Compare commits

...

7 Commits

Author SHA1 Message Date
xXmeme-machineXx
7853ad0848
Merge 1e9272f7ae into 4d05f84325 2024-06-27 06:36:19 +08:00
dirkf
4d05f84325 [PalcoMP3] Conform to new linter rule
* no space after @ in decorator
2024-06-20 20:03:49 +01:00
dirkf
e0094e63c3 [jsinterp] Various tweaks
* treat Infinity like NaN
* cache operator list
2024-06-20 20:03:49 +01:00
dirkf
fd8242e3ef [jsinterp] Fix and improve expression parsing
* improve BODMAS (fixes https://github.com/ytdl-org/youtube-dl/issues/32815)
* support more weird expressions with multiple unary ops
2024-06-20 20:03:49 +01:00
dirkf
ad01fa6cca [jsinterp] Add Debugger from yt-dlp
* https://github.com/yt-dlp/yt-dlp/commit/8f53dc4
* thx pukkandan
2024-06-20 20:03:49 +01:00
dirkf
2eac0fa379 [utils] Save orig_msg in ExtractorError 2024-06-20 20:03:49 +01:00
xXmeme-machineXx
1e9272f7ae Fix gfycat, for links redirected to redgifs (#25555)
Fixes #25555, #26631, #27543, #28330
2021-12-22 08:30:57 +00:00
7 changed files with 95 additions and 29 deletions

View File

@ -577,9 +577,11 @@ class TestJSInterpreter(unittest.TestCase):
def test_unary_operators(self): def test_unary_operators(self):
jsi = JSInterpreter('function f(){return 2 - - - 2;}') jsi = JSInterpreter('function f(){return 2 - - - 2;}')
self.assertEqual(jsi.call_function('f'), 0) self.assertEqual(jsi.call_function('f'), 0)
# fails jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
# jsi = JSInterpreter('function f(){return 2 + - + - - 2;}') self.assertEqual(jsi.call_function('f'), 0)
# self.assertEqual(jsi.call_function('f'), 0) # https://github.com/ytdl-org/youtube-dl/issues/32815
jsi = JSInterpreter('function f(){return 0 - 7 * - 6;}')
self.assertEqual(jsi.call_function('f'), 42)
""" # fails so far """ # fails so far
def test_packed(self): def test_packed(self):

View File

@ -158,6 +158,10 @@ _NSIG_TESTS = [
'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js', 'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
'_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ', '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
), ),
(
'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
'1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
),
] ]

View File

@ -3033,7 +3033,6 @@ class InfoExtractor(object):
transform_source=transform_source, default=None) transform_source=transform_source, default=None)
def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs): def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
# allow passing `transform_source` through to _find_jwplayer_data() # allow passing `transform_source` through to _find_jwplayer_data()
transform_source = kwargs.pop('transform_source', None) transform_source = kwargs.pop('transform_source', None)
kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {} kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}

View File

@ -8,6 +8,9 @@ from ..utils import (
qualities, qualities,
ExtractorError, ExtractorError,
) )
from ..compat import (
compat_urlparse,
)
class GfycatIE(InfoExtractor): class GfycatIE(InfoExtractor):
@ -63,15 +66,21 @@ class GfycatIE(InfoExtractor):
def _real_extract(self, url): def _real_extract(self, url):
video_id = self._match_id(url) video_id = self._match_id(url)
response = self._request_webpage(url, video_id)
final_url = response.geturl() # follow redirects
hostname = compat_urlparse.urlparse(final_url).netloc
if hostname == 'www.redgifs.com':
api_endpoint = 'https://api.redgifs.com/v1/gifs/%s' % video_id.lower()
else:
api_endpoint = 'https://api.gfycat.com/v1/gfycats/%s' % video_id
gfy = self._download_json( gfy = self._download_json(
'https://api.gfycat.com/v1/gfycats/%s' % video_id, api_endpoint,
video_id, 'Downloading video info') video_id, 'Downloading video info')
if 'error' in gfy: if 'error' in gfy:
raise ExtractorError('Gfycat said: ' + gfy['error'], expected=True) raise ExtractorError('Gfycat said: ' + gfy['error'], expected=True)
gfy = gfy['gfyItem'] gfy = gfy['gfyItem']
title = gfy.get('title') or gfy['gfyName'] title = gfy.get('title') or gfy.get('gifName') or gfy['gfyName']
description = gfy.get('description') description = gfy.get('description')
timestamp = int_or_none(gfy.get('createDate')) timestamp = int_or_none(gfy.get('createDate'))
uploader = gfy.get('userName') uploader = gfy.get('userName')

View File

@ -8,7 +8,7 @@ from ..compat import compat_str
from ..utils import ( from ..utils import (
int_or_none, int_or_none,
str_or_none, str_or_none,
try_get, traverse_obj,
) )
@ -109,7 +109,7 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
} }
name''' name'''
@ classmethod @classmethod
def suitable(cls, url): def suitable(cls, url):
return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url) return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url)
@ -118,7 +118,8 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist'] artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist']
def entries(): def entries():
for music in (try_get(artist, lambda x: x['musics']['nodes'], list) or []): for music in traverse_obj(artist, (
'musics', 'nodes', lambda _, m: m['musicID'])):
yield self._parse_music(music) yield self._parse_music(music)
return self.playlist_result( return self.playlist_result(
@ -137,7 +138,7 @@ class PalcoMP3VideoIE(PalcoMP3BaseIE):
'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande', 'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande',
'description': 'md5:7043342c09a224598e93546e98e49282', 'description': 'md5:7043342c09a224598e93546e98e49282',
'upload_date': '20161107', 'upload_date': '20161107',
'uploader_id': 'maiaramaraisaoficial', 'uploader_id': '@maiaramaraisaoficial',
'uploader': 'Maiara e Maraisa', 'uploader': 'Maiara e Maraisa',
} }
}] }]

View File

@ -14,6 +14,7 @@ from .utils import (
remove_quotes, remove_quotes,
unified_timestamp, unified_timestamp,
variadic, variadic,
write_string,
) )
from .compat import ( from .compat import (
compat_basestring, compat_basestring,
@ -53,15 +54,16 @@ def wraps_op(op):
# NB In principle NaN cannot be checked by membership. # NB In principle NaN cannot be checked by membership.
# Here all NaN values are actually this one, so _NaN is _NaN, # Here all NaN values are actually this one, so _NaN is _NaN,
# although _NaN != _NaN. # although _NaN != _NaN. Ditto Infinity.
_NaN = float('nan') _NaN = float('nan')
_Infinity = float('inf')
def _js_bit_op(op): def _js_bit_op(op):
def zeroise(x): def zeroise(x):
return 0 if x in (None, JS_Undefined, _NaN) else x return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x
@wraps_op(op) @wraps_op(op)
def wrapped(a, b): def wrapped(a, b):
@ -84,7 +86,7 @@ def _js_arith_op(op):
def _js_div(a, b): def _js_div(a, b):
if JS_Undefined in (a, b) or not (a or b): if JS_Undefined in (a, b) or not (a or b):
return _NaN return _NaN
return operator.truediv(a or 0, b) if b else float('inf') return operator.truediv(a or 0, b) if b else _Infinity
def _js_mod(a, b): def _js_mod(a, b):
@ -220,6 +222,42 @@ class LocalNameSpace(ChainMap):
return 'LocalNameSpace%s' % (self.maps, ) return 'LocalNameSpace%s' % (self.maps, )
class Debugger(object):
ENABLED = False
@staticmethod
def write(*args, **kwargs):
level = kwargs.get('level', 100)
def truncate_string(s, left, right=0):
if s is None or len(s) <= left + right:
return s
return '...'.join((s[:left - 3], s[-right:] if right else ''))
write_string('[debug] JS: {0}{1}\n'.format(
' ' * (100 - level),
' '.join(truncate_string(compat_str(x), 50, 50) for x in args)))
@classmethod
def wrap_interpreter(cls, f):
def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs):
if cls.ENABLED and stmt.strip():
cls.write(stmt, level=allow_recursion)
try:
ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs)
except Exception as e:
if cls.ENABLED:
if isinstance(e, ExtractorError):
e = e.orig_msg
cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion)
raise
if cls.ENABLED and stmt.strip():
if should_ret or not repr(ret) == stmt:
cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
return ret, should_ret
return interpret_statement
class JSInterpreter(object): class JSInterpreter(object):
__named_object_counter = 0 __named_object_counter = 0
@ -307,8 +345,7 @@ class JSInterpreter(object):
def __op_chars(cls): def __op_chars(cls):
op_chars = set(';,[') op_chars = set(';,[')
for op in cls._all_operators(): for op in cls._all_operators():
for c in op[0]: op_chars.update(op[0])
op_chars.add(c)
return op_chars return op_chars
def _named_object(self, namespace, obj): def _named_object(self, namespace, obj):
@ -326,9 +363,8 @@ class JSInterpreter(object):
# collections.Counter() is ~10% slower in both 2.7 and 3.9 # collections.Counter() is ~10% slower in both 2.7 and 3.9
counters = dict((k, 0) for k in _MATCHING_PARENS.values()) counters = dict((k, 0) for k in _MATCHING_PARENS.values())
start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1 start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
in_quote, escaping, skipping = None, False, 0 in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
after_op, in_regex_char_group = True, False skipping = 0
for idx, char in enumerate(expr): for idx, char in enumerate(expr):
paren_delta = 0 paren_delta = 0
if not in_quote: if not in_quote:
@ -382,10 +418,12 @@ class JSInterpreter(object):
return separated[0][1:].strip(), separated[1].strip() return separated[0][1:].strip(), separated[1].strip()
@staticmethod @staticmethod
def _all_operators(): def _all_operators(_cached=[]):
return itertools.chain( if not _cached:
# Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence _cached.extend(itertools.chain(
_SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS) # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
_SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS))
return _cached
def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
if op in ('||', '&&'): if op in ('||', '&&'):
@ -416,7 +454,7 @@ class JSInterpreter(object):
except Exception as e: except Exception as e:
if allow_undefined: if allow_undefined:
return JS_Undefined return JS_Undefined
raise self.Exception('Cannot get index {idx:.100}'.format(**locals()), expr=repr(obj), cause=e) raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
def _dump(self, obj, namespace): def _dump(self, obj, namespace):
try: try:
@ -438,6 +476,7 @@ class JSInterpreter(object):
_FINALLY_RE = re.compile(r'finally\s*\{') _FINALLY_RE = re.compile(r'finally\s*\{')
_SWITCH_RE = re.compile(r'switch\s*\(') _SWITCH_RE = re.compile(r'switch\s*\(')
@Debugger.wrap_interpreter
def interpret_statement(self, stmt, local_vars, allow_recursion=100): def interpret_statement(self, stmt, local_vars, allow_recursion=100):
if allow_recursion < 0: if allow_recursion < 0:
raise self.Exception('Recursion limit reached') raise self.Exception('Recursion limit reached')
@ -511,7 +550,6 @@ class JSInterpreter(object):
expr = self._dump(inner, local_vars) + outer expr = self._dump(inner, local_vars) + outer
if expr.startswith('('): if expr.startswith('('):
m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr) m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr)
if m: if m:
# short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig` # short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig`
@ -693,7 +731,7 @@ class JSInterpreter(object):
(?P<op>{_OPERATOR_RE})? (?P<op>{_OPERATOR_RE})?
=(?!=)(?P<expr>.*)$ =(?!=)(?P<expr>.*)$
)|(?P<return> )|(?P<return>
(?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$ (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$
)|(?P<indexing> )|(?P<indexing>
(?P<in>{_NAME_RE})\[(?P<idx>.+)\]$ (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
)|(?P<attribute> )|(?P<attribute>
@ -727,11 +765,12 @@ class JSInterpreter(object):
raise JS_Break() raise JS_Break()
elif expr == 'continue': elif expr == 'continue':
raise JS_Continue() raise JS_Continue()
elif expr == 'undefined': elif expr == 'undefined':
return JS_Undefined, should_return return JS_Undefined, should_return
elif expr == 'NaN': elif expr == 'NaN':
return _NaN, should_return return _NaN, should_return
elif expr == 'Infinity':
return _Infinity, should_return
elif md.get('return'): elif md.get('return'):
return local_vars[m.group('name')], should_return return local_vars[m.group('name')], should_return
@ -760,18 +799,28 @@ class JSInterpreter(object):
right_expr = separated.pop() right_expr = separated.pop()
# handle operators that are both unary and binary, minimal BODMAS # handle operators that are both unary and binary, minimal BODMAS
if op in ('+', '-'): if op in ('+', '-'):
# simplify/adjust consecutive instances of these operators
undone = 0 undone = 0
while len(separated) > 1 and not separated[-1].strip(): while len(separated) > 1 and not separated[-1].strip():
undone += 1 undone += 1
separated.pop() separated.pop()
if op == '-' and undone % 2 != 0: if op == '-' and undone % 2 != 0:
right_expr = op + right_expr right_expr = op + right_expr
elif op == '+':
while len(separated) > 1 and separated[-1].strip() 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] left_val = separated[-1]
for dm_op in ('*', '%', '/', '**'): for dm_op in ('*', '%', '/', '**'):
bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim)) bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim))
if len(bodmas) > 1 and not bodmas[-1].strip(): if len(bodmas) > 1 and not bodmas[-1].strip():
expr = op.join(separated) + op + right_expr expr = op.join(separated) + op + right_expr
right_expr = None 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 break
if right_expr is None: if right_expr is None:
continue continue
@ -797,6 +846,8 @@ class JSInterpreter(object):
def eval_method(): def eval_method():
if (variable, member) == ('console', 'debug'): if (variable, member) == ('console', 'debug'):
if Debugger.ENABLED:
Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion))
return return
types = { types = {
'String': compat_str, 'String': compat_str,

View File

@ -2406,7 +2406,7 @@ class ExtractorError(YoutubeDLError):
""" tb, if given, is the original traceback (so that it can be printed out). """ tb, if given, is the original traceback (so that it can be printed out).
If expected is set, this is a normal error message and most likely not a bug in youtube-dl. If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
""" """
self.orig_msg = msg
if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError): if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
expected = True expected = True
if video_id is not None: if video_id is not None: