diff options
Diffstat (limited to 'inc/3rdparty/site_config/standard/wired.com.txt')
| -rwxr-xr-x[-rw-r--r--] | inc/3rdparty/site_config/standard/wired.com.txt | 47 |
1 files changed, 25 insertions, 22 deletions
diff --git a/inc/3rdparty/site_config/standard/wired.com.txt b/inc/3rdparty/site_config/standard/wired.com.txt index 69bbf5b7..f5a72d14 100644..100755 --- a/inc/3rdparty/site_config/standard/wired.com.txt +++ b/inc/3rdparty/site_config/standard/wired.com.txt | |||
| @@ -1,22 +1,25 @@ | |||
| 1 | title: //meta[@property="og:title"]/@content | 1 | title: //meta[@name='Title']/@content |
| 2 | title: //h1 | 2 | author: //meta[@name='Author']/@content |
| 3 | title: //*[@class='posttitle'] | 3 | date: //meta[@name='DisplayDate']/@content |
| 4 | author: //*[@class='entryAuthor']/a[1] | 4 | body: //div[@class='entry'] |
| 5 | author://*[@class='member-title'] | 5 | strip: //p[contains(., 'Pages:') and contains(., 'View All')] |
| 6 | author://li[@class='author']/a[contains(@href, '/author/')] | 6 | strip: //p[@class='caption'] |
| 7 | date: substring-after(//div[@class='entryAuthor'], '·') | 7 | strip: //div[@class='desc' or @class='slide' or @id='slide-info'] |
| 8 | date: substring-before(//*[@class='entryDate'], '|') | 8 | |
| 9 | body: //div[@class='entry'] | 9 | strip_id_or_class: pullquote |
| 10 | strip: //span[contains(@class, 'nextprev')] | 10 | strip_id_or_class: left_rail |
| 11 | #strip_id_or_class: ngg-galleryoverview | 11 | strip_id_or_class: related-container |
| 12 | # ngg-galleryoverview is the whole content sometimes, e.g. http://www.wired.com/underwire/2011/12/best-mixtapes-of-2011/?pid=5736&viewall=true | 12 | strip_id_or_class: radvert-caption-wrap |
| 13 | 13 | ||
| 14 | strip: //p[span[contains(@class, 'contentjump')]] | 14 | # Remove gallery? |
| 15 | strip: //text()[contains(., 'nextpage')] | 15 | strip_id_or_class: wpgallery |
| 16 | 16 | ||
| 17 | prune: no | 17 | #strip: //text()[contains(., 'nextpage')] |
| 18 | 18 | ||
| 19 | single_page_link: //a[contains(@href, '/all/1') and contains(@class, 'contentjumpall')] | 19 | prune: no |
| 20 | 20 | ||
| 21 | test_url: http://www.wired.com/cloudline/2011/10/meet-arms-cortex-a15-the-future-of-the-ipad-and-possibly-the-macbook-air/ | 21 | single_page_link: //a[.='View All' and contains(@href, '/all/')] |
| 22 | test_url: http://www.wired.com/threatlevel/2012/05/ff_counterfeiter/all/1 \ No newline at end of file | 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/ | ||