diff --git a/recipes/arcadia.recipe b/recipes/arcadia.recipe new file mode 100644 index 0000000000..980baa6ab8 --- /dev/null +++ b/recipes/arcadia.recipe @@ -0,0 +1,35 @@ +# -*- coding: utf8 -*- + +from calibre.web.feeds.recipes import BasicNewsRecipe +import re + +class Arcadia_BBS(BasicNewsRecipe): + title = u'Arcadia' + __author__ = 'Masahiro Hasegawa' + language = 'ja' + encoding = 'utf8' + filter_regexps = [r'ad\.jp\.ap\.valuecommerce.com',] + timefmt = '[%Y/%m/%d]' + remove_tags_before = dict(name='a', attrs={'name':'kiji'}) + + sid_list = [] #some sotory id + + def parse_index(self): + result = [] + for sid in self.sid_list: + s_result = [] + soup = self.index_to_soup( + 'http://www.mai-net.net/bbs/sst/sst.php?act=dump&all=%d' + % sid) + sec = soup.findAll('a', attrs={'href':re.compile(r'.*?kiji')}) + for s in sec[:-2]: + s_result.append(dict(title=s.string, + url="http://www.mai-net.net" + s['href'], + date=s.parent.parent.parent.findAll('td')[3].string[:-6], + description='', content='')) + result.append((s_result[0]['title'], s_result)) + return result + + + +