__init__.py

import os
import warnings

from chainerx import _build_info


if _build_info.build_chainerx:
    from chainerx import _core
    _available = True
else:
    _available = False


if _available:
    from numpy import dtype  # NOQA
    from numpy import (  # NOQA
        Inf, Infinity, NAN, NINF, NZERO, NaN, PINF, PZERO,
        e, euler_gamma,
        inf, infty, nan,
        newaxis,
        pi)
    from numpy import (
        bool_, int8, int16, int32, int64, uint8, float16, float32, float64)  # NOQA
    all_dtypes = (
        bool_, int8, int16, int32, int64, uint8, float16, float32, float64)

    from chainerx._core import *  # NOQA
    from chainerx._core import _to_cupy  # NOQA

    from builtins import bool, int, float  # NOQA

    from chainerx import _device  # NOQA

    from chainerx.creation.from_data import asanyarray  # NOQA
    from chainerx.creation.from_data import fromfile  # NOQA
    from chainerx.creation.from_data import fromfunction  # NOQA
    from chainerx.creation.from_data import fromiter  # NOQA
    from chainerx.creation.from_data import fromstring  # NOQA
    from chainerx.creation.from_data import loadtxt  # NOQA

    from chainerx.manipulation.shape import ravel  # NOQA

    from chainerx.math.misc import clip  # NOQA

    from chainerx import random  # NOQA

    _global_context = _core.Context()
    _core.set_global_default_context(_global_context)

    # Implements ndarray methods in Python
    from chainerx import _ndarray
    _ndarray.populate()

    # Temporary workaround implementations that fall back to NumPy/CuPy's
    # respective functions.
    from chainerx import _fallback_workarounds
    _fallback_workarounds.populate()

    # Dynamically inject docstrings
    from chainerx import _docs
    _docs.set_docs()

    from chainerx import _cuda
    # Share memory pool with CuPy.
    if bool(int(os.getenv('CHAINERX_CUDA_CUPY_SHARE_ALLOCATOR', '0'))):
        _cuda.cupy_share_allocator()
else:
    class ndarray(object):

        """Dummy class for type testing."""

        def __init__(self, *args, **kwargs):
            raise RuntimeError('chainerx is not available.')


def is_available():
    return _available


if _available and _core._is_debug():
    # Warn if the ChainerX core binary is built in debug mode
    warnings.warn(
        'ChainerX core binary is built in debug mode.', stacklevel=2)