Directory structure:
====================

    model		containing CRF word segmenting model trained on 8000 Vietnamese sentences
    samples		containing sample input data files
    src			containing source code files
    Makefile		make file
    dist		containing binary libarary 	- the jar output file	
    JVnSegmenter.jar	binary libarary (output of compiling process)
    README		this one

For Netbean:
====================
    nbproject		containing neccessary files for the Netbean project	
    build		containing classes file if compiling with Netbean
    dist		containing binary libarary 	- the jar output file	
    build.xml		    	
    manifest.mf
    
How to compile in Linux:
=======================

    make -f [Make file name] clean
    make -f [Make file name] all

How to compile in Window:
========================
    Open project with Netbean and build it again	

How to test:
============

    make test
    
    after testing, go to samples directory to see the outputs. 
    the output data files are in UTF-8 and have words bounded by [ and ]
    

How to use JVnSegmenter:
=====================

    command line:
    
    case 1:
	
	java -mx512MB -classpath <CLASSPATH> JVnSegmenter.JVnSegmenter -modeldir <MODELDIR> -inputfile <INPUTFILE>
	
    case 2:
    
	java -mx512MB -classpath <CLASSPATH> JVnSegmenter.JVnSegmenter -modeldir <MODELDIR> -inputdir <INPUTDIR>
	
    where:
	
	-mx512MB: to extend heap memory for JVM
	
	<CLASSPATH> is the path to the "JVnSegmenter.jar" file. For instance, if we are currently at the home 
	directory of JVnSegmenter, the <CLASSPATH> is "dist/JVnSegmenter.jar"
	
	<MODELDIR> is the path to the trained model. For instance, if we are currently at the home directory
	of CRFChunker, the <MODELDIR> is "./model"
	
	<INPUTFILE> is the input data file with the format as follows:
	    - text file in UTF-8 format
	    - one sentence on each line
	    
	<INPUTDIR> is the directory contain multiple input data files. This option is useful when we 
	need to tag a lot of documents without loading the model multiple times. This helps reduce 
	computational time dramatically.
	

    example:
    
	if we are currently at the top (home) directory of JVnSegmenter and we would like to tag an input
	data file named "input.txt" which is also located at the top directory of JVnSegmenter, we perform:
	
	java -mx512MB -classpath ./JVnSegmenter.jar JVnSegmenter.JVnSegmenter -modeldir ./model -inputfile ./input.txt.pos
	
	the output will be "input.txt.vwseg"
	
	