[youtube] Add fallback metadata extraction from videoDetails (closes #18052)
authorSergey M․ <dstftw@gmail.com>
Fri, 2 Nov 2018 23:26:16 +0000 (06:26 +0700)
committerSergey M․ <dstftw@gmail.com>
Fri, 2 Nov 2018 23:26:16 +0000 (06:26 +0700)
youtube_dl/extractor/youtube.py

index 78203ef84a9efe7fd3046bf025cf871e1d26aead..abadfa5455f95a9270f3a933fd5222d6f9c5bb4a 100644 (file)
@@ -41,6 +41,7 @@ from ..utils import (
     remove_quotes,
     remove_start,
     smuggle_url,
+    str_or_none,
     str_to_int,
     try_get,
     unescapeHTML,
@@ -501,6 +502,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'categories': ['Science & Technology'],
                 'tags': ['youtube-dl'],
                 'duration': 10,
+                'view_count': int,
                 'like_count': int,
                 'dislike_count': int,
                 'start_time': 1,
@@ -583,6 +585,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'categories': ['Science & Technology'],
                 'tags': ['youtube-dl'],
                 'duration': 10,
+                'view_count': int,
                 'like_count': int,
                 'dislike_count': int,
             },
@@ -1538,6 +1541,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         def extract_view_count(v_info):
             return int_or_none(try_get(v_info, lambda x: x['view_count'][0]))
 
+        player_response = {}
+
         # Get video info
         embed_webpage = None
         if re.search(r'player-age-gate-content">', video_webpage) is not None:
@@ -1580,6 +1585,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 if args.get('livestream') == '1' or args.get('live_playback') == 1:
                     is_live = True
                 sts = ytplayer_config.get('sts')
+                if not player_response:
+                    pl_response = str_or_none(args.get('player_response'))
+                    if pl_response:
+                        pl_response = self._parse_json(pl_response, video_id, fatal=False)
+                        if isinstance(pl_response, dict):
+                            player_response = pl_response
             if not video_info or self._downloader.params.get('youtube_include_dash_manifest', True):
                 # We also try looking in get_video_info since it may contain different dashmpd
                 # URL that points to a DASH manifest with possibly different itag set (some itags
@@ -1608,6 +1619,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     if not video_info_webpage:
                         continue
                     get_video_info = compat_parse_qs(video_info_webpage)
+                    if not player_response:
+                        pl_response = get_video_info.get('player_response', [None])[0]
+                        if isinstance(pl_response, dict):
+                            player_response = pl_response
                     add_dash_mpd(get_video_info)
                     if view_count is None:
                         view_count = extract_view_count(get_video_info)
@@ -1653,9 +1668,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     '"token" parameter not in video info for unknown reason',
                     video_id=video_id)
 
+        video_details = try_get(
+            player_response, lambda x: x['videoDetails'], dict) or {}
+
         # title
         if 'title' in video_info:
             video_title = video_info['title'][0]
+        elif 'title' in player_response:
+            video_title = video_details['title']
         else:
             self._downloader.report_warning('Unable to extract video title')
             video_title = '_'
@@ -1718,6 +1738,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         if view_count is None:
             view_count = extract_view_count(video_info)
+        if view_count is None and video_details:
+            view_count = int_or_none(video_details.get('viewCount'))
 
         # Check for "rental" videos
         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
@@ -1898,7 +1920,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
 
         # uploader
-        video_uploader = try_get(video_info, lambda x: x['author'][0], compat_str)
+        video_uploader = try_get(
+            video_info, lambda x: x['author'][0],
+            compat_str) or str_or_none(video_details.get('author'))
         if video_uploader:
             video_uploader = compat_urllib_parse_unquote_plus(video_uploader)
         else:
@@ -2011,12 +2035,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         like_count = _extract_count('like')
         dislike_count = _extract_count('dislike')
 
+        if view_count is None:
+            view_count = str_to_int(self._search_regex(
+                r'<[^>]+class=["\']watch-view-count[^>]+>\s*([\d,\s]+)', video_webpage,
+                'view count', default=None))
+
         # subtitles
         video_subtitles = self.extract_subtitles(video_id, video_webpage)
         automatic_captions = self.extract_automatic_captions(video_id, video_webpage)
 
         video_duration = try_get(
             video_info, lambda x: int_or_none(x['length_seconds'][0]))
+        if not video_duration:
+            video_duration = int_or_none(video_details.get('lengthSeconds'))
         if not video_duration:
             video_duration = parse_duration(self._html_search_meta(
                 'duration', video_webpage, 'video duration'))
@@ -2244,6 +2275,7 @@ class YoutubePlaylistIE(YoutubePlaylistBaseInfoExtractor):
             'description': 'md5:507cdcb5a49ac0da37a920ece610be80',
             'categories': ['People & Blogs'],
             'tags': list,
+            'view_count': int,
             'like_count': int,
             'dislike_count': int,
         },