Menu

Custom OpenNLP-UIMA annotator

Help
Kamil
2010-10-04
2013-04-16
  • Kamil

    Kamil - 2010-10-04

    Hi everyone,

    I'd like to use the OpenNLP Sentence detector, Tokenizer and POS Tagger as part of my UIMA process.
    The sample PEAR  of the opennlp.uima package is going well but I'm not able to customize the
    annotator, i.e. to use models for diferent languages.

    What I've done so far:

    1) new project with uima nature
    2) opennlp-uima.jar on project's buid-path
    3) Copy&Paste Descriptors to desc folder
    4) In tab resources: add Model an bind to ModelName

    When I load the descriptor to CAS Visual Debugger I always get an ResourceInitializationException.

    Is there any kind of documentation available for this problem additional to the readme file included in the package?
    I am thankful for any advise.

    Kamil

     
  • Joern Kottmann

    Joern Kottmann - 2010-10-04

    Did not reproduced your steps, but there is most likely something wrong with the way the resources are mapped.

    I suggest that you take the existing generated pear (for 1.5). Its build when you do a mvn install.
    Unzip the pear and modify it to your needs.
    Which models do you want to replace ?

    Otherwise please post the exception, than we can help you to get that model loaded correctly.

    Right now there is not more documentation, but I will work on it.

    Jörn

     
  • Kamil

    Kamil - 2010-10-05

    Thank you very much, I followed your steps and it works perfectly now.

    Regards
    Kamil

     

Log in to post a comment.