From 4e067ceabd705201a16b4c92cf4b23f3b990326c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Nicolas=20L=C5=93uillet?= Date: Sun, 13 Jul 2014 10:15:40 +0200 Subject: updated specific configuration for parsing --- inc/3rdparty/site_config/standard/cnnsi.com.txt | 50 ++++++++++++------------- 1 file changed, 25 insertions(+), 25 deletions(-) mode change 100644 => 100755 inc/3rdparty/site_config/standard/cnnsi.com.txt (limited to 'inc/3rdparty/site_config/standard/cnnsi.com.txt') diff --git a/inc/3rdparty/site_config/standard/cnnsi.com.txt b/inc/3rdparty/site_config/standard/cnnsi.com.txt old mode 100644 new mode 100755 index 6a2c2b80..ac49aef9 --- a/inc/3rdparty/site_config/standard/cnnsi.com.txt +++ b/inc/3rdparty/site_config/standard/cnnsi.com.txt @@ -1,26 +1,26 @@ -# main sportsillustrated.com articles - -body: //div[@id="cnnStoryContent"] -title: //div[@id="cnnStoryHeadline"]//h1 -author: //div[@id="cnnSubBanner"]//strong -date: substring-after(//div[@id="cnnTimeStamp"], "Updated: ") -date: substring-after(//div[@id="cnnTimeStamp"], "Posted: ") - -# kill ugly font buttons -strip: //div[@id="cnnSCFontButtons"] - -# kill misc filler videos & etc -strip: //div[@class="cnnDivideContent"] -strip: //*[@class="cnnTMbox"] - -# si vault articles -# ------------- -body: //div[@class="siv_artPara"] -title: //div[@class="siv_artHeader"]//h1 -author: //div[@class="byline"] -date: //div[@class="date"] - -next_page_link: //div[@id='cnnStoryContinue']/a -strip_id_or_class: cnnstorypagination - +# main sportsillustrated.com articles + +body: //div[@id="cnnStoryContent"] +title: //div[@id="cnnStoryHeadline"]//h1 +author: //div[@id="cnnSubBanner"]//strong +date: substring-after(//div[@id="cnnTimeStamp"], "Updated: ") +date: substring-after(//div[@id="cnnTimeStamp"], "Posted: ") + +# kill ugly font buttons +strip: //div[@id="cnnSCFontButtons"] + +# kill misc filler videos & etc +strip: //div[@class="cnnDivideContent"] +strip: //*[@class="cnnTMbox"] + +# si vault articles +# ------------- +body: //div[@class="siv_artPara"] +title: //div[@class="siv_artHeader"]//h1 +author: //div[@class="byline"] +date: //div[@class="date"] + +next_page_link: //div[@id='cnnStoryContinue']/a +strip_id_or_class: cnnstorypagination + test_url: http://cnnsi.com/2012/writers/peter_king/01/08/wild.card.round/index.html \ No newline at end of file -- cgit v1.2.3