[python] Re-enable scikit-learn 0.22+ support (#2949)

* Revert "specify the last supported version of scikit-learn (#2637)"

This reverts commit d1002776.

* ban scikit-learn 0.22.0 and skip broken test

* fix updated test

* fix lint test

* Revert "fix lint test"

This reverts commit 8b4db0805fe7a9e7f7eb0be3eac231f85026d196.

.ci/test.sh

@@ -74,7 +74,7 @@ if [[ $TASK == "r-package" ]]; then
     exit 0
 fi
 
-conda install -q -y -n $CONDA_ENV joblib matplotlib numpy pandas psutil pytest python-graphviz "scikit-learn<=0.21.3" scipy
+conda install -q -y -n $CONDA_ENV joblib matplotlib numpy pandas psutil pytest python-graphviz scikit-learn scipy
 
 if [[ $OS_NAME == "macos" ]] && [[ $COMPILER == "clang" ]]; then
     # fix "OMP: Error #15: Initializing libiomp5.dylib, but found libomp.dylib already initialized." (OpenMP library conflict due to conda's MKL)

.ci/test_windows.ps1

@@ -17,7 +17,7 @@ conda init powershell
 conda activate
 conda config --set always_yes yes --set changeps1 no
 conda update -q -y conda
-conda create -q -y -n $env:CONDA_ENV python=$env:PYTHON_VERSION joblib matplotlib numpy pandas psutil pytest python-graphviz "scikit-learn<=0.21.3" scipy ; Check-Output $?
+conda create -q -y -n $env:CONDA_ENV python=$env:PYTHON_VERSION joblib matplotlib numpy pandas psutil pytest python-graphviz scikit-learn scipy ; Check-Output $?
 conda activate $env:CONDA_ENV
 
 if ($env:TASK -eq "regular") {

docker/dockerfile-python

@@ -18,7 +18,7 @@ RUN apt-get update && \
     export PATH="$CONDA_DIR/bin:$PATH" && \
     conda config --set always_yes yes --set changeps1 no && \
     # lightgbm
-    conda install -q -y numpy scipy "scikit-learn<=0.21.3" pandas && \
+    conda install -q -y numpy scipy scikit-learn pandas && \
     git clone --recursive --branch stable --depth 1 https://github.com/Microsoft/LightGBM && \
     cd LightGBM/python-package && python setup.py install && \
     # clean

docker/gpu/dockerfile.gpu

@@ -75,8 +75,8 @@ RUN echo "export PATH=$CONDA_DIR/bin:"'$PATH' > /etc/profile.d/conda.sh && \
     rm ~/miniconda.sh
 
 RUN conda config --set always_yes yes --set changeps1 no && \
-    conda create -y -q -n py2 python=2.7 mkl numpy scipy "scikit-learn<=0.21.3" jupyter notebook ipython pandas matplotlib && \
-    conda create -y -q -n py3 python=3.6 mkl numpy scipy "scikit-learn<=0.21.3" jupyter notebook ipython pandas matplotlib
+    conda create -y -q -n py2 python=2.7 mkl numpy scipy scikit-learn jupyter notebook ipython pandas matplotlib && \
+    conda create -y -q -n py3 python=3.6 mkl numpy scipy scikit-learn jupyter notebook ipython pandas matplotlib
 
 #################################################################################################################
 #           LightGBM

docs/GPU-Tutorial.rst

-LightGBM GPU Tutorial
+LightGBM GPU Tutorial
 =====================
 
 The purpose of this document is to give you a quick step-by-step tutorial on GPU training.
@@ -78,7 +78,7 @@ If you want to use the Python interface of LightGBM, you can install it now (alo
 ::
 
     sudo apt-get -y install python-pip
-    sudo -H pip install setuptools numpy scipy "scikit-learn<=0.21.3" -U
+    sudo -H pip install setuptools numpy scipy scikit-learn -U
     cd python-package/
     sudo python setup.py install --precompile
     cd ..

docs/Python-API.rst

@@ -24,10 +24,6 @@ Training API
 Scikit-learn API
 ----------------
 
-.. warning::
-
-    The last supported version of scikit-learn is ``0.21.3``. Our estimators are incompatible with newer versions.
-
 .. autosummary::
     :toctree: pythonapi/
 

docs/Python-Intro.rst

@@ -15,11 +15,11 @@ Install
 -------
 
 Install Python-package dependencies,
-``setuptools``, ``wheel``, ``numpy`` and ``scipy`` are required, ``scikit-learn<=0.21.3`` is required for sklearn interface and recommended:
+``setuptools``, ``wheel``, ``numpy`` and ``scipy`` are required, ``scikit-learn`` is required for sklearn interface and recommended:
 
 ::
 
-    pip install setuptools wheel numpy scipy "scikit-learn<=0.21.3" -U
+    pip install setuptools wheel numpy scipy scikit-learn -U
 
 Refer to `Python-package`_ folder for the installation guide.
 

examples/python-guide/README.md

@@ -8,7 +8,7 @@ You should install LightGBM [Python-package](https://github.com/microsoft/LightG
 You also need scikit-learn, pandas, matplotlib (only for plot example), and scipy (only for logistic regression example) to run the examples, but they are not required for the package itself. You can install them with pip:
 
 ```
-pip install "scikit-learn<=0.21.3" pandas matplotlib scipy -U
+pip install scikit-learn pandas matplotlib scipy -U
 ```
 
 Now you can run examples in this folder, for example:

python-package/lightgbm/compat.py

@@ -116,16 +116,24 @@ try:
     from sklearn.preprocessing import LabelEncoder
     from sklearn.utils.class_weight import compute_sample_weight
     from sklearn.utils.multiclass import check_classification_targets
-    from sklearn.utils.validation import (assert_all_finite, check_X_y,
-                                          check_array, check_consistent_length)
+    from sklearn.utils.validation import assert_all_finite, check_X_y, check_array
     try:
         from sklearn.model_selection import StratifiedKFold, GroupKFold
         from sklearn.exceptions import NotFittedError
     except ImportError:
         from sklearn.cross_validation import StratifiedKFold, GroupKFold
         from sklearn.utils.validation import NotFittedError
+    try:
+        from sklearn.utils.validation import _check_sample_weight
+    except ImportError:
+        from sklearn.utils.validation import check_consistent_length
+
+        # dummy function to support older version of scikit-learn
+        def _check_sample_weight(sample_weight, X, dtype=None):
+            check_consistent_length(sample_weight, X)
+            return sample_weight
+
     SKLEARN_INSTALLED = True
-    from sklearn import __version__ as SKLEARN_VERSION
     _LGBMModelBase = BaseEstimator
     _LGBMRegressorBase = RegressorMixin
     _LGBMClassifierBase = ClassifierMixin
@@ -135,13 +143,12 @@ try:
     _LGBMGroupKFold = GroupKFold
     _LGBMCheckXY = check_X_y
     _LGBMCheckArray = check_array
-    _LGBMCheckConsistentLength = check_consistent_length
+    _LGBMCheckSampleWeight = _check_sample_weight
     _LGBMAssertAllFinite = assert_all_finite
     _LGBMCheckClassificationTargets = check_classification_targets
     _LGBMComputeSampleWeight = compute_sample_weight
 except ImportError:
     SKLEARN_INSTALLED = False
-    SKLEARN_VERSION = '0.0.0'
     _LGBMModelBase = object
     _LGBMClassifierBase = object
     _LGBMRegressorBase = object
@@ -151,7 +158,7 @@ except ImportError:
     _LGBMGroupKFold = None
     _LGBMCheckXY = None
     _LGBMCheckArray = None
-    _LGBMCheckConsistentLength = None
+    _LGBMCheckSampleWeight = None
     _LGBMAssertAllFinite = None
     _LGBMCheckClassificationTargets = None
     _LGBMComputeSampleWeight = None

python-package/lightgbm/sklearn.py

@@ -7,9 +7,9 @@ import warnings
 import numpy as np
 
 from .basic import Dataset, LightGBMError, _ConfigAliases
-from .compat import (SKLEARN_INSTALLED, SKLEARN_VERSION, _LGBMClassifierBase,
+from .compat import (SKLEARN_INSTALLED, _LGBMClassifierBase,
                      LGBMNotFittedError, _LGBMLabelEncoder, _LGBMModelBase,
-                     _LGBMRegressorBase, _LGBMCheckXY, _LGBMCheckArray, _LGBMCheckConsistentLength,
+                     _LGBMRegressorBase, _LGBMCheckXY, _LGBMCheckArray, _LGBMCheckSampleWeight,
                      _LGBMAssertAllFinite, _LGBMCheckClassificationTargets, _LGBMComputeSampleWeight,
                      argc_, range_, zip_, string_type, DataFrame, DataTable)
 from .engine import train
@@ -298,9 +298,6 @@ class LGBMModel(_LGBMModelBase):
         """
         if not SKLEARN_INSTALLED:
             raise LightGBMError('Scikit-learn is required for this module')
-        elif SKLEARN_VERSION > '0.21.3':
-            raise RuntimeError("The last supported version of scikit-learn is 0.21.3.\n"
-                               "Found version: {0}.".format(SKLEARN_VERSION))
 
         self.boosting_type = boosting_type
         self.objective = objective
@@ -547,7 +544,8 @@ class LGBMModel(_LGBMModelBase):
 
         if not isinstance(X, (DataFrame, DataTable)):
             _X, _y = _LGBMCheckXY(X, y, accept_sparse=True, force_all_finite=False, ensure_min_samples=2)
-            _LGBMCheckConsistentLength(_X, _y, sample_weight)
+            if sample_weight is not None:
+                sample_weight = _LGBMCheckSampleWeight(sample_weight, _X)
         else:
             _X, _y = X, y
 

python-package/setup.py

@@ -276,7 +276,7 @@ if __name__ == "__main__":
           install_requires=[
               'numpy',
               'scipy',
-              'scikit-learn<=0.21.3'
+              'scikit-learn!=0.22.0'
           ],
           maintainer='Guolin Ke',
           maintainer_email='guolin.ke@microsoft.com',

tests/python_package_test/test_sklearn.py

@@ -293,6 +293,11 @@ class TestSklearn(unittest.TestCase):
                 check_name = check.func.__name__ if hasattr(check, 'func') else check.__name__
                 if check_name == 'check_estimators_nan_inf':
                     continue  # skip test because LightGBM deals with nan
+                elif check_name == "check_no_attributes_set_in_init":
+                    # skip test because scikit-learn incorrectly asserts that
+                    # private attributes cannot be set in __init__
+                    # (see https://github.com/microsoft/LightGBM/issues/2628)
+                    continue
                 try:
                     check(name, estimator)
                 except SkipTest as message: