Investigate UGRID conventions for tri meshes

[ianthomas23]

There are standards for defining unstructured grids (UGRIDs), see documentation at https://ugrid-conventions.github.io/ugrid-conventions/ and corresponding github repo at https://github.com/ugrid-conventions/ugrid-conventions/#contributing-to-ugrid.

It would be useful if holoviews (and other holoviz projects) could provide easier use of grids that follow those standards. Initially some investigation will be needed to understand what overlap there is between those conventions and the existing holoviews functionality.

[philippjfr]

See related issues:

#2853
#3258
#3812

[rsignell-usgs]

There are several python packages that attempt to utilize UGRID conventions, but probably the one to look at first would be uxarray, as it is the latest (and they knew about the others). Do you agree @erogluorhan?

Blog post about uxarray.

[erogluorhan]

Hey @rsignell-usgs , yes that's correct! In addition to the resources you provided, here are some other updates and information:

We are exploring the Datashader stack for big data visualization, and we already had some good outcomes so far. I'd recommend to visit our [AGU poster]
https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1182328 if you are planning to be there.

More background info
UXarray supports not only UGRID but also MPAS (on the way), SCRIP, and Exodus. UXarray unifies all of these conventions at the data loading step by representing grids in the UGRID convention regardless of the original grid type that is read in from the file.

UXarray is still in the early development stages though, and we are seeking options about how to move forward to best comply with Xarray expecially.