Merge pull request #170 from JuliaAI/dev

For a 0.7.0 release

.github/workflows/ci.yml

@@ -19,7 +19,7 @@ jobs:
       fail-fast: false
       matrix:
         version:
-          - '1.3'
+          - '1.6'
           - '1'
         os:
           - ubuntu-latest

Project.toml

@@ -1,7 +1,7 @@
 name = "MLJTuning"
 uuid = "03970b2e-30c4-11ea-3135-d1576263f10f"
 authors = ["Anthony D. Blaom <anthony.blaom@gmail.com>"]
-version = "0.6.16"
+version = "0.7.0"
 
 [deps]
 ComputationalResources = "ed09eef8-17a6-5b46-8889-db040fac31e3"
@@ -17,27 +17,7 @@ RecipesBase = "3cdcf5f2-1ef4-517c-9805-6587b60abb01"
 ComputationalResources = "0.3"
 Distributions = "0.22,0.23,0.24, 0.25"
 LatinHypercubeSampling = "1.7.2"
-MLJBase = "0.18.19, 0.19"
-MLJModelInterface = "0.4.1, 1.1.1"
+MLJBase = "0.20"
 ProgressMeter = "1.7.1"
 RecipesBase = "0.8,0.9,1"
-julia = "1.3"
-
-[extras]
-CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
-DecisionTree = "7806a523-6efd-50cb-b5f6-3fa6f1930dbb"
-Distances = "b4f34e82-e78d-54a5-968a-f98e89d6e8f7"
-Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
-LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
-MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
-MultivariateStats = "6f286f6a-111f-5878-ab1e-185364afe411"
-NearestNeighbors = "b8a86587-4115-5ab1-83bc-aa920d37bbce"
-ScientificTypes = "321657f4-b219-11e9-178b-2701a2544e81"
-StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
-Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
-StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
-Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
-Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
-
-[targets]
-test = ["CategoricalArrays", "DecisionTree", "Distances", "Distributions", "LinearAlgebra", "MLJModelInterface", "MultivariateStats", "NearestNeighbors", "ScientificTypes", "StableRNGs", "Statistics", "StatsBase", "Tables", "Test"]
+julia = "1.6"

src/MLJTuning.jl

@@ -21,7 +21,8 @@ export learning_curve!, learning_curve
 import MLJBase
 using MLJBase
 import MLJBase: Bounded, Unbounded, DoublyUnbounded,
-    LeftUnbounded, RightUnbounded, _process_accel_settings, chunks
+    LeftUnbounded, RightUnbounded, _process_accel_settings, chunks,
+    restore, save
 using RecipesBase
 using Distributed
 import Distributions

src/learning_curves.jl

@@ -92,13 +92,6 @@ Other key-word options are documented at [`TunedModel`](@ref).
 learning_curve(mach::Machine{<:Supervised}; kwargs...) =
     learning_curve(mach.model, mach.args...; kwargs...)
 
-# for backwards compatibility
-function learning_curve!(mach::Machine{<:Supervised}; kwargs...)
-    Base.depwarn("`learning_curve!` is deprecated, use `learning_curve` instead. ",
-            Core.Typeof(learning_curve!).name.mt.name)
-    learning_curve(mach; kwargs...)
-end
-
 function learning_curve(model::Supervised, args...;
                         resolution=30,
                         resampling=Holdout(),
@@ -299,8 +292,12 @@ end
 
     n_threads = Threads.nthreads()
     if n_threads == 1
-        return _tuning_results(rngs, CPU1(),
-                         tuned, rng_name, verbosity)
+        return _tuning_results(rngs,
+                               CPU1(),
+                               tuned,
+                               rows,
+                               rng_name,
+                               verbosity)
     end
 
     old_rng = recursive_getproperty(tuned.model.model, rng_name)

src/serialization.jl

@@ -0,0 +1,7 @@
+MLJModelInterface.save(::MLJTuning.EitherTunedModel, fitresult::Machine) =
+    serializable(fitresult)
+
+function MLJModelInterface.restore(::MLJTuning.EitherTunedModel, fitresult)
+    fitresult.fitresult = restore(fitresult.model, fitresult.fitresult)
+    return fitresult
+end

src/tuned_models.jl

@@ -67,7 +67,7 @@ const EitherTunedModel{T,M} =
 #todo update:
 """
     tuned_model = TunedModel(; model=<model to be mutated>,
-                             tuning=Grid(),
+                             tuning=RandomSearch(),
                              resampling=Holdout(),
                              range=nothing,
                              measure=nothing,
@@ -173,7 +173,10 @@ plus other key/value pairs specific to the `tuning` strategy.
 - `models`: Alternatively, an iterator of MLJ models to be explicitly
   evaluated. These may have varying types.
 
-- `tuning=Grid()`: tuning strategy to be applied (eg, `RandomSearch()`)
+- `tuning=RandomSearch()`: tuning strategy to be applied (eg, `Grid()`). See
+  the [Tuning
+  Models](https://alan-turing-institute.github.io/MLJ.jl/dev/tuning_models/#Tuning-Models)
+  section of the MLJ manual for a complete list of options.
 
 - `resampling=Holdout()`: resampling strategy (eg, `Holdout()`, `CV()`),
   `StratifiedCV()`) to be applied in performance evaluations
@@ -253,7 +256,7 @@ function TunedModel(; model=nothing,
             throw(ERR_NEED_EXPLICIT)
         end
     else
-        tuning === nothing && (tuning = Grid())
+        tuning === nothing && (tuning = RandomSearch())
     end
 
     # either a `model` is specified or we are in the case

test/Project.toml

@@ -0,0 +1,36 @@
+[deps]
+CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
+ComputationalResources = "ed09eef8-17a6-5b46-8889-db040fac31e3"
+DecisionTree = "7806a523-6efd-50cb-b5f6-3fa6f1930dbb"
+Distances = "b4f34e82-e78d-54a5-968a-f98e89d6e8f7"
+Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"
+Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
+LatinHypercubeSampling = "a5e1c1ea-c99a-51d3-a14d-a9a37257b02d"
+LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
+MLJBase = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
+MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
+MultivariateStats = "6f286f6a-111f-5878-ab1e-185364afe411"
+NearestNeighbors = "b8a86587-4115-5ab1-83bc-aa920d37bbce"
+Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
+ScientificTypes = "321657f4-b219-11e9-178b-2701a2544e81"
+Serialization = "9e88b42a-f829-5b0c-bbe9-9e923198166b"
+StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
+Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
+StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
+Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
+Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
+
+[compat]
+CategoricalArrays = "0.10"
+ComputationalResources = "0.3"
+DecisionTree = "0.10"
+Distances = "0.10"
+Distributions = "0.25"
+MLJBase = "0.20"
+MLJModelInterface = "1.3"
+MultivariateStats = "0.9"
+NearestNeighbors = "0.4"
+ScientificTypes = "3.0"
+StableRNGs = "1.0"
+StatsBase = "0.33"
+Tables = "1.6"

test/learning_curves.jl

@@ -206,16 +206,5 @@ end
 
 end
 
-@testset "deprecation of learning_curve!" begin
-    atom = KNNRegressor()
-    mach = machine(atom, X, y)
-    r = range(atom, :K, lower=1, upper=2)
-    @test_deprecated learning_curve!(mach;
-                                     range=r,
-                                     measure=LPLoss(),
-                                     verbosity=0)
-
-end
-
 end # module
 true

test/runtests.jl

@@ -59,6 +59,10 @@ end
         @test include("learning_curves.jl")
 end
 
+@testset "Serialization" begin
+    @test include("serialization.jl")
+end
+
 # @testset "julia bug" begin
 #     @test include("julia_bug.jl")
 # end

test/serialization.jl

@@ -0,0 +1,73 @@
+
+module TestSerialization
+
+using Test
+using MLJBase
+using Serialization
+using MLJTuning
+using ..Models
+
+function test_args(mach)
+    # Check source nodes are empty if any
+    for arg in mach.args
+        if arg isa Source 
+            @test arg == source()
+        end
+    end
+end
+
+function test_data(mach)
+    @test !isdefined(mach, :old_rows)
+    @test !isdefined(mach, :data)
+    @test !isdefined(mach, :resampled_data)
+    @test !isdefined(mach, :cache)
+end
+
+function generic_tests(mach₁, mach₂)
+    test_args(mach₂)
+    test_data(mach₂)
+    @test mach₂.state == -1
+    for field in (:frozen, :model, :old_model, :old_upstream_state, :fit_okay)
+        @test getfield(mach₁, field) == getfield(mach₂, field)
+    end
+end
+
+
+@testset "Test TunedModel" begin
+    filename = "tuned_model.jls"
+    X, y = make_regression(100)
+    base_model = DecisionTreeRegressor()
+    tuned_model = TunedModel(
+        model=base_model,
+        tuning=Grid(),
+        range=[range(base_model, :min_samples_split, values=[2,3,4])],
+    )
+    mach = machine(tuned_model, X, y)
+    fit!(mach, rows=1:50, verbosity=0)
+    smach = MLJBase.serializable(mach)
+    @test smach.fitresult isa Machine
+    @test smach.report == mach.report
+    generic_tests(mach, smach)
+
+    Serialization.serialize(filename, smach)
+    smach = Serialization.deserialize(filename)
+    MLJBase.restore!(smach)
+
+    @test MLJBase.predict(smach, X) == MLJBase.predict(mach, X)
+    @test fitted_params(smach) isa NamedTuple
+    @test report(smach) == report(mach)
+
+    rm(filename)
+
+    # End to end
+    MLJBase.save(filename, mach)
+    smach = machine(filename)
+    @test predict(smach, X) == predict(mach, X)
+
+    rm(filename)
+
+end
+
+end
+
+true