|
1 """Open an arbitrary URL. |
|
2 |
|
3 See the following document for more info on URLs: |
|
4 "Names and Addresses, URIs, URLs, URNs, URCs", at |
|
5 http://www.w3.org/pub/WWW/Addressing/Overview.html |
|
6 |
|
7 See also the HTTP spec (from which the error codes are derived): |
|
8 "HTTP - Hypertext Transfer Protocol", at |
|
9 http://www.w3.org/pub/WWW/Protocols/ |
|
10 |
|
11 Related standards and specs: |
|
12 - RFC1808: the "relative URL" spec. (authoritative status) |
|
13 - RFC1738 - the "URL standard". (authoritative status) |
|
14 - RFC1630 - the "URI spec". (informational status) |
|
15 |
|
16 The object returned by URLopener().open(file) will differ per |
|
17 protocol. All you know is that is has methods read(), readline(), |
|
18 readlines(), fileno(), close() and info(). The read*(), fileno() |
|
19 and close() methods work like those of open files. |
|
20 The info() method returns a mimetools.Message object which can be |
|
21 used to query various info about the object, if available. |
|
22 (mimetools.Message objects are queried with the getheader() method.) |
|
23 """ |
|
24 |
|
25 import string |
|
26 import socket |
|
27 import os |
|
28 import time |
|
29 import sys |
|
30 from urlparse import urljoin as basejoin |
|
31 |
|
32 __all__ = ["urlopen", "URLopener", "FancyURLopener", "urlretrieve", |
|
33 "urlcleanup", "quote", "quote_plus", "unquote", "unquote_plus", |
|
34 "urlencode", "url2pathname", "pathname2url", "splittag", |
|
35 "localhost", "thishost", "ftperrors", "basejoin", "unwrap", |
|
36 "splittype", "splithost", "splituser", "splitpasswd", "splitport", |
|
37 "splitnport", "splitquery", "splitattr", "splitvalue", |
|
38 "splitgophertype", "getproxies"] |
|
39 |
|
40 __version__ = '1.17' # XXX This version is not always updated :-( |
|
41 |
|
42 MAXFTPCACHE = 10 # Trim the ftp cache beyond this size |
|
43 |
|
44 # Helper for non-unix systems |
|
45 if os.name == 'mac': |
|
46 from macurl2path import url2pathname, pathname2url |
|
47 elif os.name == 'nt': |
|
48 from nturl2path import url2pathname, pathname2url |
|
49 elif os.name == 'riscos': |
|
50 from rourl2path import url2pathname, pathname2url |
|
51 else: |
|
52 def url2pathname(pathname): |
|
53 """OS-specific conversion from a relative URL of the 'file' scheme |
|
54 to a file system path; not recommended for general use.""" |
|
55 return unquote(pathname) |
|
56 |
|
57 def pathname2url(pathname): |
|
58 """OS-specific conversion from a file system path to a relative URL |
|
59 of the 'file' scheme; not recommended for general use.""" |
|
60 return quote(pathname) |
|
61 |
|
62 # This really consists of two pieces: |
|
63 # (1) a class which handles opening of all sorts of URLs |
|
64 # (plus assorted utilities etc.) |
|
65 # (2) a set of functions for parsing URLs |
|
66 # XXX Should these be separated out into different modules? |
|
67 |
|
68 |
|
69 # Shortcut for basic usage |
|
70 _urlopener = None |
|
71 def urlopen(url, data=None, proxies=None): |
|
72 """urlopen(url [, data]) -> open file-like object""" |
|
73 global _urlopener |
|
74 if proxies is not None: |
|
75 opener = FancyURLopener(proxies=proxies) |
|
76 elif not _urlopener: |
|
77 opener = FancyURLopener() |
|
78 _urlopener = opener |
|
79 else: |
|
80 opener = _urlopener |
|
81 if data is None: |
|
82 return opener.open(url) |
|
83 else: |
|
84 return opener.open(url, data) |
|
85 def urlretrieve(url, filename=None, reporthook=None, data=None): |
|
86 global _urlopener |
|
87 if not _urlopener: |
|
88 _urlopener = FancyURLopener() |
|
89 return _urlopener.retrieve(url, filename, reporthook, data) |
|
90 def urlcleanup(): |
|
91 if _urlopener: |
|
92 _urlopener.cleanup() |
|
93 |
|
94 # exception raised when downloaded size does not match content-length |
|
95 class ContentTooShortError(IOError): |
|
96 def __init__(self, message, content): |
|
97 IOError.__init__(self, message) |
|
98 self.content = content |
|
99 |
|
100 ftpcache = {} |
|
101 class URLopener: |
|
102 """Class to open URLs. |
|
103 This is a class rather than just a subroutine because we may need |
|
104 more than one set of global protocol-specific options. |
|
105 Note -- this is a base class for those who don't want the |
|
106 automatic handling of errors type 302 (relocated) and 401 |
|
107 (authorization needed).""" |
|
108 |
|
109 __tempfiles = None |
|
110 |
|
111 version = "Python-urllib/%s" % __version__ |
|
112 |
|
113 # Constructor |
|
114 def __init__(self, proxies=None, **x509): |
|
115 if proxies is None: |
|
116 proxies = getproxies() |
|
117 assert hasattr(proxies, 'has_key'), "proxies must be a mapping" |
|
118 self.proxies = proxies |
|
119 self.key_file = x509.get('key_file') |
|
120 self.cert_file = x509.get('cert_file') |
|
121 self.addheaders = [('User-Agent', self.version)] |
|
122 self.__tempfiles = [] |
|
123 self.__unlink = os.unlink # See cleanup() |
|
124 self.tempcache = None |
|
125 # Undocumented feature: if you assign {} to tempcache, |
|
126 # it is used to cache files retrieved with |
|
127 # self.retrieve(). This is not enabled by default |
|
128 # since it does not work for changing documents (and I |
|
129 # haven't got the logic to check expiration headers |
|
130 # yet). |
|
131 self.ftpcache = ftpcache |
|
132 # Undocumented feature: you can use a different |
|
133 # ftp cache by assigning to the .ftpcache member; |
|
134 # in case you want logically independent URL openers |
|
135 # XXX This is not threadsafe. Bah. |
|
136 |
|
137 def __del__(self): |
|
138 self.close() |
|
139 |
|
140 def close(self): |
|
141 self.cleanup() |
|
142 |
|
143 def cleanup(self): |
|
144 # This code sometimes runs when the rest of this module |
|
145 # has already been deleted, so it can't use any globals |
|
146 # or import anything. |
|
147 if self.__tempfiles: |
|
148 for file in self.__tempfiles: |
|
149 try: |
|
150 self.__unlink(file) |
|
151 except OSError: |
|
152 pass |
|
153 del self.__tempfiles[:] |
|
154 if self.tempcache: |
|
155 self.tempcache.clear() |
|
156 |
|
157 def addheader(self, *args): |
|
158 """Add a header to be used by the HTTP interface only |
|
159 e.g. u.addheader('Accept', 'sound/basic')""" |
|
160 self.addheaders.append(args) |
|
161 |
|
162 # External interface |
|
163 def open(self, fullurl, data=None): |
|
164 """Use URLopener().open(file) instead of open(file, 'r').""" |
|
165 fullurl = unwrap(toBytes(fullurl)) |
|
166 if self.tempcache and fullurl in self.tempcache: |
|
167 filename, headers = self.tempcache[fullurl] |
|
168 fp = open(filename, 'rb') |
|
169 return addinfourl(fp, headers, fullurl) |
|
170 urltype, url = splittype(fullurl) |
|
171 if not urltype: |
|
172 urltype = 'file' |
|
173 if urltype in self.proxies: |
|
174 proxy = self.proxies[urltype] |
|
175 urltype, proxyhost = splittype(proxy) |
|
176 host, selector = splithost(proxyhost) |
|
177 url = (host, fullurl) # Signal special case to open_*() |
|
178 else: |
|
179 proxy = None |
|
180 name = 'open_' + urltype |
|
181 self.type = urltype |
|
182 name = name.replace('-', '_') |
|
183 if not hasattr(self, name): |
|
184 if proxy: |
|
185 return self.open_unknown_proxy(proxy, fullurl, data) |
|
186 else: |
|
187 return self.open_unknown(fullurl, data) |
|
188 try: |
|
189 if data is None: |
|
190 return getattr(self, name)(url) |
|
191 else: |
|
192 return getattr(self, name)(url, data) |
|
193 except socket.error, msg: |
|
194 raise IOError, ('socket error', msg), sys.exc_info()[2] |
|
195 |
|
196 def open_unknown(self, fullurl, data=None): |
|
197 """Overridable interface to open unknown URL type.""" |
|
198 type, url = splittype(fullurl) |
|
199 raise IOError, ('url error', 'unknown url type', type) |
|
200 |
|
201 def open_unknown_proxy(self, proxy, fullurl, data=None): |
|
202 """Overridable interface to open unknown URL type.""" |
|
203 type, url = splittype(fullurl) |
|
204 raise IOError, ('url error', 'invalid proxy for %s' % type, proxy) |
|
205 |
|
206 # External interface |
|
207 def retrieve(self, url, filename=None, reporthook=None, data=None): |
|
208 """retrieve(url) returns (filename, headers) for a local object |
|
209 or (tempfilename, headers) for a remote object.""" |
|
210 url = unwrap(toBytes(url)) |
|
211 if self.tempcache and url in self.tempcache: |
|
212 return self.tempcache[url] |
|
213 type, url1 = splittype(url) |
|
214 if filename is None and (not type or type == 'file'): |
|
215 try: |
|
216 fp = self.open_local_file(url1) |
|
217 hdrs = fp.info() |
|
218 del fp |
|
219 return url2pathname(splithost(url1)[1]), hdrs |
|
220 except IOError, msg: |
|
221 pass |
|
222 fp = self.open(url, data) |
|
223 headers = fp.info() |
|
224 if filename: |
|
225 tfp = open(filename, 'wb') |
|
226 else: |
|
227 import tempfile |
|
228 garbage, path = splittype(url) |
|
229 garbage, path = splithost(path or "") |
|
230 path, garbage = splitquery(path or "") |
|
231 path, garbage = splitattr(path or "") |
|
232 suffix = os.path.splitext(path)[1] |
|
233 (fd, filename) = tempfile.mkstemp(suffix) |
|
234 self.__tempfiles.append(filename) |
|
235 tfp = os.fdopen(fd, 'wb') |
|
236 result = filename, headers |
|
237 if self.tempcache is not None: |
|
238 self.tempcache[url] = result |
|
239 bs = 1024*8 |
|
240 size = -1 |
|
241 read = 0 |
|
242 blocknum = 0 |
|
243 if reporthook: |
|
244 if "content-length" in headers: |
|
245 size = int(headers["Content-Length"]) |
|
246 reporthook(blocknum, bs, size) |
|
247 while 1: |
|
248 block = fp.read(bs) |
|
249 if block == "": |
|
250 break |
|
251 read += len(block) |
|
252 tfp.write(block) |
|
253 blocknum += 1 |
|
254 if reporthook: |
|
255 reporthook(blocknum, bs, size) |
|
256 fp.close() |
|
257 tfp.close() |
|
258 del fp |
|
259 del tfp |
|
260 |
|
261 # raise exception if actual size does not match content-length header |
|
262 if size >= 0 and read < size: |
|
263 raise ContentTooShortError("retrieval incomplete: got only %i out " |
|
264 "of %i bytes" % (read, size), result) |
|
265 |
|
266 return result |
|
267 |
|
268 # Each method named open_<type> knows how to open that type of URL |
|
269 |
|
270 def open_http(self, url, data=None): |
|
271 """Use HTTP protocol.""" |
|
272 import httplib |
|
273 user_passwd = None |
|
274 proxy_passwd= None |
|
275 if isinstance(url, str): |
|
276 host, selector = splithost(url) |
|
277 if host: |
|
278 user_passwd, host = splituser(host) |
|
279 host = unquote(host) |
|
280 realhost = host |
|
281 else: |
|
282 host, selector = url |
|
283 # check whether the proxy contains authorization information |
|
284 proxy_passwd, host = splituser(host) |
|
285 # now we proceed with the url we want to obtain |
|
286 urltype, rest = splittype(selector) |
|
287 url = rest |
|
288 user_passwd = None |
|
289 if urltype.lower() != 'http': |
|
290 realhost = None |
|
291 else: |
|
292 realhost, rest = splithost(rest) |
|
293 if realhost: |
|
294 user_passwd, realhost = splituser(realhost) |
|
295 if user_passwd: |
|
296 selector = "%s://%s%s" % (urltype, realhost, rest) |
|
297 if proxy_bypass(realhost): |
|
298 host = realhost |
|
299 |
|
300 #print "proxy via http:", host, selector |
|
301 if not host: raise IOError, ('http error', 'no host given') |
|
302 |
|
303 if proxy_passwd: |
|
304 import base64 |
|
305 proxy_auth = base64.b64encode(proxy_passwd).strip() |
|
306 else: |
|
307 proxy_auth = None |
|
308 |
|
309 if user_passwd: |
|
310 import base64 |
|
311 auth = base64.b64encode(user_passwd).strip() |
|
312 else: |
|
313 auth = None |
|
314 h = httplib.HTTP(host) |
|
315 if data is not None: |
|
316 h.putrequest('POST', selector) |
|
317 h.putheader('Content-Type', 'application/x-www-form-urlencoded') |
|
318 h.putheader('Content-Length', '%d' % len(data)) |
|
319 else: |
|
320 h.putrequest('GET', selector) |
|
321 if proxy_auth: h.putheader('Proxy-Authorization', 'Basic %s' % proxy_auth) |
|
322 if auth: h.putheader('Authorization', 'Basic %s' % auth) |
|
323 if realhost: h.putheader('Host', realhost) |
|
324 for args in self.addheaders: h.putheader(*args) |
|
325 h.endheaders() |
|
326 if data is not None: |
|
327 h.send(data) |
|
328 errcode, errmsg, headers = h.getreply() |
|
329 if errcode == -1: |
|
330 # something went wrong with the HTTP status line |
|
331 raise IOError, ('http protocol error', 0, |
|
332 'got a bad status line', None) |
|
333 fp = h.getfile() |
|
334 if errcode == 200: |
|
335 return addinfourl(fp, headers, "http:" + url) |
|
336 else: |
|
337 if data is None: |
|
338 return self.http_error(url, fp, errcode, errmsg, headers) |
|
339 else: |
|
340 return self.http_error(url, fp, errcode, errmsg, headers, data) |
|
341 |
|
342 def http_error(self, url, fp, errcode, errmsg, headers, data=None): |
|
343 """Handle http errors. |
|
344 Derived class can override this, or provide specific handlers |
|
345 named http_error_DDD where DDD is the 3-digit error code.""" |
|
346 # First check if there's a specific handler for this error |
|
347 name = 'http_error_%d' % errcode |
|
348 if hasattr(self, name): |
|
349 method = getattr(self, name) |
|
350 if data is None: |
|
351 result = method(url, fp, errcode, errmsg, headers) |
|
352 else: |
|
353 result = method(url, fp, errcode, errmsg, headers, data) |
|
354 if result: return result |
|
355 return self.http_error_default(url, fp, errcode, errmsg, headers) |
|
356 |
|
357 def http_error_default(self, url, fp, errcode, errmsg, headers): |
|
358 """Default error handler: close the connection and raise IOError.""" |
|
359 void = fp.read() |
|
360 fp.close() |
|
361 raise IOError, ('http error', errcode, errmsg, headers) |
|
362 |
|
363 if hasattr(socket, "ssl"): |
|
364 def open_https(self, url, data=None): |
|
365 """Use HTTPS protocol.""" |
|
366 import httplib |
|
367 user_passwd = None |
|
368 proxy_passwd = None |
|
369 if isinstance(url, str): |
|
370 host, selector = splithost(url) |
|
371 if host: |
|
372 user_passwd, host = splituser(host) |
|
373 host = unquote(host) |
|
374 realhost = host |
|
375 else: |
|
376 host, selector = url |
|
377 # here, we determine, whether the proxy contains authorization information |
|
378 proxy_passwd, host = splituser(host) |
|
379 urltype, rest = splittype(selector) |
|
380 url = rest |
|
381 user_passwd = None |
|
382 if urltype.lower() != 'https': |
|
383 realhost = None |
|
384 else: |
|
385 realhost, rest = splithost(rest) |
|
386 if realhost: |
|
387 user_passwd, realhost = splituser(realhost) |
|
388 if user_passwd: |
|
389 selector = "%s://%s%s" % (urltype, realhost, rest) |
|
390 #print "proxy via https:", host, selector |
|
391 if not host: raise IOError, ('https error', 'no host given') |
|
392 if proxy_passwd: |
|
393 import base64 |
|
394 proxy_auth = base64.b64encode(proxy_passwd).strip() |
|
395 else: |
|
396 proxy_auth = None |
|
397 if user_passwd: |
|
398 import base64 |
|
399 auth = base64.b64encode(user_passwd).strip() |
|
400 else: |
|
401 auth = None |
|
402 h = httplib.HTTPS(host, 0, |
|
403 key_file=self.key_file, |
|
404 cert_file=self.cert_file) |
|
405 if data is not None: |
|
406 h.putrequest('POST', selector) |
|
407 h.putheader('Content-Type', |
|
408 'application/x-www-form-urlencoded') |
|
409 h.putheader('Content-Length', '%d' % len(data)) |
|
410 else: |
|
411 h.putrequest('GET', selector) |
|
412 if proxy_auth: h.putheader('Proxy-Authorization', 'Basic %s' % proxy_auth) |
|
413 if auth: h.putheader('Authorization', 'Basic %s' % auth) |
|
414 if realhost: h.putheader('Host', realhost) |
|
415 for args in self.addheaders: h.putheader(*args) |
|
416 h.endheaders() |
|
417 if data is not None: |
|
418 h.send(data) |
|
419 errcode, errmsg, headers = h.getreply() |
|
420 if errcode == -1: |
|
421 # something went wrong with the HTTP status line |
|
422 raise IOError, ('http protocol error', 0, |
|
423 'got a bad status line', None) |
|
424 fp = h.getfile() |
|
425 if errcode == 200: |
|
426 return addinfourl(fp, headers, "https:" + url) |
|
427 else: |
|
428 if data is None: |
|
429 return self.http_error(url, fp, errcode, errmsg, headers) |
|
430 else: |
|
431 return self.http_error(url, fp, errcode, errmsg, headers, |
|
432 data) |
|
433 |
|
434 def open_gopher(self, url): |
|
435 """Use Gopher protocol.""" |
|
436 if not isinstance(url, str): |
|
437 raise IOError, ('gopher error', 'proxy support for gopher protocol currently not implemented') |
|
438 import gopherlib |
|
439 host, selector = splithost(url) |
|
440 if not host: raise IOError, ('gopher error', 'no host given') |
|
441 host = unquote(host) |
|
442 type, selector = splitgophertype(selector) |
|
443 selector, query = splitquery(selector) |
|
444 selector = unquote(selector) |
|
445 if query: |
|
446 query = unquote(query) |
|
447 fp = gopherlib.send_query(selector, query, host) |
|
448 else: |
|
449 fp = gopherlib.send_selector(selector, host) |
|
450 return addinfourl(fp, noheaders(), "gopher:" + url) |
|
451 |
|
452 def open_file(self, url): |
|
453 """Use local file or FTP depending on form of URL.""" |
|
454 if not isinstance(url, str): |
|
455 raise IOError, ('file error', 'proxy support for file protocol currently not implemented') |
|
456 if url[:2] == '//' and url[2:3] != '/' and url[2:12].lower() != 'localhost/': |
|
457 return self.open_ftp(url) |
|
458 else: |
|
459 return self.open_local_file(url) |
|
460 |
|
461 def open_local_file(self, url): |
|
462 """Use local file.""" |
|
463 import mimetypes, mimetools, email.Utils |
|
464 try: |
|
465 from cStringIO import StringIO |
|
466 except ImportError: |
|
467 from StringIO import StringIO |
|
468 host, file = splithost(url) |
|
469 localname = url2pathname(file) |
|
470 try: |
|
471 stats = os.stat(localname) |
|
472 except OSError, e: |
|
473 raise IOError(e.errno, e.strerror, e.filename) |
|
474 size = stats.st_size |
|
475 modified = email.Utils.formatdate(stats.st_mtime, usegmt=True) |
|
476 mtype = mimetypes.guess_type(url)[0] |
|
477 headers = mimetools.Message(StringIO( |
|
478 'Content-Type: %s\nContent-Length: %d\nLast-modified: %s\n' % |
|
479 (mtype or 'text/plain', size, modified))) |
|
480 if not host: |
|
481 urlfile = file |
|
482 if file[:1] == '/': |
|
483 urlfile = 'file://' + file |
|
484 return addinfourl(open(localname, 'rb'), |
|
485 headers, urlfile) |
|
486 host, port = splitport(host) |
|
487 if not port \ |
|
488 and socket.gethostbyname(host) in (localhost(), thishost()): |
|
489 urlfile = file |
|
490 if file[:1] == '/': |
|
491 urlfile = 'file://' + file |
|
492 return addinfourl(open(localname, 'rb'), |
|
493 headers, urlfile) |
|
494 raise IOError, ('local file error', 'not on local host') |
|
495 |
|
496 def open_ftp(self, url): |
|
497 """Use FTP protocol.""" |
|
498 if not isinstance(url, str): |
|
499 raise IOError, ('ftp error', 'proxy support for ftp protocol currently not implemented') |
|
500 import mimetypes, mimetools |
|
501 try: |
|
502 from cStringIO import StringIO |
|
503 except ImportError: |
|
504 from StringIO import StringIO |
|
505 host, path = splithost(url) |
|
506 if not host: raise IOError, ('ftp error', 'no host given') |
|
507 host, port = splitport(host) |
|
508 user, host = splituser(host) |
|
509 if user: user, passwd = splitpasswd(user) |
|
510 else: passwd = None |
|
511 host = unquote(host) |
|
512 user = unquote(user or '') |
|
513 passwd = unquote(passwd or '') |
|
514 host = socket.gethostbyname(host) |
|
515 if not port: |
|
516 import ftplib |
|
517 port = ftplib.FTP_PORT |
|
518 else: |
|
519 port = int(port) |
|
520 path, attrs = splitattr(path) |
|
521 path = unquote(path) |
|
522 dirs = path.split('/') |
|
523 dirs, file = dirs[:-1], dirs[-1] |
|
524 if dirs and not dirs[0]: dirs = dirs[1:] |
|
525 if dirs and not dirs[0]: dirs[0] = '/' |
|
526 key = user, host, port, '/'.join(dirs) |
|
527 # XXX thread unsafe! |
|
528 if len(self.ftpcache) > MAXFTPCACHE: |
|
529 # Prune the cache, rather arbitrarily |
|
530 for k in self.ftpcache.keys(): |
|
531 if k != key: |
|
532 v = self.ftpcache[k] |
|
533 del self.ftpcache[k] |
|
534 v.close() |
|
535 try: |
|
536 if not key in self.ftpcache: |
|
537 self.ftpcache[key] = \ |
|
538 ftpwrapper(user, passwd, host, port, dirs) |
|
539 if not file: type = 'D' |
|
540 else: type = 'I' |
|
541 for attr in attrs: |
|
542 attr, value = splitvalue(attr) |
|
543 if attr.lower() == 'type' and \ |
|
544 value in ('a', 'A', 'i', 'I', 'd', 'D'): |
|
545 type = value.upper() |
|
546 (fp, retrlen) = self.ftpcache[key].retrfile(file, type) |
|
547 mtype = mimetypes.guess_type("ftp:" + url)[0] |
|
548 headers = "" |
|
549 if mtype: |
|
550 headers += "Content-Type: %s\n" % mtype |
|
551 if retrlen is not None and retrlen >= 0: |
|
552 headers += "Content-Length: %d\n" % retrlen |
|
553 headers = mimetools.Message(StringIO(headers)) |
|
554 return addinfourl(fp, headers, "ftp:" + url) |
|
555 except ftperrors(), msg: |
|
556 raise IOError, ('ftp error', msg), sys.exc_info()[2] |
|
557 |
|
558 def open_data(self, url, data=None): |
|
559 """Use "data" URL.""" |
|
560 if not isinstance(url, str): |
|
561 raise IOError, ('data error', 'proxy support for data protocol currently not implemented') |
|
562 # ignore POSTed data |
|
563 # |
|
564 # syntax of data URLs: |
|
565 # dataurl := "data:" [ mediatype ] [ ";base64" ] "," data |
|
566 # mediatype := [ type "/" subtype ] *( ";" parameter ) |
|
567 # data := *urlchar |
|
568 # parameter := attribute "=" value |
|
569 import mimetools |
|
570 try: |
|
571 from cStringIO import StringIO |
|
572 except ImportError: |
|
573 from StringIO import StringIO |
|
574 try: |
|
575 [type, data] = url.split(',', 1) |
|
576 except ValueError: |
|
577 raise IOError, ('data error', 'bad data URL') |
|
578 if not type: |
|
579 type = 'text/plain;charset=US-ASCII' |
|
580 semi = type.rfind(';') |
|
581 if semi >= 0 and '=' not in type[semi:]: |
|
582 encoding = type[semi+1:] |
|
583 type = type[:semi] |
|
584 else: |
|
585 encoding = '' |
|
586 msg = [] |
|
587 msg.append('Date: %s'%time.strftime('%a, %d %b %Y %T GMT', |
|
588 time.gmtime(time.time()))) |
|
589 msg.append('Content-type: %s' % type) |
|
590 if encoding == 'base64': |
|
591 import base64 |
|
592 data = base64.decodestring(data) |
|
593 else: |
|
594 data = unquote(data) |
|
595 msg.append('Content-Length: %d' % len(data)) |
|
596 msg.append('') |
|
597 msg.append(data) |
|
598 msg = '\n'.join(msg) |
|
599 f = StringIO(msg) |
|
600 headers = mimetools.Message(f, 0) |
|
601 #f.fileno = None # needed for addinfourl |
|
602 return addinfourl(f, headers, url) |
|
603 |
|
604 |
|
605 class FancyURLopener(URLopener): |
|
606 """Derived class with handlers for errors we can handle (perhaps).""" |
|
607 |
|
608 def __init__(self, *args, **kwargs): |
|
609 URLopener.__init__(self, *args, **kwargs) |
|
610 self.auth_cache = {} |
|
611 self.tries = 0 |
|
612 self.maxtries = 10 |
|
613 |
|
614 def http_error_default(self, url, fp, errcode, errmsg, headers): |
|
615 """Default error handling -- don't raise an exception.""" |
|
616 return addinfourl(fp, headers, "http:" + url) |
|
617 |
|
618 def http_error_302(self, url, fp, errcode, errmsg, headers, data=None): |
|
619 """Error 302 -- relocated (temporarily).""" |
|
620 self.tries += 1 |
|
621 if self.maxtries and self.tries >= self.maxtries: |
|
622 if hasattr(self, "http_error_500"): |
|
623 meth = self.http_error_500 |
|
624 else: |
|
625 meth = self.http_error_default |
|
626 self.tries = 0 |
|
627 return meth(url, fp, 500, |
|
628 "Internal Server Error: Redirect Recursion", headers) |
|
629 result = self.redirect_internal(url, fp, errcode, errmsg, headers, |
|
630 data) |
|
631 self.tries = 0 |
|
632 return result |
|
633 |
|
634 def redirect_internal(self, url, fp, errcode, errmsg, headers, data): |
|
635 if 'location' in headers: |
|
636 newurl = headers['location'] |
|
637 elif 'uri' in headers: |
|
638 newurl = headers['uri'] |
|
639 else: |
|
640 return |
|
641 void = fp.read() |
|
642 fp.close() |
|
643 # In case the server sent a relative URL, join with original: |
|
644 newurl = basejoin(self.type + ":" + url, newurl) |
|
645 return self.open(newurl) |
|
646 |
|
647 def http_error_301(self, url, fp, errcode, errmsg, headers, data=None): |
|
648 """Error 301 -- also relocated (permanently).""" |
|
649 return self.http_error_302(url, fp, errcode, errmsg, headers, data) |
|
650 |
|
651 def http_error_303(self, url, fp, errcode, errmsg, headers, data=None): |
|
652 """Error 303 -- also relocated (essentially identical to 302).""" |
|
653 return self.http_error_302(url, fp, errcode, errmsg, headers, data) |
|
654 |
|
655 def http_error_307(self, url, fp, errcode, errmsg, headers, data=None): |
|
656 """Error 307 -- relocated, but turn POST into error.""" |
|
657 if data is None: |
|
658 return self.http_error_302(url, fp, errcode, errmsg, headers, data) |
|
659 else: |
|
660 return self.http_error_default(url, fp, errcode, errmsg, headers) |
|
661 |
|
662 def http_error_401(self, url, fp, errcode, errmsg, headers, data=None): |
|
663 """Error 401 -- authentication required. |
|
664 This function supports Basic authentication only.""" |
|
665 if not 'www-authenticate' in headers: |
|
666 URLopener.http_error_default(self, url, fp, |
|
667 errcode, errmsg, headers) |
|
668 stuff = headers['www-authenticate'] |
|
669 import re |
|
670 match = re.match('[ \t]*([^ \t]+)[ \t]+realm="([^"]*)"', stuff) |
|
671 if not match: |
|
672 URLopener.http_error_default(self, url, fp, |
|
673 errcode, errmsg, headers) |
|
674 scheme, realm = match.groups() |
|
675 if scheme.lower() != 'basic': |
|
676 URLopener.http_error_default(self, url, fp, |
|
677 errcode, errmsg, headers) |
|
678 name = 'retry_' + self.type + '_basic_auth' |
|
679 if data is None: |
|
680 return getattr(self,name)(url, realm) |
|
681 else: |
|
682 return getattr(self,name)(url, realm, data) |
|
683 |
|
684 def http_error_407(self, url, fp, errcode, errmsg, headers, data=None): |
|
685 """Error 407 -- proxy authentication required. |
|
686 This function supports Basic authentication only.""" |
|
687 if not 'proxy-authenticate' in headers: |
|
688 URLopener.http_error_default(self, url, fp, |
|
689 errcode, errmsg, headers) |
|
690 stuff = headers['proxy-authenticate'] |
|
691 import re |
|
692 match = re.match('[ \t]*([^ \t]+)[ \t]+realm="([^"]*)"', stuff) |
|
693 if not match: |
|
694 URLopener.http_error_default(self, url, fp, |
|
695 errcode, errmsg, headers) |
|
696 scheme, realm = match.groups() |
|
697 if scheme.lower() != 'basic': |
|
698 URLopener.http_error_default(self, url, fp, |
|
699 errcode, errmsg, headers) |
|
700 name = 'retry_proxy_' + self.type + '_basic_auth' |
|
701 if data is None: |
|
702 return getattr(self,name)(url, realm) |
|
703 else: |
|
704 return getattr(self,name)(url, realm, data) |
|
705 |
|
706 def retry_proxy_http_basic_auth(self, url, realm, data=None): |
|
707 host, selector = splithost(url) |
|
708 newurl = 'http://' + host + selector |
|
709 proxy = self.proxies['http'] |
|
710 urltype, proxyhost = splittype(proxy) |
|
711 proxyhost, proxyselector = splithost(proxyhost) |
|
712 i = proxyhost.find('@') + 1 |
|
713 proxyhost = proxyhost[i:] |
|
714 user, passwd = self.get_user_passwd(proxyhost, realm, i) |
|
715 if not (user or passwd): return None |
|
716 proxyhost = quote(user, safe='') + ':' + quote(passwd, safe='') + '@' + proxyhost |
|
717 self.proxies['http'] = 'http://' + proxyhost + proxyselector |
|
718 if data is None: |
|
719 return self.open(newurl) |
|
720 else: |
|
721 return self.open(newurl, data) |
|
722 |
|
723 def retry_proxy_https_basic_auth(self, url, realm, data=None): |
|
724 host, selector = splithost(url) |
|
725 newurl = 'https://' + host + selector |
|
726 proxy = self.proxies['https'] |
|
727 urltype, proxyhost = splittype(proxy) |
|
728 proxyhost, proxyselector = splithost(proxyhost) |
|
729 i = proxyhost.find('@') + 1 |
|
730 proxyhost = proxyhost[i:] |
|
731 user, passwd = self.get_user_passwd(proxyhost, realm, i) |
|
732 if not (user or passwd): return None |
|
733 proxyhost = quote(user, safe='') + ':' + quote(passwd, safe='') + '@' + proxyhost |
|
734 self.proxies['https'] = 'https://' + proxyhost + proxyselector |
|
735 if data is None: |
|
736 return self.open(newurl) |
|
737 else: |
|
738 return self.open(newurl, data) |
|
739 |
|
740 def retry_http_basic_auth(self, url, realm, data=None): |
|
741 host, selector = splithost(url) |
|
742 i = host.find('@') + 1 |
|
743 host = host[i:] |
|
744 user, passwd = self.get_user_passwd(host, realm, i) |
|
745 if not (user or passwd): return None |
|
746 host = quote(user, safe='') + ':' + quote(passwd, safe='') + '@' + host |
|
747 newurl = 'http://' + host + selector |
|
748 if data is None: |
|
749 return self.open(newurl) |
|
750 else: |
|
751 return self.open(newurl, data) |
|
752 |
|
753 def retry_https_basic_auth(self, url, realm, data=None): |
|
754 host, selector = splithost(url) |
|
755 i = host.find('@') + 1 |
|
756 host = host[i:] |
|
757 user, passwd = self.get_user_passwd(host, realm, i) |
|
758 if not (user or passwd): return None |
|
759 host = quote(user, safe='') + ':' + quote(passwd, safe='') + '@' + host |
|
760 newurl = 'https://' + host + selector |
|
761 if data is None: |
|
762 return self.open(newurl) |
|
763 else: |
|
764 return self.open(newurl, data) |
|
765 |
|
766 def get_user_passwd(self, host, realm, clear_cache = 0): |
|
767 key = realm + '@' + host.lower() |
|
768 if key in self.auth_cache: |
|
769 if clear_cache: |
|
770 del self.auth_cache[key] |
|
771 else: |
|
772 return self.auth_cache[key] |
|
773 user, passwd = self.prompt_user_passwd(host, realm) |
|
774 if user or passwd: self.auth_cache[key] = (user, passwd) |
|
775 return user, passwd |
|
776 |
|
777 def prompt_user_passwd(self, host, realm): |
|
778 """Override this in a GUI environment!""" |
|
779 import getpass |
|
780 try: |
|
781 user = raw_input("Enter username for %s at %s: " % (realm, |
|
782 host)) |
|
783 passwd = getpass.getpass("Enter password for %s in %s at %s: " % |
|
784 (user, realm, host)) |
|
785 return user, passwd |
|
786 except KeyboardInterrupt: |
|
787 print |
|
788 return None, None |
|
789 |
|
790 |
|
791 # Utility functions |
|
792 |
|
793 _localhost = None |
|
794 def localhost(): |
|
795 """Return the IP address of the magic hostname 'localhost'.""" |
|
796 global _localhost |
|
797 if _localhost is None: |
|
798 _localhost = socket.gethostbyname('localhost') |
|
799 return _localhost |
|
800 |
|
801 _thishost = None |
|
802 def thishost(): |
|
803 """Return the IP address of the current host.""" |
|
804 global _thishost |
|
805 if _thishost is None: |
|
806 _thishost = socket.gethostbyname(socket.gethostname()) |
|
807 return _thishost |
|
808 |
|
809 _ftperrors = None |
|
810 def ftperrors(): |
|
811 """Return the set of errors raised by the FTP class.""" |
|
812 global _ftperrors |
|
813 if _ftperrors is None: |
|
814 import ftplib |
|
815 _ftperrors = ftplib.all_errors |
|
816 return _ftperrors |
|
817 |
|
818 _noheaders = None |
|
819 def noheaders(): |
|
820 """Return an empty mimetools.Message object.""" |
|
821 global _noheaders |
|
822 if _noheaders is None: |
|
823 import mimetools |
|
824 try: |
|
825 from cStringIO import StringIO |
|
826 except ImportError: |
|
827 from StringIO import StringIO |
|
828 _noheaders = mimetools.Message(StringIO(), 0) |
|
829 _noheaders.fp.close() # Recycle file descriptor |
|
830 return _noheaders |
|
831 |
|
832 |
|
833 # Utility classes |
|
834 |
|
835 class ftpwrapper: |
|
836 """Class used by open_ftp() for cache of open FTP connections.""" |
|
837 |
|
838 def __init__(self, user, passwd, host, port, dirs): |
|
839 self.user = user |
|
840 self.passwd = passwd |
|
841 self.host = host |
|
842 self.port = port |
|
843 self.dirs = dirs |
|
844 self.init() |
|
845 |
|
846 def init(self): |
|
847 import ftplib |
|
848 self.busy = 0 |
|
849 self.ftp = ftplib.FTP() |
|
850 self.ftp.connect(self.host, self.port) |
|
851 self.ftp.login(self.user, self.passwd) |
|
852 for dir in self.dirs: |
|
853 self.ftp.cwd(dir) |
|
854 |
|
855 def retrfile(self, file, type): |
|
856 import ftplib |
|
857 self.endtransfer() |
|
858 if type in ('d', 'D'): cmd = 'TYPE A'; isdir = 1 |
|
859 else: cmd = 'TYPE ' + type; isdir = 0 |
|
860 try: |
|
861 self.ftp.voidcmd(cmd) |
|
862 except ftplib.all_errors: |
|
863 self.init() |
|
864 self.ftp.voidcmd(cmd) |
|
865 conn = None |
|
866 if file and not isdir: |
|
867 # Try to retrieve as a file |
|
868 try: |
|
869 cmd = 'RETR ' + file |
|
870 conn = self.ftp.ntransfercmd(cmd) |
|
871 except ftplib.error_perm, reason: |
|
872 if str(reason)[:3] != '550': |
|
873 raise IOError, ('ftp error', reason), sys.exc_info()[2] |
|
874 if not conn: |
|
875 # Set transfer mode to ASCII! |
|
876 self.ftp.voidcmd('TYPE A') |
|
877 # Try a directory listing |
|
878 if file: cmd = 'LIST ' + file |
|
879 else: cmd = 'LIST' |
|
880 conn = self.ftp.ntransfercmd(cmd) |
|
881 self.busy = 1 |
|
882 # Pass back both a suitably decorated object and a retrieval length |
|
883 return (addclosehook(conn[0].makefile('rb'), |
|
884 self.endtransfer), conn[1]) |
|
885 def endtransfer(self): |
|
886 if not self.busy: |
|
887 return |
|
888 self.busy = 0 |
|
889 try: |
|
890 self.ftp.voidresp() |
|
891 except ftperrors(): |
|
892 pass |
|
893 |
|
894 def close(self): |
|
895 self.endtransfer() |
|
896 try: |
|
897 self.ftp.close() |
|
898 except ftperrors(): |
|
899 pass |
|
900 |
|
901 class addbase: |
|
902 """Base class for addinfo and addclosehook.""" |
|
903 |
|
904 def __init__(self, fp): |
|
905 self.fp = fp |
|
906 self.read = self.fp.read |
|
907 self.readline = self.fp.readline |
|
908 if hasattr(self.fp, "readlines"): self.readlines = self.fp.readlines |
|
909 if hasattr(self.fp, "fileno"): |
|
910 self.fileno = self.fp.fileno |
|
911 else: |
|
912 self.fileno = lambda: None |
|
913 if hasattr(self.fp, "__iter__"): |
|
914 self.__iter__ = self.fp.__iter__ |
|
915 if hasattr(self.fp, "next"): |
|
916 self.next = self.fp.next |
|
917 |
|
918 def __repr__(self): |
|
919 return '<%s at %r whose fp = %r>' % (self.__class__.__name__, |
|
920 id(self), self.fp) |
|
921 |
|
922 def close(self): |
|
923 self.read = None |
|
924 self.readline = None |
|
925 self.readlines = None |
|
926 self.fileno = None |
|
927 if self.fp: self.fp.close() |
|
928 self.fp = None |
|
929 |
|
930 class addclosehook(addbase): |
|
931 """Class to add a close hook to an open file.""" |
|
932 |
|
933 def __init__(self, fp, closehook, *hookargs): |
|
934 addbase.__init__(self, fp) |
|
935 self.closehook = closehook |
|
936 self.hookargs = hookargs |
|
937 |
|
938 def close(self): |
|
939 addbase.close(self) |
|
940 if self.closehook: |
|
941 self.closehook(*self.hookargs) |
|
942 self.closehook = None |
|
943 self.hookargs = None |
|
944 |
|
945 class addinfo(addbase): |
|
946 """class to add an info() method to an open file.""" |
|
947 |
|
948 def __init__(self, fp, headers): |
|
949 addbase.__init__(self, fp) |
|
950 self.headers = headers |
|
951 |
|
952 def info(self): |
|
953 return self.headers |
|
954 |
|
955 class addinfourl(addbase): |
|
956 """class to add info() and geturl() methods to an open file.""" |
|
957 |
|
958 def __init__(self, fp, headers, url): |
|
959 addbase.__init__(self, fp) |
|
960 self.headers = headers |
|
961 self.url = url |
|
962 |
|
963 def info(self): |
|
964 return self.headers |
|
965 |
|
966 def geturl(self): |
|
967 return self.url |
|
968 |
|
969 |
|
970 # Utilities to parse URLs (most of these return None for missing parts): |
|
971 # unwrap('<URL:type://host/path>') --> 'type://host/path' |
|
972 # splittype('type:opaquestring') --> 'type', 'opaquestring' |
|
973 # splithost('//host[:port]/path') --> 'host[:port]', '/path' |
|
974 # splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]' |
|
975 # splitpasswd('user:passwd') -> 'user', 'passwd' |
|
976 # splitport('host:port') --> 'host', 'port' |
|
977 # splitquery('/path?query') --> '/path', 'query' |
|
978 # splittag('/path#tag') --> '/path', 'tag' |
|
979 # splitattr('/path;attr1=value1;attr2=value2;...') -> |
|
980 # '/path', ['attr1=value1', 'attr2=value2', ...] |
|
981 # splitvalue('attr=value') --> 'attr', 'value' |
|
982 # splitgophertype('/Xselector') --> 'X', 'selector' |
|
983 # unquote('abc%20def') -> 'abc def' |
|
984 # quote('abc def') -> 'abc%20def') |
|
985 |
|
986 try: |
|
987 unicode |
|
988 except NameError: |
|
989 def _is_unicode(x): |
|
990 return 0 |
|
991 else: |
|
992 def _is_unicode(x): |
|
993 return isinstance(x, unicode) |
|
994 |
|
995 def toBytes(url): |
|
996 """toBytes(u"URL") --> 'URL'.""" |
|
997 # Most URL schemes require ASCII. If that changes, the conversion |
|
998 # can be relaxed |
|
999 if _is_unicode(url): |
|
1000 try: |
|
1001 url = url.encode("ASCII") |
|
1002 except UnicodeError: |
|
1003 raise UnicodeError("URL " + repr(url) + |
|
1004 " contains non-ASCII characters") |
|
1005 return url |
|
1006 |
|
1007 def unwrap(url): |
|
1008 """unwrap('<URL:type://host/path>') --> 'type://host/path'.""" |
|
1009 url = url.strip() |
|
1010 if url[:1] == '<' and url[-1:] == '>': |
|
1011 url = url[1:-1].strip() |
|
1012 if url[:4] == 'URL:': url = url[4:].strip() |
|
1013 return url |
|
1014 |
|
1015 _typeprog = None |
|
1016 def splittype(url): |
|
1017 """splittype('type:opaquestring') --> 'type', 'opaquestring'.""" |
|
1018 global _typeprog |
|
1019 if _typeprog is None: |
|
1020 import re |
|
1021 _typeprog = re.compile('^([^/:]+):') |
|
1022 |
|
1023 match = _typeprog.match(url) |
|
1024 if match: |
|
1025 scheme = match.group(1) |
|
1026 return scheme.lower(), url[len(scheme) + 1:] |
|
1027 return None, url |
|
1028 |
|
1029 _hostprog = None |
|
1030 def splithost(url): |
|
1031 """splithost('//host[:port]/path') --> 'host[:port]', '/path'.""" |
|
1032 global _hostprog |
|
1033 if _hostprog is None: |
|
1034 import re |
|
1035 _hostprog = re.compile('^//([^/?]*)(.*)$') |
|
1036 |
|
1037 match = _hostprog.match(url) |
|
1038 if match: return match.group(1, 2) |
|
1039 return None, url |
|
1040 |
|
1041 _userprog = None |
|
1042 def splituser(host): |
|
1043 """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'.""" |
|
1044 global _userprog |
|
1045 if _userprog is None: |
|
1046 import re |
|
1047 _userprog = re.compile('^(.*)@(.*)$') |
|
1048 |
|
1049 match = _userprog.match(host) |
|
1050 if match: return map(unquote, match.group(1, 2)) |
|
1051 return None, host |
|
1052 |
|
1053 _passwdprog = None |
|
1054 def splitpasswd(user): |
|
1055 """splitpasswd('user:passwd') -> 'user', 'passwd'.""" |
|
1056 global _passwdprog |
|
1057 if _passwdprog is None: |
|
1058 import re |
|
1059 _passwdprog = re.compile('^([^:]*):(.*)$') |
|
1060 |
|
1061 match = _passwdprog.match(user) |
|
1062 if match: return match.group(1, 2) |
|
1063 return user, None |
|
1064 |
|
1065 # splittag('/path#tag') --> '/path', 'tag' |
|
1066 _portprog = None |
|
1067 def splitport(host): |
|
1068 """splitport('host:port') --> 'host', 'port'.""" |
|
1069 global _portprog |
|
1070 if _portprog is None: |
|
1071 import re |
|
1072 _portprog = re.compile('^(.*):([0-9]+)$') |
|
1073 |
|
1074 match = _portprog.match(host) |
|
1075 if match: return match.group(1, 2) |
|
1076 return host, None |
|
1077 |
|
1078 _nportprog = None |
|
1079 def splitnport(host, defport=-1): |
|
1080 """Split host and port, returning numeric port. |
|
1081 Return given default port if no ':' found; defaults to -1. |
|
1082 Return numerical port if a valid number are found after ':'. |
|
1083 Return None if ':' but not a valid number.""" |
|
1084 global _nportprog |
|
1085 if _nportprog is None: |
|
1086 import re |
|
1087 _nportprog = re.compile('^(.*):(.*)$') |
|
1088 |
|
1089 match = _nportprog.match(host) |
|
1090 if match: |
|
1091 host, port = match.group(1, 2) |
|
1092 try: |
|
1093 if not port: raise ValueError, "no digits" |
|
1094 nport = int(port) |
|
1095 except ValueError: |
|
1096 nport = None |
|
1097 return host, nport |
|
1098 return host, defport |
|
1099 |
|
1100 _queryprog = None |
|
1101 def splitquery(url): |
|
1102 """splitquery('/path?query') --> '/path', 'query'.""" |
|
1103 global _queryprog |
|
1104 if _queryprog is None: |
|
1105 import re |
|
1106 _queryprog = re.compile('^(.*)\?([^?]*)$') |
|
1107 |
|
1108 match = _queryprog.match(url) |
|
1109 if match: return match.group(1, 2) |
|
1110 return url, None |
|
1111 |
|
1112 _tagprog = None |
|
1113 def splittag(url): |
|
1114 """splittag('/path#tag') --> '/path', 'tag'.""" |
|
1115 global _tagprog |
|
1116 if _tagprog is None: |
|
1117 import re |
|
1118 _tagprog = re.compile('^(.*)#([^#]*)$') |
|
1119 |
|
1120 match = _tagprog.match(url) |
|
1121 if match: return match.group(1, 2) |
|
1122 return url, None |
|
1123 |
|
1124 def splitattr(url): |
|
1125 """splitattr('/path;attr1=value1;attr2=value2;...') -> |
|
1126 '/path', ['attr1=value1', 'attr2=value2', ...].""" |
|
1127 words = url.split(';') |
|
1128 return words[0], words[1:] |
|
1129 |
|
1130 _valueprog = None |
|
1131 def splitvalue(attr): |
|
1132 """splitvalue('attr=value') --> 'attr', 'value'.""" |
|
1133 global _valueprog |
|
1134 if _valueprog is None: |
|
1135 import re |
|
1136 _valueprog = re.compile('^([^=]*)=(.*)$') |
|
1137 |
|
1138 match = _valueprog.match(attr) |
|
1139 if match: return match.group(1, 2) |
|
1140 return attr, None |
|
1141 |
|
1142 def splitgophertype(selector): |
|
1143 """splitgophertype('/Xselector') --> 'X', 'selector'.""" |
|
1144 if selector[:1] == '/' and selector[1:2]: |
|
1145 return selector[1], selector[2:] |
|
1146 return None, selector |
|
1147 |
|
1148 _hextochr = dict(('%02x' % i, chr(i)) for i in range(256)) |
|
1149 _hextochr.update(('%02X' % i, chr(i)) for i in range(256)) |
|
1150 |
|
1151 def unquote(s): |
|
1152 """unquote('abc%20def') -> 'abc def'.""" |
|
1153 res = s.split('%') |
|
1154 for i in xrange(1, len(res)): |
|
1155 item = res[i] |
|
1156 try: |
|
1157 res[i] = _hextochr[item[:2]] + item[2:] |
|
1158 except KeyError: |
|
1159 res[i] = '%' + item |
|
1160 except UnicodeDecodeError: |
|
1161 res[i] = unichr(int(item[:2], 16)) + item[2:] |
|
1162 return "".join(res) |
|
1163 |
|
1164 def unquote_plus(s): |
|
1165 """unquote('%7e/abc+def') -> '~/abc def'""" |
|
1166 s = s.replace('+', ' ') |
|
1167 return unquote(s) |
|
1168 |
|
1169 always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ' |
|
1170 'abcdefghijklmnopqrstuvwxyz' |
|
1171 '0123456789' '_.-') |
|
1172 _safemaps = {} |
|
1173 |
|
1174 def quote(s, safe = '/'): |
|
1175 """quote('abc def') -> 'abc%20def' |
|
1176 |
|
1177 Each part of a URL, e.g. the path info, the query, etc., has a |
|
1178 different set of reserved characters that must be quoted. |
|
1179 |
|
1180 RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists |
|
1181 the following reserved characters. |
|
1182 |
|
1183 reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | |
|
1184 "$" | "," |
|
1185 |
|
1186 Each of these characters is reserved in some component of a URL, |
|
1187 but not necessarily in all of them. |
|
1188 |
|
1189 By default, the quote function is intended for quoting the path |
|
1190 section of a URL. Thus, it will not encode '/'. This character |
|
1191 is reserved, but in typical usage the quote function is being |
|
1192 called on a path where the existing slash characters are used as |
|
1193 reserved characters. |
|
1194 """ |
|
1195 cachekey = (safe, always_safe) |
|
1196 try: |
|
1197 safe_map = _safemaps[cachekey] |
|
1198 except KeyError: |
|
1199 safe += always_safe |
|
1200 safe_map = {} |
|
1201 for i in range(256): |
|
1202 c = chr(i) |
|
1203 safe_map[c] = (c in safe) and c or ('%%%02X' % i) |
|
1204 _safemaps[cachekey] = safe_map |
|
1205 res = map(safe_map.__getitem__, s) |
|
1206 return ''.join(res) |
|
1207 |
|
1208 def quote_plus(s, safe = ''): |
|
1209 """Quote the query fragment of a URL; replacing ' ' with '+'""" |
|
1210 if ' ' in s: |
|
1211 s = quote(s, safe + ' ') |
|
1212 return s.replace(' ', '+') |
|
1213 return quote(s, safe) |
|
1214 |
|
1215 def urlencode(query,doseq=0): |
|
1216 """Encode a sequence of two-element tuples or dictionary into a URL query string. |
|
1217 |
|
1218 If any values in the query arg are sequences and doseq is true, each |
|
1219 sequence element is converted to a separate parameter. |
|
1220 |
|
1221 If the query arg is a sequence of two-element tuples, the order of the |
|
1222 parameters in the output will match the order of parameters in the |
|
1223 input. |
|
1224 """ |
|
1225 |
|
1226 if hasattr(query,"items"): |
|
1227 # mapping objects |
|
1228 query = query.items() |
|
1229 else: |
|
1230 # it's a bother at times that strings and string-like objects are |
|
1231 # sequences... |
|
1232 try: |
|
1233 # non-sequence items should not work with len() |
|
1234 # non-empty strings will fail this |
|
1235 if len(query) and not isinstance(query[0], tuple): |
|
1236 raise TypeError |
|
1237 # zero-length sequences of all types will get here and succeed, |
|
1238 # but that's a minor nit - since the original implementation |
|
1239 # allowed empty dicts that type of behavior probably should be |
|
1240 # preserved for consistency |
|
1241 except TypeError: |
|
1242 ty,va,tb = sys.exc_info() |
|
1243 raise TypeError, "not a valid non-string sequence or mapping object", tb |
|
1244 |
|
1245 l = [] |
|
1246 if not doseq: |
|
1247 # preserve old behavior |
|
1248 for k, v in query: |
|
1249 k = quote_plus(str(k)) |
|
1250 v = quote_plus(str(v)) |
|
1251 l.append(k + '=' + v) |
|
1252 else: |
|
1253 for k, v in query: |
|
1254 k = quote_plus(str(k)) |
|
1255 if isinstance(v, str): |
|
1256 v = quote_plus(v) |
|
1257 l.append(k + '=' + v) |
|
1258 elif _is_unicode(v): |
|
1259 # is there a reasonable way to convert to ASCII? |
|
1260 # encode generates a string, but "replace" or "ignore" |
|
1261 # lose information and "strict" can raise UnicodeError |
|
1262 v = quote_plus(v.encode("ASCII","replace")) |
|
1263 l.append(k + '=' + v) |
|
1264 else: |
|
1265 try: |
|
1266 # is this a sufficient test for sequence-ness? |
|
1267 x = len(v) |
|
1268 except TypeError: |
|
1269 # not a sequence |
|
1270 v = quote_plus(str(v)) |
|
1271 l.append(k + '=' + v) |
|
1272 else: |
|
1273 # loop over the sequence |
|
1274 for elt in v: |
|
1275 l.append(k + '=' + quote_plus(str(elt))) |
|
1276 return '&'.join(l) |
|
1277 |
|
1278 # Proxy handling |
|
1279 def getproxies_environment(): |
|
1280 """Return a dictionary of scheme -> proxy server URL mappings. |
|
1281 |
|
1282 Scan the environment for variables named <scheme>_proxy; |
|
1283 this seems to be the standard convention. If you need a |
|
1284 different way, you can pass a proxies dictionary to the |
|
1285 [Fancy]URLopener constructor. |
|
1286 |
|
1287 """ |
|
1288 proxies = {} |
|
1289 for name, value in os.environ.items(): |
|
1290 name = name.lower() |
|
1291 if value and name[-6:] == '_proxy': |
|
1292 proxies[name[:-6]] = value |
|
1293 return proxies |
|
1294 |
|
1295 if sys.platform == 'darwin': |
|
1296 def getproxies_internetconfig(): |
|
1297 """Return a dictionary of scheme -> proxy server URL mappings. |
|
1298 |
|
1299 By convention the mac uses Internet Config to store |
|
1300 proxies. An HTTP proxy, for instance, is stored under |
|
1301 the HttpProxy key. |
|
1302 |
|
1303 """ |
|
1304 try: |
|
1305 import ic |
|
1306 except ImportError: |
|
1307 return {} |
|
1308 |
|
1309 try: |
|
1310 config = ic.IC() |
|
1311 except ic.error: |
|
1312 return {} |
|
1313 proxies = {} |
|
1314 # HTTP: |
|
1315 if 'UseHTTPProxy' in config and config['UseHTTPProxy']: |
|
1316 try: |
|
1317 value = config['HTTPProxyHost'] |
|
1318 except ic.error: |
|
1319 pass |
|
1320 else: |
|
1321 proxies['http'] = 'http://%s' % value |
|
1322 # FTP: XXXX To be done. |
|
1323 # Gopher: XXXX To be done. |
|
1324 return proxies |
|
1325 |
|
1326 def proxy_bypass(x): |
|
1327 return 0 |
|
1328 |
|
1329 def getproxies(): |
|
1330 return getproxies_environment() or getproxies_internetconfig() |
|
1331 |
|
1332 elif os.name == 'nt': |
|
1333 def getproxies_registry(): |
|
1334 """Return a dictionary of scheme -> proxy server URL mappings. |
|
1335 |
|
1336 Win32 uses the registry to store proxies. |
|
1337 |
|
1338 """ |
|
1339 proxies = {} |
|
1340 try: |
|
1341 import _winreg |
|
1342 except ImportError: |
|
1343 # Std module, so should be around - but you never know! |
|
1344 return proxies |
|
1345 try: |
|
1346 internetSettings = _winreg.OpenKey(_winreg.HKEY_CURRENT_USER, |
|
1347 r'Software\Microsoft\Windows\CurrentVersion\Internet Settings') |
|
1348 proxyEnable = _winreg.QueryValueEx(internetSettings, |
|
1349 'ProxyEnable')[0] |
|
1350 if proxyEnable: |
|
1351 # Returned as Unicode but problems if not converted to ASCII |
|
1352 proxyServer = str(_winreg.QueryValueEx(internetSettings, |
|
1353 'ProxyServer')[0]) |
|
1354 if '=' in proxyServer: |
|
1355 # Per-protocol settings |
|
1356 for p in proxyServer.split(';'): |
|
1357 protocol, address = p.split('=', 1) |
|
1358 # See if address has a type:// prefix |
|
1359 import re |
|
1360 if not re.match('^([^/:]+)://', address): |
|
1361 address = '%s://%s' % (protocol, address) |
|
1362 proxies[protocol] = address |
|
1363 else: |
|
1364 # Use one setting for all protocols |
|
1365 if proxyServer[:5] == 'http:': |
|
1366 proxies['http'] = proxyServer |
|
1367 else: |
|
1368 proxies['http'] = 'http://%s' % proxyServer |
|
1369 proxies['ftp'] = 'ftp://%s' % proxyServer |
|
1370 internetSettings.Close() |
|
1371 except (WindowsError, ValueError, TypeError): |
|
1372 # Either registry key not found etc, or the value in an |
|
1373 # unexpected format. |
|
1374 # proxies already set up to be empty so nothing to do |
|
1375 pass |
|
1376 return proxies |
|
1377 |
|
1378 def getproxies(): |
|
1379 """Return a dictionary of scheme -> proxy server URL mappings. |
|
1380 |
|
1381 Returns settings gathered from the environment, if specified, |
|
1382 or the registry. |
|
1383 |
|
1384 """ |
|
1385 return getproxies_environment() or getproxies_registry() |
|
1386 |
|
1387 def proxy_bypass(host): |
|
1388 try: |
|
1389 import _winreg |
|
1390 import re |
|
1391 except ImportError: |
|
1392 # Std modules, so should be around - but you never know! |
|
1393 return 0 |
|
1394 try: |
|
1395 internetSettings = _winreg.OpenKey(_winreg.HKEY_CURRENT_USER, |
|
1396 r'Software\Microsoft\Windows\CurrentVersion\Internet Settings') |
|
1397 proxyEnable = _winreg.QueryValueEx(internetSettings, |
|
1398 'ProxyEnable')[0] |
|
1399 proxyOverride = str(_winreg.QueryValueEx(internetSettings, |
|
1400 'ProxyOverride')[0]) |
|
1401 # ^^^^ Returned as Unicode but problems if not converted to ASCII |
|
1402 except WindowsError: |
|
1403 return 0 |
|
1404 if not proxyEnable or not proxyOverride: |
|
1405 return 0 |
|
1406 # try to make a host list from name and IP address. |
|
1407 rawHost, port = splitport(host) |
|
1408 host = [rawHost] |
|
1409 try: |
|
1410 addr = socket.gethostbyname(rawHost) |
|
1411 if addr != rawHost: |
|
1412 host.append(addr) |
|
1413 except socket.error: |
|
1414 pass |
|
1415 try: |
|
1416 fqdn = socket.getfqdn(rawHost) |
|
1417 if fqdn != rawHost: |
|
1418 host.append(fqdn) |
|
1419 except socket.error: |
|
1420 pass |
|
1421 # make a check value list from the registry entry: replace the |
|
1422 # '<local>' string by the localhost entry and the corresponding |
|
1423 # canonical entry. |
|
1424 proxyOverride = proxyOverride.split(';') |
|
1425 i = 0 |
|
1426 while i < len(proxyOverride): |
|
1427 if proxyOverride[i] == '<local>': |
|
1428 proxyOverride[i:i+1] = ['localhost', |
|
1429 '127.0.0.1', |
|
1430 socket.gethostname(), |
|
1431 socket.gethostbyname( |
|
1432 socket.gethostname())] |
|
1433 i += 1 |
|
1434 # print proxyOverride |
|
1435 # now check if we match one of the registry values. |
|
1436 for test in proxyOverride: |
|
1437 test = test.replace(".", r"\.") # mask dots |
|
1438 test = test.replace("*", r".*") # change glob sequence |
|
1439 test = test.replace("?", r".") # change glob char |
|
1440 for val in host: |
|
1441 # print "%s <--> %s" %( test, val ) |
|
1442 if re.match(test, val, re.I): |
|
1443 return 1 |
|
1444 return 0 |
|
1445 |
|
1446 else: |
|
1447 # By default use environment variables |
|
1448 getproxies = getproxies_environment |
|
1449 |
|
1450 def proxy_bypass(host): |
|
1451 return 0 |
|
1452 |
|
1453 # Test and time quote() and unquote() |
|
1454 def test1(): |
|
1455 s = '' |
|
1456 for i in range(256): s = s + chr(i) |
|
1457 s = s*4 |
|
1458 t0 = time.time() |
|
1459 qs = quote(s) |
|
1460 uqs = unquote(qs) |
|
1461 t1 = time.time() |
|
1462 if uqs != s: |
|
1463 print 'Wrong!' |
|
1464 print repr(s) |
|
1465 print repr(qs) |
|
1466 print repr(uqs) |
|
1467 print round(t1 - t0, 3), 'sec' |
|
1468 |
|
1469 |
|
1470 def reporthook(blocknum, blocksize, totalsize): |
|
1471 # Report during remote transfers |
|
1472 print "Block number: %d, Block size: %d, Total size: %d" % ( |
|
1473 blocknum, blocksize, totalsize) |
|
1474 |
|
1475 # Test program |
|
1476 def test(args=[]): |
|
1477 if not args: |
|
1478 args = [ |
|
1479 '/etc/passwd', |
|
1480 'file:/etc/passwd', |
|
1481 'file://localhost/etc/passwd', |
|
1482 'ftp://ftp.gnu.org/pub/README', |
|
1483 ## 'gopher://gopher.micro.umn.edu/1/', |
|
1484 'http://www.python.org/index.html', |
|
1485 ] |
|
1486 if hasattr(URLopener, "open_https"): |
|
1487 args.append('https://synergy.as.cmu.edu/~geek/') |
|
1488 try: |
|
1489 for url in args: |
|
1490 print '-'*10, url, '-'*10 |
|
1491 fn, h = urlretrieve(url, None, reporthook) |
|
1492 print fn |
|
1493 if h: |
|
1494 print '======' |
|
1495 for k in h.keys(): print k + ':', h[k] |
|
1496 print '======' |
|
1497 fp = open(fn, 'rb') |
|
1498 data = fp.read() |
|
1499 del fp |
|
1500 if '\r' in data: |
|
1501 table = string.maketrans("", "") |
|
1502 data = data.translate(table, "\r") |
|
1503 print data |
|
1504 fn, h = None, None |
|
1505 print '-'*40 |
|
1506 finally: |
|
1507 urlcleanup() |
|
1508 |
|
1509 def main(): |
|
1510 import getopt, sys |
|
1511 try: |
|
1512 opts, args = getopt.getopt(sys.argv[1:], "th") |
|
1513 except getopt.error, msg: |
|
1514 print msg |
|
1515 print "Use -h for help" |
|
1516 return |
|
1517 t = 0 |
|
1518 for o, a in opts: |
|
1519 if o == '-t': |
|
1520 t = t + 1 |
|
1521 if o == '-h': |
|
1522 print "Usage: python urllib.py [-t] [url ...]" |
|
1523 print "-t runs self-test;", |
|
1524 print "otherwise, contents of urls are printed" |
|
1525 return |
|
1526 if t: |
|
1527 if t > 1: |
|
1528 test1() |
|
1529 test(args) |
|
1530 else: |
|
1531 if not args: |
|
1532 print "Use -h for help" |
|
1533 for url in args: |
|
1534 print urlopen(url).read(), |
|
1535 |
|
1536 # Run test program when run as a script |
|
1537 if __name__ == '__main__': |
|
1538 main() |