Rule learning: RIPPER, decision lists, and models that output human-readable rules

I’m validating the current package versions and executing the demo end to end so the code and printed rules match the APIs described.