Project description;;;

    The project is called MineSoft (short for Mining Software). 
    The project deals with knowledge extraction from software repositories.
    There are two techniques implemented here to extract knowledge from the code:

        (1) SWordNet: Inferring semantically related words from software context, EMSE Journal 2014
        (2) VerbNet: Automatically Mining Software-Based,Semantically-Similar Words from Comment-Code Mappings, MSR 2013

    Mainly you can extract pairs of semantically related words from the corpus (repositories)

How to run this program;;;

        This main script is the entry point for project. Usually called as:

        python3.4 main.py [-h] -method method_name -input input_dir -output output_dir

        -h, --help            show this help message and exit
        -method method_name, --method method_name
                            which method you want to use (either swordnet or verbnet)
        -input input_dir, --input input_dir
                            which directory is input (e.g. eclipse repo root path)
        -output output_dir, --output output_dir
                            which directory is output (i.e. where pairs will be stored)

Prerequisite;;;

	(1) Dependencies as defined in Gupta's POSSE:

		1. Perl required 
		   Tested on v5.14.2 

		2.WordNet should be installed. For Linux just do ``apt install wordnet''.

		Please provide the path to the WordNet binary in the script:

		./POSSE/Scripts/getWordNetType.sh

		Change line#2 

		/usr/bin/wordnet $1 | grep "Information available for \(noun\|verb\|adj\|adv\) $1" | cut -d " " -f4

		TO

		<pathToWordNetBinary> $1 | grep "Information available for \(noun\|verb\|adj\|adv\) $1" | cut -d " " -f4

	(2) Python 3 required 

	(3) Following Python libraries needed:
		
		nltk
		regex
		string

	(4) Java 1.8

	(5) nltk stanford pos tagger need nltk internals to be pointed to java 1.8
	
		Find the line in the verbnet.py that says:
		os.environ['JAVAHOME'] = "/usr/lib/jvm/java-1.8.0-openjdk-amd64/bin/java"
	
		And change the path to where your java 1.8 is residing

What does the output look like;;;

	The output is a csv file that has following columns:

		term1, term2, number of contexts, context1, context2, context3, ...

	term1 = the first term in the pair of related terms
	term2 = the second term in the pair of related terms
	number of contexts = in how many occurrences these terms term1 and term2 are found to be similar
	context1, context2, context3, ... = comma separated pair of contexts in which term1 appeared and term2 appeared