YAXArrays.jl
YAXArrays.jl copied to clipboard
JuliaDataCubes
Unstructured grids
AFAIU Python's uxarray is an extension of xarray for unstructured grids, and allows to do a bunch of stuff that is not easy with "plain" xarray.
Two features I could use from the top of my head are (i) plotting grid cells that cross the meridian, and (ii) taking zonal means or meridional slices correctly.
My understanding as an outsider is that there are no plans to deal with unstructured grids here, but I just wanted to point to uxarray's existence, in case there are some useful things to bring into YAXArrays.jl.
There has been some discussion about unstructured grids, but this would happen on the DimensionalData level, because this is where all the indexing happens. There is a recent issue in DimensionalData discussing this: https://github.com/rafaqz/DimensionalData.jl/issues/1105 Can you elaborate a bit on your second point?
If you mean taking a zonal statistic based on vector data that works with Rasters.zonal and that would give you a Vector datacube.
Ok great thanks!
About the 2nd point I think there might be some naming conventions clashing about what "zonal" means. For me, in climate and oceanography, "zonal" and "meridional" mean along longitudes and along latitudes, respectively. The issue with unstructured grids and curvilinear grids is that taking a zonal/meridional average/slice requires some interpolation as the grid is not aligned with longitudes and latitudes. It seems to me that uxarray has that functionality.