Open j08lue opened 3 weeks ago
@j08lue This change of collection id is better handled by creating a new collection because the id is an index and the database and the collection id maps 1:1 to the s3 prefix in the production data store. If we decide to do this, et's instead:
veda-data-store/OMI_trno2-COG
to veda-data-store/no2-yearly
no2-yearly.json
here in veda-data along with a (b) discovery items config file (it should be easy to produce following how the original collection and discovery were configured. (c) We will also need a transfer-config json that shows how the UAH covid-eo-dashboard/OMI_trno2-COG collection maps to no2-yearly in production (I believe this is an ongoing collection?)Great to see the process laid out this clearly, @anayeaye, and sounds like a very solid approach. Btw, we should copy this to operations docs somewhere - like "How to rename collections" and "How to deprecate collections"!
Not sure it is worth the effort for the OMI rename, though? What do others think - @anayeaye, @slesaad, @smohiudd?
Btw, we should copy this to operations docs somewhere - like "How to rename collections" and "How to deprecate collections"!
Yes! We do also have some work planned on covering how to remove VEDA data safely #116, #117 Maybe we could sneak in some differentiation on collection id vs name in that documentation, too? It is easy to change the collection name but that index is a more significant change. On the same note, maybe we should add waaaaay more emphasis on choosing a collection id in our contributing docs https://nasa-impact.github.io/veda-docs/contributing/dataset-ingestion/. I think it is not initially clear how valuable a human readable meaningful STAC Collection id is.
We currently have three NO2 products in our catalog:
no2
- monthly datano2-monthly-difference
- monthly anomaly (difference from some baseline)OMI_trno2-COG
- yearly dataThey present the same type of property (NO2 concentrations), just different derived products. Would it make sense to rename the
OMI_trno2-COG
collection tono2-yearly
or so, to harmonize the naming pattern between the three?We would need to update it in VEDA Earthdata Dashboard and the EO Dashboard, but I think that is easy enough to do.
Acceptance criteria
OMI_trno2-COG
tono2-yearly