calibre/recipes/pc_advisor.recipe

88 lines
5.7 KiB
Python

#!/usr/bin/env python
__license__ = 'GPL v3'
__author__ = 'Lorenzo Vigentini'
__copyright__ = '2009, Lorenzo Vigentini <l.vigentini at gmail.com>'
__version__ = 'v1.01'
__date__ = '14, January 2010'
__description__ = 'PC Advisor delivers expert advice you can trust to business and home PC users who want to buy the best-value equipment and make the most out of the equipment they already own.'
'''
http://www.pcadvisor.co.uk/
'''
from calibre.web.feeds.news import BasicNewsRecipe
class pcAdvisor(BasicNewsRecipe):
__author__ = 'Lorenzo Vigentini'
description = 'PC Advisor delivers expert advice you can trust to business and home PC users who want to buy the best-value equipment and make the most out of the equipment they already own.'
cover_url = 'http://media.pcadvisor.co.uk/images/spacer.gif'
title = 'Pc Advisor '
publisher = 'IDG Communication'
category = 'PC, computing, product reviews, UK'
language = 'en'
encoding = 'cp1252'
timefmt = '[%a, %d %b, %Y]'
oldest_article = 15
max_articles_per_feed = 25
use_embedded_content = False
recursion = 10
remove_javascript = True
no_stylesheets = True
keep_only_tags = [
dict(name='div', attrs={'id':'articlecontent'})
]
remove_tags = [
dict(name='div', attrs={'id':['crosssitesignup','submitarticle','dontPrint','commentsForm','userReviewFormContainer','reevooContainerId']}),
dict(name='div', attrs={'class':'mpu'}),
dict(name='p', attrs={'id':'articlePageList'}),
dict(name='div', attrs={'style':['margin: 0pt 10px 5px;','margin: 0pt 10px 5px;']}),
dict(name='p', attrs={'class':'dontPrint'}),
dict(name='h2', attrs={'class':'sectionTitle'}),
dict(name='a', attrs={'title':'Subscribe to PC Advisor'}),
dict(name='a', attrs={'name':'revooContent'}),
{'name':['form','script','link']}
]
remove_tags_after = [
dict(name='p', attrs={'id':'crosssitesignup'})
]
def get_article_url(self, article):
return article.get('guid', None)
feeds = [
(u'News Headlines', u'http://www.pcadvisor.co.uk/rss/feeds/pcanews.xml'),
(u'Reviews', u'http://www.pcadvisor.co.uk/rss/feeds/pcareviews.xml'),
(u'New Products', u'http://www.pcadvisor.co.uk/rss/feeds/blog18.xml'),
(u'PC Advisor Blog', u'http://www.pcadvisor.co.uk/rss/feeds/blog4.xml'),
(u'PC Security', u'http://www.pcadvisor.co.uk/rss/feeds/pca-security.xml'),
(u'Laptops', u'http://www.pcadvisor.co.uk/rss/feeds/pca-laptop.xml'),
(u'Green Computing', u'http://www.pcadvisor.co.uk/rss/feeds/pca-green-computing.xml'),
(u'Internet and broadband', u'http://www.pcadvisor.co.uk/rss/feeds/pca-internet.xml'),
(u'Prones and PDAs', u'http://www.pcadvisor.co.uk/rss/feeds/pca-phones.xml'),
(u'Software', u'http://www.pcadvisor.co.uk/rss/feeds/pca-software.xml'),
(u'Small Business', u'http://www.pcadvisor.co.uk/rss/feeds/pca-small-business.xml'),
(u'Photo and video', u'http://www.pcadvisor.co.uk/rss/feeds/pca-photo-video.xml'),
(u'Mac News', u'http://www.pcadvisor.co.uk/rss/feeds/pca-mac.xml'),
(u'Linux', u'http://www.pcadvisor.co.uk/rss/feeds/pca-linux.xml'),
(u'WiFi and Networking', u'http://www.pcadvisor.co.uk/rss/feeds/pca-networking.xml'),
(u'Gadgets', u'http://www.pcadvisor.co.uk/rss/feeds/pca-gadgets.xml')
]
extra_css = '''
h1 {font-family:"Trebuchet MS",Arial,Helvetica,sans-serif; font-size:20px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold; line-height:18px;}
h2 {font-family:"Trebuchet MS",Arial,Helvetica,sans-serif; font-size:18px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold; line-height:16px; }
h3 {color:#333333;font-family:"Trebuchet MS",Arial,Helvetica,sans-serif; font-size:16px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold; line-height:14px;}
h4 {color:#333333; font-family:"Trebuchet MS",Arial,Helvetica,sans-serif;font-size:16px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold; line-height:14px; }
h5 {color:#333333; font-family:"Trebuchet MS",Arial,Helvetica,sans-serif; font-size:12px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold; line-height:14px; text-transform:uppercase;}
.newsdate {color:#333333;font-family:"Trebuchet MS",Arial,Helvetica,sans-serif;font-size:10px; font-size-adjust:none; font-stretch:normal; font-style:italic; font-variant:normal; font-weight:bold; line-height:10px; text-decoration:none;}
.author {color:#333333;font-family:"Trebuchet MS",Arial,Helvetica,sans-serif;font-size:10px; font-size-adjust:none; font-stretch:normal; font-style:bold; font-variant:normal; font-weight:bold; line-height:10px; text-decoration:none;}
p {font-family:"Trebuchet MS",Arial,Helvetica,sans-serif; font-size:10px;}
'''