Merge pull request #34 from JuliaAI/scitype-issue

Resolve scitype name clashes

Author: EssamWisam

Project.toml

@@ -14,6 +14,7 @@ MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
 OrderedCollections = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
 Parameters = "d96e819e-fc66-5662-9728-84c9c7592b0a"
 ScientificTypes = "321657f4-b219-11e9-178b-2701a2544e81"
+ScientificTypesBase = "30f210dd-8aff-4c5f-94ba-8e64358c1161"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 TableOperations = "ab02a1b2-a7df-11e8-156e-fb1833f50b87"
@@ -22,14 +23,15 @@ Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 [compat]
 BitBasis = "0.9"
 CategoricalArrays = "0.10"
-MLJModelInterface = "1.11"
 Combinatorics = "1"
 Dates = "1"
 Distributions = "0.25"
 LinearAlgebra = "1"
+MLJModelInterface = "1.11"
 OrderedCollections = "1"
 Parameters = "0.12"
-ScientificTypes = "3.0"
+ScientificTypes = "3.1.0"
+ScientificTypesBase = "3.0.0"
 Statistics = "1"
 StatsBase = "0.34"
 TableOperations = "1.2"
@@ -38,11 +40,11 @@ julia = "1.10"
 
 [extras]
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
-StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
 MLJBase = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
-Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
+StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
 StatsModels = "3eaba693-59b7-5ba5-a881-562e759f1c8d"
+Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
 test = ["Test", "DataFrames", "MLJBase", "Random", "StableRNGs", "StatsModels"]

src/MLJTransforms.jl

@@ -1,9 +1,12 @@
 module MLJTransforms
 using Tables
-using ScientificTypes
-using ScientificTypes: scitype
+# Note: The `scitype` in
+# MLJModelInterface clashes with the `scitype` in ScientificTypes. See also
+# https://github.com/JuliaAI/MLJBase.jl/issues/1002
+import ScientificTypes: elscitype, schema, coerce, ScientificTimeType
+using MLJModelInterface # exports `scitype`, which will call `ScientificTypes.scitype`,
+                        # once MLJBase is loaded (but this is not a dependency!)
 using CategoricalArrays
-using MLJModelInterface
 using TableOperations
 using StatsBase
 using LinearAlgebra
@@ -15,7 +18,6 @@ using Parameters
 using Dates
 using OrderedCollections
 
-
 const MMI = MLJModelInterface
 
 # Functions of generic use across transformers

test/encoders/contrast_encoder.jl

@@ -195,16 +195,16 @@ end
 
     df = DataFrame(X)
 
-    mf = ModelFrame(
-        @formula(age ~ (name + height + favnum)),
+    mf = StatsModels.ModelFrame(
+        StatsModels.@formula(age ~ (name + height + favnum)),
         df,
         contrasts = Dict(
             :name => StatsModels.ContrastsCoding(buildrandomcontrast(nothing, 3)),
             :favnum => StatsModels.ContrastsCoding(buildrandomcontrast(nothing, 4)),
         ),
     )
 
-    X_tr_sm = ModelMatrix(mf).m[:, 2:end]
+    X_tr_sm = StatsModels.ModelMatrix(mf).m[:, 2:end]
 
     @test X_tr_mlj == X_tr_sm
 end
@@ -221,24 +221,24 @@ end
     X_tr_mlj = Tables.matrix(X_tr)[:, 1:end-1]
     df = DataFrame(X)
 
-    mf = ModelFrame(
-        @formula(age ~ (name + height + favnum)),
+    mf = StatsModels.ModelFrame(
+        StatsModels.@formula(age ~ (name + height + favnum)),
         df,
         contrasts = Dict(
-            :name => HypothesisCoding(
+            :name => StatsModels.HypothesisCoding(
                 buildrandomhypothesis(nothing, 3);
                 levels = levels(X.name),
                 labels = [],
             ),
-            :favnum => HypothesisCoding(
+            :favnum => StatsModels.HypothesisCoding(
                 buildrandomhypothesis(nothing, 4);
                 levels = levels(X.favnum),
                 labels = [],
             ),
         ),
     )
 
-    X_tr_sm = ModelMatrix(mf).m[:, 2:end]
+    X_tr_sm = StatsModels.ModelMatrix(mf).m[:, 2:end]
 
     @test X_tr_mlj == X_tr_sm
 end
@@ -257,11 +257,11 @@ end
     for ind in 1:6
         stats_models(k, ind) = [
             StatsModels.ContrastsCoding(buildrandomcontrast(nothing, k)),
-            DummyCoding(; base = (k == 3) ? "Mary" : 10),
-            EffectsCoding(; base = (k == 3) ? "Mary" : 10),
-            SeqDiffCoding(),
-            HelmertCoding(),
-            HypothesisCoding(
+            StatsModels.DummyCoding(; base = (k == 3) ? "Mary" : 10),
+            StatsModels.EffectsCoding(; base = (k == 3) ? "Mary" : 10),
+            StatsModels.SeqDiffCoding(),
+            StatsModels.HelmertCoding(),
+            StatsModels.HypothesisCoding(
                 buildrandomhypothesis(nothing, k);
                 levels = (k == 3) ? levels(X.name) : levels(X.favnum),
                 labels = [],
@@ -277,8 +277,8 @@ end
 
         df = DataFrame(X)
 
-        mf = ModelFrame(
-            @formula(age ~ (name + height + favnum)),
+        mf = StatsModels.ModelFrame(
+            StatsModels.@formula(age ~ (name + height + favnum)),
             df,
             contrasts = Dict(
                 :name => stats_models(3, ind),
@@ -287,7 +287,7 @@ end
         )
 
         X_tr_mlj = Tables.matrix(X_tr)[:, 1:end-1]
-        X_tr_sm = ModelMatrix(mf).m[:, 2:end]
+        X_tr_sm = StatsModels.ModelMatrix(mf).m[:, 2:end]
         @test X_tr_mlj ≈ X_tr_sm
     end
 end
@@ -298,11 +298,11 @@ end
         for ind2 in 2:5
             stats_models(k, ind) = [
                 StatsModels.ContrastsCoding(buildrandomcontrast(nothing, k)),
-                DummyCoding(; base = (k == 3) ? "Mary" : 10),
-                EffectsCoding(; base = (k == 3) ? "Mary" : 10),
-                SeqDiffCoding(),
-                HelmertCoding(),
-                HypothesisCoding(
+                StatsModels.DummyCoding(; base = (k == 3) ? "Mary" : 10),
+                StatsModels.EffectsCoding(; base = (k == 3) ? "Mary" : 10),
+                StatsModels.SeqDiffCoding(),
+                StatsModels.HelmertCoding(),
+                StatsModels.HypothesisCoding(
                     buildrandomhypothesis(nothing, k);
                     levels = (k == 3) ? levels(X.name) : levels(X.favnum),
                     labels = [],
@@ -331,8 +331,8 @@ end
 
             df = DataFrame(X)
 
-            mf = ModelFrame(
-                @formula(age ~ (name + height + favnum)),
+            mf = StatsModels.ModelFrame(
+                StatsModels.@formula(age ~ (name + height + favnum)),
                 df,
                 contrasts = Dict(
                     :name => stats_models(3, ind1),
@@ -341,7 +341,7 @@ end
             )
 
             X_tr_mlj = Tables.matrix(X_tr)[:, 1:end-1]
-            X_tr_sm = ModelMatrix(mf).m[:, 2:end]
+            X_tr_sm = StatsModels.ModelMatrix(mf).m[:, 2:end]
 
             @test X_tr_mlj ≈ X_tr_sm
         end
@@ -358,7 +358,7 @@ end
     encoder = ContrastEncoder(ignore = true, ordered_factor = false)
     mach = machine(encoder, X)
     fit!(mach)
-    Xnew_transf = MMI.transform(mach, X)
+    Xnew_transf = MLJBase.transform(mach, X)
 
     # same output
     @test X_transf == Xnew_transf
@@ -392,7 +392,7 @@ end
             buildmatrix = matrix_func[i],
         )
         mach = fit!(machine(encoder, X))
-        Xnew = MMI.transform(mach, X)
+        Xnew = MLJBase.transform(mach, X)
 
         # Test Consistency with Types
         scs = schema(Xnew).scitypes
@@ -406,4 +406,4 @@ end
         @test last_type <: Integer && isconcretetype(last_type)
         @test last_sctype <: Count
     end
-end
+end

test/encoders/frequency_encoder.jl

@@ -5,7 +5,7 @@ using MLJTransforms: frequency_encoder_fit, frequency_encoder_transform
 
     X = dataset_forms[1]
     normalize = [false, true]
-    A_col, C_col, D_col, F_col = MMI.selectcols(X, [1, 3, 4, 6])
+    A_col, C_col, D_col, F_col = selectcols(X, [1, 3, 4, 6])
     for norm in normalize
         result = frequency_encoder_fit(X; normalize = norm)[:statistic_given_feat_val]
         enc =
@@ -72,7 +72,7 @@ end
         encoder = FrequencyEncoder(ignore = true, ordered_factor = false)
         mach = machine(encoder, X)
         fit!(mach)
-        Xnew_transf = MMI.transform(mach, X)
+        Xnew_transf = MLJBase.transform(mach, X)
 
         # same output
         @test X_transf == Xnew_transf
@@ -111,7 +111,7 @@ end
 
     encoder = FrequencyEncoder(ordered_factor = false, normalize = false)
     mach = fit!(machine(encoder, X))
-    Xnew = MMI.transform(mach, X)
+    Xnew = MLJBase.transform(mach, X)
 
 
     scs = schema(Xnew).scitypes

test/encoders/missingness_encoding.jl

@@ -170,7 +170,7 @@ end
     encoder = MissingnessEncoder(ignore = true, ordered_factor = false)
     mach = machine(encoder, X)
     fit!(mach)
-    Xnew_transf = MMI.transform(mach, X)
+    Xnew_transf =  MLJBase.transform(mach, X)
 
     # same output
     @test isequal(X_transf, Xnew_transf)
@@ -197,7 +197,7 @@ end
 
     encoder = MissingnessEncoder()
     mach = fit!(machine(encoder, Xm))
-    Xnew = MMI.transform(mach, Xm)
+    Xnew = MLJBase.transform(mach, Xm)
 
     schema(Xm)
     schema(Xnew)

test/encoders/ordinal_encoding.jl

@@ -15,7 +15,7 @@ push!(
     @test ordinal_encoder_fit(dataset_forms[1]) == ordinal_encoder_fit(dataset_forms[2])
     X = dataset_forms[1]
     result = ordinal_encoder_fit(X)[:index_given_feat_level]
-    A_col, C_col, D_col, F_col = MMI.selectcols(X, [1, 3, 4, 6])
+    A_col, C_col, D_col, F_col = selectcols(X, [1, 3, 4, 6])
     true_output = Dict{Symbol, Dict{Any, AbstractFloat}}(
         :F => Dict(
             "m" => findfirst(==("m"), levels(F_col)),
@@ -70,7 +70,7 @@ end
         encoder = OrdinalEncoder(ignore = true, ordered_factor = false)
         mach = machine(encoder, X)
         fit!(mach)
-        Xnew_transf = MMI.transform(mach, X)
+        Xnew_transf = MLJBase.transform(mach, X)
 
         # same output
         @test X_transf == Xnew_transf
@@ -108,7 +108,7 @@ end
 
     encoder = OrdinalEncoder(ordered_factor = false)
     mach = fit!(machine(encoder, X))
-    Xnew = MMI.transform(mach, X)
+    Xnew = MLJBase.transform(mach, X)
 
     scs = schema(Xnew).scitypes
     ts  = schema(Xnew).types
@@ -123,7 +123,7 @@ end
     ## Int32 case
     encoder = OrdinalEncoder(ordered_factor = false, output_type = Int32)
     mach = fit!(machine(encoder, X))
-    Xnew = MMI.transform(mach, X)
+    Xnew = MLJBase.transform(mach, X)
     scs = schema(Xnew).scitypes
     ts = schema(Xnew).types
     # Check scitypes for previously categorical features

test/encoders/target_encoding.jl

@@ -63,7 +63,7 @@ end
     X, y = classification_forms[1]
     n = length(y)
 
-    A_col, C_col, D_col, F_col = MMI.selectcols(X, [1, 3, 4, 6])
+    A_col, C_col, D_col, F_col = selectcols(X, [1, 3, 4, 6])
     true_output = Dict{Symbol, Dict{Any, AbstractFloat}}(
         :F => Dict(
             "m" => sum(y[F_col.=="m"] .== 0) / length(y[F_col.=="m"]),
@@ -119,7 +119,7 @@ end
     n = length(y)
     μ̂ = mean(y)
 
-    A_col, C_col, D_col, F_col = MMI.selectcols(X, [1, 3, 4, 6])
+    A_col, C_col, D_col, F_col = selectcols(X, [1, 3, 4, 6])
     true_output = Dict{Symbol, Dict{Any, AbstractFloat}}(
         :F => Dict(
             "m" => mean(y[F_col.=="m"]),
@@ -172,7 +172,7 @@ end
     y_classes = classes(y)
     n = length(y)
 
-    A_col, C_col, D_col, F_col = MMI.selectcols(X, [1, 3, 4, 6])
+    A_col, C_col, D_col, F_col = selectcols(X, [1, 3, 4, 6])
     true_output = Dict{Symbol, Dict{Any, AbstractVector{AbstractFloat}}}(
         :F => Dict(
             "m" =>
@@ -320,7 +320,7 @@ end
             TargetEncoder(ignore = true, ordered_factor = false, lambda = 0.5, m = 1.0)
         mach = machine(encoder, X, y)
         fit!(mach)
-        Xnew_transf = MMI.transform(mach, X)
+        Xnew_transf = MLJBase.transform(mach, X)
 
         # same output
         @test X_transf == Xnew_transf
@@ -368,7 +368,7 @@ end
     D = [true, false, true, false, true]
     E = [1, 2, 3, 4, 5]
 
-    # Define the target variable 
+    # Define the target variable
     y = ["c1", "c2", "c3", "c1", "c2"]
 
     # Combine into a named tuple
@@ -386,7 +386,7 @@ end
 
     encoder = TargetEncoder(ordered_factor = false, lambda = 1.0, m = 0)
     mach = fit!(machine(encoder, X, y))
-    Xnew = MMI.transform(mach, X)
+    Xnew = MLJBase.transform(mach, X)
 
     scs = schema(Xnew).scitypes
     ts  = schema(Xnew).types
@@ -396,4 +396,3 @@ end
     @test scs[end] === schema(X).scitypes[end]
     @test ts[end] == schema(X).types[end]
 end
-

test/generic.jl

@@ -142,7 +142,7 @@ end
 
 @testset "Test generic fit output" begin
     X = dataset_forms[1]
-    A_col, C_col, D_col, F_col = MMI.selectcols(X, [1, 3, 4, 6])
+    A_col, C_col, D_col, F_col = selectcols(X, [1, 3, 4, 6])
     result = dummy_encoder_fit(X)[:hash_given_feat_val]
     enc = (col, level) -> (hash(level))
     true_output = Dict{Symbol, Dict{Any, Any}}(

test/runtests.jl

@@ -1,21 +1,16 @@
 using MLJTransforms
 using Test
 using DataFrames
-using ScientificTypes
 using CategoricalArrays
-using MLJModelInterface
 using MLJBase
 using StatsBase
 using LinearAlgebra
-using StatsModels
+import StatsModels
 using Random
-const MMI = MLJModelInterface
 using LinearAlgebra
-using StatsModels
 
 # Other transformers
 using Tables, CategoricalArrays
-using ScientificTypes: scitype, schema
 using Statistics
 using StableRNGs
 stable_rng = StableRNGs.StableRNG(123)
@@ -40,4 +35,4 @@ include("transformers/other_transformers/interaction_transformer.jl")
 include("transformers/other_transformers/continuous_encoder.jl")
 include("transformers/other_transformers/univariate_boxcox_transformer.jl")
 include("transformers/other_transformers/standardizer.jl")
-include("transformers/other_transformers/univariate_discretizer.jl")
+include("transformers/other_transformers/univariate_discretizer.jl")