[MRG] Update joblib to 0.10.2 (scikit-learn#7290)

* Update joblib to 0.10.2

* Exclude sklearn/externals from diff for flake8

Author: ogrisel

build_tools/travis/flake8_diff.sh

@@ -14,6 +14,8 @@
 #     turn-around
 
 set -e
+# pipefail is necessary to propagate exit codes
+set -o pipefail
 
 PROJECT=scikit-learn/scikit-learn
 PROJECT_URL=https://github.com/$PROJECT.git
@@ -79,7 +81,20 @@ echo -e '\nRunning flake8 on the diff in the range'\
      "($(git rev-list $COMMIT.. | wc -l) commit(s)):"
 echo '--------------------------------------------------------------------------------'
 
-# Conservative approach: diff without context so that code that was
-# not changed does not create failures
-git diff --unified=0 $COMMIT | flake8 --diff --show-source
+# We ignore files from sklearn/externals. Unfortunately there is no
+# way to do it with flake8 directly (the --exclude does not seem to
+# work with --diff). We could use the exclude magic in the git pathspec
+# ':!sklearn/externals' but it is only available on git 1.9 and Travis
+# uses git 1.8.
+# We need the following command to exit with 0 hence the echo in case
+# there is no match
+MODIFIED_FILES=$(git diff --name-only $COMMIT | grep -v 'sklearn/externals' || echo "no_match")
+
+if [[ "$MODIFIED_FILES" == "no_match" ]]; then
+    echo "No file outside sklearn/externals has been modified"
+else
+    # Conservative approach: diff without context so that code that
+    # was not changed does not create failures
+    git diff --unified=0 $COMMIT -- $MODIFIED_FILES | flake8 --diff --show-source
+fi
 echo -e "No problem detected by flake8\n"

sklearn/externals/joblib/__init__.py

@@ -115,7 +115,7 @@
 # Dev branch marker is: 'X.Y.dev' or 'X.Y.devN' where N is an integer.
 # 'X.Y.dev0' is the canonical version of 'X.Y.dev'
 #
-__version__ = '0.10.0'
+__version__ = '0.10.2'
 
 
 from .memory import Memory, MemorizedResult

sklearn/externals/joblib/_memory_helpers.py

@@ -102,4 +102,4 @@ def open_py_source(filename):
         buffer.seek(0)
         text = TextIOWrapper(buffer, encoding, line_buffering=True)
         text.mode = 'r'
-        return text
+        return text

sklearn/externals/joblib/_parallel_backends.py

@@ -25,9 +25,9 @@ class ParallelBackendBase(with_metaclass(ABCMeta)):
     def effective_n_jobs(self, n_jobs):
         """Determine the number of jobs that can actually run in parallel
 
-        n_jobs is the is the number of workers requested by the callers.
-        Passing n_jobs=-1 means requesting all available workers for instance
-        matching the number of CPU cores on the worker host(s).
+        n_jobs is the number of workers requested by the callers. Passing
+        n_jobs=-1 means requesting all available workers for instance matching
+        the number of CPU cores on the worker host(s).
 
         This method should return a guesstimate of the number of workers that
         can actually perform work concurrently. The primary use case is to make

sklearn/externals/joblib/format_stack.py

@@ -273,8 +273,10 @@ def linereader(file=file, lnum=[lnum], getline=linecache.getline):
             # enclosing scope.
             for token in generate_tokens(linereader):
                 tokeneater(*token)
-        except (IndexError, UnicodeDecodeError):
+        except (IndexError, UnicodeDecodeError, SyntaxError):
             # signals exit of tokenizer
+            # SyntaxError can happen when trying to tokenize
+            # a compiled (e.g. .so or .pyd) extension
             pass
         except tokenize.TokenError as msg:
             _m = ("An unexpected error occurred while tokenizing input file %s\n"

sklearn/externals/joblib/func_inspect.py

@@ -50,8 +50,7 @@ def get_func_code(func):
             line_no = 1
             if source_file.startswith('<doctest '):
                 source_file, line_no = re.match(
-                            '\<doctest (.*\.rst)\[(.*)\]\>',
-                            source_file).groups()
+                    '\<doctest (.*\.rst)\[(.*)\]\>', source_file).groups()
                 line_no = int(line_no)
                 source_file = '<doctest %s>' % source_file
             return source_code, source_file, line_no
@@ -312,7 +311,7 @@ def filter_args(func, ignore_lst, args=(), kwargs=dict()):
                              "function %s"
                              % (item,
                                 _signature_str(name, arg_spec))
-            )
+                             )
     # XXX: Return a sorted list of pairs?
     return arg_dict
 
@@ -352,4 +351,4 @@ def format_call(func, args, kwargs, object_name="Memory"):
                                           path, signature)
     return msg
     # XXX: Not using logging framework
-    #self.debug(msg)
+    # self.debug(msg)

sklearn/externals/joblib/numpy_pickle_utils.py

@@ -14,7 +14,7 @@
 import contextlib
 from contextlib import closing
 
-from ._compat import PY3_OR_LATER, PY26, PY27
+from ._compat import PY3_OR_LATER, PY26, PY27, _basestring
 
 try:
     from threading import RLock
@@ -299,7 +299,7 @@ def __init__(self, filename, mode="rb", compresslevel=9):
         else:
             raise ValueError("Invalid mode: %r" % (mode,))
 
-        if isinstance(filename, (str, bytes)):
+        if isinstance(filename, _basestring):
             self._fp = open(filename, mode)
             self._closefp = True
             self._mode = mode_code

sklearn/externals/joblib/parallel.py

@@ -32,6 +32,10 @@
 from ._compat import _basestring
 from .func_inspect import getfullargspec
 
+# Make sure that those two classes are part of the public joblib.parallel API
+# so that 3rd party backend implementers can import them from here.
+from ._parallel_backends import AutoBatchingMixin  # noqa
+from ._parallel_backends import ParallelBackendBase  # noqa
 
 BACKENDS = {
     'multiprocessing': MultiprocessingBackend,