bbsmenuから板一覧を取得する

bbsmenu.py

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

import re
import urllib
import sys
import pprint

DEFAULT_BBSMENU_URL = 'http://menu.2ch.net/bbsmenu.html'

def get_2ch_bbs_menu(bbsmenu_url=DEFAULT_BBSMENU_URL) :
  res = urllib.urlopen(bbsmenu_url)
  
  p1 = re.compile(r'<BR><BR><B>(.*)</B><BR>')
  p2 = re.compile(r'^<A HREF=([0-9a-zA-Z_:/\.]*\.2ch\.net/[0-9a-zA-Z_]*/)>(.*)</A><br>$')
  
  bbs_menu = {}
  for x in res.read().decode('sjis').splitlines() :
    m1 = p1.match(x)
    m2 = p2.match(x)
    if m1 :
      category = m1.group(1)
      bbs_menu[category] = {}
    elif m2 :
      board_name = m2.group(2)
      board_url = m2.group(1)
      try :
        bbs_menu[category][board_name] = board_url
      except NameError :
        pass
  
  for x in bbs_menu.keys() :
    if bbs_menu[x] == [] :
      del bbs_menu[x]
  
  return bbs_menu

if __name__ == "__main__" :
  pprint.pprint(get_2ch_bbs_menu())
  sys.exit()