Renamed AutoEncoderMLJ to AutoEncoder and started isolating MLJ model…

…s in their own submodule

src/BetaML.jl

@@ -30,6 +30,7 @@ include("Clustering/Clustering.jl") # Clustering (hard) algorithms
 include("GMM/GMM.jl")               # GMM-based learners (clustering, fitter, regression) 
 include("Imputation/Imputation.jl") # (Missing) imputation algorithms
 include("Utils/Utils_extra.jl")     # Utility functions that depend on some BetaML functionality. Set them here to avoid recursive dependence
+include("Bmlj/Bmlj.jl")               # MLJ Interface module
 
 # "Merging" of the modules...
 @force    using .Api
@@ -50,6 +51,7 @@ include("Utils/Utils_extra.jl")     # Utility functions that depend on some Beta
 @reexport using .GMM
 @force    using .Imputation
 @reexport using .Imputation
+import .Bmlj # some MLJ models have the same name as BetaML models, set them in a separate interface submodule
 
 # ------------------------------------------------------------------------------
 #MLJ interface...
@@ -58,7 +60,7 @@ const MLJ_TREES_MODELS      = (DecisionTreeClassifier, DecisionTreeRegressor, Ra
 const MLJ_CLUSTERING_MODELS = (KMeans, KMedoids, GaussianMixtureClusterer)
 const MLJ_IMPUTERS_MODELS   = (SimpleImputer, GaussianMixtureImputer, RandomForestImputer,GeneralImputer) # these are the name of the MLJ models, not the BetaML ones...
 const MLJ_NN_MODELS         = (NeuralNetworkRegressor,MultitargetNeuralNetworkRegressor, NeuralNetworkClassifier)
-const MLJ_OTHER_MODELS      = (GaussianMixtureRegressor,MultitargetGaussianMixtureRegressor,AutoEncoderMLJ)
+const MLJ_OTHER_MODELS      = (GaussianMixtureRegressor,MultitargetGaussianMixtureRegressor,Bmlj.AutoEncoder)
 const MLJ_INTERFACED_MODELS = (MLJ_PERCEPTRON_MODELS..., MLJ_TREES_MODELS..., MLJ_CLUSTERING_MODELS..., MLJ_IMPUTERS_MODELS..., MLJ_NN_MODELS..., MLJ_OTHER_MODELS...) 
 
 

src/Utils/Utils_extra.jl

@@ -265,9 +265,6 @@ function inverse_predict(m::AutoEncoder,X)
     end
     return xtemp|> makematrix
 end
-
-include("Utils_MLJ.jl")     # Utility functions that depend on some BetaML functionality. Set them here to avoid recursive dependence
-
 end
 
 

test/Project.toml

@@ -9,6 +9,7 @@ DocStringExtensions = "ffbed154-4ef7-542d-bbb7-c09d3a79fcae"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 MLDatasets = "eb30cadb-4394-5ae3-aed4-317e484a6458"
 MLJBase = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
+MLJTestInterface = "72560011-54dd-4dc2-94f3-c5de45b75ecd"
 PDMats = "90014a1f-27ba-587c-ab20-58faa44d9150"
 Pipe = "b98c9c47-44ae-5843-9183-064241ee97a0"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"

test/Utils_tests.jl

@@ -726,15 +726,28 @@ s2  = silhouette(pd,[1,1,2,2])
 # NEW TEST
 println("Testing MLJ interface for Utils....")
 import MLJBase
-const Mlj = MLJBase
-
-X, y        = Mlj.@load_iris
-model       = AutoEncoderMLJ(outdims=2,rng=copy(TESTRNG))
-ae          = Mlj.machine(model, X)
-Mlj.fit!(ae)
-X_latent    = Mlj.transform(ae, X)
-X_recovered = Mlj.inverse_transform(ae,X_latent)
-@test relative_mean_error(Mlj.matrix(X),X_recovered) < 0.05
+const MLJ = MLJBase
+
+X, y        = MLJ.@load_iris
+model       = BetaML.Bmlj.AutoEncoder(outdims=2,rng=copy(TESTRNG))
+ae          = MLJ.machine(model, X)
+MLJ.fit!(ae)
+X_latent    = MLJ.transform(ae, X)
+X_recovered = MLJ.inverse_transform(ae,X_latent)
+@test relative_mean_error(MLJ.matrix(X),X_recovered) < 0.05
+
+import MLJTestInterface
+
+@testset "generic mlj interface test" begin
+    fails, summary = MLJTestInterface.test(
+        [BetaML.Bmlj.AutoEncoder,],
+        MLJTestInterface.make_regression()[1];
+        mod=@__MODULE__,
+        verbosity=0, # bump to debug
+        throw=false, # set to true to debug
+    )
+    @test isempty(fails)
+end
 
 #=
 using Random, StableRNGs