inc/3rdparty/site_config/standard/wired.com.txt

   1 title: //meta[@name='Title']/@content
   2 author: //meta[@name='Author']/@content
   3 date: //meta[@name='DisplayDate']/@content
   4 body: //div[@class='entry']
   5 strip: //p[contains(., 'Pages:') and contains(., 'View All')]
   6 strip: //p[@class='caption']
   7 strip: //div[@class='desc' or @class='slide' or @id='slide-info']
   8
   9 strip_id_or_class: pullquote
  10 strip_id_or_class: left_rail
  11 strip_id_or_class: related-container
  12 strip_id_or_class: radvert-caption-wrap
  13
  14 # Remove gallery?
  15 strip_id_or_class: wpgallery
  16
  17 #strip: //text()[contains(., 'nextpage')]
  18
  19 prune: no
  20
  21 single_page_link: //a[.='View All' and contains(@href, '/all/')]
  22
  23 test_url: http://www.wired.com/cloudline/2011/10/meet-arms-cortex-a15-the-future-of-the-ipad-and-possibly-the-macbook-air/
  24 test_url: http://www.wired.com/wiredenterprise/2013/09/docker/
  25 test_url: http://www.wired.com/threatlevel/2012/05/ff_counterfeiter/all/