Todo: incorporate the extracted ngrams and link them to the given document.

See mat-parsing/test.py for more info about how it works.

See test.py for examples.

Thought of an improved design pattern for the tagger, for a better integration in the current workflow.

Changed the models a little.

Added a Resource database object.

Allows parsing from various sources.

For now, two taggers can be used, using a common interface:
* TreeTagger
* NltkTagger

Towards node db