Update tests

Author: lhnguyen-vn

test/parameters.jl

@@ -1,45 +1,61 @@
 @testset "Parameters" begin
+    # Swarm size
     n = 3
+    # Test ranges
     r1 = range(String, :r1; values=["a", "b", "c"])
     r2 = range(Int, :r2; lower=1, upper=3, scale=exp10)
     r3 = range(Float64, :r3; lower=1, upper=Inf, origin=4, unit=1)
     r4 = range(Float32, :r4; lower=-Inf, upper=Inf, origin=0, unit=1)
+    rs = (r1, r2, r3, r4)
+    # Test distribution types
+    Ds = (Dirichlet, Uniform, Gamma, Normal)
+    # Manually fitted distributions for test ranges
     d1 = Dirichlet(ones(3))
     d2 = Uniform(1, 3)
     d3 = truncated(Gamma(16, 0.25), 1, Inf)
     d4 = Normal(0, 1)
+    ds = (d1, d2, d3, d4)
+    # Test range representation lengths
+    lengths = (3, 1, 1, 1)
+    # Test range's corresponding indices in internal representation state.X
+    indices = (1:3, 4, 5, 6)
+    # Generated samples from d1, d2, d3, d4 with StableRNG(1234)
     X1 = [0.14280010160187237 0.49409071076694583 0.3631091876311819;
           0.295010672512568   0.4534584876713112  0.25153083981612073;
           0.12881930462550284 0.2617407494915029  0.6094399458829942]
     X2 = [2.7429797605672808, 2.3976392099947, 2.5742724788985445]
     X3 = [3.9372495283243105, 3.6569395920512977, 3.6354556967115146]
     X4 = [-0.8067647083847199, 0.420991611378423, 0.6736019046580138]
+    Xs = (X1, X2, X3, X4)
 
     @testset "Initializer" begin
-        @test PSO._initializer(r1) === Dirichlet
-        @test PSO._initializer(r2) === Uniform
-        @test PSO._initializer(r3) === Gamma
-        @test PSO._initializer(r4) === Normal
+        for (r, D) in zip(rs, Ds)
+            @test PSO._initializer(r) === D
+        end
     end
 
     @testset "Initialize with distribution types" begin
         rng = StableRNG(1234)
-        @test PSO._initialize(rng, r1, Dirichlet, n)[[1,3,4]] == (r1, 3, X1)
-        @test PSO._initialize(rng, r2, Uniform, n)[[1,3,4]] == (r2, 1, X2)
-        @test PSO._initialize(rng, r3, Gamma, n)[[1,3,4]] == (r3, 1, X3)
-        @test PSO._initialize(rng, r4, Normal, n)[[1,3,4]] == (r4, 1, X4)
+        for (r, D, l, X) in zip(rs, Ds, lengths, Xs)
+            r̂, l̂, X̂ = PSO._initialize(rng, r, D, n)[[1,3,4]]
+            @test r̂ === r
+            @test l̂ == l
+            @test X̂ ≈ X
+        end
     end
 
     @testset "Initialize with distributions" begin
         rng = StableRNG(1234)
-        @test PSO._initialize(rng, r1, d1, n)[[1,3,4]] == (r1, 3, X1)
-        @test PSO._initialize(rng, r2, d2, n)[[1,3,4]] == (r2, 1, X2)
-        @test PSO._initialize(rng, r3, d3, n)[[1,3,4]] == (r3, 1, X3)
-        @test PSO._initialize(rng, r4, d4, n)[[1,3,4]] == (r4, 1, X4)
+        for (r, d, l, X) in zip(rs, ds, lengths, Xs)
+            r̂, l̂, X̂ = PSO._initialize(rng, r, d, n)[[1,3,4]]
+            @test r̂ === r
+            @test l̂ == l
+            @test X̂ ≈ X
+        end
     end
 
     @testset "Range Indices" begin
-        @test PSO._to_indices((3,1,1,1)) == (1:3, 4, 5, 6)
+        @test PSO._to_indices(lengths) == indices
     end
 
     @testset "Unsupported distributions" begin
@@ -52,24 +68,26 @@
     end
 
     @testset "Initialize one range" begin
-        ps = ParticleSwarm(n_particles=3, rng=StableRNG(1234))
-        state = PSO.initialize(r1, ps)
-        @test state.ranges == (r1,)
-        @test state.indices == (1:3,)
-        @test state.X == X1
+        ps = ParticleSwarm(n_particles=n, rng=StableRNG(1234))
+        for (r, l, i, X) in zip(rs, lengths, indices, Xs)
+            state = PSO.initialize(r, ps)
+            @test state.ranges == (r,)
+            @test state.indices == (l == 1 ? 1 : 1:l,)
+            @test state.X ≈ X
+        end
     end
 
     @testset "Initialize multiple ranges" begin
-        ps = ParticleSwarm(n_particles=3, rng=StableRNG(1234))
+        ps = ParticleSwarm(n_particles=n, rng=StableRNG(1234))
         ranges = [r1, (r2, Uniform), (r3, d3), r4]
         state = PSO.initialize(ranges, ps)
-        @test state.ranges == (r1, r2, r3, r4)
-        @test state.indices == (1:3, 4, 5, 6)
-        @test state.X == hcat(X1, X2, X3, X4)
+        @test state.ranges == rs
+        @test state.indices == indices
+        @test state.X ≈ hcat(Xs...)
     end
 
     @testset "Retrieve parameters" begin
-        ps = ParticleSwarm(n_particles=3, rng=StableRNG(1234))
+        ps = ParticleSwarm(n_particles=n, rng=StableRNG(1234))
         ranges = [r1, (r2, Uniform), (r3, d3), r4]
         state = PSO.initialize(ranges, ps)
         PSO.retrieve!(state, ps)