Merge pull request #63 from itsdfish/auto-juliaformatter-pr

Automatic JuliaFormatter.jl run

Examples/Binomial_ABC.jl

@@ -6,15 +6,15 @@ prior_loglike(θ) = logpdf(Beta(1, 1), θ)
 
 sample_prior() = rand(Beta(1, 1))
 
-bounds = ((0,1),)
+bounds = ((0, 1),)
 names = (:θ,)
 
 N = 10
-k = rand(Binomial(N, .5))
-data = (N = N,k = k)
+k = rand(Binomial(N, 0.5))
+data = (N = N, k = k)
 
 function loglike(data, θ)
-    (;N,k) = data
+    (; N, k) = data
     n_sim = 10^4
     counter(_) = rand(Binomial(N, θ)) == k ? 1 : 0
     cnt = mapreduce(counter, +, 1:n_sim)
@@ -23,14 +23,14 @@ end
 
 # loglike(θ, data) = logpdf(Binomial(data.N, θ), data.k)
 
-model = DEModel(; 
-    sample_prior, 
-    prior_loglike, 
-    loglike, 
+model = DEModel(;
+    sample_prior,
+    prior_loglike,
+    loglike,
     data,
     names
 )
 
-de = DE(;sample_prior, bounds, burnin=1000, Np=3, σ=.01)
+de = DE(; sample_prior, bounds, burnin = 1000, Np = 3, σ = 0.01)
 n_iter = 2000
-chains = sample(model, de, MCMCThreads(), n_iter, progress=true)
+chains = sample(model, de, MCMCThreads(), n_iter, progress = true)

Examples/Discrete_Example.jl

@@ -1,36 +1,33 @@
 # Note: poor performance
 
-
-
 # this may not be working properly
 
-
 cd(@__DIR__)
 using DifferentialEvolutionMCMC, Random, StatsBase, Distributions
 
 N = 30
 # Parameters for each cluster, we assume that each cluster is Gaussian distributed in the example.
 μs = [-3.5, 0.0]
 # Construct the data points.
-data = mapreduce(c -> rand(MvNormal([μs[c], μs[c]], 1.), N), hcat, 1:2)
+data = mapreduce(c -> rand(MvNormal([μs[c], μs[c]], 1.0), N), hcat, 1:2)
 
 priors = (
-    idx = (Categorical([.5,.5]),N),
-    μs = (Normal(0, 1),2),
+    idx = (Categorical([0.5, 0.5]), N),
+    μs = (Normal(0, 1), 2)
 )
 
-bounds = ((1,2),(0.0,Inf))
+bounds = ((1, 2), (0.0, Inf))
 
 function loglike(data, idx, μs)
     LL = 0.0
-    for (i,id) in enumerate(idx)
-        LL += logpdf(MvNormal([μs[id], μs[id]], 1.0), data[:,i])
+    for (i, id) in enumerate(idx)
+        LL += logpdf(MvNormal([μs[id], μs[id]], 1.0), data[:, i])
     end
     return LL
 end
 
-model = DEModel(priors=priors, model=loglike, data=data)
+model = DEModel(priors = priors, model = loglike, data = data)
 
-de = DE(bounds=bounds, burnin=1000, priors=priors, Np=15)
+de = DE(bounds = bounds, burnin = 1000, priors = priors, Np = 15)
 n_iter = 2000
-@elapsed chains = sample(model, de, n_iter, progress=true)
+@elapsed chains = sample(model, de, n_iter, progress = true)

Examples/Gaussian_Example.jl

@@ -19,7 +19,7 @@ end
 function sample_prior()
     μ = rand(Normal(0, 1))
     σ = rand(truncated(Cauchy(0, 1), 0, Inf))
-    return [μ,σ]
+    return [μ, σ]
 end
 
 # likelihood function 
@@ -34,24 +34,24 @@ data = rand(Normal(0.0, 1.0), 50)
 #                             configure sampler
 ###################################################################################
 # parameter names
-names = (:μ,:σ)
+names = (:μ, :σ)
 # parameter bounds
-bounds = ((-Inf,Inf),(0.0,Inf))
+bounds = ((-Inf, Inf), (0.0, Inf))
 
 # model object
-model = DEModel(; 
-    sample_prior, 
-    prior_loglike, 
-    loglike, 
+model = DEModel(;
+    sample_prior,
+    prior_loglike,
+    loglike,
     data,
     names
 )
 
 # DEMCMC sampler object
-de = DE(;sample_prior, bounds, burnin = 1000, Np = 6)
+de = DE(; sample_prior, bounds, burnin = 1000, Np = 6)
 # number of interations per particle
 n_iter = 2000
 ###################################################################################
 #                             estimate parameters
 ###################################################################################
-chains = sample(model, de, MCMCThreads(), n_iter, progress=true)
+chains = sample(model, de, MCMCThreads(), n_iter, progress = true)

Examples/Guassian_Example_Vector.jl

@@ -13,10 +13,10 @@ end
 function sample_prior()
     μ = rand(Normal(0, 1))
     σ = rand(truncated(Cauchy(0, 1), 0, Inf))
-    return [μ,σ]
+    return [μ, σ]
 end
 
-bounds = ((-Inf,Inf),(0.0,Inf))
+bounds = ((-Inf, Inf), (0.0, Inf))
 
 data = rand(Normal(0.0, 1.0), 50)
 
@@ -25,17 +25,16 @@ function loglike(data, θ...)
     return sum(logpdf.(Normal(μ, σ), data))
 end
 
-names = (:μ,:σ)
+names = (:μ, :σ)
 
-model = DEModel(; 
-    sample_prior, 
-    prior_loglike, 
-    loglike, 
+model = DEModel(;
+    sample_prior,
+    prior_loglike,
+    loglike,
     data,
     names
 )
 
-
-de = DE(;sample_prior, bounds, burnin=1000, Np=6)
+de = DE(; sample_prior, bounds, burnin = 1000, Np = 6)
 n_iter = 2000
-chains = sample(model, de, MCMCThreads(), n_iter, progress=true)
+chains = sample(model, de, MCMCThreads(), n_iter, progress = true)

Examples/Hierarchical_Example.jl

@@ -35,8 +35,8 @@ end
 # likelihood function 
 function loglike(data, μβ0, σβ0, β0, σ)
     LL = 0.0
-    for s in 1:length(data)
-        μ = μβ0 + β0[s] 
+    for s = 1:length(data)
+        μ = μβ0 + β0[s]
         y = data[s] .- μ
         LL += sum(logpdf.(Normal(0, σ), y))
     end
@@ -55,29 +55,29 @@ n_subj = 50
 σβ0 = 1.0
 # subject intercept deviations
 β0 = rand(Normal(0.0, σβ0), n_subj)
-σ = .5
+σ = 0.5
 
 sim(μβ0, β0, σ, n_data) = rand(Normal(μβ0 + β0, σ), n_data)
-data = [sim(μβ0, β0[i], σ, n_data) for i in 1:n_subj]
+data = [sim(μβ0, β0[i], σ, n_data) for i = 1:n_subj]
 ###################################################################################
 #                             configure sampler
 ###################################################################################
 # parameter bounds
 bounds = (
-    (-Inf,Inf),
-    (0.0,Inf),
-    (-Inf,Inf),
-    (0.0,Inf),
+    (-Inf, Inf),
+    (0.0, Inf),
+    (-Inf, Inf),
+    (0.0, Inf)
 )
 
 # parameter names
 names = (:μβ0, :σβ0, :β0, :σ)
 
 # model object 
 model = DEModel(;
-    sample_prior, 
-    prior_loglike, 
-    loglike, 
+    sample_prior,
+    prior_loglike,
+    loglike,
     data,
     names
 )
@@ -86,8 +86,8 @@ model = DEModel(;
 # update hyper parameters first
 # update lower level parameters second
 blocks = [
-    [true,true,fill(false, n_subj),true],
-    [false,false,fill(true, n_subj),false],
+    [true, true, fill(false, n_subj), true],
+    [false, false, fill(true, n_subj), false]
 ]
 blocks = as_union.(blocks)
 
@@ -97,36 +97,35 @@ blocks = as_union.(blocks)
 # push!(blocks, subj_blocks...)
 # blocks = as_union.(blocks)
 
-
 # use block updating on each iteration 
 blocking_on = x -> true
 
 # sampler object
 de = DE(;
     sample_prior,
-    bounds, 
+    bounds,
     sample = resample,
-    burnin = 20_000, 
+    burnin = 20_000,
     n_initial = (n_subj + 1) * 4,
     Np = 3,
     n_groups = 2,
     θsnooker = 0.1,
     blocking_on,
-    blocks,
+    blocks
     #generate_proposal = variable_gamma
 )
 ###################################################################################
 #                             estimate parameters
 ###################################################################################
 n_iter = 40_000
-chains = sample(model, de, MCMCThreads(), n_iter, progress=true)
+chains = sample(model, de, MCMCThreads(), n_iter, progress = true)
 ###################################################################################
 #                             Stan
 ###################################################################################
 using StanSample
 tmpdir = pwd() * "/tmp"
 
-idx = repeat(1:n_subj, inner=n_data)
+idx = repeat(1:n_subj, inner = n_data)
 y = vcat(data...)
 #xrep = repeat(x, inner=n_subj)
 stan_data = Dict(
@@ -135,13 +134,13 @@ stan_data = Dict(
     "n" => n_data * n_subj,
     "idx" => idx,
     "y" => y,
-    "x" => fill(0.0, n_data * n_subj),
+    "x" => fill(0.0, n_data * n_subj)
 )
 ###################################################################################################
 #                                        Define Model
 ###################################################################################################
 # load the Stan model file
-stream = open("temp.stan","r")
+stream = open("temp.stan", "r")
 stan_model = read(stream, String)
 close(stream)
 stan_model = SampleModel("temp", stan_model, tmpdir)
@@ -159,4 +158,4 @@ stan_sample(
 )
 samples = read_samples(stan_model, :mcmcchains)
 #rhats = filter(!isnan, MCMCChains.ess_rhat(samples).nt.rhat)
-#mean(rhats .> 1.01)
+#mean(rhats .> 1.01)

Examples/KDE.jl

@@ -1,4 +1,4 @@
 using KernelDensity, Distributions, Interpolations
 import KernelDensity: kernel_dist
 kernel_dist(::Type{Epanechnikov}, w::Float64) = Epanechnikov(0.0, w)
-kernel(data) = kde(data; kernel=Epanechnikov)
+kernel(data) = kde(data; kernel = Epanechnikov)

Examples/KDE_Example.jl

@@ -22,28 +22,27 @@ end
 function sample_prior()
     μ = rand(Normal(0, 1))
     σ = rand(truncated(Cauchy(0, 1), 0, Inf))
-    return [μ,σ]
+    return [μ, σ]
 end
 
 # parameter names
-names = (:μ,:σ)
+names = (:μ, :σ)
 # parameter bounds
-bounds = ((-Inf,Inf),(0.0,Inf))
+bounds = ((-Inf, Inf), (0.0, Inf))
 
 data = rand(Normal(0.0, 1.0), 50)
 
-
 # model object
-model = DEModel(; 
-    sample_prior, 
-    prior_loglike, 
-    loglike, 
+model = DEModel(;
+    sample_prior,
+    prior_loglike,
+    loglike,
     data,
     names
 )
 
 # DEMCMC sampler object
-de = DE(;sample_prior, bounds, burnin = 1000, Np = 6)
+de = DE(; sample_prior, bounds, burnin = 1000, Np = 6)
 # number of interations per particle
 n_iter = 2000
-chains = sample(model, de, MCMCThreads(), n_iter, progress=true)
+chains = sample(model, de, MCMCThreads(), n_iter, progress = true)

Examples/Multivariate_Guassian_Example.jl

@@ -16,7 +16,7 @@ data = rand(MvNormal(μs, 1.0 * I), n_d)
 function sample_prior()
     μ = rand(Normal(0, 1), n_μ)
     σ = rand(truncated(Cauchy(0, 1), 0, Inf))
-    return as_union([μ,σ])
+    return as_union([μ, σ])
 end
 
 # returns prior log likelihood
@@ -33,30 +33,30 @@ function loglike(data, μs, σ)
 end
 
 # upper and lower bounds of parameters
-bounds = ((-Inf,Inf),(0.0,Inf))
+bounds = ((-Inf, Inf), (0.0, Inf))
 # parameter names 
-names = (:μ,:σ)
+names = (:μ, :σ)
 
 # model object
-model = DEModel(; 
-    sample_prior, 
-    prior_loglike, 
-    loglike, 
+model = DEModel(;
+    sample_prior,
+    prior_loglike,
+    loglike,
     data,
     names
 )
 
 # DEMCMC sampler 
 de = DE(;
     sample_prior,
-    bounds, 
+    bounds,
     sample = resample,
-    burnin = 5000, 
+    burnin = 5000,
     n_initial = (n_μ + 1) * 4,
     Np = 3,
     n_groups = 1,
-    θsnooker = 0.1,
+    θsnooker = 0.1
 )
 # sample from the posterior distribution 
 n_iter = 50_000
-chains = sample(model, de, MCMCThreads(), n_iter, progress=true)
+chains = sample(model, de, MCMCThreads(), n_iter, progress = true)