Merge 3302588c77 into 12d8ea8246

Closes #9554
Authored by: coletdjnz, bashonly

Co-authored-by: bashonly <88596187+bashonly@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
@@ -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/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/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/facebook.py

@@ -445,7 +445,9 @@ def _perform_login(self, username, password):
         try:
             login_results = self._download_webpage(request, None,
                                                    note='Logging in', errnote='unable to fetch login page')
-            if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
+            if 'Your Request Couldn' in login_results:
+                self.raise_login_required('Failed to login with credentials', method='cookies')
+            elif re.search(r'<form[^>]*name="login"[^<]*</form>', login_results):
                 error = self._html_search_regex(
                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
                     login_results, 'login error', default=None, group='error')
@@ -478,13 +480,41 @@ def _perform_login(self, username, password):
             return
 
     def _extract_from_url(self, url, video_id):
+        cookies = self._get_cookies(url)
+        # user passed logged-in cookies or attempted to login
+        login_data = cookies.get('c_user') and cookies.get('xs')
+        logged_in = False
+
         webpage = self._download_webpage(
             url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id)
 
+        sjs_data = [self._parse_json(j, video_id, fatal=False) for j in re.findall(
+            r'data-sjs>({.*?ScheduledServerJS.*?})</script>', webpage)]
+
+        if login_data:
+            logged_in = get_first(sjs_data, (
+                'require', ..., ..., ..., '__bbox', 'define',
+                lambda _, v: 'CurrentUserInitialData' in v, ..., 'ACCOUNT_ID'), default='0') != '0'
+            if logged_in and (info := get_first(sjs_data, ('require', ..., ..., ..., '__bbox', 'require', ..., ..., ...,
+                                                           '__bbox', 'result', 'data', (('ufac_client', 'state',
+                                                                                         (('set_contact_point_state_renderer', 'title'),
+                                                                                          ('intro_state_renderer', 'header_title'))),
+                                                                                        ('epsilon_checkpoint', 'screen', 'title'))))):
+                if any(content in info for content in ['days left to appeal', 'suspended your account']):
+                    raise ExtractorError('Your account is suspended', expected=True)
+                if 'Enter mobile number' == info:
+                    raise ExtractorError('Facebook is requiring mobile number confirmation', expected=True)
+                if 'your account has been locked' in info:
+                    raise ExtractorError('Your account has been locked', expected=True)
+
+        if props := get_first(sjs_data, (
+                'require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., 'rootView',
+                lambda _, v: v['title'].startswith('This content isn\'t available'))):
+            raise ExtractorError(
+                f'Content unavailable. Facebook said: {props.get("body") or props["title"]}', expected=True)
+
         def extract_metadata(webpage):
-            post_data = [self._parse_json(j, video_id, fatal=False) for j in re.findall(
-                r'data-sjs>({.*?ScheduledServerJS.*?})</script>', webpage)]
-            post = traverse_obj(post_data, (
+            post = traverse_obj(sjs_data, (
                 ..., 'require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., '__bbox', 'result', 'data'), expected_type=dict) or []
             media = traverse_obj(post, (..., 'attachments', ..., lambda k, v: (
                 k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')), expected_type=dict)
@@ -795,6 +825,10 @@ def parse_attachment(attachment, key='media'):
             video_data = extract_from_jsmods_instances(tahoe_js_data)
 
         if not video_data:
+            if not login_data:
+                raise ExtractorError('Cannot parse data. Try logging in.', expected=True)
+            if not logged_in:
+                raise ExtractorError('Failed to login with provided data.', expected=True)
             raise ExtractorError('Cannot parse data')
 
         if len(video_data) > 1:

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']}},
             },
         },
     ]
@@ -3317,7 +3318,36 @@ 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 +3428,39 @@ 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 +3553,7 @@ 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 +3577,7 @@ 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 +3586,7 @@ 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
@@ -3614,8 +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:
@@ -3631,19 +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
 
@@ -3862,6 +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 = {
@@ -3874,7 +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
@@ -3892,8 +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 '')