Imports and reasoning thing

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Imports and reasoning thing

Rinke Hoekstra-4
Hi,

(Perhaps this should have gone to the OWLAPI or Pellet list)

I'm experiencing some decidedly odd behaviour when reasoning over an  
ontology that imports another.

Case 1:
I have an ontology that defines some classes + role inclusion axioms,  
and has a test case specified as individuals [1]

Reasoner behaviour: ontology is consistent, all expected inferences  
are shown. No warning or error, logs are as they are supposed to be.

Case 1:
The same ontology is split into one ontology with basic class and role  
inclusion axioms, and another that extends these definitions for the  
test case + the test case itself ([2] and [3]). The third ontology  
imports the second.

Reasoner behaviour: Running the reasoner once while viewing classes:  
everything is fine. Going to the individuals tab, and selecting an  
individual will trigger the reasoner (again?), and it will either get  
stuck at some arbitrary percentage, or it will be really quick and  
continue without warning. In both cases, some inferences I would  
expect are missing. A peek at the logs gives me:

ERROR org.mindswap.pellet.exceptions.InternalReasonerException:  
Invalid type not(self(http://www.leibnizcenter.org/ontostore/process.owl#is_causal_change)
) for individual http://www.leibnizcenter.org/ontostore/process-example.owl#c2 
   org.protege.editor.owl.ui.frame.AbstractOWLFrame[AWT-EventQueue-0]

This is incomprehensible to me, as 1) the individual is _supposed_ to  
be having an is_causal_change property with itself, there's nothing in  
the class axioms that prevents it from having it (there are no  
disjoints defined anywhere) and 2) it works fine when I run the exact  
same set of axioms in a single ontology (!), and 3) no little red  
triangle thingy.

What in the name of @&#$*# is going on? ;)

Cheers,

        Rinke

[1] http://www.leibnizcenter.org/ontostore/process-test.owl
[2] http://www.leibnizcenter.org/ontostore/process.owl
[3] http://www.leibnizcenter.org/ontostore/process-example.owl


-----------------------------------------------
Drs. Rinke Hoekstra

Email: [hidden email]    Skype:  rinkehoekstra
Phone: +31-20-5253499     Fax:   +31-20-5253495
Web:   http://www.leibnizcenter.org/users/rinke

Leibniz Center for Law,          Faculty of Law
University of Amsterdam,            PO Box 1030
1000 BA  Amsterdam,             The Netherlands
-----------------------------------------------



_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: Imports and reasoning thing

Timothy Redmond

I don't think that this problem has anything to do with imports.  The  
same exception is getting thrown without the import.  Also many  
inferences are missing from the inferred individuals view in the case  
without the import.  My guess is that they are probably the same issue  
manifesting  in slightly different ways.  I am going to put a query to  
the pellet mailing list in a bit to see if they can help us get to the  
bottom of this.

-Timothy


On Aug 15, 2008, at 12:13 AM, Rinke Hoekstra wrote:

> Hi,
>
> (Perhaps this should have gone to the OWLAPI or Pellet list)
>
> I'm experiencing some decidedly odd behaviour when reasoning over an
> ontology that imports another.
>
> Case 1:
> I have an ontology that defines some classes + role inclusion axioms,
> and has a test case specified as individuals [1]
>
> Reasoner behaviour: ontology is consistent, all expected inferences
> are shown. No warning or error, logs are as they are supposed to be.
>
> Case 1:
> The same ontology is split into one ontology with basic class and role
> inclusion axioms, and another that extends these definitions for the
> test case + the test case itself ([2] and [3]). The third ontology
> imports the second.
>
> Reasoner behaviour: Running the reasoner once while viewing classes:
> everything is fine. Going to the individuals tab, and selecting an
> individual will trigger the reasoner (again?), and it will either get
> stuck at some arbitrary percentage, or it will be really quick and
> continue without warning. In both cases, some inferences I would
> expect are missing. A peek at the logs gives me:
>
> ERROR org.mindswap.pellet.exceptions.InternalReasonerException:
> Invalid type not(self(http://www.leibnizcenter.org/ontostore/process.owl#is_causal_change 
> )
> ) for individual http://www.leibnizcenter.org/ontostore/process-example.owl#c2
>   org.protege.editor.owl.ui.frame.AbstractOWLFrame[AWT-EventQueue-0]
>
> This is incomprehensible to me, as 1) the individual is _supposed_ to
> be having an is_causal_change property with itself, there's nothing in
> the class axioms that prevents it from having it (there are no
> disjoints defined anywhere) and 2) it works fine when I run the exact
> same set of axioms in a single ontology (!), and 3) no little red
> triangle thingy.
>
> What in the name of @&#$*# is going on? ;)
>
> Cheers,
>
> Rinke
>
> [1] http://www.leibnizcenter.org/ontostore/process-test.owl
> [2] http://www.leibnizcenter.org/ontostore/process.owl
> [3] http://www.leibnizcenter.org/ontostore/process-example.owl
>
>
> -----------------------------------------------
> Drs. Rinke Hoekstra
>
> Email: [hidden email]    Skype:  rinkehoekstra
> Phone: +31-20-5253499     Fax:   +31-20-5253495
> Web:   http://www.leibnizcenter.org/users/rinke
>
> Leibniz Center for Law,          Faculty of Law
> University of Amsterdam,            PO Box 1030
> 1000 BA  Amsterdam,             The Netherlands
> -----------------------------------------------
>
>
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: Imports and reasoning thing

Rinke Hoekstra-4
Hi Timothy,

Thanks for that. Indeed, I hadn't noticed the exception when not using  
the import. Still p1 gets classified correctly in the no-imports case,  
but it isn't when using imports. Odd.

Thanks,

        Rinke


On 16 aug 2008, at 03:48, Timothy Redmond wrote:

>
> I don't think that this problem has anything to do with imports.  The
> same exception is getting thrown without the import.  Also many
> inferences are missing from the inferred individuals view in the case
> without the import.  My guess is that they are probably the same issue
> manifesting  in slightly different ways.  I am going to put a query to
> the pellet mailing list in a bit to see if they can help us get to the
> bottom of this.
>
> -Timothy
>
>
> On Aug 15, 2008, at 12:13 AM, Rinke Hoekstra wrote:
>
>> Hi,
>>
>> (Perhaps this should have gone to the OWLAPI or Pellet list)
>>
>> I'm experiencing some decidedly odd behaviour when reasoning over an
>> ontology that imports another.
>>
>> Case 1:
>> I have an ontology that defines some classes + role inclusion axioms,
>> and has a test case specified as individuals [1]
>>
>> Reasoner behaviour: ontology is consistent, all expected inferences
>> are shown. No warning or error, logs are as they are supposed to be.
>>
>> Case 1:
>> The same ontology is split into one ontology with basic class and  
>> role
>> inclusion axioms, and another that extends these definitions for the
>> test case + the test case itself ([2] and [3]). The third ontology
>> imports the second.
>>
>> Reasoner behaviour: Running the reasoner once while viewing classes:
>> everything is fine. Going to the individuals tab, and selecting an
>> individual will trigger the reasoner (again?), and it will either get
>> stuck at some arbitrary percentage, or it will be really quick and
>> continue without warning. In both cases, some inferences I would
>> expect are missing. A peek at the logs gives me:
>>
>> ERROR org.mindswap.pellet.exceptions.InternalReasonerException:
>> Invalid type not(self(http://www.leibnizcenter.org/ontostore/process.owl#is_causal_change
>> )
>> ) for individual http://www.leibnizcenter.org/ontostore/process-example.owl#c2
>>  org.protege.editor.owl.ui.frame.AbstractOWLFrame[AWT-EventQueue-0]
>>
>> This is incomprehensible to me, as 1) the individual is _supposed_ to
>> be having an is_causal_change property with itself, there's nothing  
>> in
>> the class axioms that prevents it from having it (there are no
>> disjoints defined anywhere) and 2) it works fine when I run the exact
>> same set of axioms in a single ontology (!), and 3) no little red
>> triangle thingy.
>>
>> What in the name of @&#$*# is going on? ;)
>>
>> Cheers,
>>
>> Rinke
>>
>> [1] http://www.leibnizcenter.org/ontostore/process-test.owl
>> [2] http://www.leibnizcenter.org/ontostore/process.owl
>> [3] http://www.leibnizcenter.org/ontostore/process-example.owl
>>
>>
>> -----------------------------------------------
>> Drs. Rinke Hoekstra
>>
>> Email: [hidden email]    Skype:  rinkehoekstra
>> Phone: +31-20-5253499     Fax:   +31-20-5253495
>> Web:   http://www.leibnizcenter.org/users/rinke
>>
>> Leibniz Center for Law,          Faculty of Law
>> University of Amsterdam,            PO Box 1030
>> 1000 BA  Amsterdam,             The Netherlands
>> -----------------------------------------------
>>
>>
>>
>> _______________________________________________
>> p4-feedback mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback

-----------------------------------------------
Drs. Rinke Hoekstra

Email: [hidden email]    Skype:  rinkehoekstra
Phone: +31-20-5253499     Fax:   +31-20-5253495
Web:   http://www.leibnizcenter.org/users/rinke

Leibniz Center for Law,          Faculty of Law
University of Amsterdam,            PO Box 1030
1000 BA  Amsterdam,             The Netherlands
-----------------------------------------------



_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback