Metadata

The templates have no any metadata. It is created in the pipeline process so no limits in terms of metadata addition is set.

PRESENT:

Currently for GRIB files DE pipeline structure works fine due to file format. Also grib files have a very structured metadata which in this case facilitates the works.

However, NC files cannot be treated in the same way due to Earthkit data fields. Also metadata provided by NC is not as structured as GRIB but more extense. Currently NC metadata is retrieved by original metadata and translating some keys to grib metadata.

TO DISCUSS:

1. Should we create a standard metadata which can be filled directly from files information? Should we create a standard metadata translating the metadata info to a standard naming and structure
2. Should we prepare the pipeline avoiding earthkit? It could be easier to work just in xr datasets