Show warning message if initial_parameters is passed instead of initial_params (#176)

* Show warning message if initial_parameters is passed instead of initial_params

* run JuliaFormatter

* fix CI

* remove initial_parameters from kwargs in multi-chain functions

* format

* remove reduntant comment

* minor version bump

* add utility function for initial_parameters warning

Author: shravanngoswamii

Project.toml

@@ -3,7 +3,7 @@ uuid = "80f14c24-f653-4e6a-9b94-39d6b0f70001"
 keywords = ["markov chain monte carlo", "probabilistic programming"]
 license = "MIT"
 desc = "A lightweight interface for common MCMC methods."
-version = "5.8.0"
+version = "5.8.1"
 
 [deps]
 BangBang = "198e06fe-97b7-11e9-32a5-e1d131e6ad66"

src/sample.jl

@@ -105,6 +105,20 @@ function StatsBase.sample(
     return mcmcsample(rng, model, sampler, parallel, N, nchains; kwargs...)
 end
 
+# Utility function to check and warn about common kwargs mistakes
+function _check_initial_params_kwarg(kwargs)
+    if haskey(kwargs, :initial_parameters)
+        @warn "The `initial_parameters` keyword argument is not recognised; please use `initial_params` instead."
+        return true
+    end
+    return false
+end
+
+# Utility function to remove initial_parameters from kwargs after warning
+function _filter_initial_params_kwarg(kwargs)
+    return pairs((; (k => v for (k, v) in pairs(kwargs) if k !== :initial_parameters)...))
+end
+
 # Default implementations of regular and parallel sampling.
 function mcmcsample(
     rng::Random.AbstractRNG,
@@ -121,6 +135,9 @@ function mcmcsample(
     initial_state=nothing,
     kwargs...,
 )
+    # Warn if initial_parameters is passed instead of initial_params
+    _check_initial_params_kwarg(kwargs)
+
     # Check the number of requested samples.
     N > 0 || error("the number of samples must be ≥ 1")
     discard_initial >= 0 ||
@@ -405,6 +422,11 @@ function mcmcsample(
     initial_state=nothing,
     kwargs...,
 )
+    # Warn if initial_parameters is passed instead of initial_params and remove it from kwargs
+    if _check_initial_params_kwarg(kwargs)
+        kwargs = _filter_initial_params_kwarg(kwargs)
+    end
+
     # Check if actually multiple threads are used.
     if Threads.nthreads() == 1
         @warn "Only a single thread available: MCMC chains are not sampled in parallel"
@@ -588,6 +610,11 @@ function mcmcsample(
     initial_state=nothing,
     kwargs...,
 )
+    # Warn if initial_parameters is passed instead of initial_params and remove it from kwargs
+    if _check_initial_params_kwarg(kwargs)
+        kwargs = _filter_initial_params_kwarg(kwargs)
+    end
+
     # Check if actually multiple processes are used.
     if Distributed.nworkers() == 1
         @warn "Only a single process available: MCMC chains are not sampled in parallel"
@@ -727,6 +754,11 @@ function mcmcsample(
     initial_state=nothing,
     kwargs...,
 )
+    # Warn if initial_parameters is passed instead of initial_params and remove it from kwargs
+    if _check_initial_params_kwarg(kwargs)
+        kwargs = _filter_initial_params_kwarg(kwargs)
+    end
+
     # Check if the number of chains is larger than the number of samples
     if nchains > N
         @warn "Number of chains ($nchains) is greater than number of samples per chain ($N)"

test/sample.jl

@@ -31,6 +31,23 @@
             )
             @test chain[1].a == -1.8
             @test chain[1].b == 3.2
+
+            # test warning for initial_parameters (typo) in single-chain sampling
+            @test_logs (:warn, r"initial_parameters.*not recognised.*initial_params") sample(
+                MyModel(), MySampler(), 3; progress=false, initial_parameters=(b=1.0, a=2.0)
+            )
+
+            # test warning for initial_parameters (typo) in multi-chain sampling
+            @test_logs (:warn, r"initial_parameters.*not recognised.*initial_params") match_mode =
+                :any sample(
+                MyModel(),
+                MySampler(),
+                MCMCThreads(),
+                3,
+                2;
+                progress=false,
+                initial_parameters=(b=1.0, a=2.0),
+            )
         end
 
         @testset "IJulia" begin
@@ -282,6 +299,17 @@
             MyModel(), MySampler(), MCMCDistributed(), 5, 10; chain_type=MyChain
         )
 
+        # Test warning for initial_parameters (typo)
+        @test_logs (:warn, r"initial_parameters.*not recognised.*initial_params") sample(
+            MyModel(),
+            MySampler(),
+            MCMCDistributed(),
+            3,
+            2;
+            progress=false,
+            initial_parameters=(b=1.0, a=2.0),
+        )
+
         # Suppress output.
         logs, _ = collect_test_logs(; min_level=Logging.LogLevel(-1)) do
             sample(
@@ -408,6 +436,17 @@
             MyModel(), MySampler(), MCMCSerial(), 5, 10; chain_type=MyChain
         )
 
+        # Test warning for initial_parameters (typo)
+        @test_logs (:warn, r"initial_parameters.*not recognised.*initial_params") sample(
+            MyModel(),
+            MySampler(),
+            MCMCSerial(),
+            3,
+            2;
+            progress=false,
+            initial_parameters=(b=1.0, a=2.0),
+        )
+
         # Suppress output.
         logs, _ = collect_test_logs(; min_level=Logging.LogLevel(-1)) do
             sample(