aboutsummaryrefslogtreecommitdiffhomepage
path: root/inc/3rdparty/site_config/standard/thedaily.com.txt
diff options
context:
space:
mode:
Diffstat (limited to 'inc/3rdparty/site_config/standard/thedaily.com.txt')
-rw-r--r--inc/3rdparty/site_config/standard/thedaily.com.txt24
1 files changed, 24 insertions, 0 deletions
diff --git a/inc/3rdparty/site_config/standard/thedaily.com.txt b/inc/3rdparty/site_config/standard/thedaily.com.txt
new file mode 100644
index 00000000..24ebbbac
--- /dev/null
+++ b/inc/3rdparty/site_config/standard/thedaily.com.txt
@@ -0,0 +1,24 @@
1#keep all body text
2prune: no
3
4#title, body, metadata
5title: //div[@class='story_header']/h1
6body: //div[@id='content']
7author: substring-after(//span[@class='byline'], "by ")
8author: substring-after(//span[@class='byline'], "By ")
9author: //span[@class='byline']
10date: //span[@class='date']
11
12#formatting
13convert_double_br_tags: yes
14dissolve: //div[@class='slides_full']/ul/li
15
16# cleanup
17strip: //a[@id='story_note']
18strip: //br
19strip: //div[@class='intro']
20strip: //div[@class='share-block']
21strip: //div[@class='sidebar-social']
22strip: //div[@class='top-stories']
23strip: //div[@class='prevnext']
24test_url: http://www.thedaily.com/page/2012/01/09/010912-news-college-costs-1-5/ \ No newline at end of file