lukasmasuch/best-of-ml-python — reverse-engineered prompt

Build me a GitHub project that is a clean ranked directory of the best Python machine learning libraries. I want it to feel like an awesome list, but with scores and useful details so people can quickly compare tools.

The main page should group projects into categories like machine learning frameworks, data visualization, NLP, image data, AutoML, experiment tracking, deployment, interpretability, and PyTorch or TensorFlow utilities. Each project should show its name, link, short description, license, stars, activity status, contributor info, package links, install commands, and small symbols for things like trending, inactive, newly added, or framework related.

Keep the project data in a simple YAML file so people can add or update entries easily. Generate a polished README from that data, with a contents section, explanation of the icons, collapsible project details, contribution notes, latest changes, and history. Set it up so the rankings and metadata can be refreshed weekly from GitHub and package sources.