[ndr] Fix thumbnail extraction
authorSergey M․ <dstftw@gmail.com>
Sat, 17 May 2014 10:58:37 +0000 (17:58 +0700)
committerSergey M․ <dstftw@gmail.com>
Sat, 17 May 2014 10:58:37 +0000 (17:58 +0700)
youtube_dl/extractor/ndr.py

index 0650f956481c9011032a278fc1a9375b98e26539..0905d7bb84236b64384c564cc7260a12e8a3e8c8 100644 (file)
@@ -66,10 +66,12 @@ class NDRIE(InfoExtractor):
 
         video_url = re.search(r'''3: {src:'(?P<video>.+?)\.hi\.mp4', type:"video/mp4"},''', page)
         if video_url:
-            thumbnail = self._html_search_regex(r'(?m)title: "NDR PLAYER",\s*poster: "([^"]+)",',
-                page, 'thumbnail', fatal=False)
-            if thumbnail:
-                thumbnail = 'http://www.ndr.de' + thumbnail
+            thumbnails = re.findall(r'''\d+: {src: "([^"]+)"(?: \|\| '[^']+')?, quality: '([^']+)'}''', page)
+            if thumbnails:
+                QUALITIES = ['xs', 's', 'm', 'l', 'xl']
+                thumbnails.sort(key=lambda thumb: QUALITIES.index(thumb[1]))
+                thumbnail = 'http://www.ndr.de' + thumbnails[-1][0]
+
             for format_id in ['lo', 'hi', 'hq']:
                 formats.append({
                     'url': '%s.%s.mp4' % (video_url.group('video'), format_id),