Telegram chatbot

Telegram chatbot created with deep learning model (LSTM) and telebot library.

Description

This program will allow you to create very easily a custom chatbot capable of sending texts, images or videos.
The deep learning model used is a Long Short Term Memory (LSTM) but you can change his structure if you want.

Getting Started

Install the libraries

Execute the following command : pip install -r requirements.txt

Create a bot in telegram

To do that, you have to open your telegram application. Then speak to this user : @BotFather. It is a bot created by telegram itself that allows you to manage the creation and the editing of your bots. Just follow the instructions and get the API token of your bot.
You can follow this tutorial : tutorial

Change the settings

Once you got the api token of your bot, in the my_config.py file change the value of TOKEN.

Last step

Before you can run this program, you will have to complete the most important file : ìntents.json In this file you will have to write all the sentences you would like the chatbot learn.

• tag : title of the question/answer. It does not matter for the model. This is just for you.
• type : You have 4 possibilities
  • text : the chatbot will answer only with a text message
  • file : the chatbot will send a text message followed by a document
  • photos : the chatbot will send a text message followed by a photo.
  • multiple_photos : the chatbot will send text messages with pictures attached.
• patterns : write all possible turns of phrase for a question
• responses : write the chatbot's answers. You can write more than one. The chatbot will choose randomly one of them.
• link : path to your documents and photos

Launch

Execute the following command in a terminal : python main.py

Heroku

You can add this program on HEROKU to let it work without interruption for free. The file Procfile is there for that.

You can now speak to your bot in your telegram application !