UIMA with Ruby?

We 开发者_Go百科are looking into doing some NLP with Ruby taking multiple document formats and using these for similarity testing.

We could use existing parts like Standford Parser, ruby stemmer etc to put this together but UIMA seems to be an industrial strength solution containing document import and preprocessing with indexing.

Has anybody used this? and in particular made any attempt to incorporate it into Ruby?

All opinions welcome.

You might want to write your NLP pipelines in UIMA and expose UIMA as a REST service. Another solution would be to use JRuby.