staticFeatureProcessor.py

#!/usr/bin/env python

__author__ = 'AR'

import os, hashlib, json, sys, time
from collections import Counter
from datetime  import datetime as dt

# Usage Message
if len(os.sys.argv) < 5:
    print "Usage: ./staticfeatureProcessor -i <source to folder with strace files> -o <destination to store output files>"
    os.sys.exit()

for i in xrange(len(sys.argv)):
    if sys.argv[i] == '-i':
        sourcePath = sys.argv[i + 1]  # Path to folder containing strace files
    if sys.argv[i] == '-o':
        destPath = sys.argv[i + 1]

def main():
 	files = [x for x in os.listdir(sourcePath) if x.endswith('.static')]

 	featureDict = {}

 	# Define the keys for static features with values as list 
 	if files:
 		for f in files:
 			with open(sourcePath + '\\' + f, 'r') as f:
 				for line in f.readlines():
 					if ':' in line:
 						featureDict[line[:line.index(':')]] = [] #line[33:-1]

 	# Adding values to the key's list
 	if files:
 		for f in files:
 			with open(sourcePath + '\\' + f, 'r') as f:
 				for line in f.readlines():
 					if ':' in line:
 						if line[33:-1] not in featureDict[line[:line.index(':')]]:
 							m = hashlib.sha256()
 							m.update(str(line[33:-1]))
	 						featureDict[line[:line.index(':')]].append(m.hexdigest())

	# Dumping the dictionaty to the 
	with open(destPath + 'StaticfeatureDict.json', 'w') as featureListFile:
		json.dump(featureDict, featureListFile)
 	#print featureDict



if __name__=="__main__":
    print '{:*^70}'.format(" Machine Learning and Malware Classification ")
    main()
    print '{:*^70}'.format('All Strace stored to ' + destPath + ' directory')
    print
    print '{:*^70}'.format(" EOP ")