OpenNLP

[Michael Parco <33...@cua.edu>]

Greetings,

I am working with the openNLP libraries in the datafu-pig jar; however, I
am having trouble getting any of the openNLP UDFs to run properly. It
appears from the errors I am seeing that the training models ('data/
da-token.bin <http://opennlp.sourceforge.net/models-1.5/da-token.bin>')
etc. need to be applied to the UDFs, but are not provided through the
datafu jar itself? I have downloaded some of the trained models separately,
but still not luck. Any guidance for usage here?

Re: OpenNLP

[Matthew Hayes <ma...@gmail.com>]

Correct, the models aren't provided by the JAR itself. It wouldn't be practical to do so. You have to download them separately from the page you referenced and upload them to HDFS. If you then provide the path to the model in HDFS to the UDF it should work. Let me know if you get it working. I'll file a task to improve documentation.

-Matt

> On Mar 10, 2016, at 6:57 AM, Michael Parco <33...@cua.edu> wrote:
> 
> Greetings,
> 
> I am working with the openNLP libraries in the datafu-pig jar; however, I
> am having trouble getting any of the openNLP UDFs to run properly. It
> appears from the errors I am seeing that the training models ('data/
> da-token.bin <http://opennlp.sourceforge.net/models-1.5/da-token.bin>')
> etc. need to be applied to the UDFs, but are not provided through the
> datafu jar itself? I have downloaded some of the trained models separately,
> but still not luck. Any guidance for usage here?