[Pellet-users] Seperating A & T boxes in Pellet

[Matt Williams]

Dear Evren,

Thanks for the reply.

I was asking because I'm trying to slim down the search space for an 
rdf-based search; I know the terms I'm looking for are in the A-Box, so 
if I can split the ontology into an A and T-Box, then the search space 
goes down (by a lot, I suspect).

This is in light of my previous questions about negated ground roles - 
AFAIK, I'm going to have to search for these myself, in the rdf.

Does that make sense?

Thanks,

Matt

Evren Sirin wrote:
> On 3/10/07 6:13 PM, Matt Williams wrote:
>> Dear Evren,
>>
>> Looking at the API for the 1.4 rc2, I can see stuff relating to A and 
>> T boxes. I was wondering:
>>
>> 1: Is it's possible to split the A & T box into separate structures, 
>> to handle them differently.
>>
>>   
> Do you mean serialize them separately? It is possible but there is no 
> built-in function to do that for you.
>> 2: If so, how to do it
>> 3: The role of KnowledgeBase - and how you generate one from a Jena 
>> model (I'm guessing there's a .getKB() method somewhere I haven't seen).
>>
>> I'm also a bit confused about the aterm references in some of the 
>> methods - for example, some of the methods in
>>
>> org.mindswap.pellet.query.impl.QueryImpl
>>
>> require aterms to be passed - should I be "casting" an rdf node as an 
>> aterm, or is it more complex than this?
>>   
> If you are going to use the Jena interface you don't really need to 
> worry about the KB or ATerms. You simply create a Jena model, set Pellet 
> as its reasoner and use only Jena specific functions. All the reasoning, 
> transformations, etc. will be done in the background for you. See the 
> sample programs provided in the examples directory of the distribution 
> package (esp. JenaReasoner.java, PelletQueryExample.java).
> 
> Cheers,
> Evren
>> Thanks a lot,
>>
>> Matt
>>
>>
>>   
>