bazarr/libs/subliminal_patch/core.py

# coding=utf-8
from __future__ import absolute_import
import codecs
import json
import re
import os
import logging
import socket
import traceback
import time
import operator

import itertools
from six.moves.http_client import ResponseNotReady

import rarfile
import requests

from collections import defaultdict
from bs4 import UnicodeDammit
from babelfish import LanguageReverseError
from guessit.jsonutils import GuessitEncoder
from subliminal import ProviderError, refiner_manager

from .extensions import provider_registry
from subliminal.exceptions import ServiceUnavailable, DownloadLimitExceeded
from subliminal.score import compute_score as default_compute_score
from subliminal.utils import hash_napiprojekt, hash_opensubtitles, hash_shooter, hash_thesubdb
from subliminal.video import VIDEO_EXTENSIONS, Video, Episode, Movie
from subliminal.core import guessit, ProviderPool, io, is_windows_special_path, \
    ThreadPoolExecutor, check_video
from subliminal_patch.exceptions import TooManyRequests, APIThrottled

from subzero.language import Language, ENDSWITH_LANGUAGECODE_RE, FULL_LANGUAGE_LIST
try:
    from os import scandir
    _scandir_generic = scandir
except ImportError:
    from scandir import scandir, scandir_generic as _scandir_generic
import six

logger = logging.getLogger(__name__)

# may be absolute or relative paths; set to selected options
CUSTOM_PATHS = []
INCLUDE_EXOTIC_SUBS = True

DOWNLOAD_TRIES = 0
DOWNLOAD_RETRY_SLEEP = 6

# fixme: this may be overkill
REMOVE_CRAP_FROM_FILENAME = re.compile(r"(?i)(?:([\s_-]+(?:obfuscated|scrambled|nzbgeek|chamele0n|buymore|xpost|postbot"
                                       r"|asrequested)(?:\[.+\])?)|([\s_-]\w{2,})(\[.+\]))(?=\.\w+$|$)")

SUBTITLE_EXTENSIONS = ('.srt', '.sub', '.smi', '.txt', '.ssa', '.ass', '.mpl', '.vtt')


def remove_crap_from_fn(fn):
    # in case of the second regex part, the legit release group name will be in group(2), if it's followed by [string]
    # otherwise replace fully, because the first part matched
    def repl(m):
        return m.group(2) if len(m.groups()) == 3 else ""

    return REMOVE_CRAP_FROM_FILENAME.sub(repl, fn)


class SZProviderPool(ProviderPool):
    def __init__(self, providers=None, provider_configs=None, blacklist=None, throttle_callback=None,
                 pre_download_hook=None, post_download_hook=None, language_hook=None):
        #: Name of providers to use
        self.providers = providers

        #: Provider configuration
        self.provider_configs = provider_configs or {}

        #: Initialized providers
        self.initialized_providers = {}

        #: Discarded providers
        self.discarded_providers = set()

        self.blacklist = blacklist or []

        self.throttle_callback = throttle_callback

        self.pre_download_hook = pre_download_hook
        self.post_download_hook = post_download_hook
        self.language_hook = language_hook

        if not self.throttle_callback:
            self.throttle_callback = lambda x, y: x

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_value, traceback):
        self.terminate()

    def __getitem__(self, name):
        if name not in self.providers:
            raise KeyError
        if name not in self.initialized_providers:
            logger.info('Initializing provider %s', name)
            provider = provider_registry[name](**self.provider_configs.get(name, {}))
            provider.initialize()
            self.initialized_providers[name] = provider

        return self.initialized_providers[name]

    def __delitem__(self, name):
        if name not in self.initialized_providers:
            raise KeyError(name)

        try:
            logger.info('Terminating provider %s', name)
            self.initialized_providers[name].terminate()
        except (requests.Timeout, socket.timeout) as e:
            logger.error('Provider %r timed out, improperly terminated', name)
            self.throttle_callback(name, e)
        except Exception as e:
            logger.exception('Provider %r terminated unexpectedly', name)
            self.throttle_callback(name, e)

        del self.initialized_providers[name]

    def list_subtitles_provider(self, provider, video, languages):
        """List subtitles with a single provider.

        The video and languages are checked against the provider.
        
        patch: add traceback info

        :param str provider: name of the provider.
        :param video: video to list subtitles for.
        :type video: :class:`~subliminal.video.Video`
        :param languages: languages to search for.
        :type languages: set of :class:`~babelfish.language.Language`
        :return: found subtitles.
        :rtype: list of :class:`~subliminal.subtitle.Subtitle` or None

        """
        if self.language_hook:
            languages_search_base = self.language_hook(provider)
        else:
            languages_search_base = languages

        # check video validity
        if not provider_registry[provider].check(video):
            logger.info('Skipping provider %r: not a valid video', provider)
            return []

        # check whether we want to search this provider for the languages
        use_languages = languages_search_base & languages
        if not use_languages:
            logger.info('Skipping provider %r: no language to search for (advanced: %r, requested: %r)', provider,
                        languages_search_base, languages)
            return []

        # check supported languages
        provider_languages = provider_registry[provider].languages & use_languages
        if not provider_languages:
            logger.info('Skipping provider %r: no language to search for', provider)
            return []

        # list subtitles
        logger.info('Listing subtitles with provider %r and languages %r', provider, provider_languages)
        results = []
        try:
            try:
                results = self[provider].list_subtitles(video, provider_languages)
            except ResponseNotReady:
                logger.error('Provider %r response error, reinitializing', provider)
                try:
                    self[provider].terminate()
                    self[provider].initialize()
                    results = self[provider].list_subtitles(video, provider_languages)
                except:
                    logger.error('Provider %r reinitialization error: %s', provider, traceback.format_exc())

            seen = []
            out = []
            for s in results:
                if (str(provider), str(s.id)) in self.blacklist:
                    logger.info("Skipping blacklisted subtitle: %s", s)
                    continue
                if s.id in seen:
                    continue
                s.plex_media_fps = float(video.fps) if video.fps else None
                out.append(s)
                seen.append(s.id)

            return out

        except (requests.Timeout, socket.timeout) as e:
            logger.error('Provider %r timed out', provider)
            self.throttle_callback(provider, e)

        except Exception as e:
            logger.exception('Unexpected error in provider %r: %s', provider, traceback.format_exc())
            self.throttle_callback(provider, e)

    def list_subtitles(self, video, languages):
        """List subtitles.
        
        patch: handle LanguageReverseError

        :param video: video to list subtitles for.
        :type video: :class:`~subliminal.video.Video`
        :param languages: languages to search for.
        :type languages: set of :class:`~babelfish.language.Language`
        :return: found subtitles.
        :rtype: list of :class:`~subliminal.subtitle.Subtitle`

        """
        subtitles = []

        for name in self.providers:
            # check discarded providers
            if name in self.discarded_providers:
                logger.debug('Skipping discarded provider %r', name)
                continue

            # list subtitles
            try:
                provider_subtitles = self.list_subtitles_provider(name, video, languages)
            except LanguageReverseError:
                logger.exception("Unexpected language reverse error in %s, skipping. Error: %s", name,
                                 traceback.format_exc())
                continue

            if provider_subtitles is None:
                logger.info('Discarding provider %s', name)
                self.discarded_providers.add(name)
                continue

            # add the subtitles
            subtitles.extend(provider_subtitles)

        return subtitles

    def download_subtitle(self, subtitle):
        """Download `subtitle`'s :attr:`~subliminal.subtitle.Subtitle.content`.
        
        patch: add retry functionality
        
        :param subtitle: subtitle to download.
        :type subtitle: :class:`~subliminal.subtitle.Subtitle`
        :return: `True` if the subtitle has been successfully downloaded, `False` otherwise.
        :rtype: bool
        """
        # check discarded providers
        if subtitle.provider_name in self.discarded_providers:
            logger.warning('Provider %r is discarded', subtitle.provider_name)
            return False

        logger.info('Downloading subtitle %r', subtitle)
        tries = 0

        # retry downloading on failure until settings' download retry limit hit
        while True:
            tries += 1
            try:
                if self.pre_download_hook:
                    self.pre_download_hook(subtitle)

                self[subtitle.provider_name].download_subtitle(subtitle)
                if self.post_download_hook:
                    self.post_download_hook(subtitle)

                break
            except (requests.ConnectionError,
                    requests.exceptions.ProxyError,
                    requests.exceptions.SSLError,
                    requests.Timeout,
                    socket.timeout) as e:
                logger.error('Provider %r connection error', subtitle.provider_name)
                self.throttle_callback(subtitle.provider_name, e)

            except ResponseNotReady as e:
                logger.error('Provider %r response error, reinitializing', subtitle.provider_name)
                try:
                    self[subtitle.provider_name].terminate()
                    self[subtitle.provider_name].initialize()
                except:
                    logger.error('Provider %r reinitialization error: %s', subtitle.provider_name,
                                 traceback.format_exc())
                    self.throttle_callback(subtitle.provider_name, e)

            except rarfile.BadRarFile:
                logger.error('Malformed RAR file from provider %r, skipping subtitle.', subtitle.provider_name)
                logger.debug("RAR Traceback: %s", traceback.format_exc())
                return False

            except Exception as e:
                logger.exception('Unexpected error in provider %r, Traceback: %s', subtitle.provider_name,
                                 traceback.format_exc())
                self.throttle_callback(subtitle.provider_name, e)
                self.discarded_providers.add(subtitle.provider_name)
                return False

            if tries == DOWNLOAD_TRIES:
                self.discarded_providers.add(subtitle.provider_name)
                logger.error('Maximum retries reached for provider %r, discarding it', subtitle.provider_name)
                return False

            # don't hammer the provider
            logger.debug('Errors while downloading subtitle, retrying provider %r in %s seconds',
                         subtitle.provider_name, DOWNLOAD_RETRY_SLEEP)
            time.sleep(DOWNLOAD_RETRY_SLEEP)

        # check subtitle validity
        if not subtitle.is_valid():
            logger.error('Invalid subtitle')
            return False

        if not os.environ.get("SZ_KEEP_ENCODING", False):
            subtitle.normalize()

        return True

    def download_best_subtitles(self, subtitles, video, languages, min_score=0, hearing_impaired=False, only_one=False,
                                compute_score=None):
        """Download the best matching subtitles.
        
        patch: 
            - hearing_impaired is now string
            - add .score to subtitle
            - move all languages check further to the top (still necessary?)

        :param subtitles: the subtitles to use.
        :type subtitles: list of :class:`~subliminal.subtitle.Subtitle`
        :param video: video to download subtitles for.
        :type video: :class:`~subliminal.video.Video`
        :param languages: languages to download.
        :type languages: set of :class:`~babelfish.language.Language`
        :param int min_score: minimum score for a subtitle to be downloaded.
        :param bool hearing_impaired: hearing impaired preference.
        :param bool only_one: download only one subtitle, not one per language.
        :param compute_score: function that takes `subtitle` and `video` as positional arguments,
            `hearing_impaired` as keyword argument and returns the score.
        :return: downloaded subtitles.
        :rtype: list of :class:`~subliminal.subtitle.Subtitle`

        """
        compute_score = compute_score or default_compute_score
        use_hearing_impaired = hearing_impaired in ("prefer", "force HI")

        is_episode = isinstance(video, Episode)

        # sort subtitles by score
        unsorted_subtitles = []

        for s in subtitles:
            # get the matches
            if s.language not in languages:
                logger.debug("%r: Skipping, language not searched for", s)
                continue

            try:
                matches = s.get_matches(video)
            except AttributeError:
                logger.error("%r: Match computation failed: %s", s, traceback.format_exc())
                continue

            orig_matches = matches.copy()

            logger.debug('%r: Found matches %r', s, matches)
            score, score_without_hash = compute_score(matches, s, video, hearing_impaired=use_hearing_impaired)
            unsorted_subtitles.append(
                (s, score, score_without_hash, matches, orig_matches))

        # sort subtitles by score
        scored_subtitles = sorted(unsorted_subtitles, key=operator.itemgetter(1, 2), reverse=True)

        # download best subtitles, falling back on the next on error
        downloaded_subtitles = []
        for subtitle, score, score_without_hash, matches, orig_matches in scored_subtitles:
            # check score
            if score < min_score:
                logger.info('%r: Score %d is below min_score (%d)', subtitle, score, min_score)
                break

            # stop when all languages are downloaded
            if set(s.language for s in downloaded_subtitles) == languages:
                logger.debug('All languages downloaded')
                break

            # check downloaded languages
            if subtitle.language in set(s.language for s in downloaded_subtitles):
                logger.debug('%r: Skipping subtitle: already downloaded', subtitle.language)
                continue

            # bail out if hearing_impaired was wrong
            if subtitle.hearing_impaired_verifiable and "hearing_impaired" not in matches and \
                            hearing_impaired in ("force HI", "force non-HI"):
                logger.debug('%r: Skipping subtitle with score %d because hearing-impaired set to %s', subtitle,
                             score, hearing_impaired)
                continue

            if is_episode:
                can_verify_series = True
                if not subtitle.hash_verifiable and "hash" in matches:
                    can_verify_series = False

                matches_series = False
                if {"season", "episode"}.issubset(orig_matches) and \
                                ("series" in orig_matches or "imdb_id" in orig_matches):
                    matches_series = True

                if can_verify_series and not matches_series:
                    logger.debug("%r: Skipping subtitle with score %d, because it doesn't match our series/episode",
                                 subtitle, score)
                    continue

            # download
            logger.debug("%r: Trying to download subtitle with matches %s, score: %s; release(s): %s", subtitle, matches,
                         score, subtitle.release_info)
            if self.download_subtitle(subtitle):
                subtitle.score = score
                downloaded_subtitles.append(subtitle)

            # stop if only one subtitle is requested
            if only_one:
                logger.debug('Only one subtitle downloaded')
                break

        return downloaded_subtitles


class SZAsyncProviderPool(SZProviderPool):
    """Subclass of :class:`ProviderPool` with asynchronous support for :meth:`~ProviderPool.list_subtitles`.

    :param int max_workers: maximum number of threads to use. If `None`, :attr:`max_workers` will be set
        to the number of :attr:`~ProviderPool.providers`.

    """
    def __init__(self, max_workers=None, *args, **kwargs):
        super(SZAsyncProviderPool, self).__init__(*args, **kwargs)

        #: Maximum number of threads to use
        self.max_workers = max_workers or len(self.providers)
        logger.info("Using %d threads for %d providers (%s)", self.max_workers, len(self.providers), self.providers)

    def list_subtitles_provider(self, provider, video, languages):
        # list subtitles
        provider_subtitles = None
        try:
            provider_subtitles = super(SZAsyncProviderPool, self).list_subtitles_provider(provider, video, languages)
        except LanguageReverseError:
            logger.exception("Unexpected language reverse error in %s, skipping. Error: %s", provider,
                             traceback.format_exc())

        return provider, provider_subtitles

    def list_subtitles(self, video, languages, blacklist=None):
        if is_windows_special_path:
            return super(SZAsyncProviderPool, self).list_subtitles(video, languages)

        subtitles = []

        with ThreadPoolExecutor(self.max_workers) as executor:
            for provider, provider_subtitles in executor.map(self.list_subtitles_provider, self.providers,
                                                             itertools.repeat(video, len(self.providers)),
                                                             itertools.repeat(languages, len(self.providers))):
                # discard provider that failed
                if provider_subtitles is None:
                    logger.info('Discarding provider %s', provider)
                    self.discarded_providers.add(provider)
                    continue

                # add subtitles
                subtitles.extend(provider_subtitles)

        return subtitles


if is_windows_special_path:
    SZAsyncProviderPool = SZProviderPool


def scan_video(path, dont_use_actual_file=False, hints=None, providers=None, skip_hashing=False, hash_from=None):
    """Scan a video from a `path`.

    patch:
        - allow passing of hints/options to guessit
        - allow dry-run with dont_use_actual_file
        - add crap removal (obfuscated/scrambled)
        - trust plex's movie name

    :param str path: existing path to the video.
    :return: the scanned video.
    :rtype: :class:`~subliminal.video.Video`

    """
    hints = hints or {}
    video_type = hints.get("type")

    # check for non-existing path
    if not dont_use_actual_file and not os.path.exists(path):
        raise ValueError('Path does not exist')

    # check video extension
    if not path.lower().endswith(VIDEO_EXTENSIONS):
        raise ValueError('%r is not a valid video extension' % os.path.splitext(path)[1])

    dirpath, filename = os.path.split(path)
    logger.info('Determining basic video properties for %r in %r', filename, dirpath)

    # hint guessit the filename itself and its 2 parent directories if we're an episode (most likely
    # Series name/Season/filename), else only one
    split_path = os.path.normpath(path).split(os.path.sep)[-3 if video_type == "episode" else -2:]

    # remove crap from folder names
    if video_type == "episode":
        if len(split_path) > 2:
            split_path[-3] = remove_crap_from_fn(split_path[-3])
    else:
        if len(split_path) > 1:
            split_path[-2] = remove_crap_from_fn(split_path[-2])

    guess_from = os.path.join(*split_path)

    # remove crap from file name
    guess_from = remove_crap_from_fn(guess_from)

    # guess
    hints["single_value"] = True
    if "title" in hints:
        hints["expected_title"] = [hints["title"]]

    guessed_result = guessit(guess_from, options=hints)

    logger.debug('GuessIt found: %s', json.dumps(guessed_result, cls=GuessitEncoder, indent=4, ensure_ascii=False))
    video = Video.fromguess(path, guessed_result)
    video.hints = hints

    # get possibly alternative title from the filename itself
    alt_guess = guessit(filename, options=hints)
    if "title" in alt_guess and alt_guess["title"] != guessed_result["title"]:
        if video_type == "episode":
            video.alternative_series.append(alt_guess["title"])
        else:
            video.alternative_titles.append(alt_guess["title"])
        logger.debug("Adding alternative title: %s", alt_guess["title"])

    if dont_use_actual_file and not hash_from:
        return video

    # size and hashes
    if not skip_hashing:
        hash_path = hash_from or path
        video.size = os.path.getsize(hash_path)
        if video.size > 10485760:
            logger.debug('Size is %d', video.size)
            osub_hash = None

            if "bsplayer" in providers:
                video.hashes['bsplayer'] = osub_hash = hash_opensubtitles(hash_path)

            if "opensubtitles" in providers:
                video.hashes['opensubtitles'] = osub_hash = osub_hash or hash_opensubtitles(hash_path)

            if "opensubtitlescom" in providers:
                video.hashes['opensubtitlescom'] = osub_hash = osub_hash or hash_opensubtitles(hash_path)

            if "shooter" in providers:
                video.hashes['shooter'] = hash_shooter(hash_path)

            if "thesubdb" in providers:
                video.hashes['thesubdb'] = hash_thesubdb(hash_path)

            if "napiprojekt" in providers:
                try:
                    video.hashes['napiprojekt'] = hash_napiprojekt(hash_path)
                except MemoryError:
                    logger.warning(u"Couldn't compute napiprojekt hash for %s", hash_path)

            if "napisy24" in providers:
                # Napisy24 uses the same hash as opensubtitles
                video.hashes['napisy24'] = osub_hash or hash_opensubtitles(hash_path)

            logger.debug('Computed hashes %r', video.hashes)
        else:
            logger.warning('Size is lower than 10MB: hashes not computed')

    return video


def _search_external_subtitles(path, languages=None, only_one=False, scandir_generic=False, match_strictness="strict"):
    dirpath, filename = os.path.split(path)
    dirpath = dirpath or '.'
    fn_no_ext, fileext = os.path.splitext(filename)
    fn_no_ext_lower = fn_no_ext.lower()
    subtitles = {}
    _scandir = _scandir_generic if scandir_generic else scandir

    for entry in _scandir(dirpath):
        if (not entry.name or entry.name in ('\x0c', '$', ',', '\x7f')) and not scandir_generic:
            logger.debug('Could not determine the name of the file, retrying with scandir_generic')
            return _search_external_subtitles(path, languages, only_one, True)
        if not entry.is_file(follow_symlinks=False):
            continue

        p = entry.name

        # keep only valid subtitle filenames
        if not p.lower().endswith(SUBTITLE_EXTENSIONS):
            continue

        # not p.lower().startswith(fileroot.lower()) or not

        p_root, p_ext = os.path.splitext(p)
        if not INCLUDE_EXOTIC_SUBS and p_ext not in (".srt", ".ass", ".ssa", ".vtt"):
            continue

        if p_root.lower() == fn_no_ext_lower:
            # skip check for language code if the subtitle file name is the same as the video name
            subtitles[p] = None
            continue

        # extract potential forced/normal/default/hi tag
        # fixme: duplicate from subtitlehelpers
        split_tag = p_root.rsplit('.', 1)
        adv_tag = None
        if len(split_tag) > 1:
            adv_tag = split_tag[1].lower()
            if adv_tag in ['forced', 'normal', 'default', 'embedded', 'embedded-forced', 'custom', 'hi', 'cc', 'sdh']:
                p_root = split_tag[0]

        forced = False
        if adv_tag:
            forced = "forced" in adv_tag

        hi = False
        if adv_tag:
            hi_tag = ["hi", "cc", "sdh"]
            hi = any(i for i in hi_tag if i in adv_tag)

        #add simplified/traditional chinese detection
        simplified_chinese = ["chs", "sc", "zhs", "hans", "gb", u"简", u"双语"]
        traditional_chinese = ["cht", "tc", "zht", "hant", "big5", u"繁", u"雙語"]
        FULL_LANGUAGE_LIST.extend(simplified_chinese)
        FULL_LANGUAGE_LIST.extend(traditional_chinese)
        p_root = p_root.replace('zh-TW', 'zht')

        # remove possible language code for matching
        p_root_bare = ENDSWITH_LANGUAGECODE_RE.sub(
            lambda m: "" if str(m.group(1)).lower() in FULL_LANGUAGE_LIST else m.group(0), p_root)

        p_root_lower = p_root_bare.lower()

        filename_matches = p_root_lower == fn_no_ext_lower
        filename_contains = p_root_lower in fn_no_ext_lower

        if not filename_matches:
            if match_strictness == "strict" or (match_strictness == "loose" and not filename_contains):
                continue

        language = None

        # extract the potential language code
        try:
            language_code = p_root.rsplit(".", 1)[1].replace('_', '-')
            try:
                language = Language.fromietf(language_code)      
                language.forced = forced
                language.hi = hi
            except (ValueError, LanguageReverseError):
                #add simplified/traditional chinese detection
                if any(ext in str(language_code) for ext in simplified_chinese):
                    language = Language.fromietf('zh')
                    language.forced = forced
                    language.hi = hi
                elif any(ext in str(language_code) for ext in traditional_chinese):
                    language = Language.fromietf('zh') 
                    language.forced = forced
                    language.hi = hi
                else:
                    logger.error('Cannot parse language code %r', language_code)
                    language_code = None
        except IndexError:
                language_code = None

        if not language and not language_code and only_one:
            language = Language.rebuild(list(languages)[0], forced=forced, hi=hi)

        subtitles[p] = language

    logger.debug('Found subtitles %r', subtitles)

    return subtitles


def search_external_subtitles(path, languages=None, only_one=False, match_strictness="strict"):
    """
    wrap original search_external_subtitles function to search multiple paths for one given video
    # todo: cleanup and merge with _search_external_subtitles
    """
    video_path, video_filename = os.path.split(path)
    subtitles = {}
    for folder_or_subfolder in [video_path] + CUSTOM_PATHS:
        # folder_or_subfolder may be a relative path or an absolute one
        try:
            abspath = six.text_type(os.path.abspath(
                os.path.join(*[video_path if not os.path.isabs(folder_or_subfolder) else "", folder_or_subfolder,
                               video_filename])))
        except Exception as e:
            logger.error("skipping path %s because of %s", repr(folder_or_subfolder), e)
            continue
        logger.debug("external subs: scanning path %s", abspath)

        if os.path.isdir(os.path.dirname(abspath)):
            try:
                subtitles.update(_search_external_subtitles(abspath, languages=languages,
                                                            only_one=only_one, match_strictness=match_strictness))
            except OSError:
                subtitles.update(_search_external_subtitles(abspath, languages=languages,
                                                            only_one=only_one, match_strictness=match_strictness,
                                                            scandir_generic=True))
    logger.debug("external subs: found %s", subtitles)
    return subtitles


def list_all_subtitles(videos, languages, **kwargs):
    """List all available subtitles.
    
    patch: remove video check, it has been done before

    The `videos` must pass the `languages` check of :func:`check_video`.

    All other parameters are passed onwards to the :class:`ProviderPool` constructor.

    :param videos: videos to list subtitles for.
    :type videos: set of :class:`~subliminal.video.Video`
    :param languages: languages to search for.
    :type languages: set of :class:`~babelfish.language.Language`
    :return: found subtitles per video.
    :rtype: dict of :class:`~subliminal.video.Video` to list of :class:`~subliminal.subtitle.Subtitle`

    """
    listed_subtitles = defaultdict(list)

    # return immediatly if no video passed the checks
    if not videos:
        return listed_subtitles

    # list subtitles
    with SZProviderPool(**kwargs) as pool:
        for video in videos:
            logger.info('Listing subtitles for %r', video)
            subtitles = pool.list_subtitles(video, languages - video.subtitle_languages)
            listed_subtitles[video].extend(subtitles)
            logger.info('Found %d subtitle(s)', len(subtitles))

    return listed_subtitles


def download_subtitles(subtitles, pool_class=ProviderPool, **kwargs):
    """Download :attr:`~subliminal.subtitle.Subtitle.content` of `subtitles`.

    :param subtitles: subtitles to download.
    :type subtitles: list of :class:`~subliminal.subtitle.Subtitle`
    :param pool_class: class to use as provider pool.
    :type pool_class: :class:`ProviderPool`, :class:`AsyncProviderPool` or similar
    :param \*\*kwargs: additional parameters for the provided `pool_class` constructor.

    """
    with pool_class(**kwargs) as pool:
        for subtitle in subtitles:
            logger.info('Downloading subtitle %r with score %s', subtitle, subtitle.score)
            pool.download_subtitle(subtitle)


def download_best_subtitles(videos, languages, min_score=0, hearing_impaired=False, only_one=False, compute_score=None,
                            pool_class=ProviderPool, throttle_time=0, **kwargs):
    """List and download the best matching subtitles.

    The `videos` must pass the `languages` and `undefined` (`only_one`) checks of :func:`check_video`.

    :param videos: videos to download subtitles for.
    :type videos: set of :class:`~subliminal.video.Video`
    :param languages: languages to download.
    :type languages: set of :class:`~babelfish.language.Language`
    :param int min_score: minimum score for a subtitle to be downloaded.
    :param bool hearing_impaired: hearing impaired preference.
    :param bool only_one: download only one subtitle, not one per language.
    :param compute_score: function that takes `subtitle` and `video` as positional arguments,
        `hearing_impaired` as keyword argument and returns the score.
    :param pool_class: class to use as provider pool.
    :type pool_class: :class:`ProviderPool`, :class:`AsyncProviderPool` or similar
    :param \*\*kwargs: additional parameters for the provided `pool_class` constructor.
    :return: downloaded subtitles per video.
    :rtype: dict of :class:`~subliminal.video.Video` to list of :class:`~subliminal.subtitle.Subtitle`

    """
    downloaded_subtitles = defaultdict(list)

    # check videos
    checked_videos = []
    for video in videos:
        if not check_video(video, languages=languages, undefined=only_one):
            logger.info('Skipping video %r', video)
            continue
        checked_videos.append(video)

    # return immediately if no video passed the checks
    if not checked_videos:
        return downloaded_subtitles

    got_multiple = len(checked_videos) > 1

    # download best subtitles
    with pool_class(**kwargs) as pool:
        for video in checked_videos:
            logger.info('Downloading best subtitles for %r', video)
            subtitles = pool.download_best_subtitles(pool.list_subtitles(video, languages - video.subtitle_languages),
                                                     video, languages, min_score=min_score,
                                                     hearing_impaired=hearing_impaired, only_one=only_one,
                                                     compute_score=compute_score)
            logger.info('Downloaded %d subtitle(s)', len(subtitles))
            downloaded_subtitles[video].extend(subtitles)

            if got_multiple and throttle_time:
                logger.debug("Waiting %ss before continuing ...", throttle_time)
                time.sleep(throttle_time)

    return downloaded_subtitles


def get_subtitle_path(video_path, language=None, extension='.srt', forced_tag=False, hi_tag=False, tags=None):
    """Get the subtitle path using the `video_path` and `language`.

    :param str video_path: path to the video.
    :param language: language of the subtitle to put in the path.
    :type language: :class:`~babelfish.language.Language`
    :param str extension: extension of the subtitle.
    :return: path of the subtitle.
    :rtype: str

    """
    subtitle_root = os.path.splitext(video_path)[0]
    tags = tags or []
    if forced_tag:
        tags.append("forced")

    elif hi_tag:
        tags.append("hi")

    if language:
        subtitle_root += '.' + str(language.basename)

    if tags:
        subtitle_root += ".%s" % "-".join(tags)

    return subtitle_root + extension


def save_subtitles(file_path, subtitles, single=False, directory=None, chmod=None, formats=("srt",),
                   tags=None, path_decoder=None, debug_mods=False):
    """Save subtitles on filesystem.

    Subtitles are saved in the order of the list. If a subtitle with a language has already been saved, other subtitles
    with the same language are silently ignored.

    The extension used is `.lang.srt` by default or `.srt` is `single` is `True`, with `lang` being the IETF code for
    the :attr:`~subliminal.subtitle.Subtitle.language` of the subtitle.

    :param file_path: video file path
    :param formats: list of "srt" and "vtt"
    :param subtitles: subtitles to save.
    :type subtitles: list of :class:`~subliminal.subtitle.Subtitle`
    :param bool single: save a single subtitle, default is to save one subtitle per language.
    :param str directory: path to directory where to save the subtitles, default is next to the video.
    :return: the saved subtitles
    :rtype: list of :class:`~subliminal.subtitle.Subtitle`

    patch: unicode path problems
    """

    logger.debug("Subtitle formats requested: %r", formats)

    saved_subtitles = []
    for subtitle in subtitles:
        # check content
        if subtitle.content is None:
            logger.error('Skipping subtitle %r: no content', subtitle)
            continue

        # check language
        if subtitle.language in set(s.language for s in saved_subtitles):
            logger.debug('Skipping subtitle %r: language already saved', subtitle)
            continue

        # create subtitle path
        subtitle_path = get_subtitle_path(file_path, None if single else subtitle.language,
                                          forced_tag=subtitle.language.forced, hi_tag=subtitle.language.hi, tags=tags)
        if directory is not None:
            subtitle_path = os.path.join(directory, os.path.split(subtitle_path)[1])

        if path_decoder:
            subtitle_path = path_decoder(subtitle_path)

        # force unicode
        subtitle_path = UnicodeDammit(subtitle_path).unicode_markup

        subtitle.storage_path = subtitle_path

        for format in formats:
            if format != "srt":
                subtitle_path = os.path.splitext(subtitle_path)[0] + (u".%s" % format)

            logger.debug(u"Saving %r to %r", subtitle, subtitle_path)
            content = subtitle.get_modified_content(format=format, debug=debug_mods)
            if content:
                if os.path.exists(subtitle_path):
                    os.remove(subtitle_path)

                with open(subtitle_path, 'wb') as f:
                    f.write(content)
                subtitle.storage_path = subtitle_path
            else:
                logger.error(u"Something went wrong when getting modified subtitle for %s", subtitle)

        # change chmod if requested
        if chmod:
            os.chmod(subtitle_path, chmod)

        saved_subtitles.append(subtitle)

        # check single
        if single:
            break

    return saved_subtitles


def refine(video, episode_refiners=None, movie_refiners=None, **kwargs):
    """Refine a video using :ref:`refiners`.
    
    patch: add traceback logging

    .. note::

        Exceptions raised in refiners are silently passed and logged.

    :param video: the video to refine.
    :type video: :class:`~subliminal.video.Video`
    :param tuple episode_refiners: refiners to use for episodes.
    :param tuple movie_refiners: refiners to use for movies.
    :param \*\*kwargs: additional parameters for the :func:`~subliminal.refiners.refine` functions.

    """
    refiners = ()
    if isinstance(video, Episode):
        refiners = episode_refiners or ('metadata', 'tvdb', 'omdb')
    elif isinstance(video, Movie):
        refiners = movie_refiners or ('metadata', 'omdb')
    for refiner in refiners:
        logger.info('Refining video with %s', refiner)
        try:
            refiner_manager[refiner].plugin(video, **kwargs)
        except:
            logger.error('Failed to refine video: %s', traceback.format_exc())