calibre/recipes/hankyoreh.recipe
2015-09-16 21:18:51 -07:00

41 lines
1.7 KiB
Plaintext

# -*- coding: utf-8 -*-
__license__ = 'GPL v3'
__copyright__ = '2011, Seongkyoun Yoo <seongkyoun.yoo at gmail.com>'
'''
Profile to download The Hankyoreh
'''
from calibre.web.feeds.news import BasicNewsRecipe
class Hankyoreh(BasicNewsRecipe):
language = 'ko'
title = u'한겨례'
description = u'The Hankyoreh News articles'
__author__ = 'Seongkyoun Yoo'
oldest_article = 7
max_articles_per_feed = 10
no_stylesheets = True
remove_javascript = True
keep_only_tags = [
dict(name='div', attrs ={'class':['article-head']}),
dict(name='div', attrs ={'class':['article-text']}),
]
remove_tags = [
dict(name='p', attrs={'class':['category']}),
]
remove_tags_after = dict(id={'ad_box01'})
feeds = [
#(u'전체기사', 'http://www.hani.co.kr/ilram/rss/hkr_news_list_all.xml'),
(u'정치', 'http://www.hani.co.kr/ilram/rss/hkr_news_list_politics.xml'),
#(u'경제', 'http://www.hani.co.kr/rss/economy/'),
(u'사회', 'http://www.hani.co.kr/ilram/rss/hkr_news_list_society.xml'),
#(u'국제', 'http://www.hani.co.kr/rss/international/'),
(u'문화', 'http://www.hani.co.kr/ilram/rss/hkr_news_list_culture.xml'),
(u'스포츠', 'http://www.hani.co.kr/ilram/rss/hkr_news_list_sports.xml'),
#(u'과학', 'http://www.hani.co.kr/rss/science/'),
(u'사설·칼럼','http://www.hani.co.kr/ilram/rss/hkr_news_list_opinion.xml'),
(u'만화만평', 'http://www.hani.co.kr/ilram/rss/hkr_news_list_cartoon.xml'),
#(u'한겨례섹션','http://www.hani.co.kr/rss/specialsection/'),
]