[OBO-Edit Working Group] Can't delete genus? [was Re: Re-testing 2.1-rc4]

[Nomi H. <nlh...@lb...>]

Hi Ramona,

Thanks for testing. I have been unable to reproduce the problem you described. I followed exactly what you said, and I tried it with and without Autocommit turned on, and in all cases it worked--when I deleted the genus with the upper-left trash can, it stayed gone.

I did notice a reasoner bug during my testing. If I tried "Create new genus-differentia child" with the reasoner on, it didn't create the child, and it threw an exception.

org.oboedit.verify.CheckTask - currentCheck: DISJOINTEDNESS_CHECK: condition=14
org.oboedit.verify.CheckTask - Check DISJOINTEDNESS_CHECK: condition=14 needs reasoner--reasoner state is currently org.obo.reasoner.rbr.RuleBasedReasoner@44f6925f
obo.reasoner.impl.ReasonerOperationModel - reasoner not instance of RBR
obo.reasoner.impl.ReasonerOperationModel - reasoner not instance of RBR
obo.reasoner.impl.ReasonerOperationModel - reasoner not instance of RBR
obo.reasoner.impl.ReasonerOperationModel - reasoner not instance of RBR
 - Uncaught event dispatch exception
java.lang.UnsupportedOperationException
	at java.util.AbstractCollection.add(AbstractCollection.java:221)
	at org.obo.reasoner.rbr.IntersectionRule.getNewInferences(IntersectionRule.java:200)
	at org.obo.reasoner.rbr.RuleBasedReasoner.doReasoning(RuleBasedReasoner.java:270)
	at org.obo.reasoner.rbr.RuleBasedReasoner.findNewImplications(RuleBasedReasoner.java:249)

Interestingly, even if I then turned OFF the reasoner and tried "Create new genus-differentia child", it still failed. I had to turn off the reasoner and restart OBO-Edit with the reasoner off to stop that error.

I confirmed that this reasoner bug is not new; it existed at least as far back as 2.1-b7. I have filed a new bug report.


So, has anyone else encountered the reappearing-deleted-genus issue that Ramona reported?

	Nomi


On Dec 16, 2011, at 4:49 AM, Walls, Ramona wrote:

> I have been trying out the cross product editor, and I am still having a
> problem deleting the cross product. I didn't get to try this in 2.1-rc4
> before you did this re-release, so I don't know if it was happening before.
> 
> First I made a new term using "create genus differentia child" with the
> reasoner off. Then I specified the genus and differentia terms. Then I
> turned on the reason to see that the correct children were inferred, and
> they were. Next I deleted the differentia using the trash can. That worked
> fine, and the inferred children went away in the tree editor. Then I
> deleted the genus using the trash can in the upper left. It was deleted
> from the cross-product editor, but still showed up in the tree editor.
> When I clicked on another term, then went back to the first one, the genus
> was there in the cross-product editor again. I tried the same thing by
> creating a regular child, and then adding the cross product, and I got the
> same result -- I can delete the differentia, but not the genus. This is
> the same, whether I have the reasoner on or off.
> 
> Sorry to be the bearer of bad news!
> 
> Ramona
> 
> On 12/15/11 3:24 PM, "Nomi Harris" <nlh...@lb...> wrote:
> 
>> Not wanting to do yet another pre-release candidate, I sort of violated
>> my own protocol and re-released 2.1-rc4 with two minor changes (some help
>> page updates, and the removal of a jar file in runtime/ that was keeping
>> Thomas's plugin from working). Can some of you (especially Thomas)
>> download the latest version from
>> https://sourceforge.net/projects/geneontology/files/OBO-Edit%202%20%5BLate
>> st%20versions%5D/oboedit2.1-rc4/ and make sure it's ok? (Thomas, please
>> check that your plugin now works.)
>> 
>> Karen, can you get someone at Stanford to test it again in the pipeline?
>> It should perform exactly the same as the previous release candidate, but
>> I just want to be sure.
>> 
>> Thanks,
>>   Nomi
>