Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [dev] (MM/DD/YYYY)

### Added
* Added `mkl_random` patching for NumPy, with `mkl_random` context manager, `is_patched` query, and `patch_numpy_random` and `restore_numpy_random` calls to replace `numpy.random` calls with calls from `mkl_random.interfaces.numpy_random` [gh-90](https://github.com/IntelPython/mkl_random/pull/90)

* Added `mkl_random.interfaces` with `mkl_random.interfaces.numpy_random` interface, which aliases `mkl_random` functionality to more strictly adhere to NumPy's API (i.e., drops arguments and functions which are not part of standard NumPy) [gh-92](https://github.com/IntelPython/mkl_random/pull/92)

### Removed
Expand Down
12 changes: 12 additions & 0 deletions mkl_random/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -95,6 +95,14 @@

from mkl_random import interfaces

from ._patch_numpy import (
is_patched,
mkl_random,
patch_numpy_random,
patched_names,
restore_numpy_random,
)

__all__ = [
"MKLRandomState",
"RandomState",
Expand Down Expand Up @@ -147,6 +155,10 @@
"shuffle",
"permutation",
"interfaces",
"patch_numpy_random",
"restore_numpy_random",
"is_patched",
"patched_names",
]

del _init_helper
280 changes: 280 additions & 0 deletions mkl_random/_patch_numpy.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,280 @@
# Copyright (c) 2019, Intel Corporation
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
# * Redistributions of source code must retain the above copyright notice,
# this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of Intel Corporation nor the names of its contributors
# may be used to endorse or promote products derived from this software
# without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

"""Define functions for patching NumPy with MKL-based NumPy interface."""

import warnings
from contextlib import ContextDecorator
from threading import Lock, local

import numpy as _np

import mkl_random.interfaces.numpy_random as _nrand

_DEFAULT_NAMES = tuple(_nrand.__all__)


class _GlobalPatch:
def __init__(self):
self._lock = Lock()
self._patch_count = 0
self._restore_dict = {}
self._patched_functions = tuple(_DEFAULT_NAMES)
self._numpy_module = None
self._requested_names = None
self._active_names = ()
self._patched = ()
self._tls = local()

def _normalize_names(self, names):
if names is None:
names = _DEFAULT_NAMES
return tuple(names)

def _validate_module(self, numpy_module):
if not hasattr(numpy_module, "random"):
raise TypeError(
"Expected a numpy-like module with a `.random` attribute."
)

def _register_func(self, name, func):
if name not in self._patched_functions:
raise ValueError(f"{name} not an mkl_random function.")
np_random = self._numpy_module.random
if name not in self._restore_dict:
self._restore_dict[name] = getattr(np_random, name)
setattr(np_random, name, func)

def _restore_func(self, name, verbose=False):
if name not in self._patched_functions:
raise ValueError(f"{name} not an mkl_random function.")
try:
val = self._restore_dict[name]
except KeyError:
if verbose:
print(f"failed to restore {name}")
return
else:
if verbose:
print(f"found and restoring {name}...")
np_random = self._numpy_module.random
setattr(np_random, name, val)

def _initialize_patch(self, numpy_module, names, strict):
self._validate_module(numpy_module)
np_random = numpy_module.random
missing = []
patchable = []
for name in names:
if name not in self._patched_functions:
missing.append(name)
continue
if not hasattr(np_random, name) or not hasattr(_nrand, name):
missing.append(name)
continue
patchable.append(name)

if strict and missing:
raise AttributeError(
"Could not patch these names (missing on numpy.random or "
"mkl_random.interfaces.numpy_random): "
+ ", ".join(str(x) for x in missing)
)

self._numpy_module = numpy_module
self._requested_names = names
self._active_names = tuple(patchable)
self._patched = tuple(patchable)

def do_patch(
self,
numpy_module=None,
names=None,
strict=False,
verbose=False,
):
if numpy_module is None:
numpy_module = _np
names = self._normalize_names(names)
strict = bool(strict)

with self._lock:
local_count = getattr(self._tls, "local_count", 0)
if self._patch_count == 0:
self._initialize_patch(numpy_module, names, strict)
if verbose:
print(
"Now patching NumPy random submodule with mkl_random "
"NumPy interface."
)
print(
"Please direct bug reports to "
"https://github.com/IntelPython/mkl_random"
)
for name in self._active_names:
self._register_func(name, getattr(_nrand, name))
else:
if self._numpy_module is not numpy_module:
raise RuntimeError(
"Already patched a different numpy module; "
"call restore() first."
)
if names != self._requested_names:
raise RuntimeError(
"Already patched with a different names set; "
"call restore() first."
)
self._patch_count += 1
self._tls.local_count = local_count + 1

def do_restore(self, verbose=False):
with self._lock:
local_count = getattr(self._tls, "local_count", 0)
if local_count <= 0:
if verbose:
warnings.warn(
"Warning: restore_numpy_random called more times than "
"patch_numpy_random in this thread.",
stacklevel=2,
)
return

self._tls.local_count = local_count - 1
self._patch_count -= 1
if self._patch_count == 0:
if verbose:
print("Now restoring original NumPy random submodule.")
for name in tuple(self._restore_dict):
self._restore_func(name, verbose=verbose)
self._restore_dict.clear()
self._numpy_module = None
self._requested_names = None
self._active_names = ()
self._patched = ()

def is_patched(self):
with self._lock:
return self._patch_count > 0

def patched_names(self):
with self._lock:
return list(self._patched)


_patch = _GlobalPatch()


def patch_numpy_random(
numpy_module=None,
names=None,
strict=False,
verbose=False,
):
"""
Patch NumPy's random submodule with mkl_random's NumPy interface.

Parameters
----------
numpy_module : module, optional
NumPy-like module to patch. Defaults to imported NumPy.
names : iterable[str], optional
Attributes under `numpy_module.random` to patch.
strict : bool, optional
Raise if any requested symbol cannot be patched.
verbose : bool, optional
Print messages when starting the patching process.

Examples
--------
>>> import numpy as np
>>> import mkl_random
>>> mkl_random.is_patched()
False
>>> mkl_random.patch_numpy_random(np)
>>> mkl_random.is_patched()
True
>>> mkl_random.restore()
>>> mkl_random.is_patched()
False
"""
_patch.do_patch(
numpy_module=numpy_module,
names=names,
strict=bool(strict),
verbose=bool(verbose),
)


def restore_numpy_random(verbose=False):
"""
Restore NumPy's random submodule to its original implementations.

Parameters
----------
verbose : bool, optional
Print message when starting restoration process.
"""
_patch.do_restore(verbose=bool(verbose))


def is_patched():
"""Return whether NumPy has been patched with mkl_random."""
return _patch.is_patched()


def patched_names():
"""Return names actually patched in `numpy.random`."""
return _patch.patched_names()


class mkl_random(ContextDecorator):
"""
Context manager and decorator to temporarily patch NumPy random submodule
with MKL-based implementations.

Examples
--------
>>> import numpy as np
>>> import mkl_random
>>> with mkl_random.mkl_random(np):
... x = np.random.normal(size=10)
"""

def __init__(self, numpy_module=None, names=None, strict=False):
self._numpy_module = numpy_module
self._names = names
self._strict = strict

def __enter__(self):
patch_numpy_random(
numpy_module=self._numpy_module,
names=self._names,
strict=self._strict,
)
return self

def __exit__(self, *exc):
restore_numpy_random()
return False
Loading
Loading