Merge 1f92b9ec7a into 96da952504

Authored by: seproDev, pukkandan

test/test_networking.py

@@ -785,6 +785,25 @@ class TestHTTPImpersonateRequestHandler(TestRequestHandlerBase):
                 assert res.status == 200
                 assert std_headers['user-agent'].lower() not in res.read().decode().lower()
 
+    def test_response_extensions(self, handler):
+        with handler() as rh:
+            for target in rh.supported_targets:
+                request = Request(
+                    f'http://127.0.0.1:{self.http_port}/gen_200', extensions={'impersonate': target})
+                res = validate_and_send(rh, request)
+                assert res.extensions['impersonate'] == rh._get_request_target(request)
+
+    def test_http_error_response_extensions(self, handler):
+        with handler() as rh:
+            for target in rh.supported_targets:
+                request = Request(
+                    f'http://127.0.0.1:{self.http_port}/gen_404', extensions={'impersonate': target})
+                try:
+                    validate_and_send(rh, request)
+                except HTTPError as e:
+                    res = e.response
+                assert res.extensions['impersonate'] == rh._get_request_target(request)
+
 
 class TestRequestHandlerMisc:
     """Misc generic tests for request handlers, not related to request or validation testing"""

yt_dlp/YoutubeDL.py

@@ -2136,6 +2136,11 @@ class YoutubeDL:
 
     def _check_formats(self, formats):
         for f in formats:
+            working = f.get('__working')
+            if working is not None:
+                if working:
+                    yield f
+                continue
             self.to_screen('[info] Testing format %s' % f['format_id'])
             path = self.get_output_path('temp')
             if not self._ensure_dir_exists(f'{path}/'):
@@ -2152,33 +2157,44 @@ class YoutubeDL:
                         os.remove(temp_file.name)
                     except OSError:
                         self.report_warning('Unable to delete temporary file "%s"' % temp_file.name)
+            f['__working'] = success
             if success:
                 yield f
             else:
                 self.to_screen('[info] Unable to download format %s. Skipping...' % f['format_id'])
 
+    def _select_formats(self, formats, selector):
+        return list(selector({
+            'formats': formats,
+            'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
+            'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats)  # No formats with video
+                                   or all(f.get('acodec') == 'none' for f in formats)),  # OR, No formats with audio
+        }))
+
     def _default_format_spec(self, info_dict, download=True):
+        download = download and not self.params.get('simulate')
+        prefer_best = download and (
+            self.params['outtmpl']['default'] == '-'
+            or info_dict.get('is_live') and not self.params.get('live_from_start'))
 
         def can_merge():
             merger = FFmpegMergerPP(self)
             return merger.available and merger.can_merge()
 
-        prefer_best = (
-            not self.params.get('simulate')
-            and download
-            and (
-                not can_merge()
-                or info_dict.get('is_live') and not self.params.get('live_from_start')
-                or self.params['outtmpl']['default'] == '-'))
-        compat = (
-            prefer_best
-            or self.params.get('allow_multiple_audio_streams', False)
-            or 'format-spec' in self.params['compat_opts'])
+        if not prefer_best and download and not can_merge():
+            prefer_best = True
+            formats = self._get_formats(info_dict)
+            evaluate_formats = lambda spec: self._select_formats(formats, self.build_format_selector(spec))
+            if evaluate_formats('b/bv+ba') != evaluate_formats('bv*+ba/b'):
+                self.report_warning('ffmpeg not found. The downloaded format may not be the best available. '
+                                    'Installing ffmpeg is strongly recommended: https://github.com/yt-dlp/yt-dlp#dependencies')
 
-        return (
-            'best/bestvideo+bestaudio' if prefer_best
-            else 'bestvideo*+bestaudio/best' if not compat
-            else 'bestvideo+bestaudio/best')
+        compat = (self.params.get('allow_multiple_audio_streams')
+                  or 'format-spec' in self.params['compat_opts'])
+
+        return ('best/bestvideo+bestaudio' if prefer_best
+                else 'bestvideo+bestaudio/best' if compat
+                else 'bestvideo*+bestaudio/best')
 
     def build_format_selector(self, format_spec):
         def syntax_error(note, start):
@@ -2928,12 +2944,7 @@ class YoutubeDL:
                 self.write_debug(f'Default format spec: {req_format}')
                 format_selector = self.build_format_selector(req_format)
 
-            formats_to_download = list(format_selector({
-                'formats': formats,
-                'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
-                'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats)  # No formats with video
-                                       or all(f.get('acodec') == 'none' for f in formats)),  # OR, No formats with audio
-            }))
+            formats_to_download = self._select_formats(formats, format_selector)
             if interactive_format_selection and not formats_to_download:
                 self.report_error('Requested format is not available', tb=False, is_error=False)
                 continue

yt_dlp/extractor/patreon.py

@@ -219,7 +219,29 @@ class PatreonIE(PatreonBaseIE):
             'thumbnail': r're:^https?://.+',
         },
         'params': {'skip_download': 'm3u8'},
+    }, {
+        # multiple attachments/embeds
+        'url': 'https://www.patreon.com/posts/holy-wars-solos-100601977',
+        'playlist_count': 3,
+        'info_dict': {
+            'id': '100601977',
+            'title': '"Holy Wars" (Megadeth) Solos Transcription & Lesson/Analysis',
+            'description': 'md5:d099ab976edfce6de2a65c2b169a88d3',
+            'uploader': 'Bradley Hall',
+            'uploader_id': '24401883',
+            'uploader_url': 'https://www.patreon.com/bradleyhallguitar',
+            'channel_id': '3193932',
+            'channel_url': 'https://www.patreon.com/bradleyhallguitar',
+            'channel_follower_count': int,
+            'timestamp': 1710777855,
+            'upload_date': '20240318',
+            'like_count': int,
+            'comment_count': int,
+            'thumbnail': r're:^https?://.+',
+        },
+        'skip': 'Patron-only content',
     }]
+    _RETURN_TYPE = 'video'
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
@@ -234,58 +256,54 @@ class PatreonIE(PatreonBaseIE):
                 'include': 'audio,user,user_defined_tags,campaign,attachments_media',
             })
         attributes = post['data']['attributes']
-        title = attributes['title'].strip()
-        image = attributes.get('image') or {}
-        info = {
-            'id': video_id,
-            'title': title,
-            'description': clean_html(attributes.get('content')),
-            'thumbnail': image.get('large_url') or image.get('url'),
-            'timestamp': parse_iso8601(attributes.get('published_at')),
-            'like_count': int_or_none(attributes.get('like_count')),
-            'comment_count': int_or_none(attributes.get('comment_count')),
-        }
-        can_view_post = traverse_obj(attributes, 'current_user_can_view')
-        if can_view_post and info['comment_count']:
-            info['__post_extractor'] = self.extract_comments(video_id)
+        info = traverse_obj(attributes, {
+            'title': ('title', {str.strip}),
+            'description': ('content', {clean_html}),
+            'thumbnail': ('image', ('large_url', 'url'), {url_or_none}, any),
+            'timestamp': ('published_at', {parse_iso8601}),
+            'like_count': ('like_count', {int_or_none}),
+            'comment_count': ('comment_count', {int_or_none}),
+        })
 
-        for i in post.get('included', []):
-            i_type = i.get('type')
-            if i_type == 'media':
-                media_attributes = i.get('attributes') or {}
-                download_url = media_attributes.get('download_url')
+        entries = []
+        idx = 0
+        for include in traverse_obj(post, ('included', lambda _, v: v['type'])):
+            include_type = include['type']
+            if include_type == 'media':
+                media_attributes = traverse_obj(include, ('attributes', {dict})) or {}
+                download_url = url_or_none(media_attributes.get('download_url'))
                 ext = mimetype2ext(media_attributes.get('mimetype'))
 
                 # if size_bytes is None, this media file is likely unavailable
                 # See: https://github.com/yt-dlp/yt-dlp/issues/4608
                 size_bytes = int_or_none(media_attributes.get('size_bytes'))
                 if download_url and ext in KNOWN_EXTENSIONS and size_bytes is not None:
-                    # XXX: what happens if there are multiple attachments?
-                    return {
-                        **info,
+                    idx += 1
+                    entries.append({
+                        'id': f'{video_id}-{idx}',
                         'ext': ext,
                         'filesize': size_bytes,
                         'url': download_url,
-                    }
-            elif i_type == 'user':
-                user_attributes = i.get('attributes')
-                if user_attributes:
-                    info.update({
-                        'uploader': user_attributes.get('full_name'),
-                        'uploader_id': str_or_none(i.get('id')),
-                        'uploader_url': user_attributes.get('url'),
                     })
 
-            elif i_type == 'post_tag':
-                info.setdefault('tags', []).append(traverse_obj(i, ('attributes', 'value')))
+            elif include_type == 'user':
+                info.update(traverse_obj(include, {
+                    'uploader': ('attributes', 'full_name', {str}),
+                    'uploader_id': ('id', {str_or_none}),
+                    'uploader_url': ('attributes', 'url', {url_or_none}),
+                }))
 
-            elif i_type == 'campaign':
-                info.update({
-                    'channel': traverse_obj(i, ('attributes', 'title')),
-                    'channel_id': str_or_none(i.get('id')),
-                    'channel_url': traverse_obj(i, ('attributes', 'url')),
-                    'channel_follower_count': int_or_none(traverse_obj(i, ('attributes', 'patron_count'))),
-                })
+            elif include_type == 'post_tag':
+                if post_tag := traverse_obj(include, ('attributes', 'value', {str})):
+                    info.setdefault('tags', []).append(post_tag)
+
+            elif include_type == 'campaign':
+                info.update(traverse_obj(include, {
+                    'channel': ('attributes', 'title', {str}),
+                    'channel_id': ('id', {str_or_none}),
+                    'channel_url': ('attributes', 'url', {url_or_none}),
+                    'channel_follower_count': ('attributes', 'patron_count', {int_or_none}),
+                }))
 
         # handle Vimeo embeds
         if traverse_obj(attributes, ('embed', 'provider')) == 'Vimeo':
@@ -296,36 +314,50 @@ class PatreonIE(PatreonBaseIE):
                     v_url, video_id, 'Checking Vimeo embed URL',
                     headers={'Referer': 'https://patreon.com/'},
                     fatal=False, errnote=False):
-                return self.url_result(
+                entries.append(self.url_result(
                     VimeoIE._smuggle_referrer(v_url, 'https://patreon.com/'),
-                    VimeoIE, url_transparent=True, **info)
+                    VimeoIE, url_transparent=True))
 
         embed_url = traverse_obj(attributes, ('embed', 'url', {url_or_none}))
         if embed_url and self._request_webpage(embed_url, video_id, 'Checking embed URL', fatal=False, errnote=False):
-            return self.url_result(embed_url, **info)
+            entries.append(self.url_result(embed_url))
 
-        post_file = traverse_obj(attributes, 'post_file')
+        post_file = traverse_obj(attributes, ('post_file', {dict}))
         if post_file:
             name = post_file.get('name')
             ext = determine_ext(name)
             if ext in KNOWN_EXTENSIONS:
-                return {
-                    **info,
+                entries.append({
+                    'id': video_id,
                     'ext': ext,
                     'url': post_file['url'],
-                }
+                })
             elif name == 'video' or determine_ext(post_file.get('url')) == 'm3u8':
                 formats, subtitles = self._extract_m3u8_formats_and_subtitles(post_file['url'], video_id)
-                return {
-                    **info,
+                entries.append({
+                    'id': video_id,
                     'formats': formats,
                     'subtitles': subtitles,
-                }
+                })
 
-        if can_view_post is False:
+        can_view_post = traverse_obj(attributes, 'current_user_can_view')
+        comments = None
+        if can_view_post and info.get('comment_count'):
+            comments = self.extract_comments(video_id)
+
+        if not entries and can_view_post is False:
             self.raise_no_formats('You do not have access to this post', video_id=video_id, expected=True)
-        else:
+        elif not entries:
             self.raise_no_formats('No supported media found in this post', video_id=video_id, expected=True)
+        elif len(entries) == 1:
+            info.update(entries[0])
+        else:
+            for entry in entries:
+                entry.update(info)
+            return self.playlist_result(entries, video_id, **info, __post_extractor=comments)
+
+        info['id'] = video_id
+        info['__post_extractor'] = comments
         return info
 
     def _get_comments(self, post_id):

yt_dlp/extractor/zdf.py

@@ -14,6 +14,7 @@ from ..utils import (
     parse_codecs,
     qualities,
     traverse_obj,
+    try_call,
     try_get,
     unified_timestamp,
     update_url_query,
@@ -184,12 +185,20 @@ class ZDFIE(ZDFBaseIE):
         'info_dict': {
             'id': '151025_magie_farben2_tex',
             'ext': 'mp4',
+            'duration': 2615.0,
             'title': 'Die Magie der Farben (2/2)',
             'description': 'md5:a89da10c928c6235401066b60a6d5c1a',
-            'duration': 2615,
             'timestamp': 1465021200,
-            'upload_date': '20160604',
             'thumbnail': 'https://www.zdf.de/assets/mauve-im-labor-100~768x432?cb=1464909117806',
+            'upload_date': '20160604',
+            'episode': 'Die Magie der Farben (2/2)',
+            'episode_id': 'POS_954f4170-36a5-4a41-a6cf-78f1f3b1f127',
+            'season': 'Staffel 1',
+            'series': 'Die Magie der Farben',
+            'season_number': 1,
+            'seriesId': 'a39900dd-cdbd-4a6a-a413-44e8c6ae18bc',
+            'season_id': '5a92e619-8a0f-4410-a3d5-19c76fbebb37',
+            'episode_number': 2,
         },
     }, {
         'url': 'https://www.zdf.de/funk/druck-11790/funk-alles-ist-verzaubert-102.html',
@@ -197,12 +206,13 @@ class ZDFIE(ZDFBaseIE):
         'info_dict': {
             'ext': 'mp4',
             'id': 'video_funk_1770473',
-            'duration': 1278,
-            'description': 'Die Neue an der Schule verdreht Ismail den Kopf.',
+            'duration': 1278.0,
             'title': 'Alles ist verzaubert',
+            'description': 'Die Neue an der Schule verdreht Ismail den Kopf.',
             'timestamp': 1635520560,
-            'upload_date': '20211029',
             'thumbnail': 'https://www.zdf.de/assets/teaser-funk-alles-ist-verzaubert-102~1920x1080?cb=1663848412907',
+            'upload_date': '20211029',
+            'episode': 'Alles ist verzaubert',
         },
     }, {
         # Same as https://www.phoenix.de/sendungen/dokumentationen/gesten-der-maechtigen-i-a-89468.html?ref=suche
@@ -248,19 +258,48 @@ class ZDFIE(ZDFBaseIE):
             'timestamp': 1675160100,
             'upload_date': '20230131',
             'thumbnail': 'https://epg-image.zdf.de/fotobase-webdelivery/images/e2d7e55a-09f0-424e-ac73-6cac4dd65f35?layout=2400x1350',
+            'series': 'SOKO Stuttgart',
+            'seriesId': 'f862ce9a-6dd1-4388-a698-22b36ac4c9e9',
+            'season': 'Staffel 11',
+            'season_number': 11,
+            'season_id': 'ae1b4990-6d87-4970-a571-caccf1ba2879',
+            'episode': 'Das Geld anderer Leute',
+            'episode_number': 10,
+            'episode_id': 'POS_7f367934-f2f0-45cb-9081-736781ff2d23',
         },
     }, {
         'url': 'https://www.zdf.de/dokumentation/terra-x/unser-gruener-planet-wuesten-doku-100.html',
         'info_dict': {
-            'id': '220605_dk_gruener_planet_wuesten_tex',
+            'id': '220525_green_planet_makingof_1_tropen_tex',
             'ext': 'mp4',
-            'title': 'Unser grüner Planet - Wüsten',
-            'description': 'md5:4fc647b6f9c3796eea66f4a0baea2862',
-            'duration': 2613.0,
-            'timestamp': 1654450200,
-            'upload_date': '20220605',
-            'format_note': 'uhd, main',
-            'thumbnail': 'https://www.zdf.de/assets/saguaro-kakteen-102~3840x2160?cb=1655910690796',
+            'title': 'Making-of Unser grüner Planet - Tropen',
+            'description': 'md5:d7c6949dc7c75c73c4ad51c785fb0b79',
+            'duration': 435.0,
+            'timestamp': 1653811200,
+            'upload_date': '20220529',
+            'format_note': 'hd, main',
+            'thumbnail': 'https://www.zdf.de/assets/unser-gruener-planet-making-of-1-tropen-100~3840x2160?cb=1653493335577',
+            'episode': 'Making-of Unser grüner Planet - Tropen',
+        },
+    }, {
+        'url': 'https://www.zdf.de/serien/northern-lights/begegnung-auf-der-bruecke-100.html',
+        'info_dict': {
+            'id': '240319_2310_sendung_not',
+            'ext': 'mp4',
+            'title': 'Begegnung auf der Brücke',
+            'description': 'Lloyd begegnet an einem verregneten Abend in Dublin einer jungen Frau, die völlig durchnässt auf der Grattan Bridge steht. Es ist der Beginn einer außergewöhnlichen Freundschaft.',
+            'thumbnail': 'https://epg-image.zdf.de/fotobase-webdelivery/images/c5ff1d1f-f5c8-4468-86ac-1b2f1dbecc76?layout=2400x1350',
+            'upload_date': '20240319',
+            'duration': 3083.0,
+            'timestamp': 1710886200,
+            'seriesId': '1d7a1879-01ee-4468-8237-c6b4ecd633c7',
+            'series': 'Northern Lights',
+            'season': 'Staffel 1',
+            'season_number': 1,
+            'season_id': '22ac26a2-4ea2-4055-ac0b-98b755cdf718',
+            'episode': 'Begegnung auf der Brücke',
+            'episode_number': 1,
+            'episode_id': 'POS_71049438-024b-471f-b472-4fe2e490d1fb',
         },
     }]
 
@@ -306,19 +345,34 @@ class ZDFIE(ZDFBaseIE):
             'title': chap.get('anchorLabel')
         } for chap, next_chap in zip(chapter_marks, chapter_marks[1:])]
 
+        target_info = traverse_obj(content, ('programmeItem', 0, 'http://zdf.de/rels/target'))
         return merge_dicts(info, {
             'title': title,
             'description': content.get('leadParagraph') or content.get('teasertext'),
             'duration': int_or_none(t.get('duration')),
             'timestamp': unified_timestamp(content.get('editorialDate')),
             'thumbnails': thumbnails,
-            'chapters': chapters or None
+            'chapters': chapters or None,
+            'seriesId': traverse_obj(target_info, ('http://zdf.de/rels/cmdm/series', 'seriesUuid')),
+            'series': traverse_obj(target_info, ('http://zdf.de/rels/cmdm/series', 'seriesTitle')),
+            'season': traverse_obj(target_info, ('http://zdf.de/rels/cmdm/season', 'seasonTitle')),
+            'season_number': traverse_obj(target_info, ('http://zdf.de/rels/cmdm/season', 'seasonNumber')),
+            'season_id': traverse_obj(target_info, ('http://zdf.de/rels/cmdm/season', 'seasonUuid')),
+            'episode': title,
+            'episode_number': traverse_obj(target_info, 'episodeNumber'),
+            'episode_id': traverse_obj(target_info, 'contentId'),
         })
 
     def _extract_regular(self, url, player, video_id):
-        content = self._call_api(
-            player['content'], video_id, 'content', player['apiToken'], url)
-        return self._extract_entry(player['content'], player, content, video_id)
+        player_content_v2 = player['content']
+        player_content_v3 = update_url_query(player_content_v2, {'profile': 'player-3'})
+
+        content = try_call(
+            lambda: self._call_api(player_content_v3, video_id, 'content', player['apiToken'], url),
+            lambda: self._call_api(player_content_v2, video_id, 'content', player['apiToken'], url),
+        )
+
+        return self._extract_entry(player_content_v2, player, content, video_id)
 
     def _extract_mobile(self, video_id):
         video = self._download_json(

yt_dlp/networking/_curlcffi.py

@@ -132,6 +132,16 @@ class CurlCFFIRH(ImpersonateRequestHandler, InstanceStoreMixin):
         extensions.pop('cookiejar', None)
         extensions.pop('timeout', None)
 
+    def send(self, request: Request) -> Response:
+        target = self._get_request_target(request)
+        try:
+            response = super().send(request)
+        except HTTPError as e:
+            e.response.extensions['impersonate'] = target
+            raise
+        response.extensions['impersonate'] = target
+        return response
+
     def _send(self, request: Request):
         max_redirects_exceeded = False
         session: curl_cffi.requests.Session = self._get_instance(

yt_dlp/networking/common.py

@@ -497,6 +497,7 @@ class Response(io.IOBase):
     @param headers: response headers.
     @param status: Response HTTP status code. Default is 200 OK.
     @param reason: HTTP status reason. Will use built-in reasons based on status code if not provided.
+    @param extensions: Dictionary of handler-specific response extensions.
     """
 
     def __init__(
@@ -505,7 +506,9 @@ class Response(io.IOBase):
             url: str,
             headers: Mapping[str, str],
             status: int = 200,
-            reason: str = None):
+            reason: str = None,
+            extensions: dict = None
+    ):
 
         self.fp = fp
         self.headers = Message()
@@ -517,6 +520,7 @@ class Response(io.IOBase):
             self.reason = reason or HTTPStatus(status).phrase
         except ValueError:
             self.reason = None
+        self.extensions = extensions or {}
 
     def readable(self):
         return self.fp.readable()