From 105b5ba58e113c54c030913159af05fc0553691d Mon Sep 17 00:00:00 2001 From: nicole trinity Date: Thu, 1 Aug 2024 11:56:05 -0400 Subject: [PATCH] applied @dirkf's patches for nsig extraction on youtube --- test/test_youtube_signature.py | 4 ++++ youtube_dl/extractor/youtube.py | 19 ++++++++++++++----- 2 files changed, 18 insertions(+), 5 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 5b4aa3aa0..5bb89710c 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -174,6 +174,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/5604538d/player_ias.vflset/en_US/base.js', '7X-he4jjvMx7BCX', 'sViSydX8IHtdWA', ), + ( + 'https://www.youtube.com/s/player/20dfca59/player_ias.vflset/en_US/base.js', + '-fLCxedkAk4LUTK2', 'O8kfRq1y1eyHGw', + ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 84371ff06..390294966 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1659,15 +1659,24 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_n_function_name(self, jscode): func_name, idx = self._search_regex( # 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)s + # 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) # old: .get("n"))&&(b=nfunc[idx](b) # older: .get("n"))&&(b=nfunc(b) r'''(?x) - (?:\(\s*(?P[a-z])\s*=\s*(?: + (?:\((?:[\w$()\s]+,)*?\s*(?P[a-z])\s*=\s*(?: String\s*\.\s*fromCharCode\s*\(\s*110\s*\)| - "n+"\[\s*\+?s*[\w$.]+\s*] - )\s*,(?P[a-z])\s*=\s*[a-z]\s*)? - \.\s*get\s*\(\s*(?(b)(?P=b)|"n{1,2}")(?:\s*\)){2}\s*&&\s*\(\s*(?(c)(?P=c)|b)\s*=\s* + "n+"\[\s*\+?s*[\w$.]+\s*]| + (?P(?:[\w$]+\s*\.\s*)+n\b(?:(?!&&).)+\)) + )\s* + (?(b1) + &&\s*\(\s*(?P=b)| + (?: + ,(?P[a-z])\s*=\s*[a-z]\s*)? + \.\s*get\s*\(\s*(?(b)(?P=b)|"n{1,2}")(?:\s*\)){2}\s* + &&\s*\(\s*(?(c)(?P=c)|(?P=b)) + ) + )\s*=\s* (?P[a-zA-Z_$][\w$]*)(?:\s*\[(?P\d+)\])?\s*\(\s*[\w$]+\s*\) ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx')) if not idx: