Can a PEAR component be used as a Collection Reader?

["Chavan, Girish" <ch...@upmc.edu>]

I am trying to use a PEAR component as a collection reader in a CPE descriptor.

I am wondering if I can use the PEAR descriptor as the descriptor for the collectionReader section of the CPE descriptor.

Thanks,
Girish
-----Original Message-----
From: Michael Baessler [mailto:mba@michael-baessler.de] 
Sent: Wednesday, January 23, 2008 8:57 AM
To: uima-user@incubator.apache.org
Subject: Re: Programmatically check for correctness of Aggregate-TAE

What about checking the aggregate descriptor settings before you 
initialize it... like:

       File aggDesc = new File("aggregateDescriptor.xml");
       XMLInputSource in = new XMLInputSource(aggDesc);
       AnalysisEngineDescription aeDesc = 
UIMAFramework.getXMLParser().parseAnalysisEngineDescription(in);
       ConfigurationParameterSettings paramSettings = 
aeDesc.getAnalysisEngineMetaData().getConfigurationParameterSettings();

So you can check if the necessary settings are available if not you can 
throw an exception. If all is fine you can instantiate the ae.

Does this help?

-- Michael

Christoph Büscher wrote:
> Hi List,
>
> I would like to be able to check the parameter settings of one of my 
> custom aggregate TAEs on initialization.
> The general idea is that the component engines of the aggregate are 
> configured in the aggregates XML descriptor via "overrides". However 
> there are a few constraints on the parameter settings that I would 
> like to check when the aggregate is initialized.
> So far I don't know if and how it is possible to provide an own Java 
> implementation for the aggregate TAE. If this is possible, the 
> appropriate place would be the aggregates initialize() method, since 
> the configuration is accessible there. Can anyone tell me if it is 
> possible to have an implementation for an aggregate or not and if so, 
> how? If this isn't possible, where would be another good place to do a 
> sanity check on the aggregates parameter settings?
>
> Many thanks in advance,
>
> Christoph

Re: Can a PEAR component be used as a Collection Reader?

[Michael Baessler <mb...@michael-baessler.de>]

No that's currently not possible... but I agree, it would be nice.

-- Michael

Chavan, Girish wrote:
> I am trying to use a PEAR component as a collection reader in a CPE descriptor.
>
> I am wondering if I can use the PEAR descriptor as the descriptor for the collectionReader section of the CPE descriptor.
>
> Thanks,
> Girish
> -----Original Message-----
> From: Michael Baessler [mailto:mba@michael-baessler.de] 
> Sent: Wednesday, January 23, 2008 8:57 AM
> To: uima-user@incubator.apache.org
> Subject: Re: Programmatically check for correctness of Aggregate-TAE
>
> What about checking the aggregate descriptor settings before you 
> initialize it... like:
>
>        File aggDesc = new File("aggregateDescriptor.xml");
>        XMLInputSource in = new XMLInputSource(aggDesc);
>        AnalysisEngineDescription aeDesc = 
> UIMAFramework.getXMLParser().parseAnalysisEngineDescription(in);
>        ConfigurationParameterSettings paramSettings = 
> aeDesc.getAnalysisEngineMetaData().getConfigurationParameterSettings();
>
> So you can check if the necessary settings are available if not you can 
> throw an exception. If all is fine you can instantiate the ae.
>
> Does this help?
>
> -- Michael
>
> Christoph Büscher wrote:
>   
>> Hi List,
>>
>> I would like to be able to check the parameter settings of one of my 
>> custom aggregate TAEs on initialization.
>> The general idea is that the component engines of the aggregate are 
>> configured in the aggregates XML descriptor via "overrides". However 
>> there are a few constraints on the parameter settings that I would 
>> like to check when the aggregate is initialized.
>> So far I don't know if and how it is possible to provide an own Java 
>> implementation for the aggregate TAE. If this is possible, the 
>> appropriate place would be the aggregates initialize() method, since 
>> the configuration is accessible there. Can anyone tell me if it is 
>> possible to have an implementation for an aggregate or not and if so, 
>> how? If this isn't possible, where would be another good place to do a 
>> sanity check on the aggregates parameter settings?
>>
>> Many thanks in advance,
>>
>> Christoph
>>