diff options
| -rw-r--r-- | arstechnica.com.txt | 3 | ||||
| -rw-r--r-- | businessinsider.com.txt | 8 | ||||
| -rw-r--r-- | wired.com.txt | 6 |
3 files changed, 13 insertions, 4 deletions
diff --git a/arstechnica.com.txt b/arstechnica.com.txt index c2ffb6a..767f680 100644 --- a/arstechnica.com.txt +++ b/arstechnica.com.txt @@ -2,6 +2,7 @@ author: //p[@class='byline']/a body: //div[contains(@class,'article-content')] strip: //h2[@class='title'] strip_id_or_class: byline +strip_id_or_class: story-sidebar prune: no date: //div[@class='byline']/span[@class='posted']//abbr/@original-title @@ -13,4 +14,4 @@ strip: //div[@class='pager'] next_page_link: //nav//a[span/@class='next']/@href test_url: http://arstechnica.com/tech-policy/news/2012/02/gigabit-internet-for-80-the-unlikely-success-of-californias-sonicnet.ars -test_url: http://arstechnica.com/apple/2005/04/macosx-10-4/
\ No newline at end of file +test_url: http://arstechnica.com/apple/2005/04/macosx-10-4/ diff --git a/businessinsider.com.txt b/businessinsider.com.txt index f847687..39eb742 100644 --- a/businessinsider.com.txt +++ b/businessinsider.com.txt @@ -1,12 +1,16 @@ title://div[@class="sl-layout-post"]/h1 -body: //div[contains(@class, 'post-content') or contains(@class, 'KonaBody')] +body: //div[contains(@class, 'post-content') or contains(@class, 'slide-module') or contains(@class, 'KonaBody')] strip: //div[contains(@class, "post-sidebar")] strip: //div[@id='related-links'] +strip: //div[@class='related-links-container'] +strip: //p[@class='source'] author://div[@class="byline"]/a date://div[@class="byline"]/span[@class="date"] prune: no +single_page_link: //a[contains(text(), 'View as one page')] + strip://*[contains(@class,'sponsored-text')] strip: //div[@id='post_footer'] -test_url: http://www.businessinsider.com/microsoft-just-put-one-of-its-hardcore-technical-geniuses-on-xbox-2012-1
\ No newline at end of file +test_url: http://www.businessinsider.com/microsoft-just-put-one-of-its-hardcore-technical-geniuses-on-xbox-2012-1 diff --git a/wired.com.txt b/wired.com.txt index e06ad0e..f5a72d1 100644 --- a/wired.com.txt +++ b/wired.com.txt @@ -3,9 +3,13 @@ author: //meta[@name='Author']/@content date: //meta[@name='DisplayDate']/@content body: //div[@class='entry'] strip: //p[contains(., 'Pages:') and contains(., 'View All')] +strip: //p[@class='caption'] +strip: //div[@class='desc' or @class='slide' or @id='slide-info'] strip_id_or_class: pullquote strip_id_or_class: left_rail +strip_id_or_class: related-container +strip_id_or_class: radvert-caption-wrap # Remove gallery? strip_id_or_class: wpgallery @@ -18,4 +22,4 @@ single_page_link: //a[.='View All' and contains(@href, '/all/')] test_url: http://www.wired.com/cloudline/2011/10/meet-arms-cortex-a15-the-future-of-the-ipad-and-possibly-the-macbook-air/ test_url: http://www.wired.com/wiredenterprise/2013/09/docker/ -test_url: http://www.wired.com/threatlevel/2012/05/ff_counterfeiter/all/
\ No newline at end of file +test_url: http://www.wired.com/threatlevel/2012/05/ff_counterfeiter/all/ |