Overview
Hi everyone! I am Corrado and today I'll be sharing with you how to create a Telegram chat bot that integrates with the OpenAI API. In this dev.to post, we'll go through the step-by-step process of building a fully functional chatbot that uses artificial intelligence to respond to user inputs. Here's what we're going to cover:
- Setting up a Flask app on your local machine
- Create a Telegram Bot
- Connecting the Flask app to the Telegram Bot
- Writing a function to make requests to OpenAI API
- Deploying the app on Deta.space
By following these steps, you'll be able to create a Telegram bot that can hold intelligent conversations with users by leveraging the power of OpenAI. So let's get started!
1. Creating the App locally
mkdir telegrambot
cd telegrambot
touch main.py
# Open the file in the editor that u prefer
Now you can decide if you want to work in a virtual environment or just install all modules directly on your machine.
Just to show you how to do that, these are the commands that we can use to setup the virtual env:
python -m venv <name_of_virtual_environment>
# if you use bash
source <name_of_virtual_environment>
# if you use fish
source <name_of_virtual_environment>/bin/activate.fish
Let's now open the main.py file and write a basic flask app
main.py
from flask import Flask, request
app = Flask(__name__)
@app.route("/", methods=["GET"])
def index():
return "Hello World!"
if __name__ == "__main__":
port = 5000
app.run(debug=True, host='0.0.0.0', port=port)
And now we need to install flask trough pip: pip3 install flask.
After installing we can just run the app with: python3 main.py
We see that we successfully created the app! Nice job
2. Create the telegram bot
To connect our Flask app to Telegram Bot API, we first need to create a new bot on Telegram. We can easily do so by following these simple steps:
Open Telegram and search for "@botfather" in the search bar.
Start a conversation with BotFather by sending the command "/start".
To create a new bot, type the command "/newbot" and follow the instructions provided by BotFather.
Give your bot a name and a unique username that ends with "bot".
Once you've successfully created your bot, BotFather will give you an authorization token that you'll need to use to communicate with the Telegram Bot API.
By creating a new bot on Telegram, we can now link it to our Flask app and start building the chatbot's functionalities.
3. After creating the bot and getting a bot token, we can now link it with our Flask app.
In order to achieve that, we can use the requests module. We can easily install this package by running the command pip3 install requests.
Here is the updated code of main.py:
import requests
from flask import Flask, request
app = Flask(__name__)
BOT_TOKEN = <your_token_value>
def sendmessage(chatid, message):
url = f"https://api.telegram.org/bot{BOT_TOKEN}/sendMessage"
payload = {
"text": message,
"chat_id": chatid
}
print(requests.get(url, params=payload))
@app.route("/", methods=["POST", "GET"])
def index():
if (request.method == "POST"):
resp = request.get_json()
print(resp)
# We get the content of the message
msgtext = resp["message"]["text"]
# We get the id of the chat
chatid = resp["message"]["chat"]["id"]
# The message is sent back
sendmessage(chatid, msgtext)
return "Done"
else:
return "Hello World!"
if __name__ == "__main__":
port = 5000
app.run(debug=True, host='0.0.0.0', port=port)
Before trying out if the bot works we need to use ngrok to setup an http tunnel to localhost. We can do that using the command ngrok http 5000
We can see that the url works by using curl
# telegrambot> curl https://a577-131-114-63-2.eu.ngrok.io
Hello World!⏎
Now we can set the webHook of the telegram bot:
curl "https://api.telegram.org/bot<TOKEN>/setwebHook?url=https://a577-131-114-63-2.eu.ngrok.io"
{"ok":true,"result":true,"description":"Webhook was set"}⏎
And we can see that the bot now works perfectly
4. Writing a function that calls the openai API
In order to get an API KEY we need to create an account on OpenAI. After that we go to this url https://platform.openai.com/account/api-keys
to get our own key.
Now we can update the function sendmessage to be like this:
def sendmessage(chatid, message):
headers = {"Content-Type": "application/json",
"Authorization": f"Bearer {OPEN_AI}"}
info = {
# The model used (you can use gpt-4 if you have access to it)
"model": "gpt-3.5-turbo",
"messages": [{"role": "user", "content": message}],
# Randomness of the result
"temperature": 0.7,
"max_tokens": 1024
}
post_request = requests.post("https://api.openai.com/v1/chat/completions", headers=headers, json=info)
response_api = post_request.json()["choices"][0]["message"]["content"]
url = f"https://api.telegram.org/bot{BOT_TOKEN}/sendMessage"
payload = {
"text": response_api,
"chat_id": chatid
}
Deploy on deta.space
In order to deploy on deta.space we need to follow these steps that we can see on the documentation
- Create an account
- Download the Space Cli
curl -fsSL https://get.deta.dev/space-cli.sh | sh - Login on the cli with
space login - Create the project with
space new - Updated a few lines of the code to be like this
import os
OPEN_AI = os.getenv("OPEN_AI")
BOT_TOKEN = os.getenv("BOT_TOKEN")
- Creating a Spacefile how to run the app
# Spacefile Docs: https://go.deta.dev/docs/spacefile/v0
v: 0
app_name: telegrambot
micros:
- name: telegrambot
src: .
engine: python3.9
primary: true
public_routes:
- "/"
presets:
env:
- name: BOT_TOKEN
description: Telegram Bot Key
default: "<BOT_TOKEN>"
- name: OPEN_AI
description: Open AI Key
default: "<OPEN_AI_API>"
- Push the app to deta.space with
space push - Changing the telegram hook with the same commands but with the url of the deta.space app
Finished
Great job, guys! We've created a beautiful bot that uses artificial intelligence to assist you. If you have any questions or projects that you'd like to suggest, please leave a comment below to let me know. Bye for now! :D
Top comments (0)