Add sort_mesh call to build_mali_mesh framework function.

Sorting the global indicies will improve memory performance of
`cellsOnCell` and such. Additionally, sorting will help with
PIO decompostion and ice-sheet-coupler communication.

compass/landice/mesh.py

@@ -13,6 +13,7 @@
 from mpas_tools.logging import check_call
 from mpas_tools.mesh.conversion import convert, cull
 from mpas_tools.mesh.creation import build_planar_mesh
+from mpas_tools.mesh.creation.sort_mesh import sort_mesh
 from netCDF4 import Dataset
 from scipy.interpolate import NearestNDInterpolator, interpn
 
@@ -682,6 +683,8 @@ def build_mali_mesh(self, cell_width, x1, y1, geom_points,
     dsMesh = xarray.open_dataset('base_mesh.nc')
     logger.info('culling mesh')
     dsMesh = cull(dsMesh, logger=logger)
+    logger.info('soring mesh')
+    dsMesh = sort_mesh(dsMesh)
     logger.info('converting to MPAS mesh')
     dsMesh = convert(dsMesh, logger=logger)
     logger.info('writing grid_converted.nc')