import numpy as np
import pandas as pd


def load_csv():
    return pd.read_csv("H:\Misc\MSc\Data\ForDiscrete.csv")

def create_quartiles(df):
    curr_df = df

    for each_variable in curr_df:
        values = curr_df[each_variable]

        # find quartiles
        Q1 = np.percentile(values,25)
        Q2 = np.percentile(values,50)
        Q3 = np.percentile(values,75)

        # replace values with quartile range
        factor_df = pd.DataFrame(curr_df[each_variable])

        for index,row in factor_df.iterrows():
            if float(row) <= Q1:
                factor_df.loc[index, each_variable] = 1
            elif float(row) >= Q1 and float(row) <= Q2:
                factor_df.loc[index, each_variable] = 2
            elif float(row) >= Q2 and float(row) <= Q3:
                factor_df.loc[index, each_variable] = 3
            elif float(row) >= Q3:
                factor_df.loc[index, each_variable] = 4
            else:
                print('check error')

        curr_df[each_variable] = factor_df[each_variable].values

    return curr_df

def write_to_csv(df):
    loc_df = pd.DataFrame(df)
    print(loc_df.to_string())
    #loc_df.to_csv(path_or_buf="C:\\Users\\cdogan\\DataOutput")


if __name__ == '__main__':
    discrete_df = create_quartiles(load_csv())
    write_to_csv(discrete_df)