1 """Parse (absolute and relative) URLs.
2
3 urlparse module is based upon the following RFC specifications.
4
5 RFC 3986 (STD66): "Uniform Resource Identifiers" by T. Berners-Lee, R. Fielding
6 and L. Masinter, January 2005.
7
8 RFC 2732 : "Format for Literal IPv6 Addresses in URL's by R.Hinden, B.Carpenter
9 and L.Masinter, December 1999.
10
11 RFC 2396: "Uniform Resource Identifiers (URI)": Generic Syntax by T.
12 Berners-Lee, R. Fielding, and L. Masinter, August 1998.
13
14 RFC 2368: "The mailto URL scheme", by P.Hoffman , L Masinter, J. Zawinski, July 1998.
15
16 RFC 1808: "Relative Uniform Resource Locators", by R. Fielding, UC Irvine, June
17 1995.
18
19 RFC 1738: "Uniform Resource Locators (URL)" by T. Berners-Lee, L. Masinter, M.
20 McCahill, December 1994
21
22 RFC 3986 is considered the current standard and any future changes to
23 urlparse module should conform with it. The urlparse module is
24 currently not entirely compliant with this RFC due to defacto
25 scenarios for parsing, and for backward compatibility purposes, some
26 parsing quirks from older RFCs are retained. The testcases in
27 test_urlparse.py provides a good indicator of parsing behavior.
28
29 The WHATWG URL Parser spec should also be considered. We are not compliant with
30 it either due to existing user code API behavior expectations (Hyrum's Law).
31 It serves as a useful guide when making changes.
32 """
33
34 from collections import namedtuple
35 import functools
36 import re
37 import sys
38 import types
39 import warnings
40 import ipaddress
41
42 __all__ = ["urlparse", "urlunparse", "urljoin", "urldefrag",
43 "urlsplit", "urlunsplit", "urlencode", "parse_qs",
44 "parse_qsl", "quote", "quote_plus", "quote_from_bytes",
45 "unquote", "unquote_plus", "unquote_to_bytes",
46 "DefragResult", "ParseResult", "SplitResult",
47 "DefragResultBytes", "ParseResultBytes", "SplitResultBytes"]
48
49 # A classification of schemes.
50 # The empty string classifies URLs with no scheme specified,
51 # being the default value returned by “urlsplit” and “urlparse”.
52
53 uses_relative = ['', 'ftp', 'http', 'gopher', 'nntp', 'imap',
54 'wais', 'file', 'https', 'shttp', 'mms',
55 'prospero', 'rtsp', 'rtsps', 'rtspu', 'sftp',
56 'svn', 'svn+ssh', 'ws', 'wss']
57
58 uses_netloc = ['', 'ftp', 'http', 'gopher', 'nntp', 'telnet',
59 'imap', 'wais', 'file', 'mms', 'https', 'shttp',
60 'snews', 'prospero', 'rtsp', 'rtsps', 'rtspu', 'rsync',
61 'svn', 'svn+ssh', 'sftp', 'nfs', 'git', 'git+ssh',
62 'ws', 'wss']
63
64 uses_params = ['', 'ftp', 'hdl', 'prospero', 'http', 'imap',
65 'https', 'shttp', 'rtsp', 'rtsps', 'rtspu', 'sip',
66 'sips', 'mms', 'sftp', 'tel']
67
68 # These are not actually used anymore, but should stay for backwards
69 # compatibility. (They are undocumented, but have a public-looking name.)
70
71 non_hierarchical = ['gopher', 'hdl', 'mailto', 'news',
72 'telnet', 'wais', 'imap', 'snews', 'sip', 'sips']
73
74 uses_query = ['', 'http', 'wais', 'imap', 'https', 'shttp', 'mms',
75 'gopher', 'rtsp', 'rtsps', 'rtspu', 'sip', 'sips']
76
77 uses_fragment = ['', 'ftp', 'hdl', 'http', 'gopher', 'news',
78 'nntp', 'wais', 'https', 'shttp', 'snews',
79 'file', 'prospero']
80
81 # Characters valid in scheme names
82 scheme_chars = ('abcdefghijklmnopqrstuvwxyz'
83 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'
84 '0123456789'
85 '+-.')
86
87 # Leading and trailing C0 control and space to be stripped per WHATWG spec.
88 # == "".join([chr(i) for i in range(0, 0x20 + 1)])
89 _WHATWG_C0_CONTROL_OR_SPACE = '\x00\x01\x02\x03\x04\x05\x06\x07\x08\t\n\x0b\x0c\r\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f '
90
91 # Unsafe bytes to be removed per WHATWG spec
92 _UNSAFE_URL_BYTES_TO_REMOVE = ['\t', '\r', '\n']
93
94 def clear_cache():
95 """Clear internal performance caches. Undocumented; some tests want it."""
96 urlsplit.cache_clear()
97 _byte_quoter_factory.cache_clear()
98
99 # Helpers for bytes handling
100 # For 3.2, we deliberately require applications that
101 # handle improperly quoted URLs to do their own
102 # decoding and encoding. If valid use cases are
103 # presented, we may relax this by using latin-1
104 # decoding internally for 3.3
105 _implicit_encoding = 'ascii'
106 _implicit_errors = 'strict'
107
108 def _noop(obj):
109 return obj
110
111 def _encode_result(obj, encoding=_implicit_encoding,
112 errors=_implicit_errors):
113 return obj.encode(encoding, errors)
114
115 def _decode_args(args, encoding=_implicit_encoding,
116 errors=_implicit_errors):
117 return tuple(x.decode(encoding, errors) if x else '' for x in args)
118
119 def _coerce_args(*args):
120 # Invokes decode if necessary to create str args
121 # and returns the coerced inputs along with
122 # an appropriate result coercion function
123 # - noop for str inputs
124 # - encoding function otherwise
125 str_input = isinstance(args[0], str)
126 for arg in args[1:]:
127 # We special-case the empty string to support the
128 # "scheme=''" default argument to some functions
129 if arg and isinstance(arg, str) != str_input:
130 raise TypeError("Cannot mix str and non-str arguments")
131 if str_input:
132 return args + (_noop,)
133 return _decode_args(args) + (_encode_result,)
134
135 # Result objects are more helpful than simple tuples
136 class ESC[4;38;5;81m_ResultMixinStr(ESC[4;38;5;149mobject):
137 """Standard approach to encoding parsed results from str to bytes"""
138 __slots__ = ()
139
140 def encode(self, encoding='ascii', errors='strict'):
141 return self._encoded_counterpart(*(x.encode(encoding, errors) for x in self))
142
143
144 class ESC[4;38;5;81m_ResultMixinBytes(ESC[4;38;5;149mobject):
145 """Standard approach to decoding parsed results from bytes to str"""
146 __slots__ = ()
147
148 def decode(self, encoding='ascii', errors='strict'):
149 return self._decoded_counterpart(*(x.decode(encoding, errors) for x in self))
150
151
152 class ESC[4;38;5;81m_NetlocResultMixinBase(ESC[4;38;5;149mobject):
153 """Shared methods for the parsed result objects containing a netloc element"""
154 __slots__ = ()
155
156 @property
157 def username(self):
158 return self._userinfo[0]
159
160 @property
161 def password(self):
162 return self._userinfo[1]
163
164 @property
165 def hostname(self):
166 hostname = self._hostinfo[0]
167 if not hostname:
168 return None
169 # Scoped IPv6 address may have zone info, which must not be lowercased
170 # like http://[fe80::822a:a8ff:fe49:470c%tESt]:1234/keys
171 separator = '%' if isinstance(hostname, str) else b'%'
172 hostname, percent, zone = hostname.partition(separator)
173 return hostname.lower() + percent + zone
174
175 @property
176 def port(self):
177 port = self._hostinfo[1]
178 if port is not None:
179 if port.isdigit() and port.isascii():
180 port = int(port)
181 else:
182 raise ValueError(f"Port could not be cast to integer value as {port!r}")
183 if not (0 <= port <= 65535):
184 raise ValueError("Port out of range 0-65535")
185 return port
186
187 __class_getitem__ = classmethod(types.GenericAlias)
188
189
190 class ESC[4;38;5;81m_NetlocResultMixinStr(ESC[4;38;5;149m_NetlocResultMixinBase, ESC[4;38;5;149m_ResultMixinStr):
191 __slots__ = ()
192
193 @property
194 def _userinfo(self):
195 netloc = self.netloc
196 userinfo, have_info, hostinfo = netloc.rpartition('@')
197 if have_info:
198 username, have_password, password = userinfo.partition(':')
199 if not have_password:
200 password = None
201 else:
202 username = password = None
203 return username, password
204
205 @property
206 def _hostinfo(self):
207 netloc = self.netloc
208 _, _, hostinfo = netloc.rpartition('@')
209 _, have_open_br, bracketed = hostinfo.partition('[')
210 if have_open_br:
211 hostname, _, port = bracketed.partition(']')
212 _, _, port = port.partition(':')
213 else:
214 hostname, _, port = hostinfo.partition(':')
215 if not port:
216 port = None
217 return hostname, port
218
219
220 class ESC[4;38;5;81m_NetlocResultMixinBytes(ESC[4;38;5;149m_NetlocResultMixinBase, ESC[4;38;5;149m_ResultMixinBytes):
221 __slots__ = ()
222
223 @property
224 def _userinfo(self):
225 netloc = self.netloc
226 userinfo, have_info, hostinfo = netloc.rpartition(b'@')
227 if have_info:
228 username, have_password, password = userinfo.partition(b':')
229 if not have_password:
230 password = None
231 else:
232 username = password = None
233 return username, password
234
235 @property
236 def _hostinfo(self):
237 netloc = self.netloc
238 _, _, hostinfo = netloc.rpartition(b'@')
239 _, have_open_br, bracketed = hostinfo.partition(b'[')
240 if have_open_br:
241 hostname, _, port = bracketed.partition(b']')
242 _, _, port = port.partition(b':')
243 else:
244 hostname, _, port = hostinfo.partition(b':')
245 if not port:
246 port = None
247 return hostname, port
248
249
250 _DefragResultBase = namedtuple('DefragResult', 'url fragment')
251 _SplitResultBase = namedtuple(
252 'SplitResult', 'scheme netloc path query fragment')
253 _ParseResultBase = namedtuple(
254 'ParseResult', 'scheme netloc path params query fragment')
255
256 _DefragResultBase.__doc__ = """
257 DefragResult(url, fragment)
258
259 A 2-tuple that contains the url without fragment identifier and the fragment
260 identifier as a separate argument.
261 """
262
263 _DefragResultBase.url.__doc__ = """The URL with no fragment identifier."""
264
265 _DefragResultBase.fragment.__doc__ = """
266 Fragment identifier separated from URL, that allows indirect identification of a
267 secondary resource by reference to a primary resource and additional identifying
268 information.
269 """
270
271 _SplitResultBase.__doc__ = """
272 SplitResult(scheme, netloc, path, query, fragment)
273
274 A 5-tuple that contains the different components of a URL. Similar to
275 ParseResult, but does not split params.
276 """
277
278 _SplitResultBase.scheme.__doc__ = """Specifies URL scheme for the request."""
279
280 _SplitResultBase.netloc.__doc__ = """
281 Network location where the request is made to.
282 """
283
284 _SplitResultBase.path.__doc__ = """
285 The hierarchical path, such as the path to a file to download.
286 """
287
288 _SplitResultBase.query.__doc__ = """
289 The query component, that contains non-hierarchical data, that along with data
290 in path component, identifies a resource in the scope of URI's scheme and
291 network location.
292 """
293
294 _SplitResultBase.fragment.__doc__ = """
295 Fragment identifier, that allows indirect identification of a secondary resource
296 by reference to a primary resource and additional identifying information.
297 """
298
299 _ParseResultBase.__doc__ = """
300 ParseResult(scheme, netloc, path, params, query, fragment)
301
302 A 6-tuple that contains components of a parsed URL.
303 """
304
305 _ParseResultBase.scheme.__doc__ = _SplitResultBase.scheme.__doc__
306 _ParseResultBase.netloc.__doc__ = _SplitResultBase.netloc.__doc__
307 _ParseResultBase.path.__doc__ = _SplitResultBase.path.__doc__
308 _ParseResultBase.params.__doc__ = """
309 Parameters for last path element used to dereference the URI in order to provide
310 access to perform some operation on the resource.
311 """
312
313 _ParseResultBase.query.__doc__ = _SplitResultBase.query.__doc__
314 _ParseResultBase.fragment.__doc__ = _SplitResultBase.fragment.__doc__
315
316
317 # For backwards compatibility, alias _NetlocResultMixinStr
318 # ResultBase is no longer part of the documented API, but it is
319 # retained since deprecating it isn't worth the hassle
320 ResultBase = _NetlocResultMixinStr
321
322 # Structured result objects for string data
323 class ESC[4;38;5;81mDefragResult(ESC[4;38;5;149m_DefragResultBase, ESC[4;38;5;149m_ResultMixinStr):
324 __slots__ = ()
325 def geturl(self):
326 if self.fragment:
327 return self.url + '#' + self.fragment
328 else:
329 return self.url
330
331 class ESC[4;38;5;81mSplitResult(ESC[4;38;5;149m_SplitResultBase, ESC[4;38;5;149m_NetlocResultMixinStr):
332 __slots__ = ()
333 def geturl(self):
334 return urlunsplit(self)
335
336 class ESC[4;38;5;81mParseResult(ESC[4;38;5;149m_ParseResultBase, ESC[4;38;5;149m_NetlocResultMixinStr):
337 __slots__ = ()
338 def geturl(self):
339 return urlunparse(self)
340
341 # Structured result objects for bytes data
342 class ESC[4;38;5;81mDefragResultBytes(ESC[4;38;5;149m_DefragResultBase, ESC[4;38;5;149m_ResultMixinBytes):
343 __slots__ = ()
344 def geturl(self):
345 if self.fragment:
346 return self.url + b'#' + self.fragment
347 else:
348 return self.url
349
350 class ESC[4;38;5;81mSplitResultBytes(ESC[4;38;5;149m_SplitResultBase, ESC[4;38;5;149m_NetlocResultMixinBytes):
351 __slots__ = ()
352 def geturl(self):
353 return urlunsplit(self)
354
355 class ESC[4;38;5;81mParseResultBytes(ESC[4;38;5;149m_ParseResultBase, ESC[4;38;5;149m_NetlocResultMixinBytes):
356 __slots__ = ()
357 def geturl(self):
358 return urlunparse(self)
359
360 # Set up the encode/decode result pairs
361 def _fix_result_transcoding():
362 _result_pairs = (
363 (DefragResult, DefragResultBytes),
364 (SplitResult, SplitResultBytes),
365 (ParseResult, ParseResultBytes),
366 )
367 for _decoded, _encoded in _result_pairs:
368 _decoded._encoded_counterpart = _encoded
369 _encoded._decoded_counterpart = _decoded
370
371 _fix_result_transcoding()
372 del _fix_result_transcoding
373
374 def urlparse(url, scheme='', allow_fragments=True):
375 """Parse a URL into 6 components:
376 <scheme>://<netloc>/<path>;<params>?<query>#<fragment>
377
378 The result is a named 6-tuple with fields corresponding to the
379 above. It is either a ParseResult or ParseResultBytes object,
380 depending on the type of the url parameter.
381
382 The username, password, hostname, and port sub-components of netloc
383 can also be accessed as attributes of the returned object.
384
385 The scheme argument provides the default value of the scheme
386 component when no scheme is found in url.
387
388 If allow_fragments is False, no attempt is made to separate the
389 fragment component from the previous component, which can be either
390 path or query.
391
392 Note that % escapes are not expanded.
393 """
394 url, scheme, _coerce_result = _coerce_args(url, scheme)
395 splitresult = urlsplit(url, scheme, allow_fragments)
396 scheme, netloc, url, query, fragment = splitresult
397 if scheme in uses_params and ';' in url:
398 url, params = _splitparams(url)
399 else:
400 params = ''
401 result = ParseResult(scheme, netloc, url, params, query, fragment)
402 return _coerce_result(result)
403
404 def _splitparams(url):
405 if '/' in url:
406 i = url.find(';', url.rfind('/'))
407 if i < 0:
408 return url, ''
409 else:
410 i = url.find(';')
411 return url[:i], url[i+1:]
412
413 def _splitnetloc(url, start=0):
414 delim = len(url) # position of end of domain part of url, default is end
415 for c in '/?#': # look for delimiters; the order is NOT important
416 wdelim = url.find(c, start) # find first of this delim
417 if wdelim >= 0: # if found
418 delim = min(delim, wdelim) # use earliest delim position
419 return url[start:delim], url[delim:] # return (domain, rest)
420
421 def _checknetloc(netloc):
422 if not netloc or netloc.isascii():
423 return
424 # looking for characters like \u2100 that expand to 'a/c'
425 # IDNA uses NFKC equivalence, so normalize for this check
426 import unicodedata
427 n = netloc.replace('@', '') # ignore characters already included
428 n = n.replace(':', '') # but not the surrounding text
429 n = n.replace('#', '')
430 n = n.replace('?', '')
431 netloc2 = unicodedata.normalize('NFKC', n)
432 if n == netloc2:
433 return
434 for c in '/?#@:':
435 if c in netloc2:
436 raise ValueError("netloc '" + netloc + "' contains invalid " +
437 "characters under NFKC normalization")
438
439 # Valid bracketed hosts are defined in
440 # https://www.rfc-editor.org/rfc/rfc3986#page-49 and https://url.spec.whatwg.org/
441 def _check_bracketed_host(hostname):
442 if hostname.startswith('v'):
443 if not re.match(r"\Av[a-fA-F0-9]+\..+\Z", hostname):
444 raise ValueError(f"IPvFuture address is invalid")
445 else:
446 ip = ipaddress.ip_address(hostname) # Throws Value Error if not IPv6 or IPv4
447 if isinstance(ip, ipaddress.IPv4Address):
448 raise ValueError(f"An IPv4 address cannot be in brackets")
449
450 # typed=True avoids BytesWarnings being emitted during cache key
451 # comparison since this API supports both bytes and str input.
452 @functools.lru_cache(typed=True)
453 def urlsplit(url, scheme='', allow_fragments=True):
454 """Parse a URL into 5 components:
455 <scheme>://<netloc>/<path>?<query>#<fragment>
456
457 The result is a named 5-tuple with fields corresponding to the
458 above. It is either a SplitResult or SplitResultBytes object,
459 depending on the type of the url parameter.
460
461 The username, password, hostname, and port sub-components of netloc
462 can also be accessed as attributes of the returned object.
463
464 The scheme argument provides the default value of the scheme
465 component when no scheme is found in url.
466
467 If allow_fragments is False, no attempt is made to separate the
468 fragment component from the previous component, which can be either
469 path or query.
470
471 Note that % escapes are not expanded.
472 """
473
474 url, scheme, _coerce_result = _coerce_args(url, scheme)
475 # Only lstrip url as some applications rely on preserving trailing space.
476 # (https://url.spec.whatwg.org/#concept-basic-url-parser would strip both)
477 url = url.lstrip(_WHATWG_C0_CONTROL_OR_SPACE)
478 scheme = scheme.strip(_WHATWG_C0_CONTROL_OR_SPACE)
479
480 for b in _UNSAFE_URL_BYTES_TO_REMOVE:
481 url = url.replace(b, "")
482 scheme = scheme.replace(b, "")
483
484 allow_fragments = bool(allow_fragments)
485 netloc = query = fragment = ''
486 i = url.find(':')
487 if i > 0 and url[0].isascii() and url[0].isalpha():
488 for c in url[:i]:
489 if c not in scheme_chars:
490 break
491 else:
492 scheme, url = url[:i].lower(), url[i+1:]
493 if url[:2] == '//':
494 netloc, url = _splitnetloc(url, 2)
495 if (('[' in netloc and ']' not in netloc) or
496 (']' in netloc and '[' not in netloc)):
497 raise ValueError("Invalid IPv6 URL")
498 if '[' in netloc and ']' in netloc:
499 bracketed_host = netloc.partition('[')[2].partition(']')[0]
500 _check_bracketed_host(bracketed_host)
501 if allow_fragments and '#' in url:
502 url, fragment = url.split('#', 1)
503 if '?' in url:
504 url, query = url.split('?', 1)
505 _checknetloc(netloc)
506 v = SplitResult(scheme, netloc, url, query, fragment)
507 return _coerce_result(v)
508
509 def urlunparse(components):
510 """Put a parsed URL back together again. This may result in a
511 slightly different, but equivalent URL, if the URL that was parsed
512 originally had redundant delimiters, e.g. a ? with an empty query
513 (the draft states that these are equivalent)."""
514 scheme, netloc, url, params, query, fragment, _coerce_result = (
515 _coerce_args(*components))
516 if params:
517 url = "%s;%s" % (url, params)
518 return _coerce_result(urlunsplit((scheme, netloc, url, query, fragment)))
519
520 def urlunsplit(components):
521 """Combine the elements of a tuple as returned by urlsplit() into a
522 complete URL as a string. The data argument can be any five-item iterable.
523 This may result in a slightly different, but equivalent URL, if the URL that
524 was parsed originally had unnecessary delimiters (for example, a ? with an
525 empty query; the RFC states that these are equivalent)."""
526 scheme, netloc, url, query, fragment, _coerce_result = (
527 _coerce_args(*components))
528 if netloc or (scheme and scheme in uses_netloc and url[:2] != '//'):
529 if url and url[:1] != '/': url = '/' + url
530 url = '//' + (netloc or '') + url
531 if scheme:
532 url = scheme + ':' + url
533 if query:
534 url = url + '?' + query
535 if fragment:
536 url = url + '#' + fragment
537 return _coerce_result(url)
538
539 def urljoin(base, url, allow_fragments=True):
540 """Join a base URL and a possibly relative URL to form an absolute
541 interpretation of the latter."""
542 if not base:
543 return url
544 if not url:
545 return base
546
547 base, url, _coerce_result = _coerce_args(base, url)
548 bscheme, bnetloc, bpath, bparams, bquery, bfragment = \
549 urlparse(base, '', allow_fragments)
550 scheme, netloc, path, params, query, fragment = \
551 urlparse(url, bscheme, allow_fragments)
552
553 if scheme != bscheme or scheme not in uses_relative:
554 return _coerce_result(url)
555 if scheme in uses_netloc:
556 if netloc:
557 return _coerce_result(urlunparse((scheme, netloc, path,
558 params, query, fragment)))
559 netloc = bnetloc
560
561 if not path and not params:
562 path = bpath
563 params = bparams
564 if not query:
565 query = bquery
566 return _coerce_result(urlunparse((scheme, netloc, path,
567 params, query, fragment)))
568
569 base_parts = bpath.split('/')
570 if base_parts[-1] != '':
571 # the last item is not a directory, so will not be taken into account
572 # in resolving the relative path
573 del base_parts[-1]
574
575 # for rfc3986, ignore all base path should the first character be root.
576 if path[:1] == '/':
577 segments = path.split('/')
578 else:
579 segments = base_parts + path.split('/')
580 # filter out elements that would cause redundant slashes on re-joining
581 # the resolved_path
582 segments[1:-1] = filter(None, segments[1:-1])
583
584 resolved_path = []
585
586 for seg in segments:
587 if seg == '..':
588 try:
589 resolved_path.pop()
590 except IndexError:
591 # ignore any .. segments that would otherwise cause an IndexError
592 # when popped from resolved_path if resolving for rfc3986
593 pass
594 elif seg == '.':
595 continue
596 else:
597 resolved_path.append(seg)
598
599 if segments[-1] in ('.', '..'):
600 # do some post-processing here. if the last segment was a relative dir,
601 # then we need to append the trailing '/'
602 resolved_path.append('')
603
604 return _coerce_result(urlunparse((scheme, netloc, '/'.join(
605 resolved_path) or '/', params, query, fragment)))
606
607
608 def urldefrag(url):
609 """Removes any existing fragment from URL.
610
611 Returns a tuple of the defragmented URL and the fragment. If
612 the URL contained no fragments, the second element is the
613 empty string.
614 """
615 url, _coerce_result = _coerce_args(url)
616 if '#' in url:
617 s, n, p, a, q, frag = urlparse(url)
618 defrag = urlunparse((s, n, p, a, q, ''))
619 else:
620 frag = ''
621 defrag = url
622 return _coerce_result(DefragResult(defrag, frag))
623
624 _hexdig = '0123456789ABCDEFabcdef'
625 _hextobyte = None
626
627 def unquote_to_bytes(string):
628 """unquote_to_bytes('abc%20def') -> b'abc def'."""
629 # Note: strings are encoded as UTF-8. This is only an issue if it contains
630 # unescaped non-ASCII characters, which URIs should not.
631 if not string:
632 # Is it a string-like object?
633 string.split
634 return b''
635 if isinstance(string, str):
636 string = string.encode('utf-8')
637 bits = string.split(b'%')
638 if len(bits) == 1:
639 return string
640 res = [bits[0]]
641 append = res.append
642 # Delay the initialization of the table to not waste memory
643 # if the function is never called
644 global _hextobyte
645 if _hextobyte is None:
646 _hextobyte = {(a + b).encode(): bytes.fromhex(a + b)
647 for a in _hexdig for b in _hexdig}
648 for item in bits[1:]:
649 try:
650 append(_hextobyte[item[:2]])
651 append(item[2:])
652 except KeyError:
653 append(b'%')
654 append(item)
655 return b''.join(res)
656
657 _asciire = re.compile('([\x00-\x7f]+)')
658
659 def unquote(string, encoding='utf-8', errors='replace'):
660 """Replace %xx escapes by their single-character equivalent. The optional
661 encoding and errors parameters specify how to decode percent-encoded
662 sequences into Unicode characters, as accepted by the bytes.decode()
663 method.
664 By default, percent-encoded sequences are decoded with UTF-8, and invalid
665 sequences are replaced by a placeholder character.
666
667 unquote('abc%20def') -> 'abc def'.
668 """
669 if isinstance(string, bytes):
670 return unquote_to_bytes(string).decode(encoding, errors)
671 if '%' not in string:
672 string.split
673 return string
674 if encoding is None:
675 encoding = 'utf-8'
676 if errors is None:
677 errors = 'replace'
678 bits = _asciire.split(string)
679 res = [bits[0]]
680 append = res.append
681 for i in range(1, len(bits), 2):
682 append(unquote_to_bytes(bits[i]).decode(encoding, errors))
683 append(bits[i + 1])
684 return ''.join(res)
685
686
687 def parse_qs(qs, keep_blank_values=False, strict_parsing=False,
688 encoding='utf-8', errors='replace', max_num_fields=None, separator='&'):
689 """Parse a query given as a string argument.
690
691 Arguments:
692
693 qs: percent-encoded query string to be parsed
694
695 keep_blank_values: flag indicating whether blank values in
696 percent-encoded queries should be treated as blank strings.
697 A true value indicates that blanks should be retained as
698 blank strings. The default false value indicates that
699 blank values are to be ignored and treated as if they were
700 not included.
701
702 strict_parsing: flag indicating what to do with parsing errors.
703 If false (the default), errors are silently ignored.
704 If true, errors raise a ValueError exception.
705
706 encoding and errors: specify how to decode percent-encoded sequences
707 into Unicode characters, as accepted by the bytes.decode() method.
708
709 max_num_fields: int. If set, then throws a ValueError if there
710 are more than n fields read by parse_qsl().
711
712 separator: str. The symbol to use for separating the query arguments.
713 Defaults to &.
714
715 Returns a dictionary.
716 """
717 parsed_result = {}
718 pairs = parse_qsl(qs, keep_blank_values, strict_parsing,
719 encoding=encoding, errors=errors,
720 max_num_fields=max_num_fields, separator=separator)
721 for name, value in pairs:
722 if name in parsed_result:
723 parsed_result[name].append(value)
724 else:
725 parsed_result[name] = [value]
726 return parsed_result
727
728
729 def parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
730 encoding='utf-8', errors='replace', max_num_fields=None, separator='&'):
731 """Parse a query given as a string argument.
732
733 Arguments:
734
735 qs: percent-encoded query string to be parsed
736
737 keep_blank_values: flag indicating whether blank values in
738 percent-encoded queries should be treated as blank strings.
739 A true value indicates that blanks should be retained as blank
740 strings. The default false value indicates that blank values
741 are to be ignored and treated as if they were not included.
742
743 strict_parsing: flag indicating what to do with parsing errors. If
744 false (the default), errors are silently ignored. If true,
745 errors raise a ValueError exception.
746
747 encoding and errors: specify how to decode percent-encoded sequences
748 into Unicode characters, as accepted by the bytes.decode() method.
749
750 max_num_fields: int. If set, then throws a ValueError
751 if there are more than n fields read by parse_qsl().
752
753 separator: str. The symbol to use for separating the query arguments.
754 Defaults to &.
755
756 Returns a list, as G-d intended.
757 """
758 qs, _coerce_result = _coerce_args(qs)
759 separator, _ = _coerce_args(separator)
760
761 if not separator or (not isinstance(separator, (str, bytes))):
762 raise ValueError("Separator must be of type string or bytes.")
763
764 # If max_num_fields is defined then check that the number of fields
765 # is less than max_num_fields. This prevents a memory exhaustion DOS
766 # attack via post bodies with many fields.
767 if max_num_fields is not None:
768 num_fields = 1 + qs.count(separator) if qs else 0
769 if max_num_fields < num_fields:
770 raise ValueError('Max number of fields exceeded')
771
772 r = []
773 query_args = qs.split(separator) if qs else []
774 for name_value in query_args:
775 if not name_value and not strict_parsing:
776 continue
777 nv = name_value.split('=', 1)
778 if len(nv) != 2:
779 if strict_parsing:
780 raise ValueError("bad query field: %r" % (name_value,))
781 # Handle case of a control-name with no equal sign
782 if keep_blank_values:
783 nv.append('')
784 else:
785 continue
786 if len(nv[1]) or keep_blank_values:
787 name = nv[0].replace('+', ' ')
788 name = unquote(name, encoding=encoding, errors=errors)
789 name = _coerce_result(name)
790 value = nv[1].replace('+', ' ')
791 value = unquote(value, encoding=encoding, errors=errors)
792 value = _coerce_result(value)
793 r.append((name, value))
794 return r
795
796 def unquote_plus(string, encoding='utf-8', errors='replace'):
797 """Like unquote(), but also replace plus signs by spaces, as required for
798 unquoting HTML form values.
799
800 unquote_plus('%7e/abc+def') -> '~/abc def'
801 """
802 string = string.replace('+', ' ')
803 return unquote(string, encoding, errors)
804
805 _ALWAYS_SAFE = frozenset(b'ABCDEFGHIJKLMNOPQRSTUVWXYZ'
806 b'abcdefghijklmnopqrstuvwxyz'
807 b'0123456789'
808 b'_.-~')
809 _ALWAYS_SAFE_BYTES = bytes(_ALWAYS_SAFE)
810
811 def __getattr__(name):
812 if name == 'Quoter':
813 warnings.warn('Deprecated in 3.11. '
814 'urllib.parse.Quoter will be removed in Python 3.14. '
815 'It was not intended to be a public API.',
816 DeprecationWarning, stacklevel=2)
817 return _Quoter
818 raise AttributeError(f'module {__name__!r} has no attribute {name!r}')
819
820 class ESC[4;38;5;81m_Quoter(ESC[4;38;5;149mdict):
821 """A mapping from bytes numbers (in range(0,256)) to strings.
822
823 String values are percent-encoded byte values, unless the key < 128, and
824 in either of the specified safe set, or the always safe set.
825 """
826 # Keeps a cache internally, via __missing__, for efficiency (lookups
827 # of cached keys don't call Python code at all).
828 def __init__(self, safe):
829 """safe: bytes object."""
830 self.safe = _ALWAYS_SAFE.union(safe)
831
832 def __repr__(self):
833 return f"<Quoter {dict(self)!r}>"
834
835 def __missing__(self, b):
836 # Handle a cache miss. Store quoted string in cache and return.
837 res = chr(b) if b in self.safe else '%{:02X}'.format(b)
838 self[b] = res
839 return res
840
841 def quote(string, safe='/', encoding=None, errors=None):
842 """quote('abc def') -> 'abc%20def'
843
844 Each part of a URL, e.g. the path info, the query, etc., has a
845 different set of reserved characters that must be quoted. The
846 quote function offers a cautious (not minimal) way to quote a
847 string for most of these parts.
848
849 RFC 3986 Uniform Resource Identifier (URI): Generic Syntax lists
850 the following (un)reserved characters.
851
852 unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
853 reserved = gen-delims / sub-delims
854 gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
855 sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
856 / "*" / "+" / "," / ";" / "="
857
858 Each of the reserved characters is reserved in some component of a URL,
859 but not necessarily in all of them.
860
861 The quote function %-escapes all characters that are neither in the
862 unreserved chars ("always safe") nor the additional chars set via the
863 safe arg.
864
865 The default for the safe arg is '/'. The character is reserved, but in
866 typical usage the quote function is being called on a path where the
867 existing slash characters are to be preserved.
868
869 Python 3.7 updates from using RFC 2396 to RFC 3986 to quote URL strings.
870 Now, "~" is included in the set of unreserved characters.
871
872 string and safe may be either str or bytes objects. encoding and errors
873 must not be specified if string is a bytes object.
874
875 The optional encoding and errors parameters specify how to deal with
876 non-ASCII characters, as accepted by the str.encode method.
877 By default, encoding='utf-8' (characters are encoded with UTF-8), and
878 errors='strict' (unsupported characters raise a UnicodeEncodeError).
879 """
880 if isinstance(string, str):
881 if not string:
882 return string
883 if encoding is None:
884 encoding = 'utf-8'
885 if errors is None:
886 errors = 'strict'
887 string = string.encode(encoding, errors)
888 else:
889 if encoding is not None:
890 raise TypeError("quote() doesn't support 'encoding' for bytes")
891 if errors is not None:
892 raise TypeError("quote() doesn't support 'errors' for bytes")
893 return quote_from_bytes(string, safe)
894
895 def quote_plus(string, safe='', encoding=None, errors=None):
896 """Like quote(), but also replace ' ' with '+', as required for quoting
897 HTML form values. Plus signs in the original string are escaped unless
898 they are included in safe. It also does not have safe default to '/'.
899 """
900 # Check if ' ' in string, where string may either be a str or bytes. If
901 # there are no spaces, the regular quote will produce the right answer.
902 if ((isinstance(string, str) and ' ' not in string) or
903 (isinstance(string, bytes) and b' ' not in string)):
904 return quote(string, safe, encoding, errors)
905 if isinstance(safe, str):
906 space = ' '
907 else:
908 space = b' '
909 string = quote(string, safe + space, encoding, errors)
910 return string.replace(' ', '+')
911
912 # Expectation: A typical program is unlikely to create more than 5 of these.
913 @functools.lru_cache
914 def _byte_quoter_factory(safe):
915 return _Quoter(safe).__getitem__
916
917 def quote_from_bytes(bs, safe='/'):
918 """Like quote(), but accepts a bytes object rather than a str, and does
919 not perform string-to-bytes encoding. It always returns an ASCII string.
920 quote_from_bytes(b'abc def\x3f') -> 'abc%20def%3f'
921 """
922 if not isinstance(bs, (bytes, bytearray)):
923 raise TypeError("quote_from_bytes() expected bytes")
924 if not bs:
925 return ''
926 if isinstance(safe, str):
927 # Normalize 'safe' by converting to bytes and removing non-ASCII chars
928 safe = safe.encode('ascii', 'ignore')
929 else:
930 # List comprehensions are faster than generator expressions.
931 safe = bytes([c for c in safe if c < 128])
932 if not bs.rstrip(_ALWAYS_SAFE_BYTES + safe):
933 return bs.decode()
934 quoter = _byte_quoter_factory(safe)
935 return ''.join([quoter(char) for char in bs])
936
937 def urlencode(query, doseq=False, safe='', encoding=None, errors=None,
938 quote_via=quote_plus):
939 """Encode a dict or sequence of two-element tuples into a URL query string.
940
941 If any values in the query arg are sequences and doseq is true, each
942 sequence element is converted to a separate parameter.
943
944 If the query arg is a sequence of two-element tuples, the order of the
945 parameters in the output will match the order of parameters in the
946 input.
947
948 The components of a query arg may each be either a string or a bytes type.
949
950 The safe, encoding, and errors parameters are passed down to the function
951 specified by quote_via (encoding and errors only if a component is a str).
952 """
953
954 if hasattr(query, "items"):
955 query = query.items()
956 else:
957 # It's a bother at times that strings and string-like objects are
958 # sequences.
959 try:
960 # non-sequence items should not work with len()
961 # non-empty strings will fail this
962 if len(query) and not isinstance(query[0], tuple):
963 raise TypeError
964 # Zero-length sequences of all types will get here and succeed,
965 # but that's a minor nit. Since the original implementation
966 # allowed empty dicts that type of behavior probably should be
967 # preserved for consistency
968 except TypeError as err:
969 raise TypeError("not a valid non-string sequence "
970 "or mapping object") from err
971
972 l = []
973 if not doseq:
974 for k, v in query:
975 if isinstance(k, bytes):
976 k = quote_via(k, safe)
977 else:
978 k = quote_via(str(k), safe, encoding, errors)
979
980 if isinstance(v, bytes):
981 v = quote_via(v, safe)
982 else:
983 v = quote_via(str(v), safe, encoding, errors)
984 l.append(k + '=' + v)
985 else:
986 for k, v in query:
987 if isinstance(k, bytes):
988 k = quote_via(k, safe)
989 else:
990 k = quote_via(str(k), safe, encoding, errors)
991
992 if isinstance(v, bytes):
993 v = quote_via(v, safe)
994 l.append(k + '=' + v)
995 elif isinstance(v, str):
996 v = quote_via(v, safe, encoding, errors)
997 l.append(k + '=' + v)
998 else:
999 try:
1000 # Is this a sufficient test for sequence-ness?
1001 x = len(v)
1002 except TypeError:
1003 # not a sequence
1004 v = quote_via(str(v), safe, encoding, errors)
1005 l.append(k + '=' + v)
1006 else:
1007 # loop over the sequence
1008 for elt in v:
1009 if isinstance(elt, bytes):
1010 elt = quote_via(elt, safe)
1011 else:
1012 elt = quote_via(str(elt), safe, encoding, errors)
1013 l.append(k + '=' + elt)
1014 return '&'.join(l)
1015
1016
1017 def to_bytes(url):
1018 warnings.warn("urllib.parse.to_bytes() is deprecated as of 3.8",
1019 DeprecationWarning, stacklevel=2)
1020 return _to_bytes(url)
1021
1022
1023 def _to_bytes(url):
1024 """to_bytes(u"URL") --> 'URL'."""
1025 # Most URL schemes require ASCII. If that changes, the conversion
1026 # can be relaxed.
1027 # XXX get rid of to_bytes()
1028 if isinstance(url, str):
1029 try:
1030 url = url.encode("ASCII").decode()
1031 except UnicodeError:
1032 raise UnicodeError("URL " + repr(url) +
1033 " contains non-ASCII characters")
1034 return url
1035
1036
1037 def unwrap(url):
1038 """Transform a string like '<URL:scheme://host/path>' into 'scheme://host/path'.
1039
1040 The string is returned unchanged if it's not a wrapped URL.
1041 """
1042 url = str(url).strip()
1043 if url[:1] == '<' and url[-1:] == '>':
1044 url = url[1:-1].strip()
1045 if url[:4] == 'URL:':
1046 url = url[4:].strip()
1047 return url
1048
1049
1050 def splittype(url):
1051 warnings.warn("urllib.parse.splittype() is deprecated as of 3.8, "
1052 "use urllib.parse.urlparse() instead",
1053 DeprecationWarning, stacklevel=2)
1054 return _splittype(url)
1055
1056
1057 _typeprog = None
1058 def _splittype(url):
1059 """splittype('type:opaquestring') --> 'type', 'opaquestring'."""
1060 global _typeprog
1061 if _typeprog is None:
1062 _typeprog = re.compile('([^/:]+):(.*)', re.DOTALL)
1063
1064 match = _typeprog.match(url)
1065 if match:
1066 scheme, data = match.groups()
1067 return scheme.lower(), data
1068 return None, url
1069
1070
1071 def splithost(url):
1072 warnings.warn("urllib.parse.splithost() is deprecated as of 3.8, "
1073 "use urllib.parse.urlparse() instead",
1074 DeprecationWarning, stacklevel=2)
1075 return _splithost(url)
1076
1077
1078 _hostprog = None
1079 def _splithost(url):
1080 """splithost('//host[:port]/path') --> 'host[:port]', '/path'."""
1081 global _hostprog
1082 if _hostprog is None:
1083 _hostprog = re.compile('//([^/#?]*)(.*)', re.DOTALL)
1084
1085 match = _hostprog.match(url)
1086 if match:
1087 host_port, path = match.groups()
1088 if path and path[0] != '/':
1089 path = '/' + path
1090 return host_port, path
1091 return None, url
1092
1093
1094 def splituser(host):
1095 warnings.warn("urllib.parse.splituser() is deprecated as of 3.8, "
1096 "use urllib.parse.urlparse() instead",
1097 DeprecationWarning, stacklevel=2)
1098 return _splituser(host)
1099
1100
1101 def _splituser(host):
1102 """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'."""
1103 user, delim, host = host.rpartition('@')
1104 return (user if delim else None), host
1105
1106
1107 def splitpasswd(user):
1108 warnings.warn("urllib.parse.splitpasswd() is deprecated as of 3.8, "
1109 "use urllib.parse.urlparse() instead",
1110 DeprecationWarning, stacklevel=2)
1111 return _splitpasswd(user)
1112
1113
1114 def _splitpasswd(user):
1115 """splitpasswd('user:passwd') -> 'user', 'passwd'."""
1116 user, delim, passwd = user.partition(':')
1117 return user, (passwd if delim else None)
1118
1119
1120 def splitport(host):
1121 warnings.warn("urllib.parse.splitport() is deprecated as of 3.8, "
1122 "use urllib.parse.urlparse() instead",
1123 DeprecationWarning, stacklevel=2)
1124 return _splitport(host)
1125
1126
1127 # splittag('/path#tag') --> '/path', 'tag'
1128 _portprog = None
1129 def _splitport(host):
1130 """splitport('host:port') --> 'host', 'port'."""
1131 global _portprog
1132 if _portprog is None:
1133 _portprog = re.compile('(.*):([0-9]*)', re.DOTALL)
1134
1135 match = _portprog.fullmatch(host)
1136 if match:
1137 host, port = match.groups()
1138 if port:
1139 return host, port
1140 return host, None
1141
1142
1143 def splitnport(host, defport=-1):
1144 warnings.warn("urllib.parse.splitnport() is deprecated as of 3.8, "
1145 "use urllib.parse.urlparse() instead",
1146 DeprecationWarning, stacklevel=2)
1147 return _splitnport(host, defport)
1148
1149
1150 def _splitnport(host, defport=-1):
1151 """Split host and port, returning numeric port.
1152 Return given default port if no ':' found; defaults to -1.
1153 Return numerical port if a valid number is found after ':'.
1154 Return None if ':' but not a valid number."""
1155 host, delim, port = host.rpartition(':')
1156 if not delim:
1157 host = port
1158 elif port:
1159 if port.isdigit() and port.isascii():
1160 nport = int(port)
1161 else:
1162 nport = None
1163 return host, nport
1164 return host, defport
1165
1166
1167 def splitquery(url):
1168 warnings.warn("urllib.parse.splitquery() is deprecated as of 3.8, "
1169 "use urllib.parse.urlparse() instead",
1170 DeprecationWarning, stacklevel=2)
1171 return _splitquery(url)
1172
1173
1174 def _splitquery(url):
1175 """splitquery('/path?query') --> '/path', 'query'."""
1176 path, delim, query = url.rpartition('?')
1177 if delim:
1178 return path, query
1179 return url, None
1180
1181
1182 def splittag(url):
1183 warnings.warn("urllib.parse.splittag() is deprecated as of 3.8, "
1184 "use urllib.parse.urlparse() instead",
1185 DeprecationWarning, stacklevel=2)
1186 return _splittag(url)
1187
1188
1189 def _splittag(url):
1190 """splittag('/path#tag') --> '/path', 'tag'."""
1191 path, delim, tag = url.rpartition('#')
1192 if delim:
1193 return path, tag
1194 return url, None
1195
1196
1197 def splitattr(url):
1198 warnings.warn("urllib.parse.splitattr() is deprecated as of 3.8, "
1199 "use urllib.parse.urlparse() instead",
1200 DeprecationWarning, stacklevel=2)
1201 return _splitattr(url)
1202
1203
1204 def _splitattr(url):
1205 """splitattr('/path;attr1=value1;attr2=value2;...') ->
1206 '/path', ['attr1=value1', 'attr2=value2', ...]."""
1207 words = url.split(';')
1208 return words[0], words[1:]
1209
1210
1211 def splitvalue(attr):
1212 warnings.warn("urllib.parse.splitvalue() is deprecated as of 3.8, "
1213 "use urllib.parse.parse_qsl() instead",
1214 DeprecationWarning, stacklevel=2)
1215 return _splitvalue(attr)
1216
1217
1218 def _splitvalue(attr):
1219 """splitvalue('attr=value') --> 'attr', 'value'."""
1220 attr, delim, value = attr.partition('=')
1221 return attr, (value if delim else None)