opennlp
opennlp copied to clipboard
OPENNLP-957: [WIP] Adds a normalizer aggregate feature generator.
Lets discuss this a bit first, and that should be done on the dev list. There is also the case that multiple tokens should be joined or a token has to be split. And I soḿehow assumed it is possible to do this with the existing interface. Adding explicit normalization support makes it more complex, do we really need that?