julia lsp too slow to use in helix

[haxfn]

output of hx -v --log log.txt:

2025-06-22T18:58:43.316 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"initialize","params":{"capabilities":{"general":{"positionEncodings":["utf-8","utf-32","utf-16"]},"textDocument":{"codeAction":{"codeActionLiteralSupport":{"codeActionKind":{"valueSet":["","quickfix","refactor","refactor.extract","refactor.inline","refactor.rewrite","source","source.organizeImports"]}},"dataSupport":true,"disabledSupport":true,"isPreferredSupport":true,"resolveSupport":{"properties":["edit","command"]}},"completion":{"completionItem":{"deprecatedSupport":true,"insertReplaceSupport":true,"resolveSupport":{"properties":["documentation","detail","additionalTextEdits"]},"snippetSupport":true,"tagSupport":{"valueSet":[1]}},"completionItemKind":{}},"formatting":{"dynamicRegistration":false},"hover":{"contentFormat":["markdown"]},"inlayHint":{"dynamicRegistration":false},"publishDiagnostics":{"tagSupport":{"valueSet":[1,2]},"versionSupport":true},"rename":{"dynamicRegistration":false,"honorsChangeAnnotations":false,"prepareSupport":true},"signatureHelp":{"signatureInformation":{"activeParameterSupport":true,"documentationFormat":["markdown"],"parameterInformation":{"labelOffsetSupport":true}}}},"window":{"workDoneProgress":true},"workspace":{"applyEdit":true,"configuration":true,"didChangeConfiguration":{"dynamicRegistration":false},"didChangeWatchedFiles":{"dynamicRegistration":true,"relativePatternSupport":false},"executeCommand":{"dynamicRegistration":false},"fileOperations":{"didRename":true,"willRename":true},"inlayHint":{"refreshSupport":false},"symbol":{"dynamicRegistration":false},"workspaceEdit":{"documentChanges":true,"failureHandling":"abort","normalizesLineEndings":false,"resourceOperations":["create","rename","delete"]},"workspaceFolders":true}},"clientInfo":{"name":"helix","version":"25.01.1 (171dfc60)"},"processId":61198,"rootPath":"/home/vishal/Documents/Repos/RLExperiments","rootUri":"file:///home/vishal/Documents/Repos/RLExperiments","workspaceFolders":[{"name":"RLExperiments","uri":"file:///home/vishal/Documents/Repos/RLExperiments"}]},"id":0}
2025-06-22T18:59:01.481 helix_lsp::transport [INFO] julia <- {"id":0,"jsonrpc":"2.0","result":{"capabilities":{"textDocumentSync":{"openClose":true,"change":1,"willSave":false,"willSaveWaitUntil":false,"save":{"includeText":true}},"completionProvider":{"resolveProvider":false,"triggerCharacters":[".","@","\"","^"]},"hoverProvider":true,"signatureHelpProvider":{"triggerCharacters":["(",","]},"declarationProvider":false,"definitionProvider":true,"typeDefinitionProvider":false,"implementationProvider":false,"referencesProvider":true,"documentHighlightProvider":true,"documentSymbolProvider":true,"codeActionProvider":true,"colorProvider":false,"documentFormattingProvider":true,"documentRangeFormattingProvider":false,"renameProvider":{"prepareProvider":true},"foldingRangeProvider":false,"executeCommandProvider":{"commands":["FixMissingRef","ReexportModule","ExpandFunction","ExplicitPackageVarImport"]},"selectionRangeProvider":true,"workspaceSymbolProvider":true,"workspace":{"workspaceFolders":{"supported":true,"changeNotifications":true}}}}}
2025-06-22T18:59:01.481 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"initialized","params":{}}
2025-06-22T18:59:01.481 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didOpen","params":{"textDocument":{"languageId":"julia","text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\n\\pi\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n","uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":3}}}
2025-06-22T18:59:04.637 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/notebooks/Easy21.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/notebooks/MCControl.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/notebooks/MCPrediction.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/Control.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/Easy21.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/RLExperiments.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/RLExperimentsUtils.jl","version":0,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_view::handlers::lsp [INFO] Version (0) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (3)), dropping PublishDiagnostic notification
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia <- {"method":"workspace/configuration","id":"865febf3-6729-4dd7-bee6-5038a08a4c72","params":{"items":[{"section":"julia.format.indent"},{"section":"julia.format.indents"},{"section":"julia.format.ops"},{"section":"julia.format.tuples"},{"section":"julia.format.curly"},{"section":"julia.format.calls"},{"section":"julia.format.iterOps"},{"section":"julia.format.comments"},{"section":"julia.format.docs"},{"section":"julia.format.lineends"},{"section":"julia.format.keywords"},{"section":"julia.format.kwarg"},{"section":"julia.lint.call"},{"section":"julia.lint.iter"},{"section":"julia.lint.nothingcomp"},{"section":"julia.lint.constif"},{"section":"julia.lint.lazy"},{"section":"julia.lint.datadecl"},{"section":"julia.lint.typeparam"},{"section":"julia.lint.modname"},{"section":"julia.lint.pirates"},{"section":"julia.lint.useoffuncargs"},{"section":"julia.lint.run"},{"section":"julia.lint.missingrefs"},{"section":"julia.lint.disabledDirs"},{"section":"julia.completionmode"}]},"jsonrpc":"2.0"}
2025-06-22T18:59:04.814 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","result":[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null],"id":"865febf3-6729-4dd7-bee6-5038a08a4c72"}
2025-06-22T18:59:05.439 helix_lsp::transport [INFO] julia <- {"method":"window/workDoneProgress/create","id":"069a4101-1d26-4c49-a992-6b67fa108b9d","params":{"token":"4bff701e-72a0-4af0-8b53-aba19895340e"},"jsonrpc":"2.0"}
2025-06-22T18:59:05.439 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","result":null,"id":"069a4101-1d26-4c49-a992-6b67fa108b9d"}
2025-06-22T18:59:06.108 helix_lsp::transport [INFO] julia <- {"method":"$/progress","params":{"token":"4bff701e-72a0-4af0-8b53-aba19895340e","value":{"kind":"begin","title":"Julia Language Server","message":"Indexing packages..."}},"jsonrpc":"2.0"}
2025-06-22T18:59:06.108 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":3,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:08.065 helix_lsp::transport [ERROR] julia err <- "ERROR: MethodError: no method matching (::LanguageServer.var\"#93#95\"{LanguageServerInstance})(::String, ::Int64)\n"
2025-06-22T18:59:08.187 helix_lsp::transport [ERROR] julia err <- "The function `#93` exists, but no method is defined for this combination of argument types.\n"
2025-06-22T18:59:08.187 helix_lsp::transport [ERROR] julia err <- "\n"
2025-06-22T18:59:08.187 helix_lsp::transport [ERROR] julia err <- "Closest candidates are:\n"
2025-06-22T18:59:08.187 helix_lsp::transport [ERROR] julia err <- "  (::LanguageServer.var\"#93#95\")(::Any)\n"
2025-06-22T18:59:08.187 helix_lsp::transport [ERROR] julia err <- "   @ LanguageServer ~/.julia/packages/LanguageServer/JrIEf/src/languageserverinstance.jl:185\n"
2025-06-22T18:59:08.199 helix_lsp::transport [ERROR] julia err <- "\n"
2025-06-22T18:59:08.199 helix_lsp::transport [ERROR] julia err <- "Stacktrace:\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [1] (::SymbolServer.var\"#70#72\"{LanguageServer.var\"#93#95\"{LanguageServerInstance}})(msg::String, p::Int64)\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ SymbolServer ~/.julia/packages/SymbolServer/2jiVV/src/SymbolServer.jl:151\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [2] (::SymbolServer.var\"#65#67\"{SymbolServer.SymbolServerInstance, String, SymbolServer.var\"#70#72\"{LanguageServer.var\"#93#95\"{LanguageServerInstance}}})(download_dir::String)\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ SymbolServer ~/.julia/packages/SymbolServer/2jiVV/src/SymbolServer.jl:126\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [3] mktempdir(fn::SymbolServer.var\"#65#67\"{SymbolServer.SymbolServerInstance, String, SymbolServer.var\"#70#72\"{LanguageServer.var\"#93#95\"{LanguageServerInstance}}}, parent::String; prefix::String)\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ Base.Filesystem ./file.jl:819\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [4] mktempdir\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ ./file.jl:815 [inlined]\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [5] download_cache_files\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ ~/.julia/packages/SymbolServer/2jiVV/src/SymbolServer.jl:99 [inlined]\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [6] getstore(ssi::SymbolServer.SymbolServerInstance, environment_path::String, progress_callback::LanguageServer.var\"#93#95\"{LanguageServerInstance}, error_handler::Nothing; download::Bool)\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ SymbolServer ~/.julia/packages/SymbolServer/2jiVV/src/SymbolServer.jl:156\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [7] getstore\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ ~/.julia/packages/SymbolServer/2jiVV/src/SymbolServer.jl:149 [inlined]\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- " [8] (::LanguageServer.var\"#92#94\"{LanguageServerInstance})()\n"
2025-06-22T18:59:08.216 helix_lsp::transport [ERROR] julia err <- "   @ LanguageServer ~/.julia/packages/LanguageServer/JrIEf/src/languageserverinstance.jl:182\n"
2025-06-22T18:59:13.790 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\n\\p\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":4}}}
2025-06-22T18:59:13.803 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:13.912 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":2,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":1}
2025-06-22T18:59:13.958 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\n\\\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":5}}}
2025-06-22T18:59:13.971 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:14.080 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":1,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":2}
2025-06-22T18:59:14.113 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\n\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":6}}}
2025-06-22T18:59:14.127 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:14.235 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":0,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":3}
2025-06-22T18:59:14.436 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nm\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":7}}}
2025-06-22T18:59:14.449 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:14.557 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":1,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":4}
2025-06-22T18:59:14.580 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmo\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":8}}}
2025-06-22T18:59:14.653 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":4,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:14.653 helix_view::handlers::lsp [INFO] Version (4) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (8)), dropping PublishDiagnostic notification
2025-06-22T18:59:14.700 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":2,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":5}
2025-06-22T18:59:14.778 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmod\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":9}}}
2025-06-22T18:59:14.899 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":3,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":6}
2025-06-22T18:59:14.912 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmodu\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":10}}}
2025-06-22T18:59:14.925 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:15.003 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmodul\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":11}}}
2025-06-22T18:59:15.124 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":5,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":7}
2025-06-22T18:59:15.250 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmodule\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":12}}}
2025-06-22T18:59:15.263 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:15.371 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":6,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":8}
2025-06-22T18:59:15.490 helix_lsp::transport [INFO] julia <- {"id":1,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.490 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(1)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":5,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":2,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(2)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":6,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":3,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(3)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":7,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":4,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(4)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":8,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":5,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(5)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":9,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":6,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (5) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.498 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(6)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":10,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (6) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":11,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (7) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":7,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(7)), original request likely timed out
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":12,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (8) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.498 helix_lsp::transport [INFO] julia <- {"id":8,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (9) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (10) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.498 helix_view::handlers::lsp [INFO] Version (11) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (12)), dropping PublishDiagnostic notification
2025-06-22T18:59:15.501 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/completion","params":{"context":{"triggerKind":1},"position":{"character":6,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":9}
2025-06-22T18:59:19.960 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmodul\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":13}}}
2025-06-22T18:59:19.974 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:20.010 helix_lsp::transport [INFO] julia <- {"id":9,"jsonrpc":"2.0","result":{"isIncomplete":true,"items":[{"label":"module","kind":14,"documentation":"module","insertTextFormat":2,"textEdit":{"range":{"start":{"line":3,"character":0},"end":{"line":3,"character":6}},"newText":"module $1\n\t$0\nend"}}]}}
2025-06-22T18:59:20.012 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":13,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:22.843 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmodule\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":17}}}
2025-06-22T18:59:22.845 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nmodule \n\t\nend\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":18}}}
2025-06-22T18:59:22.847 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":17,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:22.849 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":18,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:22.855 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:22.857 helix_view::handlers::lsp [INFO] Version (17) is out of date for File("/home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl") (expected (18)), dropping PublishDiagnostic notification
2025-06-22T18:59:24.515 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":19}}}
2025-06-22T18:59:24.519 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":19,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:24.528 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:27.046 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didChange","params":{"contentChanges":[{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\n\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n"}],"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":20}}}
2025-06-22T18:59:27.049 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":20,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:27.059 helix_vcs::diff::worker [INFO] Diff computation timed out, update of diffs might appear delayed
2025-06-22T18:59:27.167 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/signatureHelp","params":{"position":{"character":0,"line":3},"textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}},"id":10}
2025-06-22T18:59:27.167 helix_lsp::transport [INFO] julia <- {"id":10,"jsonrpc":"2.0","result":{"signatures":[],"activeSignature":0,"activeParameter":0}}
2025-06-22T18:59:29.428 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"textDocument/didSave","params":{"text":"# todo: write modular components, and make places for those boxes to be fitted:\n# gpi[pe,pi], pe[visit,on/ffline], pi[greedy/egreedy]\nmodule MCPrediction\n\nimport Distributions\nimport GLMakie\n\ninclude(\"./RLExperimentsUtils.jl\")\ninclude(\"./Easy21.jl\")\n\nimport .RLExperimentsUtils:@do_while,sample,isconverged\nimport .Easy21\n\nPolicy = Matrix{Float64}\nStateValue = Vector{Float64}\nActionValue = Matrix{Float64}\n\nfunction createrandompolicy()\n\t# π = rand(length(Easy21.states), length(Easy21.actions))\n\t# π ./= sum(π, dims=2)\n\tπ = Matrix((rand(Distributions.Dirichlet(length(Easy21.actions), 1), length(Easy21.states)))')\n\t# samples from triangle rather than cube -> triangle projection\n\t# # verify random policy\n\t# begin\n\t# \tπ = createrandompolicy()\n\t# \tisvalidpolicy = all(isapprox.(sum(π, dims=2), 1.0))\n\t# \tstring(\"ispolicyvalid ?= \", isvalidpolicy)\n\t# end\n\tπ\nend\n\nfunction createepisode(s₁::Easy21.StartState, π::Policy)::Easy21.Episode\n\tepisode = []\n\ts = s₁\n\tr = 0\n\twhile r == 0\n\t\tsi = Easy21.stateindex(s)\n\t\tπ_s = π[si, :]\n\t\ta = sample(Easy21.actions, π_s)\n\t\t(s′, r) = Easy21.step(s, a)\n\t\tpush!(episode, (s, a, r))\n\t\ts = s′\n\tend\n\t\n\tepisode\nend\n\nfunction mcpolicyevaluation(q₁::ActionValue, π::Policy; k=UInt64)::ActionValue\n\tγ = 1.0  # should I include this inside Env?\n\n\tq̂_π = q₁\n\tq_π = zeros((size(q₁))...)\n\n\tg = zeros((size(q₁))...)\n\tn = zeros((size(q₁))...)\n\n\tnum_iter = 0\n\t@do_while begin\n\t\tnum_iter += 1\n\t\tq_π = q̂_π\n\t\t\n\t\ts = Easy21.initialstate()\n\t\tepisode = createepisode(s, π)\n\t\t\n\t\tgₜ = 0\n\t\tfor t in length(episode):-1:1\n\t\t\tsₜ, aₜ, r = episode[t]\n\t\t\tgₜ = + r + γ * gₜ\n\t\t\t(si, ai) = (Easy21.stateindex(sₜ), Easy21.actionindex(aₜ))\n\t\t\tn[si, ai] += 1\n\t\t\tg[si, ai] += gₜ\n\t\tend\n\t\tq̂_π = g ./ n\n\n\t\tisactionvaluenotconverged = !(isconverged(q_π, q̂_π))\n\t\t(isactionvaluenotconverged) ? nothing : println(\"ActionValue Converged! num_iter = \", num_iter)\n\tend ((num_iter < k) && (isactionvaluenotconverged))\n\n\tq_π\nend\n\nfunction main()\n\ts₁ = Easy21.initialstate()\n\tπ₁ = createrandompolicy()\n\tepisode = createepisode(s₁, π₁)\n\n\tq₁ = zeros(length(Easy21.states), length(Easy21.actions))\n\tq_π₁ = mcpolicyevaluation(q₁, π₁; k=2000000)\n\tv_π₁ = vec(maximum(q_π₁, dims=2))\n\n\tplot(v_π₁)\nend\n\n# PLOTTING-STATEVALUE === {{{\nfunction plot(v_π₁::StateValue)\n\txs = zeros(length(Easy21.states))\n\tys = zeros(length(Easy21.states))\n\tzs = zeros(length(Easy21.states))\n\t\n\tfor si in 1:length(Easy21.states)\n\t\ts = Easy21.states[si]\n\t\t(y, x) = (dealerfirstcard, playersum) = s\n\t\tz = v_π₁[si]\n\t\n\t\txs[si] = x\n\t\tys[si] = y\n\t\tzs[si] = z\n\tend\n\n\tfig = GLMakie.Figure()\n\tax3 = GLMakie.Axis3(\n\t\tfig[1, 1],\n\t\txlabel = \"Dealer Showing\", \n\t\tylabel = \"Player Sum\", \n\t\tzlabel = \"v_π₁(s)\",\n\t\ttitle = \"mcpolicyevaluaiton\"\n\t)\n\t# GLMakie.limits!(ax3, (1, 21), (1, 10), (-1, 1))\n\tGLMakie.surface!(ax3, xs, ys, zs)\n\n\tfig\nend\n# }}} === PLOTTING-STATEVALUE\n\n\n# PRETTY-PRINTING === {{{\n\nfunction Base.show(io::IO, ::MIME\"text/plain\", actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n\nfunction Base.show(io::IO, actionvalue::ActionValue)\n\tprint(io, \"[\\n\")\n\tfor i in 1:size(actionvalue)[1]\n\t    print(io, i, ' ', actionvalue[i, :], '\\n')  # todo, format: right-align the state number column\n    end\n\tprint(io, \"]\\n\")\nend\n# }}} === PRETTY-PRINTING\n\nend # module MCPrediction\n","textDocument":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl"}}}
2025-06-22T18:59:29.429 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"shutdown","id":11}
2025-06-22T18:59:29.585 helix_lsp::transport [INFO] julia <- {"method":"textDocument/publishDiagnostics","params":{"uri":"file:///home/vishal/Documents/Repos/RLExperiments/src/MCPrediction.jl","version":20,"diagnostics":[]},"jsonrpc":"2.0"}
2025-06-22T18:59:29.585 helix_lsp::transport [INFO] julia <- {"id":11,"jsonrpc":"2.0","result":null}
2025-06-22T18:59:29.585 helix_lsp::transport [INFO] julia -> {"jsonrpc":"2.0","method":"exit"}

[haxfn]

Other information:

julia version 1.11.5
LanguageServer v4.1.0
helix 25.01.1 (171dfc6)

Linux turing 6.15.2-arch1-1 #1 SMP PREEMPT_DYNAMIC Tue, 10 Jun 2025 21:32:33 +0000 x86_64 GNU/Linux

using no config in languages.toml for julia

Configured language servers:
✓ julia: /home/vishal/.juliaup/bin/julia
Configured debug adapter: None
Configured formatter: None
Tree-sitter parser: ✓
Highlight queries: ✓
Textobject queries: ✓
Indent queries: ✓