def startEpisode(self, episode):
self.tmdb.lang = guiLanguage().alpha2
if episode.get('series') is None:
raise SmewtException("TVDBMetadataProvider: Episode doesn't contain 'series' field: %s", episode)
name = episode.series.title
name = name.replace(',', ' ')
matching_series = self.getSeries(name)
# Try first with the languages from guessit, and then with english
languages = tolist(episode.get('language', [])) + ['en']
# Sort the series by id (stupid heuristic about most popular series
# might have been added sooner to the db and the db id
# follows the insertion order)
# TODO: we should do something smarter like comparing series name distance,
# episodes count and/or episodes names
#print '\n'.join(['%s %s --> %f [%s] %s' % (x[1], name, textutils.levenshtein(x[1], name), x[2], x[0]) for x in matching_series])
matching_series.sort(key=lambda x: (textutils.levenshtein(x[1], name), int(x[0])))
series = None
language = 'en'
for lang in languages:
try:
language = lang
ind = zip(*matching_series)[2].index(lang)
series = matching_series[ind][0]
break
except ValueError, e:
language = matching_series[0][2]
series = matching_series[0][0]
def fuzzyMatch2(baseGuess, md):
for p1, p2 in zip(baseGuess.unique_key(), md.unique_key()):
if type(p1) == str or type(p1) == unicode:
# TODO: levenshtein doesn't cut it here, we need a better string distance
if levenshtein(p1.lower(), p2.lower()) > 80:
return False
elif isinstance(p1, Metadata):
if not fuzzyMatch2(p1, p2):
return False
else:
if p1 != p2:
return False
return True
def fuzzyMatch2(baseGuess, md):
for p1, p2 in zip(baseGuess.unique_key(), md.unique_key()):
if type(p1) == str or type(p1) == unicode:
# TODO: levenshtein doesn't cut it here, we need a better string distance
if levenshtein(p1.lower(), p2.lower()) > 80:
return False
elif isinstance(p1, Metadata):
if not fuzzyMatch2(p1, p2):
return False
else:
if p1 != p2:
return False
return True
def startEpisode(self, episode):
self.tmdb.lang = guiLanguage().alpha2
if episode.get('series') is None:
raise SmewtException(
"TVDBMetadataProvider: Episode doesn't contain 'series' field: %s",
episode)
name = episode.series.title
name = name.replace(',', ' ')
matching_series = self.getSeries(name)
# Try first with the languages from guessit, and then with english
languages = tolist(episode.get('language', [])) + ['en']
# Sort the series by id (stupid heuristic about most popular series
# might have been added sooner to the db and the db id
# follows the insertion order)
# TODO: we should do something smarter like comparing series name distance,
# episodes count and/or episodes names
#print '\n'.join(['%s %s --> %f [%s] %s' % (x[1], name, textutils.levenshtein(x[1], name), x[2], x[0]) for x in matching_series])
matching_series.sort(
key=lambda x: (textutils.levenshtein(x[1], name), int(x[0])))
series = None
language = 'en'
for lang in languages:
try:
language = lang
ind = zip(*matching_series)[2].index(lang)
series = matching_series[ind][0]
break
except ValueError, e:
language = matching_series[0][2]
series = matching_series[0][0]