import pandas as pd

from sklearn.metrics.pairwise import cosine_similarity

from sklearn.feature_extraction.text import CountVectorizer

from sklearn.feature_extraction.text import TfidfVectorizer

from nltk.corpus import stopwords

import random

import json

import pickle

import numpy as np

import string

import timeit

import csv


import nltk

from nltk.stem import WordNetLemmatizer



class AIChatbot:


    def clean_up_sentence(self, sentence):

        lemmatizer = WordNetLemmatizer()

        sentence_words = nltk.word_tokenize(sentence)

        sentence_words = [lemmatizer.lemmatize(word) for word in sentence_words]

        return sentence_word



    def train_chat(self, json_file_path, tfidf_vectorizer_pikle_path , tfidf_matrix_train_pikle_path):

        

        questions = []

        with open(json_file_path,'r') as data:

            for line in csv.reader(data):

                questions.append(line[0])

        
        start = timeit.default_timer()


        tfidf_vectorizer = TfidfVectorizer()

        tfidf_matrix_train = tfidf_vectorizer.fit_transform(questions)

        y = tfidf_matrix_train.toarray()


        stop = timeit.default_timer()

        print ("training time took was : ")

        print (stop - start)

       
        f = open(tfidf_vectorizer_pikle_path, 'wb')

        pickle.dump(tfidf_vectorizer, f)

        f.close()


        f = open(tfidf_matrix_train_pikle_path, 'wb')

        pickle.dump(tfidf_matrix_train, f)

        f.close()


        return tfidf_vectorizer , tfidf_matrix_train

       


    def talk_to_cb_primary(self, message, minimum_score, tfidf_vectorizer_pikle_path ,tfidf_matrix_train_pikle_path):

    
        json_file_path = "/root/chatbot/path/QA-WA.csv"

 
        tfidf_vectorizer_pikle_path = "/root/chatbot/contents/ai_vectorizer.pickle"

        tfidf_matrix_train_pikle_path ="/root/chatbot/contents/ai_matrix_train.pickle"


        test_set = (message, "")
        

        try:

            
            f = open(tfidf_vectorizer_pikle_path, 'rb')

            tfidf_vectorizer = pickle.load(f)

            f.close()


            f = open(tfidf_matrix_train_pikle_path, 'rb')

            tfidf_matrix_train = pickle.load(f)

            f.close()


        except:

            # ---------------to train------------------#

            tfidf_vectorizer , tfidf_matrix_train = self.train_chat(json_file_path , tfidf_vectorizer_pikle_path , tfidf_matrix_train_pikle_path)


        tfidf_matrix_test = tfidf_vectorizer.transform(test_set).toarray()

        cosine = cosine_similarity(tfidf_matrix_test, tfidf_matrix_train)

        max1 = cosine.max()

        response_index = 0


        if (max1 > minimum_score):


            row1 = cosine[0]


            response_index = np.where(row1 == max1)


            count = 0

            with open(json_file_path, "r") as data :

                for line in csv.reader(data, delimiter='"'):

                    if count == response_index[0][0]:

                   # print(line[0], line[1])

                        return line[1]

                    else:

                        count = count + 1

        
        else:

     
            return "If you want to connect to our agent, Please reply with Y or Yes."

            
        count = 0

        with open(json_file_path, "r") as data :

            for line in csv.reader(data, delimiter='"'):

                if count == response_index[0][0]:

                    return line[1]

                else:

                    count = count + 1

   
    def previous_chats(self, query):

        minimum_score = 0.55


        tfidf_vectorizer_pikle_path = "aii_previous_vectorizer.pickle"

        tfidf_matrix_train_path = "aii_previous_matrix_train.pickle"

        query_response= self.talk_to_cb_primary(query , minimum_score , tfidf_vectorizer_pikle_path , tfidf_matrix_train_path)

        return query_response


    def runbot(self, msg):

        return self.previous_chats(msg)