Merge 6fdeb8b411 into 8e15177b41

Closes #9358
Authored by: jakeogh, minamotorin, shoxie007, bbilly1

Co-authored-by: minamotorin <76122224+minamotorin@users.noreply.github.com>
Co-authored-by: shoxie007 <74592022+shoxie007@users.noreply.github.com>
Co-authored-by: Simon <35427372+bbilly1@users.noreply.github.com>

README.md

@@ -666,7 +666,7 @@ ## Filesystem Options:
                                     The name of the browser to load cookies
                                     from. Currently supported browsers are:
                                     brave, chrome, chromium, edge, firefox,
-                                    opera, safari, vivaldi. Optionally, the
+                                    opera, safari, vivaldi, whale. Optionally, the
                                     KEYRING used for decrypting Chromium cookies
                                     on Linux, the name/path of the PROFILE to
                                     load cookies from, and the CONTAINER name

yt_dlp/cookies.py

@@ -46,7 +46,7 @@
 from .utils._utils import _YDLLogger
 from .utils.networking import normalize_url
 
-CHROMIUM_BASED_BROWSERS = {'brave', 'chrome', 'chromium', 'edge', 'opera', 'vivaldi'}
+CHROMIUM_BASED_BROWSERS = {'brave', 'chrome', 'chromium', 'edge', 'opera', 'vivaldi', 'whale'}
 SUPPORTED_BROWSERS = CHROMIUM_BASED_BROWSERS | {'firefox', 'safari'}
 
 
@@ -219,6 +219,7 @@ def _get_chromium_based_browser_settings(browser_name):
             'edge': os.path.join(appdata_local, R'Microsoft\Edge\User Data'),
             'opera': os.path.join(appdata_roaming, R'Opera Software\Opera Stable'),
             'vivaldi': os.path.join(appdata_local, R'Vivaldi\User Data'),
+            'whale': os.path.join(appdata_local, R'Naver\Naver Whale\User Data'),
         }[browser_name]
 
     elif sys.platform == 'darwin':
@@ -230,6 +231,7 @@ def _get_chromium_based_browser_settings(browser_name):
             'edge': os.path.join(appdata, 'Microsoft Edge'),
             'opera': os.path.join(appdata, 'com.operasoftware.Opera'),
             'vivaldi': os.path.join(appdata, 'Vivaldi'),
+            'whale': os.path.join(appdata, 'Naver/Whale'),
         }[browser_name]
 
     else:
@@ -241,6 +243,7 @@ def _get_chromium_based_browser_settings(browser_name):
             'edge': os.path.join(config, 'microsoft-edge'),
             'opera': os.path.join(config, 'opera'),
             'vivaldi': os.path.join(config, 'vivaldi'),
+            'whale': os.path.join(config, 'naver-whale'),
         }[browser_name]
 
     # Linux keyring names can be determined by snooping on dbus while opening the browser in KDE:
@@ -252,6 +255,7 @@ def _get_chromium_based_browser_settings(browser_name):
         'edge': 'Microsoft Edge' if sys.platform == 'darwin' else 'Chromium',
         'opera': 'Opera' if sys.platform == 'darwin' else 'Chromium',
         'vivaldi': 'Vivaldi' if sys.platform == 'darwin' else 'Chrome',
+        'whale': 'Whale',
     }[browser_name]
 
     browsers_without_profiles = {'opera'}

yt_dlp/extractor/bbc.py

@@ -602,7 +602,7 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
         'url': 'http://www.bbc.com/news/world-europe-32668511',
         'info_dict': {
             'id': 'world-europe-32668511',
-            'title': 'Russia stages massive WW2 parade',
+            'title': 'Russia stages massive WW2 parade despite Western boycott',
             'description': 'md5:00ff61976f6081841f759a08bf78cc9c',
         },
         'playlist_count': 2,
@@ -623,6 +623,7 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
         'info_dict': {
             'id': '3662a707-0af9-3149-963f-47bea720b460',
             'title': 'BUGGER',
+            'description': r're:BUGGER  The recent revelations by the whistleblower Edward Snowden were fascinating. .{211}\.{3}$',
         },
         'playlist_count': 18,
     }, {
@@ -631,14 +632,14 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
         'info_dict': {
             'id': 'p02mprgb',
             'ext': 'mp4',
-            'title': 'Aerial footage showed the site of the crash in the Alps - courtesy BFM TV',
-            'description': 'md5:2868290467291b37feda7863f7a83f54',
+            'title': 'Germanwings crash site aerial video',
+            'description': r're:(?s)Aerial video showed the site where the Germanwings flight 4U 9525, .{156} BFM TV\.$',
             'duration': 47,
             'timestamp': 1427219242,
             'upload_date': '20150324',
+            'thumbnail': 'https://ichef.bbci.co.uk/news/1024/media/images/81879000/jpg/_81879090_81879089.jpg',
         },
         'params': {
-            # rtmp download
             'skip_download': True,
         }
     }, {
@@ -656,21 +657,24 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
         },
         'params': {
             'skip_download': True,
-        }
+        },
+        'skip': 'now SIMORGH_DATA with no video',
     }, {
         # single video embedded with data-playable containing XML playlists (regional section)
         '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',
+            'display_id': '150619_video_honduras_militares_hospitales_corrupcion_aw',
             'ext': 'mp4',
             'title': 'Honduras militariza sus hospitales por nuevo escándalo de corrupción',
-            'description': 'md5:1525f17448c4ee262b64b8f0c9ce66c8',
+            'description': 'Honduras militariza sus hospitales por nuevo escándalo de corrupción',
             'timestamp': 1434713142,
             'upload_date': '20150619',
+            'thumbnail': 'https://a.files.bbci.co.uk/worldservice/live/assets/images/2015/06/19/150619132146_honduras_hsopitales_militares_640x360_aptn_nocredit.jpg',
         },
         'params': {
             'skip_download': True,
-        }
+        },
     }, {
         # single video from video playlist embedded with vxp-playlist-data JSON
         'url': 'http://www.bbc.com/news/video_and_audio/must_see/33376376',
@@ -683,22 +687,21 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
         },
         'params': {
             'skip_download': True,
-        }
+        },
+        'skip': '404 Not Found',
     }, {
-        # single video story with digitalData
+        # single video story with __PWA_PRELOADED_STATE__
         'url': 'http://www.bbc.com/travel/story/20150625-sri-lankas-spicy-secret',
         'info_dict': {
             'id': 'p02q6gc4',
-            'ext': 'flv',
-            'title': 'Sri Lanka’s spicy secret',
-            'description': 'As a new train line to Jaffna opens up the country’s north, travellers can experience a truly distinct slice of Tamil culture.',
-            'timestamp': 1437674293,
-            'upload_date': '20150723',
+            'ext': 'mp4',
+            'title': 'Tasting the spice of life in Jaffna',
+            'description': r're:(?s)BBC Travel Show’s Henry Golding explores the city of Jaffna .{151} aftertaste\.$',
+            'timestamp': 1646058397,
+            'upload_date': '20220228',
+            'duration': 255,
+            'thumbnail': 'https://ichef.bbci.co.uk/images/ic/1920xn/p02vxvkn.jpg',
         },
-        'params': {
-            # rtmp download
-            'skip_download': True,
-        }
     }, {
         # single video story without digitalData
         'url': 'http://www.bbc.com/autos/story/20130513-hyundais-rock-star',
@@ -710,12 +713,10 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'timestamp': 1415867444,
             'upload_date': '20141113',
         },
-        'params': {
-            # rtmp download
-            'skip_download': True,
-        }
+        'skip': 'redirects to TopGear home page',
     }, {
         # single video embedded with Morph
+        # TODO: replacement test page
         'url': 'http://www.bbc.co.uk/sport/live/olympics/36895975',
         'info_dict': {
             'id': 'p041vhd0',
@@ -726,27 +727,22 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'uploader': 'BBC Sport',
             'uploader_id': 'bbc_sport',
         },
-        'params': {
-            # m3u8 download
-            'skip_download': True,
-        },
-        'skip': 'Georestricted to UK',
+        'skip': 'Video no longer in page',
     }, {
-        # single video with playlist.sxml URL in playlist param
+        # single video in __INITIAL_DATA__
         'url': 'http://www.bbc.com/sport/0/football/33653409',
         'info_dict': {
             'id': 'p02xycnp',
             'ext': 'mp4',
-            'title': 'Transfers: Cristiano Ronaldo to Man Utd, Arsenal to spend?',
-            'description': 'BBC Sport\'s David Ornstein has the latest transfer gossip, including rumours of a Manchester United return for Cristiano Ronaldo.',
+            'title': 'Ronaldo to Man Utd, Arsenal to spend?',
+            'description': r're:(?s)BBC Sport\'s David Ornstein rounds up the latest transfer reports, .{359} here\.$',
+            'timestamp': 1437750175,
+            'upload_date': '20150724',
+            'thumbnail': r're:https?://.+/.+media/images/69320000/png/_69320754_mmgossipcolumnextraaugust18.png',
             'duration': 140,
         },
-        'params': {
-            # rtmp download
-            'skip_download': True,
-        }
     }, {
-        # article with multiple videos embedded with playlist.sxml in playlist param
+        # article with multiple videos embedded with Morph.setPayload
         'url': 'http://www.bbc.com/sport/0/football/34475836',
         'info_dict': {
             'id': '34475836',
@@ -754,6 +750,21 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'description': 'Fast-paced football, wit, wisdom and a ready smile - why Liverpool fans should come to love new boss Jurgen Klopp.',
         },
         'playlist_count': 3,
+    }, {
+        # Testing noplaylist
+        'url': 'http://www.bbc.com/sport/0/football/34475836',
+        'info_dict': {
+            'id': 'p034ppnv',
+            'ext': 'mp4',
+            'title': 'All you need to know about Jurgen Klopp',
+            'timestamp': 1444335081,
+            'upload_date': '20151008',
+            'duration': 122.0,
+            'thumbnail': 'https://ichef.bbci.co.uk/onesport/cps/976/cpsprodpb/7542/production/_85981003_klopp.jpg',
+        },
+        'params': {
+            'noplaylist': True,
+        },
     }, {
         # school report article with single video
         'url': 'http://www.bbc.co.uk/schoolreport/35744779',
@@ -762,6 +773,7 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'title': 'School which breaks down barriers in Jerusalem',
         },
         'playlist_count': 1,
+        'skip': 'redirects to Young Reporter home page https://www.bbc.co.uk/news/topics/cg41ylwv43pt',
     }, {
         # single video with playlist URL from weather section
         'url': 'http://www.bbc.com/weather/features/33601775',
@@ -778,18 +790,33 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'thumbnail': r're:https?://.+/.+\.jpg',
             'timestamp': 1437785037,
             'upload_date': '20150725',
+            'duration': 105,
         },
     }, {
         # video with window.__INITIAL_DATA__ and value as JSON string
         'url': 'https://www.bbc.com/news/av/world-europe-59468682',
         'info_dict': {
-            'id': 'p0b71qth',
+            'id': 'p0b779gc',
             'ext': 'mp4',
             'title': 'Why France is making this woman a national hero',
-            'description': 'md5:7affdfab80e9c3a1f976230a1ff4d5e4',
+            'description': r're:(?s)France is honouring the US-born 20th Century singer and activist Josephine .{208} Second World War.',
             'thumbnail': r're:https?://.+/.+\.jpg',
-            'timestamp': 1638230731,
-            'upload_date': '20211130',
+            'timestamp': 1638215626,
+            'upload_date': '20211129',
+            'duration': 125,
+        },
+    }, {
+        # video with script id __NEXT_DATA__ and value as JSON string
+        'url': 'https://www.bbc.com/news/uk-68546268',
+        'info_dict': {
+            'id': 'p0hj0lq7',
+            'ext': 'mp4',
+            'title': 'Nasser Hospital doctor describes his treatment by IDF',
+            'description': r're:(?s)Doctor Abu Sabha said he was detained by Israeli forces after .{276} hostages\."$',
+            'thumbnail': r're:https?://.+/.+\.jpg',
+            'timestamp': 1710188248,
+            'upload_date': '20240311',
+            'duration': 104,
         },
     }, {
         # single video article embedded with data-media-vpid
@@ -817,6 +844,7 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'uploader': 'Radio 3',
             'uploader_id': 'bbc_radio_three',
         },
+        'skip': '404 Not Found',
     }, {
         'url': 'http://www.bbc.co.uk/learningenglish/chinese/features/lingohack/ep-181227',
         'info_dict': {
@@ -824,6 +852,7 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'ext': 'mp4',
             'title': 'md5:2fabf12a726603193a2879a055f72514',
             'description': 'Learn English words and phrases from this story',
+            'thumbnail': 'https://ichef.bbci.co.uk/images/ic/1200x675/p06pq9gk.jpg',
         },
         'add_ie': [BBCCoUkIE.ie_key()],
     }, {
@@ -832,28 +861,30 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
         'info_dict': {
             'id': 'p07c6sb9',
             'ext': 'mp4',
-            'title': 'How positive thinking is harming your happiness',
-            'alt_title': 'The downsides of positive thinking',
-            'description': 'md5:fad74b31da60d83b8265954ee42d85b4',
+            'title': 'The downsides of positive thinking',
+            'description': 'The downsides of positive thinking',
             'duration': 235,
-            'thumbnail': r're:https?://.+/p07c9dsr.jpg',
-            'upload_date': '20190604',
-            'categories': ['Psychology'],
+            'thumbnail': r're:https?://.+/p07c9dsr\.(?:jpg|webp|png)',
+            'upload_date': '20220223',
+            'timestamp': 1645632746,
         },
     }, {
         # BBC Sounds
-        'url': 'https://www.bbc.co.uk/sounds/play/m001q78b',
+        'url': 'https://www.bbc.co.uk/sounds/play/w3ct5rgx',
         'info_dict': {
-            'id': 'm001q789',
+            'id': 'p0hrw4nr',
             'ext': 'mp4',
-            'title': 'The Night Tracks Mix - Music for the darkling hour',
-            'thumbnail': 'https://ichef.bbci.co.uk/images/ic/raw/p0c00hym.jpg',
-            'chapters': 'count:8',
-            'description': 'md5:815fb51cbdaa270040aab8145b3f1d67',
-            'uploader': 'Radio 3',
-            'duration': 1800,
-            'uploader_id': 'bbc_radio_three',
-        },
+            'title': 'Are our coastlines being washed away?',
+            'description': r're:(?s)Around the world, coastlines are constantly changing .{2000,} Images\)$',
+            'timestamp': 1713556800,
+            'upload_date': '20240419',
+            'duration': 1588,
+            'thumbnail': 'https://ichef.bbci.co.uk/images/ic/raw/p0hrnxbl.jpg',
+            'uploader': 'World Service',
+            'uploader_id': 'bbc_world_service',
+            'series': 'CrowdScience',
+            'chapters': [],
+        }
     }, {  # onion routes
         'url': 'https://www.bbcnewsd73hkzno2ini43t4gblxvycyac5aw4gnv7t2rccijh7745uqd.onion/news/av/world-europe-63208576',
         'only_matching': True,
@@ -1008,8 +1039,7 @@ def _real_extract(self, url):
             webpage, 'group id', default=None)
         if group_id:
             return self.url_result(
-                'https://www.bbc.co.uk/programmes/%s' % group_id,
-                ie=BBCCoUkIE.ie_key())
+                f'https://www.bbc.co.uk/programmes/{group_id}', BBCCoUkIE)
 
         # single video story (e.g. http://www.bbc.com/travel/story/20150625-sri-lankas-spicy-secret)
         programme_id = self._search_regex(
@@ -1069,83 +1099,133 @@ def _real_extract(self, url):
                 }
 
         # Morph based embed (e.g. http://www.bbc.co.uk/sport/live/olympics/36895975)
-        # There are several setPayload calls may be present but the video
-        # seems to be always related to the first one
-        morph_payload = self._parse_json(
-            self._search_regex(
-                r'Morph\.setPayload\([^,]+,\s*({.+?})\);',
-                webpage, 'morph payload', default='{}'),
-            playlist_id, fatal=False)
+        # Several setPayload calls may be present but the video(s)
+        # should be in one that mentions leadMedia or videoData
+        morph_payload = self._search_json(
+            r'\bMorph\s*\.\s*setPayload\s*\([^,]+,', webpage, 'morph payload', playlist_id,
+            contains_pattern=r'{(?s:(?:(?!</script>).)+(?:"leadMedia"|\\"videoData\\")\s*:.+)}',
+            default={})
         if morph_payload:
-            components = try_get(morph_payload, lambda x: x['body']['components'], list) or []
-            for component in components:
-                if not isinstance(component, dict):
-                    continue
-                lead_media = try_get(component, lambda x: x['props']['leadMedia'], dict)
-                if not lead_media:
-                    continue
-                identifiers = lead_media.get('identifiers')
-                if not identifiers or not isinstance(identifiers, dict):
-                    continue
-                programme_id = identifiers.get('vpid') or identifiers.get('playablePid')
+            for lead_media in traverse_obj(morph_payload, (
+                    'body', 'components', ..., 'props', 'leadMedia', {dict})):
+                programme_id = traverse_obj(lead_media, ('identifiers', ('vpid', 'playablePid'), {str}, any))
                 if not programme_id:
                     continue
-                title = lead_media.get('title') or self._og_search_title(webpage)
                 formats, subtitles = self._download_media_selector(programme_id)
-                description = lead_media.get('summary')
-                uploader = lead_media.get('masterBrand')
-                uploader_id = lead_media.get('mid')
-                duration = None
-                duration_d = lead_media.get('duration')
-                if isinstance(duration_d, dict):
-                    duration = parse_duration(dict_get(
-                        duration_d, ('rawDuration', 'formattedDuration', 'spokenDuration')))
                 return {
                     'id': programme_id,
-                    'title': title,
-                    'description': description,
-                    'duration': duration,
-                    'uploader': uploader,
-                    'uploader_id': uploader_id,
+                    'title': lead_media.get('title') or self._og_search_title(webpage),
+                    **traverse_obj(lead_media, {
+                        'description': ('summary', {str}),
+                        'duration': ('duration', ('rawDuration', 'formattedDuration', 'spokenDuration'), {parse_duration}),
+                        'uploader': ('masterBrand', {str}),
+                        'uploader_id': ('mid', {str}),
+                    }),
                     'formats': formats,
                     'subtitles': subtitles,
                 }
+            body = self._parse_json(traverse_obj(morph_payload, (
+                'body', 'content', 'article', 'body')), playlist_id, fatal=False)
+            for video_data in traverse_obj(body, (lambda _, v: v['videoData']['pid'], 'videoData')):
+                if video_data.get('vpid'):
+                    video_id = video_data['vpid']
+                    formats, subtitles = self._download_media_selector(video_id)
+                    entry = {
+                        'id': video_id,
+                        'formats': formats,
+                        'subtitles': subtitles,
+                    }
+                else:
+                    video_id = video_data['pid']
+                    entry = self.url_result(
+                        f'https://www.bbc.co.uk/programmes/{video_id}', BBCCoUkIE,
+                        video_id, url_transparent=True)
+                entry.update({
+                    'timestamp': traverse_obj(morph_payload, (
+                        'body', 'content', 'article', 'dateTimeInfo', 'dateTime', {parse_iso8601})
+                    ),
+                    **traverse_obj(video_data, {
+                        'thumbnail': (('iChefImage', 'image'), {url_or_none}, any),
+                        'title': (('title', 'caption'), {str}, any),
+                        'duration': ('duration', {parse_duration}),
+                    }),
+                })
+                if video_data.get('isLead') and not self._yes_playlist(playlist_id, video_id):
+                    return entry
+                entries.append(entry)
+            if entries:
+                playlist_title = traverse_obj(morph_payload, (
+                    'body', 'content', 'article', 'headline', {str})) or playlist_title
+                return self.playlist_result(
+                    entries, playlist_id, playlist_title, playlist_description)
 
-        preload_state = self._parse_json(self._search_regex(
-            r'window\.__PRELOADED_STATE__\s*=\s*({.+?});', webpage,
-            'preload state', default='{}'), playlist_id, fatal=False)
-        if preload_state:
-            current_programme = preload_state.get('programmes', {}).get('current') or {}
-            programme_id = current_programme.get('id')
-            if current_programme and programme_id and current_programme.get('type') == 'playable_item':
-                title = current_programme.get('titles', {}).get('tertiary') or playlist_title
-                formats, subtitles = self._download_media_selector(programme_id)
-                synopses = current_programme.get('synopses') or {}
-                network = current_programme.get('network') or {}
-                duration = int_or_none(
-                    current_programme.get('duration', {}).get('value'))
-                thumbnail = None
-                image_url = current_programme.get('image_url')
-                if image_url:
-                    thumbnail = image_url.replace('{recipe}', 'raw')
+        # various PRELOADED_STATE JSON
+        preload_state = self._search_json(
+            r'window\.__(?:PWA_)?PRELOADED_STATE__\s*=', webpage,
+            'preload state', playlist_id, transform_source=js_to_json, default={})
+        # PRELOADED_STATE with current programmme
+        current_programme = traverse_obj(preload_state, ('programmes', 'current', {dict}))
+        programme_id = traverse_obj(current_programme, ('id', {str}))
+        if programme_id and current_programme.get('type') == 'playable_item':
+            title = traverse_obj(current_programme, ('titles', ('tertiary', 'secondary'), {str}, any)) or playlist_title
+            formats, subtitles = self._download_media_selector(programme_id)
+            return {
+                'id': programme_id,
+                'title': title,
+                'formats': formats,
+                **traverse_obj(current_programme, {
+                    'description': ('synopses', ('long', 'medium', 'short'), {str}, any),
+                    'thumbnail': ('image_url', {lambda u: url_or_none(u.replace('{recipe}', 'raw'))}),
+                    'duration': ('duration', 'value', {int_or_none}),
+                    'uploader': ('network', 'short_title', {str}),
+                    'uploader_id': ('network', 'id', {str}),
+                    'timestamp': ((('availability', 'from'), ('release', 'date')), {parse_iso8601}, any),
+                    'series': ('titles', 'primary', {str}),
+                }),
+                'subtitles': subtitles,
+                'chapters': traverse_obj(preload_state, (
+                    'tracklist', 'tracks', lambda _, v: float(v['offset']['start']), {
+                        'title': ('titles', {lambda x: join_nonempty(
+                            'primary', 'secondary', 'tertiary', delim=' - ', from_dict=x)}),
+                        'start_time': ('offset', 'start', {float_or_none}),
+                        'end_time': ('offset', 'end', {float_or_none}),
+                    })
+                ),
+            }
+
+        # PWA_PRELOADED_STATE with article video asset
+        asset_id = traverse_obj(preload_state, (
+            'entities', 'articles', lambda k, _: k.rsplit('/', 1)[-1] == playlist_id,
+            'assetVideo', 0, {str}, any))
+        if asset_id:
+            video_id = traverse_obj(preload_state, ('entities', 'videos', asset_id, 'vpid', {str}))
+            if video_id:
+                article = traverse_obj(preload_state, (
+                    'entities', 'articles', lambda _, v: v['assetVideo'][0] == asset_id, any))
+
+                def image_url(image_id):
+                    return traverse_obj(preload_state, (
+                        'entities', 'images', image_id, 'url',
+                        {lambda u: url_or_none(u.replace('$recipe', 'raw'))}))
+
+                formats, subtitles = self._download_media_selector(video_id)
                 return {
-                    'id': programme_id,
-                    'title': title,
-                    'description': dict_get(synopses, ('long', 'medium', 'short')),
-                    'thumbnail': thumbnail,
-                    'duration': duration,
-                    'uploader': network.get('short_title'),
-                    'uploader_id': network.get('id'),
+                    'id': video_id,
+                    **traverse_obj(preload_state, ('entities', 'videos', asset_id, {
+                        'title': ('title', {str}),
+                        'description': (('synopsisLong', 'synopsisMedium', 'synopsisShort'), {str}, any),
+                        'thumbnail': (0, {image_url}),
+                        'duration': ('duration', {int_or_none}),
+                    })),
                     'formats': formats,
                     'subtitles': subtitles,
-                    'chapters': traverse_obj(preload_state, (
-                        'tracklist', 'tracks', lambda _, v: float_or_none(v['offset']['start']), {
-                            'title': ('titles', {lambda x: join_nonempty(
-                                'primary', 'secondary', 'tertiary', delim=' - ', from_dict=x)}),
-                            'start_time': ('offset', 'start', {float_or_none}),
-                            'end_time': ('offset', 'end', {float_or_none}),
-                        })) or None,
+                    'timestamp': traverse_obj(article, ('displayDate', {parse_iso8601})),
                 }
+            else:
+                return self.url_result(
+                    f'https://www.bbc.co.uk/programmes/{asset_id}', BBCCoUkIE,
+                    asset_id, playlist_title, display_id=playlist_id,
+                    description=playlist_description)
 
         bbc3_config = self._parse_json(
             self._search_regex(
@@ -1191,6 +1271,28 @@ def _real_extract(self, url):
                 return self.playlist_result(
                     entries, playlist_id, playlist_title, playlist_description)
 
+        def parse_model(model):
+            """Extract single video from model structure"""
+            item_id = traverse_obj(model, ('versions', 0, 'versionId', {str}))
+            if not item_id:
+                return
+            formats, subtitles = self._download_media_selector(item_id)
+            return {
+                'id': item_id,
+                'formats': formats,
+                'subtitles': subtitles,
+                **traverse_obj(model, {
+                    'title': ('title', {str}),
+                    'thumbnail': ('imageUrl', {lambda u: urljoin(url, u.replace('$recipe', 'raw'))}),
+                    'description': ('synopses', ('long', 'medium', 'short'), {str}, {lambda x: x or None}, any),
+                    'duration': ('versions', 0, 'duration', {int}),
+                    'timestamp': ('versions', 0, 'availableFrom', {functools.partial(int_or_none, scale=1000)}),
+                })
+            }
+
+        def is_type(*types):
+            return lambda _, v: v['type'] in types
+
         initial_data = self._search_regex(
             r'window\.__INITIAL_DATA__\s*=\s*("{.+?}")\s*;', webpage,
             'quoted preload state', default=None)
@@ -1202,6 +1304,19 @@ def _real_extract(self, url):
             initial_data = self._parse_json(initial_data or '"{}"', playlist_id, fatal=False)
         initial_data = self._parse_json(initial_data, playlist_id, fatal=False)
         if initial_data:
+            for video_data in traverse_obj(initial_data, (
+                    'stores', 'article', 'articleBodyContent', is_type('video'))):
+                model = traverse_obj(video_data, (
+                    'model', 'blocks', is_type('aresMedia'),
+                    'model', 'blocks', is_type('aresMediaMetadata'),
+                    'model', {dict}, any))
+                entry = parse_model(model)
+                if entry:
+                    entries.append(entry)
+            if entries:
+                return self.playlist_result(
+                    entries, playlist_id, playlist_title, playlist_description)
+
             def parse_media(media):
                 if not media:
                     return
@@ -1234,27 +1349,90 @@ def parse_media(media):
                         'subtitles': subtitles,
                         'timestamp': item_time,
                         'description': strip_or_none(item_desc),
+                        'duration': int_or_none(item.get('duration')),
                     })
-            for resp in (initial_data.get('data') or {}).values():
-                name = resp.get('name')
+
+            for resp in traverse_obj(initial_data, ('data', lambda _, v: v['name'])):
+                name = resp['name']
                 if name == 'media-experience':
                     parse_media(try_get(resp, lambda x: x['data']['initialItem']['mediaItem'], dict))
                 elif name == 'article':
-                    for block in (try_get(resp,
-                                          (lambda x: x['data']['blocks'],
-                                           lambda x: x['data']['content']['model']['blocks'],),
-                                          list) or []):
-                        if block.get('type') not in ['media', 'video']:
-                            continue
-                        parse_media(block.get('model'))
+                    for block in traverse_obj(resp, (
+                            'data', (None, ('content', 'model')), 'blocks',
+                            is_type('media', 'video'), 'model', {dict})):
+                        parse_media(block)
             return self.playlist_result(
                 entries, playlist_id, playlist_title, playlist_description)
 
+        # extract from SIMORGH_DATA hydration JSON
+        simorgh_data = self._search_json(
+            r'window\s*\.\s*SIMORGH_DATA\s*=', webpage,
+            'simorgh data', playlist_id, default={})
+        if simorgh_data:
+            done = False
+            for video_data in traverse_obj(simorgh_data, (
+                    'pageData', 'content', 'model', 'blocks', is_type('video', 'legacyMedia'))):
+                model = traverse_obj(video_data, (
+                    'model', 'blocks', is_type('aresMedia'),
+                    'model', 'blocks', is_type('aresMediaMetadata'),
+                    'model', {dict}, any))
+                if video_data['type'] == 'video':
+                    entry = parse_model(model)
+                else:  # legacyMedia: no duration, subtitles
+                    block_id, entry = traverse_obj(model, ('blockId', {str})), None
+                    media_data = traverse_obj(simorgh_data, (
+                        'pageData', 'promo', 'media',
+                        {lambda x: x if x['id'] == block_id else None}))
+                    formats = traverse_obj(media_data, ('playlist', lambda _, v: url_or_none(v['url']), {
+                        'url': ('url', {url_or_none}),
+                        'ext': ('format', {str}),
+                        'tbr': ('bitrate', {functools.partial(int_or_none, scale=1000)}),
+                    }))
+                    if formats:
+                        entry = {
+                            'id': block_id,
+                            'display_id': playlist_id,
+                            'formats': formats,
+                            'description': traverse_obj(simorgh_data, ('pageData', 'promo', 'summary', {str})),
+                            **traverse_obj(model, {
+                                'title': ('title', {str}),
+                                'thumbnail': ('imageUrl', {lambda u: urljoin(url, u.replace('$recipe', 'raw'))}),
+                                'description': ('synopses', ('long', 'medium', 'short'), {str}, any),
+                                'timestamp': ('firstPublished', {functools.partial(int_or_none, scale=1000)}),
+                            }),
+                        }
+                        done = True
+                if entry:
+                    entries.append(entry)
+                if done:
+                    break
+            if entries:
+                return self.playlist_result(
+                    entries, playlist_id, playlist_title, playlist_description)
+
         def extract_all(pattern):
             return list(filter(None, map(
                 lambda s: self._parse_json(s, playlist_id, fatal=False),
                 re.findall(pattern, webpage))))
 
+        # US accessed article with single embedded video (e.g.
+        # https://www.bbc.com/news/uk-68546268)
+        next_data = traverse_obj(self._search_nextjs_data(webpage, playlist_id, default={}),
+                                 ('props', 'pageProps', 'page'))
+        model = traverse_obj(next_data, (
+            ..., 'contents', is_type('video'),
+            'model', 'blocks', is_type('media'),
+            'model', 'blocks', is_type('mediaMetadata'),
+            'model', {dict}, any))
+        if model and (entry := parse_model(model)):
+            if not entry.get('timestamp'):
+                entry['timestamp'] = traverse_obj(next_data, (
+                    ..., 'contents', is_type('timestamp'), 'model',
+                    'timestamp', {functools.partial(int_or_none, scale=1000)}, any))
+            entries.append(entry)
+            return self.playlist_result(
+                entries, playlist_id, playlist_title, playlist_description)
+
         # Multiple video article (e.g.
         # http://www.bbc.co.uk/blogs/adamcurtis/entries/3662a707-0af9-3149-963f-47bea720b460)
         EMBED_URL = r'https?://(?:www\.)?bbc\.co\.uk/(?:[^/]+/)+%s(?:\b[^"]+)?' % self._ID_REGEX

yt_dlp/extractor/common.py

@@ -957,7 +957,8 @@ def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=
         if urlh is False:
             assert not fatal
             return False
-        content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
+        content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal,
+                                             encoding=encoding, data=data)
         return (content, urlh)
 
     @staticmethod
@@ -1005,8 +1006,10 @@ def __check_blocked(self, content):
                 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
                 expected=True)
 
-    def _request_dump_filename(self, url, video_id):
-        basen = f'{video_id}_{url}'
+    def _request_dump_filename(self, url, video_id, data=None):
+        if data is not None:
+            data = hashlib.md5(data).hexdigest()
+        basen = join_nonempty(video_id, data, url, delim='_')
         trim_length = self.get_param('trim_file_name') or 240
         if len(basen) > trim_length:
             h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
@@ -1028,16 +1031,18 @@ def __decode_webpage(self, webpage_bytes, encoding, headers):
         except LookupError:
             return webpage_bytes.decode('utf-8', 'replace')
 
-    def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
+    def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True,
+                              prefix=None, encoding=None, data=None):
         webpage_bytes = urlh.read()
         if prefix is not None:
             webpage_bytes = prefix + webpage_bytes
+        url_or_request = self._create_request(url_or_request, data)
         if self.get_param('dump_intermediate_pages', False):
             self.to_screen('Dumping request to ' + urlh.url)
             dump = base64.b64encode(webpage_bytes).decode('ascii')
             self._downloader.to_screen(dump)
         if self.get_param('write_pages'):
-            filename = self._request_dump_filename(urlh.url, video_id)
+            filename = self._request_dump_filename(urlh.url, video_id, url_or_request.data)
             self.to_screen(f'Saving request to {filename}')
             with open(filename, 'wb') as outf:
                 outf.write(webpage_bytes)
@@ -1098,7 +1103,7 @@ def download_content(self, url_or_request, video_id, note=note, errnote=errnote,
                              impersonate=None, require_impersonation=False):
             if self.get_param('load_pages'):
                 url_or_request = self._create_request(url_or_request, data, headers, query)
-                filename = self._request_dump_filename(url_or_request.url, video_id)
+                filename = self._request_dump_filename(url_or_request.url, video_id, url_or_request.data)
                 self.to_screen(f'Loading request from {filename}')
                 try:
                     with open(filename, 'rb') as dumpf:

yt_dlp/extractor/markiza.py

@@ -1,76 +1,128 @@
 import re
 
 from .common import InfoExtractor
-from ..compat import compat_str
 from ..utils import (
-    orderedSet,
-    parse_duration,
-    try_get,
+    clean_html,
+    determine_ext,
+    int_or_none,
+    js_to_json,
+    traverse_obj,
+    unified_strdate,
+    url_or_none,
 )
 
 
-class MarkizaIE(InfoExtractor):
-    _WORKING = False
-    _VALID_URL = r'https?://(?:www\.)?videoarchiv\.markiza\.sk/(?:video/(?:[^/]+/)*|embed/)(?P<id>\d+)(?:[_/]|$)'
+class NovaEmbedIE(InfoExtractor):
+    _VALID_URL = r'https?://media(?:tn)?\.cms\.markiza\.sk/embed/(?P<id>[^/?#&]+)'
     _TESTS = [{
-        'url': 'http://videoarchiv.markiza.sk/video/oteckovia/84723_oteckovia-109',
-        'md5': 'ada4e9fad038abeed971843aa028c7b0',
+        'url': 'https://media.cms.markiza.sk/embed/8o0n0r?autoplay=1',
         'info_dict': {
-            'id': '139078',
+            'id': '8o0n0r',
+            'title': '2180. díl',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'duration': 2578,
+        },
+        'params': {
+            'skip_download': True,
+            'ignore_no_formats_error': True,
+        },
+        'expected_warnings': ['DRM protected', 'Requested format is not available'],
+    }, {
+        'url': 'https://media.cms.markiza.sk/embed/KybpWYvcgOa',
+        'info_dict': {
+            'id': 'KybpWYvcgOa',
             'ext': 'mp4',
-            'title': 'Oteckovia 109',
-            'description': 'md5:d41d8cd98f00b204e9800998ecf8427e',
-            'thumbnail': r're:^https?://.*\.jpg$',
-            'duration': 2760,
+            'title': 'Borhyová oslavila 60? Soutěžící z pořadu odboural moderátora Ondřeje Sokola',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'duration': 114,
         },
-    }, {
-        'url': 'http://videoarchiv.markiza.sk/video/televizne-noviny/televizne-noviny/85430_televizne-noviny',
-        'info_dict': {
-            'id': '85430',
-            'title': 'Televízne noviny',
-        },
-        'playlist_count': 23,
-    }, {
-        'url': 'http://videoarchiv.markiza.sk/video/oteckovia/84723',
-        'only_matching': True,
-    }, {
-        'url': 'http://videoarchiv.markiza.sk/video/84723',
-        'only_matching': True,
-    }, {
-        'url': 'http://videoarchiv.markiza.sk/video/filmy/85190_kamenak',
-        'only_matching': True,
-    }, {
-        'url': 'http://videoarchiv.markiza.sk/video/reflex/zo-zakulisia/84651_pribeh-alzbetky',
-        'only_matching': True,
-    }, {
-        'url': 'http://videoarchiv.markiza.sk/embed/85295',
-        'only_matching': True,
+        'params': {'skip_download': 'm3u8'},
     }]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        data = self._download_json(
-            'http://videoarchiv.markiza.sk/json/video_jwplayer7.json',
-            video_id, query={'id': video_id})
+        webpage = self._download_webpage(url, video_id)
 
-        info = self._parse_jwplayer_data(data, m3u8_id='hls', mpd_id='dash')
+        has_drm = False
+        duration = None
+        formats = []
 
-        if info.get('_type') == 'playlist':
-            info.update({
-                'id': video_id,
-                'title': try_get(
-                    data, lambda x: x['details']['name'], compat_str),
-            })
-        else:
-            info['duration'] = parse_duration(
-                try_get(data, lambda x: x['details']['duration'], compat_str))
-        return info
+        def process_format_list(format_list, format_id=""):
+            nonlocal formats, has_drm
+            if not isinstance(format_list, list):
+                format_list = [format_list]
+            for format_dict in format_list:
+                if not isinstance(format_dict, dict):
+                    continue
+                if (not self.get_param('allow_unplayable_formats')
+                        and traverse_obj(format_dict, ('drm', 'keySystem'))):
+                    has_drm = True
+                    continue
+                format_url = url_or_none(format_dict.get('src'))
+                format_type = format_dict.get('type')
+                ext = determine_ext(format_url)
+                if (format_type == 'application/x-mpegURL'
+                        or format_id == 'HLS' or ext == 'm3u8'):
+                    formats.extend(self._extract_m3u8_formats(
+                        format_url, video_id, 'mp4',
+                        entry_protocol='m3u8_native', m3u8_id='hls',
+                        fatal=False))
+                elif (format_type == 'application/dash+xml'
+                      or format_id == 'DASH' or ext == 'mpd'):
+                    formats.extend(self._extract_mpd_formats(
+                        format_url, video_id, mpd_id='dash', fatal=False))
+                else:
+                    formats.append({
+                        'url': format_url,
+                    })
+
+        player = self._search_json(
+            r'player:', webpage, 'player', video_id, fatal=False, end_pattern=r';\s*</script>')
+        if player:
+            for src in traverse_obj(player, ('lib', 'source', 'sources', ...)):
+                process_format_list(src)
+            duration = traverse_obj(player, ('sourceInfo', 'duration', {int_or_none}))
+        if not formats and not has_drm:
+            # older code path, in use before August 2023
+            player = self._parse_json(
+                self._search_regex(
+                    (r'(?:(?:replacePlaceholders|processAdTagModifier).*?:\s*)?(?:replacePlaceholders|processAdTagModifier)\s*\(\s*(?P<json>{.*?})\s*\)(?:\s*\))?\s*,',
+                     r'Player\.init\s*\([^,]+,(?P<cndn>\s*\w+\s*\?)?\s*(?P<json>{(?(cndn).+?|.+)})\s*(?(cndn):|,\s*{.+?}\s*\)\s*;)'),
+                    webpage, 'player', group='json'), video_id)
+            if player:
+                for format_id, format_list in player['tracks'].items():
+                    process_format_list(format_list, format_id)
+                duration = int_or_none(player.get('duration'))
+
+        if not formats and has_drm:
+            self.report_drm(video_id)
+
+        title = self._og_search_title(
+            webpage, default=None) or self._search_regex(
+            (r'<value>(?P<title>[^<]+)',
+             r'videoTitle\s*:\s*(["\'])(?P<value>(?:(?!\1).)+)\1'), webpage,
+            'title', group='value')
+        thumbnail = self._og_search_thumbnail(
+            webpage, default=None) or self._search_regex(
+            r'poster\s*:\s*(["\'])(?P<value>(?:(?!\1).)+)\1', webpage,
+            'thumbnail', fatal=False, group='value')
+        duration = int_or_none(self._search_regex(
+            r'videoDuration\s*:\s*(\d+)', webpage, 'duration',
+            default=duration))
+
+        return {
+            'id': video_id,
+            'title': title,
+            'thumbnail': thumbnail,
+            'duration': duration,
+            'formats': formats,
+        }
 
 
-class MarkizaPageIE(InfoExtractor):
-    _WORKING = False
-    _VALID_URL = r'https?://(?:www\.)?(?:(?:[^/]+\.)?markiza|tvnoviny)\.sk/(?:[^/]+/)*(?P<id>\d+)_'
+class NovaIE(InfoExtractor):
+    IE_DESC = 'tvnoviny.sk, markiza.sk'
+    _VALID_URL = r'https?://(?:www\.)?(?:(?:[^/]+\.)?markiza|tvnoviny)\.sk/(?:[^/]+/)+(?P<id>[^/]+?)(?:\.html|/|$)'
     _TESTS = [{
         'url': 'http://www.markiza.sk/soubiz/zahranicny/1923705_oteckovia-maju-svoj-den-ti-slavni-nie-su-o-nic-menej-rozkosni',
         'md5': 'ada4e9fad038abeed971843aa028c7b0',
@@ -102,23 +154,116 @@ class MarkizaPageIE(InfoExtractor):
         'only_matching': True,
     }]
 
-    @classmethod
-    def suitable(cls, url):
-        return False if MarkizaIE.suitable(url) else super(MarkizaPageIE, cls).suitable(url)
-
     def _real_extract(self, url):
-        playlist_id = self._match_id(url)
+        mobj = self._match_valid_url(url)
+        display_id = mobj.group('id')
+        site = mobj.group('site')
 
-        webpage = self._download_webpage(
-            # Downloading for some hosts (e.g. dajto, doma) fails with 500
-            # although everything seems to be OK, so considering 500
-            # status code to be expected.
-            url, playlist_id, expected_status=500)
+        webpage = self._download_webpage(url, display_id)
 
-        entries = [
-            self.url_result('http://videoarchiv.markiza.sk/video/%s' % video_id)
-            for video_id in orderedSet(re.findall(
-                r'(?:initPlayer_|data-entity=["\']|id=["\']player_)(\d+)',
-                webpage))]
+        description = clean_html(self._og_search_description(webpage, default=None))
+        if site == 'novaplus':
+            upload_date = unified_strdate(self._search_regex(
+                r'(\d{1,2}-\d{1,2}-\d{4})$', display_id, 'upload date', default=None))
+        elif site == 'fanda':
+            upload_date = unified_strdate(self._search_regex(
+                r'<span class="date_time">(\d{1,2}\.\d{1,2}\.\d{4})', webpage, 'upload date', default=None))
+        else:
+            upload_date = None
 
-        return self.playlist_result(entries, playlist_id)
+        # novaplus
+        embed_id = self._search_regex(
+            r'<iframe[^>]+\bsrc=["\'](?:https?:)?//media(?:tn)?\.cms\.markiza\.sk/embed/([^/?#&"\']+)',
+            webpage, 'embed url', default=None)
+        if embed_id:
+            return {
+                '_type': 'url_transparent',
+                'url': 'https://media.cms.markiza.sk/embed/%s' % embed_id,
+                'ie_key': NovaEmbedIE.ie_key(),
+                'id': embed_id,
+                'description': description,
+                'upload_date': upload_date
+            }
+
+        video_id = self._search_regex(
+            [r"(?:media|video_id)\s*:\s*'(\d+)'",
+             r'media=(\d+)',
+             r'id="article_video_(\d+)"',
+             r'id="player_(\d+)"'],
+            webpage, 'video id')
+
+        config_url = self._search_regex(
+            r'src="(https?://(?:tn|api)\.markiza\.sk/bin/player/videojs/config\.php\?[^"]+)"',
+            webpage, 'config url', default=None)
+        config_params = {}
+
+        if not config_url:
+            player = self._parse_json(
+                self._search_regex(
+                    r'(?s)Player\s*\(.+?\s*,\s*({.+?\bmedia\b["\']?\s*:\s*["\']?\d+.+?})\s*\)', webpage,
+                    'player', default='{}'),
+                video_id, transform_source=js_to_json, fatal=False)
+            if player:
+                config_url = url_or_none(player.get('configUrl'))
+                params = player.get('configParams')
+                if isinstance(params, dict):
+                    config_params = params
+
+        if not config_url:
+            DEFAULT_SITE_ID = '23000'
+            SITES = {
+                'tvnoviny': DEFAULT_SITE_ID,
+                'novaplus': DEFAULT_SITE_ID,
+                'vymena': DEFAULT_SITE_ID,
+                'krasna': DEFAULT_SITE_ID,
+                'fanda': '30',
+                'tn': '30',
+                'doma': '30',
+            }
+
+            site_id = self._search_regex(
+                r'site=(\d+)', webpage, 'site id', default=None) or SITES.get(
+                site, DEFAULT_SITE_ID)
+
+            config_url = 'https://api.nova.cz/bin/player/videojs/config.php'
+            config_params = {
+                'site': site_id,
+                'media': video_id,
+                'quality': 3,
+                'version': 1,
+            }
+
+        config = self._download_json(
+            config_url, display_id,
+            'Downloading config JSON', query=config_params,
+            transform_source=lambda s: s[s.index('{'):s.rindex('}') + 1])
+
+        mediafile = config['mediafile']
+        video_url = mediafile['src']
+
+        m = re.search(r'^(?P<url>rtmpe?://[^/]+/(?P<app>[^/]+?))/&*(?P<playpath>.+)$', video_url)
+        if m:
+            formats = [{
+                'url': m.group('url'),
+                'app': m.group('app'),
+                'play_path': m.group('playpath'),
+                'player_path': 'http://tvnoviny.nova.cz/static/shared/app/videojs/video-js.swf',
+                'ext': 'flv',
+            }]
+        else:
+            formats = [{
+                'url': video_url,
+            }]
+
+        title = mediafile.get('meta', {}).get('title') or self._og_search_title(webpage)
+        thumbnail = config.get('poster')
+
+        return {
+            'id': video_id,
+            'display_id': display_id,
+            'title': title,
+            'description': description,
+            'upload_date': upload_date,
+            'thumbnail': thumbnail,
+            'formats': formats,
+        }

yt_dlp/extractor/youtube.py

@@ -3317,7 +3317,36 @@ def _extract_heatmap(self, data):
                 'value': ('intensityScoreNormalized', {float_or_none}),
             })) or None
 
-    def _extract_comment(self, comment_renderer, parent=None):
+    def _extract_comment(self, entities, parent=None):
+        comment_entity_payload = get_first(entities, ('payload', 'commentEntityPayload', {dict}))
+        if not (comment_id := traverse_obj(comment_entity_payload, ('properties', 'commentId', {str}))):
+            return
+
+        toolbar_entity_payload = get_first(entities, ('payload', 'engagementToolbarStateEntityPayload', {dict}))
+        time_text = traverse_obj(comment_entity_payload, ('properties', 'publishedTime', {str})) or ''
+
+        return {
+            'id': comment_id,
+            'parent': parent or 'root',
+            **traverse_obj(comment_entity_payload, {
+                'text': ('properties', 'content', 'content', {str}),
+                'like_count': ('toolbar', 'likeCountA11y', {parse_count}),
+                'author_id': ('author', 'channelId', {self.ucid_or_none}),
+                'author': ('author', 'displayName', {str}),
+                'author_thumbnail': ('author', 'avatarThumbnailUrl', {url_or_none}),
+                'author_is_uploader': ('author', 'isCreator', {bool}),
+                'author_is_verified': ('author', 'isVerified', {bool}),
+                'author_url': ('author', 'channelCommand', 'innertubeCommand', (
+                    ('browseEndpoint', 'canonicalBaseUrl'), ('commandMetadata', 'webCommandMetadata', 'url')
+                ), {lambda x: urljoin('https://www.youtube.com', x)}),
+            }, get_all=False),
+            'is_favorited': (None if toolbar_entity_payload is None else
+                             toolbar_entity_payload.get('heartState') == 'TOOLBAR_HEART_STATE_HEARTED'),
+            '_time_text': time_text,  # FIXME: non-standard, but we need a way of showing that it is an estimate.
+            'timestamp': self._parse_time_text(time_text),
+        }
+
+    def _extract_comment_old(self, comment_renderer, parent=None):
         comment_id = comment_renderer.get('commentId')
         if not comment_id:
             return
@@ -3398,21 +3427,39 @@ def extract_header(contents):
                 break
             return _continuation
 
-        def extract_thread(contents):
+        def extract_thread(contents, entity_payloads):
             if not parent:
                 tracker['current_page_thread'] = 0
             for content in contents:
                 if not parent and tracker['total_parent_comments'] >= max_parents:
                     yield
                 comment_thread_renderer = try_get(content, lambda x: x['commentThreadRenderer'])
-                comment_renderer = get_first(
-                    (comment_thread_renderer, content), [['commentRenderer', ('comment', 'commentRenderer')]],
-                    expected_type=dict, default={})
 
-                comment = self._extract_comment(comment_renderer, parent)
+                # old comment format
+                if not entity_payloads:
+                    comment_renderer = get_first(
+                        (comment_thread_renderer, content), [['commentRenderer', ('comment', 'commentRenderer')]],
+                        expected_type=dict, default={})
+
+                    comment = self._extract_comment_old(comment_renderer, parent)
+
+                # new comment format
+                else:
+                    view_model = (
+                        traverse_obj(comment_thread_renderer, ('commentViewModel', 'commentViewModel', {dict}))
+                        or traverse_obj(content, ('commentViewModel', {dict})))
+                    comment_keys = traverse_obj(view_model, (('commentKey', 'toolbarStateKey'), {str}))
+                    if not comment_keys:
+                        continue
+                    entities = traverse_obj(entity_payloads, lambda _, v: v['entityKey'] in comment_keys)
+                    comment = self._extract_comment(entities, parent)
+                    if comment:
+                        comment['is_pinned'] = traverse_obj(view_model, ('pinnedText', {str})) is not None
+
                 if not comment:
                     continue
                 comment_id = comment['id']
+
                 if comment.get('is_pinned'):
                     tracker['pinned_comment_ids'].add(comment_id)
                 # Sometimes YouTube may break and give us infinite looping comments.
@@ -3505,7 +3552,7 @@ def extract_thread(contents):
             check_get_keys = None
             if not is_forced_continuation and not (tracker['est_total'] == 0 and tracker['running_total'] == 0):
                 check_get_keys = [[*continuation_items_path, ..., (
-                    'commentsHeaderRenderer' if is_first_continuation else ('commentThreadRenderer', 'commentRenderer'))]]
+                    'commentsHeaderRenderer' if is_first_continuation else ('commentThreadRenderer', 'commentViewModel', 'commentRenderer'))]]
             try:
                 response = self._extract_response(
                     item_id=None, query=continuation,
@@ -3529,6 +3576,7 @@ def extract_thread(contents):
                 raise
             is_forced_continuation = False
             continuation = None
+            mutations = traverse_obj(response, ('frameworkUpdates', 'entityBatchUpdate', 'mutations', ..., {dict}))
             for continuation_items in traverse_obj(response, continuation_items_path, expected_type=list, default=[]):
                 if is_first_continuation:
                     continuation = extract_header(continuation_items)
@@ -3537,7 +3585,7 @@ def extract_thread(contents):
                         break
                     continue
 
-                for entry in extract_thread(continuation_items):
+                for entry in extract_thread(continuation_items, mutations):
                     if not entry:
                         return
                     yield entry