Asset key names recently changed for certain collections, and unfortunately are inconsistent across Items. For example:
GEDI
asset keys for main H5 data appear to be a combination of version and file path so are non-unique across items:
https://cmr.earthdata.nasa.gov/cloudstac/LPCLOUD/collections/GEDI02_A_002/items/GEDI02_A_2019108002012_O01959_01_T03909_02_003_01_V002
has asset key "002/GEDI02_A_2019108002012_O01959_01_T03909_02_003_01_V002/GEDI02_A_2019108002012_O01959_01_T03909_02_003_01_V002"
but https://cmr.earthdata.nasa.gov/cloudstac/LPCLOUD/collections/GEDI02_A_002/items/GEDI02_A_2019108063057_O01963_04_T01067_02_003_01_V002
has ""002/GEDI02_A_2019108063057_O01963_04_T01067_02_003_01_V002/GEDI02_A_2019108063057_O01963_04_T01067_02_003_01_V002"
ICESat-2
Similarly icesat-2 has asset keys appear to a different substring of the storage path:
https://cmr.earthdata.nasa.gov/cloudstac/NSIDC_CPRD/collections/ATL06_007/items/ATL06_20181014001049_02350102_007_01.h5
has:
"gov/nsidc-cumulus-prod-protected/ATLAS/ATL06/007/2018/10/14/ATL06_20181014001049_02350102_007_01"
Proposal
I was expecting all collections to have simple, standard asset keys (B01, B02, B03... etc) like HLS:
https://cmr.earthdata.nasa.gov/cloudstac/LPCLOUD/collections/HLSL30_2.0/items/HLS.L30.T59WPT.2013101T001445.v2.0
Even something generic like "Data" or "H5file" I think would be more useful compared to the current IS2 and GEDI schemes.
More discussion of consistent asset key naming within a collection as a STAC best practice here:
radiantearth/stac-best-practices#7
And I noticed this related issue here showing how this has been an annoyance in the past for other datasets. cc @chuckwondo
#326
Asset key names recently changed for certain collections, and unfortunately are inconsistent across Items. For example:
GEDI
asset keys for main H5 data appear to be a combination of version and file path so are non-unique across items:
https://cmr.earthdata.nasa.gov/cloudstac/LPCLOUD/collections/GEDI02_A_002/items/GEDI02_A_2019108002012_O01959_01_T03909_02_003_01_V002
has asset key "002/GEDI02_A_2019108002012_O01959_01_T03909_02_003_01_V002/GEDI02_A_2019108002012_O01959_01_T03909_02_003_01_V002"
but https://cmr.earthdata.nasa.gov/cloudstac/LPCLOUD/collections/GEDI02_A_002/items/GEDI02_A_2019108063057_O01963_04_T01067_02_003_01_V002
has ""002/GEDI02_A_2019108063057_O01963_04_T01067_02_003_01_V002/GEDI02_A_2019108063057_O01963_04_T01067_02_003_01_V002"
ICESat-2
Similarly icesat-2 has asset keys appear to a different substring of the storage path:
https://cmr.earthdata.nasa.gov/cloudstac/NSIDC_CPRD/collections/ATL06_007/items/ATL06_20181014001049_02350102_007_01.h5
has:
"gov/nsidc-cumulus-prod-protected/ATLAS/ATL06/007/2018/10/14/ATL06_20181014001049_02350102_007_01"
Proposal
I was expecting all collections to have simple, standard asset keys (B01, B02, B03... etc) like HLS:
https://cmr.earthdata.nasa.gov/cloudstac/LPCLOUD/collections/HLSL30_2.0/items/HLS.L30.T59WPT.2013101T001445.v2.0
Even something generic like "Data" or "H5file" I think would be more useful compared to the current IS2 and GEDI schemes.
More discussion of consistent asset key naming within a collection as a STAC best practice here:
radiantearth/stac-best-practices#7
And I noticed this related issue here showing how this has been an annoyance in the past for other datasets. cc @chuckwondo
#326