Merge 6608d38afe into 96da952504

Authored by: seproDev, pukkandan

test/test_networking.py

@@ -785,6 +785,25 @@ class TestHTTPImpersonateRequestHandler(TestRequestHandlerBase):
                 assert res.status == 200
                 assert std_headers['user-agent'].lower() not in res.read().decode().lower()
 
+    def test_response_extensions(self, handler):
+        with handler() as rh:
+            for target in rh.supported_targets:
+                request = Request(
+                    f'http://127.0.0.1:{self.http_port}/gen_200', extensions={'impersonate': target})
+                res = validate_and_send(rh, request)
+                assert res.extensions['impersonate'] == rh._get_request_target(request)
+
+    def test_http_error_response_extensions(self, handler):
+        with handler() as rh:
+            for target in rh.supported_targets:
+                request = Request(
+                    f'http://127.0.0.1:{self.http_port}/gen_404', extensions={'impersonate': target})
+                try:
+                    validate_and_send(rh, request)
+                except HTTPError as e:
+                    res = e.response
+                assert res.extensions['impersonate'] == rh._get_request_target(request)
+
 
 class TestRequestHandlerMisc:
     """Misc generic tests for request handlers, not related to request or validation testing"""

yt_dlp/YoutubeDL.py

@@ -2136,6 +2136,11 @@ class YoutubeDL:
 
     def _check_formats(self, formats):
         for f in formats:
+            working = f.get('__working')
+            if working is not None:
+                if working:
+                    yield f
+                continue
             self.to_screen('[info] Testing format %s' % f['format_id'])
             path = self.get_output_path('temp')
             if not self._ensure_dir_exists(f'{path}/'):
@@ -2152,33 +2157,44 @@ class YoutubeDL:
                         os.remove(temp_file.name)
                     except OSError:
                         self.report_warning('Unable to delete temporary file "%s"' % temp_file.name)
+            f['__working'] = success
             if success:
                 yield f
             else:
                 self.to_screen('[info] Unable to download format %s. Skipping...' % f['format_id'])
 
+    def _select_formats(self, formats, selector):
+        return list(selector({
+            'formats': formats,
+            'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
+            'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats)  # No formats with video
+                                   or all(f.get('acodec') == 'none' for f in formats)),  # OR, No formats with audio
+        }))
+
     def _default_format_spec(self, info_dict, download=True):
+        download = download and not self.params.get('simulate')
+        prefer_best = download and (
+            self.params['outtmpl']['default'] == '-'
+            or info_dict.get('is_live') and not self.params.get('live_from_start'))
 
         def can_merge():
             merger = FFmpegMergerPP(self)
             return merger.available and merger.can_merge()
 
-        prefer_best = (
-            not self.params.get('simulate')
-            and download
-            and (
-                not can_merge()
-                or info_dict.get('is_live') and not self.params.get('live_from_start')
-                or self.params['outtmpl']['default'] == '-'))
-        compat = (
-            prefer_best
-            or self.params.get('allow_multiple_audio_streams', False)
-            or 'format-spec' in self.params['compat_opts'])
+        if not prefer_best and download and not can_merge():
+            prefer_best = True
+            formats = self._get_formats(info_dict)
+            evaluate_formats = lambda spec: self._select_formats(formats, self.build_format_selector(spec))
+            if evaluate_formats('b/bv+ba') != evaluate_formats('bv*+ba/b'):
+                self.report_warning('ffmpeg not found. The downloaded format may not be the best available. '
+                                    'Installing ffmpeg is strongly recommended: https://github.com/yt-dlp/yt-dlp#dependencies')
 
-        return (
-            'best/bestvideo+bestaudio' if prefer_best
-            else 'bestvideo*+bestaudio/best' if not compat
-            else 'bestvideo+bestaudio/best')
+        compat = (self.params.get('allow_multiple_audio_streams')
+                  or 'format-spec' in self.params['compat_opts'])
+
+        return ('best/bestvideo+bestaudio' if prefer_best
+                else 'bestvideo+bestaudio/best' if compat
+                else 'bestvideo*+bestaudio/best')
 
     def build_format_selector(self, format_spec):
         def syntax_error(note, start):
@@ -2928,12 +2944,7 @@ class YoutubeDL:
                 self.write_debug(f'Default format spec: {req_format}')
                 format_selector = self.build_format_selector(req_format)
 
-            formats_to_download = list(format_selector({
-                'formats': formats,
-                'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
-                'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats)  # No formats with video
-                                       or all(f.get('acodec') == 'none' for f in formats)),  # OR, No formats with audio
-            }))
+            formats_to_download = self._select_formats(formats, format_selector)
             if interactive_format_selection and not formats_to_download:
                 self.report_error('Requested format is not available', tb=False, is_error=False)
                 continue

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',
             '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()],
     }, {
@@ -836,24 +865,26 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             'alt_title': 'The downsides of positive thinking',
             'description': 'md5:fad74b31da60d83b8265954ee42d85b4',
             'duration': 235,
-            'thumbnail': r're:https?://.+/p07c9dsr.jpg',
-            'upload_date': '20190604',
+            'thumbnail': r're:https?://.+/p07c9dsr\.(?:jpg|webp|png)',
+            'upload_date': '20220223',
             'categories': ['Psychology'],
         },
     }, {
         # 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 .{2153} 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',
+        }
     }, {  # onion routes
         'url': 'https://www.bbcnewsd73hkzno2ini43t4gblxvycyac5aw4gnv7t2rccijh7745uqd.onion/news/av/world-europe-63208576',
         'only_matching': True,
@@ -1008,8 +1039,8 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             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}',
+                ie=BBCCoUkIE)
 
         # single video story (e.g. http://www.bbc.com/travel/story/20150625-sri-lankas-spicy-secret)
         programme_id = self._search_regex(
@@ -1069,83 +1100,133 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
                 }
 
         # 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'\{(?:(?!</script>)[\s\S])+?(?:"leadMedia"|\\"videoData\\")\s*:(?:(?!</script>)[\s\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 +1272,30 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
                 return self.playlist_result(
                     entries, playlist_id, playlist_title, playlist_description)
 
+        k_int_or_none = functools.partial(int_or_none, scale=1000)
+
+        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}, any),
+                    'duration': ('versions', 0, 'duration', {int}),
+                    'timestamp': ('versions', 0, 'availableFrom', {k_int_or_none}),
+                })
+            }
+
+        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 +1307,19 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
             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
@@ -1235,26 +1353,89 @@ class BBCIE(BBCCoUkIE):  # XXX: Do not subclass from concrete IE
                         'timestamp': item_time,
                         'description': strip_or_none(item_desc),
                     })
-            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', {k_int_or_none}),
+                    }))
+                    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', {k_int_or_none}),
+                            }),
+                        }
+                        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:
+            if entry := parse_model(model):
+                if not entry.get('timestamp'):
+                    entry['timestamp'] = traverse_obj(next_data, (
+                        ..., 'contents', is_type('timestamp'),
+                        'model', 'timestamp', {k_int_or_none}, 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/patreon.py

@@ -219,7 +219,29 @@ class PatreonIE(PatreonBaseIE):
             'thumbnail': r're:^https?://.+',
         },
         'params': {'skip_download': 'm3u8'},
+    }, {
+        # multiple attachments/embeds
+        'url': 'https://www.patreon.com/posts/holy-wars-solos-100601977',
+        'playlist_count': 3,
+        'info_dict': {
+            'id': '100601977',
+            'title': '"Holy Wars" (Megadeth) Solos Transcription & Lesson/Analysis',
+            'description': 'md5:d099ab976edfce6de2a65c2b169a88d3',
+            'uploader': 'Bradley Hall',
+            'uploader_id': '24401883',
+            'uploader_url': 'https://www.patreon.com/bradleyhallguitar',
+            'channel_id': '3193932',
+            'channel_url': 'https://www.patreon.com/bradleyhallguitar',
+            'channel_follower_count': int,
+            'timestamp': 1710777855,
+            'upload_date': '20240318',
+            'like_count': int,
+            'comment_count': int,
+            'thumbnail': r're:^https?://.+',
+        },
+        'skip': 'Patron-only content',
     }]
+    _RETURN_TYPE = 'video'
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
@@ -234,58 +256,54 @@ class PatreonIE(PatreonBaseIE):
                 'include': 'audio,user,user_defined_tags,campaign,attachments_media',
             })
         attributes = post['data']['attributes']
-        title = attributes['title'].strip()
-        image = attributes.get('image') or {}
-        info = {
-            'id': video_id,
-            'title': title,
-            'description': clean_html(attributes.get('content')),
-            'thumbnail': image.get('large_url') or image.get('url'),
-            'timestamp': parse_iso8601(attributes.get('published_at')),
-            'like_count': int_or_none(attributes.get('like_count')),
-            'comment_count': int_or_none(attributes.get('comment_count')),
-        }
-        can_view_post = traverse_obj(attributes, 'current_user_can_view')
-        if can_view_post and info['comment_count']:
-            info['__post_extractor'] = self.extract_comments(video_id)
+        info = traverse_obj(attributes, {
+            'title': ('title', {str.strip}),
+            'description': ('content', {clean_html}),
+            'thumbnail': ('image', ('large_url', 'url'), {url_or_none}, any),
+            'timestamp': ('published_at', {parse_iso8601}),
+            'like_count': ('like_count', {int_or_none}),
+            'comment_count': ('comment_count', {int_or_none}),
+        })
 
-        for i in post.get('included', []):
-            i_type = i.get('type')
-            if i_type == 'media':
-                media_attributes = i.get('attributes') or {}
-                download_url = media_attributes.get('download_url')
+        entries = []
+        idx = 0
+        for include in traverse_obj(post, ('included', lambda _, v: v['type'])):
+            include_type = include['type']
+            if include_type == 'media':
+                media_attributes = traverse_obj(include, ('attributes', {dict})) or {}
+                download_url = url_or_none(media_attributes.get('download_url'))
                 ext = mimetype2ext(media_attributes.get('mimetype'))
 
                 # if size_bytes is None, this media file is likely unavailable
                 # See: https://github.com/yt-dlp/yt-dlp/issues/4608
                 size_bytes = int_or_none(media_attributes.get('size_bytes'))
                 if download_url and ext in KNOWN_EXTENSIONS and size_bytes is not None:
-                    # XXX: what happens if there are multiple attachments?
-                    return {
-                        **info,
+                    idx += 1
+                    entries.append({
+                        'id': f'{video_id}-{idx}',
                         'ext': ext,
                         'filesize': size_bytes,
                         'url': download_url,
-                    }
-            elif i_type == 'user':
-                user_attributes = i.get('attributes')
-                if user_attributes:
-                    info.update({
-                        'uploader': user_attributes.get('full_name'),
-                        'uploader_id': str_or_none(i.get('id')),
-                        'uploader_url': user_attributes.get('url'),
                     })
 
-            elif i_type == 'post_tag':
-                info.setdefault('tags', []).append(traverse_obj(i, ('attributes', 'value')))
+            elif include_type == 'user':
+                info.update(traverse_obj(include, {
+                    'uploader': ('attributes', 'full_name', {str}),
+                    'uploader_id': ('id', {str_or_none}),
+                    'uploader_url': ('attributes', 'url', {url_or_none}),
+                }))
 
-            elif i_type == 'campaign':
-                info.update({
-                    'channel': traverse_obj(i, ('attributes', 'title')),
-                    'channel_id': str_or_none(i.get('id')),
-                    'channel_url': traverse_obj(i, ('attributes', 'url')),
-                    'channel_follower_count': int_or_none(traverse_obj(i, ('attributes', 'patron_count'))),
-                })
+            elif include_type == 'post_tag':
+                if post_tag := traverse_obj(include, ('attributes', 'value', {str})):
+                    info.setdefault('tags', []).append(post_tag)
+
+            elif include_type == 'campaign':
+                info.update(traverse_obj(include, {
+                    'channel': ('attributes', 'title', {str}),
+                    'channel_id': ('id', {str_or_none}),
+                    'channel_url': ('attributes', 'url', {url_or_none}),
+                    'channel_follower_count': ('attributes', 'patron_count', {int_or_none}),
+                }))
 
         # handle Vimeo embeds
         if traverse_obj(attributes, ('embed', 'provider')) == 'Vimeo':
@@ -296,36 +314,50 @@ class PatreonIE(PatreonBaseIE):
                     v_url, video_id, 'Checking Vimeo embed URL',
                     headers={'Referer': 'https://patreon.com/'},
                     fatal=False, errnote=False):
-                return self.url_result(
+                entries.append(self.url_result(
                     VimeoIE._smuggle_referrer(v_url, 'https://patreon.com/'),
-                    VimeoIE, url_transparent=True, **info)
+                    VimeoIE, url_transparent=True))
 
         embed_url = traverse_obj(attributes, ('embed', 'url', {url_or_none}))
         if embed_url and self._request_webpage(embed_url, video_id, 'Checking embed URL', fatal=False, errnote=False):
-            return self.url_result(embed_url, **info)
+            entries.append(self.url_result(embed_url))
 
-        post_file = traverse_obj(attributes, 'post_file')
+        post_file = traverse_obj(attributes, ('post_file', {dict}))
         if post_file:
             name = post_file.get('name')
             ext = determine_ext(name)
             if ext in KNOWN_EXTENSIONS:
-                return {
-                    **info,
+                entries.append({
+                    'id': video_id,
                     'ext': ext,
                     'url': post_file['url'],
-                }
+                })
             elif name == 'video' or determine_ext(post_file.get('url')) == 'm3u8':
                 formats, subtitles = self._extract_m3u8_formats_and_subtitles(post_file['url'], video_id)
-                return {
-                    **info,
+                entries.append({
+                    'id': video_id,
                     'formats': formats,
                     'subtitles': subtitles,
-                }
+                })
 
-        if can_view_post is False:
+        can_view_post = traverse_obj(attributes, 'current_user_can_view')
+        comments = None
+        if can_view_post and info.get('comment_count'):
+            comments = self.extract_comments(video_id)
+
+        if not entries and can_view_post is False:
             self.raise_no_formats('You do not have access to this post', video_id=video_id, expected=True)
-        else:
+        elif not entries:
             self.raise_no_formats('No supported media found in this post', video_id=video_id, expected=True)
+        elif len(entries) == 1:
+            info.update(entries[0])
+        else:
+            for entry in entries:
+                entry.update(info)
+            return self.playlist_result(entries, video_id, **info, __post_extractor=comments)
+
+        info['id'] = video_id
+        info['__post_extractor'] = comments
         return info
 
     def _get_comments(self, post_id):

yt_dlp/networking/_curlcffi.py

@@ -132,6 +132,16 @@ class CurlCFFIRH(ImpersonateRequestHandler, InstanceStoreMixin):
         extensions.pop('cookiejar', None)
         extensions.pop('timeout', None)
 
+    def send(self, request: Request) -> Response:
+        target = self._get_request_target(request)
+        try:
+            response = super().send(request)
+        except HTTPError as e:
+            e.response.extensions['impersonate'] = target
+            raise
+        response.extensions['impersonate'] = target
+        return response
+
     def _send(self, request: Request):
         max_redirects_exceeded = False
         session: curl_cffi.requests.Session = self._get_instance(

yt_dlp/networking/common.py

@@ -497,6 +497,7 @@ class Response(io.IOBase):
     @param headers: response headers.
     @param status: Response HTTP status code. Default is 200 OK.
     @param reason: HTTP status reason. Will use built-in reasons based on status code if not provided.
+    @param extensions: Dictionary of handler-specific response extensions.
     """
 
     def __init__(
@@ -505,7 +506,9 @@ class Response(io.IOBase):
             url: str,
             headers: Mapping[str, str],
             status: int = 200,
-            reason: str = None):
+            reason: str = None,
+            extensions: dict = None
+    ):
 
         self.fp = fp
         self.headers = Message()
@@ -517,6 +520,7 @@ class Response(io.IOBase):
             self.reason = reason or HTTPStatus(status).phrase
         except ValueError:
             self.reason = None
+        self.extensions = extensions or {}
 
     def readable(self):
         return self.fp.readable()