Fix sklearn v0.18 cross_validation compat

README.rst

@@ -55,7 +55,7 @@ Example
     [5 rows x 65 columns]
 
     # split to training and test data
-    >>> train_df, test_df = df.cross_validation.train_test_split()
+    >>> train_df, test_df = df.model_selection.train_test_split()
 
     # create estimator (accessor is mapped to sklearn namespace)
     >>> estimator = df.svm.LinearSVC()

doc/source/sklearn.rst

@@ -42,36 +42,41 @@ You can create ``ModelFrame`` instance from ``scikit-learn`` datasets directly.
 
 Following table shows ``scikit-learn`` module and corresponding ``ModelFrame`` module. Some accessors has its abbreviated versions.
 
-================================  ==========================================
+================================  ======================================================
 ``scikit-learn``                  ``ModelFrame`` accessor
-================================  ==========================================
+================================  ======================================================
+``sklearn.calibration``           ``ModelFrame.calibration``
 ``sklearn.cluster``               ``ModelFrame.cluster``
 ``sklearn.covariance``            ``ModelFrame.covariance``
 ``sklearn.cross_decomposition``   ``ModelFrame.cross_decomposition``
-``sklearn.cross_validation``      ``ModelFrame.cross_validation``, ``crv``
+``sklearn.cross_validation``      ``ModelFrame.cross_validation``, ``crv`` (deprecated)
 ``sklearn.datasets``              (not accesible from accessor)
 ``sklearn.decomposition``         ``ModelFrame.decomposition``
+``sklearn.discriminant_analysis`` ``ModelFrame.discriminant_analysis``, ``da``
 ``sklearn.dummy``                 ``ModelFrame.dummy``
 ``sklearn.ensemble``              ``ModelFrame.ensemble``
 ``sklearn.feature_extraction``    ``ModelFrame.feature_extraction``
 ``sklearn.feature_selection``     ``ModelFrame.feature_selection``
-``sklearn.gaussian_process``      ``ModelFrame.gaussian_process``
-``sklearn.grid_search``           ``ModelFrame.grid_search``
+``sklearn.gaussian_process``      ``ModelFrame.gaussian_process``, ``gp``
+``sklearn.grid_search``           ``ModelFrame.grid_search`` (deprecated)
 ``sklearn.isotonic``              ``ModelFrame.isotonic``
 ``sklearn.kernel_approximation``  ``ModelFrame.kernel_approximation``
-``sklearn.lda``                   ``ModelFrame.lda``
-``sklearn.learning_curve``        ``ModelFrame.learning_curve``
+``sklearn.kernel_ridge``          ``ModelFrame.kernel_ridge``
+``sklearn.lda``                   ``ModelFrame.lda`` (deprecated)
+``sklearn.learning_curve``        ``ModelFrame.learning_curve`` (deprecated)
 ``sklearn.linear_model``          ``ModelFrame.linear_model``, ``lm``
 ``sklearn.manifold``              ``ModelFrame.manifold``
 ``sklearn.metrics``               ``ModelFrame.metrics``
 ``sklearn.mixture``               ``ModelFrame.mixture``
+``sklearn.model_selection``       ``ModelFrame.model_selection``, ``ms``
 ``sklearn.multiclass``            ``ModelFrame.multiclass``
+``sklearn.multioutput``           ``ModelFrame.multioutput``
 ``sklearn.naive_bayes``           ``ModelFrame.naive_bayes``
 ``sklearn.neighbors``             ``ModelFrame.neighbors``
 ``sklearn.neural_network``        ``ModelFrame.neural_network``
 ``sklearn.pipeline``              ``ModelFrame.pipeline``
 ``sklearn.preprocessing``         ``ModelFrame.preprocessing``, ``pp``
-``sklearn.qda``                   ``ModelFrame.qda``
+``sklearn.qda``                   ``ModelFrame.qda`` (deprecated)
 ``sklearn.semi_supervised``       ``ModelFrame.semi_supervised``
 ``sklearn.svm``                   ``ModelFrame.svm``
 ``sklearn.tree``                  ``ModelFrame.tree``
@@ -277,7 +282,7 @@ Cross Validation
 
 .. code-block:: python
 
-   >>> train_df, test_df = df.cross_validation.train_test_split()
+   >>> train_df, test_df = df.model_selection.train_test_split()
    >>> train_df
         .target  sepal length  sepal width  petal length  petal width
    124        2           6.7          3.3           5.7          2.1
@@ -312,30 +317,17 @@ Cross Validation
    [38 rows x 5 columns]
 
 
-Also, there are some iterative classes which returns indexes for training sets and test sets. You can slice ``ModelFrame`` using these indexes.
-
-.. code-block:: python
-
-   >>> kf = df.cross_validation.KFold(n=150, n_folds=3)
-   >>> for train_index, test_index in kf:
-   ...    print('training set shape: ', df.iloc[train_index, :].shape,
-   ...          'test set shape: ', df.iloc[test_index, :].shape)
-   ('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
-   ('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
-   ('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
-
-
-For further simplification, ``ModelFrame.cross_validation.iterate`` can accept such iterators and returns ``ModelFrame`` corresponding to training and test data.
+You can iterate over Splitter classes via ``ModelFrame.model_selection.iterate`` which returns ``ModelFrame`` corresponding to training and test data.
 
 .. code-block:: python
 
-   >>> kf = df.cross_validation.KFold(n=150, n_folds=3)
+   >>> kf = df.model_selection.KFold(n_splits=3)
    >>> for train_df, test_df in df.cross_validation.iterate(kf):
    ...    print('training set shape: ', train_df.shape,
    ...          'test set shape: ', test_df.shape)
-   ('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
-   ('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
-   ('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
+   training set shape:  (112, 5) test set shape:  (38, 5)
+   training set shape:  (112, 5) test set shape:  (38, 5)
+   training set shape:  (112, 5) test set shape:  (38, 5)
 
 Grid Search
 -----------
@@ -349,8 +341,8 @@ You can perform grid search using ``ModelFrame.fit``.
    ...                    {'kernel': ['linear'], 'C': [1, 10, 100]}]
 
    >>> df = pdml.ModelFrame(datasets.load_digits())
-   >>> cv = df.grid_search.GridSearchCV(df.svm.SVC(C=1), tuned_parameters,
-   ...                                  cv=5, scoring='precision')
+   >>> cv = df.model_selection.GridSearchCV(df.svm.SVC(C=1), tuned_parameters,
+   ...                                      cv=5)
 
    >>> df.fit(cv)
 
@@ -363,7 +355,7 @@ In addition, ``ModelFrame.grid_search`` has a ``describe`` function to organize
 
 .. code-block:: python
 
-   >>> df.grid_search.describe(cv)
+   >>> df.model_selection.describe(cv)
           mean       std    C   gamma  kernel
    0  0.974108  0.013139    1  0.0010     rbf
    1  0.951416  0.020010    1  0.0001     rbf

doc/source/whatsnew.rst

@@ -8,7 +8,7 @@ v0.4.0
 Enhancement
 ^^^^^^^^^^^
 
-- Support scikit-learn v0.17.x.
+- Support scikit-learn v0.17.x and v0.18.0.
 - Support imbalanced-learn via ``.imbalance`` accessor.
 
 Bug Fix

pandas_ml/core/frame.py

@@ -575,11 +575,15 @@ def _cross_decomposition(self):
     @property
     @Appender(_shared_docs['skaccessor'] % dict(module='cross_validation'))
     def cross_validation(self):
+        msg = '.cross_validation is deprecated. Use .ms or .model_selection'
+        warnings.warn(msg, DeprecationWarning, stacklevel=2)
         return self._cross_validation
 
     @property
     @Appender(_shared_docs['skaccessor'] % dict(module='cross_validation'))
     def crv(self):
+        msg = '.crv is deprecated. Use .ms or .model_selection'
+        warnings.warn(msg, DeprecationWarning, stacklevel=2)
         return self._cross_validation
 
     @cache_readonly
@@ -666,6 +670,8 @@ def _gaussian_process(self):
     @property
     @Appender(_shared_docs['skaccessor'] % dict(module='grid_search'))
     def grid_search(self):
+        msg = '.grid_search is deprecated. Use .ms or .model_selection'
+        warnings.warn(msg, DeprecationWarning, stacklevel=2)
         return self._grid_search
 
     @cache_readonly
@@ -722,6 +728,8 @@ def lda(self):
     @property
     @Appender(_shared_docs['skaccessor'] % dict(module='learning_curve'))
     def learning_curve(self):
+        msg = '.learning_curve is deprecated. Use .ms or .model_selection'
+        warnings.warn(msg, DeprecationWarning, stacklevel=2)
         return self._learning_curve
 
     @cache_readonly

pandas_ml/skaccessors/cross_validation.py

@@ -9,7 +9,7 @@
 
 class CrossValidationMethods(_AccessorMethods):
     """
-    Accessor to ``sklearn.cross_validation``.
+    Deprecated. Accessor to ``sklearn.cross_validation``.
     """
 
     _module_name = 'sklearn.cross_validation'

pandas_ml/skaccessors/grid_search.py

@@ -7,7 +7,7 @@
 
 class GridSearchMethods(_AccessorMethods):
     """
-    Accessor to ``sklearn.grid_search``.
+    Deprecated. Accessor to ``sklearn.grid_search``.
     """
 
     _module_name = 'sklearn.grid_search'

pandas_ml/skaccessors/learning_curve.py

@@ -5,7 +5,7 @@
 
 class LearningCurveMethods(_AccessorMethods):
     """
-    Accessor to ``sklearn.learning_curve``.
+    Deprecated. Accessor to ``sklearn.learning_curve``.
     """
 
     _module_name = 'sklearn.learning_curve'

pandas_ml/skaccessors/model_selection.py

@@ -23,8 +23,11 @@ def StratifiedShuffleSplit(self, *args, **kwargs):
 
         - ``y``: ``ModelFrame.target``
         """
-        target = self._target
-        return self._module.StratifiedShuffleSplit(target.values, *args, **kwargs)
+        if _SKLEARN_ge_018:
+            return self._module.StratifiedShuffleSplit(*args, **kwargs)
+        else:
+            target = self._target
+            return self._module.StratifiedShuffleSplit(target.values, *args, **kwargs)
 
     def iterate(self, cv, reset_index=False):
         """
@@ -45,7 +48,12 @@ def iterate(self, cv, reset_index=False):
                 msg = "{0} is not a subclass of BaseCrossValidator"
                 warnings.warn(msg.format(cv.__class__.__name__))
 
-            for train_index, test_index in cv.split(self._df.index):
+            if isinstance(cv, self._module.StratifiedShuffleSplit):
+                gen = cv.split(self._df.data.values, self._df.target.values)
+            else:
+                gen = cv.split(self._df.index)
+
+            for train_index, test_index in gen:
                     train_df = self._df.iloc[train_index, :]
                     test_df = self._df.iloc[test_index, :]
                     if reset_index:

pandas_ml/skaccessors/test/test_model_selection.py

@@ -239,19 +239,22 @@ def test_check_cv(self):
         self.assertIsInstance(result, ms.KFold)
 
     def test_StratifiedShuffleSplit(self):
-        return
-
         iris = datasets.load_iris()
         df = pdml.ModelFrame(iris)
         sf1 = df.model_selection.StratifiedShuffleSplit(random_state=self.random_state)
-        sf2 = ms.StratifiedShuffleSplit(iris.target, random_state=self.random_state)
+        sf2 = ms.StratifiedShuffleSplit(random_state=self.random_state)
 
         # consume generator
         ind1 = [x for x in sf1.split(df.data.values, df.target.values)]
         ind2 = [x for x in sf2.split(iris.data, iris.target)]
 
-        self.assert_numpy_array_equal(ind1[0], ind1[0])
-        self.assert_numpy_array_equal(ind1[1], ind2[1])
+        for i1, i2 in zip(ind1, ind2):
+            self.assertIsInstance(i1, tuple)
+            self.assertEqual(len(i1), 2)
+            self.assertIsInstance(i2, tuple)
+            self.assertEqual(len(i2), 2)
+            self.assert_numpy_array_equal(i1[0], i1[0])
+            self.assert_numpy_array_equal(i1[1], i2[1])
 
         sf1 = df.model_selection.StratifiedShuffleSplit(random_state=self.random_state)
         with tm.assert_produces_warning(UserWarning):