MAINT Remove imports from sklearn.utils._joblib (scikit-learn#13676)

* Remove sklearn.utils._joblib imports

* Lint

* More fixes

sklearn/cluster/k_means_.py

@@ -15,6 +15,7 @@
 
 import numpy as np
 import scipy.sparse as sp
+from joblib import Parallel, delayed, effective_n_jobs
 
 from ..base import BaseEstimator, ClusterMixin, TransformerMixin
 from ..metrics.pairwise import euclidean_distances
@@ -28,9 +29,6 @@
 from ..utils import check_random_state
 from ..utils.validation import check_is_fitted
 from ..utils.validation import FLOAT_DTYPES
-from ..utils._joblib import Parallel
-from ..utils._joblib import delayed
-from ..utils._joblib import effective_n_jobs
 from ..exceptions import ConvergenceWarning
 from . import _k_means
 from ._k_means_elkan import k_means_elkan

sklearn/cluster/mean_shift_.py

@@ -16,15 +16,14 @@
 
 import numpy as np
 import warnings
+from joblib import Parallel, delayed
 
 from collections import defaultdict
 from ..utils.validation import check_is_fitted
 from ..utils import check_random_state, gen_batches, check_array
 from ..base import BaseEstimator, ClusterMixin
 from ..neighbors import NearestNeighbors
 from ..metrics.pairwise import pairwise_distances_argmin
-from ..utils._joblib import Parallel
-from ..utils._joblib import delayed
 
 
 def estimate_bandwidth(X, quantile=0.3, n_samples=None, random_state=0,

sklearn/compose/_column_transformer.py

@@ -11,9 +11,9 @@
 
 import numpy as np
 from scipy import sparse
+from joblib import Parallel, delayed
 
 from ..base import clone, TransformerMixin
-from ..utils._joblib import Parallel, delayed
 from ..pipeline import _fit_transform_one, _transform_one, _name_estimators
 from ..preprocessing import FunctionTransformer
 from ..utils import Bunch

sklearn/covariance/graph_lasso_.py

@@ -13,6 +13,7 @@
 
 import numpy as np
 from scipy import linalg
+from joblib import Parallel, delayed
 
 from .empirical_covariance_ import (empirical_covariance, EmpiricalCovariance,
                                     log_likelihood)
@@ -22,7 +23,6 @@
 from ..linear_model import cd_fast
 from ..linear_model import lars_path_gram
 from ..model_selection import check_cv, cross_val_score
-from ..utils._joblib import Parallel, delayed
 
 
 # Helper functions to compute the objective and dual objective functions

sklearn/datasets/california_housing.py

@@ -28,12 +28,13 @@
 import numpy as np
 import logging
 
+import joblib
+
 from .base import get_data_home
 from .base import _fetch_remote
 from .base import _pkl_filepath
 from .base import RemoteFileMetadata
 from ..utils import Bunch
-from ..utils import _joblib
 
 # The original data can be found at:
 # https://www.dcc.fc.up.pt/~ltorgo/Regression/cal_housing.tgz
@@ -124,11 +125,11 @@ def fetch_california_housing(data_home=None, download_if_missing=True,
             columns_index = [8, 7, 2, 3, 4, 5, 6, 1, 0]
             cal_housing = cal_housing[:, columns_index]
 
-            _joblib.dump(cal_housing, filepath, compress=6)
+            joblib.dump(cal_housing, filepath, compress=6)
         remove(archive_path)
 
     else:
-        cal_housing = _joblib.load(filepath)
+        cal_housing = joblib.load(filepath)
 
     feature_names = ["MedInc", "HouseAge", "AveRooms", "AveBedrms",
                      "Population", "AveOccup", "Latitude", "Longitude"]

sklearn/datasets/covtype.py

@@ -20,13 +20,13 @@
 from os import remove, makedirs
 
 import numpy as np
+import joblib
 
 from .base import get_data_home
 from .base import _fetch_remote
 from .base import RemoteFileMetadata
 from ..utils import Bunch
 from .base import _pkl_filepath
-from ..utils import _joblib
 from ..utils import check_random_state
 
 # The original data can be found in:
@@ -117,16 +117,16 @@ def fetch_covtype(data_home=None, download_if_missing=True,
         X = Xy[:, :-1]
         y = Xy[:, -1].astype(np.int32, copy=False)
 
-        _joblib.dump(X, samples_path, compress=9)
-        _joblib.dump(y, targets_path, compress=9)
+        joblib.dump(X, samples_path, compress=9)
+        joblib.dump(y, targets_path, compress=9)
 
     elif not available and not download_if_missing:
         raise IOError("Data not found and `download_if_missing` is False")
     try:
         X, y
     except NameError:
-        X = _joblib.load(samples_path)
-        y = _joblib.load(targets_path)
+        X = joblib.load(samples_path)
+        y = joblib.load(targets_path)
 
     if shuffle:
         ind = np.arange(X.shape[0])

sklearn/datasets/kddcup99.py

@@ -15,13 +15,12 @@
 from os.path import dirname, exists, join
 
 import numpy as np
-
+import joblib
 
 from .base import _fetch_remote
 from .base import get_data_home
 from .base import RemoteFileMetadata
 from ..utils import Bunch
-from ..utils import _joblib
 from ..utils import check_random_state
 from ..utils import shuffle as shuffle_method
 
@@ -284,17 +283,17 @@ def _fetch_brute_kddcup99(data_home=None,
         # (error: 'Incorrect data length while decompressing[...] the file
         #  could be corrupted.')
 
-        _joblib.dump(X, samples_path, compress=0)
-        _joblib.dump(y, targets_path, compress=0)
+        joblib.dump(X, samples_path, compress=0)
+        joblib.dump(y, targets_path, compress=0)
     elif not available:
         if not download_if_missing:
             raise IOError("Data not found and `download_if_missing` is False")
 
     try:
         X, y
     except NameError:
-        X = _joblib.load(samples_path)
-        y = _joblib.load(targets_path)
+        X = joblib.load(samples_path)
+        y = joblib.load(targets_path)
 
     return Bunch(data=X, target=y)
 

sklearn/datasets/lfw.py

@@ -15,11 +15,11 @@
 from distutils.version import LooseVersion
 
 import numpy as np
+import joblib
+from joblib import Memory
 
 from .base import get_data_home, _fetch_remote, RemoteFileMetadata
 from ..utils import Bunch
-from ..utils._joblib import Memory
-from ..utils import _joblib
 
 logger = logging.getLogger(__name__)
 
@@ -303,7 +303,7 @@ def fetch_lfw_people(data_home=None, funneled=True, resize=0.5,
 
     # wrap the loader in a memoizing function that will return memmaped data
     # arrays for optimal memory usage
-    if LooseVersion(_joblib.__version__) < LooseVersion('0.12'):
+    if LooseVersion(joblib.__version__) < LooseVersion('0.12'):
         # Deal with change of API in joblib
         m = Memory(cachedir=lfw_home, compress=6, verbose=0)
     else:
@@ -474,7 +474,7 @@ def fetch_lfw_pairs(subset='train', data_home=None, funneled=True, resize=0.5,
 
     # wrap the loader in a memoizing function that will return memmaped data
     # arrays for optimal memory usage
-    if LooseVersion(_joblib.__version__) < LooseVersion('0.12'):
+    if LooseVersion(joblib.__version__) < LooseVersion('0.12'):
         # Deal with change of API in joblib
         m = Memory(cachedir=lfw_home, compress=6, verbose=0)
     else:

sklearn/datasets/olivetti_faces.py

@@ -18,12 +18,12 @@
 
 import numpy as np
 from scipy.io.matlab import loadmat
+import joblib
 
 from .base import get_data_home
 from .base import _fetch_remote
 from .base import RemoteFileMetadata
 from .base import _pkl_filepath
-from ..utils import _joblib
 from ..utils import check_random_state, Bunch
 
 # The original data can be found at:
@@ -104,10 +104,10 @@ def fetch_olivetti_faces(data_home=None, shuffle=False, random_state=0,
         remove(mat_path)
 
         faces = mfile['faces'].T.copy()
-        _joblib.dump(faces, filepath, compress=6)
+        joblib.dump(faces, filepath, compress=6)
         del mfile
     else:
-        faces = _joblib.load(filepath)
+        faces = joblib.load(filepath)
 
     # We want floating point data, but float32 is enough (there is only
     # one byte of precision in the original uint8s anyway)

sklearn/datasets/rcv1.py

@@ -16,12 +16,12 @@
 
 import numpy as np
 import scipy.sparse as sp
+import joblib
 
 from .base import get_data_home
 from .base import _pkl_filepath
 from .base import _fetch_remote
 from .base import RemoteFileMetadata
-from ..utils import _joblib
 from .svmlight_format import load_svmlight_files
 from ..utils import shuffle as shuffle_
 from ..utils import Bunch
@@ -181,16 +181,16 @@ def fetch_rcv1(data_home=None, subset='all', download_if_missing=True,
         sample_id = np.hstack((Xy[9], Xy[1], Xy[3], Xy[5], Xy[7]))
         sample_id = sample_id.astype(np.uint32, copy=False)
 
-        _joblib.dump(X, samples_path, compress=9)
-        _joblib.dump(sample_id, sample_id_path, compress=9)
+        joblib.dump(X, samples_path, compress=9)
+        joblib.dump(sample_id, sample_id_path, compress=9)
 
         # delete archives
         for f in files:
             f.close()
             remove(f.name)
     else:
-        X = _joblib.load(samples_path)
-        sample_id = _joblib.load(sample_id_path)
+        X = joblib.load(samples_path)
+        sample_id = joblib.load(sample_id_path)
 
     # load target (y), categories, and sample_id_bis
     if download_if_missing and (not exists(sample_topics_path) or
@@ -240,11 +240,11 @@ def fetch_rcv1(data_home=None, subset='all', download_if_missing=True,
         categories = categories[order]
         y = sp.csr_matrix(y[:, order])
 
-        _joblib.dump(y, sample_topics_path, compress=9)
-        _joblib.dump(categories, topics_path, compress=9)
+        joblib.dump(y, sample_topics_path, compress=9)
+        joblib.dump(categories, topics_path, compress=9)
     else:
-        y = _joblib.load(sample_topics_path)
-        categories = _joblib.load(topics_path)
+        y = joblib.load(sample_topics_path)
+        categories = joblib.load(topics_path)
 
     if subset == 'all':
         pass

sklearn/datasets/species_distributions.py

@@ -41,16 +41,16 @@
 from os import makedirs, remove
 from os.path import exists
 
-
 import logging
 import numpy as np
 
+import joblib
+
 from .base import get_data_home
 from .base import _fetch_remote
 from .base import RemoteFileMetadata
 from ..utils import Bunch
 from .base import _pkl_filepath
-from ..utils import _joblib
 
 # The original data can be found at:
 # https://biodiversityinformatics.amnh.org/open_source/maxent/samples.zip
@@ -257,8 +257,8 @@ def fetch_species_distributions(data_home=None,
                       test=test,
                       train=train,
                       **extra_params)
-        _joblib.dump(bunch, archive_path, compress=9)
+        joblib.dump(bunch, archive_path, compress=9)
     else:
-        bunch = _joblib.load(archive_path)
+        bunch = joblib.load(archive_path)
 
     return bunch

sklearn/datasets/twenty_newsgroups.py

@@ -35,6 +35,7 @@
 
 import numpy as np
 import scipy.sparse as sp
+import joblib
 
 from .base import get_data_home
 from .base import load_files
@@ -43,7 +44,6 @@
 from .base import RemoteFileMetadata
 from ..feature_extraction.text import CountVectorizer
 from ..preprocessing import normalize
-from ..utils import _joblib
 from ..utils import check_random_state, Bunch
 
 logger = logging.getLogger(__name__)
@@ -398,12 +398,12 @@ def fetch_20newsgroups_vectorized(subset="train", remove=(), data_home=None,
                                    download_if_missing=download_if_missing)
 
     if os.path.exists(target_file):
-        X_train, X_test = _joblib.load(target_file)
+        X_train, X_test = joblib.load(target_file)
     else:
         vectorizer = CountVectorizer(dtype=np.int16)
         X_train = vectorizer.fit_transform(data_train.data).tocsr()
         X_test = vectorizer.transform(data_test.data).tocsr()
-        _joblib.dump((X_train, X_test), target_file, compress=9)
+        joblib.dump((X_train, X_test), target_file, compress=9)
 
     # the data is stored as int16 for compactness
     # but normalize needs floats

sklearn/decomposition/dict_learning.py

@@ -11,9 +11,9 @@
 
 import numpy as np
 from scipy import linalg
+from joblib import Parallel, delayed, effective_n_jobs
 
 from ..base import BaseEstimator, TransformerMixin
-from ..utils._joblib import Parallel, delayed, effective_n_jobs
 from ..utils import (check_array, check_random_state, gen_even_slices,
                      gen_batches)
 from ..utils.extmath import randomized_svd, row_norms

sklearn/decomposition/online_lda.py

@@ -14,14 +14,14 @@
 import numpy as np
 import scipy.sparse as sp
 from scipy.special import gammaln
+from joblib import Parallel, delayed, effective_n_jobs
 
 from ..base import BaseEstimator, TransformerMixin
 from ..utils import (check_random_state, check_array,
                      gen_batches, gen_even_slices)
 from ..utils.fixes import logsumexp
 from ..utils.validation import check_non_negative
 from ..utils.validation import check_is_fitted
-from ..utils._joblib import Parallel, delayed, effective_n_jobs
 
 from ._online_lda import (mean_change, _dirichlet_expectation_1d,
                           _dirichlet_expectation_2d)

sklearn/decomposition/tests/test_sparse_pca.py

@@ -142,15 +142,15 @@ def test_mini_batch_fit_transform():
     U1 = spca_lars.transform(Y)
     # Test multiple CPUs
     if sys.platform == 'win32':  # fake parallelism for win32
-        import sklearn.utils._joblib.parallel as joblib_par
-        _mp = joblib_par.multiprocessing
-        joblib_par.multiprocessing = None
+        import joblib
+        _mp = joblib.parallel.multiprocessing
+        joblib.parallel.multiprocessing = None
         try:
             spca = MiniBatchSparsePCA(n_components=3, n_jobs=2, alpha=alpha,
                                       random_state=0)
             U2 = spca.fit(Y).transform(Y)
         finally:
-            joblib_par.multiprocessing = _mp
+            joblib.parallel.multiprocessing = _mp
     else:  # we can efficiently use parallelism
         spca = MiniBatchSparsePCA(n_components=3, n_jobs=2, alpha=alpha,
                                   random_state=0)

sklearn/ensemble/bagging.py

@@ -10,9 +10,10 @@
 from abc import ABCMeta, abstractmethod
 from warnings import warn
 
+from joblib import Parallel, delayed
+
 from .base import BaseEnsemble, _partition_estimators
 from ..base import ClassifierMixin, RegressorMixin
-from ..utils._joblib import Parallel, delayed
 from ..metrics import r2_score, accuracy_score
 from ..tree import DecisionTreeClassifier, DecisionTreeRegressor
 from ..utils import check_random_state, check_X_y, check_array, column_or_1d