eurec4a / eurec4a-intake

Intake catalogue for EUREC4A field campaign datasets
17 stars 19 forks source link

Non-descriptive grid names in intake catalog (simulations) #123

Closed suessspeise closed 2 years ago

suessspeise commented 2 years ago

This concerns the intake catalog of the grids for the EUREC4A LES simulations.

import eurec4a
grids = eurec4a.get_intake_catalog().simulations.grids
list(grids)

fetches this rather cryptic list:

['6b59890b-99f3-939b-e76a-0a3ad2e43140',
 '3c7523ba-f78c-c23e-6654-012805ba3300',
 'ecf22d17-dcee-1510-a807-11ae4a612be0']

Better desriptions are available:

for grid in list(grids):
    print(grids[grid].description)
ICON LES horizontal grid of DOM01 (EUREC4A_PR1250m_DOM01)
ICON LES horizontal grid of DOM02 (EUREC4A_PR1250m_DOM02)
ICON LES vertical grid of DOM01 (EUREC4A_PR1250m_DOM01)

But the identifiers could also be more informative.

d70-t commented 2 years ago

The catalog intentionally contains the UUIDs of the grids. They are unique identifiers and are not meant to be interpreted in any way. Instead they should be used like:

data = get_some_dataset()
hgrid = grids[data.uuidOfHGrid].to_dask()
vgrid = grids[data.uuidOfVGrid].to_dask()

By using non-descriptive opaque identifiers, it's easier not to accidentally have two different grids which end up having the same name.

suessspeise commented 2 years ago

fair enough 😢