[utils] Add support for quoted string literals in --match-filter (closes #8050, closes #12142, closes #12144)

test/test_YoutubeDL.py

@@ -1,4 +1,5 @@
 #!/usr/bin/env python
+# coding: utf-8
 
 from __future__ import unicode_literals
 
@@ -606,6 +607,8 @@ def _match_entry(self, info_dict, incomplete):
             'duration': 30,
             'filesize': 10 * 1024,
             'playlist_id': '42',
+            'uploader': "變態妍字幕版 太妍 тест",
+            'creator': "тест ' 123 ' тест--",
         }
         second = {
             'id': '2',
@@ -616,6 +619,7 @@ def _match_entry(self, info_dict, incomplete):
             'description': 'foo',
             'filesize': 5 * 1024,
             'playlist_id': '43',
+            'uploader': "тест 123",
         }
         videos = [first, second]
 
@@ -656,6 +660,26 @@ def f(v):
         res = get_videos(f)
         self.assertEqual(res, ['1'])
 
+        f = match_filter_func('uploader = "變態妍字幕版 太妍 тест"')
+        res = get_videos(f)
+        self.assertEqual(res, ['1'])
+
+        f = match_filter_func('uploader != "變態妍字幕版 太妍 тест"')
+        res = get_videos(f)
+        self.assertEqual(res, ['2'])
+
+        f = match_filter_func('creator = "тест \' 123 \' тест--"')
+        res = get_videos(f)
+        self.assertEqual(res, ['1'])
+
+        f = match_filter_func("creator = 'тест \\' 123 \\' тест--'")
+        res = get_videos(f)
+        self.assertEqual(res, ['1'])
+
+        f = match_filter_func(r"creator = 'тест \' 123 \' тест--' & duration > 30")
+        res = get_videos(f)
+        self.assertEqual(res, [])
+
     def test_playlist_items_selection(self):
         entries = [{
             'id': compat_str(i),

youtube_dl/utils.py

@@ -2383,6 +2383,7 @@ def _match_one(filter_part, dct):
         \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
         (?:
             (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
+            (?P<quote>["\'])(?P<quotedstrval>(?:\\.|(?!(?P=quote)|\\).)+?)(?P=quote)|
             (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
         )
         \s*$
@@ -2391,7 +2392,8 @@ def _match_one(filter_part, dct):
     if m:
         op = COMPARISON_OPERATORS[m.group('op')]
         actual_value = dct.get(m.group('key'))
-        if (m.group('strval') is not None or
+        if (m.group('quotedstrval') is not None or
+            m.group('strval') is not None or
             # If the original field is a string and matching comparisonvalue is
             # a number we should respect the origin of the original field
             # and process comparison value as a string (see
@@ -2401,7 +2403,10 @@ def _match_one(filter_part, dct):
             if m.group('op') not in ('=', '!='):
                 raise ValueError(
                     'Operator %s does not support string values!' % m.group('op'))
-            comparison_value = m.group('strval') or m.group('intval')
+            comparison_value = m.group('quotedstrval') or m.group('strval') or m.group('intval')
+            quote = m.group('quote')
+            if quote is not None:
+                comparison_value = comparison_value.replace(r'\%s' % quote, quote)
         else:
             try:
                 comparison_value = int(m.group('intval'))