On 02/07/2014 07:49 PM, William Karl Thompson wrote:
OpenNLP has some code that takes brat annotation files and creates "BratAnnotation"
object instances. I've taken the code and modified it (simplified in some ways) to generate cTAKES
annotations, using a "BratAnnotator" analysis engine that reads in brat annotation files.
I would be happy to share that code with anyone who wants to look at it and make it better!
We worked quite a bit with brat for training the OpenNLP Name Finder. As
part of this project we the format support was developed.
In my opinion it works best if you just support the format your training
data is in and then train all the models on it directly, rather than
using converters to transform it into a different format for training.
It would probably be very useful for some to have a nice UIMA
integration to read in a brat corpus for training and validation.
In case you found any bugs, or have some suggestion please consider to
contribute your code changes back to OpenNLP.
Jörn
