[Task]: Robust testing of temporal averaging APIs (including missing data) with real world datasets

[tomvothecoder]

This post keeps a dedicated track record for testing temporal averaging APIs with real world datasets.
The discussion stems from this comment in #275, which includes some temporal averaging testing by @lee1043.

We have some notebooks here for earlier versions of xcdat that test using a few datasets from PMP demo data:

• https://github.com/xCDAT/xcdat_test/tree/main/validation/v0.2.0/temporal_average
• https://github.com/xCDAT/xcdat_test/tree/main/validation/v1.0.0/temporal_average -- note, not actually v1.0.

However, more robust testing is needed by looping over models and ensuring edge cases are covered (e.g., missing data).

FYI @lee1043, @msahn, @pochedls.

[tomvothecoder]

0.3.3 was just released which includes #350, drop leap days for specific CF calendar types for daily climatology and departures.

@lee1043, @msahn, @gleckler1, you can pull the latest version and test xcdat's temporal averaging with PMP now! Thanks!

Changelog: #359

[lee1043]

@tomvothecoder Excellent, thank you for the release!