Merge pull request #98 from jeremiahpslewis/jpsl/devfix

Update devcontainer

.devcontainer/devcontainer.json

@@ -1,7 +1,17 @@
 {
-    "extensions": [
-      "julialang.language-julia"
-    ],
-    "runArgs": ["--privileged"],
-    "dockerFile": "Dockerfile"
-}
+	"image": "mcr.microsoft.com/vscode/devcontainers/base:ubuntu-24.04",
+    "features": {
+		"ghcr.io/julialang/devcontainer-features/julia": "1.10"
+    },
+	// Set *default* container specific settings.json values on container create.
+	"settings": {
+		"terminal.integrated.fontFamily": "monospace",
+		"julia.environmentPath": "."
+	},
+
+	// Add the IDs of extensions you want installed when the container is created.
+	"extensions": [
+		"quarto.quarto",
+		"GitHub.copilot",
+	],	
+}

src/AIAPC2020/hooks.jl

@@ -56,7 +56,11 @@ function _best_action_lookup(state_, table)
     best_action = 1
     max_value = table[1, state_]
 
-    for i = 2:size(table, 1)
+    for i = axes(table, 1)
+        if i == 1
+            continue
+        end
+
         value = table[i, state_]
         if value > max_value
             max_value = value

src/multiprocessing_template_dddc.jl

@@ -1,6 +1,7 @@
 using AlgorithmicCompetition
 using Statistics
-using DataFrames CSV
+using DataFrames
+using CSV
 using Distributed
 using Dates
 

test/integration.jl

@@ -568,7 +568,7 @@ end
     ξ = 0.1
     δ = 0.95
     n_prices = 15
-    max_iter = 1000000
+    max_iter = Int64(5e6)
     price_index = 1:n_prices
 
     competition_params_dict = Dict(
@@ -584,11 +584,11 @@ end
         δ,
         max_iter,
         competition_solution_dict;
-        convergence_threshold = 100,
+        convergence_threshold = 500,
     )
 
 
-    c_out = run(hyperparameters; stop_on_convergence = false, debug = true)
+    c_out = run(hyperparameters; stop_on_convergence = true, debug = true)
 
     # ensure that the policy is updated by the learner
     @test sum(c_out.policy[Player(1)].policy.learner.approximator.model .!= 0) != 0
@@ -607,15 +607,12 @@ end
           c_out.hook[Player(2)][1].best_response_vector
 
 
-    @test mean(
-        c_out.hook[Player(1)][2].rewards[(end-2):end] .!=
-        c_out.hook[Player(2)][2].rewards[(end-2):end],
-    ) >= 0.3
+    @test mean(argmax(c_out.policy[Player(1)].policy.learner.approximator.model, dims=1) .!= argmax(c_out.policy[Player(2)].policy.learner.approximator.model, dims=1)) < 0.98
 
     for i in [Player(1), Player(2)]
         @test c_out.hook[i][1].convergence_duration >= 0
         @test c_out.hook[i][1].is_converged
-        @test c_out.hook[i][1].convergence_threshold == 100
+        @test c_out.hook[i][1].convergence_threshold == 500
         @test sum(c_out.hook[i][2].rewards .== 0) == 0
     end
 
@@ -712,8 +709,8 @@ end
     for debug in [true, false]
         AlgorithmicCompetition.run_dddc(
             n_parameter_iterations = 1,
-            max_iter = Int(2e5),
-            convergence_threshold = Int(1e5),
+            max_iter = Int(1e5),
+            convergence_threshold = Int(1e2),
             n_grid_increments = 3,
             debug = debug,
         )

viz/a1_viz.jl

@@ -22,7 +22,7 @@ csv_files = filter!(
 
 df_ = DataFrame.(CSV.File.(csv_files))
 
-for i = 1:length(df_)
+for i = eachindex(df_, 2)
     df_[i][!, "metadata"] .= csv_files[i]
 end
 df = vcat(df_...)

viz/dddc_viz.jl

@@ -39,7 +39,7 @@ df_raw_ = arrow_files[.!is_summary_file]
 
 if rebuild_summary_files
     rm.(df_summary_)
-    @showprogress for i in 1:length(df_raw_)
+    @showprogress for i in eachindex(df_raw_)
         df = DataFrame(Arrow.Table(df_raw_[i]))
         if nrow(df) > 0
             df = expand_and_extract_dddc(df)
@@ -52,7 +52,7 @@ end
 if rebuild_overall_summary
     arrows_ = DataFrame.(Arrow.Table.(df_summary_))
 
-    for i = 1:length(arrows_)
+    for i = eachindex(arrows_)
         arrows_[i][!, "metadata"] .= df_summary_[i]
         if "signal_is_strong" in names(arrows_[i])
             arrows_[i] = select!(arrows_[i], Not(:signal_is_strong))