shiweifu

def down_page_callable(**kw):
    try:
        _task_queue = kw["task_queue"]
        _url_list = kw["url_list"]
        _log_queue = kw["log_queue"]
        _deeph = kw["deeph"] 
        _filter_exts = kw["filter_exts"],
        _root = kw["root"]
    except KeyError, ke:

shiweifu

# -*- coding: UTF-8 -*-
import httplib
import urllib2
import socket
import traceback
import re
import BeautifulSoup
import threading
from threadpool import ThreadPool
from Queue import Queue, Empty

shiweifu

class FileWrapper:
    """Wrapper to convert file-like objects to iterables"""

    def __init__(self, filelike, blksize=8192):
        self.filelike = filelike
        self.blksize = blksize
        if hasattr(filelike,'close'):
            self.close = filelike.close

shiweifu

socket = require("socket")
httpcode = require("httpcode")
require("mime")
getMime = mime.getMime

assert(getMime)


function main( p )
	local port

shiweifu

mime = {}

local mime_table = 
{
				[".html"] = {
								["mime"] = "text/html",
								["bin"] = false
							},
				[".xml"] = {
							["mime"] = "application/xml",

shiweifu

function main()
	regex = [[false"><a title="(.-)"%starget=".-"%shref="(.-)%.html"]]

	f = io.open("test.html", "r")
	html = f:read("*a")
	f:close()

	s_idx, e_idx, title, url = string.find(html, regex, 0)

shiweifu

-- trim implementations

function trim1(s)
  return (s:gsub("^%s*(.-)%s*$", "%1"))
end
-- from PiL2 20.4

function trim2(s)
  return s:match "^%s*(.-)%s*$"
end

shiweifu

w3m/0.4.1
Emacs-w3m/1.4.4 w3m/0.5.1+cvs-1.946
Mozilla/4.0 (compatible; MSIE 4.01; Windows CE; MSN Companion 2.0; 800x600; Compaq)  
DonutP; Windows98SE  
MoonBrowser (version 0.41 Beta4)  
Moonbrowser - IE based browser (Japan)  
Mozilla/4.0 (compatible; MSIE 4.01; Windows CE; MSN Companion 2.0; 800x600; Compaq)  
Mozilla/1.1 (compatible; MSPIE 2.0; Windows CE)  
Mozilla/1.22 (compatible; MSIE 2.0d; Windows NT)  
Mozilla/2.0 (compatible; MSIE 2.1; Mac_PowerPC)  

shiweifu

UA_LIST = [a.split("\n")[0].strip() for a in open("user_agent", "r").readlines()]

random_value = lambda lst: random.randint(0, len(lst))

def down_page(url):
    global UA_LIST
    headers = { "User-Agent" : random_value(UA_LIST)}
    req = urllib2.Request(url, headers=headers)
    response = urllib2.urlopen(req)
    the_page = response.read()

shiweifu

#!/usr/bin/env python
#coding: utf-8

#author: shiweifu
#mail  : [email protected]



import logging