import re
import numpy as np
from scipy import spatial as sp

amount_of_words = 0
amount_of_sentences = 0
sentences = []
freq = {}
text = open("2.3.1.txt", 'r')

for line in text:
    amount_of_sentences += 1
    current = re.split('[^a-z]', line.lower())
    result = []
    for cur in current:
        if cur != '':
            if freq.get(cur) is None:
                amount_of_words += 1
                cur_list = [0 for i in range(amount_of_sentences - 1)]
                cur_list.insert(amount_of_sentences - 1, 1)
                freq[cur] = cur_list
            else:
                cur_list = freq[cur]
                length = len(cur_list)
                if length < amount_of_sentences:
                    for i in range(length, amount_of_sentences - length - 1):
                        cur_list.insert(i, 0)
                    cur_list.insert(amount_of_sentences - 1, 1)
                else:
                    cur_list[amount_of_sentences - 1] += 1
                freq[cur] = cur_list
            result.append(cur)
    sentences.append(result)
text.close()

matrix_freq = np.zeros((amount_of_sentences, amount_of_words))
i = 0

for value in freq.values():
    length = len(value)
    for j in range(length):
        matrix_freq[j, i] = value[j]
    for j in range(length, amount_of_sentences):
        matrix_freq[j, i] = 0
    i += 1

results = []
comp = matrix_freq[0]
min1 = 1
min2 = 1
min1_index = 0
min2_index = 0
for i in range(1, amount_of_sentences):
    x = sp.distance.cosine(comp, matrix_freq[i])
    results.insert(i, x)
    if x <= min1:
        min1 = x
        min1_index = i
    elif x <= min2:
        min2 = x
        min2_index = i

print(freq.keys())
for i in matrix_freq:
    print(i)
# print(results)
# results.sort()
# print(results)
# print(min1, min2)
# print(min1_index, min2_index)

res = open('2.3.1.result.txt', 'w')
res.write(str(min1_index) + ' ' + str(min2_index))
res.close()