compareTool.py

import csv
import logging
from DataCleaning.library.functions.changeDirectory import *
from floatFormatter import *
from DataCleaning.library.functions.addPath import *

addPath("config")
from config import *

def compareTool():
    fname1 = "proof.csv"
    fname2 = outputFileName
    ind1 = [0, 1, 2]
    ind2 = [0, 1, 2]


    # read file proof.csv
    development2Proof()
    f = open(fname1, 'r', encoding='UTF-8')
    csvreader = csv.reader(f)
    file1 = list(csvreader)
    f.close()
    proof2Development()

    # read .txt file
    development2Processed()
    with open(fname2, "r",  encoding='utf-8') as f:
        file2 = f.readlines()
    processed2Development()

    # create two empty sets
    set1 = set()
    set2 = set()

    # file 1
    for i in range(len(file1)):
        tempList = file1[i]
        tempList0 = []
        for j in ind1:
            tempElement = tempList[j]
            tempList0.append(tempElement)
        tempText = "|".join(tempList0)
        set1.add(tempText)

    # file 2
    for i in range(len(file2)):
        tempList = file2[i].strip("
").split("|")
        tempList0 = []
        for j in ind2:
            tempElement = tempList[j]
            tempList0.append(tempElement)
        tempText = "|".join(tempList0)
        set2.add(tempText)
    return set1, set2