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
33 __all__
= ["urlopen", "URLopener", "FancyURLopener", "urlretrieve",
34 "urlcleanup", "quote", "quote_plus", "unquote", "unquote_plus",
35 "urlencode", "url2pathname", "pathname2url", "splittag",
36 "localhost", "thishost", "ftperrors", "basejoin", "unwrap",
37 "splittype", "splithost", "splituser", "splitpasswd", "splitport",
38 "splitnport", "splitquery", "splitattr", "splitvalue",
41 __version__
= '1.17' # XXX This version is not always updated :-(
43 MAXFTPCACHE
= 10 # Trim the ftp cache beyond this size
45 # Helper for non-unix systems
47 from macurl2path
import url2pathname
, pathname2url
49 from nturl2path
import url2pathname
, pathname2url
50 elif os
.name
== 'riscos':
51 from rourl2path
import url2pathname
, pathname2url
53 def url2pathname(pathname
):
54 """OS-specific conversion from a relative URL of the 'file' scheme
55 to a file system path; not recommended for general use."""
56 return unquote(pathname
)
58 def pathname2url(pathname
):
59 """OS-specific conversion from a file system path to a relative URL
60 of the 'file' scheme; not recommended for general use."""
61 return quote(pathname
)
63 # This really consists of two pieces:
64 # (1) a class which handles opening of all sorts of URLs
65 # (plus assorted utilities etc.)
66 # (2) a set of functions for parsing URLs
67 # XXX Should these be separated out into different modules?
70 # Shortcut for basic usage
72 def urlopen(url
, data
=None, proxies
=None):
73 """Create a file-like object for the specified URL to read from."""
74 from warnings
import warnpy3k
75 warnings
.warnpy3k("urllib.urlopen() has been removed in Python 3.0 in "
76 "favor of urllib2.urlopen()", stacklevel
=2)
79 if proxies
is not None:
80 opener
= FancyURLopener(proxies
=proxies
)
82 opener
= FancyURLopener()
87 return opener
.open(url
)
89 return opener
.open(url
, data
)
90 def urlretrieve(url
, filename
=None, reporthook
=None, data
=None):
93 _urlopener
= FancyURLopener()
94 return _urlopener
.retrieve(url
, filename
, reporthook
, data
)
107 # exception raised when downloaded size does not match content-length
108 class ContentTooShortError(IOError):
109 def __init__(self
, message
, content
):
110 IOError.__init
__(self
, message
)
111 self
.content
= content
115 """Class to open URLs.
116 This is a class rather than just a subroutine because we may need
117 more than one set of global protocol-specific options.
118 Note -- this is a base class for those who don't want the
119 automatic handling of errors type 302 (relocated) and 401
120 (authorization needed)."""
124 version
= "Python-urllib/%s" % __version__
127 def __init__(self
, proxies
=None, **x509
):
129 proxies
= getproxies()
130 assert hasattr(proxies
, 'has_key'), "proxies must be a mapping"
131 self
.proxies
= proxies
132 self
.key_file
= x509
.get('key_file')
133 self
.cert_file
= x509
.get('cert_file')
134 self
.addheaders
= [('User-Agent', self
.version
)]
135 self
.__tempfiles
= []
136 self
.__unlink
= os
.unlink
# See cleanup()
137 self
.tempcache
= None
138 # Undocumented feature: if you assign {} to tempcache,
139 # it is used to cache files retrieved with
140 # self.retrieve(). This is not enabled by default
141 # since it does not work for changing documents (and I
142 # haven't got the logic to check expiration headers
144 self
.ftpcache
= ftpcache
145 # Undocumented feature: you can use a different
146 # ftp cache by assigning to the .ftpcache member;
147 # in case you want logically independent URL openers
148 # XXX This is not threadsafe. Bah.
157 # This code sometimes runs when the rest of this module
158 # has already been deleted, so it can't use any globals
159 # or import anything.
161 for file in self
.__tempfiles
:
166 del self
.__tempfiles
[:]
168 self
.tempcache
.clear()
170 def addheader(self
, *args
):
171 """Add a header to be used by the HTTP interface only
172 e.g. u.addheader('Accept', 'sound/basic')"""
173 self
.addheaders
.append(args
)
176 def open(self
, fullurl
, data
=None):
177 """Use URLopener().open(file) instead of open(file, 'r')."""
178 fullurl
= unwrap(toBytes(fullurl
))
179 # percent encode url, fixing lame server errors for e.g, like space
181 fullurl
= quote(fullurl
, safe
="%/:=&?~#+!$,;'@()*[]")
182 if self
.tempcache
and fullurl
in self
.tempcache
:
183 filename
, headers
= self
.tempcache
[fullurl
]
184 fp
= open(filename
, 'rb')
185 return addinfourl(fp
, headers
, fullurl
)
186 urltype
, url
= splittype(fullurl
)
189 if urltype
in self
.proxies
:
190 proxy
= self
.proxies
[urltype
]
191 urltype
, proxyhost
= splittype(proxy
)
192 host
, selector
= splithost(proxyhost
)
193 url
= (host
, fullurl
) # Signal special case to open_*()
196 name
= 'open_' + urltype
198 name
= name
.replace('-', '_')
199 if not hasattr(self
, name
):
201 return self
.open_unknown_proxy(proxy
, fullurl
, data
)
203 return self
.open_unknown(fullurl
, data
)
206 return getattr(self
, name
)(url
)
208 return getattr(self
, name
)(url
, data
)
209 except socket
.error
, msg
:
210 raise IOError, ('socket error', msg
), sys
.exc_info()[2]
212 def open_unknown(self
, fullurl
, data
=None):
213 """Overridable interface to open unknown URL type."""
214 type, url
= splittype(fullurl
)
215 raise IOError, ('url error', 'unknown url type', type)
217 def open_unknown_proxy(self
, proxy
, fullurl
, data
=None):
218 """Overridable interface to open unknown URL type."""
219 type, url
= splittype(fullurl
)
220 raise IOError, ('url error', 'invalid proxy for %s' % type, proxy
)
223 def retrieve(self
, url
, filename
=None, reporthook
=None, data
=None):
224 """retrieve(url) returns (filename, headers) for a local object
225 or (tempfilename, headers) for a remote object."""
226 url
= unwrap(toBytes(url
))
227 if self
.tempcache
and url
in self
.tempcache
:
228 return self
.tempcache
[url
]
229 type, url1
= splittype(url
)
230 if filename
is None and (not type or type == 'file'):
232 fp
= self
.open_local_file(url1
)
235 return url2pathname(splithost(url1
)[1]), hdrs
238 fp
= self
.open(url
, data
)
242 tfp
= open(filename
, 'wb')
245 garbage
, path
= splittype(url
)
246 garbage
, path
= splithost(path
or "")
247 path
, garbage
= splitquery(path
or "")
248 path
, garbage
= splitattr(path
or "")
249 suffix
= os
.path
.splitext(path
)[1]
250 (fd
, filename
) = tempfile
.mkstemp(suffix
)
251 self
.__tempfiles
.append(filename
)
252 tfp
= os
.fdopen(fd
, 'wb')
254 result
= filename
, headers
255 if self
.tempcache
is not None:
256 self
.tempcache
[url
] = result
262 if "content-length" in headers
:
263 size
= int(headers
["Content-Length"])
264 reporthook(blocknum
, bs
, size
)
273 reporthook(blocknum
, bs
, size
)
279 # raise exception if actual size does not match content-length header
280 if size
>= 0 and read
< size
:
281 raise ContentTooShortError("retrieval incomplete: got only %i out "
282 "of %i bytes" % (read
, size
), result
)
286 # Each method named open_<type> knows how to open that type of URL
288 def open_http(self
, url
, data
=None):
289 """Use HTTP protocol."""
293 if isinstance(url
, str):
294 host
, selector
= splithost(url
)
296 user_passwd
, host
= splituser(host
)
301 # check whether the proxy contains authorization information
302 proxy_passwd
, host
= splituser(host
)
303 # now we proceed with the url we want to obtain
304 urltype
, rest
= splittype(selector
)
307 if urltype
.lower() != 'http':
310 realhost
, rest
= splithost(rest
)
312 user_passwd
, realhost
= splituser(realhost
)
314 selector
= "%s://%s%s" % (urltype
, realhost
, rest
)
315 if proxy_bypass(realhost
):
318 #print "proxy via http:", host, selector
319 if not host
: raise IOError, ('http error', 'no host given')
323 proxy_auth
= base64
.b64encode(proxy_passwd
).strip()
329 auth
= base64
.b64encode(user_passwd
).strip()
332 h
= httplib
.HTTP(host
)
334 h
.putrequest('POST', selector
)
335 h
.putheader('Content-Type', 'application/x-www-form-urlencoded')
336 h
.putheader('Content-Length', '%d' % len(data
))
338 h
.putrequest('GET', selector
)
339 if proxy_auth
: h
.putheader('Proxy-Authorization', 'Basic %s' % proxy_auth
)
340 if auth
: h
.putheader('Authorization', 'Basic %s' % auth
)
341 if realhost
: h
.putheader('Host', realhost
)
342 for args
in self
.addheaders
: h
.putheader(*args
)
344 errcode
, errmsg
, headers
= h
.getreply()
348 # something went wrong with the HTTP status line
349 raise IOError, ('http protocol error', 0,
350 'got a bad status line', None)
351 # According to RFC 2616, "2xx" code indicates that the client's
352 # request was successfully received, understood, and accepted.
353 if (200 <= errcode
< 300):
354 return addinfourl(fp
, headers
, "http:" + url
, errcode
)
357 return self
.http_error(url
, fp
, errcode
, errmsg
, headers
)
359 return self
.http_error(url
, fp
, errcode
, errmsg
, headers
, data
)
361 def http_error(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
362 """Handle http errors.
363 Derived class can override this, or provide specific handlers
364 named http_error_DDD where DDD is the 3-digit error code."""
365 # First check if there's a specific handler for this error
366 name
= 'http_error_%d' % errcode
367 if hasattr(self
, name
):
368 method
= getattr(self
, name
)
370 result
= method(url
, fp
, errcode
, errmsg
, headers
)
372 result
= method(url
, fp
, errcode
, errmsg
, headers
, data
)
373 if result
: return result
374 return self
.http_error_default(url
, fp
, errcode
, errmsg
, headers
)
376 def http_error_default(self
, url
, fp
, errcode
, errmsg
, headers
):
377 """Default error handler: close the connection and raise IOError."""
380 raise IOError, ('http error', errcode
, errmsg
, headers
)
383 def open_https(self
, url
, data
=None):
384 """Use HTTPS protocol."""
389 if isinstance(url
, str):
390 host
, selector
= splithost(url
)
392 user_passwd
, host
= splituser(host
)
397 # here, we determine, whether the proxy contains authorization information
398 proxy_passwd
, host
= splituser(host
)
399 urltype
, rest
= splittype(selector
)
402 if urltype
.lower() != 'https':
405 realhost
, rest
= splithost(rest
)
407 user_passwd
, realhost
= splituser(realhost
)
409 selector
= "%s://%s%s" % (urltype
, realhost
, rest
)
410 #print "proxy via https:", host, selector
411 if not host
: raise IOError, ('https error', 'no host given')
414 proxy_auth
= base64
.b64encode(proxy_passwd
).strip()
419 auth
= base64
.b64encode(user_passwd
).strip()
422 h
= httplib
.HTTPS(host
, 0,
423 key_file
=self
.key_file
,
424 cert_file
=self
.cert_file
)
426 h
.putrequest('POST', selector
)
427 h
.putheader('Content-Type',
428 'application/x-www-form-urlencoded')
429 h
.putheader('Content-Length', '%d' % len(data
))
431 h
.putrequest('GET', selector
)
432 if proxy_auth
: h
.putheader('Proxy-Authorization', 'Basic %s' % proxy_auth
)
433 if auth
: h
.putheader('Authorization', 'Basic %s' % auth
)
434 if realhost
: h
.putheader('Host', realhost
)
435 for args
in self
.addheaders
: h
.putheader(*args
)
437 errcode
, errmsg
, headers
= h
.getreply()
441 # something went wrong with the HTTP status line
442 raise IOError, ('http protocol error', 0,
443 'got a bad status line', None)
444 # According to RFC 2616, "2xx" code indicates that the client's
445 # request was successfully received, understood, and accepted.
446 if (200 <= errcode
< 300):
447 return addinfourl(fp
, headers
, "https:" + url
, errcode
)
450 return self
.http_error(url
, fp
, errcode
, errmsg
, headers
)
452 return self
.http_error(url
, fp
, errcode
, errmsg
, headers
,
455 def open_file(self
, url
):
456 """Use local file or FTP depending on form of URL."""
457 if not isinstance(url
, str):
458 raise IOError, ('file error', 'proxy support for file protocol currently not implemented')
459 if url
[:2] == '//' and url
[2:3] != '/' and url
[2:12].lower() != 'localhost/':
460 return self
.open_ftp(url
)
462 return self
.open_local_file(url
)
464 def open_local_file(self
, url
):
465 """Use local file."""
466 import mimetypes
, mimetools
, email
.utils
468 from cStringIO
import StringIO
470 from StringIO
import StringIO
471 host
, file = splithost(url
)
472 localname
= url2pathname(file)
474 stats
= os
.stat(localname
)
476 raise IOError(e
.errno
, e
.strerror
, e
.filename
)
478 modified
= email
.utils
.formatdate(stats
.st_mtime
, usegmt
=True)
479 mtype
= mimetypes
.guess_type(url
)[0]
480 headers
= mimetools
.Message(StringIO(
481 'Content-Type: %s\nContent-Length: %d\nLast-modified: %s\n' %
482 (mtype
or 'text/plain', size
, modified
)))
486 urlfile
= 'file://' + file
487 return addinfourl(open(localname
, 'rb'),
489 host
, port
= splitport(host
)
491 and socket
.gethostbyname(host
) in (localhost(), thishost()):
494 urlfile
= 'file://' + file
495 return addinfourl(open(localname
, 'rb'),
497 raise IOError, ('local file error', 'not on local host')
499 def open_ftp(self
, url
):
500 """Use FTP protocol."""
501 if not isinstance(url
, str):
502 raise IOError, ('ftp error', 'proxy support for ftp protocol currently not implemented')
503 import mimetypes
, mimetools
505 from cStringIO
import StringIO
507 from StringIO
import StringIO
508 host
, path
= splithost(url
)
509 if not host
: raise IOError, ('ftp error', 'no host given')
510 host
, port
= splitport(host
)
511 user
, host
= splituser(host
)
512 if user
: user
, passwd
= splitpasswd(user
)
515 user
= unquote(user
or '')
516 passwd
= unquote(passwd
or '')
517 host
= socket
.gethostbyname(host
)
520 port
= ftplib
.FTP_PORT
523 path
, attrs
= splitattr(path
)
525 dirs
= path
.split('/')
526 dirs
, file = dirs
[:-1], dirs
[-1]
527 if dirs
and not dirs
[0]: dirs
= dirs
[1:]
528 if dirs
and not dirs
[0]: dirs
[0] = '/'
529 key
= user
, host
, port
, '/'.join(dirs
)
531 if len(self
.ftpcache
) > MAXFTPCACHE
:
532 # Prune the cache, rather arbitrarily
533 for k
in self
.ftpcache
.keys():
539 if not key
in self
.ftpcache
:
540 self
.ftpcache
[key
] = \
541 ftpwrapper(user
, passwd
, host
, port
, dirs
)
542 if not file: type = 'D'
545 attr
, value
= splitvalue(attr
)
546 if attr
.lower() == 'type' and \
547 value
in ('a', 'A', 'i', 'I', 'd', 'D'):
549 (fp
, retrlen
) = self
.ftpcache
[key
].retrfile(file, type)
550 mtype
= mimetypes
.guess_type("ftp:" + url
)[0]
553 headers
+= "Content-Type: %s\n" % mtype
554 if retrlen
is not None and retrlen
>= 0:
555 headers
+= "Content-Length: %d\n" % retrlen
556 headers
= mimetools
.Message(StringIO(headers
))
557 return addinfourl(fp
, headers
, "ftp:" + url
)
558 except ftperrors(), msg
:
559 raise IOError, ('ftp error', msg
), sys
.exc_info()[2]
561 def open_data(self
, url
, data
=None):
562 """Use "data" URL."""
563 if not isinstance(url
, str):
564 raise IOError, ('data error', 'proxy support for data protocol currently not implemented')
567 # syntax of data URLs:
568 # dataurl := "data:" [ mediatype ] [ ";base64" ] "," data
569 # mediatype := [ type "/" subtype ] *( ";" parameter )
571 # parameter := attribute "=" value
574 from cStringIO
import StringIO
576 from StringIO
import StringIO
578 [type, data
] = url
.split(',', 1)
580 raise IOError, ('data error', 'bad data URL')
582 type = 'text/plain;charset=US-ASCII'
583 semi
= type.rfind(';')
584 if semi
>= 0 and '=' not in type[semi
:]:
585 encoding
= type[semi
+1:]
590 msg
.append('Date: %s'%time
.strftime('%a, %d %b %Y %T GMT',
591 time
.gmtime(time
.time())))
592 msg
.append('Content-type: %s' % type)
593 if encoding
== 'base64':
595 data
= base64
.decodestring(data
)
598 msg
.append('Content-Length: %d' % len(data
))
603 headers
= mimetools
.Message(f
, 0)
604 #f.fileno = None # needed for addinfourl
605 return addinfourl(f
, headers
, url
)
608 class FancyURLopener(URLopener
):
609 """Derived class with handlers for errors we can handle (perhaps)."""
611 def __init__(self
, *args
, **kwargs
):
612 URLopener
.__init
__(self
, *args
, **kwargs
)
617 def http_error_default(self
, url
, fp
, errcode
, errmsg
, headers
):
618 """Default error handling -- don't raise an exception."""
619 return addinfourl(fp
, headers
, "http:" + url
, errcode
)
621 def http_error_302(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
622 """Error 302 -- relocated (temporarily)."""
624 if self
.maxtries
and self
.tries
>= self
.maxtries
:
625 if hasattr(self
, "http_error_500"):
626 meth
= self
.http_error_500
628 meth
= self
.http_error_default
630 return meth(url
, fp
, 500,
631 "Internal Server Error: Redirect Recursion", headers
)
632 result
= self
.redirect_internal(url
, fp
, errcode
, errmsg
, headers
,
637 def redirect_internal(self
, url
, fp
, errcode
, errmsg
, headers
, data
):
638 if 'location' in headers
:
639 newurl
= headers
['location']
640 elif 'uri' in headers
:
641 newurl
= headers
['uri']
646 # In case the server sent a relative URL, join with original:
647 newurl
= basejoin(self
.type + ":" + url
, newurl
)
648 return self
.open(newurl
)
650 def http_error_301(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
651 """Error 301 -- also relocated (permanently)."""
652 return self
.http_error_302(url
, fp
, errcode
, errmsg
, headers
, data
)
654 def http_error_303(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
655 """Error 303 -- also relocated (essentially identical to 302)."""
656 return self
.http_error_302(url
, fp
, errcode
, errmsg
, headers
, data
)
658 def http_error_307(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
659 """Error 307 -- relocated, but turn POST into error."""
661 return self
.http_error_302(url
, fp
, errcode
, errmsg
, headers
, data
)
663 return self
.http_error_default(url
, fp
, errcode
, errmsg
, headers
)
665 def http_error_401(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
666 """Error 401 -- authentication required.
667 This function supports Basic authentication only."""
668 if not 'www-authenticate' in headers
:
669 URLopener
.http_error_default(self
, url
, fp
,
670 errcode
, errmsg
, headers
)
671 stuff
= headers
['www-authenticate']
673 match
= re
.match('[ \t]*([^ \t]+)[ \t]+realm="([^"]*)"', stuff
)
675 URLopener
.http_error_default(self
, url
, fp
,
676 errcode
, errmsg
, headers
)
677 scheme
, realm
= match
.groups()
678 if scheme
.lower() != 'basic':
679 URLopener
.http_error_default(self
, url
, fp
,
680 errcode
, errmsg
, headers
)
681 name
= 'retry_' + self
.type + '_basic_auth'
683 return getattr(self
,name
)(url
, realm
)
685 return getattr(self
,name
)(url
, realm
, data
)
687 def http_error_407(self
, url
, fp
, errcode
, errmsg
, headers
, data
=None):
688 """Error 407 -- proxy authentication required.
689 This function supports Basic authentication only."""
690 if not 'proxy-authenticate' in headers
:
691 URLopener
.http_error_default(self
, url
, fp
,
692 errcode
, errmsg
, headers
)
693 stuff
= headers
['proxy-authenticate']
695 match
= re
.match('[ \t]*([^ \t]+)[ \t]+realm="([^"]*)"', stuff
)
697 URLopener
.http_error_default(self
, url
, fp
,
698 errcode
, errmsg
, headers
)
699 scheme
, realm
= match
.groups()
700 if scheme
.lower() != 'basic':
701 URLopener
.http_error_default(self
, url
, fp
,
702 errcode
, errmsg
, headers
)
703 name
= 'retry_proxy_' + self
.type + '_basic_auth'
705 return getattr(self
,name
)(url
, realm
)
707 return getattr(self
,name
)(url
, realm
, data
)
709 def retry_proxy_http_basic_auth(self
, url
, realm
, data
=None):
710 host
, selector
= splithost(url
)
711 newurl
= 'http://' + host
+ selector
712 proxy
= self
.proxies
['http']
713 urltype
, proxyhost
= splittype(proxy
)
714 proxyhost
, proxyselector
= splithost(proxyhost
)
715 i
= proxyhost
.find('@') + 1
716 proxyhost
= proxyhost
[i
:]
717 user
, passwd
= self
.get_user_passwd(proxyhost
, realm
, i
)
718 if not (user
or passwd
): return None
719 proxyhost
= quote(user
, safe
='') + ':' + quote(passwd
, safe
='') + '@' + proxyhost
720 self
.proxies
['http'] = 'http://' + proxyhost
+ proxyselector
722 return self
.open(newurl
)
724 return self
.open(newurl
, data
)
726 def retry_proxy_https_basic_auth(self
, url
, realm
, data
=None):
727 host
, selector
= splithost(url
)
728 newurl
= 'https://' + host
+ selector
729 proxy
= self
.proxies
['https']
730 urltype
, proxyhost
= splittype(proxy
)
731 proxyhost
, proxyselector
= splithost(proxyhost
)
732 i
= proxyhost
.find('@') + 1
733 proxyhost
= proxyhost
[i
:]
734 user
, passwd
= self
.get_user_passwd(proxyhost
, realm
, i
)
735 if not (user
or passwd
): return None
736 proxyhost
= quote(user
, safe
='') + ':' + quote(passwd
, safe
='') + '@' + proxyhost
737 self
.proxies
['https'] = 'https://' + proxyhost
+ proxyselector
739 return self
.open(newurl
)
741 return self
.open(newurl
, data
)
743 def retry_http_basic_auth(self
, url
, realm
, data
=None):
744 host
, selector
= splithost(url
)
745 i
= host
.find('@') + 1
747 user
, passwd
= self
.get_user_passwd(host
, realm
, i
)
748 if not (user
or passwd
): return None
749 host
= quote(user
, safe
='') + ':' + quote(passwd
, safe
='') + '@' + host
750 newurl
= 'http://' + host
+ selector
752 return self
.open(newurl
)
754 return self
.open(newurl
, data
)
756 def retry_https_basic_auth(self
, url
, realm
, data
=None):
757 host
, selector
= splithost(url
)
758 i
= host
.find('@') + 1
760 user
, passwd
= self
.get_user_passwd(host
, realm
, i
)
761 if not (user
or passwd
): return None
762 host
= quote(user
, safe
='') + ':' + quote(passwd
, safe
='') + '@' + host
763 newurl
= 'https://' + host
+ selector
765 return self
.open(newurl
)
767 return self
.open(newurl
, data
)
769 def get_user_passwd(self
, host
, realm
, clear_cache
= 0):
770 key
= realm
+ '@' + host
.lower()
771 if key
in self
.auth_cache
:
773 del self
.auth_cache
[key
]
775 return self
.auth_cache
[key
]
776 user
, passwd
= self
.prompt_user_passwd(host
, realm
)
777 if user
or passwd
: self
.auth_cache
[key
] = (user
, passwd
)
780 def prompt_user_passwd(self
, host
, realm
):
781 """Override this in a GUI environment!"""
784 user
= raw_input("Enter username for %s at %s: " % (realm
,
786 passwd
= getpass
.getpass("Enter password for %s in %s at %s: " %
789 except KeyboardInterrupt:
798 """Return the IP address of the magic hostname 'localhost'."""
800 if _localhost
is None:
801 _localhost
= socket
.gethostbyname('localhost')
806 """Return the IP address of the current host."""
808 if _thishost
is None:
809 _thishost
= socket
.gethostbyname(socket
.gethostname())
814 """Return the set of errors raised by the FTP class."""
816 if _ftperrors
is None:
818 _ftperrors
= ftplib
.all_errors
823 """Return an empty mimetools.Message object."""
825 if _noheaders
is None:
828 from cStringIO
import StringIO
830 from StringIO
import StringIO
831 _noheaders
= mimetools
.Message(StringIO(), 0)
832 _noheaders
.fp
.close() # Recycle file descriptor
839 """Class used by open_ftp() for cache of open FTP connections."""
841 def __init__(self
, user
, passwd
, host
, port
, dirs
,
842 timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
):
848 self
.timeout
= timeout
854 self
.ftp
= ftplib
.FTP()
855 self
.ftp
.connect(self
.host
, self
.port
, self
.timeout
)
856 self
.ftp
.login(self
.user
, self
.passwd
)
857 for dir in self
.dirs
:
860 def retrfile(self
, file, type):
863 if type in ('d', 'D'): cmd
= 'TYPE A'; isdir
= 1
864 else: cmd
= 'TYPE ' + type; isdir
= 0
866 self
.ftp
.voidcmd(cmd
)
867 except ftplib
.all_errors
:
869 self
.ftp
.voidcmd(cmd
)
871 if file and not isdir
:
872 # Try to retrieve as a file
875 conn
= self
.ftp
.ntransfercmd(cmd
)
876 except ftplib
.error_perm
, reason
:
877 if str(reason
)[:3] != '550':
878 raise IOError, ('ftp error', reason
), sys
.exc_info()[2]
880 # Set transfer mode to ASCII!
881 self
.ftp
.voidcmd('TYPE A')
882 # Try a directory listing. Verify that directory exists.
888 except ftplib
.error_perm
, reason
:
889 raise IOError, ('ftp error', reason
), sys
.exc_info()[2]
895 conn
= self
.ftp
.ntransfercmd(cmd
)
897 # Pass back both a suitably decorated object and a retrieval length
898 return (addclosehook(conn
[0].makefile('rb'),
899 self
.endtransfer
), conn
[1])
900 def endtransfer(self
):
917 """Base class for addinfo and addclosehook."""
919 def __init__(self
, fp
):
921 self
.read
= self
.fp
.read
922 self
.readline
= self
.fp
.readline
923 if hasattr(self
.fp
, "readlines"): self
.readlines
= self
.fp
.readlines
924 if hasattr(self
.fp
, "fileno"):
925 self
.fileno
= self
.fp
.fileno
927 self
.fileno
= lambda: None
928 if hasattr(self
.fp
, "__iter__"):
929 self
.__iter
__ = self
.fp
.__iter
__
930 if hasattr(self
.fp
, "next"):
931 self
.next
= self
.fp
.next
934 return '<%s at %r whose fp = %r>' % (self
.__class
__.__name
__,
940 self
.readlines
= None
942 if self
.fp
: self
.fp
.close()
945 class addclosehook(addbase
):
946 """Class to add a close hook to an open file."""
948 def __init__(self
, fp
, closehook
, *hookargs
):
949 addbase
.__init
__(self
, fp
)
950 self
.closehook
= closehook
951 self
.hookargs
= hookargs
956 self
.closehook(*self
.hookargs
)
957 self
.closehook
= None
960 class addinfo(addbase
):
961 """class to add an info() method to an open file."""
963 def __init__(self
, fp
, headers
):
964 addbase
.__init
__(self
, fp
)
965 self
.headers
= headers
970 class addinfourl(addbase
):
971 """class to add info() and geturl() methods to an open file."""
973 def __init__(self
, fp
, headers
, url
, code
=None):
974 addbase
.__init
__(self
, fp
)
975 self
.headers
= headers
989 # Utilities to parse URLs (most of these return None for missing parts):
990 # unwrap('<URL:type://host/path>') --> 'type://host/path'
991 # splittype('type:opaquestring') --> 'type', 'opaquestring'
992 # splithost('//host[:port]/path') --> 'host[:port]', '/path'
993 # splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'
994 # splitpasswd('user:passwd') -> 'user', 'passwd'
995 # splitport('host:port') --> 'host', 'port'
996 # splitquery('/path?query') --> '/path', 'query'
997 # splittag('/path#tag') --> '/path', 'tag'
998 # splitattr('/path;attr1=value1;attr2=value2;...') ->
999 # '/path', ['attr1=value1', 'attr2=value2', ...]
1000 # splitvalue('attr=value') --> 'attr', 'value'
1001 # unquote('abc%20def') -> 'abc def'
1002 # quote('abc def') -> 'abc%20def')
1011 return isinstance(x
, unicode)
1014 """toBytes(u"URL") --> 'URL'."""
1015 # Most URL schemes require ASCII. If that changes, the conversion
1017 if _is_unicode(url
):
1019 url
= url
.encode("ASCII")
1020 except UnicodeError:
1021 raise UnicodeError("URL " + repr(url
) +
1022 " contains non-ASCII characters")
1026 """unwrap('<URL:type://host/path>') --> 'type://host/path'."""
1028 if url
[:1] == '<' and url
[-1:] == '>':
1029 url
= url
[1:-1].strip()
1030 if url
[:4] == 'URL:': url
= url
[4:].strip()
1035 """splittype('type:opaquestring') --> 'type', 'opaquestring'."""
1037 if _typeprog
is None:
1039 _typeprog
= re
.compile('^([^/:]+):')
1041 match
= _typeprog
.match(url
)
1043 scheme
= match
.group(1)
1044 return scheme
.lower(), url
[len(scheme
) + 1:]
1049 """splithost('//host[:port]/path') --> 'host[:port]', '/path'."""
1051 if _hostprog
is None:
1053 _hostprog
= re
.compile('^//([^/?]*)(.*)$')
1055 match
= _hostprog
.match(url
)
1056 if match
: return match
.group(1, 2)
1060 def splituser(host
):
1061 """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'."""
1063 if _userprog
is None:
1065 _userprog
= re
.compile('^(.*)@(.*)$')
1067 match
= _userprog
.match(host
)
1068 if match
: return map(unquote
, match
.group(1, 2))
1072 def splitpasswd(user
):
1073 """splitpasswd('user:passwd') -> 'user', 'passwd'."""
1075 if _passwdprog
is None:
1077 _passwdprog
= re
.compile('^([^:]*):(.*)$',re
.S
)
1079 match
= _passwdprog
.match(user
)
1080 if match
: return match
.group(1, 2)
1083 # splittag('/path#tag') --> '/path', 'tag'
1085 def splitport(host
):
1086 """splitport('host:port') --> 'host', 'port'."""
1088 if _portprog
is None:
1090 _portprog
= re
.compile('^(.*):([0-9]+)$')
1092 match
= _portprog
.match(host
)
1093 if match
: return match
.group(1, 2)
1097 def splitnport(host
, defport
=-1):
1098 """Split host and port, returning numeric port.
1099 Return given default port if no ':' found; defaults to -1.
1100 Return numerical port if a valid number are found after ':'.
1101 Return None if ':' but not a valid number."""
1103 if _nportprog
is None:
1105 _nportprog
= re
.compile('^(.*):(.*)$')
1107 match
= _nportprog
.match(host
)
1109 host
, port
= match
.group(1, 2)
1111 if not port
: raise ValueError, "no digits"
1116 return host
, defport
1119 def splitquery(url
):
1120 """splitquery('/path?query') --> '/path', 'query'."""
1122 if _queryprog
is None:
1124 _queryprog
= re
.compile('^(.*)\?([^?]*)$')
1126 match
= _queryprog
.match(url
)
1127 if match
: return match
.group(1, 2)
1132 """splittag('/path#tag') --> '/path', 'tag'."""
1134 if _tagprog
is None:
1136 _tagprog
= re
.compile('^(.*)#([^#]*)$')
1138 match
= _tagprog
.match(url
)
1139 if match
: return match
.group(1, 2)
1143 """splitattr('/path;attr1=value1;attr2=value2;...') ->
1144 '/path', ['attr1=value1', 'attr2=value2', ...]."""
1145 words
= url
.split(';')
1146 return words
[0], words
[1:]
1149 def splitvalue(attr
):
1150 """splitvalue('attr=value') --> 'attr', 'value'."""
1152 if _valueprog
is None:
1154 _valueprog
= re
.compile('^([^=]*)=(.*)$')
1156 match
= _valueprog
.match(attr
)
1157 if match
: return match
.group(1, 2)
1160 _hextochr
= dict(('%02x' % i
, chr(i
)) for i
in range(256))
1161 _hextochr
.update(('%02X' % i
, chr(i
)) for i
in range(256))
1164 """unquote('abc%20def') -> 'abc def'."""
1166 for i
in xrange(1, len(res
)):
1169 res
[i
] = _hextochr
[item
[:2]] + item
[2:]
1172 except UnicodeDecodeError:
1173 res
[i
] = unichr(int(item
[:2], 16)) + item
[2:]
1176 def unquote_plus(s
):
1177 """unquote('%7e/abc+def') -> '~/abc def'"""
1178 s
= s
.replace('+', ' ')
1181 always_safe
= ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
1182 'abcdefghijklmnopqrstuvwxyz'
1186 def quote(s
, safe
= '/'):
1187 """quote('abc def') -> 'abc%20def'
1189 Each part of a URL, e.g. the path info, the query, etc., has a
1190 different set of reserved characters that must be quoted.
1192 RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
1193 the following reserved characters.
1195 reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
1198 Each of these characters is reserved in some component of a URL,
1199 but not necessarily in all of them.
1201 By default, the quote function is intended for quoting the path
1202 section of a URL. Thus, it will not encode '/'. This character
1203 is reserved, but in typical usage the quote function is being
1204 called on a path where the existing slash characters are used as
1205 reserved characters.
1207 cachekey
= (safe
, always_safe
)
1209 safe_map
= _safemaps
[cachekey
]
1213 for i
in range(256):
1215 safe_map
[c
] = (c
in safe
) and c
or ('%%%02X' % i
)
1216 _safemaps
[cachekey
] = safe_map
1217 res
= map(safe_map
.__getitem
__, s
)
1220 def quote_plus(s
, safe
= ''):
1221 """Quote the query fragment of a URL; replacing ' ' with '+'"""
1223 s
= quote(s
, safe
+ ' ')
1224 return s
.replace(' ', '+')
1225 return quote(s
, safe
)
1227 def urlencode(query
,doseq
=0):
1228 """Encode a sequence of two-element tuples or dictionary into a URL query string.
1230 If any values in the query arg are sequences and doseq is true, each
1231 sequence element is converted to a separate parameter.
1233 If the query arg is a sequence of two-element tuples, the order of the
1234 parameters in the output will match the order of parameters in the
1238 if hasattr(query
,"items"):
1240 query
= query
.items()
1242 # it's a bother at times that strings and string-like objects are
1245 # non-sequence items should not work with len()
1246 # non-empty strings will fail this
1247 if len(query
) and not isinstance(query
[0], tuple):
1249 # zero-length sequences of all types will get here and succeed,
1250 # but that's a minor nit - since the original implementation
1251 # allowed empty dicts that type of behavior probably should be
1252 # preserved for consistency
1254 ty
,va
,tb
= sys
.exc_info()
1255 raise TypeError, "not a valid non-string sequence or mapping object", tb
1259 # preserve old behavior
1261 k
= quote_plus(str(k
))
1262 v
= quote_plus(str(v
))
1263 l
.append(k
+ '=' + v
)
1266 k
= quote_plus(str(k
))
1267 if isinstance(v
, str):
1269 l
.append(k
+ '=' + v
)
1270 elif _is_unicode(v
):
1271 # is there a reasonable way to convert to ASCII?
1272 # encode generates a string, but "replace" or "ignore"
1273 # lose information and "strict" can raise UnicodeError
1274 v
= quote_plus(v
.encode("ASCII","replace"))
1275 l
.append(k
+ '=' + v
)
1278 # is this a sufficient test for sequence-ness?
1282 v
= quote_plus(str(v
))
1283 l
.append(k
+ '=' + v
)
1285 # loop over the sequence
1287 l
.append(k
+ '=' + quote_plus(str(elt
)))
1291 def getproxies_environment():
1292 """Return a dictionary of scheme -> proxy server URL mappings.
1294 Scan the environment for variables named <scheme>_proxy;
1295 this seems to be the standard convention. If you need a
1296 different way, you can pass a proxies dictionary to the
1297 [Fancy]URLopener constructor.
1301 for name
, value
in os
.environ
.items():
1303 if value
and name
[-6:] == '_proxy':
1304 proxies
[name
[:-6]] = value
1307 def proxy_bypass_environment(host
):
1308 """Test if proxies should not be used for a particular host.
1310 Checks the environment for a variable named no_proxy, which should
1311 be a list of DNS suffixes separated by commas, or '*' for all hosts.
1313 no_proxy
= os
.environ
.get('no_proxy', '') or os
.environ
.get('NO_PROXY', '')
1314 # '*' is special case for always bypass
1317 # strip port off host
1318 hostonly
, port
= splitport(host
)
1319 # check if the host ends with any of the DNS suffixes
1320 for name
in no_proxy
.split(','):
1321 if name
and (hostonly
.endswith(name
) or host
.endswith(name
)):
1323 # otherwise, don't bypass
1327 if sys
.platform
== 'darwin':
1328 from _scproxy
import _get_proxy_settings
, _get_proxies
1330 def proxy_bypass_macosx_sysconf(host
):
1332 Return True iff this host shouldn't be accessed using a proxy
1334 This function uses the MacOSX framework SystemConfiguration
1335 to fetch the proxy information.
1339 from fnmatch
import fnmatch
1341 hostonly
, port
= splitport(host
)
1344 parts
= ipAddr
.split('.')
1345 parts
= map(int, parts
)
1347 parts
= (parts
+ [0, 0, 0, 0])[:4]
1348 return (parts
[0] << 24) |
(parts
[1] << 16) |
(parts
[2] << 8) | parts
[3]
1350 proxy_settings
= _get_proxy_settings()
1352 # Check for simple host names:
1354 if proxy_settings
['exclude_simple']:
1359 for value
in proxy_settings
.get('exceptions', ()):
1360 # Items in the list are strings like these: *.local, 169.254/16
1361 if not value
: continue
1363 m
= re
.match(r
"(\d+(?:\.\d+)*)(/\d+)?", value
)
1367 hostIP
= socket
.gethostbyname(hostonly
)
1368 hostIP
= ip2num(hostIP
)
1369 except socket
.error
:
1372 base
= ip2num(m
.group(1))
1373 mask
= int(m
.group(2)[1:])
1376 if (hostIP
>> mask
) == (base
>> mask
):
1379 elif fnmatch(host
, value
):
1385 def getproxies_macosx_sysconf():
1386 """Return a dictionary of scheme -> proxy server URL mappings.
1388 This function uses the MacOSX framework SystemConfiguration
1389 to fetch the proxy information.
1391 return _get_proxies()
1395 def proxy_bypass(host
):
1396 if getproxies_environment():
1397 return proxy_bypass_environment(host
)
1399 return proxy_bypass_macosx_sysconf(host
)
1402 return getproxies_environment() or getproxies_macosx_sysconf()
1404 elif os
.name
== 'nt':
1405 def getproxies_registry():
1406 """Return a dictionary of scheme -> proxy server URL mappings.
1408 Win32 uses the registry to store proxies.
1415 # Std module, so should be around - but you never know!
1418 internetSettings
= _winreg
.OpenKey(_winreg
.HKEY_CURRENT_USER
,
1419 r
'Software\Microsoft\Windows\CurrentVersion\Internet Settings')
1420 proxyEnable
= _winreg
.QueryValueEx(internetSettings
,
1423 # Returned as Unicode but problems if not converted to ASCII
1424 proxyServer
= str(_winreg
.QueryValueEx(internetSettings
,
1426 if '=' in proxyServer
:
1427 # Per-protocol settings
1428 for p
in proxyServer
.split(';'):
1429 protocol
, address
= p
.split('=', 1)
1430 # See if address has a type:// prefix
1432 if not re
.match('^([^/:]+)://', address
):
1433 address
= '%s://%s' % (protocol
, address
)
1434 proxies
[protocol
] = address
1436 # Use one setting for all protocols
1437 if proxyServer
[:5] == 'http:':
1438 proxies
['http'] = proxyServer
1440 proxies
['http'] = 'http://%s' % proxyServer
1441 proxies
['ftp'] = 'ftp://%s' % proxyServer
1442 internetSettings
.Close()
1443 except (WindowsError, ValueError, TypeError):
1444 # Either registry key not found etc, or the value in an
1445 # unexpected format.
1446 # proxies already set up to be empty so nothing to do
1451 """Return a dictionary of scheme -> proxy server URL mappings.
1453 Returns settings gathered from the environment, if specified,
1457 return getproxies_environment() or getproxies_registry()
1459 def proxy_bypass_registry(host
):
1464 # Std modules, so should be around - but you never know!
1467 internetSettings
= _winreg
.OpenKey(_winreg
.HKEY_CURRENT_USER
,
1468 r
'Software\Microsoft\Windows\CurrentVersion\Internet Settings')
1469 proxyEnable
= _winreg
.QueryValueEx(internetSettings
,
1471 proxyOverride
= str(_winreg
.QueryValueEx(internetSettings
,
1472 'ProxyOverride')[0])
1473 # ^^^^ Returned as Unicode but problems if not converted to ASCII
1474 except WindowsError:
1476 if not proxyEnable
or not proxyOverride
:
1478 # try to make a host list from name and IP address.
1479 rawHost
, port
= splitport(host
)
1482 addr
= socket
.gethostbyname(rawHost
)
1485 except socket
.error
:
1488 fqdn
= socket
.getfqdn(rawHost
)
1491 except socket
.error
:
1493 # make a check value list from the registry entry: replace the
1494 # '<local>' string by the localhost entry and the corresponding
1496 proxyOverride
= proxyOverride
.split(';')
1497 # now check if we match one of the registry values.
1498 for test
in proxyOverride
:
1499 if test
== '<local>':
1500 if '.' not in rawHost
:
1502 test
= test
.replace(".", r
"\.") # mask dots
1503 test
= test
.replace("*", r
".*") # change glob sequence
1504 test
= test
.replace("?", r
".") # change glob char
1506 # print "%s <--> %s" %( test, val )
1507 if re
.match(test
, val
, re
.I
):
1511 def proxy_bypass(host
):
1512 """Return a dictionary of scheme -> proxy server URL mappings.
1514 Returns settings gathered from the environment, if specified,
1518 if getproxies_environment():
1519 return proxy_bypass_environment(host
)
1521 return proxy_bypass_registry(host
)
1524 # By default use environment variables
1525 getproxies
= getproxies_environment
1526 proxy_bypass
= proxy_bypass_environment
1528 # Test and time quote() and unquote()
1531 for i
in range(256): s
= s
+ chr(i
)
1542 print round(t1
- t0
, 3), 'sec'
1545 def reporthook(blocknum
, blocksize
, totalsize
):
1546 # Report during remote transfers
1547 print "Block number: %d, Block size: %d, Total size: %d" % (
1548 blocknum
, blocksize
, totalsize
)
1556 'file://localhost/etc/passwd',
1557 'ftp://ftp.gnu.org/pub/README',
1558 'http://www.python.org/index.html',
1560 if hasattr(URLopener
, "open_https"):
1561 args
.append('https://synergy.as.cmu.edu/~geek/')
1564 print '-'*10, url
, '-'*10
1565 fn
, h
= urlretrieve(url
, None, reporthook
)
1569 for k
in h
.keys(): print k
+ ':', h
[k
]
1571 with
open(fn
, 'rb') as fp
:
1574 table
= string
.maketrans("", "")
1575 data
= data
.translate(table
, "\r")
1585 opts
, args
= getopt
.getopt(sys
.argv
[1:], "th")
1586 except getopt
.error
, msg
:
1588 print "Use -h for help"
1595 print "Usage: python urllib.py [-t] [url ...]"
1596 print "-t runs self-test;",
1597 print "otherwise, contents of urls are printed"
1605 print "Use -h for help"
1607 print urlopen(url
).read(),
1609 # Run test program when run as a script
1610 if __name__
== '__main__':