Finalize evaluation module

[remrama]

A first draft has been merged, but some final upkeep needs to happen before yasa.evaluation is an official part of YASA.

• Properly handle sleep statistics that are identical between scorers in SleepStatsAgreement
• Add SleepStatsAgreement.plot_blandaltman
• Add pytests
• Update changelog
• Review docstrings (especially the examples to make sure they did not fall behind code updates)
• Add tutorial with sample data

I think that all those tasks are required except the Bland-Altman plots, which might be technically optional though strongly preferred. Also, they should be implemented without Pinguoin, as we've managed to avoid that dependency everywhere else, plus the Bland-Altman plots here will need to have additional flexibility anyways (e.g., to accommodate the regression-based measures).

Sidenote: I'm still not sure what the best way to validate this pipeline is, as I'm not aware of an existing (public) dataset that has ground-truth values.