diff options
Diffstat (limited to 'techcrunch.com.txt')
| -rw-r--r-- | techcrunch.com.txt | 18 |
1 files changed, 18 insertions, 0 deletions
diff --git a/techcrunch.com.txt b/techcrunch.com.txt new file mode 100644 index 0000000..1509c46 --- /dev/null +++ b/techcrunch.com.txt @@ -0,0 +1,18 @@ +body: //div[contains(@class, 'media-container') or contains(@class, 'body-copy')] + +author: //a[@class="name"] + +date: //div[@class="post-time"] + +title: //h1[@class="headline"] +strip_id_or_class: module-crunchbase + +# The following is for the mobile site +body: //div[@id="singlentry"] +author: substring-after(//span[@class="single-post-meta-top"],'rsaquo; ') +date: substring-before(//div[@class="single-post-meta-top"],' @') +title: //a[@class="sh2"] + +prune: no + +test_url: http://techcrunch.com/2011/10/18/apples-insanely-great-q1-2012/
\ No newline at end of file |