Add more metadata in Satpy reader YAML-files for categorical datasets

[sjoro]

Feature Request

When reading a categorical dataset, e.g. SEVIRI Cloud Mask product, the user typically ends up with a data-array with some integers in it. In the example case here, seviri_l2_grib does not provide any information on what these cloud mask values (between 0 and 3) mean and the user needs to find this information online in a user manual.

Describe the solution you'd like
Satpy-reader should provide the user the needed information for the categorical datasets mapping the different categories to the corresponding values in the dataset. CF convention suggests using flag_values and flag_meanings (and flag_masks in case of bit field notating a unique bit set for s specific meaning), which could be added to the readers where necessary.
https://cfconventions.org/cf-conventions/cf-conventions.html#flags

Describe any changes to existing user workflow
None. The information could be added in the reader YAML-files for affected datasets.

Additional context
Just opening up the conversation on this here in case there's a better solution for the problem.

[djhoese]

Your suggestion is exactly what I was going to suggest. In most readers that are NetCDF-based this is already done since most NetCDF creators try to follow (or almost follow) CF conventions. Parts of Satpy and Pyresample even treat datasets differently if they have flag_meanings defined. I think if the file doesn't provide that information then putting it in the YAML makes sense.

[strandgren]

Just for the record this functionality is briefly mentioned in issue #1921 as well.