Merge remote-tracking branch 'origin/master' into pr/8968

yt_dlp/YoutubeDL.py

@@ -690,7 +690,6 @@ class YoutubeDL:
         self.params['http_headers'] = HTTPHeaderDict(std_headers, self.params.get('http_headers'))
         self._load_cookies(self.params['http_headers'].get('Cookie'))  # compat
         self.params['http_headers'].pop('Cookie', None)
-        self._request_director = self.build_request_director(_REQUEST_HANDLERS.values(), _RH_PREFERENCES)
 
         if auto_init and auto_init != 'no_verbose_header':
             self.print_debug_header()
@@ -964,6 +963,7 @@ class YoutubeDL:
     def close(self):
         self.save_cookies()
         self._request_director.close()
+        del self._request_director
 
     def trouble(self, message=None, tb=None, is_error=True):
         """Determine action to take when a download problem appears.
@@ -4160,6 +4160,10 @@ class YoutubeDL:
             director.preferences.add(lambda rh, _: 500 if rh.RH_KEY == 'Urllib' else 0)
         return director
 
+    @functools.cached_property
+    def _request_director(self):
+        return self.build_request_director(_REQUEST_HANDLERS.values(), _RH_PREFERENCES)
+
     def encode(self, s):
         if isinstance(s, bytes):
             return s  # Already encoded

yt_dlp/extractor/archiveorg.py

@@ -300,7 +300,7 @@ class ArchiveOrgIE(InfoExtractor):
             is_logged_in = bool(self._get_cookies('https://archive.org').get('logged-in-sig'))
             if extension in KNOWN_EXTENSIONS and (not f.get('private') or is_logged_in):
                 entry['formats'].append({
-                    'url': 'https://archive.org/download/' + identifier + '/' + f['name'],
+                    'url': 'https://archive.org/download/' + identifier + '/' + urllib.parse.quote(f['name']),
                     'format': f.get('format'),
                     'width': int_or_none(f.get('width')),
                     'height': int_or_none(f.get('height')),

yt_dlp/extractor/cloudflarestream.py

@@ -4,27 +4,25 @@ from .common import InfoExtractor
 
 
 class CloudflareStreamIE(InfoExtractor):
+    _SUBDOMAIN_RE = r'(?:(?:watch|iframe|customer-\w+)\.)?'
     _DOMAIN_RE = r'(?:cloudflarestream\.com|(?:videodelivery|bytehighway)\.net)'
-    _EMBED_RE = r'embed\.%s/embed/[^/]+\.js\?.*?\bvideo=' % _DOMAIN_RE
+    _EMBED_RE = rf'embed\.{_DOMAIN_RE}/embed/[^/]+\.js\?.*?\bvideo='
     _ID_RE = r'[\da-f]{32}|[\w-]+\.[\w-]+\.[\w-]+'
-    _VALID_URL = r'''(?x)
-                    https?://
-                        (?:
-                            (?:watch\.)?%s/|
-                            %s
-                        )
-                        (?P<id>%s)
-                    ''' % (_DOMAIN_RE, _EMBED_RE, _ID_RE)
-    _EMBED_REGEX = [fr'<script[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//{_EMBED_RE}(?:{_ID_RE}).*?)\1']
+    _VALID_URL = rf'https?://(?:{_SUBDOMAIN_RE}{_DOMAIN_RE}/|{_EMBED_RE})(?P<id>{_ID_RE})'
+    _EMBED_REGEX = [
+        rf'<script[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//{_EMBED_RE}(?:{_ID_RE}).*?)\1',
+        rf'<iframe[^>]+\bsrc=["\'](?P<url>https?://{_SUBDOMAIN_RE}{_DOMAIN_RE}/[\da-f]{{32}})',
+    ]
     _TESTS = [{
         'url': 'https://embed.cloudflarestream.com/embed/we4g.fla9.latest.js?video=31c9291ab41fac05471db4e73aa11717',
         'info_dict': {
             'id': '31c9291ab41fac05471db4e73aa11717',
             'ext': 'mp4',
             'title': '31c9291ab41fac05471db4e73aa11717',
+            'thumbnail': 'https://videodelivery.net/31c9291ab41fac05471db4e73aa11717/thumbnails/thumbnail.jpg',
         },
         'params': {
-            'skip_download': True,
+            'skip_download': 'm3u8',
         },
     }, {
         'url': 'https://watch.cloudflarestream.com/9df17203414fd1db3e3ed74abbe936c1',
@@ -35,6 +33,21 @@ class CloudflareStreamIE(InfoExtractor):
     }, {
         'url': 'https://embed.videodelivery.net/embed/r4xu.fla9.latest.js?video=81d80727f3022488598f68d323c1ad5e',
         'only_matching': True,
+    }, {
+        'url': 'https://customer-aw5py76sw8wyqzmh.cloudflarestream.com/2463f6d3e06fa29710a337f5f5389fd8/iframe',
+        'only_matching': True,
+    }]
+    _WEBPAGE_TESTS = [{
+        'url': 'https://upride.cc/incident/shoulder-pass-at-light/',
+        'info_dict': {
+            'id': 'eaef9dea5159cf968be84241b5cedfe7',
+            'ext': 'mp4',
+            'title': 'eaef9dea5159cf968be84241b5cedfe7',
+            'thumbnail': 'https://videodelivery.net/eaef9dea5159cf968be84241b5cedfe7/thumbnails/thumbnail.jpg',
+        },
+        'params': {
+            'skip_download': 'm3u8',
+        },
     }]
 
     def _real_extract(self, url):

yt_dlp/extractor/ntvru.py

@@ -35,6 +35,7 @@ class NTVRuIE(InfoExtractor):
             'duration': 172,
             'view_count': int,
         },
+        'skip': '404 Not Found',
     }, {
         'url': 'http://www.ntv.ru/peredacha/segodnya/m23700/o232416',
         'md5': '82dbd49b38e3af1d00df16acbeab260c',
@@ -78,7 +79,8 @@ class NTVRuIE(InfoExtractor):
     }]
 
     _VIDEO_ID_REGEXES = [
-        r'<meta property="og:url" content="http://www\.ntv\.ru/video/(\d+)',
+        r'<meta property="og:url" content="https?://www\.ntv\.ru/video/(\d+)',
+        r'<meta property="og:video:(?:url|iframe)" content="https?://www\.ntv\.ru/embed/(\d+)',
         r'<video embed=[^>]+><id>(\d+)</id>',
         r'<video restriction[^>]+><key>(\d+)</key>',
     ]

yt_dlp/extractor/swearnet.py

@@ -1,5 +1,5 @@
 from .common import InfoExtractor
-from ..utils import int_or_none, traverse_obj
+from ..utils import ExtractorError, int_or_none, traverse_obj
 
 
 class SwearnetEpisodeIE(InfoExtractor):
@@ -51,7 +51,13 @@ class SwearnetEpisodeIE(InfoExtractor):
         display_id, season_number, episode_number = self._match_valid_url(url).group('id', 'season_num', 'episode_num')
         webpage = self._download_webpage(url, display_id)
 
-        external_id = self._search_regex(r'externalid\s*=\s*"([^"]+)', webpage, 'externalid')
+        try:
+            external_id = self._search_regex(r'externalid\s*=\s*"([^"]+)', webpage, 'externalid')
+        except ExtractorError:
+            if 'Upgrade Now' in webpage:
+                self.raise_login_required()
+            raise
+
         json_data = self._download_json(
             f'https://play.vidyard.com/player/{external_id}.json', display_id)['payload']['chapters'][0]
 

yt_dlp/extractor/vimeo.py

@@ -48,17 +48,15 @@ class VimeoBaseInfoExtractor(InfoExtractor):
         return url, data, headers
 
     def _perform_login(self, username, password):
-        webpage = self._download_webpage(
-            self._LOGIN_URL, None, 'Downloading login page')
-        token, vuid = self._extract_xsrft_and_vuid(webpage)
+        viewer = self._download_json('https://vimeo.com/_next/viewer', None, 'Downloading login token')
         data = {
             'action': 'login',
             'email': username,
             'password': password,
             'service': 'vimeo',
-            'token': token,
+            'token': viewer['xsrft'],
         }
-        self._set_vimeo_cookie('vuid', vuid)
+        self._set_vimeo_cookie('vuid', viewer['vuid'])
         try:
             self._download_webpage(
                 self._LOGIN_URL, None, 'Logging in',

yt_dlp/extractor/youtube.py

@@ -3640,15 +3640,28 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         return orderedSet(requested_clients)
 
+    def _invalid_player_response(self, pr, video_id):
+        # YouTube may return a different video player response than expected.
+        # See: https://github.com/TeamNewPipe/NewPipe/issues/8713
+        if (pr_id := traverse_obj(pr, ('videoDetails', 'videoId'))) != video_id:
+            return pr_id
+
     def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg, smuggled_data):
         initial_pr = None
         if webpage:
             initial_pr = self._search_json(
                 self._YT_INITIAL_PLAYER_RESPONSE_RE, webpage, 'initial player response', video_id, fatal=False)
 
+        prs = []
+        if initial_pr and not self._invalid_player_response(initial_pr, video_id):
+            # Android player_response does not have microFormats which are needed for
+            # extraction of some data. So we return the initial_pr with formats
+            # stripped out even if not requested by the user
+            # See: https://github.com/yt-dlp/yt-dlp/issues/501
+            prs.append({**initial_pr, 'streamingData': None})
+
         all_clients = set(clients)
         clients = clients[::-1]
-        prs = []
 
         def append_client(*client_names):
             """ Append the first client name that exists but not already used """
@@ -3660,18 +3673,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                         all_clients.add(actual_client)
                         return
 
-        # Android player_response does not have microFormats which are needed for
-        # extraction of some data. So we return the initial_pr with formats
-        # stripped out even if not requested by the user
-        # See: https://github.com/yt-dlp/yt-dlp/issues/501
-        if initial_pr:
-            pr = dict(initial_pr)
-            pr['streamingData'] = None
-            prs.append(pr)
-
-        last_error = None
         tried_iframe_fallback = False
         player_url = None
+        skipped_clients = {}
         while clients:
             client, base_client, variant = _split_innertube_client(clients.pop())
             player_ytcfg = master_ytcfg if client == 'web' else {}
@@ -3692,26 +3696,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 pr = initial_pr if client == 'web' and initial_pr else self._extract_player_response(
                     client, video_id, player_ytcfg or master_ytcfg, player_ytcfg, player_url if require_js_player else None, initial_pr, smuggled_data)
             except ExtractorError as e:
-                if last_error:
-                    self.report_warning(last_error)
-                last_error = e
+                self.report_warning(e)
                 continue
 
-            if pr:
-                # YouTube may return a different video player response than expected.
-                # See: https://github.com/TeamNewPipe/NewPipe/issues/8713
-                pr_video_id = traverse_obj(pr, ('videoDetails', 'videoId'))
-                if pr_video_id and pr_video_id != video_id:
-                    self.report_warning(
-                        f'Skipping player response from {client} client (got player response for video "{pr_video_id}" instead of "{video_id}")' + bug_reports_message())
-                else:
-                    # Save client name for introspection later
-                    name = short_client_name(client)
-                    sd = traverse_obj(pr, ('streamingData', {dict})) or {}
-                    sd[STREAMING_DATA_CLIENT_NAME] = name
-                    for f in traverse_obj(sd, (('formats', 'adaptiveFormats'), ..., {dict})):
-                        f[STREAMING_DATA_CLIENT_NAME] = name
-                    prs.append(pr)
+            if pr_id := self._invalid_player_response(pr, video_id):
+                skipped_clients[client] = pr_id
+            elif pr:
+                # Save client name for introspection later
+                name = short_client_name(client)
+                sd = traverse_obj(pr, ('streamingData', {dict})) or {}
+                sd[STREAMING_DATA_CLIENT_NAME] = name
+                for f in traverse_obj(sd, (('formats', 'adaptiveFormats'), ..., {dict})):
+                    f[STREAMING_DATA_CLIENT_NAME] = name
+                prs.append(pr)
 
             # creator clients can bypass AGE_VERIFICATION_REQUIRED if logged in
             if variant == 'embedded' and self._is_unplayable(pr) and self.is_authenticated:
@@ -3722,10 +3719,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 elif not variant:
                     append_client(f'tv_embedded.{base_client}', f'{base_client}_embedded')
 
-        if last_error:
-            if not len(prs):
-                raise last_error
-            self.report_warning(last_error)
+        if skipped_clients:
+            self.report_warning(
+                f'Skipping player responses from {"/".join(skipped_clients)} clients '
+                f'(got player responses for video "{"/".join(set(skipped_clients.values()))}" instead of "{video_id}")')
+            if not prs:
+                raise ExtractorError(
+                    'All player responses are invalid. Your IP is likely being blocked by Youtube', expected=True)
+        elif not prs:
+            raise ExtractorError('Failed to extract any player response')
         return prs, player_url
 
     def _needs_live_processing(self, live_status, duration):

yt_dlp/networking/common.py

@@ -68,6 +68,7 @@ class RequestDirector:
     def close(self):
         for handler in self.handlers.values():
             handler.close()
+        self.handlers = {}
 
     def add_handler(self, handler: RequestHandler):
         """Add a handler. If a handler of the same RH_KEY exists, it will overwrite it"""