[ie] Make _search_nextjs_data non fatal (#8937)

Authored by: Grub4K
This commit is contained in:
Simon Sawicki 2024-04-21 13:40:38 +02:00 committed by GitHub
parent e3b42d8b1b
commit 3ee1194288
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
5 changed files with 22 additions and 9 deletions

View File

@ -1906,6 +1906,15 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/
expected_status=TEAPOT_RESPONSE_STATUS) expected_status=TEAPOT_RESPONSE_STATUS)
self.assertEqual(content, TEAPOT_RESPONSE_BODY) self.assertEqual(content, TEAPOT_RESPONSE_BODY)
def test_search_nextjs_data(self):
data = '<script id="__NEXT_DATA__" type="application/json">{"props":{}}</script>'
self.assertEqual(self.ie._search_nextjs_data(data, None), {'props': {}})
self.assertEqual(self.ie._search_nextjs_data('', None, fatal=False), {})
self.assertEqual(self.ie._search_nextjs_data('', None, default=None), None)
self.assertEqual(self.ie._search_nextjs_data('', None, default={}), {})
with self.assertRaises(DeprecationWarning):
self.assertEqual(self.ie._search_nextjs_data('', None, default='{}'), {})
if __name__ == '__main__': if __name__ == '__main__':
unittest.main() unittest.main()

View File

@ -105,7 +105,7 @@ class AsobiStageIE(InfoExtractor):
video_type = {'archive': 'archives', 'player': 'broadcasts'}[type_] video_type = {'archive': 'archives', 'player': 'broadcasts'}[type_]
webpage = self._download_webpage(url, video_id) webpage = self._download_webpage(url, video_id)
event_data = traverse_obj( event_data = traverse_obj(
self._search_nextjs_data(webpage, video_id, default='{}'), self._search_nextjs_data(webpage, video_id, default={}),
('props', 'pageProps', 'eventCMSData', { ('props', 'pageProps', 'eventCMSData', {
'title': ('event_name', {str}), 'title': ('event_name', {str}),
'thumbnail': ('event_thumbnail_image', {url_or_none}), 'thumbnail': ('event_thumbnail_image', {url_or_none}),

View File

@ -1738,12 +1738,16 @@ class InfoExtractor:
traverse_json_ld(json_ld) traverse_json_ld(json_ld)
return filter_dict(info) return filter_dict(info)
def _search_nextjs_data(self, webpage, video_id, *, transform_source=None, fatal=True, **kw): def _search_nextjs_data(self, webpage, video_id, *, fatal=True, default=NO_DEFAULT, **kw):
return self._parse_json( if default == '{}':
self._search_regex( self._downloader.deprecation_warning('using `default=\'{}\'` is deprecated, use `default={}` instead')
r'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^>]*>([^<]+)</script>', default = {}
webpage, 'next.js data', fatal=fatal, **kw), if default is not NO_DEFAULT:
video_id, transform_source=transform_source, fatal=fatal) fatal = False
return self._search_json(
r'<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^>]*>', webpage, 'next.js data',
video_id, end_pattern='</script>', fatal=fatal, default=default, **kw)
def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__', *, fatal=True, traverse=('data', 0)): def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__', *, fatal=True, traverse=('data', 0)):
"""Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function""" """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""

View File

@ -41,7 +41,7 @@ class STVPlayerIE(InfoExtractor):
ptype, video_id = self._match_valid_url(url).groups() ptype, video_id = self._match_valid_url(url).groups()
webpage = self._download_webpage(url, video_id, fatal=False) or '' webpage = self._download_webpage(url, video_id, fatal=False) or ''
props = self._search_nextjs_data(webpage, video_id, default='{}').get('props') or {} props = self._search_nextjs_data(webpage, video_id, default={}).get('props') or {}
player_api_cache = try_get( player_api_cache = try_get(
props, lambda x: x['initialReduxState']['playerApiCache']) or {} props, lambda x: x['initialReduxState']['playerApiCache']) or {}

View File

@ -776,7 +776,7 @@ class TikTokIE(TikTokBaseIE):
status = traverse_obj(sigi_data, ('VideoPage', 'statusCode', {int})) or 0 status = traverse_obj(sigi_data, ('VideoPage', 'statusCode', {int})) or 0
video_data = traverse_obj(sigi_data, ('ItemModule', video_id, {dict})) video_data = traverse_obj(sigi_data, ('ItemModule', video_id, {dict}))
elif next_data := self._search_nextjs_data(webpage, video_id, default='{}'): elif next_data := self._search_nextjs_data(webpage, video_id, default={}):
self.write_debug('Found next.js data') self.write_debug('Found next.js data')
status = traverse_obj(next_data, ('props', 'pageProps', 'statusCode', {int})) or 0 status = traverse_obj(next_data, ('props', 'pageProps', 'statusCode', {int})) or 0
video_data = traverse_obj(next_data, ('props', 'pageProps', 'itemInfo', 'itemStruct', {dict})) video_data = traverse_obj(next_data, ('props', 'pageProps', 'itemInfo', 'itemStruct', {dict}))