import csv

import pymorphy2

import re

import psycopg2

conn = psycopg2.connect(dbname='energy', user='mao', password='daring', host='localhost')

cursor = conn.cursor()

morph = pymorphy2.MorphAnalyzer(lang='ru')

answer_id=[] 

answer = dict()

cursor.execute('SELECT id, answer FROM app.chats_answer;')

records = cursor.fetchall()

for row in records:

 answer[row[0]]=row[1]

questions=[] 

cursor.execute('SELECT question, answer_id FROM app.chats_question;')

records = cursor.fetchall()

transform=0

for row in records:

 if row[0]>"":

  if row[1]>0:

   phrases=row[0]

   words=phrases.split(' ')

   phrase=""

   for word in words:

    word = morph.parse(word)[0].normal_form  

    phrase = phrase + word + " "

   if (len(phrase)>0):

    questions.append(phrase.strip())

    answer_id.append(row[1])

    transform=transform+1

#print (questions)

#print (answer)

#print (answer_id)

cursor.close()

conn.close()

import sklearn

from sklearn.feature_extraction.text import TfidfVectorizer

from sklearn.decomposition import TruncatedSVD

vectorizer_q = TfidfVectorizer()

vectorizer_q.fit(questions)

matrix_big_q = vectorizer_q.transform(questions)

print ("Размер матрицы: ")

print (matrix_big_q.shape)

if transform>200:

 transform=200

print(transform)

svd_q = TruncatedSVD(n_components=transform)

svd_q.fit(matrix_big_q)

matrix_small_q = svd_q.transform(matrix_big_q)

print ("Коэффициент уменьшения матрицы: ")

print ( svd_q.explained_variance_ratio_.sum())

# тело программы k=5, temperature=10.0 можно подбирать

import numpy as np

from sklearn.neighbors import BallTree

from sklearn.base import BaseEstimator

def softmax(x):

  #создание вероятностного распределения

  proba = np.exp(-x)

  return proba / sum(proba)

class NeighborSampler(BaseEstimator):

  def __init__(self, k=5, temperature=10.0):

    self.k=k

    self.temperature = temperature

  def fit(self, X, y):

    self.tree_ = BallTree(X)

    self.y_ = np.array(y)

  def predict(self, X, random_state=None):

    distances, indices = self.tree_.query(X, return_distance=True, k=self.k)

    result = []

    for distance, index in zip(distances, indices):

      result.append(np.random.choice(index, p=softmax(distance * self.temperature)))

    return self.y_[result]

from sklearn.pipeline import make_pipeline

ns_q = NeighborSampler()

ns_q.fit(matrix_small_q, answer_id) 

pipe_q = make_pipeline(vectorizer_q, svd_q, ns_q)

import re

import telebot

telebot.apihelper.ENABLE_MIDDLEWARE = True

bot = telebot.TeleBot("299999999:sdfgnreognrtgortgmrtgmrtgm")

@bot.message_handler(commands=['start'])

def start_message(message):

	bot.send_message(message.from_user.id, " Здравствуйте. Я виртуальный помощник Mao?")

@bot.message_handler(func=lambda message: True)

def get_text_messages(message):

	request=message.text

	words= re.split('\W',request)

	phrase=""

	for word in words:

		word = morph.parse(word)[0].normal_form  

		phrase = phrase + word + " "

	reply_id    = int(pipe_q.predict([phrase.strip()]))

	bot.send_message(message.from_user.id, answer[reply_id])

	print("Запрос:", request, " \n\tНормализованный: ", phrase, " \n\t\tОтвет :", answer[reply_id])

bot.infinity_polling(none_stop=True, interval=1)

print("Пишите ваш вопрос, слова exit или выход для выхода")

request=""

while request not in ['exit', 'выход']:

 request=input()

 words= re.split('\W',request)

 phrase=""

 for word in words:

  word = morph.parse(word)[0].normal_form  

  phrase = phrase + word + " "

 reply_id    = int(pipe_q.predict([phrase.strip()]))

 print (answer[reply_id])