expand compat for Turing, increment version

Author: itsdfish

Project.toml

@@ -1,7 +1,7 @@
 name = "SequentialSamplingModels"
 uuid = "0e71a2a6-2b30-4447-8742-d083a85e82d1"
 authors = ["itsdfish"]
-version = "0.12.3"
+version = "0.12.4c"
 
 [deps]
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
@@ -44,7 +44,7 @@ SpecialFunctions = "1,2"
 Statistics = "1"
 StatsAPI = "1.0.0"
 StatsBase = "0.33.0,0.34.0"
-Turing = "0.29.0,0.30.0,0.31.0,0.32,0.33,0.34.0,0.35.0,0.36.0,0.37,0.38.0"
+Turing = "0.35.0,0.36.0,0.37,0.38.0,0.39.0"
 julia = "1"
 
 [extras]

docs/Project.toml

@@ -28,4 +28,4 @@ Plots = "1.0.0"
 StatsBase = "0.33.0,0.34.0"
 StatsModels = "0.7.0"
 StatsPlots = "0.15.0"
-Turing = "0.33.0,0.34,0.35.0,0.36.0, 0.37.0"
+Turing = "0.33.0,0.34,0.35.0,0.36.0, 0.37.0, 0.38.0, 0.39.0"

docs/src/predictive_distributions.md

@@ -2,6 +2,52 @@
 
 This tutorial explains the steps required for constructing and plotting prior and posterior predictive distributions of a sequential sampling models (SSMs). The primary function we will be using is `predict_distribution`, which allows you to generate prior or posterior predictive distributions from a given model. 
 
+## Full Code 
+
+You can reveal copy-and-pastable version of the full code by clicking the ▶ below.
+
+```@raw html
+<details>
+<summary><b>Show Full Code</b></summary>
+```
+```julia
+using Distributions
+using Plots
+using Random
+using SequentialSamplingModels
+using Turing 
+Random.seed!(1124)
+
+n_samples = 50
+rts = rand(Wald(ν=1.5, α=.8, τ=.3), n_samples)
+
+@model function wald_model(rts)
+    ν ~ truncated(Normal(1.5, 1), 0, Inf)
+    α ~ truncated(Normal(.8, 1), 0, Inf)
+    τ = 0.3
+    rts ~ Wald(ν, α, τ)
+    return (;ν, α, τ)
+end
+
+model = wald_model(rts)
+
+prior_chain = sample(model, Prior(), 1000)
+
+pred_model = predict_distribution(Wald; model, func=mean, n_samples)
+prior_preds = generated_quantities(pred_model, prior_chain)
+
+post_chain = sample(model, NUTS(1000, .85), 1000)
+post_preds = generated_quantities(pred_model, post_chain)
+
+histogram(prior_preds[:], xlims=(0,4), xlabel="Mean RT", ylabel="Density", norm=true, 
+    color=:grey, label="prior", grid=false)
+histogram!(post_preds[:], alpha=.7, color=:darkred, norm=true, label="posterior", grid=false)
+vline!([mean(rts)], linestyle=:dash, color=:black, linewidth=2, label="data")
+```
+```@raw html
+</details>
+```
+
 # Example 
 The first step is to load the required packages and set the seed for the random number generator.
 

ext/TuringExt.jl

@@ -1,7 +1,5 @@
 module TuringExt
 
-import DynamicPPL: reconstruct
-import DynamicPPL: vectorize
 import DynamicPPL: tovec
 using SequentialSamplingModels
 import SequentialSamplingModels: predict_distribution
@@ -31,8 +29,6 @@ Generates a predictive distribution for a statistic defined by `func`.
     return func(sim_data, args...; kwargs...)
 end
 
-vectorize(d::SSM2D, r::NamedTuple) = [r...]
-reconstruct(d::SSM2D, v::NamedTuple) = deepcopy(v)
 tovec(x::@NamedTuple{choice::C, rt::R}) where {C <: Integer, R <: AbstractFloat} =
     [x.choice, x.rt]
 end