Sahil Rajput

Posted on Nov 23, 2018

Build your first ChatBot in 5 minutes

#chatbot #python #easy

I was searching the internet on "How to build a Chatbot?" and I discovered ChatterBot which is a machine learning, conversational dialog engine for creating chat bots

How ChatterBot works

Image source: [ChatterBot](https://chatterbot.readthedocs.io/en/stable/#process-flow-diagram)

In this article we will see how to build a chatbot with ChatterBot in just 5 minutes.

Let's start

First Install ChatterBot

pip install ChatterBot

Create a file chat.py

#import ChatBot
from chatterbot import ChatBot

Create a new chat bot with the name of your choice(I am using "Candice").

bot = ChatBot('Candice')

Your bot is created but at this point your bot has no knowledge, for that you have to train it on some data.
Also, by default the ChatterBot library will create a sqlite database to build up statements of the chats.

Train your bot

#import ListTrainer
from chatterbot.trainers import ListTrainer

bot.set_trainer(ListTrainer)
# Training 
bot.train(['What is your name?', 'My name is Candice'])
bot.train(['Who are you?', 'I am a bot, created by you' ])

Your bot is now trained on 2 statements. When you ask your bot "what is your name", it will reply back with "My name is Candice".

You can also trained it on multiple statements like

bot.train(['Do you know me?', 'Yes, you created me', 'No', 'Sahil?', 'No idea'])

As you can see it is difficult to train the bot on every single statements. So, we will use ChatterBotCorpusTrainer to train our bot on the large dataset.

from chatterbot.trainers import ChatterBotCorpusTrainer
# Create a new trainer for the chatbot
trainer = ChatterBotCorpusTrainer(bot)

# Train the chatbot based on the english corpus
trainer.train("chatterbot.corpus.english")

# Get a response to an input statement
chatbot.get_response("Hello, how are you today?")

or you can download the dataset of your language and train your bot in your language.

I downloaded the english language dataset and we can train our bot like this

for files in os.listdir('./english/'):
    data=open('./english/'+files,'r').readlines()
    bot.train(data)

NOTE: Make sure the dataset and the program file is on same folder, otherwise edit the path.

Chat feature

# To exit say "Bye"
while True:
        # Input from user
    message=input('\t\t\tYou:')
        #if message is not "Bye"
    if message.strip()!='Bye':
        reply=bot.get_response(message)
        print('Candice:',reply)
        # if message is "Bye"
    if message.strip()=='Bye':
        print('Candice: Bye')
        break

To run it, Go to terminal

python chat.py

It will train your bot first and then you can start chatting.

Source code

#import libraries
from chatterbot import ChatBot
from chatterbot.trainers import ListTrainer
import os

#Create a chatbot
bot=ChatBot('Candice')
bot.set_trainer(ListTrainer)

#training on english dataset
for files in os.listdir('./english/'):
    data=open('./english/'+files,'r').readlines()
    bot.train(data)

#chat feature
while True:
    message=input('\t\t\tYou:')
    if message.strip()!='Bye':
        reply=bot.get_response(message)
        print('Candice:',reply)
    if message.strip()=='Bye':
        print('Candice: Bye')
        break

Top comments (11)

kpspersonal • Jul 16 '19 • Edited

do pip install chatterbot-corpus
Then
Make the following changes to your code
remove the following lines
bot.set_trainer(ListTrainer)
bot.set_trainer(ChatterBotCorpusTrainer)
bot.train("chatterbot.corpus.english")

Add the following
trainer = ChatterBotCorpusTrainer(bot)
trainer.train('chatterbot.corpus.english')

Remove the following
training on english dataset
for files in os.listdir('./english/'):
data=open('./english/'+files,'r').readlines()
data=open('./english/'+files,'r').readlines()
trainer.train(data)
It should work now.

KarthiJi • May 13 '19 • Edited

im not able to execute this code, cud u help me out with that..? can dm me abt it..?

Sahil Rajput • May 13 '19

What error you are facing?

KarthiJi • May 13 '19

cant import error

Sahil Rajput • May 17 '19

Download all the libraries.

Oshan Wisumperuma • Feb 17 '19

I faced an issue when training using corpus files through "ListTrainer". seems even "categories" and "conversations" tags in yml file also get inserted into the data source.