def download(obj, provider, language, age, directory, encoding, single, force, hearing_impaired, min_score, max_workers,
verbose, path):
"""Download best subtitles.
PATH can be an directory containing videos, a video file path or a video file name. It can be used multiple times.
If an existing subtitle is detected (external or embedded) in the correct language, the download is skipped for
the associated video.
"""
# process parameters
language = set(language)
# scan videos
videos = []
ignored_videos = []
errored_paths = []
with click.progressbar(path, label='Collecting videos', item_show_func=lambda p: p or '') as bar:
for p in bar:
logger.debug('Collecting path %s', p)
# non-existing
if not os.path.exists(p):
try:
video = Video.fromname(p)
except:
logger.exception('Unexpected error while collecting non-existing path %s', p)
errored_paths.append(p)
continue
videos.append(video)
continue
# directories
if os.path.isdir(p):
try:
scanned_videos = scan_videos(p, subtitles=not force, embedded_subtitles=not force,
subtitles_dir=directory, age=age)
except:
logger.exception('Unexpected error while collecting directory path %s', p)
errored_paths.append(p)
continue
for video in scanned_videos:
if check_video(video, languages=language, age=age, undefined=single):
videos.append(video)
else:
ignored_videos.append(video)
continue
# other inputs
try:
video = scan_video(p, subtitles=not force, embedded_subtitles=not force, subtitles_dir=directory)
except:
logger.exception('Unexpected error while collecting path %s', p)
errored_paths.append(p)
continue
if check_video(video, languages=language, age=age, undefined=single):
videos.append(video)
else:
ignored_videos.append(video)
# output errored paths
if verbose > 0:
for p in errored_paths:
click.secho('%s errored' % p, fg='red')
# output ignored videos
if verbose > 1:
for video in ignored_videos:
click.secho('%s ignored - subtitles: %s / age: %d day%s' % (
os.path.split(video.name)[1],
', '.join(str(s) for s in video.subtitle_languages) or 'none',
video.age.days,
's' if video.age.days > 1 else ''
), fg='yellow')
# report collected videos
click.echo('%s video%s collected / %s video%s ignored / %s error%s' % (
click.style(str(len(videos)), bold=True, fg='green' if videos else None),
's' if len(videos) > 1 else '',
click.style(str(len(ignored_videos)), bold=True, fg='yellow' if ignored_videos else None),
's' if len(ignored_videos) > 1 else '',
click.style(str(len(errored_paths)), bold=True, fg='red' if errored_paths else None),
's' if len(errored_paths) > 1 else '',
))
# exit if no video collected
if not videos:
return
# download best subtitles
downloaded_subtitles = defaultdict(list)
with AsyncProviderPool(max_workers=max_workers, providers=provider, provider_configs=obj['provider_configs']) as p:
with click.progressbar(videos, label='Downloading subtitles',
item_show_func=lambda v: os.path.split(v.name)[1] if v is not None else '') as bar:
for v in bar:
scores = get_scores(v)
subtitles = p.download_best_subtitles(p.list_subtitles(v, language - v.subtitle_languages),
v, language, min_score=scores['hash'] * min_score / 100,
hearing_impaired=hearing_impaired, only_one=single)
downloaded_subtitles[v] = subtitles
# save subtitles
total_subtitles = 0
for v, subtitles in downloaded_subtitles.items():
saved_subtitles = save_subtitles(v, subtitles, single=single, directory=directory, encoding=encoding)
total_subtitles += len(saved_subtitles)
if verbose > 0:
click.echo('%s subtitle%s downloaded for %s' % (click.style(str(len(saved_subtitles)), bold=True),
's' if len(saved_subtitles) > 1 else '',
os.path.split(v.name)[1]))
if verbose > 1:
for s in saved_subtitles:
matches = s.get_matches(v)
score = compute_score(s, v)
# score color
score_color = None
scores = get_scores(v)
if isinstance(v, Movie):
if score < scores['title']:
score_color = 'red'
elif score < scores['title'] + scores['year'] + scores['release_group']:
score_color = 'yellow'
else:
score_color = 'green'
elif isinstance(v, Episode):
if score < scores['series'] + scores['season'] + scores['episode']:
score_color = 'red'
elif score < scores['series'] + scores['season'] + scores['episode'] + scores['release_group']:
score_color = 'yellow'
else:
score_color = 'green'
# scale score from 0 to 100 taking out preferences
scaled_score = score
if s.hearing_impaired == hearing_impaired:
scaled_score -= scores['hearing_impaired']
scaled_score *= 100 / scores['hash']
# echo some nice colored output
click.echo(' - [{score}] {language} subtitle from {provider_name} (match on {matches})'.format(
score=click.style('{:5.1f}'.format(scaled_score), fg=score_color, bold=score >= scores['hash']),
language=s.language.name if s.language.country is None else '%s (%s)' % (s.language.name,
s.language.country.name),
provider_name=s.provider_name,
matches=', '.join(sorted(matches, key=scores.get, reverse=True))
))
if verbose == 0:
click.echo('Downloaded %s subtitle%s' % (click.style(str(total_subtitles), bold=True),
's' if total_subtitles > 1 else ''))
def compute_score(matches,
subtitle,
video,
hearing_impaired=None,
score_obj=None):
"""Compute the score of the `subtitle` against the `video` with `hearing_impaired` preference.
patch:
- remove upper bounds of score
- re-add matches argument and remove get_matches from here
:func:`compute_score` uses the :meth:`Subtitle.get_matches <subliminal.subtitle.Subtitle.get_matches>` method and
applies the scores (either from :data:`episode_scores` or :data:`movie_scores`) after some processing.
:param subtitle: the subtitle to compute the score of.
:type subtitle: :class:`~subliminal.subtitle.Subtitle`
:param video: the video to compute the score against.
:type video: :class:`~subliminal.video.Video`
:param bool hearing_impaired: hearing impaired preference.
:return: score of the subtitle.
:rtype: int
"""
logger.info('%r: Computing score for video %r with %r', subtitle, video,
dict(hearing_impaired=hearing_impaired))
if score_obj is not None:
scores = score_obj.scores
score_obj.check_custom_profiles(subtitle, matches)
else:
scores = get_scores(video)
is_episode = isinstance(video, Episode)
is_movie = isinstance(video, Movie)
episode_hash_valid_if = {"series", "season", "episode", "source"}
movie_hash_valid_if = {"video_codec", "source"}
orig_matches = matches.copy()
# on hash match, discard everything else
if subtitle.hash_verifiable and 'hash' in matches:
# hash is error-prone, try to fix that
hash_valid_if = episode_hash_valid_if if is_episode else movie_hash_valid_if
# don't validate hashes of specials, as season and episode tend to be wrong
if is_movie or not video.is_special:
if hash_valid_if <= set(matches):
# series, season and episode matched, hash is valid
logger.debug(
'%r: Using valid hash, as %s are correct (%r) and (%r)',
subtitle, hash_valid_if, matches, video)
matches &= {'hash'}
else:
# no match, invalidate hash
logger.debug(
'%r: Ignoring hash as other matches are wrong (missing: %r) and (%r)',
subtitle, hash_valid_if - matches, video)
matches -= {"hash"}
elif 'hash' in matches:
logger.debug('%r: Hash not verifiable for this provider. Keeping it',
subtitle)
matches &= {'hash'}
# handle equivalent matches
eq_matches = set()
if is_episode:
_episode_checks(video, eq_matches, matches)
elif is_movie and 'imdb_id' in matches:
logger.debug('Adding imdb_id match equivalents')
eq_matches |= {'title', 'year'}
matches |= eq_matches
# handle hearing impaired
if hearing_impaired is not None and subtitle.hearing_impaired == hearing_impaired:
logger.debug('Matched hearing_impaired')
matches.add('hearing_impaired')
orig_matches.add('hearing_impaired')
# compute the score
score = sum((scores.get(match, 0) for match in matches))
logger.info('%r: Computed score %r with final matches %r', subtitle, score,
matches)
score_without_hash = sum((scores.get(match, 0)
for match in orig_matches | eq_matches
if match != "hash"))
return score, score_without_hash
def download(obj, provider, language, age, directory, encoding, single, force,
hearing_impaired, min_score, max_workers, verbose, path):
"""Download best subtitles.
PATH can be an directory containing videos, a video file path or a video file name. It can be used multiple times.
If an existing subtitle is detected (external or embedded) in the correct language, the download is skipped for
the associated video.
"""
# process parameters
language = set(language)
# scan videos
videos = []
ignored_videos = []
errored_paths = []
with click.progressbar(path,
label='Collecting videos',
item_show_func=lambda p: p or '') as bar:
for p in bar:
logger.debug('Collecting path %s', p)
# non-existing
if not os.path.exists(p):
try:
video = Video.fromname(p)
except:
logger.exception(
'Unexpected error while collecting non-existing path %s',
p)
errored_paths.append(p)
continue
videos.append(video)
continue
# directories
if os.path.isdir(p):
try:
scanned_videos = scan_videos(p,
subtitles=not force,
embedded_subtitles=not force,
subtitles_dir=directory,
age=age)
except:
logger.exception(
'Unexpected error while collecting directory path %s',
p)
errored_paths.append(p)
continue
for video in scanned_videos:
if check_video(video,
languages=language,
age=age,
undefined=single):
videos.append(video)
else:
ignored_videos.append(video)
continue
# other inputs
try:
video = scan_video(p,
subtitles=not force,
embedded_subtitles=not force,
subtitles_dir=directory)
except:
logger.exception('Unexpected error while collecting path %s',
p)
errored_paths.append(p)
continue
if check_video(video,
languages=language,
age=age,
undefined=single):
videos.append(video)
else:
ignored_videos.append(video)
# output errored paths
if verbose > 0:
for p in errored_paths:
click.secho('%s errored' % p, fg='red')
# output ignored videos
if verbose > 1:
for video in ignored_videos:
click.secho(
'%s ignored - subtitles: %s / age: %d day%s' %
(os.path.split(video.name)[1],
', '.join(str(s) for s in video.subtitle_languages)
or 'none', video.age.days, 's' if video.age.days > 1 else ''),
fg='yellow')
# report collected videos
click.echo('%s video%s collected / %s video%s ignored / %s error%s' % (
click.style(
str(len(videos)), bold=True, fg='green' if videos else None),
's' if len(videos) > 1 else '',
click.style(str(len(ignored_videos)),
bold=True,
fg='yellow' if ignored_videos else None),
's' if len(ignored_videos) > 1 else '',
click.style(str(len(errored_paths)),
bold=True,
fg='red' if errored_paths else None),
's' if len(errored_paths) > 1 else '',
))
# exit if no video collected
if not videos:
return
# download best subtitles
downloaded_subtitles = defaultdict(list)
with AsyncProviderPool(max_workers=max_workers,
providers=provider,
provider_configs=obj['provider_configs']) as p:
with click.progressbar(
videos,
label='Downloading subtitles',
item_show_func=lambda v: os.path.split(v.name)[1]
if v is not None else '') as bar:
for v in bar:
scores = get_scores(v)
subtitles = p.download_best_subtitles(
p.list_subtitles(v, language - v.subtitle_languages),
v,
language,
min_score=scores['hash'] * min_score / 100,
hearing_impaired=hearing_impaired,
only_one=single)
downloaded_subtitles[v] = subtitles
# save subtitles
total_subtitles = 0
for v, subtitles in downloaded_subtitles.items():
saved_subtitles = save_subtitles(v,
subtitles,
single=single,
directory=directory,
encoding=encoding)
total_subtitles += len(saved_subtitles)
if verbose > 0:
click.echo(
'%s subtitle%s downloaded for %s' %
(click.style(str(len(saved_subtitles)), bold=True), 's' if
len(saved_subtitles) > 1 else '', os.path.split(v.name)[1]))
if verbose > 1:
for s in saved_subtitles:
matches = s.get_matches(v)
score = compute_score(s, v)
# score color
score_color = None
scores = get_scores(v)
if isinstance(v, Movie):
if score < scores['title']:
score_color = 'red'
elif score < scores['title'] + scores['year'] + scores[
'release_group']:
score_color = 'yellow'
else:
score_color = 'green'
elif isinstance(v, Episode):
if score < scores['series'] + scores['season'] + scores[
'episode']:
score_color = 'red'
elif score < scores['series'] + scores['season'] + scores[
'episode'] + scores['release_group']:
score_color = 'yellow'
else:
score_color = 'green'
# scale score from 0 to 100 taking out preferences
scaled_score = score
if s.hearing_impaired == hearing_impaired:
scaled_score -= scores['hearing_impaired']
scaled_score *= 100 / scores['hash']
# echo some nice colored output
click.echo(
' - [{score}] {language} subtitle from {provider_name} (match on {matches})'
.format(score=click.style('{:5.1f}'.format(scaled_score),
fg=score_color,
bold=score >= scores['hash']),
language=s.language.name
if s.language.country is None else '%s (%s)' %
(s.language.name, s.language.country.name),
provider_name=s.provider_name,
matches=', '.join(
sorted(matches, key=scores.get,
reverse=True))))
if verbose == 0:
click.echo('Downloaded %s subtitle%s' %
(click.style(str(total_subtitles), bold=True),
's' if total_subtitles > 1 else ''))