diff --git a/recipes/hankyoreh.recipe b/recipes/hankyoreh.recipe new file mode 100644 index 0000000000..3afd3dd5dd --- /dev/null +++ b/recipes/hankyoreh.recipe @@ -0,0 +1,50 @@ +__license__ = 'GPL v3' +__copyright__ = '2011, Seongkyoun Yoo ' +''' +Profile to download The Hankyoreh +''' +import re +from calibre.web.feeds.news import BasicNewsRecipe +from calibre.ebooks.BeautifulSoup import BeautifulSoup + + +class Hankyoreh(BasicNewsRecipe): + title = u'Hankyoreh' + language = 'ko' + description = u'The Hankyoreh News articles' + __author__ = 'Seongkyoun Yoo' + oldest_article = 5 + recursions = 1 + max_articles_per_feed = 5 + no_stylesheets = True + keep_only_tags = [ + dict(name='tr', attrs={'height':['60px']}), + dict(id=['fontSzArea']) + ] + remove_tags = [ + dict(target='_blank'), + dict(name='td', attrs={'style':['padding: 10px 8px 5px 8px;']}), + dict(name='iframe', attrs={'width':['590']}), + ] + remove_tags_after = [ + dict(target='_top') + ] + feeds = [ + ('All News','http://www.hani.co.kr/rss/'), + ('Politics','http://www.hani.co.kr/rss/politics/'), + ('Economy','http://www.hani.co.kr/rss/economy/'), + ('Society','http://www.hani.co.kr/rss/society/'), + ('International','http://www.hani.co.kr/rss/international/'), + ('Culture','http://www.hani.co.kr/rss/culture/'), + ('Sports','http://www.hani.co.kr/rss/sports/'), + ('Science','http://www.hani.co.kr/rss/science/'), + ('Opinion','http://www.hani.co.kr/rss/opinion/'), + ('Cartoon','http://www.hani.co.kr/rss/cartoon/'), + ('English Edition','http://www.hani.co.kr/rss/english_edition/'), + ('Specialsection','http://www.hani.co.kr/rss/specialsection/'), + ('Hanionly','http://www.hani.co.kr/rss/hanionly/'), + ('Hkronly','http://www.hani.co.kr/rss/hkronly/'), + ('Multihani','http://www.hani.co.kr/rss/multihani/'), + ('Lead','http://www.hani.co.kr/rss/lead/'), + ('Newsrank','http://www.hani.co.kr/rss/newsrank/'), + ] \ No newline at end of file