update rzeczpospolita recipe

This commit is contained in:
Tomasz Długosz 2018-10-09 01:02:07 +02:00
parent fc439f3d82
commit c7fc794a08

View File

@ -5,7 +5,6 @@ class RzeczpospolitaRecipe(BasicNewsRecipe):
__license__ = 'GPL v3'
__author__ = u'kwetal, Tomasz Dlugosz, adrianf0'
language = 'pl'
version = 2
title = u'Rzeczpospolita OnLine'
publisher = u'Presspublica Sp.'
@ -25,28 +24,15 @@ class RzeczpospolitaRecipe(BasicNewsRecipe):
feeds.append((u"Prawo", u'http://www.rp.pl/rss/1037')) # Prawo
keep_only_tags = []
keep_only_tags.append(dict(name='div', attrs={'class': 'article-content'}))
keep_only_tags.append(dict(name='h1', attrs={'id': 'article-title'}))
keep_only_tags.append(dict(name='img', attrs={'class': 'img-responsive article__image'}))
keep_only_tags.append(dict(name='div', attrs={'class': ['article-content', 'article__lead js-voice-read', 'article__content js-voice-read','article__image-desc','article__image-author']}))
remove_tags = []
remove_tags.append(dict(name='div', attrs={'id': 'article-copyright-box'}))
remove_tags.append(dict(name='div', attrs={'class': 'article-footer'}))
remove_tags.append(dict(name='div', attrs={'class': 'article-tags'}))
remove_tags.append(dict(name='div', attrs={'class': 'related-articles__wrapper'}))
remove_tags.append(dict(name='span', attrs={'class': ['article__premium-player','ad-label']}))
extra_css = '''
body {font-family: verdana, arial, helvetica, geneva, sans-serif ;}
h1{text-align: left;}
h2{font-size: medium; font-weight: bold;}
p.lead {font-weight: bold; text-align: left;}
.authordate {font-size: small; color: #696969;}
.fot{font-size: x-small; color: #666666;}
div.article__image-desc {font-style:italic; font-size:70%;text-align:right}
div.article__image-author {font-size:60%;text-align:right}
'''
# def skip_ad_pages(self, soup):
# if ('advertisement' in soup.find('title').string.lower()):
# href = soup.find('a').get('href')
# return self.index_to_soup(href, raw=True)
# else:
# return None
def print_version(self, url):
return url + '?template=printart'