Fix missing fallback argument for InitFromParams(::ModeResult)

HISTORY.md

@@ -1,3 +1,9 @@
+# 0.42.1
+
+`InitFromParams(mode_estimate)`, where `mode_estimate` was obtained from an optimisation on a Turing model, now accepts a second optional argument which provides a fallback initialisation strategy if some parameters are missing from `mode_estimate`.
+
+This also means that you can now invoke `returned(model, mode_estimate)` to calculate a model's return values given the parameters in `mode_estimate`.
+
 # 0.42.0
 
 ## DynamicPPL 0.39

Project.toml

@@ -1,6 +1,6 @@
 name = "Turing"
 uuid = "fce5fe82-541a-59a6-adf8-730c64b5f9a0"
-version = "0.42.0"
+version = "0.42.1"
 
 [deps]
 ADTypes = "47edcb42-4c32-4615-8424-f2b9edc5f35b"

src/optimisation/Optimisation.jl

@@ -128,12 +128,18 @@ function Base.show(io::IO, m::ModeResult)
 end
 
 """
-    InitFromParams(m::ModeResult)
+    InitFromParams(
+        m::ModeResult,
+        fallback::Union{AbstractInitStrategy,Nothing}=InitFromPrior()
+    )
 
-Initialize a model from the parameters stored in a `ModeResult`.
+Initialize a model from the parameters stored in a `ModeResult`. The `fallback` is used if
+some parameters are missing from the `ModeResult`.
 """
-function DynamicPPL.InitFromParams(m::ModeResult)
-    return DynamicPPL.InitFromParams(m.params)
+function DynamicPPL.InitFromParams(
+    m::ModeResult, fallback::Union{DynamicPPL.AbstractInitStrategy,Nothing}=InitFromPrior()
+)
+    return DynamicPPL.InitFromParams(m.params, fallback)
 end
 
 # Various StatsBase methods for ModeResult

test/Project.toml

@@ -53,7 +53,7 @@ Combinatorics = "1"
 Distributions = "0.25"
 DistributionsAD = "0.6.3"
 DynamicHMC = "2.1.6, 3.0"
-DynamicPPL = "0.39"
+DynamicPPL = "0.39.2"
 FiniteDifferences = "0.10.8, 0.11, 0.12"
 ForwardDiff = "0.10.12 - 0.10.32, 0.10, 1"
 HypothesisTests = "0.11"

test/optimisation/Optimisation.jl

@@ -587,6 +587,19 @@ using Turing
         @test chain[:σ][1] == mle.params[@varname(σ)]
     end
 
+    @testset "returned on ModeResult" begin
+        @model function f()
+            x ~ Normal()
+            2.0 ~ Normal(x)
+            return x + 1.0
+        end
+        model = f()
+        result = maximum_a_posteriori(model)
+        @test returned(model, result) == result.params[@varname(x)] + 1.0
+        result = maximum_likelihood(model)
+        @test returned(model, result) == result.params[@varname(x)] + 1.0
+    end
+
     # Issue: https://discourse.julialang.org/t/turing-mixture-models-with-dirichlet-weightings/112910
     @testset "Optimization with different linked dimensionality" begin
         @model demo_dirichlet() = x ~ Dirichlet(2 * ones(3))