Added extra type parameter in QGradMonomialBasis

src/QGradMonomialBases.jl

@@ -1,31 +1,31 @@
 
 # Types and constructors
 
-struct QGradMonomialBasis{P,G,D} <: TensorPolynomialBasis{P,P,G}
+struct QGradMonomialBasis{P,V,G,D} <: TensorPolynomialBasis{P,V,G}
   order::Int
   terms::CartesianIndices{D}
   perms::Matrix{Int}
 end
 
-function (::Type{QGradMonomialBasis{P}})(order::Int) where P
+function (::Type{QGradMonomialBasis{P,V}})(order::Int) where {P,V}
   D = _length(P)
-  G = _gradient_type(P,P)
+  @assert _length(V) == D
+  G = _gradient_type(V,P)
   n1d = order+1
   t = fill(n1d,D)
   t[1] = order
   terms = CartesianIndices(tuple(t...))
   perms = _prepare_perms(D)
-  QGradMonomialBasis{P,G,D}(order,terms,perms)
+  QGradMonomialBasis{P,V,G,D}(order,terms,perms)
 end
 
 # Implementation of the interface
 
-length(b::QGradMonomialBasis{P,G,D}) where {P,G,D} = D*b.order*(b.order+1)^(D-1)
+length(b::QGradMonomialBasis{P,V,G,D}) where {P,V,G,D} = D*b.order*(b.order+1)^(D-1)
 
-ndims(b::QGradMonomialBasis{P,G,D}) where {P,G,D} = D
+ndims(b::QGradMonomialBasis{P,V,G,D}) where {P,V,G,D} = D
 
-function ScratchData(b::QGradMonomialBasis{P}) where P
-  V = P
+function ScratchData(b::QGradMonomialBasis{P,V}) where {P,V}
   T = eltype(V)
   dim = _length(P)
   n1d = b.order+1
@@ -35,19 +35,18 @@ function ScratchData(b::QGradMonomialBasis{P}) where P
 end
 
 function evaluate!(
-  v::AbstractVector{P},
-  b::QGradMonomialBasis{P,G,D},
+  v::AbstractVector{V},
+  b::QGradMonomialBasis{P,V,G,D},
   x::P,
-  cache::MonomialBasisCache) where {P,G,D}
+  cache::MonomialBasisCache) where {P,V,G,D}
   _evaluate_nd_qgrad!(v,x,b.order,b.terms,b.perms,cache.c)
 end
 
 function gradient!(
   v::AbstractVector{G},
-  b::QGradMonomialBasis{P,G,D},
+  b::QGradMonomialBasis{P,V,G,D},
   x::P,
-  cache::MonomialBasisCache) where {P,G,D}
-  V = P
+  cache::MonomialBasisCache) where {P,V,G,D}
   _gradient_nd_qgrad!(v,x,b.order,b.terms,b.perms,cache.c,cache.g,V)
 end
 

test/QGradMonomialBasesTests.jl

@@ -4,13 +4,14 @@ using Test
 using TensorPolynomialBases
 using TensorValues
 
-P = VectorValue{3,Float64}
+P = VectorValue{3}
+V = VectorValue{3,Float64}
 order = 1
-basis = QGradMonomialBasis{P}(order)
+basis = QGradMonomialBasis{P,V}(order)
 G = gradient_type(basis)
 
-x = VectorValue(2.0,3.0,5.0)
-v = P[
+x = VectorValue(2,3,5)
+v = V[
   (1.0, 0.0, 0.0), (0.0, 1.0, 0.0), (0.0, 0.0, 1.0),
   (3.0, 0.0, 0.0), (0.0, 5.0, 0.0), (0.0, 0.0, 2.0),
   (5.0, 0.0, 0.0), (0.0, 2.0, 0.0), (0.0, 0.0, 3.0),
@@ -30,14 +31,15 @@ g = G[
   (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 3.0, 2.0, 0.0)]
 test_polynomial_basis(basis,x,v,g)
 
-P = VectorValue{2,Float64}
+P = VectorValue{2}
+V = VectorValue{2,Float64}
 order = 2
-basis = QGradMonomialBasis{P}(order)
+basis = QGradMonomialBasis{P,V}(order)
 G = gradient_type(basis)
 
 x = VectorValue(2.0,3.0)
 
-v = P[
+v = V[
   (1.0, 0.0), (0.0, 1.0), (2.0, 0.0), (0.0, 3.0),
   (3.0, 0.0), (0.0, 2.0), (6.0, 0.0), (0.0, 6.0),
   (9.0, 0.0), (0.0, 4.0), (18.0, 0.0), (0.0, 12.0)]