gistfile1.py

#!/usr/bin/python2
#  -*- coding: utf-8 -*-
from __future__ import unicode_literals

import re

from BeautifulSoup import BeautifulSoup
import chardet
from tornado.httpclient import AsyncHTTPClient
import tornado.ioloop
from tornado.options import define, options
import tornado.web
import zlib


def tag_editable_selector(tag):
    return tag.get('contenteditable', None) not in [None, 'false']


def tag_selector(tag):
    if tag.name in (
        '[document]', 'html', 'head', 'meta', 'code', 'style', 'script'
    ):
        return False
    if tag.findParent(tag_editable_selector):
        return False
    return True


def prank_data(data):
    soup = BeautifulSoup(data)
    for txt in soup.findAll(text=True):
        if not tag_selector(txt.parent):
            continue
        newtext = re.sub(r"(?u)\b(\w{6})\b", r"\1™", txt)
        txt.replaceWith(newtext)
    return str(soup)


class MainHandler(tornado.web.RequestHandler):
    def handle_response(self, resp):
        if resp.error:
            self.write_error(resp.code)
        else:
            page = resp.body
            if 'gzip' in resp.headers.get('Content-Encoding', ''):
                page = zlib.decompress(page, 16+zlib.MAX_WBITS)
            content_type = resp.headers.get('Content-Type', None)
            print 'Content-Type:', content_type
            if content_type is not None:
                content_type = re.sub(
                    r"(charset\=)(.*)", r"\1utf-8", content_type)
                self.set_header('Content-Type', content_type)
            if 'text/html' in (content_type or ''):
                enc = chardet.detect(page)['encoding']
                page = page.decode(enc)
                page = prank_data(page)
            self.write(page)
            self.finish()

    @tornado.web.asynchronous
    def get(self, path):
        httpclient = AsyncHTTPClient()
        url = 'http://'+options.host+'/'+path
        print url
        httpclient.fetch(url, self.handle_response)


application = tornado.web.Application([
    (r"/(.*)", MainHandler),
])


if __name__ == "__main__":
    define("host", default="habrahabr.ru", help="Source host to tm-sorce")
    define("port", default=8888, help="Port to listen on", type=int)
    options.parse_command_line()
    print options.host, options.port
    application.listen(options.port)
    tornado.ioloop.IOLoop.current().start()

Just my version of this.