Strip separators from matches

This is useful in advanced mode for raw values like title, episode_title and release_group to have separators stripped off.

guessit/rules/common/__init__.py

@@ -6,6 +6,7 @@
 import re
 
 seps = r' [](){}+*|=-_~#/\\.,;:'  # list of tags/words separators
+seps_no_groups = seps.replace('[](){}', '')
 seps_no_fs = seps.replace('/', '').replace('\\', '')
 
 title_seps = r'-+/\|'  # separators for title

guessit/rules/processors.py

@@ -9,6 +9,8 @@
 import six
 
 from rebulk import Rebulk, Rule, CustomRule, POST_PROCESS, PRE_PROCESS, AppendMatch, RemoveMatch
+
+from guessit.rules.common import seps_no_groups
 from guessit.rules.common.words import iter_words
 from .common.formatters import cleanup
 from .common.comparators import marker_sorted
@@ -205,11 +207,31 @@ def then(self, matches, when_response, context):  # pragma: no cover
         pass
 
 
+class StripSeparators(CustomRule):
+    """
+    Strip separators from matches. Keep separators if they are from acronyms, like in ".S.H.I.E.L.D."
+    """
+    priority = POST_PROCESS
+
+    def when(self, matches, context):
+        return matches
+
+    def then(self, matches, when_response, context):  # pragma: no cover
+        for match in matches:
+            for _ in range(0, len(match.span)):
+                if match.raw[0] in seps_no_groups and (len(match.raw) < 3 or match.raw[2] not in seps_no_groups):
+                    match.raw_start += 1
+
+            for _ in reversed(range(0, len(match.span))):
+                if match.raw[-1] in seps_no_groups and (len(match.raw) < 3 or match.raw[-3] not in seps_no_groups):
+                    match.raw_end -= 1
+
+
 def processors():
     """
     Builder for rebulk object.
     :return: Created Rebulk object
     :rtype: Rebulk
     """
     return Rebulk().rules(EnlargeGroupMatches, EquivalentHoles, RemoveLessSpecificSeasonEpisode,
-                          RemoveAmbiguous, SeasonYear, Processors)
+                          RemoveAmbiguous, SeasonYear, Processors, StripSeparators)

guessit/rules/properties/container.py

@@ -6,6 +6,8 @@
 from rebulk.remodule import re
 
 from rebulk import Rebulk
+
+from guessit.rules.common import seps
 from ..common.validators import seps_surround
 from ...reutils import build_or_pattern
 
@@ -18,7 +20,7 @@ def container():
     """
     rebulk = Rebulk().regex_defaults(flags=re.IGNORECASE).string_defaults(ignore_case=True)
     rebulk.defaults(name='container',
-                    formatter=lambda value: value[1:],
+                    formatter=lambda value: value.strip(seps),
                     tags=['extension'],
                     conflict_solver=lambda match, other: other
                     if other.name in ['format', 'video_codec'] or

guessit/test/rules/processors_test.py

@@ -0,0 +1,46 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# pylint: disable=no-self-use, pointless-statement, missing-docstring, invalid-name, pointless-string-statement
+
+from rebulk.match import Matches, Match
+
+from guessit.rules.processors import StripSeparators
+
+
+def test_strip_separators():
+    strip_separators = StripSeparators()
+
+    matches = Matches()
+
+    m = Match(3, 11, input_string="pre.ABCDEF.post")
+
+    assert m.raw == '.ABCDEF.'
+    matches.append(m)
+
+    returned_matches = strip_separators.when(matches, None)
+    assert returned_matches == matches
+
+    strip_separators.then(matches, returned_matches, None)
+
+    assert m.raw == 'ABCDEF'
+
+
+def test_strip_separators_keep_acronyms():
+    strip_separators = StripSeparators()
+
+    matches = Matches()
+
+    m = Match(0, 13, input_string=".S.H.I.E.L.D.")
+    m2 = Match(0, 22, input_string=".Agent.Of.S.H.I.E.L.D.")
+
+    assert m.raw == '.S.H.I.E.L.D.'
+    matches.append(m)
+    matches.append(m2)
+
+    returned_matches = strip_separators.when(matches, None)
+    assert returned_matches == matches
+
+    strip_separators.then(matches, returned_matches, None)
+
+    assert m.raw == '.S.H.I.E.L.D.'
+    assert m2.raw == 'Agent.Of.S.H.I.E.L.D.'