def guess_language(string, node=None, options=None):
allowed_languages = None
if options and 'allowed_languages' in options:
allowed_languages = options.get('allowed_languages')
directory = list(filter(lambda x: x.category == 'path', node.ancestors))[0]
if len(directory.clean_value) <= 3:
# skip if we have a langage code as directory
return None
guess = search_language(string, allowed_languages)
return guess
def guess_language(string, node=None, options=None):
allowed_languages = None
if options and 'allowed_languages' in options:
allowed_languages = options.get('allowed_languages')
directory = list(filter(lambda x: x.category == 'path',
node.ancestors))[0]
if len(directory.clean_value) <= 3:
# skip if we have a langage code as directory
return None
guess = search_language(string, allowed_languages)
return guess
def guess_language(string):
language, span, confidence = search_language(string)
if language:
# is it a subtitle language?
if 'sub' in clean_string(string[:span[0]]).lower().split(' '):
return (Guess({'subtitleLanguage': language},
confidence=confidence),
span)
else:
return (Guess({'language': language},
confidence=confidence),
span)
return None, None
def guess_language(string, node, skip=None):
if skip:
relative_skip = []
for entry in skip:
node_idx = entry['node_idx']
span = entry['span']
if node_idx == node.node_idx[:len(node_idx)]:
relative_span = (span[0] - node.offset + 1, span[1] - node.offset + 1)
relative_skip.append(relative_span)
skip = relative_skip
language, span, confidence = search_language(string, skip=skip)
if language:
return (Guess({'language': language},
confidence=confidence,
raw= string[span[0]:span[1]]),
span)
return None, None
def guess_language(string, node, skip=None):
if skip:
relative_skip = []
for entry in skip:
node_idx = entry['node_idx']
span = entry['span']
if node_idx == node.node_idx[:len(node_idx)]:
relative_span = (span[0] - node.offset + 1,
span[1] - node.offset + 1)
relative_skip.append(relative_span)
skip = relative_skip
language, span, confidence = search_language(string, skip=skip)
if language:
return (Guess({'language': language},
confidence=confidence,
raw=string[span[0]:span[1]]), span)
return None, None
def guess_language(string):
guess = search_language(string)
return guess
def guess_language(self, string, node=None, options=None):
allowed_languages = None
if options and 'allowed_languages' in options:
allowed_languages = options.get('allowed_languages')
guess = search_language(string, allowed_languages)
return guess
def guess_language(string):
language, span, confidence = search_language(string)
if language:
return (Guess({'language': language}, confidence=confidence), span)
return None, None
def guess_language(self, string, node=None, options=None):
guess = search_language(string)
return guess
def guess_language(self, string, node=None, options=None):
guess = search_language(string)
return guess
def guess_language(string, node=None, options=None):
allowed_languages = None
if options and 'allowed_languages' in options:
allowed_languages = options.get('allowed_languages')
guess = search_language(string, allowed_languages)
return guess
def guess_language(string):
language, span, confidence = search_language(string)
if language:
return (Guess({"language": language}, confidence=confidence), span)
return None, None
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)