I am using a commercial software for tagging entities and relations.
That software generates offset text annotations in an own proprietary XML format.
I need to be able to do analytics on it using KNIME, therefore I needed to import it.
In less than 1 hour I've then extended the org.knime.ext.textprocessing package with a new org.knime.ext.textprocessing.nodes.source.parser that extends org.xml.sax.helpers.DefaultHandler and implements org.knime.ext.textprocessing.nodes.source.parser.DocumentParser, and that appropriately handles my data.
I've then packaged it "by hand" and put the classes in the knime-textprocessing.jar, and gave it a try ... and it works.
Now, I would like to be able to share that piece of code with others ... and I was wondering if you could instruct me on how at best proceed. Furthermore, I would like to know how to debug it and test it: I did it "by hand" and "it works", however I would prefer to know more "guidelines" ...
More in general, I was also wondering if this strategy is the one that you recommend, or if you would prefer / advice instead that I create a separate plugin that depends upon your textprocessing package.
Looking forward your advice, and compliments for the framework you've created ! really useful and easy to extend !