斧頭幫第三關 http://axe.g0v.tw/level/3

axe3.py

#!/usr/bin/env python
# encoding: utf-8
 
# http://axe.g0v.tw/level/3
 
import urllib, urllib2, re
import cookielib
 
lines = []
jar = cookielib.FileCookieJar("cookies")
opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(jar))
 
def parse_page(index):
	url = "http://axe-level-1.herokuapp.com/lv3/" if index == 1 else "http://axe-level-1.herokuapp.com/lv3/?page=next"
	f = opener.open(url)
	html = f.read()
	pattern = r"<tr>\s*<td>(.*)</td>\s*<td>(.*)</td>\s*<td>(.*)</td>\s*</tr>"
	results = re.findall(pattern, html, re.MULTILINE)[1:]
	line = '{"town": "%s", "village": "%s", "name" : "%s"}'
	for result in results:
		lines.append(line % tuple(result))
 
for i in range(1, 77):
	parse_page(i)
 
json_text = "[%s]" % ",\n".join(lines)
 
with open("test.txt", "w") as f:
	f.write(json_text)