# gzip_find_nearest.py

# gzip_find_nearest.py *** fixed!

import gzip

def predict(test_set, training_set):
    predicted = 0
    test_number = 0
    for x1 in test_set:
        distance_from_x1 = []
        for x2 in training_set:
            Cx2 = len(gzip.compress(x2.encode()))
            x1x2 = " ".join([x1, x2])
            Cx1x2 = len(gzip.compress(x1x2.encode()))
            ncd = (Cx1x2 - min(len(x1.encode()), Cx2)) / max(len(x1.encode()), Cx2)
            distance_from_x1.append((ncd, x2))
        sorted_distances = sorted(distance_from_x1)

        guess = sorted_distances[0][1]

        print(f"\nAltered Test: {x1}")
        match_percentage = round((100-sorted_distances[0][0]),6)
        print(f"\tGuessing Nearest Match: {guess} at {match_percentage}%")
        if answers[x1] == guess:
            print("\t+++ CORRECT +++")
            predicted = predicted + 1
        else:
            print(f"\t\tExpected Answer: {answers[x1]}")
        test_number = test_number + 1
        print(f"Predicted: {predicted} out of {test_number}  Errors: {test_number - predicted}")

# for testing
import random
import string

letters = list(string.ascii_letters)
L = len(letters)

tests = []
def generate_tests(num_tests):
    i = 0
    for _ in range(num_tests):
        test = ''
        for _ in range(9):
            s = letters.pop(i%11)
            letters.append(s)
            i += 1
            test += s
        tests.append(test)
        s = letters.pop(i%7)
        letters.append(s)

def alter_test(test):
    pos = random.sample(range(9), 2) # choose two positions to change
    new_test = list(test)
    new_test[pos[0]] = letters[(letters.index(new_test[pos[0]]) + random.randint(1, L - 1)) % L]
    new_test[pos[1]] = letters[(letters.index(new_test[pos[1]]) + random.randint(1, L - 1)) % L]
    return ''.join(new_test)

# generate a list of n random n-letter tests
generate_tests(1000)

# make a copy of each test with two letters changed
altered_tests = [alter_test(test) for test in tests]
answers = dict(zip(altered_tests, tests))
random.shuffle(tests)

# predict nearest matches for each altered test
predict(altered_tests, tests)