[youtube] fix: playlist

youtube_dlc/extractor/youtube.py

@@ -2891,6 +2891,77 @@ def _extract_playlist(self, playlist_id):
         url = self._TEMPLATE_URL % playlist_id
         page = self._download_webpage(url, playlist_id)
 
+        yt_initial = self._get_yt_initial_data('', page)
+        if yt_initial:
+            playlist_items = try_get(yt_initial, lambda x: x['contents']['twoColumnBrowseResultsRenderer']['tabs'][0]['tabRenderer']['content']['sectionListRenderer']['contents'][0]['itemSectionRenderer']['contents'][0]['playlistVideoListRenderer']['contents'], list)
+            video_ids = []
+            entries = []
+            playlist_page = 1
+            api_key = self._search_regex(
+                r'"INNERTUBE_API_KEY":"([^"]+)"',
+                page, 'api key', default="AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8", fatal=False)
+            api_client_version = self._search_regex(
+                r'"INNERTUBE_CONTEXT_CLIENT_VERSION":"([^"]+)"',
+                page, 'client version', fatal=False)
+            while playlist_items:
+                item = playlist_items.pop(0)
+
+                item_video = try_get(item, lambda x: x['playlistVideoRenderer'], dict)
+                if item_video:
+                    video_id = try_get(item_video, lambda x: x['videoId'], compat_str)
+                    if video_id in video_ids:
+                        continue
+                    else:
+                        video_ids.append(video_id)
+                    entry = {
+                        '_type': 'url',
+                        'duration': int_or_none(try_get(item_video, lambda x: x['lengthSeconds'], compat_str)),
+                        'id': video_id,
+                        'ie_key': 'Youtube',
+                        # 'thumbnails': try_get(item_video, lambda x: x['thumbnail']['thumbnails'], list),
+                        'title': try_get(item_video, lambda x: x['title']['runs'][0]['text'], compat_str),
+                        'url': video_id
+                    }
+                    entries.append(entry)
+
+                item_continue = try_get(item, lambda x: x['continuationItemRenderer'], dict)
+                if item_continue:
+                    playlist_page += 1
+                    continuation_token = try_get(item_continue, lambda x: x['continuationEndpoint']['continuationCommand']['token'], compat_str)
+                    request_data = {
+                        'context': {
+                            'client': {
+                                'clientName': 'WEB',
+                                'clientVersion': api_client_version
+                            }
+                        },
+                        'continuation': continuation_token
+                    }
+                    response = self._download_json(
+                        'https://www.youtube.com/youtubei/v1/browse?key=%s' % api_key,
+                        data=json.dumps(request_data).encode('utf8'),
+                        errnote='Unable to download playlist page', fatal=False,
+                        headers={'Content-Type': 'application/json'},
+                        note='Downloading page %s' % playlist_page,
+                        video_id=playlist_id)
+                    playlist_items_new = try_get(response, lambda x: x['onResponseReceivedActions'][0]['appendContinuationItemsAction']['continuationItems'], list)
+                    if playlist_items_new:
+                        # load more pages until we get a page of all videos already in the playlist (some playlists loop)
+                        video_ids_new = [try_get(i, lambda x: x['playlistVideoRenderer']['videoId'], compat_str) for i in playlist_items_new]      
+                        video_ids_new = [i for i in video_ids_new if i and i not in video_ids]
+                        if video_ids_new:
+                            playlist_items.extend(playlist_items_new)
+
+            playlist_title = try_get(yt_initial, lambda x: x['microformat']['microformatDataRenderer']['title'], compat_str)
+            playlist_description = try_get(yt_initial, lambda x: x['microformat']['microformatDataRenderer']['description'], compat_str)
+            playlist = self.playlist_result(
+                entries,
+                playlist_id=playlist_id,
+                playlist_title=playlist_title,
+                playlist_description=playlist_description)
+            has_videos = bool(entries)
+            return has_videos, playlist
+
         # the yt-alert-message now has tabindex attribute (see https://github.com/ytdl-org/youtube-dl/issues/11604)
         for match in re.findall(r'<div class="yt-alert-message"[^>]*>([^<]+)</div>', page):
             match = match.strip()