Merge a47fc807bd 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/_extractors.py

@@ -1906,6 +1906,7 @@ from .syvdk import SYVDKIE
 from .syfy import SyfyIE
 from .sztvhu import SztvHuIE
 from .tagesschau import TagesschauIE
+from .taptap import TapTapIE
 from .tass import TassIE
 from .tbs import TBSIE
 from .tbsjp import (

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/taptap.py

@@ -0,0 +1,148 @@
+from .common import InfoExtractor
+from ..utils import (
+    clean_html,
+    int_or_none,
+    parse_qs,
+    str_or_none,
+    traverse_obj,
+    url_or_none,
+)
+
+
+class TapTapIE(InfoExtractor):
+    _VALID_URL = r'https?://www\.taptap\.cn/(?P<section>moment|app)/(?P<id>\d+)'
+    _TESTS = [{
+        'url': 'https://www.taptap.cn/moment/194618230982052443',
+        'info_dict': {
+            'id': 'moment_194618230982052443',
+            "title": "《崩坏3》开放世界「后崩坏书」新篇章 于淹没之地仰视辰星",
+            "description": "md5:cf66f7819d413641b8b28c8543f4ecda",
+            "timestamp": 1633453402,
+            "upload_date": "20211005",
+            "uploader": "乌酱",
+        },
+        'playlist_count': 1,
+        'playlist': [{
+            'info_dict': {
+                'id': '2202584',
+                "ext": "mp4",
+                "title": "《崩坏3》开放世界「后崩坏书」新篇章 于淹没之地仰视辰星",
+                "description": "md5:cf66f7819d413641b8b28c8543f4ecda",
+                "duration": 66,
+                "timestamp": 1633453402,
+                "upload_date": "20211005",
+                "uploader": "乌酱",
+                'thumbnail': r're:^https?://.*\.(png|jpg)',
+            }
+        }]
+    }, {
+        'url': 'https://www.taptap.cn/moment/521630629209573493',
+        'info_dict': {
+            'id': 'moment_521630629209573493',
+            "title": "《崩坏：星穹铁道》黄泉角色PV——「你的颜色」",
+            "description": "md5:2c81245da864428c904d53ae4ad2182b",
+            "timestamp": 1711425600,
+            "upload_date": "20240326",
+            "uploader": "崩坏：星穹铁道",
+        },
+        'playlist_count': 1,
+        'playlist': [{
+            'info_dict': {
+                'id': '4006511',
+                "ext": "mp4",
+                "title": "《崩坏：星穹铁道》黄泉角色PV——「你的颜色」",
+                "description": "md5:2c81245da864428c904d53ae4ad2182b",
+                "duration": 173,
+                "timestamp": 1711425600,
+                "upload_date": "20240326",
+                "uploader": "崩坏：星穹铁道",
+                'thumbnail': r're:^https?://.*\.(png|jpg)',
+            }
+        }]
+    }, {
+        'url': 'https://www.taptap.cn/app/168332',
+        'info_dict': {
+            'id': 'app_168332',
+            "title": "原神",
+            "description": "md5:e345f39a5fea5de2a46923f70d5f76ab",
+        },
+        'playlist_count': 2,
+        'playlist': [{
+            'info_dict': {
+                'id': '4058443',
+                "ext": "mp4",
+                "title": "原神",
+                "description": "md5:e345f39a5fea5de2a46923f70d5f76ab",
+                "duration": 26,
+                'thumbnail': r're:^https?://.*\.(png|jpg)',
+            }
+        }, {
+            'info_dict': {
+                'id': '4058462',
+                "ext": "mp4",
+                "title": "原神",
+                "description": "md5:e345f39a5fea5de2a46923f70d5f76ab",
+                "duration": 295,
+                'thumbnail': r're:^https?://.*\.(png|jpg)',
+            }
+        }]
+
+    }]
+
+    def _deserialize_nuxt_data(self, serialized_nuxt):
+        for row in serialized_nuxt:
+            if isinstance(row, dict):
+                for key, value_or_ref in row.items():
+                    if isinstance(value_or_ref, int):
+                        row[key] = serialized_nuxt[value_or_ref]
+            elif isinstance(row, list):
+                for index, value_or_ref in tuple(enumerate(row)):
+                    if isinstance(value_or_ref, int):
+                        row[index] = serialized_nuxt[value_or_ref]
+        return serialized_nuxt[0]
+
+    def _extract_video(self, video_id, x_ua):
+        data = self._download_json(
+            'https://www.taptap.cn/webapiv2/video-resource/v1/multi-get', video_id,
+            query={'video_ids': video_id, 'X-UA': x_ua})
+
+        video = traverse_obj(data, ('data', 'list', 0, {
+            'id': ('video_id', {str_or_none}),
+            'url': ('play_url', ('url', 'url_h265'), {url_or_none}),
+            'duration': ('info', 'duration', {int_or_none}),
+            'thumbnail': ('thumbnail', ('original_url', 'url'), {url_or_none}),
+        }), get_all=False)
+        if '.m3u8' in video['url']:
+            video['formats'] = self._extract_m3u8_formats(video.pop('url'), video_id)
+        return video
+
+    def _real_extract(self, url):
+        section, list_id = self._match_valid_url(url).groups()
+        list_id = f'{section}_{list_id}'
+
+        webpage = self._download_webpage(url, list_id)
+        nuxt_data = self._deserialize_nuxt_data(self._search_json(
+            r'<script[^>]+\bid=["\']__NUXT_DATA__["\'][^>]*>', webpage,
+            'nuxt data', list_id, contains_pattern=r'\[(?s:.+)\]'))[1]
+        x_ua = traverse_obj(nuxt_data, (
+            'state', '$sbff', ..., {lambda x: parse_qs(x)['X-UA']}, ...), get_all=False)
+
+        if section == 'moment':
+            moment_data = traverse_obj(nuxt_data, ('data', ..., 'moment'), get_all=False)
+            video_ids = traverse_obj(moment_data, ('topic', (('videos', ...), 'pin_video'), 'video_id'))
+            metainfo = traverse_obj(moment_data, {
+                'timestamp': ('created_time', {int_or_none}),
+                'uploader': ('author', 'user', 'name', {str}),
+                'title': ('topic', 'title', {str}),
+                'description': ('topic', 'summary', {str}),
+            })
+        elif section == 'app':
+            video_ids = traverse_obj(nuxt_data, ('data', ..., ('app_videos', 'videos'), ..., 'video_id'))
+            metainfo = traverse_obj(nuxt_data, ('data', ..., {
+                'title': ('title', {str}),
+                'description': ('description', 'text', {str}, {clean_html}),
+            }), get_all=False)
+
+        entries = [self._extract_video(video_id, x_ua) for video_id in set(video_ids)]
+
+        return self.playlist_result([{**metainfo, **e} for e in entries], **metainfo, id=list_id)

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()