Merge 3918da628d1848ed3ec0fe98c242b5c25defd6b1 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

@@ -1264,6 +1264,7 @@ from .telebruxelles import TeleBruxellesIE
 from .telecinco import TelecincoIE
 from .telegraaf import TelegraafIE
 from .telemb import TeleMBIE
+from .teleportal import TeleportalIE
 from .telequebec import (
     TeleQuebecIE,
     TeleQuebecSquatIE,

youtube_dl/extractor/teleportal.py

@@ -0,0 +1,89 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+from .common import InfoExtractor
+from ..utils import (
+    clean_html,
+    float_or_none,
+    parse_iso8601,
+    str_or_none,
+    str_to_int,
+    try_get,
+    url_or_none,
+)
+
+
+class TeleportalIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?teleportal\.ua/(?:ua/)?(?P<id>[0-9a-z-]+(?:/[0-9a-z-]+)*)'
+    _TEST = {
+        'url': 'https://teleportal.ua/ua/show/stb/master-cheff/bitva-sezonov/vypusk-3',
+        # no permanent check on file contents as HLS may vary
+        'info_dict': {
+            'id': 'show/stb/master-cheff/bitva-sezonov/vypusk-3',
+            'ext': 'mp4',
+            'title': 'МастерШеф. Битва сезонів 3 випуск: найогидніший випуск сезону!',
+            'display_id': '2618466',
+            'description': 'md5:4179bcc3a12edfa2f655888cd741ac09',
+            'timestamp': 1644102480,
+            'upload_date': '20220205',
+            'thumbnail': r're:^https?://.+\.jpg$',
+            'release_timestamp': 1643994000,
+            'duration': 11254.0,
+            'series_id': '20632',
+            'series': 'МастерШеф. Битва сезонів 3 випуск: найогидніший випуск сезону!',
+            'season': 'Битва сезонів',
+            'episode': 'Найогидніший випуск сезону!',
+            'episode_num': 3,
+            'categories': ['Шоу'],
+        },
+        'params': {
+            'hls_prefer_native': True,
+            # 'skip_download': True,
+        },
+    }
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        backend_url = 'https://tp-back.starlight.digital/ua/' + video_id
+        series_metadata = self._download_json(backend_url, video_id) or {}
+        title = series_metadata['title']
+        _hash = series_metadata.get('hash', '')
+        api_url = 'https://vcms-api2.starlight.digital/player-api/' + _hash
+        api_metadata = self._download_json(
+            api_url, video_id,
+            query={
+                'referer': 'https://teleportal.ua/',
+                'lang': 'ua',
+            }
+        )
+        video_info = api_metadata['video'][0]
+        formats = []
+        for media in ('mediaHlsNoAdv', 'mediaHls'):
+            media = url_or_none(try_get(video_info, lambda x: x[media]))
+            if not media:
+                continue
+            formats.extend(self._extract_m3u8_formats(media, video_id, 'mp4', fatal=False))
+            break
+        self._sort_formats(formats)
+
+        thumbnail = url_or_none(video_info.get('poster'))
+        category = series_metadata.get('typeTitle')
+
+        return {
+            'id': video_id,
+            'title': title,
+            'formats': formats,
+            'description': clean_html(series_metadata.get('description')) or series_metadata.get('seoDescription'),
+            'display_id': str_or_none(video_info.get('vcmsId')),
+            'hash': _hash,
+            'thumbnail': thumbnail,
+            'timestamp': parse_iso8601(video_info.get('time_upload_video'), delimiter=' '),
+            'release_timestamp': parse_iso8601(video_info.get('publishDate'), delimiter=' '),
+            'duration': float_or_none(video_info.get('duration')),
+            'series_id': str_or_none(series_metadata.get('id')),
+            'series': series_metadata.get('title'),
+            'season': video_info.get('seasonName') or series_metadata.get('seasonGallery', {}).get('title'),
+            'episode': video_info.get('name'),
+            'episode_num': str_to_int(series_metadata.get('seriesTitle')),
+            'categories': [category] if category else None,
+        }

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)