Feature/new datasets

[mariahpope]

These are some helpful changes and/or additions that were necessary as I tested out the anemoi-datasets module on a few different datasets. Please let me know if you have any questions! Thanks!

Bug Fix

1. src/anemoi/datasets/create/functions/sources/xarray/variable.py
   Bug fix - the arg in the class Variable is "variable" not "var". This caused a failure for me when working with levels until fixed.

Additions I found helpful

2. src/anemoi/datasets/create/init.py
   Add the ability to work with cftime when needed.

3. src/anemoi/datasets/create/functions/sources/xarray/fieldlist.py
   This is an update that was helpful when working with a curvilinear grid (but will not always be necessary - it depends on the dataset). It came after I noticed that some dimensions are not always set as coordinates/variables in the dataset. For example. a dataset with a curvilinear grid could have lat/lon as coordinates, but the relevant x and y index dimensions are not set as coordinates or variables. Therefore, this is a check to make sure that if dimension in a dataset was identified in a "flavour" that it is also set as a coordinate within the dataset.

Ability to read from Microsoft's Planetary Computer

There are a number of super useful datasets on planetary computer, and these updates make it possible to read most zarr stores from there (e.g. CONUS404 example: https://planetarycomputer.microsoft.com/dataset/conus404#Example-Notebook). The URL that a user of anemoi-datasets would use in a yaml file is the STAC collection URL associated with the dataset. I chose to set it up this way because this url appeared to be the most consistent way to reference zarr datasets within planetary computer.

4. src/anemoi/datasets/create/functions/sources/xarray/init.py
   Elif statement specifically for planetary computer. There are a few extra hoops to jump through for open_zarr to work.

5. src/anemoi/datasets/data/stores.py
   Add class PlanetaryComputerStore() that mimics the already existing S3 store and HTTPS store.

Unit test addition

6. tests/xarray/test_zarr.py
   I added two unit tests for additional datasets. The updates within this PR were able to make anemoi-datasets work with both of these datasets.

[FussyDuck]

All committers have signed the CLA.

[b8raoult]

Can you add the new dependencies in the pyproject.toml file, as optional dependecies? Namely pandas, pystac_client and planetary_computer.

[mchantry]

Thanks for the great contribution!