I've used PyLint to harden the styling according to PEP8 as much as I could but looking for feedback on anything I've missed. Also some questions below:

  1. Passing argparse values directly to other functions helped me cut down on the number of local variables in main(). Is this a good practice regarding argparse?
  2. Should this module be broken into at least two scripts with one that holds the test_proxy() function or is the module length okay?
  3. I've used the function test_proxy() for a daemon thread, is it better to make it a Class?
  4. Is there enough exception handling here or are some obvious ones missing?
  5. Should the Request package be used instead of urllib.request?

Code here:

""" A multithreaded proxy checker

Given a file containing proxies, per line, in the form of ip:port, will attempt
to establish a connection through each proxy to a provided URL. Duration of
connection attempts is governed by a passed in timeout value. Additionally,
spins off a number of daemon threads to speed up processing using a passed in
threads parameter. Proxies that passed the test are written out to a file
called results.txt


    goodproxy.py [-h] -file FILE -url URL [-timeout TIMEOUT] [-threads THREADS]


    -file    -- filename containing a list of ip:port per line
    -url     -- URL to test connections against
    -timeout -- attempt time before marking that proxy as bad (default 1.0)
    -threads -- number of threads to spin off (default 16)


    get_proxy_list_size  -- returns the current size of the proxy holdingQueue
    test_proxy            -- does the actual connecting to the URL via a proxy
    main                 -- creates daemon threads, write results to a file

import argparse
import queue
import socket
import sys
import threading
import time
import urllib.request

def get_proxy_list_size(proxy_list):
    """ Return the current Queue size holding a list of proxy ip:ports """

    return proxy_list.qsize()

def test_proxy(url, url_timeout, proxy_list, lock, good_proxies, bad_proxies):
    """ Attempt to establish a connection to a passed in URL through a proxy.

    This function is used in a daemon thread and will loop continuously while
    waiting for available proxies in the proxy_list. Once proxy_list contains
    a proxy, this function will extract that proxy. This action automatically
    lock the queue until this thread is done with it. Builds a urllib.request
    opener and configures it with the proxy. Attempts to open the URL and if
    successsful then saves the good proxy into the good_proxies list. If an
    exception is thrown, writes the bad proxy to a bodproxies list. The call
    to task_done() at the end unlocks the queue for further processing.


    while True:

        # take an item from the proxy list queue; get() auto locks the
        # queue for use by this thread
        proxy_ip = proxy_list.get()

        # configure urllib.request to use proxy
        proxy = urllib.request.ProxyHandler({'http': proxy_ip})
        opener = urllib.request.build_opener(proxy)

        # some sites block frequent querying from generic headers
        request = urllib.request.Request(
            url, headers={'User-Agent': 'Proxy Tester'})

            # attempt to establish a connection
            urllib.request.urlopen(request, timeout=float(url_timeout))

            # if all went well save the good proxy to the list
            with lock:

        except (urllib.request.URLError,

            # handle any error related to connectivity (timeouts, refused
            # connections, HTTPError, URLError, etc)
            with lock:

            proxy_list.task_done()  # release the queue

def main(argv):
    """ Main Function

    Uses argparse to process input parameters. File and URL are required while
    the timeout and thread values are optional. Uses threading to create a
    number of daemon threads each of which monitors a Queue for available
    proxies to test. Once the Queue begins populating, the waiting daemon
    threads will start picking up the proxies and testing them. Successful
    results are written out to a results.txt file.


    proxy_list = queue.Queue()  # Hold a list of proxy ip:ports
    lock = threading.Lock()  # locks good_proxies, bad_proxies lists
    good_proxies = []  # proxies that passed connectivity tests
    bad_proxies = []  # proxies that failed connectivity tests

    # Process input parameters
    parser = argparse.ArgumentParser(description='Proxy Checker')

        '-file', help='a text file with a list of proxy:port per line',
        '-url', help='URL for connection attempts', required=True)
        type=float, help='timeout in seconds (defaults to 1', default=1)
        '-threads', type=int, help='number of threads (defaults to 16)',

    args = parser.parse_args(argv)

    # setup daemons ^._.^
    for _ in range(args.threads):
        worker = threading.Thread(

    start = time.time()

    # load a list of proxies from the proxy file
    with open(args.file) as proxyfile:
        for line in proxyfile:

    # block main thread until the proxy list queue becomes empty

    # save results to file
    with open("result.txt", 'w') as result_file:

    # some metrics
    print("Runtime: {0:.2f}s".format(time.time() - start))

if __name__ == "__main__":

1 Answer 1


Generally LTGM. Few suggestions anyway:

  • The worker loop termination

    As coded, it is implicit. Being explicit is always better, along the lines of:

            while True:
        except Empty:
  • Order of operations

    The workers are started before the proxy list is populated. Doing so surely raises some eyebrows, and is possibly unsafe. You seem to rely on the global lock, but are you sure that the main thread would not be blocked on open(args.file)? Loading the proxy file first seems more natural.

  • \$\begingroup\$ I've expanded the code into a hopefully useful app now at github. Thank you for your suggestions! \$\endgroup\$ Commented Jul 25, 2015 at 1:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.