Update guessit with unicode fix

libs/guessit/transfo/__init__.py

@@ -45,7 +45,7 @@ def format_guess(guess):
         elif isinstance(value, base_text_type):
             if prop in ('edition',):
                 value = clean_string(value)
-            guess[prop] = canonical_form(value)
+            guess[prop] = canonical_form(value).replace('\\', '')
 
     return guess
 

libs/guessit/transfo/guess_country.py

@@ -19,24 +19,30 @@
 #
 
 from __future__ import unicode_literals
-#from guessit.transfo import SingleNodeGuesser
-#from guessit.date import search_year
 from guessit.country import Country
 from guessit import Guess
 import logging
 
 log = logging.getLogger(__name__)
 
+# list of common words which could be interpreted as countries, but which
+# are far too common to be able to say they represent a country
+country_common_words = frozenset([ 'bt', 'bb' ])
 
 def process(mtree):
     for node in mtree.unidentified_leaves():
-        # only keep explicit groups (enclosed in parentheses/brackets)
         if len(node.node_idx) == 2:
-            try:
-                country = Country(node.value[1:-1], strict=True)
-                if node.value[0] + node.value[-1] not in ['()', '[]', '{}']:
-                    continue
-                node.guess = Guess(country=country, confidence=1.0)
+            c = node.value[1:-1].lower()
+            if c in country_common_words:
+                continue
 
+            # only keep explicit groups (enclosed in parentheses/brackets)
+            if node.value[0] + node.value[-1] not in ['()', '[]', '{}']:
+                continue
+
+            try:
+                country = Country(c, strict=True)
             except ValueError:
-                pass
+                continue
+
+            node.guess = Guess(country=country, confidence=1.0)

libs/guessit/transfo/guess_filetype.py

@@ -21,7 +21,7 @@
 from __future__ import unicode_literals
 from guessit import Guess
 from guessit.patterns import (subtitle_exts, video_exts, episode_rexps,
-                              find_properties, canonical_form)
+                              find_properties, compute_canonical_form)
 from guessit.date import valid_year
 from guessit.textutils import clean_string
 import os.path
@@ -89,7 +89,7 @@ def guess_filetype(mtree, filetype):
 
     # check whether we are in a 'Movies', 'Tv Shows', ... folder
     folder_rexps = [ (r'Movies?', upgrade_movie),
-                     (r'Tv ?Shows?', upgrade_episode),
+                     (r'Tv[ _-]?Shows?', upgrade_episode),
                      (r'Series', upgrade_episode)
                      ]
     for frexp, upgrade_func in folder_rexps:
@@ -142,7 +142,7 @@ def guess_filetype(mtree, filetype):
                 upgrade_episode()
                 break
 
-            elif canonical_form(value) == 'DVB':
+            elif compute_canonical_form('format', value) == 'DVB':
                 upgrade_episode()
                 break
 

libs/guessit/transfo/guess_language.py

@@ -22,7 +22,7 @@ from __future__ import unicode_literals
 from guessit import Guess
 from guessit.transfo import SingleNodeGuesser
 from guessit.language import search_language
-from guessit.textutils import clean_string
+from guessit.textutils import clean_string, find_words
 import logging
 
 log = logging.getLogger(__name__)
@@ -31,18 +31,13 @@ log = logging.getLogger(__name__)
 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 (Guess({'language': language},
+                      confidence=confidence),
+                span)
 
     return None, None
 
 
 def process(mtree):
     SingleNodeGuesser(guess_language, None, log).process(mtree)
+    # Note: 'language' is promoted to 'subtitleLanguage' in the post_process transfo

libs/guessit/transfo/guess_movie_title_from_position.py

@@ -20,6 +20,7 @@
 
 from __future__ import unicode_literals
 from guessit import Guess
+import unicodedata
 import logging
 
 log = logging.getLogger(__name__)

libs/guessit/transfo/guess_release_group.py

@@ -20,49 +20,51 @@
 
 from __future__ import unicode_literals
 from guessit.transfo import SingleNodeGuesser
-from guessit.patterns import properties, canonical_form
+from guessit.patterns import prop_multi, compute_canonical_form, _dash, _psep
 import re
 import logging
 
 log = logging.getLogger(__name__)
 
+def get_patterns(property_name):
+    return [ p.replace(_dash, _psep) for patterns in prop_multi[property_name].values() for p in patterns  ]
 
-CODECS = properties['videoCodec']
-FORMATS = properties['format']
+CODECS = get_patterns('videoCodec')
+FORMATS = get_patterns('format')
+
+GROUP_NAMES = [ r'(?P<videoCodec>' + codec + r')-?(?P<releaseGroup>.*?)[ \.]'
+                for codec in CODECS ]
+GROUP_NAMES += [ r'(?P<format>' + fmt + r')-?(?P<releaseGroup>.*?)[ \.]'
+                 for fmt in FORMATS ]
+
+GROUP_NAMES2 = [ r'\.(?P<videoCodec>' + codec + r')-(?P<releaseGroup>.*?)(-(.*?))?[ \.]'
+                 for codec in CODECS ]
+GROUP_NAMES2 += [ r'\.(?P<format>' + fmt + r')-(?P<releaseGroup>.*?)(-(.*?))?[ \.]'
+                  for fmt in FORMATS ]
+
+GROUP_NAMES = [ re.compile(r, re.IGNORECASE) for r in GROUP_NAMES ]
+GROUP_NAMES2 = [ re.compile(r, re.IGNORECASE) for r in GROUP_NAMES2 ]
 
 def adjust_metadata(md):
-    codec = canonical_form(md['videoCodec'])
-    if codec in FORMATS:
-        md['format'] = codec
-        del md['videoCodec']
-    return md
+    return dict((property_name, compute_canonical_form(property_name, value) or value)
+                for property_name, value in md.items())
 
 
 def guess_release_group(string):
-    group_names = [ r'\.(Xvid)-(?P<releaseGroup>.*?)[ \.]',
-                    r'\.(DivX)-(?P<releaseGroup>.*?)[\. ]',
-                    r'\.(DVDivX)-(?P<releaseGroup>.*?)[\. ]',
-                    ]
-
     # first try to see whether we have both a known codec and a known release group
-    group_names = [ r'\.(?P<videoCodec>' + codec + r')-(?P<releaseGroup>.*?)[ \.]'
-                    for codec in (CODECS + FORMATS) ]
-
-    for rexp in group_names:
-        match = re.search(rexp, string, re.IGNORECASE)
+    for rexp in GROUP_NAMES:
+        match = rexp.search(string)
         if match:
             metadata = match.groupdict()
-            if canonical_form(metadata['releaseGroup']) in properties['releaseGroup']:
+            release_group = compute_canonical_form('releaseGroup', metadata['releaseGroup'])
+            if release_group:
                 return adjust_metadata(metadata), (match.start(1), match.end(2))
 
     # pick anything as releaseGroup as long as we have a codec in front
     # this doesn't include a potential dash ('-') ending the release group
     # eg: [...].X264-HiS@SiLUHD-English.[...]
-    group_names = [ r'\.(?P<videoCodec>' + codec + r')-(?P<releaseGroup>.*?)(-(.*?))?[ \.]'
-                    for codec in (CODECS + FORMATS) ]
-
-    for rexp in group_names:
-        match = re.search(rexp, string, re.IGNORECASE)
+    for rexp in GROUP_NAMES2:
+        match = rexp.search(string)
         if match:
             return adjust_metadata(match.groupdict()), (match.start(1), match.end(2))
 

libs/guessit/transfo/post_process.py

@@ -20,6 +20,7 @@
 
 from __future__ import unicode_literals
 from guessit.patterns import subtitle_exts
+from guessit.textutils import reorder_title, find_words
 import logging
 
 log = logging.getLogger(__name__)
@@ -45,6 +46,15 @@ def process(mtree):
             node == mtree.leaves()[-2]):
             promote_subtitle()
 
+        # - if we find the word 'sub' before the language, and in the same explicit
+        #   group, then upgrade the language
+        explicit_group = mtree.node_at(node.node_idx[:2])
+        group_str = explicit_group.value.lower()
+
+        if ('sub' in find_words(group_str) and
+            0 <= group_str.find('sub') < (node.span[0] - explicit_group.span[0])):
+            promote_subtitle()
+
         # - if a language is in an explicit group just preceded by "st",
         #   it is a subtitle language (eg: '...st[fr-eng]...')
         try:
@@ -60,11 +70,4 @@ def process(mtree):
         if 'series' not in node.guess:
             continue
 
-        series = node.guess['series']
-        lseries = series.lower()
-
-        if lseries[-4:] == ',the':
-            node.guess['series'] = 'The ' + series[:-4]
-
-        if lseries[-5:] == ', the':
-            node.guess['series'] = 'The ' + series[:-5]
+        node.guess['series'] = reorder_title(node.guess['series'])

libs/guessit/transfo/split_on_dash.py

@@ -38,15 +38,5 @@ def process(mtree):
             indices.extend([ span[0], span[1] ])
             match = pattern.search(node.value, span[1])
 
-        didx = node.value.find('-')
-        while didx > 0:
-            if (didx > 10 and
-                (didx - 1 not in indices and
-                 didx + 2 not in indices)):
-
-                indices.extend([ didx, didx + 1 ])
-
-            didx = node.value.find('-', didx + 1)
-
         if indices:
             node.partition(indices)