kedro-org/kedro ? reverse-engineered prompt

Build me a Python toolbox for production ready data science and machine learning pipelines.

I want it to help people structure projects properly, with a standard project template, a way to register and load data from different places like local files, cloud storage, and other file systems, and a clean pipeline system where plain Python functions connect through their inputs and outputs automatically.

The app should make pipelines easy to reuse, test, and understand, and it should support versioning for files and models when that makes sense. It would be great if there’s a simple way to visualise pipelines too, so people can see how the pieces fit together.

Please make it feel like a real open source library, with good docs, solid tests, clear logging, and sensible defaults. If you need current reference docs for the best implementation details, look them up online first.