Adds xarray backend to support loading unstructured grids

[lukecampbell]

This commit adds a new xarray backend as a plugin to support loading FVCOM and other unstructured gridded data from DAP and netCDF sources. The backend enables clients to pass a mapping of variables to rename, the variables are renamed after the xarray objects are loaded from the store but before assembled into a dataset which aleviates issues arising from unstructured grids that have coordinate variables with multiple dimensions.

Pull Request Reminders

• [ ] Make sure the docs notebooks models.ipynb and ts_work.ipynb still run.
• [ ] Add tests for the new functionality.
• [ ] Add a bullet to docs/whats_new.rst describing your new work.

[codecov[bot]]

Codecov Report

Merging #62 (68e56d6) into main (d269b7f) will decrease coverage by 1.78%. The diff coverage is 40.62%.

@@            Coverage Diff             @@
##             main      #62      +/-   ##
==========================================
- Coverage   89.92%   88.14%   -1.79%     
==========================================
  Files          13       15       +2     
  Lines         854      886      +32     
==========================================
+ Hits          768      781      +13     
- Misses         86      105      +19     

Flag	Coverage Δ
unittests	88.14% <40.62%> (-1.79%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
extract_model/xr/triangular_mesh_netcdf.py	38.70% <38.70%> (ø)
extract_model/xr/__init__.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update d269b7f...68e56d6. Read the comment docs.

[lukecampbell]

Superseded by https://github.com/axiom-data-science/extract_model/pull/63