RP dev

[pernak18]

quick commit to address #103

[RobertPincus]

@pernak18 Can you please change the target of this PR from master to develop?

[pernak18]

@RobertPincus done. sorry -- i misunderstood the original directions.

[RobertPincus]

Hi @pernak18 So the CI is failing because the encoding of the variables differs between LW and SW. In the LW they are char but in the SW they are strings which I guess is how xarray writes them.

[Mini-Me-2:~/Codes/rte-rrtmgp] robert% ncdump -h rrtmgp/data/rrtmgp-data-lw-g256-2018-12-04.nc | grep string_len
	string_len = 32 ;
	char gas_minor(minor_absorber, string_len) ;
	char gas_names(absorber, string_len) ;
	char identifier_minor(minor_absorber, string_len) ;
	char minor_gases_lower(minor_absorber_intervals_lower, string_len) ;
	char minor_gases_upper(minor_absorber_intervals_upper, string_len) ;
	char scaling_gas_lower(minor_absorber_intervals_lower, string_len) ;
	char scaling_gas_upper(minor_absorber_intervals_upper, string_len) ;
[Mini-Me-2:~/Codes/rte-rrtmgp] robert% ncdump -h rrtmgp/data/rrtmgp-data-sw-g224-2018-12-04.nc | grep string_len
	string_len = 32 ;
	string gas_minor(minor_absorber, string_len) ;
	string gas_names(absorber, string_len) ;
	string identifier_minor(minor_absorber, string_len) ;
	string minor_gases_lower(minor_absorber_intervals_lower, string_len) ;
	string minor_gases_upper(minor_absorber_intervals_upper, string_len) ;
	string scaling_gas_lower(minor_absorber_intervals_lower, string_len) ;
	string scaling_gas_upper(minor_absorber_intervals_upper, string_len) ;

[RobertPincus]

I think we should provide files that are in every way except data identical in the LW and the SW

[pernak18]

damnit. yes, my intent was identical LW and SW. the link i provided in the issue does warn against this behavior, and i didn't double check the array type. i was so focused on string_len and this phantom string1 that xarray.Dataset.to_netcdf kept putting in for character arrays. i'll look into it more when i have some time

[RobertPincus]

@pernak18 Thanks for making any progress at all, I'm grateful. I used a lower level (netCDF4) interface some years ago - maybe there's something helpful here? https://github.com/RobertPincus/RFMIP-IRF-Scripts/blob/master/generate-output-file-templates.py

[pernak18]

@RobertPincus how about now?

[RobertPincus]

Thanks kindly, @pernak18. This looks great. Can you address the file format issue in #103 in this same PR?

[pernak18]

@RobertPincus how about now