scikit-learn · amueller · Dec 20, 2018 · Jun 10, 2017 · Jun 10, 2017 · Jul 27, 2018
diff --git a/doc/modules/compose.rst b/doc/modules/compose.rst
@@ -76,13 +76,13 @@ filling in the names automatically::
 
 The estimators of a pipeline are stored as a list in the ``steps`` attribute::
 
-    >>> pipe.steps[0]
+    >>> pipe.steps[0]  # doctest: +NORMALIZE_WHITESPACE
     ('reduce_dim', PCA(copy=True, iterated_power='auto', n_components=None, random_state=None,
       svd_solver='auto', tol=0.0, whiten=False))
 
 and as a ``dict`` in ``named_steps``::
 
-    >>> pipe.named_steps['reduce_dim']
+    >>> pipe.named_steps['reduce_dim']  # doctest: +NORMALIZE_WHITESPACE
     PCA(copy=True, iterated_power='auto', n_components=None, random_state=None,
       svd_solver='auto', tol=0.0, whiten=False)
 

diff --git a/doc/modules/linear_model.rst b/doc/modules/linear_model.rst
@@ -185,7 +185,7 @@ for another implementation::
 
     >>> from sklearn import linear_model
     >>> reg = linear_model.Lasso(alpha=0.1)
-    >>> reg.fit([[0, 0], [1, 1]], [0, 1])
+    >>> reg.fit([[0, 0], [1, 1]], [0, 1])  # doctest: +NORMALIZE_WHITESPACE
     Lasso(alpha=0.1, copy_X=True, fit_intercept=True, max_iter=1000,
        normalize=False, positive=False, precompute=False, random_state=None,
        selection='cyclic', tol=0.0001, warm_start=False)
@@ -639,7 +639,7 @@ Bayesian Ridge Regression is used for regression::
     >>> X = [[0., 0.], [1., 1.], [2., 2.], [3., 3.]]
     >>> Y = [0., 1., 2., 3.]
     >>> reg = linear_model.BayesianRidge()
-    >>> reg.fit(X, Y)
+    >>> reg.fit(X, Y)  # doctest: +NORMALIZE_WHITESPACE
     BayesianRidge(alpha_1=1e-06, alpha_2=1e-06, compute_score=False, copy_X=True,
            fit_intercept=True, lambda_1=1e-06, lambda_2=1e-06, n_iter=300,
            normalize=False, tol=0.001, verbose=False)

diff --git a/doc/modules/model_evaluation.rst b/doc/modules/model_evaluation.rst
@@ -979,7 +979,7 @@ with a svm classifier in a binary class problem::
   >>> X = [[0], [1]]
   >>> y = [-1, 1]
   >>> est = svm.LinearSVC(random_state=0)
-  >>> est.fit(X, y)
+  >>> est.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
   LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,
        intercept_scaling=1, loss='squared_hinge', max_iter=1000,
        multi_class='ovr', penalty='l2', random_state=0, tol=0.0001,
@@ -997,7 +997,7 @@ with a svm classifier in a multiclass problem::
   >>> Y = np.array([0, 1, 2, 3])
   >>> labels = np.array([0, 1, 2, 3])
   >>> est = svm.LinearSVC()
-  >>> est.fit(X, Y)
+  >>> est.fit(X, Y)  # doctest: +NORMALIZE_WHITESPACE
   LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,
        intercept_scaling=1, loss='squared_hinge', max_iter=1000,
        multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,

diff --git a/doc/modules/preprocessing.rst b/doc/modules/preprocessing.rst
@@ -488,7 +488,7 @@ Continuing the example above::
 
   >>> enc = preprocessing.OneHotEncoder()
   >>> X = [['male', 'from US', 'uses Safari'], ['female', 'from Europe', 'uses Firefox']]
-  >>> enc.fit(X)  # doctest: +ELLIPSIS
+  >>> enc.fit(X)  # doctest: +ELLIPSIS +NORMALIZE_WHITESPACE
   OneHotEncoder(categorical_features=None, categories=None,
          dtype=<... 'numpy.float64'>, handle_unknown='error',
          n_values=None, sparse=True)
@@ -514,7 +514,7 @@ dataset::
     >>> # Note that for there are missing categorical values for the 2nd and 3rd
     >>> # feature
     >>> X = [['male', 'from US', 'uses Safari'], ['female', 'from Europe', 'uses Firefox']]
-    >>> enc.fit(X) # doctest: +ELLIPSIS
+    >>> enc.fit(X) # doctest: +ELLIPSIS +NORMALIZE_WHITESPACE
     OneHotEncoder(categorical_features=None,
            categories=[...],
            dtype=<... 'numpy.float64'>, handle_unknown='error',
@@ -532,7 +532,7 @@ columns for this feature will be all zeros
 
     >>> enc = preprocessing.OneHotEncoder(handle_unknown='ignore')
     >>> X = [['male', 'from US', 'uses Safari'], ['female', 'from Europe', 'uses Firefox']]
-    >>> enc.fit(X) # doctest: +ELLIPSIS
+    >>> enc.fit(X) # doctest: +ELLIPSIS  +NORMALIZE_WHITESPACE
     OneHotEncoder(categorical_features=None, categories=None,
            dtype=<... 'numpy.float64'>, handle_unknown='ignore',
            n_values=None, sparse=True)

diff --git a/doc/tutorial/statistical_inference/model_selection.rst b/doc/tutorial/statistical_inference/model_selection.rst
@@ -267,7 +267,7 @@ parameter automatically by cross-validation::
     >>> diabetes = datasets.load_diabetes()
     >>> X_diabetes = diabetes.data
     >>> y_diabetes = diabetes.target
-    >>> lasso.fit(X_diabetes, y_diabetes)
+    >>> lasso.fit(X_diabetes, y_diabetes)  # doctest: +NORMALIZE_WHITESPACE
     LassoCV(alphas=None, copy_X=True, cv=3, eps=0.001, fit_intercept=True,
         max_iter=1000, n_alphas=100, n_jobs=None, normalize=False,
         positive=False, precompute='auto', random_state=None,

diff --git a/doc/tutorial/statistical_inference/supervised_learning.rst b/doc/tutorial/statistical_inference/supervised_learning.rst
@@ -334,6 +334,7 @@ application of Occam's razor: *prefer simpler models*.
     >>> best_alpha = alphas[scores.index(max(scores))]
     >>> regr.alpha = best_alpha
     >>> regr.fit(diabetes_X_train, diabetes_y_train)
+    ... # doctest: +NORMALIZE_WHITESPACE
     Lasso(alpha=0.025118864315095794, copy_X=True, fit_intercept=True,
        max_iter=1000, normalize=False, positive=False, precompute=False,
        random_state=None, selection='cyclic', tol=0.0001, warm_start=False)

diff --git a/doc/tutorial/statistical_inference/unsupervised_learning.rst b/doc/tutorial/statistical_inference/unsupervised_learning.rst
@@ -274,7 +274,7 @@ data by projecting on a principal subspace.
 
     >>> from sklearn import decomposition
     >>> pca = decomposition.PCA()
-    >>> pca.fit(X)
+    >>> pca.fit(X)  # doctest: +NORMALIZE_WHITESPACE
     PCA(copy=True, iterated_power='auto', n_components=None, random_state=None,
       svd_solver='auto', tol=0.0, whiten=False)
     >>> print(pca.explained_variance_)  # doctest: +SKIP

diff --git a/doc/whats_new/v0.21.rst b/doc/whats_new/v0.21.rst
@@ -185,12 +185,19 @@ Support for Python 3.4 and below has been officially dropped.
   ``max_depth`` by 1 while expanding the tree if ``max_leaf_nodes`` and
   ``max_depth`` were both specified by the user. Please note that this also
   affects all ensemble methods using decision trees.
-  :pr:`12344` by :user:`Adrin Jalali <adrinjalali>`.
+  :issue:`12344` by :user:`Adrin Jalali <adrinjalali>`.
 
 
 Multiple modules
 ................
 
+- The `__repr__()` method of all estimators (used when calling
+  `print(estimator)`) has been entirely re-written, building on Python's
+  pretty printing standard library. All parameters are printed by default,
+  but this can be altered with the ``print_changed_only`` option in
+  :func:`sklearn.set_config`. :issue:`11705` by :user:`Nicolas Hug
+  <NicolasHug>`.
+
 Changes to estimator checks
 ---------------------------
 

diff --git a/examples/plot_changed_only_pprint_parameter.py b/examples/plot_changed_only_pprint_parameter.py
@@ -0,0 +1,30 @@
+"""
+=================================
+Compact estimator representations
+=================================
+
+This example illustrates the use of the print_changed_only global parameter.
+
+Setting print_changed_only to True will alterate the representation of
+estimators to only show the parameters that have been set to non-default
+values. This can be used to have more compact representations.
+"""
+print(__doc__)
+
+from sklearn.linear_model import LogisticRegression
+from sklearn import set_config
+
+
+lr = LogisticRegression(penalty='l1')
+print('Default representation:')
+print(lr)
+# LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
+#                    intercept_scaling=1, l1_ratio=None, max_iter=100,
+#                    multi_class='warn', n_jobs=None, penalty='l1',
+#                    random_state=None, solver='warn', tol=0.0001, verbose=0,
+#                    warm_start=False)
+
+set_config(print_changed_only=True)
+print('\nWith changed_only option:')
+print(lr)
+# LogisticRegression(penalty='l1')
diff --git a/sklearn/_config.py b/sklearn/_config.py
@@ -5,7 +5,8 @@
 
 _global_config = {
     'assume_finite': bool(os.environ.get('SKLEARN_ASSUME_FINITE', False)),
-    'working_memory': int(os.environ.get('SKLEARN_WORKING_MEMORY', 1024))
+    'working_memory': int(os.environ.get('SKLEARN_WORKING_MEMORY', 1024)),
+    'print_changed_only': False,
 }
 
 
@@ -20,7 +21,8 @@ def get_config():
     return _global_config.copy()
 
 
-def set_config(assume_finite=None, working_memory=None):
+def set_config(assume_finite=None, working_memory=None,
+               print_changed_only=None):
     """Set global scikit-learn configuration
 
     .. versionadded:: 0.19
@@ -43,11 +45,21 @@ def set_config(assume_finite=None, working_memory=None):
 
         .. versionadded:: 0.20
 
+    print_changed_only : bool, optional
+        If True, only the parameters that were set to non-default
+        values will be printed when printing an estimator. For example,
+        ``print(SVC())`` while True will only print 'SVC()' while the default
+        behaviour would be to print 'SVC(C=1.0, cache_size=200, ...)' with
+        all the non-changed parameters.
+
+        .. versionadded:: 0.21
     """
     if assume_finite is not None:
         _global_config['assume_finite'] = assume_finite
     if working_memory is not None:
         _global_config['working_memory'] = working_memory
+    if print_changed_only is not None:
+        _global_config['print_changed_only'] = print_changed_only
 
 
 @contextmanager

diff --git a/sklearn/base.py b/sklearn/base.py
@@ -224,9 +224,23 @@ def set_params(self, **params):
         return self
 
     def __repr__(self):
-        class_name = self.__class__.__name__
-        return '%s(%s)' % (class_name, _pprint(self.get_params(deep=False),
-                                               offset=len(class_name),),)
+        from .utils._pprint import _EstimatorPrettyPrinter
+
+        N_CHAR_MAX = 700  # number of non-whitespace or newline chars
+        N_MAX_ELEMENTS_TO_SHOW = 30  # number of elements to show in sequences
+
+        # use ellipsis for sequences with a lot of elements
+        pp = _EstimatorPrettyPrinter(
+            compact=True, indent=1, indent_at_name=True,
+            n_max_elements_to_show=N_MAX_ELEMENTS_TO_SHOW)
+
+        repr_ = pp.pformat(self)
+
+        # Use bruteforce ellipsis if string is very long
+        if len(''.join(repr_.split())) > N_CHAR_MAX:  # check non-blank chars
+            lim = N_CHAR_MAX // 2
+            repr_ = repr_[:lim] + '...' + repr_[-lim:]
+        return repr_
 
     def __getstate__(self):
         try:

diff --git a/sklearn/cluster/hierarchical.py b/sklearn/cluster/hierarchical.py
@@ -919,7 +919,7 @@ class FeatureAgglomeration(AgglomerativeClustering, AgglomerationTransform):
     >>> images = digits.images
     >>> X = np.reshape(images, (len(images), -1))
     >>> agglo = cluster.FeatureAgglomeration(n_clusters=32)
-    >>> agglo.fit(X) # doctest: +ELLIPSIS
+    >>> agglo.fit(X) # doctest: +ELLIPSIS +NORMALIZE_WHITESPACE
     FeatureAgglomeration(affinity='euclidean', compute_full_tree='auto',
                connectivity=None, linkage='ward', memory=None, n_clusters=32,
                pooling_func=...)

diff --git a/sklearn/decomposition/pca.py b/sklearn/decomposition/pca.py
@@ -276,7 +276,7 @@ class PCA(_BasePCA):
     >>> from sklearn.decomposition import PCA
     >>> X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [3, 2]])
     >>> pca = PCA(n_components=2)
-    >>> pca.fit(X)
+    >>> pca.fit(X)  # doctest: +NORMALIZE_WHITESPACE
     PCA(copy=True, iterated_power='auto', n_components=2, random_state=None,
       svd_solver='auto', tol=0.0, whiten=False)
     >>> print(pca.explained_variance_ratio_)  # doctest: +ELLIPSIS
@@ -294,7 +294,7 @@ class PCA(_BasePCA):
     [6.30061... 0.54980...]
 
     >>> pca = PCA(n_components=1, svd_solver='arpack')
-    >>> pca.fit(X)
+    >>> pca.fit(X)  # doctest: +NORMALIZE_WHITESPACE
     PCA(copy=True, iterated_power='auto', n_components=1, random_state=None,
       svd_solver='arpack', tol=0.0, whiten=False)
     >>> print(pca.explained_variance_ratio_)  # doctest: +ELLIPSIS

diff --git a/sklearn/discriminant_analysis.py b/sklearn/discriminant_analysis.py
@@ -242,7 +242,7 @@ class LinearDiscriminantAnalysis(BaseEstimator, LinearClassifierMixin,
     >>> X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [3, 2]])
     >>> y = np.array([1, 1, 1, 2, 2, 2])
     >>> clf = LinearDiscriminantAnalysis()
-    >>> clf.fit(X, y)
+    >>> clf.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     LinearDiscriminantAnalysis(n_components=None, priors=None, shrinkage=None,
                   solver='svd', store_covariance=False, tol=0.0001)
     >>> print(clf.predict([[-0.8, -1]]))

diff --git a/sklearn/ensemble/forest.py b/sklearn/ensemble/forest.py
@@ -956,7 +956,7 @@ class labels (multi-output problem).
     ...                            random_state=0, shuffle=False)
     >>> clf = RandomForestClassifier(n_estimators=100, max_depth=2,
     ...                              random_state=0)
-    >>> clf.fit(X, y)
+    >>> clf.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
                 max_depth=2, max_features='auto', max_leaf_nodes=None,
                 min_impurity_decrease=0.0, min_impurity_split=None,
@@ -1208,7 +1208,7 @@ class RandomForestRegressor(ForestRegressor):
     ...                        random_state=0, shuffle=False)
     >>> regr = RandomForestRegressor(max_depth=2, random_state=0,
     ...                              n_estimators=100)
-    >>> regr.fit(X, y)
+    >>> regr.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=2,
                max_features='auto', max_leaf_nodes=None,
                min_impurity_decrease=0.0, min_impurity_split=None,
@@ -1235,7 +1235,7 @@ class RandomForestRegressor(ForestRegressor):
     search of the best split. To obtain a deterministic behaviour during
     fitting, ``random_state`` has to be fixed.
 
-    The default value ``max_features="auto"`` uses ``n_features`` 
+    The default value ``max_features="auto"`` uses ``n_features``
     rather than ``n_features / 3``. The latter was originally suggested in
     [1], whereas the former was more recently justified empirically in [2].
 
@@ -1244,7 +1244,7 @@ class RandomForestRegressor(ForestRegressor):
 
     .. [1] L. Breiman, "Random Forests", Machine Learning, 45(1), 5-32, 2001.
 
-    .. [2] P. Geurts, D. Ernst., and L. Wehenkel, "Extremely randomized 
+    .. [2] P. Geurts, D. Ernst., and L. Wehenkel, "Extremely randomized
            trees", Machine Learning, 63(1), 3-42, 2006.
 
     See also
@@ -1496,7 +1496,7 @@ class labels (multi-output problem).
     References
     ----------
 
-    .. [1] P. Geurts, D. Ernst., and L. Wehenkel, "Extremely randomized 
+    .. [1] P. Geurts, D. Ernst., and L. Wehenkel, "Extremely randomized
            trees", Machine Learning, 63(1), 3-42, 2006.
 
     See also

diff --git a/sklearn/feature_extraction/dict_vectorizer.py b/sklearn/feature_extraction/dict_vectorizer.py
@@ -345,6 +345,7 @@ def restrict(self, support, indices=False):
         >>> v.get_feature_names()
         ['bar', 'baz', 'foo']
         >>> v.restrict(support.get_support()) # doctest: +ELLIPSIS
+        ...                                   # doctest: +NORMALIZE_WHITESPACE
         DictVectorizer(dtype=..., separator='=', sort=True,
                 sparse=True)
         >>> v.get_feature_names()

diff --git a/sklearn/impute.py b/sklearn/impute.py
@@ -464,7 +464,7 @@ class MissingIndicator(BaseEstimator, TransformerMixin):
     ...                [np.nan, 2, 3],
     ...                [2, 4, 0]])
     >>> indicator = MissingIndicator()
-    >>> indicator.fit(X1)
+    >>> indicator.fit(X1)  # doctest: +NORMALIZE_WHITESPACE
     MissingIndicator(error_on_new=True, features='missing-only',
              missing_values=nan, sparse='auto')
     >>> X2_tr = indicator.transform(X2)

diff --git a/sklearn/kernel_approximation.py b/sklearn/kernel_approximation.py
@@ -163,7 +163,7 @@ class SkewedChi2Sampler(BaseEstimator, TransformerMixin):
     ...                                  random_state=0)
     >>> X_features = chi2_feature.fit_transform(X, y)
     >>> clf = SGDClassifier(max_iter=10, tol=1e-3)
-    >>> clf.fit(X_features, y)
+    >>> clf.fit(X_features, y)  # doctest: +NORMALIZE_WHITESPACE
     SGDClassifier(alpha=0.0001, average=False, class_weight=None,
            early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,
            l1_ratio=0.15, learning_rate='optimal', loss='hinge', max_iter=10,
@@ -283,7 +283,7 @@ class AdditiveChi2Sampler(BaseEstimator, TransformerMixin):
     >>> chi2sampler = AdditiveChi2Sampler(sample_steps=2)
     >>> X_transformed = chi2sampler.fit_transform(X, y)
     >>> clf = SGDClassifier(max_iter=5, random_state=0, tol=1e-3)
-    >>> clf.fit(X_transformed, y)
+    >>> clf.fit(X_transformed, y)  # doctest: +NORMALIZE_WHITESPACE
     SGDClassifier(alpha=0.0001, average=False, class_weight=None,
            early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,
            l1_ratio=0.15, learning_rate='optimal', loss='hinge', max_iter=5,

diff --git a/sklearn/linear_model/coordinate_descent.py b/sklearn/linear_model/coordinate_descent.py
@@ -623,7 +623,7 @@ class ElasticNet(LinearModel, RegressorMixin):
 
     >>> X, y = make_regression(n_features=2, random_state=0)
     >>> regr = ElasticNet(random_state=0)
-    >>> regr.fit(X, y)
+    >>> regr.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     ElasticNet(alpha=1.0, copy_X=True, fit_intercept=True, l1_ratio=0.5,
           max_iter=1000, normalize=False, positive=False, precompute=False,
           random_state=0, selection='cyclic', tol=0.0001, warm_start=False)
@@ -903,6 +903,7 @@ class Lasso(ElasticNet):
     >>> from sklearn import linear_model
     >>> clf = linear_model.Lasso(alpha=0.1)
     >>> clf.fit([[0,0], [1, 1], [2, 2]], [0, 1, 2])
+    ... # doctest: +NORMALIZE_WHITESPACE
     Lasso(alpha=0.1, copy_X=True, fit_intercept=True, max_iter=1000,
        normalize=False, positive=False, precompute=False, random_state=None,
        selection='cyclic', tol=0.0001, warm_start=False)
@@ -1552,7 +1553,7 @@ class ElasticNetCV(LinearModelCV, RegressorMixin):
 
     >>> X, y = make_regression(n_features=2, random_state=0)
     >>> regr = ElasticNetCV(cv=5, random_state=0)
-    >>> regr.fit(X, y)
+    >>> regr.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     ElasticNetCV(alphas=None, copy_X=True, cv=5, eps=0.001, fit_intercept=True,
            l1_ratio=0.5, max_iter=1000, n_alphas=100, n_jobs=None,
            normalize=False, positive=False, precompute='auto', random_state=0,
@@ -1907,6 +1908,7 @@ class MultiTaskLasso(MultiTaskElasticNet):
     >>> from sklearn import linear_model
     >>> clf = linear_model.MultiTaskLasso(alpha=0.1)
     >>> clf.fit([[0,0], [1, 1], [2, 2]], [[0, 0], [1, 1], [2, 2]])
+    ... # doctest: +NORMALIZE_WHITESPACE
     MultiTaskLasso(alpha=0.1, copy_X=True, fit_intercept=True, max_iter=1000,
             normalize=False, random_state=None, selection='cyclic', tol=0.0001,
             warm_start=False)

diff --git a/sklearn/linear_model/passive_aggressive.py b/sklearn/linear_model/passive_aggressive.py
@@ -141,7 +141,7 @@ class PassiveAggressiveClassifier(BaseSGDClassifier):
     >>> X, y = make_classification(n_features=4, random_state=0)
     >>> clf = PassiveAggressiveClassifier(max_iter=1000, random_state=0,
     ... tol=1e-3)
-    >>> clf.fit(X, y)
+    >>> clf.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     PassiveAggressiveClassifier(C=1.0, average=False, class_weight=None,
                   early_stopping=False, fit_intercept=True, loss='hinge',
                   max_iter=1000, n_iter=None, n_iter_no_change=5, n_jobs=None,
@@ -380,7 +380,7 @@ class PassiveAggressiveRegressor(BaseSGDRegressor):
     >>> X, y = make_regression(n_features=4, random_state=0)
     >>> regr = PassiveAggressiveRegressor(max_iter=100, random_state=0,
     ... tol=1e-3)
-    >>> regr.fit(X, y)
+    >>> regr.fit(X, y)  # doctest: +NORMALIZE_WHITESPACE
     PassiveAggressiveRegressor(C=1.0, average=False, early_stopping=False,
                   epsilon=0.1, fit_intercept=True, loss='epsilon_insensitive',
                   max_iter=100, n_iter=None, n_iter_no_change=5,