1 title: //h1[@class='post-title']
2 author: //div[@class='post-byline']/a
3 date: substring-before(//div[@class='post-byline'], ', by')
5 body: //div[@class='post-body']
7 test_url: http://sf.curbed.com/archives/2011/10/17/lower_haight_loft_would_really_really_really_like_a_buyer.php