Mercurial > repos > public > sbplib_julia
diff test/Grids/mapped_grid_test.jl @ 1835:a6f28a8b8f3f refactor/lazy_tensors/elementwise_ops
Merge default
author | Jonatan Werpers <jonatan@werpers.com> |
---|---|
date | Thu, 09 Jan 2025 12:40:49 +0100 |
parents | 43c0bfc13de3 |
children | 2b5f81e288f1 d91a9f47380f |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/test/Grids/mapped_grid_test.jl Thu Jan 09 12:40:49 2025 +0100 @@ -0,0 +1,381 @@ +using Diffinitive.Grids +using Diffinitive.RegionIndices +using Test +using StaticArrays +using LinearAlgebra + + +_skew_mapping(a,b) = (ξ̄->ξ̄[1]*a + ξ̄[2]*b, ξ̄->[a b]) + +function _partially_curved_mapping() + x̄((ξ, η)) = @SVector[ξ, η*(1+ξ*(ξ-1))] + J((ξ, η)) = @SMatrix[ + 1 0; + η*(2ξ-1) 1+ξ*(ξ-1); + ] + + return x̄, J +end + +function _fully_curved_mapping() + x̄((ξ, η)) = @SVector[2ξ + η*(1-η), 3η+(1+η/2)*ξ^2] + J((ξ, η)) = @SMatrix[ + 2 1-2η; + (2+η)*ξ 3+1/2*ξ^2; + ] + + return x̄, J +end + +@testset "MappedGrid" begin + @testset "Constructor" begin + lg = equidistant_grid((0,0), (1,1), 11, 21) + + x̄ = map(ξ̄ -> 2ξ̄, lg) + J = map(ξ̄ -> @SArray(fill(2., 2, 2)), lg) + mg = MappedGrid(lg, x̄, J) + + @test mg isa Grid{SVector{2, Float64},2} + @test jacobian(mg) isa Array{<:AbstractMatrix} + @test logical_grid(mg) isa Grid + + @test collect(mg) == x̄ + @test jacobian(mg) == J + @test logical_grid(mg) == lg + + + x̄ = map(ξ̄ -> @SVector[ξ̄[1],ξ̄[2], ξ̄[1] + ξ̄[2]], lg) + J = map(ξ̄ -> @SMatrix[1 0; 0 1; 1 1], lg) + mg = MappedGrid(lg, x̄, J) + + @test mg isa Grid{SVector{3, Float64},2} + @test jacobian(mg) isa Array{<:AbstractMatrix} + @test logical_grid(mg) isa Grid + + @test collect(mg) == x̄ + @test jacobian(mg) == J + @test logical_grid(mg) == lg + + sz1 = (10,11) + sz2 = (10,12) + @test_throws ArgumentError("Sizes must match") MappedGrid( + equidistant_grid((0,0), (1,1), sz2...), + rand(SVector{2},sz1...), + rand(SMatrix{2,2},sz1...), + ) + + @test_throws ArgumentError("Sizes must match") MappedGrid( + equidistant_grid((0,0), (1,1), sz1...), + rand(SVector{2},sz2...), + rand(SMatrix{2,2},sz1...), + ) + + @test_throws ArgumentError("Sizes must match") MappedGrid( + equidistant_grid((0,0), (1,1), sz1...), + rand(SVector{2},sz1...), + rand(SMatrix{2,2},sz2...), + ) + + err_str = "The size of the jacobian must match the dimensions of the grid and coordinates" + @test_throws ArgumentError(err_str) MappedGrid( + equidistant_grid((0,0), (1,1), 10, 11), + rand(SVector{3}, 10, 11), + rand(SMatrix{3,4}, 10, 11), + ) + + @test_throws ArgumentError(err_str) MappedGrid( + equidistant_grid((0,0), (1,1), 10, 11), + rand(SVector{3}, 10, 11), + rand(SMatrix{4,2}, 10, 11), + ) + end + + @testset "Indexing Interface" begin + lg = equidistant_grid((0,0), (1,1), 11, 21) + x̄ = map(ξ̄ -> 2ξ̄, lg) + J = map(ξ̄ -> @SArray(fill(2., 2, 2)), lg) + mg = MappedGrid(lg, x̄, J) + @test mg[1,1] == [0.0, 0.0] + @test mg[4,2] == [0.6, 0.1] + @test mg[6,10] == [1., 0.9] + + @test mg[begin, begin] == [0.0, 0.0] + @test mg[end,end] == [2.0, 2.0] + @test mg[begin,end] == [0., 2.] + + @test axes(mg) == (1:11, 1:21) + + @testset "cartesian indexing" begin + cases = [ + (1,1) , + (3,5) , + (10,6), + (1,1) , + (3,2) , + ] + + @testset "i = $is" for (lg, is) ∈ cases + @test mg[CartesianIndex(is...)] == mg[is...] + end + end + + @testset "eachindex" begin + @test eachindex(mg) == CartesianIndices((11,21)) + end + + @testset "firstindex" begin + @test firstindex(mg, 1) == 1 + @test firstindex(mg, 2) == 1 + end + + @testset "lastindex" begin + @test lastindex(mg, 1) == 11 + @test lastindex(mg, 2) == 21 + end + end + + @testset "Iterator interface" begin + lg = equidistant_grid((0,0), (1,1), 11, 21) + x̄ = map(ξ̄ -> 2ξ̄, lg) + J = map(ξ̄ -> @SArray(fill(2., 2, 2)), lg) + + mg = MappedGrid(lg, x̄, J) + + lg2 = equidistant_grid((0,0), (1,1), 15, 11) + sg = MappedGrid( + equidistant_grid((0,0), (1,1), 15, 11), + map(ξ̄ -> @SArray[ξ̄[1], ξ̄[2], -ξ̄[1]], lg2), rand(SMatrix{3,2,Float64},15,11) + ) + + @test eltype(mg) == SVector{2,Float64} + @test eltype(sg) == SVector{3,Float64} + + @test eltype(typeof(mg)) == SVector{2,Float64} + @test eltype(typeof(sg)) == SVector{3,Float64} + + @test size(mg) == (11,21) + @test size(sg) == (15,11) + + @test size(mg,2) == 21 + @test size(sg,2) == 11 + + @test length(mg) == 231 + @test length(sg) == 165 + + @test Base.IteratorSize(mg) == Base.HasShape{2}() + @test Base.IteratorSize(typeof(mg)) == Base.HasShape{2}() + + @test Base.IteratorSize(sg) == Base.HasShape{2}() + @test Base.IteratorSize(typeof(sg)) == Base.HasShape{2}() + + element, state = iterate(mg) + @test element == lg[1,1].*2 + element, _ = iterate(mg, state) + @test element == lg[2,1].*2 + + element, state = iterate(sg) + @test element == sg.physicalcoordinates[1,1] + element, _ = iterate(sg, state) + @test element == sg.physicalcoordinates[2,1] + + @test collect(mg) == 2 .* lg + end + + @testset "Base" begin + lg = equidistant_grid((0,0), (1,1), 11, 21) + x̄ = map(ξ̄ -> 2ξ̄, lg) + J = map(ξ̄ -> @SArray(fill(2., 2, 2)), lg) + mg = MappedGrid(lg, x̄, J) + + @test ndims(mg) == 2 + end + + @testset "==" begin + sz = (15,11) + lg = equidistant_grid((0,0), (1,1), sz...) + x = rand(SVector{3,Float64}, sz...) + J = rand(SMatrix{3,2,Float64}, sz...) + + sg = MappedGrid(lg, x, J) + + sg1 = MappedGrid(equidistant_grid((0,0), (1,1), sz...), copy(x), copy(J)) + + sz2 = (15,12) + lg2 = equidistant_grid((0,0), (1,1), sz2...) + x2 = rand(SVector{3,Float64}, sz2...) + J2 = rand(SMatrix{3,2,Float64}, sz2...) + sg2 = MappedGrid(lg2, x2, J2) + + sg3 = MappedGrid(lg, rand(SVector{3,Float64}, sz...), J) + sg4 = MappedGrid(lg, x, rand(SMatrix{3,2,Float64}, sz...)) + + @test sg == sg1 + @test sg != sg2 # Different size + @test sg != sg3 # Different coordinates + @test sg != sg4 # Different jacobian + end + + @testset "boundary_identifiers" begin + lg = equidistant_grid((0,0), (1,1), 11, 15) + x̄ = map(ξ̄ -> 2ξ̄, lg) + J = map(ξ̄ -> @SArray(fill(2., 2, 2)), lg) + mg = MappedGrid(lg, x̄, J) + @test boundary_identifiers(mg) == boundary_identifiers(lg) + end + + @testset "boundary_indices" begin + lg = equidistant_grid((0,0), (1,1), 11, 15) + x̄ = map(ξ̄ -> 2ξ̄, lg) + J = map(ξ̄ -> @SArray(fill(2., 2, 2)), lg) + mg = MappedGrid(lg, x̄, J) + + @test boundary_indices(mg, CartesianBoundary{1,LowerBoundary}()) == boundary_indices(lg,CartesianBoundary{1,LowerBoundary}()) + @test boundary_indices(mg, CartesianBoundary{2,LowerBoundary}()) == boundary_indices(lg,CartesianBoundary{2,LowerBoundary}()) + @test boundary_indices(mg, CartesianBoundary{1,UpperBoundary}()) == boundary_indices(lg,CartesianBoundary{1,UpperBoundary}()) + end + + @testset "boundary_grid" begin + x̄, J = _partially_curved_mapping() + mg = mapped_grid(x̄, J, 10, 11) + J1((ξ, η)) = @SMatrix[ + 1 ; + η*(2ξ-1); + ] + J2((ξ, η)) = @SMatrix[ + 0; + 1+ξ*(ξ-1); + ] + + function expected_bg(mg, bId, Jb) + lg = logical_grid(mg) + return MappedGrid( + boundary_grid(lg, bId), + map(x̄, boundary_grid(lg, bId)), + map(Jb, boundary_grid(lg, bId)), + ) + end + + let bid = TensorGridBoundary{1, LowerBoundary}() + @test boundary_grid(mg, bid) == expected_bg(mg, bid, J2) + end + + let bid = TensorGridBoundary{1, UpperBoundary}() + @test boundary_grid(mg, bid) == expected_bg(mg, bid, J2) + end + + let bid = TensorGridBoundary{2, LowerBoundary}() + @test boundary_grid(mg, bid) == expected_bg(mg, bid, J1) + end + + let bid = TensorGridBoundary{2, UpperBoundary}() + @test boundary_grid(mg, bid) == expected_bg(mg, bid, J1) + end + end +end + +@testset "mapped_grid" begin + x̄, J = _partially_curved_mapping() + mg = mapped_grid(x̄, J, 10, 11) + @test mg isa MappedGrid{SVector{2,Float64}, 2} + + lg = equidistant_grid((0,0), (1,1), 10, 11) + @test logical_grid(mg) == lg + @test collect(mg) == map(x̄, lg) + + @test mapped_grid(lg, x̄, J) == mg +end + +@testset "metric_tensor" begin + x̄((ξ, η)) = @SVector[ξ*η, ξ + η^2] + J((ξ, η)) = @SMatrix[ + η ξ; + 1 2η; + ] + + g = mapped_grid(x̄, J, 10, 11) + G = map(logical_grid(g)) do (ξ,η) + @SMatrix[ + 1+η^2 ξ*η+2η; + ξ*η+2η ξ^2 + 4η^2; + ] + end + @test metric_tensor(g) ≈ G +end + +@testset "min_spacing" begin + let g = mapped_grid(identity, x->@SMatrix[1], 11) + @test min_spacing(g) ≈ 0.1 + end + + let g = mapped_grid(x->x+x.^2/2, x->@SMatrix[1 .+ x], 11) + @test min_spacing(g) ≈ 0.105 + end + + let g = mapped_grid(x->x + x.*(1 .- x)/2, x->@SMatrix[1.5 .- x], 11) + @test min_spacing(g) ≈ 0.055 + end + + let g = mapped_grid(identity, x->@SMatrix[1 0; 0 1], 11,11) + @test min_spacing(g) ≈ 0.1 + end + + let g = mapped_grid(identity, x->@SMatrix[1 0; 0 1], 11,21) + @test min_spacing(g) ≈ 0.05 + end + + + @testset let a = @SVector[1,0], b = @SVector[1,1]/√2 + g = mapped_grid(_skew_mapping(a,b)...,11,11) + + @test min_spacing(g) ≈ 0.1*norm(b-a) + end + + @testset let a = @SVector[1,0], b = @SVector[-1,1]/√2 + g = mapped_grid(_skew_mapping(a,b)...,11,11) + + @test min_spacing(g) ≈ 0.1*norm(a+b) + end +end + +@testset "normal" begin + g = mapped_grid(_partially_curved_mapping()...,10, 11) + + @test normal(g, CartesianBoundary{1,LowerBoundary}()) == fill(@SVector[-1,0], 11) + @test normal(g, CartesianBoundary{1,UpperBoundary}()) == fill(@SVector[1,0], 11) + @test normal(g, CartesianBoundary{2,LowerBoundary}()) == fill(@SVector[0,-1], 10) + @test normal(g, CartesianBoundary{2,UpperBoundary}()) ≈ map(boundary_grid(g,CartesianBoundary{2,UpperBoundary}())|>logical_grid) do ξ̄ + α = 1-2ξ̄[1] + @SVector[α,1]/√(α^2 + 1) + end + + g = mapped_grid(_fully_curved_mapping()...,5,4) + + unit(v) = v/norm(v) + @testset let bId = CartesianBoundary{1,LowerBoundary}() + lbg = boundary_grid(logical_grid(g), bId) + @test normal(g, bId) ≈ map(lbg) do (ξ, η) + -unit(@SVector[1/2, η/3-1/6]) + end + end + + @testset let bId = CartesianBoundary{1,UpperBoundary}() + lbg = boundary_grid(logical_grid(g), bId) + @test normal(g, bId) ≈ map(lbg) do (ξ, η) + unit(@SVector[7/2, 2η-1]/(5 + 3η + 2η^2)) + end + end + + @testset let bId = CartesianBoundary{2,LowerBoundary}() + lbg = boundary_grid(logical_grid(g), bId) + @test normal(g, bId) ≈ map(lbg) do (ξ, η) + -unit(@SVector[-2ξ, 2]/(6 + ξ^2 - 2ξ)) + end + end + + @testset let bId = CartesianBoundary{2,UpperBoundary}() + lbg = boundary_grid(logical_grid(g), bId) + @test normal(g, bId) ≈ map(lbg) do (ξ, η) + unit(@SVector[-3ξ, 2]/(6 + ξ^2 + 3ξ)) + end + end +end