Skip to main content.
home | support | download

Back to List Archive

Re: [swish-e] Swish-e prefs to ignore certain sentences

From: <jp(at)not-real.bobschem.com>
Date: Thu Apr 17 2008 - 21:46:38 GMT
Jordan's <!-- SwishCommand noindex --><!-- SwishCommand index -->  
suggestion is pretty cool but, as he said, it seems useful only in a  
case of template use or PHP includes where one file gets the tag (or a  
limited number of static html pages).  Is there any way to centralize  
this sentence restriction within the swish-e configuration files so  
hundreds of html files don't need to be edited (and dozens of work  
hours lost) each time a new phrase/sentence has to be eliminated from  
swish-e search results?

JP





On Apr 17, 2008, at 9:43 AM, Jordan Hayes wrote:

I've got a website with roughly 500 html and php pages.  On roughly
400-425 of those pages, there are two "welcome" sentences at the top
of the page.  Those sentences have been given CSS class="subtitle".

One way to do this, especially if you've been using a templating system
to generate these pages, is to take something like

    <p>Welcome to ABC, voted the best widget company!<hr/>

and make it look like this:

    <!-- SwishCommand noindex -->
    <p>Welcome to ABC, voted the best widget company!<hr/>
    <!-- SwishCommand index -->

/jordan

_______________________________________________
Users mailing list
Users@lists.swish-e.org
http://lists.swish-e.org/listinfo/users



I've got a website with roughly 500 html and php pages.  On roughly
400-425 of those pages, there are two "welcome" sentences at the top
of the page.  Those sentences have been given CSS class="subtitle".
Additionally, there's another random sentence on quite a number of
pages, like "Find out more about this subject."  These sentences are
not given any special font class.  An example of a welcome sentence
would be 'Welcome to ABC, voted the best widget company."

Swish-e search results are pulling the 'welcome' sentences and placing
them at the very beginning of each search result.  When someone
searches for a key word, the first 15-25 words of each result are the
welcome sentences.  The random 'find out more...' sentences also often
show up in search results, making the swish-e results contaminated
with useless data.

I am familiar with configuring swish-e to ignore individual words
(with the stopwords file) and to even ignore meta names like <h1> or
<h3>, which I have done successfully.

I didn't find any info on how to ignore a particular sentence or
multi-word phrase, especially sentences that aren't using a tag like
<h3>, but rather a CSS class or no class at all.
_______________________________________________
Users mailing list
Users@lists.swish-e.org
http://lists.swish-e.org/listinfo/users
Received on Thu Apr 17 17:41:52 2008