Open tomkralidis opened 1 year ago
I've been thinking on this.
We can reportedly create the new BUFR by creating and saving a template and then by specifying this template in the call to codes_bufr_new_from_samples
. This will not work with the workflow for a generic BUFR pipeline as it requires prior knowledge of the BUFR to be processed and the creation of a template before we start processing the data. One option around this may to be the creation of a template the first time a sequence is encountered but we then need to decide how to manage these templates. I think this may then be defeating the purpose of the generic BUFR pipeline.
The following may be helpful but I haven't checked how to use these in the python library yet.
@efucile to discuss with ECMWF to add additional templates to ecCodes. @david-i-berry to provide suggestions, based on:
Needs:
Agree to lower priority.
For efficiency, in
wis2box/data/bufr4.py
, do not create new BUFR from samples (i.e. not fromcodes_bufr_new_from_samples("BUFR4")
).