rely on mixins to set estimator_type, removed unused imports

python-package/xgboost/compat.py

@@ -47,24 +47,17 @@ def lazy_isinstance(instance: Any, module: str, name: str) -> bool:
     from sklearn.base import BaseEstimator as XGBModelBase
     from sklearn.base import ClassifierMixin as XGBClassifierBase
     from sklearn.base import RegressorMixin as XGBRegressorBase
-    from sklearn.preprocessing import LabelEncoder
 
     try:
-        from sklearn.model_selection import KFold as XGBKFold
         from sklearn.model_selection import StratifiedKFold as XGBStratifiedKFold
     except ImportError:
-        from sklearn.cross_validation import KFold as XGBKFold
         from sklearn.cross_validation import StratifiedKFold as XGBStratifiedKFold
 
     # sklearn.utils Tags types can be imported unconditionally once
     # xgboost's minimum scikit-learn version is 1.6 or higher
     try:
-        from sklearn.utils import ClassifierTags as _sklearn_ClassifierTags
-        from sklearn.utils import RegressorTags as _sklearn_RegressorTags
         from sklearn.utils import Tags as _sklearn_Tags
     except ImportError:
-        _sklearn_ClassifierTags = object
-        _sklearn_RegressorTags = object
         _sklearn_Tags = object
 
     SKLEARN_INSTALLED = True
@@ -82,14 +75,8 @@ class XGBClassifierBase:  # type: ignore[no-redef]
     class XGBRegressorBase:  # type: ignore[no-redef]
         """Dummy class for sklearn.base.RegressorMixin."""
 
-    class LabelEncoder:  # type: ignore[no-redef]
-        """Dummy class for sklearn.preprocessing.LabelEncoder."""
-
-    XGBKFold = None
     XGBStratifiedKFold = None
 
-    _sklearn_ClassifierTags = object
-    _sklearn_RegressorTags = object
     _sklearn_Tags = object
     _sklearn_version = object
 

python-package/xgboost/sklearn.py

@@ -35,8 +35,6 @@
     XGBClassifierBase,
     XGBModelBase,
     XGBRegressorBase,
-    _sklearn_ClassifierTags,
-    _sklearn_RegressorTags,
     _sklearn_Tags,
     _sklearn_version,
     import_cupy,
@@ -840,7 +838,7 @@ def __sklearn_tags__(self) -> _sklearn_Tags:
         # take whatever tags are provided by BaseEstimator, then modify
         # them with XGBoost-specific values
         return self._update_sklearn_tags_from_dict(
-            tags=XGBModelBase.__sklearn_tags__(self),  # pylint: disable=no-member
+            tags=super().__sklearn_tags__(),  # pylint: disable=no-member
             tags_dict=self._more_tags(),
         )
 
@@ -1554,12 +1552,9 @@ def _more_tags(self) -> Dict[str, bool]:
         return tags
 
     def __sklearn_tags__(self) -> _sklearn_Tags:
-        tags = XGBModel.__sklearn_tags__(self)
-        tags.estimator_type = "classifier"
+        tags = super().__sklearn_tags__()
         tags_dict = self._more_tags()
-        tags.classifier_tags = _sklearn_ClassifierTags(
-            multi_label=tags_dict["multilabel"]
-        )
+        tags.classifier_tags.multi_label = tags_dict["multilabel"]
         return tags
 
     @_deprecate_positional_args
@@ -1849,10 +1844,8 @@ def _more_tags(self) -> Dict[str, bool]:
         return tags
 
     def __sklearn_tags__(self) -> _sklearn_Tags:
-        tags = XGBModel.__sklearn_tags__(self)
-        tags.estimator_type = "regressor"
+        tags = super().__sklearn_tags__()
         tags_dict = self._more_tags()
-        tags.regressor_tags = _sklearn_RegressorTags()
         tags.target_tags.multi_output = tags_dict["multioutput"]
         tags.target_tags.single_output = not tags_dict["multioutput_only"]
         return tags

tests/python/test_with_sklearn.py

@@ -826,6 +826,32 @@ def get_tm(clf: xgb.XGBClassifier) -> str:
     assert clf.get_params()["tree_method"] is None
 
 
+def test_get_params_works_as_expected():
+    # XGBModel -> BaseEstimator
+    params = xgb.XGBModel(max_depth=2).get_params()
+    assert params["max_depth"] == 2
+    # 'objective' defaults to None in the signature of XGBModel
+    assert params["objective"] is None
+
+    # XGBRegressor -> XGBModel -> BaseEstimator
+    params = xgb.XGBRegressor(max_depth=3).get_params()
+    assert params["max_depth"] == 3
+    # 'objective' defaults to 'reg:squarederror' in the signature of XGBRegressor
+    assert params["objective"] == "reg:squarederror"
+    # 'colsample_bynode' defaults to 'None' for XGBModel (which XGBRegressor inherits from), so it
+    # should be in get_params() output
+    assert params["colsample_bynode"] is None
+
+    # XGBRFRegressor -> XGBRegressor -> XGBModel -> BaseEstimator
+    params = xgb.XGBRFRegressor(max_depth=4, objective="reg:tweedie").get_params()
+    assert params["max_depth"] == 4
+    # 'objective' is a keyword argument for XGBRegressor, so it should be in get_params() output
+    # ... but values passed through kwargs should override the default from the signature of XGBRegressor
+    assert params["objective"] == "reg:tweedie"
+    # 'colsample_bynode' defaults to 0.8 for XGBRFRegressor...that should be preferred to the None from XGBRegressor
+    assert params["colsample_bynode"] == 0.8
+
+
 def test_kwargs_error():
     params = {'updater': 'grow_gpu_hist', 'subsample': .5, 'n_jobs': -1}
     with pytest.raises(TypeError):