Project Description : PowerfulBI: ask question about data in database

The project implements OpenAI and langchain framwork to query against your database. OpenAI has been given read only permission to access the data, so there is no risk of DML(insertion,deletion etc.). Furthermore, AWS cognito has been used to authenticate the users. The idea is whenever the authenticated user ask the question reagarding data, the app responses with the facts and analytics from database while unauthenticated user's question is answered from llm model with generic answers.

Use this repo

• create the Aws MySQL database and use world.sql to create schema and populate the data(refer populate_aws_rds.py for implementation).
• For database new user credentials for read access refer read_permission_LLM.py. use .env file from step 5 to use the python code mentioned in step 1 and step 2.
• create Aws Cognito pool and users for the Flask application.
• clone this repo.
• create .env file in root path and provide the environment variables: OPENAI_API_KEY,DB_HOST,DB_PORT,DB_USER,DB_PASSWORD,DB_NAME,SQL_FILE,READ_USER,READ_USER_PASSWORD,COGNITO_POOL_ID,COGNITO_CLIENT_ID,COGNITO_REGION,FLASK_SECRET_KEY,ACCESS_KEY_ID,SECRET_ACCESS_KEY,ENVIRONMENT. Here ACCESS_KEY_ID,SECRET_ACCESS_KEY have been used which is not ideal. Use Service principal or iam role instead. If you are embedding the ACCESS_KEY_ID,SECRET_ACCESS_KEY directly, don't share to anywhere public.
• Finally, run docker compose up
• Here are two images of powerfulBI chatapp

  Unauthenticated user interaction

  Authenticated user interaction