Merge 1b92d6fab7 into 5904853ae5

Closes #7442
Authored by: bashonly

yt_dlp/extractor/_extractors.py

@@ -820,6 +820,7 @@
     InstagramUserIE,
     InstagramTagIE,
     InstagramStoryIE,
+    InstagramLiveIE,
 )
 from .internazionale import InternazionaleIE
 from .internetvideoarchive import InternetVideoArchiveIE

yt_dlp/extractor/cbc.py

@@ -151,7 +151,7 @@ def _real_extract(self, url):
 
 class CBCPlayerIE(InfoExtractor):
     IE_NAME = 'cbc.ca:player'
-    _VALID_URL = r'(?:cbcplayer:|https?://(?:www\.)?cbc\.ca/(?:player/play/|i/caffeine/syndicate/\?mediaId=))(?P<id>(?:\d\.)?\d+)'
+    _VALID_URL = r'(?:cbcplayer:|https?://(?:www\.)?cbc\.ca/(?:player/play/(?:video/)?|i/caffeine/syndicate/\?mediaId=))(?P<id>(?:\d\.)?\d+)'
     _TESTS = [{
         'url': 'http://www.cbc.ca/player/play/2683190193',
         'md5': '64d25f841ddf4ddb28a235338af32e2c',
@@ -277,6 +277,28 @@ class CBCPlayerIE(InfoExtractor):
             'location': 'Canada',
             'media_type': 'Full Program',
         },
+    }, {
+        'url': 'https://www.cbc.ca/player/play/video/1.7194274',
+        'md5': '188b96cf6bdcb2540e178a6caa957128',
+        'info_dict': {
+            'id': '2334524995812',
+            'ext': 'mp4',
+            'title': '#TheMoment a rare white spirit moose was spotted in Alberta',
+            'description': 'md5:18ae269a2d0265c5b0bbe4b2e1ac61a3',
+            'timestamp': 1714788791,
+            'duration': 77.678,
+            'subtitles': {'eng': [{'ext': 'vtt', 'protocol': 'm3u8_native'}]},
+            'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/201/543/THE_MOMENT.jpg',
+            'uploader': 'CBCC-NEW',
+            'chapters': 'count:0',
+            'upload_date': '20240504',
+            'categories': 'count:3',
+            'series': 'The National',
+            'tags': 'count:15',
+            'creators': ['encoder'],
+            'location': 'Canada',
+            'media_type': 'Excerpt',
+        },
     }, {
         'url': 'cbcplayer:1.7159484',
         'only_matching': True,

yt_dlp/extractor/crunchyroll.py

@@ -53,15 +53,19 @@ def _set_auth_info(self, response):
         CrunchyrollBaseIE._AUTH_EXPIRY = time_seconds(seconds=traverse_obj(response, ('expires_in', {float_or_none}), default=300) - 10)
 
     def _request_token(self, headers, data, note='Requesting token', errnote='Failed to request token'):
-        try:  # TODO: Add impersonation support here
+        try:
             return self._download_json(
                 f'{self._BASE_URL}/auth/v1/token', None, note=note, errnote=errnote,
-                headers=headers, data=urlencode_postdata(data))
+                headers=headers, data=urlencode_postdata(data), impersonate=True)
         except ExtractorError as error:
             if not isinstance(error.cause, HTTPError) or error.cause.status != 403:
                 raise
+            if target := error.cause.response.extensions.get('impersonate'):
+                raise ExtractorError(f'Got HTTP Error 403 when using impersonate target "{target}"')
             raise ExtractorError(
-                'Request blocked by Cloudflare; navigate to Crunchyroll in your browser, '
+                'Request blocked by Cloudflare. '
+                'Install the required impersonation dependency if possible, '
+                'or else navigate to Crunchyroll in your browser, '
                 'then pass the fresh cookies (with --cookies-from-browser or --cookies) '
                 'and your browser\'s User-Agent (with --user-agent)', expected=True)
 

yt_dlp/extractor/instagram.py

@@ -8,6 +8,7 @@
 from ..networking.exceptions import HTTPError
 from ..utils import (
     ExtractorError,
+    UserNotLive,
     decode_base_n,
     encode_base_n,
     filter_dict,
@@ -733,3 +734,59 @@ def _real_extract(self, url):
                     **filter_dict(highlight_data),
                 })
         return self.playlist_result(info_data, playlist_id=story_id, playlist_title=story_title)
+
+
+class InstagramLiveIE(InstagramBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<user>[^/]+)/live(?:/|/(?P<id>\d+))?'
+    IE_NAME = 'instagram:live'
+
+    _TESTS = [{
+        'url': 'https://www.instagram.com/answer/live/42',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        username, video_id = self._match_valid_url(url).groups()
+        video_id = video_id or username
+
+        metadata_str = self._html_search_regex(
+            r'data-sjs>({.*?ScheduledServerJS.*?target_user_id.*?})</script>',
+            self._download_webpage(url, video_id), 'live metadata', default=None)
+        if not metadata_str:
+            raise UserNotLive(video_id=video_id)
+
+        metadata = self._parse_json(metadata_str, video_id)
+
+        target_user_id = traverse_obj(
+            metadata,
+            ('require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., 'rootView', 'props', 'target_user_id', {str}),
+            ('require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., 'hostableView', 'props', 'target_user_id', {str}),
+            get_all=False)
+        if not target_user_id:
+            raise ExtractorError('Missing "target_user_id" in metadata')
+
+        app_id = traverse_obj(
+            metadata,
+            ('require', ..., ..., ..., '__bbox', 'define', ..., ..., 'customHeaders', 'X-IG-App-ID', {str}),
+            ('require', ..., ..., ..., '__bbox', 'define', ..., ..., 'appId', {int}),
+            ('require', ..., ..., ..., '__bbox', 'define', ..., ..., 'APP_ID', {str}),
+            ('require', ..., ..., ..., '__bbox', 'define', ..., ..., 'app_id', {str}),
+            get_all=False)
+        if not app_id:
+            raise ExtractorError('Missing "app_id" in metadata')
+
+        web_info_json = self._download_json(
+            'https://www.instagram.com/api/v1/live/web_info/', video_id,
+            query={'target_user_id': target_user_id}, headers={'x-ig-app-id': app_id},
+            note='Downloading web_info', errnote='Unable to download web_info')
+
+        video_id = traverse_obj(web_info_json, ('id'))
+
+        return {
+            'id': video_id,
+            'formats': self._extract_mpd_formats(
+                traverse_obj(web_info_json, ('dash_abr_playback_url', {url_or_none})), video_id),
+            'uploader': traverse_obj(web_info_json, ('broadcast_owner', 'full_name')),
+            'uploader_id': username,
+            'live_status': 'is_live',
+        }

yt_dlp/extractor/youporn.py

@@ -72,15 +72,15 @@ class YouPornIE(InfoExtractor):
             'id': '16290308',
             'age_limit': 18,
             'categories': [],
-            'description': 'md5:00ea70f642f431c379763c17c2f396bc',
+            'description': str,  # TODO: detect/remove SEO spam description in ytdl backport
             'display_id': 'tinderspecial-trailer1',
             'duration': 298.0,
             'ext': 'mp4',
             'upload_date': '20201123',
             'uploader': 'Ersties',
             'tags': [],
-            'thumbnail': 'https://fi1.ypncdn.com/202011/23/16290308/original/8/tinderspecial-trailer1-8(m=eaAaaEPbaaaa).jpg',
-            'timestamp': 1606089600,
+            'thumbnail': r're:https://.+\.jpg',
+            'timestamp': 1606147564,
             'title': 'Tinder In Real Life',
             'view_count': int,
         }
@@ -88,11 +88,17 @@ class YouPornIE(InfoExtractor):
 
     def _real_extract(self, url):
         video_id, display_id = self._match_valid_url(url).group('id', 'display_id')
-        definitions = self._download_json(
-            f'https://www.youporn.com/api/video/media_definitions/{video_id}/', display_id or video_id)
+        self._set_cookie('.youporn.com', 'age_verified', '1')
+        webpage = self._download_webpage(f'https://www.youporn.com/watch/{video_id}', video_id)
+        definitions = self._search_json(r'\bplayervars\s*:', webpage, 'player vars', video_id)['mediaDefinitions']
 
-        def get_format_data(data, f):
-            return traverse_obj(data, lambda _, v: v['format'] == f and url_or_none(v['videoUrl']))
+        def get_format_data(data, stream_type):
+            info_url = traverse_obj(data, (lambda _, v: v['format'] == stream_type, 'videoUrl', {url_or_none}, any))
+            if not info_url:
+                return []
+            return traverse_obj(
+                self._download_json(info_url, video_id, f'Downloading {stream_type} info JSON', fatal=False),
+                lambda _, v: v['format'] == stream_type and url_or_none(v['videoUrl']))
 
         formats = []
         # Try to extract only the actual master m3u8 first, avoiding the duplicate single resolution "master" m3u8s
@@ -123,10 +129,6 @@ def get_format_data(data, f):
             f['height'] = height
             formats.append(f)
 
-        webpage = self._download_webpage(
-            'http://www.youporn.com/watch/%s' % video_id, display_id,
-            headers={'Cookie': 'age_verified=1'})
-
         title = self._html_search_regex(
             r'(?s)<div[^>]+class=["\']watchVideoTitle[^>]+>(.+?)</div>',
             webpage, 'title', default=None) or self._og_search_title(