Merge 1ef35f1c00 into c699bafc50

Closes #11537

Authored by: bashonly

test/test_traversal.py

@@ -481,7 +481,7 @@ class TestTraversalHelpers:
             'id': 'name',
             'data': 'content',
             'url': 'url',
-        }, all, {subs_list_to_dict}]) == {
+        }, all, {subs_list_to_dict(lang=None)}]) == {
             'de': [{'url': 'https://example.com/subs/de.ass'}],
             'en': [{'data': 'content'}],
         }, 'subs with mandatory items missing should be filtered'
@@ -507,6 +507,54 @@ class TestTraversalHelpers:
             {'url': 'https://example.com/subs/en1', 'ext': 'ext'},
             {'url': 'https://example.com/subs/en2', 'ext': 'ext'},
         ]}, '`quality` key should sort subtitle list accordingly'
+        assert traverse_obj([
+            {'name': 'de', 'url': 'https://example.com/subs/de.ass'},
+            {'name': 'de'},
+            {'name': 'en', 'content': 'content'},
+            {'url': 'https://example.com/subs/en'},
+        ], [..., {
+            'id': 'name',
+            'url': 'url',
+            'data': 'content',
+        }, all, {subs_list_to_dict(lang='en')}]) == {
+            'de': [{'url': 'https://example.com/subs/de.ass'}],
+            'en': [
+                {'data': 'content'},
+                {'url': 'https://example.com/subs/en'},
+            ],
+        }, 'optionally provided lang should be used if no id available'
+        assert traverse_obj([
+            {'name': 1, 'url': 'https://example.com/subs/de1'},
+            {'name': {}, 'url': 'https://example.com/subs/de2'},
+            {'name': 'de', 'ext': 1, 'url': 'https://example.com/subs/de3'},
+            {'name': 'de', 'ext': {}, 'url': 'https://example.com/subs/de4'},
+        ], [..., {
+            'id': 'name',
+            'url': 'url',
+            'ext': 'ext',
+        }, all, {subs_list_to_dict(lang=None)}]) == {
+            'de': [
+                {'url': 'https://example.com/subs/de3'},
+                {'url': 'https://example.com/subs/de4'},
+            ],
+        }, 'non str types should be ignored for id and ext'
+        assert traverse_obj([
+            {'name': 1, 'url': 'https://example.com/subs/de1'},
+            {'name': {}, 'url': 'https://example.com/subs/de2'},
+            {'name': 'de', 'ext': 1, 'url': 'https://example.com/subs/de3'},
+            {'name': 'de', 'ext': {}, 'url': 'https://example.com/subs/de4'},
+        ], [..., {
+            'id': 'name',
+            'url': 'url',
+            'ext': 'ext',
+        }, all, {subs_list_to_dict(lang='de')}]) == {
+            'de': [
+                {'url': 'https://example.com/subs/de1'},
+                {'url': 'https://example.com/subs/de2'},
+                {'url': 'https://example.com/subs/de3'},
+                {'url': 'https://example.com/subs/de4'},
+            ],
+        }, 'non str types should be replaced by default id'
 
     def test_trim_str(self):
         with pytest.raises(TypeError):
@@ -525,7 +573,7 @@ class TestTraversalHelpers:
     def test_unpack(self):
         assert unpack(lambda *x: ''.join(map(str, x)))([1, 2, 3]) == '123'
         assert unpack(join_nonempty)([1, 2, 3]) == '1-2-3'
-        assert unpack(join_nonempty(delim=' '))([1, 2, 3]) == '1 2 3'
+        assert unpack(join_nonempty, delim=' ')([1, 2, 3]) == '1 2 3'
         with pytest.raises(TypeError):
             unpack(join_nonempty)()
         with pytest.raises(TypeError):

test/test_utils.py

@@ -72,7 +72,6 @@ from yt_dlp.utils import (
     intlist_to_bytes,
     iri_to_uri,
     is_html,
-    join_nonempty,
     js_to_json,
     limit_length,
     locked_file,
@@ -2158,10 +2157,6 @@ Line 1
         assert int_or_none(v=10) == 10, 'keyword passed positional should call function'
         assert int_or_none(scale=0.1)(10) == 100, 'call after partial application should call the function'
 
-        assert callable(join_nonempty(delim=', ')), 'varargs positional should apply partially'
-        assert callable(join_nonempty()), 'varargs positional should apply partially'
-        assert join_nonempty(None, delim=', ') == '', 'passed varargs should call the function'
-
 
 if __name__ == '__main__':
     unittest.main()

yt_dlp/YoutubeDL.py

@@ -4381,7 +4381,9 @@ class YoutubeDL:
             return None
 
         for idx, t in list(enumerate(thumbnails))[::-1]:
-            thumb_ext = (f'{t["id"]}.' if multiple else '') + determine_ext(t['url'], 'jpg')
+            thumb_ext = t.get('ext') or determine_ext(t['url'], 'jpg')
+            if multiple:
+                thumb_ext = f'{t["id"]}.{thumb_ext}'
             thumb_display_id = f'{label} thumbnail {t["id"]}'
             thumb_filename = replace_extension(filename, thumb_ext, info_dict.get('ext'))
             thumb_filename_final = replace_extension(thumb_filename_base, thumb_ext, info_dict.get('ext'))

yt_dlp/extractor/afreecatv.py

@@ -66,6 +66,14 @@ class AfreecaTVBaseIE(InfoExtractor):
             extensions={'legacy_ssl': True}), display_id,
             'Downloading API JSON', 'Unable to download API JSON')
 
+    @staticmethod
+    def _fixup_thumb(thumb_url):
+        if not url_or_none(thumb_url):
+            return None
+        # Core would determine_ext as 'php' from the url, so we need to provide the real ext
+        # See: https://github.com/yt-dlp/yt-dlp/issues/11537
+        return [{'url': thumb_url, 'ext': 'jpg'}]
+
 
 class AfreecaTVIE(AfreecaTVBaseIE):
     IE_NAME = 'soop'
@@ -155,7 +163,7 @@ class AfreecaTVIE(AfreecaTVBaseIE):
             'uploader': ('writer_nick', {str}),
             'uploader_id': ('bj_id', {str}),
             'duration': ('total_file_duration', {int_or_none(scale=1000)}),
-            'thumbnail': ('thumb', {url_or_none}),
+            'thumbnails': ('thumb', {self._fixup_thumb}),
         })
 
         entries = []
@@ -226,8 +234,7 @@ class AfreecaTVCatchStoryIE(AfreecaTVBaseIE):
 
         return self.playlist_result(self._entries(data), video_id)
 
-    @staticmethod
+    def _entries(self, data):
-    def _entries(data):
         # 'files' is always a list with 1 element
         yield from traverse_obj(data, (
             'data', lambda _, v: v['story_type'] == 'catch',
@@ -238,7 +245,7 @@ class AfreecaTVCatchStoryIE(AfreecaTVBaseIE):
                 'title': ('title', {str}),
                 'uploader': ('writer_nick', {str}),
                 'uploader_id': ('writer_id', {str}),
-                'thumbnail': ('thumb', {url_or_none}),
+                'thumbnails': ('thumb', {self._fixup_thumb}),
                 'timestamp': ('write_timestamp', {int_or_none}),
             }))
 

yt_dlp/extractor/common.py

@@ -279,6 +279,7 @@ class InfoExtractor:
     thumbnails:     A list of dictionaries, with the following entries:
                         * "id" (optional, string) - Thumbnail format ID
                         * "url"
+                        * "ext" (optional, string) - actual image extension if not given in URL
                         * "preference" (optional, int) - quality of the image
                         * "width" (optional, int)
                         * "height" (optional, int)

yt_dlp/extractor/googledrive.py

@@ -3,6 +3,7 @@ import urllib.parse
 
 from .common import InfoExtractor
 from .youtube import YoutubeIE
+from ..networking.exceptions import HTTPError
 from ..utils import (
     ExtractorError,
     bug_reports_message,
@@ -12,6 +13,7 @@ from ..utils import (
     get_element_html_by_id,
     int_or_none,
     lowercase_escape,
+    traverse_obj,
     try_get,
     update_url_query,
 )
@@ -51,6 +53,17 @@ class GoogleDriveIE(InfoExtractor):
             'duration': 184,
             'thumbnail': 'https://drive.google.com/thumbnail?id=1IP0o8dHcQrIHGgVyp0Ofvx2cGfLzyO1x',
         },
+    }, {
+        # shortcut url
+        'url': 'https://drive.google.com/file/d/1_n3-8ZwEUV4OniMsLAJ_C1JEjuT2u5Pk/view?usp=drivesdk',
+        'md5': '43d34f7be1acc0262f337a039d1ad12d',
+        'info_dict': {
+            'id': '1J1RCw2jcgUngrZRdpza-IHXYkardZ-4l',
+            'ext': 'webm',
+            'title': 'Forrest walk with Best Mind Refresh Music Mithran [tEvJKrE4cS0].webm',
+            'duration': 512,
+            'thumbnail': 'https://drive.google.com/thumbnail?id=1J1RCw2jcgUngrZRdpza-IHXYkardZ-4l',
+        },
     }, {
         # video can't be watched anonymously due to view count limit reached,
         # but can be downloaded (see https://github.com/ytdl-org/youtube-dl/issues/14046)
@@ -166,6 +179,17 @@ class GoogleDriveIE(InfoExtractor):
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
+        try:
+            _, webpage_urlh = self._download_webpage_handle(url, video_id)
+        except ExtractorError as e:
+            if isinstance(e.cause, HTTPError):
+                if e.cause.status in (401, 403):
+                    self.raise_login_required('Access Denied')
+                raise
+        if webpage_urlh.url != url:
+            url = webpage_urlh.url
+            video_id = self._match_id(url)
+
         video_info = urllib.parse.parse_qs(self._download_webpage(
             'https://drive.google.com/get_video_info',
             video_id, 'Downloading video webpage', query={'docid': video_id}))
@@ -289,7 +313,7 @@ class GoogleDriveIE(InfoExtractor):
 
 class GoogleDriveFolderIE(InfoExtractor):
     IE_NAME = 'GoogleDrive:Folder'
-    _VALID_URL = r'https?://(?:docs|drive)\.google\.com/drive/folders/(?P<id>[\w-]{28,})'
+    _VALID_URL = r'https?://(?:docs|drive)\.google\.com/drive/(?:folders/(?P<id>[\w-]{19,})|my-drive)'
     _TESTS = [{
         'url': 'https://drive.google.com/drive/folders/1dQ4sx0-__Nvg65rxTSgQrl7VyW_FZ9QI',
         'info_dict': {
@@ -297,47 +321,83 @@ class GoogleDriveFolderIE(InfoExtractor):
             'title': 'Forrest',
         },
         'playlist_count': 3,
+    }, {
+        'note': 'Contains various formats and a subfolder, folder name was formerly mismatched.'
+                'also contains loop shortcut, shortcut to non-downloadable files, etc.',
+        'url': 'https://docs.google.com/drive/folders/1jjrhqi94d8TSHSVMSdBjD49MOiHYpHfF',
+        'info_dict': {
+            'id': '1jjrhqi94d8TSHSVMSdBjD49MOiHYpHfF',
+            'title': '], sideChannel: {}});',
+        },
+        'playlist_count': 8,
     }]
-    _BOUNDARY = '=====vc17a3rwnndj====='
-    _REQUEST = "/drive/v2beta/files?openDrive=true&reason=102&syncType=0&errorRecovery=false&q=trashed%20%3D%20false%20and%20'{folder_id}'%20in%20parents&fields=kind%2CnextPageToken%2Citems(kind%2CmodifiedDate%2CmodifiedByMeDate%2ClastViewedByMeDate%2CfileSize%2Cowners(kind%2CpermissionId%2Cid)%2ClastModifyingUser(kind%2CpermissionId%2Cid)%2ChasThumbnail%2CthumbnailVersion%2Ctitle%2Cid%2CresourceKey%2Cshared%2CsharedWithMeDate%2CuserPermission(role)%2CexplicitlyTrashed%2CmimeType%2CquotaBytesUsed%2Ccopyable%2CfileExtension%2CsharingUser(kind%2CpermissionId%2Cid)%2Cspaces%2Cversion%2CteamDriveId%2ChasAugmentedPermissions%2CcreatedDate%2CtrashingUser(kind%2CpermissionId%2Cid)%2CtrashedDate%2Cparents(id)%2CshortcutDetails(targetId%2CtargetMimeType%2CtargetLookupStatus)%2Ccapabilities(canCopy%2CcanDownload%2CcanEdit%2CcanAddChildren%2CcanDelete%2CcanRemoveChildren%2CcanShare%2CcanTrash%2CcanRename%2CcanReadTeamDrive%2CcanMoveTeamDriveItem)%2Clabels(starred%2Ctrashed%2Crestricted%2Cviewed))%2CincompleteSearch&appDataFilter=NO_APP_DATA&spaces=drive&pageToken={page_token}&maxResults=50&supportsTeamDrives=true&includeItemsFromAllDrives=true&corpora=default&orderBy=folder%2Ctitle_natural%20asc&retryCount=0&key={key} HTTP/1.1"
-    _DATA = f'''--{_BOUNDARY}
-content-type: application/http
-content-transfer-encoding: binary
 
-GET %s
+    def _extract_json_meta(self, webpage, video_id, dsval=None, hashval=None, name=None, **kwargs):
-
+        """
---{_BOUNDARY}
+        Uses regex to search for json metadata with 'ds' value(0-5) or 'hash' value(1-6)
-'''
+        from the webpage.
-
+            logged out  folder info:ds0hash1;          items:ds4hash6
-    def _call_api(self, folder_id, key, data, **kwargs):
+            logged in   folder info:ds0hash1;          items:ds5hash6
-        response = self._download_webpage(
+            my-drive    folder info:ds0hash1/ds0hash4; items:ds5hash6
-            'https://clients6.google.com/batch/drive/v2beta',
+        For example, if the webpage contains the line below, the empty data array
-            folder_id, data=data.encode(),
+        can be got by passing dsval=3 or hashval=2 to this method.
-            headers={
+            AF_initDataCallback({key: 'ds:3', hash: '2', data:[], sideChannel: {}});
-                'Content-Type': 'text/plain;charset=UTF-8;',
+        """
-                'Origin': 'https://drive.google.com',
+        _ARRAY_RE = r'\[(?s:.+)\]'
-            }, query={
+        _META_END_RE = r', sideChannel: \{\}\}\);'  # greedy match to deal with the 2nd test case
-                '$ct': f'multipart/mixed; boundary="{self._BOUNDARY}"',
+        if dsval is not None:
-                'key': key,
+            if not name:
-            }, **kwargs)
+                name = f'webpage JSON metadata ds:{dsval}'
-        return self._search_json('', response, 'api response', folder_id, **kwargs) or {}
+            return self._search_json(
-
+                rf'''key\s*?:\s*?(['"])ds:\s*?{dsval}\1,[^\[]*?data:''', webpage, name, video_id,
-    def _get_folder_items(self, folder_id, key):
+                end_pattern=_META_END_RE, contains_pattern=_ARRAY_RE, **kwargs)
-        page_token = ''
+        elif hashval is not None:
-        while page_token is not None:
+            if not name:
-            request = self._REQUEST.format(folder_id=folder_id, page_token=page_token, key=key)
+                name = f'webpage JSON metadata hash:{hashval}'
-            page = self._call_api(folder_id, key, self._DATA % request)
+            return self._search_json(
-            yield from page['items']
+                rf'''hash\s*?:\s*?(['"]){hashval}\1,[^\[]*?data:''', webpage, name, video_id,
-            page_token = page.get('nextPageToken')
+                end_pattern=_META_END_RE, contains_pattern=_ARRAY_RE, **kwargs)
 
     def _real_extract(self, url):
-        folder_id = self._match_id(url)
+        def item_url_getter(item, video_id):
+            if not isinstance(item, list):
+                return None
+            available_IEs = (GoogleDriveFolderIE, GoogleDriveIE)  # subfolder or item
+            if 'application/vnd.google-apps.shortcut' in item:  # extract real link
+                entry_url = traverse_obj(
+                    item,
+                    (..., ..., lambda _, v: any(ie.suitable(v) for ie in available_IEs), any))
+            else:
+                entry_url = traverse_obj(
+                    item,
+                    (lambda _, v: any(ie.suitable(v) for ie in available_IEs), any))
+            if not entry_url:
+                return None
+            return self.url_result(entry_url, video_id=video_id, video_title=item[2])
 
-        webpage = self._download_webpage(url, folder_id)
+        folder_id = self._match_id(url) or 'my-drive'
-        key = self._search_regex(r'"(\w{39})"', webpage, 'key')
+        headers = self.geo_verification_headers()
 
-        folder_info = self._call_api(folder_id, key, self._DATA % f'/drive/v2beta/files/{folder_id} HTTP/1.1', fatal=False)
+        try:
+            webpage, urlh = self._download_webpage_handle(url, folder_id, headers=headers)
+        except ExtractorError as e:
+            if isinstance(e.cause, HTTPError):
+                if e.cause.status == 404:
+                    self.raise_no_formats(e.cause.msg, expected=True)
+                elif e.cause.status == 403:
+                    # logged in with an account without access
+                    self.raise_login_required('Access Denied')
+            raise
+        if urllib.parse.urlparse(urlh.url).netloc == 'accounts.google.com':
+            # not logged in when visiting a private folder
+            self.raise_login_required('Access Denied')
 
-        return self.playlist_from_matches(
+        title = self._extract_json_meta(webpage, folder_id, dsval=0, name='folder info')[1][2]
-            self._get_folder_items(folder_id, key), folder_id, folder_info.get('title'),
+        items = self._extract_json_meta(webpage, folder_id, hashval=6, name='folder items')[-1]
-            ie=GoogleDriveIE, getter=lambda item: f'https://drive.google.com/file/d/{item["id"]}')
+
+        if items is False:  # empty folder
+            return self.playlist_result([], folder_id, title)
+
+        return self.playlist_result(
+            (entry for item in items if (entry := item_url_getter(item, folder_id))),
+            folder_id, title)

yt_dlp/utils/_utils.py

@@ -216,7 +216,7 @@ def partial_application(func):
     sig = inspect.signature(func)
     required_args = [
         param.name for param in sig.parameters.values()
-        if param.kind in (inspect.Parameter.POSITIONAL_ONLY, inspect.Parameter.POSITIONAL_OR_KEYWORD, inspect.Parameter.VAR_POSITIONAL)
+        if param.kind in (inspect.Parameter.POSITIONAL_ONLY, inspect.Parameter.POSITIONAL_OR_KEYWORD)
         if param.default is inspect.Parameter.empty
     ]
 
@@ -4837,7 +4837,6 @@ def number_of_digits(number):
     return len('%d' % number)
 
 
-@partial_application
 def join_nonempty(*values, delim='-', from_dict=None):
     if from_dict is not None:
         values = (traversal.traverse_obj(from_dict, variadic(v)) for v in values)

yt_dlp/utils/traversal.py

@@ -332,14 +332,14 @@ class _RequiredError(ExtractorError):
 
 
 @typing.overload
-def subs_list_to_dict(*, ext: str | None = None) -> collections.abc.Callable[[list[dict]], dict[str, list[dict]]]: ...
+def subs_list_to_dict(*, lang: str | None = 'und', ext: str | None = None) -> collections.abc.Callable[[list[dict]], dict[str, list[dict]]]: ...
 
 
 @typing.overload
-def subs_list_to_dict(subs: list[dict] | None, /, *, ext: str | None = None) -> dict[str, list[dict]]: ...
+def subs_list_to_dict(subs: list[dict] | None, /, *, lang: str | None = 'und', ext: str | None = None) -> dict[str, list[dict]]: ...
 
 
-def subs_list_to_dict(subs: list[dict] | None = None, /, *, ext=None):
+def subs_list_to_dict(subs: list[dict] | None = None, /, *, lang='und', ext=None):
     """
     Convert subtitles from a traversal into a subtitle dict.
     The path should have an `all` immediately before this function.
@@ -352,7 +352,7 @@ def subs_list_to_dict(subs: list[dict] | None = None, /, *, ext=None):
     `quality`  The sort order for each subtitle
     """
     if subs is None:
-        return functools.partial(subs_list_to_dict, ext=ext)
+        return functools.partial(subs_list_to_dict, lang=lang, ext=ext)
 
     result = collections.defaultdict(list)
 
@@ -360,10 +360,16 @@ def subs_list_to_dict(subs: list[dict] | None = None, /, *, ext=None):
         if not url_or_none(sub.get('url')) and not sub.get('data'):
             continue
         sub_id = sub.pop('id', None)
-        if sub_id is None:
+        if not isinstance(sub_id, str):
-            continue
+            if not lang:
-        if ext is not None and not sub.get('ext'):
+                continue
-            sub['ext'] = ext
+            sub_id = lang
+        sub_ext = sub.get('ext')
+        if not isinstance(sub_ext, str):
+            if not ext:
+                sub.pop('ext', None)
+            else:
+                sub['ext'] = ext
         result[sub_id].append(sub)
     result = dict(result)
 
@@ -452,9 +458,9 @@ def trim_str(*, start=None, end=None):
     return trim
 
 
-def unpack(func):
+def unpack(func, **kwargs):
     @functools.wraps(func)
-    def inner(items, **kwargs):
+    def inner(items):
         return func(*items, **kwargs)
 
     return inner