Version for Protege 5.2.0?

[msinclair2]

Would you be able to update jfact for the latest version of Protege (5.2.0)?

I inquired about this on the protege user list, and got the following response from Csongor:

This is a question to the jFact plug-in developer(s), really. Apparently they already have a plug-in that you can download and install manually. JFact v4.0.0 is available for download at [1], and you can download the source of version 4.0.2 from [2] and build it yourself. I tried both versions, 4.0.0 and 4.0.2, both with Protege 5.2.0 and 5.0.0, but in all cases I got this error message:
Plugin: jfact (4.0.0) was not successfully started. Please see the Protégé log for more details.
and
Plugin: jfact (4.0.2) was not successfully started. Please see the Protégé log for more details.
respectively. There was no further information in the log files, to shed more light the problem.

So, your best bet is to contact ignazio1977 on github, or whoever maintains the plug-in at the moment.

Csongor

[1] https://sourceforge.net/projects/jfact/
[2] https://github.com/owlcs/jfact/releases

On 09/11/2017 12:43 PM, Michael Sinclair wrote:
Is there a version of the JFact reasoner plugin for the latest protégé (5.2.0)? If so can it be made available on the repository?

Thanks.

[ignazio1977]

I have not tried JFact 4.x with Protege 5.2, so there might be some bug to fix; in the meantime, the latest 4.x release is 4.0.4, available on Maven Central - you can try deploying that manually and reporting any exceptions you see (the fact it did not get started successfully is not much in terms of debug info...)

[msinclair2]

Hi Ignazio, I downloaded 4.0.4, and it worked! However, the version name still shows up as 4.0.3 in the menus, so the display name was not updated. I also noticed there is 5.0.0 and 5.0.1, what are the major differences between 4 and 5? Thanks, Michael

…

On Mon, Sep 11, 2017 at 4:10 PM, Ignazio Palmisano ***@***.*** > wrote: I have not tried JFact 4.x with Protege 5.2, so there might be some bug to fix; in the meantime, the latest 4.x release is 4.0.4, available on Maven Central - you can try deploying that manually and reporting any exceptions you see (the fact it did not get started successfully is not much in terms of debug info...) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#18 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATWldh50Td6M5Y1rZUJcKcFyBYrmTdC3ks5sha_PgaJpZM4PT0U7> .

[ignazio1977]

Version 5 is for owlapi 5, however there are currently reported bugs against it, I wouldn't recommend using it for the moment.

[msinclair2]

Ignazio, I am running through the traditional pizza ontology with JFact 4.0.4 on Protege 5.2.0 and there are classification errors which aren't happening with FaCT++. I'm attaching the example ontology. JFact thinks that HighCaloriePizza == Pizza == LowCaloriePizza, while FaCT++ gets it right. Do you know why this could be happening? Also, ProbeInconsistentTopping shows up twice (redundantly) under owl:Nothing. Sometimes I need to stop and start the reasoner with the inferred hierarchy open in order to get the inferred hierarchy to show otherwise it only shows the incosistent owl:Nothing results and nothing else. Thanks, Michael

…

On Mon, Sep 11, 2017 at 11:32 PM, Ignazio Palmisano < ***@***.***> wrote: Version 5 is for owlapi 5, however there are currently reported bugs against it, I wouldn't recommend using it for the moment. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#18 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATWldrNSug1lL7PpUlwBcPigSA5XwT1Pks5shhdlgaJpZM4PT0U7> .

[ignazio1977]

Likely a bug, thanks for reporting it.

Returning the owl:Nothing result only is harder to analyse, I'll try replicating but that will take time.

[ignazio1977]

The attachment didn't make it through from email to comment, can you attach it to the issue or paste the definition of the classes?

[msinclair2]

Sorry, I guess attachments don't work if you attach them directly to your email reply from within gmail.

I'm trying to attach it to my response on github now, but it won't accept a ".owl" filetype, only a limited range of file types. So, I'm changing it from "pizza.owl" to "pizza.txt", and when you download it, you can manually change the extension back if necessary.

pizza.txt

Perhaps it has to do with the syntax of datatype property expressions, as this seems to have changed between versions, e.g. specifying "xsd:integer" instead of just "integer".