[youtube] Improve error detection (#16445)

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

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index b2c714505232e8d93836a4464533746e6b3a1b88..44740cb75c0867d4d4710ffebf23bff044dcb660 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -1700,6 +1700,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         def extract_token(v_info):
             return dict_get(v_info, ('account_playback_token', 'accountPlaybackToken', 'token'))
 
+        def extract_player_response(player_response, video_id):
+            pl_response = str_or_none(player_response)
+            if not pl_response:
+                return
+            pl_response = self._parse_json(pl_response, video_id, fatal=False)
+            if isinstance(pl_response, dict):
+                add_dash_mpd_pr(pl_response)
+                return pl_response
+
         player_response = {}
 
         # Get video info
@@ -1722,7 +1731,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 note='Refetching age-gated info webpage',
                 errnote='unable to download video info webpage')
             video_info = compat_parse_qs(video_info_webpage)
+            pl_response = video_info.get('player_response', [None])[0]
+            player_response = extract_player_response(pl_response, video_id)
             add_dash_mpd(video_info)
+            view_count = extract_view_count(video_info)
         else:
             age_gate = False
             video_info = None
@@ -1745,11 +1757,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     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
+                    player_response = extract_player_response(args.get('player_response'), video_id)
             if not video_info or self._downloader.params.get('youtube_include_dash_manifest', True):
                 add_dash_mpd_pr(player_response)
                 # We also try looking in get_video_info since it may contain different dashmpd
@@ -1781,9 +1789,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     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_pr(player_response)
+                        player_response = extract_player_response(pl_response, video_id)
                     add_dash_mpd(get_video_info)
                     if view_count is None:
                         view_count = extract_view_count(get_video_info)
@@ -1807,7 +1813,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         def extract_unavailable_message():
             return self._html_search_regex(
-                r'(?s)<h1[^>]+id="unavailable-message"[^>]*>(.+?)</h1>',
+                (r'(?s)<div[^>]+id=["\']unavailable-submessage["\'][^>]+>(.+?)</div',
+                 r'(?s)<h1[^>]+id=["\']unavailable-message["\'][^>]*>(.+?)</h1>'),
                 video_webpage, 'unavailable message', default=None)
 
         if not video_info:
@@ -1820,16 +1827,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         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:
+        video_title = video_info.get('title', [None])[0] or video_details.get('title')
+        if not video_title:
             self._downloader.report_warning('Unable to extract video title')
             video_title = '_'
 
-        # description
         description_original = video_description = get_element_by_id("eow-description", video_webpage)
         if video_description:
 
@@ -1854,11 +1856,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             ''', replace_url, video_description)
             video_description = clean_html(video_description)
         else:
-            fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
-            if fd_mobj:
-                video_description = unescapeHTML(fd_mobj.group(1))
-            else:
-                video_description = ''
+            video_description = self._html_search_meta('description', video_webpage) or video_details.get('shortDescription')
 
         if not smuggled_data.get('force_singlefeed', False):
             if not self._downloader.params.get('noplaylist'):
@@ -2101,9 +2099,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     a_format.setdefault('http_headers', {})['Youtubedl-no-compression'] = 'True'
                     formats.append(a_format)
             else:
-                error_message = clean_html(video_info.get('reason', [None])[0])
+                error_message = extract_unavailable_message()
+                if not error_message:
+                    error_message = clean_html(try_get(
+                        player_response, lambda x: x['playabilityStatus']['reason'],
+                        compat_str))
                 if not error_message:
-                    error_message = extract_unavailable_message()
+                    error_message = clean_html(
+                        try_get(video_info, lambda x: x['reason'][0], compat_str))
                 if error_message:
                     raise ExtractorError(error_message, expected=True)
                 raise ExtractorError('no conn, hlsvp, hlsManifestUrl or url_encoded_fmt_stream_map information found in video info')