Read UGRID files with transposed connectivity (i.e. fesom.mesh.diag)

[philipc2]

Closes #282 #1026

Overview

• Updates UGRID reader to correctly parse grid files that contain transposed connectivity information
  • The fesom.mesh.diag.nc grid file stores connectivity information in a transpose order ([3, n_face] instead of [n_face, 3]
• Fixes a bug where the start index was not correctly adjusted due to a negative fill value.

Plot of fesom.mesh.diag.nc

[rajeeja]

Good fix, FESOM stores elements each node is connected to aka adjacency and happy to see that we had the diag file in main for over a year. It wasn't tested.
test/grid_geoflow.exo is removed ? It shouldn't have been there in the first place. Anyways. I'll approve as this will enable newer changes.

[philipc2]

Good fix, FESOM stores elements each node is connected to aka adjacency and happy to see that we had the diag file in main for over a year. It wasn't tested. test/grid_geoflow.exo is removed ? It shouldn't have been there in the first place. Anyways. I'll approve as this will enable newer changes.

Yeah, one of our testing scrips produces the .exo file and it must of been included in a previous commit.

[rajeeja]

Good fix, FESOM stores elements each node is connected to aka adjacency and happy to see that we had the diag file in main for over a year. It wasn't tested. test/grid_geoflow.exo is removed ? It shouldn't have been there in the first place. Anyways. I'll approve as this will enable newer changes.

Yeah, one of our testing scrips produces the .exo file and it must of been included in a previous commit.

One minor thing that could make it good would be to have something about transpose in the README

[philipc2]

Good fix, FESOM stores elements each node is connected to aka adjacency and happy to see that we had the diag file in main for over a year. It wasn't tested. test/grid_geoflow.exo is removed ? It shouldn't have been there in the first place. Anyways. I'll approve as this will enable newer changes.

Yeah, one of our testing scrips produces the .exo file and it must of been included in a previous commit.

One minor thing that could make it good would be to have something about transpose in the README

Good idea. I'll add a note to it.

[philipc2]

@fluidnumerics-joe

I'm not sure if you've seen this PR, but we've already added and released support for reading the UGRID-compliant FESOM grid files.

I haven't followed up with #1013 since merging this. However, I'll work on getting it in soon.

[fluidnumerics-joe]

Sounds good. I haven't been able to get uxarray working for fesom2 netcdf grids from the latest pypi release, but I can from your philipc2/fesom branch. Curious to know the timeline to have these features in the pypi index .

[philipc2]

Sounds good. I haven't been able to get uxarray working for fesom2 netcdf grids from the latest pypi release, but I can from your philipc2/fesom branch. Curious to know the timeline to have these features in the pypi index .

I'm going to wrap up #1013 this week and aim to release it next week.