Merge e0556cd956b62b9ea10e2a9478c1e1fad856c6e6 into da7223d4aa42ff9fc680b0951d043dd03cec2d30

* improve extraction of global "useful data" Array from player JS
* also handle tv-player and add tests: thx seproDev (yt-dlp/yt-dlp#12684)

Co-Authored-By: sepro <sepro@sepr0.com>

test/test_youtube_signature.py

@@ -232,8 +232,32 @@ _NSIG_TESTS = [
         'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
     ),
     (
-        'https://www.youtube.com/s/player/91201489/player_ias_tce.vflset/en_US/base.js',
-        'W9HJZKktxuYoDTqW', 'U48vOZHaeYS6vO',
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
+    ),
+    (
+        'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js',
+        'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ',
+    ),
+    (
+        'https://www.youtube.com/s/player/d50f54ef/player_ias_tce.vflset/en_US/base.js',
+        'Ha7507LzRmH3Utygtj', 'XFTb2HoeOE5MHg',
+    ),
+    (
+        'https://www.youtube.com/s/player/074a8365/player_ias_tce.vflset/en_US/base.js',
+        'Ha7507LzRmH3Utygtj', 'ufTsrE0IVYrkl8v',
+    ),
+    (
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'N5uAlLqm0eg1GyHO', 'dCBQOejdq5s-ww',
+    ),
+    (
+        'https://www.youtube.com/s/player/69f581a5/tv-player-ias.vflset/tv-player-ias.js',
+        '-qIP447rVlTTwaZjY', 'KNcGOksBAvwqQg',
+    ),
+    (
+        'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
+        'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA',
     ),
 ]
 

youtube_dl/extractor/canvas.py

@@ -1,7 +1,6 @@
 from __future__ import unicode_literals
 
 import re
-import json
 
 from .common import InfoExtractor
 from .gigya import GigyaBaseIE
@@ -17,6 +16,7 @@ from ..utils import (
     str_or_none,
     strip_or_none,
     url_or_none,
+    urlencode_postdata
 )
 
 
@@ -259,7 +259,7 @@ class VrtNUIE(GigyaBaseIE):
         'expected_warnings': ['Unable to download asset JSON', 'is not a supported codec', 'Unknown MIME type'],
     }]
     _NETRC_MACHINE = 'vrtnu'
-    _APIKEY = '3_0Z2HujMtiWq_pkAjgnS2Md2E11a1AwZjYiBETtwNE-EoEHDINgtnvcAOpNgmrVGy'
+    _APIKEY = '3_qhEcPa5JGFROVwu5SWKqJ4mVOIkwlFNMSKwzPDAh8QZOtHqu6L4nD5Q7lk0eXOOG'
     _CONTEXT_ID = 'R3595707040'
 
     def _real_initialize(self):
@@ -270,35 +270,38 @@ class VrtNUIE(GigyaBaseIE):
         if username is None:
             return
 
-        auth_data = {
-            'APIKey': self._APIKEY,
-            'targetEnv': 'jssdk',
-            'loginID': username,
-            'password': password,
-            'authMode': 'cookie',
-        }
-
-        auth_info = self._gigya_login(auth_data)
+        auth_info = self._download_json(
+            'https://accounts.vrt.be/accounts.login', None,
+            note='Login data', errnote='Could not get Login data',
+            headers={}, data=urlencode_postdata({
+                'loginID': username,
+                'password': password,
+                'sessionExpiration': '-2',
+                'APIKey': self._APIKEY,
+                'targetEnv': 'jssdk',
+            }))
 
         # Sometimes authentication fails for no good reason, retry
         login_attempt = 1
         while login_attempt <= 3:
             try:
-                # When requesting a token, no actual token is returned, but the
-                # necessary cookies are set.
+                self._request_webpage('https://token.vrt.be/vrtnuinitlogin',
+                                      None, note='Requesting XSRF Token', errnote='Could not get XSRF Token',
+                                      query={'provider': 'site', 'destination': 'https://www.vrt.be/vrtnu/'})
+
+                post_data = {
+                    'UID': auth_info['UID'],
+                    'UIDSignature': auth_info['UIDSignature'],
+                    'signatureTimestamp': auth_info['signatureTimestamp'],
+                    'client_id': 'vrtnu-site',
+                    '_csrf': self._get_cookies('https://login.vrt.be/').get('OIDCXSRF').value
+                }
+
                 self._request_webpage(
-                    'https://token.vrt.be',
+                    'https://login.vrt.be/perform_login',
                     None, note='Requesting a token', errnote='Could not get a token',
-                    headers={
-                        'Content-Type': 'application/json',
-                        'Referer': 'https://www.vrt.be/vrtnu/',
-                    },
-                    data=json.dumps({
-                        'uid': auth_info['UID'],
-                        'uidsig': auth_info['UIDSignature'],
-                        'ts': auth_info['signatureTimestamp'],
-                        'email': auth_info['profile']['email'],
-                    }).encode('utf-8'))
+                    headers={}, data=urlencode_postdata(post_data))
+
             except ExtractorError as e:
                 if isinstance(e.cause, compat_HTTPError) and e.cause.code == 401:
                     login_attempt += 1

youtube_dl/extractor/youtube.py

@@ -91,12 +91,12 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'IOS',
-                    'clientVersion': '19.45.4',
+                    'clientVersion': '20.10.4',
                     'deviceMake': 'Apple',
                     'deviceModel': 'iPhone16,2',
-                    'userAgent': 'com.google.ios.youtube/19.45.4 (iPhone16,2; U; CPU iOS 18_1_0 like Mac OS X;)',
+                    'userAgent': 'com.google.ios.youtube/20.10.4 (iPhone16,2; U; CPU iOS 18_3_2 like Mac OS X;)',
                     'osName': 'iPhone',
-                    'osVersion': '18.1.0.22B83',
+                    'osVersion': '18.3.2.22D82',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 5,
@@ -109,7 +109,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'MWEB',
-                    'clientVersion': '2.20241202.07.00',
+                    'clientVersion': '2.20250311.03.00',
                     # mweb previously did not require PO Token with this UA
                     'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)',
                 },
@@ -122,7 +122,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'TVHTML5',
-                    'clientVersion': '7.20250120.19.00',
+                    'clientVersion': '7.20250312.16.00',
                     'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/Version',
                 },
             },
@@ -133,7 +133,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'WEB',
-                    'clientVersion': '2.20241126.01.00',
+                    'clientVersion': '2.20250312.04.00',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 1,
@@ -692,7 +692,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         'invidious': '|'.join(_INVIDIOUS_SITES),
     }
     _PLAYER_INFO_RE = (
-        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/player',
+        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})//(?:tv-)?player',
         r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
         r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
     )
@@ -1857,7 +1857,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
     def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None):
 
         var_ay = self._search_regex(
-            r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"[^"]+"\s*\.\s*split\("\{"\))(?=\s*[,;])',
+            r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"(?:\\"|[^"])+"\s*\.\s*split\("\W+"\))(?=\s*[,;])',
             jsi.code, 'useful values', default='')
 
         func_name = self._extract_n_function_name(jsi.code)