DuckDuckWhale/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-12-16 14:36:47 +00:00
Code Issues Releases Wiki Activity

[Youtube] Rework n function extraction pattern

Browse Source 
Now also succeeds with player b12cc44b
This commit is contained in:
dirkf 2024-08-06 20:59:09 +01:00
parent dbc08fba83
commit c5098961b0
1 changed files with 26 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
youtube_dl/extractor/youtube.py
View File

@ -1661,23 +1661,33 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
# new: (b=String.fromCharCode(110),c=a.get(b))&&c=nfunc[idx](c) # new: (b=String.fromCharCode(110),c=a.get(b))&&c=nfunc[idx](c)
# or: (b="nn"[+a.D],c=a.get(b))&&(c=nfunc[idx](c) # or: (b="nn"[+a.D],c=a.get(b))&&(c=nfunc[idx](c)
# or: (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b) # or: (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b)
# old: .get("n"))&&(b=nfunc[idx](b) # or: (b="nn"[+a.D],vL(a),c=a.j[b]||null)&&(c=narray[idx](c),a.set(b,c),narray.length||nfunc("")
# older: .get("n"))&&(b=nfunc(b) # old: (b=a.get("n"))&&(b=nfunc[idx](b)(?P<c>[a-z])\s*=\s*[a-z]\s*
# older: (b=a.get("n"))&&(b=nfunc(b)
r'''(?x) r'''(?x)
(?:\((?:[\w$()\s]+,)*?\s*(?P<b>[a-z])\s*=\s*(?: \((?:[\w$()\s]+,)*?\s* # (
String\s*\.\s*fromCharCode\s*\(\s*110\s*\)| (?P<b>[a-z])\s*=\s* # b=
"n+"\[\s*\+?s*[\w$.]+\s*]| (?:
(?P<b1>(?:[\w$]+\s*\.\s*)+n\b(?:(?!&&).)+\)) (?: # expect ,c=a.get(b) (etc)
)\s* String\s*\.\s*fromCharCode\s*\(\s*110\s*\)|
(?(b1) "n+"\[\s*\+?s*[\w$.]+\s*]
&&\s*\(\s*(?P=b)| )\s*(?:,[\w$()\s]+(?=,))*|
(?: (?P<old>[\w$]+) # a (old[er])
,(?P<c>[a-z])\s*=\s*[a-z]\s*)? )\s*
\.\s*get\s*\(\s*(?(b)(?P=b)|"n{1,2}")(?:\s*\)){2}\s* (?(old)
&&\s*\(\s*(?(c)(?P=c)|(?P=b)) # b.get("n")
) (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*?
)\s*=\s* (?:\.\s*n|\[\s*"n"\s*]|\.\s*get\s*\(\s*"n"\s*\))
(?P<nfunc>[a-zA-Z_$][\w$]*)(?:\s*\[(?P<idx>\d+)\])?\s*\(\s*[\w$]+\s*\) | # ,c=a.get(b)
,\s*(?P<c>[a-z])\s*=\s*[a-z]\s*
(?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*?
(?:\[\s*(?P=b)\s*]|\.\s*get\s*\(\s*(?P=b)\s*\))
)
# interstitial junk
\s*(?:\|\|\s*null\s*)?(?:\)\s*)?&&\s*(?:\(\s*)?
(?(c)(?P=c)|(?P=b))\s*=\s* # [c|b]=
# nfunc|nfunc[idx]
(?P<nfunc>[a-zA-Z_$][\w$]*)(?:\s*\[(?P<idx>\d+)\])?\s*\(\s*[\w$]+\s*\)
''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'),
default=(None, None)) default=(None, None))
# thx bashonly: yt-dlp/yt-dlp/pull/10611 # thx bashonly: yt-dlp/yt-dlp/pull/10611
@ -1690,7 +1700,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
\s*\{(?:(?!};).)+?["']enhanced_except_ \s*\{(?:(?!};).)+?["']enhanced_except_
''', jscode, 'Initial JS player n function name', group='name') ''', jscode, 'Initial JS player n function name', group='name')
if not idx: if not idx:
self.report_warning('Falling back to generic n function search')
return func_name return func_name
return self._parse_json(self._search_regex( return self._parse_json(self._search_regex(