Merge 5860937e17a249d80739efd07629dacd2392eba2 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/extractors.py

@@ -1430,6 +1430,7 @@ from .dlive import (
 from .umg import UMGDeIE
 from .unistra import UnistraIE
 from .unity import UnityIE
+from .uno import UNOIE
 from .uol import UOLIE
 from .uplynk import (
     UplynkIE,

youtube_dl/extractor/uno.py

@@ -0,0 +1,64 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+from .common import InfoExtractor
+from ..utils import (
+    unified_timestamp,
+    url_or_none,
+)
+
+
+class UNOIE(InfoExtractor):
+    _VALID_URL = r'https?://media\.un\.org/(?:\w+/)+(?P<id>k\d[\w]+)'
+    _TESTS = [{
+        'url': 'https://media.un.org/en/asset/k1r/k1r3vy9ikk',
+        'md5': '981c41cb283227f079d1e5059fd0d30c',
+        'info_dict': {
+            'id': '1_r3vy9ikk',
+            'ext': 'mp4',
+            'title': 'md5:abde2a46d396051535e5e6fd6f627a19',
+            'description': 'md5:2cba11ee153ae3e6ae2c629e7c4e39b0',
+            'thumbnail': 're:https?://.+/thumbnail/.+',
+            'duration': 5768,
+            'timestamp': 1625216872,
+            'upload_date': '20210702',
+            'uploader_id': 'UNWebTV_New_York',
+        }
+    }, {
+        'url': 'https://media.un.org/en/asset/k12/k12gpkg3qx',
+        'md5': '5978503ca886a922a0f00cf5a7e82395',
+        'info_dict': {
+            'id': '1_vohfjqkj',
+            'ext': 'mp4',
+            'title': '1851st Meeting, 81st session Committee on the Elimination of Discrimination Against Women (CEDAW)',
+            'description': 'Informal meeting with NGOs and human rights institutions - 1851st Meeting, 81st session CEDAW',
+            'thumbnail': 're:https?://.+/thumbnail/.+',
+            'duration': 3502,
+            'timestamp': 1644235332,
+            'upload_date': '20220207',
+            'uploader_id': 'nathalie.minard@un.org',
+        },
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+        partner_id = self._search_regex(r'partnerId\s*:\s*(\d+)\b', webpage, 'Partner ID')
+        video_id = self._search_regex(r'/p/%s(?:/\w+)+?/entry_id/(\w+)/' % (partner_id, ), webpage, 'Kaltura ID')
+        title = (
+            self._html_search_meta(('title', 'og:title'), webpage)
+            or self._html_search_regex(r'<title\b[^>]*>([^<]+)</title\b', webpage, 'title').rsplit('|', 1)[0]).strip()
+        result = self.url_result(
+            'kaltura:%s:%s' % (partner_id, video_id), 'Kaltura',
+            video_title=title,
+            video_id=video_id)
+        if result:
+            result.update({
+                '_type': 'url_transparent',
+                'description': self._html_search_meta(('description', 'og:description'), webpage, 'description'),
+                'creator': self._html_search_meta('author', webpage),
+                'upoader_id': self._html_search_meta('publisher', webpage),
+                'thumbnail': url_or_none(self._og_search_thumbnail(webpage)),
+                'timestamp': unified_timestamp(self._og_search_property('updated_time', webpage)),
+            })
+        return result

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)