diff options
Diffstat (limited to 'theatlantic.com.txt')
| -rw-r--r-- | theatlantic.com.txt | 18 |
1 files changed, 18 insertions, 0 deletions
diff --git a/theatlantic.com.txt b/theatlantic.com.txt new file mode 100644 index 0000000..061f872 --- /dev/null +++ b/theatlantic.com.txt @@ -0,0 +1,18 @@ +title: //div[@id='article']/h1 +title: //h1 + +body: //div[@class='articleText'] +body: //div[@class='articleContent'] +body: //div[@id='article'] +date: //*[contains(@class, 'date')] +author: //div[@id='profile']//*[@class='authors']//a[1] +author: //*[@class='author']/span +prune: no + +strip: //div[@class='moreOnBoxWithImages'] + +single_page_link: //a[@class='print'] + +test_url: http://www.theatlantic.com/technology/archive/2011/04/want-to-see-how-crazy-a-bot-run-market-can-be/237773/ +test_url: http://www.theatlantic.com/magazine/archive/2007/11/the-autumn-of-the-multitaskers/6342/ +test_url: http://www.theatlantic.com/entertainment/archive/2012/04/30-rock-live-a-funny-reminder-of-why-sitcoms-arent-shot-live-anymore/256447/
\ No newline at end of file |