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 | |
| 11 | RFC2396: "Uniform Resource Identifiers (URI)": Generic Syntax by T. |
| 12 | Berners-Lee, R. Fielding, and L. Masinter, August 1998. |
| 13 | |
| 14 | RFC2368: "The mailto URL scheme", by P.Hoffman , L Masinter, J. Zwinski, July 1998. |
| 15 | |
| 16 | RFC 1808: "Relative Uniform Resource Locators", by R. Fielding, UC Irvine, June |
| 17 | 1995. |
| 18 | |
| 19 | RFC1738: "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 changes to urlparse module |
| 23 | should conform to this. urlparse module is not entirely compliant with this. |
| 24 | The defacto scenarios of parsing are considered sometimes and for backward |
| 25 | compatiblity purposes, older RFC uses of parsing are retained. The testcases in |
| 26 | test_urlparse.py provides a good indicator of parsing behavior. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 27 | """ |
| 28 | |
Facundo Batista | 2ac5de2 | 2008-07-07 18:24:11 +0000 | [diff] [blame] | 29 | import sys |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 30 | import collections |
Facundo Batista | 2ac5de2 | 2008-07-07 18:24:11 +0000 | [diff] [blame] | 31 | |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 32 | __all__ = ["urlparse", "urlunparse", "urljoin", "urldefrag", |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 33 | "urlsplit", "urlunsplit", "parse_qs", "parse_qsl", |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 34 | "quote", "quote_plus", "quote_from_bytes", |
| 35 | "unquote", "unquote_plus", "unquote_to_bytes"] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 36 | |
| 37 | # A classification of schemes ('' means apply by default) |
| 38 | uses_relative = ['ftp', 'http', 'gopher', 'nntp', 'imap', |
| 39 | 'wais', 'file', 'https', 'shttp', 'mms', |
| 40 | 'prospero', 'rtsp', 'rtspu', '', 'sftp'] |
| 41 | uses_netloc = ['ftp', 'http', 'gopher', 'nntp', 'telnet', |
| 42 | 'imap', 'wais', 'file', 'mms', 'https', 'shttp', |
| 43 | 'snews', 'prospero', 'rtsp', 'rtspu', 'rsync', '', |
Senthil Kumaran | eaaec27 | 2009-03-30 21:54:41 +0000 | [diff] [blame] | 44 | 'svn', 'svn+ssh', 'sftp','nfs'] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 45 | non_hierarchical = ['gopher', 'hdl', 'mailto', 'news', |
| 46 | 'telnet', 'wais', 'imap', 'snews', 'sip', 'sips'] |
| 47 | uses_params = ['ftp', 'hdl', 'prospero', 'http', 'imap', |
| 48 | 'https', 'shttp', 'rtsp', 'rtspu', 'sip', 'sips', |
| 49 | 'mms', '', 'sftp'] |
| 50 | uses_query = ['http', 'wais', 'imap', 'https', 'shttp', 'mms', |
| 51 | 'gopher', 'rtsp', 'rtspu', 'sip', 'sips', ''] |
| 52 | uses_fragment = ['ftp', 'hdl', 'http', 'gopher', 'news', |
| 53 | 'nntp', 'wais', 'https', 'shttp', 'snews', |
| 54 | 'file', 'prospero', ''] |
| 55 | |
| 56 | # Characters valid in scheme names |
| 57 | scheme_chars = ('abcdefghijklmnopqrstuvwxyz' |
| 58 | 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' |
| 59 | '0123456789' |
| 60 | '+-.') |
| 61 | |
| 62 | MAX_CACHE_SIZE = 20 |
| 63 | _parse_cache = {} |
| 64 | |
| 65 | def clear_cache(): |
Antoine Pitrou | 2df5fc7 | 2009-12-08 19:38:17 +0000 | [diff] [blame] | 66 | """Clear the parse cache and the quoters cache.""" |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 67 | _parse_cache.clear() |
Antoine Pitrou | 2df5fc7 | 2009-12-08 19:38:17 +0000 | [diff] [blame] | 68 | _safe_quoters.clear() |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 69 | |
| 70 | |
| 71 | class ResultMixin(object): |
| 72 | """Shared methods for the parsed result objects.""" |
| 73 | |
| 74 | @property |
| 75 | def username(self): |
| 76 | netloc = self.netloc |
| 77 | if "@" in netloc: |
| 78 | userinfo = netloc.rsplit("@", 1)[0] |
| 79 | if ":" in userinfo: |
| 80 | userinfo = userinfo.split(":", 1)[0] |
| 81 | return userinfo |
| 82 | return None |
| 83 | |
| 84 | @property |
| 85 | def password(self): |
| 86 | netloc = self.netloc |
| 87 | if "@" in netloc: |
| 88 | userinfo = netloc.rsplit("@", 1)[0] |
| 89 | if ":" in userinfo: |
| 90 | return userinfo.split(":", 1)[1] |
| 91 | return None |
| 92 | |
| 93 | @property |
| 94 | def hostname(self): |
Senthil Kumaran | ad02d23 | 2010-04-16 03:02:13 +0000 | [diff] [blame] | 95 | netloc = self.netloc.split('@')[-1] |
| 96 | if '[' in netloc and ']' in netloc: |
| 97 | return netloc.split(']')[0][1:].lower() |
Senthil Kumaran | ad02d23 | 2010-04-16 03:02:13 +0000 | [diff] [blame] | 98 | elif ':' in netloc: |
| 99 | return netloc.split(':')[0].lower() |
| 100 | elif netloc == '': |
| 101 | return None |
| 102 | else: |
| 103 | return netloc.lower() |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 104 | |
| 105 | @property |
| 106 | def port(self): |
Senthil Kumaran | ad02d23 | 2010-04-16 03:02:13 +0000 | [diff] [blame] | 107 | netloc = self.netloc.split('@')[-1].split(']')[-1] |
| 108 | if ':' in netloc: |
| 109 | port = netloc.split(':')[1] |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 110 | return int(port, 10) |
Senthil Kumaran | ad02d23 | 2010-04-16 03:02:13 +0000 | [diff] [blame] | 111 | else: |
| 112 | return None |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 113 | |
| 114 | from collections import namedtuple |
| 115 | |
| 116 | class SplitResult(namedtuple('SplitResult', 'scheme netloc path query fragment'), ResultMixin): |
| 117 | |
| 118 | __slots__ = () |
| 119 | |
| 120 | def geturl(self): |
| 121 | return urlunsplit(self) |
| 122 | |
| 123 | |
| 124 | class ParseResult(namedtuple('ParseResult', 'scheme netloc path params query fragment'), ResultMixin): |
| 125 | |
| 126 | __slots__ = () |
| 127 | |
| 128 | def geturl(self): |
| 129 | return urlunparse(self) |
| 130 | |
| 131 | |
| 132 | def urlparse(url, scheme='', allow_fragments=True): |
| 133 | """Parse a URL into 6 components: |
| 134 | <scheme>://<netloc>/<path>;<params>?<query>#<fragment> |
| 135 | Return a 6-tuple: (scheme, netloc, path, params, query, fragment). |
| 136 | Note that we don't break the components up in smaller bits |
| 137 | (e.g. netloc is a single string) and we don't expand % escapes.""" |
| 138 | tuple = urlsplit(url, scheme, allow_fragments) |
| 139 | scheme, netloc, url, query, fragment = tuple |
| 140 | if scheme in uses_params and ';' in url: |
| 141 | url, params = _splitparams(url) |
| 142 | else: |
| 143 | params = '' |
| 144 | return ParseResult(scheme, netloc, url, params, query, fragment) |
| 145 | |
| 146 | def _splitparams(url): |
| 147 | if '/' in url: |
| 148 | i = url.find(';', url.rfind('/')) |
| 149 | if i < 0: |
| 150 | return url, '' |
| 151 | else: |
| 152 | i = url.find(';') |
| 153 | return url[:i], url[i+1:] |
| 154 | |
| 155 | def _splitnetloc(url, start=0): |
| 156 | delim = len(url) # position of end of domain part of url, default is end |
| 157 | for c in '/?#': # look for delimiters; the order is NOT important |
| 158 | wdelim = url.find(c, start) # find first of this delim |
| 159 | if wdelim >= 0: # if found |
| 160 | delim = min(delim, wdelim) # use earliest delim position |
| 161 | return url[start:delim], url[delim:] # return (domain, rest) |
| 162 | |
| 163 | def urlsplit(url, scheme='', allow_fragments=True): |
| 164 | """Parse a URL into 5 components: |
| 165 | <scheme>://<netloc>/<path>?<query>#<fragment> |
| 166 | Return a 5-tuple: (scheme, netloc, path, query, fragment). |
| 167 | Note that we don't break the components up in smaller bits |
| 168 | (e.g. netloc is a single string) and we don't expand % escapes.""" |
| 169 | allow_fragments = bool(allow_fragments) |
| 170 | key = url, scheme, allow_fragments, type(url), type(scheme) |
| 171 | cached = _parse_cache.get(key, None) |
| 172 | if cached: |
| 173 | return cached |
| 174 | if len(_parse_cache) >= MAX_CACHE_SIZE: # avoid runaway growth |
| 175 | clear_cache() |
| 176 | netloc = query = fragment = '' |
| 177 | i = url.find(':') |
| 178 | if i > 0: |
| 179 | if url[:i] == 'http': # optimize the common case |
| 180 | scheme = url[:i].lower() |
| 181 | url = url[i+1:] |
| 182 | if url[:2] == '//': |
| 183 | netloc, url = _splitnetloc(url, 2) |
Senthil Kumaran | 7a1e09f | 2010-04-22 12:19:46 +0000 | [diff] [blame^] | 184 | if (('[' in netloc and ']' not in netloc) or |
| 185 | (']' in netloc and '[' not in netloc)): |
| 186 | raise ValueError("Invalid IPv6 URL") |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 187 | if allow_fragments and '#' in url: |
| 188 | url, fragment = url.split('#', 1) |
| 189 | if '?' in url: |
| 190 | url, query = url.split('?', 1) |
| 191 | v = SplitResult(scheme, netloc, url, query, fragment) |
| 192 | _parse_cache[key] = v |
| 193 | return v |
| 194 | for c in url[:i]: |
| 195 | if c not in scheme_chars: |
| 196 | break |
| 197 | else: |
| 198 | scheme, url = url[:i].lower(), url[i+1:] |
Senthil Kumaran | 6be85c5 | 2010-02-19 07:42:50 +0000 | [diff] [blame] | 199 | if url[:2] == '//': |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 200 | netloc, url = _splitnetloc(url, 2) |
Senthil Kumaran | 7a1e09f | 2010-04-22 12:19:46 +0000 | [diff] [blame^] | 201 | if (('[' in netloc and ']' not in netloc) or |
| 202 | (']' in netloc and '[' not in netloc)): |
| 203 | raise ValueError("Invalid IPv6 URL") |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 204 | if allow_fragments and scheme in uses_fragment and '#' in url: |
| 205 | url, fragment = url.split('#', 1) |
| 206 | if scheme in uses_query and '?' in url: |
| 207 | url, query = url.split('?', 1) |
| 208 | v = SplitResult(scheme, netloc, url, query, fragment) |
| 209 | _parse_cache[key] = v |
| 210 | return v |
| 211 | |
| 212 | def urlunparse(components): |
| 213 | """Put a parsed URL back together again. This may result in a |
| 214 | slightly different, but equivalent URL, if the URL that was parsed |
| 215 | originally had redundant delimiters, e.g. a ? with an empty query |
| 216 | (the draft states that these are equivalent).""" |
| 217 | scheme, netloc, url, params, query, fragment = components |
| 218 | if params: |
| 219 | url = "%s;%s" % (url, params) |
| 220 | return urlunsplit((scheme, netloc, url, query, fragment)) |
| 221 | |
| 222 | def urlunsplit(components): |
| 223 | scheme, netloc, url, query, fragment = components |
| 224 | if netloc or (scheme and scheme in uses_netloc and url[:2] != '//'): |
| 225 | if url and url[:1] != '/': url = '/' + url |
| 226 | url = '//' + (netloc or '') + url |
| 227 | if scheme: |
| 228 | url = scheme + ':' + url |
| 229 | if query: |
| 230 | url = url + '?' + query |
| 231 | if fragment: |
| 232 | url = url + '#' + fragment |
| 233 | return url |
| 234 | |
| 235 | def urljoin(base, url, allow_fragments=True): |
| 236 | """Join a base URL and a possibly relative URL to form an absolute |
| 237 | interpretation of the latter.""" |
| 238 | if not base: |
| 239 | return url |
| 240 | if not url: |
| 241 | return base |
| 242 | bscheme, bnetloc, bpath, bparams, bquery, bfragment = \ |
| 243 | urlparse(base, '', allow_fragments) |
| 244 | scheme, netloc, path, params, query, fragment = \ |
| 245 | urlparse(url, bscheme, allow_fragments) |
| 246 | if scheme != bscheme or scheme not in uses_relative: |
| 247 | return url |
| 248 | if scheme in uses_netloc: |
| 249 | if netloc: |
| 250 | return urlunparse((scheme, netloc, path, |
| 251 | params, query, fragment)) |
| 252 | netloc = bnetloc |
| 253 | if path[:1] == '/': |
| 254 | return urlunparse((scheme, netloc, path, |
| 255 | params, query, fragment)) |
Facundo Batista | 23e3856 | 2008-08-14 16:55:14 +0000 | [diff] [blame] | 256 | if not path: |
| 257 | path = bpath |
| 258 | if not params: |
| 259 | params = bparams |
| 260 | else: |
| 261 | path = path[:-1] |
| 262 | return urlunparse((scheme, netloc, path, |
| 263 | params, query, fragment)) |
| 264 | if not query: |
| 265 | query = bquery |
| 266 | return urlunparse((scheme, netloc, path, |
| 267 | params, query, fragment)) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 268 | segments = bpath.split('/')[:-1] + path.split('/') |
| 269 | # XXX The stuff below is bogus in various ways... |
| 270 | if segments[-1] == '.': |
| 271 | segments[-1] = '' |
| 272 | while '.' in segments: |
| 273 | segments.remove('.') |
| 274 | while 1: |
| 275 | i = 1 |
| 276 | n = len(segments) - 1 |
| 277 | while i < n: |
| 278 | if (segments[i] == '..' |
| 279 | and segments[i-1] not in ('', '..')): |
| 280 | del segments[i-1:i+1] |
| 281 | break |
| 282 | i = i+1 |
| 283 | else: |
| 284 | break |
| 285 | if segments == ['', '..']: |
| 286 | segments[-1] = '' |
| 287 | elif len(segments) >= 2 and segments[-1] == '..': |
| 288 | segments[-2:] = [''] |
| 289 | return urlunparse((scheme, netloc, '/'.join(segments), |
| 290 | params, query, fragment)) |
| 291 | |
| 292 | def urldefrag(url): |
| 293 | """Removes any existing fragment from URL. |
| 294 | |
| 295 | Returns a tuple of the defragmented URL and the fragment. If |
| 296 | the URL contained no fragments, the second element is the |
| 297 | empty string. |
| 298 | """ |
| 299 | if '#' in url: |
| 300 | s, n, p, a, q, frag = urlparse(url) |
| 301 | defrag = urlunparse((s, n, p, a, q, '')) |
| 302 | return defrag, frag |
| 303 | else: |
| 304 | return url, '' |
| 305 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 306 | def unquote_to_bytes(string): |
| 307 | """unquote_to_bytes('abc%20def') -> b'abc def'.""" |
| 308 | # Note: strings are encoded as UTF-8. This is only an issue if it contains |
| 309 | # unescaped non-ASCII characters, which URIs should not. |
| 310 | if isinstance(string, str): |
| 311 | string = string.encode('utf-8') |
| 312 | res = string.split(b'%') |
| 313 | res[0] = res[0] |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 314 | for i in range(1, len(res)): |
| 315 | item = res[i] |
| 316 | try: |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 317 | res[i] = bytes([int(item[:2], 16)]) + item[2:] |
| 318 | except ValueError: |
| 319 | res[i] = b'%' + item |
| 320 | return b''.join(res) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 321 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 322 | def unquote(string, encoding='utf-8', errors='replace'): |
| 323 | """Replace %xx escapes by their single-character equivalent. The optional |
| 324 | encoding and errors parameters specify how to decode percent-encoded |
| 325 | sequences into Unicode characters, as accepted by the bytes.decode() |
| 326 | method. |
| 327 | By default, percent-encoded sequences are decoded with UTF-8, and invalid |
| 328 | sequences are replaced by a placeholder character. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 329 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 330 | unquote('abc%20def') -> 'abc def'. |
| 331 | """ |
| 332 | if encoding is None: encoding = 'utf-8' |
| 333 | if errors is None: errors = 'replace' |
| 334 | # pct_sequence: contiguous sequence of percent-encoded bytes, decoded |
| 335 | # (list of single-byte bytes objects) |
| 336 | pct_sequence = [] |
| 337 | res = string.split('%') |
| 338 | for i in range(1, len(res)): |
| 339 | item = res[i] |
| 340 | try: |
| 341 | if not item: raise ValueError |
| 342 | pct_sequence.append(bytes.fromhex(item[:2])) |
| 343 | rest = item[2:] |
| 344 | except ValueError: |
| 345 | rest = '%' + item |
| 346 | if not rest: |
| 347 | # This segment was just a single percent-encoded character. |
| 348 | # May be part of a sequence of code units, so delay decoding. |
| 349 | # (Stored in pct_sequence). |
| 350 | res[i] = '' |
| 351 | else: |
| 352 | # Encountered non-percent-encoded characters. Flush the current |
| 353 | # pct_sequence. |
| 354 | res[i] = b''.join(pct_sequence).decode(encoding, errors) + rest |
| 355 | pct_sequence = [] |
| 356 | if pct_sequence: |
| 357 | # Flush the final pct_sequence |
| 358 | # res[-1] will always be empty if pct_sequence != [] |
| 359 | assert not res[-1], "string=%r, res=%r" % (string, res) |
| 360 | res[-1] = b''.join(pct_sequence).decode(encoding, errors) |
| 361 | return ''.join(res) |
| 362 | |
Georg Brandl | 3d6575d | 2009-09-16 14:36:22 +0000 | [diff] [blame] | 363 | def parse_qs(qs, keep_blank_values=False, strict_parsing=False): |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 364 | """Parse a query given as a string argument. |
| 365 | |
| 366 | Arguments: |
| 367 | |
| 368 | qs: URL-encoded query string to be parsed |
| 369 | |
| 370 | keep_blank_values: flag indicating whether blank values in |
| 371 | URL encoded queries should be treated as blank strings. |
| 372 | A true value indicates that blanks should be retained as |
| 373 | blank strings. The default false value indicates that |
| 374 | blank values are to be ignored and treated as if they were |
| 375 | not included. |
| 376 | |
| 377 | strict_parsing: flag indicating what to do with parsing errors. |
| 378 | If false (the default), errors are silently ignored. |
| 379 | If true, errors raise a ValueError exception. |
| 380 | """ |
| 381 | dict = {} |
| 382 | for name, value in parse_qsl(qs, keep_blank_values, strict_parsing): |
| 383 | if name in dict: |
| 384 | dict[name].append(value) |
| 385 | else: |
| 386 | dict[name] = [value] |
| 387 | return dict |
| 388 | |
Georg Brandl | 3d6575d | 2009-09-16 14:36:22 +0000 | [diff] [blame] | 389 | def parse_qsl(qs, keep_blank_values=False, strict_parsing=False): |
Facundo Batista | c469d4c | 2008-09-03 22:49:01 +0000 | [diff] [blame] | 390 | """Parse a query given as a string argument. |
| 391 | |
| 392 | Arguments: |
| 393 | |
| 394 | qs: URL-encoded query string to be parsed |
| 395 | |
| 396 | keep_blank_values: flag indicating whether blank values in |
| 397 | URL encoded queries should be treated as blank strings. A |
| 398 | true value indicates that blanks should be retained as blank |
| 399 | strings. The default false value indicates that blank values |
| 400 | are to be ignored and treated as if they were not included. |
| 401 | |
| 402 | strict_parsing: flag indicating what to do with parsing errors. If |
| 403 | false (the default), errors are silently ignored. If true, |
| 404 | errors raise a ValueError exception. |
| 405 | |
| 406 | Returns a list, as G-d intended. |
| 407 | """ |
| 408 | pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')] |
| 409 | r = [] |
| 410 | for name_value in pairs: |
| 411 | if not name_value and not strict_parsing: |
| 412 | continue |
| 413 | nv = name_value.split('=', 1) |
| 414 | if len(nv) != 2: |
| 415 | if strict_parsing: |
| 416 | raise ValueError("bad query field: %r" % (name_value,)) |
| 417 | # Handle case of a control-name with no equal sign |
| 418 | if keep_blank_values: |
| 419 | nv.append('') |
| 420 | else: |
| 421 | continue |
| 422 | if len(nv[1]) or keep_blank_values: |
| 423 | name = unquote(nv[0].replace('+', ' ')) |
| 424 | value = unquote(nv[1].replace('+', ' ')) |
| 425 | r.append((name, value)) |
| 426 | |
| 427 | return r |
| 428 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 429 | def unquote_plus(string, encoding='utf-8', errors='replace'): |
| 430 | """Like unquote(), but also replace plus signs by spaces, as required for |
| 431 | unquoting HTML form values. |
| 432 | |
| 433 | unquote_plus('%7e/abc+def') -> '~/abc def' |
| 434 | """ |
| 435 | string = string.replace('+', ' ') |
| 436 | return unquote(string, encoding, errors) |
| 437 | |
| 438 | _ALWAYS_SAFE = frozenset(b'ABCDEFGHIJKLMNOPQRSTUVWXYZ' |
| 439 | b'abcdefghijklmnopqrstuvwxyz' |
| 440 | b'0123456789' |
| 441 | b'_.-') |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 442 | _safe_quoters= {} |
| 443 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 444 | class Quoter(collections.defaultdict): |
| 445 | """A mapping from bytes (in range(0,256)) to strings. |
| 446 | |
| 447 | String values are percent-encoded byte values, unless the key < 128, and |
| 448 | in the "safe" set (either the specified safe set, or default set). |
| 449 | """ |
| 450 | # Keeps a cache internally, using defaultdict, for efficiency (lookups |
| 451 | # of cached keys don't call Python code at all). |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 452 | def __init__(self, safe): |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 453 | """safe: bytes object.""" |
| 454 | self.safe = _ALWAYS_SAFE.union(c for c in safe if c < 128) |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 455 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 456 | def __repr__(self): |
| 457 | # Without this, will just display as a defaultdict |
| 458 | return "<Quoter %r>" % dict(self) |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 459 | |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 460 | def __missing__(self, b): |
| 461 | # Handle a cache miss. Store quoted string in cache and return. |
| 462 | res = b in self.safe and chr(b) or ('%%%02X' % b) |
| 463 | self[b] = res |
| 464 | return res |
| 465 | |
| 466 | def quote(string, safe='/', encoding=None, errors=None): |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 467 | """quote('abc def') -> 'abc%20def' |
| 468 | |
| 469 | Each part of a URL, e.g. the path info, the query, etc., has a |
| 470 | different set of reserved characters that must be quoted. |
| 471 | |
| 472 | RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists |
| 473 | the following reserved characters. |
| 474 | |
| 475 | reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | |
| 476 | "$" | "," |
| 477 | |
| 478 | Each of these characters is reserved in some component of a URL, |
| 479 | but not necessarily in all of them. |
| 480 | |
| 481 | By default, the quote function is intended for quoting the path |
| 482 | section of a URL. Thus, it will not encode '/'. This character |
| 483 | is reserved, but in typical usage the quote function is being |
| 484 | called on a path where the existing slash characters are used as |
| 485 | reserved characters. |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 486 | |
| 487 | string and safe may be either str or bytes objects. encoding must |
| 488 | not be specified if string is a str. |
| 489 | |
| 490 | The optional encoding and errors parameters specify how to deal with |
| 491 | non-ASCII characters, as accepted by the str.encode method. |
| 492 | By default, encoding='utf-8' (characters are encoded with UTF-8), and |
| 493 | errors='strict' (unsupported characters raise a UnicodeEncodeError). |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 494 | """ |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 495 | if isinstance(string, str): |
| 496 | if encoding is None: |
| 497 | encoding = 'utf-8' |
| 498 | if errors is None: |
| 499 | errors = 'strict' |
| 500 | string = string.encode(encoding, errors) |
| 501 | else: |
| 502 | if encoding is not None: |
| 503 | raise TypeError("quote() doesn't support 'encoding' for bytes") |
| 504 | if errors is not None: |
| 505 | raise TypeError("quote() doesn't support 'errors' for bytes") |
| 506 | return quote_from_bytes(string, safe) |
| 507 | |
| 508 | def quote_plus(string, safe='', encoding=None, errors=None): |
| 509 | """Like quote(), but also replace ' ' with '+', as required for quoting |
| 510 | HTML form values. Plus signs in the original string are escaped unless |
| 511 | they are included in safe. It also does not have safe default to '/'. |
| 512 | """ |
Jeremy Hylton | f819886 | 2009-03-26 16:55:08 +0000 | [diff] [blame] | 513 | # Check if ' ' in string, where string may either be a str or bytes. If |
| 514 | # there are no spaces, the regular quote will produce the right answer. |
| 515 | if ((isinstance(string, str) and ' ' not in string) or |
| 516 | (isinstance(string, bytes) and b' ' not in string)): |
| 517 | return quote(string, safe, encoding, errors) |
| 518 | if isinstance(safe, str): |
| 519 | space = ' ' |
| 520 | else: |
| 521 | space = b' ' |
Georg Brandl | faf4149 | 2009-05-26 18:31:11 +0000 | [diff] [blame] | 522 | string = quote(string, safe + space, encoding, errors) |
Jeremy Hylton | f819886 | 2009-03-26 16:55:08 +0000 | [diff] [blame] | 523 | return string.replace(' ', '+') |
Guido van Rossum | 52dbbb9 | 2008-08-18 21:44:30 +0000 | [diff] [blame] | 524 | |
| 525 | def quote_from_bytes(bs, safe='/'): |
| 526 | """Like quote(), but accepts a bytes object rather than a str, and does |
| 527 | not perform string-to-bytes encoding. It always returns an ASCII string. |
| 528 | quote_from_bytes(b'abc def\xab') -> 'abc%20def%AB' |
| 529 | """ |
| 530 | if isinstance(safe, str): |
| 531 | # Normalize 'safe' by converting to bytes and removing non-ASCII chars |
| 532 | safe = safe.encode('ascii', 'ignore') |
| 533 | cachekey = bytes(safe) # In case it was a bytearray |
| 534 | if not (isinstance(bs, bytes) or isinstance(bs, bytearray)): |
| 535 | raise TypeError("quote_from_bytes() expected a bytes") |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 536 | try: |
| 537 | quoter = _safe_quoters[cachekey] |
| 538 | except KeyError: |
| 539 | quoter = Quoter(safe) |
| 540 | _safe_quoters[cachekey] = quoter |
Jeremy Hylton | f819886 | 2009-03-26 16:55:08 +0000 | [diff] [blame] | 541 | return ''.join([quoter[char] for char in bs]) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 542 | |
Georg Brandl | 3d6575d | 2009-09-16 14:36:22 +0000 | [diff] [blame] | 543 | def urlencode(query, doseq=False): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 544 | """Encode a sequence of two-element tuples or dictionary into a URL query string. |
| 545 | |
| 546 | If any values in the query arg are sequences and doseq is true, each |
| 547 | sequence element is converted to a separate parameter. |
| 548 | |
| 549 | If the query arg is a sequence of two-element tuples, the order of the |
| 550 | parameters in the output will match the order of parameters in the |
| 551 | input. |
| 552 | """ |
| 553 | |
Jeremy Hylton | a4de60a | 2009-03-26 14:49:26 +0000 | [diff] [blame] | 554 | if hasattr(query, "items"): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 555 | query = query.items() |
| 556 | else: |
Jeremy Hylton | 230feba | 2009-03-26 16:56:59 +0000 | [diff] [blame] | 557 | # It's a bother at times that strings and string-like objects are |
| 558 | # sequences. |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 559 | try: |
| 560 | # non-sequence items should not work with len() |
| 561 | # non-empty strings will fail this |
| 562 | if len(query) and not isinstance(query[0], tuple): |
| 563 | raise TypeError |
Jeremy Hylton | 230feba | 2009-03-26 16:56:59 +0000 | [diff] [blame] | 564 | # Zero-length sequences of all types will get here and succeed, |
| 565 | # but that's a minor nit. Since the original implementation |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 566 | # allowed empty dicts that type of behavior probably should be |
| 567 | # preserved for consistency |
| 568 | except TypeError: |
Jeremy Hylton | a4de60a | 2009-03-26 14:49:26 +0000 | [diff] [blame] | 569 | ty, va, tb = sys.exc_info() |
| 570 | raise TypeError("not a valid non-string sequence " |
| 571 | "or mapping object").with_traceback(tb) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 572 | |
| 573 | l = [] |
| 574 | if not doseq: |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 575 | for k, v in query: |
| 576 | k = quote_plus(str(k)) |
| 577 | v = quote_plus(str(v)) |
| 578 | l.append(k + '=' + v) |
| 579 | else: |
| 580 | for k, v in query: |
| 581 | k = quote_plus(str(k)) |
| 582 | if isinstance(v, str): |
| 583 | v = quote_plus(v) |
| 584 | l.append(k + '=' + v) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 585 | else: |
| 586 | try: |
Jeremy Hylton | 230feba | 2009-03-26 16:56:59 +0000 | [diff] [blame] | 587 | # Is this a sufficient test for sequence-ness? |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 588 | x = len(v) |
| 589 | except TypeError: |
| 590 | # not a sequence |
| 591 | v = quote_plus(str(v)) |
| 592 | l.append(k + '=' + v) |
| 593 | else: |
| 594 | # loop over the sequence |
| 595 | for elt in v: |
| 596 | l.append(k + '=' + quote_plus(str(elt))) |
| 597 | return '&'.join(l) |
| 598 | |
| 599 | # Utilities to parse URLs (most of these return None for missing parts): |
| 600 | # unwrap('<URL:type://host/path>') --> 'type://host/path' |
| 601 | # splittype('type:opaquestring') --> 'type', 'opaquestring' |
| 602 | # splithost('//host[:port]/path') --> 'host[:port]', '/path' |
| 603 | # splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]' |
| 604 | # splitpasswd('user:passwd') -> 'user', 'passwd' |
| 605 | # splitport('host:port') --> 'host', 'port' |
| 606 | # splitquery('/path?query') --> '/path', 'query' |
| 607 | # splittag('/path#tag') --> '/path', 'tag' |
| 608 | # splitattr('/path;attr1=value1;attr2=value2;...') -> |
| 609 | # '/path', ['attr1=value1', 'attr2=value2', ...] |
| 610 | # splitvalue('attr=value') --> 'attr', 'value' |
| 611 | # urllib.parse.unquote('abc%20def') -> 'abc def' |
| 612 | # quote('abc def') -> 'abc%20def') |
| 613 | |
Georg Brandl | 13e8946 | 2008-07-01 19:56:00 +0000 | [diff] [blame] | 614 | def to_bytes(url): |
| 615 | """to_bytes(u"URL") --> 'URL'.""" |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 616 | # Most URL schemes require ASCII. If that changes, the conversion |
| 617 | # can be relaxed. |
Georg Brandl | 13e8946 | 2008-07-01 19:56:00 +0000 | [diff] [blame] | 618 | # XXX get rid of to_bytes() |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 619 | if isinstance(url, str): |
| 620 | try: |
| 621 | url = url.encode("ASCII").decode() |
| 622 | except UnicodeError: |
| 623 | raise UnicodeError("URL " + repr(url) + |
| 624 | " contains non-ASCII characters") |
| 625 | return url |
| 626 | |
| 627 | def unwrap(url): |
| 628 | """unwrap('<URL:type://host/path>') --> 'type://host/path'.""" |
| 629 | url = str(url).strip() |
| 630 | if url[:1] == '<' and url[-1:] == '>': |
| 631 | url = url[1:-1].strip() |
| 632 | if url[:4] == 'URL:': url = url[4:].strip() |
| 633 | return url |
| 634 | |
| 635 | _typeprog = None |
| 636 | def splittype(url): |
| 637 | """splittype('type:opaquestring') --> 'type', 'opaquestring'.""" |
| 638 | global _typeprog |
| 639 | if _typeprog is None: |
| 640 | import re |
| 641 | _typeprog = re.compile('^([^/:]+):') |
| 642 | |
| 643 | match = _typeprog.match(url) |
| 644 | if match: |
| 645 | scheme = match.group(1) |
| 646 | return scheme.lower(), url[len(scheme) + 1:] |
| 647 | return None, url |
| 648 | |
| 649 | _hostprog = None |
| 650 | def splithost(url): |
| 651 | """splithost('//host[:port]/path') --> 'host[:port]', '/path'.""" |
| 652 | global _hostprog |
| 653 | if _hostprog is None: |
| 654 | import re |
| 655 | _hostprog = re.compile('^//([^/?]*)(.*)$') |
| 656 | |
| 657 | match = _hostprog.match(url) |
| 658 | if match: return match.group(1, 2) |
| 659 | return None, url |
| 660 | |
| 661 | _userprog = None |
| 662 | def splituser(host): |
| 663 | """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'.""" |
| 664 | global _userprog |
| 665 | if _userprog is None: |
| 666 | import re |
| 667 | _userprog = re.compile('^(.*)@(.*)$') |
| 668 | |
| 669 | match = _userprog.match(host) |
Guido van Rossum | df9f1ec | 2008-08-06 19:31:34 +0000 | [diff] [blame] | 670 | if match: return map(unquote, match.group(1, 2)) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 671 | return None, host |
| 672 | |
| 673 | _passwdprog = None |
| 674 | def splitpasswd(user): |
| 675 | """splitpasswd('user:passwd') -> 'user', 'passwd'.""" |
| 676 | global _passwdprog |
| 677 | if _passwdprog is None: |
| 678 | import re |
Senthil Kumaran | eaaec27 | 2009-03-30 21:54:41 +0000 | [diff] [blame] | 679 | _passwdprog = re.compile('^([^:]*):(.*)$',re.S) |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 680 | |
| 681 | match = _passwdprog.match(user) |
| 682 | if match: return match.group(1, 2) |
| 683 | return user, None |
| 684 | |
| 685 | # splittag('/path#tag') --> '/path', 'tag' |
| 686 | _portprog = None |
| 687 | def splitport(host): |
| 688 | """splitport('host:port') --> 'host', 'port'.""" |
| 689 | global _portprog |
| 690 | if _portprog is None: |
| 691 | import re |
| 692 | _portprog = re.compile('^(.*):([0-9]+)$') |
| 693 | |
| 694 | match = _portprog.match(host) |
| 695 | if match: return match.group(1, 2) |
| 696 | return host, None |
| 697 | |
| 698 | _nportprog = None |
| 699 | def splitnport(host, defport=-1): |
| 700 | """Split host and port, returning numeric port. |
| 701 | Return given default port if no ':' found; defaults to -1. |
| 702 | Return numerical port if a valid number are found after ':'. |
| 703 | Return None if ':' but not a valid number.""" |
| 704 | global _nportprog |
| 705 | if _nportprog is None: |
| 706 | import re |
| 707 | _nportprog = re.compile('^(.*):(.*)$') |
| 708 | |
| 709 | match = _nportprog.match(host) |
| 710 | if match: |
| 711 | host, port = match.group(1, 2) |
| 712 | try: |
| 713 | if not port: raise ValueError("no digits") |
| 714 | nport = int(port) |
| 715 | except ValueError: |
| 716 | nport = None |
| 717 | return host, nport |
| 718 | return host, defport |
| 719 | |
| 720 | _queryprog = None |
| 721 | def splitquery(url): |
| 722 | """splitquery('/path?query') --> '/path', 'query'.""" |
| 723 | global _queryprog |
| 724 | if _queryprog is None: |
| 725 | import re |
| 726 | _queryprog = re.compile('^(.*)\?([^?]*)$') |
| 727 | |
| 728 | match = _queryprog.match(url) |
| 729 | if match: return match.group(1, 2) |
| 730 | return url, None |
| 731 | |
| 732 | _tagprog = None |
| 733 | def splittag(url): |
| 734 | """splittag('/path#tag') --> '/path', 'tag'.""" |
| 735 | global _tagprog |
| 736 | if _tagprog is None: |
| 737 | import re |
| 738 | _tagprog = re.compile('^(.*)#([^#]*)$') |
| 739 | |
| 740 | match = _tagprog.match(url) |
| 741 | if match: return match.group(1, 2) |
| 742 | return url, None |
| 743 | |
| 744 | def splitattr(url): |
| 745 | """splitattr('/path;attr1=value1;attr2=value2;...') -> |
| 746 | '/path', ['attr1=value1', 'attr2=value2', ...].""" |
| 747 | words = url.split(';') |
| 748 | return words[0], words[1:] |
| 749 | |
| 750 | _valueprog = None |
| 751 | def splitvalue(attr): |
| 752 | """splitvalue('attr=value') --> 'attr', 'value'.""" |
| 753 | global _valueprog |
| 754 | if _valueprog is None: |
| 755 | import re |
| 756 | _valueprog = re.compile('^([^=]*)=(.*)$') |
| 757 | |
| 758 | match = _valueprog.match(attr) |
| 759 | if match: return match.group(1, 2) |
| 760 | return attr, None |
| 761 | |
| 762 | test_input = """ |
| 763 | http://a/b/c/d |
| 764 | |
| 765 | g:h = <URL:g:h> |
| 766 | http:g = <URL:http://a/b/c/g> |
| 767 | http: = <URL:http://a/b/c/d> |
| 768 | g = <URL:http://a/b/c/g> |
| 769 | ./g = <URL:http://a/b/c/g> |
| 770 | g/ = <URL:http://a/b/c/g/> |
| 771 | /g = <URL:http://a/g> |
| 772 | //g = <URL:http://g> |
| 773 | ?y = <URL:http://a/b/c/d?y> |
| 774 | g?y = <URL:http://a/b/c/g?y> |
| 775 | g?y/./x = <URL:http://a/b/c/g?y/./x> |
| 776 | . = <URL:http://a/b/c/> |
| 777 | ./ = <URL:http://a/b/c/> |
| 778 | .. = <URL:http://a/b/> |
| 779 | ../ = <URL:http://a/b/> |
| 780 | ../g = <URL:http://a/b/g> |
| 781 | ../.. = <URL:http://a/> |
| 782 | ../../g = <URL:http://a/g> |
| 783 | ../../../g = <URL:http://a/../g> |
| 784 | ./../g = <URL:http://a/b/g> |
| 785 | ./g/. = <URL:http://a/b/c/g/> |
| 786 | /./g = <URL:http://a/./g> |
| 787 | g/./h = <URL:http://a/b/c/g/h> |
| 788 | g/../h = <URL:http://a/b/c/h> |
| 789 | http:g = <URL:http://a/b/c/g> |
| 790 | http: = <URL:http://a/b/c/d> |
| 791 | http:?y = <URL:http://a/b/c/d?y> |
| 792 | http:g?y = <URL:http://a/b/c/g?y> |
| 793 | http:g?y/./x = <URL:http://a/b/c/g?y/./x> |
| 794 | """ |
| 795 | |
| 796 | def test(): |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 797 | base = '' |
| 798 | if sys.argv[1:]: |
| 799 | fn = sys.argv[1] |
| 800 | if fn == '-': |
| 801 | fp = sys.stdin |
| 802 | else: |
| 803 | fp = open(fn) |
| 804 | else: |
| 805 | from io import StringIO |
| 806 | fp = StringIO(test_input) |
| 807 | for line in fp: |
| 808 | words = line.split() |
| 809 | if not words: |
| 810 | continue |
| 811 | url = words[0] |
| 812 | parts = urlparse(url) |
| 813 | print('%-10s : %s' % (url, parts)) |
| 814 | abs = urljoin(base, url) |
| 815 | if not base: |
| 816 | base = abs |
| 817 | wrapped = '<URL:%s>' % abs |
| 818 | print('%-10s = %s' % (url, wrapped)) |
| 819 | if len(words) == 3 and words[1] == '=': |
| 820 | if wrapped != words[2]: |
| 821 | print('EXPECTED', words[2], '!!!!!!!!!!') |
| 822 | |
| 823 | if __name__ == '__main__': |
| 824 | test() |