]>
Commit | Line | Data |
---|---|---|
4e067cea NL |
1 | title: //div[contains(@class, 'articleHead')]//h1 |
2 | ||
3bb6a8ed | 3 | body: //div[@itemprop='articleBody'] |
4e067cea NL |
4 | body: //div[@class='articleText'] |
5 | body: //div[@class='articleContent'] | |
6 | body: //div[@id='article'] | |
7 | date: //*[contains(@class, 'date')] | |
8 | author: //div[@id='profile']//*[@class='authors']//a[1] | |
9 | author: //*[@class='author']/span | |
10 | prune: no | |
11 | ||
12 | strip: //div[@class='moreOnBoxWithImages'] | |
13 | strip: //p[contains(., 'This article available online at:')] | |
14 | strip: //p[contains(., 'This article available online at:')]/following::* | |
15 | strip: //div[@class='earthbox'] | |
16 | ||
3bb6a8ed | 17 | single_page_link: //div[contains(@class, 'article-tools')]//a[contains(@class, 'print')] |
4e067cea | 18 | |
90a1a78b NL |
19 | native_ad_clue: //meta[@property="og:url" and contains(@content, '/sponsored/')] |
20 | ||
3bb6a8ed NL |
21 | #multi-page article |
22 | test_url: http://www.theatlantic.com/magazine/archive/2014/12/the-real-roots-of-midlife-crisis/382235/ | |
23 | test_contains: The curve tends to evince itself | |
24 | ||
4e067cea NL |
25 | test_url: http://www.theatlantic.com/technology/archive/2011/04/want-to-see-how-crazy-a-bot-run-market-can-be/237773/ |
26 | test_url: http://www.theatlantic.com/magazine/archive/2007/11/the-autumn-of-the-multitaskers/6342/ | |
3bb6a8ed | 27 | test_url: http://www.theatlantic.com/entertainment/archive/2012/04/30-rock-live-a-funny-reminder-of-why-sitcoms-arent-shot-live-anymore/256447/ |