Tabular approx tests

jeremiahpslewis · web-flow · commit 5d59969f6c55 · 2024-03-11T17:43:45.000+01:00
diff --git a/src/ReinforcementLearningCore/test/policies/learners/approximators.jl b/src/ReinforcementLearningCore/test/policies/learners/approximators.jl
diff --git a/src/ReinforcementLearningCore/test/policies/learners/learners.jl b/src/ReinforcementLearningCore/test/policies/learners/learners.jl
@@ -0,0 +1,2 @@
+include("approximator.jl")
+include("tabular_approximator.jl")
diff --git a/src/ReinforcementLearningCore/test/policies/learners/tabular_approximator.jl b/src/ReinforcementLearningCore/test/policies/learners/tabular_approximator.jl
@@ -0,0 +1,21 @@
+
+using Test
+using ReinforcementLearningCore
+using Flux
+
+@testset "Constructors" begin
+    @test TabularApproximator(fill(1, 10, 10), fill(1, 10)) isa TabularApproximator
+    @test TabularVApproximator(n_state = 10) isa
+          TabularApproximator{Vector{Float64},InvDecay}
+    @test TabularQApproximator(n_state = 10, n_action = 10) isa
+          TabularApproximator{Matrix{Float64},InvDecay}
+end
+
+@testset "RLCore.forward" begin
+    v_approx = TabularVApproximator(n_state = 10)
+    @test RLCore.forward(v_approx, 1) == 0.0
+
+    q_approx = TabularQApproximator(n_state = 5, n_action = 10)
+    @test RLCore.forward(q_approx, 1) == zeros(Float64, 10)
+    @test RLCore.forward(q_approx, 1, 5) == 0.0
+end
diff --git a/src/ReinforcementLearningCore/test/policies/policies.jl b/src/ReinforcementLearningCore/test/policies/policies.jl
@@ -1,3 +1,4 @@
 include("agent.jl")
 include("multi_agent.jl")
-include("approximators.jl")
+include("learners/learners.jl")
+

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	`+include("approximator.jl")`
	`2`	`+include("tabular_approximator.jl")`