今動くかは不明

tumblr_rss.py

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

import re
from time import sleep
import datetime
import mechanize
from pyquery import PyQuery as pq
import PyRSS2Gen

login_id = ''
password = ''

trgt_root = "http://www.tumblr.com"
login_url = "/login"
target = "/dashboard"
pages = 5

def main():
    br = mechanize.Browser()
    br.set_handle_robots(False)
    br.open(trgt_root + login_url)

    try:
        br.select_form(nr = 1)
        br['email'] = login_id
        br['password'] = password
        
        br.submit()
    except e:
        print 'login failed' + e

    posts = []
    for i in range(pages):
        page = br.open(trgt_root + target + (i == 0 and "" or "/%d" % i))
        d = pq(page.read().decode('utf-8'))
        posts += d('li.photo.not_mine')
        sleep(2)

    rss_items = []
    for post in posts:
        d = pq(post)
        
        rss_items.append( PyRSS2Gen.RSSItem(
                title = "Tumblr",
                link = "%s" % d('a.permalink')[0].attrib['href'],
                description = '<a href="%s"><img src="%s" alt="%s" /></a>' % (
                    d('a.permalink')[0].attrib['href'],
                    re.sub("_100", "_500", d("div.post_content img")[0].attrib['src']),
                    d('a.permalink')[0].attrib['title'],
                    ),
                guid = PyRSS2Gen.Guid( d('a.permalink')[0].attrib['href']),
                pubDate = datetime.datetime.now(),))

    rss = PyRSS2Gen.RSS2(
        title = "Tumblr/dashboard",
        link = trgt_root + target,
        description = "Tumblr",
        lastBuildDate = datetime.datetime.now(),
        items = rss_items)

    rss.write_xml(open("", "w+"), "utf-8")


if __name__ == '__main__':
    main()