Filter sentinel-5p netCDF files for testing

netcdf4testfiler.py

import netCDF4
import sys

USED_VARS = set([
    'carbonmonoxide_total_column',
    'longitude',
    'latitude',
    'qa_value',
    'delta_time'])

def main():
    if len(sys.argv) != 3:
        print(f'{sys.argv[0]} infile outfile')
        sys.exit()
    infile = sys.argv[1]
    outfile = sys.argv[2]
    with netCDF4.Dataset(infile, 'r') as read_f:
        p_grp = read_f.groups['PRODUCT']
        with netCDF4.Dataset(outfile, 'w') as f:
            f.time_reference = read_f.time_reference
            products = f.createGroup('/PRODUCT')
            for d in p_grp.dimensions.values():
                products.createDimension(d.name, d.size)
            for v in p_grp.variables.values():
                if v.name in USED_VARS:
                    print(f'importing {v.name}')
                else:
                    print(f'skipping {v.name}')
                var = products.createVariable(v.name, v.dtype, v.dimensions)
                for a in v.ncattrs():
                    if not a.startswith('_'):
                        var.setncattr(a, v.getncattr(a))
                var[:] = v[:]

if __name__ == '__main__':
    main()