[youtube:channel] Fix the extraction of autogenerated channels

[youtube-dl] / youtube_dl / extractor / youtube.py

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index c992cba978441081a50e5f134a4a694c83a6685c..8c0e6f2520e02e5d34a24f01a20637ea4a90a6f9 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -1019,6 +1019,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
         """Turn the encrypted s field into a working signature"""
 
         if player_url is not None:
+            if player_url.startswith(u'//'):
+                player_url = u'https:' + player_url
             try:
                 player_id = (player_url, len(s))
                 if player_id not in self._player_cache:
@@ -1098,7 +1100,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
             params = compat_urllib_parse.urlencode({
                 'lang': lang,
                 'v': video_id,
-                'fmt': self._downloader.params.get('subtitlesformat'),
+                'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
                 'name': l[0].encode('utf-8'),
             })
             url = u'http://www.youtube.com/api/timedtext?' + params
@@ -1111,7 +1113,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
     def _get_available_automatic_caption(self, video_id, webpage):
         """We need the webpage for getting the captions url, pass it as an
            argument to speed up the process."""
-        sub_format = self._downloader.params.get('subtitlesformat')
+        sub_format = self._downloader.params.get('subtitlesformat', 'srt')
         self.to_screen(u'%s: Looking for automatic captions' % video_id)
         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
         err_msg = u'Couldn\'t find automatic captions for %s' % video_id
@@ -1592,20 +1594,31 @@ class YoutubeChannelIE(InfoExtractor):
         # Download channel page
         channel_id = mobj.group(1)
         video_ids = []
+        url = 'https://www.youtube.com/channel/%s/videos' % channel_id
+        channel_page = self._download_webpage(url, channel_id)
+        if re.search(r'channel-header-autogenerated-label', channel_page) is not None:
+            autogenerated = True
+        else:
+            autogenerated = False
 
-        # Download all channel pages using the json-based channel_ajax query
-        for pagenum in itertools.count(1):
-            url = self._MORE_PAGES_URL % (pagenum, channel_id)
-            page = self._download_webpage(url, channel_id,
-                                          u'Downloading page #%s' % pagenum)
-
-            page = json.loads(page)
-
-            ids_in_page = self.extract_videos_from_page(page['content_html'])
-            video_ids.extend(ids_in_page)
-
-            if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
-                break
+        if autogenerated:
+            # The videos are contained in a single page
+            # the ajax pages can't be used, they are empty
+            video_ids = self.extract_videos_from_page(channel_page)
+        else:
+            # Download all channel pages using the json-based channel_ajax query
+            for pagenum in itertools.count(1):
+                url = self._MORE_PAGES_URL % (pagenum, channel_id)
+                page = self._download_webpage(url, channel_id,
+                                              u'Downloading page #%s' % pagenum)
+    
+                page = json.loads(page)
+    
+                ids_in_page = self.extract_videos_from_page(page['content_html'])
+                video_ids.extend(ids_in_page)
+    
+                if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
+                    break
 
         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))