1 # urllib2 opener to connection through a proxy using the CONNECT method, (useful for SSL)

     2 # tested with python 2.4

     3 

     4 import urllib2

     5 import urllib

     6 import httplib

     7 import socket

     8 

     9 

    10 class ProxyHTTPConnection(httplib.HTTPConnection):

    11 

    12     _ports = {'http' : 80, 'https' : 443}

    13 

    14 

    15     def request(self, method, url, body=None, headers={}):

    16         #request is called before connect, so can interpret url and get

    17         #real host/port to be used to make CONNECT request to proxy

    18         proto, rest = urllib.splittype(url)

    19         if proto is None:

    20             raise ValueError, "unknown URL type: %s" % url

    21         #get host

    22         host, rest = urllib.splithost(rest)

    23         #try to get port

    24         host, port = urllib.splitport(host)

    25         #if port is not defined try to get from proto

    26         if port is None:

    27             try:

    28                 port = self._ports[proto]

    29             except KeyError:

    30                 raise ValueError, "unknown protocol for: %s" % url

    31         self._real_host = host

    32         self._real_port = port

    33         httplib.HTTPConnection.request(self, method, url, body, headers)

    34         

    35 

    36     def connect(self):

    37         httplib.HTTPConnection.connect(self)

    38         #send proxy CONNECT request

    39         self.send("CONNECT %s:%d HTTP/1.0\r\n\r\n" % (self._real_host, self._real_port))

    40         #expect a HTTP/1.0 200 Connection established

    41         response = self.response_class(self.sock, strict=self.strict, method=self._method)

    42         (version, code, message) = response._read_status()

    43         #probably here we can handle auth requests...

    44         if code != 200:

    45             #proxy returned and error, abort connection, and raise exception

    46             self.close()

    47             raise socket.error, "Proxy connection failed: %d %s" % (code, message.strip())

    48         #eat up header block from proxy....

    49         while True:

    50             #should not use directly fp probablu

    51             line = response.fp.readline()

    52             if line == '\r\n': break

    53 

    54 

    55 class ProxyHTTPSConnection(ProxyHTTPConnection):

    56     

    57     default_port = 443

    58 

    59     def __init__(self, host, port = None, key_file = None, cert_file = None, strict = None):

    60         ProxyHTTPConnection.__init__(self, host, port)

    61         self.key_file = key_file

    62         self.cert_file = cert_file

    63     

    64     def connect(self):

    65         ProxyHTTPConnection.connect(self)

    66         #make the sock ssl-aware

    67         ssl = socket.ssl(self.sock, self.key_file, self.cert_file)

    68         self.sock = httplib.FakeSocket(self.sock, ssl)

    69         

    70                                        

    71 class ConnectHTTPHandler(urllib2.HTTPHandler):

    72 

    73     def do_open(self, http_class, req):

    74         return urllib2.HTTPHandler.do_open(self, ProxyHTTPConnection, req)

    75 

    76 

    77 class ConnectHTTPSHandler(urllib2.HTTPSHandler):

    78 

    79     def do_open(self, http_class, req):

    80         return urllib2.HTTPSHandler.do_open(self, ProxyHTTPSConnection, req)

    81 

    82 

    83 if __name__ == '__main__':

    84     

    85     import sys

    86     

    87     opener = urllib2.build_opener(ConnectHTTPHandler, ConnectHTTPSHandler)

    88     urllib2.install_opener(opener)

    89     req = urllib2.Request(url='http://google.com')

    90     req.set_proxy('10.101.1.1:80', 'http')

    91     f = urllib2.urlopen(req)

    92     print f.read()

    93