Remove ase dependency and update ROY dataset

[GardevoirX]

This PR aims at removing the dependency on the ase library. The update of ase leads to issue #231. Since by now only the load of the ROY dataset depends on ase, getting rid of ase is easy and reasonable.

This PR deletes the xyz file and saves the data of the xyz file that is necessary for the example notebook in beran_roy_properties.npz, together with the features originally stored in beran_roy_features.npz. Because of the removal of the xyz file, the visualization of the structures in the example is no longer usable. Perhaps we can consider preserving the xyz file and letting users to open it with ase and then do the visualization.

Contributor (creator of PR) checklist

• Tests updated (for new features and bugfixes)?
• Documentation updated (for new features)?
• Issue referenced (for PRs that solve an issue)?

For Reviewer

• CHANGELOG updated if important change?

---

📚 Documentation preview 📚: https://scikit-matter--232.org.readthedocs.build/en/232/

[PicoCentauri]

Nice! For me this is perfect.

Can you now also remove ase from the dependencies in pyproject.toml

[PicoCentauri]

I also would like to ask other contributors what they think about removing the roy xyz file.

Especially because the .xyz files is used in for example in https://github.com/lab-cosmo/software-cookbook/blob/chemiscope-output/examples/roy-gch/roy-gch.py

[rosecers]

This in general looks very good. Future iterations can make use of recent chemiscope developments from @Luthaf and @ceriottm to use the json file type to load the structures, rather than rely on xyz. However, this will lead to a larger file type, so worth discussion in the future.