tidy

Author: penelopeysm

test/ad.jl

@@ -253,8 +253,9 @@ end
                 conditioned_model = Turing.Inference.make_conditional(
                     model, varnames, deepcopy(global_vi)
                 )
-                rng = StableRNG(123)
-                @test run_ad(model, adtype; test=true, benchmark=false) isa Any
+                @test run_ad(
+                    model, adtype; rng=StableRNG(123), test=true, benchmark=false
+                ) isa Any
             end
         end
     end

test/ext/dynamichmc.jl

@@ -11,9 +11,8 @@ using StableRNGs: StableRNG
 using Turing
 
 @testset "TuringDynamicHMCExt" begin
-    rng = StableRNG(100)
     spl = externalsampler(DynamicHMC.NUTS())
-    chn = sample(rng, gdemo_default, spl, 10_000)
+    chn = sample(StableRNG(100), gdemo_default, spl, 10_000)
     check_gdemo(chn)
 end
 

test/mcmc/callbacks.jl

@@ -8,7 +8,6 @@ using Test, Turing, AbstractMCMC, Random, Distributions, LinearAlgebra
 end
 
 @testset "AbstractMCMC Callbacks Interface" begin
-    rng = Random.default_rng()
     model = test_normals()
 
     samplers = [
@@ -24,7 +23,10 @@ end
     for (name, sampler) in samplers
         @testset "$name" begin
             t1, s1 = AbstractMCMC.step(
-                rng, model, sampler; initial_params=Turing.Inference.init_strategy(sampler)
+                Random.default_rng(),
+                model,
+                sampler;
+                initial_params=Turing.Inference.init_strategy(sampler),
             )
 
             # ParamsWithStats returns named params (not θ[i])
@@ -46,6 +48,7 @@ end
     # NUTS second step has full AHMC transition metrics
     @testset "NUTS Transition Metrics" begin
         sampler = NUTS(10, 0.65)
+        rng = Random.default_rng()
         t1, s1 = AbstractMCMC.step(
             rng, model, sampler; initial_params=Turing.Inference.init_strategy(sampler)
         )

test/mcmc/emcee.jl

@@ -11,13 +11,10 @@ using Turing
 
 @testset "emcee.jl" begin
     @testset "gdemo" begin
-        rng = StableRNG(9876)
-
         n_samples = 1000
         n_walkers = 250
-
         spl = Emcee(n_walkers, 2.0)
-        chain = sample(rng, gdemo_default, spl, n_samples)
+        chain = sample(StableRNG(9876), gdemo_default, spl, n_samples)
         check_gdemo(chain)
     end
 

test/mcmc/gibbs.jl

@@ -578,9 +578,8 @@ end
                 # Sampler to use for Gibbs components.
                 hmc = HMC(0.1, 32)
                 sampler = Gibbs(@varname(s) => hmc, @varname(m) => hmc)
-                rng = StableRNG(42)
                 chain = sample(
-                    rng,
+                    StableRNG(42),
                     model,
                     sampler,
                     MCMCThreads(),
@@ -625,14 +624,13 @@ end
     end
 
     @testset "multiple varnames" begin
-        rng = Random.default_rng()
-
         @testset "with both `s` and `m` as random" begin
             model = gdemo(1.5, 2.0)
             vns = (@varname(s), @varname(m))
             spl = Gibbs(vns => MH())
 
             # `step`
+            rng = Random.default_rng()
             transition, state = AbstractMCMC.step(rng, model, spl)
             check_transition_varnames(transition, vns)
             for _ in 1:5
@@ -641,8 +639,7 @@ end
             end
 
             # `sample`
-            rng = StableRNG(42)
-            chain = sample(rng, model, spl, 1_000; progress=false)
+            chain = sample(StableRNG(42), model, spl, 1_000; progress=false)
             check_numerical(chain, [:s, :m], [49 / 24, 7 / 6]; atol=0.4)
         end
 
@@ -652,6 +649,7 @@ end
             spl = Gibbs(vns => MH())
 
             # `step`
+            rng = Random.default_rng()
             transition, state = AbstractMCMC.step(rng, model, spl)
             check_transition_varnames(transition, vns)
             for _ in 1:5
@@ -695,7 +693,6 @@ end
     end
 
     @testset "CSMC + ESS" begin
-        rng = Random.default_rng()
         model = MoGtest_default
         spl = Gibbs(
             (@varname(z1), @varname(z2), @varname(z3), @varname(z4)) => CSMC(15),
@@ -711,6 +708,7 @@ end
             @varname(mu2)
         )
         # `step`
+        rng = Random.default_rng()
         transition, state = AbstractMCMC.step(rng, model, spl)
         check_transition_varnames(transition, vns)
         for _ in 1:5
@@ -719,17 +717,16 @@ end
         end
 
         # Sample!
-        rng = StableRNG(42)
-        chain = sample(rng, MoGtest_default, spl, 1000; progress=false)
+        chain = sample(StableRNG(42), MoGtest_default, spl, 1000; progress=false)
         check_MoGtest_default(chain; atol=0.2)
     end
 
     @testset "CSMC + ESS (usage of implicit varname)" begin
-        rng = Random.default_rng()
         model = MoGtest_default_z_vector
         spl = Gibbs(@varname(z) => CSMC(15), @varname(mu1) => ESS(), @varname(mu2) => ESS())
         vns = (@varname(z), @varname(mu1), @varname(mu2))
         # `step`
+        rng = Random.default_rng()
         transition, state = AbstractMCMC.step(rng, model, spl)
         check_transition_varnames(transition, vns)
         for _ in 1:5
@@ -738,8 +735,7 @@ end
         end
 
         # Sample!
-        rng = StableRNG(42)
-        chain = sample(rng, model, spl, 1000; progress=false)
+        chain = sample(StableRNG(42), model, spl, 1000; progress=false)
         check_MoGtest_default_z_vector(chain; atol=0.2)
     end
 
@@ -775,9 +771,14 @@ end
         ]
         @testset "$(sampler_inner)" for sampler_inner in samplers_inner
             sampler = Gibbs(@varname(m1) => sampler_inner, @varname(m2) => sampler_inner)
-            rng = StableRNG(42)
             chain = sample(
-                rng, model, sampler, 1000; discard_initial=1000, thinning=10, n_adapts=0
+                StableRNG(42),
+                model,
+                sampler,
+                1000;
+                discard_initial=1000,
+                thinning=10,
+                n_adapts=0,
             )
             check_numerical(chain, [:m1, :m2], [-0.2, 0.6]; atol=0.1)
             check_logp_correct(sampler_inner)

test/mcmc/sghmc.jl

@@ -21,9 +21,8 @@ using Turing
     end
 
     @testset "sghmc inference" begin
-        rng = StableRNG(123)
         alg = SGHMC(; learning_rate=0.02, momentum_decay=0.5)
-        chain = sample(rng, gdemo_default, alg, 10_000)
+        chain = sample(StableRNG(123), gdemo_default, alg, 10_000)
         check_gdemo(chain; atol=0.1)
     end
 
@@ -39,9 +38,9 @@ end
     end
 
     @testset "sgld inference" begin
-        rng = StableRNG(1)
-
-        chain = sample(rng, gdemo_default, SGLD(; stepsize=PolynomialStepsize(0.5)), 20_000)
+        chain = sample(
+            StableRNG(1), gdemo_default, SGLD(; stepsize=PolynomialStepsize(0.5)), 20_000
+        )
         check_gdemo(chain; atol=0.25)
 
         # Weight samples by step sizes (cf section 4.2 in the paper by Welling and Teh)