This is a simple implementation of a chat interface using Azure OpenAI APIs and Python (Flask). The aim of the implementation is to keep it as simple and self-contained as possible to allow easy deployment of a private GPT chatbot.
Websockets are used to allow for streaming of response for a more "chat-like" experience. The base implementation for the sockets and chat UI is based on https://github.com/miguelgrinberg/Flask-SocketIO/blob/main/example/app.py
Ensure the following three environment variables are set:
OPENAI_API_BASE: The base of your Azure OpenAI APIOPENAI_API_KEY: Your Azure OpenAI API keyOPENAI_DEPLOYMENT_NAME: The deployment name of your Azure OpenAI instance
All three of the above are available in the Azure Portal.