Merge 43655bc01b into 12d8ea8246

Closes #9554
Authored by: coletdjnz, bashonly

Co-authored-by: bashonly <88596187+bashonly@users.noreply.github.com>

README.md

@@ -1760,7 +1760,7 @@ # EXTRACTOR ARGUMENTS
 #### youtube
 * `lang`: Prefer translated metadata (`title`, `description` etc) of this language code (case-sensitive). By default, the video primary language metadata is preferred, with a fallback to `en` translated. See [youtube.py](https://github.com/yt-dlp/yt-dlp/blob/c26f9b991a0681fd3ea548d535919cec1fbbd430/yt_dlp/extractor/youtube.py#L381-L390) for list of supported content language codes
 * `skip`: One or more of `hls`, `dash` or `translated_subs` to skip extraction of the m3u8 manifests, dash manifests and [auto-translated subtitles](https://github.com/yt-dlp/yt-dlp/issues/4090#issuecomment-1158102032) respectively
-* `player_client`: Clients to extract video data from. The main clients are `web`, `android` and `ios` with variants `_music`, `_embedded`, `_embedscreen`, `_creator` (e.g. `web_embedded`); and `mweb`, `mweb_embedscreen`, `mediaconnect` and `tv_embedded` (agegate bypass) with no variants. By default, `ios,android,web` is used, but `tv_embedded` and `creator` variants are added as required for age-gated videos. Similarly, the music variants are added for `music.youtube.com` urls. You can use `all` to use all the clients, and `default` for the default clients.
+* `player_client`: Clients to extract video data from. The main clients are `web`, `ios` and `android`, with variants `_music`, `_embedded`, `_embedscreen`, `_creator` (e.g. `web_embedded`); and `mweb`, `mweb_embedscreen` and `tv_embedded` (agegate bypass) with no variants. By default, `ios,web` is used, but `tv_embedded` and `creator` variants are added as required for age-gated videos. Similarly, the music variants are added for `music.youtube.com` urls. The `android` clients will always be given lowest priority since their formats are broken. You can use `all` to use all the clients, and `default` for the default clients.
 * `player_skip`: Skip some network requests that are generally needed for robust extraction. One or more of `configs` (skip client configs), `webpage` (skip initial webpage), `js` (skip js player). While these options can help reduce the number of requests needed or avoid some rate-limiting, they could cause some issues. See [#860](https://github.com/yt-dlp/yt-dlp/pull/860) for more details
 * `player_params`: YouTube player parameters to use for player requests. Will overwrite any default ones set by yt-dlp.
 * `comment_sort`: `top` or `new` (default) - choose comment sorting mode (on YouTube's side)

yt_dlp/extractor/pornhub.py

@@ -138,12 +138,13 @@ class PornHubIE(PornHubBaseIE):
     _EMBED_REGEX = [r'<iframe[^>]+?src=["\'](?P<url>(?:https?:)?//(?:www\.)?pornhub(?:premium)?\.(?:com|net|org)/embed/[\da-z]+)']
     _TESTS = [{
         'url': 'http://www.pornhub.com/view_video.php?viewkey=648719015',
-        'md5': 'a6391306d050e4547f62b3f485dd9ba9',
+        'md5': '4d4a4e9178b655776f86cf89ecaf0edf',
         'info_dict': {
             'id': '648719015',
             'ext': 'mp4',
             'title': 'Seductive Indian beauty strips down and fingers her pink pussy',
-            'uploader': 'Babes',
+            'uploader_id': str,
+            'uploader': 'BABES-COM',
             'upload_date': '20130628',
             'timestamp': 1372447216,
             'duration': 361,
@@ -155,6 +156,8 @@ class PornHubIE(PornHubBaseIE):
             'tags': list,
             'categories': list,
             'cast': list,
+            'thumbnail': str,
+            'production': list,
         },
     }, {
         # non-ASCII title
@@ -208,11 +211,23 @@ class PornHubIE(PornHubBaseIE):
         'url': 'http://www.pornhub.com/view_video.php?viewkey=ph601dc30bae19a',
         'info_dict': {
             'id': 'ph601dc30bae19a',
+            'ext': 'mp4',
             'uploader': 'Projekt Melody',
             'uploader_id': 'projekt-melody',
             'upload_date': '20210205',
             'title': '"Welcome to My Pussy Mansion" - CB Stream (02/03/21)',
             'thumbnail': r're:https?://.+',
+            'duration': 8173,
+            'timestamp': 1612564932,
+            'age_limit': 18,
+            'tags': list,
+            'like_count': int,
+            'view_count': int,
+            'categories': list,
+            'dislike_count': int,
+            'production': list,
+            'cast': list,
+            'comment_count': int,
         },
     }, {
         'url': 'http://www.pornhub.com/view_video.php?viewkey=ph557bbb6676d2d',
@@ -263,7 +278,33 @@ class PornHubIE(PornHubBaseIE):
     }, {
         'url': 'http://pornhubvybmsymdol4iibwgwtkpwmeyd6luq2gxajgjzfjvotyt5zhyd.onion/view_video.php?viewkey=ph5a9813bfa7156',
         'only_matching': True,
-    }]
+    }, {
+        # language spoken, model attributes, production
+        'url': 'https://www.pornhub.com/view_video.php?viewkey=65a6ca42725f2',
+        'info_dict': {
+            'id': '65a6ca42725f2',
+            'ext': 'mp4',
+            'title': 'Busty blonde sucks the juice out of me',
+            'uploader_id': 'egon-kowalski',
+            'cast': [],
+            'thumbnail': 'https://ei.phncdn.com/videos/202401/16/446618441/thumbs_14/(m=eaAaGwObaaaa)(mh=dIONqa8IT3Sw1RqL)5.jpg',
+            'uploader': 'Egon Kowalski',
+            'upload_date': '20240116',
+            'timestamp': 1705429963,
+            'duration': 358,
+            'view_count': int,
+            'like_count': int,
+            'dislike_count': int,
+            'comment_count': int,
+            'age_limit': 18,
+            'tags': list,
+            'categories': list,
+            'language_spoken': ['german'],
+            'model_attributes': ['tattoos', 'white', 'no piercing'],
+            'production': ['homemade'],
+        }
+    }
+    ]
 
     def _extract_count(self, pattern, webpage, name):
         return str_to_int(self._search_regex(pattern, webpage, '%s count' % name, default=None))
@@ -488,6 +529,15 @@ def extract_list(meta_key):
             if div:
                 return [clean_html(x).strip() for x in re.findall(r'(?s)<a[^>]+\bhref=[^>]+>.+?</a>', div)]
 
+        def get_model_attributes():
+            model_attributes = re.findall(
+                r'<a[^>]*\bdata-label=["\']model_attributes["\'][^>]*>(.*?)<\/a>',
+                webpage,
+                re.DOTALL)
+
+            if model_attributes:
+                return [x.strip().lower() for x in model_attributes]
+
         info = self._search_json_ld(webpage, video_id, default={})
         # description provided in JSON-LD is irrelevant
         info['description'] = None
@@ -509,6 +559,9 @@ def extract_list(meta_key):
             'tags': extract_list('tags'),
             'categories': extract_list('categories'),
             'cast': extract_list('pornstars'),
+            'production': extract_list('production'),
+            'language_spoken': extract_list('langSpoken'),
+            'model_attributes': get_model_attributes(),
             'subtitles': subtitles,
         }, info)
 

yt_dlp/extractor/youtube.py

@@ -2353,6 +2353,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'format': '17',  # 3gp format available on android
                 'extractor_args': {'youtube': {'player_client': ['android']}},
             },
+            'skip': 'android client broken',
         },
         {
             # Skip download of additional client configs (remix client config in this case)
@@ -2730,7 +2731,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'heatmap': 'count:100',
             },
             'params': {
-                'extractor_args': {'youtube': {'player_client': ['android'], 'player_skip': ['webpage']}},
+                'extractor_args': {'youtube': {'player_client': ['ios'], 'player_skip': ['webpage']}},
             },
         },
     ]
@@ -3662,8 +3663,6 @@ def _extract_player_response(self, client, video_id, master_ytcfg, player_ytcfg,
         yt_query = {
             'videoId': video_id,
         }
-        if _split_innertube_client(client)[0] in ('android', 'android_embedscreen'):
-            yt_query['params'] = 'CgIIAQ=='
 
         pp_arg = self._configuration_arg('player_params', [None], casesense=True)[0]
         if pp_arg:
@@ -3679,19 +3678,24 @@ def _extract_player_response(self, client, video_id, master_ytcfg, player_ytcfg,
 
     def _get_requested_clients(self, url, smuggled_data):
         requested_clients = []
-        default = ['ios', 'android', 'web']
+        android_clients = []
+        default = ['ios', 'web']
         allowed_clients = sorted(
             (client for client in INNERTUBE_CLIENTS.keys() if client[:1] != '_'),
             key=lambda client: INNERTUBE_CLIENTS[client]['priority'], reverse=True)
         for client in self._configuration_arg('player_client'):
-            if client in allowed_clients:
-                requested_clients.append(client)
-            elif client == 'default':
+            if client == 'default':
                 requested_clients.extend(default)
             elif client == 'all':
                 requested_clients.extend(allowed_clients)
-            else:
+            elif client not in allowed_clients:
                 self.report_warning(f'Skipping unsupported client {client}')
+            elif client.startswith('android'):
+                android_clients.append(client)
+            else:
+                requested_clients.append(client)
+        # Force deprioritization of broken Android clients for format de-duplication
+        requested_clients.extend(android_clients)
         if not requested_clients:
             requested_clients = default
 
@@ -3910,6 +3914,14 @@ def build_fragments(f):
                     f'{video_id}: Some formats are possibly damaged. They will be deprioritized', only_once=True)
 
             client_name = fmt.get(STREAMING_DATA_CLIENT_NAME)
+            # Android client formats are broken due to integrity check enforcement
+            # Ref: https://github.com/yt-dlp/yt-dlp/issues/9554
+            is_broken = client_name and client_name.startswith(short_client_name('android'))
+            if is_broken:
+                self.report_warning(
+                    f'{video_id}: Android client formats are broken and may yield HTTP Error 403. '
+                    'They will be deprioritized', only_once=True)
+
             name = fmt.get('qualityLabel') or quality.replace('audio_quality_', '') or ''
             fps = int_or_none(fmt.get('fps')) or 0
             dct = {
@@ -3922,7 +3934,7 @@ def build_fragments(f):
                     name, fmt.get('isDrc') and 'DRC',
                     try_get(fmt, lambda x: x['projectionType'].replace('RECTANGULAR', '').lower()),
                     try_get(fmt, lambda x: x['spatialAudioType'].replace('SPATIAL_AUDIO_TYPE_', '').lower()),
-                    throttled and 'THROTTLED', is_damaged and 'DAMAGED',
+                    throttled and 'THROTTLED', is_damaged and 'DAMAGED', is_broken and 'BROKEN',
                     (self.get_param('verbose') or all_formats) and client_name,
                     delim=', '),
                 # Format 22 is likely to be damaged. See https://github.com/yt-dlp/yt-dlp/issues/3372
@@ -3940,8 +3952,8 @@ def build_fragments(f):
                 'language': join_nonempty(audio_track.get('id', '').split('.')[0],
                                           'desc' if language_preference < -1 else '') or None,
                 'language_preference': language_preference,
-                # Strictly de-prioritize damaged and 3gp formats
-                'preference': -10 if is_damaged else -2 if itag == '17' else None,
+                # Strictly de-prioritize broken, damaged and 3gp formats
+                'preference': -20 if is_broken else -10 if is_damaged else -2 if itag == '17' else None,
             }
             mime_mobj = re.match(
                 r'((?:[^/]+)/(?:[^;]+))(?:;\s*codecs="([^"]+)")?', fmt.get('mimeType') or '')