The purpose of this repository is to provide the necessary compute and storage to develop an intelligent agent that can be used to query a relational database using plain English to produce analytic reports.
The intent is twofold: allow agent developers to concentrate on developing the value add agent, rather than the compute infrastructure, and to understand the obstacles in creating a production quality agent.
This project is implemented using the infrastructure-as-aservice (IaaS) AWS Cloud Development Kit (CDK), implemented using the Python SDK.
It consists of two stacks, located in the ./iac
directory:
Data Stack - This consists of a single S3 bucket and loads a sample dataset consisting of a single file (./data/chinook.sql
). This is a simple relational data set made up of albums, songs, tracks, artists, etc.
Database Stack - This consists of an AWS Aurora RDS Postgres deployment, a single lambda to load the sample dataset, and all the permissioning needed to execute both in a private VPC.