Grok

A whole bunch of stuff was cut out of this release. The preprocessing
components now use the new XML structure of NLPDocuments appropriately.

Improvements to the Grokling frontend, resurrection of parsing, updates for changes to unification architecture of OpenNLP, use of XML in preprocessing components conforms with JDOM, many improvements to sentence detection code, and new samples in the source tree.

A problem in the build.xml file that was causing the maxent models to be corrupted when the grok jar was created was fixed. The infrastructure for creating and releasing the Grok homepage was improved.

Improvements were made to the sentence detection code. All maxent models were
converted to the new single file format. The part-of-speech tagging model is
trained on much more data and uses a number of word stems to avoid some errors.
Grok was updated for changes in the OpenNLP API. Grok now uses Ant to build.
The directory structure has been changed considerably to facilitate the
creation of distribution files, and to ensure that source code and supporting
libraries are provided with distributions. Note: The lexicon system is broken
in this version, and therefore it cannot be used to parse.

Initial release.