From b9643eed7c9b081c91e72257b380ccbd92555254 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jaime=20Marqui=CC=81nez=20Ferra=CC=81ndiz?= Date: Fri, 15 Nov 2013 11:51:45 +0100 Subject: [PATCH] [youtube:channel] Fix the extraction of autogenerated channels The ajax pages are empty, now it looks directly in the channel's /videos page --- youtube_dl/extractor/youtube.py | 37 +++++++++++++++++++++------------ 1 file changed, 24 insertions(+), 13 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 7d7aeb461..8c0e6f252 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1594,20 +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)))