Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 1 | """Parse (absolute and relative) URLs. |
| 2 | |
Senthil Kumaran | fd41e08 | 2010-04-17 14:44:14 +0000 | [diff] [blame] | 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 | |
Benjamin Peterson | d7c3ed5 | 2010-06-27 22:32:30 +0000 | [diff] [blame] | 11 | RFC 2396: "Uniform Resource Identifiers (URI)": Generic Syntax by T. |
Senthil Kumaran | fd41e08 | 2010-04-17 14:44:14 +0000 | [diff] [blame] | 12 | Berners-Lee, R. Fielding, and L. Masinter, August 1998. |
| 13 | |
David Malcolm | ee25568 | 2010-12-02 16:41:00 +0000 | [diff] [blame] | 14 | RFC 2368: "The mailto URL scheme", by P.Hoffman , L Masinter, J. Zawinski, July 1998. |
Senthil Kumaran | fd41e08 | 2010-04-17 14:44:14 +0000 | [diff] [blame] | 15 | |
| 16 | RFC 1808: "Relative Uniform Resource Locators", by R. Fielding, UC Irvine, June |
| 17 | 1995. |
| 18 | |
Benjamin Peterson | d7c3ed5 | 2010-06-27 22:32:30 +0000 | [diff] [blame] | 19 | RFC 1738: "Uniform Resource Locators (URL)" by T. Berners-Lee, L. Masinter, M. |
Senthil Kumaran | fd41e08 | 2010-04-17 14:44:14 +0000 | [diff] [blame] | 20 | McCahill, December 1994 |
| 21 | |
Benjamin Peterson | d7c3ed5 | 2010-06-27 22:32:30 +0000 | [diff] [blame] | 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 |
Senthil Kumaran | fd41e08 | 2010-04-17 14:44:14 +0000 | [diff] [blame] | 27 | test_urlparse.py provides a good indicator of parsing behavior. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 28 | """ |
| 29 | |
Facundo Batista | 2ac5de2 | 2008-07-07 18:24:11 +0000 | [diff] [blame] | 30 | import sys |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 31 | import collections |
Facundo Batista | 2ac5de2 | 2008-07-07 18:24:11 +0000 | [diff] [blame] | 32 | |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 33 | __all__ = ["urlparse", "urlunparse", "urljoin", "urldefrag", |
Senthil Kumaran | 0256b2a | 2010-10-25 16:36:20 +0000 | [diff] [blame] | 34 | "urlsplit", "urlunsplit", "urlencode", "parse_qs", |
| 35 | "parse_qsl", "quote", "quote_plus", "quote_from_bytes", |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 36 | "unquote", "unquote_plus", "unquote_to_bytes"] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 37 | |
| 38 | # A classification of schemes ('' means apply by default) |
| 39 | uses_relative = ['ftp', 'http', 'gopher', 'nntp', 'imap', |
| 40 | 'wais', 'file', 'https', 'shttp', 'mms', |
| 41 | 'prospero', 'rtsp', 'rtspu', '', 'sftp'] |
| 42 | uses_netloc = ['ftp', 'http', 'gopher', 'nntp', 'telnet', |
| 43 | 'imap', 'wais', 'file', 'mms', 'https', 'shttp', |
| 44 | 'snews', 'prospero', 'rtsp', 'rtspu', 'rsync', '', |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 45 | 'svn', 'svn+ssh', 'sftp', 'nfs', 'git', 'git+ssh'] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 46 | non_hierarchical = ['gopher', 'hdl', 'mailto', 'news', |
| 47 | 'telnet', 'wais', 'imap', 'snews', 'sip', 'sips'] |
| 48 | uses_params = ['ftp', 'hdl', 'prospero', 'http', 'imap', |
| 49 | 'https', 'shttp', 'rtsp', 'rtspu', 'sip', 'sips', |
| 50 | 'mms', '', 'sftp'] |
| 51 | uses_query = ['http', 'wais', 'imap', 'https', 'shttp', 'mms', |
| 52 | 'gopher', 'rtsp', 'rtspu', 'sip', 'sips', ''] |
| 53 | uses_fragment = ['ftp', 'hdl', 'http', 'gopher', 'news', |
| 54 | 'nntp', 'wais', 'https', 'shttp', 'snews', |
| 55 | 'file', 'prospero', ''] |
| 56 | |
| 57 | # Characters valid in scheme names |
| 58 | scheme_chars = ('abcdefghijklmnopqrstuvwxyz' |
| 59 | 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' |
| 60 | '0123456789' |
| 61 | '+-.') |
| 62 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 63 | # XXX: Consider replacing with functools.lru_cache |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 64 | MAX_CACHE_SIZE = 20 |
| 65 | _parse_cache = {} |
| 66 | |
| 67 | def clear_cache(): |
Antoine Pitrou | 2df5fc7 | 2009-12-08 19:38:17 +0000 | [diff] [blame] | 68 | """Clear the parse cache and the quoters cache.""" |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 69 | _parse_cache.clear() |
Antoine Pitrou | 2df5fc7 | 2009-12-08 19:38:17 +0000 | [diff] [blame] | 70 | _safe_quoters.clear() |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 71 | |
| 72 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 73 | # Helpers for bytes handling |
| 74 | # For 3.2, we deliberately require applications that |
| 75 | # handle improperly quoted URLs to do their own |
| 76 | # decoding and encoding. If valid use cases are |
| 77 | # presented, we may relax this by using latin-1 |
| 78 | # decoding internally for 3.3 |
| 79 | _implicit_encoding = 'ascii' |
| 80 | _implicit_errors = 'strict' |
| 81 | |
| 82 | def _noop(obj): |
| 83 | return obj |
| 84 | |
| 85 | def _encode_result(obj, encoding=_implicit_encoding, |
| 86 | errors=_implicit_errors): |
| 87 | return obj.encode(encoding, errors) |
| 88 | |
| 89 | def _decode_args(args, encoding=_implicit_encoding, |
| 90 | errors=_implicit_errors): |
| 91 | return tuple(x.decode(encoding, errors) if x else '' for x in args) |
| 92 | |
| 93 | def _coerce_args(*args): |
| 94 | # Invokes decode if necessary to create str args |
| 95 | # and returns the coerced inputs along with |
| 96 | # an appropriate result coercion function |
| 97 | # - noop for str inputs |
| 98 | # - encoding function otherwise |
| 99 | str_input = isinstance(args[0], str) |
| 100 | for arg in args[1:]: |
| 101 | # We special-case the empty string to support the |
| 102 | # "scheme=''" default argument to some functions |
| 103 | if arg and isinstance(arg, str) != str_input: |
| 104 | raise TypeError("Cannot mix str and non-str arguments") |
| 105 | if str_input: |
| 106 | return args + (_noop,) |
| 107 | return _decode_args(args) + (_encode_result,) |
| 108 | |
| 109 | # Result objects are more helpful than simple tuples |
| 110 | class _ResultMixinStr(object): |
| 111 | """Standard approach to encoding parsed results from str to bytes""" |
| 112 | __slots__ = () |
| 113 | |
| 114 | def encode(self, encoding='ascii', errors='strict'): |
| 115 | return self._encoded_counterpart(*(x.encode(encoding, errors) for x in self)) |
| 116 | |
| 117 | |
| 118 | class _ResultMixinBytes(object): |
| 119 | """Standard approach to decoding parsed results from bytes to str""" |
| 120 | __slots__ = () |
| 121 | |
| 122 | def decode(self, encoding='ascii', errors='strict'): |
| 123 | return self._decoded_counterpart(*(x.decode(encoding, errors) for x in self)) |
| 124 | |
| 125 | |
| 126 | class _NetlocResultMixinBase(object): |
| 127 | """Shared methods for the parsed result objects containing a netloc element""" |
| 128 | __slots__ = () |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 129 | |
| 130 | @property |
| 131 | def username(self): |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 132 | return self._userinfo[0] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 133 | |
| 134 | @property |
| 135 | def password(self): |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 136 | return self._userinfo[1] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 137 | |
| 138 | @property |
| 139 | def hostname(self): |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 140 | hostname = self._hostinfo[0] |
| 141 | if not hostname: |
| 142 | hostname = None |
| 143 | elif hostname is not None: |
| 144 | hostname = hostname.lower() |
| 145 | return hostname |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 146 | |
| 147 | @property |
| 148 | def port(self): |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 149 | port = self._hostinfo[1] |
| 150 | if port is not None: |
| 151 | port = int(port, 10) |
| 152 | return port |
| 153 | |
| 154 | |
| 155 | class _NetlocResultMixinStr(_NetlocResultMixinBase, _ResultMixinStr): |
| 156 | __slots__ = () |
| 157 | |
| 158 | @property |
| 159 | def _userinfo(self): |
| 160 | netloc = self.netloc |
| 161 | userinfo, have_info, hostinfo = netloc.rpartition('@') |
| 162 | if have_info: |
| 163 | username, have_password, password = userinfo.partition(':') |
| 164 | if not have_password: |
| 165 | password = None |
Senthil Kumaran | ad02d23 | 2010-04-16 03:02:13 +0000 | [diff] [blame] | 166 | else: |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 167 | username = password = None |
| 168 | return username, password |
| 169 | |
| 170 | @property |
| 171 | def _hostinfo(self): |
| 172 | netloc = self.netloc |
| 173 | _, _, hostinfo = netloc.rpartition('@') |
| 174 | _, have_open_br, bracketed = hostinfo.partition('[') |
| 175 | if have_open_br: |
| 176 | hostname, _, port = bracketed.partition(']') |
| 177 | _, have_port, port = port.partition(':') |
| 178 | else: |
| 179 | hostname, have_port, port = hostinfo.partition(':') |
| 180 | if not have_port: |
| 181 | port = None |
| 182 | return hostname, port |
| 183 | |
| 184 | |
| 185 | class _NetlocResultMixinBytes(_NetlocResultMixinBase, _ResultMixinBytes): |
| 186 | __slots__ = () |
| 187 | |
| 188 | @property |
| 189 | def _userinfo(self): |
| 190 | netloc = self.netloc |
| 191 | userinfo, have_info, hostinfo = netloc.rpartition(b'@') |
| 192 | if have_info: |
| 193 | username, have_password, password = userinfo.partition(b':') |
| 194 | if not have_password: |
| 195 | password = None |
| 196 | else: |
| 197 | username = password = None |
| 198 | return username, password |
| 199 | |
| 200 | @property |
| 201 | def _hostinfo(self): |
| 202 | netloc = self.netloc |
| 203 | _, _, hostinfo = netloc.rpartition(b'@') |
| 204 | _, have_open_br, bracketed = hostinfo.partition(b'[') |
| 205 | if have_open_br: |
| 206 | hostname, _, port = bracketed.partition(b']') |
| 207 | _, have_port, port = port.partition(b':') |
| 208 | else: |
| 209 | hostname, have_port, port = hostinfo.partition(b':') |
| 210 | if not have_port: |
| 211 | port = None |
| 212 | return hostname, port |
| 213 | |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 214 | |
| 215 | from collections import namedtuple |
| 216 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 217 | _DefragResultBase = namedtuple('DefragResult', 'url fragment') |
| 218 | _SplitResultBase = namedtuple('SplitResult', 'scheme netloc path query fragment') |
| 219 | _ParseResultBase = namedtuple('ParseResult', 'scheme netloc path params query fragment') |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 220 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 221 | # For backwards compatibility, alias _NetlocResultMixinStr |
| 222 | # ResultBase is no longer part of the documented API, but it is |
| 223 | # retained since deprecating it isn't worth the hassle |
| 224 | ResultBase = _NetlocResultMixinStr |
| 225 | |
| 226 | # Structured result objects for string data |
| 227 | class DefragResult(_DefragResultBase, _ResultMixinStr): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 228 | __slots__ = () |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 229 | def geturl(self): |
| 230 | if self.fragment: |
| 231 | return self.url + '#' + self.fragment |
| 232 | else: |
| 233 | return self.url |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 234 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 235 | class SplitResult(_SplitResultBase, _NetlocResultMixinStr): |
| 236 | __slots__ = () |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 237 | def geturl(self): |
| 238 | return urlunsplit(self) |
| 239 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 240 | class ParseResult(_ParseResultBase, _NetlocResultMixinStr): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 241 | __slots__ = () |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 242 | def geturl(self): |
| 243 | return urlunparse(self) |
| 244 | |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 245 | # Structured result objects for bytes data |
| 246 | class DefragResultBytes(_DefragResultBase, _ResultMixinBytes): |
| 247 | __slots__ = () |
| 248 | def geturl(self): |
| 249 | if self.fragment: |
| 250 | return self.url + b'#' + self.fragment |
| 251 | else: |
| 252 | return self.url |
| 253 | |
| 254 | class SplitResultBytes(_SplitResultBase, _NetlocResultMixinBytes): |
| 255 | __slots__ = () |
| 256 | def geturl(self): |
| 257 | return urlunsplit(self) |
| 258 | |
| 259 | class ParseResultBytes(_ParseResultBase, _NetlocResultMixinBytes): |
| 260 | __slots__ = () |
| 261 | def geturl(self): |
| 262 | return urlunparse(self) |
| 263 | |
| 264 | # Set up the encode/decode result pairs |
| 265 | def _fix_result_transcoding(): |
| 266 | _result_pairs = ( |
| 267 | (DefragResult, DefragResultBytes), |
| 268 | (SplitResult, SplitResultBytes), |
| 269 | (ParseResult, ParseResultBytes), |
| 270 | ) |
| 271 | for _decoded, _encoded in _result_pairs: |
| 272 | _decoded._encoded_counterpart = _encoded |
| 273 | _encoded._decoded_counterpart = _decoded |
| 274 | |
| 275 | _fix_result_transcoding() |
| 276 | del _fix_result_transcoding |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 277 | |
| 278 | def urlparse(url, scheme='', allow_fragments=True): |
| 279 | """Parse a URL into 6 components: |
| 280 | <scheme>://<netloc>/<path>;<params>?<query>#<fragment> |
| 281 | Return a 6-tuple: (scheme, netloc, path, params, query, fragment). |
| 282 | Note that we don't break the components up in smaller bits |
| 283 | (e.g. netloc is a single string) and we don't expand % escapes.""" |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 284 | url, scheme, _coerce_result = _coerce_args(url, scheme) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 285 | tuple = urlsplit(url, scheme, allow_fragments) |
| 286 | scheme, netloc, url, query, fragment = tuple |
| 287 | if scheme in uses_params and ';' in url: |
| 288 | url, params = _splitparams(url) |
| 289 | else: |
| 290 | params = '' |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 291 | result = ParseResult(scheme, netloc, url, params, query, fragment) |
| 292 | return _coerce_result(result) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 293 | |
| 294 | def _splitparams(url): |
| 295 | if '/' in url: |
| 296 | i = url.find(';', url.rfind('/')) |
| 297 | if i < 0: |
| 298 | return url, '' |
| 299 | else: |
| 300 | i = url.find(';') |
| 301 | return url[:i], url[i+1:] |
| 302 | |
| 303 | def _splitnetloc(url, start=0): |
| 304 | delim = len(url) # position of end of domain part of url, default is end |
| 305 | for c in '/?#': # look for delimiters; the order is NOT important |
| 306 | wdelim = url.find(c, start) # find first of this delim |
| 307 | if wdelim >= 0: # if found |
| 308 | delim = min(delim, wdelim) # use earliest delim position |
| 309 | return url[start:delim], url[delim:] # return (domain, rest) |
| 310 | |
| 311 | def urlsplit(url, scheme='', allow_fragments=True): |
| 312 | """Parse a URL into 5 components: |
| 313 | <scheme>://<netloc>/<path>?<query>#<fragment> |
| 314 | Return a 5-tuple: (scheme, netloc, path, query, fragment). |
| 315 | Note that we don't break the components up in smaller bits |
| 316 | (e.g. netloc is a single string) and we don't expand % escapes.""" |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 317 | url, scheme, _coerce_result = _coerce_args(url, scheme) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 318 | allow_fragments = bool(allow_fragments) |
| 319 | key = url, scheme, allow_fragments, type(url), type(scheme) |
| 320 | cached = _parse_cache.get(key, None) |
| 321 | if cached: |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 322 | return _coerce_result(cached) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 323 | if len(_parse_cache) >= MAX_CACHE_SIZE: # avoid runaway growth |
| 324 | clear_cache() |
| 325 | netloc = query = fragment = '' |
| 326 | i = url.find(':') |
| 327 | if i > 0: |
| 328 | if url[:i] == 'http': # optimize the common case |
| 329 | scheme = url[:i].lower() |
| 330 | url = url[i+1:] |
| 331 | if url[:2] == '//': |
| 332 | netloc, url = _splitnetloc(url, 2) |
Senthil Kumaran | 7a1e09f | 2010-04-22 12:19:46 +0000 | [diff] [blame] | 333 | if (('[' in netloc and ']' not in netloc) or |
| 334 | (']' in netloc and '[' not in netloc)): |
| 335 | raise ValueError("Invalid IPv6 URL") |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 336 | if allow_fragments and '#' in url: |
| 337 | url, fragment = url.split('#', 1) |
| 338 | if '?' in url: |
| 339 | url, query = url.split('?', 1) |
| 340 | v = SplitResult(scheme, netloc, url, query, fragment) |
| 341 | _parse_cache[key] = v |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 342 | return _coerce_result(v) |
Senthil Kumaran | 84c7d9f | 2010-08-04 04:50:44 +0000 | [diff] [blame] | 343 | if url.endswith(':') or not url[i+1].isdigit(): |
| 344 | for c in url[:i]: |
| 345 | if c not in scheme_chars: |
| 346 | break |
| 347 | else: |
| 348 | scheme, url = url[:i].lower(), url[i+1:] |
Senthil Kumaran | 6be85c5 | 2010-02-19 07:42:50 +0000 | [diff] [blame] | 349 | if url[:2] == '//': |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 350 | netloc, url = _splitnetloc(url, 2) |
Senthil Kumaran | 7a1e09f | 2010-04-22 12:19:46 +0000 | [diff] [blame] | 351 | if (('[' in netloc and ']' not in netloc) or |
| 352 | (']' in netloc and '[' not in netloc)): |
| 353 | raise ValueError("Invalid IPv6 URL") |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 354 | if allow_fragments and scheme in uses_fragment and '#' in url: |
| 355 | url, fragment = url.split('#', 1) |
| 356 | if scheme in uses_query and '?' in url: |
| 357 | url, query = url.split('?', 1) |
| 358 | v = SplitResult(scheme, netloc, url, query, fragment) |
| 359 | _parse_cache[key] = v |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 360 | return _coerce_result(v) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 361 | |
| 362 | def urlunparse(components): |
| 363 | """Put a parsed URL back together again. This may result in a |
| 364 | slightly different, but equivalent URL, if the URL that was parsed |
| 365 | originally had redundant delimiters, e.g. a ? with an empty query |
| 366 | (the draft states that these are equivalent).""" |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 367 | scheme, netloc, url, params, query, fragment, _coerce_result = ( |
| 368 | _coerce_args(*components)) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 369 | if params: |
| 370 | url = "%s;%s" % (url, params) |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 371 | return _coerce_result(urlunsplit((scheme, netloc, url, query, fragment))) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 372 | |
| 373 | def urlunsplit(components): |
Senthil Kumaran | 8749a63 | 2010-06-28 14:08:00 +0000 | [diff] [blame] | 374 | """Combine the elements of a tuple as returned by urlsplit() into a |
| 375 | complete URL as a string. The data argument can be any five-item iterable. |
| 376 | This may result in a slightly different, but equivalent URL, if the URL that |
| 377 | was parsed originally had unnecessary delimiters (for example, a ? with an |
| 378 | empty query; the RFC states that these are equivalent).""" |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 379 | scheme, netloc, url, query, fragment, _coerce_result = ( |
| 380 | _coerce_args(*components)) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 381 | if netloc or (scheme and scheme in uses_netloc and url[:2] != '//'): |
| 382 | if url and url[:1] != '/': url = '/' + url |
| 383 | url = '//' + (netloc or '') + url |
| 384 | if scheme: |
| 385 | url = scheme + ':' + url |
| 386 | if query: |
| 387 | url = url + '?' + query |
| 388 | if fragment: |
| 389 | url = url + '#' + fragment |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 390 | return _coerce_result(url) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 391 | |
| 392 | def urljoin(base, url, allow_fragments=True): |
| 393 | """Join a base URL and a possibly relative URL to form an absolute |
| 394 | interpretation of the latter.""" |
| 395 | if not base: |
| 396 | return url |
| 397 | if not url: |
| 398 | return base |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 399 | base, url, _coerce_result = _coerce_args(base, url) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 400 | bscheme, bnetloc, bpath, bparams, bquery, bfragment = \ |
| 401 | urlparse(base, '', allow_fragments) |
| 402 | scheme, netloc, path, params, query, fragment = \ |
| 403 | urlparse(url, bscheme, allow_fragments) |
| 404 | if scheme != bscheme or scheme not in uses_relative: |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 405 | return _coerce_result(url) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 406 | if scheme in uses_netloc: |
| 407 | if netloc: |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 408 | return _coerce_result(urlunparse((scheme, netloc, path, |
| 409 | params, query, fragment))) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 410 | netloc = bnetloc |
| 411 | if path[:1] == '/': |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 412 | return _coerce_result(urlunparse((scheme, netloc, path, |
| 413 | params, query, fragment))) |
Senthil Kumaran | dca5b86 | 2010-12-17 04:48:45 +0000 | [diff] [blame] | 414 | if not path and not params: |
Facundo Batista | 23e3856 | 2008-08-14 16:55:14 +0000 | [diff] [blame] | 415 | path = bpath |
Senthil Kumaran | dca5b86 | 2010-12-17 04:48:45 +0000 | [diff] [blame] | 416 | params = bparams |
Facundo Batista | 23e3856 | 2008-08-14 16:55:14 +0000 | [diff] [blame] | 417 | if not query: |
| 418 | query = bquery |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 419 | return _coerce_result(urlunparse((scheme, netloc, path, |
| 420 | params, query, fragment))) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 421 | segments = bpath.split('/')[:-1] + path.split('/') |
| 422 | # XXX The stuff below is bogus in various ways... |
| 423 | if segments[-1] == '.': |
| 424 | segments[-1] = '' |
| 425 | while '.' in segments: |
| 426 | segments.remove('.') |
| 427 | while 1: |
| 428 | i = 1 |
| 429 | n = len(segments) - 1 |
| 430 | while i < n: |
| 431 | if (segments[i] == '..' |
| 432 | and segments[i-1] not in ('', '..')): |
| 433 | del segments[i-1:i+1] |
| 434 | break |
| 435 | i = i+1 |
| 436 | else: |
| 437 | break |
| 438 | if segments == ['', '..']: |
| 439 | segments[-1] = '' |
| 440 | elif len(segments) >= 2 and segments[-1] == '..': |
| 441 | segments[-2:] = [''] |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 442 | return _coerce_result(urlunparse((scheme, netloc, '/'.join(segments), |
| 443 | params, query, fragment))) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 444 | |
| 445 | def urldefrag(url): |
| 446 | """Removes any existing fragment from URL. |
| 447 | |
| 448 | Returns a tuple of the defragmented URL and the fragment. If |
| 449 | the URL contained no fragments, the second element is the |
| 450 | empty string. |
| 451 | """ |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 452 | url, _coerce_result = _coerce_args(url) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 453 | if '#' in url: |
| 454 | s, n, p, a, q, frag = urlparse(url) |
| 455 | defrag = urlunparse((s, n, p, a, q, '')) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 456 | else: |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 457 | frag = '' |
| 458 | defrag = url |
| 459 | return _coerce_result(DefragResult(defrag, frag)) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 460 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 461 | def unquote_to_bytes(string): |
| 462 | """unquote_to_bytes('abc%20def') -> b'abc def'.""" |
| 463 | # Note: strings are encoded as UTF-8. This is only an issue if it contains |
| 464 | # unescaped non-ASCII characters, which URIs should not. |
Florent Xicluna | 82a3f8a | 2010-08-14 18:30:35 +0000 | [diff] [blame] | 465 | if not string: |
| 466 | # Is it a string-like object? |
| 467 | string.split |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 468 | return b'' |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 469 | if isinstance(string, str): |
| 470 | string = string.encode('utf-8') |
| 471 | res = string.split(b'%') |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 472 | if len(res) == 1: |
| 473 | return string |
| 474 | string = res[0] |
| 475 | for item in res[1:]: |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 476 | try: |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 477 | string += bytes([int(item[:2], 16)]) + item[2:] |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 478 | except ValueError: |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 479 | string += b'%' + item |
| 480 | return string |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 481 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 482 | def unquote(string, encoding='utf-8', errors='replace'): |
| 483 | """Replace %xx escapes by their single-character equivalent. The optional |
| 484 | encoding and errors parameters specify how to decode percent-encoded |
| 485 | sequences into Unicode characters, as accepted by the bytes.decode() |
| 486 | method. |
| 487 | By default, percent-encoded sequences are decoded with UTF-8, and invalid |
| 488 | sequences are replaced by a placeholder character. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 489 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 490 | unquote('abc%20def') -> 'abc def'. |
| 491 | """ |
Florent Xicluna | c049fca | 2010-07-31 08:56:55 +0000 | [diff] [blame] | 492 | if string == '': |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 493 | return string |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 494 | res = string.split('%') |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 495 | if len(res) == 1: |
| 496 | return string |
| 497 | if encoding is None: |
| 498 | encoding = 'utf-8' |
| 499 | if errors is None: |
| 500 | errors = 'replace' |
Florent Xicluna | 0f78a94 | 2010-05-17 18:01:22 +0000 | [diff] [blame] | 501 | # pct_sequence: contiguous sequence of percent-encoded bytes, decoded |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 502 | pct_sequence = b'' |
| 503 | string = res[0] |
| 504 | for item in res[1:]: |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 505 | try: |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 506 | if not item: |
| 507 | raise ValueError |
| 508 | pct_sequence += bytes.fromhex(item[:2]) |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 509 | rest = item[2:] |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 510 | if not rest: |
| 511 | # This segment was just a single percent-encoded character. |
| 512 | # May be part of a sequence of code units, so delay decoding. |
| 513 | # (Stored in pct_sequence). |
| 514 | continue |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 515 | except ValueError: |
| 516 | rest = '%' + item |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 517 | # Encountered non-percent-encoded characters. Flush the current |
| 518 | # pct_sequence. |
| 519 | string += pct_sequence.decode(encoding, errors) + rest |
| 520 | pct_sequence = b'' |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 521 | if pct_sequence: |
| 522 | # Flush the final pct_sequence |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 523 | string += pct_sequence.decode(encoding, errors) |
| 524 | return string |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 525 | |
Victor Stinner | ac71c54 | 2011-01-14 12:52:12 +0000 | [diff] [blame^] | 526 | def parse_qs(qs, keep_blank_values=False, strict_parsing=False, |
| 527 | encoding='utf-8', errors='replace'): |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 528 | """Parse a query given as a string argument. |
| 529 | |
| 530 | Arguments: |
| 531 | |
Senthil Kumaran | 30e86a4 | 2010-08-09 20:01:35 +0000 | [diff] [blame] | 532 | qs: percent-encoded query string to be parsed |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 533 | |
| 534 | keep_blank_values: flag indicating whether blank values in |
Senthil Kumaran | 30e86a4 | 2010-08-09 20:01:35 +0000 | [diff] [blame] | 535 | percent-encoded queries should be treated as blank strings. |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 536 | A true value indicates that blanks should be retained as |
| 537 | blank strings. The default false value indicates that |
| 538 | blank values are to be ignored and treated as if they were |
| 539 | not included. |
| 540 | |
| 541 | strict_parsing: flag indicating what to do with parsing errors. |
| 542 | If false (the default), errors are silently ignored. |
| 543 | If true, errors raise a ValueError exception. |
Victor Stinner | ac71c54 | 2011-01-14 12:52:12 +0000 | [diff] [blame^] | 544 | |
| 545 | encoding and errors: specify how to decode percent-encoded sequences |
| 546 | into Unicode characters, as accepted by the bytes.decode() method. |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 547 | """ |
| 548 | dict = {} |
Victor Stinner | ac71c54 | 2011-01-14 12:52:12 +0000 | [diff] [blame^] | 549 | pairs = parse_qsl(qs, keep_blank_values, strict_parsing, |
| 550 | encoding=encoding, errors=errors) |
| 551 | for name, value in pairs: |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 552 | if name in dict: |
| 553 | dict[name].append(value) |
| 554 | else: |
| 555 | dict[name] = [value] |
| 556 | return dict |
| 557 | |
Victor Stinner | ac71c54 | 2011-01-14 12:52:12 +0000 | [diff] [blame^] | 558 | def parse_qsl(qs, keep_blank_values=False, strict_parsing=False, |
| 559 | encoding='utf-8', errors='replace'): |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 560 | """Parse a query given as a string argument. |
| 561 | |
| 562 | Arguments: |
| 563 | |
Senthil Kumaran | 30e86a4 | 2010-08-09 20:01:35 +0000 | [diff] [blame] | 564 | qs: percent-encoded query string to be parsed |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 565 | |
| 566 | keep_blank_values: flag indicating whether blank values in |
Senthil Kumaran | 30e86a4 | 2010-08-09 20:01:35 +0000 | [diff] [blame] | 567 | percent-encoded queries should be treated as blank strings. A |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 568 | true value indicates that blanks should be retained as blank |
| 569 | strings. The default false value indicates that blank values |
| 570 | are to be ignored and treated as if they were not included. |
| 571 | |
| 572 | strict_parsing: flag indicating what to do with parsing errors. If |
| 573 | false (the default), errors are silently ignored. If true, |
| 574 | errors raise a ValueError exception. |
| 575 | |
Victor Stinner | ac71c54 | 2011-01-14 12:52:12 +0000 | [diff] [blame^] | 576 | encoding and errors: specify how to decode percent-encoded sequences |
| 577 | into Unicode characters, as accepted by the bytes.decode() method. |
| 578 | |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 579 | Returns a list, as G-d intended. |
| 580 | """ |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 581 | qs, _coerce_result = _coerce_args(qs) |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 582 | pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')] |
| 583 | r = [] |
| 584 | for name_value in pairs: |
| 585 | if not name_value and not strict_parsing: |
| 586 | continue |
| 587 | nv = name_value.split('=', 1) |
| 588 | if len(nv) != 2: |
| 589 | if strict_parsing: |
| 590 | raise ValueError("bad query field: %r" % (name_value,)) |
| 591 | # Handle case of a control-name with no equal sign |
| 592 | if keep_blank_values: |
| 593 | nv.append('') |
| 594 | else: |
| 595 | continue |
| 596 | if len(nv[1]) or keep_blank_values: |
Victor Stinner | ac71c54 | 2011-01-14 12:52:12 +0000 | [diff] [blame^] | 597 | name = nv[0].replace('+', ' ') |
| 598 | name = unquote(name, encoding=encoding, errors=errors) |
| 599 | name = _coerce_result(name) |
| 600 | value = nv[1].replace('+', ' ') |
| 601 | value = unquote(value, encoding=encoding, errors=errors) |
| 602 | value = _coerce_result(value) |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 603 | r.append((name, value)) |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 604 | return r |
| 605 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 606 | def unquote_plus(string, encoding='utf-8', errors='replace'): |
| 607 | """Like unquote(), but also replace plus signs by spaces, as required for |
| 608 | unquoting HTML form values. |
| 609 | |
| 610 | unquote_plus('%7e/abc+def') -> '~/abc def' |
| 611 | """ |
| 612 | string = string.replace('+', ' ') |
| 613 | return unquote(string, encoding, errors) |
| 614 | |
| 615 | _ALWAYS_SAFE = frozenset(b'ABCDEFGHIJKLMNOPQRSTUVWXYZ' |
| 616 | b'abcdefghijklmnopqrstuvwxyz' |
| 617 | b'0123456789' |
| 618 | b'_.-') |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 619 | _ALWAYS_SAFE_BYTES = bytes(_ALWAYS_SAFE) |
| 620 | _safe_quoters = {} |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 621 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 622 | class Quoter(collections.defaultdict): |
| 623 | """A mapping from bytes (in range(0,256)) to strings. |
| 624 | |
| 625 | String values are percent-encoded byte values, unless the key < 128, and |
| 626 | in the "safe" set (either the specified safe set, or default set). |
| 627 | """ |
| 628 | # Keeps a cache internally, using defaultdict, for efficiency (lookups |
| 629 | # of cached keys don't call Python code at all). |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 630 | def __init__(self, safe): |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 631 | """safe: bytes object.""" |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 632 | self.safe = _ALWAYS_SAFE.union(safe) |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 633 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 634 | def __repr__(self): |
| 635 | # Without this, will just display as a defaultdict |
| 636 | return "<Quoter %r>" % dict(self) |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 637 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 638 | def __missing__(self, b): |
| 639 | # Handle a cache miss. Store quoted string in cache and return. |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 640 | res = chr(b) if b in self.safe else '%{:02X}'.format(b) |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 641 | self[b] = res |
| 642 | return res |
| 643 | |
| 644 | def quote(string, safe='/', encoding=None, errors=None): |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 645 | """quote('abc def') -> 'abc%20def' |
| 646 | |
| 647 | Each part of a URL, e.g. the path info, the query, etc., has a |
| 648 | different set of reserved characters that must be quoted. |
| 649 | |
| 650 | RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists |
| 651 | the following reserved characters. |
| 652 | |
| 653 | reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | |
| 654 | "$" | "," |
| 655 | |
| 656 | Each of these characters is reserved in some component of a URL, |
| 657 | but not necessarily in all of them. |
| 658 | |
| 659 | By default, the quote function is intended for quoting the path |
| 660 | section of a URL. Thus, it will not encode '/'. This character |
| 661 | is reserved, but in typical usage the quote function is being |
| 662 | called on a path where the existing slash characters are used as |
| 663 | reserved characters. |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 664 | |
| 665 | string and safe may be either str or bytes objects. encoding must |
| 666 | not be specified if string is a str. |
| 667 | |
| 668 | The optional encoding and errors parameters specify how to deal with |
| 669 | non-ASCII characters, as accepted by the str.encode method. |
| 670 | By default, encoding='utf-8' (characters are encoded with UTF-8), and |
| 671 | errors='strict' (unsupported characters raise a UnicodeEncodeError). |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 672 | """ |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 673 | if isinstance(string, str): |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 674 | if not string: |
| 675 | return string |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 676 | if encoding is None: |
| 677 | encoding = 'utf-8' |
| 678 | if errors is None: |
| 679 | errors = 'strict' |
| 680 | string = string.encode(encoding, errors) |
| 681 | else: |
| 682 | if encoding is not None: |
| 683 | raise TypeError("quote() doesn't support 'encoding' for bytes") |
| 684 | if errors is not None: |
| 685 | raise TypeError("quote() doesn't support 'errors' for bytes") |
| 686 | return quote_from_bytes(string, safe) |
| 687 | |
| 688 | def quote_plus(string, safe='', encoding=None, errors=None): |
| 689 | """Like quote(), but also replace ' ' with '+', as required for quoting |
| 690 | HTML form values. Plus signs in the original string are escaped unless |
| 691 | they are included in safe. It also does not have safe default to '/'. |
| 692 | """ |
Jeremy Hylton | f819886 | 2009-03-26 16:55:08 +0000 | [diff] [blame] | 693 | # Check if ' ' in string, where string may either be a str or bytes. If |
| 694 | # there are no spaces, the regular quote will produce the right answer. |
| 695 | if ((isinstance(string, str) and ' ' not in string) or |
| 696 | (isinstance(string, bytes) and b' ' not in string)): |
| 697 | return quote(string, safe, encoding, errors) |
| 698 | if isinstance(safe, str): |
| 699 | space = ' ' |
| 700 | else: |
| 701 | space = b' ' |
Georg Brandl | faf4149 | 2009-05-26 18:31:11 +0000 | [diff] [blame] | 702 | string = quote(string, safe + space, encoding, errors) |
Jeremy Hylton | f819886 | 2009-03-26 16:55:08 +0000 | [diff] [blame] | 703 | return string.replace(' ', '+') |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 704 | |
| 705 | def quote_from_bytes(bs, safe='/'): |
| 706 | """Like quote(), but accepts a bytes object rather than a str, and does |
| 707 | not perform string-to-bytes encoding. It always returns an ASCII string. |
| 708 | quote_from_bytes(b'abc def\xab') -> 'abc%20def%AB' |
| 709 | """ |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 710 | if not isinstance(bs, (bytes, bytearray)): |
| 711 | raise TypeError("quote_from_bytes() expected bytes") |
| 712 | if not bs: |
| 713 | return '' |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 714 | if isinstance(safe, str): |
| 715 | # Normalize 'safe' by converting to bytes and removing non-ASCII chars |
| 716 | safe = safe.encode('ascii', 'ignore') |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 717 | else: |
| 718 | safe = bytes([c for c in safe if c < 128]) |
| 719 | if not bs.rstrip(_ALWAYS_SAFE_BYTES + safe): |
| 720 | return bs.decode() |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 721 | try: |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 722 | quoter = _safe_quoters[safe] |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 723 | except KeyError: |
Florent Xicluna | c7b8e86 | 2010-05-17 17:33:07 +0000 | [diff] [blame] | 724 | _safe_quoters[safe] = quoter = Quoter(safe).__getitem__ |
| 725 | return ''.join([quoter(char) for char in bs]) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 726 | |
Senthil Kumaran | df022da | 2010-07-03 17:48:22 +0000 | [diff] [blame] | 727 | def urlencode(query, doseq=False, safe='', encoding=None, errors=None): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 728 | """Encode a sequence of two-element tuples or dictionary into a URL query string. |
| 729 | |
| 730 | If any values in the query arg are sequences and doseq is true, each |
| 731 | sequence element is converted to a separate parameter. |
| 732 | |
| 733 | If the query arg is a sequence of two-element tuples, the order of the |
| 734 | parameters in the output will match the order of parameters in the |
| 735 | input. |
Senthil Kumaran | df022da | 2010-07-03 17:48:22 +0000 | [diff] [blame] | 736 | |
| 737 | The query arg may be either a string or a bytes type. When query arg is a |
| 738 | string, the safe, encoding and error parameters are sent the quote_plus for |
| 739 | encoding. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 740 | """ |
| 741 | |
Jeremy Hylton | a4de60a | 2009-03-26 14:49:26 +0000 | [diff] [blame] | 742 | if hasattr(query, "items"): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 743 | query = query.items() |
| 744 | else: |
Jeremy Hylton | 230feba | 2009-03-26 16:56:59 +0000 | [diff] [blame] | 745 | # It's a bother at times that strings and string-like objects are |
| 746 | # sequences. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 747 | try: |
| 748 | # non-sequence items should not work with len() |
| 749 | # non-empty strings will fail this |
| 750 | if len(query) and not isinstance(query[0], tuple): |
| 751 | raise TypeError |
Jeremy Hylton | 230feba | 2009-03-26 16:56:59 +0000 | [diff] [blame] | 752 | # Zero-length sequences of all types will get here and succeed, |
| 753 | # but that's a minor nit. Since the original implementation |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 754 | # allowed empty dicts that type of behavior probably should be |
| 755 | # preserved for consistency |
| 756 | except TypeError: |
Jeremy Hylton | a4de60a | 2009-03-26 14:49:26 +0000 | [diff] [blame] | 757 | ty, va, tb = sys.exc_info() |
| 758 | raise TypeError("not a valid non-string sequence " |
| 759 | "or mapping object").with_traceback(tb) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 760 | |
| 761 | l = [] |
| 762 | if not doseq: |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 763 | for k, v in query: |
Senthil Kumaran | df022da | 2010-07-03 17:48:22 +0000 | [diff] [blame] | 764 | if isinstance(k, bytes): |
| 765 | k = quote_plus(k, safe) |
| 766 | else: |
| 767 | k = quote_plus(str(k), safe, encoding, errors) |
| 768 | |
| 769 | if isinstance(v, bytes): |
| 770 | v = quote_plus(v, safe) |
| 771 | else: |
| 772 | v = quote_plus(str(v), safe, encoding, errors) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 773 | l.append(k + '=' + v) |
| 774 | else: |
| 775 | for k, v in query: |
Senthil Kumaran | df022da | 2010-07-03 17:48:22 +0000 | [diff] [blame] | 776 | if isinstance(k, bytes): |
| 777 | k = quote_plus(k, safe) |
| 778 | else: |
| 779 | k = quote_plus(str(k), safe, encoding, errors) |
| 780 | |
| 781 | if isinstance(v, bytes): |
| 782 | v = quote_plus(v, safe) |
| 783 | l.append(k + '=' + v) |
| 784 | elif isinstance(v, str): |
| 785 | v = quote_plus(v, safe, encoding, errors) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 786 | l.append(k + '=' + v) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 787 | else: |
| 788 | try: |
Jeremy Hylton | 230feba | 2009-03-26 16:56:59 +0000 | [diff] [blame] | 789 | # Is this a sufficient test for sequence-ness? |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 790 | x = len(v) |
| 791 | except TypeError: |
| 792 | # not a sequence |
Senthil Kumaran | df022da | 2010-07-03 17:48:22 +0000 | [diff] [blame] | 793 | v = quote_plus(str(v), safe, encoding, errors) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 794 | l.append(k + '=' + v) |
| 795 | else: |
| 796 | # loop over the sequence |
| 797 | for elt in v: |
Senthil Kumaran | df022da | 2010-07-03 17:48:22 +0000 | [diff] [blame] | 798 | if isinstance(elt, bytes): |
| 799 | elt = quote_plus(elt, safe) |
| 800 | else: |
| 801 | elt = quote_plus(str(elt), safe, encoding, errors) |
| 802 | l.append(k + '=' + elt) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 803 | return '&'.join(l) |
| 804 | |
| 805 | # Utilities to parse URLs (most of these return None for missing parts): |
| 806 | # unwrap('<URL:type://host/path>') --> 'type://host/path' |
| 807 | # splittype('type:opaquestring') --> 'type', 'opaquestring' |
| 808 | # splithost('//host[:port]/path') --> 'host[:port]', '/path' |
| 809 | # splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]' |
| 810 | # splitpasswd('user:passwd') -> 'user', 'passwd' |
| 811 | # splitport('host:port') --> 'host', 'port' |
| 812 | # splitquery('/path?query') --> '/path', 'query' |
| 813 | # splittag('/path#tag') --> '/path', 'tag' |
| 814 | # splitattr('/path;attr1=value1;attr2=value2;...') -> |
| 815 | # '/path', ['attr1=value1', 'attr2=value2', ...] |
| 816 | # splitvalue('attr=value') --> 'attr', 'value' |
| 817 | # urllib.parse.unquote('abc%20def') -> 'abc def' |
| 818 | # quote('abc def') -> 'abc%20def') |
| 819 | |
Georg Brandl | 13e8946 | 2008-07-01 19:56:00 +0000 | [diff] [blame] | 820 | def to_bytes(url): |
| 821 | """to_bytes(u"URL") --> 'URL'.""" |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 822 | # Most URL schemes require ASCII. If that changes, the conversion |
| 823 | # can be relaxed. |
Georg Brandl | 13e8946 | 2008-07-01 19:56:00 +0000 | [diff] [blame] | 824 | # XXX get rid of to_bytes() |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 825 | if isinstance(url, str): |
| 826 | try: |
| 827 | url = url.encode("ASCII").decode() |
| 828 | except UnicodeError: |
| 829 | raise UnicodeError("URL " + repr(url) + |
| 830 | " contains non-ASCII characters") |
| 831 | return url |
| 832 | |
| 833 | def unwrap(url): |
| 834 | """unwrap('<URL:type://host/path>') --> 'type://host/path'.""" |
| 835 | url = str(url).strip() |
| 836 | if url[:1] == '<' and url[-1:] == '>': |
| 837 | url = url[1:-1].strip() |
| 838 | if url[:4] == 'URL:': url = url[4:].strip() |
| 839 | return url |
| 840 | |
| 841 | _typeprog = None |
| 842 | def splittype(url): |
| 843 | """splittype('type:opaquestring') --> 'type', 'opaquestring'.""" |
| 844 | global _typeprog |
| 845 | if _typeprog is None: |
| 846 | import re |
| 847 | _typeprog = re.compile('^([^/:]+):') |
| 848 | |
| 849 | match = _typeprog.match(url) |
| 850 | if match: |
| 851 | scheme = match.group(1) |
| 852 | return scheme.lower(), url[len(scheme) + 1:] |
| 853 | return None, url |
| 854 | |
| 855 | _hostprog = None |
| 856 | def splithost(url): |
| 857 | """splithost('//host[:port]/path') --> 'host[:port]', '/path'.""" |
| 858 | global _hostprog |
| 859 | if _hostprog is None: |
| 860 | import re |
| 861 | _hostprog = re.compile('^//([^/?]*)(.*)$') |
| 862 | |
| 863 | match = _hostprog.match(url) |
Senthil Kumaran | c295862 | 2010-11-22 04:48:26 +0000 | [diff] [blame] | 864 | if match: |
| 865 | host_port = match.group(1) |
| 866 | path = match.group(2) |
| 867 | if path and not path.startswith('/'): |
| 868 | path = '/' + path |
| 869 | return host_port, path |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 870 | return None, url |
| 871 | |
| 872 | _userprog = None |
| 873 | def splituser(host): |
| 874 | """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'.""" |
| 875 | global _userprog |
| 876 | if _userprog is None: |
| 877 | import re |
| 878 | _userprog = re.compile('^(.*)@(.*)$') |
| 879 | |
| 880 | match = _userprog.match(host) |
Senthil Kumaran | daa29d0 | 2010-11-18 15:36:41 +0000 | [diff] [blame] | 881 | if match: return match.group(1, 2) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 882 | return None, host |
| 883 | |
| 884 | _passwdprog = None |
| 885 | def splitpasswd(user): |
| 886 | """splitpasswd('user:passwd') -> 'user', 'passwd'.""" |
| 887 | global _passwdprog |
| 888 | if _passwdprog is None: |
| 889 | import re |
Senthil Kumaran | eaaec27 | 2009-03-30 21:54:41 +0000 | [diff] [blame] | 890 | _passwdprog = re.compile('^([^:]*):(.*)$',re.S) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 891 | |
| 892 | match = _passwdprog.match(user) |
| 893 | if match: return match.group(1, 2) |
| 894 | return user, None |
| 895 | |
| 896 | # splittag('/path#tag') --> '/path', 'tag' |
| 897 | _portprog = None |
| 898 | def splitport(host): |
| 899 | """splitport('host:port') --> 'host', 'port'.""" |
| 900 | global _portprog |
| 901 | if _portprog is None: |
| 902 | import re |
| 903 | _portprog = re.compile('^(.*):([0-9]+)$') |
| 904 | |
| 905 | match = _portprog.match(host) |
| 906 | if match: return match.group(1, 2) |
| 907 | return host, None |
| 908 | |
| 909 | _nportprog = None |
| 910 | def splitnport(host, defport=-1): |
| 911 | """Split host and port, returning numeric port. |
| 912 | Return given default port if no ':' found; defaults to -1. |
| 913 | Return numerical port if a valid number are found after ':'. |
| 914 | Return None if ':' but not a valid number.""" |
| 915 | global _nportprog |
| 916 | if _nportprog is None: |
| 917 | import re |
| 918 | _nportprog = re.compile('^(.*):(.*)$') |
| 919 | |
| 920 | match = _nportprog.match(host) |
| 921 | if match: |
| 922 | host, port = match.group(1, 2) |
| 923 | try: |
| 924 | if not port: raise ValueError("no digits") |
| 925 | nport = int(port) |
| 926 | except ValueError: |
| 927 | nport = None |
| 928 | return host, nport |
| 929 | return host, defport |
| 930 | |
| 931 | _queryprog = None |
| 932 | def splitquery(url): |
| 933 | """splitquery('/path?query') --> '/path', 'query'.""" |
| 934 | global _queryprog |
| 935 | if _queryprog is None: |
| 936 | import re |
| 937 | _queryprog = re.compile('^(.*)\?([^?]*)$') |
| 938 | |
| 939 | match = _queryprog.match(url) |
| 940 | if match: return match.group(1, 2) |
| 941 | return url, None |
| 942 | |
| 943 | _tagprog = None |
| 944 | def splittag(url): |
| 945 | """splittag('/path#tag') --> '/path', 'tag'.""" |
| 946 | global _tagprog |
| 947 | if _tagprog is None: |
| 948 | import re |
| 949 | _tagprog = re.compile('^(.*)#([^#]*)$') |
| 950 | |
| 951 | match = _tagprog.match(url) |
| 952 | if match: return match.group(1, 2) |
| 953 | return url, None |
| 954 | |
| 955 | def splitattr(url): |
| 956 | """splitattr('/path;attr1=value1;attr2=value2;...') -> |
| 957 | '/path', ['attr1=value1', 'attr2=value2', ...].""" |
| 958 | words = url.split(';') |
| 959 | return words[0], words[1:] |
| 960 | |
| 961 | _valueprog = None |
| 962 | def splitvalue(attr): |
| 963 | """splitvalue('attr=value') --> 'attr', 'value'.""" |
| 964 | global _valueprog |
| 965 | if _valueprog is None: |
| 966 | import re |
| 967 | _valueprog = re.compile('^([^=]*)=(.*)$') |
| 968 | |
| 969 | match = _valueprog.match(attr) |
| 970 | if match: return match.group(1, 2) |
| 971 | return attr, None |