nhyatt/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2025-04-12 01:48:07 -05:00
Code Issues Projects Releases Wiki Activity

[YouTube] Update signature extraction for players 643afba4, 363db69b

Browse Source
This commit is contained in:
dirkf 2025-03-25 22:16:19 +00:00
parent 087d865230
commit c55dbf4838
2 changed files with 25 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
youtube_dl/extractor/youtube.py
View File

@ -1709,6 +1709,18 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
' return %s\n') % (signature_id_tuple, expr_code) ' return %s\n') % (signature_id_tuple, expr_code)
self.to_screen('Extracted signature function:\n' + code) self.to_screen('Extracted signature function:\n' + code)
def _extract_sig_fn(self, jsi, funcname):
var_ay = self._search_regex(
r'''(?:\*/|\{|\n|^)\s*(?:'[^']+'\s*;\s*)(var\s*[\w$]+\s*=\s*('|")(?:\\\2|(?!\2).)+\2\s*\.\s*split\(('|")\W+\3\))(?=\s*[,;])''',
jsi.code, 'useful values', default='')
sig_fn = jsi.extract_function_code(funcname)
if var_ay:
sig_fn = (sig_fn[0], ';\n'.join((var_ay, sig_fn[1])))
return sig_fn
def _parse_sig_js(self, jscode): def _parse_sig_js(self, jscode):
# Examples where `sig` is funcname: # Examples where `sig` is funcname:
# sig=function(a){a=a.split(""); ... ;return a.join("")}; # sig=function(a){a=a.split(""); ... ;return a.join("")};
@ -1734,8 +1746,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
jscode, 'Initial JS player signature function name', group='sig') jscode, 'Initial JS player signature function name', group='sig')
jsi = JSInterpreter(jscode) jsi = JSInterpreter(jscode)
initial_function = jsi.extract_function(funcname)
return lambda s: initial_function([s]) initial_function = self._extract_sig_fn(jsi, funcname)
func = jsi.extract_function_from_code(*initial_function)
return lambda s: func([s])
def _cached(self, func, *cache_id): def _cached(self, func, *cache_id):
def inner(*args, **kwargs): def inner(*args, **kwargs):
@ -1854,15 +1870,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None): def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None):
var_ay = self._search_regex(
r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"(?:\\"|[^"])+"\s*\.\s*split\("\W+"\))(?=\s*[,;])',
jsi.code, 'useful values', default='')
func_name = self._extract_n_function_name(jsi.code) func_name = self._extract_n_function_name(jsi.code)
func_code = jsi.extract_function_code(func_name) func_code = self._extract_sig_fn(jsi, func_name)
if var_ay:
func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
if player_id: if player_id:
self.cache.store('youtube-nsig', player_id, func_code) self.cache.store('youtube-nsig', player_id, func_code)

10
youtube_dl/jsinterp.py
View File

@ -1368,19 +1368,21 @@ class JSInterpreter(object):
code, _ = self._separate_at_paren(func_m.group('code')) # refine the match code, _ = self._separate_at_paren(func_m.group('code')) # refine the match
return self.build_arglist(func_m.group('args')), code return self.build_arglist(func_m.group('args')), code
def extract_function(self, funcname): def extract_function(self, funcname, *global_stack):
return function_with_repr( return function_with_repr(
self.extract_function_from_code(*self.extract_function_code(funcname)), self.extract_function_from_code(*itertools.chain(
self.extract_function_code(funcname), global_stack)),
'F<%s>' % (funcname,)) 'F<%s>' % (funcname,))
def extract_function_from_code(self, argnames, code, *global_stack): def extract_function_from_code(self, argnames, code, *global_stack):
local_vars = {} local_vars = {}
start = None
while True: while True:
mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code) mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code[start:])
if mobj is None: if mobj is None:
break break
start, body_start = mobj.span() start, body_start = ((start or 0) + x for x in mobj.span())
body, remaining = self._separate_at_paren(code[body_start - 1:]) body, remaining = self._separate_at_paren(code[body_start - 1:])
name = self._named_object(local_vars, self.extract_function_from_code( name = self._named_object(local_vars, self.extract_function_from_code(
[x.strip() for x in mobj.group('args').split(',')], [x.strip() for x in mobj.group('args').split(',')],