Manipulating Pellet OWL reasoner from Java

[PradeepSuresh]

Hello all,
I just started out using Pellet, the open source OWL reasoner and followed their examples to perform consistency checks, compute inferred types etc through the protege-owl api in JAVA. However all the results are either part of the OWLModel or displayed on the reasoner server window! Is there any way I can manipulate the output thrown out by the reasoner (Pellet in this case) to be shown on my JAVA console or through SWING on a Java panel??
If anyones had experience dealing with similar problems, plz help!!
Thanks in advance :)

Pradeep

[Tania Tudorache]

Pradeep,

I have already replied to your initial email. See the initial thread:

http://article.gmane.org/gmane.comp.misc.ontology.protege.owl/26228

There is no need to double post your messages.

Tania


PradeepSuresh wrote:

> Hello all,
> I just started out using Pellet, the open source OWL reasoner and followed
> their examples to perform consistency checks, compute inferred types etc
> through the protege-owl api in JAVA. However all the results are either part
> of the OWLModel or displayed on the reasoner server window! Is there any way
> I can manipulate the output thrown out by the reasoner (Pellet in this case)
> to be shown on my JAVA console or through SWING on a Java panel??
> If anyones had experience dealing with similar problems, plz help!!
> Thanks in advance :)
>
> Pradeep
>  

_______________________________________________
protege-owl mailing list
protege-owl@...
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 

[Tania Tudorache]

Pradeep,

As far as I know if the ontology is inconsistent, Pellet will stop the
consistency check after the first ontology inconsistency is found, and
it will generate one explanation for it. So, for example, if the
individual belongs to two complimentary classes, then the reasoner will
not try to reason any further. However, if only one class is
inconsistent, the reasoner can continue with other inferences.

So, if your ontology is inconsistent - which is a serious case, the
reasoner will say something like "cannot reason with inconsistent
ontologies". If only some classes are inconsistent, this won't affect
the reasoner, and you will be able to ask for the classification results.

For example, if in your code, you need to know the inferred superclasses
of all classes, then you can iterate over all classes in the ontology
and ask the reasoner for the inferred superclasses for each class.

You can also configure Pellet to generate explanations for the
inferences. You will find an example how to do this in the Pellet
documentation.

You can get access to the Pellet internal knowledge base from the
interface ProtegePelletReasoner through the method getPelletKB().

To ask if the ontology is consistent, you can call:

reasoner.getPelletReasoner().isConsistent()

and to get the explanation (e.g if the ontology is not consistent), you
call:

reasoner.getPelletKB().getExplanation()

Hope this helps.
Tania


pradeep@... wrote:

> Hi Tania,
> Thanks for the reply. I did see that part of the documentation and forums, however what I need is the ABox related inconsistencies (those thrown out by the reasoner on clicking 'I' on the Protege 3.4, e.g. "Instance A is forced to belong to Class A and its compliment", which arises when I create individuals that violate axioms imposed on parent class. What I also need is to get all such inconsistencies, i.e. not just a single instance related inconsistency since there might be more, also, preferably without checking validity of every instance in my ontology (although this is not too bad a solution if possible). In my application, I will not have any TBox related outputs such as inferred subclasses etc.
>
> Thanks in advance again!
> Pradeep
>
>
>
>
>
>
>
> Tania Tudorache wrote:
>  
>> Pradeep,
>>
>> I have already replied to your initial email. See the initial thread:
>>
>> http://article.gmane.org/gmane.comp.misc.ontology.protege.owl/26228
>>
>> There is no need to double post your messages.
>>
>> Tania
>>
>>
>> PradeepSuresh wrote:
>>    
>>> Hello all,
>>> I just started out using Pellet, the open source OWL reasoner and
>>> followed
>>> their examples to perform consistency checks, compute inferred types etc
>>> through the protege-owl api in JAVA. However all the results are either
>>> part
>>> of the OWLModel or displayed on the reasoner server window! Is there any
>>> way
>>> I can manipulate the output thrown out by the reasoner (Pellet in this
>>> case)
>>> to be shown on my JAVA console or through SWING on a Java panel??
>>> If anyones had experience dealing with similar problems, plz help!!
>>> Thanks in advance :)
>>>
>>> Pradeep
>>>  
>>>      
>> _______________________________________________
>> protege-owl mailing list
>> protege-owl@...
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03 
>>
>>
>>    
> Quoted from:
> http://www.nabble.com/Manipulating-Pellet-OWL-reasoner-from-Java-tp18407158p18540405.html
>
>  

_______________________________________________
protege-owl mailing list
protege-owl@...
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03