[ie/youtube] Detect and warn when account cookies are rotated (#13014)

Related: https://github.com/yt-dlp/yt-dlp/issues/8227

Authored by: coletdjnz

yt_dlp/extractor/twitter.py

@@ -1221,20 +1221,10 @@ class TwitterIE(TwitterBaseIE):
     }]
 
     _MEDIA_ID_RE = re.compile(r'_video/(\d+)/')
-
-    @property
-    def _GRAPHQL_ENDPOINT(self):
-        if self.is_logged_in:
-            return 'zZXycP0V6H7m-2r0mOnFcA/TweetDetail'
-        return '2ICDjqPd81tulZcYrtpTuQ/TweetResultByRestId'
+    _GRAPHQL_ENDPOINT = '2ICDjqPd81tulZcYrtpTuQ/TweetResultByRestId'
 
     def _graphql_to_legacy(self, data, twid):
-        result = traverse_obj(data, (
-            'threaded_conversation_with_injections_v2', 'instructions', 0, 'entries',
-            lambda _, v: v['entryId'] == f'tweet-{twid}', 'content', 'itemContent',
-            'tweet_results', 'result', ('tweet', None), {dict},
-        ), default={}, get_all=False) if self.is_logged_in else traverse_obj(
-            data, ('tweetResult', 'result', {dict}), default={})
+        result = traverse_obj(data, ('tweetResult', 'result', {dict})) or {}
 
         typename = result.get('__typename')
         if typename not in ('Tweet', 'TweetWithVisibilityResults', 'TweetTombstone', 'TweetUnavailable', None):
@@ -1278,37 +1268,6 @@ class TwitterIE(TwitterBaseIE):
 
     def _build_graphql_query(self, media_id):
         return {
-            'variables': {
-                'focalTweetId': media_id,
-                'includePromotedContent': True,
-                'with_rux_injections': False,
-                'withBirdwatchNotes': True,
-                'withCommunity': True,
-                'withDownvotePerspective': False,
-                'withQuickPromoteEligibilityTweetFields': True,
-                'withReactionsMetadata': False,
-                'withReactionsPerspective': False,
-                'withSuperFollowsTweetFields': True,
-                'withSuperFollowsUserFields': True,
-                'withV2Timeline': True,
-                'withVoice': True,
-            },
-            'features': {
-                'graphql_is_translatable_rweb_tweet_is_translatable_enabled': False,
-                'interactive_text_enabled': True,
-                'responsive_web_edit_tweet_api_enabled': True,
-                'responsive_web_enhance_cards_enabled': True,
-                'responsive_web_graphql_timeline_navigation_enabled': False,
-                'responsive_web_text_conversations_enabled': False,
-                'responsive_web_uc_gql_enabled': True,
-                'standardized_nudges_misinfo': True,
-                'tweet_with_visibility_results_prefer_gql_limited_actions_policy_enabled': False,
-                'tweetypie_unmention_optimization_enabled': True,
-                'unified_cards_ad_metadata_container_dynamic_card_content_query_enabled': True,
-                'verified_phone_label_enabled': False,
-                'vibe_api_enabled': True,
-            },
-        } if self.is_logged_in else {
             'variables': {
                 'tweetId': media_id,
                 'withCommunity': False,

yt_dlp/extractor/youtube/_base.py

@@ -417,6 +417,8 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
     _NETRC_MACHINE = 'youtube'
 
+    _COOKIE_HOWTO_WIKI_URL = 'https://github.com/yt-dlp/yt-dlp/wiki/Extractors#exporting-youtube-cookies'
+
     def ucid_or_none(self, ucid):
         return self._search_regex(rf'^({self._YT_CHANNEL_UCID_RE})$', ucid, 'UC-id', default=None)
 
@@ -451,17 +453,15 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
         return preferred_lang
 
     def _initialize_consent(self):
-        cookies = self._get_cookies('https://www.youtube.com/')
-        if cookies.get('__Secure-3PSID'):
+        if self._has_auth_cookies:
             return
-        socs = cookies.get('SOCS')
+        socs = self._youtube_cookies.get('SOCS')
         if socs and not socs.value.startswith('CAA'):  # not consented
             return
         self._set_cookie('.youtube.com', 'SOCS', 'CAI', secure=True)  # accept all (required for mixes)
 
     def _initialize_pref(self):
-        cookies = self._get_cookies('https://www.youtube.com/')
-        pref_cookie = cookies.get('PREF')
+        pref_cookie = self._youtube_cookies.get('PREF')
         pref = {}
         if pref_cookie:
             try:
@@ -472,8 +472,9 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
         self._set_cookie('.youtube.com', name='PREF', value=urllib.parse.urlencode(pref))
 
     def _initialize_cookie_auth(self):
-        yt_sapisid, yt_1psapisid, yt_3psapisid = self._get_sid_cookies()
-        if yt_sapisid or yt_1psapisid or yt_3psapisid:
+        self._passed_auth_cookies = False
+        if self._has_auth_cookies:
+            self._passed_auth_cookies = True
             self.write_debug('Found YouTube account cookies')
 
     def _real_initialize(self):
@@ -492,8 +493,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
     @property
     def _youtube_login_hint(self):
-        return (f'{self._login_hint(method="cookies")}. Also see  '
-                'https://github.com/yt-dlp/yt-dlp/wiki/Extractors#exporting-youtube-cookies  '
+        return (f'{self._login_hint(method="cookies")}. Also see  {self._COOKIE_HOWTO_WIKI_URL}  '
                 'for tips on effectively exporting YouTube cookies')
 
     def _check_login_required(self):
@@ -553,12 +553,16 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
         return f'{scheme} {"_".join(parts)}'
 
+    @property
+    def _youtube_cookies(self):
+        return self._get_cookies('https://www.youtube.com')
+
     def _get_sid_cookies(self):
         """
         Get SAPISID, 1PSAPISID, 3PSAPISID cookie values
         @returns sapisid, 1psapisid, 3psapisid
         """
-        yt_cookies = self._get_cookies('https://www.youtube.com')
+        yt_cookies = self._youtube_cookies
         yt_sapisid = try_call(lambda: yt_cookies['SAPISID'].value)
         yt_3papisid = try_call(lambda: yt_cookies['__Secure-3PAPISID'].value)
         yt_1papisid = try_call(lambda: yt_cookies['__Secure-1PAPISID'].value)
@@ -595,6 +599,31 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
         return ' '.join(authorizations)
 
+    @property
+    def is_authenticated(self):
+        return self._has_auth_cookies
+
+    @property
+    def _has_auth_cookies(self):
+        yt_sapisid, yt_1psapisid, yt_3psapisid = self._get_sid_cookies()
+        # YouTube doesn't appear to clear 3PSAPISID when rotating cookies (as of 2025-04-26)
+        # But LOGIN_INFO is cleared and should exist if logged in
+        has_login_info = 'LOGIN_INFO' in self._youtube_cookies
+        return bool(has_login_info and (yt_sapisid or yt_1psapisid or yt_3psapisid))
+
+    def _request_webpage(self, *args, **kwargs):
+        response = super()._request_webpage(*args, **kwargs)
+
+        # Check that we are still logged-in and cookies have not rotated after every request
+        if getattr(self, '_passed_auth_cookies', None) and not self._has_auth_cookies:
+            self.report_warning(
+                'The provided YouTube account cookies are no longer valid. '
+                'They have likely been rotated in the browser as a security measure. '
+                f'For tips on how to effectively export YouTube cookies, refer to  {self._COOKIE_HOWTO_WIKI_URL} .',
+                only_once=False)
+
+        return response
+
     def _call_api(self, ep, query, video_id, fatal=True, headers=None,
                   note='Downloading API JSON', errnote='Unable to download API page',
                   context=None, api_key=None, api_hostname=None, default_client='web'):
@@ -695,10 +724,6 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             args, [('VISITOR_DATA', ('INNERTUBE_CONTEXT', 'client', 'visitorData'), ('responseContext', 'visitorData'))],
             expected_type=str)
 
-    @functools.cached_property
-    def is_authenticated(self):
-        return bool(self._get_sid_authorization_header())
-
     def extract_ytcfg(self, video_id, webpage):
         if not webpage:
             return {}

yt_dlp/extractor/youtube/_video.py

@@ -1982,7 +1982,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
     def _player_js_cache_key(self, player_url):
         player_id = self._extract_player_info(player_url)
         player_path = remove_start(urllib.parse.urlparse(player_url).path, f'/s/player/{player_id}/')
-        variant = self._INVERSE_PLAYER_JS_VARIANT_MAP.get(player_path)
+        variant = self._INVERSE_PLAYER_JS_VARIANT_MAP.get(player_path) or next((
+            v for k, v in self._INVERSE_PLAYER_JS_VARIANT_MAP.items()
+            if re.fullmatch(re.escape(k).replace('en_US', r'[a-zA-Z0-9_]+'), player_path)), None)
         if not variant:
             self.write_debug(
                 f'Unable to determine player JS variant\n'