Re: [Virtuoso-users] full-text search

Thu, 20 May 2010 15:02:52 +0000 

Kingsley Idehen wrote:

[email protected] wrote:
I tried some queries on the sites on those links, and it looks to me like it only compares the query to each single text field. If I for example have a blog post with the author "John Doe" and title "Hello World", and then want to find this blog post with the query "john hello world". Then there is the problem that one of the words in the query is the author, while two of the words are from the title. With bif:contains I can compare "john hello world" against "John Doe" or against "Hello World", but I haven't found a way to compare it to the combination, "john doe hello world". There is perhaps a way though? 

Why use bif:contains when you can get that via Faceted Navigation?

See the /FCT UI.


Disambiguation (filtering by Entity Properties), Entity Rank, and Text 
Score Ranks all come into play.


There are even SOAP or REST style interfaces into this service.

Links:

1. http://lod.openlinksw.com/facet_doc.html

2. 
http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/VirtuosoFacetsWebService
3. 
http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/VirtuosoURIBurnerSampleTutorial 
.


Kingsley

  



In addition to the above, are you aware of xcontains? Blogs do carry 
some degree of XML based structure, so in this case (should faceted 
navigation not be the route you want to take) why not bif:xcontains?


Links:


1. 
http://virtuoso.openlinksw.com/presentations/SPARQL_Tutorials/SPARQL_Tutorials_Part_2/SPARQL_Tutorials_Part_2.html#(9)
2. 
http://docs.openlinksw.com/virtuoso/queryingxmldata.html#xcontainspredicate 
.


Kingsley

On Thu, May 20, 2010 at 3:51 PM, Kingsley Idehen 
<[email protected] <mailto:[email protected]>> wrote:


    [email protected] <mailto:[email protected]> wrote:

        When using bif:contains in Virtuoso, from what I've seen, you
        an only query a single property. In many cases it's more
        useful to query multiple properties, for example if you have a
        blog and want to do a single query that searches through both
        author and title and perhaps even the contents of the blog
        posts. I have found a blog post with some details on using

        Solr to solve this problem [1].  
        I imagine one way to do this with just Virtuoso would be to

        select the distinct words in the blog post author name and
        title, and use SPARUL to create a new triple like (blogpost,
        localhost:index, search_terms). Would this be a good way to do
        it, and are there perhaps other possibilities?

        1.
        
http://fgiasson.com/blog/index.php/2009/04/29/rdf-aggregates-and-full-text-search-on-steroids-with-solr/
        ------------------------------------------------------------------------

        
------------------------------------------------------------------------------

         
------------------------------------------------------------------------

        _______________________________________________
        Virtuoso-users mailing list
        [email protected]
        <mailto:[email protected]>
        https://lists.sourceforge.net/lists/listinfo/virtuoso-users

         


    Matt,

    Give me a Solr example so we can replicate using Virtuoso.

    I am very keen to have an example of something that Solr offers
    that can't be done with Virtuoso as is.

    BTW - http://uriburner.com/fct is an instance with a lot of Blog
    style data, you can try that vs. Solr.

    Of course you can also go to the new monster 15 Billion Triples
    instance at: http://lod.openlinksw.com .



    -- 


    Regards,

    Kingsley Idehen       President & CEO OpenLink Software     Web:
    http://www.openlinksw.com
    Weblog: http://www.openlinksw.com/blog/~kidehen
    <http://www.openlinksw.com/blog/%7Ekidehen>
    Twitter/Identi.ca: kidehen





------------------------------------------------------------------------

------------------------------------------------------------------------------


  
------------------------------------------------------------------------


_______________________________________________
Virtuoso-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/virtuoso-users

  
    




  



--

Regards,


Kingsley Idehen	      
President & CEO 
OpenLink Software     
Web: http://www.openlinksw.com

Weblog: http://www.openlinksw.com/blog/~kidehen

Twitter/Identi.ca: kidehen 





























					Reply via email to