EncodeFeatures.py

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Tue Jun  5 15:33:41 2018

@author: gowtham
"""

import os
import pandas as pd
from EncodeAFeature import encodeFeature
import ast

permissions = pd.read_csv('Permissions.csv', header = None)
permissions_l = list(permissions.iloc[:, 0].values)

uses_features = pd.read_csv('Uses-Features.csv', header = None)
uses_features_l = list(uses_features.iloc[:, 0].values)

actions = pd.read_csv('intents.csv', header = None)
actions_l = list(actions.iloc[:, 0].values)

cwd = os.getcwd()
dirs = os.listdir(cwd + "/malwareF")

dirs.remove("ENC")

x = 1

for tsv in dirs:
    print str(x) + " - " + tsv
    x = x+1
    csvf = open("malwareF/ENC/" + tsv[:-4]+"_ENC.tsv", "w")
    data = pd.read_csv("malwareF/" + tsv, sep = '\t')
    for i in encodeFeature(data, permissions_l, 4, 548):
        csvf.write(str(i) + "\t")




    FeauX = data.iloc[:, 5].values
    b = []
    i = 0
    while(not pd.isnull(FeauX[i]) and (not data.shape[0] == i+1) ):
        dic = ast.literal_eval(FeauX[i])
        if 'name' in  dic.keys():
            if dic['name'] in uses_features_l:
                b.append(uses_features_l.index(dic['name']))
            else:
                b.append(73)
        else:
            b.append(73)
        
        i = i +1
    
    encoded_lst = []
    
    for i in range(74):
        encoded_lst.append(0)
        
    for i in b:
        encoded_lst[int(i)] = encoded_lst[int(i)] + 1
    
    for i in encoded_lst:
        csvf.write(str(i) + "\t")
        
    
    
    FeauX = data.iloc[:, 0].values
    b = []
    i = 0
    while(not pd.isnull(FeauX[i]) and (not data.shape[0] == i+1)):
        i = i+1
    
    csvf.write(str(i) + '\t')
    
    for i in encodeFeature(data, actions_l, -1, 291):
        csvf.write(str(i) + "\t")
    
    
    
csvf.close()