1 """Open an arbitrary URL.
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
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/
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)
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.)
30 from urlparse import urljoin as basejoin
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"]
40 __version__ = '1.17' # XXX This version is not always updated :-(
42 MAXFTPCACHE = 10 # Trim the ftp cache beyond this size
44 # Helper for non-unix systems
46 from macurl2path import url2pathname, pathname2url
48 from nturl2path import url2pathname, pathname2url
49 elif os.name == 'riscos':
50 from rourl2path import url2pathname, pathname2url
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)
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)
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?
69 # Shortcut for basic usage
71 def urlopen(url, data=None, proxies=None):
72 """urlopen(url [, data]) -> open file-like object"""
74 if proxies is not None:
75 opener = FancyURLopener(proxies=proxies)
77 opener = FancyURLopener()
82 return opener.open(url)
84 return opener.open(url, data)
85 def urlretrieve(url, filename=None, reporthook=None, data=None):
88 _urlopener = FancyURLopener()
89 return _urlopener.retrieve(url, filename, reporthook, data)
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
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)."""
111 version = "Python-urllib/%s" % __version__
114 def __init__(self, proxies=None, **x509):
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
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.
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.
148 for file in self.__tempfiles:
153 del self.__tempfiles[:]
155 self.tempcache.clear()
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)
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)
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_*()
180 name = 'open_' + urltype
182 name = name.replace('-', '_')
183 if not hasattr(self, name):
185 return self.open_unknown_proxy(proxy, fullurl, data)
187 return self.open_unknown(fullurl, data)
190 return getattr(self, name)(url)
192 return getattr(self, name)(url, data)
193 except socket.error, msg:
194 raise IOError, ('socket error', msg), sys.exc_info()[2]
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)
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)
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'):
216 fp = self.open_local_file(url1)
219 return url2pathname(splithost(url1)[1]), hdrs
222 fp = self.open(url, data)
225 tfp = open(filename, 'wb')
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
244 if "content-length" in headers:
245 size = int(headers["Content-Length"])
246 reporthook(blocknum, bs, size)
255 reporthook(blocknum, bs, size)
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)
268 # Each method named open_<type> knows how to open that type of URL
270 def open_http(self, url, data=None):
271 """Use HTTP protocol."""
275 if isinstance(url, str):
276 host, selector = splithost(url)
278 user_passwd, host = splituser(host)
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)
289 if urltype.lower() != 'http':
292 realhost, rest = splithost(rest)
294 user_passwd, realhost = splituser(realhost)
296 selector = "%s://%s%s" % (urltype, realhost, rest)
297 if proxy_bypass(realhost):
300 #print "proxy via http:", host, selector
301 if not host: raise IOError, ('http error', 'no host given')
305 proxy_auth = base64.b64encode(proxy_passwd).strip()
311 auth = base64.b64encode(user_passwd).strip()
314 h = httplib.HTTP(host)
316 h.putrequest('POST', selector)
317 h.putheader('Content-Type', 'application/x-www-form-urlencoded')
318 h.putheader('Content-Length', '%d' % len(data))
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)
328 errcode, errmsg, headers = h.getreply()
330 # something went wrong with the HTTP status line
331 raise IOError, ('http protocol error', 0,
332 'got a bad status line', None)
335 return addinfourl(fp, headers, "http:" + url)
338 return self.http_error(url, fp, errcode, errmsg, headers)
340 return self.http_error(url, fp, errcode, errmsg, headers, data)
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)
351 result = method(url, fp, errcode, errmsg, headers)
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)
357 def http_error_default(self, url, fp, errcode, errmsg, headers):
358 """Default error handler: close the connection and raise IOError."""
361 raise IOError, ('http error', errcode, errmsg, headers)
363 if hasattr(socket, "ssl"):
364 def open_https(self, url, data=None):
365 """Use HTTPS protocol."""
369 if isinstance(url, str):
370 host, selector = splithost(url)
372 user_passwd, host = splituser(host)
377 # here, we determine, whether the proxy contains authorization information
378 proxy_passwd, host = splituser(host)
379 urltype, rest = splittype(selector)
382 if urltype.lower() != 'https':
385 realhost, rest = splithost(rest)
387 user_passwd, realhost = splituser(realhost)
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')
394 proxy_auth = base64.b64encode(proxy_passwd).strip()
399 auth = base64.b64encode(user_passwd).strip()
402 h = httplib.HTTPS(host, 0,
403 key_file=self.key_file,
404 cert_file=self.cert_file)
406 h.putrequest('POST', selector)
407 h.putheader('Content-Type',
408 'application/x-www-form-urlencoded')
409 h.putheader('Content-Length', '%d' % len(data))
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)
419 errcode, errmsg, headers = h.getreply()
421 # something went wrong with the HTTP status line
422 raise IOError, ('http protocol error', 0,
423 'got a bad status line', None)
426 return addinfourl(fp, headers, "https:" + url)
429 return self.http_error(url, fp, errcode, errmsg, headers)
431 return self.http_error(url, fp, errcode, errmsg, headers,
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')
439 host, selector = splithost(url)
440 if not host: raise IOError, ('gopher error', 'no host given')
442 type, selector = splitgophertype(selector)
443 selector, query = splitquery(selector)
444 selector = unquote(selector)
446 query = unquote(query)
447 fp = gopherlib.send_query(selector, query, host)
449 fp = gopherlib.send_selector(selector, host)
450 return addinfourl(fp, noheaders(), "gopher:" + url)
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)
459 return self.open_local_file(url)
461 def open_local_file(self, url):
462 """Use local file."""
463 import mimetypes, mimetools, email.Utils
465 from cStringIO import StringIO
467 from StringIO import StringIO
468 host, file = splithost(url)
469 localname = url2pathname(file)
471 stats = os.stat(localname)
473 raise IOError(e.errno, e.strerror, e.filename)
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)))
483 urlfile = 'file://' + file
484 return addinfourl(open(localname, 'rb'),
486 host, port = splitport(host)
488 and socket.gethostbyname(host) in (localhost(), thishost()):
491 urlfile = 'file://' + file
492 return addinfourl(open(localname, 'rb'),
494 raise IOError, ('local file error', 'not on local host')
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
502 from cStringIO import StringIO
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)
512 user = unquote(user or '')
513 passwd = unquote(passwd or '')
514 host = socket.gethostbyname(host)
517 port = ftplib.FTP_PORT
520 path, attrs = splitattr(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)
528 if len(self.ftpcache) > MAXFTPCACHE:
529 # Prune the cache, rather arbitrarily
530 for k in self.ftpcache.keys():
536 if not key in self.ftpcache:
537 self.ftpcache[key] = \
538 ftpwrapper(user, passwd, host, port, dirs)
539 if not file: type = 'D'
542 attr, value = splitvalue(attr)
543 if attr.lower() == 'type' and \
544 value in ('a', 'A', 'i', 'I', 'd', 'D'):
546 (fp, retrlen) = self.ftpcache[key].retrfile(file, type)
547 mtype = mimetypes.guess_type("ftp:" + url)[0]
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]
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')
564 # syntax of data URLs:
565 # dataurl := "data:" [ mediatype ] [ ";base64" ] "," data
566 # mediatype := [ type "/" subtype ] *( ";" parameter )
568 # parameter := attribute "=" value
571 from cStringIO import StringIO
573 from StringIO import StringIO
575 [type, data] = url.split(',', 1)
577 raise IOError, ('data error', 'bad data URL')
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:]
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':
592 data = base64.decodestring(data)
595 msg.append('Content-Length: %d' % len(data))
600 headers = mimetools.Message(f, 0)
601 #f.fileno = None # needed for addinfourl
602 return addinfourl(f, headers, url)
605 class FancyURLopener(URLopener):
606 """Derived class with handlers for errors we can handle (perhaps)."""
608 def __init__(self, *args, **kwargs):
609 URLopener.__init__(self, *args, **kwargs)
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)
618 def http_error_302(self, url, fp, errcode, errmsg, headers, data=None):
619 """Error 302 -- relocated (temporarily)."""
621 if self.maxtries and self.tries >= self.maxtries:
622 if hasattr(self, "http_error_500"):
623 meth = self.http_error_500
625 meth = self.http_error_default
627 return meth(url, fp, 500,
628 "Internal Server Error: Redirect Recursion", headers)
629 result = self.redirect_internal(url, fp, errcode, errmsg, headers,
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']
643 # In case the server sent a relative URL, join with original:
644 newurl = basejoin(self.type + ":" + url, newurl)
645 return self.open(newurl)
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)
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)
655 def http_error_307(self, url, fp, errcode, errmsg, headers, data=None):
656 """Error 307 -- relocated, but turn POST into error."""
658 return self.http_error_302(url, fp, errcode, errmsg, headers, data)
660 return self.http_error_default(url, fp, errcode, errmsg, headers)
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']
670 match = re.match('[ \t]*([^ \t]+)[ \t]+realm="([^"]*)"', stuff)
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'
680 return getattr(self,name)(url, realm)
682 return getattr(self,name)(url, realm, data)
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']
692 match = re.match('[ \t]*([^ \t]+)[ \t]+realm="([^"]*)"', stuff)
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'
702 return getattr(self,name)(url, realm)
704 return getattr(self,name)(url, realm, data)
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
719 return self.open(newurl)
721 return self.open(newurl, data)
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
736 return self.open(newurl)
738 return self.open(newurl, data)
740 def retry_http_basic_auth(self, url, realm, data=None):
741 host, selector = splithost(url)
742 i = host.find('@') + 1
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
749 return self.open(newurl)
751 return self.open(newurl, data)
753 def retry_https_basic_auth(self, url, realm, data=None):
754 host, selector = splithost(url)
755 i = host.find('@') + 1
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
762 return self.open(newurl)
764 return self.open(newurl, data)
766 def get_user_passwd(self, host, realm, clear_cache = 0):
767 key = realm + '@' + host.lower()
768 if key in self.auth_cache:
770 del self.auth_cache[key]
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)
777 def prompt_user_passwd(self, host, realm):
778 """Override this in a GUI environment!"""
781 user = raw_input("Enter username for %s at %s: " % (realm,
783 passwd = getpass.getpass("Enter password for %s in %s at %s: " %
786 except KeyboardInterrupt:
795 """Return the IP address of the magic hostname 'localhost'."""
797 if _localhost is None:
798 _localhost = socket.gethostbyname('localhost')
803 """Return the IP address of the current host."""
805 if _thishost is None:
806 _thishost = socket.gethostbyname(socket.gethostname())
811 """Return the set of errors raised by the FTP class."""
813 if _ftperrors is None:
815 _ftperrors = ftplib.all_errors
820 """Return an empty mimetools.Message object."""
822 if _noheaders is None:
825 from cStringIO import StringIO
827 from StringIO import StringIO
828 _noheaders = mimetools.Message(StringIO(), 0)
829 _noheaders.fp.close() # Recycle file descriptor
836 """Class used by open_ftp() for cache of open FTP connections."""
838 def __init__(self, user, passwd, host, port, dirs):
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:
855 def retrfile(self, file, type):
858 if type in ('d', 'D'): cmd = 'TYPE A'; isdir = 1
859 else: cmd = 'TYPE ' + type; isdir = 0
861 self.ftp.voidcmd(cmd)
862 except ftplib.all_errors:
864 self.ftp.voidcmd(cmd)
866 if file and not isdir:
867 # Try to retrieve as a 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]
875 # Set transfer mode to ASCII!
876 self.ftp.voidcmd('TYPE A')
877 # Try a directory listing
878 if file: cmd = 'LIST ' + file
880 conn = self.ftp.ntransfercmd(cmd)
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):
902 """Base class for addinfo and addclosehook."""
904 def __init__(self, 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
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
919 return '<%s at %r whose fp = %r>' % (self.__class__.__name__,
925 self.readlines = None
927 if self.fp: self.fp.close()
930 class addclosehook(addbase):
931 """Class to add a close hook to an open file."""
933 def __init__(self, fp, closehook, *hookargs):
934 addbase.__init__(self, fp)
935 self.closehook = closehook
936 self.hookargs = hookargs
941 self.closehook(*self.hookargs)
942 self.closehook = None
945 class addinfo(addbase):
946 """class to add an info() method to an open file."""
948 def __init__(self, fp, headers):
949 addbase.__init__(self, fp)
950 self.headers = headers
955 class addinfourl(addbase):
956 """class to add info() and geturl() methods to an open file."""
958 def __init__(self, fp, headers, url):
959 addbase.__init__(self, fp)
960 self.headers = headers
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')
993 return isinstance(x, unicode)
996 """toBytes(u"URL") --> 'URL'."""
997 # Most URL schemes require ASCII. If that changes, the conversion
1001 url = url.encode("ASCII")
1002 except UnicodeError:
1003 raise UnicodeError("URL " + repr(url) +
1004 " contains non-ASCII characters")
1008 """unwrap('<URL:type://host/path>') --> 'type://host/path'."""
1010 if url[:1] == '<' and url[-1:] == '>':
1011 url = url[1:-1].strip()
1012 if url[:4] == 'URL:': url = url[4:].strip()
1017 """splittype('type:opaquestring') --> 'type', 'opaquestring'."""
1019 if _typeprog is None:
1021 _typeprog = re.compile('^([^/:]+):')
1023 match = _typeprog.match(url)
1025 scheme = match.group(1)
1026 return scheme.lower(), url[len(scheme) + 1:]
1031 """splithost('//host[:port]/path') --> 'host[:port]', '/path'."""
1033 if _hostprog is None:
1035 _hostprog = re.compile('^//([^/?]*)(.*)$')
1037 match = _hostprog.match(url)
1038 if match: return match.group(1, 2)
1042 def splituser(host):
1043 """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'."""
1045 if _userprog is None:
1047 _userprog = re.compile('^(.*)@(.*)$')
1049 match = _userprog.match(host)
1050 if match: return map(unquote, match.group(1, 2))
1054 def splitpasswd(user):
1055 """splitpasswd('user:passwd') -> 'user', 'passwd'."""
1057 if _passwdprog is None:
1059 _passwdprog = re.compile('^([^:]*):(.*)$')
1061 match = _passwdprog.match(user)
1062 if match: return match.group(1, 2)
1065 # splittag('/path#tag') --> '/path', 'tag'
1067 def splitport(host):
1068 """splitport('host:port') --> 'host', 'port'."""
1070 if _portprog is None:
1072 _portprog = re.compile('^(.*):([0-9]+)$')
1074 match = _portprog.match(host)
1075 if match: return match.group(1, 2)
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."""
1085 if _nportprog is None:
1087 _nportprog = re.compile('^(.*):(.*)$')
1089 match = _nportprog.match(host)
1091 host, port = match.group(1, 2)
1093 if not port: raise ValueError, "no digits"
1098 return host, defport
1101 def splitquery(url):
1102 """splitquery('/path?query') --> '/path', 'query'."""
1104 if _queryprog is None:
1106 _queryprog = re.compile('^(.*)\?([^?]*)$')
1108 match = _queryprog.match(url)
1109 if match: return match.group(1, 2)
1114 """splittag('/path#tag') --> '/path', 'tag'."""
1116 if _tagprog is None:
1118 _tagprog = re.compile('^(.*)#([^#]*)$')
1120 match = _tagprog.match(url)
1121 if match: return match.group(1, 2)
1125 """splitattr('/path;attr1=value1;attr2=value2;...') ->
1126 '/path', ['attr1=value1', 'attr2=value2', ...]."""
1127 words = url.split(';')
1128 return words[0], words[1:]
1131 def splitvalue(attr):
1132 """splitvalue('attr=value') --> 'attr', 'value'."""
1134 if _valueprog is None:
1136 _valueprog = re.compile('^([^=]*)=(.*)$')
1138 match = _valueprog.match(attr)
1139 if match: return match.group(1, 2)
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
1148 _hextochr = dict(('%02x' % i, chr(i)) for i in range(256))
1149 _hextochr.update(('%02X' % i, chr(i)) for i in range(256))
1152 """unquote('abc%20def') -> 'abc def'."""
1154 for i in xrange(1, len(res)):
1157 res[i] = _hextochr[item[:2]] + item[2:]
1160 except UnicodeDecodeError:
1161 res[i] = unichr(int(item[:2], 16)) + item[2:]
1164 def unquote_plus(s):
1165 """unquote('%7e/abc+def') -> '~/abc def'"""
1166 s = s.replace('+', ' ')
1169 always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
1170 'abcdefghijklmnopqrstuvwxyz'
1174 def quote(s, safe = '/'):
1175 """quote('abc def') -> 'abc%20def'
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.
1180 RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
1181 the following reserved characters.
1183 reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
1186 Each of these characters is reserved in some component of a URL,
1187 but not necessarily in all of them.
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.
1195 cachekey = (safe, always_safe)
1197 safe_map = _safemaps[cachekey]
1201 for i in range(256):
1203 safe_map[c] = (c in safe) and c or ('%%%02X' % i)
1204 _safemaps[cachekey] = safe_map
1205 res = map(safe_map.__getitem__, s)
1208 def quote_plus(s, safe = ''):
1209 """Quote the query fragment of a URL; replacing ' ' with '+'"""
1211 s = quote(s, safe + ' ')
1212 return s.replace(' ', '+')
1213 return quote(s, safe)
1215 def urlencode(query,doseq=0):
1216 """Encode a sequence of two-element tuples or dictionary into a URL query string.
1218 If any values in the query arg are sequences and doseq is true, each
1219 sequence element is converted to a separate parameter.
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
1226 if hasattr(query,"items"):
1228 query = query.items()
1230 # it's a bother at times that strings and string-like objects are
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):
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
1242 ty,va,tb = sys.exc_info()
1243 raise TypeError, "not a valid non-string sequence or mapping object", tb
1247 # preserve old behavior
1249 k = quote_plus(str(k))
1250 v = quote_plus(str(v))
1251 l.append(k + '=' + v)
1254 k = quote_plus(str(k))
1255 if isinstance(v, str):
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)
1266 # is this a sufficient test for sequence-ness?
1270 v = quote_plus(str(v))
1271 l.append(k + '=' + v)
1273 # loop over the sequence
1275 l.append(k + '=' + quote_plus(str(elt)))
1279 def getproxies_environment():
1280 """Return a dictionary of scheme -> proxy server URL mappings.
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.
1289 for name, value in os.environ.items():
1291 if value and name[-6:] == '_proxy':
1292 proxies[name[:-6]] = value
1295 if sys.platform == 'darwin':
1296 def getproxies_internetconfig():
1297 """Return a dictionary of scheme -> proxy server URL mappings.
1299 By convention the mac uses Internet Config to store
1300 proxies. An HTTP proxy, for instance, is stored under
1315 if 'UseHTTPProxy' in config and config['UseHTTPProxy']:
1317 value = config['HTTPProxyHost']
1321 proxies['http'] = 'http://%s' % value
1322 # FTP: XXXX To be done.
1323 # Gopher: XXXX To be done.
1326 def proxy_bypass(x):
1330 return getproxies_environment() or getproxies_internetconfig()
1332 elif os.name == 'nt':
1333 def getproxies_registry():
1334 """Return a dictionary of scheme -> proxy server URL mappings.
1336 Win32 uses the registry to store proxies.
1343 # Std module, so should be around - but you never know!
1346 internetSettings = _winreg.OpenKey(_winreg.HKEY_CURRENT_USER,
1347 r'Software\Microsoft\Windows\CurrentVersion\Internet Settings')
1348 proxyEnable = _winreg.QueryValueEx(internetSettings,
1351 # Returned as Unicode but problems if not converted to ASCII
1352 proxyServer = str(_winreg.QueryValueEx(internetSettings,
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
1360 if not re.match('^([^/:]+)://', address):
1361 address = '%s://%s' % (protocol, address)
1362 proxies[protocol] = address
1364 # Use one setting for all protocols
1365 if proxyServer[:5] == 'http:':
1366 proxies['http'] = proxyServer
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
1379 """Return a dictionary of scheme -> proxy server URL mappings.
1381 Returns settings gathered from the environment, if specified,
1385 return getproxies_environment() or getproxies_registry()
1387 def proxy_bypass(host):
1392 # Std modules, so should be around - but you never know!
1395 internetSettings = _winreg.OpenKey(_winreg.HKEY_CURRENT_USER,
1396 r'Software\Microsoft\Windows\CurrentVersion\Internet Settings')
1397 proxyEnable = _winreg.QueryValueEx(internetSettings,
1399 proxyOverride = str(_winreg.QueryValueEx(internetSettings,
1400 'ProxyOverride')[0])
1401 # ^^^^ Returned as Unicode but problems if not converted to ASCII
1402 except WindowsError:
1404 if not proxyEnable or not proxyOverride:
1406 # try to make a host list from name and IP address.
1407 rawHost, port = splitport(host)
1410 addr = socket.gethostbyname(rawHost)
1413 except socket.error:
1416 fqdn = socket.getfqdn(rawHost)
1419 except socket.error:
1421 # make a check value list from the registry entry: replace the
1422 # '<local>' string by the localhost entry and the corresponding
1424 proxyOverride = proxyOverride.split(';')
1426 while i < len(proxyOverride):
1427 if proxyOverride[i] == '<local>':
1428 proxyOverride[i:i+1] = ['localhost',
1430 socket.gethostname(),
1431 socket.gethostbyname(
1432 socket.gethostname())]
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
1441 # print "%s <--> %s" %( test, val )
1442 if re.match(test, val, re.I):
1447 # By default use environment variables
1448 getproxies = getproxies_environment
1450 def proxy_bypass(host):
1453 # Test and time quote() and unquote()
1456 for i in range(256): s = s + chr(i)
1467 print round(t1 - t0, 3), 'sec'
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)
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',
1486 if hasattr(URLopener, "open_https"):
1487 args.append('https://synergy.as.cmu.edu/~geek/')
1490 print '-'*10, url, '-'*10
1491 fn, h = urlretrieve(url, None, reporthook)
1495 for k in h.keys(): print k + ':', h[k]
1501 table = string.maketrans("", "")
1502 data = data.translate(table, "\r")
1512 opts, args = getopt.getopt(sys.argv[1:], "th")
1513 except getopt.error, msg:
1515 print "Use -h for help"
1522 print "Usage: python urllib.py [-t] [url ...]"
1523 print "-t runs self-test;",
1524 print "otherwise, contents of urls are printed"
1532 print "Use -h for help"
1534 print urlopen(url).read(),
1536 # Run test program when run as a script
1537 if __name__ == '__main__':