extract_users.py

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

import os
import time
import random

mail_str = '[email protected]'
password_str = 'your_password'

match_str = '<tr class="tr_bg1">'
urank_file = 'urank.html'

wget_login ="wget -O temp.html --debug --cookies=on --no-check-certificate --ca-directory=/etc/ssl/certs --save-cookies=cookies.txt --keep-session-cookies --post-data='mail=%s&password=%s' http://book.akahoshitakuya.com/login" % (mail_str, password_str)
wget_urank = 'wget -O %s --keep-session-cookies --no-check-certificate --load-cookies cookies.txt "http://book.akahoshitakuya.com/urank&s=f"' % urank_file

if not os.path.exists(urank_file):
    os.system(wget_login)
    time.sleep(random.randint(5,15))
    os.system(wget_urank)

fp = open(urank_file, 'r')
for line in fp:
    line = line.strip()
    if line.find(match_str) == 0:
        break

lines = []
while True:
    line = line[len(match_str):]
    pos = line.find(match_str)
    lines.append(line[:pos])
    if pos == -1:
        break
    line = line[pos:-1]

users = []
for line in lines:
    fields = line.split('"')
    for field in fields:
        if field.find('/u/') == 0:
          users.append(field)

for user in users:
    print "http://book.akahoshitakuya.com" + user