Pythonic demonstration of BaseHTTPServer and lxml usage.

habproxy.py

#!/usr/bin/env python2.7
# encoding: utf-8
# src: https://gist.github.com/anonymous/06e0bd519490c8f03404
# habraproxy.py — это простейший http-прокси-сервер, запускаемый локально , который показывает содержимое страниц Хабра.
# С одним исключением: после каждого слова из шести букв должен стоять значок «™».
# Мне было по фану.
#
# INSTALL:
# pip install requests lxml
#
# USAGE:
# ./habproxy.py [localhost] [1666]
#
# TODO:
# tests
# different domains
# auth
# limit of threads
# more blacklisted tags
# user phantom.js to support script-populated websites
# logging
# uwsgi
# https

import BaseHTTPServer
import logging
import re
import sys
from SocketServer import ThreadingMixIn

import requests
from lxml import html


class HabrahabrModificationHandler(BaseHTTPServer.BaseHTTPRequestHandler):
    # old-style class :-(
    TARGET_URL = 'http://habrahabr.ru'  # TODO: port?  # TODO: another domains support?
    TAGS_BLACKLIST = [
        'script',
        'style',
        'img',
    ]
    TARGET_REGEXP = re.compile(ur'([^\w])([^\W\d_]{6})([^\w])', re.UNICODE)
    logger = logging.getLogger('http')

    def process_tag_text(self, text):
        # these space-magic allow to simplify regexp
        new_text = self.TARGET_REGEXP.sub(ur'\1\2™\3', ' ' + text + ' ')
        new_text = new_text[1:-1]
        return new_text

    def iter_tag(self, root):
        for tag in root.iterchildren("*"):
            tag = self.iter_tag(tag)
            if tag.text:
                if tag.tag not in self.TAGS_BLACKLIST:
                    tag.text = self.process_tag_text(unicode(tag.text))
                    for child in tag:
                        if child.tail is not None:
                            child.tail = self.process_tag_text(unicode(child.tail))
        return root

    def process_html_page(self, source_page, url=''):
        self.logger.debug("Processing page {}".format(url))
        processed_page = source_page.decode("utf-8")
        processed_page = processed_page.replace('habrahabr.ru', 'localhost:1666')
        processed_page = processed_page.replace('https://', 'http://')  # it's fine
        root = html.fromstring(processed_page)  # , html.HTMLParser(encoding='utf-8')
        root = self.iter_tag(root)
        return html.tostring(root, method="html")

    def do_IT(self, habr_resp):  # NOQA
        if habr_resp.headers.get('content-type').startswith('text/html'):  # TODO: text/xhtml and other exotic
            result = self.process_html_page(
                habr_resp.text.encode("utf-8"),
                self.path
            )
        else:
            result = habr_resp.text.encode("utf-8")
        self.send_response(habr_resp.status_code)
        self.end_headers()
        self.wfile.write(result)
        return

    def do_GET(self):  # NOQA
        habr_resp = requests.get(self.TARGET_URL + self.path)
        return self.do_IT(habr_resp)

    def do_POST(self):  # NOQA
        # TODO: never tested it. test it.
        in_data = self.rfile.read()
        habr_resp = requests.post(self.TARGET_URL + self.path, data=in_data)
        return self.do_IT(habr_resp)


class ThreadedHTTPServer(ThreadingMixIn, BaseHTTPServer.HTTPServer):
    # don't think about too much threads
    pass


def main(local_hostname, local_port):
    http_server = ThreadedHTTPServer((local_hostname, local_port), HabrahabrModificationHandler)
    print("server is ready and listening at http://{}:{}/".format(
        local_hostname,
        local_port
    ))
    while True:
        try:
            http_server.handle_request()
        except KeyboardInterrupt:
            break
    http_server.server_close()
    return


if __name__ == '__main__':
    try:
        local_port = int(sys.argv[2] if len(sys.argv) > 2 else '1666')
    except (ValueError, TypeError):
        print("Please provide integer second argument")
        exit(1)
    main(
        local_hostname=sys.argv[1] if len(sys.argv) > 1 else 'localhost',
        local_port=local_port
    )
# TODO: remove trailing whitespace on their template code