Bangla Dependency Parser
=========================

Bangla dependency parser using maltparser.

Installation and run:
=========================

1. Download Maltparesr version 1.7 from : http://www.maltparser.org/index.html

2. The system requires Java 1.5+ to be installed.
   Set JAVA_HOME to environment variable by typing 
                     export JAVA_HOME=/path/to/Java
   to the end of the file bash.bashrc

   This can be done as follows:	

   cd /etc
   su
   Enter your root password
   gedit bash.bashrc

3. Download bengali_stack.mco into your working directory and execute the following command: 
   prompt> $java -jar maltparser-1.7.1.jar -c bengali_stack -i examples/data/inputfile.conll -o outfile.conll -m parse
   
   Here is an exemple of an input sentence:

1	উৎপল	_	NP	NNP	_	_	_
2	স্যার	_	NP	NN	_	_	_
3	ভালো	_	JJP	JJ	_	_	_
4	লোক	_	NP2	NN	_	_	_
5	।	_	NULL__VGF	SYM	_	_	_

NOTE : That the columns are tab-separated and column 4 contains the POS Tag information and column 5 contains chunking information.

   Corresponding output file:

1	উৎপল	_	NP	NNP	_	5	k1	_	_
2	স্যার	_	NP	NN	_	5	k1	_	_
3	ভালো	_	JJP	JJ	_	5	k1s	_	_
4	লোক	_	NP2	NN	_	5	k1s	_	_
5	।	_	NULL__VGF	SYM	_	0	NULL__VGF	_	_

NOTE: Columns 7 and 8 contains the dependency relations of the above sentece.


Questions:
=====================

arjundas.cs@gmail.com