ScottfreeLLC/alphapy-pro ? reverse-engineered prompt

Build me a Python machine learning pipeline tool that feels simple to use but is powerful under the hood.

I want it to run from a command line command and use YAML config files so each project can define its own data, target, model settings, and feature steps without writing much code. It should support loading CSV and Parquet data, work with pandas or polars, and handle common ML workflows like encoding, scaling, feature selection, model training, cross validation, and hyperparameter tuning. Please include support for common sklearn models plus XGBoost, LightGBM, CatBoost, and Optuna.

I also want a few example projects included, like a Titanic style classification project, a ranking project, and a simple time series forecasting project, with outputs saved into timestamped run folders. Make sure it’s easy to extend, well tested, and has basic docs or a README that explains how to install it and run an example. If you need to check current library docs online, go ahead.