Azure GPT-4 Functions and Assistants

Overview

This project implements an AI-driven assistant powered by Azure OpenAI’s GPT-4 and deployable via Azure Functions . It uses Azure serverless architecture to invoke GPT-4 for tasks like answering queries, executing specific functions, and providing intelligent assistance through REST APIs.

Features

• GPT-4 Integration : Leverage GPT-4 to handle a variety of AI-powered tasks.
• Azure Functions : Serverless function calls for on-demand execution of AI logic.
• Function Calling : Integrate function calling within the assistant to handle specific tasks based on user input.
• Configuration Management : Store API keys and other sensitive configurations securely.
• Interactive Notebooks : Use Jupyter notebooks for prototyping and experimentation.

Project Structure

Here’s a breakdown of the key files in the project: ├── .gitignore
├── assistant_with_function_calling.ipynb
├── assistants.py
├── config.json
├── config.py
├── core.py
├── dark_sine.png
├── main.py
├── README.md
├── scratchpad

Git ignore file to exclude unnecessary files from version control

Jupyter notebook for developing and testing assistants with function calling

Python script defining assistant behaviors

Configuration file containing API keys and settings

Python script for loading and managing configurations

Core logic and functionality of the GPT-4 assistant

A PNG image (possibly for documentation or UI purposes)

Main script to run and deploy the project

Project documentation (this file)

A file or directory for miscellaneous notes or temporary work

File Details

• assistant_with_function_calling.ipynb : This Jupyter Notebook allows you to prototype and experiment with your assistant and test function-calling capabilities.
• assistants.py : This file defines the various assistants you are building using GPT-4. It may include logic for text generation, conversation flows, or specific assistant behaviors.
• config.json : The config.json holds sensitive data like API keys, endpoint URLs, and other configurations required to connect to Azure OpenAI and other Azure services.
• config.py : Python file responsible for loading and handling configurations from config.json, ensuring secure and flexible configuration management.
• core.py : Core functionality of the assistant and business logic resides in this file, possibly including the handling of requests and calling the GPT-4 API.
• main.py : The entry point of the project. This script could be used to initialize the Azure Functions or run the assistant service locally for testing.
• dark_sine.png : A supporting image, possibly used for the project’s documentation, branding, or UI.
• scratchpad : A file or directory used for temporary notes, code snippets, or experimental ideas during development.

Installation and Setup

Prerequisites

• Azure Subscription : Ensure you have access to Azure OpenAI and Azure Functions.
• Python 3.x : Install Python and necessary packages.
• Jupyter Notebook : Required for running the notebook files.

Setup Steps

1. Clone the Repository :

git clone (https://github.com/mphomathabathe/AGENTS)
cd azure-gpt4-assistants 

2. Install Dependencies: Install the necessary Python packages using pip :

pip install -r requirements.txt

3. Configure API Keys: Update the config.json file with your Azure OpenAI API key and any other necessary configurations.

pip install -r requirements.txt

4. Run Locally: Run main.py to test the assistant locally before deploying it :

python main.py

Jupyter Notebook Usage

• Open the assistant_with_function_calling.ipynb file in Jupyter Notebook to interactively test and develop assistant functionality. This is useful for debugging and prototyping.

Deployment

To deploy your assistants on Azure Functions :

1. Deploy to Azure : Use the Azure CLI or Visual Studio Code to deploy the functions:

func azure functionapp publish <APP_NAME>

2. Monitor Performance : Use Azure Monitor to track usage, errors, and performance for your deployed functions.

Usage

Once deployed, you can access your assistant via an HTTP endpoint. For example, using curl:

curl -X POST https://`<your-function-url>`
-H "Content-Type: application/json"
-d '{"prompt": "Tell me about Azure GPT-4"}'

License

This project is licensed under the MIT License.

---