from janome.tokenizer import Tokenizer

import random

from janome.tokenizer import Tokenizer

# テキストデータを単語に分割する

def wakati(text):

    text = text.replace('\r','') #スペースを削除

    t = Tokenizer()

    result =t.tokenize(text, wakati=True)

    return result

def generate_text(num_sentence=200):

    src = open("txtファイルのパス", "r").read()

    wordlist = wakati(src)

    markov = {}

    w1 = ""

    w2 = ""

    for word in wordlist:

        if w1 and w2:

            if (w1, w2) not in markov:

                markov[(w1, w2)] = []

            markov[(w1, w2)].append(word)

        w1, w2 = w2, word

    #タイトルの自動生成

    count_kuten = 0

    num_sentence= num_sentence

    sentence = ""

    w1, w2  = random.choice(list(markov.keys()))

    while count_kuten < num_sentence:

        tmp = random.choice(markov[(w1, w2)])

        sentence += tmp

        if(tmp=='\n'):

            count_kuten += 1

            sentence += '\n'

        w1, w2 = w2, tmp

    print(sentence)

     
if __name__ == "__main__":

    generate_text()