Merge ab81421072 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

@@ -1341,6 +1341,7 @@ from .nubilesporn import NubilesPornIE
 from .nytimes import (
     NYTimesIE,
     NYTimesArticleIE,
+    NYTimesAudioIE,
     NYTimesCookingIE,
     NYTimesCookingRecipeIE,
 )

yt_dlp/extractor/nytimes.py

@@ -10,6 +10,7 @@ from ..utils import (
     float_or_none,
     get_elements_html_by_class,
     int_or_none,
+    js_to_json,
     merge_dicts,
     mimetype2ext,
     parse_iso8601,
@@ -418,3 +419,223 @@ class NYTimesCookingRecipeIE(InfoExtractor):
             'thumbnails': [{'url': thumb_url} for thumb_url in traverse_obj(
                 recipe_data, ('image', 'crops', 'recipe', ..., {url_or_none}))],
         }
+
+
+class NYTimesAudioIE(NYTimesBaseIE):
+    _VALID_URL = r"https?://(?:www\.)?nytimes\.com/\d{4}/\d{2}/\d{2}/(?:podcasts|books)/(?:[\w-]+/)?(?P<id>[^./?#]+)(?:\.html)?"
+    _TESTS = [
+        {
+            "url": "http://www.nytimes.com/2016/10/14/podcasts/revelations-from-the-final-weeks.html",
+            "md5": "cd402e44a059c8caf3b5f514c9264d0f",
+            "info_dict": {
+                "id": "100000004709062",
+                "title": "Revelations From the Final Weeks",
+                "ext": "mp3",
+                "description": "md5:fb5c6b93b12efc51649b4847fe066ee4",
+                "timestamp": 1476448332,
+                "upload_date": "20161014",
+                "creators": [''],
+                "series": "The Run-Up",
+                "episode": "‘He Was Like an Octopus’",
+                "episode_number": 20,
+                "duration": 2130,
+                "thumbnail": r"re:https?://\w+\.nyt.com/images/.*\.jpg",
+            },
+        },
+        {
+            "url": "https://www.nytimes.com/2023/11/25/podcasts/poultry-slam.html",
+            "info_dict": {
+                "id": "100000009191248",
+                "title": "Poultry Slam",
+                "ext": "mp3",
+                "description": "md5:1e6f16b21bb9287b8a1fe563145a72fe",
+                "timestamp": 1700911084,
+                "upload_date": "20231125",
+                "creators": [],
+                "series": "This American Life",
+                "episode": "Poultry Slam",
+                "duration": 3523,
+                "thumbnail": r"re:https?://\w+\.nyt.com/images/.*\.png",
+            },
+            "params": {
+                "skip_download": True,
+            },
+        },
+        {
+            "url": "http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html",
+            "info_dict": {
+                "id": "100000004709479",
+                "title": "Inside The New York Times Book Review: The Rise of Hitler",
+                "ext": "mp3",
+                "description": "md5:288161c98c098a0c24f07a94af7108c3",
+                "timestamp": 1476461513,
+                "upload_date": "20161014",
+                "creators": ['Pamela Paul'],
+                "series": "",
+                "episode": "The Rise of Hitler",
+                "duration": 3475,
+                "thumbnail": r"re:https?://\w+\.nyt.com/images/.*\.jpg",
+            },
+            "params": {
+                "skip_download": True,
+            },
+        },
+        {
+            "url": "https://www.nytimes.com/2023/12/07/podcasts/the-daily/nikki-haley.html",
+            "info_dict": {
+                "id": "100000009214128",
+                "title": "Nikki Haley’s Moment",
+                "ext": "mp3",
+                "description": "md5:bf9f532fe689967ef1c458bcb057f3e5",
+                "timestamp": 1701946819,
+                "upload_date": "20231207",
+                "creators": [],
+                "series": "The Daily",
+                "episode": "Listen to ‘The Daily’: Nikki Haley’s Moment",
+                "duration": 1908,
+            },
+            "params": {
+                "skip_download": True,
+            },
+        },
+        {
+            "url": "https://www.nytimes.com/2023/12/18/podcasts/israel-putin.html",
+            "md5": "708b4fd393ca103280fe9e56d91b08b5",
+            "info_dict": {
+                "id": "100000009227362",
+                "title": "Pressure Mounts on Israel, and Putin Profits Off Boycott",
+                "ext": "mp3",
+                "description": "Hear the news in five minutes.",
+                "timestamp": 1702897212,
+                "upload_date": "20231218",
+                "creators": [],
+                "series": "The Headlines",
+                "episode": "The Headlines",
+                "duration": 298,
+                "thumbnail": r"re:https?://\w+\.nyt.com/images/.*\.jpg",
+            },
+        },
+    ]
+
+    def _extract_content_from_block(self, block):
+        return traverse_obj(
+            block,
+            {
+                "creators": ("data", "track", "credit", all),
+                "duration": (
+                    ("data", "media"),
+                    ("track", "length"),
+                    ("duration", None),
+                    {int_or_none},
+                ),
+                "series": (
+                    ("data", "media"),
+                    ("podcast", "podcastSeries"),
+                    ("title", None),
+                    {str_or_none},
+                ),
+                "episode": (
+                    ("data", "media"),
+                    ("track", "headline"),
+                    ("title", "default"), {str}),
+                "episode_number": (
+                    "data",
+                    "podcast",
+                    "episode",
+                    {lambda v: v.split()[1]},
+                    {int_or_none},
+                ),
+                "url": (
+                    ("data", "media"),
+                    ("track", "fileUrl"),
+                    ("source", None),
+                    {url_or_none},
+                ),
+                "vcodec": "none",
+            },
+            get_all=False,
+        )
+
+    def _real_extract(self, url):
+        page_id = self._match_id(url)
+        webpage = self._download_webpage(url, page_id)
+
+        art_json = self._search_json(
+            r"window\.__preloadedData\s*=",
+            webpage,
+            "media details",
+            page_id,
+            transform_source=js_to_json,
+        )["initialData"]["data"]["article"]
+
+        blocks = traverse_obj(
+            art_json,
+            (
+                "sprinkledBody",
+                "content",
+                lambda _, v: v["__typename"]
+                in ("InteractiveBlock", "HeaderMultimediaBlock"),
+                "media",
+            ),
+        )
+        if not blocks:
+            raise ExtractorError("Unable to extract any media blocks from webpage")
+
+        common_info = {
+            "title": remove_end(
+                self._html_extract_title(webpage), " - The New York Times"
+            ),
+            "description": self._html_search_meta(
+                ["og:description", "twitter:description"], webpage
+            ),
+            "id": traverse_obj(
+                art_json, ("sourceId")
+            ),  # poltry slam is under art_json > 'sourceId'
+            **traverse_obj(
+                art_json,
+                {
+                    "id": (
+                        "sprinkledBody",
+                        "content",
+                        ...,
+                        "media",
+                        "sourceId",
+                        any,
+                        {str},
+                    ),
+                    "title": ("headline", "default"),
+                    "description": ("summary"),
+                    "timestamp": ("firstPublished", {parse_iso8601}),
+                    "thumbnails": (
+                        "promotionalMedia",
+                        "assetCrops",
+                        ...,
+                        "renditions",
+                        ...,
+                        all,
+                        {self._extract_thumbnails},
+                    ),
+                },
+            ),
+        }
+
+        entries = []
+        for block in blocks:
+            if block.get("html"):
+                block = self._search_json(
+                    r"function\s+getFlexData\(\)\s*\{\s*return",
+                    block.get("html"),
+                    "Retrieve the inner JSON",
+                    page_id,
+                )
+            entries.append(
+                merge_dicts(self._extract_content_from_block(block), common_info)
+            )
+
+        if len(entries) > 1:
+            return self.playlist_result(entries, page_id, **common_info)
+
+        return {
+            "id": page_id,
+            **entries[0],
+        }

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/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()