[vimeo] Improve regex for the config js dict (fixes #3955, fixes #3974)
authorJaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@gmail.com>
Fri, 17 Oct 2014 13:49:16 +0000 (15:49 +0200)
committerJaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@gmail.com>
Fri, 17 Oct 2014 13:55:58 +0000 (15:55 +0200)
The javascript code contains assignments of empty dicts with the same variable name

youtube_dl/extractor/vimeo.py

index e6a86f18ec5018615344e547cd332892c81fc665..07959d3fe0e4b75a0b58d3605560c2ed450e9ba6 100644 (file)
@@ -157,6 +157,18 @@ class VimeoIE(VimeoBaseInfoExtractor, SubtitlesInfoExtractor):
                 'duration': 62,
             }
         },
+        {
+            # from https://www.ouya.tv/game/Pier-Solar-and-the-Great-Architects/
+            'url': 'https://player.vimeo.com/video/98044508',
+            'note': 'The js code contains assignments to the same variable as the config',
+            'info_dict': {
+                'id': '98044508',
+                'ext': 'mp4',
+                'title': 'Pier Solar OUYA Official Trailer',
+                'uploader': 'Tulio Gonçalves',
+                'uploader_id': 'user28849593',
+            },
+        },
     ]
 
     def _verify_video_password(self, url, video_id, webpage):
@@ -244,7 +256,7 @@ class VimeoIE(VimeoBaseInfoExtractor, SubtitlesInfoExtractor):
                 # We try to find out to which variable is assigned the config dic
                 m_variable_name = re.search('(\w)\.video\.id', webpage)
                 if m_variable_name is not None:
-                    config_re = r'%s=({.+?});' % re.escape(m_variable_name.group(1))
+                    config_re = r'%s=({[^}].+?});' % re.escape(m_variable_name.group(1))
                 else:
                     config_re = [r' = {config:({.+?}),assets:', r'(?:[abc])=({.+?});']
                 config = self._search_regex(config_re, webpage, 'info section',