Closed TomAugspurger closed 2 years ago
Here's our tentative plan for the Planetary Computer:
(acquisition datetime, tile_id)
is more valuable. (acquisition date, tile_id)
(i.e. they have never been reprocessed or they are a reprocessed item). That way we don't every workflow linking to a particular item in the Planetary Computer.So we'll be able to say that
It's worth mentioning that we'd ideally use the version extension. When an asset is reprocessed by the upstream provider we would update the STAC item with the new assets and add a version link to the item linking to the old assets. But implementing that is relatively complicated so for now we will silently update the items when the assets are reprocessed.
Currently, the items generated by this package include a production date (https://lpdaac.usgs.gov/data/get-started-data/collection-overview/missions/modis-overview/#modis-naming-conventions). In the item ID
MCD15A2H.A2021265.h00v08.061.2021320165929
, the2021320165929
portion is the production date (in Julian form).Occasionally, the upstream data provider will reprocess assets. AFAICT, the original assets are deleted and replaced with the new assets. The new assets have the same actual
datetime
/ date range and the same tile IDs. But because of how the item IDs are derived, the item created for the new assets will have a different ID than the old ones.While it isn't 100% clear to me what the right thing to do is, I think that (by default) the item IDs shouldn't include the production date. In this case, the upstream provider is (I think) deleting the old assets and replacing them with the new ones, so presumably they think the new assets should replace the old ones. And so I think the new item should replace the old one too.
Here's a couple examples:
old item: https://planetarycomputer.microsoft.com/api/stac/v1/collections/modis-15A2H-061/items/MCD15A2H.A2021265.h00v08.061.2021279164319
new item: https://planetarycomputer.microsoft.com/api/stac/v1/collections/modis-15A2H-061/items/MCD15A2H.A2021265.h00v08.061.2021320165929
old metadata (gives a 404): https://e4ftl01.cr.usgs.gov/MOTA/MCD15A2H.006/2021.09.22/MCD15A2H.A2021265.h00v08.006.2021279164319.hdf.xml
new metadata: https://e4ftl01.cr.usgs.gov/MOTA/MCD15A2H.006/2021.09.22/MCD15A2H.A2021265.h00v08.006.2021320165842.hdf.xml