mirrors/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2025-08-21 19:01:39 +00:00
Code Issues Projects Releases Wiki Activity

[ie/tiktok:user] Improve infinite loop prevention (#14077)

Browse Source 
Fix edf55e81842fcfa6c302528d7f33ccd5081b37ef

Closes #14076
Authored by: bashonly
This commit is contained in:
bashonly 2025-08-19 15:39:33 -05:00 committed by GitHub
parent df0553153e
commit 471a2b60e0
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 25 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
yt_dlp/extractor/tiktok.py
View File

@ -921,6 +921,8 @@ class TikTokUserIE(TikTokBaseIE):
'id': 'MS4wLjABAAAAepiJKgwWhulvCpSuUVsp7sgVVsFJbbNaLeQ6OQ0oAJERGDUIXhb2yxxHZedsItgT', 'id': 'MS4wLjABAAAAepiJKgwWhulvCpSuUVsp7sgVVsFJbbNaLeQ6OQ0oAJERGDUIXhb2yxxHZedsItgT',
'title': 'corgibobaa', 'title': 'corgibobaa',
}, },
'expected_warnings': ['TikTok API keeps sending the same page'],
'params': {'extractor_retries': 10},
}, { }, {
'url': 'https://www.tiktok.com/@6820838815978423302', 'url': 'https://www.tiktok.com/@6820838815978423302',
'playlist_mincount': 5, 'playlist_mincount': 5,
@ -928,6 +930,8 @@ class TikTokUserIE(TikTokBaseIE):
'id': 'MS4wLjABAAAA0tF1nBwQVVMyrGu3CqttkNgM68Do1OXUFuCY0CRQk8fEtSVDj89HqoqvbSTmUP2W', 'id': 'MS4wLjABAAAA0tF1nBwQVVMyrGu3CqttkNgM68Do1OXUFuCY0CRQk8fEtSVDj89HqoqvbSTmUP2W',
'title': '6820838815978423302', 'title': '6820838815978423302',
}, },
'expected_warnings': ['TikTok API keeps sending the same page'],
'params': {'extractor_retries': 10},
}, { }, {
'url': 'https://www.tiktok.com/@meme', 'url': 'https://www.tiktok.com/@meme',
'playlist_mincount': 593, 'playlist_mincount': 593,
@ -935,12 +939,16 @@ class TikTokUserIE(TikTokBaseIE):
'id': 'MS4wLjABAAAAiKfaDWeCsT3IHwY77zqWGtVRIy9v4ws1HbVi7auP1Vx7dJysU_hc5yRiGywojRD6', 'id': 'MS4wLjABAAAAiKfaDWeCsT3IHwY77zqWGtVRIy9v4ws1HbVi7auP1Vx7dJysU_hc5yRiGywojRD6',
'title': 'meme', 'title': 'meme',
}, },
'expected_warnings': ['TikTok API keeps sending the same page'],
'params': {'extractor_retries': 10},
}, { }, {
'url': 'tiktokuser:MS4wLjABAAAAM3R2BtjzVT-uAtstkl2iugMzC6AtnpkojJbjiOdDDrdsTiTR75-8lyWJCY5VvDrZ', 'url': 'tiktokuser:MS4wLjABAAAAM3R2BtjzVT-uAtstkl2iugMzC6AtnpkojJbjiOdDDrdsTiTR75-8lyWJCY5VvDrZ',
'playlist_mincount': 31, 'playlist_mincount': 31,
'info_dict': { 'info_dict': {
'id': 'MS4wLjABAAAAM3R2BtjzVT-uAtstkl2iugMzC6AtnpkojJbjiOdDDrdsTiTR75-8lyWJCY5VvDrZ', 'id': 'MS4wLjABAAAAM3R2BtjzVT-uAtstkl2iugMzC6AtnpkojJbjiOdDDrdsTiTR75-8lyWJCY5VvDrZ',
}, },
'expected_warnings': ['TikTok API keeps sending the same page'],
'params': {'extractor_retries': 10},
}] }]
_API_BASE_URL = 'https://www.tiktok.com/api/creator/item_list/' _API_BASE_URL = 'https://www.tiktok.com/api/creator/item_list/'
@ -979,7 +987,7 @@ class TikTokUserIE(TikTokBaseIE):
'webcast_language': 'en', 'webcast_language': 'en',
} }
def _entries(self, sec_uid, user_name): def _entries(self, sec_uid, user_name, fail_early=False):
display_id = user_name or sec_uid display_id = user_name or sec_uid
seen_ids = set() seen_ids = set()
@ -1023,10 +1031,13 @@ class TikTokUserIE(TikTokBaseIE):
if cursor < 1472706000000 or not traverse_obj(response, 'hasMorePrevious'): if cursor < 1472706000000 or not traverse_obj(response, 'hasMorePrevious'):
return return
# User directly passed sec_uid via prefix URL, bypassing our private account detection # This code path is ideally only reached when one of the following is true:
if not user_name and not seen_ids: # 1. TikTok profile is private and webpage detection was bypassed due to a tiktokuser:sec_uid URL
# 2. TikTok profile is *not* private but all of their videos are private
if fail_early and not seen_ids:
self.raise_login_required( self.raise_login_required(
'This user\'s account is likely private. Log into an account that has access') 'This user\'s account is likely either private or all of their videos are private. '
'Log into an account that has access')
def _extract_sec_uid_from_embed(self, user_name): def _extract_sec_uid_from_embed(self, user_name):
webpage = self._download_webpage( webpage = self._download_webpage(
@ -1053,8 +1064,9 @@ class TikTokUserIE(TikTokBaseIE):
def _real_extract(self, url): def _real_extract(self, url):
user_name, sec_uid = self._match_id(url), None user_name, sec_uid = self._match_id(url), None
if mobj := re.fullmatch(r'MS4wLjABAAAA[\w-]{64}', user_name): if re.fullmatch(r'MS4wLjABAAAA[\w-]{64}', user_name):
user_name, sec_uid = None, mobj.group(0) user_name, sec_uid = None, user_name
fail_early = True
else: else:
webpage = self._download_webpage( webpage = self._download_webpage(
self._UPLOADER_URL_FORMAT % user_name, user_name, self._UPLOADER_URL_FORMAT % user_name, user_name,
@ -1065,8 +1077,12 @@ class TikTokUserIE(TikTokBaseIE):
if detail.get('statusCode') == 10222: if detail.get('statusCode') == 10222:
self.raise_login_required( self.raise_login_required(
'This user\'s account is private. Log into an account that has access') 'This user\'s account is private. Log into an account that has access')
sec_uid = traverse_obj(detail, ( sec_uid = traverse_obj(detail, ('userInfo', 'user', 'secUid', {str}))
'userInfo', 'user', 'secUid', {str})) or self._extract_sec_uid_from_embed(user_name) if sec_uid:
fail_early = not traverse_obj(detail, ('userInfo', 'itemList', ...))
else:
sec_uid = self._extract_sec_uid_from_embed(user_name)
fail_early = False
if not sec_uid: if not sec_uid:
raise ExtractorError( raise ExtractorError(
@ -1074,7 +1090,7 @@ class TikTokUserIE(TikTokBaseIE):
'from a video posted by this user, try using "tiktokuser:channel_id" as the ' 'from a video posted by this user, try using "tiktokuser:channel_id" as the '
'input URL (replacing `channel_id` with its actual value)', expected=True) 'input URL (replacing `channel_id` with its actual value)', expected=True)
return self.playlist_result(self._entries(sec_uid, user_name), sec_uid, user_name) return self.playlist_result(self._entries(sec_uid, user_name, fail_early), sec_uid, user_name)
class TikTokBaseListIE(TikTokBaseIE): # XXX: Conventionally, base classes should end with BaseIE/InfoExtractor class TikTokBaseListIE(TikTokBaseIE): # XXX: Conventionally, base classes should end with BaseIE/InfoExtractor