Hyderabad Treebank Extension for TrEd

This extension allows to convert data from the Hyderabad Treebank (tested on ICON 09 data) to PML and visualise and process them in TrEd. It has two separate scripts to convert either the original *.morph data, or the converted data in CoNLL-ST format.

Morph Data

Use

bin/hydt-morph2pml.perl [--max-sentences|-s num] language file1.morph [file2.morph...]

It is advisable to split large files into smaller ones to improve TrEd's performance.

CoNLL Data

Use

bin/hydt_conll_to_pml.sh language output-dir hydt-conll-file

ICON 2010 Data

The script bin/icon2010-preprocess.perl converts the 2010 ssf format to 2009 ssf format, so you can use the above script. For Hindi data, you can use the above mentioned CoNLL version.