About

NB: Grok is in the process of being reorganized into separately usable (and separately developed) OpenNLP components. Stay tuned -- watch the news on the Grok sourceforge page for where to find the new stuff.

The Grok project is dedicated to developing a large collection of basic tools for use in natural language software. A particularly important aspect of Grok is that its natural language modules should follow specific guidelines, or interfaces, so that they may be freely exchanged with other modules of the same type. To this goal, Grok provides a library of modules that implement the interfaces specified by OpenNLP.

The most developed aspect of Grok is its preprocessing subsystem, which includes components for doing tokenization, sentence detection, part-of-speech tagging and name finding. The parsing system is currently being revamped to support new extensions to the Combinatory Categorial Grammar formalism which Grok uses.

Ultimately, we see Grok as a useful tool for the development of natural language software and linguistic theory.