100 % sparse

test/test_sparse.jl

@@ -2,8 +2,9 @@
 # License is MIT: see https://github.com/JuliaFEM/FEMBase.jl/blob/master/LICENSE
 
 using FEMBase
-using FEMBase: SparseMatrixCOO, SparseVectorCOO, optimize!,
-               resize_sparse, resize_sparsevec
+using FEMBase: SparseMatrixCOO, SparseVectorCOO, optimize!
+using FEMBase: resize_sparse, resize_sparsevec
+using FEMBase: get_nonzero_rows, get_nonzero_columns
 using Base.Test
 
 @testset "Add to SparseMatrixCOO" begin
@@ -47,3 +48,54 @@ end
     @test size(b) == (4, )
 end
 
+@testset "type conversions" begin
+    A = reshape(collect(1:9), 3, 3)
+    A2 = convert(SparseMatrixCOO, A)
+    A3 = convert(SparseMatrixCOO, sparse(A))
+    @test isapprox(A, full(A2))
+    @test isapprox(A, full(A3))
+    b = sparse(collect(1:3))
+    b2 = convert(SparseVectorCOO, b)
+    @test isapprox(b, full(b2))
+end
+
+@testset "construct sparse matrix and empty it" begin
+    A = SparseMatrixCOO()
+    push!(A, 1, 1, 1.0)
+    A_expected = sparse([1], [1], [1.0], 2, 2)
+    A2 = sparse(A, 2, 2)
+    @test isapprox(A2, A_expected)
+    empty!(A)
+    @test length(A.I) == length(A.J) == length(A.V) == 0
+end
+
+@testset "construct sparse matrix with special function" begin
+    A = SparseMatrixCOO()
+    push!(A, 1, 1, 1.0)
+    push!(A, 1, 1, 2.0)
+    B = sparse(A, 2, 2, +)
+    @test isapprox(B[1,1], 3.0)
+end
+
+@testset "normal addition to sparse vector" begin
+    A = SparseMatrixCOO()
+    add!(A, [1, 2], [1.0 2.0])
+end
+
+@testset "get nonzero rows, columns, size" begin
+    A = SparseMatrixCOO()
+    add!(A, [1, 3], [1, 4], [1.0 2.0; 3.0 4.0])
+    B = sparse(A)
+    @test get_nonzero_rows(B) == [1, 3]
+    @test get_nonzero_columns(B) == [1, 4]
+    @test size(B) == (3, 4)
+    @test size(B, 1) == 3
+end
+
+@testset "isapprox" begin
+    A = SparseMatrixCOO()
+    B = SparseMatrixCOO()
+    add!(A, [1, 3], [1, 4], [1.0 2.0; 3.0 4.0])
+    add!(B, [1, 3], [1, 4], [1.0 2.0; 3.0 4.0])
+    @test isapprox(A, B)
+end