Hello, Petter Reinholdtsen! I once reply to FTP Master to describe what is the importance of this package, sorry that I forget to reply it publicly.
The purporse of this package is that, if a user tries to import an out- dated version of gymnasium, the farama-notifications or gymnasium- notices will throw an error notification. It have a list of out-dated version, if it find the version is out-dated, it will notify the user. And, Farama Foundations may want to extend its functions over all of the packages they maintained, thus they make this package seperate. In other words, it will be needed for other packages by Farama Foundation. This package (python-farama-notifications) is a necessary dependency of upcoming package python-gymnasium, while python-gymnasium is the dependency of python-stable-baselines3. There is a change for gymnasium in to use farama-notifications instead of gymnasium-notices: https://github.com/Farama-Foundation/Gymnasium/pull/332/files The magic is in this file: https://github.com/Farama-Foundation/Gymnasium/blob/main/gymnasium/__init__.py If user tries to import an out-dated version of gymnasium, the farama- notifications or gymnasium-notices will throw an error notification. https://github.com/Farama-Foundation/gymnasium-notices/ Thus, farama-notifications is a successor of gymnasium-notices. It have a list of out-dated version, if it find the version is out-dated, it will notify the user. The old gymnasium-notices has detailed information for its function: https://pypi.org/project/gymnasium-notices/ I have something else about python-gymnasium and python-stable- baselines3 to say anyway. First of all, I am familiar with them and I use them daily. python-gymnasium and python-stable-baselines3 are important libraries for deep reinforcement learning. Deep reinforcement learning is an important tech poweres many important things like AlphaGo, AlphaFold, and ChatGPT. Gaming agents in video games like Dota 2 are usually powered by deep reinforcement too. ChatGPT is powered by an immportant technique called RLHF: Reinforcement Learning From Human Feedback. That is the reinforcement learning I mentioned before. As far as I know, libraries linked with machine learning is still rare in debian repo, thus I want to package them to debian.
