From: Sergey M․ Date: Fri, 15 May 2015 15:42:34 +0000 (+0600) Subject: [youtube:feed] Check each 'load more' portion for unique video ids X-Git-Url: http://git.bitcoin.ninja/?a=commitdiff_plain;h=62c95fd5fcb8dbea2faeb4edac4c5177cbac5912;p=youtube-dl [youtube:feed] Check each 'load more' portion for unique video ids --- diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 9096a2975..1f9940cf5 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1621,10 +1621,16 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): # for the video ids doesn't contain an index ids = [] more_widget_html = content_html = page - for page_num in itertools.count(1): matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html) - new_ids = orderedSet(matches) + + # 'recommended' feed has infinite 'load more' and each new portion spins + # the same videos in (sometimes) slightly different order, so we'll check + # for unicity and break when portion has no new videos + new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches)) + if not new_ids: + break + ids.extend(new_ids) mobj = re.search(r'data-uix-load-more-href="/?(?P[^"]+)"', more_widget_html)