Rajiv-Patil/Berka-OLAP — reverse-engineered prompt

Build me a local Berka financial analytics database that I can start with Docker and use right away for reporting and simple AI style querying.

I want it to automatically load the Berka CSV data, keep the raw import in a staging area, clean and transform it, then create a star schema for OLAP with fact tables for things like transactions and orders plus useful dimensions like dates, locations, and client details. Please also include semantic summary text columns on the dimension tables so the data is easier for an LLM or RAG workflow to understand.

Make it so I can run one command to bring everything up, wait a few seconds, and then connect on localhost port 5433 to a database called berka with the default postgres user and password. If the dataset folder is missing, grab or guide the needed CSVs into the right place. I also want the setup to be easy to reset and rebuild from scratch if the SQL changes. Look up current docs online if you need to.