import os
import numpy as np
import csv
import zipfile

"""
    You are expected to upload a e*******.zip file, inside you should have a folder Question_2 which contains 6 gradients.csv file.
    For verification, you can comment out the line10 to 13, and change the truth_path and ID to 'b', however, do note the grading process WILL contain these lines.
"""
ID = 'e012345678'
zip_ref = zipfile.ZipFile(ID+'.zip', 'r')
zip_ref.extractall('.')
zip_ref.close()

truth_path = 'the_truth_path' #change truth_path = 'b' for verification
file_name = ['Question_2/dw-100-40-4.csv', 'Question_2/db-100-40-4.csv', 'Question_2/dw-28-6-4.csv', 'Question_2/db-28-6-4.csv', 'Question_2/dw-14-28-4.csv', 'Question_2/db-14-28-4.csv']
true_file = ['true-dw-100-40-4.csv', 'true-db-100-40-4.csv', 'true-dw-28-6-4.csv', 'true-db-28-6-4.csv', 'true-dw-14-28-4.csv', 'true-db-14-28-4.csv']
threshold = 0.05

def read_file(name):
    l = list()
    with open(name) as f:
        reader = csv.reader(f)
        for row in reader:
            l.append(row)
    return l

"""
    You can try your grading function, while the function is yet to decided. However, in the ideal situation you should expect dif = 0
"""
def do_some_grading(l0, l1, th):
    dif = np.mean(np.abs(l0-l1).astype(float)/np.abs(0.1+l1))
    if dif <= th:
        return 1
    else:
        return 0

"""
    The threshold is introduced to address the numerial bias due to rounded floats, which could be as small as zero
"""
def compare(sub, true, threshold=0):
    scores = list()
    if not len(sub)==len(true):
        return 0
    for i in range(len(sub)):
        l0 = np.array(sub[i]).astype(np.float)
        l1 = np.array(true[i]).astype(np.float)
        if not len(l0)==len(l1):
            return 0
        else:
            scores.append(do_some_grading(l0, l1, threshold))
    return scores

true_grads = list()
for f in true_file:
    true_grads.append(read_file(os.path.join(truth_path,f)))

score = list()
for i, fn in enumerate(file_name):
    grads = read_file(os.path.join(ID,fn))
    s = compare(grads, true_grads[i], threshold)
    score += s
print(np.sum(score))