[youtube] Improve age-gated videos extraction in 429 error conditions (refs #24283)

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

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index d3e18a6adc7737f82e0e840b8fdd16afcfaab8dd..906988e6f0974443cb2552746befe4aa238aff3b 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -1726,6 +1726,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         embed_webpage = None
         if re.search(r'player-age-gate-content">', video_webpage) is not None:
             age_gate = True
+            video_info = None
             # We simulate the access to the video from www.youtube.com/v/{video_id}
             # this can be viewed without login into Youtube
             url = proto + '://www.youtube.com/embed/%s' % video_id
@@ -1737,15 +1738,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     r'"sts"\s*:\s*(\d+)', embed_webpage, 'sts', default=''),
             })
             video_info_url = proto + '://www.youtube.com/get_video_info?' + data
-            video_info_webpage = self._download_webpage(
-                video_info_url, video_id,
-                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)
+            try:
+                video_info_webpage = self._download_webpage(
+                    video_info_url, video_id,
+                    note='Refetching age-gated info webpage',
+                    errnote='unable to download video info webpage')
+            except ExtractorError:
+                video_info_webpage = None
+            if 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
@@ -1790,11 +1795,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                         query['el'] = el
                     if sts:
                         query['sts'] = sts
-                    video_info_webpage = self._download_webpage(
-                        '%s://www.youtube.com/get_video_info' % proto,
-                        video_id, note=False,
-                        errnote='unable to download video info webpage',
-                        fatal=False, query=query)
+                    try:
+                        video_info_webpage = self._download_webpage(
+                            '%s://www.youtube.com/get_video_info' % proto,
+                            video_id, note=False,
+                            errnote='unable to download video info webpage',
+                            query=query)
+                    except ExtractorError as e:
+                        # Skip further retries if we get 429 since solving
+                        # captcha only unblocks access to website but
+                        # not get_video_info end point
+                        if isinstance(e.cause, compat_HTTPError) and e.cause.code == 429:
+                            break
+                        continue
                     if not video_info_webpage:
                         continue
                     get_video_info = compat_parse_qs(video_info_webpage)
@@ -1833,13 +1846,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             if messages:
                 return '\n'.join(messages)
 
-        if not video_info:
+        if not video_info and not player_response:
             unavailable_message = extract_unavailable_message()
             if not unavailable_message:
                 unavailable_message = 'Unable to extract video data'
             raise ExtractorError(
                 'YouTube said: %s' % unavailable_message, expected=True, video_id=video_id)
 
+        if not isinstance(video_info, dict):
+            video_info = {}
+
         video_details = try_get(
             player_response, lambda x: x['videoDetails'], dict) or {}