uima-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Yosi Mass <YOSIM...@il.ibm.com>
Subject Re: Question on the SemanticSearch.zip - SIAPI
Date Thu, 25 Oct 2007 07:12:04 GMT
Dear Scott,

Can you please explain what you mean that the query <organization/> doesn't
return hits for organization.compnay annotations?

You probably know that a SIAPI query returns full documents as results and
not a result for each annotation so a document with <organization> and also
tags under organization will be returned only once.

Thanks, Yosi

             "Sommer, Scott                                                
             <Scott.Sommer@dst                                          To 
             o.defence.gov.au>         <uima-user@incubator.apache.org>    
             25/10/2007 08:24                                              
                                       Question on the SemanticSearch.zip  
             Please respond to         - SIAPI                             

Hi guys,

I posted a while back about about indexing using SIAPI. Since then I've
managed to get the entire process of indexing and searching working
pretty smoothly. However, I have a general question for anyone who is
familiar with SIAPI.

Basically, prior to indexing documents I have built a Index
Specification Descriptor which contains the typesystem to use for
indexing. My typesystem has hierachy in it, so there is stuff like
organization, organization.company in there. What I was hoping for is
that when I submit something like '<organization/>' as the xml search
fragment I would get hits for not only organization annotations but also
all the child types of organization including organization.company,
however, this doesn't happen. Is there anything that I can do with the
Index Specification Descriptor to define a hierachy for the SIAPI
indexer? Or does the indexer have no concept of hierachy?

If it doesn't support hierachy typesystems, I was wondering if I can
modify the xml fragment before it is submitted to basically OR all the
child types. I've seen some xml fragments for omnifind which have
<.or></.or> tags in them, but I have only seen it used to filter on
propertys, ie <book><.or><author>John Smith</author><title>XML
-Microsoft</title></.or></book>. Is there anything similar I can do with
juru but on types?

Finally, is there a workaround to my issue that I haven't thought of?

Thanks in advance,

IMPORTANT: This email remains the property of the Australian Defence
Organisation and is subject to the jurisdiction of section 70 of the CRIMES
ACT 1914.  If you have received this email in error, you are requested to
contact the sender and delete the email.

View raw message