This directory holds the Python code needed to train, test and use the models described in this work. Some prerequisites are needed, among these are the Goldberg's parser (http://www.cs.bgu.ac.il/~yoavg/software/sdparser/), and Adler's tagger (included with the previous link). 
In addition some Python packages are also needed and not included in this package. These include mainly NLTK and Numpy.

In order to train a model run:

> Python Driver.py <parseType> <epochs> <taggedFile>

Where:
	parseType is Const or Dep
	ephocs is the number of epochs to train
	taggedFile is the corpus on which training is performed, composed in the same manner as the corpus provided in this site

Pretrained models are also supplied in ./Pretrained Models/ :
	   - Const_classifier
	   - Dep_classife