Fixed online regridding code (file_regrid.py) for restart-to-restart file regridding only

[yantosca]

This PR fixes several issues with GCPy's online regridding program using xESMF in GCPy module gcpy/file_regrid.py. Several utility functions have been added to the gcpy/grid.py, gcpy/regrid.py, and gcpy/cstools.py modules.

NOTE: At this point, regridding between restart files has been tested. There are some issues with diagnostic regridding (involving the ordering of vertical levels not being consistent with the lev:positive attribute, see: https://github.com/geoschem/GCHP/issues/21). We do not recommend regridding between GCHP diagnostic files.

Regridding 4x5 GCClassic to 2x2.5 GCClassic

NOTE: The grayscale came in when converting from PDF to PNG. Not sure why.

Regridding C24 checkpoint to 4x5 GCClassic

Regridding 4x5 GCClassic to C24 checkpoint

Regridding C48 checkpoint to C24 checkpoint

Regridding C48 checkpoint to C24 stretched grid

@lizziel: Not sure if the regridding is wrong or the plotting is wrong

[lizziel]

There is definitely something weird going on with CS comparison plots, probably related to this issue: https://github.com/geoschem/gcpy/issues/167. Regridding both C24 and C48 to lat-lon would probably be better for validation. Regardless, that stretched grid plot looks strange, but maybe a separate issue.

[yantosca]

Thanks @lizziel. I will address your comments. In the meantime I made the comparison plots of c48 -> 2x25 and c24-> 2x25. Regridding looks OK there. c24 -> 2x2.5

c48 -> 2x2.5 (note: Ref is mislabeled but the resolution is OK)