def guess_filetype(filename, filetype = 'autodetect'):
other = {}
# look at the extension first
fileext = os.path.splitext(filename)[1][1:].lower()
if fileext in subtitle_exts:
if 'movie' in filetype:
filetype = 'moviesubtitle'
elif 'episode' in filetype:
filetype = 'episodesubtitle'
else:
filetype = 'subtitle'
other = { 'container': fileext }
elif fileext in video_exts:
if filetype == 'autodetect':
filetype = 'video'
other = { 'container': fileext }
else:
if filetype == 'autodetect':
filetype = 'unknown'
other = { 'extension': fileext }
# now look whether there are some specific hints for episode vs movie
if filetype in ('video', 'subtitle'):
for rexp, confidence, span_adjust in episode_rexps:
match = re.search(rexp, filename, re.IGNORECASE)
if match:
if filetype == 'video':
filetype = 'episode'
elif filetype == 'subtitle':
filetype = 'episodesubtitle'
break
for prop, value, start, end in find_properties(filename):
if canonical_form(value) == 'DVB':
if filetype == 'video':
filetype = 'episode'
elif filetype == 'subtitle':
filetype = 'episodesubtitle'
break
# if no episode info found, assume it's a movie
if filetype == 'video':
filetype = 'movie'
elif filetype == 'subtitle':
filetype = 'moviesubtitle'
return filetype, other
def guess_filetype(mtree, filetype):
# put the filetype inside a dummy container to be able to have the
# following functions work correctly as closures
# this is a workaround for python 2 which doesn't have the
# 'nonlocal' keyword (python 3 does have it)
filetype_container = [filetype]
other = {}
filename = mtree.string
def upgrade_episode():
if filetype_container[0] == 'video':
filetype_container[0] = 'episode'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'episodesubtitle'
def upgrade_movie():
if filetype_container[0] == 'video':
filetype_container[0] = 'movie'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'moviesubtitle'
def upgrade_subtitle():
if 'movie' in filetype_container[0]:
filetype_container[0] = 'moviesubtitle'
elif 'episode' in filetype_container[0]:
filetype_container[0] = 'episodesubtitle'
else:
filetype_container[0] = 'subtitle'
def upgrade(type='unknown'):
if filetype_container[0] == 'autodetect':
filetype_container[0] = type
# look at the extension first
fileext = os.path.splitext(filename)[1][1:].lower()
if fileext in subtitle_exts:
upgrade_subtitle()
other = {'container': fileext}
elif fileext in video_exts:
upgrade(type='video')
other = {'container': fileext}
else:
upgrade(type='unknown')
other = {'extension': fileext}
# check whether we are in a 'Movies', 'Tv Shows', ... folder
folder_rexps = [(r'Movies?', upgrade_movie),
(r'Tv[ _-]?Shows?', upgrade_episode),
(r'Series', upgrade_episode)]
for frexp, upgrade_func in folder_rexps:
frexp = re.compile(frexp, re.IGNORECASE)
for pathgroup in mtree.children:
if frexp.match(pathgroup.value):
upgrade_func()
# check for a few specific cases which will unintentionally make the
# following heuristics confused (eg: OSS 117 will look like an episode,
# season 1, epnum 17, when it is in fact a movie)
fname = clean_string(filename).lower()
for m in MOVIES:
if m in fname:
upgrade_movie()
for s in SERIES:
if s in fname:
upgrade_episode()
# now look whether there are some specific hints for episode vs movie
if filetype_container[0] in ('video', 'subtitle'):
# if we have an episode_rexp (eg: s02e13), it is an episode
for rexp, _, _ in episode_rexps:
match = re.search(rexp, filename, re.IGNORECASE)
if match:
upgrade_episode()
break
# if we have a 3-4 digit number that's not a year, maybe an episode
match = re.search(r'[^0-9]([0-9]{3,4})[^0-9]', filename)
if match:
fullnumber = int(match.group()[1:-1])
#season = fullnumber // 100
epnumber = fullnumber % 100
possible = True
# check for validity
if epnumber > 40:
possible = False
if valid_year(fullnumber):
possible = False
if possible:
upgrade_episode()
# if we have certain properties characteristic of episodes, it is an ep
for prop, value, _, _ in find_properties(filename):
log.debug('prop: %s = %s' % (prop, value))
if prop == 'episodeFormat':
upgrade_episode()
break
elif compute_canonical_form('format', value) == 'DVB':
upgrade_episode()
break
# origin-specific type
if 'tvu.org.ru' in filename:
upgrade_episode()
# if no episode info found, assume it's a movie
upgrade_movie()
filetype = filetype_container[0]
return filetype, other
def guess_properties(string):
try:
prop, value, pos, end = find_properties(string)[0]
return { prop: value }, (pos, end)
except IndexError:
return None, None
def guess_filetype(mtree, filetype):
# put the filetype inside a dummy container to be able to have the
# following functions work correctly as closures
# this is a workaround for python 2 which doesn't have the
# 'nonlocal' keyword (python 3 does have it)
filetype_container = [filetype]
other = {}
filename = mtree.string
def upgrade_episode():
if filetype_container[0] == 'video':
filetype_container[0] = 'episode'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'episodesubtitle'
def upgrade_movie():
if filetype_container[0] == 'video':
filetype_container[0] = 'movie'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'moviesubtitle'
def upgrade_subtitle():
if 'movie' in filetype_container[0]:
filetype_container[0] = 'moviesubtitle'
elif 'episode' in filetype_container[0]:
filetype_container[0] = 'episodesubtitle'
else:
filetype_container[0] = 'subtitle'
def upgrade(type='unknown'):
if filetype_container[0] == 'autodetect':
filetype_container[0] = type
# look at the extension first
fileext = os.path.splitext(filename)[1][1:].lower()
if fileext in subtitle_exts:
upgrade_subtitle()
other = { 'container': fileext }
elif fileext in video_exts:
upgrade(type='video')
other = { 'container': fileext }
else:
upgrade(type='unknown')
other = { 'extension': fileext }
# check whether we are in a 'Movies', 'Tv Shows', ... folder
folder_rexps = [ (r'Movies?', upgrade_movie),
(r'Tv ?Shows?', upgrade_episode),
(r'Series', upgrade_episode)
]
for frexp, upgrade_func in folder_rexps:
frexp = re.compile(frexp, re.IGNORECASE)
for pathgroup in mtree.children:
if frexp.match(pathgroup.value):
upgrade_func()
# check for a few specific cases which will unintentionally make the
# following heuristics confused (eg: OSS 117 will look like an episode,
# season 1, epnum 17, when it is in fact a movie)
fname = clean_string(filename).lower()
for m in MOVIES:
if m in fname:
upgrade_movie()
for s in SERIES:
if s in fname:
upgrade_episode()
# now look whether there are some specific hints for episode vs movie
if filetype_container[0] in ('video', 'subtitle'):
# if we have an episode_rexp (eg: s02e13), it is an episode
for rexp, _, _ in episode_rexps:
match = re.search(rexp, filename, re.IGNORECASE)
if match:
upgrade_episode()
break
# if we have a 3-4 digit number that's not a year, maybe an episode
match = re.search(r'[^0-9]([0-9]{3,4})[^0-9]', filename)
if match:
fullnumber = int(match.group()[1:-1])
#season = fullnumber // 100
epnumber = fullnumber % 100
possible = True
# check for validity
if epnumber > 40:
possible = False
if valid_year(fullnumber):
possible = False
if possible:
upgrade_episode()
# if we have certain properties characteristic of episodes, it is an ep
for prop, value, _, _ in find_properties(filename):
log.debug('prop: %s = %s' % (prop, value))
if prop == 'episodeFormat':
upgrade_episode()
break
elif canonical_form(value) == 'DVB':
upgrade_episode()
break
# origin-specific type
if 'tvu.org.ru' in filename:
upgrade_episode()
# if no episode info found, assume it's a movie
upgrade_movie()
filetype = filetype_container[0]
return filetype, other
def guess_filetype(filename, filetype):
other = {}
# look at the extension first
fileext = os.path.splitext(filename)[1][1:].lower()
if fileext in subtitle_exts:
if 'movie' in filetype:
filetype = 'moviesubtitle'
elif 'episode' in filetype:
filetype = 'episodesubtitle'
else:
filetype = 'subtitle'
other = { 'container': fileext }
elif fileext in video_exts:
if filetype == 'autodetect':
filetype = 'video'
other = { 'container': fileext }
else:
if filetype == 'autodetect':
filetype = 'unknown'
other = { 'extension': fileext }
# put the filetype inside a dummy container to be able to have the
# following functions work correctly as closures
# this is a workaround for python 2 which doesn't have the
# 'nonlocal' keyword (python 3 does have it)
filetype_container = [filetype]
def upgrade_episode():
if filetype_container[0] == 'video':
filetype_container[0] = 'episode'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'episodesubtitle'
def upgrade_movie():
if filetype_container[0] == 'video':
filetype_container[0] = 'movie'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'moviesubtitle'
# now look whether there are some specific hints for episode vs movie
if filetype in ('video', 'subtitle'):
for rexp, _, _ in episode_rexps:
match = re.search(rexp, filename, re.IGNORECASE)
if match:
upgrade_episode()
break
for prop, value, _, _ in find_properties(filename):
log.debug('prop: %s = %s' % (prop, value))
if prop == 'episodeFormat':
upgrade_episode()
break
elif canonical_form(value) == 'DVB':
upgrade_episode()
break
# if no episode info found, assume it's a movie
upgrade_movie()
filetype = filetype_container[0]
return filetype, other
def guess_filetype(filename, filetype):
other = {}
# look at the extension first
fileext = os.path.splitext(filename)[1][1:].lower()
if fileext in subtitle_exts:
if 'movie' in filetype:
filetype = 'moviesubtitle'
elif 'episode' in filetype:
filetype = 'episodesubtitle'
else:
filetype = 'subtitle'
other = {'container': fileext}
elif fileext in video_exts:
if filetype == 'autodetect':
filetype = 'video'
other = {'container': fileext}
else:
if filetype == 'autodetect':
filetype = 'unknown'
other = {'extension': fileext}
# put the filetype inside a dummy container to be able to have the
# following functions work correctly as closures
# this is a workaround for python 2 which doesn't have the
# 'nonlocal' keyword (python 3 does have it)
filetype_container = [filetype]
def upgrade_episode():
if filetype_container[0] == 'video':
filetype_container[0] = 'episode'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'episodesubtitle'
def upgrade_movie():
if filetype_container[0] == 'video':
filetype_container[0] = 'movie'
elif filetype_container[0] == 'subtitle':
filetype_container[0] = 'moviesubtitle'
# now look whether there are some specific hints for episode vs movie
if filetype in ('video', 'subtitle'):
for rexp, _, _ in episode_rexps:
match = re.search(rexp, filename, re.IGNORECASE)
if match:
upgrade_episode()
break
for prop, value, _, _ in find_properties(filename):
log.debug('prop: %s = %s' % (prop, value))
if prop == 'episodeFormat':
upgrade_episode()
break
elif canonical_form(value) == 'DVB':
upgrade_episode()
break
if 'tvu.org.ru' in filename:
upgrade_episode()
# if no episode info found, assume it's a movie
upgrade_movie()
filetype = filetype_container[0]
return filetype, other
def guess_groups(string, result, filetype):
# add sentinels so we can match a separator char at either end of
# our groups, even when they are at the beginning or end of the string
# we will adjust the span accordingly later
#
# filetype can either be movie, moviesubtitle, episode, episodesubtitle
current = " " + string + " "
regions = [] # list of (start, end) of matched regions
def guessed(match_dict, confidence):
guess = format_guess(Guess(match_dict, confidence=confidence))
result.append(guess)
log.debug("Found with confidence %.2f: %s" % (confidence, guess))
return guess
def update_found(string, guess, span, span_adjust=(0, 0)):
span = (span[0] + span_adjust[0], span[1] + span_adjust[1])
regions.append((span, guess))
return blank_region(string, span)
# try to find dates first, as they are very specific
date, span = search_date(current)
if date:
guess = guessed({"date": date}, confidence=1.0)
current = update_found(current, guess, span)
# for non episodes only, look for year information
if filetype not in ("episode", "episodesubtitle"):
year, span = search_year(current)
if year:
guess = guessed({"year": year}, confidence=1.0)
current = update_found(current, guess, span)
# specific regexps (ie: cd number, season X episode, ...)
for rexp, confidence, span_adjust in video_rexps:
match = re.search(rexp, current, re.IGNORECASE)
if match:
metadata = match.groupdict()
# is this the better place to put it? (maybe, as it is at least the soonest that we can catch it)
if "cdNumberTotal" in metadata and metadata["cdNumberTotal"] is None:
del metadata["cdNumberTotal"]
guess = guessed(metadata, confidence=confidence)
current = update_found(current, guess, match.span(), span_adjust)
if filetype in ("episode", "episodesubtitle"):
for rexp, confidence, span_adjust in episode_rexps:
match = re.search(rexp, current, re.IGNORECASE)
if match:
metadata = match.groupdict()
guess = guessed(metadata, confidence=confidence)
current = update_found(current, guess, match.span(), span_adjust)
# Now websites, but as exact string instead of regexps
clow = current.lower()
for site in websites:
pos = clow.find(site.lower())
if pos != -1:
guess = guessed({"website": site}, confidence=confidence)
current = update_found(current, guess, (pos, pos + len(site)))
clow = current.lower()
# release groups have certain constraints, cannot be included in the previous general regexps
group_names = [
r"\.(Xvid)-(?P<releaseGroup>.*?)[ \.]",
r"\.(DivX)-(?P<releaseGroup>.*?)[\. ]",
r"\.(DVDivX)-(?P<releaseGroup>.*?)[\. ]",
]
for rexp in group_names:
match = re.search(rexp, current, re.IGNORECASE)
if match:
metadata = match.groupdict()
metadata.update({"videoCodec": match.group(1)})
guess = guessed(metadata, confidence=0.8)
current = update_found(current, guess, match.span(), span_adjust=(1, -1))
# common well-defined words and regexps
confidence = 1.0 # for all of them
for prop, value, pos, end in find_properties(current):
guess = guessed({prop: value}, confidence=confidence)
current = update_found(current, guess, (pos, end))
# weak guesses for episode number, only run it if we don't have an estimate already
if filetype in ("episode", "episodesubtitle"):
if not any("episodeNumber" in match for match in result):
for rexp, _, span_adjust in weak_episode_rexps:
match = re.search(rexp, current, re.IGNORECASE)
if match:
metadata = match.groupdict()
epnum = int(metadata["episodeNumber"])
if epnum > 100:
guess = guessed({"season": epnum // 100, "episodeNumber": epnum % 100}, confidence=0.6)
else:
guess = guessed(metadata, confidence=0.3)
current = update_found(current, guess, match.span(), span_adjust)
# try to find languages now
language, span, confidence = search_language(current)
while language:
# is it a subtitle language?
if "sub" in clean_string(current[: span[0]]).lower().split(" "):
guess = guessed({"subtitleLanguage": language}, confidence=confidence)
else:
guess = guessed({"language": language}, confidence=confidence)
current = update_found(current, guess, span)
language, span, confidence = search_language(current)
# remove our sentinels now and ajust spans accordingly
assert current[0] == " " and current[-1] == " "
current = current[1:-1]
regions = [((start - 1, end - 1), guess) for (start, end), guess in regions]
# split into '-' separated subgroups (with required separator chars
# around the dash)
didx = current.find("-")
while didx > 0:
regions.append(((didx, didx), None))
didx = current.find("-", didx + 1)
# cut our final groups, and rematch the guesses to the group that created
# id, None if it is a leftover group
region_spans = [span for span, guess in regions]
string_groups = split_on_groups(string, region_spans)
remaining_groups = split_on_groups(current, region_spans)
guesses = []
pos = 0
for group in string_groups:
found = False
for span, guess in regions:
if span[0] == pos:
guesses.append(guess)
found = True
if not found:
guesses.append(None)
pos += len(group)
return zip(string_groups, remaining_groups, guesses)
