Ayush74510/Referral-Fraud-Detection-Pipeline ? reverse-engineered prompt

Build me a Python data pipeline that takes 7 raw CSV files from a data folder, cleans them up, joins them into one referral level table, and flags which referral rewards look valid or suspicious based on a set of business rules. I want it to produce a final CSV report called referral_fraud_report.csv with the important columns, plus clear audit columns that show why each row passed or failed. Please also add a simple entry script so I can run the whole thing easily, and make it work both locally and in Docker. If useful, include a small profiling script that can scan the input tables and save basic quality summaries, since this project seems to rely on checking duplicates, missing links, mixed timestamps, and odd reward values. Keep the code organized and readable, and if you need to check any current library details, look up the latest docs online.