Many files in modelmeta have incorrect metadata. It is extremely laborious to update them - usually index_metadata detects a "conflict" when a file has updated metadata, so you have to remove it from the database, update it, reindex it, and re-add it to ensembles. It is slow and fussy work.
It would be easy to automate this process with a wrapper script that calls the various other tools.This script should:
Accept a database, a a dataset file, and a .yaml.
Connect to the database and note which ensembles the file is part of
Delete the file from the database
run update_metadata on the dataset file using the *.yaml
index the file into the database using index_netcdf
associate the file to all previously-noted ensembles using associate_ensemble
Alternately, index_netcdf could be updated to handle "conflicts" in a more sophisticated fashion and allow updates to dataset metadata more easily.
Many files in modelmeta have incorrect metadata. It is extremely laborious to update them - usually
index_metadata
detects a "conflict" when a file has updated metadata, so you have to remove it from the database, update it, reindex it, and re-add it to ensembles. It is slow and fussy work.It would be easy to automate this process with a wrapper script that calls the various other tools.This script should:
update_metadata
on the dataset file using the *.yamlindex_netcdf
associate_ensemble
Alternately,
index_netcdf
could be updated to handle "conflicts" in a more sophisticated fashion and allow updates to dataset metadata more easily.