chat-gpt-celery-flask

A sample implementation of a chatbot using Flask, Celery, and GPT-3.

Overview

This repository demonstrates how to create a simple chatbot by combining the power of Flask, Celery, and GPT-3. Flask is used as the web framework to handle HTTP requests and responses, Celery is used as the task queue to run GPT-3 requests asynchronously, and GPT-3 is used as the language generation model to generate text based on the prompts provided by the user.

Requirements

• Python 3.7 or later
• Flask
• Celery
• Redis (or any other message broker supported by Celery)
• OpenAI API Key (to access GPT-3)

Usage

1. Clone the repository:

   git clone https://github.com/pobruno/chat-gpt-celery-flask.git
   

2. Install the dependencies:

   pip install -r requirements.txt
   

3. Replace example.env to .env and change YOUR_API_KEY in app.py with your OpenAI API key.
4. Start the Redis server:

   redis-server
   

5. Start the Celery worker:

   celery -A app.celery worker --loglevel=info
   

6. Start the Flask app:

   flask run
   

7. Send a POST request to http://localhost:5000/chat with a JSON body containing the prompt, for example:

   { "prompt": "What is the capital of France?" }
   

8. You will receive a JSON response containing the task id, for example:

   { "task_id": "8d5f0b5d-a5e5-4b8f-b5f5-5a5a5a5a5a5a" }
   

9. Use the task id to check the status of the task and retrieve the generated text by sending a GET request to http://localhost:5000/result/<task_id>, for example:

   http://localhost:5000/result/8d5f0b5d-a5e5-4b8f-b5f5-5a5a5a5a5a5a
   

10. To genarate image same process as 7 to 9 just POST request URL is http://localhost:5000/image_chat and GET URL is http://localhost:5000/image/<task_id>

Docker

1. Copy example.env to .env and give required variables

   cp example.env .env
   

2. Build the Docker image by running the following command in terminal:

   docker build -t gpt-chat-flask-celery .
   

3. Once the image is built, you can run the container using the following command:

   docker run -p 5000:5000 --name gpt-chat-flask-celery gpt-chat-flask-celery
   

4. You can access the application on http://localhost:5000

Limitations

This sample chatbot has some limitations and is just for demonstration purposes.

• It uses GPT-3 with a single endpoint and does not handle pagination.
• It does not handle any form of natural language processing, sentiment analysis, or other advanced features.
• It does not have any form of UI and just accepts and returns json data.

Contributions

If you find any bugs or would like to add new features, please feel free to open a pull request or an issue.

License

This repository is released under the MIT license. See LICENSE for more details.

Test Client

to test this can use this simple client

React-Client