[youtube:channel] Fix extraction (fixes #4435)

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

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index f8d5cdd5d6210477f5d286facfa52921dc5b4186..a7bdce72f1eec8eed16db1159e5a9ddae6ceaf20 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -440,9 +440,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
             'info_dict': {
                 'id': 'lqQg6PlCWgI',
                 'ext': 'mp4',
-                
+                'upload_date': '20120731',
+                'uploader_id': 'olympic',
+                'description': 'HO09  - Women -  GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games',
+                'uploader': 'Olympics',
+                'title': 'Hockey - Women -  GER-AUS - London 2012 Olympic Games',
+            },
+            'params': {
+                'skip_download': 'requires avconv',
             }
-        }
+        },
     ]
 
     def __init__(self, *args, **kwargs):
@@ -866,7 +873,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
 
         m_cat_container = self._search_regex(
             r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>',
-            video_webpage, 'categories', fatal=False)
+            video_webpage, 'categories', default=None)
         if m_cat_container:
             category = self._html_search_regex(
                 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
@@ -1010,9 +1017,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
         # Look for the DASH manifest
         if self._downloader.params.get('youtube_include_dash_manifest', True):
             dash_mpd = video_info.get('dashmpd')
-            if not dash_mpd:
-                self.report_warning('%s: DASH manifest missing' % video_id)
-            else:
+            if dash_mpd:
                 dash_manifest_url = dash_mpd[0]
                 try:
                     dash_formats = self._parse_dash_manifest(
@@ -1264,8 +1269,6 @@ class YoutubeTopListIE(YoutubePlaylistIE):
 class YoutubeChannelIE(InfoExtractor):
     IE_DESC = 'YouTube.com channels'
     _VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P<id>[0-9A-Za-z_-]+)'
-    _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
-    _MORE_PAGES_URL = 'https://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
     IE_NAME = 'youtube:channel'
     _TESTS = [{
         'note': 'paginated channel',
@@ -1302,20 +1305,27 @@ class YoutubeChannelIE(InfoExtractor):
             return self.playlist_result(entries, channel_id)
 
         def _entries():
+            more_widget_html = content_html = channel_page
             for pagenum in itertools.count(1):
-                url = self._MORE_PAGES_URL % (pagenum, channel_id)
-                page = self._download_json(
-                    url, channel_id, note='Downloading page #%s' % pagenum,
-                    transform_source=uppercase_escape)
 
-                ids_in_page = self.extract_videos_from_page(page['content_html'])
+                ids_in_page = self.extract_videos_from_page(content_html)
                 for video_id in ids_in_page:
                     yield self.url_result(
                         video_id, 'Youtube', video_id=video_id)
 
-                if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
+                mobj = re.search(
+                    r'data-uix-load-more-href="/?(?P<more>[^"]+)"',
+                    more_widget_html)
+                if not mobj:
                     break
 
+                more = self._download_json(
+                    'https://youtube.com/%s' % mobj.group('more'), channel_id,
+                    'Downloading page #%s' % (pagenum + 1),
+                    transform_source=uppercase_escape)
+                content_html = more['content_html']
+                more_widget_html = more['load_more_widget_html']
+
         return self.playlist_result(_entries(), channel_id)