Add support for training classifiers with one class present (#29)

* skip fitting the classifier if there is only one class

* test for ambiguities, non row_tables

* bump to `v0.3.3`

Author: tylerjthomas9

.github/workflows/CI.yml

@@ -40,10 +40,8 @@ jobs:
           - os: macos-latest
             version: '1'
             arch: x64
-    env:
-      PYTHON: ''
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
       - uses: julia-actions/setup-julia@v1
@@ -54,6 +52,6 @@ jobs:
       - uses: julia-actions/julia-buildpkg@v1
       - uses: julia-actions/julia-runtest@v1
       - uses: julia-actions/julia-processcoverage@v1
-      - uses: codecov/codecov-action@v2
+      - uses: codecov/codecov-action@v3
         with:
           file: lcov.info

Project.toml

@@ -1,7 +1,7 @@
 name = "CatBoost"
 uuid = "e2e10f9a-a85d-4fa9-b6b2-639a32100a12"
 authors = ["Beacon Biosignals, Inc."]
-version = "0.3.2"
+version = "0.3.3"
 
 [deps]
 MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
@@ -10,11 +10,15 @@ PythonCall = "6099a3de-0909-46bc-b1f4-468b9a2dfc0d"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
 [compat]
-Aqua = "0.5"
+Aqua = "0.8"
+DataFrames = "1.6"
+MLJBase = "1"
 MLJModelInterface = "1"
+MLJTestInterface = "0.2"
 OrderedCollections = "1.4"
 PythonCall = "0.9"
 Tables = "1.4"
+Test = "1.6"
 julia = "1.6"
 
 [extras]

format/Manifest.toml

@@ -11,82 +11,61 @@ uuid = "2a0f44e3-6c83-55bd-87e4-b1978d98bd5f"
 
 [[CSTParser]]
 deps = ["Tokenize"]
-git-tree-sha1 = "60e9121d9ea044c30a04397e59b00c5d9eb826ee"
+git-tree-sha1 = "3ddd48d200eb8ddf9cb3e0189fc059fd49b97c1f"
 uuid = "00ebfdb7-1f24-5e51-bd34-a7502290713f"
-version = "2.5.0"
+version = "3.3.6"
 
 [[CommonMark]]
-deps = ["Crayons", "JSON", "URIs"]
-git-tree-sha1 = "7632afc57f92720a01d9aedf23f413f4e5e21015"
+deps = ["Crayons", "JSON", "PrecompileTools", "URIs"]
+git-tree-sha1 = "532c4185d3c9037c0237546d817858b23cf9e071"
 uuid = "a80b9123-70ca-4bc0-993e-6e3bcb318db6"
-version = "0.8.1"
+version = "0.8.12"
 
 [[Compat]]
-deps = ["Base64", "Dates", "DelimitedFiles", "Distributed", "InteractiveUtils", "LibGit2", "Libdl", "LinearAlgebra", "Markdown", "Mmap", "Pkg", "Printf", "REPL", "Random", "SHA", "Serialization", "SharedArrays", "Sockets", "SparseArrays", "Statistics", "Test", "UUIDs", "Unicode"]
-git-tree-sha1 = "0a817fbe51c976de090aa8c997b7b719b786118d"
+deps = ["Dates", "LinearAlgebra", "UUIDs"]
+git-tree-sha1 = "886826d76ea9e72b35fcd000e535588f7b60f21d"
 uuid = "34da2185-b29b-5c13-b0c7-acf172513d20"
-version = "3.28.0"
+version = "4.10.1"
 
 [[Crayons]]
-git-tree-sha1 = "3f71217b538d7aaee0b69ab47d9b7724ca8afa0d"
+git-tree-sha1 = "249fe38abf76d48563e2f4556bebd215aa317e15"
 uuid = "a8cc5b0e-0ffa-5ad4-8c14-923d3ee1735f"
-version = "4.0.4"
+version = "4.1.1"
 
 [[DataStructures]]
 deps = ["Compat", "InteractiveUtils", "OrderedCollections"]
-git-tree-sha1 = "4437b64df1e0adccc3e5d1adbc3ac741095e4677"
+git-tree-sha1 = "ac67408d9ddf207de5cfa9a97e114352430f01ed"
 uuid = "864edb3b-99cc-5e75-8d2d-829cb0a9cfe8"
-version = "0.18.9"
+version = "0.18.16"
 
 [[Dates]]
 deps = ["Printf"]
 uuid = "ade2ca70-3891-5945-98fb-dc099432e06a"
 
-[[DelimitedFiles]]
-deps = ["Mmap"]
-uuid = "8bb1440f-4735-579b-a4ab-409b98df4dab"
-
-[[Distributed]]
-deps = ["Random", "Serialization", "Sockets"]
-uuid = "8ba89e20-285c-5b6f-9357-94700520ee1b"
-
-[[DocStringExtensions]]
-deps = ["LibGit2", "Markdown", "Pkg", "Test"]
-git-tree-sha1 = "9d4f64f79012636741cf01133158a54b24924c32"
-uuid = "ffbed154-4ef7-542d-bbb7-c09d3a79fcae"
-version = "0.8.4"
-
-[[Documenter]]
-deps = ["Base64", "Dates", "DocStringExtensions", "IOCapture", "InteractiveUtils", "JSON", "LibGit2", "Logging", "Markdown", "REPL", "Test", "Unicode"]
-git-tree-sha1 = "3ebb967819b284dc1e3c0422229b58a40a255649"
-uuid = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
-version = "0.26.3"
-
 [[Downloads]]
 deps = ["ArgTools", "LibCURL", "NetworkOptions"]
 uuid = "f43a241f-c20a-4ad4-852c-f6b1247861c6"
 
-[[IOCapture]]
-deps = ["Logging"]
-git-tree-sha1 = "377252859f740c217b936cebcd918a44f9b53b59"
-uuid = "b5f81e59-6552-4d32-b1f0-c071b021bf89"
-version = "0.1.1"
+[[Glob]]
+git-tree-sha1 = "97285bbd5230dd766e9ef6749b80fc617126d496"
+uuid = "c27321d9-0574-5035-807b-f59d2c89b15c"
+version = "1.3.1"
 
 [[InteractiveUtils]]
 deps = ["Markdown"]
 uuid = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
 
 [[JSON]]
 deps = ["Dates", "Mmap", "Parsers", "Unicode"]
-git-tree-sha1 = "81690084b6198a2e1da36fcfda16eeca9f9f24e4"
+git-tree-sha1 = "31e996f0a15c7b280ba9f76636b3ff9e2ae58c9a"
 uuid = "682c06a0-de6a-54ab-a142-c8b1cf79cde6"
-version = "0.21.1"
+version = "0.21.4"
 
 [[JuliaFormatter]]
-deps = ["CSTParser", "CommonMark", "DataStructures", "Documenter", "Pkg", "Tokenize"]
-git-tree-sha1 = "b947b46a3477e4c1ea32a7db66905d6f63dd7076"
+deps = ["CSTParser", "CommonMark", "DataStructures", "Glob", "Pkg", "PrecompileTools", "Tokenize"]
+git-tree-sha1 = "8f5295e46f594ad2d8652f1098488a77460080cd"
 uuid = "98e50ef6-434e-11e9-1051-2b60c6c9e899"
-version = "0.13.10"
+version = "1.0.45"
 
 [[LibCURL]]
 deps = ["LibCURL_jll", "MozillaCACerts_jll"]
@@ -132,20 +111,32 @@ uuid = "14a3606d-f60d-562e-9121-12d972cd8159"
 uuid = "ca575930-c2e3-43a9-ace4-1e988b2c1908"
 
 [[OrderedCollections]]
-git-tree-sha1 = "4fa2ba51070ec13fcc7517db714445b4ab986bdf"
+git-tree-sha1 = "dfdf5519f235516220579f949664f1bf44e741c5"
 uuid = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
-version = "1.4.0"
+version = "1.6.3"
 
 [[Parsers]]
-deps = ["Dates"]
-git-tree-sha1 = "c8abc88faa3f7a3950832ac5d6e690881590d6dc"
+deps = ["Dates", "PrecompileTools", "UUIDs"]
+git-tree-sha1 = "8489905bcdbcfac64d1daa51ca07c0d8f0283821"
 uuid = "69de0a69-1ddd-5017-9359-2bf0b02dc9f0"
-version = "1.1.0"
+version = "2.8.1"
 
 [[Pkg]]
 deps = ["Artifacts", "Dates", "Downloads", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "REPL", "Random", "SHA", "Serialization", "TOML", "Tar", "UUIDs", "p7zip_jll"]
 uuid = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 
+[[PrecompileTools]]
+deps = ["Preferences"]
+git-tree-sha1 = "03b4c25b43cb84cee5c90aa9b5ea0a78fd848d2f"
+uuid = "aea7be01-6a6a-4083-8856-8a6e6704d82a"
+version = "1.2.0"
+
+[[Preferences]]
+deps = ["TOML"]
+git-tree-sha1 = "00805cd429dcb4870060ff49ef443486c262e38e"
+uuid = "21216c6a-2e73-6563-6e65-726566657250"
+version = "1.4.1"
+
 [[Printf]]
 deps = ["Unicode"]
 uuid = "de0858da-6303-5e67-8744-51eddeeeb8d7"
@@ -164,21 +155,9 @@ uuid = "ea8e919c-243c-51af-8825-aaa63cd721ce"
 [[Serialization]]
 uuid = "9e88b42a-f829-5b0c-bbe9-9e923198166b"
 
-[[SharedArrays]]
-deps = ["Distributed", "Mmap", "Random", "Serialization"]
-uuid = "1a1011a3-84de-559e-8e89-a11a2f7dc383"
-
 [[Sockets]]
 uuid = "6462fe0b-24de-5631-8697-dd941f90decc"
 
-[[SparseArrays]]
-deps = ["LinearAlgebra", "Random"]
-uuid = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
-
-[[Statistics]]
-deps = ["LinearAlgebra", "SparseArrays"]
-uuid = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
-
 [[TOML]]
 deps = ["Dates"]
 uuid = "fa267f1f-6049-4f14-aa54-33bafae1ed76"
@@ -187,19 +166,15 @@ uuid = "fa267f1f-6049-4f14-aa54-33bafae1ed76"
 deps = ["ArgTools", "SHA"]
 uuid = "a4e569a6-e804-4fa4-b0f3-eef7a1d5b13e"
 
-[[Test]]
-deps = ["InteractiveUtils", "Logging", "Random", "Serialization"]
-uuid = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
-
 [[Tokenize]]
-git-tree-sha1 = "45b1932b0ec576159181bf75df71d6d86aa9c850"
+git-tree-sha1 = "3ac1ac11b09e8033ec93a7993acdb9b68252be6d"
 uuid = "0796e94c-ce3b-5d07-9a54-7f471281c624"
-version = "0.5.13"
+version = "0.5.27"
 
 [[URIs]]
-git-tree-sha1 = "97bbe755a53fe859669cd907f2d96aee8d2c1355"
+git-tree-sha1 = "67db6cc7b3821e19ebe75791a9dd19c9b1188f2b"
 uuid = "5c2747f8-b7ea-4ff2-ba2e-563bfd36b1d4"
-version = "1.3.0"
+version = "1.5.1"
 
 [[UUIDs]]
 deps = ["Random", "SHA"]

src/mlj_catboostclassifier.jl

@@ -73,14 +73,22 @@ function model_init(mlj_model::CatBoostClassifier; kw...)
 end
 
 function MMI.fit(mlj_model::CatBoostClassifier, verbosity::Int, data_pool, y_first)
-    verbose = verbosity <= 1 ? false : true
+    # Check if y_first has only one unique value
+    unique_classes = pyconvert(Vector, numpy.unique(data_pool.get_label()))
+    if length(unique_classes) == 1
+        # Skip training and store the single class
+        fitresult = (model=nothing, single_class=unique_classes[1], y_first=y_first)
+        cache = (; mlj_model=deepcopy(mlj_model))
+        report = (feature_importances=[],) # No feature importances in this case
+        return (fitresult, cache, report)
+    end
 
+    verbose = verbosity <= 1 ? false : true
     model = model_init(mlj_model; verbose)
     model.fit(data_pool)
 
     cache = (; mlj_model=deepcopy(mlj_model))
     report = (feature_importances=feature_importance(model),)
-
     fitresult = (model, y_first)
 
     return (fitresult, cache, report)
@@ -90,6 +98,14 @@ MMI.fitted_params(::CatBoostClassifier, model) = (model=model,)
 MMI.reports_feature_importances(::Type{<:CatBoostClassifier}) = true
 
 function MMI.predict(mlj_model::CatBoostClassifier, fitresult, X_pool)
+    if fitresult[1] === nothing
+        # Always predict the single class
+        n = nrow(X_pool)
+        classes = [fitresult.single_class]
+        probs = ones(n, 1)
+        return MMI.UnivariateFinite(classes, probs; pool=fitresult.y_first)
+    end
+
     model, y_first = fitresult
     classes = pyconvert(Array, model.classes_.tolist())
     py_preds = predict_proba(model, X_pool)
@@ -98,6 +114,12 @@ function MMI.predict(mlj_model::CatBoostClassifier, fitresult, X_pool)
 end
 
 function MMI.predict_mode(mlj_model::CatBoostClassifier, fitresult, X_pool)
+    if fitresult[1] === nothing
+        # Return probability 1 for the single class
+        n = nrow(X_pool)
+        return hcat(ones(n), zeros(n))
+    end
+
     model, y_first = fitresult
     py_preds = predict(model, X_pool)
     preds = pyconvert(Array, py_preds)

src/wrapper.jl

@@ -93,8 +93,9 @@ function feature_importance(py_model::Py)
     py_df_importance["importance"] = py_model.feature_importances_
     tbl_importance = pandas_to_tbl(py_df_importance)
     n_features = size(tbl_importance.name, 1)
-    feat_importance = [Symbol(tbl_importance.name[i]) => tbl_importance.importance[i] for i in
-                                                                                          1:n_features]
+    feat_importance = [Symbol(tbl_importance.name[i]) => tbl_importance.importance[i]
+                       for i in
+                           1:n_features]
     return feat_importance
 end
 

test/mlj_interface.jl

@@ -49,13 +49,12 @@
             @test isempty(failures)
         end
         @testset "CatBoostClassifier" begin
-            for data in [MLJTestInterface.make_binary(), MLJTestInterface.make_multiclass()]
+            for data in [MLJTestInterface.make_binary(),
+                         MLJTestInterface.make_multiclass(),
+                         MLJTestInterface.make_binary(; row_table=true),
+                         MLJTestInterface.make_multiclass(; row_table=false)]
                 X = data[1]
                 y = data[2]
-                # catboost fails if only 1 class is present when training
-                # MLJTestInterface splits the data down the middle, so the binary
-                # data only has one class during training
-                y[1] = y[end]
                 failures, summary = MLJTestInterface.test([CatBoostClassifier], X, y;
                                                           mod=@__MODULE__, verbosity=0, # bump to debug
                                                           throw=false)

test/runtests.jl

@@ -11,4 +11,4 @@ const MLJ_EXAMPLES_DIR = joinpath(@__DIR__, "..", "examples/mlj")
 include("wrapper.jl")
 include("mlj_interface.jl")
 
-Aqua.test_all(CatBoost; ambiguities=false)
+Aqua.test_all(CatBoost)