WSGI Proxy App to add HTTP Basic Auth to outbound requests

authproxy.py

HTTPUSER = 'mylogin'
HTTPPASS = 'mypassword'

from wsgiproxy.app import WSGIProxyApp
from base64 import encodestring

proxyapp = WSGIProxyApp('http://remote.server.domain.com/')

# Craft Basic Auth header. Don't forget to strip the trailing \n.
HTTPAUTH = 'Basic ' + encodestring(HTTPUSER+':'+HTTPPASS).strip()

# Remove "hop-by-hop" headers (as defined by RFC2613, Section 13)
# since they are not allowed by the WSGI standard.
FILTER_HEADERS = [
    'Connection',
    'Keep-Alive',
    'Proxy-Authenticate',
    'Proxy-Authorization',
    'TE',
    'Trailers',
    'Transfer-Encoding',
    'Upgrade',
    ]


def wrap_start_response(start_response):
    def wrapped_start_response(status, headers_out):
        # Remove "hop-by-hop" headers
        headers_out = [(k,v) for (k,v) in headers_out
                       if k not in FILTER_HEADERS]
        return start_response(status, headers_out)
    return wrapped_start_response


def application(environ, start_response):
    start_response = wrap_start_response(start_response)
    environ['HTTP_AUTHORIZATION'] = HTTPAUTH
    return proxyapp(environ, start_response)


if __name__ == '__main__':
    from wsgiref.simple_server import make_server
    httpd = make_server('', 8000, application)
    print "Serving on port 8000..."
    httpd.serve_forever()

supervisord.conf

[program:authproxy]
command=/home/dotcloud/env/bin/python /home/dotcloud/current/authproxy.py

Ok it works fine.

Would gunicorn be faster ?

While I would be tempted to answer « Yeah, gunicorn is awesome and will be faster ! », I think it won't be true if you only have 1 concurrent request at a time. In that specific case, I guess it's harder to be faster than « while True: handle_request() » :-)

I my case it's a proxy to solr, so maybe i'll have many requests from web, AND ( i'm sure of that) a lot a update from background workers