mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-06-23 15:30:45 -04:00
62 lines
1.6 KiB
Python
62 lines
1.6 KiB
Python
#!/usr/bin/env python
|
|
##
|
|
# Last Edited: 2018-02-13 Carlos Alves <carlosalves90@gmail.com>
|
|
##
|
|
|
|
__license__ = 'GPL v3'
|
|
__author__ = '2010, Gustavo Azambuja <hola at gazambuja.com>'
|
|
'''
|
|
180.com.uy
|
|
'''
|
|
|
|
from calibre.web.feeds.news import BasicNewsRecipe
|
|
|
|
|
|
class Noticias(BasicNewsRecipe):
|
|
title = '180.com.uy'
|
|
__author__ = 'Gustavo Azambuja'
|
|
description = 'Noticias de Uruguay'
|
|
language = 'es_UY'
|
|
timefmt = '[%a, %d %b, %Y]'
|
|
use_embedded_content = False
|
|
recursion = 5
|
|
encoding = 'utf-8'
|
|
remove_javascript = True
|
|
no_stylesheets = True
|
|
|
|
oldest_article = 2
|
|
max_articles_per_feed = 100
|
|
remove_tags_after = dict(name='article')
|
|
keep_only_tags = [
|
|
dict(name='div', attrs={'class': 'nota'}),
|
|
dict(name='h3',),
|
|
dict(name='h4',),
|
|
dict(name='article',)
|
|
]
|
|
remove_tags = [
|
|
dict(name='div', attrs={'class': 'items'})
|
|
]
|
|
|
|
remove_attributes = ['width', 'height', 'style', 'font', 'color']
|
|
|
|
extra_css = '''
|
|
h1{font-family: Georgia,"Times New Roman",Times,serif}
|
|
h3{font-family: Georgia,"Times New Roman",Times,serif}
|
|
h2{font-family: Georgia,"Times New Roman",Times,serif}
|
|
p{font-family: Verdana,Arial,Helvetica,sans-serif}
|
|
body{font-family: Verdana,Arial,Helvetica,sans-serif}
|
|
img{margin-bottom: 0.4em; display:block;}
|
|
'''
|
|
|
|
feeds = [
|
|
(u'Titulares', u'http://www.180.com.uy/feed.php')
|
|
]
|
|
|
|
def get_cover_url(self):
|
|
pass
|
|
|
|
def preprocess_html(self, soup):
|
|
for item in soup.findAll(style=True):
|
|
del item['style']
|
|
return soup
|