Parsing/Filling CF metadata as part of preprocessing

[jbusecke]

This is a follow up on #52. Over there I made the decision to rename variables manually within this package, but I would still like to use cf conventions where possible. I am thinking of checking and maybe adding cf compliant attributes to the datasets as part of combined_preprocessing.

We have basically two options to add missing cf metadata:

1. Manually parsing them as part of combined_preprocessing (This we can do right now, and I would suggest to do this as an intermediate step)
2. Ultimately I would hope that this step is actually implemented further upstream in e.g. https://github.com/roocs/daops. This would work as function that queries a database of 'fixes' for the given dataset, applies those metadata fixes. But this also requires entering a LOT of data up front, so I think this would be the desired long term solution (where 1 is phased out at some point).

[dcherian]

.cf.guess_coord_axis and .cf.add_canonical_attributes may be useful.

https://cf-xarray.readthedocs.io/en/latest/generated/xarray.DataArray.cf.guess_coord_axis.html#xarray.DataArray.cf.guess_coord_axis

https://cf-xarray.readthedocs.io/en/latest/generated/xarray.DataArray.cf.add_canonical_attributes.html#xarray.DataArray.cf.add_canonical_attributes

[jbusecke]

Ah very nice!