Well, I said it may be easier to use the factory methods of uimaFIT
if you want to use ClearTK and uimaFIT at the same time.

I am pretty sure that you can use ClearTK just with descriptors in the
traditional UIMA way, without uimaFIT. In fact, I'd be very surprised
if there was anything preventing that. I don't know if the ClearTK folks
ship such descriptors. You'd better ask on their mailing list I suppose.

But still, regarding your very first message in the error. I looked quite
right what you had, I just had the impression, that you didn't properly
configure your reader. And the part where you set up the reader wasn't
included in your code snippet.

Cheers,

-- Richard

Am 08.05.2013 um 17:28 schrieb harshal patni <[email protected]>:

> Oopps! So then is there a way to integrate ClearTK in UIMA directly?
> 
> Harshal
> 
> 
> On Wed, May 8, 2013 at 8:51 PM, Richard Eckart de Castilho <
> [email protected]> wrote:
> 
>> Am 08.05.2013 um 17:05 schrieb harshal patni <[email protected]>:
>> 
>>> Hi Richard,
>>>                Thanks a lot. That helped! I replaced the following
>>> function -
>>> 
>>> AnalysisEngineFactory.createAnalysisEngineFromPath(
>>> "desc/AggregateAnalysisEngine.xml");
>>> 
>>> Now this produces an AnalysisEngine. How do I convert from AnalysisEngine
>>> to AnalysisEngineDescription. SimplePipeline.runPipeline only takes
>>> AnalysisEngineDescription.
>>> 
>>> 
>>> Also I would like to get an annotated XML as an output. How would I do
>> that?
>> 
>> That's currently a "hole" in the uimaFIT API. You can resort to plain UIMA
>> for loading descriptors, as you did initially.
>> 
>> See also
>> http://mail-archives.apache.org/mod_mbox/uima-user/201304.mbox/%[email protected]%3e
>> 
>> I'll add some convenience here for uimaFIT 2.0.0
>> 
>> https://issues.apache.org/jira/browse/UIMA-2885
>> 
>> Cheers,
>> 
>> -- Richard

Reply via email to