Merge 97c1053ba81c29cb34066e9ce3fd5c353d1fcf84 into 673277e510ebd996b62a2fcc76169bf3cce29910

* [YouTube] Extract playlist items from LOCKUP_VIEW_MODEL_...
* resolves #33073
* thx seproDev (yt-dlp/yt-dlp#11615)

Co-authored-by: sepro <sepro@sepr0.com>

youtube_dl/extractor/bbc.py

@@ -25,6 +25,7 @@ from ..utils import (
     get_element_by_class,
     int_or_none,
     js_to_json,
+    parse_bitrate,
     parse_duration,
     parse_iso8601,
     strip_or_none,
@@ -647,9 +648,7 @@ class BBCIE(BBCCoUkIE):
             'skip_download': True,
         }
     }, {
-        # article with single video embedded with data-playable containing XML playlist
-        # with direct video links as progressiveDownloadUrl (for now these are extracted)
-        # and playlist with f4m and m3u8 as streamingUrl
+        # article with single video (formerly) embedded, now using SIMORGH_DATA JSON
         'url': 'http://www.bbc.com/turkce/haberler/2015/06/150615_telabyad_kentin_cogu',
         'info_dict': {
             'id': '150615_telabyad_kentin_cogu',
@@ -661,12 +660,13 @@ class BBCIE(BBCCoUkIE):
         },
         'params': {
             'skip_download': True,
-        }
+        },
+        'skip': 'Video no longer embedded, 2021',
     }, {
-        # single video embedded with data-playable containing XML playlists (regional section)
+        # single video embedded, legacy media, in promo object of SIMORGH_DATA JSON
         'url': 'http://www.bbc.com/mundo/video_fotos/2015/06/150619_video_honduras_militares_hospitales_corrupcion_aw',
         'info_dict': {
-            'id': '150619_video_honduras_militares_hospitales_corrupcion_aw',
+            'id': '39275083',
             'ext': 'mp4',
             'title': 'Honduras militariza sus hospitales por nuevo escándalo de corrupción',
             'description': 'md5:1525f17448c4ee262b64b8f0c9ce66c8',
@@ -845,6 +845,18 @@ class BBCIE(BBCCoUkIE):
             'upload_date': '20190604',
             'categories': ['Psychology'],
         },
+    }, {
+        # BBC World Service etc: media nested in content object of SIMORGH_DATA JSON
+        'url': 'http://www.bbc.co.uk/scotland/articles/cm49v4x1r9lo',
+        'info_dict': {
+            'id': 'p06p040v',
+            'ext': 'mp4',
+            'title': 'Five things ants can teach us about management',
+            'description': 'They may be tiny, but us humans could learn a thing or two from ants.',
+            'duration': 191,
+            'thumbnail': r're:https?://.+/p06p0qzv.jpg',
+            'upload_date': '20181016',
+        },
     }]
 
     @classmethod
@@ -1107,6 +1119,99 @@ class BBCIE(BBCCoUkIE):
                     'subtitles': subtitles,
                 }
 
+        # simorgh-based playlist (see https://github.com/bbc/simorgh)
+        # JSON assigned to window.SIMORGH_DATA in a <script> element
+        simorgh_data = self._parse_json(
+            self._search_regex(
+                r'window\.SIMORGH_DATA\s*=\s*(\{[^<]+})\s*</',
+                webpage, 'simorgh playlist', default='{}'),
+            playlist_id, fatal=False)
+        # legacy media, video in promo object (eg, http://www.bbc.com/mundo/video_fotos/2015/06/150619_video_honduras_militares_hospitales_corrupcion_aw)
+        playlist = try_get(simorgh_data, lambda x: x['pageData']['promo']['media']['playlist']) or []
+        if playlist:
+            media = simorgh_data['pageData']['promo']
+            if media['media'].get('format') == 'video':
+                media.update(media['media'])
+                title = (dict_get(media.get('headlines') or {},
+                                  ('shortHeadline', 'headline'))
+                         or playlist_title),
+                programme_id = media.get('id')
+                if programme_id and title:
+                    formats = []
+                    keys = {'url', 'format', 'format_id', 'language', 'quality', 'tbr', 'resolution'}
+                    for format in playlist:
+                        if not (format.get('url') and format.get('format')):
+                            continue
+                        bitrate = format.pop('bitrate')
+                        format['tbr'] = int_or_none(bitrate, scale=1000) or parse_bitrate(bitrate)
+                        format['language'] = media.get('language')
+                        # format id: penultimate item from the url split on _ and .
+                        (fmt,) = re.split('[_.]', format['url'])[-2:][:1]
+                        format['format_id'] = '%s_%s' % (format['format'], fmt)
+                        # try to set resolution using any available data
+                        aspect_ratio = re.split(r'[xX:]', media.get('aspectRatio') or '')
+                        if len(aspect_ratio) != 2:
+                            aspect_ratio = None
+                        else:
+                            aspect_ratio = float_or_none(aspect_ratio[0], scale=aspect_ratio[1])
+                        # these may not be present, but try anyway
+                        width = int_or_none(format.get('width'))
+                        height = int_or_none(format.get('height'))
+                        if (not height) and aspect_ratio:
+                            height = int(width / aspect_ratio)
+                        elif (not width) and aspect_ratio:
+                            width = int(height * aspect_ratio)
+                        format['resolution'] = ('%dx%d' % (width, height) if width and height
+                                                else dict_get(format, ('resolution', 'res'), default=fmt))
+                        format['quality'] = -1
+                        formats.append(dict((k, format[k]) for k in keys))
+                    self._sort_formats(formats)
+                    return {
+                        'id': programme_id,
+                        'title': title,
+                        'description': media.get('summary') or playlist_description,
+                        'formats': formats,
+                        'subtitles': None,
+                        'thumbnail': try_get(media, lambda x: x['image']['href']),
+                        'timestamp': int_or_none(media.get('timestamp'), scale=1000)
+                    }
+
+        # general case: media nested in content object
+        # test: https://www.bbc.co.uk/scotland/articles/cm49v4x1r9lo
+        if simorgh_data:
+
+            def extract_media_from_simorgh(model):
+                if not isinstance(model, dict):
+                    return
+                for block in model.get('blocks') or {}:
+                    if block.get('type') == 'aresMediaMetadata':
+                        vpid = try_get(block, lambda x: x['model']['versions'][0]['versionId'])
+                        if vpid:
+                            formats, subtitles = self._download_media_selector(vpid)
+                            self._sort_formats(formats)
+                            model = block['model']
+                            version = model['versions'][0]
+                            thumbnail = model.get('imageUrl')
+                            return {
+                                'id': vpid,
+                                'title': model.get('title') or 'unnamed clip',
+                                'description': dict_get(model.get('synopses') or {}, ('long', 'medium', 'short')),
+                                'duration': (int_or_none(version.get('duration'))
+                                             or parse_duration(version.get('durationISO8601'))),
+                                'timestamp': version.get('availableFrom'),
+                                'thumbnail': urljoin(url, thumbnail.replace('$recipe', 'raw')) if thumbnail else None,
+                                'formats': formats,
+                                'subtitles': subtitles,
+                            }
+                    else:
+                        entry = extract_media_from_simorgh(block.get('model'))
+                        if entry:
+                            return entry
+
+            playlist = extract_media_from_simorgh(try_get(simorgh_data, lambda x: x['pageData']['content']['model']))
+            if playlist:
+                return playlist
+
         preload_state = self._parse_json(self._search_regex(
             r'window\.__PRELOADED_STATE__\s*=\s*({.+?});', webpage,
             'preload state', default='{}'), playlist_id, fatal=False)

youtube_dl/extractor/youtube.py

@@ -27,6 +27,7 @@ from ..compat import (
 )
 from ..jsinterp import JSInterpreter
 from ..utils import (
+    bug_reports_message,
     clean_html,
     dict_get,
     error_to_compat_str,
@@ -65,6 +66,7 @@ from ..utils import (
     url_or_none,
     urlencode_postdata,
     urljoin,
+    variadic,
 )
 
 
@@ -460,6 +462,26 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'uploader': uploader,
         }
 
+    @staticmethod
+    def _extract_thumbnails(data, *path_list, **kw_final_key):
+        """
+        Extract thumbnails from thumbnails dict
+        @param path_list: path list to level that contains 'thumbnails' key
+        """
+        final_key = kw_final_key.get('final_key', 'thumbnails')
+
+        return traverse_obj(data, ((
+            tuple(variadic(path) + (final_key, Ellipsis)
+                  for path in path_list or [()])), {
+            'url': ('url', T(url_or_none),
+                    # Sometimes youtube gives a wrong thumbnail URL. See:
+                    # https://github.com/yt-dlp/yt-dlp/issues/233
+                    # https://github.com/ytdl-org/youtube-dl/issues/28023
+                    T(lambda u: update_url(u, query=None) if u and 'maxresdefault' in u else u)),
+            'height': ('height', T(int_or_none)),
+            'width': ('width', T(int_or_none)),
+        }, T(lambda t: t if t.get('url') else None)))
+
     def _search_results(self, query, params):
         data = {
             'context': {
@@ -3183,8 +3205,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
             expected_type=txt_or_none)
 
     def _grid_entries(self, grid_renderer):
-        for item in grid_renderer['items']:
-            if not isinstance(item, dict):
+        for item in traverse_obj(grid_renderer, ('items', Ellipsis, T(dict))):
+            lockup_view_model = traverse_obj(item, ('lockupViewModel', T(dict)))
+            if lockup_view_model:
+                entry = self._extract_lockup_view_model(lockup_view_model)
+                if entry:
+                    yield entry
                 continue
             renderer = self._extract_grid_item_renderer(item)
             if not isinstance(renderer, dict):
@@ -3268,6 +3294,25 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
                 continue
             yield self._extract_video(renderer)
 
+    def _extract_lockup_view_model(self, view_model):
+        content_id = view_model.get('contentId')
+        if not content_id:
+            return
+        content_type = view_model.get('contentType')
+        if content_type not in ('LOCKUP_CONTENT_TYPE_PLAYLIST', 'LOCKUP_CONTENT_TYPE_PODCAST'):
+            self.report_warning(
+                'Unsupported lockup view model content type "{0}"{1}'.format(content_type, bug_reports_message()), only_once=True)
+            return
+        return merge_dicts(self.url_result(
+            update_url_query('https://www.youtube.com/playlist', {'list': content_id}),
+            ie=YoutubeTabIE.ie_key(), video_id=content_id), {
+                'title': traverse_obj(view_model, (
+                    'metadata', 'lockupMetadataViewModel', 'title', 'content', T(compat_str))),
+                'thumbnails': self._extract_thumbnails(view_model, (
+                    'contentImage', 'collectionThumbnailViewModel', 'primaryThumbnail',
+                    'thumbnailViewModel', 'image'), final_key='sources'),
+        })
+
     def _video_entry(self, video_renderer):
         video_id = video_renderer.get('videoId')
         if video_id: