[utils] Improve js_to_json, align with yt-dlp

* support variable substitution, from https://github.com/yt-dlp/yt-dlp/pull/#521 etc,
  thanks ChillingPepper, Grub4k, pukkandan
* improve escape handling, from https://github.com/yt-dlp/yt-dlp/pull/#521
  thanks Grub4k
* support template strings from https://github.com/yt-dlp/yt-dlp/pull/6623
  thanks Grub4k
* add limited `!` evaluation (eg, !!0 -> false, see tests)
pull/29318/head
dirkf 2023-05-03 12:06:34 +01:00
parent 825a40744b
commit d9d07a9581
2 changed files with 186 additions and 29 deletions

View File

@ -905,6 +905,85 @@ class TestUtil(unittest.TestCase):
)
self.assertEqual(escape_url('http://vimeo.com/56015672#at=0'), 'http://vimeo.com/56015672#at=0')
def test_js_to_json_vars_strings(self):
self.assertDictEqual(
json.loads(js_to_json(
'''{
'null': a,
'nullStr': b,
'true': c,
'trueStr': d,
'false': e,
'falseStr': f,
'unresolvedVar': g,
}''',
{
'a': 'null',
'b': '"null"',
'c': 'true',
'd': '"true"',
'e': 'false',
'f': '"false"',
'g': 'var',
}
)),
{
'null': None,
'nullStr': 'null',
'true': True,
'trueStr': 'true',
'false': False,
'falseStr': 'false',
'unresolvedVar': 'var'
}
)
self.assertDictEqual(
json.loads(js_to_json(
'''{
'int': a,
'intStr': b,
'float': c,
'floatStr': d,
}''',
{
'a': '123',
'b': '"123"',
'c': '1.23',
'd': '"1.23"',
}
)),
{
'int': 123,
'intStr': '123',
'float': 1.23,
'floatStr': '1.23',
}
)
self.assertDictEqual(
json.loads(js_to_json(
'''{
'object': a,
'objectStr': b,
'array': c,
'arrayStr': d,
}''',
{
'a': '{}',
'b': '"{}"',
'c': '[]',
'd': '"[]"',
}
)),
{
'object': {},
'objectStr': '{}',
'array': [],
'arrayStr': '[]',
}
)
def test_js_to_json_realworld(self):
inp = '''{
'clip':{'provider':'pseudo'}
@ -975,10 +1054,10 @@ class TestUtil(unittest.TestCase):
!42: 42
}''')
self.assertEqual(json.loads(on), {
'a': 0,
'b': 1,
'c': 0,
'd': 42.42,
'a': True,
'b': False,
'c': False,
'd': True,
'e': [],
'f': "abc",
'g': "",
@ -1048,10 +1127,26 @@ class TestUtil(unittest.TestCase):
on = js_to_json('{ "040": "040" }')
self.assertEqual(json.loads(on), {'040': '040'})
on = js_to_json('[1,//{},\n2]')
self.assertEqual(json.loads(on), [1, 2])
on = js_to_json(r'"\^\$\#"')
self.assertEqual(json.loads(on), R'^$#', msg='Unnecessary escapes should be stripped')
on = js_to_json('\'"\\""\'')
self.assertEqual(json.loads(on), '"""', msg='Unnecessary quote escape should be escaped')
def test_js_to_json_malformed(self):
self.assertEqual(js_to_json('42a1'), '42"a1"')
self.assertEqual(js_to_json('42a-1'), '42"a"-1')
def test_js_to_json_template_literal(self):
self.assertEqual(js_to_json('`Hello ${name}`', {'name': '"world"'}), '"Hello world"')
self.assertEqual(js_to_json('`${name}${name}`', {'name': '"X"'}), '"XX"')
self.assertEqual(js_to_json('`${name}${name}`', {'name': '5'}), '"55"')
self.assertEqual(js_to_json('`${name}"${name}"`', {'name': '5'}), '"5\\"5\\""')
self.assertEqual(js_to_json('`${name}`', {}), '"name"')
def test_extract_attributes(self):
self.assertEqual(extract_attributes('<e x="y">'), {'x': 'y'})
self.assertEqual(extract_attributes("<e x='y'>"), {'x': 'y'})

View File

@ -4365,46 +4365,108 @@ def strip_jsonp(code):
r'\g<callback_data>', code)
def js_to_json(code):
COMMENT_RE = r'/\*(?:(?!\*/).)*?\*/|//[^\n]*'
def js_to_json(code, *args, **kwargs):
# vars is a dict of (var, val) pairs to substitute
vars = args[0] if len(args) > 0 else kwargs.get('vars', {})
strict = kwargs.get('strict', False)
STRING_QUOTES = '\'"`'
STRING_RE = '|'.join(r'{0}(?:\\.|[^\\{0}])*{0}'.format(q) for q in STRING_QUOTES)
COMMENT_RE = r'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
SKIP_RE = r'\s*(?:{comment})?\s*'.format(comment=COMMENT_RE)
INTEGER_TABLE = (
(r'(?s)^(0[xX][0-9a-fA-F]+){skip}:?$'.format(skip=SKIP_RE), 16),
(r'(?s)^(0+[0-7]+){skip}:?$'.format(skip=SKIP_RE), 8),
(r'(?s)^(\d+){skip}:?$'.format(skip=SKIP_RE), 10),
)
# compat candidate
JSONDecodeError = json.JSONDecodeError if 'JSONDecodeError' in dir(json) else ValueError
def process_escape(match):
JSON_PASSTHROUGH_ESCAPES = r'"\bfnrtu'
escape = match.group(1) or match.group(2)
return ('\\' + escape if escape in JSON_PASSTHROUGH_ESCAPES
else '\\u00' if escape == 'x'
else '' if escape == '\n'
else escape)
def template_substitute(match):
evaluated = js_to_json(match.group(1), vars, strict=strict)
if evaluated[0] == '"':
return json.loads(evaluated)
return evaluated
def fix_kv(m):
v = m.group(0)
if v in ('true', 'false', 'null'):
return v
elif v.startswith('/*') or v.startswith('//') or v.startswith('!') or v == ',':
return ""
elif v in ('undefined', 'void 0'):
return 'null'
elif v.startswith('/*') or v.startswith('//') or v == ',':
return ''
if v[0] in ("'", '"'):
v = re.sub(r'(?s)\\.|"', lambda m: {
'"': '\\"',
"\\'": "'",
'\\\n': '',
'\\x': '\\u00',
}.get(m.group(0), m.group(0)), v[1:-1])
else:
for regex, base in INTEGER_TABLE:
im = re.match(regex, v)
if im:
i = int(im.group(1), base)
return '"%d":' % i if v.endswith(':') else '%d' % i
if v[0] in STRING_QUOTES:
v = re.sub(r'(?s)\${([^}]+)}', template_substitute, v[1:-1]) if v[0] == '`' else v[1:-1]
escaped = re.sub(r'(?s)(")|\\(.)', process_escape, v)
return '"{0}"'.format(escaped)
return '"%s"' % v
inv = IDENTITY
im = re.split(r'^!+', v)
if len(im) > 1 and not im[-1].endswith(':'):
if (len(v) - len(im[1])) % 2 == 1:
inv = lambda x: 'true' if x == 0 else 'false'
else:
inv = lambda x: 'false' if x == 0 else 'true'
if not any(x for x in im):
return
v = im[-1]
for regex, base in INTEGER_TABLE:
im = re.match(regex, v)
if im:
i = int(im.group(1), base)
return ('"%s":' if v.endswith(':') else '%s') % inv(i)
if v in vars:
try:
if not strict:
json.loads(vars[v])
except JSONDecodeError:
return inv(json.dumps(vars[v]))
else:
return inv(vars[v])
if not strict:
v = try_call(inv, args=(v,), default=v)
if v in ('true', 'false'):
return v
return '"{0}"'.format(v)
raise ValueError('Unknown value: ' + v)
def create_map(mobj):
return json.dumps(dict(json.loads(js_to_json(mobj.group(1) or '[]', vars=vars))))
code = re.sub(r'new Map\((\[.*?\])?\)', create_map, code)
if not strict:
code = re.sub(r'new Date\((".+")\)', r'\g<1>', code)
code = re.sub(r'new \w+\((.*?)\)', lambda m: json.dumps(m.group(0)), code)
code = re.sub(r'parseInt\([^\d]+(\d+)[^\d]+\)', r'\1', code)
code = re.sub(r'\(function\([^)]*\)\s*\{[^}]*\}\s*\)\s*\(\s*(["\'][^)]*["\'])\s*\)', r'\1', code)
return re.sub(r'''(?sx)
"(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
'(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
{comment}|,(?={skip}[\]}}])|
(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_])[.a-zA-Z_0-9]*|
\b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
[0-9]+(?={skip}:)|
{str_}|
{comment}|
,(?={skip}[\]}}])|
void\s0|
!*(?:(?<!\d)[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
(?:\b|!+)0(?:[xX][\da-fA-F]+|[0-7]+)(?:{skip}:)?|
!+\d+(?:\.\d*)?(?:{skip}:)?|
[0-9]+(?:{skip}:)|
!+
'''.format(comment=COMMENT_RE, skip=SKIP_RE), fix_kv, code)
'''.format(comment=COMMENT_RE, skip=SKIP_RE, str_=STRING_RE), fix_kv, code)
def qualities(quality_ids):