[source selector] group by source, rather than import name

[beets]

Import name is an implementation detail which we shouldn't expose to users

Can we downplay that detail, and instead group by the displayed source (we could still leave it under each option in gray).

So in the screenshot below: NASA_NEXGDDP_CMIP6_Subnational --> nccs.nasa.gov [NASA_NEXGDDP_CMIP6_Subnational]NASA_Mean_CMIP6_MPI-ESM1-2-HR, P1M, Celsius --> NASA_Mean_CMIP6_MPI-ESM1-2-HR, P1M, Celsius

[beets]

relatedly, what's the difference between these two import groups (cc @pradh)

[pradh]

These are different datasets. DCP30 (finer resolution US only) is from https://www.nccs.nasa.gov/services/data-collections/land-based-products/nex-dcp30, while GDDP (coarser res than DCP30, worldwide) is from https://www.nccs.nasa.gov/services/data-collections/land-based-products/nex-gddp.

The corresponding urls in Provenance node should have these e.g., https://datacommons.org/browser/dc/17mxl62, https://datacommons.org/browser/dc/zq39qv3

[beets]

thanks, that makes sense. perhaps we should put friendlier names now that we are using it more in the UI, and returning it in the API?

[pradh]

we should think about whether to change the import-name itself, or introduce a different display name for provenance that references the dataset (or introduce a node for Dataset referenced from Provenances?). Because an import doesn't 1:1 map to a dataset, can be arbitrary depending on aggregation type, and referenced in different places (like here).

[beets]

+1 for adding a linking Dataset with a nice display name

[premr]

Adriana, check this one out.