Hi Richard,

Can you be a bit more explicit what the failing scenario is? I'm
definitely confused by wrapping a CR in an AE descriptor. Is it
possible to paste here an aggregate descriptor using sample components
from the UIMA SDK that demonstrates the problem?

Thanks,
Eddie


On Sun, Jun 10, 2012 at 2:11 PM, Richard Eckart de Castilho
<[email protected]> wrote:
> Am 10.06.2012 um 19:50 schrieb Richard Eckart de Castilho:
>
>> I guess another option should be to change CollectionReaderAdapter to create 
>> any missing initial view for sofa-unaware readers. That would not have any 
>> side other component type and it would solve the problem for my use-case as 
>> well. The problem is, that doesn't work, because the 
>> PrimitiveAnalysisEngine_impl.classAnalysisComponentProcess() already tries 
>> to access the mapped view and fails. Changing that to test if the 
>> mAnalysisComponent is a sofa-unaware CollectionReaderAdapter and creating a 
>> new view only in that case looks rather like a hack to me, although it would 
>> probably resolve the situation. I didn't test that yet, but if you think it 
>> reasonable, I can check it.
>
> Actually, the exception triggered by 
> PrimitiveAnalysisEngine_impl.classAnalysisComponentProcess() when accessing 
> the non-existing mapped CAS seems completely redundant, because if the 
> analysis engine delegate is a sofa-unaware CasMultiplier or 
> CollectionReader(Adapter) that doesn't actually use its input CAS, it doesn't 
> matter at all a that point that the mapped view does not exist. It's enough 
> if the mapped initial view is set up in any new CAS created for the 
> CasMultiplier/CollectionReader.
>
> So, there are many possible ways. I personally don't find to very attractive 
> to change the CollectionReaderDescription because I think that has quite some 
> overhead. Even if that was done, the problem would probably remain for mapped 
> CasMultipliers. I like it that UIMA internally treats all components equally, 
> so I would prefer doing something that this also works out well when sofa 
> mappings are used on components that produce new CASes and potentially do not 
> at all make used of the input CAS.
>
> -- Richard
>
> --
> -------------------------------------------------------------------
> Richard Eckart de Castilho
> Technical Lead
> Ubiquitous Knowledge Processing Lab (UKP-TUD)
> FB 20 Computer Science Department
> Technische Universität Darmstadt
> Hochschulstr. 10, D-64289 Darmstadt, Germany
> phone [+49] (0)6151 16-7477, fax -5455, room S2/02/B117
> [email protected]
> www.ukp.tu-darmstadt.de
> Web Research at TU Darmstadt (WeRC) www.werc.tu-darmstadt.de
> -------------------------------------------------------------------
>
>
>
>
>
>

Reply via email to