp7#R-zqYx)H-)v5Wq)i;6Q*hc-8b52*c_!iIF;Aw zVgdy2Nsjf}Q*Rc^p52Bd*$l;-NKc-8`hm%-mj9=D#LdbVS%{6f)9ro#X`dq`Q0($A za~O!tTsi$JwNo7%dj6^yaeXYZrEFLxW;I^8f7U+f*vU_Heun#)_0a2MhH+C4iQ13oOE)+*?VH&X1Y81V0p2cVsGYE%mT?$_j{C;$0 z-9!uGaBi95&P%QfiGt`>MdMbs#=_E;V?vEs#yehi^p_jJ@m@W|YD)H9m8_jVv&&%GQxO_1XTE|UipLjt8!}+{lu)6=Q-`2hOm_&I&y`N$`}jPC|T+&T?H3PAyR9HFFs)GjO%`RD~c%eu;Gvs?? z3vC$Ut6tZu9zL?E@G|8^(6wa-qr7tTL_$g3&{!ZHPHP2MYdk6bFrqQU9I-vOp$_v8o0PJiwU!TqyO8UVXtuKRebLn7qd^w zs!dzWhVEN!S9n7F-wndV_-T*an+A}2Zlj7}Jbwb;`*u9^-WAtx^N-vf?ePA}Mu`?y?a=dt;v zBs~B%#lccv)lbG7$$_#yF&jVJ)cv4Q!+FF2(U}Qy856G}-1mrY7>;VKK9l+JLzTw} zey)sMtM49lK7TkL9hU3rIV7LddlbZU_P+1n?XoU9f27vx{;9zZhOrNEHFeHIx8ZC) zT*27c8Jycqd)-ggzGeYw-~rdf_OkHt>(2M zrCOOCV~r^nkke*r+2C1dsAW)Vh>(Ka_#=LCI zENpAnH0{K2wMutYg82|+C1ZE{s1wp%NCxn-+c4MMbv`}OQiW8x&wTbO2X!CTMK!z; zmWHi60b8KAOHqXVq}JWLQpS@vv}|)XxUO33S7Wf=I_5k#GCic+&|Y~YZ2ZvTJ;e5;@UxY={GBm zMd7-v`G1)D%D5=ly1SMRm5@?UNnuyIYYC;LTb6F= zSmM8&=Q-znzwUn7o!{Iu*UVgV&GpB4+{EkaPg@O{AepA^+lxn5fWoS?fjwUKRg@AZ zi2xUYgtGV48;Y$leQ8<}GbBg6cOa{jAA)ZgIV46D z#S>VTMnZ9jAJr+4C4JpfH49>n1+SrOSIV_iNmm=lGl$Y`I~3pOgXG05;cR09g752r z0L0{0B$k-1%36!e)>Zm7Q&5G{nI0}JSzhZKBODYo6=Hc9#$`ND=jRr0#X)#E{yLIs z;UgWQyMF0poz|Wib4RR*wJ8*aNh&ah!Vb+V|Cts4-V(F@-{X6W#zDBR(8zzK<7{+4 zKEc;>;&@6A&d;HZQ8&6^ZGxnE_5Ub~UHN~}9HDV^}96M+?28f5+M zvuV*e89}U%x#*0hxNyxqi5)eR6BxhUQudO{lcy&FrIygDnQYch(_5-`!K{S!w_%8= zUm;wm$i@AAl`E3QD4^ZDCX^dk_a+Y*h`aBWz>nIzXmnV1QC$QgBfXZdvS=iQOLg97 zX-4-Oa^JF)3Sq0hD(Mo=GQ2X~-GdA;GkC&%eW@YP?bC4ZO3&NybMke9mD)T? zxlybdXr;-MG7-C-%hpA28Ta)Yx)uv?O%<{jb}$g8L#q8Ey5-%DR4?7k)#B8j%MIdv z4&fQivBy6Nx<}lzMr-kWl7!&jEM08QEOuh& z(jFj4*lIKccqrpBn`-k zAv_{Kh*U(NknFN;+I290LcD60*sWcXHAXJoj=ox19Tt%Lo)NGss?3BoX{Es?e;dp; z7pO}!zFotE{sqM~^zA^UK_wFI?nBATcghd4Gyi&!aNO?)vSjuzL^vHp*Ht%UIHYx8l;`Vb={hXn>X&jO=c8*{*>VSOb!eUW!28% zWW_LIXl=A|s8St@$3ptUwieviqKA<`2|h;tRGu^@7{Cp1$y znE#&UaB}Gf<8OK8#iQHbmVM4Xq%WDYyHl}DPyU7kb-$y1diA8j>TS$yxs(j53A8pYe`dg5S45vsu$V z*;_{Jac51+)&5RD^(IMwzSH*ix=W(34vmAumCWTZ8vMG9KQqG-q9PqC)^3;>Qod&R z^uy@|eSDF=-bD|cIejY+KL9yO$!6yV?q42A0~h=p@I&9YrG$dVO0bJ zW=Oem{(6_3#4A?S(84&mOTA&LI0W2{<10J;z?wm{eKhvI|+0Q z#8hMbIVuc1o*J!WT3*^7V#w+!>Dp6w zO7@kB!QYn4duKK<@`=2eQ?9pgq^KBI8Nuqe^yzDR?%ZVFxK^D|%q#Z~JXhv)syBLS z9Wu-bzx0kSx+4OZK_RV1Sq1GJ$SUxjsIEFTVC-HFwiu4zok?-}41ESb{%swObhQb~^tbg`i z^5sWT?uaeXem*EAN*+h4;W^SjWa{6KBkCyrA1$6SNy(EJ;If+pMzn_E`Z!wEp!XkN zBo1E;x$sq64czsy;B_Lvni?uy0va#A+6qH8RYS`$jnOwPfryW)t&v4-yz*GxSsd`C zKanr(#Sn)CWZR{P1A`6`UEcjeq6(ar^qO`ti30+Pz?{lh#qM4oY_6@+l5*`5Oo5#6 z>Y}Dwfb0Zef>&_T3%AAce&tWSCv&2?%Xp~WiY2GbD9?j})33H~vU2=x=FWAF$2z7# z7=D8#U)t^9-w-YQvA3mpr2LFp47Ab!*yYyaI>K9-xtKb4iW;+an2NA3t#MR901wWD zd#({Q(7szF@jrGz>1j37Q~|Zp=5_w(2q(9K)PIfPC#QvfJvrIR1AXOMUtJJXiLx=* z8kKu_C&VPAx`!y8fz>q(UsVuZ>g3&)_anXJLUAksUxrZSC?**z1E)=3K}+TlYgR+E zS4}W0o8WO|iOW+ey&D$bk41|8f-tr#3F-!n6Prm7dMfARXHXVk-Oe2+AkB}D3wy+y zlmua4(!$4-*qezewmydKejnxKNaAXvKiV9E;?4(vrzKM#pG2mw+V=lX?8Onkf+v$B~*JPSbGq0OY*!TnpNM}CbgKZ&?I%Ft}m7QP6;BNnrNcIMoAT3o3 z9$rP?!@ko0HsY2^szFw|X`CfWPw(iDjPu53F~y6uQn6IsPuJPN9XC&YNFl>e9+lwO zVdo-tAk|?B5xZmuz9PnO#i}a~{#o+g7SQ>f*6?fcH5a}e zd_2~9Bo41Cj%l0h7NL57*b8hgP-=3s%|{B2K|X=>o{^R-TtYLRY@l3*2)30;UR?hG2>*$8?v$^NDy z6R{1%TS*x&FCj{%U7;|4*LJ8R4<9Y~rz{ptGMlR;Bdo_S^Ot`Utk>PA=OE98j-~)L zuN;>_q1;?r(tpxM@relX$mFg^SIf8V* zJ#RL_`n{Ir&M0ll#+2mqWdi3Y@J@u`vzz!J-5pxUdao?Z=^5=6w?Fw?^bKS=5+UG_ z1Jr&!khxZ5?PN6Eu_%c^u?$}=kE!iSWyDv~rnBt& zLP}&rQF;YpWZES_oxmVaV~eKUfciwVWZAi7$|>b}olBO;j=$HN?5okOQ)=; zMeHI8JYGV@|Id7a(i@BpVzk^ z$xV4LO2+EPgNVHL{G=?2tc>+8p}1pG0Lp2B+oa%lUo;iSKJ-Ma`BHG6#$SOrSjsI# z2#&8V(7+bgLS$!tz9o{mzp!l2jup48<6cu@a_ucKg)r&bH?Jz(<_NIvxqws0AOCq* zN};Fw zOs+*QQ>{sRpnEW$UA|#GB4>NduE5PeB_9;+?aE7J?S~(Sl^=!MWX=Ei9Qrod)4Jyi@VFUt_;aVs)7j?4D?^=4@IOTo3)Px_>`4(Y84 zZzi0gm^xNJFQ=F4qdS7osxOn6RW~=T`>1DVfKk_Dr7CNl;Iwi z?Z<2833vwWrp1MmIe2d>5b@@C`B_eCqp`@%>EUOTO~xMOktFr;h;*0XFfw#6&;I@5 z#AHl?WaW&7bAV{;^e85tG-*H;9{c;3@>{8B=;Rpa@`sF=c&wzaOU$=YN67=+zV)2G z!K}Ov*{hzG`K@yCLddhvw-;%oHo&B+EWzcgbTJvYl!RCO4uE4*{qkKn@wz$g3{#xV zr8HFZSn4CCX9(@oG{cDL&YtxMbhxPhuZQ=*4$M3eH^U&E>33Cxl`{U^Hvdi1i^@jQVz5C?_!&*-ddJVPyS=+75sinuT&yL0_|H?%plsS0zv*tV{zwcq^=lj*db50g_ zG&NSM0ET7$M}gUIRkAL)5k5Fi$u5NVtJ9IJw8MCapgYT&eX(@8DzymkcW3NDs&fHm zkI?e|%FRvNHCsiHaOM@$m;R-0h|mqLge<=2>pvbl=0dG}O9}_HjClOyh^gVqd$PAU z+lBtrKy9Lj{Jx7COeDZ4&SOvC{E`+PjJNT0Y+ljW%^TrP3#fnOdCoDs^xKvBfqk31cl<8M?06}tg}goClpC0G?aIXwZlz2$&1Vhpg@8r zq#0j{0Q{@C5S|z3Pwc=B$jnE(lcWM#(PJ&!0^IK!nGuFO9um8!qzEsDScrSI?fKOq z)GYVAWYoFhh}l;^=2;n&@VIRQ@cd#<%}XO&x&*LfZd9Yk>Nc*0x%z10m2FMFlCXMj z{0tFFs8jc{mGVmFJ&V<5lONm)T3(MG{-dwvucjL0uX3#<; z{0%qmU^0!1Dg$WdWBCeNbDqjvjEEkM+& zH`>lgTHN>$5w208!pbDdBz5;`-EO3`P2TUxJdf8Xm7a$pvDn`v5TM}j(n{92osP5lQ(uK8RKDHwskhGdyzdslVfuO-zE^@PvqHgC**7&U1Cn> z=wVq?K!>Oa;qPE2=l(AjfSX>qNySRG!f%;tMlDVuZI`rNy+pKU`VAhOHHiHjVzw(|Vn8>rW98@*sE23V@<%w>=RUzHH`OACo!1@|DG5Ld% zsAT;L3jI?XCZ}y{RD6NbaYuc#Dx9L0+Fp&+cP<_yHlX4zz#H(=NM&8-a=s-YZ{q-E z*Z3oxi4wLXNXTLF-#AP8XzaD+UKXg1JQ@2IBZlK$ung2304~AdqjADi1r(G~eWMK! zW2WM$bLRAd??J2^RIQlg9^_yK8A|sT`Kp1PdiPbtTjYiN!z$Mji+RBVE(<-p^VLbq9f1m)2{f4x0+a72q?17qBCKcSEJ0#J$lMk(&;Q1g!XuR~!C7l)|R z3@K?_aVIb-(>#3wWQDKEJN#?3W3uO-(dM75k3fny$=B^E$B>@GA6KgnuiFW=W|Huk z3Jg(>^Z(#0S3h-~*GCtZ-t0h)=|`1s`hs{G2;O zrr7sGZU|`w+nJ!k2vR9x@{V2%`V6TUT9t(8)zqNbfEx;OIdVE>-J7R#cjjv61iuv_ zq9vo6+p*%V?Fw-1XL?)Wj49%YoGidiqc7OaV}<~Uyk4|iyMY{LJyuUnnZ8qFe=1@~iQ6FRHW{1U)&ZyTdBiZ&PX&tH1RCZA z+lwcjo{jFxl3neYhcQ-bj=VAna>v~$8IIq}lLOt<4H```5AY8c!l>F0y4oHXokOTG z2$ezt+L~UU3)t8JX_k53@2L1s>pFJ>#=?OytoLJRZzkbdO1Z2l!n`{wpV?x#HS4>; zio=k^ty-$(wF1GlFb#QLAIrhGU}6Wn1#&%S>pSCR3Mjc1ATweUG2@6A-U&LHANjCrBP* z;Df+!&uIPWQ?F0^U7z=ROy@ww|D{arjH4#cK!&t_u^V1b<@2vG5UZvF!R(e>z}PFrGAvqP=wlAFbH(E@YfCRd zVD7VD^2dtrOo-@oD%!av2ylAHnH9gTGo(N7(0zX-{LGUn?;be%;@>=dHHe=p7r63<3`y(FvV(gV$g;O5cez_$t!H07wXibG=rlaoiWs}OQeif@Ki8NSAWWd> z-^ig7kmlI3RqTd+A+TK+ryfycC{Zj)_^Uxj&?Va$wehs8o$(2mi9&9CCQtz(tAM5G zEg_@rAsMW~d@q_E?BFb)pLM~kPGQP1Xv`#DwThw;c%K=PujSrDY93#Zy+zx&8EmGq zPQ6UYQdTk7x((%)VWzw_7nd1qezMbduo5&#HmuOHj1>oOAoh&xLryw1F$xA~gVc%p z#odt6wwbmYq8%9{Z z2W3|t#hF_AFXYc3QXz3hVVXfFj&W@+-$#UYEHRUf$S*;ETYpC0hXw7N7uRrAVg1W9g zLuR~1Iq2HrAZGq16fUb87Dc(YTCrl;dg@Y?abHxzq+c+24oYRc)jKo{(Y7+ibU2Ji z1d&d4nsGVjIhONbow6`KK6G0_CYo?K}>2^R#TFY zW}$>$9e5YA`Lk9g>!z?zFsi`RT_r^?iyy6J+vvR`d;+bN-69BB3+hIx2k8wt8YSQeDj;rXecSER}Cq6o{J`*yV^i6V}7bI zfYTI0aXFy(B<@HO<=mEcVYUQwwqs6W5}r~8Kt2|))8{wA2Xqx-2seVdKw_UZgvqYo z$Hbs;xA&ez>8cg`$06-D+2nyne_zI(p#W&^R2M3D740d!RxZvT{Fs+Dp$OF+{*e#) zr<}FshQx*Uy^@Wv6g!UyvWX7nPQeYYXB981XH|sgbB*cuG_8AN+>(-4=?`#^v(CpV zpP)kC2zhh-j95)`Y)yi0bcr+7KB$I|zgXdPnZvv$i8jhMY$J)^2{WH~fcZ;nNy)fE z|KIaA!$&S==+!_JwkVB<_`?9zmh#_Z5)P*=Tg?WLoLFu8r{%88vgIM0$e;>ayBk^kqF4>+4aUz< zLyD@`ICCKOyv<@nJ|U*L?z<&iDmCZKY4KE zC#ZQOPw%Nlet;nFUli@>Kikp#(#MJXaW#KMaox{aHIePNZq^6C$0Zw&$rB#Gs`pX*JO^Cq63K#PyOiH5JQ=rcCLeq z7x9WbVFlL7!_%ajTlf&%20BE){^ok7?o$paG*2xNF5Y@p4#6^lGXw7_gH9ni!HjBm z=4Wt$9G2<7KmXplLptg}`iW27u0=g6W|0?FDz1l0S!hh2-gzz?8s^$Te1x4QUnS^V zb8^q-NZmz^{SJ}2CxwHF3u;ZSCZAlm5oVJG1z&>L`Yhn30XWCmS6J@DX2dt|t#Eb| zzsQah0EuVH!vh6i!aq-4h9dS$aRWJzuVw|`R4dvot{D1U8pbOcf8{%lP?k;c?-Qmz zZ^m64xU-*Vr_NVUqIJms`24|fb@~jP zbq%pX-_Qdy=n6O|r3ahgd!~g%u#+rOwV#;nBOE!Vjk&OLi%rA-+iLk?eWV&bSPwZ4{A=hjNC*aD1=6aio|a+PCw zS)`(XCs_K2EmDc#)SioVzqRa(_Y(VN9S4&nq??yk9Vw|u+t*-jY0PHdUeFPS6KG_8 zfR&H$j^4!};q!KIi&mbe4B8V9eKx5OJOcwW)N!vLH&HXoC!1Eo`gHa@@~Dxf zCurLd_OJb_9yvds^S(Jr56+XByK533*d!TP8mbXzJKa=Zo3f)gjzFMSKngT{qgNk} z%mh4oux)&ZVWnP#6y&{s^Pb`q>SKmt-$MC0+m+28{;_?3cr7(ay{iHi%${GHhmXD< zB5RS{OYR^mUlwEY{tfFlZHTcupZeeVvf*?4=UCLi*};;C4O8lYO9ntjdn{pXWglS{ zh)>Rn>DP~7hZA9s5WT(cT@qRSdMP26q0=nW8>TihH$DAV6`AEGQ4+-oRzyDt!SOcR zjn~9tgzn}CVT9h(4NG|OXaKGp%xgE1Cu)w%pg?OHyst2Nw zism?uO7n|n`qqLTcR#X!-77@>WE6A!j*@6<@>?!@_6wV_G`U#Z7@>_#x&xp$5?;7ItvQCd^kbANw^+|By9H zzc)6NgLv;)T+<6x%_rd+)6jn`@=HN}bTVY_AkKd22EXA({i$9c{!&%7nJ+(Xqlf_u-|M?1v6G8Ve1fI$Q*YK1p8{M~D@js+D%Ihhh{ecClwyW&w6 z9e&?~%3^jJ6?1RKgaJ>MQZK^h=4lftyHrwR2MY^PQ{aa{t9b~HU6G9jGq6Ea=1N%J z)ge)}v_o8QX^BLk&GH5u>-PtkykXM_(DLA!#nq$ve7v8hlOlY;p1EWttRX+tMc zkr(HZO7%sBn~aq61>|#iqv656RE$)iBnaIJ46~zCd_?5|eXX%HcObpH@*U=rARj>} zRnNBwIx@FUMwRxK#HOj4q<9-5BZb4G3Z&CM5shBv@z+vR|@KqcOAywm~~ z0Fdi)DFXNR+{UA|J1Mc;l$OG`hd_5w(u`IF~-tr_5v%{&4Cd+osf5ucR_8LvrC-)_Q580X0wJ-(NLst z%5B$irfPdfnWZ*e70Wcn8g_PkS6`IBr*#n{)0dCkGbr4m_DZN?U5^-$NMXGN{*hVB&ke76dr2&qd$;ef#ALaY;%MTa)-!Vt%&smfx*~r(oT`7C8%lmI zF2w zQ8h3@p~4?sXbW9wMl+XSi=!%hv6=cfj8qoPFn_z~f`@oBf_FU1L@oVqAehrpD11~2 zW{4;Nc9*&-<_kS~=FraamI=t!Gs3IIGW%68YIOLmcFp|}`+2kA?-4u|yDBQ4a#1Y* zVyNb1N%*r8zOIR_pect?Ho-rnZ1X4lV7T^$gZY{vC#cZL=-L8t^p=f8R1RJi?mZVS zcB?DZUfd-!@KPc9@QuymD>1N22>ws@NfRr|VKO!+ksN~sYIozZ=hyhaww=`LNr{kq*xVRmL%rZ+-2kPv-GEMLTz)AS+_kjKGCNT}^ zBg$q|gNecQI^iLugFerSuNA?BnyG~A{&c6tBlDBqQ%3m6+jFdAq{?D*EbFu{`(8dH z&JDl4I+CVvL1^!*tw)7j>e8i3b_4h|U>fdK3A8)w}GmoFo3`d<=_9wl^(HA8&LxR8N;RNC#H*~Oua!6Oa*5BN;p&5 zH#YlM`{6Bvr-!c67!P=%9!dvnDRnqC{)5v`dZq{f@DIPe^`5AU>7iH0;8iZ`DTfH* z{+!aaUqPiGJ0|t#jCFW_g`w<@ILGpkW<4Q|!=5M$p20esynUy-Ek#OXt%JiwK#$Ka zI>^>J1F)kdDN$D<^^NO-sl2R`w&tg}6S0(}4ZdhfawIVTx1Qt{)#6wh+ls7Gp$6^Z zo&blZoB$Wa__5{~%B<7Q{D|S!v2M4^%GsOK#|#|95zqg8*@j_&ut zOX<*g>j=b4@Qh^Yjns-?69s_U`(AB)sZ^5A^GDysUDKEhNk79o{D^uHF&s(Dk>wN? zzWCFV8R{!WFPDi-AYm{I#RDoh{Q0+ z1-*0!tGZhRRC>;U*DQr{5s76_a^~l>AEo+GlsDXK1iA5my-=BmrVq~9?U4o2ULgM{n?fv0Y%pI2ZFI3Ci%MVGugDP+5`4LIa%^v|B+ zjwELRcFy+OMq`r&r4bI_IcM(Z-Aze|?x(Mg2IAU*7z@;6E$$G|SO31mfm!Tj05Qk@uNL0x`(=;1&_Kg;hW!gqcT|0d+I%7G z;9A+VhmH&k_7tO$D!-jdx)I-Yx|RlJ`6x1`QKmY?8IxvSXsA3oi-(N9)}jbaL^!W; zk6Ug&SKPY?0xw(o?VZrsHk8aCYkX=O_n}+!Rl{gOlcK+C53y^NGhFGyqmXn>Fr*2L zg9!k|s40mC3xnV2qWwmt$85oe;hDpfzB1SQDJY*%GE{0?bg~tDhq~eZ>+zJUI53PJLV1>5NPGo4etsdgy)a%D2l@; z_+$T6?N)uX0)Vsh!}qbk7CDSRrI*m3u6>QL&38DHLVVLdaz?M{pHhQ_(JW)SsIXQl z3mmz8khnF$Xq0uuxQlL((870-0L=_VHkNS$Bfe7opu^5hQhXAWAWhvGJ5e!{QRC#H z39eyMz4L#Odg9@Kd9RbV)$H+nW@@@)$8t%wt7TPa_ds1@4s5>0OnbK7l^eWAUmIuc zx|e6<=DKH+D~GfmiHF1|6EEa=!K2#zX1@L1O4?4PR@6%8CA)Zpkk4y|i_pPXwi}uP z#`F4XXX*UK$@q}z$H|^NGJd!l4ppKqudQ-CK?MU?&CI+CKSD*ZHQ^zAi8wyov(s6f z+DIkZU4dYjHglePy(G%*hfv)LDuwF5zx1_|9lM!a8H;53b+Q8f-4tT9v~yND$6%j7 z?O{l!WA{c~02xlkjeh6Y#t+5*-JSx#K@qaKbK`}pif0eGGv~F zwNfTemJXoSoYR>#e(4X#W}ps4x*NJlsNd+Kn=5xPLgFBJh-AH*)L8?S%41gu+B`M6 zys~enJDg9n7vy?Aq)M=J}ysuURA+7 z*YZ2=_5m;tJ)r7Um9>b9(eOHnH4e9|597`MAd+odI_kIfJw9}kpRntE%4f`CWA9t6 zn#(5mr&ok1sL?S2n$N7y-G343K>aGdIsaa=%@Bu>ek5eS(q& zMPf>pQzgthOg9l;9B!fTXB7%?zE1E#IAk5W4pa@(sf#h8R@ra zt3zkb6Z#J$luoUP*?aFr{ zH>i|atI#a4vRp=+_%UIb5z#T2<$2->&P|RD9|^||q+2;y-d@h%TTix>U~*B=}1V5G3%x`DVbxc^c`b&i*sGYM|xfEW&Xi?$C_Yo*IsAl zjpYjO zU1QSI)&VIWgDq1>{3GC2TbquL$@^7?$r-byG69lgUILgUu7V_b{w>t5M>T6iY> z#*$!7vb#?^2WVI94iHxK>@TOV&y}4EX|32x7DV$mN@sDs4@ll=Q|XjxHOLubez33; zYCMNMR$q3HIrRlrWFvO2f6b?{Z!`j|q*(Z$iRlF1`t|titm%BNEFlX5z&hUdk%X<1 z0`^?ts=l~DFR!RqzwKbU=-U*sM{WE1v`Mc|kTqSw$gJH6R&#RT2 z*}*ztE(;sGHYJY|awBCJk^Hfpck?o^$CXb$22rl#$<5)hq46$_22x2@dkv}_|Bz;~ z?tIklcMb7kZN^F43ODNZi2>36a8ulsdJSDRk69N5UUk79=d-wOOZjwxH?bPssBvBT z1Z*`Z1}qA8TddsZ%mK@41T2Viwn4cNL&HM>IR8oP-;+k~^Y4{=IWJB<@wn}*x5;*= z`Nu+R;?5+03C^_M(^!%IYFahuA_T8y38wKp;0^zO#ve<+``j-3ST$4MrSai6vy$v= zf=FzRy8m3Mz!Fdckben}szp}4{*?1Bk>af6WwxRbVtnM_)Oe$;>z*_YR@dge(Otm_ zQqE1H8K@uSnJcB)?%thzG~NYFcj9qM_9mi~<3IJbPn{++5lZYbdP-g9l~ zr3m2(^RGSSYTGlW0JhgYt~IRz{@?Ik3GdnL-OUgKifXGye!~KxfbK4V8*I-W4#>Ev z;yR(f&y-H9OBvTXfu=1Y@$exDy2J(x)bf%}Sz?=O##znjuurP9Y>VH+D^k zv)_N<>4cA;VThewThnqa>0(WmC(;kznB!aZgx>l>XVxm=qp3`MP;Bk*7{!_(1tj-+ zrJdq$KI0|DHqRy1>$%Xx6jD_0Z5!_)V89D)gua#H{X@dx9D38goKpWJQ@q?n^|i_X z@rSBDuPhw_s{E<=jnL?pFG~RTrTrLOJ&E@>8pl1uKPE^8Tps=At^JxNui}sK=K=z% zc@d<*q!zoMf=T5DtxQJzroNoss3?;EqgCi%9pzszIeV`6CHlMF7|u@;&D02!*J6Go zdkbdC;*Xz&SdwkyCXfiMxITg_N`%)Kdy=cqXpqMK_yKgx+}VeOwetWa`lhc528a9% zYmyXUPVOXmBQh>g_(>P`8ze-x0W^)h?@~`ZC4DNb(M9DDvaSk+oH8jJDN4h)EOU}` zeIVM}Wt)wHa7g>dRwVnY7HBg~Jfvy1W#zJgkJB{u?-wX{j(LUR`Ec^a_mn(Ip1O0Z zZ>FI(y5Z-a0^hvg6Whr>K8+PjSn4T zFX@Ns1+^Ez_-|L}ql{tx2H%T1BP{x<7LMZ!Z^5Jlf+t)xH48QwwmSuIRR)J^kiXOQi!|Sq9GRM(Joo9y1T6>(hPc^eb?L$OgoXz zKkpefBiZGq8yd$=s(4IUq%flVeoNW@Wv34X5EyZ)S1=IrVO(jkZs*rTmph4NM2m#- zNZi~|blg#sj0&%NUD$%Lif(6$fWYCi-vVTz#lcZM3t4}gtQ&PkEknr9zUHf$AsblF zJ>*w)4nVTI7mE&yi?EsZ#o-1XQ`4Sac?+;(UY}2}3=|yX{ zu!9baW&-9|m*aqy5wuy!#ah73`YFOMrkE7)Js(}ZXV4i2^9Q79 z$zockQIiFZ)TX(-1|LNd7(B*w)gTfv)o zwY$LKT|;%$_)WBLtdTl3}4XZ z9`CpOkgP)}uME+E-P2jwB#Y7Zvso2vO0`KmNpcaa&^(EGJ=U81=e5#T3bod9Z`t-! z1eBs@S5`K+`9Hs{EL1|#jq~h|%oy*I2ypKub_V<*{Ox+eZf@X-38yo>SCC4+zn<{h z(g6+-;;bOG%bRWtD(sO`Aa|SQU7*d~;S;m+g`mBah#}5ozBX;-_wLv+>VB5U-V#lU z4SwHj%`^DtEAWQclAIYwi;4t-wQVB>9S5k5x%g8&OC^VA>zg(~D3L4)o2gd-nh)Q1 z>e}X51Yf%X6nWtc4-TdB7&iFg1d@0J>^lA3AG~1NL-H_VTS6JKHSw-rq2k{IH53A9 z^&VN{)T2z=(74nsF^nITV-qJKU+0(XU6QW zt>r|ex$ zn43Upwng{HeIZEg0p8I3ZXBv+nKsG1e>`50(0xNLaB;deYG*vC7=nKu0g{V{dDn`G zQm0FrvhnsI+=9%{dUEF^lH^U6-6?wf))c!ov%UL3*qc*|2RxL0lHC?yFIhkX;nl*S zVla%Etw?{>B4f}Xf>cG7p#vNs|SO$@P3jDpZ$wE&s9zz*GpbMpc*=G#gB+>1vtAXR1=w$x-edGa_9suG3GB`+_%gd z;yS-*7(F1NobNg9#&71CvoIyCcju;)(Y0U12CzP1%|n(Z|Kw`?$_Yl^2mwo;Vx~z= zVqG%(e2>Sf4$qRTQ_)=}B44{GGF*5~lOXAr7_?b$*I71jVV^bbfxXrvzemOtDdtpw zC+n6BnQRJrr5TzbC;09peznKyCPc-n(_D4)cSw$>vt+7F@nL#$=V(JRkwU0XV)B`2 zqkXw&4J&|iB<(IpL;gp+@7GI-?~pY-`)WT~d93Xf%hom(ik~T|o*%t``ox6v`;Uzn z`+Q7nQRixM^Cj6`)V5~^!Klfu5%TQ|Oz{iI9TVp0U{}qb&i49$IQD!EvQYpR6&)Q> z3Lq|)k5~^NKi;){$QNdp*!6n)PtGy@3GC2RS?F@@;L?2x6~auvYUCAM6MK5oQz9w2 z`^g!8_j&OOWjb!gx+A5ZXOpG=k2u1n=)bi}Ww4iPn)A5&*w!r7K*GE+5tS3U^8m=- zmdB0)&V+BDrt<`1bLg`7ryjH8UkibNQDq$;3h z(Rph3_6ef{URQT5OMKRXsh#7&EnBGH(g1r)bxsJ@7~EoW+gmWKB^Ie_2b6*<$iH8R z@0)8nu>M#t^3RF!N1X|)p*>a1ZPH_`#Y|l_MBg^|r5vJrB=sD*d!fd5@P}iN=nim< z`)H}nCi<)sKW<(8`Mr7|&Q|Hz zkx5iT=|vybJWl`padmLbe#P6?=UMkez>mFKX9r~%VbXbRhQs>~YeJcQEn@)%piMEQ zk=-BChBOQL0-D{Ck`{DA^|O7Q@&=Tz*cy=~@d<^%V=ME~&@ zY09C1)=yTFjBi>MInP^P&-AXcPw8>~IfIavCU+P1qg0j#4th9RI#82F_LTwn_YNAN0FWD#Ffl)0u#q4Dp1m*Mgai*k%J7#H@~% zn73gNeMO$C?IrM$AV_cc?ld3uTV-|VU*jHv=$bzwt@wftK({~sEb!m|Awx36*pN_D z_8_VJR|H5rv~p9xh9=2k@zhUY4=l?*NACv{==U)xErDZR9H22U56LpRQ|+Zl4LShg zQ&K-0Yaae1YEoQWh~3@`>@FVo3sPsvtLtn!%c3L?`sN5d5U^Y%4`1)PU*?0Quy;4{ zDW6=YF>!=&Zev)CvjD%{1y4KA#yU!|g_4qef9pBZhqRR3WR-slGH{&)zWtT-d9hA> zAem)fFrV};IQ;Z1hh!iugph`AYRr>A|6Fl@M}We8k4ikm9GqX^ZFq01;@s)1ZC)_s z>{-PhDs@FX(q=jT?nu7x7K(H{Nq32{>{&w0Q2ifOUm4b9-1bc)ATd%JMoELDq%=s0 zNGT0c(p?)RFhE-A5JW;yx@*V~kdp2Y5F|&}7`zwW_jA0@a~%7!Z|C(tyMCuE^0u_< zmNq$O7bhlC64WeTQdvcO_sLsjcTrSn_nPb6rY$3F8d^xCllc)2zrxXR&eDJ2r*w2T zs0Ztf1S@F}0hhkUA<8tnD2>>~BH>Qro!oHA8}jvlau5$*uZo+!YSg*CO0OW9g0G?| z5`wt=GQ1$Gxmzm|%zUPAAlVtddTfj>(3P;oICzM^bL?=k$PYSKd!(+3UD*Uv~~eB<;1kH;(0$3C`*6^?7LC-@zhMB3$n{s zXJmWF5+4kN8~>zw4~Yzu%=DG)3f*?Wjt_J*{+BPXBK6F(Nj@jB?YlkBgwPPcDGFaxgjr0^o?BbBYxLD#u&G0LM;w?>f6YIB@R(-P zz_}z~fYffoLhEt_D$kz4uCPjTH~L`j^so3B{qRq-$p*ndf}lgIuW<*&t7W5(?7ym- z`$s5!8)w~Zj#d3^PtRDm0tYl~H?sKDXRc~hrxLsmmxG68sAovT7v`*#&L=yUYCuKZvwWA8yBfKfikI&GtmVc6@=1UA{a+khhYbn2*#&xeH1(1yQ;R_%hxtT#W6tZv>> zp#okxG@A7I8qDB%@ubU+u=V+w{T;cwrV7dX6p^I`$Q#D0h&Wg+s|6hYH^Xx9pry*l9xyba~)_F`R^LQ%hcb0uj1VIlCoq!jqF5Lrz2mo zyD;qSuwpoeQif@gh3PtWC1dlUo_5z)HO<+!Jt&kVLKeV;@-!6TweUMfi1T#S6MqqVCh7j~lybDL0 z@#q#0JawjEnr&pcI<7uPl*T-2{C9Az;Fl{J_^28j!-u=BmIXOJcq||~9^<^&b>ih1 z$+StUOYkZpm%d+7Td}p~uRD5#n+w(%-JCPoV^X=O3e41Cy228oe;HAQBr1W&yqQ~* zI_$lr<-gR*`%on+;H)hTie~(64~!H?MvK#n8T-kxKec6FJ?bC?oefZ1dHrJ)0LaJx z<7tL^E5T6@t7`cmQ|h?yPStL`o$oJ4g8wcr4C{3>v{xW1MvHEw=hu3ssumj@&G@tP%@Ew=_`|>FV0mZea5jy zx;SHD|MTDOkIu+zLxhj7e9GQqy32MbULNyoE??|5Kl+`g&Zq(`=owMymM9_jLj9B6 z25BpkN^H$DWMaXPJbi_P$Pr5jpEZsoS#zRP%B2*SJ9kfJ{Rnoa3c^KxBXlr2o`9(n6o>WYmHl zwToiatNsDw@$1sL$HvSQ!a>pzVfM*u5CAeGY#cM)N)X#{H*M*zovwqtYZKT$G`vva z`iyrSw;DYZ^WIBcQuaR^I)O}LD?)rF^$u6`LqmB!IE+pq`aQlp_d6pH)b9%fH>f+Z z&e-Z|oZr6g`u4nPSjOs=N^_AQ_N({17HL-S2DlrUQkR^%_jgm{Z~F!l@;C9-!LQw9D0RWVzqhvN(!2Pm+U0i07kIz1ak^crv$CEE>v-$ECW$H zOvfnk`v-2QGg3%Q|1FQD8_}q0n(xKh^JN%;Q*FeJwDgwx0qtwY9pfU@}XRXPlBxpo-Y^Z{jRGq4R4OMNe*NIuPnk~ zr2rUa)5ZjKCixP^Qo;r#l5+dYcdE?21jU?sFR;q?Hr2)z^<#SicSrwGjo$25w)%mECR^o!-RUz0d5Xac{sn1=t#v_nc;Na}p6?i{N zKMDWbEU7?ecp~iI2(>KH(24W=*~iS{ZI^mKRq!PlZPeCfZriO}BZXM2dEQv+jK{#+ z)ra?x$9XOoM zSc3l#%SC6z%3RC;$}#5Dz6+;TwWl*Dp_bA!LB7z#R&kK2VDT&C0@4&VWBlF=C0eDR zFfRuuT33^;z`t&_%qB<~dG$xK)|si{*j*F}K%nHS6y6{Hftbp!R!=#ElPb1eCNnq< zjkLE#K*q4l+>}x3RXK}XWC~C_cCN>|E81k~NWtx0H4$$M()m)ln6u+w<~vT#%%=d) z!RFY;bj@caHk%u42=v~28!7fgDFMRL@$NqF&tZtz}?W>36lkWFYnqMk25Bf9*ah=O!r+Ec2fD;6CAV!}y@H*qE z-t6G#=t}vnxGH0G5XYg_lN=%wHl&#;8$5h30+je}RvKaR{+P-bx$kTb!X*uSZjn~A ztt{b{XKwMa=orSHI7~K;xPWCWUd)}O;gLzl;NiAf2WIu;w_HgznnnVI_NLoNFj80N<2K{t)1t8(1g+j)Jb zX6<)11>q~QwUawY2_XJ`+@|h>SuU7X-SSLCN{mVbSdlG@hGd2#nEW5IdhLG zq6hTCifAlhmChHkt<7u;mHSXvE0Iw9H~O`GNNCd`(zOb`89G@gmm9GtGS1SokRsFx z!cNMW!vQ7bT5qTo%8{=#R4Sm-Il8L_I|;}aSN=&G?j8AnCILfnchu8P2zj_X>uI>X zzxNX%CHW@<#BE+aqE5;tu!VDhbs^8Ud`48X!rxafEi#yYuVVgdyZ90t{}u9Wvw;z> z)7zMoZClxTY}8anSave&0l^XZs@94TEC~h8=nS6FStHI<&1qjt#xnh>1tap45ps#V zVW4XWvJ-U^IL& zPF@#=mgG;~41+{^rFoZ@7JZIP5VuwxG1z}dIPyv_P2AQvNN;o&+V?DwY6i-%L5j=t zJujO{*))R(!;Ih^a2nMtYaIRDA?=t2@kf^Dx=k&@m#sUn*T!}EG4Fnmq$8-#;R?9! z_}t4=7cqQ1mFZGG0LdgoKbESyv=K!v2q9UD9#qFJ>#|K>b7<{C6HnZ4aJa)ur%g8F zRdLWp-$xSJbe8l*c1&X_^@Bz2v$Jil{7fU)*tct?b&hB8yi7>BtD(dfynBlqcP4&3 z6}YPRzrQa+m4DEndND+?mDwjCar>AJF`B`(E=%I9iq4tYFA{vtuwy<}xWENBfX9xu z;9FWREM)GTs3AtG>*Vcn^KcAZ#ysBtN&S`nm@G^RbE(MVur6E3%HXy$WRcqOiUouj zC@|3#$oZUauhMwE#}=_iSS1AJ)KxvwyQQ4Ebg@Hie=+~0nD0yKazj0&Ke}&kT&L)3 zL2tRg*9A{rr&*_qU!WLg!`&uVzDP9tL2t9_y>ueB3OAVGHhrO730o5iTXGy3>bt!3 zi4><}d=!8=_I{FjZMPg@uWQRtNEv_ zXVT!h)+%19+yY{d=(HLbgSyybS=8?gsx>I{QGuSZgRph z`ND2ojEjegby|!of|k{^>FXx}VM0h=goYU|l~y95JdzgLK{c?F%UvrN-)zj2b(=1P zd^T8iU0&s^KiF;_xoj=?fD$sB8ei&wa4bq37Z$~{2T$6#yArrG{@vEOi?W6(0sJ;| zARKQ05B6Um2}J1kj+;8Qv#f6qP8?6dkKvEFbsar5j&a5eyi=#GoaelWTf%valM>x9?$)7*E%sHqX_aN7)ZP)pKn<0CSQnd|3tE( z$?IGa^>}i+d?vaqOq}nL?USAc^S-;esx9q1cL?a~`F~xzH9+1D)?F0APXeg{E7!3vSjfT*%?HZPDBR1;uTt>} zZ;@@L3zLsrg1}S4Ca1t!8KoMW5@y(`ICE2#-=Q z)^A3ympfUejfg6@s?Zc$GXw8FE4HS83W$piSZ0fy`&p&pa4bB@IIK_P_Dt5Pfb^{J z%ct{y33(kk;9;x*2ZK&|J zo&CgKY;3xT*uz0CPKBE9<*&+R6;viA)SS_niS{N$Y{My-+KcMEM|nBW1QEMLe>-e) zCNbDm{(Z{|-l4brgDtXBE`DP#5&>a|Ab{h4_^SzWF$O(a+-%IuKfO(46|f+)M|O5! zpNGzd=D*FC|1L*_9~nUNVt4v#hA8&|Kh8PMR=5j*Mw8phqjFd*KCCc2szNpAa1y1O z{?Cf(0!!=vOX!cb;v~f-6rvUdRNp`X@^K}z>oBNv{Zg4amKaMVeVuxIQIWnzfE|xD z$xDcNgo^fE<7v=u-1$8742?{F5MQ`g5f}3-(6zW5Oe;_V^*t0d7XPbBsTSz0jSBZP zIp0+&Qc)g?gnSzIwP_&QSx%~T*39G{+Qktl18B_Ke`adY}%#_8vg=gZ|aMxrQp#BLlt`GwC9q z%TIeh6(U|en5PFlXz0p`T#Tpgm@5Ga;t!X`0q<(ZJ9mrDrkJ7j7~hX+(y1LKR;or> zMWmpz@CGW;xM|}p0Wl2&7&hHkJC+Iaz1&oY-UR z%e?O=atMZ3-BNn(?qHSkeP34#41pgEA5M!)9CeC!2m1zc(N%+sZH|wG$S)U#Ne}Xc zZfV!&cp*s-<8ioK&h^(hp11Ev5>bPaY3QrSR{f_sWD@RuSHs@4nU+%^oF(QC*wq)| zDGdpS9jOe4KW@zb6zKX72cwSX|46>gH4sIiTF>#op6FC>_V3y1)ZbE7ZSYA-XLO<^ zmzRLgR%QGgq6Q{mVzFpL(Mi${#50od2jPQGxAnSXLx2{a8P^M5oMPA6*6Me^c*2YA z_T&~U+v9}~FAdnStmM`+?9~v+=;}533Pt9PAIb}ou7!ezyDnnvuD>;pDtqYrny)FI zyYpeXenRD(@?4jVUs$bBuB*VW`p}i&%WU;gvCpJcB8C%!c^%3Lyynd7ZOd_w-QpgE z+v3qQU#hVQnk_WHO6cfAf_*A2NGihC`ALc00j_ zvF=?BX3MGI9DOOR^+7r;5n3(18Ew0PIkb=(KHGC1Q~X9vi~5L-{(FBu6*`+LVHQpK zLEjFODK>2fnR#fupc=01`VHu*=J_Xhpr1n&_BXt z6pmcjTq*Lw5S0}Tt(ghQuuiKGrAdt|-lCP6k>uZpfx7Vu2`*%uBu4#~k##fIl;maZ z@;%?jdkf@SnB(#KT9l9fM%!X9N5L{Fl7DZLvVx&%(U)#YsC2ehZvz@PDk?LFDUvJJ z?a6viT?rET}3dBbuqrVYQkCwTtNM#+hwLckcihf{q3pP z?KwKfMGCli53{)rVCQ*_wYtzrNP9T?w0Za6=NkN94hM(nCI$a`^nKntgW`B|XvZ+| zQ&da>O}g`^L+naD0sV$@W7?pOfSfDEp(iT>WBUX3psq7ft%t+e``Bss=Q9~8EXt3b z)D7twQ^INBui+iAg8+sUjq;tP2Uh@8Q!|@u_c!utr@rH;*xYiq@Fv3bpM1c1)RD(! z0&^O5XPRl9V03WUNi#CSw01@T804mHNKNnB@n#k?{dL!bD+cbI{(l>kQC1l{aQf!}N+TL~Vgc-8qlB9*P%%o#!oPJn4eNcyFVKz!a1pF}`8yK3~ zkWQ9aypVf9HWd?RV>TTFDHk~{p_&_{p9_wN+4LK`fSTdOME`=`ezI&g2-63hpB}3( zz#|J1hngBu*j$zZ>urt<;H*tOHpp$=c6wqh0Dm!R*)CSSa|K={&%ZZFmL?e57Wxb9 zbk4R}l{pv_EPzx2zgSX4T-Ydu+$CPUTmL-0>_33TEde~N+Oa^~kf|RzZw>VP76;Ad zJ}q?x77?lU9O6}@JfA9?I=}1T`a5Ewf)1qd$dpR}&KBPW@QKuaVu)XKdu{R16xM9A zCO7<|a1T~GgqQjH80%K81O8<^1hnZal0odT-T`}TCe~rqAhk!`g&4-^$KQ7la2E+6 z*;?NAnaU*Y_cc`W4dF|$86RvJur&E;@8@ChL9S0$UpF{E6gG`brgkmDjz}gc2&|gS1iu_ZcW=z%})v?mr?b6Hnb#j#{j2o9~`?S!y-NjM_c>Vs{I|H8f`O4#&XIspw}G% ztJFtX_5RW;?!{@hd0J5!+0SV;isxo^zFXllHyllM@$~!1&m)*n7BKGU2*YUOI~rRv zW(PlsoE{orFvX{hB}gYYjoaggQT@@^6UKxtB5r)%k0f!^md;db&F3p&MPq!kWGmHR3J@$#DYo`+G%9 zGh1wN3|yrzqE_VeUu}2jBOU|6JFdN}tNZjM`54TsANoqzy`nmTwd<`$h$4OS*CPFX zJ_?~eOgHA}Tu0lj*%v!ul4xa}6-6Jgw4jc>jeYgQLr;2-b!OoYB z&x=(LAftaN{U)$JRJiub1>e5T}SobkL({EpUyfjhWpa-9(&WA=$QyOd&yKi zjqpiihsf59GXIpy==QTsKeFS;$tw2M{l&PG>@yq> zsAM~kYclq3Ih`R8#oF5y(@~TpV?ODnwXU_Z^z>JGo5h1Q4>l}ceT_kTW+~Bge*9)s zu;@}SG-@`H_$LW$s3N$Kk2{q(H7z&72>r&IUrnzL%%}omj&CAQU5SqypS=C80rP+9 zJ26$4(bu9{uf#WWV}+u@Mp+4ldeg^?gr10ZZ0vwku@3-6MbW~JfLkX0Q^v^t`Rh;P?Lf(kF(XmVTlf=e}d^tY(Ah4}w^oVq$)STD;a^Gt9 z!|%@@)O@aR#jyc^q#N}ISK{ZXEX~gGeRae<&h};*v&zm7hr5I!Vf+pY=UrS9 zktd|b7n1a4jNnM6`#}Mli(+Y~RhJgmkfhN1eBR7Y$9hhci1#FzhH10&c@+wHyl_eJ>^*JN= zthsotTG2?#OhpXz(@}e)7pJ^3ES2og+1C=2o&G9+#z#J#8+G)W%#h)$mKH*>PY5Wm z@!{KH%rt(7reiO$plL5C16UskXPxup#=RUJoLf#a4Y~LHKmb#G8v)JMU|yjvN~B1- zsAFd7oG(#f=tvEdsCI}_troP8@)_9&K8>c~+&tp6n0N=0h;|8-83b~YhLOnAER)ia ze_HqLZsJF2@fjls->F~^A+VP`9((667DwI{c}Xm*5&Gm3U&nhZ6l9mlpVUu23oCXK z*n|DrzxcNQ48N4sdep~w;jMXzX4r>f)9D;eA>OpFuS6k{$`9+pYoTn)WE07TLpL}xOk31bvbcXxU-Leq0OSvwo4@qLvMW>i1Zzxr$a44l#Y(54=daRZnCel^5* zGevcRtHXvabO3LKdlDSb#g8n0hr#DQSZ{r-YWJ8(T*#pCdJvIwJZiOsMSONhme;*= zCx@~?k@(d+u4JOIN^F<;hv%HjTbKX7XyAz!4w|e#cd!Gt6QL~b;JRUEWv;uzT{OaL z**&*~ER_X|5ny^|!HdtEDjb&Pc3@0xgJF_Zd}B zrB>y5H0cj49a;k^uKxT36VW*rCNeB_oJ?^M|8D9uE?#C~)*m*LbWn5udm))r>hhMP zKD`S(3l?#Xgfqfl_uUH({I=GeJTmW4)MPT4V*EZ-j?`(QTONcdr2C91kAqpPa+x{o z3Gk)Y*9tm2!B*O6j)GR7S$fU28ohzHz^T+J;f-*6I0f7TZkoHR<|izYGKs?6J5FF$ z&FL7rW^5+$<*#H5dEXyyjOUK!ylk00Z&OVY=WU{UXT5qIIq&A}gD^U=ekA$IEa2wc z-zDCSq9lII%)Lt{ua%d0e~&jSXX&mHISw%m3toa6ROb?YEb$FU%<3k{kaj=B+i^`K z);?R6){%9B@%qSF9#HOUx>*9&cqu)UsLFX!@#?nz^Sgmf#X36|6vhHk8 zw~h+Gfy$T8GWz|K_kp`EJ}lZx%*HbiA27K(=N?=~0rN%Ht!95sx?iEJzK>4Bc3iIp zrH;XmkyjFb{1$=ZBlgvlN1i}v{fH0rk@8@*Ar*V%9kj6B9XT0keMmODki|v z-~ntXuW}kR?rioARJ~4W3Ns-KGhytPlKNT#Ytv7`4Sk~_UAu<*?aN;4$8Z*O^Rnos z7X<%Pge91c7>TiG{vXU3NR%Az!7PodS_-%AU z`7p`W>w(VH&w?qqU$AV#ZQpl96KE2UvsHCbZrivOFt2`Oj>#EMPo9aAEVOEr+Iy2x z>-DvCs6~6H`Z^(idv}wSed!~Nm<;qtA*W6D1Kk1kibS7!D^c>A$%l=5(;^tJu<@~@ zF{LqTRR+>k3Z-H@cxAM`keGLJ;E@Fo_z(N&f0&&E3trMGpV|NUYyX!3QqQZ5m81o~ zs!&@h9`hR}b?)?d< z8Zvy~u%cm?+-9fiOI?+9EwW^Z?EdUtJS#M^eiUqSgGzudkx3%i$Yel2hz6O4=^9^kIQX=)@{QOu3?L9K_c+zd; zWb0t6?A(7vC#9(6Q5WxI5`lSHkh$W!#t=+Rs@+=ff$>=7aR0AAp?jfbFcS-rK>7rp zI!58~vrWYA?lKuO;W?{aRKwcC)yKE@l4qRk2u+4IOG4x4+vFp~`bgk|EYeLUqIx?G znT)BvjUasQv7)se##k@H>QHy=nP%nAlZxe1i^27Adh|=6Friq7- z*gSA9oW^cy$j?O>XWnMiw-8tV&8Qi{#iw|UF?RdME#`K}zazP>IfhR-KcB#~o<9}6 ztkP_|Az=F%dwcas&23k&v3Hm>jCMjTnInu8`O+Jpfh=?7gHs<)<%HRM@+2u5N6+3{ zCN7m<@$A8{mT8!}cW~6hJ$G@Brw1?F?>oWTL!CX!dqYKHQV6oiPM?r+*?ATJp;*$h zUPkI(nVA%h@Ysr38aa;o%36a;MF}#^smW?}&uyFRpX_U1%Q9hfWGp}>n>(UQds=P3 zq2Kh~K#8r1VQi`MMt%=}NN}QVUHhQ}$&;SnA`DDQ57u9dJli~J1BN^m@K-F?j} z9ON}=jw|^#jl*uESVT0N<3$N3R|>~%?C^R!c+2Zaf72M00zxMHel+gal7>u_d`f&j zo;yL$K1qjilwid#X)Jyh{>K01EP7<xO4y zwkgGaD6esuxn-eq75teEq>)l&Wf|V6c4YD5ONYY&@X~=cBqk}&g(_YRdwRiWOX3>+ z4tj*so*H?7?m2dPN<01yv&US66w-Uu?MC{CyjgICXnzoXkJHi3W5-UF{w5vGD@alQ zS$AKaCJh#S!mJ#xi3kZa9&GHZ1&x}`#_mZEuk!c(k=H;JcYNEudI5fuGMY`tMI$51!d{zs)>;bvV z?SYxZrjy6%%i=6Nov00_5UTepI9*6Vpq@os%sVd@Ew;dQ`|U`-&3^bBHJyT}IKmsc z<+9OcyNM=Vr93ip)tGZNeAJN0fli>epC_^9z*$-L+04#%p={VG!Ou>bh1?Uf&h3xv8e`9d z{;SdTpF^jynYnaA(3Bou&bI*}g@5jEnU877q%gUEAw1cU7LP%XS;=8{AE`Ql$>dwf zc{AL}H;at06|sNSoVD7&(x#aTbpy!LO>rrqS(x@!wmvqo%uVsQpJmp+9W}yXXKbM} z@ped1OG0t5!Xm@e7wK-gtZjW|+I2QS?h{E5FPEZv5jn+WjaH+FgruH8^gdt$oG-8n zH~PX#(e~Uc4pb6UQsSIwbbE?rSr16!$2815_M0Ell)JS&v)0P5?V%gm*d(mdJu)FC zPu|Bo*eL1;A(mJ5IZ*8nx6XI4T*h%*I?SwZYh!wr=-_iU4PTYS=e-SMAW)I5^WWqU zyh(Fgjiq@Qw&@PJn2Qni8Kx@>J#VxhYq54r4}uQf#Xl!NaQzUtI&@mv*@|Uy8CP?7 zBfHhAH>q$(sO5hM9lJ>5a&j*wh_k@DRbdxTV5e@^*y&vAb)WwE^<@kB^@!#fN^uN% zA&&IpP`w>tW~bh7ff>ViEsk$&RrAy2_;fg#ZV5P0E{bUk&abs=Y**um_HV~EP&+?n zeV^6rrG%rNwkKDd)q*vazqnpRyw4Air9}=cp3fp44AWzpKjkZ&X0ba+kj3~_sR}z! zt2r&5X5rlv`h@KPPzlImqg^qGAO<^iU?7w#h%Py>3#YQKw5M;oWeIzCqG&PDwgf5< zEAceq;2&{U?1ws9U$nX`Y>pvT^?*(ns>)$;;C1TE*voJE&TG)^eF(zl_IW?Fl@kp$ ztU~wpljJ#%74EgBHjD`0gr%*e>@oBDFPCYjLL1-jO7u~)J+(B&T+qB8VrDVN2khRm z-}t@cb#412IW>l&$IML`qXb*hp)}r)JNmC=K&SwYt6;#kII4kD*fqN0IBv`jSam+3 zK!>c2nO|W&!9TJg9=LJCp}$rD*TlcAglK9?&ma~{-hvxwpvt8frmNKLS36p5J0IJY zSf-zH)X)}}H#oFPSHtxqhNo%wB<4YhD1o2+*YAg`(8^tB<1w>M#Dyo{N8f!5lTm!X zdOFK4w_t>_Btk&DH>fN?{ zNKeUitkT9`11OFd!f{LFCovuQOIp8Xl7C3idO0Ao?M0Z)*^5&tI=2e%jPa;e>yqP5 zn%r9KQ1T#7^*-D1;6QyYnXoVfaE1^S3zzU8xG07X3)9S7`Rqm#T9`NK{=07cGxk&~ zhef)X&p9>8{sB4xb~pNs~xP=)0(8gPY|$2}XX%XrA(c4a?7WHBz3nzO!|Gq2SY&2k zUf%yht=z`hoPVD7&gje7y#te)O*Au1>L7@`f44jIw%?@__OP2%BV|GWaA0+;v-quz zzbmh zo-?)2EG`n7a}uQ852fKmFB#NWEXHFjv-9R3h=U92B=k1xW@ij)J%QQp;3)3QK>N}|MTw{+4IV_$cOY3&cU+E4DZ}Mqu4dF_yEeh!H2!KmT|sn zNf5FHmvCQcE|-L_>{6#9cW+lhVYAu|X!nPc)-r&s@l4y0ka$yxyWdvWT^&zWAFhif07RU@`H~Jv z!!^uE(enu%gw4aF=BL}mnTSlkO>(j2O4ebSnZ$iTcP@<-ymW#krYqvZ-qicP?EfWHB+P2!hk&YDwDw6u`cnMt8nRyK+9b&1ANgUyVTsd+n|+8?YXa7 z?8ve8-&ZaRMpshCz8h{WFVVt)gc}%@x0B&b{)!jEeAx3P^Mq(B4UyhB5y`QZTGZ0~xOu7CwzCdx!Sn8QhOFX5Mq8}E z;1$sUqqX=3M42|4F-(HuPZSfUPWkOC`g0$9Pv%z?_*!vuJkuc$#6m+rBKVuwn`9jH z8=(G6W{Q5(&}L=DQ)_Fy)zf$RGVQ67K@W2oO@Fw;8)Imfy<%|o1Yw(FPVjI25M9=r z>2%QRCpW}=+7kv*#)Eq{%16!4_W_kpc`=8!jVly`nQ)DWH?P}&Q7W-07t`57yF1t%wC zI}bvrkyf^~e!4Ezl_2o!GzQ5xA=OhZini`I8rBNcLB=QV&(&h|i;a*b-tCP{vs&Gk zh?K5CCi;Z=fyAuGZ!ZV9JS=KM<70Z( zwzDi{-d?=QnJ!$+vpo$=*AdtPI}Zi3I3D+6d0hLAo!Q{dZJ-}*ozJD9C-P?D+_sh3 zmjmO7H_p;=>`wuLUvdI$<(%c2YQ#oiyulX47`8l#{%gQFGjfFBPIjDsElpE%^_F1i zcO60X1A~sU^$&U&LQt$E;5(pIU_IOx#Xb`r}ut8`s|Yy=<|G6F+O0 zJTPBrv^hIxsQx=6XB(Gu76DpOQGUKZkI-5q3iZP(#{yvOob!)UeS6vED>JWymvJD} zT)gs=&UldRmZHUP{B!XiI8WrC*W~}4PZctukkY?=&#YoZzu1gvUuHtb|u2&rLosTasuc(ECxg4i?exFAdEs6u=``C^l zfw-|OCm8+5f&!oT$IEQ>v*M9TNGo+Npj1B|bAwU{@W)mW|}-7sEHU`24YaMdFI@T^-0cx`79rfN=U7NbAFB<)=XD?j%FH zdhN*>nFjWp{?VFQ6%eBsxoj>7Q(?RQ_-CuMVfD6{(OTbZlMfZP&l`6vbUAoS4oqHh zWf)nVfa4pn1sO^fsrt{{k|%4SUU87bB#s}w(x2ts0DgW_4u}W)YFrKC+EMO`{GGxu zC)`zjzg1SqbV3@_c`W_(WL9YF_+B-<03fVo0LP8ELOED3P^9|m6cf%+cs5)0!PYpT z!qp#+e1q~w~$uJ9(b;ht9IH$o8VAC&ABCA_PkfkQjWJu zmVt2%<>dClqA;9mp==$#taE#6wDYmxigzs6Q19`BKg(Viu9{aavN5PMc8%vj0H@19 zmZUX;nL`4{p=&XSjgQ%c3xj`qL%K7><$t;a5lW_Cd~YH0`1!lpwxqgPEMdoChN5je z3_ENfM*rO=MeQc^DZ5GRd*RZ!*Rb2u)c3l!7D|`wA4GSjIDZUva;Ft(eN8(F3g2h) zZV_r;wj-)Z;L1nbFnywVN00P`hCu{$4W|km_;ykzuh7C(I;p=P4$}9X`6RS$K7GaMVM#LYdqy%%ch)|b&SC{)250R8*DoTQ$9dV68)F2DoC&>uT%e9xO-#VEW}r@#=3COJ@8v0O2(=~<+M-wH|ApUhYDn2?|ASjUH5+vkE- z1Kjy%2X|}gV`&CX`MpB1@0`$26?MR;au$zJ%_oV`rek zeNWmguP<6gnoQcb=1Bl#Y_pOvMxD-d zFU`UD!`VQYmg*Jsmy=#?_1N?WFp=Y)#u<~$(reZOgII`NW?vTK-RcM#eyhC3;B&^J z8tJ{KJDl>$6+^#%Qr4sp*neoNJPEy=<+y6&-tv0Wj{@SZ=VHL)XV%7q^r^f@g-c=u zM^prNu>{8Y2M^r{Y@IzSsg{5R#`uW5BUnO`Ctkjyq^Iv?-@8@a=?S`8oI=$?yI_C@hWH5h@%e*LfKbq&}AJrx$zpdT;u$vD3 z#;5O!X-(}D;cwpQAMRU@*~wYnX?l+R1$bhg*WryLbs|J!l?oHz>6N$c|FJ6<2@#vU zxeiAKnUCF+q>_H!iQ?}<>8u`_Kd)I<>w79Sv3TG5_otJOh!E^~GJ=IN1XvBInquB0 zZt#w2=89sqQ8guQl$_kg87r4(f4=|IIK???=d*Gy(#{a9>5hK8FJH{9Z2eUsxb)5G zrIlIDbPV_pE#6Q9JTXQE@k359MDu;)R^#TMRpsD4zJ^SJIn;Ivx;XWNzwwjzCmBlb z`(LI=X0I!+j1~>a|Gi~z27Dt6P@c1Q<+x$p<6}=5R)V}DCFbC(tRW3vYPuYBN(j?Lu zBZP%Srj2ffNHS&jv|M5E@7^?RQ8q zd9R;?c^uTf27xkoFYVH~U!2VUEsN=XXPBhBKjZAOp!T-lWVTSRSiJwDh5F^~tXQ2z zB2-Ebg?^mHT;WJ2a)R%J#s>^&ixBWhWBRK%7W5E_r&oDKj>>2oLKCcDfkPFs{0A9N zCvcoG@(V8=!v`=dbmwH%WNF-J8*q>&h&A++eyB81ywM9#?$P4-`PKR9`?OM?IIFhmONWjE`Y~2O(VB@*|lDD3C$CV5De+*mi&V) z*}+%$|QdWwtqIWo__w?T2+=T;BWWpq)P@a2;xI#;!}%FO0X zbg-r;FYis)cqV>Kvg*CU3u{S>aHTuEw#a@UZ6WZ>o*s;O!&m)rNtQ8A;2l}L+Ooog z#x?lBan*a@811 z&h=+De^#|TiQqf-Bdae&bs51#^lvmKCfYx1w0@3BkqsiVf9s;kUJl5RXj6x3PP6Y5 z&sSzge3K&^T7unnQXWnR^f6Kq7A$tI<{A^FviOqk)T2nB4s}_TcM*9U6k^H{v^A;B z*TV$i#D!XF5^eF$zLT!}a~+Y+1$Aap4CxcZmt#$@k7bj=E3$E-3(yV9-n3CTksTlqw9kCXh8|7hW}d zAUcerw=Pq)En_ji4#n}ws2%w})7B?>`DN8ewFXLlRQ^=@vu}HnzZ^|&bN2cFYwtUw znrxy51EJU`ilBgsNB}{qbV3nDF;t}^RX`BwJqUzEUKK<^Lhrsvk=}cNpa>#OY9y58 z6$AxBKqZtA*eAgI?f%&xyXV`p7=CVn{)Z#I`d@?!UdY4zkgHpUI5x{6MfeJ6k^j(dYG>CxmI!EF2I zDa#EO!{pDOg^IgkqBzz#0K-q+ATe1P`RFqc6^9gQ=&wF2j=r@|-V!!n^WH{kc<^HJi#l6Rj6vk}Lz9oVJldwu*2 z3npQLS>k!Xt>u#N5q&(O{}}MH@_oU()N{le!rSNLZ6c&6o}1Q{pqDJTVn96c$yC&C zR!7Fh*$I6+VE(XhqLm}}+8{&5tjEiqPdJqe@*gxJ8wfVA9Zfr6e_PN9TJ`l7d0zLE zg;7dTPnoPIG~K8tDh(IjqZ^aq<>6_LIVBpIwe69Y*#N5wvJQ0x%#^B_ifN9p*YCXF z0lCX!>jB|i0iERKVDn+}*Uv)e$3p54t6pDLzdFM!76b0)tXbyAh60Y~voV*>>&|RM zc@e39ej9$oTmmA~5(As%v^&e+Ftglqc(RI`cBaIb8`y zmFJJk&s1{7Aj}mTb17M|#q9Sm2 zXC}DD+Jh}E9$NLfr~?7$4G3`d@{y?Sz0*E2vx7)mM&-0+l}h{PPJDu=EZvN^YOpL$ z_1=b*+{ue*TeNAHE;bhHjywa_3;uy%hwxza>_-snnXERq!)9*VcdxvidEmnH;i@2j zuo!SH6%>FZnF<3ubtu~IsS(H(lA^JL!Y(baTX{N7DJCEEIqMUI)j?7PSMO}%)brmnCOzv4A$GY8Wm{)7^`kk-^ z+%xR}N9F+>IQ~^X_t`&5sl@Y&A2NxpFX78Mmw`JL8&!Y$I~G3SbK-JLp0v3Y*H1+a zf|mls*kS8RT#c#Afpc!eTSV5*_pw`GT~XPs^ci<9R~>mINBu+S`AwH$ilPG^LIyGm z`(EmJ-&O3ArF>C_o!7GM!^K=c97x@SDDg0sB>rfSn|N(kI)hoi*Uyg{u5j5qy^l7Of&xRR>pW*F=p>>0qlM;zJaM{QAx_dppu$rztAOJ z8%`F==ozX6cp{q-a64%4gUX8yY>* zV}BwnWjQ8AzI2{J^HSGTMDmYdOFwlFxoqJv++xo5Y1eQsO*HFo`mGH9`^~0g`RBUg>0oeMG@#+V$R(`5t^& z7S9>7JP|lb_sZ3^CeNduqw8JkLn8KJ4lh*BRjZbB--m0{JM7T>^g@ALXM#O!si0|1Tb zp*R^XOe7G4a)CL{uc!d;#Ry}0u>+m@^04BP)@a=Jmb1X93j;rcd%Un&A*ZX9JhBkg zf2EQ90OdWh1G*Rajw|d~6*isD#WS)UP%HouNS2}H&l!q@YUXtRU|KT4R-mjCa?XIL z=3#r2;2h>%Mm0vYBRee0M=cw^(Oqfu=mjnexQ~fUPpZN`{F9G;E1?=11SHQwHx}={ zr1y9XOp*D86yg$Kl#iBU1A+1gRIB$XIkz*M3c(9pfsb>8Ad#diIZWKrnDG6L6G^ zKgXbikpzYa$szJm&-y@;4t+i_@W4Uiy`FV@T}q2^lU>bwOL~jjkV;CYH_-EDxZJKGM=bLP~yMTp5U^jr&IUR{K(iQ26TuA!SVB|-e1vEGS zD~|!F`~a-G??;Il+C|Ss#LTPja}hanfqcRF7=3yVmc*lgg<+aMi|R(<7;_xH4e~94 z39JQNJ=2*Z(#=-KRcx;9ow$#_!X?7K0K;HD(VHv-z?}}p2IXv>B9>YFIHo-Co zAh{MkdiXLAv8wP5+7ajw-;tXFK-RP?AgSmWI>C#s1N}z0E4uw}W>9zTXVB;mfsfl|Gpoqpt1C=1xTTAS&x#E3 zSsg7xG06`bQMb>wCw>#=;{k1O7J4Y+U*6mw1w%hGZaF$*AN#w7S^EiaL*NI%h1qZ6 z9`kw7UX2W+HH-cRu!q-}893AH*vh^aggE)JLSjP6V7EFMx#;)f$*M$7Xyv+g0|C@W|-<6MEhsKm10zz*i>HNMwqhc{YCLONJL~zQqP6Ksza3DsVbRQBiQ)gJv=x za3;M(f?GyqjF8Nt2(AZ@WA`jXG3UN=ZrT$Qn5mgeJIB8s`&xekC<*$D?|SvWwE+8| zCDUY<+kRrNAU~k?n1UExUb@ws_7RlfXgs>f?W42`T2;(7MHW|C`_b4O5FHvuj43aF z9v+9l=~_Eh8RvOW3YVP!?`#%3u+0sLAM0lv%6zw_IWk$7jU=|s>GNM`1~ta_7F zhJ?|iwE2BNn(R||`^zxaM0@CmE0PP90+QDA(oFHl$gok0QmruQUUp?tmchBSN_E=R zA*`KA{@Q9fNz1{y!{pgRg#hSB8)t($Uzs)MR#waVnWQTslf#{);CsttE?CQ~_{%Em zZOM)nrIUjdVs?}8V?KJrgM5b^3>cp!_(n9mx-H-d@Kbwwt6o!%5hWrduXm8<75e@d zVdZORzJqQ42BKNwhhnF~|=s-4qlJYVFYT#Q5q^ez+$)Oi zW$8xD_Bo)w)nB_52a158fBvjDs{X+UKOnL=UF{b9QAG;++WOJ9kK)gzx3rqX(DLiT zygi_oUMUE`l~%73?|u9DK+nO|!T(YR6`cFmhm)iJl-dLQd5Ru-=EnEEb7#CBF?~(( z2znU#ZUWlU)8@jk0c19R+`(jxR>;A-)0sUumJN6Il2<&H4{(ElAW3mj$=UGjy(Db# zCmsiPkUh(8{8Y9cB+}v{kE`{`uhLuLV0h|yOx60^X2_vP8MP9DUy&&b80i2sjj>F^mgBF!jLfU z064j)4jRx$sY;1#K64e6<8{&qn$p{{edJ(aqu!YvCHJgPS;o%Syu*P9F6k<&G3+n^ z7F8L4$17RJ(qpA|9p7U*vFfEIan}x(1QaRJt(Y%RBS?zN3Pe{CQvv-)|D9Ii>v!*G zxk)+)z4#uRjhY@xLdZH&4o)>eFjRLx=6QNZ(l(bT4VpCm)aUkGS%>X$~0LOuUqgYW_kPQy+57l}CP(-`<@+6UlC7_&8 zlBl$vdTyjT5S!ElTmQLq?SSpZa)4J59t!~>A0sZY+N_4SH5?r+8&RX27;0MEI5pm>$qa( ze#?&*094#kdAaHj7eIfvVXRrH31QcR_}NFGivMfs(tQ}yvK(Pth1ig*NTr@1wQ%g2 z&;gCLT!2#Kaiyq_W9&?d;$IhzI4G{V6JrluBfvEQUX~~xm&9&i1BflyVo7{EyRuYM zcmNQyJebWzz~nL`)3E)3h$Jvl8Akwh!k`RpOg#Hs6pRa$-S->-#|~Kq6NM#goUP>Y zIl5|0T*6LhAAJ^rO$RQ$9L=NYpyC2G^$+(f3)nNc3m`kK(QoRf;2M?gUQ~Z9@Y6*o zLZG_pA!(Ta;w-4KhN(1pfIN;!uuXC7T!7~V|9#Nx#tArQ&9kv{hxUS5*MDPH|L0nc zBahD`pN$V;|88Z!UMItC$?*_$JoH>lUQY(vCVMIHB3fMb6+D8Xey_OwJv)l2XK|sZ9vSO0$KsIINTGdfQ1-9v%11ZN;+*2;RooeKgj`r!OS-FW zr(-@Kvyet$%|%}cw1}(Xn)>#A;@-M)@}&zbD*#Az-oRD_5G>~hBI0?vVL4|Z}*_y3SgUJT}lZAv!v zj&*bMQ9B{gxhBDOjPN$UV~uH5KGqdxED$1F=~xl@+BA9stA~oWW4%;#YLc;JUG)eo z$?{lgBv(hCR+UO+5YXii?;#Nm&)@&;4~qj8Ry{IO{6)W#f1~0+n&1n}*&r{*0mgdP zTg4HV(LbhA?f8|s8+qZ!8aW!z!i8j7!ErZ+a|=K%AOdJbuAg5%|KcFB=kBF_)(&A zKTR|q*W;_q)~Mc)V`0u&0df0eAn3B0=Pl(9ys&fCg=&EeL>9paV^O>}eo`iV=W|B& z%DUvStp1bILN31y1-n+*vC}>EqDV7Ufol0gz{ccq@{|IyQ6g(?P=4iXkIdcZvu~XY z7!dKbRv^r))G{ce1Uu$04>cv9XRk+LW-ot=ZwM8IXI*$+D64wc18_a%-E>~JR-IEA zpy0mO=O}g{Vga+T`P*4xZrmgNimXY5%4~|cRje{|8N*Js+ah0C0dsNL$X{;t_%_-(v%> zOr*Szc9}N>Ul4_Ik6K}jH@JC62%qY#?mr1XpY7fdQU@Rb67XwvR3?(kJf7x}Ns187 zXoU_@SA3gj%Tu>7Xc68{LVb>fL(Z9bytgqd z5&c5^YC+0p&-U_NVAug|#S%t&3$wOMN)vz+j{Or46dP^|I4*o5?j7ie7rA zcvnCYHGZ_sF!n=$Z3YzymIeIv;%f-SLDr1g2?^Xhk&&tb?CnQN$qq|JpS9I} zFaV**DKz9X>8tf#jsQJOflU?W-854uD8a|&Q34;Gn_v2;FnU&n9{4jD9s^)Y7eCC2 zk!z0q@|l}tVeAe-{hrAG)_$PQ-Wx5ed#pm3-@=_MdSV)RJSx4}NEl}is&}ud@b{;x zzJX zUh2~hY?6!qg8LsSD+fa19NrrmeOCBEtjnLq?Hel#Yf?C_?`OMe*MA`R z?5q6ha?k65Kj&L=RGM~y=#bl|%4>x4Z{#0lGsE-oqH+g}@ zPAAdxENt70DGZ96_i{+12;|Dkf5zxo0Gtu#`C9&-^vqi$eZ=1Bs!w=Jf4^{pRZ7z( zIVVWb%c4$xSWU0CxnSES!_KSJ5bN;NJZJFBD~1=89@TPXDxqci*5>KkL9C4dc5)!FOJd0{up=qsWYLC+&F_zzq!E zjE1-IWvCij-lfAS0nB6jdnkGf-h9VSK{fQaQF>Nv7h6vj?wDqRRr3J5c&0uPl{?7L` zXJY&uW#%;W#x-MEvsXWhDS*^7d->s~_Ou`Prl~DE&pF9ht#ep8p6F=V8-YaE+SfMPnLI$+{u)3UUhDUvy>TLEYiu)8JBW#vUr0dWXp6y?gi7hyV&jA5j1g|1Lu zGLnS3O%5fW#kz9l_21Q;Ch4_Tstun(+-{%O3JUPqLx)R$S`QEY=nH#f)Z+Ab%clMK zy;@sM90jZBm?+=&;8WN5M`nCZ|E64E#MsvJehj$kg+-iKw`-@0V#oQRBFY-GPa<`N<#Bg4}Whc#hJTtmG?XPbP)~V-_ z-U%z1{Hs>Wd>;bJ%=ELOv-dfR#AWKb4PoJ--9-c@=6K`nEZH>KtZlfLzyf;gtaYYx zWx~L%Mm9^2s~x(J7C$)8uIlDW=&IHPF%w@-QR0*)x*3`jIT%~OdLomzqV7#^ts&G0 zW|CC%|J7{#vtNKnZBb%m&y4bnNi~_))b>Ey+KjU$@1}u^e&MaNw2lBwZ$*@w#~YW1LZNF66s>Qj zwOPK&Iir%jAhPhIUG*D%Z}!AwL*M;)A~_PJO}2O@HVy)@^) z{}MBhk=HLsdUshc>8JQ@`FmX2RTddV*0J%H<(}hizyEZHYpUTwD440iAUpiP0+?d~ zA5JSzSy&dJ`H`9#K>rv2xxAb%2nB;c*e~Od5V0Pvk=BzSPyw#==M1nYxzDAL1re)w z@i*rT{B(!f3lQ~>78?}r&lha7zlBMq=P9@pZry*Sv(#Felh^vk&%1Ty$+roG^%|i9 zxHO$8Sf)nV9?R~9(cB1I+VYx|laUcrS>4sIq3Ona0c~0;`;E$m;jd^22V_e}ed8gz z8y%+J;~U;pUbG49`J_Aj80}4}FP2 zK)VJzksI$I)=N!@q{Y0IWk?2E0>+JwXtN~8dxWx0nzCZ-i*Z!1xZUs(4U}5M3knl` z{PLTpZmjVZ9^IR1Jgc#gkm79Dz+#7n2TthYFEn%POSg7oLkg}EZSO44u4)7-jVc5i zb!3A;L2h={EFf&aSG8&WuSd^i%>C6&OEL|{QT%l9?IXP+{w83Fl89aTp5Nh% z8H=HebaZ=#Qu^zIYYomjSyg3c3bKqx%ZqZ##Iq&Kay%uwyd=8@)3B{}R(+wc$n8O% z^)i!>(}t_riHoC)0%#0+QG3MzX1F4^p%W7lyZB>Cts}!!(0eJ^WX5xf43xh+?wuJM zEN*eD)OXjb>fmUD`OCpj11m$%o3%Js$1YNscX{GD<=%jtfx8I-QDm=*^PnD<5TDGrW?HWvOi)* zpB;st*+_FW#BW^QYkEtT{CLh)Y0ARbHc~cu?z>3>`+X=4ORKC*nx4Gc~Is{Afk_mG=;r6e=*xeh( zSwCpjXq|*}lMbzR7Yx^{8xfzWfpKE^=iH?GhO@xczzZC*>>y-1@(c)6UzhvPehRpd zVs5yZ1~2xhOKF$4bPpV|8-42%j+Pb;u`QMflpo_h`Sxai8f-UlG4D48o~vaLqVOR% zH~jX{j9O=UgfyPY{$C-k>!P8H6kc8djSERZZ$e1O5LKh8Ad_tPwCgp`rF+lzM7Hh~ zyR*$Ci7$JG$c!EDd$x{aQQ&_O)P(!>(W8RB%7yn1d8XzV2${`<0fD|u8{DL3-K2iF zNfiapKRi{NyfG6?DP8o^LqA8K%bDBSp4fpDvZ)xMcUrUzw)QyQs^cS9)+Cea_IMr7 zfj|ljfJ6&u5btSUq@cGKZlm=UDI5tNd@|h(Y~vH{4>fT$#qIrW_Wbb>@2OZ7ir1YQ z`yo5)S{wkfVe45_%b7z`HBY8HTTR<#F6usUGHzX!fO&bybRT7VG;^o6*Zq`Yo+>YO zmc<`37LJAH!Y zEn5Aoi#ZtyYSwD@35#}kIsU$QNQD@_B-DJpzER7-y`rK3Z{M46Gjxd4J==0`Dv{#m z1_C|1NMr^%xZZ8UxEJq2DL-qC#t>4ZCEjwQnw03Q#JNfh8a4kh>>hgRi|oRU^}#K- zps^1bF%&ZYXt3(lvEy@(tzOS%Hn*zHlKa8|cW=t4HQuy62eNhZgDjnavGP`_&ZWCH zU*5XdP7_GYsK^&kq2)o*2S5$6pX@c|6r^XVK?=#L+LE`d``MBmEON!I4&i&Ee+-=u;F>LSxaLF^z z`fKx{1VJl~?ncg%EuudCmB_ zV)+(pB^bqk`vt*za7gt$=4ww8%utU^tjoDd^vvp0moxj?QZTc$2qEYOqJv+5U$@cZ zzZ2|G>(WPPP=|3{f%oyl5qGKCEI)&rY@RJqnU}YcMwXoB`x;RiWv< z!NI+t@W|s!#AUg1(L&g)@MV(E*<5x+24~!7Wd6xfc-% zV}V@sS%W;DN4Mqv{eATD069P{Ct-1gcm|DeVZ`l|Db-ix3ef=lU>) - - - - - - -Extract log probability (target) — fit-method-lp • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract log probability (target) — fit-method-lp • cmdstanr - - - - - - - - - - - - - - + + - -
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
-
- -
- -
+

Extract log probability (target)

- Source: R/fit.R + Source: R/fit.R
@@ -186,26 +114,30 @@

Extract log probability (target)

(target) accumulated in the model block of the Stan program. For variational inference the log density of the variational approximation to the posterior is also available via the $lp_approx() method.

-

See the Log Probability Increment vs. Sampling Statement +

See the Log Probability Increment vs. Sampling Statement section of the Stan Reference Manual for details on when normalizing constants are dropped from log probability calculations.

- 
lp()
-
-lp_approx()
- +
+ 
lp()
+
+lp_approx()
+
-

Value

+
+

Value

+ -

A numeric vector with length equal to the number of (post-warmup) +

A numeric vector with length equal to the number of (post-warmup) draws for MCMC and variational inference, and length equal to 1 for optimization.

-

Details

- +
+
+

Details

-

lp__ is the unnormalized log density on Stan's unconstrained space. +

lp__ is the unnormalized log density on Stan's unconstrained space. This will in general be different than the unnormalized model log density evaluated at a posterior draw (which is on the constrained space). lp__ is intended to diagnose sampling efficiency and evaluate approximations.

@@ -213,53 +145,57 @@

Details (also on the unconstrained space). It is exposed in the variational method for performing the checks described in Yao et al. (2018) and implemented in the loo package.

-

References

- +
+
+

References

Yao, Y., Vehtari, A., Simpson, D., and Gelman, A. (2018). Yes, but did it work?: Evaluating variational inference. Proceedings of the 35th International Conference on Machine Learning, PMLR 80:5581–5590.

-

See also

- -

CmdStanMCMC, CmdStanMLE, CmdStanVB

- -

Examples

- 
# \dontrun{
-fit_mcmc <- cmdstanr_example("logistic")
-head(fit_mcmc$lp())
-
#> [1] -68.9225 -69.2427 -70.5110 -65.5960 -65.2714 -64.1837

-fit_mle <- cmdstanr_example("logistic", method = "optimize")
-fit_mle$lp()
-
#> [1] -63.9218

-fit_vb <- cmdstanr_example("logistic", method = "variational")
-plot(fit_vb$lp(), fit_vb$lp_approx())
-
# }
+    

+    

+    
See also

+    
CmdStanMCMC, CmdStanMLE, CmdStanVB

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic")
+head(fit_mcmc$lp())
+#> [1] -66.8230 -66.2198 -66.0794 -65.9100 -66.1692 -66.3888
+
+fit_mle <- cmdstanr_example("logistic", method = "optimize")
+fit_mle$lp()
+#> [1] -63.9218
+
+fit_vb <- cmdstanr_example("logistic", method = "variational")
+plot(fit_vb$lp(), fit_vb$lp_approx())
+
+# }
+
+
+
+
- - + + diff --git a/docs/reference/fit-method-metadata.html b/docs/reference/fit-method-metadata.html index ebf454d11..88674bf7b 100644 --- a/docs/reference/fit-method-metadata.html +++ b/docs/reference/fit-method-metadata.html @@ -1,77 +1,14 @@ - - - - - - - -Extract metadata from CmdStan CSV files — fit-method-metadata • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract metadata from CmdStan CSV files — fit-method-metadata • cmdstanr - - - - - - - - - - - - - - + + -
-
- -
- -
+

Extract metadata from CmdStan CSV files

- Source: R/fit.R + Source: R/fit.R

The $metadata() method returns a list of information gathered from the CSV output files, including the CmdStan configuration used when -fitting the model. See Examples and read_cmdstan_csv().

+fitting the model. See Examples and read_cmdstan_csv().

- 
metadata()
- - -

See also

- -

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

+
+ 
metadata()
+
-

Examples

- 
# \dontrun{
-fit_mcmc <- cmdstanr_example("logistic", method = "sample")
-str(fit_mcmc$metadata())
-
#> List of 40
-#>  $ stan_version_major  : num 2
-#>  $ stan_version_minor  : num 29
-#>  $ stan_version_patch  : num 1
-#>  $ start_datetime      : chr "2022-03-18 18:25:53 UTC"
-#>  $ method              : chr "sample"
-#>  $ save_warmup         : num 0
-#>  $ thin                : num 1
-#>  $ gamma               : num 0.05
-#>  $ kappa               : num 0.75
-#>  $ t0                  : num 10
-#>  $ init_buffer         : num 75
-#>  $ term_buffer         : num 50
-#>  $ window              : num 25
-#>  $ algorithm           : chr "hmc"
-#>  $ engine              : chr "nuts"
-#>  $ metric              : chr "diag_e"
-#>  $ stepsize_jitter     : num 0
-#>  $ num_chains          : num 1
-#>  $ id                  : num [1:4] 1 2 3 4
-#>  $ init                : num [1:4] 2 2 2 2
-#>  $ seed                : num 1.35e+09
-#>  $ refresh             : num 100
-#>  $ sig_figs            : num -1
-#>  $ profile_file        : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181225-1-94bdc7.csv"
-#>  $ stanc_version       : chr "stanc3 v2.29.1"
-#>  $ sampler_diagnostics : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
-#>  $ variables           : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
-#>  $ step_size_adaptation: num [1:4] 0.763 0.761 0.699 0.761
-#>  $ model_name          : chr "logistic_model"
-#>  $ adapt_engaged       : num 1
-#>  $ adapt_delta         : num 0.8
-#>  $ max_treedepth       : num 10
-#>  $ step_size           : num [1:4] 1 1 1 1
-#>  $ iter_warmup         : num 1000
-#>  $ iter_sampling       : num 1000
-#>  $ threads_per_chain   : num 1
-#>  $ time                :'data.frame':	4 obs. of  4 variables:
-#>   ..$ chain_id: num [1:4] 1 2 3 4
-#>   ..$ warmup  : num [1:4] 0.026 0.032 0.058 0.028
-#>   ..$ sampling: num [1:4] 0.094 0.104 0.173 0.098
-#>   ..$ total   : num [1:4] 0.12 0.136 0.231 0.126
-#>  $ stan_variable_sizes :List of 4
-#>   ..$ lp__   : num 1
-#>   ..$ alpha  : num 1
-#>   ..$ beta   : num 3
-#>   ..$ log_lik: num 100
-#>  $ stan_variables      : chr [1:4] "lp__" "alpha" "beta" "log_lik"
-#>  $ model_params        : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...

-fit_mle <- cmdstanr_example("logistic", method = "optimize")
-str(fit_mle$metadata())
-
#> List of 30
-#>  $ stan_version_major : num 2
-#>  $ stan_version_minor : num 29
-#>  $ stan_version_patch : num 1
-#>  $ start_datetime     : chr "2022-03-18 18:25:55 UTC"
-#>  $ method             : chr "optimize"
-#>  $ algorithm          : chr "lbfgs"
-#>  $ init_alpha         : num 0.001
-#>  $ tol_obj            : num 1e-12
-#>  $ tol_rel_obj        : num 10000
-#>  $ tol_grad           : num 1e-08
-#>  $ tol_rel_grad       : num 1e+07
-#>  $ tol_param          : num 1e-08
-#>  $ history_size       : num 5
-#>  $ iter               : num 2000
-#>  $ save_iterations    : num 0
-#>  $ id                 : num 1
-#>  $ init               : num 2
-#>  $ seed               : num 4.24e+08
-#>  $ refresh            : num 100
-#>  $ sig_figs           : num -1
-#>  $ profile_file       : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181225-1-85709d.csv"
-#>  $ stanc_version      : chr "stanc3 v2.29.1"
-#>  $ sampler_diagnostics: chr(0) 
-#>  $ variables          : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
-#>  $ model_name         : chr "logistic_model"
-#>  $ threads            : num 1
-#>  $ time               :'data.frame':	0 obs. of  0 variables
-#>  $ stan_variable_sizes:List of 4
-#>   ..$ lp__   : num 1
-#>   ..$ alpha  : num 1
-#>   ..$ beta   : num 3
-#>   ..$ log_lik: num 100
-#>  $ stan_variables     : chr [1:4] "lp__" "alpha" "beta" "log_lik"
-#>  $ model_params       : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...

-fit_vb <- cmdstanr_example("logistic", method = "variational")
-str(fit_vb$metadata())
-
#> List of 29
-#>  $ stan_version_major : num 2
-#>  $ stan_version_minor : num 29
-#>  $ stan_version_patch : num 1
-#>  $ start_datetime     : chr "2022-03-18 18:25:55 UTC"
-#>  $ method             : chr "variational"
-#>  $ algorithm          : chr "meanfield"
-#>  $ iter               : num 50
-#>  $ grad_samples       : num 1
-#>  $ elbo_samples       : num 100
-#>  $ eta                : num 1
-#>  $ tol_rel_obj        : num 0.01
-#>  $ eval_elbo          : num 100
-#>  $ output_samples     : num 1000
-#>  $ id                 : num 1
-#>  $ init               : num 2
-#>  $ seed               : num 5.22e+08
-#>  $ refresh            : num 100
-#>  $ sig_figs           : num -1
-#>  $ profile_file       : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181225-1-18f1e4.csv"
-#>  $ stanc_version      : chr "stanc3 v2.29.1"
-#>  $ sampler_diagnostics: chr(0) 
-#>  $ variables          : chr [1:106] "lp__" "lp_approx__" "alpha" "beta[1]" ...
-#>  $ model_name         : chr "logistic_model"
-#>  $ adapt_engaged      : num 1
-#>  $ threads            : num 1
-#>  $ time               :'data.frame':	0 obs. of  0 variables
-#>  $ stan_variable_sizes:List of 5
-#>   ..$ lp__       : num 1
-#>   ..$ lp_approx__: num 1
-#>   ..$ alpha      : num 1
-#>   ..$ beta       : num 3
-#>   ..$ log_lik    : num 100
-#>  $ stan_variables     : chr [1:5] "lp__" "lp_approx__" "alpha" "beta" ...
-#>  $ model_params       : chr [1:106] "lp__" "lp_approx__" "alpha" "beta[1]" ...
# }
+    
 
-
+
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+str(fit_mcmc$metadata())
+#> List of 40
+#>  $ stan_version_major  : num 2
+#>  $ stan_version_minor  : num 32
+#>  $ stan_version_patch  : num 2
+#>  $ start_datetime      : chr "2023-07-25 20:36:00 UTC"
+#>  $ method              : chr "sample"
+#>  $ save_warmup         : num 0
+#>  $ thin                : num 1
+#>  $ gamma               : num 0.05
+#>  $ kappa               : num 0.75
+#>  $ t0                  : num 10
+#>  $ init_buffer         : num 75
+#>  $ term_buffer         : num 50
+#>  $ window              : num 25
+#>  $ algorithm           : chr "hmc"
+#>  $ engine              : chr "nuts"
+#>  $ metric              : chr "diag_e"
+#>  $ stepsize_jitter     : num 0
+#>  $ num_chains          : num 1
+#>  $ id                  : num [1:4] 1 2 3 4
+#>  $ init                : num [1:4] 2 2 2 2
+#>  $ seed                : num 49669263
+#>  $ refresh             : num 100
+#>  $ sig_figs            : num -1
+#>  $ profile_file        : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251436-1-352ddc.csv"
+#>  $ stanc_version       : chr "stanc3 v2.32.2"
+#>  $ sampler_diagnostics : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
+#>  $ variables           : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+#>  $ step_size_adaptation: num [1:4] 0.702 0.797 0.751 0.804
+#>  $ model_name          : chr "logistic_model"
+#>  $ adapt_engaged       : num 1
+#>  $ adapt_delta         : num 0.8
+#>  $ max_treedepth       : num 10
+#>  $ step_size           : num [1:4] 1 1 1 1
+#>  $ iter_warmup         : num 1000
+#>  $ iter_sampling       : num 1000
+#>  $ threads_per_chain   : num 1
+#>  $ time                :'data.frame':	4 obs. of  4 variables:
+#>   ..$ chain_id: num [1:4] 1 2 3 4
+#>   ..$ warmup  : num [1:4] 0.022 0.022 0.021 0.022
+#>   ..$ sampling: num [1:4] 0.068 0.065 0.073 0.067
+#>   ..$ total   : num [1:4] 0.09 0.087 0.094 0.089
+#>  $ stan_variable_sizes :List of 4
+#>   ..$ lp__   : num 1
+#>   ..$ alpha  : num 1
+#>   ..$ beta   : num 3
+#>   ..$ log_lik: num 100
+#>  $ stan_variables      : chr [1:4] "lp__" "alpha" "beta" "log_lik"
+#>  $ model_params        : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+
+fit_mle <- cmdstanr_example("logistic", method = "optimize")
+str(fit_mle$metadata())
+#> List of 31
+#>  $ stan_version_major : num 2
+#>  $ stan_version_minor : num 32
+#>  $ stan_version_patch : num 2
+#>  $ start_datetime     : chr "2023-07-25 20:36:01 UTC"
+#>  $ method             : chr "optimize"
+#>  $ algorithm          : chr "lbfgs"
+#>  $ init_alpha         : num 0.001
+#>  $ tol_obj            : num 1e-12
+#>  $ tol_rel_obj        : num 10000
+#>  $ tol_grad           : num 1e-08
+#>  $ tol_rel_grad       : num 1e+07
+#>  $ tol_param          : num 1e-08
+#>  $ history_size       : num 5
+#>  $ jacobian           : num 0
+#>  $ iter               : num 2000
+#>  $ save_iterations    : num 0
+#>  $ id                 : num 1
+#>  $ init               : num 2
+#>  $ seed               : num 1.74e+09
+#>  $ refresh            : num 100
+#>  $ sig_figs           : num -1
+#>  $ profile_file       : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251436-1-4f11df.csv"
+#>  $ stanc_version      : chr "stanc3 v2.32.2"
+#>  $ sampler_diagnostics: chr(0) 
+#>  $ variables          : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+#>  $ model_name         : chr "logistic_model"
+#>  $ threads            : num 1
+#>  $ time               :'data.frame':	0 obs. of  0 variables
+#>  $ stan_variable_sizes:List of 4
+#>   ..$ lp__   : num 1
+#>   ..$ alpha  : num 1
+#>   ..$ beta   : num 3
+#>   ..$ log_lik: num 100
+#>  $ stan_variables     : chr [1:4] "lp__" "alpha" "beta" "log_lik"
+#>  $ model_params       : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+
+fit_vb <- cmdstanr_example("logistic", method = "variational")
+str(fit_vb$metadata())
+#> List of 29
+#>  $ stan_version_major : num 2
+#>  $ stan_version_minor : num 32
+#>  $ stan_version_patch : num 2
+#>  $ start_datetime     : chr "2023-07-25 20:36:01 UTC"
+#>  $ method             : chr "variational"
+#>  $ algorithm          : chr "meanfield"
+#>  $ iter               : num 50
+#>  $ grad_samples       : num 1
+#>  $ elbo_samples       : num 100
+#>  $ eta                : num 1
+#>  $ tol_rel_obj        : num 0.01
+#>  $ eval_elbo          : num 100
+#>  $ output_samples     : num 1000
+#>  $ id                 : num 1
+#>  $ init               : num 2
+#>  $ seed               : num 2.97e+08
+#>  $ refresh            : num 100
+#>  $ sig_figs           : num -1
+#>  $ profile_file       : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251436-1-5bbf54.csv"
+#>  $ stanc_version      : chr "stanc3 v2.32.2"
+#>  $ sampler_diagnostics: chr(0) 
+#>  $ variables          : chr [1:106] "lp__" "lp_approx__" "alpha" "beta[1]" ...
+#>  $ model_name         : chr "logistic_model"
+#>  $ adapt_engaged      : num 1
+#>  $ threads            : num 1
+#>  $ time               :'data.frame':	0 obs. of  0 variables
+#>  $ stan_variable_sizes:List of 5
+#>   ..$ lp__       : num 1
+#>   ..$ lp_approx__: num 1
+#>   ..$ alpha      : num 1
+#>   ..$ beta       : num 3
+#>   ..$ log_lik    : num 100
+#>  $ stan_variables     : chr [1:5] "lp__" "lp_approx__" "alpha" "beta" ...
+#>  $ model_params       : chr [1:106] "lp__" "lp_approx__" "alpha" "beta[1]" ...
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-mle.html b/docs/reference/fit-method-mle.html index f191e2ad6..a6544a893 100644 --- a/docs/reference/fit-method-mle.html +++ b/docs/reference/fit-method-mle.html @@ -1,81 +1,18 @@ - - - - - - - -Extract (penalized) maximum likelihood estimate after optimization — fit-method-mle • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract (penalized) maximum likelihood estimate after optimization — fit-method-mle • cmdstanr - - - - - - - - - - - - - - + + -
-
- -
- -
+

Extract (penalized) maximum likelihood estimate after optimization

- Source: R/fit.R + Source: R/fit.R
-

The $mle() method is only available for CmdStanMLE objects. +

The $mle() method is only available for CmdStanMLE objects. It returns the penalized maximum likelihood estimate (posterior mode) as a numeric vector with one element per variable. The returned vector does not include lp__, the total log probability (target) accumulated in the model block of the Stan program, which is available via the -$lp() method and also included in the -$draws() method.

+$lp() method and also included in the +$draws() method.

- 
mle(variables = NULL)
+
+ 
mle(variables = NULL)
+
-

Arguments

- - - - - - -
variables

(character vector) The variables (parameters, transformed +

+

Arguments

+
variables
+

(character vector) The variables (parameters, transformed parameters, and generated quantities) to include. If NULL (the default) -then all variables are included.

- -

Value

+then all variables are included.

-

A numeric vector. See Examples.

-

See also

+
+
+

Value

+ -

CmdStanMLE

- -

Examples

- 
# \dontrun{
-fit <- cmdstanr_example("logistic", method = "optimize")
-fit$mle("alpha")
-
#>    alpha 
-#> 0.364475 
fit$mle("beta")
-
#>   beta[1]   beta[2]   beta[3] 
-#> -0.631555 -0.258975  0.648491 
fit$mle("beta[2]")
-
#>   beta[2] 
-#> -0.258975 
# }
+
A numeric vector. See Examples.

+    

+    

+    
See also

+    
CmdStanMLE

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("logistic", method = "optimize")
+fit$mle("alpha")
+#>    alpha 
+#> 0.364457 
+fit$mle("beta")
+#>   beta[1]   beta[2]   beta[3] 
+#> -0.631547 -0.258967  0.648495 
+fit$mle("beta[2]")
+#>   beta[2] 
+#> -0.258967 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-num_chains.html b/docs/reference/fit-method-num_chains.html index 2f9483247..9c1066025 100644 --- a/docs/reference/fit-method-num_chains.html +++ b/docs/reference/fit-method-num_chains.html @@ -1,75 +1,12 @@ - - - - - - - -Extract number of chains after MCMC — fit-method-num_chains • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract number of chains after MCMC — fit-method-num_chains • cmdstanr - - + + - - -
-
- -
- -
+

Extract number of chains after MCMC

- Source: R/fit.R + Source: R/fit.R
@@ -179,48 +107,52 @@

Extract number of chains after MCMC

The $num_chains() method returns the number of MCMC chains.

- 
num_chains()
- - -

Value

- -

An integer.

-

See also

+
+ 
num_chains()
+
-

CmdStanMCMC

+
+

Value

+ -

Examples

- 
# \dontrun{
-fit_mcmc <- cmdstanr_example(chains = 2)
-fit_mcmc$num_chains()
-
#> [1] 2
# }
+
An integer.

+    

+    

+    
See also

+    
CmdStanMCMC

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example(chains = 2)
+fit_mcmc$num_chains()
+#> [1] 2
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+

- - + + diff --git a/docs/reference/fit-method-output.html b/docs/reference/fit-method-output.html index 395a5186d..f83b9096e 100644 --- a/docs/reference/fit-method-output.html +++ b/docs/reference/fit-method-output.html @@ -1,80 +1,17 @@ - - - - - - - -Access console output — fit-method-output • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Access console output — fit-method-output • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Access console output

- Source: R/fit.R + Source: R/fit.R
@@ -189,226 +117,233 @@

Access console output

the console output.

- 
output(id = NULL)
- -

Arguments

- - - - - - -
id

(integer) The chain id. Ignored if the model was not fit using -MCMC.

- -

See also

+
+ 
output(id = NULL)
+
-

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

+
+

Arguments

+
id
+

(integer) The chain id. Ignored if the model was not fit using +MCMC.

-

Examples

- 
# \dontrun{
-fit_mcmc <- cmdstanr_example("logistic", method = "sample")
-fit_mcmc$output(1)
-
#> 
-#> method = sample (Default)
-#>   sample
-#>     num_samples = 1000 (Default)
-#>     num_warmup = 1000 (Default)
-#>     save_warmup = 0 (Default)
-#>     thin = 1 (Default)
-#>     adapt
-#>       engaged = 1 (Default)
-#>       gamma = 0.050000000000000003 (Default)
-#>       delta = 0.80000000000000004 (Default)
-#>       kappa = 0.75 (Default)
-#>       t0 = 10 (Default)
-#>       init_buffer = 75 (Default)
-#>       term_buffer = 50 (Default)
-#>       window = 25 (Default)
-#>     algorithm = hmc (Default)
-#>       hmc
-#>         engine = nuts (Default)
-#>           nuts
-#>             max_depth = 10 (Default)
-#>         metric = diag_e (Default)
-#>         metric_file =  (Default)
-#>         stepsize = 1 (Default)
-#>         stepsize_jitter = 0 (Default)
-#>     num_chains = 1 (Default)
-#> id = 1 (Default)
-#> data
-#>   file = /private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpBlSxPc/temp_libpathbc8d32dd446a/cmdstanr/logistic.data.json
-#> init = 2 (Default)
-#> random
-#>   seed = 1354164569
-#> output
-#>   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-1-827a18.csv
-#>   diagnostic_file =  (Default)
-#>   refresh = 100 (Default)
-#>   sig_figs = -1 (Default)
-#>   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181225-1-987172.csv
-#> num_threads = 1 (Default)
-#> 
-#> 
-#> Gradient evaluation took 2.4e-05 seconds
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.24 seconds.
-#> Adjust your expectations accordingly!
-#> 
-#> 
-#> Iteration:    1 / 2000 [  0%]  (Warmup)
-#> Iteration:  100 / 2000 [  5%]  (Warmup)
-#> Iteration:  200 / 2000 [ 10%]  (Warmup)
-#> Iteration:  300 / 2000 [ 15%]  (Warmup)
-#> Iteration:  400 / 2000 [ 20%]  (Warmup)
-#> Iteration:  500 / 2000 [ 25%]  (Warmup)
-#> Iteration:  600 / 2000 [ 30%]  (Warmup)
-#> Iteration:  700 / 2000 [ 35%]  (Warmup)
-#> Iteration:  800 / 2000 [ 40%]  (Warmup)
-#> Iteration:  900 / 2000 [ 45%]  (Warmup)
-#> Iteration: 1000 / 2000 [ 50%]  (Warmup)
-#> Iteration: 1001 / 2000 [ 50%]  (Sampling)
-#> Iteration: 1100 / 2000 [ 55%]  (Sampling)
-#> Iteration: 1200 / 2000 [ 60%]  (Sampling)
-#> Iteration: 1300 / 2000 [ 65%]  (Sampling)
-#> Iteration: 1400 / 2000 [ 70%]  (Sampling)
-#> Iteration: 1500 / 2000 [ 75%]  (Sampling)
-#> Iteration: 1600 / 2000 [ 80%]  (Sampling)
-#> Iteration: 1700 / 2000 [ 85%]  (Sampling)
-#> Iteration: 1800 / 2000 [ 90%]  (Sampling)
-#> Iteration: 1900 / 2000 [ 95%]  (Sampling)
-#> Iteration: 2000 / 2000 [100%]  (Sampling)
-#> 
-#>  Elapsed Time: 0.026 seconds (Warm-up)
-#>                0.088 seconds (Sampling)
-#>                0.114 seconds (Total)
out <- fit_mcmc$output()
-str(out)
-
#> List of 4
-#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
-#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
-#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
-#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...

-fit_mle <- cmdstanr_example("logistic", method = "optimize")
-fit_mle$output()
-
#> 
-#> method = optimize
-#>   optimize
-#>     algorithm = lbfgs (Default)
-#>       lbfgs
-#>         init_alpha = 0.001 (Default)
-#>         tol_obj = 9.9999999999999998e-13 (Default)
-#>         tol_rel_obj = 10000 (Default)
-#>         tol_grad = 1e-08 (Default)
-#>         tol_rel_grad = 10000000 (Default)
-#>         tol_param = 1e-08 (Default)
-#>         history_size = 5 (Default)
-#>     iter = 2000 (Default)
-#>     save_iterations = 0 (Default)
-#> id = 1 (Default)
-#> data
-#>   file = /private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpBlSxPc/temp_libpathbc8d32dd446a/cmdstanr/logistic.data.json
-#> init = 2 (Default)
-#> random
-#>   seed = 780877759
-#> output
-#>   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-1-1e7746.csv
-#>   diagnostic_file =  (Default)
-#>   refresh = 100 (Default)
-#>   sig_figs = -1 (Default)
-#>   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181225-1-71b1d7.csv
-#> num_threads = 1 (Default)
-#> 
-#> Initial log joint probability = -137.575
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes 
-#>        7      -63.9218   9.27364e-05   0.000115205      0.9521      0.9521       11   
-#> Optimization terminated normally: 
-#>   Convergence detected: relative gradient magnitude is below tolerance

-fit_vb <- cmdstanr_example("logistic", method = "variational")
-fit_vb$output()
-
#> 
-#> method = variational
-#>   variational
-#>     algorithm = meanfield (Default)
-#>       meanfield
-#>     iter = 10000 (Default)
-#>     grad_samples = 1 (Default)
-#>     elbo_samples = 100 (Default)
-#>     eta = 1 (Default)
-#>     adapt
-#>       engaged = 1 (Default)
-#>       iter = 50 (Default)
-#>     tol_rel_obj = 0.01 (Default)
-#>     eval_elbo = 100 (Default)
-#>     output_samples = 1000 (Default)
-#> id = 1 (Default)
-#> data
-#>   file = /private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpBlSxPc/temp_libpathbc8d32dd446a/cmdstanr/logistic.data.json
-#> init = 2 (Default)
-#> random
-#>   seed = 655253804
-#> output
-#>   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181225-1-7c6411.csv
-#>   diagnostic_file =  (Default)
-#>   refresh = 100 (Default)
-#>   sig_figs = -1 (Default)
-#>   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181225-1-8abe62.csv
-#> num_threads = 1 (Default)
-#> 
-#> ------------------------------------------------------------
-#> EXPERIMENTAL ALGORITHM:
-#>   This procedure has not been thoroughly tested and may be unstable
-#>   or buggy. The interface is subject to change.
-#> ------------------------------------------------------------
-#> 
-#> 
-#> 
-#> Gradient evaluation took 2.6e-05 seconds
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.26 seconds.
-#> Adjust your expectations accordingly!
-#> 
-#> 
-#> Begin eta adaptation.
-#> Iteration:   1 / 250 [  0%]  (Adaptation)
-#> Iteration:  50 / 250 [ 20%]  (Adaptation)
-#> Iteration: 100 / 250 [ 40%]  (Adaptation)
-#> Iteration: 150 / 250 [ 60%]  (Adaptation)
-#> Iteration: 200 / 250 [ 80%]  (Adaptation)
-#> Success! Found best value [eta = 1] earlier than expected.
-#> 
-#> Begin stochastic gradient ascent.
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes 
-#>    100          -66.196             1.000            1.000
-#>    200          -66.407             0.502            1.000
-#>    300          -66.643             0.336            0.004   MEDIAN ELBO CONVERGED
-#> 
-#> Drawing a sample of size 1000 from the approximate posterior... 
-#> COMPLETED.
# }
+
+ -
+
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$output(1)
+#> 
+#> method = sample (Default)
+#>   sample
+#>     num_samples = 1000 (Default)
+#>     num_warmup = 1000 (Default)
+#>     save_warmup = 0 (Default)
+#>     thin = 1 (Default)
+#>     adapt
+#>       engaged = 1 (Default)
+#>       gamma = 0.050000000000000003 (Default)
+#>       delta = 0.80000000000000004 (Default)
+#>       kappa = 0.75 (Default)
+#>       t0 = 10 (Default)
+#>       init_buffer = 75 (Default)
+#>       term_buffer = 50 (Default)
+#>       window = 25 (Default)
+#>     algorithm = hmc (Default)
+#>       hmc
+#>         engine = nuts (Default)
+#>           nuts
+#>             max_depth = 10 (Default)
+#>         metric = diag_e (Default)
+#>         metric_file =  (Default)
+#>         stepsize = 1 (Default)
+#>         stepsize_jitter = 0 (Default)
+#>     num_chains = 1 (Default)
+#> id = 1 (Default)
+#> data
+#>   file = /private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpKFjP9Y/temp_libpath15d4242e55c52/cmdstanr/logistic.data.json
+#> init = 2 (Default)
+#> random
+#>   seed = 810297768
+#> output
+#>   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-1-71373e.csv
+#>   diagnostic_file =  (Default)
+#>   refresh = 100 (Default)
+#>   sig_figs = -1 (Default)
+#>   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251436-1-09f0a3.csv
+#> num_threads = 1 (Default)
+#> 
+#> 
+#> Gradient evaluation took 2e-05 seconds
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.2 seconds.
+#> Adjust your expectations accordingly!
+#> 
+#> 
+#> Iteration:    1 / 2000 [  0%]  (Warmup)
+#> Iteration:  100 / 2000 [  5%]  (Warmup)
+#> Iteration:  200 / 2000 [ 10%]  (Warmup)
+#> Iteration:  300 / 2000 [ 15%]  (Warmup)
+#> Iteration:  400 / 2000 [ 20%]  (Warmup)
+#> Iteration:  500 / 2000 [ 25%]  (Warmup)
+#> Iteration:  600 / 2000 [ 30%]  (Warmup)
+#> Iteration:  700 / 2000 [ 35%]  (Warmup)
+#> Iteration:  800 / 2000 [ 40%]  (Warmup)
+#> Iteration:  900 / 2000 [ 45%]  (Warmup)
+#> Iteration: 1000 / 2000 [ 50%]  (Warmup)
+#> Iteration: 1001 / 2000 [ 50%]  (Sampling)
+#> Iteration: 1100 / 2000 [ 55%]  (Sampling)
+#> Iteration: 1200 / 2000 [ 60%]  (Sampling)
+#> Iteration: 1300 / 2000 [ 65%]  (Sampling)
+#> Iteration: 1400 / 2000 [ 70%]  (Sampling)
+#> Iteration: 1500 / 2000 [ 75%]  (Sampling)
+#> Iteration: 1600 / 2000 [ 80%]  (Sampling)
+#> Iteration: 1700 / 2000 [ 85%]  (Sampling)
+#> Iteration: 1800 / 2000 [ 90%]  (Sampling)
+#> Iteration: 1900 / 2000 [ 95%]  (Sampling)
+#> Iteration: 2000 / 2000 [100%]  (Sampling)
+#> 
+#>  Elapsed Time: 0.022 seconds (Warm-up)
+#>                0.067 seconds (Sampling)
+#>                0.089 seconds (Total)
+out <- fit_mcmc$output()
+str(out)
+#> List of 4
+#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
+#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
+#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
+#>  $ : chr [1:73] "" "method = sample (Default)" "  sample" "    num_samples = 1000 (Default)" ...
+
+fit_mle <- cmdstanr_example("logistic", method = "optimize")
+fit_mle$output()
+#> 
+#> method = optimize
+#>   optimize
+#>     algorithm = lbfgs (Default)
+#>       lbfgs
+#>         init_alpha = 0.001 (Default)
+#>         tol_obj = 9.9999999999999998e-13 (Default)
+#>         tol_rel_obj = 10000 (Default)
+#>         tol_grad = 1e-08 (Default)
+#>         tol_rel_grad = 10000000 (Default)
+#>         tol_param = 1e-08 (Default)
+#>         history_size = 5 (Default)
+#>     jacobian = 0 (Default)
+#>     iter = 2000 (Default)
+#>     save_iterations = 0 (Default)
+#> id = 1 (Default)
+#> data
+#>   file = /private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpKFjP9Y/temp_libpath15d4242e55c52/cmdstanr/logistic.data.json
+#> init = 2 (Default)
+#> random
+#>   seed = 311070905
+#> output
+#>   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-1-656287.csv
+#>   diagnostic_file =  (Default)
+#>   refresh = 100 (Default)
+#>   sig_figs = -1 (Default)
+#>   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251436-1-75d1fd.csv
+#> num_threads = 1 (Default)
+#> 
+#> Initial log joint probability = -109.352
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes 
+#>        7      -63.9218   0.000676152    0.00167181       0.941       0.941        9   
+#> Optimization terminated normally: 
+#>   Convergence detected: relative gradient magnitude is below tolerance
+
+fit_vb <- cmdstanr_example("logistic", method = "variational")
+fit_vb$output()
+#> 
+#> method = variational
+#>   variational
+#>     algorithm = meanfield (Default)
+#>       meanfield
+#>     iter = 10000 (Default)
+#>     grad_samples = 1 (Default)
+#>     elbo_samples = 100 (Default)
+#>     eta = 1 (Default)
+#>     adapt
+#>       engaged = 1 (Default)
+#>       iter = 50 (Default)
+#>     tol_rel_obj = 0.01 (Default)
+#>     eval_elbo = 100 (Default)
+#>     output_samples = 1000 (Default)
+#> id = 1 (Default)
+#> data
+#>   file = /private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpKFjP9Y/temp_libpath15d4242e55c52/cmdstanr/logistic.data.json
+#> init = 2 (Default)
+#> random
+#>   seed = 296025410
+#> output
+#>   file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-1-7927e8.csv
+#>   diagnostic_file =  (Default)
+#>   refresh = 100 (Default)
+#>   sig_figs = -1 (Default)
+#>   profile_file = /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251436-1-2bb650.csv
+#> num_threads = 1 (Default)
+#> 
+#> ------------------------------------------------------------
+#> EXPERIMENTAL ALGORITHM:
+#>   This procedure has not been thoroughly tested and may be unstable
+#>   or buggy. The interface is subject to change.
+#> ------------------------------------------------------------
+#> 
+#> 
+#> 
+#> Gradient evaluation took 2.1e-05 seconds
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.21 seconds.
+#> Adjust your expectations accordingly!
+#> 
+#> 
+#> Begin eta adaptation.
+#> Iteration:   1 / 250 [  0%]  (Adaptation)
+#> Iteration:  50 / 250 [ 20%]  (Adaptation)
+#> Iteration: 100 / 250 [ 40%]  (Adaptation)
+#> Iteration: 150 / 250 [ 60%]  (Adaptation)
+#> Iteration: 200 / 250 [ 80%]  (Adaptation)
+#> Success! Found best value [eta = 1] earlier than expected.
+#> 
+#> Begin stochastic gradient ascent.
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes 
+#>    100          -66.992             1.000            1.000
+#>    200          -66.023             0.507            1.000
+#>    300          -66.735             0.342            0.015
+#>    400          -66.204             0.258            0.015
+#>    500          -66.470             0.207            0.011
+#>    600          -66.254             0.173            0.011
+#>    700          -66.213             0.149            0.008   MEDIAN ELBO CONVERGED
+#> 
+#> Drawing a sample of size 1000 from the approximate posterior... 
+#> COMPLETED.
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-profiles.html b/docs/reference/fit-method-profiles.html index 2a17e670a..48a8cbe07 100644 --- a/docs/reference/fit-method-profiles.html +++ b/docs/reference/fit-method-profiles.html @@ -1,79 +1,16 @@ - - - - - - - -Return profiling data — fit-method-profiles • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Return profiling data — fit-method-profiles • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Return profiling data

- Source: R/fit.R + Source: R/fit.R

The $profiles() method returns a list of data frames with profiling data if any profiling data was written to the profile CSV files. -See save_profile_files() to control where the files are saved.

+See save_profile_files() to control where the files are saved.

Support for profiling Stan programs is available with CmdStan >= 2.26 and requires adding profiling statements to the Stan program.

- 
profiles()
- +
+ 
profiles()
+
-

Value

+
+

Value

+ -

A list of data frames with profiling data if the profiling CSV files +

A list of data frames with profiling data if the profiling CSV files were created.

-

See also

- -

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

- -

Examples

- 

-# \dontrun{
-# first fit a model using MCMC
-mcmc_program <- write_stan_file(
-  'data {
-    int<lower=0> N;
-    int<lower=0,upper=1> y[N];
-  }
-  parameters {
-    real<lower=0,upper=1> theta;
-  }
-  model {
-    profile("likelihood") {
-      y ~ bernoulli(theta);
-    }
-  }
-  generated quantities {
-    int y_rep[N];
-    profile("gq") {
-      y_rep = bernoulli_rng(rep_vector(theta, N));
-    }
-  }
-'
-)
-mod_mcmc <- cmdstan_model(mcmc_program)
-
#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model-d65431ed4576.stan', line 3, column 4: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc
-#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model-d65431ed4576.stan', line 14, column 4: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc

-data <- list(N = 10, y = c(1,1,0,0,0,1,0,1,0,0))
-fit <- mod_mcmc$sample(data = data, seed = 123, refresh = 0)
-
#> Running MCMC with 4 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> Chain 3 finished in 0.0 seconds.
-#> Chain 4 finished in 0.0 seconds.
-#> 
-#> All 4 chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.5 seconds.
-#> 

-fit$profiles()
-
#> [[1]]
-#>         name   thread_id  total_time forward_time reverse_time chain_stack
-#> 1         gq 0x10f232e00 0.000386068  0.000386068  0.000000000           0
-#> 2 likelihood 0x10f232e00 0.001285670  0.000903227  0.000382445        7169
-#>   no_chain_stack autodiff_calls no_autodiff_calls
-#> 1              0              0              1000
-#> 2              0           7169                 1
-#> 
-#> [[2]]
-#>         name   thread_id  total_time forward_time reverse_time chain_stack
-#> 1         gq 0x101daee00 0.000393415  0.000393415  0.000000000           0
-#> 2 likelihood 0x101daee00 0.001305930  0.000932404  0.000373525        7155
-#>   no_chain_stack autodiff_calls no_autodiff_calls
-#> 1              0              0              1000
-#> 2              0           7155                 1
-#> 
-#> [[3]]
-#>         name   thread_id  total_time forward_time reverse_time chain_stack
-#> 1         gq 0x1121b1e00 0.000534289  0.000534289  0.000000000           0
-#> 2 likelihood 0x1121b1e00 0.001384640  0.000993686  0.000390953        6879
-#>   no_chain_stack autodiff_calls no_autodiff_calls
-#> 1              0              0              1000
-#> 2              0           6879                 1
-#> 
-#> [[4]]
-#>         name   thread_id  total_time forward_time reverse_time chain_stack
-#> 1         gq 0x113ecce00 0.000356809  0.000356809  0.000000000           0
-#> 2 likelihood 0x113ecce00 0.001137030  0.000803924  0.000333102        6892
-#>   no_chain_stack autodiff_calls no_autodiff_calls
-#> 1              0              0              1000
-#> 2              0           6892                 1
-#> 
# }
+    

+    
 
-
+
+

Examples

+ 

+# \dontrun{
+# first fit a model using MCMC
+mcmc_program <- write_stan_file(
+  'data {
+    int<lower=0> N;
+    int<lower=0,upper=1> y[N];
+  }
+  parameters {
+    real<lower=0,upper=1> theta;
+  }
+  model {
+    profile("likelihood") {
+      y ~ bernoulli(theta);
+    }
+  }
+  generated quantities {
+    int y_rep[N];
+    profile("gq") {
+      y_rep = bernoulli_rng(rep_vector(theta, N));
+    }
+  }
+'
+)
+mod_mcmc <- cmdstan_model(mcmc_program)
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model-16218658ea62a.stan', line 3, column 4: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model-16218658ea62a.stan', line 14, column 4: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+
+data <- list(N = 10, y = c(1,1,0,0,0,1,0,1,0,0))
+fit <- mod_mcmc$sample(data = data, seed = 123, refresh = 0)
+#> Running MCMC with 4 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> Chain 3 finished in 0.0 seconds.
+#> Chain 4 finished in 0.0 seconds.
+#> 
+#> All 4 chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.5 seconds.
+#> 
+
+fit$profiles()
+#> [[1]]
+#>         name   thread_id total_time forward_time reverse_time chain_stack
+#> 1         gq 0x10afd5e00 0.00037530  0.000375300  0.000000000           0
+#> 2 likelihood 0x10afd5e00 0.00114227  0.000802925  0.000339345        7169
+#>   no_chain_stack autodiff_calls no_autodiff_calls
+#> 1              0              0              1000
+#> 2              0           7169                 1
+#> 
+#> [[2]]
+#>         name   thread_id total_time forward_time reverse_time chain_stack
+#> 1         gq 0x11a045e00 0.00043389  0.000433890  0.000000000           0
+#> 2 likelihood 0x11a045e00 0.00111438  0.000786092  0.000328286        7155
+#>   no_chain_stack autodiff_calls no_autodiff_calls
+#> 1              0              0              1000
+#> 2              0           7155                 1
+#> 
+#> [[3]]
+#>         name   thread_id  total_time forward_time reverse_time chain_stack
+#> 1         gq 0x10c67be00 0.000330582  0.000330582  0.000000000           0
+#> 2 likelihood 0x10c67be00 0.001059290  0.000746087  0.000313204        6879
+#>   no_chain_stack autodiff_calls no_autodiff_calls
+#> 1              0              0              1000
+#> 2              0           6879                 1
+#> 
+#> [[4]]
+#>         name   thread_id  total_time forward_time reverse_time chain_stack
+#> 1         gq 0x108f13e00 0.000365107  0.000365107  0.000000000           0
+#> 2 likelihood 0x108f13e00 0.001159310  0.000815756  0.000343553        6892
+#>   no_chain_stack autodiff_calls no_autodiff_calls
+#> 1              0              0              1000
+#> 2              0           6892                 1
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-return_codes.html b/docs/reference/fit-method-return_codes.html index 6a82b3484..9ccd42588 100644 --- a/docs/reference/fit-method-return_codes.html +++ b/docs/reference/fit-method-return_codes.html @@ -1,76 +1,13 @@ - - - - - - - -Extract return codes from CmdStan — fit-method-return_codes • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract return codes from CmdStan — fit-method-return_codes • cmdstanr - - - - + + -
-
- -
- -
+

Extract return codes from CmdStan

- Source: R/fit.R + Source: R/fit.R
@@ -181,55 +109,65 @@

Extract return codes from CmdStan

from the CmdStan run(s). A return code of 0 indicates a successful run.

- 
return_codes()
- +
+ 
return_codes()
+
-

Value

+
+

Value

+ -

An integer vector of return codes with length equal to the number of +

An integer vector of return codes with length equal to the number of CmdStan runs (number of chains for MCMC and one otherwise).

-

See also

- -

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

- -

Examples

- 
# \dontrun{
-# example with return codes all zero
-fit_mcmc <- cmdstanr_example("schools", method = "sample")
-
#> Warning: 138 of 4000 (3.0%) transitions ended with a divergence.
-#> See https://mc-stan.org/misc/warnings for details.
fit_mcmc$return_codes() # should be all zero
-
#> [1] 0 0 0 0

-# example of non-zero return code (optimization fails for hierarchical model)
-fit_opt <- cmdstanr_example("schools", method = "optimize")
-
#> Optimization terminated with error: 
#>   Line search failed to achieve a sufficient decrease, no more progress can be made
fit_opt$return_codes() # should be non-zero
-
#> [1] 1
# }
-
-
+
+ + +
+

Examples

+ 
# \dontrun{
+# example with return codes all zero
+fit_mcmc <- cmdstanr_example("schools", method = "sample")
+#> Warning: 258 of 4000 (6.0%) transitions ended with a divergence.
+#> See https://mc-stan.org/misc/warnings for details.
+#> Warning: 2 of 4 chains had an E-BFMI less than 0.2.
+#> See https://mc-stan.org/misc/warnings for details.
+fit_mcmc$return_codes() # should be all zero
+#> [1] 0 0 0 0
+
+# example of non-zero return code (optimization fails for hierarchical model)
+fit_opt <- cmdstanr_example("schools", method = "optimize")
+#> Optimization terminated with error: 
+#>   Line search failed to achieve a sufficient decrease, no more progress can be made
+fit_opt$return_codes() # should be non-zero
+#> [1] 1
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-sampler_diagnostics.html b/docs/reference/fit-method-sampler_diagnostics.html index f2c07e2cd..f36572b42 100644 --- a/docs/reference/fit-method-sampler_diagnostics.html +++ b/docs/reference/fit-method-sampler_diagnostics.html @@ -1,78 +1,15 @@ - - - - - - - -Extract sampler diagnostics after MCMC — fit-method-sampler_diagnostics • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Extract sampler diagnostics after MCMC — fit-method-sampler_diagnostics • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Extract sampler diagnostics after MCMC

- Source: R/fit.R + Source: R/fit.R
@@ -182,108 +110,112 @@

Extract sampler diagnostics after MCMC

Extract the values of sampler diagnostics for each iteration and chain of MCMC. To instead get summaries of these diagnostics and associated warning messages use the -$diagnostic_summary() method.

+$diagnostic_summary() method.

+
+ +
+ 
sampler_diagnostics(
+  inc_warmup = FALSE,
+  format = getOption("cmdstanr_draws_format", "draws_array")
+)
- 
sampler_diagnostics(
-  inc_warmup = FALSE,
-  format = getOption("cmdstanr_draws_format", "draws_array")
-)
+
+

Arguments

+
inc_warmup
+

(logical) Should warmup draws be included? Defaults to FALSE.

+ -

Arguments

- - - - - - - - - - -
inc_warmup

(logical) Should warmup draws be included? Defaults to FALSE.

format

(string) The draws format to return. See -draws for details.

+
format
+

(string) The draws format to return. See +draws for details.

-

Value

+
+
+

Value

+ -

Depends on format, but the default is a 3-D -draws_array object (iteration x chain x +

Depends on format, but the default is a 3-D +draws_array object (iteration x chain x variable). The variables for Stan's default MCMC algorithm are "accept_stat__", "stepsize__", "treedepth__", "n_leapfrog__", "divergent__", "energy__".

-

See also

- -

CmdStanMCMC

- -

Examples

- 
# \dontrun{
-fit <- cmdstanr_example("logistic")
-sampler_diagnostics <- fit$sampler_diagnostics()
-str(sampler_diagnostics)
-
#>  'draws_array' num [1:1000, 1:4, 1:6] 3 2 2 2 3 1 1 2 2 2 ...
-#>  - attr(*, "dimnames")=List of 3
-#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   ..$ variable : chr [1:6] "treedepth__" "divergent__" "energy__" "accept_stat__" ...

-library(posterior)
-as_draws_df(sampler_diagnostics)
-
#> # A draws_df: 1000 iterations, 4 chains, and 6 variables
-#>    treedepth__ divergent__ energy__ accept_stat__ stepsize__ n_leapfrog__
-#> 1            3           0       68          1.00       0.87            7
-#> 2            2           0       73          0.60       0.87            7
-#> 3            2           0       72          1.00       0.87            3
-#> 4            2           0       67          0.97       0.87            3
-#> 5            3           0       68          1.00       0.87            7
-#> 6            1           0       68          0.82       0.87            3
-#> 7            1           0       70          0.57       0.87            3
-#> 8            2           0       67          1.00       0.87            3
-#> 9            2           0       67          0.82       0.87            3
-#> 10           2           0       67          0.98       0.87            3
-#> # ... with 3990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# or specify format to get a data frame instead of calling as_draws_df
-fit$sampler_diagnostics(format = "df")
-
#> # A draws_df: 1000 iterations, 4 chains, and 6 variables
-#>    treedepth__ divergent__ energy__ accept_stat__ stepsize__ n_leapfrog__
-#> 1            3           0       68          1.00       0.87            7
-#> 2            2           0       73          0.60       0.87            7
-#> 3            2           0       72          1.00       0.87            3
-#> 4            2           0       67          0.97       0.87            3
-#> 5            3           0       68          1.00       0.87            7
-#> 6            1           0       68          0.82       0.87            3
-#> 7            1           0       70          0.57       0.87            3
-#> 8            2           0       67          1.00       0.87            3
-#> 9            2           0       67          0.82       0.87            3
-#> 10           2           0       67          0.98       0.87            3
-#> # ... with 3990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
# }
+    

+    

+    
See also

+    
CmdStanMCMC

+    

 
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("logistic")
+sampler_diagnostics <- fit$sampler_diagnostics()
+str(sampler_diagnostics)
+#>  'draws_array' num [1:1000, 1:4, 1:6] 1 3 2 2 2 3 2 2 3 2 ...
+#>  - attr(*, "dimnames")=List of 3
+#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   ..$ variable : chr [1:6] "treedepth__" "divergent__" "energy__" "accept_stat__" ...
+
+library(posterior)
+as_draws_df(sampler_diagnostics)
+#> # A draws_df: 1000 iterations, 4 chains, and 6 variables
+#>    treedepth__ divergent__ energy__ accept_stat__ stepsize__ n_leapfrog__
+#> 1            1           0       66          0.91       0.67            3
+#> 2            3           0       66          0.87       0.67            7
+#> 3            2           0       72          0.65       0.67            3
+#> 4            2           0       70          1.00       0.67            3
+#> 5            2           0       66          1.00       0.67            3
+#> 6            3           0       65          0.93       0.67            7
+#> 7            2           0       69          0.70       0.67            3
+#> 8            2           0       70          0.99       0.67            3
+#> 9            3           0       66          0.98       0.67            7
+#> 10           2           0       67          0.90       0.67            3
+#> # ... with 3990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# or specify format to get a data frame instead of calling as_draws_df
+fit$sampler_diagnostics(format = "df")
+#> # A draws_df: 1000 iterations, 4 chains, and 6 variables
+#>    treedepth__ divergent__ energy__ accept_stat__ stepsize__ n_leapfrog__
+#> 1            1           0       66          0.91       0.67            3
+#> 2            3           0       66          0.87       0.67            7
+#> 3            2           0       72          0.65       0.67            3
+#> 4            2           0       70          1.00       0.67            3
+#> 5            2           0       66          1.00       0.67            3
+#> 6            3           0       65          0.93       0.67            7
+#> 7            2           0       69          0.70       0.67            3
+#> 8            2           0       70          0.99       0.67            3
+#> 9            3           0       66          0.98       0.67            7
+#> 10           2           0       67          0.90       0.67            3
+#> # ... with 3990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-save_object.html b/docs/reference/fit-method-save_object.html index ef36b1b08..e721424a6 100644 --- a/docs/reference/fit-method-save_object.html +++ b/docs/reference/fit-method-save_object.html @@ -1,79 +1,16 @@ - - - - - - - -Save fitted model object to a file — fit-method-save_object • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Save fitted model object to a file — fit-method-save_object • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Save fitted model object to a file

- Source: R/fit.R + Source: R/fit.R
-

This method is a wrapper around base::saveRDS() that ensures +

This method is a wrapper around base::saveRDS() that ensures that all posterior draws and diagnostics are saved when saving a fitted model object. Because the contents of the CmdStan output CSV files are only read into R lazily (i.e., as needed), the $save_object() method is the safest way to guarantee that everything has been read in before saving.

- 
save_object(file, ...)
- -

Arguments

- - - - - - - - - - -
file

(string) Path where the file should be saved.

...

Other arguments to pass to base::saveRDS() besides object and file.

- -

See also

+
+ 
save_object(file, ...)
+
-

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

+
+

Arguments

+
file
+

(string) Path where the file should be saved.

-

Examples

- 
# \dontrun{
-fit <- cmdstanr_example("logistic")
 
-temp_rds_file <- tempfile(fileext = ".RDS")
-fit$save_object(file = temp_rds_file)
-rm(fit)
+
...

+
Other arguments to pass to base::saveRDS() besides object and file.

 
-fit <- readRDS(temp_rds_file)
-fit$summary()
-
#> # A tibble: 105 × 10
-#>    variable      mean  median     sd    mad       q5      q95  rhat ess_bulk
-#>    <chr>        <dbl>   <dbl>  <dbl>  <dbl>    <dbl>    <dbl> <dbl>    <dbl>
-#>  1 lp__       -65.9   -65.6   1.43   1.20   -68.8    -64.3     1.00    2291.
-#>  2 alpha        0.379   0.380 0.216  0.220    0.0214   0.735   1.00    4664.
-#>  3 beta[1]     -0.668  -0.663 0.252  0.250   -1.09    -0.259   1.00    4187.
-#>  4 beta[2]     -0.273  -0.265 0.225  0.224   -0.646    0.0856  1.00    3736.
-#>  5 beta[3]      0.682   0.679 0.261  0.258    0.260    1.10    1.00    3626.
-#>  6 log_lik[1]  -0.515  -0.509 0.0979 0.0970  -0.683   -0.366   1.00    4695.
-#>  7 log_lik[2]  -0.401  -0.380 0.144  0.136   -0.666   -0.200   1.00    4367.
-#>  8 log_lik[3]  -0.496  -0.462 0.218  0.200   -0.899   -0.206   1.00    4031.
-#>  9 log_lik[4]  -0.449  -0.432 0.150  0.144   -0.722   -0.237   1.00    3885.
-#> 10 log_lik[5]  -1.18   -1.17  0.276  0.276   -1.68    -0.769   1.00    4020.
-#> # … with 95 more rows, and 1 more variable: ess_tail <dbl>
# }
+
+ -
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("logistic")
+
+temp_rds_file <- tempfile(fileext = ".RDS")
+fit$save_object(file = temp_rds_file)
+rm(fit)
+
+fit <- readRDS(temp_rds_file)
+fit$summary()
+#> # A tibble: 105 × 10
+#>    variable      mean  median     sd    mad       q5      q95  rhat ess_bulk
+#>    <chr>        <num>   <num>  <num>  <num>    <num>    <num> <num>    <num>
+#>  1 lp__       -66.0   -65.6   1.45   1.24   -68.8    -64.3     1.00    2217.
+#>  2 alpha        0.386   0.382 0.217  0.215    0.0345   0.754   1.00    4227.
+#>  3 beta[1]     -0.659  -0.650 0.246  0.246   -1.07    -0.268   1.00    3920.
+#>  4 beta[2]     -0.275  -0.277 0.226  0.226   -0.654    0.0937  1.00    4340.
+#>  5 beta[3]      0.675   0.665 0.275  0.267    0.236    1.15    1.00    3661.
+#>  6 log_lik[1]  -0.514  -0.508 0.0974 0.0967  -0.683   -0.365   1.00    4222.
+#>  7 log_lik[2]  -0.410  -0.389 0.154  0.150   -0.690   -0.198   1.00    4320.
+#>  8 log_lik[3]  -0.500  -0.469 0.218  0.206   -0.908   -0.207   1.00    4270.
+#>  9 log_lik[4]  -0.453  -0.436 0.152  0.146   -0.729   -0.238   1.00    3898.
+#> 10 log_lik[5]  -1.18   -1.16  0.287  0.285   -1.68    -0.760   1.00    3941.
+#> # ℹ 95 more rows
+#> # ℹ 1 more variable: ess_tail <num>
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-save_output_files.html b/docs/reference/fit-method-save_output_files.html index 196500301..9f5eee91a 100644 --- a/docs/reference/fit-method-save_output_files.html +++ b/docs/reference/fit-method-save_output_files.html @@ -1,82 +1,19 @@ - - - - - - - -Save output and data files — fit-method-save_output_files • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Save output and data files — fit-method-save_output_files • cmdstanr - - - - - - - - - - - - - - + + -
-
- -
- -
+

Save output and data files

- Source: R/fit.R + Source: R/fit.R
@@ -193,76 +121,77 @@

Save output and data files

the current file paths without moving any files.

- 
save_output_files(dir = ".", basename = NULL, timestamp = TRUE, random = TRUE)
+    

+    
save_output_files(dir = ".", basename = NULL, timestamp = TRUE, random = TRUE)
+
+save_latent_dynamics_files(
+  dir = ".",
+  basename = NULL,
+  timestamp = TRUE,
+  random = TRUE
+)
+
+save_profile_files(dir = ".", basename = NULL, timestamp = TRUE, random = TRUE)
+
+save_data_file(dir = ".", basename = NULL, timestamp = TRUE, random = TRUE)
+
+output_files(include_failed = FALSE)
+
+profile_files(include_failed = FALSE)
+
+latent_dynamics_files(include_failed = FALSE)
+
+data_file()

+    

+
+    

+    
Arguments

+    
dir

+
(string) Path to directory where the files should be saved.

 
-save_latent_dynamics_files(
-  dir = ".",
-  basename = NULL,
-  timestamp = TRUE,
-  random = TRUE
-)
 
-save_profile_files(dir = ".", basename = NULL, timestamp = TRUE, random = TRUE)
+
basename

+
(string) Base filename to use. See Details.

 
-save_data_file(dir = ".", basename = NULL, timestamp = TRUE, random = TRUE)
 
-output_files(include_failed = FALSE)
+
timestamp

+
(logical) Should a timestamp be added to the file name(s)?
+Defaults to TRUE. See Details.

 
-profile_files(include_failed = FALSE)
 
-latent_dynamics_files(include_failed = FALSE)
+
random

+
(logical) Should random alphanumeric characters be added to the
+end of the file name(s)? Defaults to TRUE. See Details.

 
-data_file()
-

Arguments

- - - - - - - - - - - - - - - - - - - - - - -
dir

(string) Path to directory where the files should be saved.

basename

(string) Base filename to use. See Details.

timestamp

(logical) Should a timestamp be added to the file name(s)? -Defaults to TRUE. See Details.

random

(logical) Should random alphanumeric characters be added to the -end of the file name(s)? Defaults to TRUE. See Details.

include_failed

(logical) Should CmdStan runs that failed also be -included? The default is FALSE.

+
include_failed
+

(logical) Should CmdStan runs that failed also be +included? The default is FALSE.

-

Value

+
+
+

Value

+ -

The $save_* methods print a message with the new file paths and (invisibly) +

The $save_* methods print a message with the new file paths and (invisibly) return a character vector of the new paths (or NA for any that couldn't be copied). They also have the side effect of setting the internal paths in the fitted model object to the new paths.

+ +

The methods without the save_ prefix return character vectors of file paths without moving any files.

-

Details

- +
+
+

Details

For $save_output_files() the files moved to dir will have names of -the form basename-timestamp-id-random, where

    -

  • basename is the user's provided basename argument;

    • -

  • timestamp is of the form format(Sys.time(), "%Y%m%d%H%M");

    • +the form basename-timestamp-id-random, where

    • basename is the user's provided basename argument;

      • +

    • timestamp is of the form format(Sys.time(), "%Y%m%d%H%M");

    • id is the MCMC chain id (or 1 for non MCMC);

    • random contains six random alphanumeric characters;

      • -
    - -

    For $save_latent_dynamics_files() everything is the same as for +

For $save_latent_dynamics_files() everything is the same as for $save_output_files() except "-diagnostic-" is included in the new file name after basename.

For $save_profile_files() everything is the same as for @@ -270,64 +199,69 @@

Details file name after basename.

For $save_data_file() no id is included in the file name because even with multiple MCMC chains the data file is the same.

-

See also

- -

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

- -

Examples

- 
# \dontrun{
-fit <- cmdstanr_example()
-fit$output_files()
-
#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181226-1-549ffb.csv"
-#> [2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181226-2-549ffb.csv"
-#> [3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181226-3-549ffb.csv"
-#> [4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181226-4-549ffb.csv"
fit$data_file()
-
#> [1] "/private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpBlSxPc/temp_libpathbc8d32dd446a/cmdstanr/logistic.data.json"

-# just using tempdir for the example
-my_dir <- tempdir()
-fit$save_output_files(dir = my_dir, basename = "banana")
-
#> Moved 4 files and set internal paths to new locations:
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/banana-202203181226-1-89ddc2.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/banana-202203181226-2-89ddc2.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/banana-202203181226-3-89ddc2.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/banana-202203181226-4-89ddc2.csv
fit$save_output_files(dir = my_dir, basename = "tomato", timestamp = FALSE)
-
#> Moved 4 files and set internal paths to new locations:
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/tomato-1-07c277.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/tomato-2-07c277.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/tomato-3-07c277.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/tomato-4-07c277.csv
fit$save_output_files(dir = my_dir, basename = "lettuce", timestamp = FALSE, random = FALSE)
-
#> Moved 4 files and set internal paths to new locations:
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/lettuce-1.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/lettuce-2.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/lettuce-3.csv
-#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/lettuce-4.csv
# }
+    

+    
 
-
+
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example()
+fit$output_files()
+#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-1-3f09a2.csv"
+#> [2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-2-3f09a2.csv"
+#> [3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-3-3f09a2.csv"
+#> [4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251436-4-3f09a2.csv"
+fit$data_file()
+#> [1] "/private/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpKFjP9Y/temp_libpath15d4242e55c52/cmdstanr/logistic.data.json"
+
+# just using tempdir for the example
+my_dir <- tempdir()
+fit$save_output_files(dir = my_dir, basename = "banana")
+#> Moved 4 files and set internal paths to new locations:
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/banana-202307251436-1-1e5555.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/banana-202307251436-2-1e5555.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/banana-202307251436-3-1e5555.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/banana-202307251436-4-1e5555.csv
+fit$save_output_files(dir = my_dir, basename = "tomato", timestamp = FALSE)
+#> Moved 4 files and set internal paths to new locations:
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/tomato-1-092670.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/tomato-2-092670.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/tomato-3-092670.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/tomato-4-092670.csv
+fit$save_output_files(dir = my_dir, basename = "lettuce", timestamp = FALSE, random = FALSE)
+#> Moved 4 files and set internal paths to new locations:
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/lettuce-1.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/lettuce-2.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/lettuce-3.csv
+#> - /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/lettuce-4.csv
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-summary.html b/docs/reference/fit-method-summary.html index faf9864f2..5713a8b11 100644 --- a/docs/reference/fit-method-summary.html +++ b/docs/reference/fit-method-summary.html @@ -1,53 +1,5 @@ - - - - - - - -Compute a summary table of estimates and diagnostics — fit-method-summary • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Compute a summary table of estimates and diagnostics — fit-method-summary • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

Compute a summary table of estimates and diagnostics

- Source: R/fit.R + Source: R/fit.R

The $summary() method runs -summarise_draws() from the posterior +summarise_draws() from the posterior package and returns the output. For MCMC, only post-warmup draws are included in the summary.

There is also a $print() method that prints the same summary stats but @@ -199,130 +127,174 @@

Compute a summary table of estimates and diagnostics

to be able to return them to the user. See Examples.

- 
summary(variables = NULL, ...)
+
+ 
summary(variables = NULL, ...)
+
-

Arguments

- - - - - - - - - - -
variables

(character vector) The variables to include.

...

Optional arguments to pass to posterior::summarise_draws().

+
+

Arguments

+
variables
+

(character vector) The variables to include.

-

Value

-

The $summary() method returns the tibble data frame created by -posterior::summarise_draws().

-

The $print() method returns the fitted model object itself (invisibly), -which is the standard behavior for print methods in R.

-

See also

+
...
+

Optional arguments to pass to posterior::summarise_draws().

-

CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

+
+
+

Value

+ -

Examples

- 
# \dontrun{
-fit <- cmdstanr_example("logistic")
-fit$summary()
-
#> # A tibble: 105 × 10
-#>    variable      mean  median     sd    mad        q5     q95  rhat ess_bulk
-#>    <chr>        <dbl>   <dbl>  <dbl>  <dbl>     <dbl>   <dbl> <dbl>    <dbl>
-#>  1 lp__       -65.9   -65.6   1.41   1.24   -68.7     -64.3    1.00    2184.
-#>  2 alpha        0.376   0.378 0.218  0.220    0.00751   0.730  1.00    3633.
-#>  3 beta[1]     -0.663  -0.653 0.246  0.245   -1.09     -0.263  1.00    4022.
-#>  4 beta[2]     -0.269  -0.263 0.228  0.226   -0.649     0.107  1.00    3949.
-#>  5 beta[3]      0.676   0.673 0.265  0.271    0.244     1.12   1.00    3975.
-#>  6 log_lik[1]  -0.517  -0.511 0.0992 0.0990  -0.692    -0.367  1.00    3507.
-#>  7 log_lik[2]  -0.404  -0.384 0.147  0.140   -0.665    -0.196  1.00    4695.
-#>  8 log_lik[3]  -0.498  -0.466 0.219  0.205   -0.906    -0.205  1.00    3889.
-#>  9 log_lik[4]  -0.453  -0.432 0.156  0.151   -0.746    -0.230  1.00    3814.
-#> 10 log_lik[5]  -1.18   -1.16  0.278  0.281   -1.66     -0.756  1.00    4117.
-#> # … with 95 more rows, and 1 more variable: ess_tail <dbl>
fit$print()
-
#>    variable   mean median   sd  mad     q5    q95 rhat ess_bulk ess_tail
-#>  lp__       -65.95 -65.64 1.41 1.24 -68.75 -64.27 1.00     2184     3071
-#>  alpha        0.38   0.38 0.22 0.22   0.01   0.73 1.00     3633     3074
-#>  beta[1]     -0.66  -0.65 0.25 0.25  -1.09  -0.26 1.00     4021     2815
-#>  beta[2]     -0.27  -0.26 0.23 0.23  -0.65   0.11 1.00     3948     3214
-#>  beta[3]      0.68   0.67 0.27 0.27   0.24   1.12 1.00     3974     3343
-#>  log_lik[1]  -0.52  -0.51 0.10 0.10  -0.69  -0.37 1.00     3506     2773
-#>  log_lik[2]  -0.40  -0.38 0.15 0.14  -0.67  -0.20 1.00     4694     3405
-#>  log_lik[3]  -0.50  -0.47 0.22 0.20  -0.91  -0.21 1.00     3889     3133
-#>  log_lik[4]  -0.45  -0.43 0.16 0.15  -0.75  -0.23 1.00     3814     3282
-#>  log_lik[5]  -1.18  -1.16 0.28 0.28  -1.66  -0.76 1.00     4117     3034
-#> 
-#>  # showing 10 of 105 rows (change via 'max_rows' argument or 'cmdstanr_max_rows' option)
fit$print(max_rows = 2) # same as print(fit, max_rows = 2)
-
#>  variable   mean median   sd  mad     q5    q95 rhat ess_bulk ess_tail
-#>     lp__  -65.95 -65.64 1.41 1.24 -68.75 -64.27 1.00     2184     3071
-#>     alpha   0.38   0.38 0.22 0.22   0.01   0.73 1.00     3633     3074
-#> 
-#>  # showing 2 of 105 rows (change via 'max_rows' argument or 'cmdstanr_max_rows' option)

-# include only certain variables
-fit$summary("beta")
-
#> # A tibble: 3 × 10
-#>   variable   mean median    sd   mad     q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>  <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 beta[1]  -0.663 -0.653 0.246 0.245 -1.09  -0.263  1.00    4022.    2815.
-#> 2 beta[2]  -0.269 -0.263 0.228 0.226 -0.649  0.107  1.00    3949.    3215.
-#> 3 beta[3]   0.676  0.673 0.265 0.271  0.244  1.12   1.00    3975.    3343.
fit$print(c("alpha", "beta[2]"))
-
#>  variable  mean median   sd  mad    q5  q95 rhat ess_bulk ess_tail
-#>   alpha    0.38   0.38 0.22 0.22  0.01 0.73 1.00     3633     3074
-#>   beta[2] -0.27  -0.26 0.23 0.23 -0.65 0.11 1.00     3948     3214

-# include all variables but only certain summaries
-fit$summary(NULL, c("mean", "sd"))
-
#> # A tibble: 105 × 3
-#>    variable      mean     sd
-#>    <chr>        <dbl>  <dbl>
-#>  1 lp__       -65.9   1.41  
-#>  2 alpha        0.376 0.218 
-#>  3 beta[1]     -0.663 0.246 
-#>  4 beta[2]     -0.269 0.228 
-#>  5 beta[3]      0.676 0.265 
-#>  6 log_lik[1]  -0.517 0.0992
-#>  7 log_lik[2]  -0.404 0.147 
-#>  8 log_lik[3]  -0.498 0.219 
-#>  9 log_lik[4]  -0.453 0.156 
-#> 10 log_lik[5]  -1.18  0.278 
-#> # … with 95 more rows

-# can use functions created from formulas
-# for example, calculate Pr(beta > 0)
-fit$summary("beta", prob_gt_0 = ~ mean(. > 0))
-
#> # A tibble: 3 × 2
-#>   variable prob_gt_0
-#>   <chr>        <dbl>
-#> 1 beta[1]    0.00175
-#> 2 beta[2]    0.116  
-#> 3 beta[3]    0.997  
# }
+
The $summary() method returns the tibble data frame created by
+posterior::summarise_draws().

 
-
+ +

The $print() method returns the fitted model object itself (invisibly), +which is the standard behavior for print methods in R.

+
+ + +
+

Examples

+ 
# \dontrun{
+fit <- cmdstanr_example("logistic")
+fit$summary()
+#> # A tibble: 105 × 10
+#>    variable      mean  median     sd    mad       q5      q95  rhat ess_bulk
+#>    <chr>        <num>   <num>  <num>  <num>    <num>    <num> <num>    <num>
+#>  1 lp__       -66.0   -65.7   1.45   1.26   -68.9    -64.3     1.00    1950.
+#>  2 alpha        0.380   0.376 0.218  0.214    0.0316   0.747   1.00    4585.
+#>  3 beta[1]     -0.669  -0.662 0.254  0.252   -1.09    -0.248   1.00    4669.
+#>  4 beta[2]     -0.277  -0.273 0.231  0.232   -0.665    0.0927  1.00    4400.
+#>  5 beta[3]      0.676   0.677 0.268  0.267    0.242    1.12    1.00    4179.
+#>  6 log_lik[1]  -0.515  -0.508 0.0986 0.0967  -0.686   -0.365   1.00    4439.
+#>  7 log_lik[2]  -0.405  -0.387 0.146  0.140   -0.676   -0.197   1.00    4528.
+#>  8 log_lik[3]  -0.501  -0.464 0.221  0.207   -0.917   -0.210   1.00    4266.
+#>  9 log_lik[4]  -0.450  -0.429 0.155  0.148   -0.729   -0.229   1.00    4175.
+#> 10 log_lik[5]  -1.18   -1.16  0.281  0.278   -1.67    -0.750   1.00    4417.
+#> # ℹ 95 more rows
+#> # ℹ 1 more variable: ess_tail <num>
+fit$print()
+#>    variable   mean median   sd  mad     q5    q95 rhat ess_bulk ess_tail
+#>  lp__       -66.00 -65.68 1.45 1.26 -68.86 -64.27 1.00     1950     2912
+#>  alpha        0.38   0.38 0.22 0.21   0.03   0.75 1.00     4584     2789
+#>  beta[1]     -0.67  -0.66 0.25 0.25  -1.09  -0.25 1.00     4669     3265
+#>  beta[2]     -0.28  -0.27 0.23 0.23  -0.67   0.09 1.00     4399     2567
+#>  beta[3]      0.68   0.68 0.27 0.27   0.24   1.12 1.00     4179     3160
+#>  log_lik[1]  -0.52  -0.51 0.10 0.10  -0.69  -0.37 1.00     4439     2680
+#>  log_lik[2]  -0.40  -0.39 0.15 0.14  -0.68  -0.20 1.00     4528     3426
+#>  log_lik[3]  -0.50  -0.46 0.22 0.21  -0.92  -0.21 1.00     4265     2821
+#>  log_lik[4]  -0.45  -0.43 0.15 0.15  -0.73  -0.23 1.00     4175     2970
+#>  log_lik[5]  -1.18  -1.16 0.28 0.28  -1.67  -0.75 1.00     4417     3002
+#> 
+#>  # showing 10 of 105 rows (change via 'max_rows' argument or 'cmdstanr_max_rows' option)
+fit$print(max_rows = 2) # same as print(fit, max_rows = 2)
+#>  variable   mean median   sd  mad     q5    q95 rhat ess_bulk ess_tail
+#>     lp__  -66.00 -65.68 1.45 1.26 -68.86 -64.27 1.00     1950     2912
+#>     alpha   0.38   0.38 0.22 0.21   0.03   0.75 1.00     4584     2789
+#> 
+#>  # showing 2 of 105 rows (change via 'max_rows' argument or 'cmdstanr_max_rows' option)
+
+# include only certain variables
+fit$summary("beta")
+#> # A tibble: 3 × 10
+#>   variable   mean median    sd   mad     q5     q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>  <num>   <num> <num>    <num>    <num>
+#> 1 beta[1]  -0.669 -0.662 0.254 0.252 -1.09  -0.248   1.00    4669.    3265.
+#> 2 beta[2]  -0.277 -0.273 0.231 0.232 -0.665  0.0927  1.00    4400.    2568.
+#> 3 beta[3]   0.676  0.677 0.268 0.267  0.242  1.12    1.00    4179.    3161.
+fit$print(c("alpha", "beta[2]"))
+#>  variable  mean median   sd  mad    q5  q95 rhat ess_bulk ess_tail
+#>   alpha    0.38   0.38 0.22 0.21  0.03 0.75 1.00     4584     2789
+#>   beta[2] -0.28  -0.27 0.23 0.23 -0.67 0.09 1.00     4399     2567
+
+# include all variables but only certain summaries
+fit$summary(NULL, c("mean", "sd"))
+#> # A tibble: 105 × 3
+#>    variable      mean     sd
+#>    <chr>        <num>  <num>
+#>  1 lp__       -66.0   1.45  
+#>  2 alpha        0.380 0.218 
+#>  3 beta[1]     -0.669 0.254 
+#>  4 beta[2]     -0.277 0.231 
+#>  5 beta[3]      0.676 0.268 
+#>  6 log_lik[1]  -0.515 0.0986
+#>  7 log_lik[2]  -0.405 0.146 
+#>  8 log_lik[3]  -0.501 0.221 
+#>  9 log_lik[4]  -0.450 0.155 
+#> 10 log_lik[5]  -1.18  0.281 
+#> # ℹ 95 more rows
+
+# can use functions created from formulas
+# for example, calculate Pr(beta > 0)
+fit$summary("beta", prob_gt_0 = ~ mean(. > 0))
+#> # A tibble: 3 × 2
+#>   variable prob_gt_0
+#>   <chr>        <num>
+#> 1 beta[1]     0.0035
+#> 2 beta[2]     0.112 
+#> 3 beta[3]     0.995 
+
+# can combine user-specified functions with
+# the default summary functions
+fit$summary(variables = c("alpha", "beta"),
+  posterior::default_summary_measures()[1:4],
+  quantiles = ~ quantile2(., probs = c(0.025, 0.975)),
+  posterior::default_convergence_measures()
+  )
+#> # A tibble: 4 × 10
+#>   variable   mean median    sd   mad    q2.5  q97.5  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 alpha     0.380  0.376 0.218 0.214 -0.0392  0.823  1.00    4585.    2790.
+#> 2 beta[1]  -0.669 -0.662 0.254 0.252 -1.18   -0.180  1.00    4669.    3265.
+#> 3 beta[2]  -0.277 -0.273 0.231 0.232 -0.736   0.165  1.00    4400.    2568.
+#> 4 beta[3]   0.676  0.677 0.268 0.267  0.157   1.20   1.00    4179.    3161.
+
+# the functions need to calculate the appropriate
+# value for a matrix input
+fit$summary(variables = "alpha", dim)
+#> # A tibble: 1 × 3
+#>   variable dim.1 dim.2
+#>   <chr>    <num> <num>
+#> 1 alpha     1000     4
+
+# the usual [stats::var()] is therefore not directly suitable as it
+# will produce a covariance matrix unless the data is converted to a vector
+fit$print(c("alpha", "beta"), var2 = ~var(as.vector(.x)))
+#>  variable var2
+#>   alpha   0.05
+#>   beta[1] 0.06
+#>   beta[2] 0.05
+#>   beta[3] 0.07
+
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-time.html b/docs/reference/fit-method-time.html index dfba38e72..0f3ead7d4 100644 --- a/docs/reference/fit-method-time.html +++ b/docs/reference/fit-method-time.html @@ -1,77 +1,14 @@ - - - - - - - -Report timing of CmdStan runs — fit-method-time • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Report timing of CmdStan runs — fit-method-time • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Report timing of CmdStan runs

- Source: R/fit.R + Source: R/fit.R
@@ -183,74 +111,77 @@

Report timing of CmdStan runs

sampling phases.

- 
time()
- +
+ 
time()
+
-

Value

+
+

Value

+ -

A list with elements

    -

  • total: (scalar) The total run time. For MCMC this may be different than +

    A list with elements

    • total: (scalar) The total run time. For MCMC this may be different than the sum of the chain run times if parallelization was used.

    • chains: (data frame) For MCMC only, timing info for the individual chains. The data frame has columns "chain_id", "warmup", "sampling", and "total".

      • -
    - -

    See also

    - -

    CmdStanMCMC, CmdStanMLE, CmdStanVB, CmdStanGQ

    - -

    Examples

    - 
    # \dontrun{
-fit_mcmc <- cmdstanr_example("logistic", method = "sample")
-fit_mcmc$time()
-
    #> $total
-#> [1] 0.8284969
-#> 
-#> $chains
-#>   chain_id warmup sampling total
-#> 1        1  0.030    0.099 0.129
-#> 2        2  0.023    0.082 0.105
-#> 3        3  0.024    0.091 0.115
-#> 4        4  0.054    0.165 0.219
-#> 
    
-fit_mle <- cmdstanr_example("logistic", method = "optimize")
-fit_mle$time()
-
    #> $total
-#> [1] 0.125402
-#> 
    
-fit_vb <- cmdstanr_example("logistic", method = "variational")
-fit_vb$time()
-
    #> $total
-#> [1] 0.128371
-#> 
    # }
+
+ -
+
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$time()
+#> $total
+#> [1] 0.553201
+#> 
+#> $chains
+#>   chain_id warmup sampling total
+#> 1        1  0.021    0.068 0.089
+#> 2        2  0.022    0.070 0.092
+#> 3        3  0.022    0.076 0.098
+#> 4        4  0.023    0.071 0.094
+#> 
+
+fit_mle <- cmdstanr_example("logistic", method = "optimize")
+fit_mle$time()
+#> $total
+#> [1] 0.1247039
+#> 
+
+fit_vb <- cmdstanr_example("logistic", method = "variational")
+fit_vb$time()
+#> $total
+#> [1] 0.1268399
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/fit-method-unconstrain_draws.html b/docs/reference/fit-method-unconstrain_draws.html new file mode 100644 index 000000000..1a27f650d --- /dev/null +++ b/docs/reference/fit-method-unconstrain_draws.html @@ -0,0 +1,185 @@ + +Transform all parameter draws to the unconstrained scale — fit-method-unconstrain_draws • cmdstanr + + +
+
+ + + +
+
+
+

Transform all parameter draws to the unconstrained scale

+ Source: R/fit.R + +
+ +
+

The $unconstrain_draws() method transforms all parameter draws to the +unconstrained scale. The method returns a list for each chain, containing the parameter +values from each iteration on the unconstrained scale. If called with no arguments, then +the draws within the fit object are unconstrained. Alternatively, either an existing +draws object or a character vector of paths to CSV files can be passed.

+
+ +
+ 
unconstrain_draws(files = NULL, draws = NULL)
+
+ +
+

Arguments

+
files
+

(character vector) The paths to the CmdStan CSV files. These can +be files generated by running CmdStanR or running CmdStan directly.

+ + +
draws
+

A posterior::draws_* object.

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+
+# Unconstrain all internal draws
+unconstrained_internal_draws <- fit_mcmc$unconstrain_draws()
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+
+# Unconstrain external CmdStan CSV files
+unconstrained_csv <- fit_mcmc$unconstrain_draws(files = fit_mcmc$output_files())
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+
+# Unconstrain existing draws object
+unconstrained_draws <- fit_mcmc$unconstrain_draws(draws = fit_mcmc$draws())
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+# }
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-unconstrain_variables.html b/docs/reference/fit-method-unconstrain_variables.html new file mode 100644 index 000000000..558713b7c --- /dev/null +++ b/docs/reference/fit-method-unconstrain_variables.html @@ -0,0 +1,165 @@ + +Transform a set of parameter values to the unconstrained scale — fit-method-unconstrain_variables • cmdstanr + + +
+
+ + + +
+
+
+

Transform a set of parameter values to the unconstrained scale

+ Source: R/fit.R + +
+ +
+

The $unconstrain_variables() method transforms input parameters to +the unconstrained scale

+
+ +
+ 
unconstrain_variables(variables)
+
+ +
+

Arguments

+
variables
+

(list) A list of parameter values to transform, in the same +format as provided to the init argument of the $sample() method.

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+fit_mcmc$unconstrain_variables(list(alpha = 0.5, beta = c(0.7, 1.1, 0.2)))
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+# }
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/fit-method-variable_skeleton.html b/docs/reference/fit-method-variable_skeleton.html new file mode 100644 index 000000000..0bc3100e1 --- /dev/null +++ b/docs/reference/fit-method-variable_skeleton.html @@ -0,0 +1,171 @@ + +Return the variable skeleton needed by the utils::relist function to re-structure a +vector of constrained parameter values to a named list — fit-method-variable_skeleton • cmdstanr + + +
+
+ + + +
+
+
+

Return the variable skeleton needed by the utils::relist function to re-structure a +vector of constrained parameter values to a named list

+ Source: R/fit.R + +
+ +
+

The $variable_skeleton() method returns the variable skeleton

+
+ +
+ 
variable_skeleton(transformed_parameters = TRUE, generated_quantities = TRUE)
+
+ +
+

Arguments

+
transformed_parameters
+

(boolean) Whether to include transformed parameters +in the skeleton (defaults to TRUE)

+ + +
generated_quantities
+

(boolean) Whether to include generated quantities +in the skeleton (defaults to TRUE)

+ +
+ + +
+

Examples

+ 
# \dontrun{
+fit_mcmc <- cmdstanr_example("logistic", method = "sample")
+fit_mcmc$init_model_methods()
+#> Error: Model methods cannot be used with a pre-compiled Stan executable, the model must be compiled again
+fit_mcmc$variable_skeleton()
+#> Error: The method has not been compiled, please call `init_model_methods()` first
+# }
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/index.html b/docs/reference/index.html index cbf3081a9..7c5bfb855 100644 --- a/docs/reference/index.html +++ b/docs/reference/index.html @@ -1,74 +1,12 @@ - - - - - - - -Function reference • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Function reference • cmdstanr - - + + - - -
-
- -
- -
+

Reference

- - - - - - - - - - -
-

Package description

-

An overview of the package and how it differs from RStan.

+ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + + - - - - - - + + - - - + + - - - + + - - - + + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + + + + + + + + - - - - - - - - - - - - - - - - - - - - - - + + + + - - - - - - - - - - - - - - -
+

Package description

+

An overview of the package and how it differs from RStan.

-

cmdstanr-package

+
+

cmdstanr-package cmdstanr CmdStanR

CmdStanR: the R interface to CmdStan

-

Installing and setting the path to CmdStan

-

Install CmdStan, assuming the necessary C++ toolchain.

+
+

Installing and setting the path to CmdStan

+

Install CmdStan, assuming the necessary C++ toolchain.

+

install_cmdstan() rebuild_cmdstan() cmdstan_make_local() check_cmdstan_toolchain()

Install CmdStan or clean and rebuild an existing installation

+

set_cmdstan_path() cmdstan_path() cmdstan_version()

Get or set the file path to the CmdStan installation

-

Running CmdStan from R

-

Run CmdStan from R.

+
+

Running CmdStan from R

+

Run CmdStan from R.

+

cmdstan_model()

Create a new CmdStanModel object

+

CmdStanModel

CmdStanModel objects

+

check_syntax()

Check syntax of a Stan program

+

compile()

Compile a Stan program

+

diagnose()

Run Stan's diagnose method

+
+

expose_functions()

+

Expose Stan functions to R

format()

Run stanc's auto-formatter on the model code.

+

generate_quantities()

Run Stan's standalone generated quantities method

+

optimize()

Run Stan's optimization algorithms

+

sample()

Run Stan's MCMC algorithms

+

sample_mpi()

Run Stan's MCMC algorithms with MPI

+

variables()

Input and output variables of a Stan program

+

variational()

Run Stan's variational approximation algorithms

+

cmdstanr_example() print_example_program()

Fit models for use in examples

-

Fitted model objects and methods

+
+

Fitted model objects and methods

+

CmdStanMCMC

CmdStanMCMC objects

+

CmdStanMLE

CmdStanMLE objects

+

CmdStanVB

CmdStanVB objects

+

CmdStanGQ

CmdStanGQ objects

+

CmdStanDiagnose

CmdStanDiagnose objects

+

cmdstan_summary() cmdstan_diagnose()

Run CmdStan's stansummary and diagnose utilities

+

code()

Return Stan code

+
+

constrain_variables()

+

Transform a set of unconstrained parameter values to the constrained scale

diagnostic_summary()

Sampler diagnostic summaries and warnings

+

draws()

Extract posterior draws

+
+

grad_log_prob()

+

Calculate the log-probability and the gradient w.r.t. each input for a +given vector of unconstrained parameters

gradients()

Extract gradients after diagnostic mode

+
+

hessian()

+

Calculate the log-probability , the gradient w.r.t. each input, and the hessian +for a given vector of unconstrained parameters

init()

Extract user-specified initial values

+
+

init_model_methods()

+

Compile additional methods for accessing the model log-probability function +and parameter constraining and unconstraining.

inv_metric()

Extract inverse metric (mass matrix) after MCMC

+
+

log_prob()

+

Calculate the log-probability given a provided vector of unconstrained parameters.

loo()

Leave-one-out cross-validation (LOO-CV)

+

lp() lp_approx()

Extract log probability (target)

+

metadata()

Extract metadata from CmdStan CSV files

+

mle()

Extract (penalized) maximum likelihood estimate after optimization

+

num_chains()

Extract number of chains after MCMC

+

output()

Access console output

+

profiles()

Return profiling data

+

return_codes()

Extract return codes from CmdStan

+

sampler_diagnostics()

Extract sampler diagnostics after MCMC

+

save_object()

Save fitted model object to a file

+

save_output_files() save_latent_dynamics_files() save_profile_files() save_data_file() output_files() profile_files() latent_dynamics_files() data_file()

Save output and data files

+

summary()

Compute a summary table of estimates and diagnostics

+

time()

Report timing of CmdStan runs

-

Other tools for working with CmdStan

+
+

unconstrain_draws()

+

Transform all parameter draws to the unconstrained scale

+

unconstrain_variables()

+

Transform a set of parameter values to the unconstrained scale

+

variable_skeleton()

+

Return the variable skeleton needed by the utils::relist function to re-structure a +vector of constrained parameter values to a named list

+

expose_functions()

+

Expose Stan functions to R

+

Other tools for working with CmdStan

+

read_cmdstan_csv() as_cmdstan_fit()

Read CmdStan CSV files into R

+

write_stan_json()

Write data to a JSON file readable by CmdStan

+

write_stan_file()

Write Stan code to a file

+

draws_to_csv()

Write posterior draws objects to CSV files suitable for running standalone generated quantities with CmdStan.

-

Using CmdStanR with knitr and R Markdown

+
+

as_mcmc.list()

+

Convert CmdStanMCMC to mcmc.list

+

as_draws(<CmdStanMCMC>) as_draws(<CmdStanMLE>) as_draws(<CmdStanVB>) as_draws(<CmdStanGQ>)

+

Create a draws object from a CmdStanR fitted model object

+

Using CmdStanR with knitr and R Markdown

+

register_knitr_engine()

Register CmdStanR's knitr engine for Stan

+

eng_cmdstan()

CmdStan knitr engine for Stan

- +
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/install_cmdstan.html b/docs/reference/install_cmdstan.html index 85948b739..5995ec079 100644 --- a/docs/reference/install_cmdstan.html +++ b/docs/reference/install_cmdstan.html @@ -1,53 +1,5 @@ - - - - - - - -Install CmdStan or clean and rebuild an existing installation — install_cmdstan • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Install CmdStan or clean and rebuild an existing installation — install_cmdstan • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Install CmdStan or clean and rebuild an existing installation

- Source: R/install.R + Source: R/install.R

The install_cmdstan() function attempts to download and -install the latest release of CmdStan. +install the latest release of CmdStan. Installing a previous release or a new release candidate is also possible by specifying the version or release_url argument. See the first few sections of the CmdStan -installation guide +installation guide for details on the C++ toolchain required for installing CmdStan.

The rebuild_cmdstan() function cleans and rebuilds the CmdStan installation. Use this function in case of any issues when compiling models.

@@ -213,154 +141,161 @@

Install CmdStan or clean and rebuild an existing installation

be called directly by the user.

- 
install_cmdstan(
-  dir = NULL,
-  cores = getOption("mc.cores", 2),
-  quiet = FALSE,
-  overwrite = FALSE,
-  timeout = 1200,
-  version = NULL,
-  release_url = NULL,
-  cpp_options = list(),
-  check_toolchain = TRUE
-)
-
-rebuild_cmdstan(
-  dir = cmdstan_path(),
-  cores = getOption("mc.cores", 2),
-  quiet = FALSE,
-  timeout = 600
-)
-
-cmdstan_make_local(dir = cmdstan_path(), cpp_options = NULL, append = TRUE)
-
-check_cmdstan_toolchain(fix = FALSE, quiet = FALSE)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
dir

(string) The path to the directory in which to install CmdStan. +

+ 
install_cmdstan(
+  dir = NULL,
+  cores = getOption("mc.cores", 2),
+  quiet = FALSE,
+  overwrite = FALSE,
+  timeout = 1200,
+  version = NULL,
+  release_url = NULL,
+  cpp_options = list(),
+  check_toolchain = TRUE,
+  wsl = FALSE
+)
+
+rebuild_cmdstan(
+  dir = cmdstan_path(),
+  cores = getOption("mc.cores", 2),
+  quiet = FALSE,
+  timeout = 600
+)
+
+cmdstan_make_local(dir = cmdstan_path(), cpp_options = NULL, append = TRUE)
+
+check_cmdstan_toolchain(fix = FALSE, quiet = FALSE)
+
+ +
+

Arguments

+
dir
+

(string) The path to the directory in which to install CmdStan. The default is to install it in a directory called .cmdstan within the -user's home directory (i.e, file.path(Sys.getenv("HOME"), ".cmdstan")).

cores

(integer) The number of CPU cores to use to parallelize building +user's home directory (i.e, file.path(Sys.getenv("HOME"), ".cmdstan")).

+ + +
cores
+

(integer) The number of CPU cores to use to parallelize building CmdStan and speed up installation. If cores is not specified then the default is to look for the option "mc.cores", which can be set for an -entire R session by options(mc.cores=value). If the "mc.cores" option -has not been set then the default is 2.

quiet

(logical) For install_cmdstan(), should the verbose output +entire R session by options(mc.cores=value). If the "mc.cores" option +has not been set then the default is 2.

+ + +
quiet
+

(logical) For install_cmdstan(), should the verbose output from the system processes be suppressed when building the CmdStan binaries? The default is FALSE. For check_cmdstan_toolchain(), should the function suppress printing informational messages? The default is FALSE. -If TRUE only errors will be printed.

overwrite

(logical) Should CmdStan still be downloaded and installed +If TRUE only errors will be printed.

+ + +
overwrite
+

(logical) Should CmdStan still be downloaded and installed even if an installation of the same version is found in dir? The default is FALSE, in which case an informative error is thrown instead of -overwriting the user's installation.

timeout

(positive real) Timeout (in seconds) for the build stage of -the installation.

version

(string) The CmdStan release version to install. The default +overwriting the user's installation.

+ + +
timeout
+

(positive real) Timeout (in seconds) for the build stage of +the installation.

+ + +
version
+

(string) The CmdStan release version to install. The default is NULL, which downloads the latest stable release from -https://github.com/stan-dev/cmdstan/releases.

release_url

(string) The URL for the specific CmdStan release or -release candidate to install. See https://github.com/stan-dev/cmdstan/releases. +https://github.com/stan-dev/cmdstan/releases.

+ + +
release_url
+

(string) The URL for the specific CmdStan release or +release candidate to install. See https://github.com/stan-dev/cmdstan/releases. The URL should point to the tarball (.tar.gz. file) itself, e.g., release_url="https://github.com/stan-dev/cmdstan/releases/download/v2.25.0/cmdstan-2.25.0.tar.gz". -If both version and release_url are specified then version will be used.

cpp_options

(list) Any makefile flags/variables to be written to -the make/local file. For example, list("CXX" = "clang++") will force -the use of clang for compilation.

check_toolchain

(logical) Should install_cmdstan() attempt to check +If both version and release_url are specified then version will be used.

+ + +
cpp_options
+

(list) Any makefile flags/variables to be written to +the make/local file. For example, list("CXX" = "clang++") will force +the use of clang for compilation.

+ + +
check_toolchain
+

(logical) Should install_cmdstan() attempt to check that the required toolchain is installed and properly configured. The -default is TRUE.

append

(logical) For cmdstan_make_local(), should the listed +default is TRUE.

+ + +
wsl
+

(logical) Should CmdStan be installed and run through the Windows +Subsystem for Linux (WSL). The default is FALSE.

+ + +
append
+

(logical) For cmdstan_make_local(), should the listed makefile flags be appended to the end of the existing make/local file? -The default is TRUE. If FALSE the file is overwritten.

fix

For check_cmdstan_toolchain(), should CmdStanR attempt to fix +The default is TRUE. If FALSE the file is overwritten.

+ + +
fix
+

For check_cmdstan_toolchain(), should CmdStanR attempt to fix any detected toolchain problems? Currently this option is only available on Windows. The default is FALSE, in which case problems are only reported -along with suggested fixes.

+along with suggested fixes.

-

Value

+
+
+

Value

+ -

For cmdstan_make_local(), if cpp_options=NULL then the existing +

For cmdstan_make_local(), if cpp_options=NULL then the existing contents of make/local are returned without writing anything, otherwise the updated contents are returned.

+
-

Examples

- 
# \dontrun{
-check_cmdstan_toolchain()
-
#> The C++ toolchain required for CmdStan is setup properly!

-# install_cmdstan(cores = 4)
-
-cpp_options <- list(
-  "CXX" = "clang++",
-  "CXXFLAGS+= -march=native",
-  PRECOMPILED_HEADERS = TRUE
-)
-# cmdstan_make_local(cpp_options = cpp_options)
-# rebuild_cmdstan()
-# }
-
-
+
+

Examples

+ 
# \dontrun{
+check_cmdstan_toolchain()
+#> The C++ toolchain required for CmdStan is setup properly!
+
+# install_cmdstan(cores = 4)
+
+cpp_options <- list(
+  "CXX" = "clang++",
+  "CXXFLAGS+= -march=native",
+  PRECOMPILED_HEADERS = TRUE
+)
+# cmdstan_make_local(cpp_options = cpp_options)
+# rebuild_cmdstan()
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-check_syntax.html b/docs/reference/model-method-check_syntax.html index 933874799..1d321677e 100644 --- a/docs/reference/model-method-check_syntax.html +++ b/docs/reference/model-method-check_syntax.html @@ -1,77 +1,14 @@ - - - - - - - -Check syntax of a Stan program — model-method-check_syntax • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Check syntax of a Stan program — model-method-check_syntax • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Check syntax of a Stan program

- Source: R/model.R + Source: R/model.R
-

The $check_syntax() method of a CmdStanModel object +

The $check_syntax() method of a CmdStanModel object checks the Stan program for syntax errors and returns TRUE (invisibly) if parsing succeeds. If invalid syntax in found an error is thrown.

- 
check_syntax(
-  pedantic = FALSE,
-  include_paths = NULL,
-  stanc_options = list(),
-  quiet = FALSE
-)
+
+ 
check_syntax(
+  pedantic = FALSE,
+  include_paths = NULL,
+  stanc_options = list(),
+  quiet = FALSE
+)
+
-

Arguments

- - - - - - - - - - - - - - - - - - -
pedantic

(logical) Should pedantic mode be turned on? The default is +

+

Arguments

+
pedantic
+

(logical) Should pedantic mode be turned on? The default is FALSE. Pedantic mode attempts to warn you about potential issues in your -Stan program beyond syntax errors. For details see the Pedantic mode chapter in -the Stan Reference Manual.

include_paths

(character vector) Paths to directories where Stan +Stan program beyond syntax errors. For details see the Pedantic mode chapter in +the Stan Reference Manual.

+ + +
include_paths
+

(character vector) Paths to directories where Stan should look for files specified in #include directives in the Stan -program.

stanc_options

(list) Any other Stan-to-C++ transpiler options to be +program.

+ + +
stanc_options
+

(list) Any other Stan-to-C++ transpiler options to be used when compiling the model. See the documentation for the -$compile() method for details.

quiet

(logical) Should informational messages be suppressed? The +$compile() method for details.

+ + +
quiet
+

(logical) Should informational messages be suppressed? The default is FALSE, which will print a message if the Stan program is valid or the compiler error message if there are syntax errors. If TRUE, only -the error message will be printed.

+the error message will be printed.

-

Value

+
+
+

Value

+ -

The $check_syntax() method returns TRUE (invisibly) if the model +

The $check_syntax() method returns TRUE (invisibly) if the model is valid.

-

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other CmdStanModel methods: -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variables, -model-method-variational

- -

Examples

- 
# \dontrun{
-file <- write_stan_file("
-data {
-  int N;
-  int y[N];
-}
-parameters {
-  // should have <lower=0> but omitting to demonstrate pedantic mode
-  real lambda;
-}
-model {
-  y ~ poisson(lambda);
-}
-")
-mod <- cmdstan_model(file, compile = FALSE)
-
-# the program is syntactically correct, however...
-mod$check_syntax()
-
#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 4, column 2: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc
#> Stan program is syntactically correct

-# pedantic mode will warn that lambda should be constrained to be positive
-# and that lambda has no prior distribution
-mod$check_syntax(pedantic = TRUE)
-
#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 4, column 2: Declaration
-#>     of arrays by placing brackets after a variable name is deprecated and
-#>     will be removed in Stan 2.32.0. Instead use the array keyword before the
-#>     type. This can be changed automatically using the auto-format flag to
-#>     stanc
-#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 11, column 14: A
-#>     poisson distribution is given parameter lambda as a rate parameter
-#>     (argument 1), but lambda was not constrained to be strictly positive.
-#> Warning: The parameter lambda has no priors.
#> Stan program is syntactically correct
# }
+
The Stan and CmdStan documentation:
Other CmdStanModel methods: 
+model-method-compile,
+model-method-diagnose,
+model-method-expose_functions,
+model-method-format,
+model-method-generate-quantities,
+model-method-optimize,
+model-method-sample_mpi,
+model-method-sample,
+model-method-variables,
+model-method-variational

+
-
+
+

Examples

+ 
# \dontrun{
+file <- write_stan_file("
+data {
+  int N;
+  int y[N];
+}
+parameters {
+  // should have <lower=0> but omitting to demonstrate pedantic mode
+  real lambda;
+}
+model {
+  y ~ poisson(lambda);
+}
+")
+mod <- cmdstan_model(file, compile = FALSE)
+
+# the program is syntactically correct, however...
+mod$check_syntax()
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 4, column 2: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+#> Stan program is syntactically correct
+
+# pedantic mode will warn that lambda should be constrained to be positive
+# and that lambda has no prior distribution
+mod$check_syntax(pedantic = TRUE)
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 4, column 2: Declaration
+#>     of arrays by placing brackets after a variable name is deprecated and
+#>     will be removed in Stan 2.33.0. Instead use the array keyword before the
+#>     type. This can be changed automatically using the auto-format flag to
+#>     stanc
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model_74b584fd23ea8c78eceda86d1d425fe8.stan', line 11, column 14: A
+#>     poisson distribution is given parameter lambda as a rate parameter
+#>     (argument 1), but lambda was not constrained to be strictly positive.
+#> Warning: The parameter lambda has no priors. This means either no prior is
+#>     provided, or the prior(s) depend on data variables. In the later case,
+#>     this may be a false positive.
+#> Stan program is syntactically correct
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-compile.html b/docs/reference/model-method-compile.html index b0c7f4429..63b0fa515 100644 --- a/docs/reference/model-method-compile.html +++ b/docs/reference/model-method-compile.html @@ -1,53 +1,5 @@ - - - - - - - -Compile a Stan program — model-method-compile • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Compile a Stan program — model-method-compile • cmdstanr - - - - - - - - - - - + + - - - -
-
- -
- -
+

Compile a Stan program

- Source: R/model.R + Source: R/model.R
-

The $compile() method of a CmdStanModel object checks the +

The $compile() method of a CmdStanModel object checks the syntax of the Stan program, translates the program to C++, and creates a compiled executable. To just check the syntax of a Stan program without -compiling it use the $check_syntax() method +compiling it use the $check_syntax() method instead.

In most cases the user does not need to explicitly call the $compile() -method as compilation will occur when calling cmdstan_model(). However it -is possible to set compile=FALSE in the call to cmdstan_model() and +method as compilation will occur when calling cmdstan_model(). However it +is possible to set compile=FALSE in the call to cmdstan_model() and subsequently call the $compile() method directly.

After compilation, the paths to the executable and the .hpp file containing the generated C++ code are available via the $exe_file() and @@ -207,167 +135,195 @@

Compile a Stan program

$save_hpp_file(dir).

- 
compile(
-  quiet = TRUE,
-  dir = NULL,
-  pedantic = FALSE,
-  include_paths = NULL,
-  user_header = NULL,
-  cpp_options = list(),
-  stanc_options = list(),
-  force_recompile = getOption("cmdstanr_force_recompile", default = FALSE),
-  threads = FALSE
-)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
quiet

(logical) Should the verbose output from CmdStan during +

+ 
compile(
+  quiet = TRUE,
+  dir = NULL,
+  pedantic = FALSE,
+  include_paths = NULL,
+  user_header = NULL,
+  cpp_options = list(),
+  stanc_options = list(),
+  force_recompile = getOption("cmdstanr_force_recompile", default = FALSE),
+  compile_model_methods = FALSE,
+  compile_hessian_method = FALSE,
+  compile_standalone = FALSE,
+  threads = FALSE
+)
+
+ +
+

Arguments

+
quiet
+

(logical) Should the verbose output from CmdStan during compilation be suppressed? The default is TRUE, but if you encounter an error we recommend trying again with quiet=FALSE to see more of the -output.

dir

(string) The path to the directory in which to store the CmdStan +output.

+ + +
dir
+

(string) The path to the directory in which to store the CmdStan executable (or .hpp file if using $save_hpp_file()). The default is the -same location as the Stan program.

pedantic

(logical) Should pedantic mode be turned on? The default is +same location as the Stan program.

+ + +
pedantic
+

(logical) Should pedantic mode be turned on? The default is FALSE. Pedantic mode attempts to warn you about potential issues in your -Stan program beyond syntax errors. For details see the Pedantic mode chapter in +Stan program beyond syntax errors. For details see the Pedantic mode chapter in the Stan Reference Manual. Note: to do a pedantic check for a model without compiling it or for a model that is already compiled the -$check_syntax() method can be used instead.

include_paths

(character vector) Paths to directories where Stan +$check_syntax() method can be used instead.

+ + +
include_paths
+

(character vector) Paths to directories where Stan should look for files specified in #include directives in the Stan -program.

user_header

(string) The path to a C++ file (with a .hpp extension) -to compile with the Stan model.

cpp_options

(list) Any makefile options to be used when compiling the +program.

+ + +
user_header
+

(string) The path to a C++ file (with a .hpp extension) +to compile with the Stan model.

+ + +
cpp_options
+

(list) Any makefile options to be used when compiling the model (STAN_THREADS, STAN_MPI, STAN_OPENCL, etc.). Anything you would otherwise write in the make/local file. For an example of using threading see the Stan case study -Reduce Sum: A Minimal Example.

stanc_options

(list) Any Stan-to-C++ transpiler options to be used +Reduce Sum: A Minimal Example.

+ + +
stanc_options
+

(list) Any Stan-to-C++ transpiler options to be used when compiling the model. See the Examples section below as well as the stanc chapter of the CmdStan Guide for more details on available options: -https://mc-stan.org/docs/cmdstan-guide/stanc.html.

force_recompile

(logical) Should the model be recompiled even if was +https://mc-stan.org/docs/cmdstan-guide/stanc.html.

+ + +
force_recompile
+

(logical) Should the model be recompiled even if was not modified since last compiled. The default is FALSE. Can also be set -via a global cmdstanr_force_recompile option.

threads

Deprecated and will be removed in a future release. Please -turn on threading via cpp_options = list(stan_threads = TRUE) instead.

- -

Value

- -

The $compile() method is called for its side effect of creating the -executable and adding its path to the CmdStanModel object, but it also -returns the CmdStanModel object invisibly.

-

After compilation, the $exe_file(), $hpp_file(), and $save_hpp_file() -methods can be used and return file paths.

-

See also

- -

The $check_syntax() method to check +via a global cmdstanr_force_recompile option.

+ + +
compile_model_methods
+

(logical) Compile additional model methods +(log_prob(), grad_log_prob(), constrain_variables(), +unconstrain_variables()).

+ + +
compile_hessian_method
+

(logical) Should the (experimental) hessian() method be +be compiled with the model methods?

+ + +
compile_standalone
+

(logical) Should functions in the Stan model be +compiled for use in R? If TRUE the functions will be available via the +functions field in the compiled model object. This can also be done after +compilation using the +$expose_functions() method.

+ + +
threads
+

Deprecated and will be removed in a future release. Please +turn on threading via cpp_options = list(stan_threads = TRUE) instead.

+ +
+
+

Value

+ + +

The $compile() method is called for its side effect of creating the +executable and adding its path to the CmdStanModel object, but it also +returns the CmdStanModel object invisibly.

+ + +

After compilation, the $exe_file(), $hpp_file(), and $save_hpp_file()

+ + +

methods can be used and return file paths.

+
+
+

See also

+

The $check_syntax() method to check Stan syntax or enable pedantic model without compiling.

The CmdStanR website -(mc-stan.org/cmdstanr) for online +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other CmdStanModel methods: -model-method-check_syntax, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variables, -model-method-variational

- -

Examples

- 
# \dontrun{
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-
-# by default compilation happens when cmdstan_model() is called.
-# to delay compilation until calling the $compile() method set compile=FALSE
-mod <- cmdstan_model(file, compile = FALSE)
-mod$compile()
-mod$exe_file()
-
#> [1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1/examples/bernoulli/bernoulli"

-# turn on threading support (for using functions that support within-chain parallelization)
-mod$compile(force_recompile = TRUE, cpp_options = list(stan_threads = TRUE))
-mod$exe_file()
-
#> [1] "/Users/jgabry/.cmdstan/cmdstan-2.29.1/examples/bernoulli/bernoulli"

-# turn on pedantic mode (new in Stan v2.24)
-file_pedantic <- write_stan_file("
-parameters {
-  real sigma;  // pedantic mode will warn about missing <lower=0>
-}
-model {
-  sigma ~ exponential(1);
-}
-")
-mod <- cmdstan_model(file_pedantic, pedantic = TRUE)
-
#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model-d65457acfdf5.stan', line 6, column 2: Parameter
-#>     sigma is given a exponential distribution, which has strictly positive
-#>     support, but sigma was not constrained to be strictly positive.

-# }
-
-
+

The Stan and CmdStan documentation:

Other CmdStanModel methods: +model-method-check_syntax, +model-method-diagnose, +model-method-expose_functions, +model-method-format, +model-method-generate-quantities, +model-method-optimize, +model-method-sample_mpi, +model-method-sample, +model-method-variables, +model-method-variational

+
+ +
+

Examples

+ 
# \dontrun{
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+
+# by default compilation happens when cmdstan_model() is called.
+# to delay compilation until calling the $compile() method set compile=FALSE
+mod <- cmdstan_model(file, compile = FALSE)
+mod$compile()
+mod$exe_file()
+#> [1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2/examples/bernoulli/bernoulli"
+
+# turn on threading support (for using functions that support within-chain parallelization)
+mod$compile(force_recompile = TRUE, cpp_options = list(stan_threads = TRUE))
+mod$exe_file()
+#> [1] "/Users/jgabry/.cmdstan/cmdstan-2.32.2/examples/bernoulli/bernoulli"
+
+# turn on pedantic mode (new in Stan v2.24)
+file_pedantic <- write_stan_file("
+parameters {
+  real sigma;  // pedantic mode will warn about missing <lower=0>
+}
+model {
+  sigma ~ exponential(1);
+}
+")
+mod <- cmdstan_model(file_pedantic, pedantic = TRUE)
+#> Warning in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model-1621885c67f.stan', line 6, column 2: Parameter
+#>     sigma is given a exponential distribution, which has strictly positive
+#>     support, but sigma was not constrained to be strictly positive.
+
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-diagnose.html b/docs/reference/model-method-diagnose.html index 602c34747..548b5959b 100644 --- a/docs/reference/model-method-diagnose.html +++ b/docs/reference/model-method-diagnose.html @@ -1,79 +1,16 @@ - - - - - - - -Run Stan's diagnose method — model-method-diagnose • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run Stan's diagnose method — model-method-diagnose • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Run Stan's diagnose method

- Source: R/model.R + Source: R/model.R
-

The $diagnose() method of a CmdStanModel object +

The $diagnose() method of a CmdStanModel object runs Stan's basic diagnostic feature that will calculate the gradients of the initial state and compare them with gradients calculated by finite differences. Discrepancies between the two indicate that there is a problem with the model or initial states or else there is a bug in Stan.

- 
diagnose(
-  data = NULL,
-  seed = NULL,
-  init = NULL,
-  output_dir = NULL,
-  output_basename = NULL,
-  epsilon = NULL,
-  error = NULL
-)
+
+ 
diagnose(
+  data = NULL,
+  seed = NULL,
+  init = NULL,
+  output_dir = NULL,
+  output_basename = NULL,
+  epsilon = NULL,
+  error = NULL
+)
+
-

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
data

(multiple options) The data to use for the variables specified in -the data block of the Stan program. One of the following:

    -

  • A named list of R objects with the names corresponding to variables +

    +

    Arguments

    +
    data
    +

    (multiple options) The data to use for the variables specified in +the data block of the Stan program. One of the following:

    • A named list of R objects with the names corresponding to variables declared in the data block of the Stan program. Internally this list is then -written to JSON for CmdStan using write_stan_json(). See -write_stan_json() for details on the conversions performed on R objects +written to JSON for CmdStan using write_stan_json(). See +write_stan_json() for details on the conversions performed on R objects before they are passed to Stan.

    • A path to a data file compatible with CmdStan (JSON or R dump). See the appendices in the CmdStan guide for details on using these formats.

    • NULL or an empty list if the Stan program has no data block.

      • -
seed

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. + + + +

seed
+

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to @@ -224,19 +151,18 @@

Arg chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each -chain.

init

(multiple options) The initialization method to use for the +chain.

+ + +
init
+

(multiple options) The initialization method to use for the variables declared in the parameters block of the Stan program. One of -the following:

    -

  • A real number x>0. This initializes all parameters randomly between +the following:

    • A real number x>0. This initializes all parameters randomly between [-x,x] on the unconstrained parameter space.;

    • The number 0. This initializes all parameters to 0;

    • A character vector of paths (one per chain) to JSON or Rdump files containing initial values for all or some parameters. See -write_stan_json() to write R objects to JSON files compatible with +write_stan_json() to write R objects to JSON files compatible with CmdStan.

    • A list of lists containing initial values for all or some parameters. For MCMC the list should contain a sublist for each chain. For optimization and @@ -249,105 +175,104 @@

      Arg has argument chain_id it will be supplied with the chain id (from 1 to number of chains) when called to generate the initial values. See Examples.

      • -
output_dir

(string) A path to a directory where CmdStan should write + + + +

output_dir
+

(string) A path to a directory where CmdStan should write its output CSV files. For interactive use this can typically be left at NULL (temporary directory) since CmdStanR makes the CmdStan output (posterior draws and diagnostics) available in R via methods of the fitted -model objects. The behavior of output_dir is as follows:

    -

  • If NULL (the default), then the CSV files are written to a temporary +model objects. The behavior of output_dir is as follows:

    • If NULL (the default), then the CSV files are written to a temporary directory and only saved permanently if the user calls one of the $save_* methods of the fitted model object (e.g., -$save_output_files()). These temporary +$save_output_files()). These temporary files are removed when the fitted model object is -garbage collected (manually or automatically).

      • +garbage collected (manually or automatically).

    • If a path, then the files are created in output_dir with names corresponding to the defaults used by $save_output_files().

      • -
output_basename

(string) A string to use as a prefix for the names of + + + +

output_basename
+

(string) A string to use as a prefix for the names of the output CSV files of CmdStan. If NULL (the default), the basename of the output CSV files will be comprised from the model name, timestamp, and -5 random characters.

epsilon

(positive real) The finite difference step size. Default -value is 1e-6.

error

(positive real) The error threshold. Default value is 1e-6.

+5 random characters.

-

Value

-

A CmdStanDiagnose object.

-

See also

+
epsilon
+

(positive real) The finite difference step size. Default +value is 1e-6.

-

The CmdStanR website -(mc-stan.org/cmdstanr) for online -documentation and tutorials.

-

The Stan and CmdStan documentation:

-

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variables, -model-method-variational

+
error
+

(positive real) The error threshold. Default value is 1e-6.

-

Examples

- 
# \dontrun{
-test <- cmdstanr_example("logistic", method = "diagnose")
+

+    

+    
Value

+    
 
-# retrieve the gradients
-test$gradients()
-
#>   param_idx      value     model finite_diff        error
-#> 1         0  1.3362100  -9.33442    -9.33442  2.88445e-08
-#> 2         1  1.7816500 -32.65590   -32.65590 -2.89809e-10
-#> 3         2 -0.0718337  -2.56116    -2.56116 -2.13148e-08
-#> 4         3 -1.9334500  29.16060    29.16060 -3.65847e-08
# }
+
A CmdStanDiagnose object.

+    

+    

+    
See also

+    
The CmdStanR website
+(mc-stan.org/cmdstanr) for online
+documentation and tutorials.

+
The Stan and CmdStan documentation:
Other CmdStanModel methods: 
+model-method-check_syntax,
+model-method-compile,
+model-method-expose_functions,
+model-method-format,
+model-method-generate-quantities,
+model-method-optimize,
+model-method-sample_mpi,
+model-method-sample,
+model-method-variables,
+model-method-variational

+    

 
-
+
+

Examples

+ 
# \dontrun{
+test <- cmdstanr_example("logistic", method = "diagnose")
+
+# retrieve the gradients
+test$gradients()
+#>   param_idx     value     model finite_diff        error
+#> 1         0 -1.310990  28.37700    28.37700 -1.92695e-08
+#> 2         1 -1.931430   8.55341     8.55341 -1.60934e-08
+#> 3         2  0.751196 -12.17180   -12.17180 -2.15070e-08
+#> 4         3 -1.808550  25.15020    25.15020  6.91254e-09
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-expose_functions.html b/docs/reference/model-method-expose_functions.html new file mode 100644 index 000000000..ad674e0e1 --- /dev/null +++ b/docs/reference/model-method-expose_functions.html @@ -0,0 +1,231 @@ + +Expose Stan functions to R — model-method-expose_functions • cmdstanr + + +
+
+ + + +
+
+
+

Expose Stan functions to R

+ Source: R/model.R + +
+ +
+

The $expose_functions() method of a CmdStanModel object +will compile the functions in the Stan program's functions block and +expose them for use in R. This can also be specified via the +compile_standalone argument to the $compile() +method.

+

This method is also available for fitted model objects (CmdStanMCMC, CmdStanVB, etc.). +See Examples.

+

Note: there may be many compiler warnings emitted during compilation but +these can be ignored so long as they are warnings and not errors.

+
+ +
+ 
expose_functions(global = FALSE, verbose = FALSE)
+
+ +
+

Arguments

+
global
+

(logical) Should the functions be added to the Global +Environment? The default is FALSE, in which case the functions are +available via the functions field of the R6 object.

+ + +
verbose
+

(logical) Should detailed information about generated code be +printed to the console? Defaults to FALSE.

+ +
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online +documentation and tutorials.

+

The Stan and CmdStan documentation:

Other CmdStanModel methods: +model-method-check_syntax, +model-method-compile, +model-method-diagnose, +model-method-format, +model-method-generate-quantities, +model-method-optimize, +model-method-sample_mpi, +model-method-sample, +model-method-variables, +model-method-variational

+
+ +
+

Examples

+ 
# \dontrun{
+stan_file <- write_stan_file(
+ "
+ functions {
+   real a_plus_b(real a, real b) {
+     return a + b;
+   }
+ }
+ parameters {
+   real x;
+ }
+ model {
+   x ~ std_normal();
+ }
+ "
+)
+mod <- cmdstan_model(stan_file)
+mod$expose_functions()
+#> Compiling standalone functions...
+mod$functions$a_plus_b(1, 2)
+#> [1] 3
+
+fit <- mod$sample(refresh = 0)
+#> Running MCMC with 4 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> Chain 3 finished in 0.0 seconds.
+#> Chain 4 finished in 0.0 seconds.
+#> 
+#> All 4 chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.6 seconds.
+#> 
+fit$expose_functions() # already compiled because of above but this would compile them otherwise
+#> Functions already compiled, nothing to do!
+fit$functions$a_plus_b(1, 2)
+#> [1] 3
+# }
+
+
+
+
+
+ +
+ + +
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+
+ +
+

Site built with pkgdown 2.0.7.

+
+ +
+ + + + + + + + diff --git a/docs/reference/model-method-format.html b/docs/reference/model-method-format.html index 9bb9f867e..041779776 100644 --- a/docs/reference/model-method-format.html +++ b/docs/reference/model-method-format.html @@ -1,77 +1,14 @@ - - - - - - - -Run stanc's auto-formatter on the model code. — model-method-format • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run stanc's auto-formatter on the model code. — model-method-format • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Run stanc's auto-formatter on the model code.

- Source: R/model.R + Source: R/model.R
-

The $format() method of a CmdStanModel object +

The $format() method of a CmdStanModel object runs stanc's auto-formatter on the model code. Either saves the formatted model directly back to the file or prints it for inspection.

- 
format(
-  overwrite_file = FALSE,
-  canonicalize = FALSE,
-  backup = TRUE,
-  max_line_length = NULL,
-  quiet = FALSE
-)
+
+ 
format(
+  overwrite_file = FALSE,
+  canonicalize = FALSE,
+  backup = TRUE,
+  max_line_length = NULL,
+  quiet = FALSE
+)
+
+ +
+

Arguments

+
overwrite_file
+

(logical) Should the formatted code be written back +to the input model file. The default is FALSE.

+ -

Arguments

- - - - - - - - - - - - - - - - - - - - - - -
overwrite_file

(logical) Should the formatted code be written back -to the input model file. The default is FALSE.

canonicalize

(list or logical) Defines whether or not the compiler +

canonicalize
+

(list or logical) Defines whether or not the compiler should 'canonicalize' the Stan model, removing things like deprecated syntax. Default is FALSE. If TRUE, all canonicalizations are run. You can also supply a list of strings which represent options. In that case the options -are passed to stanc (new in Stan 2.29). See the User's guide section -for available canonicalization options.

backup

(logical) If TRUE, create stanfile.bak backups before +are passed to stanc (new in Stan 2.29). See the User's guide section +for available canonicalization options.

+ + +
backup
+

(logical) If TRUE, create stanfile.bak backups before writing to the file. Disable this option if you're sure you have other copies of the file or are using a version control system like Git. Defaults -to TRUE. The value is ignored if overwrite_file = FALSE.

max_line_length

(integer) The maximum length of a line when formatting. -The default is NULL, which defers to the default line length of stanc.

quiet

(logical) Should informational messages be suppressed? The -default is FALSE.

+to TRUE. The value is ignored if overwrite_file = FALSE.

-

Value

-

The $format() method returns TRUE (invisibly) if the model -is valid.

-

See also

+
max_line_length
+

(integer) The maximum length of a line when formatting. +The default is NULL, which defers to the default line length of stanc.

-

The CmdStanR website -(mc-stan.org/cmdstanr) for online -documentation and tutorials.

-

The Stan and CmdStan documentation:

-

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variables, -model-method-variational

+
quiet
+

(logical) Should informational messages be suppressed? The +default is FALSE.

+ +
+
+

Value

+ -

Examples

- 
# \dontrun{
-file <- write_stan_file("
-data {
-  int N;
-  int y[N];
-}
-parameters {
-  real                     lambda;
-}
-model {
-  target +=
- poisson_log(y | lambda);
-}
-")
-mod <- cmdstan_model(file, compile = FALSE)
-mod$format(canonicalize = TRUE)
-
#> Semantic error in '/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/model_44779dbadddd7550b5164157b77f4d11.stan', line 11, column 1 to column 24:
-#>    -------------------------------------------------
-#>      9:  model {
-#>     10:    target +=
-#>     11:   poisson_log(y | lambda);
-#>           ^
-#>     12:  }
-#>     13:  
-#>    -------------------------------------------------
-#> 
-#> Only functions with names ending in _lpdf, _lupdf, _lpmf, _lupmf, _cdf, _lcdf, _lccdf can make use of conditional notation.
#> Error: Syntax error found! See the message above for more information.
# }
+
The $format() method returns TRUE (invisibly) if the model
+is valid.

+    

+    

+    
See also

+    
The CmdStanR website
+(mc-stan.org/cmdstanr) for online
+documentation and tutorials.

+
The Stan and CmdStan documentation:
Other CmdStanModel methods: 
+model-method-check_syntax,
+model-method-compile,
+model-method-diagnose,
+model-method-expose_functions,
+model-method-generate-quantities,
+model-method-optimize,
+model-method-sample_mpi,
+model-method-sample,
+model-method-variables,
+model-method-variational

+    

 
-
+
+

Examples

+ 
# \dontrun{
+file <- write_stan_file("
+data {
+  int N;
+  int y[N];
+}
+parameters {
+  real                     lambda;
+}
+model {
+  target +=
+ poisson_lpmf(y | lambda);
+}
+")
+mod <- cmdstan_model(file, compile = FALSE)
+mod$format(canonicalize = TRUE)
+#> data {
+#>   int N;
+#>   array[N] int y;
+#> }
+#> parameters {
+#>   real lambda;
+#> }
+#> model {
+#>   target += poisson_lpmf(y | lambda);
+#> }
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-generate-quantities.html b/docs/reference/model-method-generate-quantities.html index 8e5b08ce5..63e57c29a 100644 --- a/docs/reference/model-method-generate-quantities.html +++ b/docs/reference/model-method-generate-quantities.html @@ -1,77 +1,14 @@ - - - - - - - -Run Stan's standalone generated quantities method — model-method-generate-quantities • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run Stan's standalone generated quantities method — model-method-generate-quantities • cmdstanr - - - - - - - - - - - + + - - - -
-
- -
- -
+

Run Stan's standalone generated quantities method

- Source: R/model.R + Source: R/model.R
-

The $generate_quantities() method of a CmdStanModel object +

The $generate_quantities() method of a CmdStanModel object runs Stan's standalone generated quantities to obtain generated quantities based on previously fitted parameters.

- 
generate_quantities(
-  fitted_params,
-  data = NULL,
-  seed = NULL,
-  output_dir = NULL,
-  output_basename = NULL,
-  sig_figs = NULL,
-  parallel_chains = getOption("mc.cores", 1),
-  threads_per_chain = NULL,
-  opencl_ids = NULL
-)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
fitted_params

(multiple options) The parameter draws to use. One of -the following:

+
+ 
generate_quantities(
+  fitted_params,
+  data = NULL,
+  seed = NULL,
+  output_dir = NULL,
+  output_basename = NULL,
+  sig_figs = NULL,
+  parallel_chains = getOption("mc.cores", 1),
+  threads_per_chain = NULL,
+  opencl_ids = NULL
+)
+
-

NOTE: if you plan on making many calls to $generate_quantities() then the +

+

Arguments

+
fitted_params
+

(multiple options) The parameter draws to use. One of +the following:

NOTE: if you plan on making many calls to $generate_quantities() then the most efficient option is to pass the paths of the CmdStan CSV output files (this avoids CmdStanR having to rewrite the draws contained in the fitted model object to CSV each time). If you no longer have the CSV files you can -use draws_to_csv() once to write them and then pass the resulting file -paths to $generate_quantities() as many times as needed.

data

(multiple options) The data to use for the variables specified in -the data block of the Stan program. One of the following:

    -

  • A named list of R objects with the names corresponding to variables +use draws_to_csv() once to write them and then pass the resulting file +paths to $generate_quantities() as many times as needed.

    + + +
    data
    +

    (multiple options) The data to use for the variables specified in +the data block of the Stan program. One of the following:

    • A named list of R objects with the names corresponding to variables declared in the data block of the Stan program. Internally this list is then -written to JSON for CmdStan using write_stan_json(). See -write_stan_json() for details on the conversions performed on R objects +written to JSON for CmdStan using write_stan_json(). See +write_stan_json() for details on the conversions performed on R objects before they are passed to Stan.

    • A path to a data file compatible with CmdStan (JSON or R dump). See the appendices in the CmdStan guide for details on using these formats.

    • NULL or an empty list if the Stan program has no data block.

      • -
seed

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. + + + +

seed
+

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to @@ -239,198 +163,200 @@

Arg chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each -chain.

output_dir

(string) A path to a directory where CmdStan should write +chain.

+ + +
output_dir
+

(string) A path to a directory where CmdStan should write its output CSV files. For interactive use this can typically be left at NULL (temporary directory) since CmdStanR makes the CmdStan output (posterior draws and diagnostics) available in R via methods of the fitted -model objects. The behavior of output_dir is as follows:

    -

  • If NULL (the default), then the CSV files are written to a temporary +model objects. The behavior of output_dir is as follows:

    • If NULL (the default), then the CSV files are written to a temporary directory and only saved permanently if the user calls one of the $save_* methods of the fitted model object (e.g., -$save_output_files()). These temporary +$save_output_files()). These temporary files are removed when the fitted model object is -garbage collected (manually or automatically).

      • +garbage collected (manually or automatically).

    • If a path, then the files are created in output_dir with names corresponding to the defaults used by $save_output_files().

      • -
output_basename

(string) A string to use as a prefix for the names of + + + +

output_basename
+

(string) A string to use as a prefix for the names of the output CSV files of CmdStan. If NULL (the default), the basename of the output CSV files will be comprised from the model name, timestamp, and -5 random characters.

sig_figs

(positive integer) The number of significant figures used +5 random characters.

+ + +
sig_figs
+

(positive integer) The number of significant figures used when storing the output values. By default, CmdStan represent the output values with 6 significant figures. The upper limit for sig_figs is 18. Increasing this value will result in larger output CSV files and thus an -increased usage of disk space.

parallel_chains

(positive integer) The maximum number of MCMC chains +increased usage of disk space.

+ + +
parallel_chains
+

(positive integer) The maximum number of MCMC chains to run in parallel. If parallel_chains is not specified then the default is to look for the option "mc.cores", which can be set for an entire R -session by options(mc.cores=value). If the "mc.cores" option has not -been set then the default is 1.

threads_per_chain

(positive integer) If the model was -compiled with threading support, the number of +session by options(mc.cores=value). If the "mc.cores" option has not +been set then the default is 1.

+ + +
threads_per_chain
+

(positive integer) If the model was +compiled with threading support, the number of threads to use in parallelized sections within an MCMC chain (e.g., when using the Stan functions reduce_sum() or map_rect()). This is in contrast with parallel_chains, which specifies the number of chains to run in parallel. The actual number of CPU cores used is parallel_chains*threads_per_chain. For an example of using threading see the Stan case study -Reduce Sum: A Minimal Example.

opencl_ids

(integer vector of length 2) The platform and +Reduce Sum: A Minimal Example.

+ + +
opencl_ids
+

(integer vector of length 2) The platform and device IDs of the OpenCL device to use for fitting. The model must be compiled with cpp_options = list(stan_opencl = TRUE) for this -argument to have an effect.

- -

Value

+argument to have an effect.

-

A CmdStanGQ object.

-

See also

+
+
+

Value

+ -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

A CmdStanGQ object.

+
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variables, -model-method-variational

- -

Examples

- 
# \dontrun{
-# first fit a model using MCMC
-mcmc_program <- write_stan_file(
-  "data {
-    int<lower=0> N;
-    int<lower=0,upper=1> y[N];
-  }
-  parameters {
-    real<lower=0,upper=1> theta;
-  }
-  model {
-    y ~ bernoulli(theta);
-  }"
-)
-mod_mcmc <- cmdstan_model(mcmc_program)
-
-data <- list(N = 10, y = c(1,1,0,0,0,1,0,1,0,0))
-fit_mcmc <- mod_mcmc$sample(data = data, seed = 123, refresh = 0)
-
#> Running MCMC with 4 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> Chain 3 finished in 0.0 seconds.
-#> Chain 4 finished in 0.0 seconds.
-#> 
-#> All 4 chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.6 seconds.
-#> 

-# stan program for standalone generated quantities
-# (could keep model block, but not necessary so removing it)
-gq_program <- write_stan_file(
-  "data {
-    int<lower=0> N;
-    int<lower=0,upper=1> y[N];
-  }
-  parameters {
-    real<lower=0,upper=1> theta;
-  }
-  generated quantities {
-    int y_rep[N] = bernoulli_rng(rep_vector(theta, N));
-  }"
-)
-
-mod_gq <- cmdstan_model(gq_program)
-fit_gq <- mod_gq$generate_quantities(fit_mcmc, data = data, seed = 123)
-
#> Running standalone generated quantities after 4 MCMC chains, 1 chain at a time ...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> Chain 3 finished in 0.0 seconds.
-#> Chain 4 finished in 0.0 seconds.
-#> 
-#> All 4 chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.5 seconds.
str(fit_gq$draws())
-
#>  'draws_array' int [1:1000, 1:4, 1:10] 0 0 0 0 1 0 1 0 0 0 ...
-#>  - attr(*, "dimnames")=List of 3
-#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   ..$ variable : chr [1:10] "y_rep[1]" "y_rep[2]" "y_rep[3]" "y_rep[4]" ...

-library(posterior)
-as_draws_df(fit_gq$draws())
-
#> # A draws_df: 1000 iterations, 4 chains, and 10 variables
-#>    y_rep[1] y_rep[2] y_rep[3] y_rep[4] y_rep[5] y_rep[6] y_rep[7] y_rep[8]
-#> 1         0        0        0        0        0        0        0        0
-#> 2         0        1        1        1        1        1        1        1
-#> 3         0        0        0        0        0        1        1        1
-#> 4         0        0        0        0        0        1        0        1
-#> 5         1        1        1        1        1        0        1        0
-#> 6         0        0        0        0        1        0        1        1
-#> 7         1        0        0        0        1        1        1        0
-#> 8         0        1        0        0        0        1        0        1
-#> 9         0        1        0        1        1        1        1        1
-#> 10        0        1        0        1        0        0        0        0
-#> # ... with 3990 more draws, and 2 more variables
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
# }
-
-
+

The Stan and CmdStan documentation:

Other CmdStanModel methods: +model-method-check_syntax, +model-method-compile, +model-method-diagnose, +model-method-expose_functions, +model-method-format, +model-method-optimize, +model-method-sample_mpi, +model-method-sample, +model-method-variables, +model-method-variational

+
+ +
+

Examples

+ 
# \dontrun{
+# first fit a model using MCMC
+mcmc_program <- write_stan_file(
+  "data {
+    int<lower=0> N;
+    int<lower=0,upper=1> y[N];
+  }
+  parameters {
+    real<lower=0,upper=1> theta;
+  }
+  model {
+    y ~ bernoulli(theta);
+  }"
+)
+mod_mcmc <- cmdstan_model(mcmc_program)
+
+data <- list(N = 10, y = c(1,1,0,0,0,1,0,1,0,0))
+fit_mcmc <- mod_mcmc$sample(data = data, seed = 123, refresh = 0)
+#> Running MCMC with 4 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> Chain 3 finished in 0.0 seconds.
+#> Chain 4 finished in 0.0 seconds.
+#> 
+#> All 4 chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.6 seconds.
+#> 
+
+# stan program for standalone generated quantities
+# (could keep model block, but not necessary so removing it)
+gq_program <- write_stan_file(
+  "data {
+    int<lower=0> N;
+    int<lower=0,upper=1> y[N];
+  }
+  parameters {
+    real<lower=0,upper=1> theta;
+  }
+  generated quantities {
+    int y_rep[N] = bernoulli_rng(rep_vector(theta, N));
+  }"
+)
+
+mod_gq <- cmdstan_model(gq_program)
+fit_gq <- mod_gq$generate_quantities(fit_mcmc, data = data, seed = 123)
+#> Running standalone generated quantities after 4 MCMC chains, 1 chain at a time ...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> Chain 3 finished in 0.0 seconds.
+#> Chain 4 finished in 0.0 seconds.
+#> 
+#> All 4 chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.5 seconds.
+str(fit_gq$draws())
+#>  'draws_array' int [1:1000, 1:4, 1:10] 0 0 0 0 1 0 1 0 0 0 ...
+#>  - attr(*, "dimnames")=List of 3
+#>   ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   ..$ variable : chr [1:10] "y_rep[1]" "y_rep[2]" "y_rep[3]" "y_rep[4]" ...
+
+library(posterior)
+as_draws_df(fit_gq$draws())
+#> # A draws_df: 1000 iterations, 4 chains, and 10 variables
+#>    y_rep[1] y_rep[2] y_rep[3] y_rep[4] y_rep[5] y_rep[6] y_rep[7] y_rep[8]
+#> 1         0        0        0        0        0        0        0        0
+#> 2         0        1        1        1        1        1        1        1
+#> 3         0        0        0        0        0        1        1        1
+#> 4         0        0        0        0        0        1        0        1
+#> 5         1        1        1        1        1        0        1        0
+#> 6         0        0        0        0        1        0        1        1
+#> 7         1        0        0        0        1        1        1        0
+#> 8         0        1        0        0        0        1        0        1
+#> 9         0        1        0        1        1        1        1        1
+#> 10        0        1        0        1        0        0        0        0
+#> # ... with 3990 more draws, and 2 more variables
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-optimize-1.png b/docs/reference/model-method-optimize-1.png index d7b8b97f8c19824bd7f4936ec52513cc8fae745f..27fd88217456b4a7d0da11a09599ea685749e06d 100644 GIT binary patch literal 15795 zcmeHu2UJtp+IGN#j)ie-BOr{@Y!pF2YOo9#5KyYrpfn{ykPaaPE90P|AR=8wL;|73 z&;kTSsRDx#Is~Ky2oPElNJ#ki!I^c}x7NRIzw5jATQ`fW6HZRq=Pl3sJn!DQf5Fsn z+jnB$!CKDfQHN+A?_Ee23`BFc8_N~X-M8I^%$xx$Wtq&!(VF}v-MIJmm5mR%TrqH zBa!fLZ0GfZTs=H8Pj46R)9)=h4&OU?izRd|@y(4K*+Lg!J@u?$U7d-cUhnYx-mvN!BbtBUKM; zhncpex%fmTzR9YMma?HP9}LFfDD?Zz$9sEw)zVrBqQ4Z7$@b~3Z_9GU7sVtLd%L>c z7I{ItL43s4zsBRYZ{Hqv^3)v?iDZXF7VB*)C@7elJn84C*b3Pn_Pysdk6Gc_!2J>$ zERv`<`x;H!Sku@m6I^g)(%{TWBrY?Qy>|oGV-(OIw zKc0&x*x)!xl%dZFOTKAC;4+n^!+)E?2YAS`)nW^~wQ^}0DcSg)@o*}#O_LJ6{q zmg`SS`y8hzVDOtc@~r9GH}Z((&B2x|4VpHA+QAVAcfY(d_@xwZ8=VhYFUI1at$+?PG(Tf{q1lM%p|dV4!o?dZ!mQ{f#?iC?)nJtwHH8 zQrO_|ovPj*g84alDI3C8&iVl^U&?xasgp7orO?J?i-<^96-jo}doeGLDPb$0*Uy2* zpZ{XWY4!)`Q$*)qK$Sm0$+8!mS;d;(|8SB4mk>nH1zFQs#58RNW-GV~2DQZ8Dik5L zz(6|Ytu8fDD{C*+;Z`$5QK%l4E*N`Nh0Lt5{)<>!i)d?nWqun+#zy;OZs_WZN!J0( zI0*3aw*nZvBW6SOuY%Xh=YEEU>F-?#VgKI@g?|kC??lo+PG5Uw%q~jA92tK%Sv%*!G0ai>eORRSV z^HeVTp=jei2=_OY!)9&e(gD`sjMTBBU{BRgCG}!p`wv%uKgRtJKK4I&`kK`?m*q~9 z&1zY392>6Ws1Xe}rm78|F!<_vr-%*7&Dr6@y=~lwXfvMVD5*n&3KDMc)DvOYp2bo_ zV6#QtPt?lg^h$ZiDg){__d1`Ne(R|S_0Yd6^6(_`21hS~0_l2z~eh4W*6{)E3>+l2p1y4JmF{FK1&GODcEhDF)Y z@+Y@%-%d>a%#$2eZ7RbPyR}(+)w3l<8F%{ObTY;&QppsBX4=xymGz-&DfST0DC?oo zm>7vmb$~d4N^F*vG#zqU^W@35%VEz^;w9Gu0!Y?2HU!brIRrxgl1w|G zY28z`rg!%}Xw1|Ja?}8pZyQ@!{=myYiVNR!R!^c{SO2VWiWnbRA84tzpCTy5im6#q zw_&iKl)DViwFabcuAD7h8Sqq3+FX9Nq|0nSiV4@S1;z5Or=b7MwZEIa|1X)&iv}Fb zvEas@~!R=$z~g4FOAmo(Yd znHg`IbQS4jhx%KLBf7?K6e?nDTYf%s@0FZT*^wDHTl>?=K7DFPc5n!smw ze@?tBr6Gwcl-oBCqz5?|K34JLVT1IGYE8>bGoW<8LQ0oO0oeIFw)YefQc_ZYAO(0{ zU0t0jf3P*I(%$rz2)CW33qw=8Cv2yu_b=*hT=(w(qiRq<`YQ< zXFHwS%#&OPi-1AdAo`)z)?#%WKHXloAtvPUV#?kJkcbHsLSe(;qFPZ{u7x zrCU*Ww-G5#NtK4BU!sUZ&0q`~P(R5BwaOy5ajALN*T$uD61s13;0bNGHan9>5ZktR zwma6AD#d@xJPZ)c8**}ALfj2%@Qa(p6Wml_g_+cU5j6FH^o0$w%5ptdOR`qyD&nBQ zYjA1+9|Puvs*ZCF<`Cbuvbb?d2>K^I6t?GUui z&O%@+1O&Q&mCdOiONjub7eAOwfXn@-)$6|zZ~xBRgj!i~adCf;-8{qe))yXbC5dnF zWtEx0^GR)zhhy{Ym~Ax^&56Ga!5mDy!yXLpDk0G&tgupcJB z2s5v%+jqP{PtPB2aqQd{YJJc6k`V0KONMu)0V!AZ=gxqEs2!v8visbL301~-qWJ9v zY=Iy502E64P@8Ux2Gxq%3BStdp2&yeP9|89hBe@j%7;DsdC<*AKx@0fpT`YMU><0A zCgNKN9No1huV_M5=$oM@y#q)3NE*CeuTl+^vG>hZ$P)5vdsGG*TI~Xk2OKL z^iQ9lSrb_>+p_p*3cmu8=T~0;j=(JqGDGz(>TZTr%|q3xiWC;+D_A;8A@wtV+OqV8 zhO@)bl~MuIgjId{c7Q|jd!8zjtyP0cY0j>kOULic#F`1i@*SX};3M58L`nLwi3dva z;Iz*_VeUe4`2_p?-q{{yI-f1jx$y;BYLAa)n3 zV^Jn;?N|K=ipaUy%VD5N;-H?UF-in=eda@$US9X-ykE98ec~NwemZv?bUm&$<8YpJ z73&nq<4ATl4}I-s5E_k!`bb4yC@^1w$D5j(PR`Ee)9;r|O|%u^0|8lbo*nrvLIcB| z0*pKb*eR{Bpx~lcZ-+n-PD`&kT?C4ON`j>o_5J&sqVm7w7Zq7E7>wKA{$W>KT!yJP z#>$fAHbdV2Tj$ySi$;G!BS;T}_P^V5jiDETIxbTO&ffh=yWXS&a@jvUL;p%z<;jzbC}sftPUBs+U&w`KjpQSTD-EF`MQjpg zu3L~2IFrKp+0hW02d|QKV5i>+l&4BAx4ssddCaV}VoEK3EBag0s(3IpD=?MATxqra z*4NhG_}`w}Up?jh=Tq_c$5t#@6krIg1^8PCQq?M&DY!k7%i!IF2QfLa2R1)m@`^aP zX@l3PhXq^ZEk{TCOD+V8GLsqB8EO8O*x@Yh@S_C3Dz za$nAO$U87|apm=tT|ewtScoy6VS96UwKqH-JCw+fi+QCO9d~cv(ta$BK}>6@WG;lv zU}$I%HZNYR2eA)c8)UyB>!CwJ0OJA0JVnwVgKfkF{@KWT z<*AU<34I0V~6V81Zzw9`CEQ|hri5xAL##X=Ae+5mzN7IzH#(fE@>c;76`_k z^}3=$w{3j>5yJ8a-;-6zX0zKpNpjZfcB&={nPm25*99^?*jClT1KmXX;2*O3{r&xS zz3B)YpESB+WgBt$RM7US*N5w*Que93yNWpuw91I=_rE|ij=-v<_~sqVwq)a;{C00c zd6EX&+p_m7wQqw{CsC#+ff_vaD*XAv-?_RQ2x!|iRCrc;p;v7k|<*N6AowEYkKmsL%v1lu@ zF`{NCMEoyYO((vT)Ra%??rt|{PbQeu=MH_zrtN^SZ$jRVIQ){RFyTFv-4|o!1G@^t z0onv!-x)y@uHfto!fQJsZvF60MeKSCo%LtqZ}VCs^S_sXa`9hj3r}<))bi0sb!l2Q z3@a67TB8DGF`aD}8}BdNGxrQU08`Lk^P&u*awRr3Z0N~=d;IYO#c2hywaxoI1NXs9 z)^ZdUBcpv3;bHGPb)YHNHB37)CLk#$yjxNKhc!gWPx3ts0_5EbJFmbEFv6#UD9Hb? z4b%_JdME(fFJ{XsO&y}&WbJQKbaP-77psxi!`M&O;zaI!)rG=IoHh}>lN=kI0+w== z^j~iJ7SvB;O8e_?*)}P-BU(293r7U_ocfH^fj#gc0_S=j}aoskeK}=vb22)TC4_(2%WpUcGdrw zM0(41UrY4-yzs&TUGhDn*m<)1^`ja5jP)6LdD2wRfYZ_1HwG?*_Mo=qT%XLA4O#hA z-`mN{eme4SA@U2=9jCgyR-*JTZ`~Tbf}yTcO|pESvzfN=Rf9zbZ;kwCSIg(cDwB|Vlb`1NfEcQ zw#svK+Tt*&6P|&GF#(^Ghx+DHo0=4FBYgTN_m#46Wl4@TcUI_{NxtWbH6-P}q4EU} zl1${H+^UBG@#sF6fOv2W{VE?n3q{vI>_GEyBsQ8qa_X=G!A9jm^E7Ef@= zMCU6useSYbaDBzG&b{H`HbUDU_ayq=yTbRvQ%uh&Ci7uUe$8XrdA&A6kBUV) z_0X{E5cW~q5B#Vr7i1k%xkpzn_DLt_0=PrNxWIczOvFhmb1LuM%<+T=<&`4#VK`B% zSi7h+mj_N~rgCI5_DDXtWXwRGu|y%%2#zkR#+~15ix@baA>>fOJ$+(|d^;A4z3(kd zRn<|nHyU!&t*uPv8;#H%iyU?GQ{;qM;cVojuJr{)*=U@MW=Upu-1%IijF?A{_(|~e%R=aNiG(EctP^|mJ3UF3 z*{qN)S9#8bFY^>l)XtOCRgMsf@6zoN2C0n+DNVUon1PltsgLSY=&u25Nqf{i)Km9M z{lK0yFY+cH8K*okA)zUcwmL(7vXf{XiY+ER-1?-KU!w3qFmU|&=c{vKk$7ojzk+}} zlPyEE8-Lyxi-gM=w(OV;2)`hjzSl8DL2uIEP&T3aE04UXcBJZpxnlI^-jh4hv0gD- zMa`Obhbl-*yLh}?^0U3eukbrGaTPTkqMS{7Vy4x;kX*O}JIxkeJf(d^8b|K610LEc zud5LL=y*4e?4X#r_+@tXKAY53;@&59oBEpgDM~`ZdwSu8>+70FMIB}C61v`o(cLd6 z>%>I~eIl~;r~2zrcELA37-U&z6ZXz}_q@v*XzDNl%Miyt@3U3$ad!R*i&aQ%_+Y$U zZ)tFY&Z}p5hJrtOswIVf_lCRAKZ;0&`816B zdXntaz5BYRe4S6bHzy~6rKWU|T{Yma+FS7^N$R9_@%PH9;I+tRYrzlwlA!AGhD^p# zBzwOyzb!mePQCYF&}94dD}x4@nNqs%Xq=54VvmB|(=Cc6FY0%{|M(zW=51W2{>_V^ zqr99=;bzO!?=X~j5{;d5G0EK|XMi+rb-3|qcXy$iaT%GsP}+jc?H;YVGmbi(h{~%_ zH`8TV#*E;za~r0*+pe?Zze6DHj{&;Ot9WBu?N05cX-dnT$uO^cBU6xwnM?ON3y*oqMdM{1(u$D4M=-j0f;BQaCgEFQrHOAaBUao5mkm!3Jc z6v5uagmR+c{0TF^iH!J}f(n`CaYJ}wteIX2g|Xq#D}%HsHj6cSYl~uCRh53T9&aic zCs-CSSJXd1CFc<=kCaXoHWmg>_vj9U58d56_CU3{iy129WHPjF0APn`*x}8i96Svo5Eve0V?eT+B)>ZpBbqW6O(-6I_SIu_SbJ4+#aGpk=fD zf&^;@gQ@rXZQL+ZfdtM)6G9E^j*#XcA7(3<;zH?{W?B9jr}xDPsMo8lrJp%<(N>yt2k6_-!DnJd>gCgt-$5=jSd2D<^ZjF71uMZ#bz z{cPj*4mltRH21fU!rfZJCu|U(oY`I)6)YYx1yH! zAh1WNeD=d@(gx+Y6R&V9eRRFz&RzH1I|nl^WNmMA6AG zpQwu3aOkbUA}RgUYeWlaIvq@t476>MNrsWbBUj2r<9 zo>ufczEnla!O_~I_Vwu=%$9aqZ`Wgt9NM)?&ptx1ULfH zhEwumi&FSp)LNxEe$7khyzIkCDHO@hI_J_8aqp4?i1`6dr3MK0-r>n=n1y2nuP2Q^ z<*^U6KR_K>0AF(4imkG#GWDI6Y*N^;IPF|a-;LV+WAqJg%7J~zxSs3fBxzI4(WiZ( z3<6Y91W&grAGj+iNo{R?vSG^}nSD6nMMKFD8H6wR;fI->JSjR750u_@^E}cUJ3LA3VvG96WfFm1mwEkrKqBR>C6Y&Kh;w9QK_Oo} zh)Rd*f7paEa&3qrlEHw5VKwJI?)2kMWO}r|`Wds;)sQa6V})UGJ;KkoPf|N)Dm~}2 zn|N29&N-K5(L)qq7;9!Fdpte5`Ju+Z^W8_SU4e<0ms@bDAbv(($C@^PimeK!ZtkIX zDp#1?-|?;6+4kgq@fwHtmp7;)o)rT2?;jZWhDcNe-tY|6gu#CM(v#!}?!QuIl<{Q> z{5;~t_YXij)cFf`Of?O1qtilSClG))cE2$Qhjwf<<6EKsy8md#-mcG7@<_b9!K`#E}w?HPxPe z2l9Xx(%r(OtP1*YtNi5^Iy@~Btjz|i;_`y0?rjsFZw{(dMRnvr2UYTB`mmh-HRY_XMt!D7cI zv-3NnaaKW67j`wiQLruZbFwwm!Q3AWz8UiFP$W;{`o#f24a4ql-<#M_6&x*@s!Ek27 zsY`&3V2<_i$rK*>de6IS*Dn_VN(u3_b1`A8Y-H1PG0iE`Llg#cLI7o5^;&#?liiUx zb|}S@lnAAQ)N-96M53M=WdPkU_9{xTArihaB(rY|CMXExnMZGtm*!M{^*s2g6qGC1 z&}cJI>mZN=V5FLv8tNG{0@&U;l3hPtEU#=5FC%?m5fo?DpqV~!(Rang^tb7c8Ylt= zW$eyQQ!p$5(a-DJfa7@Q!I1PDwOH*IkaKkh*dn-D5k$;>?Lb#>TNgNz=x$=2>yRK* zI#(mE>QeK-MyJR@kK%n%YmFO?e9mwNmb6cC7$6nUy&P;SvM6%9RM}yP-Mu++`lhK&YOvOT9qatWCjK zG(dL9u_ekG$167dy8Ecuymt8PZK(%$$vCzL?itW^fzjy?gVMBs!^GZ0RycncjpOyk z%7ARYP#stt^0i6%E<30ibD`pL53uT~YeB#?3(he;^FUUwQOy(jf;eEVq z+qRAVc1F|$08r)1s1noSAyg+~MjoQWTmThhFO>`pU5)~)gn#pat{~U{GEy?p)%6ZT zpy;EX8dSk>On~FUV@Bp1m83Y~5K#iM_&JDY)#j5fX`VXZNn>QPZStB22L~m4k(FT~ zTpo{u5(iT56=d`CWkHn-DK0dnOZh(mbsCUl{xUsXnzk~g5i6rX3-i-k84{(_;j1A) z145ende>vhQWT{Q1a+(oa=F1q1tcsMKL;h5O5uxMAoJdrgK$H{ynp{*hdZ8bTjqaF z4@fDHR+@+yz@(eQz->?5fdUc1DeJ6IEbjAN<6<%hk{toXlTYzCpCL0Vp$v>du3QNp z)(DpdwG8}i*>kK86xw*-(qtU37}p3OlU-U91aw6yxMl2v^&f$pH3Ndf#>QpP22Kqlhy)q9aAu03s+TBE8I5Lsbxwjs;PWUIL*6b=6S>G%)l~R73&+ zDFGn_a2Uae1T`dpkOU3I5FipwkPwplo;dfP-#&Zq&W^kH?%v;dFt5*f&-=dL`uUb~ zt{-tYr1kekP$B|R>xBaR5PZ!sG zyY$Y;fz7#*dk$-_(NNH;x%Fzv^+l~`G)^7evOi`2zU2Mi9X+~v+qQ4F-r^Wvjj?ZL zF`rY1;#NO7@xzVyvka0!B9+hTDK&BKmNdPLA0HYoTO_CJ3;t}jMy0YzRrqA}#Dr3O zM})*pzIlGJPZ{Td*>U>Xlm_{Y3fuvng8YBxxh+CyI6vjID+XQJS)ES^iQdQi3ZBwl z;(a2br^TsfwK8v>fHJ>XJu!;HOYaF+gUO_Ii}Wu16sU|BHF#8Ii$pklN%;nNN6YmW zmG}NSO2#`*5xsX!;YcpT05Z(y=44w27$E4!ubSm!@e*%%Oggei4<4<{(ELf(ezmju;_-;a+5|Yx|kM2UQFK?P>*fM9``G@z$ ze^M!LiQ#x~j?;FPZsg5}V90R|!c}04>_5Fzg>%HGth(hDShrI}Jz@CP#m>;s zNQnwwDk>$STBJpIDPFkZKc5eaYMm%cCb-%~?CKvtOKs@Wc5Jq4C!Ws=mP^(_3v60V zb`kyxh^L7U&;z6)`5lj5xK4#Mc+b4-s20Kkj4|Hnwit~6y+D7{#{*y((D2Klm>>zg zc*2RNTFSlJG~b@p#IIukF+5BK`ddNl>q8}m_oAP0MJPNTA7~Da@ZAA4yk_#5H@()e z9@xNg^<%2K9#!Dz6Fy+?B_6u?Xn)CIGwQV!1hq8^{N;N8yY>GC%tzmd^NFU1;pMn? z1}c^ULfftb=CWTz2}dg@x(F!v&N^V8n(<@dbmB^&iuqY}B5c0afz-dzO#dL6|J>z5 z!gA=I@rmO>AubC;4V~4(Ozmr3`1WEUjJ3n*z>3Z9(T@;S@H?mKy?0Of)^j(431g*2 zdTpH#*{%uGHoy{DkG!JBAO5!F?W!;1?f+akd|{~iUq$>eHX7KkZ4cR;wLy$72GIC2 zN~jk??kD%bz^ShesD1b{wMm>3{GjzbnR8Niy)e*4m?n3J4_g2#AEU4A!>{+^eWyO< z>mpQV4FbTnyaeF;hzn`Z#lt`m26|`D8aN#P?RfsbN=Uyq82w))Ud&hwB?+VW4+A#f z3tA0R2@Q5`DK1@dut++z6&T9+*$VvViKw7fA@)4qY7GeFNuEGbp_S3~_+al~@~PK? zPM76f3>{!6YAFVPCD8v%$`%M`^MEtP#IicBssMNq!i!GZ$NL8Mi2RiomDToPsuCmU zbT$!=sZiV5Z)4u9iWk|qSkV?cs;Rh;V@u@#3Q;e&t@Iy$6-*b|E#@y&%0oIt?XpW> z2X6(He)I1nv@ety{!1Fnza#JRPy@v~wY(P_C%!Kd-k=G+=y^W8CWaZviLF3pmDIiK z_*9Tfg7_3~3vF8E7}KZTTLH5E8lFn2p}Be$h7Efu!U|RD+4%eF2YQ1Va4htr*uZ8} zwPW(wcHvui-`ZWhQzuDqp53nt@9aPJERkL+a<9sco=@a~SQc6jD}(wx(I+aZRe0k( z{1C50BT(*fU#fih%}C|hHsr83wX=tQM9&iw)jSxj2%@kSm~VToTA5ntMdx`9Y6DiD zuK)trQb6*I3M&Ovu|aJC68wI1@TB!m3sd;yL4qqFG2$)O5T^7G7g)bh`1jUl|I*7r zaj8GrQO@C~y0kLUo!2|WcAu-`3p+1nnar0|PR#DS&`_{zD%2%vPaQB>pr8BS#^eyE z2zWh9I{Han1FK5RkCSTzdRUhK`1cjJ|I)(y|HS`K6}A5y$xnDF0ehFS-_*dFUu=}5 zmf_)0OLS>qp+U9-8Z1fnpuYmO-Bela(td(Seh^Z}+U@rMPd^e>hi|Ri4nVkQ3;2P~ z-e=>*b39O2$H|H&6y1~|-}V1sPk}332y#=?O)KGv0%SS{|!};z7MASud!7 z+G63HDk;s+M9oT1Bs5E|oXXTb?ZPyxz#mvhZ~>)Z+iSQdm2^nZSnfjdAh^cYjUEwJ z@nIMUmHvb2|97rT|3O99XLB_UQ#2Y9Oh@?z*I0&YeI@^WllvZ6tvAS#r3MAg9!PPmf@W&*3K zn(77t>_54K2g$?mu19A;6#qNAnqS6);jH5IqnX#Ac!msT0xJLQ6;P{z%a^)@r5F`P zlD?TfDQ|c}ZJaeRM*X8qe2@2eP>-QPTz6cP*JF@x3wXBv#Id06PwU1i1Ih4uXe_8B z7IuCpDILy6!1aamq5qTF#vfEV+zE%Eya4}le1~}ALqkVZJ-}O&W5O*}%NF{Qz539L zq4-yDPMdUawW~|=0eT6JVSAt1;$wKz#Z}~?mm^B|qc2x-uY!{H${)MBbLyA{su5(F zR+bN5vvOxU)S2JKDD-=7QLlr~f`LvPdUEH-!5Im^4G8OPV z1pU<)1br0`Dt($op)?3V106~mZGY)Ma$#F>Z}rJ@OzZ4dyHC8U zjNNblcfwM-P?6_$Yptl_*pF8O1 zw3mYV3T{bif9_?t+k8%zUpRgJxp$oD1X~6K)n#0gzcMH)S@#)x{`wXHOQUB2O|OJWFzO4gJr89g1j z!Nr>8Qp0wyt{Ul{m9&RP&`gZYtJ6lEIwY(pQY)b(RStGVmaIoJqSTUJ<6-@2U`eRXu*nR>2@)@~OpqU6m6 z!dZrq)GS=O@4QO%?#)QcW&6@$yWAXyg62l@mtG4@U*^CtDIjtSz=&DmmmJ{5z*5op3KRQZ?yB5aarJy?8UDY}HGl%O&f1V** zKev%N&FP$$@^Lq1t*9Dw?VrP_F(cF!ocwCk2rB%hj68U2=BF?jscb=(8T9ngE8B@| zhcWW{?$IFQWdTpv-P!(wJnd(T`LFC}t7*T+gEc(nGYP%gQvA zGe%6b=YB+Q>?%*B-*dK++8~3g46V+Y_H16BCCH*t?&5Mozz(-^C0A<12l>jm)ZK^e zcj{>LJfO&Q4NuPQ{tKPB+)lD6|bR=dbax+1ooZS2owWkwa=%_ohZ^#}*;j!KPi- z8h!W9YEqN|*@v|U+|x@-#+>c_H{N4h*oiZ5V8oXpl#FoxMc7R?f+_b16yPW}(xhhx zPFEyr>2#VJJzp|%CnLU+tE&|nr9J0B*x2)|IX&lfmk2w=ip=1YY2?}JDVL6<-mg{Z ztbIl4qjjBVRc#$gZm{Tlt{5wE{_1pmnk9@YNztax{_m+;oY;f#yMD;}viI#MPgZnB zU)>{fN{?3T>? z_ABNlqhN_-VU{g>dH(}gu*0LixiVvS#5fKm4W_vzU9YR0H%$?LtEi+E`{5^_(_U1Q z@%+GabB)slYl{f=7X}1Z5qgbpZ|F#;$g`~ZQu}d*>^*&+5#27xXrIu>OAAv^yC>c5 ze|WOJ3+uV=gRcHnQ+v4h21xwzzYTLRe08*iWI0KTxND$vJVX4_=^5|%@#8JMk?WI? zb#---?B3p9UI4)-#B^obz3nByzD}#+g7a~>go6a(4&>SNdl@yYG;VVcyCtDP$-u}u@20@GZg9%d zkYW@-aP&EG;!(uHd4<>h86PH7Rn01M3JRoLcOJ1ML(eO5ZZw2VLh)a%IkmW=j24-6 zG8czixPtJR7NQGTO;J2Q0|SG;UU#Faii)8Dt6M>5bd!*~CL6<|qebHZ1omM;GIA#} z97zxob#!#tLyZyRwpVfLMT!SwRm5+VTVqTq7ZICWj~~~ZKBJVBz6&244OZsV*48%I z+uKj^h3E!qPmip3XlO`nbf9;Mh- zwExEcWHiaqRzPv*l{W+pMh`AlGJjzf`R;`f$jV0W5dp}`DJ--yo-ADI>zN{k<{j(k3nGtDF!WjF;F=H<;luI7pOY?(DyNfTMsM+)$I ztl)koqqCS4J@tBsCuBu4qxw`m&O!Nx2Y2;ElbEgEHUl(v$hC0gbHD6g*G!|)^bTX) zc|pjDYMwg!U9qeWBd&985&WE$UseriK6yCa5j0p&xWH@@j&Uvso=9JHW0Gn+04vJM z$}%tQQDU5El1s4Cwl-6YD`1l2X0IX+S$;Po1tJJprK8hUmi^PE1_n9-o8i<#WQfx+ z5>A&JIrDU8W@e>{^JxXU8z*~G;5}6NlUJRd`7Ml=I_l_KuS}VM^(4KZJ(@|hG*k%P zyHR8wH_JrHj$uE{`?SOz-lgC$037dXaYnccw%46+N;C zK(ty}Sz*}d&PC^vk3J^?V;e-1S(6fpq^Htqi!~|5w;;g_Gk1g_ycFCpFGc9t?oJ;$ zU6O&Kc6T$Q#n%-asQZBh*bL+&a1P@y!uERD0dYP5d<3O2tTR=$PqWeY?Ab9z-p1uc zL02vMre6~S0R~)jdZwqS{U_^k#fSoCt-qrd#a8UY^0sN574xcOB{?}cr61Ac*4rah z40kKqRJ1B>(N?2x{NHPu^!qrd`sbw}oor%9R;4c0*QP!ldKB1KS=Ol*H^~NzHUT4B zAE6sszj4Ad?%j*xhD@i#(}{`F;p_EFjcr#26)Ze+ZK>ftH6}8AJb__~0K2#8g$Z+A zKGL;-4H1H$Kri{b@IA+?Rh6H?duKznCLlL>#!eRjz+%Tw9qWiqNn>?|{U~wJg#n=FFU~>SG>h|H@ z+J4;zmjVfx+>t9pM!BPbUK=v!Cxe)@rY><*AfXV4V-8`&%!GN-!aKVZAlQQzMu3%; zij43#zx?tG#+-Q*3NpIXSf75T{rj}7_puiq5FMT&)uB$qOR=JMWKDB(^Vy3YHjP!S z&9m$7F3QJV-xf1IHkNp6WG!+hX{KXSLq#>kV7?1gK!GKY%AT z0?rF38$D!ZR#u$W>-D~i6oU$Id3hv4hM?G;m**{hl~sWH*yKfuw%3?zL=+Yj7Rn@R z!WXYPmb+Tt&I~mjAEZ6X){311Nx|~;GhtyfrRkDH05`!v%RZLTg%X4y$>}P~h}E8jTMbO@Sv*zM)F%`8I-y_G?O3vkZ?A zvTT(!rp@gfmO9)GG`QRpM~z5tJ_RT&e)8lA2`kyJJ$mDZE0dR%Ec=n+!B~3$csSwB zrTW~fhCMLQ0x)j5bw+a`&|2+-aroUEv)ZVOwcYw@=M`e^00QD>r@e9u3)x2l$tIM- zND{L>IXStoikOH29v?8e^S=t-f74uO{Wdt!NK&xDYH7tfV3oCEkIFvR7TIwg*hA^P zQjYX2Ncq7Py4?nAbsa8!7`^9B!41TK!SK8x#g*!tN}N3ZNhn{}!RW%OSW?9p>=<>} z;^RoLv8`&2i=oaW8K-LkJc7jHPv;A_nc4L3=iHsFoB^7_QnIpwg&fNvO)CQQKr0YWU@V7jkT7n|W0ju<})lxLmp)j9?2yqm|x$OwD z1&rtL__G+SId@NOx2Rzt>x?aCG6&H*B$FNrfBP2xY!52>IW>bg3t(~UQvPLRq824Y zYao~Y_M@O;ecd4&R(If}OoUMHTAK`jLhT(N{Az>OR5Z|qD6lEt=A_bX<*6NNxT|eTl_E=uJgy{S{ zTA!jN?&4QeOKlfMS1z9fH8bx%kd&{=M~84yLSQ^U`H)5HSauYDQqRiDx7kreD7mjN zxawIwc0{WdZOs39cZ7r)V`%q(ptDc*I&(uSjOM1Je%eCY$A*=IG-AE=R%LIpcC1WIpXKG zBO{$|YYcihhZ~oH#umrd!vFp1nwZ`XU46T#Gc{yQia$1fFVZM2iv-N?Ae;$rSbo>| z@@(bZ0$F!@f8zE)A3fBOdYyF1N!c7Ptrj#dZr%hIniQsgW7BPB;RtDAod5K?o}(xe zpy!J!3)OPSxVuzYKEVa_7o+)scnIl+4EBg7YEHK|!?s9h7 z0~PBZDk0{wVkM@JTGJnwOkP3m1U~3vIcB)mPng?SXculwNkB@CDUJna>$evCJaTto zmsRG-cd^aCGV@2Qv|_hWLNo_LFZIdEf-U}k@^YZbmM1Q+;nolR_ync{dB9fZ diff --git a/docs/reference/model-method-optimize-2.png b/docs/reference/model-method-optimize-2.png index 093449f41694c87ed4f6a948f6425a8e02c8a550..4ed9fe5c527f0931750caf802c560ea60a2989a0 100644 GIT binary patch literal 16009 zcmeHu2~<<(y7oqjv<}c#1(B&%Dk=ghqRhc#72`k@1!Sg`83bfbfCQzLDgrVn%3M(p z0z~F91d4!w5CN0O7?dGEAYn=Zge3Rdc>1q@{r8?*?>#;D{%5U|71j>fdw=`e@B2Q_ z`@HYBbLXO|$<|G~HbD@y6?x`+GYFChfgtf8Ki>$x8F={X9N299@tn!`U<+O=Cu?zF z^M&sjD}M-5jDr8f;>PElA?O9V09x@qQKFma%k$;w-$$B!$Vzr?d4%x@R@JMEQQoBW_zaiZc)Cj-A5R>dnhN0X7ndoK>DKSaUqq``PlG4*r zh_xRIX-ulWBO#~PJv}4C!kp6Fe9F$wPTT%7K84L=&X2ZomZo~GHIfS%4fe#UjrsXC z_4P%Ii;FGMjv{-UfTM>%5SUCR^|B*jW8?`#%+mAx{D&RSTA|pU(bhQ6;;V0)o0|*k zMN3Od)SEPuqp$&>pZeSML&V;Zc(TM*>Ogjme9Pz$7jns@`g*0ee!9YUndn1(RNC;> zS+pEDdJwfiv%zrw99=ID$0xZKx`w@4iWC{KEQQ=qPIEPR92h^M7TBJ3r#Z(`H9H@% z_?(lFnoh{C7e(fe=g=^401j@3N9kdkUyxe%;r@&Z{t|p?zhCrf7ubHfZ~ji#{*BY0 zzt`}oqqAwgRlO%08s)ZZR9@M^eMTpvJG~F>eCzwp(@@;LT9((uU{ra26Wb9*2*(gJ z_>f5laQG|W@Ze;be&&!+cfpi>V{s)c`M9*??95C7=d;+?A+`?^&O3z8mSdJ*a5lYD zq!}yJwmF$4wms!+ar^ODD#hy<7(GQU<}~L`vysWP2F(~xnWnZ5xsMR*>gsxR&Ky;c z@-)h%t*oH0A4{eB>wMcWUQnU!Z8++W=FV6V>1=Cqgk=$!eP7>WFuR^;&#e6XaX%|d zOUoG9MxkNITvkz0kw>*{MaG6u&WBykr_oPq>gv2ez_P(4E%4@c4i5Bbes1qnVH~ly zp|;Pg)nN3- zPu9PSQiLJxQC`EjL2Qrj98w;$(a?h4cC7&IkQ?>ceJh0Z@g$ze_}EYxte z8$G2l0U!EreR*khE>5KgFpx$v%wg$iX>a|i3pkh|%%&N05*@eF#p!gp9Nqk6{@Zt? zJ}-;Z&Iu~bIC$mxK>EXoala|k&Thb{n&O1b`doJ+%d%bC#n!=rHa@<)m09$M>xD~t&^uli1pkJ2V0#F?F>_y!)zJZc_Fvgy3nH1 zNQ6T9h@`TGuA;Btyah?c|ACnJJHfTpMjARZr%BQ!(6NjcqV_c*KXSUF&@!xsQ;~0J zj^J@hQ>3JVXKI+`8Ss2EJAR;TiV&YyuGcDLULd$gu2oL-o}xIt!8GUE6yPh&IYeHR zw$4cAZr!DeQs4wl1>bx~=U|+xxr3f3@q-H;Tg$mcW}^@VDGGCY-W;cU5#e!-GR3i? z?eHfk^YlMPcKw}}`p2C9{0CTc7JA{QeFO{Ht4-0Cf)t*o^580dF&MX6Bb;}r3wc)g z1q6+43lWRRrd-Ek*@6>>X%^|R4xB=seV;lp)SOexuJ#fKM(om(fJP?MFXRtDsKaW2*oinfuJvic_3aA|1dU)Z6*CkPKix{5R+v2KHGly=X;1{W-(b4duN*fmD`FsH zB+ZG`YudW_UfatM zf|lC~Yw{6#Y#M^!;?CTE9J3hq1_N|t15tHugK<`K&GZu{jUa3+<0Q$0+k|!dAP&g* zpHjBsv8zbtj^q5MTyvb?Fa}fr4T6bJG~gd8|Nr8x{K<94?^nNFYh+BI4RCgke8a&Q zG7u0Usda}`UtDHpmu2h&hM2;hII+1s>RqDrjl-p-rFO#x3T2YgN@W=hYHt@8r4+tN z?re1iHAbQ;3P?k;vtBwiK<_#bnUBZ}?M$)_Il`r7wAfQ7mg;p3-{%#CXWPNl#xfOW zQz+&@_pMX7}uDVa#jn zqjN4ofn*V^k)NS%Lx5x8)!07; z-@WxsK0eImdFl*K?H(D~vN1Alt>28q)>eq%KcO+Ji#}w9FZZ{uv>%q=vat&n8`}zn zur$=mXw0G2Dt%&WCu4D)+5qyY+dPis)ilwi8*cjZjFeJ~raPm55TyDbybQ-LhfkH~ zlSH!Cgpn`L?@dfOzZaIVrF{i1rB<+vbvv3LB`wWQkq8?9y|(5gY=RnY+IZy)N8Bl- zRSxx>cSWB?VQoHXg#x2a8wgqhAZRV|=7DyV9iSyJ1yr_DoM*KyHK_J_jdQ7-g*iX& zE;Qm@VIHVgwCca^x7uWLlBg-4+5BXqhA7sIL??7lqKD{S$I9&q@__wP&+i4FJWKn^ z)(dg8uemubFc_oy?Ngw=DS<_J-dns1VE-U3t#&Bp3c4mlsR0i3tzfar_q9# ztxY5K9{jJr4l4mopy}IPUDuB8ikWVVlr-n8#V9ct3|KCMN&#*Y{hz*2V@E@+J%QO* z>bl+#z^+vRezvo<9h~N)fSwKa5^dN9$JHly$bRR!e<~RMN!`JJ-TR+bEjLj?N{djQ zsxVJx+B4-c-H*JZz8W|MK`C(78;Mx!C1g@?LPR1}%cxh@v}Tn;S_H>IkO>NuAQ9pV zs>uR9IdkG+A5ci5lFiw+7`_J7u=`8+?1p=i<|tEZLX}lz;li*T6wx%|91KVKANGS` zg=!{P2+rlm`{33z-U13UD6-Z-2lPHsSQwU^F(A+U8cwpO6q-kxQ9Vg(|*TIS?{Uj`amqBDhDSDa7cy2=*j0zs>`d+s;&>-g#Jg6R%vI3=*_E|)KC{0`zys-f~X5~*rN zeL475s!gFKI2u1r42N&*&-07j@bc6Y+>vH2eEH@)iHqEJ<|412i4156ZyQ*6C6TdY0wZh29)N)B9Pi*)DN{ zzBZ2%q^}N2poN1n(KU4jgPw1_wepi1Lec1z1!lL~QhWpdJw5W;4L?6Ka0NiWHmIA# zU7!j4MV*qk>H%B~!S!u>G^l#((I#!@rp2PmAFH>WUBOvOp2xt8w&nswl+sEv6-= zUKFrik)D8aTN35rk%dJ0pG&H)EKJmv`8Jt3`Fz#?8{ z{u66zk9Qa-ETK|kHL!TPJYznOYG+PArvhX7sLGt5@34XLTL%Fz4nW~N(Vij;=nURB z^5}~?g|tQK;cEW4qEJBkmH z0?PLW*2#Mb>HbR2!Cb+VO==846c z<69#@UEbT<3&bn$ZR7l~RyeZ$5Vmlb!-fPH`jibVSYf%I$sB3nnmNenAEUD> z_?Yr%!-5hwddwrUe z)Rop8TeTl;4ip|r&)X4w^vS+6M|RxVu{ZKB^WXk7V)N!pMfLBEH#n8Hu`-7IS<2_Q zPn?doO{9f)C59PkJsz6TW$?mK8U7h`|9UhRUDI=UjGaeK%)?&dKK^wqG>)?`M9gi! z1mHsGa=t}E_JnNukACDWsspzZ_PD4nVxmp8s$)nRdSZ~|w~UXi_I)1@hm4W?;dT%N zojv|yP=4xzb@Fj|IL}0eS8`tDe2bw;}%&@w|{I{efR8>Z{~kxi2U00NOFduzCa+D zwf#&o3w`aulRZ$~Uq4VG5{XtXvT`T)y*XxRXlV7q@kIAw#};!uWr@Ri zW`V!JoR2wut-r(Dvr+%pv4_RQ#X%ak#4i%pyWmw6E*` z4VC{D>;8uUtL53WZD?iYw(w{Q9CeOFvjC7)A!wh31pCAU%fw@)Yft&>!X4>Oew@?Y zV3h@evJ|%0N&d?6@EiQ$RPLj}-ji)jHgJH0S2|huZ=vJg-EvJdy!Zu4M!IOByA%COyqmR+ZEAxDSb!SlCa{O>0E z{|9*fYOa9Wzjlv3e(xdV%SyxMa0xzMI=z&;_*-AhlWb_a>AFQ!kki+Y-?h~ZpY6W1 z*IfCW68%yeA?V7EI_T0mGmyK@k+8j*$Aa#ukcKY>Y>Mc5ukgtnF?322_S`$0PFK?K zLT~-xqFeTxgO~HJB#6+jeeDfg3nuq9h<%(&Qp?=rGR@`~=NaD2>tWu#d*A6#KUia1 zO3Nrp=rpV|SVWI5>xGjks(b%lUXyBqi$9qs{?6gkF+;HnfTt57-tD%hPuF_CUQ3dP zU1vaG(D#ngAAP5O?8rk~QMOgJ(4_Qk?2%`*wMvW;Z^HF%tk{g?dKxf})=mZX8hazf z3eFsiHz*EA?QD^-6o8y?6|7cXN4N}o4FY4aEMxR!)8{Q zm70#pPO!G@M(#a+G{s=~`?y;?XdP0J&KXX?cbfWjliL&JcOtP1%qN#;thCzR3)-S5 zWB}3;IE;a+WQLScl_Gr6h>vLck8`>PiNASvsQSyQj)U8zZnk~ShQa7U{1fz*tHdSa4V>n?#5uLr3= ze%5?zXnLeH8J-7pXvfDr|49<|6EwH!mBZ2VKJfY0HJy1adL^@yHD z+iizn?;F~B4`-gMvT@_F2h+4H0@8Xe&Q-kI9o}|nG1~NSb*6!?@HA|Pkr6YdX)K}) z7&z2rUtd}I^u;4;6qPl28j?AOxmO5-*7_E3I} z`f>(yfmNU?uDF_VzJ*TDR;;gQ)>4BdT@|oFnSLfje}}}jWq#<>CK}Iek9*LvQA-i| zOMNEt8I1Fzy(^i!cXd9E@9L@=?E~LOwbDAoWi!kp2NOG;M`Jm}KQa#v*52IfB)Yvid$nG#gh3&8_} z;N@;uQ7GY)-F-&>Hv8&;NrJjKNMhF6y4Wk$q9uDxEf@ZoBe@mu>xG3Uzt`3^NV$SX0=Q^)zg~_nF3nu@ zSsB#L>hsF!-X}bSKT^O>&qkswdxjeIjGQICB5P`Da?->%GDrP`Eu#qR z_r*>RIYjzi{_A(hmR+NyQ*Z8ahA7(g**&qkS=*VRDpQ(`t}U2oYOsDM^PIg-?c{cT z5gzYJvKPgi&eqyIeDmhbEz;7Jn2<>~Aggr8?w?#(>P5K*1e8tc9Lio@T|MI3m8Q}* zGU6D-8np8L_K_8kOHoP~w~Ffn598ydghR*OBJKFb$RkIOCe031%BFqpGWH=+&-^~TbngreQ1uh(KSf=%QLh6YfD=${#47VQNvplo=r$YPdv7& z@Um*W<2dnHT<5A_xfRjnYP9%RrA>Z#?smb|;wK~DA6h&>wtwGg(9k$h+^W>!oNV_{ z1>LM>AJ41j5hYY_J+37A8(Uh^$_Vb%g;H;?zLM{FbQ~_h$4DheJ$qC=+c&K*y4W|e$xBXP;Vg}(t}hT zuUqUom!J^%_EMpeQiN2{OmQn1PBkOs$3}PP5mFP=9lduV!m@thfkdvD2Tj&+iWT>nR&Lf&r3!TUc3a|G{*Li&4EwQ{X?Ptxw#pb1ILE~F0<QGV1LM;~oBE z@g@~ha?l~W*}-5@qcrw@T=>j2tKuI|bs6#oWxm^(Wo~UK3D(RCY$`P>hk|Ud+|tZv z1cD}fi^Hqb$_XA$#2c8~GBR?>{a9*e9_s~qaUyK;q;flp74+K|s$&Zsny6r&J?Evm z;^qjh=Xel57a?%hs9*jt6#GC$SvVI}>eQAXXy0tiEc2&_J34+NYwLRg#F;<0&y}&Z zNT$=Ppt6g2Mg*XQa~jnusmHGO&(6x)hy2i`GkVXR@pI%*l7tUX`1@pnFPcjH#@hichKY~w@JMwa~8e3vmq1tv{I8R;SyvXupHahpBHuR>)GYLx@>$GpxYqy_c6*Al~N zQESR;Ba*%Cn~XinJ?R2IU4-?nnaZ&ZdH=I6=j~-4VjrYV1Jfv~Wt}}Dl$QquB)L{F z=uhrU(e~WiBEcPcD&bSw_Zo#=NK|QKuXV9!xJGciC5-(VrIg~g`=j?bL3}`3rU?o+ zUVe3Xm|%gQJ$D}J&$8pQ$Oen&HEL%{?nHZTc!TagEV)hq+pFH3PR#f!l+T?-CV<8r zC_-nsMzzB;afaM?Nt@l*?*(5u0OA>k7@w^i^gEX@R!#_?t$+Dh*4F-F_ORPg*fUF9 z?nKWy%?r=W+%H^(tn(H98;ZVNme3420Rl5A)w^atcoYleR9Vj{%(KX?W2r(~PapU)f2D~H#HkMcGzdDtRS|eKEou_)9p_jT%{F}BOhtavo{!tEc^RQji zptQ8KtCN#crEEQVZN4>81x7%Q?HRS!4OxQz;vw&2sa8ZxU=fkquhRC|HRJIPHH#VU z*&1kjF1S7!CtSRD@1B!jm7}}(S_?N?aXLF*Wo03mlt#dE@W_NzpZdNPwAiM%(_}HS zU@Nh$sPabp{@k(K^4_mb)7<5~=e2A)yVG7A_L-RFS4S?|=F|AS`?; zQ#)NiHYAVrdlEXoP))nfEUMLDhcWaY)Iy5S75OaSP|yP*&AZYFVycgq<)CH{j>5&f zD}JvJx2mj&ay^%?*G@!&hZxxY_Kt%|BTs9hSE75}Rep3^bv_-hADWr-`QR0HZ;C1_+t+33a22>_^LwvxCL% z(x!zjeI?Gw1Rv~TDjyb72P^lade z3==RN$e{6)Fuw;wE3`?L_o+J!T29+e9eTk&(8&;bzx6g)dUJARE(%dMF!1o8RYJ68 z{qk{PF0s$cG|hoIjVE-)n<(kd7DM|i)Fxm=)d^;OP$dlfyoObY3Qowt4L@0eX5uWc zQt#5El(9O>t8b$G{+gG3?0g}v4L~9cO>h@&b8~ZxEqDw3N-Ak3b&bUqdT-7YQ?T7j zeATa^8QRyeutLTe9Ul(4#;Rsyz=>*H6NffM2$l}o*M(I6JX5sjDhLWTNC(p~1=C^J z-wZ}EECYn6BD%WU$zo8_IdHLSvATO{cENTQ1pD+rCUxABv}!J+VfzTC<&I=c^6n^w z9xmfE$r3}Qk9{m-SHIuXK2g>DaohAiEr+tQX9laa%B|JG8#r1`5A=)z%2=Qe2Q~QM zK_kxu=o~ZO&p<0zi*-43^$KvLJzzH((0O%!1?o4r%yh8@a-V3=p4=1IaUVQ1{vS4U zv?kyL1Wq?vh>uf2W?-iC-g=aS_h2ZSSQ~ymyZaT0({)e1IVs(lrt1$A>e?bg~o^y*Mk$!TZv)OS*DZQrp>h!gVj)&@Kkl!E0$R8KBkbQ_8>L7+$xY8hLbRW3Y8IW6V z*K#s$C3ywtVqsk-J(wO|&93J!0p0|3r)Ng62my3K*K8#{!?UP*FUpW}Wmn8;(Dq#HBV&&tOKyy|ge_$& zhYyctge|nC`Tq7rU)2o|YZH%Pi-IPsoVXBQUfpl8!N%rf#aJ+oP0az;o1wyjn-az0{&u3zD)^rVnkn%z%1|5W*COZowA*dQzXete$d+@H57aDO8gEI!gR)p%=fF(XI!J5h zO!cOos3t8VnBB1fWBqcwR8|cR&^os-aes=sq{1zIR%B+pzhg=tp}DVu&X_W z^cxDKIdfuV3=6F>E}qFP}ZrAl+KBlyx_qbCJF0QQd6r+ zA1b_Zz^W#|9hP^JN+FIUKOoAz1}fC!5W)~8Fk!FZI$gXaAwwr&Cj|AM0uC{5ot~W? zVzkcad2kygmCqMiq%PEh2dR-jGCmBKP9D1{sq{(}vC*{N=i;9&N%C`szw)Faa`AG zZoelF?FF|bVa!YIGf@r?IOaY0XuKoY6$tO^RxLxpbi+$5;Z&y39Tqo3O4rv0F9QNG z)?e=mOD4hD1(^V%v+{b!G7Tr7sUUGUTv7e%G!pgQJOFd0Iv^euc>GOoWB1sSjrcPv zqYO$okOZnk`sAdW6k=@-aK_pINibmK95(rDRDud2q*<+pLMb0;jfsh=h8J(H=Yj}v zY*EB6zel#j`KGfZy7n&S94GEKQMmDa!WfW(=}XvT2P_&!4ycsUZnVhpljf^NW6veP z!V}2#T1DJ=`-48<^Rf&urx|s)Gp5Cg@d+D!$=fWeaUT5ku){R(4N%}fZ+apVVAak> zlS3Eq4uS(8VQf?3)h3BpZY;2^K7w2T>~t6utwXu@;A$yhesc|xX~%8{As4SACnwh? zE^Tlh*pjg=bNTxTa%QH=3PH50$ruH(t`0Nm>#A#UhR5Kd1Qf!VD(**jsVj=c5>Z`> z!GrODp~8SJ?2NWwdU6F%0q=vhRfjAc!UOXu9!9)svNNDfe%d^TfD4BUs_*19e~Hui z@C&F}U~z0tz?gg2&Q=VDjK*w}Mf0anxD=1_6j*wwhi8CQHFv!vkO;)z!b(oGLS_JP znt^(r4cC?MQT|jb*Ks>G#&V)FwbEgfexp@!1@_S(nvo9hY*pff90uhR11=drP1Os) zR~#39FjKgcdl^qD!EmP2{Mz@cV}ii)2fQi)7D&`{-QX?s2`hvDX^Q)^hY%okc!*B! TDKcAs_5^v#^!uW}INtg<4VB|G literal 12652 zcmeHt2Ut^CyY5z0MnM_J89_mr5=2y_NRh5%!GM6m4818$X+c`(MEw;V1ayZYCCM-t zG7cgN0!9eIf=sAlfPfG{2~8w{P(lbKXYF{}bDs16^WSsl{`a1H+|Q#cthT@PeeYM- zzJAKqa^r^W8z2bUh(2-D9)i}!K#XLNrc`+1u1C+!>b$$x2{|T$ zf!W8mad$ZVxwc$Ra*`m)Jj^H{iig3r35jE3(i$mQf8u|*)yyNy>=IN)t!SBmnmw}Y zCf!`6ki=mBYNBpr5hI6m)IY3HGLw$rWGl9tLAdYq+8xX#r!oeOjFWF0F*%Dn!RR8i z#uMkmc`p~6l!Wq9q75hQD;wEThSJIySOtb#kx~CVn4*>L`LLIp#2-!T*bDlHVm^;q zm0qZ>x{c;tff?K!111ms`k1jDdEp_Ai;775b7sAil2cYgYmvWqt4*&)+RQJI=@4YK zz7}a9@)}R1!P&3YLfTKuKd}zM!6(@K!&Rm;a1MfcjFa`sh{ieX8N>@-TZK6+V!k+f zP98|)B~bby6Vut4Y{Tf;_+|~4lSQoHD!Oeq!%JOi0&I5!%{Pcy>fl_DC~^v3$inP0 z6j+vlrQWE3f2i^!C&6uX^1LmXTfV#8HohaJ*W|=Pcaf7yDAppD^8?HoBnnewr}H*e zMQ#Q3C8%Fi95T0IRA3kIaskJ+dV?{^bC=>xZsV1)uw^@{Ib_8GHlpgp$$-iIvS4ye zn19ytkmd2Ro)1A7wRr6Dvd0#xw3`L&ZM1hQr7Xt0Tago(TTcs_taJcBz_QE#6YQa5 zkt_t=nCq+*itx)X#&-yO%*Ip)W0K+`<~9Kw#D7)`)n7L#n~h**(zgMBb8o9qh)TAI z1FIgG0Z)KKsV3_JWI5H+`$P<g?^c9W@h9SA=RN~2M-~N8snDu7)){`w-P|IM;m^XT848Xgkd4s> z@0>nWn%%b}1~VlGeAOYXe+-E4c8#xs>M!Sx`Q9=Wb|{XLcGCRt^O?;>hJY-U06E6_O6kzb6iuo72!Cff&u+s#?M8D&%Yc0DnD^$X&mvojrjm{4UWy6{x;C7S+Jl0 zfkO#m0_F|+!ho(JOYdz3!elxb%xoCdei5R!&y1?uaOAD_hd ze_o9Ln<)6_<%nk2Kc8Q;o^3c#J)2*IMZ_02x5o3}Xxl(j!~D0&x;6~GIKo4GNXWb} z#R*~|A@kQ4)_rt5ntzZ@yQb}gTJP5#OvZEzdpk@v9k7cvogM!QDp)gHaFUVXk$6U& z)!~M8&sUBWCo2J}p!7@ia;o(SH%Tp-T4@6pc3)7$DKPZZCXGn8{0U@oyRZg@prR2> zv>dV|o{X-VtH--x;WBWn)j0gSCLxC}?AU7{TB?Sz(Fnn~qVJ}kZFbh=ii3+rvbn2= zVq8!5iSllvm>`;g!d_G@P9gCN1f@uGg-`iE!i!I}kH22`3Dl1R97=E=HA0uiy_1{0 z^3=h=Rh^PuWC=o^BcP@<^x}50ua!Bq;&+pxYLIH32Kgy;@H7aNK=KOWr%6fvhjt_j zd-Xjs`>0Xn$^C>KAcCq6-6ax92cFqS(U1GWQ?b^c}Kz~5H} z`{+S`ZWbp9F6fLXlAajQjiOuM7HfDV%~w+}H$b6o`T^o5>#=@V30plv95{YcG@Fa^ z+HSTQ8ajh$8%~!0d=OtS@cPK$!Wd_-TVjX~RNsNH3nyD%aUwb!Nw?0MfdHdE#mAbC zK@tjR$n$?fVXhL?Mv>y>j(d1!Gm4>@Jk&m>wTT$i0;t%7R`S->;N)pfi=yCc3%iqa#j3n z%ZAHWxOtYHD3lt*chC0Mhf=0*YoGrOemn7K{h7Rz1lz(#YWRak8n}T!RB!wK{KaB} z+x!h0J%?`GTK??Wv#Rdb+?DwZTKQjkN<7Z<*b!y(!j^uc0D&nNt12)IpvQBKKyh&Q zT4`ikf6e}5P58#IGsQ?zq`rv#y+=WI9>HzUsVvx_H8(5IZBt$8@{dd-$a$M|k$lfX zVZ(V1VW~fUUs-0Y&h<-*-bZ&Mj?_EN{h=B)%!8>BCt7}5yafuY`VxWsgX1ro-pN2q z4gc)vk7r(zGRFI;#z+7A|87TT3=_%tao^a7k1Pz{)jDtdGI!sj@NuqTmX-lQ##>2x zaGYgl9!otok!jR+r}Ab7N=Pk^faGPQ?3%t-stju?>JfgaHmS>3Az+Nx)=!_3ay{d= zlJ%~X$C^reZPX#9{h{T^eJZ5=isELMcGv5Tga_&*VY>8JksF2*3K3%m9dGj41viES z8Xv=82T1)EEM(~Ic0^`>^Y#BiU|{WpbZW+~#;x!b+wzsGOLfH+`vSfZ=3wb$)1Jsp zL;W8`yfohPefcT;%)?sp#6DW&`?KrK9Qk#q4W3G8HXtg7Gj}u0Hl5@bJIqBdo(pe$ zTkHU>4E?WDaP(E@j!R+(sVE8p(s%OOG84jk1_#V^a-!kh;)pq29b%aZPmgdmR+*)aRVPoV%#ozX z9WtH&E+7B!n~<0>McCx(QvW-=f!9s<@WTN4H9_avlZ#b9vS{q?#z}bQ8_gJ1BF6Y> zQz=GGnWYbqFazJHN53p6`X+HbqIJ*%mH()Ze@#K5`iW5!#@a1tICmr6N!ga!H*=+U zOIY2&K4Alwa^D551^S%^1B>qBso^oY=Ba2DZ# z8j}oX0`Y+-Y#=`7%tMYQgJ(b>=j2)bWVM#(Vn1ED>Vi51ZI*ruu<7?`;&1;==%9b*MiV+j2B~E{3Twbmw6>Up4i4AajZH`*Rx# z4&R-juF!5sSbeUGYl#zXSAt~-V&TuWu1)^&t2bgKLWJtHOOK#XkRXny7FgsJsibV=qi z8}9cE;)WXAX({Yz&md>Xy6JmijIFiwg}2HVGX?zM!*yb zCt||){=`9R4Ye)$v7v!tFIbe<#%z8lbSJ>i)r7~HE^&K$Z{;yiUgwOhw8*}0Hbgkj z4k5v{qVN&_gX?SqoIsy2HOq2$gE5_Xm%+{>Plcv6k91jyi3;R53=xJ><~Uk4JBzwu3&t;M}L03;|j1Vfj8-Xvz zRsI;X5~eYtm9fKQq1q;3mG@Mpz4({zrBoju9};rRKYGS~DQIn)iWXT~YhX$%ZPkoi zt%P=^j*}jug(=ZMulWTF!PHHMB7`~HPPVWk&`3%-2RtroFp^^i? z`z%D)th_*Zl~kHXMMp2(@1m^z*=rrU_d`E5qH%F?(T?R>wuhW>VY0iFS3C1kIYZzx z7-k+=8rg97?p?1Fe6lTNHaDcN%EQ3Oh+5SB<;RVS%k{TtXvdSb) zhL&0{>}y>*UnwUi=QUUv#7dRDrHYF{2vDSr9~(w*Lb+5wco62OPJpzXFRMr>#l`mv z^75WU1qTNwQ<&6PfB#WRF@+-7gFO@F7^&SATAi%N$btJ09yl=j#Gx8f-RFf4@2Cpt zBR+x^eLtQ)b~)J!w(D~Bq9uwJ+!R#J`;Fq}#`VdW7PkCOR0O)JrKQDG?^@YfT$b{j zPJ9hJ$qx29Q680)BwjDONo8^2y@SjHS3phQ)IuLztIiOEqtfhoGAk9o9Eb@jHQa@! zFk^WNTEAiJ%BNlx9+8|Ym9jl{S<-6Xl+oVSmNfcO8DpBnv@z6dX=@Xc0KW=Qa_C*w z8d@KCPEb8niK8Q$XL(6#?VDD}nTaK>>klPw3|vNdMF0IPRTMR{e{QgREia1YfjzG= z@&?toz#EjLslGW}GC$!vqadj4SA;Hiwd=&U6V0}eGy5nM3eP;TSG*!pP#2znBaI2_ zWyS}g5?#P+b`2IY9%;;by!5`rPHPi%uIJ~M5`X|aIg`c?!S2BpO0qR$8XFr)IjG8R zH+&~EM-^x}rl!I!c4V&?)vt8nBdP}K5=GAdvqKL)-{UoI%l4K1`OIZge^P$FNI=eH zaoIK16-#%ZIp)_dC1&vpCs9T->1hgWfFGr*E*DiF1wVBQqU}OkX<+wNE^i?%zDU?w zmVbM=F@xt^lDJCSc}1l{z-9LtF_Yh0Yve>xNJeAz&|JgM>AgS{?b-IBuJ73@kRImq zVae0WvdMcGFS#ka$s;Gd*AF+Bp0bJ!hdW*A5m&poC&Q0v_=#MnL^<6{h&LrWq=wn_9NQB#@1H+ zhRT(}i6;b$^mqaHq3dLdwFbSPat)F{L~oSP;ssT3?PJl~DRO3f8Fe(zi&#*QkU(Jy zWzk~K()sM@nkgRSO!jLZvf-b?8b@D^>Q+N*j9k=5^;{+}zxhY&S*R@%gA_2|tHkD}GTISW4G3F%cnSTT(4Vpm$ja z%*FzWF8R%g_P*VGsUu^jsbD-T$&H~OgW=g4lFn47!gsm?HDd|chd&^EOx1}G4dv3v znTC&hSJzLEk1bm0>(399Gp~bP-IkO#<5rCeW+XBSrqjP~)CXvyM=kMJM1sN7Rzeaw ze0gyuhZN;Kw0W=5Wx)$VE9s3-EV^7D0mJ>bb&%4%i|M^J(Nl3DrSlG|qMkOFUc}^)31(*tj>e+Jv0h3nb%2@bQNn%gP!hKyzarvqOR8p4m6S z0c%;Vk|0>I)3xkiO6?(*)%~aO@6tz|v?>z<0|Sp>7u=*4Q;Lx7oZDvLB<1!^$0{l+ zB!2qaR*lnF-H?=`wpg3mDp~8E4)S0y)FJub0Y)tHP_whM)>|KwGeu9B@$KL43Cr`7 z8vvx#;1hHien4N@Wb1q3e%K|hR7drH)u{nw88RXQ{Qb=zbn-oFqRqJXF0Fp$lxQxN zWC7-tcYepRo==&rhQ9Px#HHTn=_3R3)yl*`A2%2lI z&*-(*AfR$dB;LO2{ZLSG*|218ZmtyA#gfW8Az3SN;tFMdA3&%xOtRwxx^nv*bCuhI z{K$K`Zz?A%$(fRyoJu?%uZo7LR4TXyDzv3U%mb4{k6bdS!Sa$?`uqF6P`&O}v4;ip z!()%MpuCQ|52sA~x|9QmxfM%O_rfQ{Zj9LHhk4#EU8nuXGn>=1EnDK4XL1=!4F(q+ zyOECqUfoew&kZp?s$kbP?GHYA0vRp@<;~tfk*q=9c~eb3!T@~M^cN81;C8uZae~l0 z`&MGUP0jiK;LPK&+$paD&L~O&C8VR1cBLBnn&2twLPK<2F6SvqvX{|5l49Lo(OPAe z%9^0*Gj4H*+iP-FFwzV%!&Rcw?Byls44STj6CR6ZCNjKu4-?V+HdJzE!|N*K;E~uG zx>fj$k$?XC`xROYGCR?3m#~XFW|2LT8h{)fcPsLjdo-T>WWnwjo*h30outmF+P9VB zR^j}LxI5Nx-@_FXe1fTk`y*SFQ9qm8y*%Ap)$G?7(==NbetkrM&GG1)-3yNWBR^XQ zg03F+cxs#vFIS^5w&4$IS_@HSCVqqx+dh9wxLRG9rbZl&*=pU7qV8f0S~LeMs0n?s zORZ|zz*8h1&X@uc=t6Qj`O9q^Sbfzm#_GbFmeZNpv>FpyX`|RYwrP3Zta~CW%h7xq;<2q>G8@|m{0W5ha)R(=gLwIZh9BspE=vcE3^UPi-ZyJB6o%8mHWd;y@P zed<1_=hSOEw{EMYwGedaQD2FS9I99-sv08z|0tpy;2fuq6lT1YQSe31z|YPnapN ziuojf{3}^00O!21PxLPv8LW&^K1i??F$W%z-AO=G1;k@jW04d^6m|NMUn$Bda*Qol z4dE`HmO{CVs4G#ZVV0tkXm%pWFbW$vCOHqwprbjQ3)5tmeHRC_@qW`*6@(|t8F27; zo5g;HEUdyyO0d~#JQ6T zcksyOZWE45>_CbL6!Z^_yT(bmJm57;6Yp6 jhb<9ELGSrG diff --git a/docs/reference/model-method-optimize.html b/docs/reference/model-method-optimize.html index ff22f0e99..b4d1ad67a 100644 --- a/docs/reference/model-method-optimize.html +++ b/docs/reference/model-method-optimize.html @@ -1,81 +1,18 @@ - - - - - - - -Run Stan's optimization algorithms — model-method-optimize • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run Stan's optimization algorithms — model-method-optimize • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

Run Stan's optimization algorithms

- Source: R/model.R + Source: R/model.R
-

The $optimize() method of a CmdStanModel object runs +

The $optimize() method of a CmdStanModel object runs Stan's optimizer to obtain a posterior mode (penalized maximum likelihood) estimate.

Any argument left as NULL will default to the default value used by the installed version of CmdStan. See the -CmdStan User’s Guide +CmdStan User’s Guide for more details.

- 
optimize(
-  data = NULL,
-  seed = NULL,
-  refresh = NULL,
-  init = NULL,
-  save_latent_dynamics = FALSE,
-  output_dir = NULL,
-  output_basename = NULL,
-  sig_figs = NULL,
-  threads = NULL,
-  opencl_ids = NULL,
-  algorithm = NULL,
-  init_alpha = NULL,
-  iter = NULL,
-  tol_obj = NULL,
-  tol_rel_obj = NULL,
-  tol_grad = NULL,
-  tol_rel_grad = NULL,
-  tol_param = NULL,
-  history_size = NULL
-)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
data

(multiple options) The data to use for the variables specified in -the data block of the Stan program. One of the following:

    -

  • A named list of R objects with the names corresponding to variables +

    + 
    optimize(
+  data = NULL,
+  seed = NULL,
+  refresh = NULL,
+  init = NULL,
+  save_latent_dynamics = FALSE,
+  output_dir = NULL,
+  output_basename = NULL,
+  sig_figs = NULL,
+  threads = NULL,
+  opencl_ids = NULL,
+  algorithm = NULL,
+  init_alpha = NULL,
+  iter = NULL,
+  tol_obj = NULL,
+  tol_rel_obj = NULL,
+  tol_grad = NULL,
+  tol_rel_grad = NULL,
+  tol_param = NULL,
+  history_size = NULL
+)
    +
    + +
    +

    Arguments

    +
    data
    +

    (multiple options) The data to use for the variables specified in +the data block of the Stan program. One of the following:

    • A named list of R objects with the names corresponding to variables declared in the data block of the Stan program. Internally this list is then -written to JSON for CmdStan using write_stan_json(). See -write_stan_json() for details on the conversions performed on R objects +written to JSON for CmdStan using write_stan_json(). See +write_stan_json() for details on the conversions performed on R objects before they are passed to Stan.

    • A path to a data file compatible with CmdStan (JSON or R dump). See the appendices in the CmdStan guide for details on using these formats.

    • NULL or an empty list if the Stan program has no data block.

      • -
seed

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. + + + +

seed
+

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to @@ -240,25 +167,24 @@

Arg chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each -chain.

refresh

(non-negative integer) The number of iterations between +chain.

+ + +
refresh
+

(non-negative integer) The number of iterations between printed screen updates. If refresh = 0, only error messages will be -printed.

init

(multiple options) The initialization method to use for the +printed.

+ + +
init
+

(multiple options) The initialization method to use for the variables declared in the parameters block of the Stan program. One of -the following:

    -

  • A real number x>0. This initializes all parameters randomly between +the following:

    • A real number x>0. This initializes all parameters randomly between [-x,x] on the unconstrained parameter space.;

    • The number 0. This initializes all parameters to 0;

    • A character vector of paths (one per chain) to JSON or Rdump files containing initial values for all or some parameters. See -write_stan_json() to write R objects to JSON files compatible with +write_stan_json() to write R objects to JSON files compatible with CmdStan.

    • A list of lists containing initial values for all or some parameters. For MCMC the list should contain a sublist for each chain. For optimization and @@ -271,494 +197,515 @@

      Arg has argument chain_id it will be supplied with the chain id (from 1 to number of chains) when called to generate the initial values. See Examples.

      • -
save_latent_dynamics

(logical) Should auxiliary diagnostic information + + + +

save_latent_dynamics
+

(logical) Should auxiliary diagnostic information about the latent dynamics be written to temporary diagnostic CSV files? This argument replaces CmdStan's diagnostic_file argument and the content written to CSV is controlled by the user's CmdStan installation and not CmdStanR (for some algorithms no content may be written). The default is FALSE, which is appropriate for almost every use case. To save the temporary files created when save_latent_dynamics=TRUE see the -$save_latent_dynamics_files() -method.

output_dir

(string) A path to a directory where CmdStan should write +$save_latent_dynamics_files() +method.

+ + +
output_dir
+

(string) A path to a directory where CmdStan should write its output CSV files. For interactive use this can typically be left at NULL (temporary directory) since CmdStanR makes the CmdStan output (posterior draws and diagnostics) available in R via methods of the fitted -model objects. The behavior of output_dir is as follows:

    -

  • If NULL (the default), then the CSV files are written to a temporary +model objects. The behavior of output_dir is as follows:

    • If NULL (the default), then the CSV files are written to a temporary directory and only saved permanently if the user calls one of the $save_* methods of the fitted model object (e.g., -$save_output_files()). These temporary +$save_output_files()). These temporary files are removed when the fitted model object is -garbage collected (manually or automatically).

      • +garbage collected (manually or automatically).

    • If a path, then the files are created in output_dir with names corresponding to the defaults used by $save_output_files().

      • -
output_basename

(string) A string to use as a prefix for the names of + + + +

output_basename
+

(string) A string to use as a prefix for the names of the output CSV files of CmdStan. If NULL (the default), the basename of the output CSV files will be comprised from the model name, timestamp, and -5 random characters.

sig_figs

(positive integer) The number of significant figures used +5 random characters.

+ + +
sig_figs
+

(positive integer) The number of significant figures used when storing the output values. By default, CmdStan represent the output values with 6 significant figures. The upper limit for sig_figs is 18. Increasing this value will result in larger output CSV files and thus an -increased usage of disk space.

threads

(positive integer) If the model was -compiled with threading support, the number of +increased usage of disk space.

+ + +
threads
+

(positive integer) If the model was +compiled with threading support, the number of threads to use in parallelized sections (e.g., when -using the Stan functions reduce_sum() or map_rect()).

opencl_ids

(integer vector of length 2) The platform and +using the Stan functions reduce_sum() or map_rect()).

+ + +
opencl_ids
+

(integer vector of length 2) The platform and device IDs of the OpenCL device to use for fitting. The model must be compiled with cpp_options = list(stan_opencl = TRUE) for this -argument to have an effect.

algorithm

(string) The optimization algorithm. One of "lbfgs", +argument to have an effect.

+ + +
algorithm
+

(string) The optimization algorithm. One of "lbfgs", "bfgs", or "newton". The control parameters below are only available for "lbfgs" and "bfgs. For their default values and more details see the CmdStan User's Guide. The default values can also be obtained by -running cmdstanr_example(method="optimize")$metadata().

init_alpha

(positive real) The initial step size parameter.

iter

(positive integer) The maximum number of iterations.

tol_obj

(positive real) Convergence tolerance on changes in objective function value.

tol_rel_obj

(positive real) Convergence tolerance on relative changes in objective function value.

tol_grad

(positive real) Convergence tolerance on the norm of the gradient.

tol_rel_grad

(positive real) Convergence tolerance on the relative norm of the gradient.

tol_param

(positive real) Convergence tolerance on changes in parameter value.

history_size

(positive integer) The size of the history used when -approximating the Hessian. Only available for L-BFGS.

- -

Value

- -

A CmdStanMLE object.

-

Details

+running cmdstanr_example(method="optimize")$metadata().

+ + +
init_alpha
+

(positive real) The initial step size parameter.

+ + +
iter
+

(positive integer) The maximum number of iterations.

+ + +
tol_obj
+

(positive real) Convergence tolerance on changes in objective function value.

+ +
tol_rel_obj
+

(positive real) Convergence tolerance on relative changes in objective function value.

+ + +
tol_grad
+

(positive real) Convergence tolerance on the norm of the gradient.

+ + +
tol_rel_grad
+

(positive real) Convergence tolerance on the relative norm of the gradient.

+ + +
tol_param
+

(positive real) Convergence tolerance on changes in parameter value.

+ + +
history_size
+

(positive integer) The size of the history used when +approximating the Hessian. Only available for L-BFGS.

+ +
+
+

Value

+ + +

A CmdStanMLE object.

+
+
+

Details

CmdStan can find the posterior mode (assuming there is one). If the posterior is not convex, there is no guarantee Stan will be able to find the global mode as opposed to a local optimum of log probability. For optimization, the mode is calculated without the Jacobian adjustment for constrained variables, which shifts the mode due to the change of variables. Thus modes correspond to modes of the model as written.

-

-- CmdStan User's Guide

-

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

-- CmdStan User's Guide

+
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-sample_mpi, -model-method-sample, -model-method-variables, -model-method-variational

- -

Examples

- 
# \dontrun{
-library(cmdstanr)
-library(posterior)
-library(bayesplot)
-color_scheme_set("brightblue")
-
-# Set path to CmdStan
-# (Note: if you installed CmdStan via install_cmdstan() with default settings
-# then setting the path is unnecessary but the default below should still work.
-# Otherwise use the `path` argument to specify the location of your
-# CmdStan installation.)
-set_cmdstan_path(path = NULL)
-
#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.29.1

-# Create a CmdStanModel object from a Stan program,
-# here using the example model that comes with CmdStan
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-mod <- cmdstan_model(file)
-mod$print()
-
#> data {
-#>   int<lower=0> N;
-#>   array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
-#>   y ~ bernoulli(theta);
-#> }

-# Data as a named list (like RStan)
-stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-
-# Run MCMC using the 'sample' method
-fit_mcmc <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  parallel_chains = 2
-)
-
#> Running MCMC with 2 parallel chains...
-#> 
-#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.2 seconds.
-#> 

-# Use 'posterior' package for summaries
-fit_mcmc$summary()
-
#> # A tibble: 2 × 10
-#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
-#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.

-# Get posterior draws
-draws <- fit_mcmc$draws()
-print(draws)
-
#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 -6.8 -6.8
-#>         2 -6.9 -6.8
-#>         3 -7.0 -7.0
-#>         4 -6.9 -7.1
-#>         5 -6.7 -7.0
-#> 
-#> , , variable = theta
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 0.28 0.21
-#>         2 0.19 0.20
-#>         3 0.16 0.17
-#>         4 0.20 0.36
-#>         5 0.25 0.34
-#> 
-#> # ... with 995 more iterations

-# Convert to data frame using posterior::as_draws_df
-as_draws_df(draws)
-
#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
-#>    lp__ theta
-#> 1  -6.8  0.28
-#> 2  -6.9  0.19
-#> 3  -7.0  0.16
-#> 4  -6.9  0.20
-#> 5  -6.7  0.25
-#> 6  -7.1  0.36
-#> 7  -9.0  0.55
-#> 8  -7.2  0.15
-#> 9  -6.8  0.23
-#> 10 -7.5  0.42
-#> # ... with 1990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# Plot posterior using bayesplot (ggplot2)
-mcmc_hist(fit_mcmc$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-# Call CmdStan's diagnose and stansummary utilities
-fit_mcmc$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181226-1-73471f.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181226-2-73471f.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit_mcmc$cmdstan_summary()
-
#> Inference for Stan model: bernoulli_model
-#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
-#> 
-#> Warmup took (0.0050, 0.0050) seconds, 0.010 seconds total
-#> Sampling took (0.015, 0.014) seconds, 0.029 seconds total
-#> 
-#>                 Mean     MCSE  StdDev     5%   50%   95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8      781    26932      1.0
-#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0  2.3e+02  8.1e+03  1.0e+00
-#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0  1.0e+00  3.5e+01  2.0e+13
-#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0  1.9e+03  6.5e+04  1.0e+00
-#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0  8.9e+01  3.1e+03  1.0e+00
-#> divergent__     0.00      nan    0.00   0.00  0.00  0.00      nan      nan      nan
-#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6  7.7e+02  2.7e+04  1.0e+00
-#> 
-#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47      796    27460      1.0
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).

-# For models fit using MCMC, if you like working with RStan's stanfit objects
-# then you can create one with rstan::read_stan_csv()
-
-# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
-
-
-# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
-# and also demonstrate specifying data as a path to a file instead of a list
-my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
-fit_optim <- mod$optimize(data = my_data_file, seed = 123)
-
#> Initial log joint probability = -9.51104 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.

-fit_optim$summary()
-
#> # A tibble: 2 × 2
-#>   variable estimate
-#>   <chr>       <dbl>
-#> 1 lp__        -5.00
-#> 2 theta        0.2 

-
-# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
-fit_vb <- mod$variational(data = stan_data, seed = 123)
-
#> ------------------------------------------------------------ 
-#> EXPERIMENTAL ALGORITHM: 
-#>   This procedure has not been thoroughly tested and may be unstable 
-#>   or buggy. The interface is subject to change. 
-#> ------------------------------------------------------------ 
-#> Gradient evaluation took 9e-06 seconds 
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.09 seconds. 
-#> Adjust your expectations accordingly! 
-#> Begin eta adaptation. 
-#> Iteration:   1 / 250 [  0%]  (Adaptation) 
-#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
-#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
-#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
-#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
-#> Success! Found best value [eta = 1] earlier than expected. 
-#> Begin stochastic gradient ascent. 
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
-#>    100           -6.262             1.000            1.000 
-#>    200           -6.263             0.500            1.000 
-#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
-#> Drawing a sample of size 1000 from the approximate posterior...  
-#> COMPLETED. 
-#> Finished in  0.1 seconds.

-fit_vb$summary()
-
#> # A tibble: 3 × 7
-#>   variable      mean median    sd   mad     q5      q95
-#>   <chr>        <dbl>  <dbl> <dbl> <dbl>  <dbl>    <dbl>
-#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
-#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
-#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  

-# Plot approximate posterior using bayesplot
-mcmc_hist(fit_vb$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-
-# Specifying initial values as a function
-fit_mcmc_w_init_fun <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function() list(theta = runif(1))
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2 <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function(chain_id) {
-    # silly but demonstrates optional use of chain_id
-    list(theta = 1 / (chain_id + 1))
-  }
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.5
-#> 
-#> 
-#> [[2]]
-#> [[2]]$theta
-#> [1] 0.3333333
-#> 
-#> 

-# Specifying initial values as a list of lists
-fit_mcmc_w_init_list <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = list(
-    list(theta = 0.75), # chain 1
-    list(theta = 0.25)  # chain 2
-  )
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.5 seconds.
-#> 
fit_optim_w_init_list <- mod$optimize(
-  data = stan_data,
-  seed = 123,
-  init = list(
-    list(theta = 0.75)
-  )
-)
-
#> Initial log joint probability = -11.6657 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.
fit_optim_w_init_list$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.75
-#> 
-#> 
# }
-
-
+

The Stan and CmdStan documentation:

Other CmdStanModel methods: +model-method-check_syntax, +model-method-compile, +model-method-diagnose, +model-method-expose_functions, +model-method-format, +model-method-generate-quantities, +model-method-sample_mpi, +model-method-sample, +model-method-variables, +model-method-variational

+
+ +
+

Examples

+ 
# \dontrun{
+library(cmdstanr)
+library(posterior)
+library(bayesplot)
+color_scheme_set("brightblue")
+
+# Set path to CmdStan
+# (Note: if you installed CmdStan via install_cmdstan() with default settings
+# then setting the path is unnecessary but the default below should still work.
+# Otherwise use the `path` argument to specify the location of your
+# CmdStan installation.)
+set_cmdstan_path(path = NULL)
+#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.32.2
+
+# Create a CmdStanModel object from a Stan program,
+# here using the example model that comes with CmdStan
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+mod <- cmdstan_model(file)
+mod$print()
+#> data {
+#>   int<lower=0> N;
+#>   array[N] int<lower=0,upper=1> y;
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
+#>   y ~ bernoulli(theta);
+#> }
+
+# Data as a named list (like RStan)
+stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
+# Run MCMC using the 'sample' method
+fit_mcmc <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  parallel_chains = 2
+)
+#> Running MCMC with 2 parallel chains...
+#> 
+#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.2 seconds.
+#> 
+
+# Use 'posterior' package for summaries
+fit_mcmc$summary()
+#> # A tibble: 2 × 10
+#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
+#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.
+
+# Get posterior draws
+draws <- fit_mcmc$draws()
+print(draws)
+#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 -6.8 -6.8
+#>         2 -6.9 -6.8
+#>         3 -7.0 -7.0
+#>         4 -6.9 -7.1
+#>         5 -6.7 -7.0
+#> 
+#> , , variable = theta
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 0.28 0.21
+#>         2 0.19 0.20
+#>         3 0.16 0.17
+#>         4 0.20 0.36
+#>         5 0.25 0.34
+#> 
+#> # ... with 995 more iterations
+
+# Convert to data frame using posterior::as_draws_df
+as_draws_df(draws)
+#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
+#>    lp__ theta
+#> 1  -6.8  0.28
+#> 2  -6.9  0.19
+#> 3  -7.0  0.16
+#> 4  -6.9  0.20
+#> 5  -6.7  0.25
+#> 6  -7.1  0.36
+#> 7  -9.0  0.55
+#> 8  -7.2  0.15
+#> 9  -6.8  0.23
+#> 10 -7.5  0.42
+#> # ... with 1990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# Plot posterior using bayesplot (ggplot2)
+mcmc_hist(fit_mcmc$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+# Call CmdStan's diagnose and stansummary utilities
+fit_mcmc$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251438-1-65b170.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251438-2-65b170.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit_mcmc$cmdstan_summary()
+#> Inference for Stan model: bernoulli_model
+#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
+#> 
+#> Warmup took (0.0040, 0.0040) seconds, 0.0080 seconds total
+#> Sampling took (0.011, 0.011) seconds, 0.022 seconds total
+#> 
+#>                 Mean     MCSE  StdDev     5%   50%   95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8    781    35502      1.0
+#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0    235    10662  1.0e+00
+#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0    1.0       46  2.0e+13
+#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0   1874    85179  1.0e+00
+#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0     89     4050  1.0e+00
+#> divergent__     0.00      nan    0.00   0.00  0.00  0.00    nan      nan      nan
+#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6    775    35215  1.0e+00
+#> 
+#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47    796    36197      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+
+# For models fit using MCMC, if you like working with RStan's stanfit objects
+# then you can create one with rstan::read_stan_csv()
+
+# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
+
+# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
+# and also demonstrate specifying data as a path to a file instead of a list
+my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
+fit_optim <- mod$optimize(data = my_data_file, seed = 123)
+#> Initial log joint probability = -9.51104 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+
+fit_optim$summary()
+#> # A tibble: 2 × 2
+#>   variable estimate
+#>   <chr>       <num>
+#> 1 lp__        -5.00
+#> 2 theta        0.2 
+
+
+# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
+fit_vb <- mod$variational(data = stan_data, seed = 123)
+#> ------------------------------------------------------------ 
+#> EXPERIMENTAL ALGORITHM: 
+#>   This procedure has not been thoroughly tested and may be unstable 
+#>   or buggy. The interface is subject to change. 
+#> ------------------------------------------------------------ 
+#> Gradient evaluation took 8e-06 seconds 
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.08 seconds. 
+#> Adjust your expectations accordingly! 
+#> Begin eta adaptation. 
+#> Iteration:   1 / 250 [  0%]  (Adaptation) 
+#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
+#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
+#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
+#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
+#> Success! Found best value [eta = 1] earlier than expected. 
+#> Begin stochastic gradient ascent. 
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
+#>    100           -6.262             1.000            1.000 
+#>    200           -6.263             0.500            1.000 
+#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
+#> Drawing a sample of size 1000 from the approximate posterior...  
+#> COMPLETED. 
+#> Finished in  0.1 seconds.
+
+fit_vb$summary()
+#> # A tibble: 3 × 7
+#>   variable      mean median    sd   mad     q5      q95
+#>   <chr>        <num>  <num> <num> <num>  <num>    <num>
+#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
+#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
+#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  
+
+# Plot approximate posterior using bayesplot
+mcmc_hist(fit_vb$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+
+# Specifying initial values as a function
+fit_mcmc_w_init_fun <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function() list(theta = runif(1))
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2 <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function(chain_id) {
+    # silly but demonstrates optional use of chain_id
+    list(theta = 1 / (chain_id + 1))
+  }
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.5
+#> 
+#> 
+#> [[2]]
+#> [[2]]$theta
+#> [1] 0.3333333
+#> 
+#> 
+
+# Specifying initial values as a list of lists
+fit_mcmc_w_init_list <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = list(
+    list(theta = 0.75), # chain 1
+    list(theta = 0.25)  # chain 2
+  )
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_optim_w_init_list <- mod$optimize(
+  data = stan_data,
+  seed = 123,
+  init = list(
+    list(theta = 0.75)
+  )
+)
+#> Initial log joint probability = -11.6657 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+fit_optim_w_init_list$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.75
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-sample-1.png b/docs/reference/model-method-sample-1.png index d7b8b97f8c19824bd7f4936ec52513cc8fae745f..27fd88217456b4a7d0da11a09599ea685749e06d 100644 GIT binary patch literal 15795 zcmeHu2UJtp+IGN#j)ie-BOr{@Y!pF2YOo9#5KyYrpfn{ykPaaPE90P|AR=8wL;|73 z&;kTSsRDx#Is~Ky2oPElNJ#ki!I^c}x7NRIzw5jATQ`fW6HZRq=Pl3sJn!DQf5Fsn z+jnB$!CKDfQHN+A?_Ee23`BFc8_N~X-M8I^%$xx$Wtq&!(VF}v-MIJmm5mR%TrqH zBa!fLZ0GfZTs=H8Pj46R)9)=h4&OU?izRd|@y(4K*+Lg!J@u?$U7d-cUhnYx-mvN!BbtBUKM; zhncpex%fmTzR9YMma?HP9}LFfDD?Zz$9sEw)zVrBqQ4Z7$@b~3Z_9GU7sVtLd%L>c z7I{ItL43s4zsBRYZ{Hqv^3)v?iDZXF7VB*)C@7elJn84C*b3Pn_Pysdk6Gc_!2J>$ zERv`<`x;H!Sku@m6I^g)(%{TWBrY?Qy>|oGV-(OIw zKc0&x*x)!xl%dZFOTKAC;4+n^!+)E?2YAS`)nW^~wQ^}0DcSg)@o*}#O_LJ6{q zmg`SS`y8hzVDOtc@~r9GH}Z((&B2x|4VpHA+QAVAcfY(d_@xwZ8=VhYFUI1at$+?PG(Tf{q1lM%p|dV4!o?dZ!mQ{f#?iC?)nJtwHH8 zQrO_|ovPj*g84alDI3C8&iVl^U&?xasgp7orO?J?i-<^96-jo}doeGLDPb$0*Uy2* zpZ{XWY4!)`Q$*)qK$Sm0$+8!mS;d;(|8SB4mk>nH1zFQs#58RNW-GV~2DQZ8Dik5L zz(6|Ytu8fDD{C*+;Z`$5QK%l4E*N`Nh0Lt5{)<>!i)d?nWqun+#zy;OZs_WZN!J0( zI0*3aw*nZvBW6SOuY%Xh=YEEU>F-?#VgKI@g?|kC??lo+PG5Uw%q~jA92tK%Sv%*!G0ai>eORRSV z^HeVTp=jei2=_OY!)9&e(gD`sjMTBBU{BRgCG}!p`wv%uKgRtJKK4I&`kK`?m*q~9 z&1zY392>6Ws1Xe}rm78|F!<_vr-%*7&Dr6@y=~lwXfvMVD5*n&3KDMc)DvOYp2bo_ zV6#QtPt?lg^h$ZiDg){__d1`Ne(R|S_0Yd6^6(_`21hS~0_l2z~eh4W*6{)E3>+l2p1y4JmF{FK1&GODcEhDF)Y z@+Y@%-%d>a%#$2eZ7RbPyR}(+)w3l<8F%{ObTY;&QppsBX4=xymGz-&DfST0DC?oo zm>7vmb$~d4N^F*vG#zqU^W@35%VEz^;w9Gu0!Y?2HU!brIRrxgl1w|G zY28z`rg!%}Xw1|Ja?}8pZyQ@!{=myYiVNR!R!^c{SO2VWiWnbRA84tzpCTy5im6#q zw_&iKl)DViwFabcuAD7h8Sqq3+FX9Nq|0nSiV4@S1;z5Or=b7MwZEIa|1X)&iv}Fb zvEas@~!R=$z~g4FOAmo(Yd znHg`IbQS4jhx%KLBf7?K6e?nDTYf%s@0FZT*^wDHTl>?=K7DFPc5n!smw ze@?tBr6Gwcl-oBCqz5?|K34JLVT1IGYE8>bGoW<8LQ0oO0oeIFw)YefQc_ZYAO(0{ zU0t0jf3P*I(%$rz2)CW33qw=8Cv2yu_b=*hT=(w(qiRq<`YQ< zXFHwS%#&OPi-1AdAo`)z)?#%WKHXloAtvPUV#?kJkcbHsLSe(;qFPZ{u7x zrCU*Ww-G5#NtK4BU!sUZ&0q`~P(R5BwaOy5ajALN*T$uD61s13;0bNGHan9>5ZktR zwma6AD#d@xJPZ)c8**}ALfj2%@Qa(p6Wml_g_+cU5j6FH^o0$w%5ptdOR`qyD&nBQ zYjA1+9|Puvs*ZCF<`Cbuvbb?d2>K^I6t?GUui z&O%@+1O&Q&mCdOiONjub7eAOwfXn@-)$6|zZ~xBRgj!i~adCf;-8{qe))yXbC5dnF zWtEx0^GR)zhhy{Ym~Ax^&56Ga!5mDy!yXLpDk0G&tgupcJB z2s5v%+jqP{PtPB2aqQd{YJJc6k`V0KONMu)0V!AZ=gxqEs2!v8visbL301~-qWJ9v zY=Iy502E64P@8Ux2Gxq%3BStdp2&yeP9|89hBe@j%7;DsdC<*AKx@0fpT`YMU><0A zCgNKN9No1huV_M5=$oM@y#q)3NE*CeuTl+^vG>hZ$P)5vdsGG*TI~Xk2OKL z^iQ9lSrb_>+p_p*3cmu8=T~0;j=(JqGDGz(>TZTr%|q3xiWC;+D_A;8A@wtV+OqV8 zhO@)bl~MuIgjId{c7Q|jd!8zjtyP0cY0j>kOULic#F`1i@*SX};3M58L`nLwi3dva z;Iz*_VeUe4`2_p?-q{{yI-f1jx$y;BYLAa)n3 zV^Jn;?N|K=ipaUy%VD5N;-H?UF-in=eda@$US9X-ykE98ec~NwemZv?bUm&$<8YpJ z73&nq<4ATl4}I-s5E_k!`bb4yC@^1w$D5j(PR`Ee)9;r|O|%u^0|8lbo*nrvLIcB| z0*pKb*eR{Bpx~lcZ-+n-PD`&kT?C4ON`j>o_5J&sqVm7w7Zq7E7>wKA{$W>KT!yJP z#>$fAHbdV2Tj$ySi$;G!BS;T}_P^V5jiDETIxbTO&ffh=yWXS&a@jvUL;p%z<;jzbC}sftPUBs+U&w`KjpQSTD-EF`MQjpg zu3L~2IFrKp+0hW02d|QKV5i>+l&4BAx4ssddCaV}VoEK3EBag0s(3IpD=?MATxqra z*4NhG_}`w}Up?jh=Tq_c$5t#@6krIg1^8PCQq?M&DY!k7%i!IF2QfLa2R1)m@`^aP zX@l3PhXq^ZEk{TCOD+V8GLsqB8EO8O*x@Yh@S_C3Dz za$nAO$U87|apm=tT|ewtScoy6VS96UwKqH-JCw+fi+QCO9d~cv(ta$BK}>6@WG;lv zU}$I%HZNYR2eA)c8)UyB>!CwJ0OJA0JVnwVgKfkF{@KWT z<*AU<34I0V~6V81Zzw9`CEQ|hri5xAL##X=Ae+5mzN7IzH#(fE@>c;76`_k z^}3=$w{3j>5yJ8a-;-6zX0zKpNpjZfcB&={nPm25*99^?*jClT1KmXX;2*O3{r&xS zz3B)YpESB+WgBt$RM7US*N5w*Que93yNWpuw91I=_rE|ij=-v<_~sqVwq)a;{C00c zd6EX&+p_m7wQqw{CsC#+ff_vaD*XAv-?_RQ2x!|iRCrc;p;v7k|<*N6AowEYkKmsL%v1lu@ zF`{NCMEoyYO((vT)Ra%??rt|{PbQeu=MH_zrtN^SZ$jRVIQ){RFyTFv-4|o!1G@^t z0onv!-x)y@uHfto!fQJsZvF60MeKSCo%LtqZ}VCs^S_sXa`9hj3r}<))bi0sb!l2Q z3@a67TB8DGF`aD}8}BdNGxrQU08`Lk^P&u*awRr3Z0N~=d;IYO#c2hywaxoI1NXs9 z)^ZdUBcpv3;bHGPb)YHNHB37)CLk#$yjxNKhc!gWPx3ts0_5EbJFmbEFv6#UD9Hb? z4b%_JdME(fFJ{XsO&y}&WbJQKbaP-77psxi!`M&O;zaI!)rG=IoHh}>lN=kI0+w== z^j~iJ7SvB;O8e_?*)}P-BU(293r7U_ocfH^fj#gc0_S=j}aoskeK}=vb22)TC4_(2%WpUcGdrw zM0(41UrY4-yzs&TUGhDn*m<)1^`ja5jP)6LdD2wRfYZ_1HwG?*_Mo=qT%XLA4O#hA z-`mN{eme4SA@U2=9jCgyR-*JTZ`~Tbf}yTcO|pESvzfN=Rf9zbZ;kwCSIg(cDwB|Vlb`1NfEcQ zw#svK+Tt*&6P|&GF#(^Ghx+DHo0=4FBYgTN_m#46Wl4@TcUI_{NxtWbH6-P}q4EU} zl1${H+^UBG@#sF6fOv2W{VE?n3q{vI>_GEyBsQ8qa_X=G!A9jm^E7Ef@= zMCU6useSYbaDBzG&b{H`HbUDU_ayq=yTbRvQ%uh&Ci7uUe$8XrdA&A6kBUV) z_0X{E5cW~q5B#Vr7i1k%xkpzn_DLt_0=PrNxWIczOvFhmb1LuM%<+T=<&`4#VK`B% zSi7h+mj_N~rgCI5_DDXtWXwRGu|y%%2#zkR#+~15ix@baA>>fOJ$+(|d^;A4z3(kd zRn<|nHyU!&t*uPv8;#H%iyU?GQ{;qM;cVojuJr{)*=U@MW=Upu-1%IijF?A{_(|~e%R=aNiG(EctP^|mJ3UF3 z*{qN)S9#8bFY^>l)XtOCRgMsf@6zoN2C0n+DNVUon1PltsgLSY=&u25Nqf{i)Km9M z{lK0yFY+cH8K*okA)zUcwmL(7vXf{XiY+ER-1?-KU!w3qFmU|&=c{vKk$7ojzk+}} zlPyEE8-Lyxi-gM=w(OV;2)`hjzSl8DL2uIEP&T3aE04UXcBJZpxnlI^-jh4hv0gD- zMa`Obhbl-*yLh}?^0U3eukbrGaTPTkqMS{7Vy4x;kX*O}JIxkeJf(d^8b|K610LEc zud5LL=y*4e?4X#r_+@tXKAY53;@&59oBEpgDM~`ZdwSu8>+70FMIB}C61v`o(cLd6 z>%>I~eIl~;r~2zrcELA37-U&z6ZXz}_q@v*XzDNl%Miyt@3U3$ad!R*i&aQ%_+Y$U zZ)tFY&Z}p5hJrtOswIVf_lCRAKZ;0&`816B zdXntaz5BYRe4S6bHzy~6rKWU|T{Yma+FS7^N$R9_@%PH9;I+tRYrzlwlA!AGhD^p# zBzwOyzb!mePQCYF&}94dD}x4@nNqs%Xq=54VvmB|(=Cc6FY0%{|M(zW=51W2{>_V^ zqr99=;bzO!?=X~j5{;d5G0EK|XMi+rb-3|qcXy$iaT%GsP}+jc?H;YVGmbi(h{~%_ zH`8TV#*E;za~r0*+pe?Zze6DHj{&;Ot9WBu?N05cX-dnT$uO^cBU6xwnM?ON3y*oqMdM{1(u$D4M=-j0f;BQaCgEFQrHOAaBUao5mkm!3Jc z6v5uagmR+c{0TF^iH!J}f(n`CaYJ}wteIX2g|Xq#D}%HsHj6cSYl~uCRh53T9&aic zCs-CSSJXd1CFc<=kCaXoHWmg>_vj9U58d56_CU3{iy129WHPjF0APn`*x}8i96Svo5Eve0V?eT+B)>ZpBbqW6O(-6I_SIu_SbJ4+#aGpk=fD zf&^;@gQ@rXZQL+ZfdtM)6G9E^j*#XcA7(3<;zH?{W?B9jr}xDPsMo8lrJp%<(N>yt2k6_-!DnJd>gCgt-$5=jSd2D<^ZjF71uMZ#bz z{cPj*4mltRH21fU!rfZJCu|U(oY`I)6)YYx1yH! zAh1WNeD=d@(gx+Y6R&V9eRRFz&RzH1I|nl^WNmMA6AG zpQwu3aOkbUA}RgUYeWlaIvq@t476>MNrsWbBUj2r<9 zo>ufczEnla!O_~I_Vwu=%$9aqZ`Wgt9NM)?&ptx1ULfH zhEwumi&FSp)LNxEe$7khyzIkCDHO@hI_J_8aqp4?i1`6dr3MK0-r>n=n1y2nuP2Q^ z<*^U6KR_K>0AF(4imkG#GWDI6Y*N^;IPF|a-;LV+WAqJg%7J~zxSs3fBxzI4(WiZ( z3<6Y91W&grAGj+iNo{R?vSG^}nSD6nMMKFD8H6wR;fI->JSjR750u_@^E}cUJ3LA3VvG96WfFm1mwEkrKqBR>C6Y&Kh;w9QK_Oo} zh)Rd*f7paEa&3qrlEHw5VKwJI?)2kMWO}r|`Wds;)sQa6V})UGJ;KkoPf|N)Dm~}2 zn|N29&N-K5(L)qq7;9!Fdpte5`Ju+Z^W8_SU4e<0ms@bDAbv(($C@^PimeK!ZtkIX zDp#1?-|?;6+4kgq@fwHtmp7;)o)rT2?;jZWhDcNe-tY|6gu#CM(v#!}?!QuIl<{Q> z{5;~t_YXij)cFf`Of?O1qtilSClG))cE2$Qhjwf<<6EKsy8md#-mcG7@<_b9!K`#E}w?HPxPe z2l9Xx(%r(OtP1*YtNi5^Iy@~Btjz|i;_`y0?rjsFZw{(dMRnvr2UYTB`mmh-HRY_XMt!D7cI zv-3NnaaKW67j`wiQLruZbFwwm!Q3AWz8UiFP$W;{`o#f24a4ql-<#M_6&x*@s!Ek27 zsY`&3V2<_i$rK*>de6IS*Dn_VN(u3_b1`A8Y-H1PG0iE`Llg#cLI7o5^;&#?liiUx zb|}S@lnAAQ)N-96M53M=WdPkU_9{xTArihaB(rY|CMXExnMZGtm*!M{^*s2g6qGC1 z&}cJI>mZN=V5FLv8tNG{0@&U;l3hPtEU#=5FC%?m5fo?DpqV~!(Rang^tb7c8Ylt= zW$eyQQ!p$5(a-DJfa7@Q!I1PDwOH*IkaKkh*dn-D5k$;>?Lb#>TNgNz=x$=2>yRK* zI#(mE>QeK-MyJR@kK%n%YmFO?e9mwNmb6cC7$6nUy&P;SvM6%9RM}yP-Mu++`lhK&YOvOT9qatWCjK zG(dL9u_ekG$167dy8Ecuymt8PZK(%$$vCzL?itW^fzjy?gVMBs!^GZ0RycncjpOyk z%7ARYP#stt^0i6%E<30ibD`pL53uT~YeB#?3(he;^FUUwQOy(jf;eEVq z+qRAVc1F|$08r)1s1noSAyg+~MjoQWTmThhFO>`pU5)~)gn#pat{~U{GEy?p)%6ZT zpy;EX8dSk>On~FUV@Bp1m83Y~5K#iM_&JDY)#j5fX`VXZNn>QPZStB22L~m4k(FT~ zTpo{u5(iT56=d`CWkHn-DK0dnOZh(mbsCUl{xUsXnzk~g5i6rX3-i-k84{(_;j1A) z145ende>vhQWT{Q1a+(oa=F1q1tcsMKL;h5O5uxMAoJdrgK$H{ynp{*hdZ8bTjqaF z4@fDHR+@+yz@(eQz->?5fdUc1DeJ6IEbjAN<6<%hk{toXlTYzCpCL0Vp$v>du3QNp z)(DpdwG8}i*>kK86xw*-(qtU37}p3OlU-U91aw6yxMl2v^&f$pH3Ndf#>QpP22Kqlhy)q9aAu03s+TBE8I5Lsbxwjs;PWUIL*6b=6S>G%)l~R73&+ zDFGn_a2Uae1T`dpkOU3I5FipwkPwplo;dfP-#&Zq&W^kH?%v;dFt5*f&-=dL`uUb~ zt{-tYr1kekP$B|R>xBaR5PZ!sG zyY$Y;fz7#*dk$-_(NNH;x%Fzv^+l~`G)^7evOi`2zU2Mi9X+~v+qQ4F-r^Wvjj?ZL zF`rY1;#NO7@xzVyvka0!B9+hTDK&BKmNdPLA0HYoTO_CJ3;t}jMy0YzRrqA}#Dr3O zM})*pzIlGJPZ{Td*>U>Xlm_{Y3fuvng8YBxxh+CyI6vjID+XQJS)ES^iQdQi3ZBwl z;(a2br^TsfwK8v>fHJ>XJu!;HOYaF+gUO_Ii}Wu16sU|BHF#8Ii$pklN%;nNN6YmW zmG}NSO2#`*5xsX!;YcpT05Z(y=44w27$E4!ubSm!@e*%%Oggei4<4<{(ELf(ezmju;_-;a+5|Yx|kM2UQFK?P>*fM9``G@z$ ze^M!LiQ#x~j?;FPZsg5}V90R|!c}04>_5Fzg>%HGth(hDShrI}Jz@CP#m>;s zNQnwwDk>$STBJpIDPFkZKc5eaYMm%cCb-%~?CKvtOKs@Wc5Jq4C!Ws=mP^(_3v60V zb`kyxh^L7U&;z6)`5lj5xK4#Mc+b4-s20Kkj4|Hnwit~6y+D7{#{*y((D2Klm>>zg zc*2RNTFSlJG~b@p#IIukF+5BK`ddNl>q8}m_oAP0MJPNTA7~Da@ZAA4yk_#5H@()e z9@xNg^<%2K9#!Dz6Fy+?B_6u?Xn)CIGwQV!1hq8^{N;N8yY>GC%tzmd^NFU1;pMn? z1}c^ULfftb=CWTz2}dg@x(F!v&N^V8n(<@dbmB^&iuqY}B5c0afz-dzO#dL6|J>z5 z!gA=I@rmO>AubC;4V~4(Ozmr3`1WEUjJ3n*z>3Z9(T@;S@H?mKy?0Of)^j(431g*2 zdTpH#*{%uGHoy{DkG!JBAO5!F?W!;1?f+akd|{~iUq$>eHX7KkZ4cR;wLy$72GIC2 zN~jk??kD%bz^ShesD1b{wMm>3{GjzbnR8Niy)e*4m?n3J4_g2#AEU4A!>{+^eWyO< z>mpQV4FbTnyaeF;hzn`Z#lt`m26|`D8aN#P?RfsbN=Uyq82w))Ud&hwB?+VW4+A#f z3tA0R2@Q5`DK1@dut++z6&T9+*$VvViKw7fA@)4qY7GeFNuEGbp_S3~_+al~@~PK? zPM76f3>{!6YAFVPCD8v%$`%M`^MEtP#IicBssMNq!i!GZ$NL8Mi2RiomDToPsuCmU zbT$!=sZiV5Z)4u9iWk|qSkV?cs;Rh;V@u@#3Q;e&t@Iy$6-*b|E#@y&%0oIt?XpW> z2X6(He)I1nv@ety{!1Fnza#JRPy@v~wY(P_C%!Kd-k=G+=y^W8CWaZviLF3pmDIiK z_*9Tfg7_3~3vF8E7}KZTTLH5E8lFn2p}Be$h7Efu!U|RD+4%eF2YQ1Va4htr*uZ8} zwPW(wcHvui-`ZWhQzuDqp53nt@9aPJERkL+a<9sco=@a~SQc6jD}(wx(I+aZRe0k( z{1C50BT(*fU#fih%}C|hHsr83wX=tQM9&iw)jSxj2%@kSm~VToTA5ntMdx`9Y6DiD zuK)trQb6*I3M&Ovu|aJC68wI1@TB!m3sd;yL4qqFG2$)O5T^7G7g)bh`1jUl|I*7r zaj8GrQO@C~y0kLUo!2|WcAu-`3p+1nnar0|PR#DS&`_{zD%2%vPaQB>pr8BS#^eyE z2zWh9I{Han1FK5RkCSTzdRUhK`1cjJ|I)(y|HS`K6}A5y$xnDF0ehFS-_*dFUu=}5 zmf_)0OLS>qp+U9-8Z1fnpuYmO-Bela(td(Seh^Z}+U@rMPd^e>hi|Ri4nVkQ3;2P~ z-e=>*b39O2$H|H&6y1~|-}V1sPk}332y#=?O)KGv0%SS{|!};z7MASud!7 z+G63HDk;s+M9oT1Bs5E|oXXTb?ZPyxz#mvhZ~>)Z+iSQdm2^nZSnfjdAh^cYjUEwJ z@nIMUmHvb2|97rT|3O99XLB_UQ#2Y9Oh@?z*I0&YeI@^WllvZ6tvAS#r3MAg9!PPmf@W&*3K zn(77t>_54K2g$?mu19A;6#qNAnqS6);jH5IqnX#Ac!msT0xJLQ6;P{z%a^)@r5F`P zlD?TfDQ|c}ZJaeRM*X8qe2@2eP>-QPTz6cP*JF@x3wXBv#Id06PwU1i1Ih4uXe_8B z7IuCpDILy6!1aamq5qTF#vfEV+zE%Eya4}le1~}ALqkVZJ-}O&W5O*}%NF{Qz539L zq4-yDPMdUawW~|=0eT6JVSAt1;$wKz#Z}~?mm^B|qc2x-uY!{H${)MBbLyA{su5(F zR+bN5vvOxU)S2JKDD-=7QLlr~f`LvPdUEH-!5Im^4G8OPV z1pU<)1br0`Dt($op)?3V106~mZGY)Ma$#F>Z}rJ@OzZ4dyHC8U zjNNblcfwM-P?6_$Yptl_*pF8O1 zw3mYV3T{bif9_?t+k8%zUpRgJxp$oD1X~6K)n#0gzcMH)S@#)x{`wXHOQUB2O|OJWFzO4gJr89g1j z!Nr>8Qp0wyt{Ul{m9&RP&`gZYtJ6lEIwY(pQY)b(RStGVmaIoJqSTUJ<6-@2U`eRXu*nR>2@)@~OpqU6m6 z!dZrq)GS=O@4QO%?#)QcW&6@$yWAXyg62l@mtG4@U*^CtDIjtSz=&DmmmJ{5z*5op3KRQZ?yB5aarJy?8UDY}HGl%O&f1V** zKev%N&FP$$@^Lq1t*9Dw?VrP_F(cF!ocwCk2rB%hj68U2=BF?jscb=(8T9ngE8B@| zhcWW{?$IFQWdTpv-P!(wJnd(T`LFC}t7*T+gEc(nGYP%gQvA zGe%6b=YB+Q>?%*B-*dK++8~3g46V+Y_H16BCCH*t?&5Mozz(-^C0A<12l>jm)ZK^e zcj{>LJfO&Q4NuPQ{tKPB+)lD6|bR=dbax+1ooZS2owWkwa=%_ohZ^#}*;j!KPi- z8h!W9YEqN|*@v|U+|x@-#+>c_H{N4h*oiZ5V8oXpl#FoxMc7R?f+_b16yPW}(xhhx zPFEyr>2#VJJzp|%CnLU+tE&|nr9J0B*x2)|IX&lfmk2w=ip=1YY2?}JDVL6<-mg{Z ztbIl4qjjBVRc#$gZm{Tlt{5wE{_1pmnk9@YNztax{_m+;oY;f#yMD;}viI#MPgZnB zU)>{fN{?3T>? z_ABNlqhN_-VU{g>dH(}gu*0LixiVvS#5fKm4W_vzU9YR0H%$?LtEi+E`{5^_(_U1Q z@%+GabB)slYl{f=7X}1Z5qgbpZ|F#;$g`~ZQu}d*>^*&+5#27xXrIu>OAAv^yC>c5 ze|WOJ3+uV=gRcHnQ+v4h21xwzzYTLRe08*iWI0KTxND$vJVX4_=^5|%@#8JMk?WI? zb#---?B3p9UI4)-#B^obz3nByzD}#+g7a~>go6a(4&>SNdl@yYG;VVcyCtDP$-u}u@20@GZg9%d zkYW@-aP&EG;!(uHd4<>h86PH7Rn01M3JRoLcOJ1ML(eO5ZZw2VLh)a%IkmW=j24-6 zG8czixPtJR7NQGTO;J2Q0|SG;UU#Faii)8Dt6M>5bd!*~CL6<|qebHZ1omM;GIA#} z97zxob#!#tLyZyRwpVfLMT!SwRm5+VTVqTq7ZICWj~~~ZKBJVBz6&244OZsV*48%I z+uKj^h3E!qPmip3XlO`nbf9;Mh- zwExEcWHiaqRzPv*l{W+pMh`AlGJjzf`R;`f$jV0W5dp}`DJ--yo-ADI>zN{k<{j(k3nGtDF!WjF;F=H<;luI7pOY?(DyNfTMsM+)$I ztl)koqqCS4J@tBsCuBu4qxw`m&O!Nx2Y2;ElbEgEHUl(v$hC0gbHD6g*G!|)^bTX) zc|pjDYMwg!U9qeWBd&985&WE$UseriK6yCa5j0p&xWH@@j&Uvso=9JHW0Gn+04vJM z$}%tQQDU5El1s4Cwl-6YD`1l2X0IX+S$;Po1tJJprK8hUmi^PE1_n9-o8i<#WQfx+ z5>A&JIrDU8W@e>{^JxXU8z*~G;5}6NlUJRd`7Ml=I_l_KuS}VM^(4KZJ(@|hG*k%P zyHR8wH_JrHj$uE{`?SOz-lgC$037dXaYnccw%46+N;C zK(ty}Sz*}d&PC^vk3J^?V;e-1S(6fpq^Htqi!~|5w;;g_Gk1g_ycFCpFGc9t?oJ;$ zU6O&Kc6T$Q#n%-asQZBh*bL+&a1P@y!uERD0dYP5d<3O2tTR=$PqWeY?Ab9z-p1uc zL02vMre6~S0R~)jdZwqS{U_^k#fSoCt-qrd#a8UY^0sN574xcOB{?}cr61Ac*4rah z40kKqRJ1B>(N?2x{NHPu^!qrd`sbw}oor%9R;4c0*QP!ldKB1KS=Ol*H^~NzHUT4B zAE6sszj4Ad?%j*xhD@i#(}{`F;p_EFjcr#26)Ze+ZK>ftH6}8AJb__~0K2#8g$Z+A zKGL;-4H1H$Kri{b@IA+?Rh6H?duKznCLlL>#!eRjz+%Tw9qWiqNn>?|{U~wJg#n=FFU~>SG>h|H@ z+J4;zmjVfx+>t9pM!BPbUK=v!Cxe)@rY><*AfXV4V-8`&%!GN-!aKVZAlQQzMu3%; zij43#zx?tG#+-Q*3NpIXSf75T{rj}7_puiq5FMT&)uB$qOR=JMWKDB(^Vy3YHjP!S z&9m$7F3QJV-xf1IHkNp6WG!+hX{KXSLq#>kV7?1gK!GKY%AT z0?rF38$D!ZR#u$W>-D~i6oU$Id3hv4hM?G;m**{hl~sWH*yKfuw%3?zL=+Yj7Rn@R z!WXYPmb+Tt&I~mjAEZ6X){311Nx|~;GhtyfrRkDH05`!v%RZLTg%X4y$>}P~h}E8jTMbO@Sv*zM)F%`8I-y_G?O3vkZ?A zvTT(!rp@gfmO9)GG`QRpM~z5tJ_RT&e)8lA2`kyJJ$mDZE0dR%Ec=n+!B~3$csSwB zrTW~fhCMLQ0x)j5bw+a`&|2+-aroUEv)ZVOwcYw@=M`e^00QD>r@e9u3)x2l$tIM- zND{L>IXStoikOH29v?8e^S=t-f74uO{Wdt!NK&xDYH7tfV3oCEkIFvR7TIwg*hA^P zQjYX2Ncq7Py4?nAbsa8!7`^9B!41TK!SK8x#g*!tN}N3ZNhn{}!RW%OSW?9p>=<>} z;^RoLv8`&2i=oaW8K-LkJc7jHPv;A_nc4L3=iHsFoB^7_QnIpwg&fNvO)CQQKr0YWU@V7jkT7n|W0ju<})lxLmp)j9?2yqm|x$OwD z1&rtL__G+SId@NOx2Rzt>x?aCG6&H*B$FNrfBP2xY!52>IW>bg3t(~UQvPLRq824Y zYao~Y_M@O;ecd4&R(If}OoUMHTAK`jLhT(N{Az>OR5Z|qD6lEt=A_bX<*6NNxT|eTl_E=uJgy{S{ zTA!jN?&4QeOKlfMS1z9fH8bx%kd&{=M~84yLSQ^U`H)5HSauYDQqRiDx7kreD7mjN zxawIwc0{WdZOs39cZ7r)V`%q(ptDc*I&(uSjOM1Je%eCY$A*=IG-AE=R%LIpcC1WIpXKG zBO{$|YYcihhZ~oH#umrd!vFp1nwZ`XU46T#Gc{yQia$1fFVZM2iv-N?Ae;$rSbo>| z@@(bZ0$F!@f8zE)A3fBOdYyF1N!c7Ptrj#dZr%hIniQsgW7BPB;RtDAod5K?o}(xe zpy!J!3)OPSxVuzYKEVa_7o+)scnIl+4EBg7YEHK|!?s9h7 z0~PBZDk0{wVkM@JTGJnwOkP3m1U~3vIcB)mPng?SXculwNkB@CDUJna>$evCJaTto zmsRG-cd^aCGV@2Qv|_hWLNo_LFZIdEf-U}k@^YZbmM1Q+;nolR_ync{dB9fZ diff --git a/docs/reference/model-method-sample-2.png b/docs/reference/model-method-sample-2.png index 093449f41694c87ed4f6a948f6425a8e02c8a550..4ed9fe5c527f0931750caf802c560ea60a2989a0 100644 GIT binary patch literal 16009 zcmeHu2~<<(y7oqjv<}c#1(B&%Dk=ghqRhc#72`k@1!Sg`83bfbfCQzLDgrVn%3M(p z0z~F91d4!w5CN0O7?dGEAYn=Zge3Rdc>1q@{r8?*?>#;D{%5U|71j>fdw=`e@B2Q_ z`@HYBbLXO|$<|G~HbD@y6?x`+GYFChfgtf8Ki>$x8F={X9N299@tn!`U<+O=Cu?zF z^M&sjD}M-5jDr8f;>PElA?O9V09x@qQKFma%k$;w-$$B!$Vzr?d4%x@R@JMEQQoBW_zaiZc)Cj-A5R>dnhN0X7ndoK>DKSaUqq``PlG4*r zh_xRIX-ulWBO#~PJv}4C!kp6Fe9F$wPTT%7K84L=&X2ZomZo~GHIfS%4fe#UjrsXC z_4P%Ii;FGMjv{-UfTM>%5SUCR^|B*jW8?`#%+mAx{D&RSTA|pU(bhQ6;;V0)o0|*k zMN3Od)SEPuqp$&>pZeSML&V;Zc(TM*>Ogjme9Pz$7jns@`g*0ee!9YUndn1(RNC;> zS+pEDdJwfiv%zrw99=ID$0xZKx`w@4iWC{KEQQ=qPIEPR92h^M7TBJ3r#Z(`H9H@% z_?(lFnoh{C7e(fe=g=^401j@3N9kdkUyxe%;r@&Z{t|p?zhCrf7ubHfZ~ji#{*BY0 zzt`}oqqAwgRlO%08s)ZZR9@M^eMTpvJG~F>eCzwp(@@;LT9((uU{ra26Wb9*2*(gJ z_>f5laQG|W@Ze;be&&!+cfpi>V{s)c`M9*??95C7=d;+?A+`?^&O3z8mSdJ*a5lYD zq!}yJwmF$4wms!+ar^ODD#hy<7(GQU<}~L`vysWP2F(~xnWnZ5xsMR*>gsxR&Ky;c z@-)h%t*oH0A4{eB>wMcWUQnU!Z8++W=FV6V>1=Cqgk=$!eP7>WFuR^;&#e6XaX%|d zOUoG9MxkNITvkz0kw>*{MaG6u&WBykr_oPq>gv2ez_P(4E%4@c4i5Bbes1qnVH~ly zp|;Pg)nN3- zPu9PSQiLJxQC`EjL2Qrj98w;$(a?h4cC7&IkQ?>ceJh0Z@g$ze_}EYxte z8$G2l0U!EreR*khE>5KgFpx$v%wg$iX>a|i3pkh|%%&N05*@eF#p!gp9Nqk6{@Zt? zJ}-;Z&Iu~bIC$mxK>EXoala|k&Thb{n&O1b`doJ+%d%bC#n!=rHa@<)m09$M>xD~t&^uli1pkJ2V0#F?F>_y!)zJZc_Fvgy3nH1 zNQ6T9h@`TGuA;Btyah?c|ACnJJHfTpMjARZr%BQ!(6NjcqV_c*KXSUF&@!xsQ;~0J zj^J@hQ>3JVXKI+`8Ss2EJAR;TiV&YyuGcDLULd$gu2oL-o}xIt!8GUE6yPh&IYeHR zw$4cAZr!DeQs4wl1>bx~=U|+xxr3f3@q-H;Tg$mcW}^@VDGGCY-W;cU5#e!-GR3i? z?eHfk^YlMPcKw}}`p2C9{0CTc7JA{QeFO{Ht4-0Cf)t*o^580dF&MX6Bb;}r3wc)g z1q6+43lWRRrd-Ek*@6>>X%^|R4xB=seV;lp)SOexuJ#fKM(om(fJP?MFXRtDsKaW2*oinfuJvic_3aA|1dU)Z6*CkPKix{5R+v2KHGly=X;1{W-(b4duN*fmD`FsH zB+ZG`YudW_UfatM zf|lC~Yw{6#Y#M^!;?CTE9J3hq1_N|t15tHugK<`K&GZu{jUa3+<0Q$0+k|!dAP&g* zpHjBsv8zbtj^q5MTyvb?Fa}fr4T6bJG~gd8|Nr8x{K<94?^nNFYh+BI4RCgke8a&Q zG7u0Usda}`UtDHpmu2h&hM2;hII+1s>RqDrjl-p-rFO#x3T2YgN@W=hYHt@8r4+tN z?re1iHAbQ;3P?k;vtBwiK<_#bnUBZ}?M$)_Il`r7wAfQ7mg;p3-{%#CXWPNl#xfOW zQz+&@_pMX7}uDVa#jn zqjN4ofn*V^k)NS%Lx5x8)!07; z-@WxsK0eImdFl*K?H(D~vN1Alt>28q)>eq%KcO+Ji#}w9FZZ{uv>%q=vat&n8`}zn zur$=mXw0G2Dt%&WCu4D)+5qyY+dPis)ilwi8*cjZjFeJ~raPm55TyDbybQ-LhfkH~ zlSH!Cgpn`L?@dfOzZaIVrF{i1rB<+vbvv3LB`wWQkq8?9y|(5gY=RnY+IZy)N8Bl- zRSxx>cSWB?VQoHXg#x2a8wgqhAZRV|=7DyV9iSyJ1yr_DoM*KyHK_J_jdQ7-g*iX& zE;Qm@VIHVgwCca^x7uWLlBg-4+5BXqhA7sIL??7lqKD{S$I9&q@__wP&+i4FJWKn^ z)(dg8uemubFc_oy?Ngw=DS<_J-dns1VE-U3t#&Bp3c4mlsR0i3tzfar_q9# ztxY5K9{jJr4l4mopy}IPUDuB8ikWVVlr-n8#V9ct3|KCMN&#*Y{hz*2V@E@+J%QO* z>bl+#z^+vRezvo<9h~N)fSwKa5^dN9$JHly$bRR!e<~RMN!`JJ-TR+bEjLj?N{djQ zsxVJx+B4-c-H*JZz8W|MK`C(78;Mx!C1g@?LPR1}%cxh@v}Tn;S_H>IkO>NuAQ9pV zs>uR9IdkG+A5ci5lFiw+7`_J7u=`8+?1p=i<|tEZLX}lz;li*T6wx%|91KVKANGS` zg=!{P2+rlm`{33z-U13UD6-Z-2lPHsSQwU^F(A+U8cwpO6q-kxQ9Vg(|*TIS?{Uj`amqBDhDSDa7cy2=*j0zs>`d+s;&>-g#Jg6R%vI3=*_E|)KC{0`zys-f~X5~*rN zeL475s!gFKI2u1r42N&*&-07j@bc6Y+>vH2eEH@)iHqEJ<|412i4156ZyQ*6C6TdY0wZh29)N)B9Pi*)DN{ zzBZ2%q^}N2poN1n(KU4jgPw1_wepi1Lec1z1!lL~QhWpdJw5W;4L?6Ka0NiWHmIA# zU7!j4MV*qk>H%B~!S!u>G^l#((I#!@rp2PmAFH>WUBOvOp2xt8w&nswl+sEv6-= zUKFrik)D8aTN35rk%dJ0pG&H)EKJmv`8Jt3`Fz#?8{ z{u66zk9Qa-ETK|kHL!TPJYznOYG+PArvhX7sLGt5@34XLTL%Fz4nW~N(Vij;=nURB z^5}~?g|tQK;cEW4qEJBkmH z0?PLW*2#Mb>HbR2!Cb+VO==846c z<69#@UEbT<3&bn$ZR7l~RyeZ$5Vmlb!-fPH`jibVSYf%I$sB3nnmNenAEUD> z_?Yr%!-5hwddwrUe z)Rop8TeTl;4ip|r&)X4w^vS+6M|RxVu{ZKB^WXk7V)N!pMfLBEH#n8Hu`-7IS<2_Q zPn?doO{9f)C59PkJsz6TW$?mK8U7h`|9UhRUDI=UjGaeK%)?&dKK^wqG>)?`M9gi! z1mHsGa=t}E_JnNukACDWsspzZ_PD4nVxmp8s$)nRdSZ~|w~UXi_I)1@hm4W?;dT%N zojv|yP=4xzb@Fj|IL}0eS8`tDe2bw;}%&@w|{I{efR8>Z{~kxi2U00NOFduzCa+D zwf#&o3w`aulRZ$~Uq4VG5{XtXvT`T)y*XxRXlV7q@kIAw#};!uWr@Ri zW`V!JoR2wut-r(Dvr+%pv4_RQ#X%ak#4i%pyWmw6E*` z4VC{D>;8uUtL53WZD?iYw(w{Q9CeOFvjC7)A!wh31pCAU%fw@)Yft&>!X4>Oew@?Y zV3h@evJ|%0N&d?6@EiQ$RPLj}-ji)jHgJH0S2|huZ=vJg-EvJdy!Zu4M!IOByA%COyqmR+ZEAxDSb!SlCa{O>0E z{|9*fYOa9Wzjlv3e(xdV%SyxMa0xzMI=z&;_*-AhlWb_a>AFQ!kki+Y-?h~ZpY6W1 z*IfCW68%yeA?V7EI_T0mGmyK@k+8j*$Aa#ukcKY>Y>Mc5ukgtnF?322_S`$0PFK?K zLT~-xqFeTxgO~HJB#6+jeeDfg3nuq9h<%(&Qp?=rGR@`~=NaD2>tWu#d*A6#KUia1 zO3Nrp=rpV|SVWI5>xGjks(b%lUXyBqi$9qs{?6gkF+;HnfTt57-tD%hPuF_CUQ3dP zU1vaG(D#ngAAP5O?8rk~QMOgJ(4_Qk?2%`*wMvW;Z^HF%tk{g?dKxf})=mZX8hazf z3eFsiHz*EA?QD^-6o8y?6|7cXN4N}o4FY4aEMxR!)8{Q zm70#pPO!G@M(#a+G{s=~`?y;?XdP0J&KXX?cbfWjliL&JcOtP1%qN#;thCzR3)-S5 zWB}3;IE;a+WQLScl_Gr6h>vLck8`>PiNASvsQSyQj)U8zZnk~ShQa7U{1fz*tHdSa4V>n?#5uLr3= ze%5?zXnLeH8J-7pXvfDr|49<|6EwH!mBZ2VKJfY0HJy1adL^@yHD z+iizn?;F~B4`-gMvT@_F2h+4H0@8Xe&Q-kI9o}|nG1~NSb*6!?@HA|Pkr6YdX)K}) z7&z2rUtd}I^u;4;6qPl28j?AOxmO5-*7_E3I} z`f>(yfmNU?uDF_VzJ*TDR;;gQ)>4BdT@|oFnSLfje}}}jWq#<>CK}Iek9*LvQA-i| zOMNEt8I1Fzy(^i!cXd9E@9L@=?E~LOwbDAoWi!kp2NOG;M`Jm}KQa#v*52IfB)Yvid$nG#gh3&8_} z;N@;uQ7GY)-F-&>Hv8&;NrJjKNMhF6y4Wk$q9uDxEf@ZoBe@mu>xG3Uzt`3^NV$SX0=Q^)zg~_nF3nu@ zSsB#L>hsF!-X}bSKT^O>&qkswdxjeIjGQICB5P`Da?->%GDrP`Eu#qR z_r*>RIYjzi{_A(hmR+NyQ*Z8ahA7(g**&qkS=*VRDpQ(`t}U2oYOsDM^PIg-?c{cT z5gzYJvKPgi&eqyIeDmhbEz;7Jn2<>~Aggr8?w?#(>P5K*1e8tc9Lio@T|MI3m8Q}* zGU6D-8np8L_K_8kOHoP~w~Ffn598ydghR*OBJKFb$RkIOCe031%BFqpGWH=+&-^~TbngreQ1uh(KSf=%QLh6YfD=${#47VQNvplo=r$YPdv7& z@Um*W<2dnHT<5A_xfRjnYP9%RrA>Z#?smb|;wK~DA6h&>wtwGg(9k$h+^W>!oNV_{ z1>LM>AJ41j5hYY_J+37A8(Uh^$_Vb%g;H;?zLM{FbQ~_h$4DheJ$qC=+c&K*y4W|e$xBXP;Vg}(t}hT zuUqUom!J^%_EMpeQiN2{OmQn1PBkOs$3}PP5mFP=9lduV!m@thfkdvD2Tj&+iWT>nR&Lf&r3!TUc3a|G{*Li&4EwQ{X?Ptxw#pb1ILE~F0<QGV1LM;~oBE z@g@~ha?l~W*}-5@qcrw@T=>j2tKuI|bs6#oWxm^(Wo~UK3D(RCY$`P>hk|Ud+|tZv z1cD}fi^Hqb$_XA$#2c8~GBR?>{a9*e9_s~qaUyK;q;flp74+K|s$&Zsny6r&J?Evm z;^qjh=Xel57a?%hs9*jt6#GC$SvVI}>eQAXXy0tiEc2&_J34+NYwLRg#F;<0&y}&Z zNT$=Ppt6g2Mg*XQa~jnusmHGO&(6x)hy2i`GkVXR@pI%*l7tUX`1@pnFPcjH#@hichKY~w@JMwa~8e3vmq1tv{I8R;SyvXupHahpBHuR>)GYLx@>$GpxYqy_c6*Al~N zQESR;Ba*%Cn~XinJ?R2IU4-?nnaZ&ZdH=I6=j~-4VjrYV1Jfv~Wt}}Dl$QquB)L{F z=uhrU(e~WiBEcPcD&bSw_Zo#=NK|QKuXV9!xJGciC5-(VrIg~g`=j?bL3}`3rU?o+ zUVe3Xm|%gQJ$D}J&$8pQ$Oen&HEL%{?nHZTc!TagEV)hq+pFH3PR#f!l+T?-CV<8r zC_-nsMzzB;afaM?Nt@l*?*(5u0OA>k7@w^i^gEX@R!#_?t$+Dh*4F-F_ORPg*fUF9 z?nKWy%?r=W+%H^(tn(H98;ZVNme3420Rl5A)w^atcoYleR9Vj{%(KX?W2r(~PapU)f2D~H#HkMcGzdDtRS|eKEou_)9p_jT%{F}BOhtavo{!tEc^RQji zptQ8KtCN#crEEQVZN4>81x7%Q?HRS!4OxQz;vw&2sa8ZxU=fkquhRC|HRJIPHH#VU z*&1kjF1S7!CtSRD@1B!jm7}}(S_?N?aXLF*Wo03mlt#dE@W_NzpZdNPwAiM%(_}HS zU@Nh$sPabp{@k(K^4_mb)7<5~=e2A)yVG7A_L-RFS4S?|=F|AS`?; zQ#)NiHYAVrdlEXoP))nfEUMLDhcWaY)Iy5S75OaSP|yP*&AZYFVycgq<)CH{j>5&f zD}JvJx2mj&ay^%?*G@!&hZxxY_Kt%|BTs9hSE75}Rep3^bv_-hADWr-`QR0HZ;C1_+t+33a22>_^LwvxCL% z(x!zjeI?Gw1Rv~TDjyb72P^lade z3==RN$e{6)Fuw;wE3`?L_o+J!T29+e9eTk&(8&;bzx6g)dUJARE(%dMF!1o8RYJ68 z{qk{PF0s$cG|hoIjVE-)n<(kd7DM|i)Fxm=)d^;OP$dlfyoObY3Qowt4L@0eX5uWc zQt#5El(9O>t8b$G{+gG3?0g}v4L~9cO>h@&b8~ZxEqDw3N-Ak3b&bUqdT-7YQ?T7j zeATa^8QRyeutLTe9Ul(4#;Rsyz=>*H6NffM2$l}o*M(I6JX5sjDhLWTNC(p~1=C^J z-wZ}EECYn6BD%WU$zo8_IdHLSvATO{cENTQ1pD+rCUxABv}!J+VfzTC<&I=c^6n^w z9xmfE$r3}Qk9{m-SHIuXK2g>DaohAiEr+tQX9laa%B|JG8#r1`5A=)z%2=Qe2Q~QM zK_kxu=o~ZO&p<0zi*-43^$KvLJzzH((0O%!1?o4r%yh8@a-V3=p4=1IaUVQ1{vS4U zv?kyL1Wq?vh>uf2W?-iC-g=aS_h2ZSSQ~ymyZaT0({)e1IVs(lrt1$A>e?bg~o^y*Mk$!TZv)OS*DZQrp>h!gVj)&@Kkl!E0$R8KBkbQ_8>L7+$xY8hLbRW3Y8IW6V z*K#s$C3ywtVqsk-J(wO|&93J!0p0|3r)Ng62my3K*K8#{!?UP*FUpW}Wmn8;(Dq#HBV&&tOKyy|ge_$& zhYyctge|nC`Tq7rU)2o|YZH%Pi-IPsoVXBQUfpl8!N%rf#aJ+oP0az;o1wyjn-az0{&u3zD)^rVnkn%z%1|5W*COZowA*dQzXete$d+@H57aDO8gEI!gR)p%=fF(XI!J5h zO!cOos3t8VnBB1fWBqcwR8|cR&^os-aes=sq{1zIR%B+pzhg=tp}DVu&X_W z^cxDKIdfuV3=6F>E}qFP}ZrAl+KBlyx_qbCJF0QQd6r+ zA1b_Zz^W#|9hP^JN+FIUKOoAz1}fC!5W)~8Fk!FZI$gXaAwwr&Cj|AM0uC{5ot~W? zVzkcad2kygmCqMiq%PEh2dR-jGCmBKP9D1{sq{(}vC*{N=i;9&N%C`szw)Faa`AG zZoelF?FF|bVa!YIGf@r?IOaY0XuKoY6$tO^RxLxpbi+$5;Z&y39Tqo3O4rv0F9QNG z)?e=mOD4hD1(^V%v+{b!G7Tr7sUUGUTv7e%G!pgQJOFd0Iv^euc>GOoWB1sSjrcPv zqYO$okOZnk`sAdW6k=@-aK_pINibmK95(rDRDud2q*<+pLMb0;jfsh=h8J(H=Yj}v zY*EB6zel#j`KGfZy7n&S94GEKQMmDa!WfW(=}XvT2P_&!4ycsUZnVhpljf^NW6veP z!V}2#T1DJ=`-48<^Rf&urx|s)Gp5Cg@d+D!$=fWeaUT5ku){R(4N%}fZ+apVVAak> zlS3Eq4uS(8VQf?3)h3BpZY;2^K7w2T>~t6utwXu@;A$yhesc|xX~%8{As4SACnwh? zE^Tlh*pjg=bNTxTa%QH=3PH50$ruH(t`0Nm>#A#UhR5Kd1Qf!VD(**jsVj=c5>Z`> z!GrODp~8SJ?2NWwdU6F%0q=vhRfjAc!UOXu9!9)svNNDfe%d^TfD4BUs_*19e~Hui z@C&F}U~z0tz?gg2&Q=VDjK*w}Mf0anxD=1_6j*wwhi8CQHFv!vkO;)z!b(oGLS_JP znt^(r4cC?MQT|jb*Ks>G#&V)FwbEgfexp@!1@_S(nvo9hY*pff90uhR11=drP1Os) zR~#39FjKgcdl^qD!EmP2{Mz@cV}ii)2fQi)7D&`{-QX?s2`hvDX^Q)^hY%okc!*B! TDKcAs_5^v#^!uW}INtg<4VB|G literal 12652 zcmeHt2Ut^CyY5z0MnM_J89_mr5=2y_NRh5%!GM6m4818$X+c`(MEw;V1ayZYCCM-t zG7cgN0!9eIf=sAlfPfG{2~8w{P(lbKXYF{}bDs16^WSsl{`a1H+|Q#cthT@PeeYM- zzJAKqa^r^W8z2bUh(2-D9)i}!K#XLNrc`+1u1C+!>b$$x2{|T$ zf!W8mad$ZVxwc$Ra*`m)Jj^H{iig3r35jE3(i$mQf8u|*)yyNy>=IN)t!SBmnmw}Y zCf!`6ki=mBYNBpr5hI6m)IY3HGLw$rWGl9tLAdYq+8xX#r!oeOjFWF0F*%Dn!RR8i z#uMkmc`p~6l!Wq9q75hQD;wEThSJIySOtb#kx~CVn4*>L`LLIp#2-!T*bDlHVm^;q zm0qZ>x{c;tff?K!111ms`k1jDdEp_Ai;775b7sAil2cYgYmvWqt4*&)+RQJI=@4YK zz7}a9@)}R1!P&3YLfTKuKd}zM!6(@K!&Rm;a1MfcjFa`sh{ieX8N>@-TZK6+V!k+f zP98|)B~bby6Vut4Y{Tf;_+|~4lSQoHD!Oeq!%JOi0&I5!%{Pcy>fl_DC~^v3$inP0 z6j+vlrQWE3f2i^!C&6uX^1LmXTfV#8HohaJ*W|=Pcaf7yDAppD^8?HoBnnewr}H*e zMQ#Q3C8%Fi95T0IRA3kIaskJ+dV?{^bC=>xZsV1)uw^@{Ib_8GHlpgp$$-iIvS4ye zn19ytkmd2Ro)1A7wRr6Dvd0#xw3`L&ZM1hQr7Xt0Tago(TTcs_taJcBz_QE#6YQa5 zkt_t=nCq+*itx)X#&-yO%*Ip)W0K+`<~9Kw#D7)`)n7L#n~h**(zgMBb8o9qh)TAI z1FIgG0Z)KKsV3_JWI5H+`$P<g?^c9W@h9SA=RN~2M-~N8snDu7)){`w-P|IM;m^XT848Xgkd4s> z@0>nWn%%b}1~VlGeAOYXe+-E4c8#xs>M!Sx`Q9=Wb|{XLcGCRt^O?;>hJY-U06E6_O6kzb6iuo72!Cff&u+s#?M8D&%Yc0DnD^$X&mvojrjm{4UWy6{x;C7S+Jl0 zfkO#m0_F|+!ho(JOYdz3!elxb%xoCdei5R!&y1?uaOAD_hd ze_o9Ln<)6_<%nk2Kc8Q;o^3c#J)2*IMZ_02x5o3}Xxl(j!~D0&x;6~GIKo4GNXWb} z#R*~|A@kQ4)_rt5ntzZ@yQb}gTJP5#OvZEzdpk@v9k7cvogM!QDp)gHaFUVXk$6U& z)!~M8&sUBWCo2J}p!7@ia;o(SH%Tp-T4@6pc3)7$DKPZZCXGn8{0U@oyRZg@prR2> zv>dV|o{X-VtH--x;WBWn)j0gSCLxC}?AU7{TB?Sz(Fnn~qVJ}kZFbh=ii3+rvbn2= zVq8!5iSllvm>`;g!d_G@P9gCN1f@uGg-`iE!i!I}kH22`3Dl1R97=E=HA0uiy_1{0 z^3=h=Rh^PuWC=o^BcP@<^x}50ua!Bq;&+pxYLIH32Kgy;@H7aNK=KOWr%6fvhjt_j zd-Xjs`>0Xn$^C>KAcCq6-6ax92cFqS(U1GWQ?b^c}Kz~5H} z`{+S`ZWbp9F6fLXlAajQjiOuM7HfDV%~w+}H$b6o`T^o5>#=@V30plv95{YcG@Fa^ z+HSTQ8ajh$8%~!0d=OtS@cPK$!Wd_-TVjX~RNsNH3nyD%aUwb!Nw?0MfdHdE#mAbC zK@tjR$n$?fVXhL?Mv>y>j(d1!Gm4>@Jk&m>wTT$i0;t%7R`S->;N)pfi=yCc3%iqa#j3n z%ZAHWxOtYHD3lt*chC0Mhf=0*YoGrOemn7K{h7Rz1lz(#YWRak8n}T!RB!wK{KaB} z+x!h0J%?`GTK??Wv#Rdb+?DwZTKQjkN<7Z<*b!y(!j^uc0D&nNt12)IpvQBKKyh&Q zT4`ikf6e}5P58#IGsQ?zq`rv#y+=WI9>HzUsVvx_H8(5IZBt$8@{dd-$a$M|k$lfX zVZ(V1VW~fUUs-0Y&h<-*-bZ&Mj?_EN{h=B)%!8>BCt7}5yafuY`VxWsgX1ro-pN2q z4gc)vk7r(zGRFI;#z+7A|87TT3=_%tao^a7k1Pz{)jDtdGI!sj@NuqTmX-lQ##>2x zaGYgl9!otok!jR+r}Ab7N=Pk^faGPQ?3%t-stju?>JfgaHmS>3Az+Nx)=!_3ay{d= zlJ%~X$C^reZPX#9{h{T^eJZ5=isELMcGv5Tga_&*VY>8JksF2*3K3%m9dGj41viES z8Xv=82T1)EEM(~Ic0^`>^Y#BiU|{WpbZW+~#;x!b+wzsGOLfH+`vSfZ=3wb$)1Jsp zL;W8`yfohPefcT;%)?sp#6DW&`?KrK9Qk#q4W3G8HXtg7Gj}u0Hl5@bJIqBdo(pe$ zTkHU>4E?WDaP(E@j!R+(sVE8p(s%OOG84jk1_#V^a-!kh;)pq29b%aZPmgdmR+*)aRVPoV%#ozX z9WtH&E+7B!n~<0>McCx(QvW-=f!9s<@WTN4H9_avlZ#b9vS{q?#z}bQ8_gJ1BF6Y> zQz=GGnWYbqFazJHN53p6`X+HbqIJ*%mH()Ze@#K5`iW5!#@a1tICmr6N!ga!H*=+U zOIY2&K4Alwa^D551^S%^1B>qBso^oY=Ba2DZ# z8j}oX0`Y+-Y#=`7%tMYQgJ(b>=j2)bWVM#(Vn1ED>Vi51ZI*ruu<7?`;&1;==%9b*MiV+j2B~E{3Twbmw6>Up4i4AajZH`*Rx# z4&R-juF!5sSbeUGYl#zXSAt~-V&TuWu1)^&t2bgKLWJtHOOK#XkRXny7FgsJsibV=qi z8}9cE;)WXAX({Yz&md>Xy6JmijIFiwg}2HVGX?zM!*yb zCt||){=`9R4Ye)$v7v!tFIbe<#%z8lbSJ>i)r7~HE^&K$Z{;yiUgwOhw8*}0Hbgkj z4k5v{qVN&_gX?SqoIsy2HOq2$gE5_Xm%+{>Plcv6k91jyi3;R53=xJ><~Uk4JBzwu3&t;M}L03;|j1Vfj8-Xvz zRsI;X5~eYtm9fKQq1q;3mG@Mpz4({zrBoju9};rRKYGS~DQIn)iWXT~YhX$%ZPkoi zt%P=^j*}jug(=ZMulWTF!PHHMB7`~HPPVWk&`3%-2RtroFp^^i? z`z%D)th_*Zl~kHXMMp2(@1m^z*=rrU_d`E5qH%F?(T?R>wuhW>VY0iFS3C1kIYZzx z7-k+=8rg97?p?1Fe6lTNHaDcN%EQ3Oh+5SB<;RVS%k{TtXvdSb) zhL&0{>}y>*UnwUi=QUUv#7dRDrHYF{2vDSr9~(w*Lb+5wco62OPJpzXFRMr>#l`mv z^75WU1qTNwQ<&6PfB#WRF@+-7gFO@F7^&SATAi%N$btJ09yl=j#Gx8f-RFf4@2Cpt zBR+x^eLtQ)b~)J!w(D~Bq9uwJ+!R#J`;Fq}#`VdW7PkCOR0O)JrKQDG?^@YfT$b{j zPJ9hJ$qx29Q680)BwjDONo8^2y@SjHS3phQ)IuLztIiOEqtfhoGAk9o9Eb@jHQa@! zFk^WNTEAiJ%BNlx9+8|Ym9jl{S<-6Xl+oVSmNfcO8DpBnv@z6dX=@Xc0KW=Qa_C*w z8d@KCPEb8niK8Q$XL(6#?VDD}nTaK>>klPw3|vNdMF0IPRTMR{e{QgREia1YfjzG= z@&?toz#EjLslGW}GC$!vqadj4SA;Hiwd=&U6V0}eGy5nM3eP;TSG*!pP#2znBaI2_ zWyS}g5?#P+b`2IY9%;;by!5`rPHPi%uIJ~M5`X|aIg`c?!S2BpO0qR$8XFr)IjG8R zH+&~EM-^x}rl!I!c4V&?)vt8nBdP}K5=GAdvqKL)-{UoI%l4K1`OIZge^P$FNI=eH zaoIK16-#%ZIp)_dC1&vpCs9T->1hgWfFGr*E*DiF1wVBQqU}OkX<+wNE^i?%zDU?w zmVbM=F@xt^lDJCSc}1l{z-9LtF_Yh0Yve>xNJeAz&|JgM>AgS{?b-IBuJ73@kRImq zVae0WvdMcGFS#ka$s;Gd*AF+Bp0bJ!hdW*A5m&poC&Q0v_=#MnL^<6{h&LrWq=wn_9NQB#@1H+ zhRT(}i6;b$^mqaHq3dLdwFbSPat)F{L~oSP;ssT3?PJl~DRO3f8Fe(zi&#*QkU(Jy zWzk~K()sM@nkgRSO!jLZvf-b?8b@D^>Q+N*j9k=5^;{+}zxhY&S*R@%gA_2|tHkD}GTISW4G3F%cnSTT(4Vpm$ja z%*FzWF8R%g_P*VGsUu^jsbD-T$&H~OgW=g4lFn47!gsm?HDd|chd&^EOx1}G4dv3v znTC&hSJzLEk1bm0>(399Gp~bP-IkO#<5rCeW+XBSrqjP~)CXvyM=kMJM1sN7Rzeaw ze0gyuhZN;Kw0W=5Wx)$VE9s3-EV^7D0mJ>bb&%4%i|M^J(Nl3DrSlG|qMkOFUc}^)31(*tj>e+Jv0h3nb%2@bQNn%gP!hKyzarvqOR8p4m6S z0c%;Vk|0>I)3xkiO6?(*)%~aO@6tz|v?>z<0|Sp>7u=*4Q;Lx7oZDvLB<1!^$0{l+ zB!2qaR*lnF-H?=`wpg3mDp~8E4)S0y)FJub0Y)tHP_whM)>|KwGeu9B@$KL43Cr`7 z8vvx#;1hHien4N@Wb1q3e%K|hR7drH)u{nw88RXQ{Qb=zbn-oFqRqJXF0Fp$lxQxN zWC7-tcYepRo==&rhQ9Px#HHTn=_3R3)yl*`A2%2lI z&*-(*AfR$dB;LO2{ZLSG*|218ZmtyA#gfW8Az3SN;tFMdA3&%xOtRwxx^nv*bCuhI z{K$K`Zz?A%$(fRyoJu?%uZo7LR4TXyDzv3U%mb4{k6bdS!Sa$?`uqF6P`&O}v4;ip z!()%MpuCQ|52sA~x|9QmxfM%O_rfQ{Zj9LHhk4#EU8nuXGn>=1EnDK4XL1=!4F(q+ zyOECqUfoew&kZp?s$kbP?GHYA0vRp@<;~tfk*q=9c~eb3!T@~M^cN81;C8uZae~l0 z`&MGUP0jiK;LPK&+$paD&L~O&C8VR1cBLBnn&2twLPK<2F6SvqvX{|5l49Lo(OPAe z%9^0*Gj4H*+iP-FFwzV%!&Rcw?Byls44STj6CR6ZCNjKu4-?V+HdJzE!|N*K;E~uG zx>fj$k$?XC`xROYGCR?3m#~XFW|2LT8h{)fcPsLjdo-T>WWnwjo*h30outmF+P9VB zR^j}LxI5Nx-@_FXe1fTk`y*SFQ9qm8y*%Ap)$G?7(==NbetkrM&GG1)-3yNWBR^XQ zg03F+cxs#vFIS^5w&4$IS_@HSCVqqx+dh9wxLRG9rbZl&*=pU7qV8f0S~LeMs0n?s zORZ|zz*8h1&X@uc=t6Qj`O9q^Sbfzm#_GbFmeZNpv>FpyX`|RYwrP3Zta~CW%h7xq;<2q>G8@|m{0W5ha)R(=gLwIZh9BspE=vcE3^UPi-ZyJB6o%8mHWd;y@P zed<1_=hSOEw{EMYwGedaQD2FS9I99-sv08z|0tpy;2fuq6lT1YQSe31z|YPnapN ziuojf{3}^00O!21PxLPv8LW&^K1i??F$W%z-AO=G1;k@jW04d^6m|NMUn$Bda*Qol z4dE`HmO{CVs4G#ZVV0tkXm%pWFbW$vCOHqwprbjQ3)5tmeHRC_@qW`*6@(|t8F27; zo5g;HEUdyyO0d~#JQ6T zcksyOZWE45>_CbL6!Z^_yT(bmJm57;6Yp6 jhb<9ELGSrG diff --git a/docs/reference/model-method-sample.html b/docs/reference/model-method-sample.html index d84264098..f656c23a7 100644 --- a/docs/reference/model-method-sample.html +++ b/docs/reference/model-method-sample.html @@ -1,82 +1,19 @@ - - - - - - - -Run Stan's MCMC algorithms — model-method-sample • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run Stan's MCMC algorithms — model-method-sample • cmdstanr - - - - - - - - - - - - + + - - -
-
- -
- -
+

Run Stan's MCMC algorithms

- Source: R/model.R + Source: R/model.R
-

The $sample() method of a CmdStanModel object runs Stan's +

The $sample() method of a CmdStanModel object runs Stan's main Markov chain Monte Carlo algorithm.

Any argument left as NULL will default to the default value used by the installed version of CmdStan. See the -CmdStan User’s Guide +CmdStan User’s Guide for more details.

After model fitting any diagnostics specified via the diagnostics argument will be checked and warnings will be printed if warranted.

- 
sample(
-  data = NULL,
-  seed = NULL,
-  refresh = NULL,
-  init = NULL,
-  save_latent_dynamics = FALSE,
-  output_dir = NULL,
-  output_basename = NULL,
-  sig_figs = NULL,
-  chains = 4,
-  parallel_chains = getOption("mc.cores", 1),
-  chain_ids = seq_len(chains),
-  threads_per_chain = NULL,
-  opencl_ids = NULL,
-  iter_warmup = NULL,
-  iter_sampling = NULL,
-  save_warmup = FALSE,
-  thin = NULL,
-  max_treedepth = NULL,
-  adapt_engaged = TRUE,
-  adapt_delta = NULL,
-  step_size = NULL,
-  metric = NULL,
-  metric_file = NULL,
-  inv_metric = NULL,
-  init_buffer = NULL,
-  term_buffer = NULL,
-  window = NULL,
-  fixed_param = FALSE,
-  show_messages = TRUE,
-  diagnostics = c("divergences", "treedepth", "ebfmi"),
-  cores = NULL,
-  num_cores = NULL,
-  num_chains = NULL,
-  num_warmup = NULL,
-  num_samples = NULL,
-  validate_csv = NULL,
-  save_extra_diagnostics = NULL,
-  max_depth = NULL,
-  stepsize = NULL
-)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
data

(multiple options) The data to use for the variables specified in -the data block of the Stan program. One of the following:

    -

  • A named list of R objects with the names corresponding to variables +

    + 
    sample(
+  data = NULL,
+  seed = NULL,
+  refresh = NULL,
+  init = NULL,
+  save_latent_dynamics = FALSE,
+  output_dir = NULL,
+  output_basename = NULL,
+  sig_figs = NULL,
+  chains = 4,
+  parallel_chains = getOption("mc.cores", 1),
+  chain_ids = seq_len(chains),
+  threads_per_chain = NULL,
+  opencl_ids = NULL,
+  iter_warmup = NULL,
+  iter_sampling = NULL,
+  save_warmup = FALSE,
+  thin = NULL,
+  max_treedepth = NULL,
+  adapt_engaged = TRUE,
+  adapt_delta = NULL,
+  step_size = NULL,
+  metric = NULL,
+  metric_file = NULL,
+  inv_metric = NULL,
+  init_buffer = NULL,
+  term_buffer = NULL,
+  window = NULL,
+  fixed_param = FALSE,
+  show_messages = TRUE,
+  show_exceptions = TRUE,
+  diagnostics = c("divergences", "treedepth", "ebfmi"),
+  cores = NULL,
+  num_cores = NULL,
+  num_chains = NULL,
+  num_warmup = NULL,
+  num_samples = NULL,
+  validate_csv = NULL,
+  save_extra_diagnostics = NULL,
+  max_depth = NULL,
+  stepsize = NULL
+)
    +
    + +
    +

    Arguments

    +
    data
    +

    (multiple options) The data to use for the variables specified in +the data block of the Stan program. One of the following:

    • A named list of R objects with the names corresponding to variables declared in the data block of the Stan program. Internally this list is then -written to JSON for CmdStan using write_stan_json(). See -write_stan_json() for details on the conversions performed on R objects +written to JSON for CmdStan using write_stan_json(). See +write_stan_json() for details on the conversions performed on R objects before they are passed to Stan.

    • A path to a data file compatible with CmdStan (JSON or R dump). See the appendices in the CmdStan guide for details on using these formats.

    • NULL or an empty list if the Stan program has no data block.

      • -
seed

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. + + + +

seed
+

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to @@ -262,25 +190,24 @@

Arg chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each -chain.

refresh

(non-negative integer) The number of iterations between +chain.

+ + +
refresh
+

(non-negative integer) The number of iterations between printed screen updates. If refresh = 0, only error messages will be -printed.

init

(multiple options) The initialization method to use for the +printed.

+ + +
init
+

(multiple options) The initialization method to use for the variables declared in the parameters block of the Stan program. One of -the following:

    -

  • A real number x>0. This initializes all parameters randomly between +the following:

    • A real number x>0. This initializes all parameters randomly between [-x,x] on the unconstrained parameter space.;

    • The number 0. This initializes all parameters to 0;

    • A character vector of paths (one per chain) to JSON or Rdump files containing initial values for all or some parameters. See -write_stan_json() to write R objects to JSON files compatible with +write_stan_json() to write R objects to JSON files compatible with CmdStan.

    • A list of lists containing initial values for all or some parameters. For MCMC the list should contain a sublist for each chain. For optimization and @@ -293,601 +220,628 @@

      Arg has argument chain_id it will be supplied with the chain id (from 1 to number of chains) when called to generate the initial values. See Examples.

      • -
save_latent_dynamics

(logical) Should auxiliary diagnostic information + + + +

save_latent_dynamics
+

(logical) Should auxiliary diagnostic information about the latent dynamics be written to temporary diagnostic CSV files? This argument replaces CmdStan's diagnostic_file argument and the content written to CSV is controlled by the user's CmdStan installation and not CmdStanR (for some algorithms no content may be written). The default is FALSE, which is appropriate for almost every use case. To save the temporary files created when save_latent_dynamics=TRUE see the -$save_latent_dynamics_files() -method.

output_dir

(string) A path to a directory where CmdStan should write +$save_latent_dynamics_files() +method.

+ + +
output_dir
+

(string) A path to a directory where CmdStan should write its output CSV files. For interactive use this can typically be left at NULL (temporary directory) since CmdStanR makes the CmdStan output (posterior draws and diagnostics) available in R via methods of the fitted -model objects. The behavior of output_dir is as follows:

    -

  • If NULL (the default), then the CSV files are written to a temporary +model objects. The behavior of output_dir is as follows:

    • If NULL (the default), then the CSV files are written to a temporary directory and only saved permanently if the user calls one of the $save_* methods of the fitted model object (e.g., -$save_output_files()). These temporary +$save_output_files()). These temporary files are removed when the fitted model object is -garbage collected (manually or automatically).

      • +garbage collected (manually or automatically).

    • If a path, then the files are created in output_dir with names corresponding to the defaults used by $save_output_files().

      • -
output_basename

(string) A string to use as a prefix for the names of + + + +

output_basename
+

(string) A string to use as a prefix for the names of the output CSV files of CmdStan. If NULL (the default), the basename of the output CSV files will be comprised from the model name, timestamp, and -5 random characters.

sig_figs

(positive integer) The number of significant figures used +5 random characters.

+ + +
sig_figs
+

(positive integer) The number of significant figures used when storing the output values. By default, CmdStan represent the output values with 6 significant figures. The upper limit for sig_figs is 18. Increasing this value will result in larger output CSV files and thus an -increased usage of disk space.

chains

(positive integer) The number of Markov chains to run. The -default is 4.

parallel_chains

(positive integer) The maximum number of MCMC chains +increased usage of disk space.

+ + +
chains
+

(positive integer) The number of Markov chains to run. The +default is 4.

+ + +
parallel_chains
+

(positive integer) The maximum number of MCMC chains to run in parallel. If parallel_chains is not specified then the default is to look for the option "mc.cores", which can be set for an entire R -session by options(mc.cores=value). If the "mc.cores" option has not -been set then the default is 1.

chain_ids

(integer vector) A vector of chain IDs. Must contain as many +session by options(mc.cores=value). If the "mc.cores" option has not +been set then the default is 1.

+ + +
chain_ids
+

(integer vector) A vector of chain IDs. Must contain as many unique positive integers as the number of chains. If not set, the default -chain IDs are used (integers starting from 1).

threads_per_chain

(positive integer) If the model was -compiled with threading support, the number of +chain IDs are used (integers starting from 1).

+ + +
threads_per_chain
+

(positive integer) If the model was +compiled with threading support, the number of threads to use in parallelized sections within an MCMC chain (e.g., when using the Stan functions reduce_sum() or map_rect()). This is in contrast with parallel_chains, which specifies the number of chains to run in parallel. The actual number of CPU cores used is parallel_chains*threads_per_chain. For an example of using threading see the Stan case study -Reduce Sum: A Minimal Example.

opencl_ids

(integer vector of length 2) The platform and +Reduce Sum: A Minimal Example.

+ + +
opencl_ids
+

(integer vector of length 2) The platform and device IDs of the OpenCL device to use for fitting. The model must be compiled with cpp_options = list(stan_opencl = TRUE) for this -argument to have an effect.

iter_warmup

(positive integer) The number of warmup iterations to run +argument to have an effect.

+ + +
iter_warmup
+

(positive integer) The number of warmup iterations to run per chain. Note: in the CmdStan User's Guide this is referred to as -num_warmup.

iter_sampling

(positive integer) The number of post-warmup iterations +num_warmup.

+ + +
iter_sampling
+

(positive integer) The number of post-warmup iterations to run per chain. Note: in the CmdStan User's Guide this is referred to as -num_samples.

save_warmup

(logical) Should warmup iterations be saved? The default -is FALSE.

thin

(positive integer) The period between saved samples. This should -typically be left at its default (no thinning) unless memory is a problem.

max_treedepth

(positive integer) The maximum allowed tree depth for +num_samples.

+ + +
save_warmup
+

(logical) Should warmup iterations be saved? The default +is FALSE.

+ + +
thin
+

(positive integer) The period between saved samples. This should +typically be left at its default (no thinning) unless memory is a problem.

+ + +
max_treedepth
+

(positive integer) The maximum allowed tree depth for the NUTS engine. See the Tree Depth section of the CmdStan User's Guide -for more details.

adapt_engaged

(logical) Do warmup adaptation? The default is TRUE. +for more details.

+ + +
adapt_engaged
+

(logical) Do warmup adaptation? The default is TRUE. If a precomputed inverse metric is specified via the inv_metric argument (or metric_file) then, if adapt_engaged=TRUE, Stan will use the provided inverse metric just as an initial guess during adaptation. To turn off adaptation when using a precomputed inverse metric set -adapt_engaged=FALSE.

adapt_delta

(real in (0,1)) The adaptation target acceptance -statistic.

step_size

(positive real) The initial step size for the discrete +adapt_engaged=FALSE.

+ + +
adapt_delta
+

(real in (0,1)) The adaptation target acceptance +statistic.

+ + +
step_size
+

(positive real) The initial step size for the discrete approximation to continuous Hamiltonian dynamics. This is further tuned -during warmup.

metric

(string) One of "diag_e", "dense_e", or "unit_e", +during warmup.

+ + +
metric
+

(string) One of "diag_e", "dense_e", or "unit_e", specifying the geometry of the base manifold. See the Euclidean Metric section of the CmdStan User's Guide for more details. To specify a -precomputed (inverse) metric, see the inv_metric argument below.

metric_file

(character vector) The paths to JSON or +precomputed (inverse) metric, see the inv_metric argument below.

+ + +
metric_file
+

(character vector) The paths to JSON or Rdump files (one per chain) compatible with CmdStan that contain precomputed inverse metrics. The metric_file argument is inherited from CmdStan but is confusing in that the entry in JSON or Rdump file(s) must be named inv_metric, referring to the inverse metric. We recommend instead using CmdStanR's inv_metric argument (see below) to specify an inverse -metric directly using a vector or matrix from your R session.

inv_metric

(vector, matrix) A vector (if metric='diag_e') or a +metric directly using a vector or matrix from your R session.

+ + +
inv_metric
+

(vector, matrix) A vector (if metric='diag_e') or a matrix (if metric='dense_e') for initializing the inverse metric. This can be used as an alternative to the metric_file argument. A vector is interpreted as a diagonal metric. The inverse metric is usually set to an estimate of the posterior covariance. See the adapt_engaged argument above for details about (and control over) how specifying a precomputed -inverse metric interacts with adaptation.

init_buffer

(nonnegative integer) Width of initial fast timestep -adaptation interval during warmup.

term_buffer

(nonnegative integer) Width of final fast timestep -adaptation interval during warmup.

window

(nonnegative integer) Initial width of slow timestep/metric -adaptation interval.

fixed_param

(logical) When TRUE, call CmdStan with argument +inverse metric interacts with adaptation.

+ + +
init_buffer
+

(nonnegative integer) Width of initial fast timestep +adaptation interval during warmup.

+ + +
term_buffer
+

(nonnegative integer) Width of final fast timestep +adaptation interval during warmup.

+ + +
window
+

(nonnegative integer) Initial width of slow timestep/metric +adaptation interval.

+ + +
fixed_param
+

(logical) When TRUE, call CmdStan with argument "algorithm=fixed_param". The default is FALSE. The fixed parameter sampler generates a new sample without changing the current state of the Markov chain; only generated quantities may change. This can be useful when, for example, trying to generate pseudo-data using the generated quantities block. If the parameters block is empty then using fixed_param=TRUE is mandatory. When fixed_param=TRUE the chains and -parallel_chains arguments will be set to 1.

show_messages

(logical) When TRUE (the default), prints all +parallel_chains arguments will be set to 1.

+ + +
show_messages
+

(logical) When TRUE (the default), prints all +output during the sampling process, such as iteration numbers and elapsed times. +If the output is silenced then the $output() method of +the resulting fit object can be used to display the silenced messages.

+ + +
show_exceptions
+

(logical) When TRUE (the default), prints all informational messages, for example rejection of the current proposal. Disable if you wish to silence these messages, but this is not usually recommended unless you are very confident that the model is correct up to numerical error. If the messages are silenced then the -$output() method of the resulting fit object can be -used to display the silenced messages.

diagnostics

(character vector) The diagnostics to automatically check +$output() method of the resulting fit object can be +used to display the silenced messages.

+ + +
diagnostics
+

(character vector) The diagnostics to automatically check and warn about after sampling. Setting this to an empty string "" or NULL can be used to prevent CmdStanR from automatically reading in the sampler diagnostics from CSV if you wish to manually read in the results -and validate them yourself, for example using read_cmdstan_csv(). The +and validate them yourself, for example using read_cmdstan_csv(). The currently available diagnostics are "divergences", "treedepth", and "ebfmi" (the default is to check all of them).

These diagnostics are also available after fitting. The -$sampler_diagnostics() method provides +$sampler_diagnostics() method provides access the diagnostic values for each iteration and the -$diagnostic_summary() method provides +$diagnostic_summary() method provides summaries of the diagnostics and can regenerate the warning messages.

Diagnostics like R-hat and effective sample size are not currently available via the diagnostics argument but can be checked after fitting -using the $summary() method.

cores, num_cores, num_chains, num_warmup, num_samples, save_extra_diagnostics, max_depth, stepsize, validate_csv

Deprecated and will be removed in a future release.

+using the $summary() method.

+ -

Value

+
cores, num_cores, num_chains, num_warmup, num_samples, save_extra_diagnostics, max_depth, stepsize, validate_csv
+

Deprecated and will be removed in a future release.

-

A CmdStanMCMC object.

-

See also

+
+
+

Value

+ -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

A CmdStanMCMC object.

+
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-variables, -model-method-variational

- -

Examples

- 
# \dontrun{
-library(cmdstanr)
-library(posterior)
-library(bayesplot)
-color_scheme_set("brightblue")
-
-# Set path to CmdStan
-# (Note: if you installed CmdStan via install_cmdstan() with default settings
-# then setting the path is unnecessary but the default below should still work.
-# Otherwise use the `path` argument to specify the location of your
-# CmdStan installation.)
-set_cmdstan_path(path = NULL)
-
#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.29.1

-# Create a CmdStanModel object from a Stan program,
-# here using the example model that comes with CmdStan
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-mod <- cmdstan_model(file)
-mod$print()
-
#> data {
-#>   int<lower=0> N;
-#>   array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
-#>   y ~ bernoulli(theta);
-#> }

-# Data as a named list (like RStan)
-stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-
-# Run MCMC using the 'sample' method
-fit_mcmc <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  parallel_chains = 2
-)
-
#> Running MCMC with 2 parallel chains...
-#> 
-#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.2 seconds.
-#> 

-# Use 'posterior' package for summaries
-fit_mcmc$summary()
-
#> # A tibble: 2 × 10
-#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
-#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.

-# Get posterior draws
-draws <- fit_mcmc$draws()
-print(draws)
-
#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 -6.8 -6.8
-#>         2 -6.9 -6.8
-#>         3 -7.0 -7.0
-#>         4 -6.9 -7.1
-#>         5 -6.7 -7.0
-#> 
-#> , , variable = theta
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 0.28 0.21
-#>         2 0.19 0.20
-#>         3 0.16 0.17
-#>         4 0.20 0.36
-#>         5 0.25 0.34
-#> 
-#> # ... with 995 more iterations

-# Convert to data frame using posterior::as_draws_df
-as_draws_df(draws)
-
#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
-#>    lp__ theta
-#> 1  -6.8  0.28
-#> 2  -6.9  0.19
-#> 3  -7.0  0.16
-#> 4  -6.9  0.20
-#> 5  -6.7  0.25
-#> 6  -7.1  0.36
-#> 7  -9.0  0.55
-#> 8  -7.2  0.15
-#> 9  -6.8  0.23
-#> 10 -7.5  0.42
-#> # ... with 1990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# Plot posterior using bayesplot (ggplot2)
-mcmc_hist(fit_mcmc$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-# Call CmdStan's diagnose and stansummary utilities
-fit_mcmc$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181226-1-05e2b0.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181226-2-05e2b0.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit_mcmc$cmdstan_summary()
-
#> Inference for Stan model: bernoulli_model
-#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
-#> 
-#> Warmup took (0.0050, 0.0050) seconds, 0.010 seconds total
-#> Sampling took (0.015, 0.014) seconds, 0.029 seconds total
-#> 
-#>                 Mean     MCSE  StdDev     5%   50%   95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8      781    26932      1.0
-#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0  2.3e+02  8.1e+03  1.0e+00
-#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0  1.0e+00  3.5e+01  2.0e+13
-#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0  1.9e+03  6.5e+04  1.0e+00
-#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0  8.9e+01  3.1e+03  1.0e+00
-#> divergent__     0.00      nan    0.00   0.00  0.00  0.00      nan      nan      nan
-#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6  7.7e+02  2.7e+04  1.0e+00
-#> 
-#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47      796    27460      1.0
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).

-# For models fit using MCMC, if you like working with RStan's stanfit objects
-# then you can create one with rstan::read_stan_csv()
-
-# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
-
-
-# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
-# and also demonstrate specifying data as a path to a file instead of a list
-my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
-fit_optim <- mod$optimize(data = my_data_file, seed = 123)
-
#> Initial log joint probability = -9.51104 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.

-fit_optim$summary()
-
#> # A tibble: 2 × 2
-#>   variable estimate
-#>   <chr>       <dbl>
-#> 1 lp__        -5.00
-#> 2 theta        0.2 

-
-# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
-fit_vb <- mod$variational(data = stan_data, seed = 123)
-
#> ------------------------------------------------------------ 
-#> EXPERIMENTAL ALGORITHM: 
-#>   This procedure has not been thoroughly tested and may be unstable 
-#>   or buggy. The interface is subject to change. 
-#> ------------------------------------------------------------ 
-#> Gradient evaluation took 8e-06 seconds 
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.08 seconds. 
-#> Adjust your expectations accordingly! 
-#> Begin eta adaptation. 
-#> Iteration:   1 / 250 [  0%]  (Adaptation) 
-#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
-#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
-#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
-#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
-#> Success! Found best value [eta = 1] earlier than expected. 
-#> Begin stochastic gradient ascent. 
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
-#>    100           -6.262             1.000            1.000 
-#>    200           -6.263             0.500            1.000 
-#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
-#> Drawing a sample of size 1000 from the approximate posterior...  
-#> COMPLETED. 
-#> Finished in  0.1 seconds.

-fit_vb$summary()
-
#> # A tibble: 3 × 7
-#>   variable      mean median    sd   mad     q5      q95
-#>   <chr>        <dbl>  <dbl> <dbl> <dbl>  <dbl>    <dbl>
-#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
-#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
-#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  

-# Plot approximate posterior using bayesplot
-mcmc_hist(fit_vb$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-
-# Specifying initial values as a function
-fit_mcmc_w_init_fun <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function() list(theta = runif(1))
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2 <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function(chain_id) {
-    # silly but demonstrates optional use of chain_id
-    list(theta = 1 / (chain_id + 1))
-  }
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.5
-#> 
-#> 
-#> [[2]]
-#> [[2]]$theta
-#> [1] 0.3333333
-#> 
-#> 

-# Specifying initial values as a list of lists
-fit_mcmc_w_init_list <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = list(
-    list(theta = 0.75), # chain 1
-    list(theta = 0.25)  # chain 2
-  )
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_optim_w_init_list <- mod$optimize(
-  data = stan_data,
-  seed = 123,
-  init = list(
-    list(theta = 0.75)
-  )
-)
-
#> Initial log joint probability = -11.6657 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.3 seconds.
fit_optim_w_init_list$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.75
-#> 
-#> 
# }
-
-
+

The Stan and CmdStan documentation:

Other CmdStanModel methods: +model-method-check_syntax, +model-method-compile, +model-method-diagnose, +model-method-expose_functions, +model-method-format, +model-method-generate-quantities, +model-method-optimize, +model-method-sample_mpi, +model-method-variables, +model-method-variational

+
+ +
+

Examples

+ 
# \dontrun{
+library(cmdstanr)
+library(posterior)
+library(bayesplot)
+color_scheme_set("brightblue")
+
+# Set path to CmdStan
+# (Note: if you installed CmdStan via install_cmdstan() with default settings
+# then setting the path is unnecessary but the default below should still work.
+# Otherwise use the `path` argument to specify the location of your
+# CmdStan installation.)
+set_cmdstan_path(path = NULL)
+#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.32.2
+
+# Create a CmdStanModel object from a Stan program,
+# here using the example model that comes with CmdStan
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+mod <- cmdstan_model(file)
+mod$print()
+#> data {
+#>   int<lower=0> N;
+#>   array[N] int<lower=0,upper=1> y;
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
+#>   y ~ bernoulli(theta);
+#> }
+
+# Data as a named list (like RStan)
+stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
+# Run MCMC using the 'sample' method
+fit_mcmc <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  parallel_chains = 2
+)
+#> Running MCMC with 2 parallel chains...
+#> 
+#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.2 seconds.
+#> 
+
+# Use 'posterior' package for summaries
+fit_mcmc$summary()
+#> # A tibble: 2 × 10
+#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
+#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.
+
+# Get posterior draws
+draws <- fit_mcmc$draws()
+print(draws)
+#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 -6.8 -6.8
+#>         2 -6.9 -6.8
+#>         3 -7.0 -7.0
+#>         4 -6.9 -7.1
+#>         5 -6.7 -7.0
+#> 
+#> , , variable = theta
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 0.28 0.21
+#>         2 0.19 0.20
+#>         3 0.16 0.17
+#>         4 0.20 0.36
+#>         5 0.25 0.34
+#> 
+#> # ... with 995 more iterations
+
+# Convert to data frame using posterior::as_draws_df
+as_draws_df(draws)
+#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
+#>    lp__ theta
+#> 1  -6.8  0.28
+#> 2  -6.9  0.19
+#> 3  -7.0  0.16
+#> 4  -6.9  0.20
+#> 5  -6.7  0.25
+#> 6  -7.1  0.36
+#> 7  -9.0  0.55
+#> 8  -7.2  0.15
+#> 9  -6.8  0.23
+#> 10 -7.5  0.42
+#> # ... with 1990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# Plot posterior using bayesplot (ggplot2)
+mcmc_hist(fit_mcmc$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+# Call CmdStan's diagnose and stansummary utilities
+fit_mcmc$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251438-1-239737.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251438-2-239737.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit_mcmc$cmdstan_summary()
+#> Inference for Stan model: bernoulli_model
+#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
+#> 
+#> Warmup took (0.0040, 0.0040) seconds, 0.0080 seconds total
+#> Sampling took (0.011, 0.011) seconds, 0.022 seconds total
+#> 
+#>                 Mean     MCSE  StdDev     5%   50%   95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8    781    35502      1.0
+#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0    235    10662  1.0e+00
+#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0    1.0       46  2.0e+13
+#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0   1874    85179  1.0e+00
+#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0     89     4050  1.0e+00
+#> divergent__     0.00      nan    0.00   0.00  0.00  0.00    nan      nan      nan
+#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6    775    35215  1.0e+00
+#> 
+#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47    796    36197      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+
+# For models fit using MCMC, if you like working with RStan's stanfit objects
+# then you can create one with rstan::read_stan_csv()
+
+# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
+
+# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
+# and also demonstrate specifying data as a path to a file instead of a list
+my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
+fit_optim <- mod$optimize(data = my_data_file, seed = 123)
+#> Initial log joint probability = -9.51104 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+
+fit_optim$summary()
+#> # A tibble: 2 × 2
+#>   variable estimate
+#>   <chr>       <num>
+#> 1 lp__        -5.00
+#> 2 theta        0.2 
+
+
+# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
+fit_vb <- mod$variational(data = stan_data, seed = 123)
+#> ------------------------------------------------------------ 
+#> EXPERIMENTAL ALGORITHM: 
+#>   This procedure has not been thoroughly tested and may be unstable 
+#>   or buggy. The interface is subject to change. 
+#> ------------------------------------------------------------ 
+#> Gradient evaluation took 9e-06 seconds 
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.09 seconds. 
+#> Adjust your expectations accordingly! 
+#> Begin eta adaptation. 
+#> Iteration:   1 / 250 [  0%]  (Adaptation) 
+#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
+#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
+#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
+#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
+#> Success! Found best value [eta = 1] earlier than expected. 
+#> Begin stochastic gradient ascent. 
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
+#>    100           -6.262             1.000            1.000 
+#>    200           -6.263             0.500            1.000 
+#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
+#> Drawing a sample of size 1000 from the approximate posterior...  
+#> COMPLETED. 
+#> Finished in  0.1 seconds.
+
+fit_vb$summary()
+#> # A tibble: 3 × 7
+#>   variable      mean median    sd   mad     q5      q95
+#>   <chr>        <num>  <num> <num> <num>  <num>    <num>
+#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
+#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
+#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  
+
+# Plot approximate posterior using bayesplot
+mcmc_hist(fit_vb$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+
+# Specifying initial values as a function
+fit_mcmc_w_init_fun <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function() list(theta = runif(1))
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2 <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function(chain_id) {
+    # silly but demonstrates optional use of chain_id
+    list(theta = 1 / (chain_id + 1))
+  }
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.5
+#> 
+#> 
+#> [[2]]
+#> [[2]]$theta
+#> [1] 0.3333333
+#> 
+#> 
+
+# Specifying initial values as a list of lists
+fit_mcmc_w_init_list <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = list(
+    list(theta = 0.75), # chain 1
+    list(theta = 0.25)  # chain 2
+  )
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_optim_w_init_list <- mod$optimize(
+  data = stan_data,
+  seed = 123,
+  init = list(
+    list(theta = 0.75)
+  )
+)
+#> Initial log joint probability = -11.6657 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+fit_optim_w_init_list$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.75
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-sample_mpi.html b/docs/reference/model-method-sample_mpi.html index 228dde82b..53d2eb1a3 100644 --- a/docs/reference/model-method-sample_mpi.html +++ b/docs/reference/model-method-sample_mpi.html @@ -1,53 +1,5 @@ - - - - - - - -Run Stan's MCMC algorithms with MPI — model-method-sample_mpi • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run Stan's MCMC algorithms with MPI — model-method-sample_mpi • cmdstanr - - + + - - -
-
- -
- -
+

Run Stan's MCMC algorithms with MPI

- Source: R/model.R + Source: R/model.R
-

The $sample_mpi() method of a CmdStanModel object is +

The $sample_mpi() method of a CmdStanModel object is identical to the $sample() method but with support for -MPI. The target audience for MPI are +MPI. The target audience for MPI are those with large computer clusters. For other users, the -$sample() method provides both parallelization of +$sample() method provides both parallelization of chains and threading support for within-chain parallelization.

In order to use MPI with Stan, an MPI implementation must be installed. For Unix systems the most commonly used implementations are MPICH and OpenMPI. The implementations provide an MPI C++ compiler wrapper (for example mpicxx), which is required to compile the model.

-

An example of compiling with MPI:

mpi_options = list(STAN_MPI=TRUE, CXX="mpicxx", TBB_CXX_TYPE="gcc")
-mod = cmdstan_model("model.stan", cpp_options = mpi_options)
-
- +

An example of compiling with MPI:

+

mpi_options = list(STAN_MPI=TRUE, CXX="mpicxx", TBB_CXX_TYPE="gcc")
+mod = cmdstan_model("model.stan", cpp_options = mpi_options)

The C++ options that must be supplied to the -compile call are:

    -

  • STAN_MPI: Enables the use of MPI with Stan if TRUE.

    • +compile call are:

    • STAN_MPI: Enables the use of MPI with Stan if TRUE.

    • CXX: The name of the MPI C++ compiler wrapper. Typically "mpicxx".

    • TBB_CXX_TYPE: The C++ compiler the MPI wrapper wraps. Typically "gcc" on Linux and "clang" on macOS.

      • -
    - -

    In the call to the $sample_mpi() method it is also possible to provide +

In the call to the $sample_mpi() method it is also possible to provide the name of the MPI launcher (mpi_cmd, defaulting to "mpiexec") and any other MPI launch arguments (mpi_args). In most cases, it is enough to only define the number of processes. To use n_procs processes specify mpi_args = list("n" = n_procs).

- 
sample_mpi(
-  data = NULL,
-  mpi_cmd = "mpiexec",
-  mpi_args = NULL,
-  seed = NULL,
-  refresh = NULL,
-  init = NULL,
-  save_latent_dynamics = FALSE,
-  output_dir = NULL,
-  output_basename = NULL,
-  chains = 1,
-  chain_ids = seq_len(chains),
-  iter_warmup = NULL,
-  iter_sampling = NULL,
-  save_warmup = FALSE,
-  thin = NULL,
-  max_treedepth = NULL,
-  adapt_engaged = TRUE,
-  adapt_delta = NULL,
-  step_size = NULL,
-  metric = NULL,
-  metric_file = NULL,
-  inv_metric = NULL,
-  init_buffer = NULL,
-  term_buffer = NULL,
-  window = NULL,
-  fixed_param = FALSE,
-  sig_figs = NULL,
-  show_messages = TRUE,
-  diagnostics = c("divergences", "treedepth", "ebfmi"),
-  validate_csv = TRUE
-)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
data

(multiple options) The data to use for the variables specified in -the data block of the Stan program. One of the following:

    -

  • A named list of R objects with the names corresponding to variables +

    + 
    sample_mpi(
+  data = NULL,
+  mpi_cmd = "mpiexec",
+  mpi_args = NULL,
+  seed = NULL,
+  refresh = NULL,
+  init = NULL,
+  save_latent_dynamics = FALSE,
+  output_dir = NULL,
+  output_basename = NULL,
+  chains = 1,
+  chain_ids = seq_len(chains),
+  iter_warmup = NULL,
+  iter_sampling = NULL,
+  save_warmup = FALSE,
+  thin = NULL,
+  max_treedepth = NULL,
+  adapt_engaged = TRUE,
+  adapt_delta = NULL,
+  step_size = NULL,
+  metric = NULL,
+  metric_file = NULL,
+  inv_metric = NULL,
+  init_buffer = NULL,
+  term_buffer = NULL,
+  window = NULL,
+  fixed_param = FALSE,
+  sig_figs = NULL,
+  show_messages = TRUE,
+  show_exceptions = TRUE,
+  diagnostics = c("divergences", "treedepth", "ebfmi"),
+  validate_csv = TRUE
+)
    +
    + +
    +

    Arguments

    +
    data
    +

    (multiple options) The data to use for the variables specified in +the data block of the Stan program. One of the following:

    • A named list of R objects with the names corresponding to variables declared in the data block of the Stan program. Internally this list is then -written to JSON for CmdStan using write_stan_json(). See -write_stan_json() for details on the conversions performed on R objects +written to JSON for CmdStan using write_stan_json(). See +write_stan_json() for details on the conversions performed on R objects before they are passed to Stan.

    • A path to a data file compatible with CmdStan (JSON or R dump). See the appendices in the CmdStan guide for details on using these formats.

    • NULL or an empty list if the Stan program has no data block.

      • -
mpi_cmd

(string) The MPI launcher used for launching MPI -processes. The default launcher is "mpiexec".

mpi_args

(list) A list of arguments to use when launching MPI + + + +

mpi_cmd
+

(string) The MPI launcher used for launching MPI +processes. The default launcher is "mpiexec".

+ + +
mpi_args
+

(list) A list of arguments to use when launching MPI processes. For example, mpi_args = list("n" = 4) launches the executable as mpiexec -n 4 model_executable, followed by CmdStan arguments for the -model executable.

seed

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. +model executable.

+ + +
seed
+

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to @@ -303,25 +227,24 @@

Arg chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each -chain.

refresh

(non-negative integer) The number of iterations between +chain.

+ + +
refresh
+

(non-negative integer) The number of iterations between printed screen updates. If refresh = 0, only error messages will be -printed.

init

(multiple options) The initialization method to use for the +printed.

+ + +
init
+

(multiple options) The initialization method to use for the variables declared in the parameters block of the Stan program. One of -the following:

    -

  • A real number x>0. This initializes all parameters randomly between +the following:

    • A real number x>0. This initializes all parameters randomly between [-x,x] on the unconstrained parameter space.;

    • The number 0. This initializes all parameters to 0;

    • A character vector of paths (one per chain) to JSON or Rdump files containing initial values for all or some parameters. See -write_stan_json() to write R objects to JSON files compatible with +write_stan_json() to write R objects to JSON files compatible with CmdStan.

    • A list of lists containing initial values for all or some parameters. For MCMC the list should contain a sublist for each chain. For optimization and @@ -334,258 +257,264 @@

      Arg has argument chain_id it will be supplied with the chain id (from 1 to number of chains) when called to generate the initial values. See Examples.

      • -
save_latent_dynamics

(logical) Should auxiliary diagnostic information + + + +

save_latent_dynamics
+

(logical) Should auxiliary diagnostic information about the latent dynamics be written to temporary diagnostic CSV files? This argument replaces CmdStan's diagnostic_file argument and the content written to CSV is controlled by the user's CmdStan installation and not CmdStanR (for some algorithms no content may be written). The default is FALSE, which is appropriate for almost every use case. To save the temporary files created when save_latent_dynamics=TRUE see the -$save_latent_dynamics_files() -method.

output_dir

(string) A path to a directory where CmdStan should write +$save_latent_dynamics_files() +method.

+ + +
output_dir
+

(string) A path to a directory where CmdStan should write its output CSV files. For interactive use this can typically be left at NULL (temporary directory) since CmdStanR makes the CmdStan output (posterior draws and diagnostics) available in R via methods of the fitted -model objects. The behavior of output_dir is as follows:

    -

  • If NULL (the default), then the CSV files are written to a temporary +model objects. The behavior of output_dir is as follows:

    • If NULL (the default), then the CSV files are written to a temporary directory and only saved permanently if the user calls one of the $save_* methods of the fitted model object (e.g., -$save_output_files()). These temporary +$save_output_files()). These temporary files are removed when the fitted model object is -garbage collected (manually or automatically).

      • +garbage collected (manually or automatically).

    • If a path, then the files are created in output_dir with names corresponding to the defaults used by $save_output_files().

      • -
output_basename

(string) A string to use as a prefix for the names of + + + +

output_basename
+

(string) A string to use as a prefix for the names of the output CSV files of CmdStan. If NULL (the default), the basename of the output CSV files will be comprised from the model name, timestamp, and -5 random characters.

chains

(positive integer) The number of Markov chains to run. The -default is 4.

chain_ids

(integer vector) A vector of chain IDs. Must contain as many +5 random characters.

+ + +
chains
+

(positive integer) The number of Markov chains to run. The +default is 4.

+ + +
chain_ids
+

(integer vector) A vector of chain IDs. Must contain as many unique positive integers as the number of chains. If not set, the default -chain IDs are used (integers starting from 1).

iter_warmup

(positive integer) The number of warmup iterations to run +chain IDs are used (integers starting from 1).

+ + +
iter_warmup
+

(positive integer) The number of warmup iterations to run per chain. Note: in the CmdStan User's Guide this is referred to as -num_warmup.

iter_sampling

(positive integer) The number of post-warmup iterations +num_warmup.

+ + +
iter_sampling
+

(positive integer) The number of post-warmup iterations to run per chain. Note: in the CmdStan User's Guide this is referred to as -num_samples.

save_warmup

(logical) Should warmup iterations be saved? The default -is FALSE.

thin

(positive integer) The period between saved samples. This should -typically be left at its default (no thinning) unless memory is a problem.

max_treedepth

(positive integer) The maximum allowed tree depth for +num_samples.

+ + +
save_warmup
+

(logical) Should warmup iterations be saved? The default +is FALSE.

+ + +
thin
+

(positive integer) The period between saved samples. This should +typically be left at its default (no thinning) unless memory is a problem.

+ + +
max_treedepth
+

(positive integer) The maximum allowed tree depth for the NUTS engine. See the Tree Depth section of the CmdStan User's Guide -for more details.

adapt_engaged

(logical) Do warmup adaptation? The default is TRUE. +for more details.

+ + +
adapt_engaged
+

(logical) Do warmup adaptation? The default is TRUE. If a precomputed inverse metric is specified via the inv_metric argument (or metric_file) then, if adapt_engaged=TRUE, Stan will use the provided inverse metric just as an initial guess during adaptation. To turn off adaptation when using a precomputed inverse metric set -adapt_engaged=FALSE.

adapt_delta

(real in (0,1)) The adaptation target acceptance -statistic.

step_size

(positive real) The initial step size for the discrete +adapt_engaged=FALSE.

+ + +
adapt_delta
+

(real in (0,1)) The adaptation target acceptance +statistic.

+ + +
step_size
+

(positive real) The initial step size for the discrete approximation to continuous Hamiltonian dynamics. This is further tuned -during warmup.

metric

(string) One of "diag_e", "dense_e", or "unit_e", +during warmup.

+ + +
metric
+

(string) One of "diag_e", "dense_e", or "unit_e", specifying the geometry of the base manifold. See the Euclidean Metric section of the CmdStan User's Guide for more details. To specify a -precomputed (inverse) metric, see the inv_metric argument below.

metric_file

(character vector) The paths to JSON or +precomputed (inverse) metric, see the inv_metric argument below.

+ + +
metric_file
+

(character vector) The paths to JSON or Rdump files (one per chain) compatible with CmdStan that contain precomputed inverse metrics. The metric_file argument is inherited from CmdStan but is confusing in that the entry in JSON or Rdump file(s) must be named inv_metric, referring to the inverse metric. We recommend instead using CmdStanR's inv_metric argument (see below) to specify an inverse -metric directly using a vector or matrix from your R session.

inv_metric

(vector, matrix) A vector (if metric='diag_e') or a +metric directly using a vector or matrix from your R session.

+ + +
inv_metric
+

(vector, matrix) A vector (if metric='diag_e') or a matrix (if metric='dense_e') for initializing the inverse metric. This can be used as an alternative to the metric_file argument. A vector is interpreted as a diagonal metric. The inverse metric is usually set to an estimate of the posterior covariance. See the adapt_engaged argument above for details about (and control over) how specifying a precomputed -inverse metric interacts with adaptation.

init_buffer

(nonnegative integer) Width of initial fast timestep -adaptation interval during warmup.

term_buffer

(nonnegative integer) Width of final fast timestep -adaptation interval during warmup.

window

(nonnegative integer) Initial width of slow timestep/metric -adaptation interval.

fixed_param

(logical) When TRUE, call CmdStan with argument +inverse metric interacts with adaptation.

+ + +
init_buffer
+

(nonnegative integer) Width of initial fast timestep +adaptation interval during warmup.

+ + +
term_buffer
+

(nonnegative integer) Width of final fast timestep +adaptation interval during warmup.

+ + +
window
+

(nonnegative integer) Initial width of slow timestep/metric +adaptation interval.

+ + +
fixed_param
+

(logical) When TRUE, call CmdStan with argument "algorithm=fixed_param". The default is FALSE. The fixed parameter sampler generates a new sample without changing the current state of the Markov chain; only generated quantities may change. This can be useful when, for example, trying to generate pseudo-data using the generated quantities block. If the parameters block is empty then using fixed_param=TRUE is mandatory. When fixed_param=TRUE the chains and -parallel_chains arguments will be set to 1.

sig_figs

(positive integer) The number of significant figures used +parallel_chains arguments will be set to 1.

+ + +
sig_figs
+

(positive integer) The number of significant figures used when storing the output values. By default, CmdStan represent the output values with 6 significant figures. The upper limit for sig_figs is 18. Increasing this value will result in larger output CSV files and thus an -increased usage of disk space.

show_messages

(logical) When TRUE (the default), prints all +increased usage of disk space.

+ + +
show_messages
+

(logical) When TRUE (the default), prints all +output during the sampling process, such as iteration numbers and elapsed times. +If the output is silenced then the $output() method of +the resulting fit object can be used to display the silenced messages.

+ + +
show_exceptions
+

(logical) When TRUE (the default), prints all informational messages, for example rejection of the current proposal. Disable if you wish to silence these messages, but this is not usually recommended unless you are very confident that the model is correct up to numerical error. If the messages are silenced then the -$output() method of the resulting fit object can be -used to display the silenced messages.

diagnostics

(character vector) The diagnostics to automatically check +$output() method of the resulting fit object can be +used to display the silenced messages.

+ + +
diagnostics
+

(character vector) The diagnostics to automatically check and warn about after sampling. Setting this to an empty string "" or NULL can be used to prevent CmdStanR from automatically reading in the sampler diagnostics from CSV if you wish to manually read in the results -and validate them yourself, for example using read_cmdstan_csv(). The +and validate them yourself, for example using read_cmdstan_csv(). The currently available diagnostics are "divergences", "treedepth", and "ebfmi" (the default is to check all of them).

These diagnostics are also available after fitting. The -$sampler_diagnostics() method provides +$sampler_diagnostics() method provides access the diagnostic values for each iteration and the -$diagnostic_summary() method provides +$diagnostic_summary() method provides summaries of the diagnostics and can regenerate the warning messages.

Diagnostics like R-hat and effective sample size are not currently available via the diagnostics argument but can be checked after fitting -using the $summary() method.

validate_csv

Deprecated and will be removed in a future release.

+using the $summary() method.

-

Value

-

A CmdStanMCMC object.

-

See also

+
validate_csv
+

Deprecated. Use diagnostics instead.

-

The CmdStanR website -(mc-stan.org/cmdstanr) for online -documentation and tutorials.

-

The Stan and CmdStan documentation:

+
+
+

Value

+ -

The Stan Math Library's MPI documentation -(mc-stan.org/math/mpi) for more +

A CmdStanMCMC object.

+
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online +documentation and tutorials.

+

The Stan and CmdStan documentation:

The Stan Math Library's MPI documentation +(mc-stan.org/math/mpi) for more details on MPI support in Stan.

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample, -model-method-variables, -model-method-variational

- -

Examples

- 
# \dontrun{
-# mpi_options <- list(STAN_MPI=TRUE, CXX="mpicxx", TBB_CXX_TYPE="gcc")
-# mod <- cmdstan_model("model.stan", cpp_options = mpi_options)
-# fit <- mod$sample_mpi(..., mpi_args = list("n" = 4))
-# }
-
-
+model-method-check_syntax, +model-method-compile, +model-method-diagnose, +model-method-expose_functions, +model-method-format, +model-method-generate-quantities, +model-method-optimize, +model-method-sample, +model-method-variables, +model-method-variational

+
+ +
+

Examples

+ 
# \dontrun{
+# mpi_options <- list(STAN_MPI=TRUE, CXX="mpicxx", TBB_CXX_TYPE="gcc")
+# mod <- cmdstan_model("model.stan", cpp_options = mpi_options)
+# fit <- mod$sample_mpi(..., mpi_args = list("n" = 4))
+# }
+
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-variables.html b/docs/reference/model-method-variables.html index bf4a9d265..f26828ae3 100644 --- a/docs/reference/model-method-variables.html +++ b/docs/reference/model-method-variables.html @@ -1,82 +1,19 @@ - - - - - - - -Input and output variables of a Stan program — model-method-variables • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Input and output variables of a Stan program — model-method-variables • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

Input and output variables of a Stan program

- Source: R/model.R + Source: R/model.R
-

The $variables() method of a CmdStanModel object returns +

The $variables() method of a CmdStanModel object returns a list, each element representing a Stan model block: data, parameters, transformed_parameters and generated_quantities.

Each element contains a list of variables, with each variables represented @@ -193,99 +121,104 @@

Input and output variables of a Stan program

part of the model's input or output.

- 
variables()
- +
+ 
variables()
+
-

Value

+
+

Value

+ -

The $variables() returns a list with information on input and +

The $variables() returns a list with information on input and output variables for each of the Stan model blocks.

-

See also

- -

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variational

- -

Examples

- 
# \dontrun{
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-
-# create a `CmdStanModel` object, compiling the model is not required
-mod <- cmdstan_model(file, compile = FALSE)
-
-mod$variables()
-
#> $parameters
-#> $parameters$theta
-#> $parameters$theta$type
-#> [1] "real"
-#> 
-#> $parameters$theta$dimensions
-#> [1] 0
-#> 
-#> 
-#> 
-#> $included_files
-#> list()
-#> 
-#> $data
-#> $data$N
-#> $data$N$type
-#> [1] "int"
-#> 
-#> $data$N$dimensions
-#> [1] 0
-#> 
-#> 
-#> $data$y
-#> $data$y$type
-#> [1] "int"
-#> 
-#> $data$y$dimensions
-#> [1] 1
-#> 
-#> 
-#> 
-#> $transformed_parameters
-#> named list()
-#> 
-#> $generated_quantities
-#> named list()
-#> 

-# }
+    

+    
 
-
+
+

Examples

+ 
# \dontrun{
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+
+# create a `CmdStanModel` object, compiling the model is not required
+mod <- cmdstan_model(file, compile = FALSE)
+
+mod$variables()
+#> $parameters
+#> $parameters$theta
+#> $parameters$theta$type
+#> [1] "real"
+#> 
+#> $parameters$theta$dimensions
+#> [1] 0
+#> 
+#> 
+#> 
+#> $included_files
+#> list()
+#> 
+#> $data
+#> $data$N
+#> $data$N$type
+#> [1] "int"
+#> 
+#> $data$N$dimensions
+#> [1] 0
+#> 
+#> 
+#> $data$y
+#> $data$y$type
+#> [1] "int"
+#> 
+#> $data$y$dimensions
+#> [1] 1
+#> 
+#> 
+#> 
+#> $transformed_parameters
+#> named list()
+#> 
+#> $generated_quantities
+#> named list()
+#> 
+
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/model-method-variational-1.png b/docs/reference/model-method-variational-1.png index d7b8b97f8c19824bd7f4936ec52513cc8fae745f..27fd88217456b4a7d0da11a09599ea685749e06d 100644 GIT binary patch literal 15795 zcmeHu2UJtp+IGN#j)ie-BOr{@Y!pF2YOo9#5KyYrpfn{ykPaaPE90P|AR=8wL;|73 z&;kTSsRDx#Is~Ky2oPElNJ#ki!I^c}x7NRIzw5jATQ`fW6HZRq=Pl3sJn!DQf5Fsn z+jnB$!CKDfQHN+A?_Ee23`BFc8_N~X-M8I^%$xx$Wtq&!(VF}v-MIJmm5mR%TrqH zBa!fLZ0GfZTs=H8Pj46R)9)=h4&OU?izRd|@y(4K*+Lg!J@u?$U7d-cUhnYx-mvN!BbtBUKM; zhncpex%fmTzR9YMma?HP9}LFfDD?Zz$9sEw)zVrBqQ4Z7$@b~3Z_9GU7sVtLd%L>c z7I{ItL43s4zsBRYZ{Hqv^3)v?iDZXF7VB*)C@7elJn84C*b3Pn_Pysdk6Gc_!2J>$ zERv`<`x;H!Sku@m6I^g)(%{TWBrY?Qy>|oGV-(OIw zKc0&x*x)!xl%dZFOTKAC;4+n^!+)E?2YAS`)nW^~wQ^}0DcSg)@o*}#O_LJ6{q zmg`SS`y8hzVDOtc@~r9GH}Z((&B2x|4VpHA+QAVAcfY(d_@xwZ8=VhYFUI1at$+?PG(Tf{q1lM%p|dV4!o?dZ!mQ{f#?iC?)nJtwHH8 zQrO_|ovPj*g84alDI3C8&iVl^U&?xasgp7orO?J?i-<^96-jo}doeGLDPb$0*Uy2* zpZ{XWY4!)`Q$*)qK$Sm0$+8!mS;d;(|8SB4mk>nH1zFQs#58RNW-GV~2DQZ8Dik5L zz(6|Ytu8fDD{C*+;Z`$5QK%l4E*N`Nh0Lt5{)<>!i)d?nWqun+#zy;OZs_WZN!J0( zI0*3aw*nZvBW6SOuY%Xh=YEEU>F-?#VgKI@g?|kC??lo+PG5Uw%q~jA92tK%Sv%*!G0ai>eORRSV z^HeVTp=jei2=_OY!)9&e(gD`sjMTBBU{BRgCG}!p`wv%uKgRtJKK4I&`kK`?m*q~9 z&1zY392>6Ws1Xe}rm78|F!<_vr-%*7&Dr6@y=~lwXfvMVD5*n&3KDMc)DvOYp2bo_ zV6#QtPt?lg^h$ZiDg){__d1`Ne(R|S_0Yd6^6(_`21hS~0_l2z~eh4W*6{)E3>+l2p1y4JmF{FK1&GODcEhDF)Y z@+Y@%-%d>a%#$2eZ7RbPyR}(+)w3l<8F%{ObTY;&QppsBX4=xymGz-&DfST0DC?oo zm>7vmb$~d4N^F*vG#zqU^W@35%VEz^;w9Gu0!Y?2HU!brIRrxgl1w|G zY28z`rg!%}Xw1|Ja?}8pZyQ@!{=myYiVNR!R!^c{SO2VWiWnbRA84tzpCTy5im6#q zw_&iKl)DViwFabcuAD7h8Sqq3+FX9Nq|0nSiV4@S1;z5Or=b7MwZEIa|1X)&iv}Fb zvEas@~!R=$z~g4FOAmo(Yd znHg`IbQS4jhx%KLBf7?K6e?nDTYf%s@0FZT*^wDHTl>?=K7DFPc5n!smw ze@?tBr6Gwcl-oBCqz5?|K34JLVT1IGYE8>bGoW<8LQ0oO0oeIFw)YefQc_ZYAO(0{ zU0t0jf3P*I(%$rz2)CW33qw=8Cv2yu_b=*hT=(w(qiRq<`YQ< zXFHwS%#&OPi-1AdAo`)z)?#%WKHXloAtvPUV#?kJkcbHsLSe(;qFPZ{u7x zrCU*Ww-G5#NtK4BU!sUZ&0q`~P(R5BwaOy5ajALN*T$uD61s13;0bNGHan9>5ZktR zwma6AD#d@xJPZ)c8**}ALfj2%@Qa(p6Wml_g_+cU5j6FH^o0$w%5ptdOR`qyD&nBQ zYjA1+9|Puvs*ZCF<`Cbuvbb?d2>K^I6t?GUui z&O%@+1O&Q&mCdOiONjub7eAOwfXn@-)$6|zZ~xBRgj!i~adCf;-8{qe))yXbC5dnF zWtEx0^GR)zhhy{Ym~Ax^&56Ga!5mDy!yXLpDk0G&tgupcJB z2s5v%+jqP{PtPB2aqQd{YJJc6k`V0KONMu)0V!AZ=gxqEs2!v8visbL301~-qWJ9v zY=Iy502E64P@8Ux2Gxq%3BStdp2&yeP9|89hBe@j%7;DsdC<*AKx@0fpT`YMU><0A zCgNKN9No1huV_M5=$oM@y#q)3NE*CeuTl+^vG>hZ$P)5vdsGG*TI~Xk2OKL z^iQ9lSrb_>+p_p*3cmu8=T~0;j=(JqGDGz(>TZTr%|q3xiWC;+D_A;8A@wtV+OqV8 zhO@)bl~MuIgjId{c7Q|jd!8zjtyP0cY0j>kOULic#F`1i@*SX};3M58L`nLwi3dva z;Iz*_VeUe4`2_p?-q{{yI-f1jx$y;BYLAa)n3 zV^Jn;?N|K=ipaUy%VD5N;-H?UF-in=eda@$US9X-ykE98ec~NwemZv?bUm&$<8YpJ z73&nq<4ATl4}I-s5E_k!`bb4yC@^1w$D5j(PR`Ee)9;r|O|%u^0|8lbo*nrvLIcB| z0*pKb*eR{Bpx~lcZ-+n-PD`&kT?C4ON`j>o_5J&sqVm7w7Zq7E7>wKA{$W>KT!yJP z#>$fAHbdV2Tj$ySi$;G!BS;T}_P^V5jiDETIxbTO&ffh=yWXS&a@jvUL;p%z<;jzbC}sftPUBs+U&w`KjpQSTD-EF`MQjpg zu3L~2IFrKp+0hW02d|QKV5i>+l&4BAx4ssddCaV}VoEK3EBag0s(3IpD=?MATxqra z*4NhG_}`w}Up?jh=Tq_c$5t#@6krIg1^8PCQq?M&DY!k7%i!IF2QfLa2R1)m@`^aP zX@l3PhXq^ZEk{TCOD+V8GLsqB8EO8O*x@Yh@S_C3Dz za$nAO$U87|apm=tT|ewtScoy6VS96UwKqH-JCw+fi+QCO9d~cv(ta$BK}>6@WG;lv zU}$I%HZNYR2eA)c8)UyB>!CwJ0OJA0JVnwVgKfkF{@KWT z<*AU<34I0V~6V81Zzw9`CEQ|hri5xAL##X=Ae+5mzN7IzH#(fE@>c;76`_k z^}3=$w{3j>5yJ8a-;-6zX0zKpNpjZfcB&={nPm25*99^?*jClT1KmXX;2*O3{r&xS zz3B)YpESB+WgBt$RM7US*N5w*Que93yNWpuw91I=_rE|ij=-v<_~sqVwq)a;{C00c zd6EX&+p_m7wQqw{CsC#+ff_vaD*XAv-?_RQ2x!|iRCrc;p;v7k|<*N6AowEYkKmsL%v1lu@ zF`{NCMEoyYO((vT)Ra%??rt|{PbQeu=MH_zrtN^SZ$jRVIQ){RFyTFv-4|o!1G@^t z0onv!-x)y@uHfto!fQJsZvF60MeKSCo%LtqZ}VCs^S_sXa`9hj3r}<))bi0sb!l2Q z3@a67TB8DGF`aD}8}BdNGxrQU08`Lk^P&u*awRr3Z0N~=d;IYO#c2hywaxoI1NXs9 z)^ZdUBcpv3;bHGPb)YHNHB37)CLk#$yjxNKhc!gWPx3ts0_5EbJFmbEFv6#UD9Hb? z4b%_JdME(fFJ{XsO&y}&WbJQKbaP-77psxi!`M&O;zaI!)rG=IoHh}>lN=kI0+w== z^j~iJ7SvB;O8e_?*)}P-BU(293r7U_ocfH^fj#gc0_S=j}aoskeK}=vb22)TC4_(2%WpUcGdrw zM0(41UrY4-yzs&TUGhDn*m<)1^`ja5jP)6LdD2wRfYZ_1HwG?*_Mo=qT%XLA4O#hA z-`mN{eme4SA@U2=9jCgyR-*JTZ`~Tbf}yTcO|pESvzfN=Rf9zbZ;kwCSIg(cDwB|Vlb`1NfEcQ zw#svK+Tt*&6P|&GF#(^Ghx+DHo0=4FBYgTN_m#46Wl4@TcUI_{NxtWbH6-P}q4EU} zl1${H+^UBG@#sF6fOv2W{VE?n3q{vI>_GEyBsQ8qa_X=G!A9jm^E7Ef@= zMCU6useSYbaDBzG&b{H`HbUDU_ayq=yTbRvQ%uh&Ci7uUe$8XrdA&A6kBUV) z_0X{E5cW~q5B#Vr7i1k%xkpzn_DLt_0=PrNxWIczOvFhmb1LuM%<+T=<&`4#VK`B% zSi7h+mj_N~rgCI5_DDXtWXwRGu|y%%2#zkR#+~15ix@baA>>fOJ$+(|d^;A4z3(kd zRn<|nHyU!&t*uPv8;#H%iyU?GQ{;qM;cVojuJr{)*=U@MW=Upu-1%IijF?A{_(|~e%R=aNiG(EctP^|mJ3UF3 z*{qN)S9#8bFY^>l)XtOCRgMsf@6zoN2C0n+DNVUon1PltsgLSY=&u25Nqf{i)Km9M z{lK0yFY+cH8K*okA)zUcwmL(7vXf{XiY+ER-1?-KU!w3qFmU|&=c{vKk$7ojzk+}} zlPyEE8-Lyxi-gM=w(OV;2)`hjzSl8DL2uIEP&T3aE04UXcBJZpxnlI^-jh4hv0gD- zMa`Obhbl-*yLh}?^0U3eukbrGaTPTkqMS{7Vy4x;kX*O}JIxkeJf(d^8b|K610LEc zud5LL=y*4e?4X#r_+@tXKAY53;@&59oBEpgDM~`ZdwSu8>+70FMIB}C61v`o(cLd6 z>%>I~eIl~;r~2zrcELA37-U&z6ZXz}_q@v*XzDNl%Miyt@3U3$ad!R*i&aQ%_+Y$U zZ)tFY&Z}p5hJrtOswIVf_lCRAKZ;0&`816B zdXntaz5BYRe4S6bHzy~6rKWU|T{Yma+FS7^N$R9_@%PH9;I+tRYrzlwlA!AGhD^p# zBzwOyzb!mePQCYF&}94dD}x4@nNqs%Xq=54VvmB|(=Cc6FY0%{|M(zW=51W2{>_V^ zqr99=;bzO!?=X~j5{;d5G0EK|XMi+rb-3|qcXy$iaT%GsP}+jc?H;YVGmbi(h{~%_ zH`8TV#*E;za~r0*+pe?Zze6DHj{&;Ot9WBu?N05cX-dnT$uO^cBU6xwnM?ON3y*oqMdM{1(u$D4M=-j0f;BQaCgEFQrHOAaBUao5mkm!3Jc z6v5uagmR+c{0TF^iH!J}f(n`CaYJ}wteIX2g|Xq#D}%HsHj6cSYl~uCRh53T9&aic zCs-CSSJXd1CFc<=kCaXoHWmg>_vj9U58d56_CU3{iy129WHPjF0APn`*x}8i96Svo5Eve0V?eT+B)>ZpBbqW6O(-6I_SIu_SbJ4+#aGpk=fD zf&^;@gQ@rXZQL+ZfdtM)6G9E^j*#XcA7(3<;zH?{W?B9jr}xDPsMo8lrJp%<(N>yt2k6_-!DnJd>gCgt-$5=jSd2D<^ZjF71uMZ#bz z{cPj*4mltRH21fU!rfZJCu|U(oY`I)6)YYx1yH! zAh1WNeD=d@(gx+Y6R&V9eRRFz&RzH1I|nl^WNmMA6AG zpQwu3aOkbUA}RgUYeWlaIvq@t476>MNrsWbBUj2r<9 zo>ufczEnla!O_~I_Vwu=%$9aqZ`Wgt9NM)?&ptx1ULfH zhEwumi&FSp)LNxEe$7khyzIkCDHO@hI_J_8aqp4?i1`6dr3MK0-r>n=n1y2nuP2Q^ z<*^U6KR_K>0AF(4imkG#GWDI6Y*N^;IPF|a-;LV+WAqJg%7J~zxSs3fBxzI4(WiZ( z3<6Y91W&grAGj+iNo{R?vSG^}nSD6nMMKFD8H6wR;fI->JSjR750u_@^E}cUJ3LA3VvG96WfFm1mwEkrKqBR>C6Y&Kh;w9QK_Oo} zh)Rd*f7paEa&3qrlEHw5VKwJI?)2kMWO}r|`Wds;)sQa6V})UGJ;KkoPf|N)Dm~}2 zn|N29&N-K5(L)qq7;9!Fdpte5`Ju+Z^W8_SU4e<0ms@bDAbv(($C@^PimeK!ZtkIX zDp#1?-|?;6+4kgq@fwHtmp7;)o)rT2?;jZWhDcNe-tY|6gu#CM(v#!}?!QuIl<{Q> z{5;~t_YXij)cFf`Of?O1qtilSClG))cE2$Qhjwf<<6EKsy8md#-mcG7@<_b9!K`#E}w?HPxPe z2l9Xx(%r(OtP1*YtNi5^Iy@~Btjz|i;_`y0?rjsFZw{(dMRnvr2UYTB`mmh-HRY_XMt!D7cI zv-3NnaaKW67j`wiQLruZbFwwm!Q3AWz8UiFP$W;{`o#f24a4ql-<#M_6&x*@s!Ek27 zsY`&3V2<_i$rK*>de6IS*Dn_VN(u3_b1`A8Y-H1PG0iE`Llg#cLI7o5^;&#?liiUx zb|}S@lnAAQ)N-96M53M=WdPkU_9{xTArihaB(rY|CMXExnMZGtm*!M{^*s2g6qGC1 z&}cJI>mZN=V5FLv8tNG{0@&U;l3hPtEU#=5FC%?m5fo?DpqV~!(Rang^tb7c8Ylt= zW$eyQQ!p$5(a-DJfa7@Q!I1PDwOH*IkaKkh*dn-D5k$;>?Lb#>TNgNz=x$=2>yRK* zI#(mE>QeK-MyJR@kK%n%YmFO?e9mwNmb6cC7$6nUy&P;SvM6%9RM}yP-Mu++`lhK&YOvOT9qatWCjK zG(dL9u_ekG$167dy8Ecuymt8PZK(%$$vCzL?itW^fzjy?gVMBs!^GZ0RycncjpOyk z%7ARYP#stt^0i6%E<30ibD`pL53uT~YeB#?3(he;^FUUwQOy(jf;eEVq z+qRAVc1F|$08r)1s1noSAyg+~MjoQWTmThhFO>`pU5)~)gn#pat{~U{GEy?p)%6ZT zpy;EX8dSk>On~FUV@Bp1m83Y~5K#iM_&JDY)#j5fX`VXZNn>QPZStB22L~m4k(FT~ zTpo{u5(iT56=d`CWkHn-DK0dnOZh(mbsCUl{xUsXnzk~g5i6rX3-i-k84{(_;j1A) z145ende>vhQWT{Q1a+(oa=F1q1tcsMKL;h5O5uxMAoJdrgK$H{ynp{*hdZ8bTjqaF z4@fDHR+@+yz@(eQz->?5fdUc1DeJ6IEbjAN<6<%hk{toXlTYzCpCL0Vp$v>du3QNp z)(DpdwG8}i*>kK86xw*-(qtU37}p3OlU-U91aw6yxMl2v^&f$pH3Ndf#>QpP22Kqlhy)q9aAu03s+TBE8I5Lsbxwjs;PWUIL*6b=6S>G%)l~R73&+ zDFGn_a2Uae1T`dpkOU3I5FipwkPwplo;dfP-#&Zq&W^kH?%v;dFt5*f&-=dL`uUb~ zt{-tYr1kekP$B|R>xBaR5PZ!sG zyY$Y;fz7#*dk$-_(NNH;x%Fzv^+l~`G)^7evOi`2zU2Mi9X+~v+qQ4F-r^Wvjj?ZL zF`rY1;#NO7@xzVyvka0!B9+hTDK&BKmNdPLA0HYoTO_CJ3;t}jMy0YzRrqA}#Dr3O zM})*pzIlGJPZ{Td*>U>Xlm_{Y3fuvng8YBxxh+CyI6vjID+XQJS)ES^iQdQi3ZBwl z;(a2br^TsfwK8v>fHJ>XJu!;HOYaF+gUO_Ii}Wu16sU|BHF#8Ii$pklN%;nNN6YmW zmG}NSO2#`*5xsX!;YcpT05Z(y=44w27$E4!ubSm!@e*%%Oggei4<4<{(ELf(ezmju;_-;a+5|Yx|kM2UQFK?P>*fM9``G@z$ ze^M!LiQ#x~j?;FPZsg5}V90R|!c}04>_5Fzg>%HGth(hDShrI}Jz@CP#m>;s zNQnwwDk>$STBJpIDPFkZKc5eaYMm%cCb-%~?CKvtOKs@Wc5Jq4C!Ws=mP^(_3v60V zb`kyxh^L7U&;z6)`5lj5xK4#Mc+b4-s20Kkj4|Hnwit~6y+D7{#{*y((D2Klm>>zg zc*2RNTFSlJG~b@p#IIukF+5BK`ddNl>q8}m_oAP0MJPNTA7~Da@ZAA4yk_#5H@()e z9@xNg^<%2K9#!Dz6Fy+?B_6u?Xn)CIGwQV!1hq8^{N;N8yY>GC%tzmd^NFU1;pMn? z1}c^ULfftb=CWTz2}dg@x(F!v&N^V8n(<@dbmB^&iuqY}B5c0afz-dzO#dL6|J>z5 z!gA=I@rmO>AubC;4V~4(Ozmr3`1WEUjJ3n*z>3Z9(T@;S@H?mKy?0Of)^j(431g*2 zdTpH#*{%uGHoy{DkG!JBAO5!F?W!;1?f+akd|{~iUq$>eHX7KkZ4cR;wLy$72GIC2 zN~jk??kD%bz^ShesD1b{wMm>3{GjzbnR8Niy)e*4m?n3J4_g2#AEU4A!>{+^eWyO< z>mpQV4FbTnyaeF;hzn`Z#lt`m26|`D8aN#P?RfsbN=Uyq82w))Ud&hwB?+VW4+A#f z3tA0R2@Q5`DK1@dut++z6&T9+*$VvViKw7fA@)4qY7GeFNuEGbp_S3~_+al~@~PK? zPM76f3>{!6YAFVPCD8v%$`%M`^MEtP#IicBssMNq!i!GZ$NL8Mi2RiomDToPsuCmU zbT$!=sZiV5Z)4u9iWk|qSkV?cs;Rh;V@u@#3Q;e&t@Iy$6-*b|E#@y&%0oIt?XpW> z2X6(He)I1nv@ety{!1Fnza#JRPy@v~wY(P_C%!Kd-k=G+=y^W8CWaZviLF3pmDIiK z_*9Tfg7_3~3vF8E7}KZTTLH5E8lFn2p}Be$h7Efu!U|RD+4%eF2YQ1Va4htr*uZ8} zwPW(wcHvui-`ZWhQzuDqp53nt@9aPJERkL+a<9sco=@a~SQc6jD}(wx(I+aZRe0k( z{1C50BT(*fU#fih%}C|hHsr83wX=tQM9&iw)jSxj2%@kSm~VToTA5ntMdx`9Y6DiD zuK)trQb6*I3M&Ovu|aJC68wI1@TB!m3sd;yL4qqFG2$)O5T^7G7g)bh`1jUl|I*7r zaj8GrQO@C~y0kLUo!2|WcAu-`3p+1nnar0|PR#DS&`_{zD%2%vPaQB>pr8BS#^eyE z2zWh9I{Han1FK5RkCSTzdRUhK`1cjJ|I)(y|HS`K6}A5y$xnDF0ehFS-_*dFUu=}5 zmf_)0OLS>qp+U9-8Z1fnpuYmO-Bela(td(Seh^Z}+U@rMPd^e>hi|Ri4nVkQ3;2P~ z-e=>*b39O2$H|H&6y1~|-}V1sPk}332y#=?O)KGv0%SS{|!};z7MASud!7 z+G63HDk;s+M9oT1Bs5E|oXXTb?ZPyxz#mvhZ~>)Z+iSQdm2^nZSnfjdAh^cYjUEwJ z@nIMUmHvb2|97rT|3O99XLB_UQ#2Y9Oh@?z*I0&YeI@^WllvZ6tvAS#r3MAg9!PPmf@W&*3K zn(77t>_54K2g$?mu19A;6#qNAnqS6);jH5IqnX#Ac!msT0xJLQ6;P{z%a^)@r5F`P zlD?TfDQ|c}ZJaeRM*X8qe2@2eP>-QPTz6cP*JF@x3wXBv#Id06PwU1i1Ih4uXe_8B z7IuCpDILy6!1aamq5qTF#vfEV+zE%Eya4}le1~}ALqkVZJ-}O&W5O*}%NF{Qz539L zq4-yDPMdUawW~|=0eT6JVSAt1;$wKz#Z}~?mm^B|qc2x-uY!{H${)MBbLyA{su5(F zR+bN5vvOxU)S2JKDD-=7QLlr~f`LvPdUEH-!5Im^4G8OPV z1pU<)1br0`Dt($op)?3V106~mZGY)Ma$#F>Z}rJ@OzZ4dyHC8U zjNNblcfwM-P?6_$Yptl_*pF8O1 zw3mYV3T{bif9_?t+k8%zUpRgJxp$oD1X~6K)n#0gzcMH)S@#)x{`wXHOQUB2O|OJWFzO4gJr89g1j z!Nr>8Qp0wyt{Ul{m9&RP&`gZYtJ6lEIwY(pQY)b(RStGVmaIoJqSTUJ<6-@2U`eRXu*nR>2@)@~OpqU6m6 z!dZrq)GS=O@4QO%?#)QcW&6@$yWAXyg62l@mtG4@U*^CtDIjtSz=&DmmmJ{5z*5op3KRQZ?yB5aarJy?8UDY}HGl%O&f1V** zKev%N&FP$$@^Lq1t*9Dw?VrP_F(cF!ocwCk2rB%hj68U2=BF?jscb=(8T9ngE8B@| zhcWW{?$IFQWdTpv-P!(wJnd(T`LFC}t7*T+gEc(nGYP%gQvA zGe%6b=YB+Q>?%*B-*dK++8~3g46V+Y_H16BCCH*t?&5Mozz(-^C0A<12l>jm)ZK^e zcj{>LJfO&Q4NuPQ{tKPB+)lD6|bR=dbax+1ooZS2owWkwa=%_ohZ^#}*;j!KPi- z8h!W9YEqN|*@v|U+|x@-#+>c_H{N4h*oiZ5V8oXpl#FoxMc7R?f+_b16yPW}(xhhx zPFEyr>2#VJJzp|%CnLU+tE&|nr9J0B*x2)|IX&lfmk2w=ip=1YY2?}JDVL6<-mg{Z ztbIl4qjjBVRc#$gZm{Tlt{5wE{_1pmnk9@YNztax{_m+;oY;f#yMD;}viI#MPgZnB zU)>{fN{?3T>? z_ABNlqhN_-VU{g>dH(}gu*0LixiVvS#5fKm4W_vzU9YR0H%$?LtEi+E`{5^_(_U1Q z@%+GabB)slYl{f=7X}1Z5qgbpZ|F#;$g`~ZQu}d*>^*&+5#27xXrIu>OAAv^yC>c5 ze|WOJ3+uV=gRcHnQ+v4h21xwzzYTLRe08*iWI0KTxND$vJVX4_=^5|%@#8JMk?WI? zb#---?B3p9UI4)-#B^obz3nByzD}#+g7a~>go6a(4&>SNdl@yYG;VVcyCtDP$-u}u@20@GZg9%d zkYW@-aP&EG;!(uHd4<>h86PH7Rn01M3JRoLcOJ1ML(eO5ZZw2VLh)a%IkmW=j24-6 zG8czixPtJR7NQGTO;J2Q0|SG;UU#Faii)8Dt6M>5bd!*~CL6<|qebHZ1omM;GIA#} z97zxob#!#tLyZyRwpVfLMT!SwRm5+VTVqTq7ZICWj~~~ZKBJVBz6&244OZsV*48%I z+uKj^h3E!qPmip3XlO`nbf9;Mh- zwExEcWHiaqRzPv*l{W+pMh`AlGJjzf`R;`f$jV0W5dp}`DJ--yo-ADI>zN{k<{j(k3nGtDF!WjF;F=H<;luI7pOY?(DyNfTMsM+)$I ztl)koqqCS4J@tBsCuBu4qxw`m&O!Nx2Y2;ElbEgEHUl(v$hC0gbHD6g*G!|)^bTX) zc|pjDYMwg!U9qeWBd&985&WE$UseriK6yCa5j0p&xWH@@j&Uvso=9JHW0Gn+04vJM z$}%tQQDU5El1s4Cwl-6YD`1l2X0IX+S$;Po1tJJprK8hUmi^PE1_n9-o8i<#WQfx+ z5>A&JIrDU8W@e>{^JxXU8z*~G;5}6NlUJRd`7Ml=I_l_KuS}VM^(4KZJ(@|hG*k%P zyHR8wH_JrHj$uE{`?SOz-lgC$037dXaYnccw%46+N;C zK(ty}Sz*}d&PC^vk3J^?V;e-1S(6fpq^Htqi!~|5w;;g_Gk1g_ycFCpFGc9t?oJ;$ zU6O&Kc6T$Q#n%-asQZBh*bL+&a1P@y!uERD0dYP5d<3O2tTR=$PqWeY?Ab9z-p1uc zL02vMre6~S0R~)jdZwqS{U_^k#fSoCt-qrd#a8UY^0sN574xcOB{?}cr61Ac*4rah z40kKqRJ1B>(N?2x{NHPu^!qrd`sbw}oor%9R;4c0*QP!ldKB1KS=Ol*H^~NzHUT4B zAE6sszj4Ad?%j*xhD@i#(}{`F;p_EFjcr#26)Ze+ZK>ftH6}8AJb__~0K2#8g$Z+A zKGL;-4H1H$Kri{b@IA+?Rh6H?duKznCLlL>#!eRjz+%Tw9qWiqNn>?|{U~wJg#n=FFU~>SG>h|H@ z+J4;zmjVfx+>t9pM!BPbUK=v!Cxe)@rY><*AfXV4V-8`&%!GN-!aKVZAlQQzMu3%; zij43#zx?tG#+-Q*3NpIXSf75T{rj}7_puiq5FMT&)uB$qOR=JMWKDB(^Vy3YHjP!S z&9m$7F3QJV-xf1IHkNp6WG!+hX{KXSLq#>kV7?1gK!GKY%AT z0?rF38$D!ZR#u$W>-D~i6oU$Id3hv4hM?G;m**{hl~sWH*yKfuw%3?zL=+Yj7Rn@R z!WXYPmb+Tt&I~mjAEZ6X){311Nx|~;GhtyfrRkDH05`!v%RZLTg%X4y$>}P~h}E8jTMbO@Sv*zM)F%`8I-y_G?O3vkZ?A zvTT(!rp@gfmO9)GG`QRpM~z5tJ_RT&e)8lA2`kyJJ$mDZE0dR%Ec=n+!B~3$csSwB zrTW~fhCMLQ0x)j5bw+a`&|2+-aroUEv)ZVOwcYw@=M`e^00QD>r@e9u3)x2l$tIM- zND{L>IXStoikOH29v?8e^S=t-f74uO{Wdt!NK&xDYH7tfV3oCEkIFvR7TIwg*hA^P zQjYX2Ncq7Py4?nAbsa8!7`^9B!41TK!SK8x#g*!tN}N3ZNhn{}!RW%OSW?9p>=<>} z;^RoLv8`&2i=oaW8K-LkJc7jHPv;A_nc4L3=iHsFoB^7_QnIpwg&fNvO)CQQKr0YWU@V7jkT7n|W0ju<})lxLmp)j9?2yqm|x$OwD z1&rtL__G+SId@NOx2Rzt>x?aCG6&H*B$FNrfBP2xY!52>IW>bg3t(~UQvPLRq824Y zYao~Y_M@O;ecd4&R(If}OoUMHTAK`jLhT(N{Az>OR5Z|qD6lEt=A_bX<*6NNxT|eTl_E=uJgy{S{ zTA!jN?&4QeOKlfMS1z9fH8bx%kd&{=M~84yLSQ^U`H)5HSauYDQqRiDx7kreD7mjN zxawIwc0{WdZOs39cZ7r)V`%q(ptDc*I&(uSjOM1Je%eCY$A*=IG-AE=R%LIpcC1WIpXKG zBO{$|YYcihhZ~oH#umrd!vFp1nwZ`XU46T#Gc{yQia$1fFVZM2iv-N?Ae;$rSbo>| z@@(bZ0$F!@f8zE)A3fBOdYyF1N!c7Ptrj#dZr%hIniQsgW7BPB;RtDAod5K?o}(xe zpy!J!3)OPSxVuzYKEVa_7o+)scnIl+4EBg7YEHK|!?s9h7 z0~PBZDk0{wVkM@JTGJnwOkP3m1U~3vIcB)mPng?SXculwNkB@CDUJna>$evCJaTto zmsRG-cd^aCGV@2Qv|_hWLNo_LFZIdEf-U}k@^YZbmM1Q+;nolR_ync{dB9fZ diff --git a/docs/reference/model-method-variational-2.png b/docs/reference/model-method-variational-2.png index 093449f41694c87ed4f6a948f6425a8e02c8a550..4ed9fe5c527f0931750caf802c560ea60a2989a0 100644 GIT binary patch literal 16009 zcmeHu2~<<(y7oqjv<}c#1(B&%Dk=ghqRhc#72`k@1!Sg`83bfbfCQzLDgrVn%3M(p z0z~F91d4!w5CN0O7?dGEAYn=Zge3Rdc>1q@{r8?*?>#;D{%5U|71j>fdw=`e@B2Q_ z`@HYBbLXO|$<|G~HbD@y6?x`+GYFChfgtf8Ki>$x8F={X9N299@tn!`U<+O=Cu?zF z^M&sjD}M-5jDr8f;>PElA?O9V09x@qQKFma%k$;w-$$B!$Vzr?d4%x@R@JMEQQoBW_zaiZc)Cj-A5R>dnhN0X7ndoK>DKSaUqq``PlG4*r zh_xRIX-ulWBO#~PJv}4C!kp6Fe9F$wPTT%7K84L=&X2ZomZo~GHIfS%4fe#UjrsXC z_4P%Ii;FGMjv{-UfTM>%5SUCR^|B*jW8?`#%+mAx{D&RSTA|pU(bhQ6;;V0)o0|*k zMN3Od)SEPuqp$&>pZeSML&V;Zc(TM*>Ogjme9Pz$7jns@`g*0ee!9YUndn1(RNC;> zS+pEDdJwfiv%zrw99=ID$0xZKx`w@4iWC{KEQQ=qPIEPR92h^M7TBJ3r#Z(`H9H@% z_?(lFnoh{C7e(fe=g=^401j@3N9kdkUyxe%;r@&Z{t|p?zhCrf7ubHfZ~ji#{*BY0 zzt`}oqqAwgRlO%08s)ZZR9@M^eMTpvJG~F>eCzwp(@@;LT9((uU{ra26Wb9*2*(gJ z_>f5laQG|W@Ze;be&&!+cfpi>V{s)c`M9*??95C7=d;+?A+`?^&O3z8mSdJ*a5lYD zq!}yJwmF$4wms!+ar^ODD#hy<7(GQU<}~L`vysWP2F(~xnWnZ5xsMR*>gsxR&Ky;c z@-)h%t*oH0A4{eB>wMcWUQnU!Z8++W=FV6V>1=Cqgk=$!eP7>WFuR^;&#e6XaX%|d zOUoG9MxkNITvkz0kw>*{MaG6u&WBykr_oPq>gv2ez_P(4E%4@c4i5Bbes1qnVH~ly zp|;Pg)nN3- zPu9PSQiLJxQC`EjL2Qrj98w;$(a?h4cC7&IkQ?>ceJh0Z@g$ze_}EYxte z8$G2l0U!EreR*khE>5KgFpx$v%wg$iX>a|i3pkh|%%&N05*@eF#p!gp9Nqk6{@Zt? zJ}-;Z&Iu~bIC$mxK>EXoala|k&Thb{n&O1b`doJ+%d%bC#n!=rHa@<)m09$M>xD~t&^uli1pkJ2V0#F?F>_y!)zJZc_Fvgy3nH1 zNQ6T9h@`TGuA;Btyah?c|ACnJJHfTpMjARZr%BQ!(6NjcqV_c*KXSUF&@!xsQ;~0J zj^J@hQ>3JVXKI+`8Ss2EJAR;TiV&YyuGcDLULd$gu2oL-o}xIt!8GUE6yPh&IYeHR zw$4cAZr!DeQs4wl1>bx~=U|+xxr3f3@q-H;Tg$mcW}^@VDGGCY-W;cU5#e!-GR3i? z?eHfk^YlMPcKw}}`p2C9{0CTc7JA{QeFO{Ht4-0Cf)t*o^580dF&MX6Bb;}r3wc)g z1q6+43lWRRrd-Ek*@6>>X%^|R4xB=seV;lp)SOexuJ#fKM(om(fJP?MFXRtDsKaW2*oinfuJvic_3aA|1dU)Z6*CkPKix{5R+v2KHGly=X;1{W-(b4duN*fmD`FsH zB+ZG`YudW_UfatM zf|lC~Yw{6#Y#M^!;?CTE9J3hq1_N|t15tHugK<`K&GZu{jUa3+<0Q$0+k|!dAP&g* zpHjBsv8zbtj^q5MTyvb?Fa}fr4T6bJG~gd8|Nr8x{K<94?^nNFYh+BI4RCgke8a&Q zG7u0Usda}`UtDHpmu2h&hM2;hII+1s>RqDrjl-p-rFO#x3T2YgN@W=hYHt@8r4+tN z?re1iHAbQ;3P?k;vtBwiK<_#bnUBZ}?M$)_Il`r7wAfQ7mg;p3-{%#CXWPNl#xfOW zQz+&@_pMX7}uDVa#jn zqjN4ofn*V^k)NS%Lx5x8)!07; z-@WxsK0eImdFl*K?H(D~vN1Alt>28q)>eq%KcO+Ji#}w9FZZ{uv>%q=vat&n8`}zn zur$=mXw0G2Dt%&WCu4D)+5qyY+dPis)ilwi8*cjZjFeJ~raPm55TyDbybQ-LhfkH~ zlSH!Cgpn`L?@dfOzZaIVrF{i1rB<+vbvv3LB`wWQkq8?9y|(5gY=RnY+IZy)N8Bl- zRSxx>cSWB?VQoHXg#x2a8wgqhAZRV|=7DyV9iSyJ1yr_DoM*KyHK_J_jdQ7-g*iX& zE;Qm@VIHVgwCca^x7uWLlBg-4+5BXqhA7sIL??7lqKD{S$I9&q@__wP&+i4FJWKn^ z)(dg8uemubFc_oy?Ngw=DS<_J-dns1VE-U3t#&Bp3c4mlsR0i3tzfar_q9# ztxY5K9{jJr4l4mopy}IPUDuB8ikWVVlr-n8#V9ct3|KCMN&#*Y{hz*2V@E@+J%QO* z>bl+#z^+vRezvo<9h~N)fSwKa5^dN9$JHly$bRR!e<~RMN!`JJ-TR+bEjLj?N{djQ zsxVJx+B4-c-H*JZz8W|MK`C(78;Mx!C1g@?LPR1}%cxh@v}Tn;S_H>IkO>NuAQ9pV zs>uR9IdkG+A5ci5lFiw+7`_J7u=`8+?1p=i<|tEZLX}lz;li*T6wx%|91KVKANGS` zg=!{P2+rlm`{33z-U13UD6-Z-2lPHsSQwU^F(A+U8cwpO6q-kxQ9Vg(|*TIS?{Uj`amqBDhDSDa7cy2=*j0zs>`d+s;&>-g#Jg6R%vI3=*_E|)KC{0`zys-f~X5~*rN zeL475s!gFKI2u1r42N&*&-07j@bc6Y+>vH2eEH@)iHqEJ<|412i4156ZyQ*6C6TdY0wZh29)N)B9Pi*)DN{ zzBZ2%q^}N2poN1n(KU4jgPw1_wepi1Lec1z1!lL~QhWpdJw5W;4L?6Ka0NiWHmIA# zU7!j4MV*qk>H%B~!S!u>G^l#((I#!@rp2PmAFH>WUBOvOp2xt8w&nswl+sEv6-= zUKFrik)D8aTN35rk%dJ0pG&H)EKJmv`8Jt3`Fz#?8{ z{u66zk9Qa-ETK|kHL!TPJYznOYG+PArvhX7sLGt5@34XLTL%Fz4nW~N(Vij;=nURB z^5}~?g|tQK;cEW4qEJBkmH z0?PLW*2#Mb>HbR2!Cb+VO==846c z<69#@UEbT<3&bn$ZR7l~RyeZ$5Vmlb!-fPH`jibVSYf%I$sB3nnmNenAEUD> z_?Yr%!-5hwddwrUe z)Rop8TeTl;4ip|r&)X4w^vS+6M|RxVu{ZKB^WXk7V)N!pMfLBEH#n8Hu`-7IS<2_Q zPn?doO{9f)C59PkJsz6TW$?mK8U7h`|9UhRUDI=UjGaeK%)?&dKK^wqG>)?`M9gi! z1mHsGa=t}E_JnNukACDWsspzZ_PD4nVxmp8s$)nRdSZ~|w~UXi_I)1@hm4W?;dT%N zojv|yP=4xzb@Fj|IL}0eS8`tDe2bw;}%&@w|{I{efR8>Z{~kxi2U00NOFduzCa+D zwf#&o3w`aulRZ$~Uq4VG5{XtXvT`T)y*XxRXlV7q@kIAw#};!uWr@Ri zW`V!JoR2wut-r(Dvr+%pv4_RQ#X%ak#4i%pyWmw6E*` z4VC{D>;8uUtL53WZD?iYw(w{Q9CeOFvjC7)A!wh31pCAU%fw@)Yft&>!X4>Oew@?Y zV3h@evJ|%0N&d?6@EiQ$RPLj}-ji)jHgJH0S2|huZ=vJg-EvJdy!Zu4M!IOByA%COyqmR+ZEAxDSb!SlCa{O>0E z{|9*fYOa9Wzjlv3e(xdV%SyxMa0xzMI=z&;_*-AhlWb_a>AFQ!kki+Y-?h~ZpY6W1 z*IfCW68%yeA?V7EI_T0mGmyK@k+8j*$Aa#ukcKY>Y>Mc5ukgtnF?322_S`$0PFK?K zLT~-xqFeTxgO~HJB#6+jeeDfg3nuq9h<%(&Qp?=rGR@`~=NaD2>tWu#d*A6#KUia1 zO3Nrp=rpV|SVWI5>xGjks(b%lUXyBqi$9qs{?6gkF+;HnfTt57-tD%hPuF_CUQ3dP zU1vaG(D#ngAAP5O?8rk~QMOgJ(4_Qk?2%`*wMvW;Z^HF%tk{g?dKxf})=mZX8hazf z3eFsiHz*EA?QD^-6o8y?6|7cXN4N}o4FY4aEMxR!)8{Q zm70#pPO!G@M(#a+G{s=~`?y;?XdP0J&KXX?cbfWjliL&JcOtP1%qN#;thCzR3)-S5 zWB}3;IE;a+WQLScl_Gr6h>vLck8`>PiNASvsQSyQj)U8zZnk~ShQa7U{1fz*tHdSa4V>n?#5uLr3= ze%5?zXnLeH8J-7pXvfDr|49<|6EwH!mBZ2VKJfY0HJy1adL^@yHD z+iizn?;F~B4`-gMvT@_F2h+4H0@8Xe&Q-kI9o}|nG1~NSb*6!?@HA|Pkr6YdX)K}) z7&z2rUtd}I^u;4;6qPl28j?AOxmO5-*7_E3I} z`f>(yfmNU?uDF_VzJ*TDR;;gQ)>4BdT@|oFnSLfje}}}jWq#<>CK}Iek9*LvQA-i| zOMNEt8I1Fzy(^i!cXd9E@9L@=?E~LOwbDAoWi!kp2NOG;M`Jm}KQa#v*52IfB)Yvid$nG#gh3&8_} z;N@;uQ7GY)-F-&>Hv8&;NrJjKNMhF6y4Wk$q9uDxEf@ZoBe@mu>xG3Uzt`3^NV$SX0=Q^)zg~_nF3nu@ zSsB#L>hsF!-X}bSKT^O>&qkswdxjeIjGQICB5P`Da?->%GDrP`Eu#qR z_r*>RIYjzi{_A(hmR+NyQ*Z8ahA7(g**&qkS=*VRDpQ(`t}U2oYOsDM^PIg-?c{cT z5gzYJvKPgi&eqyIeDmhbEz;7Jn2<>~Aggr8?w?#(>P5K*1e8tc9Lio@T|MI3m8Q}* zGU6D-8np8L_K_8kOHoP~w~Ffn598ydghR*OBJKFb$RkIOCe031%BFqpGWH=+&-^~TbngreQ1uh(KSf=%QLh6YfD=${#47VQNvplo=r$YPdv7& z@Um*W<2dnHT<5A_xfRjnYP9%RrA>Z#?smb|;wK~DA6h&>wtwGg(9k$h+^W>!oNV_{ z1>LM>AJ41j5hYY_J+37A8(Uh^$_Vb%g;H;?zLM{FbQ~_h$4DheJ$qC=+c&K*y4W|e$xBXP;Vg}(t}hT zuUqUom!J^%_EMpeQiN2{OmQn1PBkOs$3}PP5mFP=9lduV!m@thfkdvD2Tj&+iWT>nR&Lf&r3!TUc3a|G{*Li&4EwQ{X?Ptxw#pb1ILE~F0<QGV1LM;~oBE z@g@~ha?l~W*}-5@qcrw@T=>j2tKuI|bs6#oWxm^(Wo~UK3D(RCY$`P>hk|Ud+|tZv z1cD}fi^Hqb$_XA$#2c8~GBR?>{a9*e9_s~qaUyK;q;flp74+K|s$&Zsny6r&J?Evm z;^qjh=Xel57a?%hs9*jt6#GC$SvVI}>eQAXXy0tiEc2&_J34+NYwLRg#F;<0&y}&Z zNT$=Ppt6g2Mg*XQa~jnusmHGO&(6x)hy2i`GkVXR@pI%*l7tUX`1@pnFPcjH#@hichKY~w@JMwa~8e3vmq1tv{I8R;SyvXupHahpBHuR>)GYLx@>$GpxYqy_c6*Al~N zQESR;Ba*%Cn~XinJ?R2IU4-?nnaZ&ZdH=I6=j~-4VjrYV1Jfv~Wt}}Dl$QquB)L{F z=uhrU(e~WiBEcPcD&bSw_Zo#=NK|QKuXV9!xJGciC5-(VrIg~g`=j?bL3}`3rU?o+ zUVe3Xm|%gQJ$D}J&$8pQ$Oen&HEL%{?nHZTc!TagEV)hq+pFH3PR#f!l+T?-CV<8r zC_-nsMzzB;afaM?Nt@l*?*(5u0OA>k7@w^i^gEX@R!#_?t$+Dh*4F-F_ORPg*fUF9 z?nKWy%?r=W+%H^(tn(H98;ZVNme3420Rl5A)w^atcoYleR9Vj{%(KX?W2r(~PapU)f2D~H#HkMcGzdDtRS|eKEou_)9p_jT%{F}BOhtavo{!tEc^RQji zptQ8KtCN#crEEQVZN4>81x7%Q?HRS!4OxQz;vw&2sa8ZxU=fkquhRC|HRJIPHH#VU z*&1kjF1S7!CtSRD@1B!jm7}}(S_?N?aXLF*Wo03mlt#dE@W_NzpZdNPwAiM%(_}HS zU@Nh$sPabp{@k(K^4_mb)7<5~=e2A)yVG7A_L-RFS4S?|=F|AS`?; zQ#)NiHYAVrdlEXoP))nfEUMLDhcWaY)Iy5S75OaSP|yP*&AZYFVycgq<)CH{j>5&f zD}JvJx2mj&ay^%?*G@!&hZxxY_Kt%|BTs9hSE75}Rep3^bv_-hADWr-`QR0HZ;C1_+t+33a22>_^LwvxCL% z(x!zjeI?Gw1Rv~TDjyb72P^lade z3==RN$e{6)Fuw;wE3`?L_o+J!T29+e9eTk&(8&;bzx6g)dUJARE(%dMF!1o8RYJ68 z{qk{PF0s$cG|hoIjVE-)n<(kd7DM|i)Fxm=)d^;OP$dlfyoObY3Qowt4L@0eX5uWc zQt#5El(9O>t8b$G{+gG3?0g}v4L~9cO>h@&b8~ZxEqDw3N-Ak3b&bUqdT-7YQ?T7j zeATa^8QRyeutLTe9Ul(4#;Rsyz=>*H6NffM2$l}o*M(I6JX5sjDhLWTNC(p~1=C^J z-wZ}EECYn6BD%WU$zo8_IdHLSvATO{cENTQ1pD+rCUxABv}!J+VfzTC<&I=c^6n^w z9xmfE$r3}Qk9{m-SHIuXK2g>DaohAiEr+tQX9laa%B|JG8#r1`5A=)z%2=Qe2Q~QM zK_kxu=o~ZO&p<0zi*-43^$KvLJzzH((0O%!1?o4r%yh8@a-V3=p4=1IaUVQ1{vS4U zv?kyL1Wq?vh>uf2W?-iC-g=aS_h2ZSSQ~ymyZaT0({)e1IVs(lrt1$A>e?bg~o^y*Mk$!TZv)OS*DZQrp>h!gVj)&@Kkl!E0$R8KBkbQ_8>L7+$xY8hLbRW3Y8IW6V z*K#s$C3ywtVqsk-J(wO|&93J!0p0|3r)Ng62my3K*K8#{!?UP*FUpW}Wmn8;(Dq#HBV&&tOKyy|ge_$& zhYyctge|nC`Tq7rU)2o|YZH%Pi-IPsoVXBQUfpl8!N%rf#aJ+oP0az;o1wyjn-az0{&u3zD)^rVnkn%z%1|5W*COZowA*dQzXete$d+@H57aDO8gEI!gR)p%=fF(XI!J5h zO!cOos3t8VnBB1fWBqcwR8|cR&^os-aes=sq{1zIR%B+pzhg=tp}DVu&X_W z^cxDKIdfuV3=6F>E}qFP}ZrAl+KBlyx_qbCJF0QQd6r+ zA1b_Zz^W#|9hP^JN+FIUKOoAz1}fC!5W)~8Fk!FZI$gXaAwwr&Cj|AM0uC{5ot~W? zVzkcad2kygmCqMiq%PEh2dR-jGCmBKP9D1{sq{(}vC*{N=i;9&N%C`szw)Faa`AG zZoelF?FF|bVa!YIGf@r?IOaY0XuKoY6$tO^RxLxpbi+$5;Z&y39Tqo3O4rv0F9QNG z)?e=mOD4hD1(^V%v+{b!G7Tr7sUUGUTv7e%G!pgQJOFd0Iv^euc>GOoWB1sSjrcPv zqYO$okOZnk`sAdW6k=@-aK_pINibmK95(rDRDud2q*<+pLMb0;jfsh=h8J(H=Yj}v zY*EB6zel#j`KGfZy7n&S94GEKQMmDa!WfW(=}XvT2P_&!4ycsUZnVhpljf^NW6veP z!V}2#T1DJ=`-48<^Rf&urx|s)Gp5Cg@d+D!$=fWeaUT5ku){R(4N%}fZ+apVVAak> zlS3Eq4uS(8VQf?3)h3BpZY;2^K7w2T>~t6utwXu@;A$yhesc|xX~%8{As4SACnwh? zE^Tlh*pjg=bNTxTa%QH=3PH50$ruH(t`0Nm>#A#UhR5Kd1Qf!VD(**jsVj=c5>Z`> z!GrODp~8SJ?2NWwdU6F%0q=vhRfjAc!UOXu9!9)svNNDfe%d^TfD4BUs_*19e~Hui z@C&F}U~z0tz?gg2&Q=VDjK*w}Mf0anxD=1_6j*wwhi8CQHFv!vkO;)z!b(oGLS_JP znt^(r4cC?MQT|jb*Ks>G#&V)FwbEgfexp@!1@_S(nvo9hY*pff90uhR11=drP1Os) zR~#39FjKgcdl^qD!EmP2{Mz@cV}ii)2fQi)7D&`{-QX?s2`hvDX^Q)^hY%okc!*B! TDKcAs_5^v#^!uW}INtg<4VB|G literal 12652 zcmeHt2Ut^CyY5z0MnM_J89_mr5=2y_NRh5%!GM6m4818$X+c`(MEw;V1ayZYCCM-t zG7cgN0!9eIf=sAlfPfG{2~8w{P(lbKXYF{}bDs16^WSsl{`a1H+|Q#cthT@PeeYM- zzJAKqa^r^W8z2bUh(2-D9)i}!K#XLNrc`+1u1C+!>b$$x2{|T$ zf!W8mad$ZVxwc$Ra*`m)Jj^H{iig3r35jE3(i$mQf8u|*)yyNy>=IN)t!SBmnmw}Y zCf!`6ki=mBYNBpr5hI6m)IY3HGLw$rWGl9tLAdYq+8xX#r!oeOjFWF0F*%Dn!RR8i z#uMkmc`p~6l!Wq9q75hQD;wEThSJIySOtb#kx~CVn4*>L`LLIp#2-!T*bDlHVm^;q zm0qZ>x{c;tff?K!111ms`k1jDdEp_Ai;775b7sAil2cYgYmvWqt4*&)+RQJI=@4YK zz7}a9@)}R1!P&3YLfTKuKd}zM!6(@K!&Rm;a1MfcjFa`sh{ieX8N>@-TZK6+V!k+f zP98|)B~bby6Vut4Y{Tf;_+|~4lSQoHD!Oeq!%JOi0&I5!%{Pcy>fl_DC~^v3$inP0 z6j+vlrQWE3f2i^!C&6uX^1LmXTfV#8HohaJ*W|=Pcaf7yDAppD^8?HoBnnewr}H*e zMQ#Q3C8%Fi95T0IRA3kIaskJ+dV?{^bC=>xZsV1)uw^@{Ib_8GHlpgp$$-iIvS4ye zn19ytkmd2Ro)1A7wRr6Dvd0#xw3`L&ZM1hQr7Xt0Tago(TTcs_taJcBz_QE#6YQa5 zkt_t=nCq+*itx)X#&-yO%*Ip)W0K+`<~9Kw#D7)`)n7L#n~h**(zgMBb8o9qh)TAI z1FIgG0Z)KKsV3_JWI5H+`$P<g?^c9W@h9SA=RN~2M-~N8snDu7)){`w-P|IM;m^XT848Xgkd4s> z@0>nWn%%b}1~VlGeAOYXe+-E4c8#xs>M!Sx`Q9=Wb|{XLcGCRt^O?;>hJY-U06E6_O6kzb6iuo72!Cff&u+s#?M8D&%Yc0DnD^$X&mvojrjm{4UWy6{x;C7S+Jl0 zfkO#m0_F|+!ho(JOYdz3!elxb%xoCdei5R!&y1?uaOAD_hd ze_o9Ln<)6_<%nk2Kc8Q;o^3c#J)2*IMZ_02x5o3}Xxl(j!~D0&x;6~GIKo4GNXWb} z#R*~|A@kQ4)_rt5ntzZ@yQb}gTJP5#OvZEzdpk@v9k7cvogM!QDp)gHaFUVXk$6U& z)!~M8&sUBWCo2J}p!7@ia;o(SH%Tp-T4@6pc3)7$DKPZZCXGn8{0U@oyRZg@prR2> zv>dV|o{X-VtH--x;WBWn)j0gSCLxC}?AU7{TB?Sz(Fnn~qVJ}kZFbh=ii3+rvbn2= zVq8!5iSllvm>`;g!d_G@P9gCN1f@uGg-`iE!i!I}kH22`3Dl1R97=E=HA0uiy_1{0 z^3=h=Rh^PuWC=o^BcP@<^x}50ua!Bq;&+pxYLIH32Kgy;@H7aNK=KOWr%6fvhjt_j zd-Xjs`>0Xn$^C>KAcCq6-6ax92cFqS(U1GWQ?b^c}Kz~5H} z`{+S`ZWbp9F6fLXlAajQjiOuM7HfDV%~w+}H$b6o`T^o5>#=@V30plv95{YcG@Fa^ z+HSTQ8ajh$8%~!0d=OtS@cPK$!Wd_-TVjX~RNsNH3nyD%aUwb!Nw?0MfdHdE#mAbC zK@tjR$n$?fVXhL?Mv>y>j(d1!Gm4>@Jk&m>wTT$i0;t%7R`S->;N)pfi=yCc3%iqa#j3n z%ZAHWxOtYHD3lt*chC0Mhf=0*YoGrOemn7K{h7Rz1lz(#YWRak8n}T!RB!wK{KaB} z+x!h0J%?`GTK??Wv#Rdb+?DwZTKQjkN<7Z<*b!y(!j^uc0D&nNt12)IpvQBKKyh&Q zT4`ikf6e}5P58#IGsQ?zq`rv#y+=WI9>HzUsVvx_H8(5IZBt$8@{dd-$a$M|k$lfX zVZ(V1VW~fUUs-0Y&h<-*-bZ&Mj?_EN{h=B)%!8>BCt7}5yafuY`VxWsgX1ro-pN2q z4gc)vk7r(zGRFI;#z+7A|87TT3=_%tao^a7k1Pz{)jDtdGI!sj@NuqTmX-lQ##>2x zaGYgl9!otok!jR+r}Ab7N=Pk^faGPQ?3%t-stju?>JfgaHmS>3Az+Nx)=!_3ay{d= zlJ%~X$C^reZPX#9{h{T^eJZ5=isELMcGv5Tga_&*VY>8JksF2*3K3%m9dGj41viES z8Xv=82T1)EEM(~Ic0^`>^Y#BiU|{WpbZW+~#;x!b+wzsGOLfH+`vSfZ=3wb$)1Jsp zL;W8`yfohPefcT;%)?sp#6DW&`?KrK9Qk#q4W3G8HXtg7Gj}u0Hl5@bJIqBdo(pe$ zTkHU>4E?WDaP(E@j!R+(sVE8p(s%OOG84jk1_#V^a-!kh;)pq29b%aZPmgdmR+*)aRVPoV%#ozX z9WtH&E+7B!n~<0>McCx(QvW-=f!9s<@WTN4H9_avlZ#b9vS{q?#z}bQ8_gJ1BF6Y> zQz=GGnWYbqFazJHN53p6`X+HbqIJ*%mH()Ze@#K5`iW5!#@a1tICmr6N!ga!H*=+U zOIY2&K4Alwa^D551^S%^1B>qBso^oY=Ba2DZ# z8j}oX0`Y+-Y#=`7%tMYQgJ(b>=j2)bWVM#(Vn1ED>Vi51ZI*ruu<7?`;&1;==%9b*MiV+j2B~E{3Twbmw6>Up4i4AajZH`*Rx# z4&R-juF!5sSbeUGYl#zXSAt~-V&TuWu1)^&t2bgKLWJtHOOK#XkRXny7FgsJsibV=qi z8}9cE;)WXAX({Yz&md>Xy6JmijIFiwg}2HVGX?zM!*yb zCt||){=`9R4Ye)$v7v!tFIbe<#%z8lbSJ>i)r7~HE^&K$Z{;yiUgwOhw8*}0Hbgkj z4k5v{qVN&_gX?SqoIsy2HOq2$gE5_Xm%+{>Plcv6k91jyi3;R53=xJ><~Uk4JBzwu3&t;M}L03;|j1Vfj8-Xvz zRsI;X5~eYtm9fKQq1q;3mG@Mpz4({zrBoju9};rRKYGS~DQIn)iWXT~YhX$%ZPkoi zt%P=^j*}jug(=ZMulWTF!PHHMB7`~HPPVWk&`3%-2RtroFp^^i? z`z%D)th_*Zl~kHXMMp2(@1m^z*=rrU_d`E5qH%F?(T?R>wuhW>VY0iFS3C1kIYZzx z7-k+=8rg97?p?1Fe6lTNHaDcN%EQ3Oh+5SB<;RVS%k{TtXvdSb) zhL&0{>}y>*UnwUi=QUUv#7dRDrHYF{2vDSr9~(w*Lb+5wco62OPJpzXFRMr>#l`mv z^75WU1qTNwQ<&6PfB#WRF@+-7gFO@F7^&SATAi%N$btJ09yl=j#Gx8f-RFf4@2Cpt zBR+x^eLtQ)b~)J!w(D~Bq9uwJ+!R#J`;Fq}#`VdW7PkCOR0O)JrKQDG?^@YfT$b{j zPJ9hJ$qx29Q680)BwjDONo8^2y@SjHS3phQ)IuLztIiOEqtfhoGAk9o9Eb@jHQa@! zFk^WNTEAiJ%BNlx9+8|Ym9jl{S<-6Xl+oVSmNfcO8DpBnv@z6dX=@Xc0KW=Qa_C*w z8d@KCPEb8niK8Q$XL(6#?VDD}nTaK>>klPw3|vNdMF0IPRTMR{e{QgREia1YfjzG= z@&?toz#EjLslGW}GC$!vqadj4SA;Hiwd=&U6V0}eGy5nM3eP;TSG*!pP#2znBaI2_ zWyS}g5?#P+b`2IY9%;;by!5`rPHPi%uIJ~M5`X|aIg`c?!S2BpO0qR$8XFr)IjG8R zH+&~EM-^x}rl!I!c4V&?)vt8nBdP}K5=GAdvqKL)-{UoI%l4K1`OIZge^P$FNI=eH zaoIK16-#%ZIp)_dC1&vpCs9T->1hgWfFGr*E*DiF1wVBQqU}OkX<+wNE^i?%zDU?w zmVbM=F@xt^lDJCSc}1l{z-9LtF_Yh0Yve>xNJeAz&|JgM>AgS{?b-IBuJ73@kRImq zVae0WvdMcGFS#ka$s;Gd*AF+Bp0bJ!hdW*A5m&poC&Q0v_=#MnL^<6{h&LrWq=wn_9NQB#@1H+ zhRT(}i6;b$^mqaHq3dLdwFbSPat)F{L~oSP;ssT3?PJl~DRO3f8Fe(zi&#*QkU(Jy zWzk~K()sM@nkgRSO!jLZvf-b?8b@D^>Q+N*j9k=5^;{+}zxhY&S*R@%gA_2|tHkD}GTISW4G3F%cnSTT(4Vpm$ja z%*FzWF8R%g_P*VGsUu^jsbD-T$&H~OgW=g4lFn47!gsm?HDd|chd&^EOx1}G4dv3v znTC&hSJzLEk1bm0>(399Gp~bP-IkO#<5rCeW+XBSrqjP~)CXvyM=kMJM1sN7Rzeaw ze0gyuhZN;Kw0W=5Wx)$VE9s3-EV^7D0mJ>bb&%4%i|M^J(Nl3DrSlG|qMkOFUc}^)31(*tj>e+Jv0h3nb%2@bQNn%gP!hKyzarvqOR8p4m6S z0c%;Vk|0>I)3xkiO6?(*)%~aO@6tz|v?>z<0|Sp>7u=*4Q;Lx7oZDvLB<1!^$0{l+ zB!2qaR*lnF-H?=`wpg3mDp~8E4)S0y)FJub0Y)tHP_whM)>|KwGeu9B@$KL43Cr`7 z8vvx#;1hHien4N@Wb1q3e%K|hR7drH)u{nw88RXQ{Qb=zbn-oFqRqJXF0Fp$lxQxN zWC7-tcYepRo==&rhQ9Px#HHTn=_3R3)yl*`A2%2lI z&*-(*AfR$dB;LO2{ZLSG*|218ZmtyA#gfW8Az3SN;tFMdA3&%xOtRwxx^nv*bCuhI z{K$K`Zz?A%$(fRyoJu?%uZo7LR4TXyDzv3U%mb4{k6bdS!Sa$?`uqF6P`&O}v4;ip z!()%MpuCQ|52sA~x|9QmxfM%O_rfQ{Zj9LHhk4#EU8nuXGn>=1EnDK4XL1=!4F(q+ zyOECqUfoew&kZp?s$kbP?GHYA0vRp@<;~tfk*q=9c~eb3!T@~M^cN81;C8uZae~l0 z`&MGUP0jiK;LPK&+$paD&L~O&C8VR1cBLBnn&2twLPK<2F6SvqvX{|5l49Lo(OPAe z%9^0*Gj4H*+iP-FFwzV%!&Rcw?Byls44STj6CR6ZCNjKu4-?V+HdJzE!|N*K;E~uG zx>fj$k$?XC`xROYGCR?3m#~XFW|2LT8h{)fcPsLjdo-T>WWnwjo*h30outmF+P9VB zR^j}LxI5Nx-@_FXe1fTk`y*SFQ9qm8y*%Ap)$G?7(==NbetkrM&GG1)-3yNWBR^XQ zg03F+cxs#vFIS^5w&4$IS_@HSCVqqx+dh9wxLRG9rbZl&*=pU7qV8f0S~LeMs0n?s zORZ|zz*8h1&X@uc=t6Qj`O9q^Sbfzm#_GbFmeZNpv>FpyX`|RYwrP3Zta~CW%h7xq;<2q>G8@|m{0W5ha)R(=gLwIZh9BspE=vcE3^UPi-ZyJB6o%8mHWd;y@P zed<1_=hSOEw{EMYwGedaQD2FS9I99-sv08z|0tpy;2fuq6lT1YQSe31z|YPnapN ziuojf{3}^00O!21PxLPv8LW&^K1i??F$W%z-AO=G1;k@jW04d^6m|NMUn$Bda*Qol z4dE`HmO{CVs4G#ZVV0tkXm%pWFbW$vCOHqwprbjQ3)5tmeHRC_@qW`*6@(|t8F27; zo5g;HEUdyyO0d~#JQ6T zcksyOZWE45>_CbL6!Z^_yT(bmJm57;6Yp6 jhb<9ELGSrG diff --git a/docs/reference/model-method-variational.html b/docs/reference/model-method-variational.html index ce353cf1f..e5fdeb91e 100644 --- a/docs/reference/model-method-variational.html +++ b/docs/reference/model-method-variational.html @@ -1,80 +1,17 @@ - - - - - - - -Run Stan's variational approximation algorithms — model-method-variational • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Run Stan's variational approximation algorithms — model-method-variational • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

Run Stan's variational approximation algorithms

- Source: R/model.R + Source: R/model.R
-

The $variational() method of a CmdStanModel object runs +

The $variational() method of a CmdStanModel object runs Stan's variational Bayes (ADVI) algorithms.

Any argument left as NULL will default to the default value used by the installed version of CmdStan. See the -CmdStan User’s Guide +CmdStan User’s Guide for more details.

- 
variational(
-  data = NULL,
-  seed = NULL,
-  refresh = NULL,
-  init = NULL,
-  save_latent_dynamics = FALSE,
-  output_dir = NULL,
-  output_basename = NULL,
-  sig_figs = NULL,
-  threads = NULL,
-  opencl_ids = NULL,
-  algorithm = NULL,
-  iter = NULL,
-  grad_samples = NULL,
-  elbo_samples = NULL,
-  eta = NULL,
-  adapt_engaged = NULL,
-  adapt_iter = NULL,
-  tol_rel_obj = NULL,
-  eval_elbo = NULL,
-  output_samples = NULL
-)
- -

Arguments

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
data

(multiple options) The data to use for the variables specified in -the data block of the Stan program. One of the following:

    -

  • A named list of R objects with the names corresponding to variables +

    + 
    variational(
+  data = NULL,
+  seed = NULL,
+  refresh = NULL,
+  init = NULL,
+  save_latent_dynamics = FALSE,
+  output_dir = NULL,
+  output_basename = NULL,
+  sig_figs = NULL,
+  threads = NULL,
+  opencl_ids = NULL,
+  algorithm = NULL,
+  iter = NULL,
+  grad_samples = NULL,
+  elbo_samples = NULL,
+  eta = NULL,
+  adapt_engaged = NULL,
+  adapt_iter = NULL,
+  tol_rel_obj = NULL,
+  eval_elbo = NULL,
+  output_samples = NULL
+)
    +
    + +
    +

    Arguments

    +
    data
    +

    (multiple options) The data to use for the variables specified in +the data block of the Stan program. One of the following:

    • A named list of R objects with the names corresponding to variables declared in the data block of the Stan program. Internally this list is then -written to JSON for CmdStan using write_stan_json(). See -write_stan_json() for details on the conversions performed on R objects +written to JSON for CmdStan using write_stan_json(). See +write_stan_json() for details on the conversions performed on R objects before they are passed to Stan.

    • A path to a data file compatible with CmdStan (JSON or R dump). See the appendices in the CmdStan guide for details on using these formats.

    • NULL or an empty list if the Stan program has no data block.

      • -
seed

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. + + + +

seed
+

(positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to @@ -239,25 +166,24 @@

Arg chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each -chain.

refresh

(non-negative integer) The number of iterations between +chain.

+ + +
refresh
+

(non-negative integer) The number of iterations between printed screen updates. If refresh = 0, only error messages will be -printed.

init

(multiple options) The initialization method to use for the +printed.

+ + +
init
+

(multiple options) The initialization method to use for the variables declared in the parameters block of the Stan program. One of -the following:

    -

  • A real number x>0. This initializes all parameters randomly between +the following:

    • A real number x>0. This initializes all parameters randomly between [-x,x] on the unconstrained parameter space.;

    • The number 0. This initializes all parameters to 0;

    • A character vector of paths (one per chain) to JSON or Rdump files containing initial values for all or some parameters. See -write_stan_json() to write R objects to JSON files compatible with +write_stan_json() to write R objects to JSON files compatible with CmdStan.

    • A list of lists containing initial values for all or some parameters. For MCMC the list should contain a sublist for each chain. For optimization and @@ -270,500 +196,521 @@

      Arg has argument chain_id it will be supplied with the chain id (from 1 to number of chains) when called to generate the initial values. See Examples.

      • -
save_latent_dynamics

(logical) Should auxiliary diagnostic information + + + +

save_latent_dynamics
+

(logical) Should auxiliary diagnostic information about the latent dynamics be written to temporary diagnostic CSV files? This argument replaces CmdStan's diagnostic_file argument and the content written to CSV is controlled by the user's CmdStan installation and not CmdStanR (for some algorithms no content may be written). The default is FALSE, which is appropriate for almost every use case. To save the temporary files created when save_latent_dynamics=TRUE see the -$save_latent_dynamics_files() -method.

output_dir

(string) A path to a directory where CmdStan should write +$save_latent_dynamics_files() +method.

+ + +
output_dir
+

(string) A path to a directory where CmdStan should write its output CSV files. For interactive use this can typically be left at NULL (temporary directory) since CmdStanR makes the CmdStan output (posterior draws and diagnostics) available in R via methods of the fitted -model objects. The behavior of output_dir is as follows:

    -

  • If NULL (the default), then the CSV files are written to a temporary +model objects. The behavior of output_dir is as follows:

    • If NULL (the default), then the CSV files are written to a temporary directory and only saved permanently if the user calls one of the $save_* methods of the fitted model object (e.g., -$save_output_files()). These temporary +$save_output_files()). These temporary files are removed when the fitted model object is -garbage collected (manually or automatically).

      • +garbage collected (manually or automatically).

    • If a path, then the files are created in output_dir with names corresponding to the defaults used by $save_output_files().

      • -
output_basename

(string) A string to use as a prefix for the names of + + + +

output_basename
+

(string) A string to use as a prefix for the names of the output CSV files of CmdStan. If NULL (the default), the basename of the output CSV files will be comprised from the model name, timestamp, and -5 random characters.

sig_figs

(positive integer) The number of significant figures used +5 random characters.

+ + +
sig_figs
+

(positive integer) The number of significant figures used when storing the output values. By default, CmdStan represent the output values with 6 significant figures. The upper limit for sig_figs is 18. Increasing this value will result in larger output CSV files and thus an -increased usage of disk space.

threads

(positive integer) If the model was -compiled with threading support, the number of +increased usage of disk space.

+ + +
threads
+

(positive integer) If the model was +compiled with threading support, the number of threads to use in parallelized sections (e.g., when using the Stan -functions reduce_sum() or map_rect()).

opencl_ids

(integer vector of length 2) The platform and +functions reduce_sum() or map_rect()).

+ + +
opencl_ids
+

(integer vector of length 2) The platform and device IDs of the OpenCL device to use for fitting. The model must be compiled with cpp_options = list(stan_opencl = TRUE) for this -argument to have an effect.

algorithm

(string) The algorithm. Either "meanfield" or -"fullrank".

iter

(positive integer) The maximum number of iterations.

grad_samples

(positive integer) The number of samples for Monte Carlo -estimate of gradients.

elbo_samples

(positive integer) The number of samples for Monte Carlo -estimate of ELBO (objective function).

eta

(positive real) The step size weighting parameter for adaptive -step size sequence.

adapt_engaged

(logical) Do warmup adaptation?

adapt_iter

(positive integer) The maximum number of adaptation -iterations.

tol_rel_obj

(positive real) Convergence tolerance on the relative norm -of the objective.

eval_elbo

(positive integer) Evaluate ELBO every Nth iteration.

output_samples

(positive integer) Number of approximate posterior -samples to draw and save.

- -

Value

- -

A CmdStanVB object.

-

Details

+argument to have an effect.

+ + +
algorithm
+

(string) The algorithm. Either "meanfield" or +"fullrank".

+ + +
iter
+

(positive integer) The maximum number of iterations.

+ + +
grad_samples
+

(positive integer) The number of samples for Monte Carlo +estimate of gradients.

+ + +
elbo_samples
+

(positive integer) The number of samples for Monte Carlo +estimate of ELBO (objective function).

+ + +
eta
+

(positive real) The step size weighting parameter for adaptive +step size sequence.

+ +
adapt_engaged
+

(logical) Do warmup adaptation?

+ + +
adapt_iter
+

(positive integer) The maximum number of adaptation +iterations.

+ + +
tol_rel_obj
+

(positive real) Convergence tolerance on the relative norm +of the objective.

+ + +
eval_elbo
+

(positive integer) Evaluate ELBO every Nth iteration.

+ + +
output_samples
+

(positive integer) Number of approximate posterior +samples to draw and save.

+ +
+
+

Value

+ + +

A CmdStanVB object.

+
+
+

Details

CmdStan can fit a variational approximation to the posterior. The approximation is a Gaussian in the unconstrained variable space. Stan implements two variational algorithms. The algorithm="meanfield" option uses a fully factorized Gaussian for the approximation. The algorithm="fullrank" option uses a Gaussian with a full-rank covariance matrix for the approximation.

-

-- CmdStan Interface User's Guide

-

See also

- -

The CmdStanR website -(mc-stan.org/cmdstanr) for online +

-- CmdStan Interface User's Guide

+
+
+

See also

+

The CmdStanR website +(mc-stan.org/cmdstanr) for online documentation and tutorials.

-

The Stan and CmdStan documentation:

- -

Other CmdStanModel methods: -model-method-check_syntax, -model-method-compile, -model-method-diagnose, -model-method-format, -model-method-generate-quantities, -model-method-optimize, -model-method-sample_mpi, -model-method-sample, -model-method-variables

- -

Examples

- 
# \dontrun{
-library(cmdstanr)
-library(posterior)
-library(bayesplot)
-color_scheme_set("brightblue")
-
-# Set path to CmdStan
-# (Note: if you installed CmdStan via install_cmdstan() with default settings
-# then setting the path is unnecessary but the default below should still work.
-# Otherwise use the `path` argument to specify the location of your
-# CmdStan installation.)
-set_cmdstan_path(path = NULL)
-
#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.29.1

-# Create a CmdStanModel object from a Stan program,
-# here using the example model that comes with CmdStan
-file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
-mod <- cmdstan_model(file)
-mod$print()
-
#> data {
-#>   int<lower=0> N;
-#>   array[N] int<lower=0,upper=1> y; // or int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
-#>   y ~ bernoulli(theta);
-#> }

-# Data as a named list (like RStan)
-stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
-
-# Run MCMC using the 'sample' method
-fit_mcmc <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  parallel_chains = 2
-)
-
#> Running MCMC with 2 parallel chains...
-#> 
-#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
-#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
-#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
-#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
-#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
-#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
-#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
-#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
-#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
-#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
-#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
-#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
-#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
-#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
-#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
-#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
-#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
-#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
-#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
-#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
-#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
-#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.2 seconds.
-#> 

-# Use 'posterior' package for summaries
-fit_mcmc$summary()
-
#> # A tibble: 2 × 10
-#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
-#>   <chr>     <dbl>  <dbl> <dbl> <dbl>   <dbl>  <dbl> <dbl>    <dbl>    <dbl>
-#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
-#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.

-# Get posterior draws
-draws <- fit_mcmc$draws()
-print(draws)
-
#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
-#> , , variable = lp__
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 -6.8 -6.8
-#>         2 -6.9 -6.8
-#>         3 -7.0 -7.0
-#>         4 -6.9 -7.1
-#>         5 -6.7 -7.0
-#> 
-#> , , variable = theta
-#> 
-#>          chain
-#> iteration    1    2
-#>         1 0.28 0.21
-#>         2 0.19 0.20
-#>         3 0.16 0.17
-#>         4 0.20 0.36
-#>         5 0.25 0.34
-#> 
-#> # ... with 995 more iterations

-# Convert to data frame using posterior::as_draws_df
-as_draws_df(draws)
-
#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
-#>    lp__ theta
-#> 1  -6.8  0.28
-#> 2  -6.9  0.19
-#> 3  -7.0  0.16
-#> 4  -6.9  0.20
-#> 5  -6.7  0.25
-#> 6  -7.1  0.36
-#> 7  -9.0  0.55
-#> 8  -7.2  0.15
-#> 9  -6.8  0.23
-#> 10 -7.5  0.42
-#> # ... with 1990 more draws
-#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}

-# Plot posterior using bayesplot (ggplot2)
-mcmc_hist(fit_mcmc$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-# Call CmdStan's diagnose and stansummary utilities
-fit_mcmc$cmdstan_diagnose()
-
#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181227-1-1c6def.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/bernoulli-202203181227-2-1c6def.csv
-#> 
-#> Checking sampler transitions treedepth.
-#> Treedepth satisfactory for all transitions.
-#> 
-#> Checking sampler transitions for divergences.
-#> No divergent transitions found.
-#> 
-#> Checking E-BFMI - sampler transitions HMC potential energy.
-#> E-BFMI satisfactory.
-#> 
-#> Effective sample size satisfactory.
-#> 
-#> Split R-hat values satisfactory all parameters.
-#> 
-#> Processing complete, no problems detected.
fit_mcmc$cmdstan_summary()
-
#> Inference for Stan model: bernoulli_model
-#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
-#> 
-#> Warmup took (0.0050, 0.0050) seconds, 0.010 seconds total
-#> Sampling took (0.019, 0.016) seconds, 0.035 seconds total
-#> 
-#>                 Mean     MCSE  StdDev     5%   50%   95%    N_Eff  N_Eff/s    R_hat
-#> 
-#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8      781    22315      1.0
-#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0  2.3e+02  6.7e+03  1.0e+00
-#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0  1.0e+00  2.9e+01  2.0e+13
-#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0  1.9e+03  5.4e+04  1.0e+00
-#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0  8.9e+01  2.5e+03  1.0e+00
-#> divergent__     0.00      nan    0.00   0.00  0.00  0.00      nan      nan      nan
-#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6  7.7e+02  2.2e+04  1.0e+00
-#> 
-#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47      796    22752      1.0
-#> 
-#> Samples were drawn using hmc with nuts.
-#> For each parameter, N_Eff is a crude measure of effective sample size,
-#> and R_hat is the potential scale reduction factor on split chains (at 
-#> convergence, R_hat=1).

-# For models fit using MCMC, if you like working with RStan's stanfit objects
-# then you can create one with rstan::read_stan_csv()
-
-# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
-
-
-# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
-# and also demonstrate specifying data as a path to a file instead of a list
-my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
-fit_optim <- mod$optimize(data = my_data_file, seed = 123)
-
#> Initial log joint probability = -9.51104 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.

-fit_optim$summary()
-
#> # A tibble: 2 × 2
-#>   variable estimate
-#>   <chr>       <dbl>
-#> 1 lp__        -5.00
-#> 2 theta        0.2 

-
-# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
-fit_vb <- mod$variational(data = stan_data, seed = 123)
-
#> ------------------------------------------------------------ 
-#> EXPERIMENTAL ALGORITHM: 
-#>   This procedure has not been thoroughly tested and may be unstable 
-#>   or buggy. The interface is subject to change. 
-#> ------------------------------------------------------------ 
-#> Gradient evaluation took 9e-06 seconds 
-#> 1000 transitions using 10 leapfrog steps per transition would take 0.09 seconds. 
-#> Adjust your expectations accordingly! 
-#> Begin eta adaptation. 
-#> Iteration:   1 / 250 [  0%]  (Adaptation) 
-#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
-#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
-#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
-#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
-#> Success! Found best value [eta = 1] earlier than expected. 
-#> Begin stochastic gradient ascent. 
-#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
-#>    100           -6.262             1.000            1.000 
-#>    200           -6.263             0.500            1.000 
-#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
-#> Drawing a sample of size 1000 from the approximate posterior...  
-#> COMPLETED. 
-#> Finished in  0.1 seconds.

-fit_vb$summary()
-
#> # A tibble: 3 × 7
-#>   variable      mean median    sd   mad     q5      q95
-#>   <chr>        <dbl>  <dbl> <dbl> <dbl>  <dbl>    <dbl>
-#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
-#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
-#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  

-# Plot approximate posterior using bayesplot
-mcmc_hist(fit_vb$draws("theta"))
-
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

-
-# Specifying initial values as a function
-fit_mcmc_w_init_fun <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function() list(theta = runif(1))
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2 <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = function(chain_id) {
-    # silly but demonstrates optional use of chain_id
-    list(theta = 1 / (chain_id + 1))
-  }
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_mcmc_w_init_fun_2$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.5
-#> 
-#> 
-#> [[2]]
-#> [[2]]$theta
-#> [1] 0.3333333
-#> 
-#> 

-# Specifying initial values as a list of lists
-fit_mcmc_w_init_list <- mod$sample(
-  data = stan_data,
-  seed = 123,
-  chains = 2,
-  refresh = 0,
-  init = list(
-    list(theta = 0.75), # chain 1
-    list(theta = 0.25)  # chain 2
-  )
-)
-
#> Running MCMC with 2 sequential chains...
-#> 
-#> Chain 1 finished in 0.0 seconds.
-#> Chain 2 finished in 0.0 seconds.
-#> 
-#> Both chains finished successfully.
-#> Mean chain execution time: 0.0 seconds.
-#> Total execution time: 0.3 seconds.
-#> 
fit_optim_w_init_list <- mod$optimize(
-  data = stan_data,
-  seed = 123,
-  init = list(
-    list(theta = 0.75)
-  )
-)
-
#> Initial log joint probability = -11.6657 
-#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
-#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
-#> Optimization terminated normally:  
-#>   Convergence detected: relative gradient magnitude is below tolerance 
-#> Finished in  0.1 seconds.
fit_optim_w_init_list$init()
-
#> [[1]]
-#> [[1]]$theta
-#> [1] 0.75
-#> 
-#> 
# }
-
-
+

The Stan and CmdStan documentation:

Other CmdStanModel methods: +model-method-check_syntax, +model-method-compile, +model-method-diagnose, +model-method-expose_functions, +model-method-format, +model-method-generate-quantities, +model-method-optimize, +model-method-sample_mpi, +model-method-sample, +model-method-variables

+
+ +
+

Examples

+ 
# \dontrun{
+library(cmdstanr)
+library(posterior)
+library(bayesplot)
+color_scheme_set("brightblue")
+
+# Set path to CmdStan
+# (Note: if you installed CmdStan via install_cmdstan() with default settings
+# then setting the path is unnecessary but the default below should still work.
+# Otherwise use the `path` argument to specify the location of your
+# CmdStan installation.)
+set_cmdstan_path(path = NULL)
+#> CmdStan path set to: /Users/jgabry/.cmdstan/cmdstan-2.32.2
+
+# Create a CmdStanModel object from a Stan program,
+# here using the example model that comes with CmdStan
+file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.stan")
+mod <- cmdstan_model(file)
+mod$print()
+#> data {
+#>   int<lower=0> N;
+#>   array[N] int<lower=0,upper=1> y;
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   theta ~ beta(1,1);  // uniform prior on interval 0,1
+#>   y ~ bernoulli(theta);
+#> }
+
+# Data as a named list (like RStan)
+stan_data <- list(N = 10, y = c(0,1,0,0,0,0,0,0,0,1))
+
+# Run MCMC using the 'sample' method
+fit_mcmc <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  parallel_chains = 2
+)
+#> Running MCMC with 2 parallel chains...
+#> 
+#> Chain 1 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 1 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 1 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 1 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 1 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 1 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 1 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 1 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 1 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 1 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 1 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 1 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 1 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 1 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 1 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 1 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 1 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 1 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 1 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 1 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 1 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 1 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 2 Iteration:    1 / 2000 [  0%]  (Warmup) 
+#> Chain 2 Iteration:  100 / 2000 [  5%]  (Warmup) 
+#> Chain 2 Iteration:  200 / 2000 [ 10%]  (Warmup) 
+#> Chain 2 Iteration:  300 / 2000 [ 15%]  (Warmup) 
+#> Chain 2 Iteration:  400 / 2000 [ 20%]  (Warmup) 
+#> Chain 2 Iteration:  500 / 2000 [ 25%]  (Warmup) 
+#> Chain 2 Iteration:  600 / 2000 [ 30%]  (Warmup) 
+#> Chain 2 Iteration:  700 / 2000 [ 35%]  (Warmup) 
+#> Chain 2 Iteration:  800 / 2000 [ 40%]  (Warmup) 
+#> Chain 2 Iteration:  900 / 2000 [ 45%]  (Warmup) 
+#> Chain 2 Iteration: 1000 / 2000 [ 50%]  (Warmup) 
+#> Chain 2 Iteration: 1001 / 2000 [ 50%]  (Sampling) 
+#> Chain 2 Iteration: 1100 / 2000 [ 55%]  (Sampling) 
+#> Chain 2 Iteration: 1200 / 2000 [ 60%]  (Sampling) 
+#> Chain 2 Iteration: 1300 / 2000 [ 65%]  (Sampling) 
+#> Chain 2 Iteration: 1400 / 2000 [ 70%]  (Sampling) 
+#> Chain 2 Iteration: 1500 / 2000 [ 75%]  (Sampling) 
+#> Chain 2 Iteration: 1600 / 2000 [ 80%]  (Sampling) 
+#> Chain 2 Iteration: 1700 / 2000 [ 85%]  (Sampling) 
+#> Chain 2 Iteration: 1800 / 2000 [ 90%]  (Sampling) 
+#> Chain 2 Iteration: 1900 / 2000 [ 95%]  (Sampling) 
+#> Chain 2 Iteration: 2000 / 2000 [100%]  (Sampling) 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.2 seconds.
+#> 
+
+# Use 'posterior' package for summaries
+fit_mcmc$summary()
+#> # A tibble: 2 × 10
+#>   variable   mean median    sd   mad      q5    q95  rhat ess_bulk ess_tail
+#>   <chr>     <num>  <num> <num> <num>   <num>  <num> <num>    <num>    <num>
+#> 1 lp__     -7.30  -7.03  0.721 0.380 -8.82   -6.75   1.00     902.    1006.
+#> 2 theta     0.247  0.233 0.122 0.129  0.0786  0.470  1.00     762.     712.
+
+# Get posterior draws
+draws <- fit_mcmc$draws()
+print(draws)
+#> # A draws_array: 1000 iterations, 2 chains, and 2 variables
+#> , , variable = lp__
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 -6.8 -6.8
+#>         2 -6.9 -6.8
+#>         3 -7.0 -7.0
+#>         4 -6.9 -7.1
+#>         5 -6.7 -7.0
+#> 
+#> , , variable = theta
+#> 
+#>          chain
+#> iteration    1    2
+#>         1 0.28 0.21
+#>         2 0.19 0.20
+#>         3 0.16 0.17
+#>         4 0.20 0.36
+#>         5 0.25 0.34
+#> 
+#> # ... with 995 more iterations
+
+# Convert to data frame using posterior::as_draws_df
+as_draws_df(draws)
+#> # A draws_df: 1000 iterations, 2 chains, and 2 variables
+#>    lp__ theta
+#> 1  -6.8  0.28
+#> 2  -6.9  0.19
+#> 3  -7.0  0.16
+#> 4  -6.9  0.20
+#> 5  -6.7  0.25
+#> 6  -7.1  0.36
+#> 7  -9.0  0.55
+#> 8  -7.2  0.15
+#> 9  -6.8  0.23
+#> 10 -7.5  0.42
+#> # ... with 1990 more draws
+#> # ... hidden reserved variables {'.chain', '.iteration', '.draw'}
+
+# Plot posterior using bayesplot (ggplot2)
+mcmc_hist(fit_mcmc$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+# Call CmdStan's diagnose and stansummary utilities
+fit_mcmc$cmdstan_diagnose()
+#> Processing csv files: /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251438-1-4ea737.csv, /var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/bernoulli-202307251438-2-4ea737.csv
+#> 
+#> Checking sampler transitions treedepth.
+#> Treedepth satisfactory for all transitions.
+#> 
+#> Checking sampler transitions for divergences.
+#> No divergent transitions found.
+#> 
+#> Checking E-BFMI - sampler transitions HMC potential energy.
+#> E-BFMI satisfactory.
+#> 
+#> Effective sample size satisfactory.
+#> 
+#> Split R-hat values satisfactory all parameters.
+#> 
+#> Processing complete, no problems detected.
+fit_mcmc$cmdstan_summary()
+#> Inference for Stan model: bernoulli_model
+#> 2 chains: each with iter=(1000,1000); warmup=(0,0); thin=(1,1); 2000 iterations saved.
+#> 
+#> Warmup took (0.0040, 0.0040) seconds, 0.0080 seconds total
+#> Sampling took (0.011, 0.011) seconds, 0.022 seconds total
+#> 
+#>                 Mean     MCSE  StdDev     5%   50%   95%  N_Eff  N_Eff/s    R_hat
+#> 
+#> lp__            -7.3  2.6e-02    0.72   -8.8  -7.0  -6.8    781    35502      1.0
+#> accept_stat__   0.92  8.3e-03    0.13   0.64  0.97   1.0    235    10662  1.0e+00
+#> stepsize__      0.95  7.9e-02   0.079   0.87   1.0   1.0    1.0       46  2.0e+13
+#> treedepth__      1.4  1.1e-02    0.48    1.0   1.0   2.0   1874    85179  1.0e+00
+#> n_leapfrog__     2.5  1.4e-01     1.3    1.0   3.0   3.0     89     4050  1.0e+00
+#> divergent__     0.00      nan    0.00   0.00  0.00  0.00    nan      nan      nan
+#> energy__         7.8  3.6e-02    1.00    6.8   7.5   9.6    775    35215  1.0e+00
+#> 
+#> theta           0.25  4.3e-03    0.12  0.079  0.23  0.47    796    36197      1.0
+#> 
+#> Samples were drawn using hmc with nuts.
+#> For each parameter, N_Eff is a crude measure of effective sample size,
+#> and R_hat is the potential scale reduction factor on split chains (at 
+#> convergence, R_hat=1).
+
+# For models fit using MCMC, if you like working with RStan's stanfit objects
+# then you can create one with rstan::read_stan_csv()
+
+# stanfit <- rstan::read_stan_csv(fit_mcmc$output_files())
+
+
+# Run 'optimize' method to get a point estimate (default is Stan's LBFGS algorithm)
+# and also demonstrate specifying data as a path to a file instead of a list
+my_data_file <- file.path(cmdstan_path(), "examples/bernoulli/bernoulli.data.json")
+fit_optim <- mod$optimize(data = my_data_file, seed = 123)
+#> Initial log joint probability = -9.51104 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000103557   2.55661e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+
+fit_optim$summary()
+#> # A tibble: 2 × 2
+#>   variable estimate
+#>   <chr>       <num>
+#> 1 lp__        -5.00
+#> 2 theta        0.2 
+
+
+# Run 'variational' method to approximate the posterior (default is meanfield ADVI)
+fit_vb <- mod$variational(data = stan_data, seed = 123)
+#> ------------------------------------------------------------ 
+#> EXPERIMENTAL ALGORITHM: 
+#>   This procedure has not been thoroughly tested and may be unstable 
+#>   or buggy. The interface is subject to change. 
+#> ------------------------------------------------------------ 
+#> Gradient evaluation took 9e-06 seconds 
+#> 1000 transitions using 10 leapfrog steps per transition would take 0.09 seconds. 
+#> Adjust your expectations accordingly! 
+#> Begin eta adaptation. 
+#> Iteration:   1 / 250 [  0%]  (Adaptation) 
+#> Iteration:  50 / 250 [ 20%]  (Adaptation) 
+#> Iteration: 100 / 250 [ 40%]  (Adaptation) 
+#> Iteration: 150 / 250 [ 60%]  (Adaptation) 
+#> Iteration: 200 / 250 [ 80%]  (Adaptation) 
+#> Success! Found best value [eta = 1] earlier than expected. 
+#> Begin stochastic gradient ascent. 
+#>   iter             ELBO   delta_ELBO_mean   delta_ELBO_med   notes  
+#>    100           -6.262             1.000            1.000 
+#>    200           -6.263             0.500            1.000 
+#>    300           -6.307             0.336            0.007   MEDIAN ELBO CONVERGED 
+#> Drawing a sample of size 1000 from the approximate posterior...  
+#> COMPLETED. 
+#> Finished in  0.1 seconds.
+
+fit_vb$summary()
+#> # A tibble: 3 × 7
+#>   variable      mean median    sd   mad     q5      q95
+#>   <chr>        <num>  <num> <num> <num>  <num>    <num>
+#> 1 lp__        -7.18  -6.94  0.588 0.259 -8.36  -6.75   
+#> 2 lp_approx__ -0.515 -0.221 0.692 0.303 -2.06  -0.00257
+#> 3 theta        0.263  0.246 0.115 0.113  0.106  0.481  
+
+# Plot approximate posterior using bayesplot
+mcmc_hist(fit_vb$draws("theta"))
+#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
+
+
+
+# Specifying initial values as a function
+fit_mcmc_w_init_fun <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function() list(theta = runif(1))
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2 <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = function(chain_id) {
+    # silly but demonstrates optional use of chain_id
+    list(theta = 1 / (chain_id + 1))
+  }
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_mcmc_w_init_fun_2$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.5
+#> 
+#> 
+#> [[2]]
+#> [[2]]$theta
+#> [1] 0.3333333
+#> 
+#> 
+
+# Specifying initial values as a list of lists
+fit_mcmc_w_init_list <- mod$sample(
+  data = stan_data,
+  seed = 123,
+  chains = 2,
+  refresh = 0,
+  init = list(
+    list(theta = 0.75), # chain 1
+    list(theta = 0.25)  # chain 2
+  )
+)
+#> Running MCMC with 2 sequential chains...
+#> 
+#> Chain 1 finished in 0.0 seconds.
+#> Chain 2 finished in 0.0 seconds.
+#> 
+#> Both chains finished successfully.
+#> Mean chain execution time: 0.0 seconds.
+#> Total execution time: 0.3 seconds.
+#> 
+fit_optim_w_init_list <- mod$optimize(
+  data = stan_data,
+  seed = 123,
+  init = list(
+    list(theta = 0.75)
+  )
+)
+#> Initial log joint probability = -11.6657 
+#>     Iter      log prob        ||dx||      ||grad||       alpha      alpha0  # evals  Notes  
+#>        6      -5.00402   0.000237915   9.55309e-07           1           1        9    
+#> Optimization terminated normally:  
+#>   Convergence detected: relative gradient magnitude is below tolerance 
+#> Finished in  0.1 seconds.
+fit_optim_w_init_list$init()
+#> [[1]]
+#> [[1]]$theta
+#> [1] 0.75
+#> 
+#> 
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/read_cmdstan_csv.html b/docs/reference/read_cmdstan_csv.html index ea17d8efc..2106577ef 100644 --- a/docs/reference/read_cmdstan_csv.html +++ b/docs/reference/read_cmdstan_csv.html @@ -1,80 +1,17 @@ - - - - - - - -Read CmdStan CSV files into R — read_cmdstan_csv • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Read CmdStan CSV files into R — read_cmdstan_csv • cmdstanr - - - - - - - - - - - + + - - - -
-
- -
- -
+

Read CmdStan CSV files into R

- Source: R/csv.R + Source: R/csv.R

read_cmdstan_csv() is used internally by CmdStanR to read CmdStan's output CSV files into R. It can also be used by CmdStan users as -a more flexible and efficient alternative to rstan::read_stan_csv(). See +a more flexible and efficient alternative to rstan::read_stan_csv(). See the Value section for details on the structure of the returned list.

It is also possible to create CmdStanR's fitted model objects directly from CmdStan CSV files using the as_cmdstan_fit() function.

- 
read_cmdstan_csv(
-  files,
-  variables = NULL,
-  sampler_diagnostics = NULL,
-  format = getOption("cmdstanr_draws_format", NULL)
-)
+    

+    
read_cmdstan_csv(
+  files,
+  variables = NULL,
+  sampler_diagnostics = NULL,
+  format = getOption("cmdstanr_draws_format", NULL)
+)
+
+as_cmdstan_fit(
+  files,
+  check_diagnostics = TRUE,
+  format = getOption("cmdstanr_draws_format")
+)

+    

+
+    

+    
Arguments

+    
files

+
(character vector) The paths to the CmdStan CSV files. These can
+be files generated by running CmdStanR or running CmdStan directly.

 
-as_cmdstan_fit(
-  files,
-  check_diagnostics = TRUE,
-  format = getOption("cmdstanr_draws_format")
-)
-

Arguments

- - - - - - - - - - - - - - - - - - - - - - -
files

(character vector) The paths to the CmdStan CSV files. These can -be files generated by running CmdStanR or running CmdStan directly.

variables

(character vector) Optionally, the names of the variables -(parameters, transformed parameters, and generated quantities) to read in.

    -

  • If NULL (the default) then all variables are included.

    • +
    variables
    +

    (character vector) Optionally, the names of the variables +(parameters, transformed parameters, and generated quantities) to read in.

    • If NULL (the default) then all variables are included.

    • If an empty string (variables="") then none are included.

      • -

    • For non-scalar variables all elements or specific elements can be selected:

        -

      • variables = "theta" selects all elements of theta;

        • +

      • For non-scalar variables all elements or specific elements can be selected:

        • variables = "theta" selects all elements of theta;

        • variables = c("theta[1]", "theta[3]") selects only the 1st and 3rd elements.

        • -
sampler_diagnostics

(character vector) Works the same way as + + + +

sampler_diagnostics
+

(character vector) Works the same way as variables but for sampler diagnostic variables (e.g., "treedepth__", -"accept_stat__", etc.). Ignored if the model was not fit using MCMC.

format

(string) The format for storing the draws or point estimates. +"accept_stat__", etc.). Ignored if the model was not fit using MCMC.

+ + +
format
+

(string) The format for storing the draws or point estimates. The default depends on the method used to fit the model. See -draws for details, in particular the note about speed -and memory for models with many parameters.

check_diagnostics

(logical) For models fit using MCMC, should +draws for details, in particular the note about speed +and memory for models with many parameters.

+ + +
check_diagnostics
+

(logical) For models fit using MCMC, should diagnostic checks be performed after reading in the files? The default is TRUE but set to FALSE to avoid checking for problems with divergences -and treedepth.

+and treedepth.

-

Value

+
+
+

Value

+ -

as_cmdstan_fit() returns a CmdStanMCMC, CmdStanMLE, or -CmdStanVB object. Some methods typically defined for those objects will not -work (e.g. save_data_file()) but the important methods like $summary(), +

as_cmdstan_fit() returns a CmdStanMCMC, CmdStanMLE, or +CmdStanVB object. Some methods typically defined for those objects will not +work (e.g. save_data_file()) but the important methods like $summary(), $draws(), $sampler_diagnostics() and others will work fine.

-

read_cmdstan_csv() returns a named list with the following components:

    -

  • metadata: A list of the meta information from the run that produced the -CSV file(s). See Examples below.

    • -
-

The other components in the returned list depend on the method that produced + +

read_cmdstan_csv() returns a named list with the following components:

  • metadata: A list of the meta information from the run that produced the +CSV file(s). See Examples below.

    • +

The other components in the returned list depend on the method that produced the CSV file(s).

-

For sampling the returned list also includes the -following components:

    -

  • time: Run time information for the individual chains. The returned object -is the same as for the $time() method except the total run + + +

    For sampling the returned list also includes the +following components:

    • time: Run time information for the individual chains. The returned object +is the same as for the $time() method except the total run time can't be inferred from the CSV files (the chains may have been run in parallel) and is therefore NA.

    • inv_metric: A list (one element per chain) of inverse mass matrices or their diagonals, depending on the type of metric used.

    • step_size: A list (one element per chain) of the step sizes used.

    • warmup_draws: If save_warmup was TRUE when fitting the model then a -draws_array (or different format if format is +draws_array (or different format if format is specified) of warmup draws.

      • -

    • post_warmup_draws: A draws_array (or +

    • post_warmup_draws: A draws_array (or different format if format is specified) of post-warmup draws.

    • warmup_sampler_diagnostics: If save_warmup was TRUE when fitting the -model then a draws_array (or different format if +model then a draws_array (or different format if format is specified) of warmup draws of the sampler diagnostic variables.

    • post_warmup_sampler_diagnostics: A -draws_array (or different format if format is +draws_array (or different format if format is specified) of post-warmup draws of the sampler diagnostic variables.

      • -
    - -

    For optimization the returned list also includes the -following components:

      -

    • point_estimates: Point estimates for the model parameters.

      • -
    - -

    For variational inference the returned list also -includes the following components:

    For optimization the returned list also includes the +following components:

    • point_estimates: Point estimates for the model parameters.

      • +

    For variational inference the returned list also +includes the following components:

    • draws: A draws_matrix (or different format if format is specified) of draws from the approximate posterior distribution.

      • -
    - -

    For standalone generated quantities the -returned list also includes the following components:

    For standalone generated quantities the +returned list also includes the following components:

    • generated_quantities: A draws_array of the generated quantities.

      • -
    - - -

    Examples

    - 
    # \dontrun{
-# Generate some CSV files to use for demonstration
-fit1 <- cmdstanr_example("logistic", method = "sample", save_warmup = TRUE)
-csv_files <- fit1$output_files()
-print(csv_files)
-
    #> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181227-1-726257.csv"
-#> [2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181227-2-726257.csv"
-#> [3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181227-3-726257.csv"
-#> [4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-202203181227-4-726257.csv"
    
-# Creating fitting model objects
-
-# Create a CmdStanMCMC object from the CSV files
-fit2 <- as_cmdstan_fit(csv_files)
-fit2$print("beta")
-
    #>  variable  mean median   sd  mad    q5   q95 rhat ess_bulk ess_tail
-#>   beta[1] -0.67  -0.66 0.24 0.24 -1.08 -0.28 1.00     4045     3169
-#>   beta[2] -0.27  -0.27 0.22 0.22 -0.64  0.08 1.00     3894     2721
-#>   beta[3]  0.69   0.68 0.26 0.26  0.27  1.13 1.00     3754     2922
    
-# Using read_cmdstan_csv
-#
-# Read in everything
-x <- read_cmdstan_csv(csv_files)
-str(x)
-
    #> List of 8
-#>  $ metadata                       :List of 40
-#>   ..$ stan_version_major  : num 2
-#>   ..$ stan_version_minor  : num 29
-#>   ..$ stan_version_patch  : num 1
-#>   ..$ start_datetime      : chr "2022-03-18 18:27:08 UTC"
-#>   ..$ method              : chr "sample"
-#>   ..$ save_warmup         : num 1
-#>   ..$ thin                : num 1
-#>   ..$ gamma               : num 0.05
-#>   ..$ kappa               : num 0.75
-#>   ..$ t0                  : num 10
-#>   ..$ init_buffer         : num 75
-#>   ..$ term_buffer         : num 50
-#>   ..$ window              : num 25
-#>   ..$ algorithm           : chr "hmc"
-#>   ..$ engine              : chr "nuts"
-#>   ..$ metric              : chr "diag_e"
-#>   ..$ stepsize_jitter     : num 0
-#>   ..$ num_chains          : num 1
-#>   ..$ id                  : num [1:4] 1 2 3 4
-#>   ..$ init                : num [1:4] 2 2 2 2
-#>   ..$ seed                : num 27467875
-#>   ..$ refresh             : num 100
-#>   ..$ sig_figs            : num -1
-#>   ..$ profile_file        : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpmzUYEz/logistic-profile-202203181227-1-051c26.csv"
-#>   ..$ stanc_version       : chr "stanc3 v2.29.1"
-#>   ..$ sampler_diagnostics : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
-#>   ..$ variables           : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
-#>   ..$ step_size_adaptation: num [1:4] 0.729 0.767 0.747 0.752
-#>   ..$ model_name          : chr "logistic_model"
-#>   ..$ adapt_engaged       : num 1
-#>   ..$ adapt_delta         : num 0.8
-#>   ..$ max_treedepth       : num 10
-#>   ..$ step_size           : num [1:4] 1 1 1 1
-#>   ..$ iter_warmup         : num 1000
-#>   ..$ iter_sampling       : num 1000
-#>   ..$ threads_per_chain   : num 1
-#>   ..$ time                :'data.frame':	4 obs. of  4 variables:
-#>   .. ..$ chain_id: num [1:4] 1 2 3 4
-#>   .. ..$ warmup  : num [1:4] 0.093 0.087 0.149 0.092
-#>   .. ..$ sampling: num [1:4] 0.081 0.096 0.1 0.087
-#>   .. ..$ total   : num [1:4] 0.174 0.183 0.249 0.179
-#>   ..$ stan_variable_sizes :List of 4
-#>   .. ..$ lp__   : num 1
-#>   .. ..$ alpha  : num 1
-#>   .. ..$ beta   : num 3
-#>   .. ..$ log_lik: num 100
-#>   ..$ stan_variables      : chr [1:4] "lp__" "alpha" "beta" "log_lik"
-#>   ..$ model_params        : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
-#>  $ time                           :List of 2
-#>   ..$ total : int NA
-#>   ..$ chains:'data.frame':	4 obs. of  4 variables:
-#>   .. ..$ chain_id: num [1:4] 1 2 3 4
-#>   .. ..$ warmup  : num [1:4] 0.093 0.087 0.149 0.092
-#>   .. ..$ sampling: num [1:4] 0.081 0.096 0.1 0.087
-#>   .. ..$ total   : num [1:4] 0.174 0.183 0.249 0.179
-#>  $ inv_metric                     :List of 4
-#>   ..$ 1: num [1:4] 0.046 0.0637 0.0532 0.0736
-#>   ..$ 2: num [1:4] 0.0421 0.0566 0.0523 0.0756
-#>   ..$ 3: num [1:4] 0.0493 0.0528 0.0523 0.0753
-#>   ..$ 4: num [1:4] 0.0365 0.0565 0.0397 0.0632
-#>  $ step_size                      :List of 4
-#>   ..$ 1: num 0.729
-#>   ..$ 2: num 0.767
-#>   ..$ 3: num 0.747
-#>   ..$ 4: num 0.752
-#>  $ warmup_draws                   : 'draws_array' num [1:1000, 1:4, 1:105] -66.8 -66.8 -66.8 -65.8 -66.3 ...
-#>   ..- attr(*, "dimnames")=List of 3
-#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   .. ..$ variable : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
-#>  $ post_warmup_draws              : 'draws_array' num [1:1000, 1:4, 1:105] -65 -65.7 -64.4 -64.2 -65.3 ...
-#>   ..- attr(*, "dimnames")=List of 3
-#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   .. ..$ variable : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
-#>  $ warmup_sampler_diagnostics     : 'draws_array' num [1:1000, 1:4, 1:6] 1 0 0 0.941 0.933 ...
-#>   ..- attr(*, "dimnames")=List of 3
-#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   .. ..$ variable : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
-#>  $ post_warmup_sampler_diagnostics: 'draws_array' num [1:1000, 1:4, 1:6] 1 0.916 0.998 0.972 0.911 ...
-#>   ..- attr(*, "dimnames")=List of 3
-#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
-#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
-#>   .. ..$ variable : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
    
-# Don't read in any of the sampler diagnostic variables
-x <- read_cmdstan_csv(csv_files, sampler_diagnostics = "")
-
-# Don't read in any of the parameters or generated quantities
-x <- read_cmdstan_csv(csv_files, variables = "")
-
-# Read in only specific parameters and sampler diagnostics
-x <- read_cmdstan_csv(
-  csv_files,
-  variables = c("alpha", "beta[2]"),
-  sampler_diagnostics = c("n_leapfrog__", "accept_stat__")
-)
-
-# For non-scalar parameters all elements can be selected or only some elements,
-# e.g. all of the vector "beta" but only one element of the vector "log_lik"
-x <- read_cmdstan_csv(
-  csv_files,
-  variables = c("beta", "log_lik[3]")
-)
-# }
-
-
    +
+ +
+

Examples

+ 
# \dontrun{
+# Generate some CSV files to use for demonstration
+fit1 <- cmdstanr_example("logistic", method = "sample", save_warmup = TRUE)
+csv_files <- fit1$output_files()
+print(csv_files)
+#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251438-1-0afc76.csv"
+#> [2] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251438-2-0afc76.csv"
+#> [3] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251438-3-0afc76.csv"
+#> [4] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-202307251438-4-0afc76.csv"
+
+# Creating fitting model objects
+
+# Create a CmdStanMCMC object from the CSV files
+fit2 <- as_cmdstan_fit(csv_files)
+fit2$print("beta")
+#>  variable  mean median   sd  mad    q5   q95 rhat ess_bulk ess_tail
+#>   beta[1] -0.67  -0.66 0.25 0.25 -1.09 -0.26 1.00     4467     2955
+#>   beta[2] -0.27  -0.27 0.22 0.22 -0.63  0.09 1.00     4244     3209
+#>   beta[3]  0.68   0.68 0.26 0.26  0.25  1.12 1.00     4340     3315
+
+# Using read_cmdstan_csv
+#
+# Read in everything
+x <- read_cmdstan_csv(csv_files)
+str(x)
+#> List of 8
+#>  $ metadata                       :List of 40
+#>   ..$ stan_version_major  : num 2
+#>   ..$ stan_version_minor  : num 32
+#>   ..$ stan_version_patch  : num 2
+#>   ..$ start_datetime      : chr "2023-07-25 20:38:31 UTC"
+#>   ..$ method              : chr "sample"
+#>   ..$ save_warmup         : num 1
+#>   ..$ thin                : num 1
+#>   ..$ gamma               : num 0.05
+#>   ..$ kappa               : num 0.75
+#>   ..$ t0                  : num 10
+#>   ..$ init_buffer         : num 75
+#>   ..$ term_buffer         : num 50
+#>   ..$ window              : num 25
+#>   ..$ algorithm           : chr "hmc"
+#>   ..$ engine              : chr "nuts"
+#>   ..$ metric              : chr "diag_e"
+#>   ..$ stepsize_jitter     : num 0
+#>   ..$ num_chains          : num 1
+#>   ..$ id                  : num [1:4] 1 2 3 4
+#>   ..$ init                : num [1:4] 2 2 2 2
+#>   ..$ seed                : num 1.15e+09
+#>   ..$ refresh             : num 100
+#>   ..$ sig_figs            : num -1
+#>   ..$ profile_file        : chr "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/logistic-profile-202307251438-1-5569d9.csv"
+#>   ..$ stanc_version       : chr "stanc3 v2.32.2"
+#>   ..$ sampler_diagnostics : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
+#>   ..$ variables           : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+#>   ..$ step_size_adaptation: num [1:4] 0.666 0.702 0.658 0.727
+#>   ..$ model_name          : chr "logistic_model"
+#>   ..$ adapt_engaged       : num 1
+#>   ..$ adapt_delta         : num 0.8
+#>   ..$ max_treedepth       : num 10
+#>   ..$ step_size           : num [1:4] 1 1 1 1
+#>   ..$ iter_warmup         : num 1000
+#>   ..$ iter_sampling       : num 1000
+#>   ..$ threads_per_chain   : num 1
+#>   ..$ time                :'data.frame':	4 obs. of  4 variables:
+#>   .. ..$ chain_id: num [1:4] 1 2 3 4
+#>   .. ..$ warmup  : num [1:4] 0.07 0.081 0.077 0.071
+#>   .. ..$ sampling: num [1:4] 0.068 0.075 0.065 0.062
+#>   .. ..$ total   : num [1:4] 0.138 0.156 0.142 0.133
+#>   ..$ stan_variable_sizes :List of 4
+#>   .. ..$ lp__   : num 1
+#>   .. ..$ alpha  : num 1
+#>   .. ..$ beta   : num 3
+#>   .. ..$ log_lik: num 100
+#>   ..$ stan_variables      : chr [1:4] "lp__" "alpha" "beta" "log_lik"
+#>   ..$ model_params        : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+#>  $ time                           :List of 2
+#>   ..$ total : int NA
+#>   ..$ chains:'data.frame':	4 obs. of  4 variables:
+#>   .. ..$ chain_id: num [1:4] 1 2 3 4
+#>   .. ..$ warmup  : num [1:4] 0.07 0.081 0.077 0.071
+#>   .. ..$ sampling: num [1:4] 0.068 0.075 0.065 0.062
+#>   .. ..$ total   : num [1:4] 0.138 0.156 0.142 0.133
+#>  $ inv_metric                     :List of 4
+#>   ..$ 1: num [1:4] 0.0426 0.0631 0.0525 0.0739
+#>   ..$ 2: num [1:4] 0.0483 0.0658 0.0426 0.065
+#>   ..$ 3: num [1:4] 0.0473 0.0591 0.0532 0.0814
+#>   ..$ 4: num [1:4] 0.0485 0.0583 0.05 0.0707
+#>  $ step_size                      :List of 4
+#>   ..$ 1: num 0.666
+#>   ..$ 2: num 0.702
+#>   ..$ 3: num 0.658
+#>   ..$ 4: num 0.727
+#>  $ warmup_draws                   : 'draws_array' num [1:1000, 1:4, 1:105] -105.5 -105.5 -105.5 -70.7 -66.2 ...
+#>   ..- attr(*, "dimnames")=List of 3
+#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   .. ..$ variable : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+#>  $ post_warmup_draws              : 'draws_array' num [1:1000, 1:4, 1:105] -65.2 -64.4 -65.4 -64.4 -70.3 ...
+#>   ..- attr(*, "dimnames")=List of 3
+#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   .. ..$ variable : chr [1:105] "lp__" "alpha" "beta[1]" "beta[2]" ...
+#>  $ warmup_sampler_diagnostics     : 'draws_array' num [1:1000, 1:4, 1:6] 6.68e-01 0.00 1.27e-204 9.99e-01 9.89e-01 ...
+#>   ..- attr(*, "dimnames")=List of 3
+#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   .. ..$ variable : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
+#>  $ post_warmup_sampler_diagnostics: 'draws_array' num [1:1000, 1:4, 1:6] 0.997 0.935 0.852 0.98 0.644 ...
+#>   ..- attr(*, "dimnames")=List of 3
+#>   .. ..$ iteration: chr [1:1000] "1" "2" "3" "4" ...
+#>   .. ..$ chain    : chr [1:4] "1" "2" "3" "4"
+#>   .. ..$ variable : chr [1:6] "accept_stat__" "stepsize__" "treedepth__" "n_leapfrog__" ...
+
+# Don't read in any of the sampler diagnostic variables
+x <- read_cmdstan_csv(csv_files, sampler_diagnostics = "")
+
+# Don't read in any of the parameters or generated quantities
+x <- read_cmdstan_csv(csv_files, variables = "")
+
+# Read in only specific parameters and sampler diagnostics
+x <- read_cmdstan_csv(
+  csv_files,
+  variables = c("alpha", "beta[2]"),
+  sampler_diagnostics = c("n_leapfrog__", "accept_stat__")
+)
+
+# For non-scalar parameters all elements can be selected or only some elements,
+# e.g. all of the vector "beta" but only one element of the vector "log_lik"
+x <- read_cmdstan_csv(
+  csv_files,
+  variables = c("beta", "log_lik[3]")
+)
+# }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/read_sample_csv.html b/docs/reference/read_sample_csv.html index e06b05c61..af34b60aa 100644 --- a/docs/reference/read_sample_csv.html +++ b/docs/reference/read_sample_csv.html @@ -1,75 +1,12 @@ - - - - - - - -Read CmdStan CSV files from sampling into R — read_sample_csv • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Read CmdStan CSV files from sampling into R — read_sample_csv • cmdstanr - - - - + + -
-
- -
- -
+

Read CmdStan CSV files from sampling into R

- Source: R/csv.R + Source: R/csv.R
-

Deprecated. Use read_cmdstan_csv() instead.

+

Deprecated. Use read_cmdstan_csv() instead.

- 
read_sample_csv(files, variables = NULL, sampler_diagnostics = NULL)
+
+ 
read_sample_csv(files, variables = NULL, sampler_diagnostics = NULL)
+
-

Arguments

- - - - - - -
files, variables, sampler_diagnostics

Deprecated. Use -read_cmdstan_csv() instead.

+
+

Arguments

+
files, variables, sampler_diagnostics
+

Deprecated. Use +read_cmdstan_csv() instead.

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/register_knitr_engine.html b/docs/reference/register_knitr_engine.html index 10533e5c4..1e4517702 100644 --- a/docs/reference/register_knitr_engine.html +++ b/docs/reference/register_knitr_engine.html @@ -1,78 +1,15 @@ - - - - - - - -Register CmdStanR's knitr engine for Stan — register_knitr_engine • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Register CmdStanR's knitr engine for Stan — register_knitr_engine • cmdstanr - - - - - - - - - - - - - - + + -
-
- -
- -
+

Register CmdStanR's knitr engine for Stan

- Source: R/knitr.R + Source: R/knitr.R
-

Registers CmdStanR's knitr engine eng_cmdstan() for processing Stan chunks. +

Registers CmdStanR's knitr engine eng_cmdstan() for processing Stan chunks. Refer to the vignette -R Markdown CmdStan Engine +R Markdown CmdStan Engine for a demonstration.

- 
register_knitr_engine(override = TRUE)
- -

Arguments

- - - - - - -
override

(logical) Override knitr's built-in, RStan-based engine for -Stan? The default is TRUE. See Details.

+
+ 
register_knitr_engine(override = TRUE)
+
-

Details

+
+

Arguments

+
override
+

(logical) Override knitr's built-in, RStan-based engine for +Stan? The default is TRUE. See Details.

+
+
+

Details

If override = TRUE (default), this registers CmdStanR's knitr engine as the engine for stan chunks, replacing knitr's built-in, RStan-based engine. If override = FALSE, this registers a cmdstan engine so that both engines @@ -206,10 +133,10 @@

Details highlighting for the Stan language, the cmdstan chunks will have stan syntax highlighting applied to them.

See the vignette -R Markdown CmdStan Engine +R Markdown CmdStan Engine for an example.

Note: When running chunks interactively in RStudio (e.g. when using -R Notebooks), it has +R Notebooks), it has been observed that the built-in, RStan-based engine is used for stan chunks even when CmdStanR's engine has been registered in the session. When the R Markdown document is knit/rendered, the correct engine is used. As a @@ -218,40 +145,35 @@

Details

If you would like to keep stan chunks as stan chunks, it is possible to specify engine = "cmdstan" in the chunk options after registering the cmdstan engine with override = FALSE.

-

References

- +
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/set_cmdstan_path.html b/docs/reference/set_cmdstan_path.html index 891803ca4..a91b01e82 100644 --- a/docs/reference/set_cmdstan_path.html +++ b/docs/reference/set_cmdstan_path.html @@ -1,79 +1,16 @@ - - - - - - - -Get or set the file path to the CmdStan installation — set_cmdstan_path • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Get or set the file path to the CmdStan installation — set_cmdstan_path • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

Get or set the file path to the CmdStan installation

- Source: R/path.R + Source: R/path.R
@@ -187,80 +115,82 @@

Get or set the file path to the CmdStan installation

for how to avoid manually setting the path in each R session.

- 
set_cmdstan_path(path = NULL)
+    

+    
set_cmdstan_path(path = NULL)
+
+cmdstan_path()
+
+cmdstan_version(error_on_NA = TRUE)

+    

 
-cmdstan_path()
+    

+    
Arguments

+    
path

+
(string) The full file path to the CmdStan installation. If
+NULL (the default) then the path is set to the default path used by
+install_cmdstan() if it exists.

 
-cmdstan_version(error_on_NA = TRUE)
-

Arguments

- - - - - - - - - - -
path

(string) The full file path to the CmdStan installation. If -NULL (the default) then the path is set to the default path used by -install_cmdstan() if it exists.

error_on_NA

(logical) Should an error be thrown if CmdStan is not +

error_on_NA
+

(logical) Should an error be thrown if CmdStan is not found. The default is TRUE. If FALSE, cmdstan_version() returns -NULL.

+NULL.

-

Value

+
+
+

Value

+ -

A string. Either the file path to the CmdStan installation or the +

A string. Either the file path to the CmdStan installation or the CmdStan version number.

+ +

CmdStan version string if available. If CmdStan is not found and error_on_NA is FALSE, cmdstan_version() returns NULL.

-

Details

- +
+
+

Details

Before the package can be used it needs to know where the CmdStan installation is located. When the package is loaded it tries to help automate -this to avoid having to manually set the path every session:

    -

  • If the environment variable "CMDSTAN" exists at load time +this to avoid having to manually set the path every session:

    • If the environment variable "CMDSTAN" exists at load time then its value will be automatically set as the default path to CmdStan for -the R session.

      • +the R session. If the environment variable "CMDSTAN" is set, but a valid +CmdStan is not found in the supplied path, the path is treated as a top +folder that contains CmdStan installations. In that case, the CmdStan +installation with the largest version number will be set as the path to +CmdStan for the R session.

    • If no environment variable is found when loaded but any directory in the form ".cmdstan/cmdstan-[version]" (e.g., ".cmdstan/cmdstan-2.23.0"), -exists in the user's home directory (Sys.getenv("HOME"), not the current +exists in the user's home directory (Sys.getenv("HOME"), not the current working directory) then the path to the cmdstan with the largest version number will be set as the path to CmdStan for the R session. This is the -same as the default directory that install_cmdstan() would use to install +same as the default directory that install_cmdstan() would use to install the latest version of CmdStan.

      • -
    - -

    It is always possible to change the path after loading the package using +

It is always possible to change the path after loading the package using set_cmdstan_path(path).

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/stan_threads.html b/docs/reference/stan_threads.html index 633442c4a..eeeb82fae 100644 --- a/docs/reference/stan_threads.html +++ b/docs/reference/stan_threads.html @@ -1,75 +1,12 @@ - - - - - - - -Set or get the number of threads used to execute Stan models — stan_threads • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Set or get the number of threads used to execute Stan models — stan_threads • cmdstanr - + + - - - -
-
- -
- -
+

Set or get the number of threads used to execute Stan models

- Source: R/utils.R + Source: R/utils.R
@@ -179,48 +107,46 @@

Set or get the number of threads used to execute Stan models

DEPRECATED. Please use the threads_per_chain argument when fitting the model.

- 
num_threads()
-
-set_num_threads(num_threads)
+
+ 
num_threads()
+
+set_num_threads(num_threads)
+
-

Arguments

- - - - - - -
num_threads

(positive integer) The number of threads to set.

+
+

Arguments

+
num_threads
+

(positive integer) The number of threads to set.

-

Value

+
+
+

Value

+ -

The value of the environment variable STAN_NUM_THREADS.

+

The value of the environment variable STAN_NUM_THREADS.

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/write_stan_file.html b/docs/reference/write_stan_file.html index dae2b825b..9b772e61c 100644 --- a/docs/reference/write_stan_file.html +++ b/docs/reference/write_stan_file.html @@ -1,82 +1,19 @@ - - - - - - - -Write Stan code to a file — write_stan_file • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Write Stan code to a file — write_stan_file • cmdstanr - - - - - - - - - - + + - - - - -
-
- -
- -
+

Write Stan code to a file

- Source: R/example.R + Source: R/example.R

Convenience function for writing Stan code to a (possibly -temporary) file with a .stan extension. By default, the -file name is chosen deterministically based on a hash +temporary) file with a .stan extension. By default, the +file name is chosen deterministically based on a hash of the Stan code, and the file is not overwritten if it already has correct contents. This means that calling this function multiple times with the same Stan code will reuse the compiled model. This also however means that the @@ -193,121 +121,125 @@

Write Stan code to a file

should ensure thread-safety in the rare cases when it is needed.

- 
write_stan_file(
-  code,
-  dir = getOption("cmdstanr_write_stan_file_dir", tempdir()),
-  basename = NULL,
-  force_overwrite = FALSE,
-  hash_salt = ""
-)
+
+ 
write_stan_file(
+  code,
+  dir = getOption("cmdstanr_write_stan_file_dir", tempdir()),
+  basename = NULL,
+  force_overwrite = FALSE,
+  hash_salt = ""
+)
+
-

Arguments

- - - - - - - - - - - - - - - - - - - - - - -
code

(character vector) The Stan code to write to the file. This can +

+

Arguments

+
code
+

(character vector) The Stan code to write to the file. This can be a character vector of length one (a string) containing the entire Stan program or a character vector with each element containing one line of the -Stan program.

dir

(string) An optional path to the directory where the file will be +Stan program.

+ + +
dir
+

(string) An optional path to the directory where the file will be written. If omitted, a global option cmdstanr_write_stan_file_dir is -used. If the global options is not set, temporary directory -is used.

basename

(string) If dir is specified, optionally the basename to +used. If the global options is not set, temporary directory +is used.

+ + +
basename
+

(string) If dir is specified, optionally the basename to use for the file created. If not specified a file name is generated -from hashing the code.

force_overwrite

(logical) If set to TRUE the file will always be -overwritten and thus the resulting model will always be recompiled.

hash_salt

(string) Text to add to the model code prior to hashing to -determine the file name if basename is not set.

+from hashing the code.

+ -

Value

+
force_overwrite
+

(logical) If set to TRUE the file will always be +overwritten and thus the resulting model will always be recompiled.

-

The path to the file.

-

Examples

- 
# stan program as a single string
-stan_program <- "
-data {
-  int<lower=0> N;
-  int<lower=0,upper=1> y[N];
-}
-parameters {
-  real<lower=0,upper=1> theta;
-}
-model {
-  y ~ bernoulli(theta);
-}
-"
+
hash_salt

+
(string) Text to add to the model code prior to hashing to
+determine the file name if basename is not set.

 
-f <- write_stan_file(stan_program)
-print(f)
-
#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T//RtmpmzUYEz/model_08f9e456ca04f3d3244db00f16ea5748.stan"

-lines <- readLines(f)
-print(lines)
-
#>  [1] ""                               "data {"                        
-#>  [3] "  int<lower=0> N;"              "  int<lower=0,upper=1> y[N];"  
-#>  [5] "}"                              "parameters {"                  
-#>  [7] "  real<lower=0,upper=1> theta;" "}"                             
-#>  [9] "model {"                        "  y ~ bernoulli(theta);"       
-#> [11] "}"                              ""                              
cat(lines, sep = "\n")
-
#> 
-#> data {
-#>   int<lower=0> N;
-#>   int<lower=0,upper=1> y[N];
-#> }
-#> parameters {
-#>   real<lower=0,upper=1> theta;
-#> }
-#> model {
-#>   y ~ bernoulli(theta);
-#> }
-#> 

-# stan program as character vector of lines
-f2 <- write_stan_file(lines)
-identical(readLines(f), readLines(f2))
-
#> [1] TRUE

-
+
+
+

Value

+ + +

The path to the file.

+
+ +
+

Examples

+ 
# stan program as a single string
+stan_program <- "
+data {
+  int<lower=0> N;
+  int<lower=0,upper=1> y[N];
+}
+parameters {
+  real<lower=0,upper=1> theta;
+}
+model {
+  y ~ bernoulli(theta);
+}
+"
+
+f <- write_stan_file(stan_program)
+print(f)
+#> [1] "/var/folders/s0/zfzm55px2nd2v__zlw5xfj2h0000gn/T/RtmpFBtN6X/model_08f9e456ca04f3d3244db00f16ea5748.stan"
+
+lines <- readLines(f)
+print(lines)
+#>  [1] ""                               "data {"                        
+#>  [3] "  int<lower=0> N;"              "  int<lower=0,upper=1> y[N];"  
+#>  [5] "}"                              "parameters {"                  
+#>  [7] "  real<lower=0,upper=1> theta;" "}"                             
+#>  [9] "model {"                        "  y ~ bernoulli(theta);"       
+#> [11] "}"                              ""                              
+cat(lines, sep = "\n")
+#> 
+#> data {
+#>   int<lower=0> N;
+#>   int<lower=0,upper=1> y[N];
+#> }
+#> parameters {
+#>   real<lower=0,upper=1> theta;
+#> }
+#> model {
+#>   y ~ bernoulli(theta);
+#> }
+#> 
+
+# stan program as character vector of lines
+f2 <- write_stan_file(lines)
+identical(readLines(f), readLines(f2))
+#> [1] TRUE
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/write_stan_json.html b/docs/reference/write_stan_json.html index 653081d79..a241a6d9f 100644 --- a/docs/reference/write_stan_json.html +++ b/docs/reference/write_stan_json.html @@ -1,75 +1,12 @@ - - - - - - - -Write data to a JSON file readable by CmdStan — write_stan_json • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Write data to a JSON file readable by CmdStan — write_stan_json • cmdstanr - - + + - - -
-
- -
- -
+

Write data to a JSON file readable by CmdStan

- Source: R/data.R + Source: R/data.R
@@ -179,120 +107,114 @@

Write data to a JSON file readable by CmdStan

Write data to a JSON file readable by CmdStan

- 
write_stan_json(data, file, always_decimal = FALSE)
+
+ 
write_stan_json(data, file, always_decimal = FALSE)
+
+ +
+

Arguments

+
data
+

(list) A named list of R objects.

+ + +
file
+

(string) The path to where the data file should be written.

+ -

Arguments

- - - - - - - - - - - - - - -
data

(list) A named list of R objects.

file

(string) The path to where the data file should be written.

always_decimal

(logical) Force generate non-integers with decimal +

always_decimal
+

(logical) Force generate non-integers with decimal points to better distinguish between integers and floating point values. If TRUE all R objects in data intended for integers must be of integer -type.

- -

Details

+type.

+
+
+

Details

write_stan_json() performs several conversions before writing the JSON -file:

    -

  • logical -> integer (TRUE -> 1, FALSE -> 0)

    • -

  • data.frame -> matrix (via data.matrix())

    • +file:

    • logical -> integer (TRUE -> 1, FALSE -> 0)

      • +

    • data.frame -> matrix (via data.matrix())

    • list -> array

    • table -> vector, matrix, or array (depending on dimensions of table)

      • -
    - -

    The list to array conversion is intended to make it easier to prepare -the data for certain Stan declarations involving arrays:

      -

    • vector[J] v[K] (or equivalently array[K] vector[J] v as of Stan 2.27) +

    The list to array conversion is intended to make it easier to prepare +the data for certain Stan declarations involving arrays:

    • vector[J] v[K] (or equivalently array[K] vector[J] v as of Stan 2.27) can be constructed in R as a list with K elements where each element a vector of length J

    • matrix[I,J] v[K] (or equivalently array[K] matrix[I,J] m as of Stan 2.27 ) can be constructed in R as a list with K elements where each element an IxJ matrix

      • -
    - -

    These can also be passed in from R as arrays instead of lists but the list +

These can also be passed in from R as arrays instead of lists but the list option is provided for convenience. Unfortunately for arrays with more than one dimension, e.g., vector[J] v[K,L] (or equivalently array[K,L] vector[J] v as of Stan 2.27) it is not possible to use an R list and an array must be used instead. For this example the array in R should have dimensions KxLxJ.

+
-

Examples

- 
x <- matrix(rnorm(10), 5, 2)
-y <- rpois(nrow(x), lambda = 10)
-z <- c(TRUE, FALSE)
-data <- list(N = nrow(x), K = ncol(x), x = x, y = y, z = z)
-
-# write data to json file
-file <- tempfile(fileext = ".json")
-write_stan_json(data, file)
-
-# check the contents of the file
-cat(readLines(file), sep = "\n")
-
#> {
-#>   "N": 5,
-#>   "K": 2,
-#>   "x": [
-#>     [0.552235993548325, 0.721685253131368],
-#>     [1.40374671287315, -0.278162254251497],
-#>     [0.512051509623765, -0.524699374531002],
-#>     [0.660479109910809, 0.0625388737657775],
-#>     [-0.519175154088263, -0.450598871694098]
-#>   ],
-#>   "y": [16, 11, 7, 7, 6],
-#>   "z": [1, 0]
-#> }

-
-# demonstrating list to array conversion
-# suppose x is declared as `vector[3] x[2]` (or equivalently `array[2] vector[3] x`)
-# we can use a list of length 2 where each element is a vector of length 3
-data <- list(x = list(1:3, 4:6))
-file <- tempfile(fileext = ".json")
-write_stan_json(data, file)
-cat(readLines(file), sep = "\n")
-
#> {
-#>   "x": [
-#>     [1, 2, 3],
-#>     [4, 5, 6]
-#>   ]
-#> }

-
+
+

Examples

+ 
x <- matrix(rnorm(10), 5, 2)
+y <- rpois(nrow(x), lambda = 10)
+z <- c(TRUE, FALSE)
+data <- list(N = nrow(x), K = ncol(x), x = x, y = y, z = z)
+
+# write data to json file
+file <- tempfile(fileext = ".json")
+write_stan_json(data, file)
+
+# check the contents of the file
+cat(readLines(file), sep = "\n")
+#> {
+#>   "N": 5,
+#>   "K": 2,
+#>   "x": [
+#>     [1.39067866394794, 0.962441316104017],
+#>     [-0.473566508639385, -0.892862153643971],
+#>     [-2.10865772647623, 0.757598190570521],
+#>     [-0.366087824635297, -0.120402278774405],
+#>     [1.44942517823716, -0.444483690720062]
+#>   ],
+#>   "y": [5, 11, 1, 9, 11],
+#>   "z": [1, 0]
+#> }
+
+
+# demonstrating list to array conversion
+# suppose x is declared as `vector[3] x[2]` (or equivalently `array[2] vector[3] x`)
+# we can use a list of length 2 where each element is a vector of length 3
+data <- list(x = list(1:3, 4:6))
+file <- tempfile(fileext = ".json")
+write_stan_json(data, file)
+cat(readLines(file), sep = "\n")
+#> {
+#>   "x": [
+#>     [1, 2, 3],
+#>     [4, 5, 6]
+#>   ]
+#> }
+
+
+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + diff --git a/docs/reference/write_stan_tempfile.html b/docs/reference/write_stan_tempfile.html index fd650e821..4e9c4ac10 100644 --- a/docs/reference/write_stan_tempfile.html +++ b/docs/reference/write_stan_tempfile.html @@ -1,75 +1,12 @@ - - - - - - - -Write Stan code to a temporary file — write_stan_tempfile • cmdstanr - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Write Stan code to a temporary file — write_stan_tempfile • cmdstanr - + + - - - -
-
- -
- -
+

Write Stan code to a temporary file

- Source: R/example.R + Source: R/example.R
-

This function is deprecated. Please use write_stan_file() instead.

+

This function is deprecated. Please use write_stan_file() instead.

- 
write_stan_tempfile(code, dir = tempdir())
+
+ 
write_stan_tempfile(code, dir = tempdir())
+
-

Arguments

- - - - - - - - - - -
code

(character vector) The Stan code to write to the file. This can +

+

Arguments

+
code
+

(character vector) The Stan code to write to the file. This can be a character vector of length one (a string) containing the entire Stan program or a character vector with each element containing one line of the -Stan program.

dir

(string) An optional path to the directory where the file will be +Stan program.

+ + +
dir
+

(string) An optional path to the directory where the file will be written. If omitted, a global option cmdstanr_write_stan_file_dir is -used. If the global options is not set, temporary directory -is used.

+used. If the global options is not set, temporary directory +is used.

+
+
-
-
-

Developed by Jonah Gabry, Rok Češnovar.

+
+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

Site built with pkgdown 1.6.1.

+

Site built with pkgdown 2.0.7.

-
-
+
- - + + From 300030c4fe46a623b59e787c7c4940d11eecc953 Mon Sep 17 00:00:00 2001 From: rok-cesnovar Date: Wed, 26 Jul 2023 19:57:41 +0200 Subject: [PATCH 14/16] update opencl vignette --- .../articles/articles-online-only/opencl.html | 151 ++++++++++-------- .../header-attrs-2.13/header-attrs.js | 12 ++ 2 files changed, 95 insertions(+), 68 deletions(-) create mode 100644 docs/articles/articles-online-only/opencl_files/header-attrs-2.13/header-attrs.js diff --git a/docs/articles/articles-online-only/opencl.html b/docs/articles/articles-online-only/opencl.html index 274f8275e..c2af01ec9 100644 --- a/docs/articles/articles-online-only/opencl.html +++ b/docs/articles/articles-online-only/opencl.html @@ -26,8 +26,6 @@ - -
+

Running Stan on the GPU with OpenCL

-

Rok Češnovar -and Jonah Gabry

+

Rok Češnovar and Jonah +Gabry

- Source: vignettes/articles-online-only/opencl.Rmd + Source: vignettes/articles-online-only/opencl.Rmd
-
-

Introduction -

+
+

+Introduction

This vignette demonstrates how to use the OpenCL capabilities of CmdStan with CmdStanR. The functionality described in this vignette requires CmdStan 2.26.1 or newer.

@@ -162,13 +160,13 @@

Introductionprofiling, which was introduced in Stan version 2.26.0.

-
-

OpenCL runtime -

+
+

+OpenCL runtime

OpenCL is supported on most modern CPUs and GPUs. In order to use OpenCL in CmdStanR, an OpenCL runtime for the target device must be installed. A guide for the most common devices is available in the -CmdStan manual’s chapter +CmdStan manual’s chapter on parallelization.

In case of using Windows, CmdStan requires the OpenCL.lib to compile the model. If you experience issue @@ -177,17 +175,17 @@

OpenCL runtimeOpenCL.lib file on your system. If you are using CUDA, the path should be similar to the one listed here.

-path_to_opencl_lib <- "C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.5/lib/x64"
-cpp_options = list(
-  paste0("LDFLAGS+= -L\"",path_to_opencl_lib,"\" -lOpenCL")
-)
-
-cmdstanr::cmdstan_make_local(cpp_options = cpp_options)
-cmdstanr::rebuild_cmdstan()
+path_to_opencl_lib <- "C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.5/lib/x64" +cpp_options = list( + paste0("LDFLAGS+= -L\"",path_to_opencl_lib,"\" -lOpenCL") +) + +cmdstanr::cmdstan_make_local(cpp_options = cpp_options) +cmdstanr::rebuild_cmdstan()

-
-

Compiling a model with OpenCL -

+
+

+Compiling a model with OpenCL

By default, models in CmdStanR are compiled without OpenCL support. Once OpenCL support is enabled, a CmdStan model will make use of OpenCL if the functions in the model support it. Technically no @@ -215,30 +213,30 @@

Compiling a model with OpenCL
-library(cmdstanr)
-
-# Generate some fake data
-n <- 250000
-k <- 20
-X <- matrix(rnorm(n * k), ncol = k)
-y <- rbinom(n, size = 1, prob = plogis(3 * X[,1] - 2 * X[,2] + 1))
-mdata <- list(k = k, n = n, y = y, X = X)

+library(cmdstanr) + +# Generate some fake data +n <- 250000 +k <- 20 +X <- matrix(rnorm(n * k), ncol = k) +y <- rbinom(n, size = 1, prob = plogis(3 * X[,1] - 2 * X[,2] + 1)) +mdata <- list(k = k, n = n, y = y, X = X)

In this model, most of the computation will be handled by the bernoulli_logit_glm_lpmf function. Because this is a supported GPU function, it should be possible to accelerate it with -OpenCL. Check here for a +OpenCL. Check here for a list of functions with OpenCL support.

To build the model with OpenCL support, add cpp_options = list(stan_opencl = TRUE) at the compilation step.

-# Compile the model with STAN_OPENCL=TRUE
-mod_cl <- cmdstan_model("opencl-files/bernoulli_logit_glm.stan",
-                        cpp_options = list(stan_opencl = TRUE))
+# Compile the model with STAN_OPENCL=TRUE +mod_cl <- cmdstan_model("opencl-files/bernoulli_logit_glm.stan", + cpp_options = list(stan_opencl = TRUE))
-
-

Running models with OpenCL -

+ +
Running MCMC with 4 parallel chains...
+
+Chain 4 finished in 96.7 seconds.
+Chain 1 finished in 97.9 seconds.
+Chain 2 finished in 98.6 seconds.
+Chain 3 finished in 98.8 seconds.
+
+All 4 chains finished successfully.
+Mean chain execution time: 98.0 seconds.
+Total execution time: 103.0 seconds.

We’ll also run a version without OpenCL and compare the run times.

-
-# no OpenCL version
-mod <- cmdstan_model("opencl-files/bernoulli_logit_glm.stan", force_recompile = TRUE)
-fit_cpu <- mod$sample(data = mdata, chains = 4, parallel_chains = 4, refresh = 0)
-

The speedup of the OpenCL model is:

-fit_cpu$time()$total / fit_cl$time()$total
+# no OpenCL version +mod <- cmdstan_model("opencl-files/bernoulli_logit_glm.stan", force_recompile = TRUE) +fit_cpu <- mod$sample(data = mdata, chains = 4, parallel_chains = 4, refresh = 0)
+
Running MCMC with 4 parallel chains...
+
+Chain 3 finished in 487.9 seconds.
+Chain 2 finished in 491.8 seconds.
+Chain 1 finished in 514.9 seconds.
+Chain 4 finished in 518.4 seconds.
+
+All 4 chains finished successfully.
+Mean chain execution time: 503.2 seconds.
+Total execution time: 521.9 seconds.
+

The speedup of the OpenCL model is:

+
+fit_cpu$time()$total / fit_cl$time()$total
+
[1] 5.065968

This speedup will be determined by the particular GPU/CPU used, the input problem sizes (data as well as parameters) and if the model uses functions that can be run on the GPU or other OpenCL devices.

@@ -286,13 +305,11 @@

Running models with OpenCL
-

-

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

+

Developed by Jonah Gabry, Rok Češnovar, Andrew Johnson.

-

-

Site built with pkgdown 2.0.7.

+

Site built with pkgdown 1.6.1.

@@ -301,7 +318,5 @@

Running models with OpenCL diff --git a/docs/articles/articles-online-only/opencl_files/header-attrs-2.13/header-attrs.js b/docs/articles/articles-online-only/opencl_files/header-attrs-2.13/header-attrs.js new file mode 100644 index 000000000..dd57d92e0 --- /dev/null +++ b/docs/articles/articles-online-only/opencl_files/header-attrs-2.13/header-attrs.js @@ -0,0 +1,12 @@ +// Pandoc 2.9 adds attributes on both header and div. We remove the former (to +// be compatible with the behavior of Pandoc < 2.8). +document.addEventListener('DOMContentLoaded', function(e) { + var hs = document.querySelectorAll("div.section[class*='level'] > :first-child"); + var i, h, a; + for (i = 0; i < hs.length; i++) { + h = hs[i]; + if (!/^h[1-6]$/i.test(h.tagName)) continue; // it should be a header h1-h6 + a = h.attributes; + while (a.length > 0) h.removeAttribute(a[0].name); + } +}); From ae25ca4402232960bf8e1ae74e118c628ea814f7 Mon Sep 17 00:00:00 2001 From: jgabry Date: Wed, 26 Jul 2023 12:00:39 -0600 Subject: [PATCH 15/16] Update .gitignore --- .gitignore | 1 + 1 file changed, 1 insertion(+) diff --git a/.gitignore b/.gitignore index 8dc61d80a..cce013141 100644 --- a/.gitignore +++ b/.gitignore @@ -7,3 +7,4 @@ design-questions/* inst/doc dev-helpers.R +release-prep.R From 7d15ab4a244830dbb785e9c892d9e4fecd0cd50a Mon Sep 17 00:00:00 2001 From: jgabry Date: Wed, 26 Jul 2023 12:06:04 -0600 Subject: [PATCH 16/16] update Rbuildignore --- .Rbuildignore | 2 +- .gitignore | 1 - 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/.Rbuildignore b/.Rbuildignore index 691c7f148..8c1f77aae 100644 --- a/.Rbuildignore +++ b/.Rbuildignore @@ -7,8 +7,8 @@ ^docs$ ^pkgdown$ ^man-roxygen$ -^dev-helpers\.R$ ^LICENSE\.md$ ^\.appveyor\.yml$ ^\.github$ ^vignettes/articles-online-only$ +^release-prep\.R$ diff --git a/.gitignore b/.gitignore index cce013141..db3f1c32c 100644 --- a/.gitignore +++ b/.gitignore @@ -4,7 +4,6 @@ .Ruserdata .vscode/* -design-questions/* inst/doc dev-helpers.R release-prep.R