utils.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# Copyright (C) 2010 Radim Rehurek <radimrehurek@seznam.cz>
# Licensed under the GNU LGPL v2.1 - http://www.gnu.org/licenses/lgpl.html

"""Various general utility functions."""

from __future__ import with_statement
from contextlib import contextmanager
import collections.abc
import logging
import warnings
import numbers
from html.entities import name2codepoint as n2cp
import pickle as _pickle
import re
import unicodedata
import os
import random
import itertools
import tempfile
from functools import wraps
import multiprocessing
import shutil
import sys
import subprocess
import inspect
import heapq
from copy import deepcopy
from datetime import datetime
import platform

import numpy as np
import scipy.sparse
from smart_open import open

from gensim import __version__ as gensim_version

logger = logging.getLogger(__name__)

# When pickling objects for persistence, use this protocol by default.
# Note that users won't be able to load models saved with high protocols on older environments that do
# not support that protocol (e.g. Python 2).
# In the rare cases where this matters, users can explicitly pass `model.save(pickle_protocol=2)`.
# See also https://github.com/RaRe-Technologies/gensim/pull/3065
PICKLE_PROTOCOL = 4

PAT_ALPHABETIC = re.compile(r'(((?![\d])\w)+)', re.UNICODE)
RE_HTML_ENTITY = re.compile(r'&(#?)([xX]?)(\w{1,8});', re.UNICODE)

NO_CYTHON = RuntimeError(
    "Compiled extensions are unavailable. "
    "If you've installed from a package, ask the package maintainer to include compiled extensions. "
    "If you're building Gensim from source yourself, install Cython and a C compiler, and then "
    "run `python setup.py build_ext --inplace` to retry. "
)
"""An exception that gensim code raises when Cython extensions are unavailable."""

#: A default, shared numpy-Generator-based PRNG for any/all uses that don't require seeding
default_prng = np.random.default_rng()


def get_random_state(seed):
    """Generate :class:`numpy.random.RandomState` based on input seed.

    Parameters
    ----------
    seed : {None, int, array_like}
        Seed for random state.

    Returns
    -------
    :class:`numpy.random.RandomState`
        Random state.

    Raises
    ------
    AttributeError
        If seed is not {None, int, array_like}.

    Notes
    -----
    Method originally from `maciejkula/glove-python <https://github.com/maciejkula/glove-python>`_
    and written by `@joshloyal <https://github.com/joshloyal>`_.

    """
    if seed is None or seed is np.random:
        return np.random.mtrand._rand
    if isinstance(seed, (numbers.Integral, np.integer)):
        return np.random.RandomState(seed)
    if isinstance(seed, np.random.RandomState):
        return seed
    raise ValueError('%r cannot be used to seed a np.random.RandomState instance' % seed)


def synchronous(tlockname):
    """A decorator to place an instance-based lock around a method.

    Notes
    -----
    Adapted from http://code.activestate.com/recipes/577105-synchronization-decorator-for-class-methods/.

    """
    def _synched(func):
        @wraps(func)
        def _synchronizer(self, *args, **kwargs):
            tlock = getattr(self, tlockname)
            logger.debug("acquiring lock %r for %s", tlockname, func.__name__)

            with tlock:  # use lock as a context manager to perform safe acquire/release pairs
                logger.debug("acquired lock %r for %s", tlockname, func.__name__)
                result = func(self, *args, **kwargs)
                logger.debug("releasing lock %r for %s", tlockname, func.__name__)
                return result
        return _synchronizer
    return _synched


def file_or_filename(input):
    """Open a filename for reading with `smart_open`, or seek to the beginning if `input` is an already open file.

    Parameters
    ----------
    input : str or file-like
        Filename or file-like object.

    Returns
    -------
    file-like object
        An open file, positioned at the beginning.

    """
    if isinstance(input, str):
        # input was a filename: open as file
        return open(input, 'rb')
    else:
        # input already a file-like object; just reset to the beginning
        input.seek(0)
        return input


@contextmanager
def open_file(input):
    """Provide "with-like" behaviour without closing the file object.

    Parameters
    ----------
    input : str or file-like
        Filename or file-like object.

    Yields
    -------
    file
        File-like object based on input (or input if this already file-like).

    """
    mgr = file_or_filename(input)
    exc = False
    try:
        yield mgr
    except Exception:
        # Handling any unhandled exceptions from the code nested in 'with' statement.
        exc = True
        if not isinstance(input, str) or not mgr.__exit__(*sys.exc_info()):
            raise
        # Try to introspect and silence errors.
    finally:
        if not exc and isinstance(input, str):
            mgr.__exit__(None, None, None)


def deaccent(text):
    """Remove letter accents from the given string.

    Parameters
    ----------
    text : str
        Input string.

    Returns
    -------
    str
        Unicode string without accents.

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import deaccent
        >>> deaccent("Šéf chomutovských komunistů dostal poštou bílý prášek")
        u'Sef chomutovskych komunistu dostal postou bily prasek'

    """
    if not isinstance(text, str):
        # assume utf8 for byte strings, use default (strict) error handling
        text = text.decode('utf8')
    norm = unicodedata.normalize("NFD", text)
    result = ''.join(ch for ch in norm if unicodedata.category(ch) != 'Mn')
    return unicodedata.normalize("NFC", result)


def copytree_hardlink(source, dest):
    """Recursively copy a directory ala shutils.copytree, but hardlink files instead of copying.

    Parameters
    ----------
    source : str
        Path to source directory
    dest : str
        Path to destination directory

    Warnings
    --------
    Available on UNIX systems only.

    """
    copy2 = shutil.copy2
    try:
        shutil.copy2 = os.link
        shutil.copytree(source, dest)
    finally:
        shutil.copy2 = copy2


def tokenize(text, lowercase=False, deacc=False, encoding='utf8', errors="strict", to_lower=False, lower=False):
    """Iteratively yield tokens as unicode strings, optionally removing accent marks and lowercasing it.

    Parameters
    ----------
    text : str or bytes
        Input string.
    deacc : bool, optional
        Remove accentuation using :func:`~gensim.utils.deaccent`?
    encoding : str, optional
        Encoding of input string, used as parameter for :func:`~gensim.utils.to_unicode`.
    errors : str, optional
        Error handling behaviour, used as parameter for :func:`~gensim.utils.to_unicode`.
    lowercase : bool, optional
        Lowercase the input string?
    to_lower : bool, optional
        Same as `lowercase`. Convenience alias.
    lower : bool, optional
        Same as `lowercase`. Convenience alias.

    Yields
    ------
    str
        Contiguous sequences of alphabetic characters (no digits!), using :func:`~gensim.utils.simple_tokenize`

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import tokenize
        >>> list(tokenize('Nic nemůže letět rychlostí vyšší, než 300 tisíc kilometrů za sekundu!', deacc=True))
        [u'Nic', u'nemuze', u'letet', u'rychlosti', u'vyssi', u'nez', u'tisic', u'kilometru', u'za', u'sekundu']

    """
    lowercase = lowercase or to_lower or lower
    text = to_unicode(text, encoding, errors=errors)
    if lowercase:
        text = text.lower()
    if deacc:
        text = deaccent(text)
    return simple_tokenize(text)


def simple_tokenize(text):
    """Tokenize input test using :const:`gensim.utils.PAT_ALPHABETIC`.

    Parameters
    ----------
    text : str
        Input text.

    Yields
    ------
    str
        Tokens from `text`.

    """
    for match in PAT_ALPHABETIC.finditer(text):
        yield match.group()


def simple_preprocess(doc, deacc=False, min_len=2, max_len=15):
    """Convert a document into a list of lowercase tokens, ignoring tokens that are too short or too long.

    Uses :func:`~gensim.utils.tokenize` internally.

    Parameters
    ----------
    doc : str
        Input document.
    deacc : bool, optional
        Remove accent marks from tokens using :func:`~gensim.utils.deaccent`?
    min_len : int, optional
        Minimum length of token (inclusive). Shorter tokens are discarded.
    max_len : int, optional
        Maximum length of token in result (inclusive). Longer tokens are discarded.

    Returns
    -------
    list of str
        Tokens extracted from `doc`.

    """
    tokens = [
        token for token in tokenize(doc, lower=True, deacc=deacc, errors='ignore')
        if min_len <= len(token) <= max_len and not token.startswith('_')
    ]
    return tokens


def any2utf8(text, errors='strict', encoding='utf8'):
    """Convert a unicode or bytes string in the given encoding into a utf8 bytestring.

    Parameters
    ----------
    text : str
        Input text.
    errors : str, optional
        Error handling behaviour if `text` is a bytestring.
    encoding : str, optional
        Encoding of `text` if it is a bytestring.

    Returns
    -------
    str
        Bytestring in utf8.

    """

    if isinstance(text, str):
        return text.encode('utf8')
    # do bytestring -> unicode -> utf8 full circle, to ensure valid utf8
    return str(text, encoding, errors=errors).encode('utf8')


to_utf8 = any2utf8


def any2unicode(text, encoding='utf8', errors='strict'):
    """Convert `text` (bytestring in given encoding or unicode) to unicode.

    Parameters
    ----------
    text : str
        Input text.
    errors : str, optional
        Error handling behaviour if `text` is a bytestring.
    encoding : str, optional
        Encoding of `text` if it is a bytestring.

    Returns
    -------
    str
        Unicode version of `text`.

    """
    if isinstance(text, str):
        return text
    return str(text, encoding, errors=errors)


to_unicode = any2unicode


def call_on_class_only(*args, **kwargs):
    """Helper to raise `AttributeError` if a class method is called on an instance. Used internally.

    Parameters
    ----------
    *args
        Variable length argument list.
    **kwargs
        Arbitrary keyword arguments.

    Raises
    ------
    AttributeError
        If a class method is called on an instance.

    """
    raise AttributeError('This method should be called on a class object.')


class SaveLoad:
    """Serialize/deserialize objects from disk, by equipping them with the `save()` / `load()` methods.

    Warnings
    --------
    This uses pickle internally (among other techniques), so objects must not contain unpicklable attributes
    such as lambda functions etc.

    """
    def add_lifecycle_event(self, event_name, log_level=logging.INFO, **event):
        """
        Append an event into the `lifecycle_events` attribute of this object, and also
        optionally log the event at `log_level`.

        Events are important moments during the object's life, such as "model created",
        "model saved", "model loaded", etc.

        The `lifecycle_events` attribute is persisted across object's :meth:`~gensim.utils.SaveLoad.save`
        and :meth:`~gensim.utils.SaveLoad.load` operations. It has no impact on the use of the model,
        but is useful during debugging and support.

        Set `self.lifecycle_events = None` to disable this behaviour. Calls to `add_lifecycle_event()`
        will not record events into `self.lifecycle_events` then.

        Parameters
        ----------
        event_name : str
            Name of the event. Can be any label, e.g. "created", "stored" etc.
        event : dict
            Key-value mapping to append to `self.lifecycle_events`. Should be JSON-serializable, so keep it simple.
            Can be empty.

            This method will automatically add the following key-values to `event`, so you don't have to specify them:

            - `datetime`: the current date & time
            - `gensim`: the current Gensim version
            - `python`: the current Python version
            - `platform`: the current platform
            - `event`: the name of this event
        log_level : int
            Also log the complete event dict, at the specified log level. Set to False to not log at all.

        """
        # See also https://github.com/RaRe-Technologies/gensim/issues/2863

        event_dict = deepcopy(event)
        event_dict['datetime'] = datetime.now().isoformat()
        event_dict['gensim'] = gensim_version
        event_dict['python'] = sys.version
        event_dict['platform'] = platform.platform()
        event_dict['event'] = event_name

        if not hasattr(self, 'lifecycle_events'):
            # Avoid calling str(self), the object may not be fully initialized yet at this point.
            logger.debug("starting a new internal lifecycle event log for %s", self.__class__.__name__)
            self.lifecycle_events = []

        if log_level:
            logger.log(log_level, "%s lifecycle event %s", self.__class__.__name__, event_dict)

        if self.lifecycle_events is not None:
            self.lifecycle_events.append(event_dict)

    @classmethod
    def load(cls, fname, mmap=None):
        """Load an object previously saved using :meth:`~gensim.utils.SaveLoad.save` from a file.

        Parameters
        ----------
        fname : str
            Path to file that contains needed object.
        mmap : str, optional
            Memory-map option.  If the object was saved with large arrays stored separately, you can load these arrays
            via mmap (shared memory) using `mmap='r'.
            If the file being loaded is compressed (either '.gz' or '.bz2'), then `mmap=None` **must be** set.

        See Also
        --------
        :meth:`~gensim.utils.SaveLoad.save`
            Save object to file.

        Returns
        -------
        object
            Object loaded from `fname`.

        Raises
        ------
        AttributeError
            When called on an object instance instead of class (this is a class method).

        """
        logger.info("loading %s object from %s", cls.__name__, fname)

        compress, subname = SaveLoad._adapt_by_suffix(fname)

        obj = unpickle(fname)
        obj._load_specials(fname, mmap, compress, subname)
        obj.add_lifecycle_event("loaded", fname=fname)
        return obj

    def _load_specials(self, fname, mmap, compress, subname):
        """Load attributes that were stored separately, and give them the same opportunity
        to recursively load using the :class:`~gensim.utils.SaveLoad` interface.

        Parameters
        ----------
        fname : str
            Input file path.
        mmap :  {None, ‘r+’, ‘r’, ‘w+’, ‘c’}
            Memory-map options. See `numpy.load(mmap_mode)
            <https://docs.scipy.org/doc/numpy-1.14.0/reference/generated/numpy.load.html>`_.
        compress : bool
            Is the input file compressed?
        subname : str
            Attribute name. Set automatically during recursive processing.

        """
        def mmap_error(obj, filename):
            return IOError(
                'Cannot mmap compressed object %s in file %s. ' % (obj, filename)
                + 'Use `load(fname, mmap=None)` or uncompress files manually.'
            )

        for attrib in getattr(self, '__recursive_saveloads', []):
            cfname = '.'.join((fname, attrib))
            logger.info("loading %s recursively from %s.* with mmap=%s", attrib, cfname, mmap)
            with ignore_deprecation_warning():
                getattr(self, attrib)._load_specials(cfname, mmap, compress, subname)

        for attrib in getattr(self, '__numpys', []):
            logger.info("loading %s from %s with mmap=%s", attrib, subname(fname, attrib), mmap)

            if compress:
                if mmap:
                    raise mmap_error(attrib, subname(fname, attrib))

                val = np.load(subname(fname, attrib))['val']
            else:
                val = np.load(subname(fname, attrib), mmap_mode=mmap)

            with ignore_deprecation_warning():
                setattr(self, attrib, val)

        for attrib in getattr(self, '__scipys', []):
            logger.info("loading %s from %s with mmap=%s", attrib, subname(fname, attrib), mmap)
            sparse = unpickle(subname(fname, attrib))
            if compress:
                if mmap:
                    raise mmap_error(attrib, subname(fname, attrib))

                with np.load(subname(fname, attrib, 'sparse')) as f:
                    sparse.data = f['data']
                    sparse.indptr = f['indptr']
                    sparse.indices = f['indices']
            else:
                sparse.data = np.load(subname(fname, attrib, 'data'), mmap_mode=mmap)
                sparse.indptr = np.load(subname(fname, attrib, 'indptr'), mmap_mode=mmap)
                sparse.indices = np.load(subname(fname, attrib, 'indices'), mmap_mode=mmap)

            with ignore_deprecation_warning():
                setattr(self, attrib, sparse)

        for attrib in getattr(self, '__ignoreds', []):
            logger.info("setting ignored attribute %s to None", attrib)
            with ignore_deprecation_warning():
                setattr(self, attrib, None)

    @staticmethod
    def _adapt_by_suffix(fname):
        """Get compress setting and filename for numpy file compression.

        Parameters
        ----------
        fname : str
            Input filename.

        Returns
        -------
        (bool, function)
            First argument will be True if `fname` compressed.

        """
        compress, suffix = (True, 'npz') if fname.endswith('.gz') or fname.endswith('.bz2') else (False, 'npy')
        return compress, lambda *args: '.'.join(args + (suffix,))

    def _smart_save(
            self, fname,
            separately=None, sep_limit=10 * 1024**2, ignore=frozenset(), pickle_protocol=PICKLE_PROTOCOL,
        ):
        """Save the object to a file. Used internally by :meth:`gensim.utils.SaveLoad.save()`.

        Parameters
        ----------
        fname : str
            Path to file.
        separately : list, optional
            Iterable of attributes than need to store distinctly.
        sep_limit : int, optional
            Limit for separation.
        ignore : frozenset, optional
            Attributes that shouldn't be store.
        pickle_protocol : int, optional
            Protocol number for pickle.

        Notes
        -----
        If `separately` is None, automatically detect large numpy/scipy.sparse arrays in the object being stored,
        and store them into separate files. This avoids pickle memory errors and allows mmap'ing large arrays back
        on load efficiently.

        You can also set `separately` manually, in which case it must be a list of attribute names to be stored
        in separate files. The automatic check is not performed in this case.

        """
        compress, subname = SaveLoad._adapt_by_suffix(fname)

        restores = self._save_specials(
            fname, separately, sep_limit, ignore, pickle_protocol, compress, subname,
        )
        try:
            pickle(self, fname, protocol=pickle_protocol)
        finally:
            # restore attribs handled specially
            for obj, asides in restores:
                for attrib, val in asides.items():
                    with ignore_deprecation_warning():
                        setattr(obj, attrib, val)
        logger.info("saved %s", fname)

    def _save_specials(self, fname, separately, sep_limit, ignore, pickle_protocol, compress, subname):
        """Save aside any attributes that need to be handled separately, including
        by recursion any attributes that are themselves :class:`~gensim.utils.SaveLoad` instances.

        Parameters
        ----------
        fname : str
            Output filename.
        separately : list or None
            List of attributes to store separately.
        sep_limit : int
            Don't store arrays smaller than this separately. In bytes.
        ignore : iterable of str
            Attributes that shouldn't be stored at all.
        pickle_protocol : int
            Protocol number for pickle.
        compress : bool
            If True - compress output with :func:`numpy.savez_compressed`.
        subname : function
            Produced by :meth:`~gensim.utils.SaveLoad._adapt_by_suffix`

        Returns
        -------
        list of (obj, {attrib: value, ...})
            Settings that the caller should use to restore each object's attributes that were set aside
            during the default :func:`~gensim.utils.pickle`.

        """
        asides = {}
        sparse_matrices = (scipy.sparse.csr_matrix, scipy.sparse.csc_matrix)
        if separately is None:
            separately = []
            for attrib, val in self.__dict__.items():
                if isinstance(val, np.ndarray) and val.size >= sep_limit:
                    separately.append(attrib)
                elif isinstance(val, sparse_matrices) and val.nnz >= sep_limit:
                    separately.append(attrib)

        with ignore_deprecation_warning():
            # whatever's in `separately` or `ignore` at this point won't get pickled
            for attrib in separately + list(ignore):
                if hasattr(self, attrib):
                    asides[attrib] = getattr(self, attrib)
                    delattr(self, attrib)

        recursive_saveloads = []
        restores = []
        for attrib, val in self.__dict__.items():
            if hasattr(val, '_save_specials'):  # better than 'isinstance(val, SaveLoad)' if IPython reloading
                recursive_saveloads.append(attrib)
                cfname = '.'.join((fname, attrib))
                restores.extend(val._save_specials(cfname, None, sep_limit, ignore, pickle_protocol, compress, subname))

        try:
            numpys, scipys, ignoreds = [], [], []
            for attrib, val in asides.items():
                if isinstance(val, np.ndarray) and attrib not in ignore:
                    numpys.append(attrib)
                    logger.info("storing np array '%s' to %s", attrib, subname(fname, attrib))

                    if compress:
                        np.savez_compressed(subname(fname, attrib), val=np.ascontiguousarray(val))
                    else:
                        np.save(subname(fname, attrib), np.ascontiguousarray(val))

                elif isinstance(val, (scipy.sparse.csr_matrix, scipy.sparse.csc_matrix)) and attrib not in ignore:
                    scipys.append(attrib)
                    logger.info("storing scipy.sparse array '%s' under %s", attrib, subname(fname, attrib))

                    if compress:
                        np.savez_compressed(
                            subname(fname, attrib, 'sparse'),
                            data=val.data,
                            indptr=val.indptr,
                            indices=val.indices
                        )
                    else:
                        np.save(subname(fname, attrib, 'data'), val.data)
                        np.save(subname(fname, attrib, 'indptr'), val.indptr)
                        np.save(subname(fname, attrib, 'indices'), val.indices)

                    data, indptr, indices = val.data, val.indptr, val.indices
                    val.data, val.indptr, val.indices = None, None, None

                    try:
                        # store array-less object
                        pickle(val, subname(fname, attrib), protocol=pickle_protocol)
                    finally:
                        val.data, val.indptr, val.indices = data, indptr, indices
                else:
                    logger.info("not storing attribute %s", attrib)
                    ignoreds.append(attrib)

            self.__dict__['__numpys'] = numpys
            self.__dict__['__scipys'] = scipys
            self.__dict__['__ignoreds'] = ignoreds
            self.__dict__['__recursive_saveloads'] = recursive_saveloads
        except Exception:
            # restore the attributes if exception-interrupted
            for attrib, val in asides.items():
                setattr(self, attrib, val)
            raise
        return restores + [(self, asides)]

    def save(
            self, fname_or_handle,
            separately=None, sep_limit=10 * 1024**2, ignore=frozenset(), pickle_protocol=PICKLE_PROTOCOL,
        ):
        """Save the object to a file.

        Parameters
        ----------
        fname_or_handle : str or file-like
            Path to output file or already opened file-like object. If the object is a file handle,
            no special array handling will be performed, all attributes will be saved to the same file.
        separately : list of str or None, optional
            If None, automatically detect large numpy/scipy.sparse arrays in the object being stored, and store
            them into separate files. This prevent memory errors for large objects, and also allows
            `memory-mapping <https://en.wikipedia.org/wiki/Mmap>`_ the large arrays for efficient
            loading and sharing the large arrays in RAM between multiple processes.

            If list of str: store these attributes into separate files. The automated size check
            is not performed in this case.
        sep_limit : int, optional
            Don't store arrays smaller than this separately. In bytes.
        ignore : frozenset of str, optional
            Attributes that shouldn't be stored at all.
        pickle_protocol : int, optional
            Protocol number for pickle.

        See Also
        --------
        :meth:`~gensim.utils.SaveLoad.load`
            Load object from file.

        """
        self.add_lifecycle_event(
            "saving",
            fname_or_handle=str(fname_or_handle),
            separately=str(separately),
            sep_limit=sep_limit,
            ignore=ignore,
        )
        try:
            _pickle.dump(self, fname_or_handle, protocol=pickle_protocol)
            logger.info("saved %s object", self.__class__.__name__)
        except TypeError:  # `fname_or_handle` does not have write attribute
            self._smart_save(fname_or_handle, separately, sep_limit, ignore, pickle_protocol=pickle_protocol)


def identity(p):
    """Identity fnc, for flows that don't accept lambda (pickling etc).

    Parameters
    ----------
    p : object
        Input parameter.

    Returns
    -------
    object
        Same as `p`.

    """
    return p


def get_max_id(corpus):
    """Get the highest feature id that appears in the corpus.

    Parameters
    ----------
    corpus : iterable of iterable of (int, numeric)
        Collection of texts in BoW format.

    Returns
    ------
    int
        Highest feature id.

    Notes
    -----
    For empty `corpus` return -1.

    """
    maxid = -1
    for document in corpus:
        if document:
            maxid = max(maxid, max(fieldid for fieldid, _ in document))
    return maxid


class FakeDict:
    """Objects of this class act as dictionaries that map integer->str(integer), for a specified
    range of integers <0, num_terms).

    This is meant to avoid allocating real dictionaries when `num_terms` is huge, which is a waste of memory.

    """
    def __init__(self, num_terms):
        """

        Parameters
        ----------
        num_terms : int
            Number of terms.

        """
        self.num_terms = num_terms

    def __str__(self):
        return "FakeDict(num_terms=%s)" % self.num_terms

    def __getitem__(self, val):
        if 0 <= val < self.num_terms:
            return str(val)
        raise ValueError("internal id out of bounds (%s, expected <0..%s))" % (val, self.num_terms))

    def __contains__(self, val):
        return 0 <= val < self.num_terms

    def iteritems(self):
        """Iterate over all keys and values.

        Yields
        ------
        (int, str)
            Pair of (id, token).

        """
        for i in range(self.num_terms):
            yield i, str(i)

    def keys(self):
        """Override the `dict.keys()`, which is used to determine the maximum internal id of a corpus,
        i.e. the vocabulary dimensionality.

        Returns
        -------
        list of int
            Highest id, packed in list.

        Notes
        -----
        To avoid materializing the whole `range(0, self.num_terms)`,
        this returns the highest id = `[self.num_terms - 1]` only.

        """
        return [self.num_terms - 1]

    def __len__(self):
        return self.num_terms

    def get(self, val, default=None):
        if 0 <= val < self.num_terms:
            return str(val)
        return default


def dict_from_corpus(corpus):
    """Scan corpus for all word ids that appear in it, then construct a mapping
    which maps each `word_id` -> `str(word_id)`.

    Parameters
    ----------
    corpus : iterable of iterable of (int, numeric)
        Collection of texts in BoW format.

    Returns
    ------
    id2word : :class:`~gensim.utils.FakeDict`
        "Fake" mapping which maps each `word_id` -> `str(word_id)`.

    Warnings
    --------
    This function is used whenever *words* need to be displayed (as opposed to just their ids)
    but no `word_id` -> `word` mapping was provided. The resulting mapping only covers words actually
    used in the corpus, up to the highest `word_id` found.

    """
    num_terms = 1 + get_max_id(corpus)
    id2word = FakeDict(num_terms)
    return id2word


def is_corpus(obj):
    """Check whether `obj` is a corpus, by peeking at its first element. Works even on streamed generators.
    The peeked element is put back into a object returned by this function, so always use
    that returned object instead of the original `obj`.

    Parameters
    ----------
    obj : object
        An `iterable of iterable` that contains (int, numeric).

    Returns
    -------
    (bool, object)
        Pair of (is `obj` a corpus, `obj` with peeked element restored)

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import is_corpus
        >>> corpus = [[(1, 1.0)], [(2, -0.3), (3, 0.12)]]
        >>> corpus_or_not, corpus = is_corpus(corpus)

    Warnings
    --------
    An "empty" corpus (empty input sequence) is ambiguous, so in this case
    the result is forcefully defined as (False, `obj`).

    """
    try:
        if 'Corpus' in obj.__class__.__name__:  # the most common case, quick hack
            return True, obj
    except Exception:
        pass
    try:
        if hasattr(obj, 'next') or hasattr(obj, '__next__'):
            # the input is an iterator object, meaning once we call next()
            # that element could be gone forever. we must be careful to put
            # whatever we retrieve back again
            doc1 = next(obj)
            obj = itertools.chain([doc1], obj)
        else:
            doc1 = next(iter(obj))  # empty corpus is resolved to False here
        if len(doc1) == 0:  # sparse documents must have a __len__ function (list, tuple...)
            return True, obj  # the first document is empty=>assume this is a corpus

        # if obj is a 1D numpy array(scalars) instead of 2-tuples, it resolves to False here
        id1, val1 = next(iter(doc1))
        id1, val1 = int(id1), float(val1)  # must be a 2-tuple (integer, float)
    except Exception:
        return False, obj
    return True, obj


def get_my_ip():
    """Try to obtain our external ip (from the Pyro4 nameserver's point of view)

    Returns
    -------
    str
        IP address.

    Warnings
    --------
    This tries to sidestep the issue of bogus `/etc/hosts` entries and other local misconfiguration,
    which often mess up hostname resolution.
    If all else fails, fall back to simple `socket.gethostbyname()` lookup.

    """
    import socket
    try:
        from Pyro4.naming import locateNS
        # we know the nameserver must exist, so use it as our anchor point
        ns = locateNS()
        s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
        s.connect((ns._pyroUri.host, ns._pyroUri.port))
        result, port = s.getsockname()
    except Exception:
        try:
            # see what ifconfig says about our default interface
            import commands
            result = commands.getoutput("ifconfig").split("\n")[1].split()[1][5:]
            if len(result.split('.')) != 4:
                raise Exception()
        except Exception:
            # give up, leave the resolution to gethostbyname
            result = socket.gethostbyname(socket.gethostname())
    return result


class RepeatCorpus(SaveLoad):
    """Wrap a `corpus` as another corpus of length `reps`. This is achieved by repeating documents from `corpus`
    over and over again, until the requested length `len(result) == reps` is reached.
    Repetition is done on-the-fly=efficiently, via `itertools`.

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import RepeatCorpus
        >>>
        >>> corpus = [[(1, 2)], []]  # 2 documents
        >>> list(RepeatCorpus(corpus, 5))  # repeat 2.5 times to get 5 documents
        [[(1, 2)], [], [(1, 2)], [], [(1, 2)]]

    """
    def __init__(self, corpus, reps):
        """

        Parameters
        ----------
        corpus : iterable of iterable of (int, numeric)
            Input corpus.
        reps : int
            Number of repeats for documents from corpus.

        """
        self.corpus = corpus
        self.reps = reps

    def __iter__(self):
        return itertools.islice(itertools.cycle(self.corpus), self.reps)


class RepeatCorpusNTimes(SaveLoad):
    """Wrap a `corpus` and repeat it `n` times.

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import RepeatCorpusNTimes
        >>>
        >>> corpus = [[(1, 0.5)], []]
        >>> list(RepeatCorpusNTimes(corpus, 3))  # repeat 3 times
        [[(1, 0.5)], [], [(1, 0.5)], [], [(1, 0.5)], []]

    """
    def __init__(self, corpus, n):
        """

        Parameters
        ----------
        corpus : iterable of iterable of (int, numeric)
            Input corpus.
        n : int
            Number of repeats for corpus.

        """
        self.corpus = corpus
        self.n = n

    def __iter__(self):
        for _ in range(self.n):
            for document in self.corpus:
                yield document


class ClippedCorpus(SaveLoad):
    """Wrap a `corpus` and return `max_doc` element from it."""
    def __init__(self, corpus, max_docs=None):
        """

        Parameters
        ----------
        corpus : iterable of iterable of (int, numeric)
            Input corpus.
        max_docs : int
            Maximum number of documents in the wrapped corpus.

        Warnings
        --------
        Any documents after `max_docs` are ignored. This effectively limits the length of the returned corpus
        to <= `max_docs`. Set `max_docs=None` for "no limit", effectively wrapping the entire input corpus.

        """
        self.corpus = corpus
        self.max_docs = max_docs

    def __iter__(self):
        return itertools.islice(self.corpus, self.max_docs)

    def __len__(self):
        return min(self.max_docs, len(self.corpus))


class SlicedCorpus(SaveLoad):
    """Wrap `corpus` and return a slice of it."""
    def __init__(self, corpus, slice_):
        """

        Parameters
        ----------
        corpus : iterable of iterable of (int, numeric)
            Input corpus.
        slice_ : slice or iterable
            Slice for `corpus`.

        Notes
        -----
        Negative slicing can only be used if the corpus is indexable, otherwise, the corpus will be iterated over.
        Slice can also be a np.ndarray to support fancy indexing.

        Calculating the size of a SlicedCorpus is expensive when using a slice as the corpus has
        to be iterated over once. Using a list or np.ndarray does not have this drawback, but consumes more memory.

        """
        self.corpus = corpus
        self.slice_ = slice_
        self.length = None

    def __iter__(self):
        if hasattr(self.corpus, 'index') and len(self.corpus.index) > 0:
            return (self.corpus.docbyoffset(i) for i in self.corpus.index[self.slice_])
        return itertools.islice(self.corpus, self.slice_.start, self.slice_.stop, self.slice_.step)

    def __len__(self):
        # check cached length, calculate if needed
        if self.length is None:
            if isinstance(self.slice_, (list, np.ndarray)):
                self.length = len(self.slice_)
            elif isinstance(self.slice_, slice):
                (start, end, step) = self.slice_.indices(len(self.corpus.index))
                diff = end - start
                self.length = diff // step + (diff % step > 0)
            else:
                self.length = sum(1 for x in self)

        return self.length


def safe_unichr(intval):
    """Create a unicode character from its integer value. In case `unichr` fails, render the character
    as an escaped `\\U<8-byte hex value of intval>` string.

    Parameters
    ----------
    intval : int
        Integer code of character

    Returns
    -------
    string
        Unicode string of character

    """
    try:
        return chr(intval)
    except ValueError:
        # ValueError: chr() arg not in range(0x10000) (narrow Python build)
        s = "\\U%08x" % intval
        # return UTF16 surrogate pair
        return s.decode('unicode-escape')


def decode_htmlentities(text):
    """Decode all HTML entities in text that are encoded as hex, decimal or named entities.
    Adapted from `python-twitter-ircbot/html_decode.py
    <http://github.com/sku/python-twitter-ircbot/blob/321d94e0e40d0acc92f5bf57d126b57369da70de/html_decode.py>`_.

    Parameters
    ----------
    text : str
        Input HTML.

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import decode_htmlentities
        >>>
        >>> u = u'E tu vivrai nel terrore - L&#x27;aldil&#xE0; (1981)'
        >>> print(decode_htmlentities(u).encode('UTF-8'))
        E tu vivrai nel terrore - L'aldilà (1981)
        >>> print(decode_htmlentities("l&#39;eau"))
        l'eau
        >>> print(decode_htmlentities("foo &lt; bar"))
        foo < bar

    """
    def substitute_entity(match):
        try:
            ent = match.group(3)
            if match.group(1) == "#":
                # decoding by number
                if match.group(2) == '':
                    # number is in decimal
                    return safe_unichr(int(ent))
                elif match.group(2) in ['x', 'X']:
                    # number is in hex
                    return safe_unichr(int(ent, 16))
            else:
                # they were using a name
                cp = n2cp.get(ent)
                if cp:
                    return safe_unichr(cp)
                else:
                    return match.group()
        except Exception:
            # in case of errors, return original input
            return match.group()

    return RE_HTML_ENTITY.sub(substitute_entity, text)


def chunkize_serial(iterable, chunksize, as_numpy=False, dtype=np.float32):
    """Yield elements from `iterable` in "chunksize"-ed groups.

    The last returned element may be smaller if the length of collection is not divisible by `chunksize`.

    Parameters
    ----------
    iterable : iterable of object
        An iterable.
    chunksize : int
        Split iterable into chunks of this size.
    as_numpy : bool, optional
        Yield chunks as `np.ndarray` instead of lists.

    Yields
    ------
    list OR np.ndarray
        "chunksize"-ed chunks of elements from `iterable`.

    Examples
    --------
    .. sourcecode:: pycon

        >>> print(list(grouper(range(10), 3)))
        [[0, 1, 2], [3, 4, 5], [6, 7, 8], [9]]

    """
    it = iter(iterable)
    while True:
        if as_numpy:
            # convert each document to a 2d numpy array (~6x faster when transmitting
            # chunk data over the wire, in Pyro)
            wrapped_chunk = [[np.array(doc, dtype=dtype) for doc in itertools.islice(it, int(chunksize))]]
        else:
            wrapped_chunk = [list(itertools.islice(it, int(chunksize)))]
        if not wrapped_chunk[0]:
            break
        # memory opt: wrap the chunk and then pop(), to avoid leaving behind a dangling reference
        yield wrapped_chunk.pop()


grouper = chunkize_serial


class InputQueue(multiprocessing.Process):
    """Populate a queue of input chunks from a streamed corpus.

    Useful for reading and chunking corpora in the background, in a separate process,
    so that workers that use the queue are not starved for input chunks.

    """
    def __init__(self, q, corpus, chunksize, maxsize, as_numpy):
        """
        Parameters
        ----------
        q : multiprocessing.Queue
            Enqueue chunks into this queue.
        corpus : iterable of iterable of (int, numeric)
            Corpus to read and split into "chunksize"-ed groups
        chunksize : int
            Split `corpus` into chunks of this size.
        as_numpy : bool, optional
            Enqueue chunks as `numpy.ndarray` instead of lists.

        """
        super(InputQueue, self).__init__()
        self.q = q
        self.maxsize = maxsize
        self.corpus = corpus
        self.chunksize = chunksize
        self.as_numpy = as_numpy

    def run(self):
        it = iter(self.corpus)
        while True:
            chunk = itertools.islice(it, self.chunksize)
            if self.as_numpy:
                # HACK XXX convert documents to numpy arrays, to save memory.
                # This also gives a scipy warning at runtime:
                # "UserWarning: indices array has non-integer dtype (float64)"
                wrapped_chunk = [[np.asarray(doc) for doc in chunk]]
            else:
                wrapped_chunk = [list(chunk)]

            if not wrapped_chunk[0]:
                self.q.put(None, block=True)
                break

            try:
                qsize = self.q.qsize()
            except NotImplementedError:
                qsize = '?'
            logger.debug("prepared another chunk of %i documents (qsize=%s)", len(wrapped_chunk[0]), qsize)
            self.q.put(wrapped_chunk.pop(), block=True)


# Multiprocessing on Windows (and on OSX with python3.8+) uses "spawn" mode, which
# causes issues with pickling.
# So for these two platforms, use simpler serial processing in `chunkize`.
# See https://github.com/RaRe-Technologies/gensim/pull/2800#discussion_r410890171
if os.name == 'nt' or (sys.platform == "darwin" and sys.version_info >= (3, 8)):
    def chunkize(corpus, chunksize, maxsize=0, as_numpy=False):
        """Split `corpus` into fixed-sized chunks, using :func:`~gensim.utils.chunkize_serial`.

        Parameters
        ----------
        corpus : iterable of object
            An iterable.
        chunksize : int
            Split `corpus` into chunks of this size.
        maxsize : int, optional
            Ignored. For interface compatibility only.
        as_numpy : bool, optional
            Yield chunks as `np.ndarray` s instead of lists?

        Yields
        ------
        list OR np.ndarray
            "chunksize"-ed chunks of elements from `corpus`.

        """
        if maxsize > 0:
            entity = "Windows" if os.name == 'nt' else "OSX with python3.8+"
            warnings.warn("detected %s; aliasing chunkize to chunkize_serial" % entity)
        for chunk in chunkize_serial(corpus, chunksize, as_numpy=as_numpy):
            yield chunk
else:
    def chunkize(corpus, chunksize, maxsize=0, as_numpy=False):
        """Split `corpus` into fixed-sized chunks, using :func:`~gensim.utils.chunkize_serial`.

        Parameters
        ----------
        corpus : iterable of object
            An iterable.
        chunksize : int
            Split `corpus` into chunks of this size.
        maxsize : int, optional
            If > 0, prepare chunks in a background process, filling a chunk queue of size at most `maxsize`.
        as_numpy : bool, optional
            Yield chunks as `np.ndarray` instead of lists?

        Yields
        ------
        list OR np.ndarray
            "chunksize"-ed chunks of elements from `corpus`.

        Notes
        -----
        Each chunk is of length `chunksize`, except the last one which may be smaller.
        A once-only input stream (`corpus` from a generator) is ok, chunking is done efficiently via itertools.

        If `maxsize > 0`, don't wait idly in between successive chunk `yields`, but rather keep filling a short queue
        (of size at most `maxsize`) with forthcoming chunks in advance. This is realized by starting a separate process,
        and is meant to reduce I/O delays, which can be significant when `corpus` comes from a slow medium
        like HDD, database or network.

        If `maxsize == 0`, don't fool around with parallelism and simply yield the chunksize
        via :func:`~gensim.utils.chunkize_serial` (no I/O optimizations).

        Yields
        ------
        list of object OR np.ndarray
            Groups based on `iterable`

        """
        assert chunksize > 0

        if maxsize > 0:
            q = multiprocessing.Queue(maxsize=maxsize)
            worker = InputQueue(q, corpus, chunksize, maxsize=maxsize, as_numpy=as_numpy)
            worker.daemon = True
            worker.start()
            while True:
                chunk = [q.get(block=True)]
                if chunk[0] is None:
                    break
                yield chunk.pop()
        else:
            for chunk in chunkize_serial(corpus, chunksize, as_numpy=as_numpy):
                yield chunk


def smart_extension(fname, ext):
    """Append a file extension `ext` to `fname`, while keeping compressed extensions like `.bz2` or
    `.gz` (if any) at the end.

    Parameters
    ----------
    fname : str
        Filename or full path.
    ext : str
        Extension to append before any compression extensions.

    Returns
    -------
    str
        New path to file with `ext` appended.

    Examples
    --------

    .. sourcecode:: pycon

        >>> from gensim.utils import smart_extension
        >>> smart_extension("my_file.pkl.gz", ".vectors")
        'my_file.pkl.vectors.gz'

    """
    fname, oext = os.path.splitext(fname)
    if oext.endswith('.bz2'):
        fname = fname + oext[:-4] + ext + '.bz2'
    elif oext.endswith('.gz'):
        fname = fname + oext[:-3] + ext + '.gz'
    else:
        fname = fname + oext + ext

    return fname


def pickle(obj, fname, protocol=PICKLE_PROTOCOL):
    """Pickle object `obj` to file `fname`, using smart_open so that `fname` can be on S3, HDFS, compressed etc.

    Parameters
    ----------
    obj : object
        Any python object.
    fname : str
        Path to pickle file.
    protocol : int, optional
        Pickle protocol number.

    """
    with open(fname, 'wb') as fout:  # 'b' for binary, needed on Windows
        _pickle.dump(obj, fout, protocol=protocol)


def unpickle(fname):
    """Load object from `fname`, using smart_open so that `fname` can be on S3, HDFS, compressed etc.

    Parameters
    ----------
    fname : str
        Path to pickle file.

    Returns
    -------
    object
        Python object loaded from `fname`.

    """
    with open(fname, 'rb') as f:
        return _pickle.load(f, encoding='latin1')  # needed because loading from S3 doesn't support readline()


def revdict(d):
    """Reverse a dictionary mapping, i.e. `{1: 2, 3: 4}` -> `{2: 1, 4: 3}`.

    Parameters
    ----------
    d : dict
        Input dictionary.

    Returns
    -------
    dict
        Reversed dictionary mapping.

    Notes
    -----
    When two keys map to the same value, only one of them will be kept in the result (which one is kept is arbitrary).

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import revdict
        >>> d = {1: 2, 3: 4}
        >>> revdict(d)
        {2: 1, 4: 3}

    """
    return {v: k for (k, v) in dict(d).items()}


def deprecated(reason):
    """Decorator to mark functions as deprecated.

    Calling a decorated function will result in a warning being emitted, using warnings.warn.
    Adapted from https://stackoverflow.com/a/40301488/8001386.

    Parameters
    ----------
    reason : str
        Reason of deprecation.

    Returns
    -------
    function
        Decorated function

    """
    if isinstance(reason, str):
        def decorator(func):
            fmt = "Call to deprecated `{name}` ({reason})."

            @wraps(func)
            def new_func1(*args, **kwargs):
                warnings.warn(
                    fmt.format(name=func.__name__, reason=reason),
                    category=DeprecationWarning,
                    stacklevel=2
                )
                return func(*args, **kwargs)

            return new_func1
        return decorator

    elif inspect.isclass(reason) or inspect.isfunction(reason):
        func = reason
        fmt = "Call to deprecated `{name}`."

        @wraps(func)
        def new_func2(*args, **kwargs):
            warnings.warn(
                fmt.format(name=func.__name__),
                category=DeprecationWarning,
                stacklevel=2
            )
            return func(*args, **kwargs)
        return new_func2

    else:
        raise TypeError(repr(type(reason)))


@contextmanager
def ignore_deprecation_warning():
    """Contextmanager for ignoring DeprecationWarning."""
    with warnings.catch_warnings():
        warnings.filterwarnings("ignore", category=DeprecationWarning)
        yield


@deprecated("Function will be removed in 4.0.0")
def toptexts(query, texts, index, n=10):
    """Debug fnc to help inspect the top `n` most similar documents (according to a similarity index `index`),
    to see if they are actually related to the query.

    Parameters
    ----------
    query : {list of (int, number), numpy.ndarray}
        vector OR BoW (list of tuples)
    texts : str
        object that can return something insightful for each document via `texts[docid]`,
        such as its fulltext or snippet.
    index : any
        A instance from from :mod:`gensim.similarity.docsim`.

    Return
    ------
    list
        a list of 3-tuples (docid, doc's similarity to the query, texts[docid])

    """
    sims = index[query]  # perform a similarity query against the corpus
    sims = sorted(enumerate(sims), key=lambda item: -item[1])

    return [(topid, topcosine, texts[topid]) for topid, topcosine in sims[:n]]  # only consider top-n most similar docs


def randfname(prefix='gensim'):
    """Generate a random filename in temp.

    Parameters
    ----------
    prefix : str
        Prefix of filename.

    Returns
    -------
    str
        Full path in the in system's temporary folder, ending in a random filename.

    """
    randpart = hex(random.randint(0, 0xffffff))[2:]
    return os.path.join(tempfile.gettempdir(), prefix + randpart)


@deprecated("Function will be removed in 4.0.0")
def upload_chunked(server, docs, chunksize=1000, preprocess=None):
    """Memory-friendly upload of documents to a SimServer (or Pyro SimServer proxy).

    Notes
    -----
    Use this function to train or index large collections.abc -- avoid sending the
    entire corpus over the wire as a single Pyro in-memory object. The documents
    will be sent in smaller chunks, of `chunksize` documents each.

    """
    start = 0
    for chunk in grouper(docs, chunksize):
        end = start + len(chunk)
        logger.info("uploading documents %i-%i", start, end - 1)
        if preprocess is not None:
            pchunk = []
            for doc in chunk:
                doc['tokens'] = preprocess(doc['text'])
                del doc['text']
                pchunk.append(doc)
            chunk = pchunk
        server.buffer(chunk)
        start = end


def getNS(host=None, port=None, broadcast=True, hmac_key=None):
    """Get a Pyro4 name server proxy.

    Parameters
    ----------
    host : str, optional
        Name server hostname.
    port : int, optional
        Name server port.
    broadcast : bool, optional
        Use broadcast mechanism? (i.e. reach out to all Pyro nodes in the network)
    hmac_key : str, optional
        Private key.

    Raises
    ------
    RuntimeError
        When Pyro name server is not found.

    Returns
    -------
    :class:`Pyro4.core.Proxy`
        Proxy from Pyro4.

    """
    import Pyro4
    try:
        return Pyro4.locateNS(host, port, broadcast, hmac_key)
    except Pyro4.errors.NamingError:
        raise RuntimeError("Pyro name server not found")


def pyro_daemon(name, obj, random_suffix=False, ip=None, port=None, ns_conf=None):
    """Register an object with the Pyro name server.

    Start the name server if not running yet and block until the daemon is terminated.
    The object is registered under `name`, or `name`+ some random suffix if `random_suffix` is set.

    """
    if ns_conf is None:
        ns_conf = {}
    if random_suffix:
        name += '.' + hex(random.randint(0, 0xffffff))[2:]

    import Pyro4
    with getNS(**ns_conf) as ns:
        with Pyro4.Daemon(ip or get_my_ip(), port or 0) as daemon:
            # register server for remote access
            uri = daemon.register(obj, name)
            ns.remove(name)
            ns.register(name, uri)
            logger.info("%s registered with nameserver (URI '%s')", name, uri)
            daemon.requestLoop()


def mock_data_row(dim=1000, prob_nnz=0.5, lam=1.0):
    """Create a random gensim BoW vector, with the feature counts following the Poisson distribution.

    Parameters
    ----------
    dim : int, optional
        Dimension of vector.
    prob_nnz : float, optional
        Probability of each coordinate will be nonzero, will be drawn from the Poisson distribution.
    lam : float, optional
        Lambda parameter for the Poisson distribution.

    Returns
    -------
    list of (int, float)
        Vector in BoW format.

    """
    nnz = np.random.uniform(size=(dim,))
    return [(i, float(np.random.poisson(lam=lam) + 1.0)) for i in range(dim) if nnz[i] < prob_nnz]


def mock_data(n_items=1000, dim=1000, prob_nnz=0.5, lam=1.0):
    """Create a random Gensim-style corpus (BoW), using :func:`~gensim.utils.mock_data_row`.

    Parameters
    ----------
    n_items : int
        Size of corpus
    dim : int
        Dimension of vector, used for :func:`~gensim.utils.mock_data_row`.
    prob_nnz : float, optional
        Probability of each coordinate will be nonzero, will be drawn from Poisson distribution,
        used for :func:`~gensim.utils.mock_data_row`.
    lam : float, optional
        Parameter for Poisson distribution, used for :func:`~gensim.utils.mock_data_row`.

    Returns
    -------
    list of list of (int, float)
        Gensim-style corpus.

    """
    return [mock_data_row(dim=dim, prob_nnz=prob_nnz, lam=lam) for _ in range(n_items)]


def prune_vocab(vocab, min_reduce, trim_rule=None):
    """Remove all entries from the `vocab` dictionary with count smaller than `min_reduce`.

    Modifies `vocab` in place, returns the sum of all counts that were pruned.

    Parameters
    ----------
    vocab : dict
        Input dictionary.
    min_reduce : int
        Frequency threshold for tokens in `vocab`.
    trim_rule : function, optional
        Function for trimming entities from vocab, default behaviour is `vocab[w] <= min_reduce`.

    Returns
    -------
    result : int
        Sum of all counts that were pruned.

    """
    result = 0
    old_len = len(vocab)
    for w in list(vocab):  # make a copy of dict's keys
        if not keep_vocab_item(w, vocab[w], min_reduce, trim_rule):  # vocab[w] <= min_reduce:
            result += vocab[w]
            del vocab[w]
    logger.info(
        "pruned out %i tokens with count <=%i (before %i, after %i)",
        old_len - len(vocab), min_reduce, old_len, len(vocab)
    )
    return result


def trim_vocab_by_freq(vocab, topk, trim_rule=None):
    """Retain `topk` most frequent words in `vocab`.
    If there are more words with the same frequency as `topk`-th one, they will be kept.
    Modifies `vocab` in place, returns nothing.

    Parameters
    ----------
    vocab : dict
        Input dictionary.
    topk : int
        Number of words with highest frequencies to keep.
    trim_rule : function, optional
        Function for trimming entities from vocab, default behaviour is `vocab[w] <= min_count`.

    """
    if topk >= len(vocab):
        return

    min_count = heapq.nlargest(topk, vocab.values())[-1]
    prune_vocab(vocab, min_count, trim_rule=trim_rule)


def merge_counts(dict1, dict2):
    """Merge `dict1` of (word, freq1) and `dict2` of (word, freq2) into `dict1` of (word, freq1+freq2).
    Parameters
    ----------
    dict1 : dict of (str, int)
        First dictionary.
    dict2 : dict of (str, int)
        Second dictionary.
    Returns
    -------
    result : dict
        Merged dictionary with sum of frequencies as values.
    """
    for word, freq in dict2.items():
        if word in dict1:
            dict1[word] += freq
        else:
            dict1[word] = freq

    return dict1


def qsize(queue):
    """Get the (approximate) queue size where available.

    Parameters
    ----------
    queue : :class:`queue.Queue`
        Input queue.

    Returns
    -------
    int
        Queue size, -1 if `qsize` method isn't implemented (OS X).

    """
    try:
        return queue.qsize()
    except NotImplementedError:
        # OS X doesn't support qsize
        return -1


RULE_DEFAULT = 0
RULE_DISCARD = 1
RULE_KEEP = 2


def keep_vocab_item(word, count, min_count, trim_rule=None):
    """Should we keep `word` in the vocab or remove it?

    Parameters
    ----------
    word : str
        Input word.
    count : int
        Number of times that word appeared in a corpus.
    min_count : int
        Discard words with frequency smaller than this.
    trim_rule : function, optional
        Custom function to decide whether to keep or discard this word.
        If a custom `trim_rule` is not specified, the default behaviour is simply `count >= min_count`.

    Returns
    -------
    bool
        True if `word` should stay, False otherwise.

    """
    default_res = count >= min_count

    if trim_rule is None:
        return default_res
    else:
        rule_res = trim_rule(word, count, min_count)
        if rule_res == RULE_KEEP:
            return True
        elif rule_res == RULE_DISCARD:
            return False
        else:
            return default_res


def check_output(stdout=subprocess.PIPE, *popenargs, **kwargs):
    r"""Run OS command with the given arguments and return its output as a byte string.

    Backported from Python 2.7 with a few minor modifications. Widely used for :mod:`gensim.models.wrappers`.
    Behaves very similar to https://docs.python.org/2/library/subprocess.html#subprocess.check_output.

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import check_output
        >>> check_output(args=['echo', '1'])
        '1\n'

    Raises
    ------
    KeyboardInterrupt
        If Ctrl+C pressed.

    """
    try:
        logger.debug("COMMAND: %s %s", popenargs, kwargs)
        process = subprocess.Popen(stdout=stdout, *popenargs, **kwargs)
        output, unused_err = process.communicate()
        retcode = process.poll()
        if retcode:
            cmd = kwargs.get("args")
            if cmd is None:
                cmd = popenargs[0]
            error = subprocess.CalledProcessError(retcode, cmd)
            error.output = output
            raise error
        return output
    except KeyboardInterrupt:
        process.terminate()
        raise


def sample_dict(d, n=10, use_random=True):
    """Selected `n` (possibly random) items from the dictionary `d`.

    Parameters
    ----------
    d : dict
        Input dictionary.
    n : int, optional
        Number of items to select.
    use_random : bool, optional
        Select items randomly (without replacement), instead of by the natural dict iteration order?

    Returns
    -------
    list of (object, object)
        Selected items from dictionary, as a list.

    """
    selected_keys = random.sample(list(d), min(len(d), n)) if use_random else itertools.islice(d.keys(), n)
    return [(key, d[key]) for key in selected_keys]


def strided_windows(ndarray, window_size):
    """Produce a numpy.ndarray of windows, as from a sliding window.

    Parameters
    ----------
    ndarray : numpy.ndarray
        Input array
    window_size : int
        Sliding window size.

    Returns
    -------
    numpy.ndarray
        Subsequences produced by sliding a window of the given size over the `ndarray`.
        Since this uses striding, the individual arrays are views rather than copies of `ndarray`.
        Changes to one view modifies the others and the original.

    Examples
    --------
    .. sourcecode:: pycon

        >>> from gensim.utils import strided_windows
        >>> strided_windows(np.arange(5), 2)
        array([[0, 1],
               [1, 2],
               [2, 3],
               [3, 4]])
        >>> strided_windows(np.arange(10), 5)
        array([[0, 1, 2, 3, 4],
               [1, 2, 3, 4, 5],
               [2, 3, 4, 5, 6],
               [3, 4, 5, 6, 7],
               [4, 5, 6, 7, 8],
               [5, 6, 7, 8, 9]])

    """
    ndarray = np.asarray(ndarray)
    if window_size == ndarray.shape[0]:
        return np.array([ndarray])
    elif window_size > ndarray.shape[0]:
        return np.ndarray((0, 0))

    stride = ndarray.strides[0]
    return np.lib.stride_tricks.as_strided(
        ndarray, shape=(ndarray.shape[0] - window_size + 1, window_size),
        strides=(stride, stride))


def iter_windows(texts, window_size, copy=False, ignore_below_size=True, include_doc_num=False):
    """Produce a generator over the given texts using a sliding window of `window_size`.

    The windows produced are views of some subsequence of a text.
    To use deep copies instead, pass `copy=True`.

    Parameters
    ----------
    texts : list of str
        List of string sentences.
    window_size : int
        Size of sliding window.
    copy : bool, optional
        Produce deep copies.
    ignore_below_size : bool, optional
        Ignore documents that are not at least `window_size` in length?
    include_doc_num : bool, optional
        Yield the text position with `texts` along with each window?

    """
    for doc_num, document in enumerate(texts):
        for window in _iter_windows(document, window_size, copy, ignore_below_size):
            if include_doc_num:
                yield (doc_num, window)
            else:
                yield window


def _iter_windows(document, window_size, copy=False, ignore_below_size=True):
    doc_windows = strided_windows(document, window_size)
    if doc_windows.shape[0] == 0:
        if not ignore_below_size:
            yield document.copy() if copy else document
    else:
        for doc_window in doc_windows:
            yield doc_window.copy() if copy else doc_window


def flatten(nested_list):
    """Recursively flatten a nested sequence of elements.

    Parameters
    ----------
    nested_list : iterable
        Possibly nested sequence of elements to flatten.

    Returns
    -------
    list
        Flattened version of `nested_list` where any elements that are an iterable (`collections.abc.Iterable`)
        have been unpacked into the top-level list, in a recursive fashion.

    """
    return list(lazy_flatten(nested_list))


def lazy_flatten(nested_list):
    """Lazy version of :func:`~gensim.utils.flatten`.

    Parameters
    ----------
    nested_list : list
        Possibly nested list.

    Yields
    ------
    object
        Element of list

    """
    for el in nested_list:
        if isinstance(el, collections.abc.Iterable) and not isinstance(el, str):
            for sub in flatten(el):
                yield sub
        else:
            yield el


def save_as_line_sentence(corpus, filename):
    """Save the corpus in LineSentence format, i.e. each sentence on a separate line,
    tokens are separated by space.

    Parameters
    ----------
    corpus : iterable of iterables of strings

    """
    with open(filename, mode='wb', encoding='utf8') as fout:
        for sentence in corpus:
            line = any2unicode(' '.join(sentence) + '\n')
            fout.write(line)


def effective_n_jobs(n_jobs):
    """Determines the number of jobs can run in parallel.

    Just like in sklearn, passing n_jobs=-1 means using all available
    CPU cores.

    Parameters
    ----------
    n_jobs : int
        Number of workers requested by caller.

    Returns
    -------
    int
        Number of effective jobs.

    """
    if n_jobs == 0:
        raise ValueError('n_jobs == 0 in Parallel has no meaning')
    elif n_jobs is None:
        return 1
    elif n_jobs < 0:
        n_jobs = max(multiprocessing.cpu_count() + 1 + n_jobs, 1)
    return n_jobs