Merge pull request #68 from TuringLang/ml/isdone

add `state` to isdone signature

Project.toml

@@ -3,7 +3,7 @@ uuid = "80f14c24-f653-4e6a-9b94-39d6b0f70001"
 keywords = ["markov chain monte carlo", "probablistic programming"]
 license = "MIT"
 desc = "A lightweight interface for common MCMC methods."
-version = "3.0.1"
+version = "3.0.2"
 
 [deps]
 BangBang = "198e06fe-97b7-11e9-32a5-e1d131e6ad66"

README.md

@@ -44,7 +44,7 @@ StatsBase.sample(
 for regular and parallel sampling, respectively. In regular sampling, users may
 provide a function
 ```julia
-isdone(rng, model, sampler, samples, iteration; kwargs...)
+isdone(rng, model, sampler, samples, state, iteration; kwargs...)
 ```
 that returns `true` when sampling should end, and `false` otherwise, instead of
 a fixed number of samples `nsamples`. AbstractMCMC defines the abstract types

src/sample.jl

@@ -206,7 +206,7 @@ function mcmcsample(
         # Step through the sampler until stopping.
         i = 2
 
-        while !isdone(rng, model, sampler, samples, i; progress=progress, kwargs...)
+        while !isdone(rng, model, sampler, samples, state, i; progress=progress, kwargs...)
             # Discard thinned samples.
             for _ in 1:(thinning - 1)
                 # Obtain the next sample and state.

test/utils.jl

@@ -53,12 +53,13 @@ function isdone(
     model::MyModel,
     s::MySampler,
     samples,
+    state,
     iteration::Int;
     kwargs...
 )
     # Calculate the mean of x.b.
     bmean = mean(x.b for x in samples)
-    return abs(bmean) <= 0.001 || iteration >= 10_000
+    return abs(bmean) <= 0.001 || iteration >= 10_000 || state >= 10_000
 end
 
 # Set a default convergence function.