Add a missing domain method for coordinate domains, with tests. (#49)

* Add a missing domain method for coordinate domains, with tests.

* fix AbstractVector argument inference

src/transformations.jl

@@ -57,6 +57,10 @@ end
 
 domain_kind(::Type{<:CoordinateTransformations}) = :multivariate
 
+function domain(coordinate_transformations::CoordinateTransformations)
+    coordinate_domains(map(domain, coordinate_transformations.transformations))
+end
+
 function Base.Tuple(coordinate_transformations::CoordinateTransformations)
     coordinate_transformations.transformations
 end
@@ -109,9 +113,9 @@ function transform_to(domain::CoordinateDomains, ct::CoordinateTransformations,
     map((d, t, x) -> transform_to(d, t, x), domains, transformations, x)
 end
 
-function transform_to(domain::CoordinateDomains, ct::CoordinateTransformations,
-                   x::AbstractVector)
-    SVector(transform_to(domain, ct, Tuple(x)))
+function transform_to(domain::CoordinateDomains{T}, ct::CoordinateTransformations,
+                      x::AbstractVector) where T
+    SVector(transform_to(domain, ct, _ntuple_like(T, x)))
 end
 
 function transform_to(domain::CoordinateDomains, ct::CoordinateTransformations, ∂x::∂Input)
@@ -131,9 +135,9 @@ function transform_from(domain::CoordinateDomains, ct::CoordinateTransformations
     map((d, t, x) -> transform_from(d, t, x), domains, transformations, x)
 end
 
-function transform_from(domain::CoordinateDomains, ct::CoordinateTransformations,
-                        x::AbstractVector)
-    SVector(transform_from(domain, ct, Tuple(x)))
+function transform_from(domain::CoordinateDomains{T}, ct::CoordinateTransformations,
+                        x::AbstractVector) where {T}
+    SVector(transform_from(domain, ct, _ntuple_like(T, x)))
 end
 
 ####
@@ -266,6 +270,7 @@ end
 
 function domain(t::SemiInfRational)
     (; L, A) = t
+    A = float(A)
     ∞ = oftype(A, Inf)
     L > 0 ? UnivariateDomain(A, ∞) : UnivariateDomain(-∞, A)
 end

src/utilities.jl

@@ -35,3 +35,16 @@ struct SubScript
 end
 
 Base.print(io::IO, s::SubScript) = print_number(io, _SUBSCRIPT_DIGITS, s.i)
+
+"""
+$(SIGNATURES)
+
+If `T <: NTuple{N}`, convert `v` into an `NTuple{N}`.
+
+Used for ingesting `::AbstractVector` arguments in contexts where an `NTuple` or
+`SVector` is preferred.
+"""
+function _ntuple_like(::Type{T}, v::AbstractVector) where {N,T<:NTuple{N}}
+    @argcheck length(v) == N
+    NTuple{N}(v)
+end

test/test_generic_api.jl

@@ -29,7 +29,7 @@ end
     @test_throws ArgumentError linear_combination(basis, bad_θ)
 end
 
-@testset "transformed bases and linear combinations" begin
+@testset "transformed bases and linear combinations (univariate)" begin
     N = 10
     basis = Chebyshev(EndpointGrid(), N)
     t = BoundedLinear(1.0, 2.0)
@@ -48,6 +48,25 @@ end
     end
 end
 
+@testset "transformed bases and linear combinations (bivariate)" begin
+    basis0 = smolyak_basis(Chebyshev, InteriorGrid(), SmolyakParameters(2, 2), Val(2))
+    t = coordinate_transformations(BoundedLinear(1.0, 2.0), SemiInfRational(0, 1))
+    basis = basis0 ∘ t
+    @test domain(basis ∘ t) == domain(t)
+    @test dimension(basis ∘ t) == dimension(basis)
+    @test collect(grid(basis)) ==
+        [transform_from(domain(basis0), t, x) for x in grid(basis0)]
+
+    θ = randn(dimension(basis))
+    l1 = linear_combination(basis0, θ)
+    l2 = linear_combination(basis, θ)
+    l3 = linear_combination(basis0, θ) ∘ t
+    for _ in 1:20
+        x = rand(2) .+ 1.0
+        @test l1(transform_to(domain(basis0), t, x)) == l2(x) == l3(x)
+    end
+end
+
 @testset "subset fallback" begin
     @test !is_subset_basis(Chebyshev(InteriorGrid(), 4), # just test the fallback method
                            smolyak_basis(Chebyshev, InteriorGrid(), SmolyakParameters(2, 2), 2))