Mobile-env

- Bugfix in multi-agent handler to check if the simulation is finished (38 )
- Consistent use of `gymnasium` rather than `gym` everywhere

* Breaking change: Switch from `gym` to `gymnasium`, which changes the API of `step()` and `reset()`: https://gymnasium.farama.org/content/migration-guide/
* Updated notebooks
* Code style black and pre-commits + added tests

- updated demo notebook
- fixes an issue where passed configs were not merged correctly

Adds support for custom metrics to `mobile-env`. Core functionality itself unchanged.

Fixes in the CI and doc build system. `mobile-env` functionality itself unchanged.

First major release of `mobile-env`:

* Simple, lightweight simulation environment for reinforcement learning in wireless mobile networks
* Easy installation and configuration
* Many options for extension, e.g., user movement, quality of experience models, channel models, etc.
* [New interactive Google CoLab demonstrating some of these features](https://colab.research.google.com/github/stefanbschneider/mobile-env/blob/master/examples/demo.ipynb)