Mailing-List: contact uima-user-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: uima-user@incubator.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
Delivery-date: Wed, 24 Oct 2007 14:53:12 +0200
Message-ID: <471F3F53.3050808@averbis.de>
Date: Wed, 24 Oct 2007 14:49:23 +0200
From: Philipp Daumke <general@averbis.de>
User-Agent: Thunderbird 2.0.0.6 (Windows/20070728)
MIME-Version: 1.0
To: uima-user@incubator.apache.org
Subject: Re: "Global" Parameters for a CPE
References: <471CEC8F.3010300@averbis.de> <471EEBB3.2060408@gmx.de>
 <0DBCCB475CDE864F8F6086D69BFC5D9FF0626E@CALLISTO.ntdom.tk.informatik.tu-darmstadt.de>
In-Reply-To: 
 <0DBCCB475CDE864F8F6086D69BFC5D9FF0626E@CALLISTO.ntdom.tk.informatik.tu-darmstadt.de>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hi Torsten,

I had the same idea and currently do it like that. But if you work with 
millions of documents and you want to initialise (or maybe clean the 
folder)  just at  the beginning it's really a hack to write the folder 
information in millions of CASes.

Another idea that I had was whether it's possible to access the 
CollectionReaderXML from the CasConsumer automatically. Is there 
something like getCPEComponents().getCollectionReader.getMetaData() 
which you can call from any CPEComponent? I couldn't find it so far.

Thanks for your help
Philipp
>>> I want to read some files from a certain directory and 
>>>       
>> write the results
>>     
>>> back in the same folder (in a subdirectory, to be more 
>>>       
>> exactly). (How)
>>     
>>> can I define parameters that are both accessible in a 
>>>       
>> CollectionReader
>>     
>>> and in a Cas Consumer? Or do I have to define the folder parameter
>>> twice, first in the CollectionReaderXML and second in the 
>>>       
>> CasConsumerXML?
>>     
>>> Thanks for your help
>>> Philipp
>>>       
>> Excellent question.  I tried to figure this out from our
>> documentation and couldn't.  Anyone?
>>     
>
> As a workaround, one could define the parameter in the CollectionReader
> and add a document annotation that holds the path. The consumer then
> reads the path from the document annotation, so it knows where to put
> the results.
>
> -Torsten
>
>