From 451046d62a1b57cc473625c728ff89d7a66933c3 Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 24 Jul 2024 14:33:34 +0100 Subject: [PATCH 01/78] [Youtube] Make n-sig throttling diagnostic up-to-date --- youtube_dl/extractor/youtube.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 2e31a8979..bb9f48e4c 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1647,7 +1647,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): except JSInterpreter.Exception as e: self.report_warning( '%s (%s %s)' % ( - 'Unable to decode n-parameter: download likely to be throttled', + 'Unable to decode n-parameter: expect download to be blocked or throttled', error_to_compat_str(e), traceback.format_exc()), video_id=video_id) @@ -1670,7 +1670,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return func_name return self._parse_json(self._search_regex( - r'var {0}\s*=\s*(\[.+?\])\s*[,;]'.format(re.escape(func_name)), jscode, + r'var\s+{0}\s*=\s*(\[.+?\])\s*[,;]'.format(re.escape(func_name)), jscode, 'Initial JS player n function list ({0}.{1})'.format(func_name, idx)), func_name, transform_source=js_to_json)[int(idx)] From e1b3fa242cb94eb9dc949ab3f2cace91c46f11bf Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 24 Jul 2024 14:45:52 +0100 Subject: [PATCH 02/78] [Youtube] Find `n` function name in player `3400486c` Fixes #32877 --- test/test_youtube_signature.py | 8 ++++++++ youtube_dl/extractor/youtube.py | 8 ++++++-- 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index cc18d0f7b..5b4aa3aa0 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -166,6 +166,14 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/b22ef6e7/player_ias.vflset/en_US/base.js', 'b6HcntHGkvBLk_FRf', 'kNPW6A7FyP2l8A', ), + ( + 'https://www.youtube.com/s/player/3400486c/player_ias.vflset/en_US/base.js', + 'lL46g3XifCKUZn1Xfw', 'z767lhet6V2Skl', + ), + ( + 'https://www.youtube.com/s/player/5604538d/player_ias.vflset/en_US/base.js', + '7X-he4jjvMx7BCX', 'sViSydX8IHtdWA', + ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index bb9f48e4c..84371ff06 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1659,11 +1659,15 @@ 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 # old: .get("n"))&&(b=nfunc[idx](b) # older: .get("n"))&&(b=nfunc(b) r'''(?x) - (?:\(\s*(?P[a-z])\s*=\s*String\s*\.\s*fromCharCode\s*\(\s*110\s*\)\s*,(?P[a-z])\s*=\s*[a-z]\s*)? - \.\s*get\s*\(\s*(?(b)(?P=b)|"n")(?:\s*\)){2}\s*&&\s*\(\s*(?(c)(?P=c)|b)\s*=\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* (?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: From 71223bff39551a11b6959a3de2dd9e2f070f3c4f Mon Sep 17 00:00:00 2001 From: Aiur Adept <151766879+aiur-adept@users.noreply.github.com> Date: Thu, 1 Aug 2024 14:18:34 -0400 Subject: [PATCH 03/78] [Youtube] Fix nsig extraction for player 20dfca59 (#32891) * dirkf's patch for nsig extraction * add generic search per yt-dlp/yt-dlp/pull/10611 - thx bashonly --------- Co-authored-by: dirkf --- test/test_youtube_signature.py | 4 ++++ youtube_dl/extractor/youtube.py | 32 ++++++++++++++++++++++++++------ 2 files changed, 30 insertions(+), 6 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 5b4aa3aa0..1c5f667f5 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..509e374a4 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1659,18 +1659,38 @@ 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')) + ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), + default=(None, None)) + # thx bashonly: yt-dlp/yt-dlp/pull/10611 + if not func_name: + self.report_warning('Falling back to generic n function search') + return self._search_regex( + r'''(?xs) + (?:(?<=[^\w$])|^) # instead of \b, which ignores $ + (?P(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) + \s*\{(?:(?!};).)+?["']enhanced_except_ + ''', jscode, 'Initial JS player n function name', group='name') if not idx: + self.report_warning('Falling back to generic n function search') return func_name return self._parse_json(self._search_regex( From dbc08fba83a0e6c5215e8cec97eb4076f85eee5f Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 6 Aug 2024 20:44:30 +0100 Subject: [PATCH 04/78] [jsinterp] Improve slice implementation for player b12cc44b Partly taken from yt-dlp/yt-dlp#10664, thx seproDev Fixes #32896 --- test/test_jsinterp.py | 28 ++++++++++++++++++++++++++++ test/test_youtube_signature.py | 4 ++++ youtube_dl/jsinterp.py | 13 ++++++++++--- 3 files changed, 42 insertions(+), 3 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 104e766be..c7a4f2cbf 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -425,6 +425,34 @@ class TestJSInterpreter(unittest.TestCase): self._test(jsi, [''], args=['', '-']) self._test(jsi, [], args=['', '']) + def test_slice(self): + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice()}', [0, 1, 2, 3, 4, 5, 6, 7, 8]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(0)}', [0, 1, 2, 3, 4, 5, 6, 7, 8]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(5)}', [5, 6, 7, 8]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(99)}', []) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-2)}', [7, 8]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-99)}', [0, 1, 2, 3, 4, 5, 6, 7, 8]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(0, 0)}', []) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(1, 0)}', []) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(0, 1)}', [0]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(3, 6)}', [3, 4, 5]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(1, -1)}', [1, 2, 3, 4, 5, 6, 7]) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-1, 1)}', []) + self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-3, -1)}', [6, 7]) + self._test('function f(){return "012345678".slice()}', '012345678') + self._test('function f(){return "012345678".slice(0)}', '012345678') + self._test('function f(){return "012345678".slice(5)}', '5678') + self._test('function f(){return "012345678".slice(99)}', '') + self._test('function f(){return "012345678".slice(-2)}', '78') + self._test('function f(){return "012345678".slice(-99)}', '012345678') + self._test('function f(){return "012345678".slice(0, 0)}', '') + self._test('function f(){return "012345678".slice(1, 0)}', '') + self._test('function f(){return "012345678".slice(0, 1)}', '0') + self._test('function f(){return "012345678".slice(3, 6)}', '345') + self._test('function f(){return "012345678".slice(1, -1)}', '1234567') + self._test('function f(){return "012345678".slice(-1, 1)}', '') + self._test('function f(){return "012345678".slice(-3, -1)}', '67') + if __name__ == '__main__': unittest.main() diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 1c5f667f5..56e92fac5 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -178,6 +178,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/20dfca59/player_ias.vflset/en_US/base.js', '-fLCxedkAk4LUTK2', 'O8kfRq1y1eyHGw', ), + ( + 'https://www.youtube.com/s/player/b12cc44b/player_ias.vflset/en_US/base.js', + 'keLa5R2U00sR9SQK', 'N1OGyujjEwMnLw', + ), ] diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 949f77775..a616ad070 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -925,9 +925,16 @@ class JSInterpreter(object): obj.reverse() return obj elif member == 'slice': - assertion(isinstance(obj, list), 'must be applied on a list') - assertion(len(argvals) == 1, 'takes exactly one argument') - return obj[argvals[0]:] + assertion(isinstance(obj, (list, compat_str)), 'must be applied on a list or string') + # From [1]: + # .slice() - like [:] + # .slice(n) - like [n:] (not [slice(n)] + # .slice(m, n) - like [m:n] or [slice(m, n)] + # [1] https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/slice + assertion(len(argvals) <= 2, 'takes between 0 and 2 arguments') + if len(argvals) < 2: + argvals += (None,) + return obj[slice(*argvals)] elif member == 'splice': assertion(isinstance(obj, list), 'must be applied on a list') assertion(argvals, 'takes one or more arguments') From c5098961b04ce83f4615f2a846c84f803b072639 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 6 Aug 2024 20:59:09 +0100 Subject: [PATCH 05/78] [Youtube] Rework n function extraction pattern Now also succeeds with player b12cc44b --- youtube_dl/extractor/youtube.py | 43 ++++++++++++++++++++------------- 1 file changed, 26 insertions(+), 17 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 509e374a4..6fe520e9a 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1661,23 +1661,33 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # 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: (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b) - # old: .get("n"))&&(b=nfunc[idx](b) - # older: .get("n"))&&(b=nfunc(b) + # or: (b="nn"[+a.D],vL(a),c=a.j[b]||null)&&(c=narray[idx](c),a.set(b,c),narray.length||nfunc("") + # old: (b=a.get("n"))&&(b=nfunc[idx](b)(?P[a-z])\s*=\s*[a-z]\s* + # older: (b=a.get("n"))&&(b=nfunc(b) r'''(?x) - (?:\((?:[\w$()\s]+,)*?\s*(?P[a-z])\s*=\s*(?: - String\s*\.\s*fromCharCode\s*\(\s*110\s*\)| - "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*\(\s*[\w$]+\s*\) + \((?:[\w$()\s]+,)*?\s* # ( + (?P[a-z])\s*=\s* # b= + (?: + (?: # expect ,c=a.get(b) (etc) + String\s*\.\s*fromCharCode\s*\(\s*110\s*\)| + "n+"\[\s*\+?s*[\w$.]+\s*] + )\s*(?:,[\w$()\s]+(?=,))*| + (?P[\w$]+) # a (old[er]) + )\s* + (?(old) + # b.get("n") + (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*? + (?:\.\s*n|\[\s*"n"\s*]|\.\s*get\s*\(\s*"n"\s*\)) + | # ,c=a.get(b) + ,\s*(?P[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[a-zA-Z_$][\w$]*)(?:\s*\[(?P\d+)\])?\s*\(\s*[\w$]+\s*\) ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), default=(None, None)) # thx bashonly: yt-dlp/yt-dlp/pull/10611 @@ -1690,7 +1700,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor): \s*\{(?:(?!};).)+?["']enhanced_except_ ''', jscode, 'Initial JS player n function name', group='name') if not idx: - self.report_warning('Falling back to generic n function search') return func_name return self._parse_json(self._search_regex( From f28d7178e4afb58a94994796cd81c22183ca33c2 Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 04:46:33 +0000 Subject: [PATCH 06/78] [InfoExtractor] Use kwarg maxsplit for re.split * May become kw-only in future Pythons --- youtube_dl/extractor/common.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 9b0016d07..78704b557 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -3170,7 +3170,7 @@ class InfoExtractor(object): # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as # of jwplayer.flash.swf rtmp_url_parts = re.split( - r'((?:mp4|mp3|flv):)', source_url, 1) + r'((?:mp4|mp3|flv):)', source_url, maxsplit=1) if len(rtmp_url_parts) == 3: rtmp_url, prefix, play_path = rtmp_url_parts a_format.update({ From 118c6d7a17c61f1a1031e5530ca2854a526834d5 Mon Sep 17 00:00:00 2001 From: dirkf Date: Sat, 7 Dec 2024 03:37:39 +0000 Subject: [PATCH 07/78] [jsinterp] Implement `typeof` operator --- test/test_jsinterp.py | 15 +++- youtube_dl/jsinterp.py | 152 ++++++++++++++++++++++++++++------------- 2 files changed, 117 insertions(+), 50 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index c7a4f2cbf..d063bbd36 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -266,7 +266,20 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f() { return (l=[0,1,2,3], function(a, b){return a+b})((l[1], l[2]), l[3]) }', 5) def test_void(self): - self._test('function f() { return void 42; }', None) + self._test('function f() { return void 42; }', JS_Undefined) + + def test_typeof(self): + self._test('function f() { return typeof undefined; }', 'undefined') + self._test('function f() { return typeof NaN; }', 'number') + self._test('function f() { return typeof Infinity; }', 'number') + self._test('function f() { return typeof true; }', 'boolean') + self._test('function f() { return typeof null; }', 'object') + self._test('function f() { return typeof "a string"; }', 'string') + self._test('function f() { return typeof 42; }', 'number') + self._test('function f() { return typeof 42.42; }', 'number') + self._test('function f() { var g = function(){}; return typeof g; }', 'function') + self._test('function f() { return typeof {key: "value"}; }', 'object') + # not yet implemented: Symbol, BigInt def test_return_function(self): jsi = JSInterpreter(''' diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index a616ad070..acc68c533 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -5,7 +5,7 @@ import json import operator import re -from functools import update_wrapper +from functools import update_wrapper, wraps from .utils import ( error_to_compat_str, @@ -20,9 +20,11 @@ from .compat import ( compat_basestring, compat_chr, compat_collections_chain_map as ChainMap, + compat_contextlib_suppress, compat_filter as filter, compat_itertools_zip_longest as zip_longest, compat_map as map, + compat_numeric_types, compat_str, ) @@ -138,6 +140,38 @@ def _js_ternary(cndn, if_true=True, if_false=False): return if_true +def _js_unary_op(op): + + @wraps_op(op) + def wrapped(_, a): + return op(a) + + return wrapped + + +# https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/typeof +def _js_typeof(expr): + with compat_contextlib_suppress(TypeError, KeyError): + return { + JS_Undefined: 'undefined', + _NaN: 'number', + _Infinity: 'number', + True: 'boolean', + False: 'boolean', + None: 'object', + }[expr] + for t, n in ( + (compat_basestring, 'string'), + (compat_numeric_types, 'number'), + ): + if isinstance(expr, t): + return n + if callable(expr): + return 'function' + # TODO: Symbol, BigInt + return 'object' + + # (op, definition) in order of binding priority, tightest first # avoid dict to maintain order # definition None => Defined in JSInterpreter._operator @@ -176,6 +210,11 @@ _SC_OPERATORS = ( ('&&', None), ) +_UNARY_OPERATORS_X = ( + ('void', _js_unary_op(lambda _: JS_Undefined)), + ('typeof', _js_unary_op(_js_typeof)), +) + _OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS)) _NAME_RE = r'[a-zA-Z_$][\w$]*' @@ -242,6 +281,7 @@ class Debugger(object): @classmethod def wrap_interpreter(cls, f): + @wraps(f) def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs): if cls.ENABLED and stmt.strip(): cls.write(stmt, level=allow_recursion) @@ -347,6 +387,8 @@ class JSInterpreter(object): def __op_chars(cls): op_chars = set(';,[') for op in cls._all_operators(): + if op[0].isalpha(): + continue op_chars.update(op[0]) return op_chars @@ -425,7 +467,7 @@ class JSInterpreter(object): if not _cached: _cached.extend(itertools.chain( # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence - _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)) + _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS, _UNARY_OPERATORS_X)) return _cached def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): @@ -479,6 +521,52 @@ class JSInterpreter(object): _FINALLY_RE = re.compile(r'finally\s*\{') _SWITCH_RE = re.compile(r'switch\s*\(') + def handle_operators(self, expr, local_vars, allow_recursion): + + for op, _ in self._all_operators(): + # hackety: have higher priority than <>, but don't confuse them + skip_delim = (op + op) if op in '<>*?' else None + if op == '?': + skip_delim = (skip_delim, '?.') + separated = list(self._separate(expr, op, skip_delims=skip_delim)) + if len(separated) < 2: + continue + + right_expr = separated.pop() + # handle operators that are both unary and binary, minimal BODMAS + if op in ('+', '-'): + # simplify/adjust consecutive instances of these operators + undone = 0 + separated = [s.strip() for s in separated] + while len(separated) > 1 and not separated[-1]: + undone += 1 + separated.pop() + if op == '-' and undone % 2 != 0: + right_expr = op + right_expr + elif op == '+': + while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: + right_expr = separated.pop() + right_expr + if separated[-1][-1:] in self.OP_CHARS: + right_expr = separated.pop() + right_expr + # hanging op at end of left => unary + (strip) or - (push right) + left_val = separated[-1] if separated else '' + for dm_op in ('*', '%', '/', '**'): + bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim)) + if len(bodmas) > 1 and not bodmas[-1].strip(): + expr = op.join(separated) + op + right_expr + if len(separated) > 1: + separated.pop() + right_expr = op.join((left_val, right_expr)) + else: + separated = [op.join((left_val, right_expr))] + right_expr = None + break + if right_expr is None: + continue + + left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) + return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), True + @Debugger.wrap_interpreter def interpret_statement(self, stmt, local_vars, allow_recursion=100): if allow_recursion < 0: @@ -533,9 +621,15 @@ class JSInterpreter(object): else: raise self.Exception('Unsupported object {obj:.100}'.format(**locals()), expr=expr) - if expr.startswith('void '): - left = self.interpret_expression(expr[5:], local_vars, allow_recursion) - return None, should_return + for op, _ in _UNARY_OPERATORS_X: + if not expr.startswith(op): + continue + operand = expr[len(op):] + if not operand or operand[0] != ' ': + continue + op_result = self.handle_operators(expr, local_vars, allow_recursion) + if op_result: + return op_result[0], should_return if expr.startswith('{'): inner, outer = self._separate_at_paren(expr) @@ -582,7 +676,7 @@ class JSInterpreter(object): if_expr, expr = self._separate_at_paren(expr) else: # may lose ... else ... because of ll.368-374 - if_expr, expr = self._separate_at_paren(expr, delim=';') + if_expr, expr = self._separate_at_paren(' %s;' % (expr,), delim=';') else_expr = None m = re.match(r'else\s*(?P\{)?', expr) if m: @@ -790,49 +884,9 @@ class JSInterpreter(object): idx = self.interpret_expression(m.group('idx'), local_vars, allow_recursion) return self._index(val, idx), should_return - for op, _ in self._all_operators(): - # hackety: have higher priority than <>, but don't confuse them - skip_delim = (op + op) if op in '<>*?' else None - if op == '?': - skip_delim = (skip_delim, '?.') - separated = list(self._separate(expr, op, skip_delims=skip_delim)) - if len(separated) < 2: - continue - - right_expr = separated.pop() - # handle operators that are both unary and binary, minimal BODMAS - if op in ('+', '-'): - # simplify/adjust consecutive instances of these operators - undone = 0 - separated = [s.strip() for s in separated] - while len(separated) > 1 and not separated[-1]: - undone += 1 - separated.pop() - if op == '-' and undone % 2 != 0: - right_expr = op + right_expr - elif op == '+': - while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: - right_expr = separated.pop() + right_expr - if separated[-1][-1:] in self.OP_CHARS: - right_expr = separated.pop() + right_expr - # hanging op at end of left => unary + (strip) or - (push right) - left_val = separated[-1] if separated else '' - for dm_op in ('*', '%', '/', '**'): - bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim)) - if len(bodmas) > 1 and not bodmas[-1].strip(): - expr = op.join(separated) + op + right_expr - if len(separated) > 1: - separated.pop() - right_expr = op.join((left_val, right_expr)) - else: - separated = [op.join((left_val, right_expr))] - right_expr = None - break - if right_expr is None: - continue - - left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) - return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), should_return + op_result = self.handle_operators(expr, local_vars, allow_recursion) + if op_result: + return op_result[0], should_return if md.get('attribute'): variable, member, nullish = m.group('var', 'member', 'nullish') From c1a03b1ac3a453508b358b6d1bc5d158cc80a0ce Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 00:00:32 +0000 Subject: [PATCH 08/78] [jsinterp] Fix and improve loose and strict equality operations * reimplement loose equality according to MDN (eg, 1 == "1") * improve strict equality (eg, "abc" === "abc" but 'abc' is not 'abc') * add tests for above --- test/test_jsinterp.py | 35 ++++++++++++++++-- youtube_dl/jsinterp.py | 80 ++++++++++++++++++++++++++++++++++++------ 2 files changed, 101 insertions(+), 14 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index d063bbd36..86137d8e5 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -1,4 +1,5 @@ #!/usr/bin/env python +# coding: utf-8 from __future__ import unicode_literals @@ -11,7 +12,7 @@ sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) import math import re -from youtube_dl.compat import compat_str +from youtube_dl.compat import compat_str as str from youtube_dl.jsinterp import JS_Undefined, JSInterpreter NaN = object() @@ -19,7 +20,7 @@ NaN = object() class TestJSInterpreter(unittest.TestCase): def _test(self, jsi_or_code, expected, func='f', args=()): - if isinstance(jsi_or_code, compat_str): + if isinstance(jsi_or_code, str): jsi_or_code = JSInterpreter(jsi_or_code) got = jsi_or_code.call_function(func, *args) if expected is NaN: @@ -89,7 +90,35 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return 19 & 21;}', 17) self._test('function f(){return 11 >> 2;}', 2) self._test('function f(){return []? 2+3: 4;}', 5) + # equality + self._test('function f(){return 1 == 1}', True) + self._test('function f(){return 1 == 1.0}', True) + self._test('function f(){return 1 == "1"}', True) self._test('function f(){return 1 == 2}', False) + self._test('function f(){return 1 != "1"}', False) + self._test('function f(){return 1 != 2}', True) + self._test('function f(){var x = {a: 1}; var y = x; return x == y}', True) + self._test('function f(){var x = {a: 1}; return x == {a: 1}}', False) + self._test('function f(){return NaN == NaN}', False) + self._test('function f(){return null == undefined}', True) + self._test('function f(){return "spam, eggs" == "spam, eggs"}', True) + # strict equality + self._test('function f(){return 1 === 1}', True) + self._test('function f(){return 1 === 1.0}', True) + self._test('function f(){return 1 === "1"}', False) + self._test('function f(){return 1 === 2}', False) + self._test('function f(){var x = {a: 1}; var y = x; return x === y}', True) + self._test('function f(){var x = {a: 1}; return x === {a: 1}}', False) + self._test('function f(){return NaN === NaN}', False) + self._test('function f(){return null === undefined}', False) + self._test('function f(){return null === null}', True) + self._test('function f(){return undefined === undefined}', True) + self._test('function f(){return "uninterned" === "uninterned"}', True) + self._test('function f(){return 1 === 1}', True) + self._test('function f(){return 1 === "1"}', False) + self._test('function f(){return 1 !== 1}', False) + self._test('function f(){return 1 !== "1"}', True) + # expressions self._test('function f(){return 0 && 1 || 2;}', 2) self._test('function f(){return 0 ?? 42;}', 0) self._test('function f(){return "life, the universe and everything" < 42;}', False) @@ -296,7 +325,7 @@ class TestJSInterpreter(unittest.TestCase): def test_undefined(self): self._test('function f() { return undefined === undefined; }', True) self._test('function f() { return undefined; }', JS_Undefined) - self._test('function f() {return undefined ?? 42; }', 42) + self._test('function f() { return undefined ?? 42; }', 42) self._test('function f() { let v; return v; }', JS_Undefined) self._test('function f() { let v; return v**0; }', 1) self._test('function f() { let v; return [v>42, v<=42, v&&42, 42&&v]; }', diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index acc68c533..b7a3ef69a 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -1,3 +1,4 @@ +# coding: utf-8 from __future__ import unicode_literals import itertools @@ -64,6 +65,10 @@ _NaN = float('nan') _Infinity = float('inf') +class JS_Undefined(object): + pass + + def _js_bit_op(op): def zeroise(x): @@ -107,12 +112,69 @@ def _js_exp(a, b): return (a or 0) ** b -def _js_eq_op(op): +def _js_to_primitive(v): + return ( + ','.join(map(_js_toString, v)) if isinstance(v, list) + else '[object Object]' if isinstance(v, dict) + else compat_str(v) if not isinstance(v, ( + compat_numeric_types, compat_basestring, bool)) + else v + ) + + +def _js_toString(v): + return ( + 'undefined' if v is JS_Undefined + else 'Infinity' if v == _Infinity + else 'NaN' if v is _NaN + else 'null' if v is None + else compat_str(v) if isinstance(v, compat_numeric_types) + else _js_to_primitive(v)) + + +_nullish = frozenset((None, JS_Undefined)) + + +def _js_eq(a, b): + # NaN != any + if _NaN in (a, b): + return False + # Object is Object + if isinstance(a, type(b)) and isinstance(b, (dict, list)): + return operator.is_(a, b) + # general case + if a == b: + return True + # null == undefined + a_b = set((a, b)) + if a_b & _nullish: + return a_b <= _nullish + a, b = _js_to_primitive(a), _js_to_primitive(b) + if not isinstance(a, compat_basestring): + a, b = b, a + # Number to String: convert the string to a number + # Conversion failure results in ... false + if isinstance(a, compat_basestring): + return float_or_none(a) == b + return a == b + + +def _js_neq(a, b): + return not _js_eq(a, b) + + +def _js_id_op(op): @wraps_op(op) def wrapped(a, b): - if set((a, b)) <= set((None, JS_Undefined)): - return op(a, a) + if _NaN in (a, b): + return op(_NaN, None) + if not isinstance(a, (compat_basestring, compat_numeric_types)): + a, b = b, a + # strings are === if == + # why 'a' is not 'a': https://stackoverflow.com/a/1504848 + if isinstance(a, (compat_basestring, compat_numeric_types)): + return a == b if op(0, 0) else a != b return op(a, b) return wrapped @@ -187,10 +249,10 @@ _OPERATORS = ( ) _COMP_OPERATORS = ( - ('===', operator.is_), - ('!==', operator.is_not), - ('==', _js_eq_op(operator.eq)), - ('!=', _js_eq_op(operator.ne)), + ('===', _js_id_op(operator.is_)), + ('!==', _js_id_op(operator.is_not)), + ('==', _js_eq), + ('!=', _js_neq), ('<=', _js_comp_op(operator.le)), ('>=', _js_comp_op(operator.ge)), ('<', _js_comp_op(operator.lt)), @@ -222,10 +284,6 @@ _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]'))) _QUOTES = '\'"/' -class JS_Undefined(object): - pass - - class JS_Break(ExtractorError): def __init__(self): ExtractorError.__init__(self, 'Invalid break') From 81e64cacf2b8c144ec1e9d3258db792f0eb8443e Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 00:15:31 +0000 Subject: [PATCH 09/78] [jsinterp] Support multiple indexing (eg a[1][2]) * extend single indexing with improved RE (should probably use/have used _separate_at_paren()) * fix some cases that should have given undefined, not throwing * standardise RE group names * support length of objects, like {1: 2, 3: 4, length: 42} --- test/test_jsinterp.py | 10 ++++++++++ youtube_dl/jsinterp.py | 37 ++++++++++++++++++++++--------------- 2 files changed, 32 insertions(+), 15 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 86137d8e5..d2fca2997 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -366,6 +366,16 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f() { let a; return a?.qq; }', JS_Undefined) self._test('function f() { let a = {m1: 42, m2: 0 }; return a?.qq; }', JS_Undefined) + def test_indexing(self): + self._test('function f() { return [1, 2, 3, 4][3]}', 4) + self._test('function f() { return [1, [2, [3, [4]]]][1][1][1][0]}', 4) + self._test('function f() { var o = {1: 2, 3: 4}; return o[3]}', 4) + self._test('function f() { var o = {1: 2, 3: 4}; return o["3"]}', 4) + self._test('function f() { return [1, [2, {3: [4]}]][1][1]["3"][0]}', 4) + self._test('function f() { return [1, 2, 3, 4].length}', 4) + self._test('function f() { var o = {1: 2, 3: 4}; return o.length}', JS_Undefined) + self._test('function f() { var o = {1: 2, 3: 4}; o["length"] = 42; return o.length}', 42) + def test_regex(self): self._test('function f() { let a=/,,[/,913,/](,)}/; }', None) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index b7a3ef69a..248210ce4 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -549,13 +549,14 @@ class JSInterpreter(object): except Exception as e: raise self.Exception('Failed to evaluate {left_val!r:.50} {op} {right_val!r:.50}'.format(**locals()), expr, cause=e) - def _index(self, obj, idx, allow_undefined=False): - if idx == 'length': + def _index(self, obj, idx, allow_undefined=True): + if idx == 'length' and isinstance(obj, list): return len(obj) try: - return obj[int(idx)] if isinstance(obj, list) else obj[idx] - except Exception as e: + return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)] + except (TypeError, KeyError, IndexError) as e: if allow_undefined: + # when is not allowed? return JS_Undefined raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e) @@ -882,13 +883,13 @@ class JSInterpreter(object): m = re.match(r'''(?x) (?P - (?P{_NAME_RE})(?:\[(?P[^\]]+?)\])?\s* + (?P{_NAME_RE})(?:\[(?P(?:.+?\]\s*\[)*.+?)\])?\s* (?P{_OPERATOR_RE})? =(?!=)(?P.*)$ )|(?P (?!if|return|true|false|null|undefined|NaN|Infinity)(?P{_NAME_RE})$ )|(?P - (?P{_NAME_RE})\[(?P.+)\]$ + (?P{_NAME_RE})\[(?P(?:.+?\]\s*\[)*.+?)\]$ )|(?P (?P{_NAME_RE})(?:(?P\?)?\.(?P[^(]+)|\[(?P[^\]]+)\])\s* )|(?P @@ -898,19 +899,23 @@ class JSInterpreter(object): if md.get('assign'): left_val = local_vars.get(m.group('out')) - if not m.group('index'): + if not m.group('out_idx'): local_vars[m.group('out')] = self._operator( m.group('op'), left_val, m.group('expr'), expr, local_vars, allow_recursion) return local_vars[m.group('out')], should_return elif left_val in (None, JS_Undefined): raise self.Exception('Cannot index undefined variable ' + m.group('out'), expr=expr) - idx = self.interpret_expression(m.group('index'), local_vars, allow_recursion) - if not isinstance(idx, (int, float)): - raise self.Exception('List index %s must be integer' % (idx, ), expr=expr) - idx = int(idx) + indexes = re.split(r'\]\s*\[', m.group('out_idx')) + for i, idx in enumerate(indexes, 1): + idx = self.interpret_expression(idx, local_vars, allow_recursion) + if i < len(indexes): + left_val = self._index(left_val, idx) + if isinstance(idx, float): + idx = int(idx) left_val[idx] = self._operator( - m.group('op'), self._index(left_val, idx), m.group('expr'), expr, local_vars, allow_recursion) + m.group('op'), self._index(left_val, idx) if m.group('op') else None, + m.group('expr'), expr, local_vars, allow_recursion) return left_val[idx], should_return elif expr.isdigit(): @@ -939,8 +944,10 @@ class JSInterpreter(object): if md.get('indexing'): val = local_vars[m.group('in')] - idx = self.interpret_expression(m.group('idx'), local_vars, allow_recursion) - return self._index(val, idx), should_return + for idx in re.split(r'\]\s*\[', m.group('in_idx')): + idx = self.interpret_expression(idx, local_vars, allow_recursion) + val = self._index(val, idx) + return val, should_return op_result = self.handle_operators(expr, local_vars, allow_recursion) if op_result: @@ -989,7 +996,7 @@ class JSInterpreter(object): # Member access if arg_str is None: - return self._index(obj, member, nullish) + return self._index(obj, member) # Function call argvals = [ From 5dee6213edddb5ea00775db6b3b73f8355144485 Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 00:26:00 +0000 Subject: [PATCH 10/78] [jsinterp] Fix and improve arithmetic operations * addition becomes concat with a string operand * improve handling of edgier cases * arithmetic in float like JS (more places need cast to int?) * increase test coverage --- test/test_jsinterp.py | 20 ++++++++++++++++ youtube_dl/jsinterp.py | 53 +++++++++++++++++++++++++++--------------- 2 files changed, 54 insertions(+), 19 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index d2fca2997..52eaf1ed8 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -41,16 +41,27 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return 42 + 7;}', 49) self._test('function f(){return 42 + undefined;}', NaN) self._test('function f(){return 42 + null;}', 42) + self._test('function f(){return 1 + "";}', '1') + self._test('function f(){return 42 + "7";}', '427') + self._test('function f(){return false + true;}', 1) + self._test('function f(){return "false" + true;}', 'falsetrue') + self._test('function f(){return ' + '1 + "2" + [3,4] + {k: 56} + null + undefined + Infinity;}', + '123,4[object Object]nullundefinedInfinity') def test_sub(self): self._test('function f(){return 42 - 7;}', 35) self._test('function f(){return 42 - undefined;}', NaN) self._test('function f(){return 42 - null;}', 42) + self._test('function f(){return 42 - "7";}', 35) + self._test('function f(){return 42 - "spam";}', NaN) def test_mul(self): self._test('function f(){return 42 * 7;}', 294) self._test('function f(){return 42 * undefined;}', NaN) self._test('function f(){return 42 * null;}', 0) + self._test('function f(){return 42 * "7";}', 294) + self._test('function f(){return 42 * "eggs";}', NaN) def test_div(self): jsi = JSInterpreter('function f(a, b){return a / b;}') @@ -58,17 +69,26 @@ class TestJSInterpreter(unittest.TestCase): self._test(jsi, NaN, args=(JS_Undefined, 1)) self._test(jsi, float('inf'), args=(2, 0)) self._test(jsi, 0, args=(0, 3)) + self._test(jsi, 6, args=(42, 7)) + self._test(jsi, 0, args=(42, float('inf'))) + self._test(jsi, 6, args=("42", 7)) + self._test(jsi, NaN, args=("spam", 7)) def test_mod(self): self._test('function f(){return 42 % 7;}', 0) self._test('function f(){return 42 % 0;}', NaN) self._test('function f(){return 42 % undefined;}', NaN) + self._test('function f(){return 42 % "7";}', 0) + self._test('function f(){return 42 % "beans";}', NaN) def test_exp(self): self._test('function f(){return 42 ** 2;}', 1764) self._test('function f(){return 42 ** undefined;}', NaN) self._test('function f(){return 42 ** null;}', 1) + self._test('function f(){return undefined ** 0;}', 1) self._test('function f(){return undefined ** 42;}', NaN) + self._test('function f(){return 42 ** "2";}', 1764) + self._test('function f(){return 42 ** "spam";}', NaN) def test_calc(self): self._test('function f(a){return 2*a+1;}', 7, args=[3]) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 248210ce4..73934a6af 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -11,6 +11,7 @@ from functools import update_wrapper, wraps from .utils import ( error_to_compat_str, ExtractorError, + float_or_none, js_to_json, remove_quotes, unified_timestamp, @@ -81,35 +82,47 @@ def _js_bit_op(op): return wrapped -def _js_arith_op(op): +def _js_arith_op(op, div=False): @wraps_op(op) def wrapped(a, b): if JS_Undefined in (a, b): return _NaN - return op(a or 0, b or 0) + # null, "" --> 0 + a, b = (float_or_none( + (x.strip() if isinstance(x, compat_basestring) else x) or 0, + default=_NaN) for x in (a, b)) + if _NaN in (a, b): + return _NaN + try: + return op(a, b) + except ZeroDivisionError: + return _NaN if not (div and (a or b)) else _Infinity return wrapped -def _js_div(a, b): - if JS_Undefined in (a, b) or not (a or b): - return _NaN - return operator.truediv(a or 0, b) if b else _Infinity +_js_arith_add = _js_arith_op(operator.add) -def _js_mod(a, b): - if JS_Undefined in (a, b) or not b: - return _NaN - return (a or 0) % b +def _js_add(a, b): + if not (isinstance(a, compat_basestring) or isinstance(b, compat_basestring)): + return _js_arith_add(a, b) + if not isinstance(a, compat_basestring): + a = _js_toString(a) + elif not isinstance(b, compat_basestring): + b = _js_toString(b) + return operator.concat(a, b) + + +_js_mod = _js_arith_op(operator.mod) +__js_exp = _js_arith_op(operator.pow) def _js_exp(a, b): if not b: return 1 # even 0 ** 0 !! - elif JS_Undefined in (a, b): - return _NaN - return (a or 0) ** b + return __js_exp(a, b) def _js_to_primitive(v): @@ -117,7 +130,7 @@ def _js_to_primitive(v): ','.join(map(_js_toString, v)) if isinstance(v, list) else '[object Object]' if isinstance(v, dict) else compat_str(v) if not isinstance(v, ( - compat_numeric_types, compat_basestring, bool)) + compat_numeric_types, compat_basestring)) else v ) @@ -128,7 +141,9 @@ def _js_toString(v): else 'Infinity' if v == _Infinity else 'NaN' if v is _NaN else 'null' if v is None - else compat_str(v) if isinstance(v, compat_numeric_types) + # bool <= int: do this first + else ('false', 'true')[v] if isinstance(v, bool) + else '{0:.7f}'.format(v).rstrip('.0') if isinstance(v, compat_numeric_types) else _js_to_primitive(v)) @@ -240,11 +255,11 @@ def _js_typeof(expr): _OPERATORS = ( ('>>', _js_bit_op(operator.rshift)), ('<<', _js_bit_op(operator.lshift)), - ('+', _js_arith_op(operator.add)), + ('+', _js_add), ('-', _js_arith_op(operator.sub)), ('*', _js_arith_op(operator.mul)), ('%', _js_mod), - ('/', _js_div), + ('/', _js_arith_op(operator.truediv, div=True)), ('**', _js_exp), ) @@ -873,7 +888,7 @@ class JSInterpreter(object): start, end = m.span() sign = m.group('pre_sign') or m.group('post_sign') ret = local_vars[var] - local_vars[var] += 1 if sign[0] == '+' else -1 + local_vars[var] = _js_add(ret, 1 if sign[0] == '+' else -1) if m.group('pre_sign'): ret = local_vars[var] expr = expr[:start] + self._dump(ret, local_vars) + expr[end:] @@ -1023,7 +1038,7 @@ class JSInterpreter(object): if obj is compat_str: if member == 'fromCharCode': assertion(argvals, 'takes one or more arguments') - return ''.join(map(compat_chr, argvals)) + return ''.join(compat_chr(int(n)) for n in argvals) raise self.Exception('Unsupported string method ' + member, expr=expr) elif obj is float: if member == 'pow': From 94fd7746084d87a43e34b094c5db1325f91ce053 Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 04:13:06 +0000 Subject: [PATCH 11/78] [jsinterp] Fix and improve split/join * improve split/join edge cases * correctly implement regex split (not like re.split) --- test/test_jsinterp.py | 19 +++++++++++++++++ youtube_dl/jsinterp.py | 48 +++++++++++++++++++++++++++++++++++++----- 2 files changed, 62 insertions(+), 5 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 52eaf1ed8..b6e87e9f1 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -483,6 +483,13 @@ class TestJSInterpreter(unittest.TestCase): self._test(jsi, 't-e-s-t', args=[test_input, '-']) self._test(jsi, '', args=[[], '-']) + self._test('function f(){return ' + '[1, 1.0, "abc", {a: 1}, null, undefined, Infinity, NaN].join()}', + '1,1,abc,[object Object],,,Infinity,NaN') + self._test('function f(){return ' + '[1, 1.0, "abc", {a: 1}, null, undefined, Infinity, NaN].join("~")}', + '1~1~abc~[object Object]~~~Infinity~NaN') + def test_split(self): test_result = list('test') tests = [ @@ -496,6 +503,18 @@ class TestJSInterpreter(unittest.TestCase): self._test(jsi, test_result, args=['t-e-s-t', '-']) self._test(jsi, [''], args=['', '-']) self._test(jsi, [], args=['', '']) + # RegExp split + self._test('function f(){return "test".split(/(?:)/)}', + ['t', 'e', 's', 't']) + self._test('function f(){return "t-e-s-t".split(/[es-]+/)}', + ['t', 't']) + # from MDN: surrogate pairs aren't handled: case 1 fails + # self._test('function f(){return "😄😄".split(/(?:)/)}', + # ['\ud83d', '\ude04', '\ud83d', '\ude04']) + # case 2 beats Py3.2: it gets the case 1 result + if sys.version_info >= (2, 6) and not ((3, 0) <= sys.version_info < (3, 3)): + self._test('function f(){return "😄😄".split(/(?:)/u)}', + ['😄', '😄']) def test_slice(self): self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice()}', [0, 1, 2, 3, 4, 5, 6, 7, 8]) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 73934a6af..bec959946 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -397,6 +397,9 @@ class JSInterpreter(object): RE_FLAGS = { # special knowledge: Python's re flags are bitmask values, current max 128 # invent new bitmask values well above that for literal parsing + # JS 'u' flag is effectively always set (surrogate pairs aren't seen), + # but \u{...} and \p{...} escapes aren't handled); no additional JS 'v' + # features are supported # TODO: execute matches with these flags (remaining: d, y) 'd': 1024, # Generate indices for substring matches 'g': 2048, # Global search @@ -404,6 +407,7 @@ class JSInterpreter(object): 'm': re.M, # Multi-line search 's': re.S, # Allows . to match newline characters 'u': re.U, # Treat a pattern as a sequence of unicode code points + 'v': re.U, # Like 'u' with extended character class and \p{} syntax 'y': 4096, # Perform a "sticky" search that matches starting at the current position in the target string } @@ -1047,13 +1051,47 @@ class JSInterpreter(object): raise self.Exception('Unsupported Math method ' + member, expr=expr) if member == 'split': - assertion(argvals, 'takes one or more arguments') - assertion(len(argvals) == 1, 'with limit argument is not implemented') - return obj.split(argvals[0]) if argvals[0] else list(obj) + assertion(len(argvals) <= 2, 'takes at most two arguments') + if len(argvals) > 1: + limit = argvals[1] + assertion(isinstance(limit, int) and limit >= 0, 'integer limit >= 0') + if limit == 0: + return [] + else: + limit = 0 + if len(argvals) == 0: + argvals = [JS_Undefined] + elif isinstance(argvals[0], self.JS_RegExp): + # avoid re.split(), similar but not enough + + def where(): + for m in argvals[0].finditer(obj): + yield m.span(0) + yield (None, None) + + def splits(limit=limit): + i = 0 + for j, jj in where(): + if j == jj == 0: + continue + if j is None and i >= len(obj): + break + yield obj[i:j] + if jj is None or limit == 1: + break + limit -= 1 + i = jj + + return list(splits()) + return ( + obj.split(argvals[0], limit - 1) if argvals[0] and argvals[0] != JS_Undefined + else list(obj)[:limit or None]) elif member == 'join': assertion(isinstance(obj, list), 'must be applied on a list') - assertion(len(argvals) == 1, 'takes exactly one argument') - return argvals[0].join(obj) + assertion(len(argvals) <= 1, 'takes at most one argument') + return (',' if len(argvals) == 0 else argvals[0]).join( + ('' if x in (None, JS_Undefined) else _js_toString(x)) + for x in obj) elif member == 'reverse': assertion(not argvals, 'does not take any arguments') obj.reverse() From 60835ca16c052eb00bb9bccd44f8843edac66a2e Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 04:16:07 +0000 Subject: [PATCH 12/78] [jsinterp] Fix and improve "methods" * push, unshift return new length * impove edge cases for push/pop, shift/unshift, forEach, indexOf, charCodeAt * increase test coverage --- test/test_jsinterp.py | 34 ++++++++++++++++++++++++++++++++++ youtube_dl/jsinterp.py | 36 +++++++++++++++--------------------- 2 files changed, 49 insertions(+), 21 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index b6e87e9f1..07ed481d8 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -544,6 +544,40 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return "012345678".slice(-1, 1)}', '') self._test('function f(){return "012345678".slice(-3, -1)}', '67') + def test_pop(self): + # pop + self._test('function f(){var a = [0, 1, 2, 3, 4, 5, 6, 7, 8]; return [a.pop(), a]}', + [8, [0, 1, 2, 3, 4, 5, 6, 7]]) + self._test('function f(){return [].pop()}', JS_Undefined) + # push + self._test('function f(){var a = [0, 1, 2]; return [a.push(3, 4), a]}', + [5, [0, 1, 2, 3, 4]]) + self._test('function f(){var a = [0, 1, 2]; return [a.push(), a]}', + [3, [0, 1, 2]]) + + def test_shift(self): + # shift + self._test('function f(){var a = [0, 1, 2, 3, 4, 5, 6, 7, 8]; return [a.shift(), a]}', + [0, [1, 2, 3, 4, 5, 6, 7, 8]]) + self._test('function f(){return [].shift()}', JS_Undefined) + # unshift + self._test('function f(){var a = [0, 1, 2]; return [a.unshift(3, 4), a]}', + [5, [3, 4, 0, 1, 2]]) + self._test('function f(){var a = [0, 1, 2]; return [a.unshift(), a]}', + [3, [0, 1, 2]]) + + def test_forEach(self): + self._test('function f(){var ret = []; var l = [4, 2]; ' + 'var log = function(e,i,a){ret.push([e,i,a]);}; ' + 'l.forEach(log); ' + 'return [ret.length, ret[0][0], ret[1][1], ret[0][2]]}', + [2, 4, 1, [4, 2]]) + self._test('function f(){var ret = []; var l = [4, 2]; ' + 'var log = function(e,i,a){this.push([e,i,a]);}; ' + 'l.forEach(log, ret); ' + 'return [ret.length, ret[0][0], ret[1][1], ret[0][2]]}', + [2, 4, 1, [4, 2]]) + if __name__ == '__main__': unittest.main() diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index bec959946..0cfae4b28 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -1113,37 +1113,31 @@ class JSInterpreter(object): index, how_many = map(int, (argvals + [len(obj)])[:2]) if index < 0: index += len(obj) - add_items = argvals[2:] - res = [] - for _ in range(index, min(index + how_many, len(obj))): - res.append(obj.pop(index)) - for i, item in enumerate(add_items): - obj.insert(index + i, item) + res = [obj.pop(index) + for _ in range(index, min(index + how_many, len(obj)))] + obj[index:index] = argvals[2:] return res - elif member == 'unshift': - assertion(isinstance(obj, list), 'must be applied on a list') - assertion(argvals, 'takes one or more arguments') - for item in reversed(argvals): - obj.insert(0, item) - return obj - elif member == 'pop': + elif member in ('shift', 'pop'): assertion(isinstance(obj, list), 'must be applied on a list') assertion(not argvals, 'does not take any arguments') - if not obj: - return - return obj.pop() + return obj.pop(0 if member == 'shift' else -1) if len(obj) > 0 else JS_Undefined + elif member == 'unshift': + assertion(isinstance(obj, list), 'must be applied on a list') + # not enforced: assertion(argvals, 'takes one or more arguments') + obj[0:0] = argvals + return len(obj) elif member == 'push': - assertion(argvals, 'takes one or more arguments') + # not enforced: assertion(argvals, 'takes one or more arguments') obj.extend(argvals) - return obj + return len(obj) elif member == 'forEach': assertion(argvals, 'takes one or more arguments') - assertion(len(argvals) <= 2, 'takes at-most 2 arguments') + assertion(len(argvals) <= 2, 'takes at most 2 arguments') f, this = (argvals + [''])[:2] return [f((item, idx, obj), {'this': this}, allow_recursion) for idx, item in enumerate(obj)] elif member == 'indexOf': assertion(argvals, 'takes one or more arguments') - assertion(len(argvals) <= 2, 'takes at-most 2 arguments') + assertion(len(argvals) <= 2, 'takes at most 2 arguments') idx, start = (argvals + [0])[:2] try: return obj.index(idx, start) @@ -1152,7 +1146,7 @@ class JSInterpreter(object): elif member == 'charCodeAt': assertion(isinstance(obj, compat_str), 'must be applied on a string') # assertion(len(argvals) == 1, 'takes exactly one argument') # but not enforced - idx = argvals[0] if isinstance(argvals[0], int) else 0 + idx = argvals[0] if len(argvals) > 0 and isinstance(argvals[0], int) else 0 if idx >= len(obj): return None return ord(obj[idx]) From f487b4a02a74bb143a7b8f7cb26819b5b661a40b Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 04:21:53 +0000 Subject: [PATCH 13/78] [jsinterp] Strip /* comments */ when parsing * NB: _separate() is looking creaky --- test/test_jsinterp.py | 10 +++++++++- youtube_dl/jsinterp.py | 20 ++++++++++++++++++-- 2 files changed, 27 insertions(+), 3 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 07ed481d8..12e7b9b94 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -160,7 +160,6 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){var x = 20; x += 30 + 1; return x;}', 51) self._test('function f(){var x = 20; x -= 30 + 1; return x;}', -11) - @unittest.skip('Not yet fully implemented') def test_comments(self): self._test(''' function f() { @@ -179,6 +178,15 @@ class TestJSInterpreter(unittest.TestCase): } ''', 3) + self._test(''' + function f() { + var x = ( /* 1 + */ 2 + + /* 30 * 40 */ + 50); + return x; + } + ''', 52) + def test_precedence(self): self._test(''' function f() { diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 0cfae4b28..ec8674936 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -488,9 +488,18 @@ class JSInterpreter(object): skipping = 0 if skip_delims: skip_delims = variadic(skip_delims) + skip_txt = None for idx, char in enumerate(expr): + if skip_txt and idx <= skip_txt[1]: + continue paren_delta = 0 if not in_quote: + if char == '/' and expr[idx:idx + 2] == '/*': + # skip a comment + skip_txt = expr[idx:].find('*/', 2) + skip_txt = [idx, idx + skip_txt + 1] if skip_txt >= 2 else None + if skip_txt: + continue if char in _MATCHING_PARENS: counters[_MATCHING_PARENS[char]] += 1 paren_delta = 1 @@ -523,12 +532,19 @@ class JSInterpreter(object): if pos < delim_len: pos += 1 continue - yield expr[start: idx - delim_len] + if skip_txt and skip_txt[0] >= start and skip_txt[1] <= idx - delim_len: + yield expr[start:skip_txt[0]] + expr[skip_txt[1] + 1: idx - delim_len] + else: + yield expr[start: idx - delim_len] + skip_txt = None start, pos = idx + 1, 0 splits += 1 if max_split and splits >= max_split: break - yield expr[start:] + if skip_txt and skip_txt[0] >= start: + yield expr[start:skip_txt[0]] + expr[skip_txt[1] + 1:] + else: + yield expr[start:] @classmethod def _separate_at_paren(cls, expr, delim=None): From ce1e556b8fde56eb84bec1580e9f07f45e721c4f Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 04:29:13 +0000 Subject: [PATCH 14/78] [jsinterp] Add return hook for player `3bb1f723` * set var `_ytdl_do_not_return` to a specific value in the scope of a function * if an expression to be returned has that value, `return` becomes `void` --- youtube_dl/jsinterp.py | 33 +++++++++++++++++++++------------ 1 file changed, 21 insertions(+), 12 deletions(-) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index ec8674936..7835187f5 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -368,7 +368,7 @@ class Debugger(object): raise if cls.ENABLED and stmt.strip(): if should_ret or repr(ret) != stmt: - cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion) + cls.write(['->', '=>'][bool(should_ret)], repr(ret), '<-|', stmt, level=allow_recursion) return ret, should_ret return interpret_statement @@ -603,7 +603,7 @@ class JSInterpreter(object): # used below _VAR_RET_THROW_RE = re.compile(r'''(?x) - (?P(?:var|const|let)\s)|return(?:\s+|(?=["'])|$)|(?Pthrow\s+) + (?:(?Pvar|const|let)\s+|(?Preturn)(?:\s+|(?=["'])|$)|(?Pthrow)\s+) ''') _COMPOUND_RE = re.compile(r'''(?x) (?Ptry)\s*\{| @@ -683,7 +683,7 @@ class JSInterpreter(object): expr = stmt[len(m.group(0)):].strip() if m.group('throw'): raise JS_Throw(self.interpret_expression(expr, local_vars, allow_recursion)) - should_return = not m.group('var') + should_return = 'return' if m.group('ret') else False if not expr: return None, should_return @@ -968,14 +968,20 @@ class JSInterpreter(object): return _Infinity, should_return elif md.get('return'): - return local_vars[m.group('name')], should_return + ret = local_vars[m.group('name')] + # challenge may try to force returning the original value + # use an optional internal var to block this + if should_return == 'return': + if '_ytdl_do_not_return' not in local_vars: + return ret, True + return (ret, True) if ret != local_vars['_ytdl_do_not_return'] else (ret, False) + else: + return ret, should_return - try: + with compat_contextlib_suppress(ValueError): ret = json.loads(js_to_json(expr)) # strict=True) if not md.get('attribute'): return ret, should_return - except ValueError: - pass if md.get('indexing'): val = local_vars[m.group('in')] @@ -1213,7 +1219,7 @@ class JSInterpreter(object): yield self.interpret_expression(v, local_vars, allow_recursion) def extract_object(self, objname): - _FUNC_NAME_RE = r'''(?:[a-zA-Z$0-9]+|"[a-zA-Z$0-9]+"|'[a-zA-Z$0-9]+')''' + _FUNC_NAME_RE = r'''(?:{n}|"{n}"|'{n}')'''.format(n=_NAME_RE) obj = {} fields = next(filter(None, ( obj_m.group('fields') for obj_m in re.finditer( @@ -1272,6 +1278,7 @@ class JSInterpreter(object): def extract_function_from_code(self, argnames, code, *global_stack): local_vars = {} + while True: mobj = re.search(r'function\((?P[^)]*)\)\s*{', code) if mobj is None: @@ -1282,10 +1289,11 @@ class JSInterpreter(object): [x.strip() for x in mobj.group('args').split(',')], body, local_vars, *global_stack)) code = code[:start] + name + remaining + return self.build_function(argnames, code, local_vars, *global_stack) - def call_function(self, funcname, *args): - return self.extract_function(funcname)(args) + def call_function(self, funcname, *args, **kw_global_vars): + return self.extract_function(funcname)(args, kw_global_vars) @classmethod def build_arglist(cls, arg_text): @@ -1304,8 +1312,9 @@ class JSInterpreter(object): global_stack = list(global_stack) or [{}] argnames = tuple(argnames) - def resf(args, kwargs={}, allow_recursion=100): - global_stack[0].update(zip_longest(argnames, args, fillvalue=None)) + def resf(args, kwargs=None, allow_recursion=100): + kwargs = kwargs or {} + global_stack[0].update(zip_longest(argnames, args, fillvalue=JS_Undefined)) global_stack[0].update(kwargs) var_stack = LocalNameSpace(*global_stack) ret, should_abort = self.interpret_statement(code.replace('\n', ' '), var_stack, allow_recursion - 1) From b4469a0f652c450a81901795a8f522ae5457a1cf Mon Sep 17 00:00:00 2001 From: dirkf Date: Sat, 7 Dec 2024 03:39:44 +0000 Subject: [PATCH 15/78] [YouTube] Handle player `3bb1f723` * fix signature code extraction * raise if n function returns input value * add new tests from yt-dlp Co-authored-by: bashonly --- test/test_youtube_signature.py | 37 +++++++++++++++++++++++++------- youtube_dl/extractor/youtube.py | 38 ++++++++++++++++++++++----------- 2 files changed, 54 insertions(+), 21 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 56e92fac5..7d1ff90ba 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -12,6 +12,7 @@ import re import string from youtube_dl.compat import ( + compat_contextlib_suppress, compat_open as open, compat_str, compat_urlretrieve, @@ -50,23 +51,33 @@ _SIG_TESTS = [ ( 'https://s.ytimg.com/yts/jsbin/html5player-en_US-vflBb0OQx.js', 84, - '123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQ0STUVWXYZ!"#$%&\'()*+,@./:;<=>' + '123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQ0STUVWXYZ!"#$%&\'()*+,@./:;<=>', ), ( 'https://s.ytimg.com/yts/jsbin/html5player-en_US-vfl9FYC6l.js', 83, - '123456789abcdefghijklmnopqr0tuvwxyzABCDETGHIJKLMNOPQRS>UVWXYZ!"#$%&\'()*+,-./:;<=F' + '123456789abcdefghijklmnopqr0tuvwxyzABCDETGHIJKLMNOPQRS>UVWXYZ!"#$%&\'()*+,-./:;<=F', ), ( 'https://s.ytimg.com/yts/jsbin/html5player-en_US-vflCGk6yw/html5player.js', '4646B5181C6C3020DF1D9C7FCFEA.AD80ABF70C39BD369CCCAE780AFBB98FA6B6CB42766249D9488C288', - '82C8849D94266724DC6B6AF89BBFA087EACCD963.B93C07FBA084ACAEFCF7C9D1FD0203C6C1815B6B' + '82C8849D94266724DC6B6AF89BBFA087EACCD963.B93C07FBA084ACAEFCF7C9D1FD0203C6C1815B6B', ), ( 'https://s.ytimg.com/yts/jsbin/html5player-en_US-vflKjOTVq/html5player.js', '312AA52209E3623129A412D56A40F11CB0AF14AE.3EE09501CB14E3BCDC3B2AE808BF3F1D14E7FBF12', '112AA5220913623229A412D56A40F11CB0AF14AE.3EE0950FCB14EEBCDC3B2AE808BF331D14E7FBF3', - ) + ), + ( + 'https://www.youtube.com/s/player/6ed0d907/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'AOq0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xx8j7v1pDL2QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0', + ), + ( + 'https://www.youtube.com/s/player/3bb1f723/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'MyOSJXtKI3m-uME_jv7-pT12gOFC02RFkGoqWpzE0Cs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + ), ] _NSIG_TESTS = [ @@ -142,6 +153,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/5a3b6271/player_ias.vflset/en_US/base.js', 'B2j7f_UPT4rfje85Lu_e', 'm5DmNymaGQ5RdQ', ), + ( + 'https://www.youtube.com/s/player/7a062b77/player_ias.vflset/en_US/base.js', + 'NRcE3y3mVtm_cV-W', 'VbsCYUATvqlt5w', + ), ( 'https://www.youtube.com/s/player/dac945fd/player_ias.vflset/en_US/base.js', 'o8BkRxXhuYsBCWi6RplPdP', '3Lx32v_hmzTm6A', @@ -154,6 +169,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/cfa9e7cb/player_ias.vflset/en_US/base.js', 'qO0NiMtYQ7TeJnfFG2', 'k9cuJDHNS5O7kQ', ), + ( + 'https://www.youtube.com/s/player/8c7583ff/player_ias.vflset/en_US/base.js', + '1wWCVpRR96eAmMI87L', 'KSkWAVv1ZQxC3A', + ), ( 'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js', '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ', @@ -182,6 +201,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/b12cc44b/player_ias.vflset/en_US/base.js', 'keLa5R2U00sR9SQK', 'N1OGyujjEwMnLw', ), + ( + 'https://www.youtube.com/s/player/3bb1f723/player_ias.vflset/en_US/base.js', + 'gK15nzVyaXE9RsMP3z', 'ZFFWFLPWx9DEgQ', + ), ] @@ -216,11 +239,9 @@ class TestSignature(unittest.TestCase): os.mkdir(self.TESTDATA_DIR) def tearDown(self): - try: + with compat_contextlib_suppress(OSError): for f in os.listdir(self.TESTDATA_DIR): os.remove(f) - except OSError: - pass def t_factory(name, sig_func, url_pattern): @@ -258,7 +279,7 @@ def n_sig(jscode, sig_input): make_sig_test = t_factory( - 'signature', signature, re.compile(r'.*-(?P[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.[a-z]+$')) + 'signature', signature, re.compile(r'.*(?:-|/player/)(?P[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$')) for test_spec in _SIG_TESTS: make_sig_test(*test_spec) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 6fe520e9a..d633032ae 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1579,19 +1579,26 @@ class YoutubeIE(YoutubeBaseInfoExtractor): self.to_screen('Extracted signature function:\n' + code) def _parse_sig_js(self, jscode): + # Examples where `sig` is funcname: + # sig=function(a){a=a.split(""); ... ;return a.join("")}; + # ;c&&(c=sig(decodeURIComponent(c)),a.set(b,encodeURIComponent(c)));return a}; + # {var l=f,m=h.sp,n=sig(decodeURIComponent(h.s));l.set(m,encodeURIComponent(n))} + # sig=function(J){J=J.split(""); ... ;return J.join("")}; + # ;N&&(N=sig(decodeURIComponent(N)),J.set(R,encodeURIComponent(N)));return J}; + # {var H=u,k=f.sp,v=sig(decodeURIComponent(f.s));H.set(k,encodeURIComponent(v))} funcname = self._search_regex( - (r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[a-zA-Z0-9$]+)\(', + (r'\b(?P[a-zA-Z0-9$]+)&&\((?P=var)=(?P[a-zA-Z0-9$]{2,})\(decodeURIComponent\((?P=var)\)\)', + r'(?P[a-zA-Z0-9$]+)\s*=\s*function\(\s*(?P[a-zA-Z0-9$]+)\s*\)\s*{\s*(?P=arg)\s*=\s*(?P=arg)\.split\(\s*""\s*\)\s*;\s*[^}]+;\s*return\s+(?P=arg)\.join\(\s*""\s*\)', + r'(?:\b|[^a-zA-Z0-9$])(?P[a-zA-Z0-9$]{2,})\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)(?:;[a-zA-Z0-9$]{2}\.[a-zA-Z0-9$]{2}\(a,\d+\))?', + # Old patterns + r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[a-zA-Z0-9$]+)\(', r'\b[a-zA-Z0-9]+\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[a-zA-Z0-9$]+)\(', r'\bm=(?P[a-zA-Z0-9$]{2,})\(decodeURIComponent\(h\.s\)\)', - r'\bc&&\(c=(?P[a-zA-Z0-9$]{2,})\(decodeURIComponent\(c\)\)', - r'(?:\b|[^a-zA-Z0-9$])(?P[a-zA-Z0-9$]{2,})\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)(?:;[a-zA-Z0-9$]{2}\.[a-zA-Z0-9$]{2}\(a,\d+\))?', - r'(?P[a-zA-Z0-9$]+)\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)', # Obsolete patterns r'("|\')signature\1\s*,\s*(?P[a-zA-Z0-9$]+)\(', r'\.sig\|\|(?P[a-zA-Z0-9$]+)\(', r'yt\.akamaized\.net/\)\s*\|\|\s*.*?\s*[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?:encodeURIComponent\s*\()?\s*(?P[a-zA-Z0-9$]+)\(', r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?P[a-zA-Z0-9$]+)\(', - r'\b[a-zA-Z0-9]+\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*(?P[a-zA-Z0-9$]+)\(', r'\bc\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*\([^)]*\)\s*\(\s*(?P[a-zA-Z0-9$]+)\('), jscode, 'Initial JS player signature function name', group='sig') @@ -1658,6 +1665,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_n_function_name(self, jscode): func_name, idx = self._search_regex( + # (R="nn"[+J.Z],mW(J),N=J.K[R]||null)&&(N=narray[idx](N),J.set(R,N))}}; # 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: (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b) @@ -1666,7 +1674,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # older: (b=a.get("n"))&&(b=nfunc(b) r'''(?x) \((?:[\w$()\s]+,)*?\s* # ( - (?P[a-z])\s*=\s* # b= + (?P[a-zA-Z])\s*=\s* # b=, R= (?: (?: # expect ,c=a.get(b) (etc) String\s*\.\s*fromCharCode\s*\(\s*110\s*\)| @@ -1679,7 +1687,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*? (?:\.\s*n|\[\s*"n"\s*]|\.\s*get\s*\(\s*"n"\s*\)) | # ,c=a.get(b) - ,\s*(?P[a-z])\s*=\s*[a-z]\s* + ,\s*(?P[a-zA-Z])\s*=\s*[a-zA-Z]\s* (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*? (?:\[\s*(?P=b)\s*]|\.\s*get\s*\(\s*(?P=b)\s*\)) ) @@ -1697,15 +1705,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor): r'''(?xs) (?:(?<=[^\w$])|^) # instead of \b, which ignores $ (?P(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) - \s*\{(?:(?!};).)+?["']enhanced_except_ + \s*\{(?:(?!};).)+?(?: + ["']enhanced_except_ | + return\s*(?P"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[a-zA-Z0-9_$]+ + ) ''', jscode, 'Initial JS player n function name', group='name') if not idx: return func_name - return self._parse_json(self._search_regex( - r'var\s+{0}\s*=\s*(\[.+?\])\s*[,;]'.format(re.escape(func_name)), jscode, - 'Initial JS player n function list ({0}.{1})'.format(func_name, idx)), - func_name, transform_source=js_to_json)[int(idx)] + return self._search_json( + r'var\s+{0}\s*='.format(re.escape(func_name)), jscode, + 'Initial JS player n function list ({0}.{1})'.format(func_name, idx), + func_name, contains_pattern=r'\[[\s\S]+\]', end_pattern='[,;]', + transform_source=js_to_json)[int(idx)] def _extract_n_function_code(self, video_id, player_url): player_id = self._extract_player_info(player_url) @@ -1734,7 +1746,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): except Exception as e: raise JSInterpreter.Exception(traceback.format_exc(), cause=e) - if ret.startswith('enhanced_except_'): + if ret.startswith('enhanced_except_') or ret.endswith(s): raise JSInterpreter.Exception('Signature function returned an exception') return ret From eed784e15f6066b152a3cce8db6fe3f059290b22 Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 12 Dec 2024 04:38:23 +0000 Subject: [PATCH 16/78] [YouTube] Pass nsig value as return hook, fixes player `3bb1f723` --- test/test_youtube_signature.py | 4 +++- youtube_dl/extractor/youtube.py | 2 +- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 7d1ff90ba..1c2ba848e 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -1,4 +1,5 @@ #!/usr/bin/env python +# coding: utf-8 from __future__ import unicode_literals @@ -275,7 +276,8 @@ def signature(jscode, sig_input): def n_sig(jscode, sig_input): funcname = YoutubeIE(FakeYDL())._extract_n_function_name(jscode) - return JSInterpreter(jscode).call_function(funcname, sig_input) + return JSInterpreter(jscode).call_function( + funcname, sig_input, _ytdl_do_not_return=sig_input) make_sig_test = t_factory( diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index d633032ae..55881f4d1 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1740,7 +1740,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def extract_nsig(s): try: - ret = func([s]) + ret = func([s], kwargs={'_ytdl_do_not_return': s}) except JSInterpreter.Exception: raise except Exception as e: From cd7c7b5edbf020730f106aabaa15ca1a7de22e44 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 13 Dec 2024 03:09:29 +0000 Subject: [PATCH 17/78] [YouTube] Simplify pattern for nsig function name extraction --- test/test_youtube_signature.py | 4 +++ youtube_dl/extractor/youtube.py | 46 ++++++++++++++------------------- 2 files changed, 23 insertions(+), 27 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 1c2ba848e..1d6bce67c 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -206,6 +206,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/3bb1f723/player_ias.vflset/en_US/base.js', 'gK15nzVyaXE9RsMP3z', 'ZFFWFLPWx9DEgQ', ), + ( + 'https://www.youtube.com/s/player/f8f53e1a/player_ias.vflset/en_US/base.js', + 'VTQOUOv0mCIeJ7i8kZB', 'kcfD8wy0sNLyNQ', + ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 55881f4d1..f59d1e36a 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1665,37 +1665,29 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_n_function_name(self, jscode): func_name, idx = self._search_regex( + # (y=NuD(),Mw(k),q=k.Z[y]||null)&&(q=narray[idx](q),k.set(y,q),k.V||NuD(''))}}; # (R="nn"[+J.Z],mW(J),N=J.K[R]||null)&&(N=narray[idx](N),J.set(R,N))}}; - # 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: (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b) + # or: (b=String.fromCharCode(110),c=a.get(b))&&c=narray[idx](c) + # or: (b="nn"[+a.D],c=a.get(b))&&(c=narray[idx](c) + # or: (PL(a),b=a.j.n||null)&&(b=narray[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("") - # old: (b=a.get("n"))&&(b=nfunc[idx](b)(?P[a-z])\s*=\s*[a-z]\s* + # old: (b=a.get("n"))&&(b=narray[idx](b)(?P[a-z])\s*=\s*[a-z]\s* # older: (b=a.get("n"))&&(b=nfunc(b) r'''(?x) - \((?:[\w$()\s]+,)*?\s* # ( - (?P[a-zA-Z])\s*=\s* # b=, R= - (?: - (?: # expect ,c=a.get(b) (etc) - String\s*\.\s*fromCharCode\s*\(\s*110\s*\)| - "n+"\[\s*\+?s*[\w$.]+\s*] - )\s*(?:,[\w$()\s]+(?=,))*| - (?P[\w$]+) # a (old[er]) - )\s* - (?(old) - # b.get("n") - (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*? - (?:\.\s*n|\[\s*"n"\s*]|\.\s*get\s*\(\s*"n"\s*\)) - | # ,c=a.get(b) - ,\s*(?P[a-zA-Z])\s*=\s*[a-zA-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[a-zA-Z_$][\w$]*)(?:\s*\[(?P\d+)\])?\s*\(\s*[\w$]+\s*\) + # (expr, ..., + \((?:(?:\s*[\w$]+\s*=)?(?:[\w$"+\.\s(\[]+(?:[)\]]\s*)?),)* + # b=... + (?P[\w$]+)\s*=\s*(?!(?P=b)[^\w$])[\w$]+\s*(?:(?: + \.\s*[\w$]+ | + \[\s*[\w$]+\s*\] | + \.\s*get\s*\(\s*[\w$"]+\s*\) + )\s*){,2}(?:\s*\|\|\s*null(?=\s*\)))?\s* + \)\s*&&\s*\( # ...)&&( + # b = nfunc, b = narray[idx] + (?P=b)\s*=\s*(?P[\w$]+)\s* + (?:\[\s*(?P[\w$]+)\s*\]\s*)? + # (...) + \(\s*[\w$]+\s*\) ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), default=(None, None)) # thx bashonly: yt-dlp/yt-dlp/pull/10611 From eeafbbc3e59bcd6513053f5c942a55e2ff309488 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 13 Dec 2024 12:16:31 +0000 Subject: [PATCH 18/78] [YouTube] Fix signature function extraction for `2f1832d2` * `_` was omitted from patterns * thx yt-dlp/yt-dlp#11801 Co-authored-by: bashonly --- test/test_youtube_signature.py | 9 +++++++++ youtube_dl/extractor/youtube.py | 24 ++++++++++++------------ 2 files changed, 21 insertions(+), 12 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 1d6bce67c..fcbc9d7a8 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -79,6 +79,11 @@ _SIG_TESTS = [ '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', 'MyOSJXtKI3m-uME_jv7-pT12gOFC02RFkGoqWpzE0Cs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', ), + ( + 'https://www.youtube.com/s/player/2f1832d2/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xxAj7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJ2OySqa0q', + ), ] _NSIG_TESTS = [ @@ -210,6 +215,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/f8f53e1a/player_ias.vflset/en_US/base.js', 'VTQOUOv0mCIeJ7i8kZB', 'kcfD8wy0sNLyNQ', ), + ( + 'https://www.youtube.com/s/player/2f1832d2/player_ias.vflset/en_US/base.js', + 'YWt1qdbe8SAfkoPHW5d', 'RrRjWQOJmBiP', + ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index f59d1e36a..7e0ba04c0 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1587,19 +1587,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # ;N&&(N=sig(decodeURIComponent(N)),J.set(R,encodeURIComponent(N)));return J}; # {var H=u,k=f.sp,v=sig(decodeURIComponent(f.s));H.set(k,encodeURIComponent(v))} funcname = self._search_regex( - (r'\b(?P[a-zA-Z0-9$]+)&&\((?P=var)=(?P[a-zA-Z0-9$]{2,})\(decodeURIComponent\((?P=var)\)\)', - r'(?P[a-zA-Z0-9$]+)\s*=\s*function\(\s*(?P[a-zA-Z0-9$]+)\s*\)\s*{\s*(?P=arg)\s*=\s*(?P=arg)\.split\(\s*""\s*\)\s*;\s*[^}]+;\s*return\s+(?P=arg)\.join\(\s*""\s*\)', - r'(?:\b|[^a-zA-Z0-9$])(?P[a-zA-Z0-9$]{2,})\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)(?:;[a-zA-Z0-9$]{2}\.[a-zA-Z0-9$]{2}\(a,\d+\))?', + (r'\b(?P[\w$]+)&&\((?P=var)=(?P[\w$]{2,})\(decodeURIComponent\((?P=var)\)\)', + r'(?P[\w$]+)\s*=\s*function\(\s*(?P[\w$]+)\s*\)\s*{\s*(?P=arg)\s*=\s*(?P=arg)\.split\(\s*""\s*\)\s*;\s*[^}]+;\s*return\s+(?P=arg)\.join\(\s*""\s*\)', + r'(?:\b|[^\w$])(?P[\w$]{2,})\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)(?:;[\w$]{2}\.[\w$]{2}\(a,\d+\))?', # Old patterns - r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[a-zA-Z0-9$]+)\(', - r'\b[a-zA-Z0-9]+\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[a-zA-Z0-9$]+)\(', - r'\bm=(?P[a-zA-Z0-9$]{2,})\(decodeURIComponent\(h\.s\)\)', + r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[\w$]+)\(', + r'\b[\w]+\s*&&\s*[\w]+\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P[\w$]+)\(', + r'\bm=(?P[\w$]{2,})\(decodeURIComponent\(h\.s\)\)', # Obsolete patterns - r'("|\')signature\1\s*,\s*(?P[a-zA-Z0-9$]+)\(', - r'\.sig\|\|(?P[a-zA-Z0-9$]+)\(', - r'yt\.akamaized\.net/\)\s*\|\|\s*.*?\s*[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?:encodeURIComponent\s*\()?\s*(?P[a-zA-Z0-9$]+)\(', - r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?P[a-zA-Z0-9$]+)\(', - r'\bc\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*\([^)]*\)\s*\(\s*(?P[a-zA-Z0-9$]+)\('), + r'("|\')signature\1\s*,\s*(?P[\w$]+)\(', + r'\.sig\|\|(?P[\w$]+)\(', + r'yt\.akamaized\.net/\)\s*\|\|\s*.*?\s*[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?:encodeURIComponent\s*\()?\s*(?P[\w$]+)\(', + r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?P[\w$]+)\(', + r'\bc\s*&&\s*[\w]+\.set\([^,]+\s*,\s*\([^)]*\)\s*\(\s*(?P[\w$]+)\('), jscode, 'Initial JS player signature function name', group='sig') jsi = JSInterpreter(jscode) @@ -1699,7 +1699,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): (?P(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) \s*\{(?:(?!};).)+?(?: ["']enhanced_except_ | - return\s*(?P"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[a-zA-Z0-9_$]+ + return\s*(?P"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[\w$]+ ) ''', jscode, 'Initial JS player n function name', group='name') if not idx: From d55d1f423d6473ae6a9e13462d94bad1d71d28e0 Mon Sep 17 00:00:00 2001 From: dirkf Date: Sat, 14 Dec 2024 11:18:34 +0000 Subject: [PATCH 19/78] [YouTube] Always extract using MWEB API client * temporary fix-up for 403 on download * MWEB parameters from yt-dlp 2024-12-06 --- youtube_dl/extractor/youtube.py | 84 ++++++++++++++++++++++++++++++--- 1 file changed, 77 insertions(+), 7 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 7e0ba04c0..1f83acf7c 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -3,11 +3,13 @@ from __future__ import unicode_literals import collections +import hashlib import itertools import json import os.path import random import re +import time import traceback from .common import InfoExtractor, SearchInfoExtractor @@ -290,6 +292,33 @@ class YoutubeBaseInfoExtractor(InfoExtractor): _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*=\s*({.+?})\s*;' _YT_INITIAL_BOUNDARY_RE = r'(?:var\s+meta| Date: Tue, 31 Dec 2024 14:51:29 +0000 Subject: [PATCH 20/78] [YouTube] Avoid early crash if webpage can't be read * see issue #33013 --- youtube_dl/extractor/youtube.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 1f83acf7c..7ea30fd40 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1951,7 +1951,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): pb_context = {'html5Preference': 'HTML5_PREF_WANTS'} player_url = self._extract_player_url(webpage) - ytcfg = self._extract_ytcfg(video_id, webpage) + ytcfg = self._extract_ytcfg(video_id, webpage or '') sts = self._extract_signature_timestamp(video_id, player_url, ytcfg) if sts: pb_context['signatureTimestamp'] = sts From ab7c61ca29ed1d1216d463d01794eb112a9144d5 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 6 Jan 2025 01:22:16 +0000 Subject: [PATCH 21/78] [YouTube] Apply code style changes, trailing commas, etc --- youtube_dl/extractor/youtube.py | 60 +++++++++++++++++---------------- 1 file changed, 31 insertions(+), 29 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 7ea30fd40..e9603d155 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -9,6 +9,7 @@ import json import os.path import random import re +import string import time import traceback @@ -67,6 +68,7 @@ from ..utils import ( class YoutubeBaseInfoExtractor(InfoExtractor): """Provide base functions for Youtube extractors""" + _LOGIN_URL = 'https://accounts.google.com/ServiceLogin' _TWOFACTOR_URL = 'https://accounts.google.com/signin/challenge' @@ -138,7 +140,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): [2, 1, None, 1, 'https://accounts.google.com/ServiceLogin?passive=true&continue=https%3A%2F%2Fwww.youtube.com%2Fsignin%3Fnext%3D%252F%26action_handle_signin%3Dtrue%26hl%3Den%26app%3Ddesktop%26feature%3Dsign_in_button&hl=en&service=youtube&uilel=3&requestPath=%2FServiceLogin&Page=PasswordSeparationSignIn', None, [], 4], - 1, [None, None, []], None, None, None, True + 1, [None, None, []], None, None, None, True, ], username, ] @@ -160,7 +162,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): None, 1, None, [1, None, None, None, [password, None, True]], [ None, None, [2, 1, None, 1, 'https://accounts.google.com/ServiceLogin?passive=true&continue=https%3A%2F%2Fwww.youtube.com%2Fsignin%3Fnext%3D%252F%26action_handle_signin%3Dtrue%26hl%3Den%26app%3Ddesktop%26feature%3Dsign_in_button&hl=en&service=youtube&uilel=3&requestPath=%2FServiceLogin&Page=PasswordSeparationSignIn', None, [], 4], - 1, [None, None, []], None, None, None, True + 1, [None, None, []], None, None, None, True, ]] challenge_results = req( @@ -213,7 +215,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): user_hash, None, 2, None, [ 9, None, None, None, None, None, None, None, - [None, tfa_code, True, 2] + [None, tfa_code, True, 2], ]] tfa_results = req( @@ -284,7 +286,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'client': { 'clientName': 'WEB', 'clientVersion': '2.20201021.03.00', - } + }, }, } @@ -385,7 +387,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'client': { 'clientName': 'WEB', 'clientVersion': '2.20201021.03.00', - } + }, }, 'query': query, } @@ -462,7 +464,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): # (HTML, videodetails, metadata, renderers) 'name': ('content', 'author', (('ownerChannelName', None), 'title'), ['text']), 'url': ('href', 'ownerProfileUrl', 'vanityChannelUrl', - ['navigationEndpoint', 'browseEndpoint', 'canonicalBaseUrl']) + ['navigationEndpoint', 'browseEndpoint', 'canonicalBaseUrl']), } if any((videodetails, metadata, renderers)): result = ( @@ -671,7 +673,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/8KVIDEO', 'description': '', 'uploader': '8KVIDEO', - 'title': 'UHDTV TEST 8K VIDEO.mp4' + 'title': 'UHDTV TEST 8K VIDEO.mp4', }, 'params': { 'youtube_include_dash_manifest': True, @@ -711,7 +713,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'uploader_url': r're:https?://(?:www\.)?youtube\.com/@theamazingatheist', 'title': 'Burning Everyone\'s Koran', 'description': 'SUBSCRIBE: http://www.youtube.com/saturninefilms \r\n\r\nEven Obama has taken a stand against freedom on this issue: http://www.huffingtonpost.com/2010/09/09/obama-gma-interview-quran_n_710282.html', - } + }, }, # Age-gated videos { @@ -839,7 +841,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }, 'expected_warnings': [ 'DASH manifest missing', - ] + ], }, # Olympics (https://github.com/ytdl-org/youtube-dl/issues/4431) { @@ -1820,8 +1822,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # cpn generation algorithm is reverse engineered from base.js. # In fact it works even with dummy cpn. - CPN_ALPHABET = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-_' - cpn = ''.join(CPN_ALPHABET[random.randint(0, 256) & 63] for _ in range(0, 16)) + CPN_ALPHABET = string.ascii_letters + string.digits + '-_' + cpn = ''.join(CPN_ALPHABET[random.randint(0, 256) & 63] for _ in range(16)) # more consistent results setting it to right before the end qs = parse_qs(playback_url) @@ -1881,8 +1883,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): mobj = re.match(cls._VALID_URL, url, re.VERBOSE) if mobj is None: raise ExtractorError('Invalid URL: %s' % url) - video_id = mobj.group(2) - return video_id + return mobj.group(2) def _extract_chapters_from_json(self, data, video_id, duration): chapters_list = try_get( @@ -2035,7 +2036,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): headers = { 'X-YouTube-Client-Name': '85', 'X-YouTube-Client-Version': '2.0', - 'Origin': 'https://www.youtube.com' + 'Origin': 'https://www.youtube.com', } video_info = self._call_api('player', query, video_id, fatal=False, headers=headers) @@ -2064,8 +2065,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return ''.join([r['text'] for r in runs if isinstance(r.get('text'), compat_str)]) search_meta = ( - lambda x: self._html_search_meta(x, webpage, default=None)) \ - if webpage else lambda x: None + (lambda x: self._html_search_meta(x, webpage, default=None)) + if webpage else lambda _: None) video_details = player_response.get('videoDetails') or {} microformat = try_get( @@ -2137,7 +2138,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def build_fragments(f): return LazyList({ 'url': update_url_query(f['url'], { - 'range': '{0}-{1}'.format(range_start, min(range_start + CHUNK_SIZE - 1, f['filesize'])) + 'range': '{0}-{1}'.format(range_start, min(range_start + CHUNK_SIZE - 1, f['filesize'])), }) } for range_start in range(0, f['filesize'], CHUNK_SIZE)) @@ -2236,7 +2237,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'protocol': 'http_dash_segments', 'fragments': build_fragments(dct), } if dct['filesize'] else { - 'downloader_options': {'http_chunk_size': CHUNK_SIZE} # No longer useful? + 'downloader_options': {'http_chunk_size': CHUNK_SIZE}, # No longer useful? }) formats.append(dct) @@ -2454,7 +2455,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): info['subtitles'] = subtitles parsed_url = compat_urllib_parse_urlparse(url) - for component in [parsed_url.fragment, parsed_url.query]: + for component in (parsed_url.fragment, parsed_url.query): query = compat_parse_qs(component) for k, v in query.items(): for d_k, s_ks in [('start', ('start', 't')), ('end', ('end',))]: @@ -2684,7 +2685,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'title': 'Super Cooper Shorts - Shorts', 'uploader': 'Super Cooper Shorts', 'uploader_id': '@SuperCooperShorts', - } + }, }, { # Channel that does not have a Shorts tab. Test should just download videos on Home tab instead 'url': 'https://www.youtube.com/@emergencyawesome/shorts', @@ -2738,7 +2739,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'description': 'md5:609399d937ea957b0f53cbffb747a14c', 'uploader': 'ThirstForScience', 'uploader_id': '@ThirstForScience', - } + }, }, { 'url': 'https://www.youtube.com/c/ChristophLaimer/playlists', 'only_matching': True, @@ -3037,7 +3038,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'uploader': '3Blue1Brown', 'uploader_id': '@3blue1brown', 'channel_id': 'UCYO_jab_esuFRV4b17AJtAw', - } + }, }] @classmethod @@ -3335,7 +3336,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'client': { 'clientName': 'WEB', 'clientVersion': client_version, - } + }, } visitor_data = try_get(context, lambda x: x['client']['visitorData'], compat_str) @@ -3354,7 +3355,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): headers['x-goog-visitor-id'] = visitor_data data['continuation'] = continuation['continuation'] data['clickTracking'] = { - 'clickTrackingParams': continuation['itct'] + 'clickTrackingParams': continuation['itct'], } count = 0 retries = 3 @@ -3613,7 +3614,7 @@ class YoutubePlaylistIE(InfoExtractor): 'uploader': 'milan', 'uploader_id': '@milan5503', 'channel_id': 'UCEI1-PVPcYXjB73Hfelbmaw', - } + }, }, { 'url': 'http://www.youtube.com/embed/_xDOZElKyNU?list=PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl', 'playlist_mincount': 455, @@ -3623,7 +3624,7 @@ class YoutubePlaylistIE(InfoExtractor): 'uploader': 'LBK', 'uploader_id': '@music_king', 'channel_id': 'UC21nz3_MesPLqtDqwdvnoxA', - } + }, }, { 'url': 'TLGGrESM50VT6acwMjAyMjAxNw', 'only_matching': True, @@ -3734,7 +3735,7 @@ class YoutubeSearchIE(SearchInfoExtractor, YoutubeBaseInfoExtractor): 'info_dict': { 'id': 'youtube-dl test video', 'title': 'youtube-dl test video', - } + }, }] def _get_n_results(self, query, n): @@ -3754,7 +3755,7 @@ class YoutubeSearchDateIE(YoutubeSearchIE): 'info_dict': { 'id': 'youtube-dl test video', 'title': 'youtube-dl test video', - } + }, }] @@ -3769,7 +3770,7 @@ class YoutubeSearchURLIE(YoutubeBaseInfoExtractor): 'id': 'youtube-dl test video', 'title': 'youtube-dl test video', }, - 'params': {'playlistend': 5} + 'params': {'playlistend': 5}, }, { 'url': 'https://www.youtube.com/results?q=test&sp=EgQIBBgB', 'only_matching': True, @@ -3785,6 +3786,7 @@ class YoutubeSearchURLIE(YoutubeBaseInfoExtractor): class YoutubeFeedsInfoExtractor(YoutubeTabIE): """ Base class for feed extractors + Subclasses must define the _FEED_NAME property. """ _LOGIN_REQUIRED = True From 00ad2b8ca12d4f9b830ed83876d0d1ab3d698675 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 6 Jan 2025 01:24:30 +0000 Subject: [PATCH 22/78] [YouTube] Refactor subtitle processing * move to internal function * use `traverse-obj()` --- youtube_dl/extractor/youtube.py | 46 +++++++++++++++++---------------- 1 file changed, 24 insertions(+), 22 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index e9603d155..56957a661 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -2415,9 +2415,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'is_live': is_live, } - pctr = try_get( + pctr = traverse_obj( player_response, - lambda x: x['captions']['playerCaptionsTracklistRenderer'], dict) + ('captions', 'playerCaptionsTracklistRenderer', T(dict))) if pctr: def process_language(container, base_url, lang_code, query): lang_subs = [] @@ -2431,28 +2431,30 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }) container[lang_code] = lang_subs - subtitles = {} - for caption_track in (pctr.get('captionTracks') or []): - base_url = caption_track.get('baseUrl') - if not base_url: - continue - if caption_track.get('kind') != 'asr': - lang_code = caption_track.get('languageCode') - if not lang_code: + def process_subtitles(): + subtitles = {} + for caption_track in traverse_obj(pctr, ( + 'captionTracks', lambda _, v: v.get('baseUrl'))): + if not base_url: continue - process_language( - subtitles, base_url, lang_code, {}) - continue - automatic_captions = {} - for translation_language in (pctr.get('translationLanguages') or []): - translation_language_code = translation_language.get('languageCode') - if not translation_language_code: + if caption_track.get('kind') != 'asr': + lang_code = caption_track.get('languageCode') + if not lang_code: + continue + process_language( + subtitles, base_url, lang_code, {}) continue - process_language( - automatic_captions, base_url, translation_language_code, - {'tlang': translation_language_code}) - info['automatic_captions'] = automatic_captions - info['subtitles'] = subtitles + automatic_captions = {} + for translation_language in traverse_obj(pctr, ( + 'translationLanguages', lambda _, v: v.get('languageCode'))): + translation_language_code = translation_language['languageCode'] + process_language( + automatic_captions, base_url, translation_language_code, + {'tlang': translation_language_code}) + info['automatic_captions'] = automatic_captions + info['subtitles'] = subtitles + + process_subtitles() parsed_url = compat_urllib_parse_urlparse(url) for component in (parsed_url.fragment, parsed_url.query): From 1036478d130c5f2001eca2d7d12558abe601d933 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 6 Jan 2025 01:39:04 +0000 Subject: [PATCH 23/78] [YouTube] Endure subtitle URLs are complete * WEB URLs are, MWEB not * resolves #33017 --- youtube_dl/extractor/youtube.py | 1 + 1 file changed, 1 insertion(+) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 56957a661..6171df84a 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -2435,6 +2435,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): subtitles = {} for caption_track in traverse_obj(pctr, ( 'captionTracks', lambda _, v: v.get('baseUrl'))): + base_url = self._yt_urljoin(caption_track['baseUrl']) if not base_url: continue if caption_track.get('kind') != 'asr': From 21fff051217d1c14a7d50fa752052eadbcafee4e Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 15 Jan 2025 03:19:15 +0000 Subject: [PATCH 24/78] [YouTube] Switch to TV API client * thx yt-dlp/yt-dlp#12059 --- youtube_dl/extractor/youtube.py | 102 ++++++++++++++++++++++++++------ 1 file changed, 83 insertions(+), 19 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 6171df84a..1424277ac 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -31,7 +31,9 @@ from ..utils import ( dict_get, error_to_compat_str, ExtractorError, + filter_dict, float_or_none, + get_first, extract_attributes, get_element_by_attribute, int_or_none, @@ -82,6 +84,34 @@ class YoutubeBaseInfoExtractor(InfoExtractor): _PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM)' + _INNERTUBE_CLIENTS = { + # mweb has 'ultralow' formats + # See: https://github.com/yt-dlp/yt-dlp/pull/557 + 'mweb': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'MWEB', + 'clientVersion': '2.20241202.07.00', + # mweb previously did not require PO Token with this UA + 'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 2, + 'REQUIRE_PO_TOKEN': True, + 'SUPPORTS_COOKIES': True, + }, + 'tv': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'TVHTML5', + 'clientVersion': '7.20241201.18.00', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 7, + 'SUPPORTS_COOKIES': True, + }, + } + def _login(self): """ Attempt to log in to YouTube. @@ -321,19 +351,24 @@ class YoutubeBaseInfoExtractor(InfoExtractor): '{0} {1} {2}'.format(time_now, self._SAPISID, origin).encode('utf-8')).hexdigest() return 'SAPISIDHASH {0}_{1}'.format(time_now, sapisidhash) - def _call_api(self, ep, query, video_id, fatal=True, headers=None): + def _call_api(self, ep, query, video_id, fatal=True, headers=None, + note='Downloading API JSON'): data = self._DEFAULT_API_DATA.copy() data.update(query) real_headers = {'content-type': 'application/json'} if headers: real_headers.update(headers) + # was: 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8' + api_key = self.get_param('youtube_innertube_key') return self._download_json( 'https://www.youtube.com/youtubei/v1/%s' % ep, video_id=video_id, - note='Downloading API JSON', errnote='Unable to download API page', + note=note, errnote='Unable to download API page', data=json.dumps(data).encode('utf8'), fatal=fatal, - headers=real_headers, - query={'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8'}) + headers=real_headers, query=filter_dict({ + 'key': api_key, + 'prettyPrint': 'false', + })) def _extract_yt_initial_data(self, video_id, webpage): return self._parse_json( @@ -342,6 +377,22 @@ class YoutubeBaseInfoExtractor(InfoExtractor): self._YT_INITIAL_DATA_RE), webpage, 'yt initial data'), video_id) + def _extract_visitor_data(self, *args): + """ + Extract visitorData from an API response or ytcfg + + Appears to be used to track session state + """ + visitor_data = self.get_param('youtube_visitor_data') + if visitor_data: + return visitor_data + + return get_first( + args, (('VISITOR_DATA', + ('INNERTUBE_CONTEXT', 'client', 'visitorData'), + ('responseContext', 'visitorData')), + T(compat_str))) + def _extract_ytcfg(self, video_id, webpage): return self._parse_json( self._search_regex( @@ -1957,6 +2008,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if sts: pb_context['signatureTimestamp'] = sts + client = traverse_obj(self._INNERTUBE_CLIENTS, ( + lambda _, v: not v.get('REQUIRE_PO_TOKEN')), + get_all=False) + query = { 'playbackContext': { 'contentPlaybackContext': pb_context, @@ -1964,30 +2019,39 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'racyCheckOk': True, }, 'context': { - 'client': { - 'clientName': 'MWEB', - 'clientVersion': '2.20241202.07.00', - 'hl': 'en', - 'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)', - 'timeZone': 'UTC', - 'utcOffsetMinutes': 0, - }, + 'client': merge_dicts( + traverse_obj(client, ('INNERTUBE_CONTEXT', 'client')), { + 'hl': 'en', + 'timeZone': 'UTC', + 'utcOffsetMinutes': 0, + }), }, 'videoId': video_id, } - headers = { - 'X-YouTube-Client-Name': '2', - 'X-YouTube-Client-Version': '2.20241202.07.00', - 'Origin': origin, + + headers = merge_dicts({ 'Sec-Fetch-Mode': 'navigate', - 'User-Agent': query['context']['client']['userAgent'], - } + 'Origin': origin, + # 'X-Goog-Visitor-Id': self._extract_visitor_data(ytcfg) or '', + }, traverse_obj(client, { + 'X-YouTube-Client-Name': 'INNERTUBE_CONTEXT_CLIENT_NAME', + 'X-YouTube-Client-Version': ( + 'INNERTUBE_CONTEXT', 'client', 'clientVersion'), + 'User-Agent': ( + 'INNERTUBE_CONTEXT', 'client', 'userAgent'), + })) + auth = self._generate_sapisidhash_header(origin) if auth is not None: headers['Authorization'] = auth headers['X-Origin'] = origin - player_response = self._call_api('player', query, video_id, fatal=False, headers=headers) + player_response = self._call_api( + 'player', query, video_id, fatal=False, headers=headers, + note=join_nonempty( + 'Downloading', traverse_obj(query, ( + 'context', 'client', 'clientName')), + 'API JSON', delim=' ')) def is_agegated(playability): if not isinstance(playability, dict): From 55ad8a24cacee03a91fe70d8d48aa9a02cc0ab11 Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 15 Jan 2025 03:22:56 +0000 Subject: [PATCH 25/78] [YouTube] Support `... /feeds/videos.xml?playlist_id={pl_id}` --- youtube_dl/extractor/youtube.py | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 1424277ac..f0406b357 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -3601,10 +3601,23 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): def _real_extract(self, url): item_id = self._match_id(url) url = update_url(url, netloc='www.youtube.com') - # Handle both video/playlist URLs qs = parse_qs(url) - video_id = qs.get('v', [None])[0] - playlist_id = qs.get('list', [None])[0] + + def qs_get(key, default=None): + return qs.get(key, [default])[-1] + + # Go around for /feeds/videos.xml?playlist_id={pl_id} + if item_id == 'feeds' and '/feeds/videos.xml?' in url: + playlist_id = qs_get('playlist_id') + if playlist_id: + return self.url_result( + update_url_query('https://www.youtube.com/playlist', { + 'list': playlist_id, + }), ie=self.ie_key(), video_id=playlist_id) + + # Handle both video/playlist URLs + video_id = qs_get('v') + playlist_id = qs_get('list') if video_id and playlist_id: if self._downloader.params.get('noplaylist'): self.to_screen('Downloading just video %s because of --no-playlist' % video_id) From b09442a2f4a8d255569abf0bb6b4867c53d0c2e9 Mon Sep 17 00:00:00 2001 From: dirkf Date: Sun, 19 Jan 2025 01:18:34 +0000 Subject: [PATCH 26/78] [YouTube] Also use ios client when is_live --- youtube_dl/extractor/youtube.py | 118 ++++++++++++++++++++++---------- 1 file changed, 81 insertions(+), 37 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index f0406b357..32e836d49 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -85,6 +85,22 @@ class YoutubeBaseInfoExtractor(InfoExtractor): _PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM)' _INNERTUBE_CLIENTS = { + 'ios': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'IOS', + 'clientVersion': '19.45.4', + 'deviceMake': 'Apple', + 'deviceModel': 'iPhone16,2', + 'userAgent': 'com.google.ios.youtube/19.45.4 (iPhone16,2; U; CPU iOS 18_1_0 like Mac OS X;)', + 'osName': 'iPhone', + 'osVersion': '18.1.0.22B83', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 5, + 'REQUIRE_JS_PLAYER': False, + 'REQUIRE_PO_TOKEN': True, + }, # mweb has 'ultralow' formats # See: https://github.com/yt-dlp/yt-dlp/pull/557 'mweb': { @@ -110,6 +126,17 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'INNERTUBE_CONTEXT_CLIENT_NAME': 7, 'SUPPORTS_COOKIES': True, }, + 'web': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'WEB', + 'clientVersion': '2.20241126.01.00', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 1, + 'REQUIRE_PO_TOKEN': True, + 'SUPPORTS_COOKIES': True, + }, } def _login(self): @@ -1995,6 +2022,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): player_response = self._extract_yt_initial_variable( webpage, self._YT_INITIAL_PLAYER_RESPONSE_RE, video_id, 'initial player response') + is_live = traverse_obj(player_response, ('videoDetails', 'isLive')) + if False and not player_response: player_response = self._call_api( 'player', {'videoId': video_id}, video_id) @@ -2008,50 +2037,65 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if sts: pb_context['signatureTimestamp'] = sts - client = traverse_obj(self._INNERTUBE_CLIENTS, ( - lambda _, v: not v.get('REQUIRE_PO_TOKEN')), - get_all=False) + client_names = traverse_obj(self._INNERTUBE_CLIENTS, ( + T(dict.items), lambda _, k_v: not k_v[1].get('REQUIRE_PO_TOKEN'), + 0))[:1] - query = { - 'playbackContext': { - 'contentPlaybackContext': pb_context, - 'contentCheckOk': True, - 'racyCheckOk': True, - }, - 'context': { - 'client': merge_dicts( - traverse_obj(client, ('INNERTUBE_CONTEXT', 'client')), { - 'hl': 'en', - 'timeZone': 'UTC', - 'utcOffsetMinutes': 0, - }), - }, - 'videoId': video_id, - } + if is_live and 'ios' not in client_names: + client_names.append('ios') - headers = merge_dicts({ + headers = { 'Sec-Fetch-Mode': 'navigate', 'Origin': origin, # 'X-Goog-Visitor-Id': self._extract_visitor_data(ytcfg) or '', - }, traverse_obj(client, { - 'X-YouTube-Client-Name': 'INNERTUBE_CONTEXT_CLIENT_NAME', - 'X-YouTube-Client-Version': ( - 'INNERTUBE_CONTEXT', 'client', 'clientVersion'), - 'User-Agent': ( - 'INNERTUBE_CONTEXT', 'client', 'userAgent'), - })) - + } auth = self._generate_sapisidhash_header(origin) if auth is not None: headers['Authorization'] = auth headers['X-Origin'] = origin - player_response = self._call_api( - 'player', query, video_id, fatal=False, headers=headers, - note=join_nonempty( - 'Downloading', traverse_obj(query, ( - 'context', 'client', 'clientName')), - 'API JSON', delim=' ')) + for client in traverse_obj(self._INNERTUBE_CLIENTS, (client_names, T(dict))): + + query = { + 'playbackContext': { + 'contentPlaybackContext': pb_context, + 'contentCheckOk': True, + 'racyCheckOk': True, + }, + 'context': { + 'client': merge_dicts( + traverse_obj(client, ('INNERTUBE_CONTEXT', 'client')), { + 'hl': 'en', + 'timeZone': 'UTC', + 'utcOffsetMinutes': 0, + }), + }, + 'videoId': video_id, + } + + api_headers = merge_dicts(headers, traverse_obj(client, { + 'X-YouTube-Client-Name': 'INNERTUBE_CONTEXT_CLIENT_NAME', + 'X-YouTube-Client-Version': ( + 'INNERTUBE_CONTEXT', 'client', 'clientVersion'), + 'User-Agent': ( + 'INNERTUBE_CONTEXT', 'client', 'userAgent'), + })) + + api_player_response = self._call_api( + 'player', query, video_id, fatal=False, headers=api_headers, + note=join_nonempty( + 'Downloading', traverse_obj(query, ( + 'context', 'client', 'clientName')), + 'API JSON', delim=' ')) + + hls = [ + traverse_obj( + resp, ('streamingData', 'hlsManifestUrl', T(url_or_none))) + for resp in (player_response, api_player_response)] + if not hls[0] and hls[1]: + player_response['streamingData']['hlsManifestUrl'] = hls[1] + else: + player_response.update(api_player_response or {}) def is_agegated(playability): if not isinstance(playability, dict): @@ -2194,6 +2238,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): itag_qualities = {} q = qualities(['tiny', 'small', 'medium', 'large', 'hd720', 'hd1080', 'hd1440', 'hd2160', 'hd2880', 'highres']) CHUNK_SIZE = 10 << 20 + is_live = video_details.get('isLive') streaming_data = player_response.get('streamingData') or {} streaming_formats = streaming_data.get('formats') or [] @@ -2338,7 +2383,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): hls_manifest_url = streaming_data.get('hlsManifestUrl') if hls_manifest_url: for f in self._extract_m3u8_formats( - hls_manifest_url, video_id, 'mp4', fatal=False): + hls_manifest_url, video_id, 'mp4', + entry_protocol='m3u8_native', live=is_live, fatal=False): if process_manifest_format( f, 'hls', None, self._search_regex( r'/itag/(\d+)', f['url'], 'itag', default=None)): @@ -2444,8 +2490,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # Strictly de-prioritize damaged formats f['preference'] = -10 - is_live = video_details.get('isLive') - owner_profile_url = self._yt_urljoin(self._extract_author_var( webpage, 'url', videodetails=video_details, metadata=microformat)) From 63fb0fc4159397618b12fa115f957b9ba70f3f88 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 20 Jan 2025 13:23:54 +0000 Subject: [PATCH 27/78] [YouTube] Retain .videoDetails members from all player responses --- youtube_dl/extractor/youtube.py | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 32e836d49..edaae5bd3 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -2095,7 +2095,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if not hls[0] and hls[1]: player_response['streamingData']['hlsManifestUrl'] = hls[1] else: + video_details = merge_dicts(*traverse_obj( + (player_response, api_player_response), + (Ellipsis, 'videoDetails', T(dict)))) player_response.update(api_player_response or {}) + player_response['videoDetails'] = video_details def is_agegated(playability): if not isinstance(playability, dict): From 5975d7bb96095fae7c35e7cfcd819255a5b57087 Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 22 Jan 2025 06:52:40 +0000 Subject: [PATCH 28/78] [YouTube] Use X-Goog-Visitor-Id * required with tv player client * resolves #33030 --- youtube_dl/extractor/youtube.py | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index edaae5bd3..c93a2a1f9 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -142,6 +142,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): def _login(self): """ Attempt to log in to YouTube. + True is returned if successful or skipped. False is returned if login failed. @@ -2040,6 +2041,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): client_names = traverse_obj(self._INNERTUBE_CLIENTS, ( T(dict.items), lambda _, k_v: not k_v[1].get('REQUIRE_PO_TOKEN'), 0))[:1] + if 'web' not in client_names: + # webpage links won't download: ignore links and playability + player_response = filter_dict( + player_response or {}, + lambda k, _: k not in ('streamingData', 'playabilityStatus')) if is_live and 'ios' not in client_names: client_names.append('ios') @@ -2047,7 +2053,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): headers = { 'Sec-Fetch-Mode': 'navigate', 'Origin': origin, - # 'X-Goog-Visitor-Id': self._extract_visitor_data(ytcfg) or '', + 'X-Goog-Visitor-Id': self._extract_visitor_data(ytcfg) or '', } auth = self._generate_sapisidhash_header(origin) if auth is not None: @@ -2059,9 +2065,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): query = { 'playbackContext': { 'contentPlaybackContext': pb_context, - 'contentCheckOk': True, - 'racyCheckOk': True, }, + 'contentCheckOk': True, + 'racyCheckOk': True, 'context': { 'client': merge_dicts( traverse_obj(client, ('INNERTUBE_CONTEXT', 'client')), { @@ -2088,11 +2094,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'context', 'client', 'clientName')), 'API JSON', delim=' ')) - hls = [ - traverse_obj( - resp, ('streamingData', 'hlsManifestUrl', T(url_or_none))) - for resp in (player_response, api_player_response)] - if not hls[0] and hls[1]: + hls = traverse_obj( + (player_response, api_player_response), + (Ellipsis, 'streamingData', 'hlsManifestUrl', T(url_or_none))) + if len(hls) == 2 and not hls[0] and hls[1]: player_response['streamingData']['hlsManifestUrl'] = hls[1] else: video_details = merge_dicts(*traverse_obj( @@ -3467,7 +3472,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): if not continuation: break if visitor_data: - headers['x-goog-visitor-id'] = visitor_data + headers['X-Goog-Visitor-Id'] = visitor_data data['continuation'] = continuation['continuation'] data['clickTracking'] = { 'clickTrackingParams': continuation['itct'], From 26b6f15d14e30dc3349abe2933edb051b46be4b8 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 3 Feb 2025 18:10:15 +0000 Subject: [PATCH 29/78] [compat] Make casefold private * if required, not supported: `from youtube_dl.casefold import _casefold as casefold` --- youtube_dl/casefold.py | 12 ++++-------- youtube_dl/compat.py | 2 +- 2 files changed, 5 insertions(+), 9 deletions(-) diff --git a/youtube_dl/casefold.py b/youtube_dl/casefold.py index ad9c66f8e..712b2e7fa 100644 --- a/youtube_dl/casefold.py +++ b/youtube_dl/casefold.py @@ -10,9 +10,10 @@ from .compat import ( # https://github.com/unicode-org/icu/blob/main/icu4c/source/data/unidata/CaseFolding.txt # In case newly foldable Unicode characters are defined, paste the new version # of the text inside the ''' marks. -# The text is expected to have only blank lines andlines with 1st character #, +# The text is expected to have only blank lines and lines with 1st character #, # all ignored, and fold definitions like this: -# `from_hex_code; space_separated_to_hex_code_list; comment` +# `from_hex_code; status; space_separated_to_hex_code_list; comment` +# Only `status` C/F are used. _map_str = ''' # CaseFolding-15.0.0.txt @@ -1657,11 +1658,6 @@ _map = dict( del _map_str -def casefold(s): +def _casefold(s): assert isinstance(s, compat_str) return ''.join((_map.get(c, c) for c in s)) - - -__all__ = [ - 'casefold', -] diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index ed1a33cf2..c621f7476 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -44,7 +44,7 @@ try: compat_str.casefold compat_casefold = lambda s: s.casefold() except AttributeError: - from .casefold import casefold as compat_casefold + from .casefold import _casefold as compat_casefold try: import collections.abc as compat_collections_abc From 711e72c292327674c4a0593fdbb83d6347738ec9 Mon Sep 17 00:00:00 2001 From: dirkf Date: Thu, 6 Feb 2025 21:09:00 +0000 Subject: [PATCH 30/78] [JSInterp] Fix bit-shift coercion for player 9c6dfc4a --- test/test_jsinterp.py | 4 ++++ test/test_youtube_signature.py | 4 ++++ youtube_dl/compat.py | 5 +++++ youtube_dl/jsinterp.py | 27 +++++++++++++++++++++------ 4 files changed, 34 insertions(+), 6 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 12e7b9b94..6c34bc896 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -459,6 +459,10 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return undefined >> 5}', 0) self._test('function f(){return 42 << NaN}', 42) self._test('function f(){return 42 << Infinity}', 42) + self._test('function f(){return 0.0 << null}', 0) + self._test('function f(){return NaN << 42}', 0) + self._test('function f(){return "21.9" << 1}', 42) + self._test('function f(){return 21 << 4294967297}', 42) def test_negative(self): self._test('function f(){return 2 * -2.0 ;}', -4) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index fcbc9d7a8..67ef75fde 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -219,6 +219,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/2f1832d2/player_ias.vflset/en_US/base.js', 'YWt1qdbe8SAfkoPHW5d', 'RrRjWQOJmBiP', ), + ( + 'https://www.youtube.com/s/player/9c6dfc4a/player_ias.vflset/en_US/base.js', + 'jbu7ylIosQHyJyJV', 'uwI0ESiynAmhNg', + ), ] diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index c621f7476..26b655fb6 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -3116,17 +3116,21 @@ else: compat_kwargs = lambda kwargs: kwargs +# compat_numeric_types try: compat_numeric_types = (int, float, long, complex) except NameError: # Python 3 compat_numeric_types = (int, float, complex) +# compat_integer_types try: compat_integer_types = (int, long) except NameError: # Python 3 compat_integer_types = (int, ) +# compat_int +compat_int = compat_integer_types[-1] if sys.version_info < (2, 7): def compat_socket_create_connection(address, timeout, source_address=None): @@ -3532,6 +3536,7 @@ __all__ = [ 'compat_http_client', 'compat_http_server', 'compat_input', + 'compat_int', 'compat_integer_types', 'compat_itertools_count', 'compat_itertools_zip_longest', diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 7835187f5..2859bc734 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -24,6 +24,8 @@ from .compat import ( compat_collections_chain_map as ChainMap, compat_contextlib_suppress, compat_filter as filter, + compat_int, + compat_integer_types, compat_itertools_zip_longest as zip_longest, compat_map as map, compat_numeric_types, @@ -70,14 +72,27 @@ class JS_Undefined(object): pass -def _js_bit_op(op): +def _js_bit_op(op, is_shift=False): - def zeroise(x): - return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x + def zeroise(x, is_shift_arg=False): + if isinstance(x, compat_integer_types): + return (x % 32) if is_shift_arg else (x & 0xffffffff) + try: + x = float(x) + if is_shift_arg: + x = int(x % 32) + elif x < 0: + x = -compat_int(-x % 0xffffffff) + else: + x = compat_int(x % 0xffffffff) + except (ValueError, TypeError): + # also here for int(NaN), including float('inf') % 32 + x = 0 + return x @wraps_op(op) def wrapped(a, b): - return op(zeroise(a), zeroise(b)) & 0xffffffff + return op(zeroise(a), zeroise(b, is_shift)) & 0xffffffff return wrapped @@ -253,8 +268,8 @@ def _js_typeof(expr): # avoid dict to maintain order # definition None => Defined in JSInterpreter._operator _OPERATORS = ( - ('>>', _js_bit_op(operator.rshift)), - ('<<', _js_bit_op(operator.lshift)), + ('>>', _js_bit_op(operator.rshift, True)), + ('<<', _js_bit_op(operator.lshift, True)), ('+', _js_add), ('-', _js_arith_op(operator.sub)), ('*', _js_arith_op(operator.mul)), From 91b1569f68471d685382b738806b2e07d8f52707 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 28 Feb 2025 00:02:10 +0000 Subject: [PATCH 31/78] [YouTube] Fix channel playlist extraction (#33074) * [YouTube] Extract playlist items from LOCKUP_VIEW_MODEL_... * resolves #33073 * thx seproDev (yt-dlp/yt-dlp#11615) Co-authored-by: sepro --- youtube_dl/extractor/youtube.py | 49 +++++++++++++++++++++++++++++++-- 1 file changed, 47 insertions(+), 2 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index c93a2a1f9..cc84a193a 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -27,6 +27,7 @@ from ..compat import ( ) from ..jsinterp import JSInterpreter from ..utils import ( + bug_reports_message, clean_html, dict_get, error_to_compat_str, @@ -65,6 +66,7 @@ from ..utils import ( url_or_none, urlencode_postdata, urljoin, + variadic, ) @@ -460,6 +462,26 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'uploader': uploader, } + @staticmethod + def _extract_thumbnails(data, *path_list, **kw_final_key): + """ + Extract thumbnails from thumbnails dict + @param path_list: path list to level that contains 'thumbnails' key + """ + final_key = kw_final_key.get('final_key', 'thumbnails') + + return traverse_obj(data, (( + tuple(variadic(path) + (final_key, Ellipsis) + for path in path_list or [()])), { + 'url': ('url', T(url_or_none), + # Sometimes youtube gives a wrong thumbnail URL. See: + # https://github.com/yt-dlp/yt-dlp/issues/233 + # https://github.com/ytdl-org/youtube-dl/issues/28023 + T(lambda u: update_url(u, query=None) if u and 'maxresdefault' in u else u)), + 'height': ('height', T(int_or_none)), + 'width': ('width', T(int_or_none)), + }, T(lambda t: t if t.get('url') else None))) + def _search_results(self, query, params): data = { 'context': { @@ -3183,8 +3205,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): expected_type=txt_or_none) def _grid_entries(self, grid_renderer): - for item in grid_renderer['items']: - if not isinstance(item, dict): + for item in traverse_obj(grid_renderer, ('items', Ellipsis, T(dict))): + lockup_view_model = traverse_obj(item, ('lockupViewModel', T(dict))) + if lockup_view_model: + entry = self._extract_lockup_view_model(lockup_view_model) + if entry: + yield entry continue renderer = self._extract_grid_item_renderer(item) if not isinstance(renderer, dict): @@ -3268,6 +3294,25 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): continue yield self._extract_video(renderer) + def _extract_lockup_view_model(self, view_model): + content_id = view_model.get('contentId') + if not content_id: + return + content_type = view_model.get('contentType') + if content_type not in ('LOCKUP_CONTENT_TYPE_PLAYLIST', 'LOCKUP_CONTENT_TYPE_PODCAST'): + self.report_warning( + 'Unsupported lockup view model content type "{0}"{1}'.format(content_type, bug_reports_message()), only_once=True) + return + return merge_dicts(self.url_result( + update_url_query('https://www.youtube.com/playlist', {'list': content_id}), + ie=YoutubeTabIE, video_id=content_id), { + 'title': traverse_obj(view_model, ( + 'metadata', 'lockupMetadataViewModel', 'title', 'content', T(compat_str))), + 'thumbnails': self._extract_thumbnails(view_model, ( + 'contentImage', 'collectionThumbnailViewModel', 'primaryThumbnail', + 'thumbnailViewModel', 'image'), final_key='sources'), + }) + def _video_entry(self, video_renderer): video_id = video_renderer.get('videoId') if video_id: From 673277e510ebd996b62a2fcc76169bf3cce29910 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 28 Feb 2025 01:02:20 +0000 Subject: [PATCH 32/78] [YouTube] Fix 91b1569 --- youtube_dl/extractor/youtube.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index cc84a193a..5f8c08201 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -3305,7 +3305,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): return return merge_dicts(self.url_result( update_url_query('https://www.youtube.com/playlist', {'list': content_id}), - ie=YoutubeTabIE, video_id=content_id), { + ie=YoutubeTabIE.ie_key(), video_id=content_id), { 'title': traverse_obj(view_model, ( 'metadata', 'lockupMetadataViewModel', 'title', 'content', T(compat_str))), 'thumbnails': self._extract_thumbnails(view_model, ( From cecaa18b80e33323193915ef9fbd2f68d94d7bce Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 23:03:17 +0000 Subject: [PATCH 33/78] [compat] Clean-up * make workaround_optparse_bug9161 private * add comments * avoid leaving test objects behind --- youtube_dl/__init__.py | 4 +- youtube_dl/compat.py | 172 ++++++++++++++++++++++++++++++----------- 2 files changed, 129 insertions(+), 47 deletions(-) diff --git a/youtube_dl/__init__.py b/youtube_dl/__init__.py index 06bdfb689..3c1272e7b 100644 --- a/youtube_dl/__init__.py +++ b/youtube_dl/__init__.py @@ -18,7 +18,7 @@ from .compat import ( compat_getpass, compat_register_utf8, compat_shlex_split, - workaround_optparse_bug9161, + _workaround_optparse_bug9161, ) from .utils import ( _UnsafeExtensionError, @@ -50,7 +50,7 @@ def _real_main(argv=None): # Compatibility fix for Windows compat_register_utf8() - workaround_optparse_bug9161() + _workaround_optparse_bug9161() setproctitle('youtube-dl') diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index 26b655fb6..e617dd511 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -16,7 +16,6 @@ import os import platform import re import shlex -import shutil import socket import struct import subprocess @@ -24,11 +23,15 @@ import sys import types import xml.etree.ElementTree +_IDENTITY = lambda x: x + # naming convention # 'compat_' + Python3_name.replace('.', '_') # other aliases exist for convenience and/or legacy +# wrap disposable test values in type() to reclaim storage -# deal with critical unicode/str things first +# deal with critical unicode/str things first: +# compat_str, compat_basestring, compat_chr try: # Python 2 compat_str, compat_basestring, compat_chr = ( @@ -39,18 +42,23 @@ except NameError: str, (str, bytes), chr ) -# casefold + +# compat_casefold try: compat_str.casefold compat_casefold = lambda s: s.casefold() except AttributeError: from .casefold import _casefold as compat_casefold + +# compat_collections_abc try: import collections.abc as compat_collections_abc except ImportError: import collections as compat_collections_abc + +# compat_urllib_request try: import urllib.request as compat_urllib_request except ImportError: # Python 2 @@ -79,11 +87,15 @@ except TypeError: _add_init_method_arg(compat_urllib_request.Request) del _add_init_method_arg + +# compat_urllib_error try: import urllib.error as compat_urllib_error except ImportError: # Python 2 import urllib2 as compat_urllib_error + +# compat_urllib_parse try: import urllib.parse as compat_urllib_parse except ImportError: # Python 2 @@ -98,17 +110,23 @@ except ImportError: # Python 2 compat_urlparse = compat_urllib_parse compat_urllib_parse_urlparse = compat_urllib_parse.urlparse + +# compat_urllib_response try: import urllib.response as compat_urllib_response except ImportError: # Python 2 import urllib as compat_urllib_response + +# compat_urllib_response.addinfourl try: compat_urllib_response.addinfourl.status except AttributeError: # .getcode() is deprecated in Py 3. compat_urllib_response.addinfourl.status = property(lambda self: self.getcode()) + +# compat_http_cookiejar try: import http.cookiejar as compat_cookiejar except ImportError: # Python 2 @@ -127,12 +145,16 @@ else: compat_cookiejar_Cookie = compat_cookiejar.Cookie compat_http_cookiejar_Cookie = compat_cookiejar_Cookie + +# compat_http_cookies try: import http.cookies as compat_cookies except ImportError: # Python 2 import Cookie as compat_cookies compat_http_cookies = compat_cookies + +# compat_http_cookies_SimpleCookie if sys.version_info[0] == 2 or sys.version_info < (3, 3): class compat_cookies_SimpleCookie(compat_cookies.SimpleCookie): def load(self, rawdata): @@ -155,11 +177,15 @@ else: compat_cookies_SimpleCookie = compat_cookies.SimpleCookie compat_http_cookies_SimpleCookie = compat_cookies_SimpleCookie + +# compat_html_entities, probably useless now try: import html.entities as compat_html_entities except ImportError: # Python 2 import htmlentitydefs as compat_html_entities + +# compat_html_entities_html5 try: # Python >= 3.3 compat_html_entities_html5 = compat_html_entities.html5 except AttributeError: @@ -2408,18 +2434,24 @@ except AttributeError: # Py < 3.1 compat_http_client.HTTPResponse.getcode = lambda self: self.status + +# compat_urllib_HTTPError try: from urllib.error import HTTPError as compat_HTTPError except ImportError: # Python 2 from urllib2 import HTTPError as compat_HTTPError compat_urllib_HTTPError = compat_HTTPError + +# compat_urllib_request_urlretrieve try: from urllib.request import urlretrieve as compat_urlretrieve except ImportError: # Python 2 from urllib import urlretrieve as compat_urlretrieve compat_urllib_request_urlretrieve = compat_urlretrieve + +# compat_html_parser_HTMLParser, compat_html_parser_HTMLParseError try: from HTMLParser import ( HTMLParser as compat_HTMLParser, @@ -2432,22 +2464,33 @@ except ImportError: # Python 3 # HTMLParseError was deprecated in Python 3.3 and removed in # Python 3.5. Introducing dummy exception for Python >3.5 for compatible # and uniform cross-version exception handling + class compat_HTMLParseError(Exception): pass + compat_html_parser_HTMLParser = compat_HTMLParser compat_html_parser_HTMLParseError = compat_HTMLParseError + +# compat_subprocess_get_DEVNULL try: _DEVNULL = subprocess.DEVNULL compat_subprocess_get_DEVNULL = lambda: _DEVNULL except AttributeError: compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w') + +# compat_http_server try: import http.server as compat_http_server except ImportError: import BaseHTTPServer as compat_http_server + +# compat_urllib_parse_unquote_to_bytes, +# compat_urllib_parse_unquote, compat_urllib_parse_unquote_plus, +# compat_urllib_parse_urlencode, +# compat_urllib_parse_parse_qs try: from urllib.parse import unquote_to_bytes as compat_urllib_parse_unquote_to_bytes from urllib.parse import unquote as compat_urllib_parse_unquote @@ -2598,6 +2641,8 @@ except ImportError: # Python 2 compat_urllib_parse_parse_qs = compat_parse_qs + +# compat_urllib_request_DataHandler try: from urllib.request import DataHandler as compat_urllib_request_DataHandler except ImportError: # Python < 3.4 @@ -2632,16 +2677,20 @@ except ImportError: # Python < 3.4 return compat_urllib_response.addinfourl(io.BytesIO(data), headers, url) + +# compat_xml_etree_ElementTree_ParseError try: from xml.etree.ElementTree import ParseError as compat_xml_parse_error except ImportError: # Python 2.6 from xml.parsers.expat import ExpatError as compat_xml_parse_error compat_xml_etree_ElementTree_ParseError = compat_xml_parse_error -etree = xml.etree.ElementTree + +# compat_xml_etree_ElementTree_Element +_etree = xml.etree.ElementTree -class _TreeBuilder(etree.TreeBuilder): +class _TreeBuilder(_etree.TreeBuilder): def doctype(self, name, pubid, system): pass @@ -2650,7 +2699,7 @@ try: # xml.etree.ElementTree.Element is a method in Python <=2.6 and # the following will crash with: # TypeError: isinstance() arg 2 must be a class, type, or tuple of classes and types - isinstance(None, etree.Element) + isinstance(None, _etree.Element) from xml.etree.ElementTree import Element as compat_etree_Element except TypeError: # Python <=2.6 from xml.etree.ElementTree import _ElementInterface as compat_etree_Element @@ -2658,12 +2707,12 @@ compat_xml_etree_ElementTree_Element = compat_etree_Element if sys.version_info[0] >= 3: def compat_etree_fromstring(text): - return etree.XML(text, parser=etree.XMLParser(target=_TreeBuilder())) + return _etree.XML(text, parser=_etree.XMLParser(target=_TreeBuilder())) else: # python 2.x tries to encode unicode strings with ascii (see the # XMLParser._fixtext method) try: - _etree_iter = etree.Element.iter + _etree_iter = _etree.Element.iter except AttributeError: # Python <=2.6 def _etree_iter(root): for el in root.findall('*'): @@ -2675,27 +2724,29 @@ else: # 2.7 source def _XML(text, parser=None): if not parser: - parser = etree.XMLParser(target=_TreeBuilder()) + parser = _etree.XMLParser(target=_TreeBuilder()) parser.feed(text) return parser.close() def _element_factory(*args, **kwargs): - el = etree.Element(*args, **kwargs) + el = _etree.Element(*args, **kwargs) for k, v in el.items(): if isinstance(v, bytes): el.set(k, v.decode('utf-8')) return el def compat_etree_fromstring(text): - doc = _XML(text, parser=etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory))) + doc = _XML(text, parser=_etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory))) for el in _etree_iter(doc): if el.text is not None and isinstance(el.text, bytes): el.text = el.text.decode('utf-8') return doc -if hasattr(etree, 'register_namespace'): - compat_etree_register_namespace = etree.register_namespace -else: + +# compat_xml_etree_register_namespace +try: + compat_etree_register_namespace = _etree.register_namespace +except AttributeError: def compat_etree_register_namespace(prefix, uri): """Register a namespace prefix. The registry is global, and any existing mapping for either the @@ -2704,14 +2755,16 @@ else: attributes in this namespace will be serialized with prefix if possible. ValueError is raised if prefix is reserved or is invalid. """ - if re.match(r"ns\d+$", prefix): - raise ValueError("Prefix format reserved for internal use") - for k, v in list(etree._namespace_map.items()): + if re.match(r'ns\d+$', prefix): + raise ValueError('Prefix format reserved for internal use') + for k, v in list(_etree._namespace_map.items()): if k == uri or v == prefix: - del etree._namespace_map[k] - etree._namespace_map[uri] = prefix + del _etree._namespace_map[k] + _etree._namespace_map[uri] = prefix compat_xml_etree_register_namespace = compat_etree_register_namespace + +# compat_xpath, compat_etree_iterfind if sys.version_info < (2, 7): # Here comes the crazy part: In 2.6, if the xpath is a unicode, # .//node does not match if a node is a direct child of . ! @@ -2898,7 +2951,6 @@ if sys.version_info < (2, 7): def __init__(self, root): self.root = root - ## # Generate all matching objects. def compat_etree_iterfind(elem, path, namespaces=None): @@ -2933,13 +2985,15 @@ if sys.version_info < (2, 7): else: - compat_xpath = lambda xpath: xpath compat_etree_iterfind = lambda element, match: element.iterfind(match) + compat_xpath = _IDENTITY +# compat_os_name compat_os_name = os._name if os.name == 'java' else os.name +# compat_shlex_quote if compat_os_name == 'nt': def compat_shlex_quote(s): return s if re.match(r'^[-_\w./]+$', s) else '"%s"' % s.replace('"', '\\"') @@ -2954,6 +3008,7 @@ else: return "'" + s.replace("'", "'\"'\"'") + "'" +# compat_shlex.split try: args = shlex.split('中文') assert (isinstance(args, list) @@ -2969,6 +3024,7 @@ except (AssertionError, UnicodeEncodeError): return list(map(lambda s: s.decode('utf-8'), shlex.split(s, comments, posix))) +# compat_ord def compat_ord(c): if isinstance(c, int): return c @@ -2976,6 +3032,7 @@ def compat_ord(c): return ord(c) +# compat_getenv, compat_os_path_expanduser, compat_setenv if sys.version_info >= (3, 0): compat_getenv = os.getenv compat_expanduser = os.path.expanduser @@ -3063,6 +3120,7 @@ else: compat_os_path_expanduser = compat_expanduser +# compat_os_path_realpath if compat_os_name == 'nt' and sys.version_info < (3, 8): # os.path.realpath on Windows does not follow symbolic links # prior to Python 3.8 (see https://bugs.python.org/issue9949) @@ -3076,6 +3134,7 @@ else: compat_os_path_realpath = compat_realpath +# compat_print if sys.version_info < (3, 0): def compat_print(s): from .utils import preferredencoding @@ -3086,6 +3145,7 @@ else: print(s) +# compat_getpass_getpass if sys.version_info < (3, 0) and sys.platform == 'win32': def compat_getpass(prompt, *args, **kwargs): if isinstance(prompt, compat_str): @@ -3098,22 +3158,22 @@ else: compat_getpass_getpass = compat_getpass +# compat_input try: compat_input = raw_input except NameError: # Python 3 compat_input = input +# compat_kwargs # Python < 2.6.5 require kwargs to be bytes try: - def _testfunc(x): - pass - _testfunc(**{'x': 0}) + (lambda x: x)(**{'x': 0}) except TypeError: def compat_kwargs(kwargs): return dict((bytes(k), v) for k, v in kwargs.items()) else: - compat_kwargs = lambda kwargs: kwargs + compat_kwargs = _IDENTITY # compat_numeric_types @@ -3132,6 +3192,8 @@ except NameError: # Python 3 # compat_int compat_int = compat_integer_types[-1] + +# compat_socket_create_connection if sys.version_info < (2, 7): def compat_socket_create_connection(address, timeout, source_address=None): host, port = address @@ -3158,6 +3220,7 @@ else: compat_socket_create_connection = socket.create_connection +# compat_contextlib_suppress try: from contextlib import suppress as compat_contextlib_suppress except ImportError: @@ -3200,12 +3263,12 @@ except AttributeError: # repeated .close() is OK, but just in case with compat_contextlib_suppress(EnvironmentError): f.close() - popen.wait() + popen.wait() # Fix https://github.com/ytdl-org/youtube-dl/issues/4223 # See http://bugs.python.org/issue9161 for what is broken -def workaround_optparse_bug9161(): +def _workaround_optparse_bug9161(): op = optparse.OptionParser() og = optparse.OptionGroup(op, 'foo') try: @@ -3224,9 +3287,10 @@ def workaround_optparse_bug9161(): optparse.OptionGroup.add_option = _compat_add_option -if hasattr(shutil, 'get_terminal_size'): # Python >= 3.3 - compat_get_terminal_size = shutil.get_terminal_size -else: +# compat_shutil_get_terminal_size +try: + from shutil import get_terminal_size as compat_get_terminal_size # Python >= 3.3 +except ImportError: _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines']) def compat_get_terminal_size(fallback=(80, 24)): @@ -3256,27 +3320,33 @@ else: columns = _columns if lines is None or lines <= 0: lines = _lines + return _terminal_size(columns, lines) +compat_shutil_get_terminal_size = compat_get_terminal_size + +# compat_itertools_count try: - itertools.count(start=0, step=1) + type(itertools.count(start=0, step=1)) compat_itertools_count = itertools.count -except TypeError: # Python 2.6 +except TypeError: # Python 2.6 lacks step def compat_itertools_count(start=0, step=1): while True: yield start start += step +# compat_tokenize_tokenize if sys.version_info >= (3, 0): from tokenize import tokenize as compat_tokenize_tokenize else: from tokenize import generate_tokens as compat_tokenize_tokenize +# compat_struct_pack, compat_struct_unpack, compat_Struct try: - struct.pack('!I', 0) + type(struct.pack('!I', 0)) except TypeError: # In Python 2.6 and 2.7.x < 2.7.7, struct requires a bytes argument # See https://bugs.python.org/issue19099 @@ -3308,8 +3378,10 @@ else: compat_Struct = struct.Struct -# compat_map/filter() returning an iterator, supposedly the -# same versioning as for zip below +# builtins returning an iterator + +# compat_map, compat_filter +# supposedly the same versioning as for zip below try: from future_builtins import map as compat_map except ImportError: @@ -3326,6 +3398,7 @@ except ImportError: except ImportError: compat_filter = filter +# compat_zip try: from future_builtins import zip as compat_zip except ImportError: # not 2.6+ or is 3.x @@ -3335,6 +3408,7 @@ except ImportError: # not 2.6+ or is 3.x compat_zip = zip +# compat_itertools_zip_longest # method renamed between Py2/3 try: from itertools import zip_longest as compat_itertools_zip_longest @@ -3342,7 +3416,8 @@ except ImportError: from itertools import izip_longest as compat_itertools_zip_longest -# new class in collections +# compat_collections_chain_map +# collections.ChainMap: new class try: from collections import ChainMap as compat_collections_chain_map # Py3.3's ChainMap is deficient @@ -3405,12 +3480,14 @@ except ImportError: return compat_collections_chain_map(*(self.maps[1:])) +# compat_re_Pattern, compat_re_Match # Pythons disagree on the type of a pattern (RegexObject, _sre.SRE_Pattern, Pattern, ...?) compat_re_Pattern = type(re.compile('')) # and on the type of a match compat_re_Match = type(re.match('a', 'a')) +# compat_base64_b64decode if sys.version_info < (3, 3): def compat_b64decode(s, *args, **kwargs): if isinstance(s, compat_str): @@ -3422,6 +3499,7 @@ else: compat_base64_b64decode = compat_b64decode +# compat_ctypes_WINFUNCTYPE if platform.python_implementation() == 'PyPy' and sys.pypy_version_info < (5, 4, 0): # PyPy2 prior to version 5.4.0 expects byte strings as Windows function # names, see the original PyPy issue [1] and the youtube-dl one [2]. @@ -3440,6 +3518,7 @@ else: return ctypes.WINFUNCTYPE(*args, **kwargs) +# compat_open if sys.version_info < (3, 0): # open(file, mode='r', buffering=- 1, encoding=None, errors=None, newline=None, closefd=True) not: opener=None def compat_open(file_, *args, **kwargs): @@ -3467,12 +3546,15 @@ except AttributeError: def compat_datetime_timedelta_total_seconds(td): return (td.microseconds + (td.seconds + td.days * 24 * 3600) * 10**6) / 10**6 + # optional decompression packages +# compat_brotli # PyPi brotli package implements 'br' Content-Encoding try: import brotli as compat_brotli except ImportError: compat_brotli = None +# compat_ncompress # PyPi ncompress package implements 'compress' Content-Encoding try: import ncompress as compat_ncompress @@ -3495,6 +3577,7 @@ legacy = [ 'compat_getpass', 'compat_parse_qs', 'compat_realpath', + 'compat_shlex_split', 'compat_urllib_parse_parse_qs', 'compat_urllib_parse_unquote', 'compat_urllib_parse_unquote_plus', @@ -3508,8 +3591,6 @@ legacy = [ __all__ = [ - 'compat_html_parser_HTMLParseError', - 'compat_html_parser_HTMLParser', 'compat_Struct', 'compat_base64_b64decode', 'compat_basestring', @@ -3518,13 +3599,9 @@ __all__ = [ 'compat_chr', 'compat_collections_abc', 'compat_collections_chain_map', - 'compat_datetime_timedelta_total_seconds', - 'compat_http_cookiejar', - 'compat_http_cookiejar_Cookie', - 'compat_http_cookies', - 'compat_http_cookies_SimpleCookie', 'compat_contextlib_suppress', 'compat_ctypes_WINFUNCTYPE', + 'compat_datetime_timedelta_total_seconds', 'compat_etree_fromstring', 'compat_etree_iterfind', 'compat_filter', @@ -3533,6 +3610,12 @@ __all__ = [ 'compat_getpass_getpass', 'compat_html_entities', 'compat_html_entities_html5', + 'compat_html_parser_HTMLParseError', + 'compat_html_parser_HTMLParser', + 'compat_http_cookiejar', + 'compat_http_cookiejar_Cookie', + 'compat_http_cookies', + 'compat_http_cookies_SimpleCookie', 'compat_http_client', 'compat_http_server', 'compat_input', @@ -3555,7 +3638,7 @@ __all__ = [ 'compat_register_utf8', 'compat_setenv', 'compat_shlex_quote', - 'compat_shlex_split', + 'compat_shutil_get_terminal_size', 'compat_socket_create_connection', 'compat_str', 'compat_struct_pack', @@ -3575,5 +3658,4 @@ __all__ = [ 'compat_xml_etree_register_namespace', 'compat_xpath', 'compat_zip', - 'workaround_optparse_bug9161', ] From 8738407d77f6da843f8f5ded1ccad73172b4abac Mon Sep 17 00:00:00 2001 From: dirkf Date: Sun, 2 Mar 2025 13:36:05 +0000 Subject: [PATCH 34/78] [compat] Support zstd Content-Encoding * see RFC 8878 7.2 --- youtube_dl/compat.py | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index e617dd511..6cd7abd24 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -3561,6 +3561,13 @@ try: except ImportError: compat_ncompress = None +# compat_zstandard +# PyPi zstandard package implements 'zstd' Content-Encoding (RFC 8878 7.2) +try: + import zstandard as compat_zstandard +except ImportError: + compat_zstandard = None + legacy = [ 'compat_HTMLParseError', @@ -3658,4 +3665,5 @@ __all__ = [ 'compat_xml_etree_register_namespace', 'compat_xpath', 'compat_zip', + 'compat_zstandard', ] From 974c7d7f349831cf32026ec57e75bc821843a07b Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 19:17:18 +0000 Subject: [PATCH 35/78] [compat] Fix inheriting from compat_collections_chain_map * see ytdl-org/youtube-dl#33079#issuecomment-2704038049 --- youtube_dl/compat.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index 6cd7abd24..8910a4dac 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -3473,11 +3473,12 @@ except ImportError: def new_child(self, m=None, **kwargs): m = m or {} m.update(kwargs) - return compat_collections_chain_map(m, *self.maps) + # support inheritance ! + return type(self)(m, *self.maps) @property def parents(self): - return compat_collections_chain_map(*(self.maps[1:])) + return type(self)(*(self.maps[1:])) # compat_re_Pattern, compat_re_Match From 94849bc997d232b344b0f3666198feec7b004b43 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 19:32:56 +0000 Subject: [PATCH 36/78] [JSInterp] Improve Date processing * add JS_Date class implementing JS Date * support constructor args other than date string * support static methods of Date * Date objects are still automatically coerced to timestamp before using in JS. --- test/test_jsinterp.py | 22 ++++++++++++ youtube_dl/jsinterp.py | 76 +++++++++++++++++++++++++++++++++++++++++- 2 files changed, 97 insertions(+), 1 deletion(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 6c34bc896..4c5256c4b 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -11,6 +11,7 @@ sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) import math import re +import time from youtube_dl.compat import compat_str as str from youtube_dl.jsinterp import JS_Undefined, JSInterpreter @@ -208,6 +209,27 @@ class TestJSInterpreter(unittest.TestCase): self._test(jsi, 86000, args=['12/31/1969 18:01:26 MDT']) # epoch 0 self._test(jsi, 0, args=['1 January 1970 00:00:00 UTC']) + # undefined + self._test(jsi, NaN, args=[JS_Undefined]) + # y,m,d, ... - may fail with older dates lacking DST data + jsi = JSInterpreter('function f() { return new Date(%s); }' + % ('2024, 5, 29, 2, 52, 12, 42',)) + self._test(jsi, 1719625932042) + # no arg + self.assertAlmostEqual(JSInterpreter( + 'function f() { return new Date() - 0; }').call_function('f'), + time.time() * 1000, delta=100) + # Date.now() + self.assertAlmostEqual(JSInterpreter( + 'function f() { return Date.now(); }').call_function('f'), + time.time() * 1000, delta=100) + # Date.parse() + jsi = JSInterpreter('function f(dt) { return Date.parse(dt); }') + self._test(jsi, 0, args=['1 January 1970 00:00:00 UTC']) + # Date.UTC() + jsi = JSInterpreter('function f() { return Date.UTC(%s); }' + % ('1970, 0, 1, 0, 0, 0, 0',)) + self._test(jsi, 0) def test_call(self): jsi = JSInterpreter(''' diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 2859bc734..c3ee3bb03 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -1,10 +1,12 @@ # coding: utf-8 from __future__ import unicode_literals +import calendar import itertools import json import operator import re +import time from functools import update_wrapper, wraps @@ -12,8 +14,10 @@ from .utils import ( error_to_compat_str, ExtractorError, float_or_none, + int_or_none, js_to_json, remove_quotes, + str_or_none, unified_timestamp, variadic, write_string, @@ -475,6 +479,73 @@ class JSInterpreter(object): flags |= cls.RE_FLAGS[ch] return flags, expr[idx + 1:] + class JS_Date(object): + _t = None + + @staticmethod + def __ymd_etc(*args, **kw_is_utc): + # args: year, monthIndex, day, hours, minutes, seconds, milliseconds + is_utc = kw_is_utc.get('is_utc', False) + + args = list(args[:7]) + args += [0] * (9 - len(args)) + args[1] += 1 # month 0..11 -> 1..12 + ms = args[6] + for i in range(6, 9): + args[i] = -1 # don't know + if is_utc: + args[-1] = 1 + # TODO: [MDN] When a segment overflows or underflows its expected + # range, it usually "carries over to" or "borrows from" the higher segment. + try: + mktime = calendar.timegm if is_utc else time.mktime + return mktime(time.struct_time(args)) * 1000 + ms + except (OverflowError, ValueError): + return None + + @classmethod + def UTC(cls, *args): + t = cls.__ymd_etc(*args, is_utc=True) + return _NaN if t is None else t + + @staticmethod + def parse(date_str, **kw_is_raw): + is_raw = kw_is_raw.get('is_raw', False) + + t = unified_timestamp(str_or_none(date_str), False) + return int(t * 1000) if t is not None else t if is_raw else _NaN + + @staticmethod + def now(**kw_is_raw): + is_raw = kw_is_raw.get('is_raw', False) + + t = time.time() + return int(t * 1000) if t is not None else t if is_raw else _NaN + + def __init__(self, *args): + if not args: + args = [self.now(is_raw=True)] + if len(args) == 1: + if isinstance(args[0], JSInterpreter.JS_Date): + self._t = int_or_none(args[0].valueOf(), default=None) + else: + arg_type = _js_typeof(args[0]) + if arg_type == 'string': + self._t = self.parse(args[0], is_raw=True) + elif arg_type == 'number': + self._t = int(args[0]) + else: + self._t = self.__ymd_etc(*args) + + def toString(self): + try: + return time.strftime('%a %b %0d %Y %H:%M:%S %Z%z', self._t).rstrip() + except TypeError: + return "Invalid Date" + + def valueOf(self): + return _NaN if self._t is None else self._t + @classmethod def __op_chars(cls): op_chars = set(';,[') @@ -715,7 +786,7 @@ class JSInterpreter(object): new_kw, _, obj = expr.partition('new ') if not new_kw: - for klass, konstr in (('Date', lambda x: int(unified_timestamp(x, False) * 1000)), + for klass, konstr in (('Date', lambda *x: self.JS_Date(*x).valueOf()), ('RegExp', self.JS_RegExp), ('Error', self.Exception)): if not obj.startswith(klass + '('): @@ -1034,6 +1105,7 @@ class JSInterpreter(object): 'String': compat_str, 'Math': float, 'Array': list, + 'Date': self.JS_Date, } obj = local_vars.get(variable) if obj in (JS_Undefined, None): @@ -1086,6 +1158,8 @@ class JSInterpreter(object): assertion(len(argvals) == 2, 'takes two arguments') return argvals[0] ** argvals[1] raise self.Exception('Unsupported Math method ' + member, expr=expr) + elif obj is self.JS_Date: + return getattr(obj, member)(*argvals) if member == 'split': assertion(len(argvals) <= 2, 'takes at most two arguments') From af049e309bfa47141a9788cd1730dd50dad6176d Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 19:37:56 +0000 Subject: [PATCH 37/78] [JSInterp] Handle undefined, etc, passed to JS_RegExp and Exception --- youtube_dl/jsinterp.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index c3ee3bb03..9b4157a43 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -408,6 +408,7 @@ class JSInterpreter(object): class Exception(ExtractorError): def __init__(self, msg, *args, **kwargs): expr = kwargs.pop('expr', None) + msg = str_or_none(msg, default='"None"') if expr is not None: msg = '{0} in: {1!r:.100}'.format(msg.rstrip(), expr) super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs) @@ -435,6 +436,7 @@ class JSInterpreter(object): flags, _ = self.regex_flags(flags) # First, avoid https://github.com/python/cpython/issues/74534 self.__self = None + pattern_txt = str_or_none(pattern_txt) or '(?:)' self.__pattern_txt = pattern_txt.replace('[[', r'[\[') self.__flags = flags From 1dc27e1c3bda9cb8f44b805c89918aa7d11ffcdc Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 19:40:53 +0000 Subject: [PATCH 38/78] [JSInterp] Make indexing error handling more conformant * by default TypeError -> undefined, else raise * set allow_undefined=True/False to override --- youtube_dl/jsinterp.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 9b4157a43..5a45fbb03 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -672,14 +672,15 @@ class JSInterpreter(object): except Exception as e: raise self.Exception('Failed to evaluate {left_val!r:.50} {op} {right_val!r:.50}'.format(**locals()), expr, cause=e) - def _index(self, obj, idx, allow_undefined=True): + def _index(self, obj, idx, allow_undefined=None): if idx == 'length' and isinstance(obj, list): return len(obj) try: return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)] except (TypeError, KeyError, IndexError) as e: - if allow_undefined: - # when is not allowed? + # allow_undefined is None gives correct behaviour + if allow_undefined or ( + allow_undefined is None and not isinstance(e, TypeError)): return JS_Undefined raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e) From 422b1b31cf398d60b4606fa57be8e39c1181932f Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 20:00:58 +0000 Subject: [PATCH 39/78] [YouTube] Temporarily redirect from tce-style player JS --- youtube_dl/extractor/youtube.py | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 5f8c08201..9e200105e 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1607,9 +1607,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor): webpage or '', 'player URL', fatal=False) if player_url: ytcfgs = ytcfgs + ({'PLAYER_JS_URL': player_url},) - return traverse_obj( + player_url = traverse_obj( ytcfgs, (Ellipsis, 'PLAYER_JS_URL'), (Ellipsis, 'WEB_PLAYER_CONTEXT_CONFIGS', Ellipsis, 'jsUrl'), get_all=False, expected_type=lambda u: urljoin('https://www.youtube.com', u)) + nplayer_url, is_tce = re.subn(r'(?<=/player_ias)_tce(?=\.vflset/)', '', player_url or '') + if is_tce: + # TODO: Add proper support for the 'tce' variant players + # See https://github.com/yt-dlp/yt-dlp/issues/12398 + self.write_debug('Modifying tce player URL: {0}'.format(player_url)) + return nplayer_url + return player_url def _download_player_url(self, video_id, fatal=False): res = self._download_webpage( From 283dca56feb9f23978733810ab155472d6473c38 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 21:02:28 +0000 Subject: [PATCH 40/78] [YouTube] Initially support tce-style player JS * resolves #33079 --- test/test_youtube_signature.py | 21 +++++++++++++++++---- youtube_dl/extractor/youtube.py | 23 +++++++++++++---------- 2 files changed, 30 insertions(+), 14 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 67ef75fde..166614e62 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -223,6 +223,18 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/9c6dfc4a/player_ias.vflset/en_US/base.js', 'jbu7ylIosQHyJyJV', 'uwI0ESiynAmhNg', ), + ( + 'https://www.youtube.com/s/player/f6e09c70/player_ias.vflset/en_US/base.js', + 'W9HJZKktxuYoDTqW', 'jHbbkcaxm54', + ), + ( + 'https://www.youtube.com/s/player/f6e09c70/player_ias_tce.vflset/en_US/base.js', + 'W9HJZKktxuYoDTqW', 'jHbbkcaxm54', + ), + ( + 'https://www.youtube.com/s/player/91201489/player_ias_tce.vflset/en_US/base.js', + 'W9HJZKktxuYoDTqW', 'U48vOZHaeYS6vO', + ), ] @@ -284,7 +296,7 @@ def t_factory(name, sig_func, url_pattern): def signature(jscode, sig_input): - func = YoutubeIE(FakeYDL())._parse_sig_js(jscode) + func = YoutubeIE(FakeYDL({'cachedir': False}))._parse_sig_js(jscode) src_sig = ( compat_str(string.printable[:sig_input]) if isinstance(sig_input, int) else sig_input) @@ -292,9 +304,10 @@ def signature(jscode, sig_input): def n_sig(jscode, sig_input): - funcname = YoutubeIE(FakeYDL())._extract_n_function_name(jscode) - return JSInterpreter(jscode).call_function( - funcname, sig_input, _ytdl_do_not_return=sig_input) + ie = YoutubeIE(FakeYDL({'cachedir': False})) + jsi = JSInterpreter(jscode) + jsi, _, func_code = ie._extract_n_function_code_jsi(sig_input, jsi) + return ie._extract_n_function_from_code(jsi, func_code)(sig_input) make_sig_test = t_factory( diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 9e200105e..11bed6cae 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1607,16 +1607,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): webpage or '', 'player URL', fatal=False) if player_url: ytcfgs = ytcfgs + ({'PLAYER_JS_URL': player_url},) - player_url = traverse_obj( + return traverse_obj( ytcfgs, (Ellipsis, 'PLAYER_JS_URL'), (Ellipsis, 'WEB_PLAYER_CONTEXT_CONFIGS', Ellipsis, 'jsUrl'), get_all=False, expected_type=lambda u: urljoin('https://www.youtube.com', u)) - nplayer_url, is_tce = re.subn(r'(?<=/player_ias)_tce(?=\.vflset/)', '', player_url or '') - if is_tce: - # TODO: Add proper support for the 'tce' variant players - # See https://github.com/yt-dlp/yt-dlp/issues/12398 - self.write_debug('Modifying tce player URL: {0}'.format(player_url)) - return nplayer_url - return player_url def _download_player_url(self, video_id, fatal=False): res = self._download_webpage( @@ -1858,12 +1851,22 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if func_code: return jsi, player_id, func_code + return self._extract_n_function_code_jsi(video_id, jsi, player_id) - func_name = self._extract_n_function_name(jscode) + 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\("\{"\))(?=\s*[,;])', + jsi.code, 'useful values', default='') + + func_name = self._extract_n_function_name(jsi.code) func_code = jsi.extract_function_code(func_name) + if var_ay: + func_code = (func_code[0], ';\n'.join((var_ay, func_code[1]))) - self.cache.store('youtube-nsig', player_id, func_code) + if player_id: + self.cache.store('youtube-nsig', player_id, func_code) return jsi, player_id, func_code def _extract_n_function_from_code(self, jsi, func_code): From 32f89de92b652bf246aa458a552c9bb397abef77 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 7 Mar 2025 21:03:54 +0000 Subject: [PATCH 41/78] [YouTube] Update TVHTML5 client parameters * resolves #33078 --- youtube_dl/extractor/youtube.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 11bed6cae..6364e3aee 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -122,7 +122,8 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'INNERTUBE_CONTEXT': { 'client': { 'clientName': 'TVHTML5', - 'clientVersion': '7.20241201.18.00', + 'clientVersion': '7.20250120.19.00', + 'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/Version', }, }, 'INNERTUBE_CONTEXT_CLIENT_NAME': 7, From 420d53387cff54ea1fccca061438d59bdb50a39c Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 10 Mar 2025 11:44:06 +0000 Subject: [PATCH 42/78] [JSInterp] Improve tests * from yt-dlp/yt-dlp#12313 * also fix d7c2708 --- test/test_jsinterp.py | 21 ++++++++++++++++++--- youtube_dl/jsinterp.py | 3 ++- 2 files changed, 20 insertions(+), 4 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 4c5256c4b..3c9650ab6 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -212,9 +212,16 @@ class TestJSInterpreter(unittest.TestCase): # undefined self._test(jsi, NaN, args=[JS_Undefined]) # y,m,d, ... - may fail with older dates lacking DST data - jsi = JSInterpreter('function f() { return new Date(%s); }' - % ('2024, 5, 29, 2, 52, 12, 42',)) - self._test(jsi, 1719625932042) + jsi = JSInterpreter( + 'function f() { return new Date(%s); }' + % ('2024, 5, 29, 2, 52, 12, 42',)) + self._test(jsi, ( + 1719625932042 # UK value + + ( + + 3600 # back to GMT + + (time.altzone if time.daylight # host's DST + else time.timezone) + ) * 1000)) # no arg self.assertAlmostEqual(JSInterpreter( 'function f() { return new Date() - 0; }').call_function('f'), @@ -485,6 +492,14 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return NaN << 42}', 0) self._test('function f(){return "21.9" << 1}', 42) self._test('function f(){return 21 << 4294967297}', 42) + self._test('function f(){return true << "5";}', 32) + self._test('function f(){return true << true;}', 2) + self._test('function f(){return "19" & "21.9";}', 17) + self._test('function f(){return "19" & false;}', 0) + self._test('function f(){return "11.0" >> "2.1";}', 2) + self._test('function f(){return 5 ^ 9;}', 12) + self._test('function f(){return 0.0 << NaN}', 0) + self._test('function f(){return null << undefined}', 0) def test_negative(self): self._test('function f(){return 2 * -2.0 ;}', -4) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 5a45fbb03..d9b33fa44 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -154,6 +154,7 @@ def _js_to_primitive(v): ) +# more exact: yt-dlp/yt-dlp#12110 def _js_toString(v): return ( 'undefined' if v is JS_Undefined @@ -162,7 +163,7 @@ def _js_toString(v): else 'null' if v is None # bool <= int: do this first else ('false', 'true')[v] if isinstance(v, bool) - else '{0:.7f}'.format(v).rstrip('.0') if isinstance(v, compat_numeric_types) + else re.sub(r'(?<=\d)\.?0*$', '', '{0:.7f}'.format(v)) if isinstance(v, compat_numeric_types) else _js_to_primitive(v)) From 37c2440d6a9a2a7226f6548ebc2b59160408e79a Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 21 Mar 2025 16:13:24 +0000 Subject: [PATCH 43/78] [YouTube] Update player client data thx seproDev (yt-dlp/yt-dlp#12603) Co-authored-by: sepro --- youtube_dl/extractor/youtube.py | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 6364e3aee..e58b48abc 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -91,12 +91,12 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'INNERTUBE_CONTEXT': { 'client': { 'clientName': 'IOS', - 'clientVersion': '19.45.4', + 'clientVersion': '20.10.4', 'deviceMake': 'Apple', 'deviceModel': 'iPhone16,2', - 'userAgent': 'com.google.ios.youtube/19.45.4 (iPhone16,2; U; CPU iOS 18_1_0 like Mac OS X;)', + 'userAgent': 'com.google.ios.youtube/20.10.4 (iPhone16,2; U; CPU iOS 18_3_2 like Mac OS X;)', 'osName': 'iPhone', - 'osVersion': '18.1.0.22B83', + 'osVersion': '18.3.2.22D82', }, }, 'INNERTUBE_CONTEXT_CLIENT_NAME': 5, @@ -109,7 +109,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'INNERTUBE_CONTEXT': { 'client': { 'clientName': 'MWEB', - 'clientVersion': '2.20241202.07.00', + 'clientVersion': '2.20250311.03.00', # mweb previously did not require PO Token with this UA 'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)', }, @@ -122,7 +122,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'INNERTUBE_CONTEXT': { 'client': { 'clientName': 'TVHTML5', - 'clientVersion': '7.20250120.19.00', + 'clientVersion': '7.20250312.16.00', 'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/Version', }, }, @@ -133,7 +133,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'INNERTUBE_CONTEXT': { 'client': { 'clientName': 'WEB', - 'clientVersion': '2.20241126.01.00', + 'clientVersion': '2.20250312.04.00', }, }, 'INNERTUBE_CONTEXT_CLIENT_NAME': 1, From da7223d4aa42ff9fc680b0951d043dd03cec2d30 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 21 Mar 2025 16:26:25 +0000 Subject: [PATCH 44/78] [YouTube] Improve support for tce-style player JS * improve extraction of global "useful data" Array from player JS * also handle tv-player and add tests: thx seproDev (yt-dlp/yt-dlp#12684) Co-Authored-By: sepro --- test/test_youtube_signature.py | 28 ++++++++++++++++++++++++++-- youtube_dl/extractor/youtube.py | 4 ++-- 2 files changed, 28 insertions(+), 4 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 166614e62..7726b1943 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -232,8 +232,32 @@ _NSIG_TESTS = [ 'W9HJZKktxuYoDTqW', 'jHbbkcaxm54', ), ( - 'https://www.youtube.com/s/player/91201489/player_ias_tce.vflset/en_US/base.js', - 'W9HJZKktxuYoDTqW', 'U48vOZHaeYS6vO', + 'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js', + 'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw', + ), + ( + 'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js', + 'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ', + ), + ( + 'https://www.youtube.com/s/player/d50f54ef/player_ias_tce.vflset/en_US/base.js', + 'Ha7507LzRmH3Utygtj', 'XFTb2HoeOE5MHg', + ), + ( + 'https://www.youtube.com/s/player/074a8365/player_ias_tce.vflset/en_US/base.js', + 'Ha7507LzRmH3Utygtj', 'ufTsrE0IVYrkl8v', + ), + ( + 'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js', + 'N5uAlLqm0eg1GyHO', 'dCBQOejdq5s-ww', + ), + ( + 'https://www.youtube.com/s/player/69f581a5/tv-player-ias.vflset/tv-player-ias.js', + '-qIP447rVlTTwaZjY', 'KNcGOksBAvwqQg', + ), + ( + 'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', + 'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA', ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index e58b48abc..415354811 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -692,7 +692,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'invidious': '|'.join(_INVIDIOUS_SITES), } _PLAYER_INFO_RE = ( - r'/s/player/(?P[a-zA-Z0-9_-]{8,})/player', + r'/s/player/(?P[a-zA-Z0-9_-]{8,})//(?:tv-)?player', r'/(?P[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$', r'\b(?Pvfl[a-zA-Z0-9_-]+)\b.*?\.js$', ) @@ -1857,7 +1857,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None): var_ay = self._search_regex( - r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"[^"]+"\s*\.\s*split\("\{"\))(?=\s*[,;])', + 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) From 2239ee796584ece265edef2df20782618a615056 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 21:39:31 +0000 Subject: [PATCH 45/78] [YouTube] Get subtitles/automatic captions from both web and API responses --- youtube_dl/extractor/youtube.py | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 415354811..a3d3acc4d 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -2136,7 +2136,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): video_details = merge_dicts(*traverse_obj( (player_response, api_player_response), (Ellipsis, 'videoDetails', T(dict)))) - player_response.update(api_player_response or {}) + player_response.update(filter_dict( + api_player_response or {}, cndn=lambda k, _: k != 'captions')) player_response['videoDetails'] = video_details def is_agegated(playability): @@ -2566,8 +2567,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): } pctr = traverse_obj( - player_response, - ('captions', 'playerCaptionsTracklistRenderer', T(dict))) + (player_response, api_player_response), + (Ellipsis, 'captions', 'playerCaptionsTracklistRenderer', T(dict))) if pctr: def process_language(container, base_url, lang_code, query): lang_subs = [] @@ -2584,20 +2585,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def process_subtitles(): subtitles = {} for caption_track in traverse_obj(pctr, ( - 'captionTracks', lambda _, v: v.get('baseUrl'))): + Ellipsis, 'captionTracks', lambda _, v: ( + v.get('baseUrl') and v.get('languageCode')))): base_url = self._yt_urljoin(caption_track['baseUrl']) if not base_url: continue + lang_code = caption_track['languageCode'] if caption_track.get('kind') != 'asr': - lang_code = caption_track.get('languageCode') - if not lang_code: - continue process_language( subtitles, base_url, lang_code, {}) continue automatic_captions = {} for translation_language in traverse_obj(pctr, ( - 'translationLanguages', lambda _, v: v.get('languageCode'))): + Ellipsis, 'translationLanguages', lambda _, v: v.get('languageCode'))): translation_language_code = translation_language['languageCode'] process_language( automatic_captions, base_url, translation_language_code, From 7dca08eff01c0e8a25be38ec76c63170f26ac118 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 21:41:15 +0000 Subject: [PATCH 46/78] [YouTube] Also get original of translated automatic captions --- youtube_dl/extractor/youtube.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index a3d3acc4d..50527c98c 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -2596,6 +2596,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): subtitles, base_url, lang_code, {}) continue automatic_captions = {} + process_language( + automatic_captions, base_url, lang_code, {}) for translation_language in traverse_obj(pctr, ( Ellipsis, 'translationLanguages', lambda _, v: v.get('languageCode'))): translation_language_code = translation_language['languageCode'] From a464c159e61ab5c2c3babc8a4eb1d3e7923f1fe0 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 21:49:21 +0000 Subject: [PATCH 47/78] [YouTube] Make `_extract_player_info()` use `_search_regex()` --- test/test_youtube_signature.py | 3 ++- youtube_dl/extractor/youtube.py | 16 +++++++--------- 2 files changed, 9 insertions(+), 10 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 7726b1943..24d84dc20 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -280,8 +280,9 @@ class TestPlayerInfo(unittest.TestCase): ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflXGBaUN.js', 'vflXGBaUN'), ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflKjOTVq/html5player.js', 'vflKjOTVq'), ) + ie = YoutubeIE(FakeYDL({'cachedir': False})) for player_url, expected_player_id in PLAYER_URLS: - player_id = YoutubeIE._extract_player_info(player_url) + player_id = ie._extract_player_info(player_url) self.assertEqual(player_id, expected_player_id) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 50527c98c..a0adee7b8 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1626,15 +1626,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor): """ Return a string representation of a signature """ return '.'.join(compat_str(len(part)) for part in example_sig.split('.')) - @classmethod - def _extract_player_info(cls, player_url): - for player_re in cls._PLAYER_INFO_RE: - id_m = re.search(player_re, player_url) - if id_m: - break - else: - raise ExtractorError('Cannot identify player %r' % player_url) - return id_m.group('id') + def _extract_player_info(self, player_url): + try: + return self._search_regex( + self._PLAYER_INFO_RE, player_url, 'player info', group='id') + except ExtractorError as e: + raise ExtractorError( + 'Cannot identify player %r' % (player_url,), cause=e) def _load_player(self, video_id, player_url, fatal=True, player_id=None): if not player_id: From a4fc1151f125a2dfb89b2d1505838bcbb71aacc9 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 21:56:32 +0000 Subject: [PATCH 48/78] [JSInterp] Improve indexing * catch invalid list index with `ValueError` (eg [1, 2]['ab'] -> undefined) * allow assignment outside existing list (eg var l = [1,2]; l[9] = 0;) --- youtube_dl/jsinterp.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index d9b33fa44..a6d4f8e75 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -678,7 +678,7 @@ class JSInterpreter(object): return len(obj) try: return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)] - except (TypeError, KeyError, IndexError) as e: + except (TypeError, KeyError, IndexError, ValueError) as e: # allow_undefined is None gives correct behaviour if allow_undefined or ( allow_undefined is None and not isinstance(e, TypeError)): @@ -1038,6 +1038,10 @@ class JSInterpreter(object): left_val = self._index(left_val, idx) if isinstance(idx, float): idx = int(idx) + if isinstance(left_val, list) and len(left_val) <= int_or_none(idx, default=-1): + # JS Array is a sparsely assignable list + # TODO: handle extreme sparsity without memory bloat, eg using auxiliary dict + left_val.extend((idx - len(left_val) + 1) * [JS_Undefined]) left_val[idx] = self._operator( m.group('op'), self._index(left_val, idx) if m.group('op') else None, m.group('expr'), expr, local_vars, allow_recursion) From 087d8652302270764bf9679c7621db0d20ad5572 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 22:14:08 +0000 Subject: [PATCH 49/78] [YouTube] Support new player URL patterns --- test/test_youtube_signature.py | 12 ++++++++---- youtube_dl/extractor/youtube.py | 6 +++--- 2 files changed, 11 insertions(+), 7 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 24d84dc20..385b718e0 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -302,8 +302,8 @@ class TestSignature(unittest.TestCase): def t_factory(name, sig_func, url_pattern): def make_tfunc(url, sig_input, expected_sig): m = url_pattern.match(url) - assert m, '%r should follow URL format' % url - test_id = m.group('id') + assert m, '{0!r} should follow URL format'.format(url) + test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id')) def test_func(self): basename = 'player-{0}-{1}.js'.format(name, test_id) @@ -336,12 +336,16 @@ def n_sig(jscode, sig_input): make_sig_test = t_factory( - 'signature', signature, re.compile(r'.*(?:-|/player/)(?P[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$')) + 'signature', signature, + re.compile(r'''(?x) + .+/(?P
html5)?player(?(h5)(?:-en_US)?-|/)(?P[a-zA-Z0-9/._-]+) + (?(h5)/(?:watch_as3|html5player))?\.js$ + ''')) for test_spec in _SIG_TESTS: make_sig_test(*test_spec) make_nsig_test = t_factory( - 'nsig', n_sig, re.compile(r'.+/player/(?P[a-zA-Z0-9_-]+)/.+.js$')) + 'nsig', n_sig, re.compile(r'.+/player/(?P[a-zA-Z0-9_/.-]+)\.js$')) for test_spec in _NSIG_TESTS: make_nsig_test(*test_spec) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index a0adee7b8..4347663ad 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -692,9 +692,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'invidious': '|'.join(_INVIDIOUS_SITES), } _PLAYER_INFO_RE = ( - r'/s/player/(?P[a-zA-Z0-9_-]{8,})//(?:tv-)?player', - r'/(?P[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$', - r'\b(?Pvfl[a-zA-Z0-9_-]+)\b.*?\.js$', + r'/s/player/(?P[a-zA-Z0-9_-]{8,})/(?:tv-)?player', + r'/(?P[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$', + r'\b(?Pvfl[a-zA-Z0-9_-]{6,})\b.*?\.js$', ) _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt') From c55dbf48384f9721ae8eb07fd914ccf2d3c823b0 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 22:16:19 +0000 Subject: [PATCH 50/78] [YouTube] Update signature extraction for players `643afba4`, `363db69b` --- youtube_dl/extractor/youtube.py | 28 +++++++++++++++++++--------- youtube_dl/jsinterp.py | 10 ++++++---- 2 files changed, 25 insertions(+), 13 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 4347663ad..5addfe7c6 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1709,6 +1709,18 @@ class YoutubeIE(YoutubeBaseInfoExtractor): ' return %s\n') % (signature_id_tuple, expr_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): # Examples where `sig` is funcname: # 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') 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 inner(*args, **kwargs): @@ -1854,15 +1870,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 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_code = jsi.extract_function_code(func_name) - if var_ay: - func_code = (func_code[0], ';\n'.join((var_ay, func_code[1]))) + func_code = self._extract_sig_fn(jsi, func_name) if player_id: self.cache.store('youtube-nsig', player_id, func_code) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index a6d4f8e75..186d2594a 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -1368,19 +1368,21 @@ class JSInterpreter(object): code, _ = self._separate_at_paren(func_m.group('code')) # refine the match 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( - 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,)) def extract_function_from_code(self, argnames, code, *global_stack): local_vars = {} + start = None while True: - mobj = re.search(r'function\((?P[^)]*)\)\s*{', code) + mobj = re.search(r'function\((?P[^)]*)\)\s*{', code[start:]) if mobj is None: 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:]) name = self._named_object(local_vars, self.extract_function_from_code( [x.strip() for x in mobj.group('args').split(',')], From b982d77d0b7bd21ead6d8b3210fd656881197453 Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 22:18:47 +0000 Subject: [PATCH 51/78] [YouTube] Align signature tests with yt-dlp thx bashonly, yt-dlp/yt-dlp#12725 --- test/test_youtube_signature.py | 28 ++++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 385b718e0..696f832e4 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -84,6 +84,16 @@ _SIG_TESTS = [ '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', '0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xxAj7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJ2OySqa0q', ), + ( + 'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'AAOAOq0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xx8j7vgpDL0QwbdV06sCIEzpWqMGkFR20CFOS21Tp-7vj_EMu-m37KtXJoOy1', + ), + ( + 'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + ), ] _NSIG_TESTS = [ @@ -153,7 +163,7 @@ _NSIG_TESTS = [ ), ( 'https://www.youtube.com/s/player/c57c113c/player_ias.vflset/en_US/base.js', - '-Txvy6bT5R6LqgnQNx', 'dcklJCnRUHbgSg', + 'M92UUMHa8PdvPd3wyM', '3hPqLJsiNZx7yA', ), ( 'https://www.youtube.com/s/player/5a3b6271/player_ias.vflset/en_US/base.js', @@ -173,7 +183,7 @@ _NSIG_TESTS = [ ), ( 'https://www.youtube.com/s/player/cfa9e7cb/player_ias.vflset/en_US/base.js', - 'qO0NiMtYQ7TeJnfFG2', 'k9cuJDHNS5O7kQ', + 'aCi3iElgd2kq0bxVbQ', 'QX1y8jGb2IbZ0w', ), ( 'https://www.youtube.com/s/player/8c7583ff/player_ias.vflset/en_US/base.js', @@ -231,10 +241,6 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/f6e09c70/player_ias_tce.vflset/en_US/base.js', 'W9HJZKktxuYoDTqW', 'jHbbkcaxm54', ), - ( - 'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js', - 'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw', - ), ( 'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js', 'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ', @@ -259,6 +265,14 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', 'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA', ), + ( + 'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js', + 'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw', + ), + ( + 'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js', + 'eWYu5d5YeY_4LyEDc', 'XJQqf-N7Xra3gg', + ), ] @@ -271,6 +285,8 @@ class TestPlayerInfo(unittest.TestCase): ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-en_US.vflset/base.js', '64dddad9'), ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-de_DE.vflset/base.js', '64dddad9'), ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-tablet-en_US.vflset/base.js', '64dddad9'), + ('https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js', 'e7567ecf'), + ('https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', '643afba4'), # obsolete ('https://www.youtube.com/yts/jsbin/player_ias-vfle4-e03/en_US/base.js', 'vfle4-e03'), ('https://www.youtube.com/yts/jsbin/player_ias-vfl49f_g4/en_US/base.js', 'vfl49f_g4'), From 1bc45b8b6c0c853e0f3d74346551fec0708c733d Mon Sep 17 00:00:00 2001 From: dirkf Date: Tue, 25 Mar 2025 22:30:08 +0000 Subject: [PATCH 52/78] [JSInterp] Use `,` for join() with null/undefined argument Eg: [1,2,3].join(null) -> '1,2,3' --- youtube_dl/jsinterp.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 186d2594a..f0c4fa928 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -1208,9 +1208,10 @@ class JSInterpreter(object): elif member == 'join': assertion(isinstance(obj, list), 'must be applied on a list') assertion(len(argvals) <= 1, 'takes at most one argument') - return (',' if len(argvals) == 0 else argvals[0]).join( - ('' if x in (None, JS_Undefined) else _js_toString(x)) - for x in obj) + return (',' if len(argvals) == 0 or argvals[0] in (None, JS_Undefined) + else argvals[0]).join( + ('' if x in (None, JS_Undefined) else _js_toString(x)) + for x in obj) elif member == 'reverse': assertion(not argvals, 'does not take any arguments') obj.reverse() From 2b4fbfce25902d557b86b003cf48f738129efce4 Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 26 Mar 2025 02:27:25 +0000 Subject: [PATCH 53/78] [YouTube] Support player `4fcd6e4a` thx seproDev, bashonly: yt-dlp/yt-dlp#12748 --- test/test_youtube_signature.py | 13 +++++++++++++ youtube_dl/extractor/youtube.py | 9 +++++++-- 2 files changed, 20 insertions(+), 2 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 696f832e4..8fdcce8d4 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -94,6 +94,11 @@ _SIG_TESTS = [ '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', ), + ( + 'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0', + ), ] _NSIG_TESTS = [ @@ -273,6 +278,14 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js', 'eWYu5d5YeY_4LyEDc', 'XJQqf-N7Xra3gg', ), + ( + 'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js', + 'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A', + ), + ( + 'https://www.youtube.com/s/player/4fcd6e4a/tv-player-ias.vflset/tv-player-ias.js', + 'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A', + ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 5addfe7c6..ce97fd75b 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1711,8 +1711,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 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='') + r'''(?x) + (?:\*/|\{|\n|^)\s*(?:'[^']+'\s*;\s*) + (var\s*[\w$]+\s*=\s*(?: + ('|")(?:\\\2|(?!\2).)+\2\s*\.\s*split\(\s*('|")\W+\3\s*\)| + \[\s*(?:('|")(?:\\\4|(?!\4).)*\4\s*(?:(?=\])|,\s*))+\] + ))(?=\s*[,;]) + ''', jsi.code, 'useful values', default='') sig_fn = jsi.extract_function_code(funcname) From c1ea7f5a242dbaad3a4646d73bb1cae22439a788 Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 26 Mar 2025 12:17:49 +0000 Subject: [PATCH 54/78] [ITV] Mark ITVX not working * update old shim * correct [_]IE_DESC --- youtube_dl/extractor/itv.py | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-) diff --git a/youtube_dl/extractor/itv.py b/youtube_dl/extractor/itv.py index c64af3be6..2510ad887 100644 --- a/youtube_dl/extractor/itv.py +++ b/youtube_dl/extractor/itv.py @@ -35,15 +35,6 @@ from ..utils import ( class ITVBaseIE(InfoExtractor): - def _search_nextjs_data(self, webpage, video_id, **kw): - transform_source = kw.pop('transform_source', None) - fatal = kw.pop('fatal', True) - return self._parse_json( - self._search_regex( - r''']+\bid=('|")__NEXT_DATA__\1[^>]*>(?P[^<]+)''', - webpage, 'next.js data', group='js', fatal=fatal, **kw), - video_id, transform_source=transform_source, fatal=fatal) - def __handle_request_webpage_error(self, err, video_id=None, errnote=None, fatal=True): if errnote is False: return False @@ -109,7 +100,9 @@ class ITVBaseIE(InfoExtractor): class ITVIE(ITVBaseIE): _VALID_URL = r'https?://(?:www\.)?itv\.com/(?:(?Pwatch)|hub)/[^/]+/(?(w)[\w-]+/)(?P\w+)' - _IE_DESC = 'ITVX' + IE_DESC = 'ITVX' + _WORKING = False + _TESTS = [{ 'note': 'Hub URLs redirect to ITVX', 'url': 'https://www.itv.com/hub/liar/2a4547a0012', @@ -270,7 +263,7 @@ class ITVIE(ITVBaseIE): 'ext': determine_ext(href, 'vtt'), }) - next_data = self._search_nextjs_data(webpage, video_id, fatal=False, default='{}') + next_data = self._search_nextjs_data(webpage, video_id, fatal=False, default={}) video_data.update(traverse_obj(next_data, ('props', 'pageProps', ('title', 'episode')), expected_type=dict)[0] or {}) title = traverse_obj(video_data, 'headerTitle', 'episodeTitle') info = self._og_extract(webpage, require_title=not title) @@ -323,7 +316,7 @@ class ITVIE(ITVBaseIE): class ITVBTCCIE(ITVBaseIE): _VALID_URL = r'https?://(?:www\.)?itv\.com/(?!(?:watch|hub)/)(?:[^/]+/)+(?P[^/?#&]+)' - _IE_DESC = 'ITV articles: News, British Touring Car Championship' + IE_DESC = 'ITV articles: News, British Touring Car Championship' _TESTS = [{ 'note': 'British Touring Car Championship', 'url': 'https://www.itv.com/btcc/articles/btcc-2018-all-the-action-from-brands-hatch', From 4e714f9df1ed2cccd51df60d45ff5504abe827b7 Mon Sep 17 00:00:00 2001 From: dirkf Date: Wed, 26 Mar 2025 12:47:19 +0000 Subject: [PATCH 55/78] [Misc] Correct [_]IE_DESC/NAME in a few IEs * thx seproDev, yt-dlp/yt-dlp/pull/12694/commits/ae69e3c * also add documenting comment in `InfoExtractor` --- youtube_dl/extractor/bokecc.py | 2 +- youtube_dl/extractor/cloudy.py | 2 +- youtube_dl/extractor/common.py | 2 ++ youtube_dl/extractor/senateisvp.py | 2 +- 4 files changed, 5 insertions(+), 3 deletions(-) diff --git a/youtube_dl/extractor/bokecc.py b/youtube_dl/extractor/bokecc.py index 6017e8344..4b8bef391 100644 --- a/youtube_dl/extractor/bokecc.py +++ b/youtube_dl/extractor/bokecc.py @@ -32,7 +32,7 @@ class BokeCCBaseIE(InfoExtractor): class BokeCCIE(BokeCCBaseIE): - _IE_DESC = 'CC视频' + IE_DESC = 'CC视频' _VALID_URL = r'https?://union\.bokecc\.com/playvideo\.bo\?(?P.*)' _TESTS = [{ diff --git a/youtube_dl/extractor/cloudy.py b/youtube_dl/extractor/cloudy.py index 85ca20ecc..d39a9a5c2 100644 --- a/youtube_dl/extractor/cloudy.py +++ b/youtube_dl/extractor/cloudy.py @@ -9,7 +9,7 @@ from ..utils import ( class CloudyIE(InfoExtractor): - _IE_DESC = 'cloudy.ec' + IE_DESC = 'cloudy.ec' _VALID_URL = r'https?://(?:www\.)?cloudy\.ec/(?:v/|embed\.php\?.*?\bid=)(?P[A-Za-z0-9]+)' _TESTS = [{ 'url': 'https://www.cloudy.ec/v/af511e2527aac', diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 78704b557..cb67b976d 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -422,6 +422,8 @@ class InfoExtractor(object): _GEO_COUNTRIES = None _GEO_IP_BLOCKS = None _WORKING = True + # supply this in public subclasses: used in supported sites list, etc + # IE_DESC = 'short description of IE' def __init__(self, downloader=None): """Constructor. Receives an optional downloader.""" diff --git a/youtube_dl/extractor/senateisvp.py b/youtube_dl/extractor/senateisvp.py index db5ef8b57..b8ac58713 100644 --- a/youtube_dl/extractor/senateisvp.py +++ b/youtube_dl/extractor/senateisvp.py @@ -47,7 +47,7 @@ class SenateISVPIE(InfoExtractor): ['vetaff', '76462', 'http://vetaff-f.akamaihd.net'], ['arch', '', 'http://ussenate-f.akamaihd.net/'] ] - _IE_NAME = 'senate.gov' + IE_NAME = 'senate.gov' _VALID_URL = r'https?://(?:www\.)?senate\.gov/isvp/?\?(?P.+)' _TESTS = [{ 'url': 'http://www.senate.gov/isvp/?comm=judiciary&type=live&stt=&filename=judiciary031715&auto_play=false&wmode=transparent&poster=http%3A%2F%2Fwww.judiciary.senate.gov%2Fthemes%2Fjudiciary%2Fimages%2Fvideo-poster-flash-fit.png', From 3eb8d22ddb8982ca4fb56bb7a8d6517538bf14c6 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 31 Mar 2025 04:21:09 +0100 Subject: [PATCH 56/78] [JSInterp] Temporary fix for #33102 --- youtube_dl/jsinterp.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index f0c4fa928..69c8f77ca 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -686,6 +686,8 @@ class JSInterpreter(object): raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e) def _dump(self, obj, namespace): + if obj is JS_Undefined: + return 'undefined' try: return json.dumps(obj) except TypeError: From 9a6ddece4dca4d5b7a8374a82f72d778750d6b88 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 04:04:01 +0100 Subject: [PATCH 57/78] [core] Refactor message routines to align better with yt-dlp * in particular, support `only_once` in the same methods --- youtube_dl/YoutubeDL.py | 37 ++++++++++++++++++++-------------- youtube_dl/extractor/common.py | 14 ++----------- 2 files changed, 24 insertions(+), 27 deletions(-) diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py index 9e5620eef..8367b6e53 100755 --- a/youtube_dl/YoutubeDL.py +++ b/youtube_dl/YoutubeDL.py @@ -540,10 +540,14 @@ class YoutubeDL(object): """Print message to stdout if not in quiet mode.""" return self.to_stdout(message, skip_eol, check_quiet=True) - def _write_string(self, s, out=None): + def _write_string(self, s, out=None, only_once=False, _cache=set()): + if only_once and s in _cache: + return write_string(s, out=out, encoding=self.params.get('encoding')) + if only_once: + _cache.add(s) - def to_stdout(self, message, skip_eol=False, check_quiet=False): + def to_stdout(self, message, skip_eol=False, check_quiet=False, only_once=False): """Print message to stdout if not in quiet mode.""" if self.params.get('logger'): self.params['logger'].debug(message) @@ -552,9 +556,9 @@ class YoutubeDL(object): terminator = ['\n', ''][skip_eol] output = message + terminator - self._write_string(output, self._screen_file) + self._write_string(output, self._screen_file, only_once=only_once) - def to_stderr(self, message): + def to_stderr(self, message, only_once=False): """Print message to stderr.""" assert isinstance(message, compat_str) if self.params.get('logger'): @@ -562,7 +566,7 @@ class YoutubeDL(object): else: message = self._bidi_workaround(message) output = message + '\n' - self._write_string(output, self._err_file) + self._write_string(output, self._err_file, only_once=only_once) def to_console_title(self, message): if not self.params.get('consoletitle', False): @@ -641,18 +645,11 @@ class YoutubeDL(object): raise DownloadError(message, exc_info) self._download_retcode = 1 - def report_warning(self, message, only_once=False, _cache={}): + def report_warning(self, message, only_once=False): ''' Print the message to stderr, it will be prefixed with 'WARNING:' If stderr is a tty file the 'WARNING:' will be colored ''' - if only_once: - m_hash = hash((self, message)) - m_cnt = _cache.setdefault(m_hash, 0) - _cache[m_hash] = m_cnt + 1 - if m_cnt > 0: - return - if self.params.get('logger') is not None: self.params['logger'].warning(message) else: @@ -663,7 +660,7 @@ class YoutubeDL(object): else: _msg_header = 'WARNING:' warning_message = '%s %s' % (_msg_header, message) - self.to_stderr(warning_message) + self.to_stderr(warning_message, only_once=only_once) def report_error(self, message, *args, **kwargs): ''' @@ -677,6 +674,16 @@ class YoutubeDL(object): kwargs['message'] = '%s %s' % (_msg_header, message) self.trouble(*args, **kwargs) + def write_debug(self, message, only_once=False): + '''Log debug message or Print message to stderr''' + if not self.params.get('verbose', False): + return + message = '[debug] {0}'.format(message) + if self.params.get('logger'): + self.params['logger'].debug(message) + else: + self.to_stderr(message, only_once) + def report_unscoped_cookies(self, *args, **kwargs): # message=None, tb=False, is_error=False if len(args) <= 2: @@ -2514,7 +2521,7 @@ class YoutubeDL(object): self.get_encoding())) write_string(encoding_str, encoding=None) - writeln_debug = lambda *s: self._write_string('[debug] %s\n' % (''.join(s), )) + writeln_debug = lambda *s: self.write_debug(''.join(s)) writeln_debug('youtube-dl version ', __version__) if _LAZY_LOADER: writeln_debug('Lazy loading extractors enabled') diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index cb67b976d..cbc4e968a 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -976,19 +976,9 @@ class InfoExtractor(object): """Print msg to screen, prefixing it with '[ie_name]'""" self._downloader.to_screen(self.__ie_msg(msg)) - def write_debug(self, msg, only_once=False, _cache=[]): + def write_debug(self, msg, only_once=False): '''Log debug message or Print message to stderr''' - if not self.get_param('verbose', False): - return - message = '[debug] ' + self.__ie_msg(msg) - logger = self.get_param('logger') - if logger: - logger.debug(message) - else: - if only_once and hash(message) in _cache: - return - self._downloader.to_stderr(message) - _cache.append(hash(message)) + self._downloader.write_debug(self.__ie_msg(msg), only_once=only_once) # name, default=None, *args, **kwargs def get_param(self, name, *args, **kwargs): From add46228701d78444e20afdde4beafa3268344f2 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 04:09:56 +0100 Subject: [PATCH 58/78] [compat] Add compat_os_makedirs * support exists_ok parameter in Py < 3.2 --- youtube_dl/compat.py | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index 8910a4dac..72b2f8f78 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -3120,6 +3120,21 @@ else: compat_os_path_expanduser = compat_expanduser +# compat_os_makedirs +try: + os.makedirs('.', exist_ok=True) + compat_os_makedirs = os.makedirs +except TypeError: # < Py3.2 + from errno import EEXIST as _errno_EEXIST + + def compat_os_makedirs(name, mode=0o777, exist_ok=False): + try: + return os.makedirs(name, mode=mode) + except OSError as ose: + if not (exist_ok and ose.errno == _errno_EEXIST): + raise + + # compat_os_path_realpath if compat_os_name == 'nt' and sys.version_info < (3, 8): # os.path.realpath on Windows does not follow symbolic links @@ -3637,6 +3652,7 @@ __all__ = [ 'compat_numeric_types', 'compat_open', 'compat_ord', + 'compat_os_makedirs', 'compat_os_name', 'compat_os_path_expanduser', 'compat_os_path_realpath', From cedeeed56f6ed61646a33e06397f9a059dbc9d90 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 04:42:14 +0100 Subject: [PATCH 59/78] [cache] Align further with yt-dlp * use compat_os_makedirs * support non-ASCII characters in cache key * improve logging --- test/test_cache.py | 2 +- youtube_dl/cache.py | 70 ++++++++++++++++++++++++++------------------- 2 files changed, 41 insertions(+), 31 deletions(-) diff --git a/test/test_cache.py b/test/test_cache.py index 931074aa1..21dfb6cfc 100644 --- a/test/test_cache.py +++ b/test/test_cache.py @@ -63,7 +63,7 @@ class TestCache(unittest.TestCase): obj = {'x': 1, 'y': ['ä', '\\a', True]} c.store('test_cache', 'k.', obj) self.assertEqual(c.load('test_cache', 'k.', min_ver='1970.01.01'), obj) - new_version = '.'.join(('%d' % ((v + 1) if i == 0 else v, )) for i, v in enumerate(version_tuple(__version__))) + new_version = '.'.join(('%0.2d' % ((v + 1) if i == 0 else v, )) for i, v in enumerate(version_tuple(__version__))) self.assertIs(c.load('test_cache', 'k.', min_ver=new_version), None) diff --git a/youtube_dl/cache.py b/youtube_dl/cache.py index 54123da0e..6169461b9 100644 --- a/youtube_dl/cache.py +++ b/youtube_dl/cache.py @@ -1,6 +1,6 @@ +# coding: utf-8 from __future__ import unicode_literals -import errno import json import os import re @@ -8,14 +8,17 @@ import shutil import traceback from .compat import ( + compat_contextlib_suppress, compat_getenv, compat_open as open, + compat_os_makedirs, + compat_urllib_parse, ) from .utils import ( error_to_compat_str, expand_path, is_outdated_version, - try_get, + traverse_obj, write_json_file, ) from .version import __version__ @@ -30,23 +33,36 @@ class Cache(object): def __init__(self, ydl): self._ydl = ydl + def _write_debug(self, *args, **kwargs): + self._ydl.write_debug(*args, **kwargs) + + def _report_warning(self, *args, **kwargs): + self._ydl.report_warning(*args, **kwargs) + + def _to_screen(self, *args, **kwargs): + self._ydl.to_screen(*args, **kwargs) + + def _get_params(self, k, default=None): + return self._ydl.params.get(k, default) + def _get_root_dir(self): - res = self._ydl.params.get('cachedir') + res = self._get_params('cachedir') if res is None: cache_root = compat_getenv('XDG_CACHE_HOME', '~/.cache') res = os.path.join(cache_root, self._YTDL_DIR) return expand_path(res) def _get_cache_fn(self, section, key, dtype): - assert re.match(r'^[a-zA-Z0-9_.-]+$', section), \ + assert re.match(r'^[\w.-]+$', section), \ 'invalid section %r' % section - assert re.match(r'^[a-zA-Z0-9_.-]+$', key), 'invalid key %r' % key + key = compat_urllib_parse.quote( + key, safe='').replace('%', ',') # encode non-ascii characters return os.path.join( self._get_root_dir(), section, '%s.%s' % (key, dtype)) @property def enabled(self): - return self._ydl.params.get('cachedir') is not False + return self._get_params('cachedir') is not False def store(self, section, key, data, dtype='json'): assert dtype in ('json',) @@ -56,61 +72,55 @@ class Cache(object): fn = self._get_cache_fn(section, key, dtype) try: - try: - os.makedirs(os.path.dirname(fn)) - except OSError as ose: - if ose.errno != errno.EEXIST: - raise + compat_os_makedirs(os.path.dirname(fn), exist_ok=True) + self._write_debug('Saving {section}.{key} to cache'.format(section=section, key=key)) write_json_file({self._VERSION_KEY: __version__, 'data': data}, fn) except Exception: tb = traceback.format_exc() - self._ydl.report_warning( - 'Writing cache to %r failed: %s' % (fn, tb)) + self._report_warning('Writing cache to {fn!r} failed: {tb}'.format(fn=fn, tb=tb)) def _validate(self, data, min_ver): - version = try_get(data, lambda x: x[self._VERSION_KEY]) + version = traverse_obj(data, self._VERSION_KEY) if not version: # Backward compatibility data, version = {'data': data}, self._DEFAULT_VERSION if not is_outdated_version(version, min_ver or '0', assume_new=False): return data['data'] - self._ydl.to_screen( - 'Discarding old cache from version {version} (needs {min_ver})'.format(**locals())) + self._write_debug('Discarding old cache from version {version} (needs {min_ver})'.format(version=version, min_ver=min_ver)) - def load(self, section, key, dtype='json', default=None, min_ver=None): + def load(self, section, key, dtype='json', default=None, **kw_min_ver): assert dtype in ('json',) + min_ver = kw_min_ver.get('min_ver') if not self.enabled: return default cache_fn = self._get_cache_fn(section, key, dtype) - try: + with compat_contextlib_suppress(IOError): # If no cache available try: - with open(cache_fn, 'r', encoding='utf-8') as cachef: + with open(cache_fn, encoding='utf-8') as cachef: + self._write_debug('Loading {section}.{key} from cache'.format(section=section, key=key), only_once=True) return self._validate(json.load(cachef), min_ver) - except ValueError: + except (ValueError, KeyError): try: file_size = os.path.getsize(cache_fn) except (OSError, IOError) as oe: file_size = error_to_compat_str(oe) - self._ydl.report_warning( - 'Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) - except IOError: - pass # No cache available + self._report_warning('Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) return default def remove(self): if not self.enabled: - self._ydl.to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)') + self._to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)') return cachedir = self._get_root_dir() if not any((term in cachedir) for term in ('cache', 'tmp')): - raise Exception('Not removing directory %s - this does not look like a cache dir' % cachedir) + raise Exception('Not removing directory %s - this does not look like a cache dir' % (cachedir,)) - self._ydl.to_screen( - 'Removing cache dir %s .' % cachedir, skip_eol=True) + self._to_screen( + 'Removing cache dir %s .' % (cachedir,), skip_eol=True, ), if os.path.exists(cachedir): - self._ydl.to_screen('.', skip_eol=True) + self._to_screen('.', skip_eol=True) shutil.rmtree(cachedir) - self._ydl.to_screen('.') + self._to_screen('.') From 7e136639dba8a6d0e966ccc41b3dd8231587c67b Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 7 Apr 2025 15:51:03 +0100 Subject: [PATCH 60/78] [compat] Improve Py2 compatibility for URL Quoting --- youtube_dl/compat.py | 69 +++++++++++++++++++++++++++++++++++++------- 1 file changed, 58 insertions(+), 11 deletions(-) diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index 72b2f8f78..ebe22bdf9 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -2498,8 +2498,7 @@ try: from urllib.parse import urlencode as compat_urllib_parse_urlencode from urllib.parse import parse_qs as compat_parse_qs except ImportError: # Python 2 - _asciire = (compat_urllib_parse._asciire if hasattr(compat_urllib_parse, '_asciire') - else re.compile(r'([\x00-\x7f]+)')) + _asciire = getattr(compat_urllib_parse, '_asciire', None) or re.compile(r'([\x00-\x7f]+)') # HACK: The following are the correct unquote_to_bytes, unquote and unquote_plus # implementations from cpython 3.4.3's stdlib. Python 2's version @@ -2567,24 +2566,21 @@ except ImportError: # Python 2 # Possible solutions are to either port it from python 3 with all # the friends or manually ensure input query contains only byte strings. # We will stick with latter thus recursively encoding the whole query. - def compat_urllib_parse_urlencode(query, doseq=0, encoding='utf-8'): + def compat_urllib_parse_urlencode(query, doseq=0, safe='', encoding='utf-8', errors='strict'): + def encode_elem(e): if isinstance(e, dict): e = encode_dict(e) elif isinstance(e, (list, tuple,)): - list_e = encode_list(e) - e = tuple(list_e) if isinstance(e, tuple) else list_e + e = type(e)(encode_elem(el) for el in e) elif isinstance(e, compat_str): - e = e.encode(encoding) + e = e.encode(encoding, errors) return e def encode_dict(d): - return dict((encode_elem(k), encode_elem(v)) for k, v in d.items()) + return tuple((encode_elem(k), encode_elem(v)) for k, v in d.items()) - def encode_list(l): - return [encode_elem(e) for e in l] - - return compat_urllib_parse._urlencode(encode_elem(query), doseq=doseq) + return compat_urllib_parse._urlencode(encode_elem(query), doseq=doseq).decode('ascii') # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib. # Python 2's version is apparently totally broken @@ -2639,6 +2635,57 @@ except ImportError: # Python 2 ('parse_qs', compat_parse_qs)): setattr(compat_urllib_parse, name, fix) + try: + all(chr(i) in b'' for i in range(256)) + except TypeError: + # not all chr(i) are str: patch Python2 quote + + _safemaps = getattr(compat_urllib_parse, '_safemaps', {}) + _always_safe = frozenset(compat_urllib_parse.always_safe) + + def _quote(s, safe='/'): + """quote('abc def') -> 'abc%20def'""" + + if not s and s is not None: # fast path + return s + safe = frozenset(safe) + cachekey = (safe, _always_safe) + try: + safe_map = _safemaps[cachekey] + except KeyError: + safe = _always_safe | safe + safe_map = {} + for i in range(256): + c = chr(i) + safe_map[c] = ( + c if (i < 128 and c in safe) + else b'%{0:02X}'.format(i)) + _safemaps[cachekey] = safe_map + + if safe.issuperset(s): + return s + return ''.join(safe_map[c] for c in s) + + # linked code + def _quote_plus(s, safe=''): + return ( + _quote(s, safe + b' ').replace(b' ', b'+') if b' ' in s + else _quote(s, safe)) + + # linked code + def _urlcleanup(): + if compat_urllib_parse._urlopener: + compat_urllib_parse._urlopener.cleanup() + _safemaps.clear() + compat_urllib_parse.ftpcache.clear() + + for name, fix in ( + ('quote', _quote), + ('quote_plus', _quote_plus), + ('urlcleanup', _urlcleanup)): + setattr(compat_urllib_parse, '_' + name, getattr(compat_urllib_parse, name)) + setattr(compat_urllib_parse, name, fix) + compat_urllib_parse_parse_qs = compat_parse_qs From 2190e892603d238e1a1fc40477bf30d131c22acc Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 7 Apr 2025 16:00:11 +0100 Subject: [PATCH 61/78] [utils] Support optional `safe` argument for `escape_rfc3986()` --- youtube_dl/utils.py | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index ac1e78002..c4262936e 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -4204,12 +4204,16 @@ def lowercase_escape(s): s) -def escape_rfc3986(s): +def escape_rfc3986(s, safe=None): """Escape non-ASCII characters as suggested by RFC 3986""" if sys.version_info < (3, 0): s = _encode_compat_str(s, 'utf-8') + if safe is not None: + safe = _encode_compat_str(safe, 'utf-8') + if safe is None: + safe = b"%/;:@&=+$,!~*'()?#[]" # ensure unicode: after quoting, it can always be converted - return compat_str(compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")) + return compat_str(compat_urllib_parse.quote(s, safe)) def escape_url(url): From 570b8680780eba23d9e991eff99983af499f7ff0 Mon Sep 17 00:00:00 2001 From: dirkf Date: Mon, 7 Apr 2025 16:02:10 +0100 Subject: [PATCH 62/78] [cache] Use `esc_rfc3986` to encode cache key --- youtube_dl/cache.py | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/youtube_dl/cache.py b/youtube_dl/cache.py index 6169461b9..54c24b7e4 100644 --- a/youtube_dl/cache.py +++ b/youtube_dl/cache.py @@ -12,10 +12,10 @@ from .compat import ( compat_getenv, compat_open as open, compat_os_makedirs, - compat_urllib_parse, ) from .utils import ( error_to_compat_str, + escape_rfc3986, expand_path, is_outdated_version, traverse_obj, @@ -55,8 +55,7 @@ class Cache(object): def _get_cache_fn(self, section, key, dtype): assert re.match(r'^[\w.-]+$', section), \ 'invalid section %r' % section - key = compat_urllib_parse.quote( - key, safe='').replace('%', ',') # encode non-ascii characters + key = escape_rfc3986(key, safe='').replace('%', ',') # encode non-ascii characters return os.path.join( self._get_root_dir(), section, '%s.%s' % (key, dtype)) From 32b8d317805b7f947b62af812e64789d1c5728a0 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 10:55:32 +0100 Subject: [PATCH 63/78] [YouTube] Support shorts playlist * only 1..100: yt-dlp/yt-dlp#11130 --- youtube_dl/extractor/youtube.py | 39 ++++++++++++++++++++++++++++----- 1 file changed, 33 insertions(+), 6 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index ce97fd75b..730e50d8e 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -3339,6 +3339,20 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'thumbnailViewModel', 'image'), final_key='sources'), }) + def _extract_shorts_lockup_view_model(self, view_model): + content_id = traverse_obj(view_model, ( + 'onTap', 'innertubeCommand', 'reelWatchEndpoint', 'videoId', + T(lambda v: v if YoutubeIE.suitable(v) else None))) + if not content_id: + return + return merge_dicts(self.url_result( + content_id, ie=YoutubeIE.ie_key(), video_id=content_id), { + 'title': traverse_obj(view_model, ( + 'overlayMetadata', 'primaryText', 'content', T(compat_str))), + 'thumbnails': self._extract_thumbnails( + view_model, 'thumbnail', final_key='sources'), + }) + def _video_entry(self, video_renderer): video_id = video_renderer.get('videoId') if video_id: @@ -3385,10 +3399,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): yield entry def _rich_grid_entries(self, contents): - for content in contents: - content = traverse_obj( - content, ('richItemRenderer', 'content'), - expected_type=dict) or {} + for content in traverse_obj( + contents, (Ellipsis, 'richItemRenderer', 'content'), + expected_type=dict): video_renderer = traverse_obj( content, 'videoRenderer', 'reelItemRenderer', expected_type=dict) @@ -3396,6 +3409,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): entry = self._video_entry(video_renderer) if entry: yield entry + # shorts item + shorts_lockup_view_model = content.get('shortsLockupViewModel') + if shorts_lockup_view_model: + entry = self._extract_shorts_lockup_view_model(shorts_lockup_view_model) + if entry: + yield entry # playlist renderer = traverse_obj( content, 'playlistRenderer', expected_type=dict) or {} @@ -3499,6 +3518,13 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): entry = self._video_entry(renderer) if entry: yield entry + renderer = isr_content.get('richGridRenderer') + if renderer: + for from_ in self._rich_grid_entries( + traverse_obj(renderer, ('contents', Ellipsis, T(dict)))): + yield from_ + continuation = self._extract_continuation(renderer) + continue if not continuation: continuation = self._extract_continuation(is_renderer) @@ -3508,8 +3534,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): rich_grid_renderer = tab_content.get('richGridRenderer') if not rich_grid_renderer: return - for entry in self._rich_grid_entries(rich_grid_renderer.get('contents') or []): - yield entry + for from_ in self._rich_grid_entries( + traverse_obj(rich_grid_renderer, ('contents', Ellipsis, T(dict)))): + yield from_ continuation = self._extract_continuation(rich_grid_renderer) From 1b08d3281d3343a62007aacdc5b669c263bcebeb Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 10:58:01 +0100 Subject: [PATCH 64/78] [YouTube] Fix playlist continuation extraction * thx coletdjnz, bashonly: yt-dlp/yt-dlp#12777 --- youtube_dl/extractor/youtube.py | 28 ++++++++++++---------------- 1 file changed, 12 insertions(+), 16 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 730e50d8e..556ef7577 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -3453,23 +3453,15 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): next_continuation = cls._extract_next_continuation_data(renderer) if next_continuation: return next_continuation - contents = [] - for key in ('contents', 'items'): - contents.extend(try_get(renderer, lambda x: x[key], list) or []) - for content in contents: - if not isinstance(content, dict): - continue - continuation_ep = try_get( - content, lambda x: x['continuationItemRenderer']['continuationEndpoint'], - dict) - if not continuation_ep: - continue - continuation = try_get( - continuation_ep, lambda x: x['continuationCommand']['token'], compat_str) + for command in traverse_obj(renderer, ( + ('contents', 'items', 'rows'), Ellipsis, 'continuationItemRenderer', + ('continuationEndpoint', ('button', 'buttonRenderer', 'command')), + (('commandExecutorCommand', 'commands', Ellipsis), None), T(dict))): + continuation = traverse_obj(command, ('continuationCommand', 'token', T(compat_str))) if not continuation: continue - ctp = continuation_ep.get('clickTrackingParams') - return YoutubeTabIE._build_continuation_query(continuation, ctp) + ctp = command.get('clickTrackingParams') + return cls._build_continuation_query(continuation, ctp) def _entries(self, tab, item_id, webpage): tab_content = try_get(tab, lambda x: x['content'], dict) @@ -3582,8 +3574,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): # Downloading page may result in intermittent 5xx HTTP error # that is usually worked around with a retry response = self._download_json( - 'https://www.youtube.com/youtubei/v1/browse?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'https://www.youtube.com/youtubei/v1/browse', None, 'Downloading page %d%s' % (page_num, ' (retry #%d)' % count if count else ''), + query={ + # 'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'prettyPrint': 'false', + }, headers=headers, data=json.dumps(data).encode('utf8')) break except ExtractorError as e: From 6c40d9f8472b5e98ca3a3880cfce01a8403f3a9f Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 11:10:29 +0100 Subject: [PATCH 65/78] [YouTube] Remove remaining hard-coded API keys * no longer required for these cases --- youtube_dl/extractor/youtube.py | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 556ef7577..7f8c0c532 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -342,14 +342,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): if not self._login(): return - _DEFAULT_API_DATA = { - 'context': { - 'client': { - 'clientName': 'WEB', - 'clientVersion': '2.20201021.03.00', - }, - }, - } + _DEFAULT_API_DATA = {'context': _INNERTUBE_CLIENTS['web']['INNERTUBE_CONTEXT']} _YT_INITIAL_DATA_RE = r'(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*=\s*({.+?})\s*;' _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*=\s*({.+?})\s*;' @@ -497,11 +490,15 @@ class YoutubeBaseInfoExtractor(InfoExtractor): data['params'] = params for page_num in itertools.count(1): search = self._download_json( - 'https://www.youtube.com/youtubei/v1/search?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'https://www.youtube.com/youtubei/v1/search', video_id='query "%s"' % query, note='Downloading page %s' % page_num, errnote='Unable to download API page', fatal=False, data=json.dumps(data).encode('utf8'), + query={ + # 'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'prettyPrint': 'false', + }, headers={'content-type': 'application/json'}) if not search: break From 6eb6d6dff5add8b17babebb3a1a3da7948e5ee18 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 11:36:35 +0100 Subject: [PATCH 66/78] [InfoExtractor] Use local variants for remaining parent method calls * ... where defined --- youtube_dl/cache.py | 6 +-- youtube_dl/extractor/common.py | 70 +++++++++++++++++----------------- 2 files changed, 38 insertions(+), 38 deletions(-) diff --git a/youtube_dl/cache.py b/youtube_dl/cache.py index 54c24b7e4..ba12d0373 100644 --- a/youtube_dl/cache.py +++ b/youtube_dl/cache.py @@ -42,11 +42,11 @@ class Cache(object): def _to_screen(self, *args, **kwargs): self._ydl.to_screen(*args, **kwargs) - def _get_params(self, k, default=None): + def _get_param(self, k, default=None): return self._ydl.params.get(k, default) def _get_root_dir(self): - res = self._get_params('cachedir') + res = self._get_param('cachedir') if res is None: cache_root = compat_getenv('XDG_CACHE_HOME', '~/.cache') res = os.path.join(cache_root, self._YTDL_DIR) @@ -61,7 +61,7 @@ class Cache(object): @property def enabled(self): - return self._get_params('cachedir') is not False + return self._get_param('cachedir') is not False def store(self, section, key, data, dtype='json'): assert dtype in ('json',) diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index cbc4e968a..6f4a48bcd 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -505,7 +505,7 @@ class InfoExtractor(object): if not self._x_forwarded_for_ip: # Geo bypass mechanism is explicitly disabled by user - if not self._downloader.params.get('geo_bypass', True): + if not self.get_param('geo_bypass', True): return if not geo_bypass_context: @@ -527,7 +527,7 @@ class InfoExtractor(object): # Explicit IP block specified by user, use it right away # regardless of whether extractor is geo bypassable or not - ip_block = self._downloader.params.get('geo_bypass_ip_block', None) + ip_block = self.get_param('geo_bypass_ip_block', None) # Otherwise use random IP block from geo bypass context but only # if extractor is known as geo bypassable @@ -538,8 +538,8 @@ class InfoExtractor(object): if ip_block: self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block) - if self._downloader.params.get('verbose', False): - self._downloader.to_screen( + if self.get_param('verbose', False): + self.to_screen( '[debug] Using fake IP %s as X-Forwarded-For.' % self._x_forwarded_for_ip) return @@ -548,7 +548,7 @@ class InfoExtractor(object): # Explicit country code specified by user, use it right away # regardless of whether extractor is geo bypassable or not - country = self._downloader.params.get('geo_bypass_country', None) + country = self.get_param('geo_bypass_country', None) # Otherwise use random country code from geo bypass context but # only if extractor is known as geo bypassable @@ -559,8 +559,8 @@ class InfoExtractor(object): if country: self._x_forwarded_for_ip = GeoUtils.random_ipv4(country) - if self._downloader.params.get('verbose', False): - self._downloader.to_screen( + if self.get_param('verbose', False): + self.to_screen( '[debug] Using fake IP %s (%s) as X-Forwarded-For.' % (self._x_forwarded_for_ip, country.upper())) @@ -586,9 +586,9 @@ class InfoExtractor(object): raise ExtractorError('An extractor error has occurred.', cause=e) def __maybe_fake_ip_and_retry(self, countries): - if (not self._downloader.params.get('geo_bypass_country', None) + if (not self.get_param('geo_bypass_country', None) and self._GEO_BYPASS - and self._downloader.params.get('geo_bypass', True) + and self.get_param('geo_bypass', True) and not self._x_forwarded_for_ip and countries): country_code = random.choice(countries) @@ -698,7 +698,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(errmsg, sys.exc_info()[2], cause=err) else: - self._downloader.report_warning(errmsg) + self.report_warning(errmsg) return False def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None): @@ -770,11 +770,11 @@ class InfoExtractor(object): webpage_bytes = prefix + webpage_bytes if not encoding: encoding = self._guess_encoding_from_content(content_type, webpage_bytes) - if self._downloader.params.get('dump_intermediate_pages', False): + if self.get_param('dump_intermediate_pages', False): self.to_screen('Dumping request to ' + urlh.geturl()) dump = base64.b64encode(webpage_bytes).decode('ascii') - self._downloader.to_screen(dump) - if self._downloader.params.get('write_pages', False): + self.to_screen(dump) + if self.get_param('write_pages', False): basen = '%s_%s' % (video_id, urlh.geturl()) if len(basen) > 240: h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest() @@ -1074,7 +1074,7 @@ class InfoExtractor(object): if mobj: break - if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): + if not self.get_param('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): _name = '\033[0;34m%s\033[0m' % name else: _name = name @@ -1092,7 +1092,7 @@ class InfoExtractor(object): elif fatal: raise RegexNotFoundError('Unable to extract %s' % _name) else: - self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message()) + self.report_warning('unable to extract %s' % _name + bug_reports_message()) return None def _search_json(self, start_pattern, string, name, video_id, **kwargs): @@ -1162,7 +1162,7 @@ class InfoExtractor(object): username = None password = None - if self._downloader.params.get('usenetrc', False): + if self.get_param('usenetrc', False): try: netrc_machine = netrc_machine or self._NETRC_MACHINE info = netrc.netrc().authenticators(netrc_machine) @@ -1173,7 +1173,7 @@ class InfoExtractor(object): raise netrc.NetrcParseError( 'No authenticators for %s' % netrc_machine) except (AttributeError, IOError, netrc.NetrcParseError) as err: - self._downloader.report_warning( + self.report_warning( 'parsing .netrc: %s' % error_to_compat_str(err)) return username, password @@ -1210,10 +1210,10 @@ class InfoExtractor(object): """ if self._downloader is None: return None - downloader_params = self._downloader.params - if downloader_params.get('twofactor') is not None: - return downloader_params['twofactor'] + twofactor = self.get_param('twofactor') + if twofactor is not None: + return twofactor return compat_getpass('Type %s and press [Return]: ' % note) @@ -1348,7 +1348,7 @@ class InfoExtractor(object): elif fatal: raise RegexNotFoundError('Unable to extract JSON-LD') else: - self._downloader.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) + self.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) return {} def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None): @@ -1579,7 +1579,7 @@ class InfoExtractor(object): if f.get('vcodec') == 'none': # audio only preference -= 50 - if self._downloader.params.get('prefer_free_formats'): + if self.get_param('prefer_free_formats'): ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus'] else: ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a'] @@ -1591,7 +1591,7 @@ class InfoExtractor(object): else: if f.get('acodec') == 'none': # video only preference -= 40 - if self._downloader.params.get('prefer_free_formats'): + if self.get_param('prefer_free_formats'): ORDER = ['flv', 'mp4', 'webm'] else: ORDER = ['webm', 'flv', 'mp4'] @@ -1657,7 +1657,7 @@ class InfoExtractor(object): """ Either "http:" or "https:", depending on the user's preferences """ return ( 'http:' - if self._downloader.params.get('prefer_insecure', False) + if self.get_param('prefer_insecure', False) else 'https:') def _proto_relative_url(self, url, scheme=None): @@ -3189,7 +3189,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(msg) else: - self._downloader.report_warning(msg) + self.report_warning(msg) return res def _float(self, v, name, fatal=False, **kwargs): @@ -3199,7 +3199,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(msg) else: - self._downloader.report_warning(msg) + self.report_warning(msg) return res def _set_cookie(self, domain, name, value, expire_time=None, port=None, @@ -3208,12 +3208,12 @@ class InfoExtractor(object): 0, name, value, port, port is not None, domain, True, domain.startswith('.'), path, True, secure, expire_time, discard, None, None, rest) - self._downloader.cookiejar.set_cookie(cookie) + self.cookiejar.set_cookie(cookie) def _get_cookies(self, url): """ Return a compat_cookies_SimpleCookie with the cookies for the url """ req = sanitized_Request(url) - self._downloader.cookiejar.add_cookie_header(req) + self.cookiejar.add_cookie_header(req) return compat_cookies_SimpleCookie(req.get_header('Cookie')) def _apply_first_set_cookie_header(self, url_handle, cookie): @@ -3273,8 +3273,8 @@ class InfoExtractor(object): return not any_restricted def extract_subtitles(self, *args, **kwargs): - if (self._downloader.params.get('writesubtitles', False) - or self._downloader.params.get('listsubtitles')): + if (self.get_param('writesubtitles', False) + or self.get_param('listsubtitles')): return self._get_subtitles(*args, **kwargs) return {} @@ -3303,8 +3303,8 @@ class InfoExtractor(object): return target def extract_automatic_captions(self, *args, **kwargs): - if (self._downloader.params.get('writeautomaticsub', False) - or self._downloader.params.get('listsubtitles')): + if (self.get_param('writeautomaticsub', False) + or self.get_param('listsubtitles')): return self._get_automatic_captions(*args, **kwargs) return {} @@ -3312,9 +3312,9 @@ class InfoExtractor(object): raise NotImplementedError('This method must be implemented by subclasses') def mark_watched(self, *args, **kwargs): - if (self._downloader.params.get('mark_watched', False) + if (self.get_param('mark_watched', False) and (self._get_login_info()[0] is not None - or self._downloader.params.get('cookiefile') is not None)): + or self.get_param('cookiefile') is not None)): self._mark_watched(*args, **kwargs) def _mark_watched(self, *args, **kwargs): @@ -3322,7 +3322,7 @@ class InfoExtractor(object): def geo_verification_headers(self): headers = {} - geo_verification_proxy = self._downloader.params.get('geo_verification_proxy') + geo_verification_proxy = self.get_param('geo_verification_proxy') if geo_verification_proxy: headers['Ytdl-request-proxy'] = geo_verification_proxy return headers From 67dbfa65f220e358de713977bed8d818059f7f27 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 11:43:32 +0100 Subject: [PATCH 67/78] [InfoExtractor] Fix merging subtitles to empty target --- youtube_dl/extractor/common.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 6f4a48bcd..a64fcfccc 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -3295,7 +3295,11 @@ class InfoExtractor(object): """ Merge subtitle dictionaries, language by language. """ # ..., * , target=None - target = kwargs.get('target') or dict(subtitle_dict1) + target = kwargs.get('target') + if target is None: + target = dict(subtitle_dict1) + else: + subtitle_dicts = (subtitle_dict1,) + subtitle_dicts for subtitle_dict in subtitle_dicts: for lang in subtitle_dict: From 75134137947d1cd58b73f4186b3693d032c5bb66 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 12:15:38 +0100 Subject: [PATCH 68/78] [JSInterp] Reorganise some declarations to align better with yt-dlp --- test/test_jsinterp.py | 39 ++++++++++++++++++++------------------- youtube_dl/jsinterp.py | 22 +++++++++++----------- 2 files changed, 31 insertions(+), 30 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 3c9650ab6..b0ac3a704 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -7,6 +7,7 @@ from __future__ import unicode_literals import os import sys import unittest + sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) import math @@ -146,6 +147,25 @@ class TestJSInterpreter(unittest.TestCase): # https://github.com/ytdl-org/youtube-dl/issues/32815 self._test('function f(){return 0 - 7 * - 6;}', 42) + def test_bitwise_operators_typecast(self): + # madness + self._test('function f(){return null << 5}', 0) + self._test('function f(){return undefined >> 5}', 0) + self._test('function f(){return 42 << NaN}', 42) + self._test('function f(){return 42 << Infinity}', 42) + self._test('function f(){return 0.0 << null}', 0) + self._test('function f(){return NaN << 42}', 0) + self._test('function f(){return "21.9" << 1}', 42) + self._test('function f(){return true << "5";}', 32) + self._test('function f(){return true << true;}', 2) + self._test('function f(){return "19" & "21.9";}', 17) + self._test('function f(){return "19" & false;}', 0) + self._test('function f(){return "11.0" >> "2.1";}', 2) + self._test('function f(){return 5 ^ 9;}', 12) + self._test('function f(){return 0.0 << NaN}', 0) + self._test('function f(){return null << undefined}', 0) + self._test('function f(){return 21 << 4294967297}', 42) + def test_array_access(self): self._test('function f(){var x = [1,2,3]; x[0] = 4; x[0] = 5; x[2.0] = 7; return x;}', [5, 2, 7]) @@ -482,25 +502,6 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return -524999584 << 5}', 379882496) self._test('function f(){return 1236566549 << 5}', 915423904) - def test_bitwise_operators_typecast(self): - # madness - self._test('function f(){return null << 5}', 0) - self._test('function f(){return undefined >> 5}', 0) - self._test('function f(){return 42 << NaN}', 42) - self._test('function f(){return 42 << Infinity}', 42) - self._test('function f(){return 0.0 << null}', 0) - self._test('function f(){return NaN << 42}', 0) - self._test('function f(){return "21.9" << 1}', 42) - self._test('function f(){return 21 << 4294967297}', 42) - self._test('function f(){return true << "5";}', 32) - self._test('function f(){return true << true;}', 2) - self._test('function f(){return "19" & "21.9";}', 17) - self._test('function f(){return "19" & false;}', 0) - self._test('function f(){return "11.0" >> "2.1";}', 2) - self._test('function f(){return 5 ^ 9;}', 12) - self._test('function f(){return 0.0 << NaN}', 0) - self._test('function f(){return null << undefined}', 0) - def test_negative(self): self._test('function f(){return 2 * -2.0 ;}', -4) self._test('function f(){return 2 - - -2 ;}', 0) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 69c8f77ca..971387df2 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -283,17 +283,6 @@ _OPERATORS = ( ('**', _js_exp), ) -_COMP_OPERATORS = ( - ('===', _js_id_op(operator.is_)), - ('!==', _js_id_op(operator.is_not)), - ('==', _js_eq), - ('!=', _js_neq), - ('<=', _js_comp_op(operator.le)), - ('>=', _js_comp_op(operator.ge)), - ('<', _js_comp_op(operator.lt)), - ('>', _js_comp_op(operator.gt)), -) - _LOG_OPERATORS = ( ('|', _js_bit_op(operator.or_)), ('^', _js_bit_op(operator.xor)), @@ -314,6 +303,17 @@ _UNARY_OPERATORS_X = ( _OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS)) +_COMP_OPERATORS = ( + ('===', _js_id_op(operator.is_)), + ('!==', _js_id_op(operator.is_not)), + ('==', _js_eq), + ('!=', _js_neq), + ('<=', _js_comp_op(operator.le)), + ('>=', _js_comp_op(operator.ge)), + ('<', _js_comp_op(operator.lt)), + ('>', _js_comp_op(operator.gt)), +) + _NAME_RE = r'[a-zA-Z_$][\w$]*' _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]'))) _QUOTES = '\'"/' From d21717978cff5321ca279593340051d4bc73c129 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 12:20:15 +0100 Subject: [PATCH 69/78] [JSInterp] Improve JS classes, etc --- test/test_jsinterp.py | 1 + youtube_dl/jsinterp.py | 85 ++++++++++++++++++++++++++++++++++-------- 2 files changed, 71 insertions(+), 15 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index b0ac3a704..30c48adfc 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -455,6 +455,7 @@ class TestJSInterpreter(unittest.TestCase): def test_regex(self): self._test('function f() { let a=/,,[/,913,/](,)}/; }', None) + self._test('function f() { let a=/,,[/,913,/](,)}/; return a.source; }', ',,[/,913,/](,)}') jsi = JSInterpreter(''' function x() { let a=/,,[/,913,/](,)}/; "".replace(a, ""); return a; } diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 971387df2..4a464743f 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -353,7 +353,7 @@ class LocalNameSpace(ChainMap): raise NotImplementedError('Deleting is not supported') def __repr__(self): - return 'LocalNameSpace%s' % (self.maps, ) + return 'LocalNameSpace({0!r})'.format(self.maps) class Debugger(object): @@ -374,6 +374,9 @@ class Debugger(object): @classmethod def wrap_interpreter(cls, f): + if not cls.ENABLED: + return f + @wraps(f) def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs): if cls.ENABLED and stmt.strip(): @@ -414,7 +417,17 @@ class JSInterpreter(object): msg = '{0} in: {1!r:.100}'.format(msg.rstrip(), expr) super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs) - class JS_RegExp(object): + class JS_Object(object): + def __getitem__(self, key): + if hasattr(self, key): + return getattr(self, key) + raise KeyError(key) + + def dump(self): + """Serialise the instance""" + raise NotImplementedError + + class JS_RegExp(JS_Object): RE_FLAGS = { # special knowledge: Python's re flags are bitmask values, current max 128 # invent new bitmask values well above that for literal parsing @@ -435,16 +448,24 @@ class JSInterpreter(object): def __init__(self, pattern_txt, flags=0): if isinstance(flags, compat_str): flags, _ = self.regex_flags(flags) - # First, avoid https://github.com/python/cpython/issues/74534 self.__self = None pattern_txt = str_or_none(pattern_txt) or '(?:)' - self.__pattern_txt = pattern_txt.replace('[[', r'[\[') + # escape unintended embedded flags + pattern_txt = re.sub( + r'(\(\?)([aiLmsux]*)(-[imsx]+:|(? Date: Fri, 4 Apr 2025 12:25:17 +0100 Subject: [PATCH 70/78] [JSInterp] Add `_separate_at_op()` --- youtube_dl/jsinterp.py | 126 ++++++++++++++++++++++++----------------- 1 file changed, 75 insertions(+), 51 deletions(-) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 4a464743f..3824a0f6b 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -704,6 +704,68 @@ class JSInterpreter(object): _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS, _UNARY_OPERATORS_X)) return _cached + def _separate_at_op(self, expr, max_split=None): + + for op, _ in self._all_operators(): + # hackety: have higher priority than <>, but don't confuse them + skip_delim = (op + op) if op in '<>*?' else None + if op == '?': + skip_delim = (skip_delim, '?.') + separated = list(self._separate(expr, op, skip_delims=skip_delim)) + if len(separated) < 2: + continue + + right_expr = separated.pop() + # handle operators that are both unary and binary, minimal BODMAS + if op in ('+', '-'): + # simplify/adjust consecutive instances of these operators + undone = 0 + separated = [s.strip() for s in separated] + while len(separated) > 1 and not separated[-1]: + undone += 1 + separated.pop() + if op == '-' and undone % 2 != 0: + right_expr = op + right_expr + elif op == '+': + while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: + right_expr = separated.pop() + right_expr + if separated[-1][-1:] in self.OP_CHARS: + right_expr = separated.pop() + right_expr + # hanging op at end of left => unary + (strip) or - (push right) + separated.append(right_expr) + dm_ops = ('*', '%', '/', '**') + dm_chars = set(''.join(dm_ops)) + + def yield_terms(s): + skip = False + for i, term in enumerate(s[:-1]): + if skip: + skip = False + continue + if not (dm_chars & set(term)): + yield term + continue + for dm_op in dm_ops: + bodmas = list(self._separate(term, dm_op, skip_delims=skip_delim)) + if len(bodmas) > 1 and not bodmas[-1].strip(): + bodmas[-1] = (op if op == '-' else '') + s[i + 1] + yield dm_op.join(bodmas) + skip = True + break + else: + if term: + yield term + + if not skip and s[-1]: + yield s[-1] + + separated = list(yield_terms(separated)) + right_expr = separated.pop() if len(separated) > 1 else None + expr = op.join(separated) + if right_expr is None: + continue + return op, separated, right_expr + def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): if op in ('||', '&&'): if (op == '&&') ^ _js_ternary(left_val): @@ -759,51 +821,9 @@ class JSInterpreter(object): _FINALLY_RE = re.compile(r'finally\s*\{') _SWITCH_RE = re.compile(r'switch\s*\(') - def handle_operators(self, expr, local_vars, allow_recursion): - - for op, _ in self._all_operators(): - # hackety: have higher priority than <>, but don't confuse them - skip_delim = (op + op) if op in '<>*?' else None - if op == '?': - skip_delim = (skip_delim, '?.') - separated = list(self._separate(expr, op, skip_delims=skip_delim)) - if len(separated) < 2: - continue - - right_expr = separated.pop() - # handle operators that are both unary and binary, minimal BODMAS - if op in ('+', '-'): - # simplify/adjust consecutive instances of these operators - undone = 0 - separated = [s.strip() for s in separated] - while len(separated) > 1 and not separated[-1]: - undone += 1 - separated.pop() - if op == '-' and undone % 2 != 0: - right_expr = op + right_expr - elif op == '+': - while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: - right_expr = separated.pop() + right_expr - if separated[-1][-1:] in self.OP_CHARS: - right_expr = separated.pop() + right_expr - # hanging op at end of left => unary + (strip) or - (push right) - left_val = separated[-1] if separated else '' - for dm_op in ('*', '%', '/', '**'): - bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim)) - if len(bodmas) > 1 and not bodmas[-1].strip(): - expr = op.join(separated) + op + right_expr - if len(separated) > 1: - separated.pop() - right_expr = op.join((left_val, right_expr)) - else: - separated = [op.join((left_val, right_expr))] - right_expr = None - break - if right_expr is None: - continue - - left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) - return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), True + def _eval_operator(self, op, left_expr, right_expr, expr, local_vars, allow_recursion): + left_val = self.interpret_expression(left_expr, local_vars, allow_recursion) + return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion) @Debugger.wrap_interpreter def interpret_statement(self, stmt, local_vars, allow_recursion=100): @@ -865,9 +885,12 @@ class JSInterpreter(object): operand = expr[len(op):] if not operand or operand[0] != ' ': continue - op_result = self.handle_operators(expr, local_vars, allow_recursion) - if op_result: - return op_result[0], should_return + separated = self._separate_at_op(operand, max_split=1) + if separated: + next_op, separated, right_expr = separated + separated.append(right_expr) + operand = next_op.join(separated) + return self._eval_operator(op, operand, '', expr, local_vars, allow_recursion), should_return if expr.startswith('{'): inner, outer = self._separate_at_paren(expr) @@ -1138,9 +1161,10 @@ class JSInterpreter(object): val = self._index(val, idx) return val, should_return - op_result = self.handle_operators(expr, local_vars, allow_recursion) - if op_result: - return op_result[0], should_return + separated = self._separate_at_op(expr) + if separated: + op, separated, right_expr = separated + return self._eval_operator(op, op.join(separated), right_expr, expr, local_vars, allow_recursion), should_return if md.get('attribute'): variable, member, nullish = m.group('var', 'member', 'nullish') From bd2ded59f218bed637514f3aaf11787fa7b87ccf Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 12:29:28 +0100 Subject: [PATCH 71/78] [JSInterp] Improve unary operators; add `!` --- test/test_jsinterp.py | 7 +++++++ youtube_dl/jsinterp.py | 9 ++++++--- 2 files changed, 13 insertions(+), 3 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 30c48adfc..95bac75d7 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -371,6 +371,13 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f() { a=5; return (a -= 1, a+=3, a); }', 7) self._test('function f() { return (l=[0,1,2,3], function(a, b){return a+b})((l[1], l[2]), l[3]) }', 5) + def test_not(self): + self._test('function f() { return ! undefined; }', True) + self._test('function f() { return !0; }', True) + self._test('function f() { return !!0; }', False) + self._test('function f() { return ![]; }', False) + self._test('function f() { return !0 !== false; }', True) + def test_void(self): self._test('function f() { return void 42; }', JS_Undefined) diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 3824a0f6b..ff9894f52 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -240,7 +240,7 @@ def _js_ternary(cndn, if_true=True, if_false=False): def _js_unary_op(op): @wraps_op(op) - def wrapped(_, a): + def wrapped(a, _): return op(a) return wrapped @@ -299,6 +299,8 @@ _SC_OPERATORS = ( _UNARY_OPERATORS_X = ( ('void', _js_unary_op(lambda _: JS_Undefined)), ('typeof', _js_unary_op(_js_typeof)), + # avoid functools.partial here since Py2 update_wrapper(partial) -> no __module__ + ('!', _js_unary_op(lambda x: _js_ternary(x, if_true=False, if_false=True))), ) _OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS)) @@ -776,7 +778,7 @@ class JSInterpreter(object): elif op == '?': right_expr = _js_ternary(left_val, *self._separate(right_expr, ':', 1)) - right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) + right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) if right_expr else left_val opfunc = op and next((v for k, v in self._all_operators() if k == op), None) if not opfunc: return right_val @@ -879,11 +881,12 @@ class JSInterpreter(object): else: raise self.Exception('Unsupported object {obj:.100}'.format(**locals()), expr=expr) + # apply unary operators (see new above) for op, _ in _UNARY_OPERATORS_X: if not expr.startswith(op): continue operand = expr[len(op):] - if not operand or operand[0] != ' ': + if not operand or (op.isalpha() and operand[0] != ' '): continue separated = self._separate_at_op(operand, max_split=1) if separated: From c052a16f72af7dd7671d4dd62826de71cd99dfb6 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 4 Apr 2025 12:53:28 +0100 Subject: [PATCH 72/78] [JSInterp] Add tests and relevant functionality from yt-dlp * thx seproDev, bashonly: yt-dlp/yt-dlp#12760, yt-dlp/yt-dlp#12761: - Improve nested attribute support - Pass global stack when extracting objects - interpret_statement: Match attribute before indexing - Fix assignment to array elements with nested brackets - Add new signature tests - Invalidate JS function cache - Avoid testdata dupes now that we cache by URL * rework nsig function name search * fully fixes #33102 * update cache required versions * update program version --- test/test_jsinterp.py | 16 +++++++++ test/test_youtube_signature.py | 64 +++++++++++++++++++++++++++++++-- youtube_dl/extractor/youtube.py | 33 ++++++++++++++--- youtube_dl/jsinterp.py | 36 +++++++++++-------- youtube_dl/version.py | 2 +- 5 files changed, 129 insertions(+), 22 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 95bac75d7..479cb43a0 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -180,6 +180,7 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){var x = 20; x = 30 + 1; return x;}', 31) self._test('function f(){var x = 20; x += 30 + 1; return x;}', 51) self._test('function f(){var x = 20; x -= 30 + 1; return x;}', -11) + self._test('function f(){var x = 2; var y = ["a", "b"]; y[x%y["length"]]="z"; return y}', ['z', 'b']) def test_comments(self): self._test(''' @@ -552,6 +553,8 @@ class TestJSInterpreter(unittest.TestCase): test_result = list('test') tests = [ 'function f(a, b){return a.split(b)}', + 'function f(a, b){return a["split"](b)}', + 'function f(a, b){let x = ["split"]; return a[x[0]](b)}', 'function f(a, b){return String.prototype.split.call(a, b)}', 'function f(a, b){return String.prototype.split.apply(a, [b])}', ] @@ -602,6 +605,9 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return "012345678".slice(-1, 1)}', '') self._test('function f(){return "012345678".slice(-3, -1)}', '67') + def test_splice(self): + self._test('function f(){var T = ["0", "1", "2"]; T["splice"](2, 1, "0")[0]; return T }', ['0', '1', '0']) + def test_pop(self): # pop self._test('function f(){var a = [0, 1, 2, 3, 4, 5, 6, 7, 8]; return [a.pop(), a]}', @@ -636,6 +642,16 @@ class TestJSInterpreter(unittest.TestCase): 'return [ret.length, ret[0][0], ret[1][1], ret[0][2]]}', [2, 4, 1, [4, 2]]) + def test_extract_function(self): + jsi = JSInterpreter('function a(b) { return b + 1; }') + func = jsi.extract_function('a') + self.assertEqual(func([2]), 3) + + def test_extract_function_with_global_stack(self): + jsi = JSInterpreter('function c(d) { return d + e + f + g; }') + func = jsi.extract_function('c', {'e': 10}, {'f': 100, 'g': 1000}) + self.assertEqual(func([1]), 1111) + if __name__ == '__main__': unittest.main() diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 8fdcce8d4..a2d3a41ae 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -94,11 +94,51 @@ _SIG_TESTS = [ '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', ), + ( + 'https://www.youtube.com/s/player/363db69b/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + ), ( 'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js', '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', 'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0', ), + ( + 'https://www.youtube.com/s/player/4fcd6e4a/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0', + ), + ( + 'https://www.youtube.com/s/player/20830619/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/20830619/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-phone-en_US.vflset/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-tablet-en_US.vflset/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'IAOAOq0QJ8wRAAgXmPlOPSBkkUs1bYFYlJCfe29xx8j7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_E2u-m37KtXJoOySqa0', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/tv-player-es6.vflset/tv-player-es6.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'IAOAOq0QJ8wRAAgXmPlOPSBkkUs1bYFYlJCfe29xx8j7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_E2u-m37KtXJoOySqa0', + ), ] _NSIG_TESTS = [ @@ -272,7 +312,7 @@ _NSIG_TESTS = [ ), ( 'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js', - 'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw', + 'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA', ), ( 'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js', @@ -286,6 +326,26 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/4fcd6e4a/tv-player-ias.vflset/tv-player-ias.js', 'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A', ), + ( + 'https://www.youtube.com/s/player/20830619/tv-player-ias.vflset/tv-player-ias.js', + 'ir9-V6cdbCiyKxhr', '9YE85kNjZiS4', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-phone-en_US.vflset/base.js', + 'ir9-V6cdbCiyKxhr', '9YE85kNjZiS4', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-tablet-en_US.vflset/base.js', + 'ir9-V6cdbCiyKxhr', '9YE85kNjZiS4', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/player_ias_tce.vflset/en_US/base.js', + 'MiBYeXx_vRREbiCCmh', 'RtZYMVvmkE0JE', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/tv-player-es6.vflset/tv-player-es6.js', + 'MiBYeXx_vRREbiCCmh', 'RtZYMVvmkE0JE', + ), ] @@ -335,7 +395,7 @@ def t_factory(name, sig_func, url_pattern): test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id')) def test_func(self): - basename = 'player-{0}-{1}.js'.format(name, test_id) + basename = 'player-{0}.js'.format(test_id) fn = os.path.join(self.TESTDATA_DIR, basename) if not os.path.exists(fn): diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 7f8c0c532..54073ef86 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1652,7 +1652,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): assert os.path.basename(func_id) == func_id self.write_debug('Extracting signature function {0}'.format(func_id)) - cache_spec, code = self.cache.load('youtube-sigfuncs', func_id), None + cache_spec, code = self.cache.load('youtube-sigfuncs', func_id, min_ver='2025.04.07'), None if not cache_spec: code = self._load_player(video_id, player_url, player_id) @@ -1813,6 +1813,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return ret def _extract_n_function_name(self, jscode): + func_name, idx = None, None + # these special cases are redundant and probably obsolete (2025-04): + # they make the tests run ~10% faster without fallback warnings + r""" func_name, idx = self._search_regex( # (y=NuD(),Mw(k),q=k.Z[y]||null)&&(q=narray[idx](q),k.set(y,q),k.V||NuD(''))}}; # (R="nn"[+J.Z],mW(J),N=J.K[R]||null)&&(N=narray[idx](N),J.set(R,N))}}; @@ -1839,9 +1843,28 @@ class YoutubeIE(YoutubeBaseInfoExtractor): \(\s*[\w$]+\s*\) ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), default=(None, None)) + """ + + if not func_name: + # nfunc=function(x){...}|function nfunc(x); ... + # ... var y=[nfunc]|y[idx]=nfunc); + # obvious REs hang, so use a two-stage tactic + for m in re.finditer(r'''(?x) + [\n;]var\s(?:(?:(?!,).)+,|\s)*?(?!\d)[\w$]+(?:\[(?P\d+)\])?\s*=\s* + (?(idx)|\[\s*)(?P(?!\d)[\w$]+)(?(idx)|\s*\]) + \s*?[;\n] + ''', jscode): + func_name = self._search_regex( + r'[;,]\s*(function\s+)?({0})(?(1)|\s*=\s*function)\s*\((?!\d)[\w$]+\)\s*\{1}(?!\s*return\s)'.format( + re.escape(m.group('nfunc')), '{'), + jscode, 'Initial JS player n function name (2)', group=2, default=None) + if func_name: + idx = m.group('idx') + break + # thx bashonly: yt-dlp/yt-dlp/pull/10611 if not func_name: - self.report_warning('Falling back to generic n function search') + self.report_warning('Falling back to generic n function search', only_once=True) return self._search_regex( r'''(?xs) (?:(?<=[^\w$])|^) # instead of \b, which ignores $ @@ -1855,14 +1878,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return func_name return self._search_json( - r'var\s+{0}\s*='.format(re.escape(func_name)), jscode, + r'(?', _js_comp_op(operator.gt)), ) +_OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS + _SC_OPERATORS)) + _NAME_RE = r'[a-zA-Z_$][\w$]*' _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]'))) _QUOTES = '\'"/' +_NESTED_BRACKETS = r'[^[\]]+(?:\[[^[\]]+(?:\[[^\]]+\])?\])?' class JS_Break(ExtractorError): @@ -1088,15 +1089,18 @@ class JSInterpreter(object): m = re.match(r'''(?x) (?P - (?P{_NAME_RE})(?:\[(?P(?:.+?\]\s*\[)*.+?)\])?\s* + (?P{_NAME_RE})(?P(?:\[{_NESTED_BRACKETS}\])+)?\s* (?P{_OPERATOR_RE})? =(?!=)(?P.*)$ )|(?P (?!if|return|true|false|null|undefined|NaN|Infinity)(?P{_NAME_RE})$ - )|(?P - (?P{_NAME_RE})\[(?P(?:.+?\]\s*\[)*.+?)\]$ )|(?P - (?P{_NAME_RE})(?:(?P\?)?\.(?P[^(]+)|\[(?P[^\]]+)\])\s* + (?P{_NAME_RE})(?: + (?P\?)?\.(?P[^(]+)| + \[(?P{_NESTED_BRACKETS})\] + )\s* + )|(?P + (?P{_NAME_RE})(?P\[.+\])$ )|(?P (?P{_NAME_RE})\((?P.*)\)$ )'''.format(**globals()), expr) @@ -1111,10 +1115,11 @@ class JSInterpreter(object): elif left_val in (None, JS_Undefined): raise self.Exception('Cannot index undefined variable ' + m.group('out'), expr=expr) - indexes = re.split(r'\]\s*\[', m.group('out_idx')) - for i, idx in enumerate(indexes, 1): + indexes = md['out_idx'] + while indexes: + idx, indexes = self._separate_at_paren(indexes) idx = self.interpret_expression(idx, local_vars, allow_recursion) - if i < len(indexes): + if indexes: left_val = self._index(left_val, idx) if isinstance(idx, float): idx = int(idx) @@ -1159,7 +1164,9 @@ class JSInterpreter(object): if md.get('indexing'): val = local_vars[m.group('in')] - for idx in re.split(r'\]\s*\[', m.group('in_idx')): + indexes = m.group('in_idx') + while indexes: + idx, indexes = self._separate_at_paren(indexes) idx = self.interpret_expression(idx, local_vars, allow_recursion) val = self._index(val, idx) return val, should_return @@ -1204,7 +1211,7 @@ class JSInterpreter(object): if obj is JS_Undefined: try: if variable not in self._objects: - self._objects[variable] = self.extract_object(variable) + self._objects[variable] = self.extract_object(variable, local_vars) obj = self._objects[variable] except self.Exception: if not nullish: @@ -1215,7 +1222,7 @@ class JSInterpreter(object): # Member access if arg_str is None: - return self._index(obj, member) + return self._index(obj, member, nullish) # Function call argvals = [ @@ -1400,7 +1407,7 @@ class JSInterpreter(object): for v in self._separate(list_txt): yield self.interpret_expression(v, local_vars, allow_recursion) - def extract_object(self, objname): + def extract_object(self, objname, *global_stack): _FUNC_NAME_RE = r'''(?:{n}|"{n}"|'{n}')'''.format(n=_NAME_RE) obj = {} fields = next(filter(None, ( @@ -1421,7 +1428,8 @@ class JSInterpreter(object): fields): argnames = self.build_arglist(f.group('args')) name = remove_quotes(f.group('key')) - obj[name] = function_with_repr(self.build_function(argnames, f.group('code')), 'F<{0}>'.format(name)) + obj[name] = function_with_repr( + self.build_function(argnames, f.group('code'), *global_stack), 'F<{0}>'.format(name)) return obj diff --git a/youtube_dl/version.py b/youtube_dl/version.py index b82fbc702..c70d9d2af 100644 --- a/youtube_dl/version.py +++ b/youtube_dl/version.py @@ -1,3 +1,3 @@ from __future__ import unicode_literals -__version__ = '2021.12.17' +__version__ = '2025.04.07' From ec75141bf03b66ffc183992ec6bb50c4ff392dd8 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 2 May 2025 13:17:18 +0100 Subject: [PATCH 73/78] [Cache] Add `clear` function --- test/test_cache.py | 12 ++++++++++++ youtube_dl/cache.py | 42 +++++++++++++++++++++++++++++++----------- 2 files changed, 43 insertions(+), 11 deletions(-) diff --git a/test/test_cache.py b/test/test_cache.py index 21dfb6cfc..0431f4f15 100644 --- a/test/test_cache.py +++ b/test/test_cache.py @@ -66,6 +66,18 @@ class TestCache(unittest.TestCase): new_version = '.'.join(('%0.2d' % ((v + 1) if i == 0 else v, )) for i, v in enumerate(version_tuple(__version__))) self.assertIs(c.load('test_cache', 'k.', min_ver=new_version), None) + def test_cache_clear(self): + ydl = FakeYDL({ + 'cachedir': self.test_dir, + }) + c = Cache(ydl) + c.store('test_cache', 'k.', 'kay') + c.store('test_cache', 'l.', 'ell') + self.assertEqual(c.load('test_cache', 'k.'), 'kay') + c.clear('test_cache', 'k.') + self.assertEqual(c.load('test_cache', 'k.'), None) + self.assertEqual(c.load('test_cache', 'l.'), 'ell') + if __name__ == '__main__': unittest.main() diff --git a/youtube_dl/cache.py b/youtube_dl/cache.py index ba12d0373..eb0a729c2 100644 --- a/youtube_dl/cache.py +++ b/youtube_dl/cache.py @@ -1,6 +1,7 @@ # coding: utf-8 from __future__ import unicode_literals +import errno import json import os import re @@ -8,7 +9,6 @@ import shutil import traceback from .compat import ( - compat_contextlib_suppress, compat_getenv, compat_open as open, compat_os_makedirs, @@ -78,6 +78,22 @@ class Cache(object): tb = traceback.format_exc() self._report_warning('Writing cache to {fn!r} failed: {tb}'.format(fn=fn, tb=tb)) + def clear(self, section, key, dtype='json'): + + if not self.enabled: + return + + fn = self._get_cache_fn(section, key, dtype) + self._write_debug('Clearing {section}.{key} from cache'.format(section=section, key=key)) + try: + os.remove(fn) + except Exception as e: + if getattr(e, 'errno') == errno.ENOENT: + # file not found + return + tb = traceback.format_exc() + self._report_warning('Clearing cache from {fn!r} failed: {tb}'.format(fn=fn, tb=tb)) + def _validate(self, data, min_ver): version = traverse_obj(data, self._VERSION_KEY) if not version: # Backward compatibility @@ -94,17 +110,21 @@ class Cache(object): return default cache_fn = self._get_cache_fn(section, key, dtype) - with compat_contextlib_suppress(IOError): # If no cache available + try: + with open(cache_fn, encoding='utf-8') as cachef: + self._write_debug('Loading {section}.{key} from cache'.format(section=section, key=key), only_once=True) + return self._validate(json.load(cachef), min_ver) + except (ValueError, KeyError): try: - with open(cache_fn, encoding='utf-8') as cachef: - self._write_debug('Loading {section}.{key} from cache'.format(section=section, key=key), only_once=True) - return self._validate(json.load(cachef), min_ver) - except (ValueError, KeyError): - try: - file_size = os.path.getsize(cache_fn) - except (OSError, IOError) as oe: - file_size = error_to_compat_str(oe) - self._report_warning('Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) + file_size = 'size: %d' % os.path.getsize(cache_fn) + except (OSError, IOError) as oe: + file_size = error_to_compat_str(oe) + self._report_warning('Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) + except Exception as e: + if getattr(e, 'errno') == errno.ENOENT: + # no cache available + return + self._report_warning('Cache retrieval from %s failed' % (cache_fn,)) return default From 3a42f6ad372935733150159b0447e790c037c1af Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 2 May 2025 13:31:45 +0100 Subject: [PATCH 74/78] [YouTube] Cache signature timestamp from player JS * if the YT webpage can't be loaded, getting the `sts` requires loading the player JS: this caches it * based on yt-dlp/yt-dlp#13047, thx bashonly --- youtube_dl/extractor/youtube.py | 142 +++++++++++++++++++++++--------- 1 file changed, 104 insertions(+), 38 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 54073ef86..a1c9a6eae 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -49,6 +49,7 @@ from ..utils import ( parse_duration, parse_qs, qualities, + remove_end, remove_start, smuggle_url, str_or_none, @@ -1584,6 +1585,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): '397': {'acodec': 'none', 'vcodec': 'av01.0.05M.08'}, } + _PLAYER_JS_VARIANT_MAP = ( + ('main', 'player_ias.vflset/en_US/base.js'), + ('tce', 'player_ias_tce.vflset/en_US/base.js'), + ('tv', 'tv-player-ias.vflset/tv-player-ias.js'), + ('tv_es6', 'tv-player-es6.vflset/tv-player-es6.js'), + ('phone', 'player-plasma-ias-phone-en_US.vflset/base.js'), + ('tablet', 'player-plasma-ias-tablet-en_US.vflset/base.js'), + ) + @classmethod def suitable(cls, url): if parse_qs(url).get('list', [None])[0]: @@ -1631,36 +1641,83 @@ class YoutubeIE(YoutubeBaseInfoExtractor): raise ExtractorError( 'Cannot identify player %r' % (player_url,), cause=e) - def _load_player(self, video_id, player_url, fatal=True, player_id=None): - if not player_id: + def _player_js_cache_key(self, player_url, extra_id=None, _cache={}): + if player_url not in _cache: player_id = self._extract_player_info(player_url) - if player_id not in self._code_cache: + player_path = remove_start( + compat_urllib_parse.urlparse(player_url).path, + '/s/player/{0}/'.format(player_id)) + variant = next((k for k, v in self._PLAYER_JS_VARIANT_MAP + if v == player_path), None) + if not variant: + variant = next( + (k for k, v in self._PLAYER_JS_VARIANT_MAP + if re.match(re.escape(v).replace('en_US', r'\w+') + '$', player_path)), + None) + if not variant: + self.write_debug( + 'Unable to determine player JS variant\n' + ' player = {0}'.format(player_url), only_once=True) + variant = re.sub(r'[^a-zA-Z0-9]', '_', remove_end(player_path, '.js')) + _cache[player_url] = join_nonempty(player_id, variant) + + if extra_id: + extra_id = '-'.join((_cache[player_url], extra_id)) + assert os.path.basename(extra_id) == extra_id + return extra_id + return _cache[player_url] + + def _load_player(self, video_id, player_url, fatal=True): + player_js_key = self._player_js_cache_key(player_url) + if player_js_key not in self._code_cache: code = self._download_webpage( player_url, video_id, fatal=fatal, - note='Downloading player ' + player_id, - errnote='Download of %s failed' % player_url) + note='Downloading player {0}'.format(player_js_key), + errnote='Download of {0} failed'.format(player_url)) if code: - self._code_cache[player_id] = code - return self._code_cache[player_id] if fatal else self._code_cache.get(player_id) + self._code_cache[player_js_key] = code + return self._code_cache.get(player_js_key) + + def _load_player_data_from_cache(self, name, player_url, extra_id=None): + cache_id = ('youtube-{0}'.format(name), self._player_js_cache_key(player_url, extra_id)) + data = self._player_cache.get(cache_id) + if data: + return data + + data = self.cache.load(*cache_id, min_ver='2025.04.07') + if data: + self._player_cache[cache_id] = data + return data + + def _store_player_data_to_cache(self, name, player_url, data, extra_id=None): + cache_id = ('youtube-{0}'.format(name), self._player_js_cache_key(player_url, extra_id)) + + if cache_id not in self._player_cache: + self.cache.store(cache_id[0], cache_id[1], data) + self._player_cache[cache_id] = data def _extract_signature_function(self, video_id, player_url, example_sig): - player_id = self._extract_player_info(player_url) + # player_id = self._extract_player_info(player_url) # Read from filesystem cache - func_id = 'js_{0}_{1}'.format( - player_id, self._signature_cache_id(example_sig)) - assert os.path.basename(func_id) == func_id - - self.write_debug('Extracting signature function {0}'.format(func_id)) - cache_spec, code = self.cache.load('youtube-sigfuncs', func_id, min_ver='2025.04.07'), None + extra_id = self._signature_cache_id(example_sig) + self.write_debug('Extracting signature function {0}-{1}'.format(player_url, extra_id)) + cache_spec, code = self._load_player_data_from_cache( + 'sigfuncs', player_url, extra_id=extra_id, min_ver='2025.04.07' + ), None if not cache_spec: - code = self._load_player(video_id, player_url, player_id) - if code: - res = self._parse_sig_js(code) - test_string = ''.join(map(compat_chr, range(len(example_sig)))) - cache_spec = [ord(c) for c in res(test_string)] - self.cache.store('youtube-sigfuncs', func_id, cache_spec) + code = self._load_player(video_id, player_url) + if code: + res = self._parse_sig_js(code) + test_string = ''.join(map(compat_chr, range(len(example_sig)))) + cache_spec = [ord(c) for c in res(test_string)] + self._store_player_data_to_cache( + 'sigfuncs', player_url, cache_spec, extra_id=extra_id) + else: + self.report_warning( + 'Failed to compute signature function {0}-{1}'.format( + player_url, extra_id)) return lambda s: ''.join(s[i] for i in cache_spec) @@ -1885,22 +1942,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_n_function_code(self, video_id, player_url): player_id = self._extract_player_info(player_url) - func_code = self.cache.load('youtube-nsig', player_id, min_ver='2025.04.07') + func_code = self._load_player_data_from_cache('nsig', player_url) jscode = func_code or self._load_player(video_id, player_url) jsi = JSInterpreter(jscode) if func_code: return jsi, player_id, func_code - return self._extract_n_function_code_jsi(video_id, jsi, player_id) - def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None): + return self._extract_n_function_code_jsi(video_id, jsi, player_id, player_url) + def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None, player_url=None): func_name = self._extract_n_function_name(jsi.code) func_code = self._extract_sig_fn(jsi, func_name) - - if player_id: - self.cache.store('youtube-nsig', player_id, func_code) + if player_url: + self._store_player_data_to_cache('nsig', player_url, func_code) return jsi, player_id, func_code def _extract_n_function_from_code(self, jsi, func_code): @@ -1944,18 +2000,28 @@ class YoutubeIE(YoutubeBaseInfoExtractor): Required to tell API what sig/player version is in use. """ sts = traverse_obj(ytcfg, 'STS', expected_type=int) - if not sts: - # Attempt to extract from player - if player_url is None: - error_msg = 'Cannot extract signature timestamp without player_url.' - if fatal: - raise ExtractorError(error_msg) - self.report_warning(error_msg) - return - code = self._load_player(video_id, player_url, fatal=fatal) - sts = int_or_none(self._search_regex( - r'(?:signatureTimestamp|sts)\s*:\s*(?P[0-9]{5})', code or '', - 'JS player signature timestamp', group='sts', fatal=fatal)) + if sts: + return sts + + if not player_url: + error_msg = 'Cannot extract signature timestamp without player url' + if fatal: + raise ExtractorError(error_msg) + self.report_warning(error_msg) + return None + + sts = self._load_player_data_from_cache('sts', player_url) + if sts: + return sts + + # Attempt to extract from player + code = self._load_player(video_id, player_url, fatal=fatal) + sts = int_or_none(self._search_regex( + r'(?:signatureTimestamp|sts)\s*:\s*(?P[0-9]{5})', code or '', + 'JS player signature timestamp', group='sts', fatal=fatal)) + if sts: + self._store_player_data_to_cache('sts', player_url, sts) + return sts def _mark_watched(self, video_id, player_response): From 4a31290ae14705100814b88f158e7876f83a5b67 Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 2 May 2025 13:38:32 +0100 Subject: [PATCH 75/78] [YouTube] Delete cached problem nsig cache data on descrambling error * inspired by yt-dlp/yt-dlp#12750 --- youtube_dl/extractor/youtube.py | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index a1c9a6eae..0d6ffa3f2 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1696,6 +1696,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor): self.cache.store(cache_id[0], cache_id[1], data) self._player_cache[cache_id] = data + def _remove_player_data_from_cache(self, name, player_url, extra_id=None): + cache_id = ('youtube-{0}'.format(name), self._player_js_cache_key(player_url, extra_id)) + + if cache_id in self._player_cache: + self.cache.clear(*cache_id) + self._player_cache.pop(cache_id, None) + def _extract_signature_function(self, video_id, player_url, example_sig): # player_id = self._extract_player_info(player_url) @@ -1989,7 +1996,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): n_param = n_param[-1] n_response = decrypt_nsig(n_param)(n_param, video_id, player_url) if n_response is None: - # give up if descrambling failed + # give up and forget cached data if descrambling failed + self._remove_player_data_from_cache('nsig', player_url) break fmt['url'] = update_url_query(fmt['url'], {'n': n_response}) From 680069a14958187cf576ed91c96d2f273aea711f Mon Sep 17 00:00:00 2001 From: dirkf Date: Fri, 2 May 2025 13:49:05 +0100 Subject: [PATCH 76/78] [YouTube] Improve n-sig function extraction for player `aa3fc80b` Resolves #33123 --- test/test_youtube_signature.py | 4 ++++ youtube_dl/extractor/youtube.py | 34 +++++++++++++++++++++------------ 2 files changed, 26 insertions(+), 12 deletions(-) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index a2d3a41ae..98221b9c2 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -346,6 +346,10 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/8a8ac953/tv-player-es6.vflset/tv-player-es6.js', 'MiBYeXx_vRREbiCCmh', 'RtZYMVvmkE0JE', ), + ( + 'https://www.youtube.com/s/player/aa3fc80b/player_ias.vflset/en_US/base.js', + '0qY9dal2uzOnOGwa-48hha', 'VSh1KDfQMk-eag', + ), ] diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 0d6ffa3f2..7290ae813 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1878,6 +1878,20 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_n_function_name(self, jscode): func_name, idx = None, None + + def generic_n_function_search(func_name=None): + return self._search_regex( + r'''(?xs) + (?:(?<=[^\w$])|^) # instead of \b, which ignores $ + (?P%s)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) + \s*\{(?:(?!};).)+?(?: + ["']enhanced_except_ | + return\s*(?P"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[\w$]+ + ) + ''' % (func_name or r'(?!\d)[a-zA-Z\d_$]+',), jscode, + 'Initial JS player n function name', group='name', + default=None if func_name else NO_DEFAULT) + # these special cases are redundant and probably obsolete (2025-04): # they make the tests run ~10% faster without fallback warnings r""" @@ -1918,26 +1932,22 @@ class YoutubeIE(YoutubeBaseInfoExtractor): (?(idx)|\[\s*)(?P(?!\d)[\w$]+)(?(idx)|\s*\]) \s*?[;\n] ''', jscode): - func_name = self._search_regex( + fn = self._search_regex( r'[;,]\s*(function\s+)?({0})(?(1)|\s*=\s*function)\s*\((?!\d)[\w$]+\)\s*\{1}(?!\s*return\s)'.format( re.escape(m.group('nfunc')), '{'), jscode, 'Initial JS player n function name (2)', group=2, default=None) - if func_name: + if fn: + func_name = fn idx = m.group('idx') - break + if generic_n_function_search(func_name): + # don't look any further + break # thx bashonly: yt-dlp/yt-dlp/pull/10611 if not func_name: self.report_warning('Falling back to generic n function search', only_once=True) - return self._search_regex( - r'''(?xs) - (?:(?<=[^\w$])|^) # instead of \b, which ignores $ - (?P(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) - \s*\{(?:(?!};).)+?(?: - ["']enhanced_except_ | - return\s*(?P"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[\w$]+ - ) - ''', jscode, 'Initial JS player n function name', group='name') + return generic_n_function_search() + if not idx: return func_name From e102b9993af6defb4e03699840b01c63e5623276 Mon Sep 17 00:00:00 2001 From: dirkf Date: Sat, 3 May 2025 18:33:39 +0100 Subject: [PATCH 77/78] [workflows/ci.yml] Move pinned Ubuntu runner images from withdrawn 20.4 to 22.04 * fix consequent missing `python-is-python2` package --- .github/workflows/ci.yml | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index d3b9ae016..8234e0ccb 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -116,7 +116,7 @@ jobs: strategy: fail-fast: true matrix: - os: [ubuntu-20.04] + os: [ubuntu-22.04] python-version: ${{ fromJSON(needs.select.outputs.cpython-versions) }} python-impl: [cpython] ytdl-test-set: ${{ fromJSON(needs.select.outputs.test-set) }} @@ -133,12 +133,12 @@ jobs: ytdl-test-set: ${{ contains(needs.select.outputs.test-set, 'download') && 'download' || 'nodownload' }} run-tests-ext: bat # jython - - os: ubuntu-20.04 + - os: ubuntu-22.04 python-version: 2.7 python-impl: jython ytdl-test-set: ${{ contains(needs.select.outputs.test-set, 'core') && 'core' || 'nocore' }} run-tests-ext: sh - - os: ubuntu-20.04 + - os: ubuntu-22.04 python-version: 2.7 python-impl: jython ytdl-test-set: ${{ contains(needs.select.outputs.test-set, 'download') && 'download' || 'nodownload' }} @@ -160,7 +160,7 @@ jobs: # NB may run apt-get install in Linux uses: ytdl-org/setup-python@v1 env: - # Temporary workaround for Python 3.5 failures - May 2024 + # Temporary (?) workaround for Python 3.5 failures - May 2024 PIP_TRUSTED_HOST: "pypi.python.org pypi.org files.pythonhosted.org" with: python-version: ${{ matrix.python-version }} @@ -240,7 +240,10 @@ jobs: # install 2.7 shell: bash run: | - sudo apt-get install -y python2 python-is-python2 + # Ubuntu 22.04 no longer has python-is-python2: fetch it + curl -L "http://launchpadlibrarian.net/474693132/python-is-python2_2.7.17-4_all.deb" -o python-is-python2.deb + sudo apt-get install -y python2 + sudo dpkg --force-breaks -i python-is-python2.deb echo "PYTHONHOME=/usr" >> "$GITHUB_ENV" #-------- Python 2.6 -- - name: Set up Python 2.6 environment From a084c80f7bac9ae343075a97cc0fb2c1c96ade89 Mon Sep 17 00:00:00 2001 From: dirkf Date: Sun, 4 May 2025 12:51:54 +0100 Subject: [PATCH 78/78] [YouTube] Fix 680069a, excess `min_ver` Resolves #33125. --- youtube_dl/extractor/youtube.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 7290ae813..b31798729 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1710,8 +1710,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): extra_id = self._signature_cache_id(example_sig) self.write_debug('Extracting signature function {0}-{1}'.format(player_url, extra_id)) cache_spec, code = self._load_player_data_from_cache( - 'sigfuncs', player_url, extra_id=extra_id, min_ver='2025.04.07' - ), None + 'sigfuncs', player_url, extra_id=extra_id), None if not cache_spec: code = self._load_player(video_id, player_url)