Rel 0.16.0 - Fixes for StanPathfinder example

notebooks/Pathfinder/StanPathfinder.jl

@@ -0,0 +1,189 @@
+### A Pluto.jl notebook ###
+# v0.19.37
+
+using Markdown
+using InteractiveUtils
+
+# ╔═╡ 5084b8f0-65ac-4704-b1fc-2a9008132bd7
+using Pkg
+
+# ╔═╡ d7753cf6-7452-421a-a3ec-76e07646f808
+Pkg.activate(expanduser("~/.julia/dev/SR2StanPluto"))
+
+# ╔═╡ 550371ad-d411-4e66-9d63-7329322c6ea1
+begin
+    # Specific to this notebook
+	using StanSample
+	using StanOptimize
+    using StanPathfinder
+	using Distributions
+    using DataFrames
+	using CairoMakie
+end
+
+# ╔═╡ eb7ea04a-da52-4e69-ac3e-87dc7f014652
+md"## Pathfinder example"
+
+# ╔═╡ cf39df58-3371-4535-88e4-f3f6c0404500
+md" ###### Widen the cells."
+
+# ╔═╡ 0616ece8-ccf8-4281-bfed-9c1192edf88e
+html"""
+<style>
+    main {
+        margin: 0 auto;
+        max-width: 2000px;
+        padding-left: max(160px, 10%);
+        padding-right: max(160px, 15%);
+    }
+</style>
+"""
+
+# ╔═╡ 4755dab0-d228-41d3-934a-56f2863a5652
+md"###### A typical set of Julia packages to include in notebooks."
+
+# ╔═╡ da214635-a5ff-4c64-8930-683b11bd8531
+bernoulli_model = "
+data { 
+  int<lower=1> N; 
+  array[N] int<lower=0,upper=1> y;
+} 
+parameters {
+  real<lower=0,upper=1> theta;
+} 
+model {
+  theta ~ beta(1,1);
+  y ~ bernoulli(theta);
+}
+";
+
+# ╔═╡ 104b926e-5a36-41d7-a587-aeafa7bcd9c5
+data = Dict("N" => 10, "y" => [0, 1, 0, 1, 0, 0, 0, 0, 0, 1]);
+
+# ╔═╡ eb4a78c4-0522-44ae-b3ab-fe489c135026
+md" 
+!!! note
+
+Keep tmpdir across multiple runs to prevent re-compilation."
+
+# ╔═╡ 1ae7dd74-40c5-439c-a95c-3d31eecc8d89
+tmpdir = joinpath(pwd(), "tmp")
+
+# ╔═╡ edfb88bc-a66d-4035-af24-2b06e3253171
+begin
+	sm = SampleModel("bernoulli_sm", bernoulli_model)
+	rc_sm = stan_sample(sm; data)
+	if success(rc_sm)
+		df_sm = read_samples(sm, :dataframe)
+	end
+	describe(df_sm)
+end
+
+# ╔═╡ c67f204e-1cbc-4fa9-924f-2d8e4da0cb87
+d_normal = fit(Normal{Float64}, df_sm.theta)
+
+# ╔═╡ 06fe53e6-bb72-40dc-ac55-4e5cb363f2a3
+d_mle = fit_mle(Normal{Float64}, df_sm.theta)
+
+# ╔═╡ 9d9e6493-07e4-44aa-8552-f9edc72a4142
+begin
+	om = OptimizeModel("bernoulli_om", bernoulli_model)
+	rc_om = stan_optimize(om; data)
+	if success(rc_om)
+  		map, cnames = read_optimize(om)
+	end
+	map
+end
+
+# ╔═╡ cc7174d3-fa01-4ed7-a49e-7cd5ecd6d450
+begin
+	θ̂ = mean(map["theta"])
+	d_map = Normal(θ̂, θ̂ * (1-θ̂))
+end
+
+# ╔═╡ a46d3d38-3c56-4b43-a7ee-57323dbb1e3b
+let
+	global pm2 = PathfinderModel("bernoulli_pf", bernoulli_model)
+	rc = stan_pathfinder(pm2; data, num_chains=4, save_cmdstan_config=true)
+	if all(success.(rc))
+		a3d, colnames = read_pathfinder(pm2)
+		global dfa = StanSample.convert_a3d(a3d, colnames, Val(:dataframes))
+	end
+end
+
+# ╔═╡ 8b95d355-691a-47a5-8a10-ddac5fb3655a
+println(StanPathfinder.cmdline(pm2, 1))
+
+# ╔═╡ 886f5fee-07b2-4d7d-a7fd-3805570ce1b1
+log_file=StanBase.log_file_path(pm2.output_base, 1)
+
+# ╔═╡ c200e687-9967-4ebc-8b7b-6f1a14aadaa2
+sample_file=StanBase.sample_file_path(pm2.output_base, 1)
+
+# ╔═╡ 98a87e92-be9d-4735-9dba-b3b1d8f313d9
+pm2.num_chains=4
+
+# ╔═╡ 60490358-75a0-464e-b1e9-869219ceb2c9
+pm2.num_threads
+
+# ╔═╡ 037133fa-7c32-4bb7-993a-7251c63885dd
+let
+	str = read(joinpath(pm2.tmpdir, "$(pm2.name)_log_1.log"), String)
+    findfirst("Path [", str)
+    str = split(str[findfirst("Path [", str)[1]:end], "\n")[1:end-1]
+end
+
+# ╔═╡ d57e4518-c4ea-41c8-93bf-f34ecf35987c
+let
+	str = read(joinpath(pm2.tmpdir, "$(pm2.name)_log_2.log"), String)
+    findfirst("Path [", str)
+    str = split(str[findfirst("Path [", str)[1]:end], "\n")[1:end-1]
+end
+
+# ╔═╡ aa891570-ad41-43d2-81c9-dd4303343f71
+res2 = create_pathfinder_profile_df(pm2)
+
+# ╔═╡ fcc59e30-c156-42cb-8351-7d54b5c4bee9
+let
+	f = Figure(; size=(1000, 400))
+	pm = PathfinderModel("bernoulli_plot", bernoulli_model)
+	for i in 1:4
+		ax = Axis(f[1, i]; title="Density theta (pathfinder run $i)\n 1000 psis samples")
+		density!(dfa[i].theta)
+	end
+	for i in 1:4
+		rc = stan_sample(sm; data)
+		df_sm2 = read_samples(sm, :dataframe)
+		ax = Axis(f[2, i]; title="Density theta (mcmc run $i)\n 4000 draws")
+		density!(df_sm2.theta)
+	end
+	f
+end
+
+# ╔═╡ Cell order:
+# ╟─eb7ea04a-da52-4e69-ac3e-87dc7f014652
+# ╟─cf39df58-3371-4535-88e4-f3f6c0404500
+# ╠═0616ece8-ccf8-4281-bfed-9c1192edf88e
+# ╟─4755dab0-d228-41d3-934a-56f2863a5652
+# ╠═5084b8f0-65ac-4704-b1fc-2a9008132bd7
+# ╠═d7753cf6-7452-421a-a3ec-76e07646f808
+# ╠═550371ad-d411-4e66-9d63-7329322c6ea1
+# ╠═da214635-a5ff-4c64-8930-683b11bd8531
+# ╠═104b926e-5a36-41d7-a587-aeafa7bcd9c5
+# ╟─eb4a78c4-0522-44ae-b3ab-fe489c135026
+# ╠═1ae7dd74-40c5-439c-a95c-3d31eecc8d89
+# ╠═edfb88bc-a66d-4035-af24-2b06e3253171
+# ╠═c67f204e-1cbc-4fa9-924f-2d8e4da0cb87
+# ╠═06fe53e6-bb72-40dc-ac55-4e5cb363f2a3
+# ╠═9d9e6493-07e4-44aa-8552-f9edc72a4142
+# ╠═cc7174d3-fa01-4ed7-a49e-7cd5ecd6d450
+# ╠═a46d3d38-3c56-4b43-a7ee-57323dbb1e3b
+# ╠═8b95d355-691a-47a5-8a10-ddac5fb3655a
+# ╠═886f5fee-07b2-4d7d-a7fd-3805570ce1b1
+# ╠═c200e687-9967-4ebc-8b7b-6f1a14aadaa2
+# ╠═98a87e92-be9d-4735-9dba-b3b1d8f313d9
+# ╠═60490358-75a0-464e-b1e9-869219ceb2c9
+# ╠═037133fa-7c32-4bb7-993a-7251c63885dd
+# ╠═d57e4518-c4ea-41c8-93bf-f34ecf35987c
+# ╠═aa891570-ad41-43d2-81c9-dd4303343f71
+# ╠═fcc59e30-c156-42cb-8351-7d54b5c4bee9