Page 1 of 1

LPCLOUD STAC search fails

Posted: Wed Feb 22, 2023 3:45 pm America/New_York
by waltersdan
Hello,
I'm trying to query the HLSL30 and HLSS30 STACs from the LPCLOUD endpoint using pystac_client. For some reason all of my searches have started returning an API error:

APIError: {"errors":["Invalid intersects parameter detected. Please verify it is a valid GeoJSON geometry."]}

This includes when I submit geometries that worked earlier today. Have there been any updates to STAC protocol or changes to the status of the LPCLOUD STACs? Is there a good way to check STAC status?

Thanks!

Re: LPCLOUD STAC search fails

Posted: Thu Feb 23, 2023 7:04 am America/New_York
by hrodman1
I don't have a solution to waltersdan's problem but I have done some digging and at least found the new collection IDs

I noticed a big change yesterday, too! Now the collection IDs listed in the LPCLOUD catalog look like this: "C2076115306-LPCLOUD". I crawled through all of the collection descriptions and found the new collection IDs that seem to correspond to the HLS collections. HLSL30 is "C2021957657-LPCLOUD" and HLSS30 is "C2021957295-LPCLOUD". If you navigate to the items level of the STAC endpoint for the HLSL30 collection in your browser (https://cmr.earthdata.nasa.gov/stac/LPCLOUD/collections/C2021957657-LPCLOUD/items), the items appear to be present but if you go to one of the items the record appears to be incomplete (https://cmr.earthdata.nasa.gov/stac/LPCLOUD/collections/C2021957657-LPCLOUD/items/G2247061770-LPCLOUD). That one only has B09 listed in the assets dictionary.

When I create any search via pystac_client, it fails it fails to even return any items from one of the catalogs that I specified!

For example, this query returns what appear to be a bunch of MODIS items:
https://cmr.earthdata.nasa.gov/stac/LPCLOUD/search?limit=250&bbox=-91.67481115470213,47.86318699029263,-91.52471085746416,47.93118792159696&datetime=2022-01-01T00:00:00Z/2023-01-01T23:59:59Z&collections=C2021957657-LPCLOUD,C2021957295-LPCLOUD

Something seems to be very wrong with the LPCLOUD STAC now, and as the original poster mentioned all of this was working on the morning of 2/22/2023!

# Here is the Python code I ran to find the new collection IDs (apologies for the poor formatting, code blocks don't seem to work very well in this forum):
import pystac_client

CMR_STAC_URL = "https://cmr.earthdata.nasa.gov/stac/LPCLOUD"
catalog = pystac_client.Client.open(CMR_STAC_URL)

# crawl through collection descriptions and check for the sensor labels
HLS_COLLECTION_IDS = []
for collection in catalog.get_children():
if any(sensor in collection.description for sensor in ["HLSL30", "HLSS30"]):
HLS_COLLECTION_IDS.append(collection.id)

# Here is the pystac_client search that I tried that seems to be unable to find items from the requested collection IDs:
hls_history_search = catalog.search(
collections=HLS_COLLECTION_IDS,
bbox=(-91.67481, 47.863, -91.524, 47.931),
limit=100,
)

Re: LPCLOUD STAC search fails

Posted: Thu Feb 23, 2023 9:03 am America/New_York
by LP DAAC-EDL - dgolon
Hi @hrodman1 and @waltersdan I'm having out developers look into this. I'll report back when we have more information. Thank you.

Re: LPCLOUD STAC search fails

Posted: Thu Feb 23, 2023 10:42 am America/New_York
by LP DAAC-EDL - dgolon
It looks like this is a known issue. We are working on a resolution.

Re: LPCLOUD STAC search fails

Posted: Thu Feb 23, 2023 11:34 am America/New_York
by hrodman1
LP DAAC - dgolon wrote:
> It looks like this is a known issue. We are working on a resolution.
Great, thanks for the update!

Re: LPCLOUD STAC search fails

Posted: Thu Feb 23, 2023 3:02 pm America/New_York
by jteague1
We apologize for the inconvenience. We released a new version of CMR-STAC yesterday, 2/22,23, which revealed some issues with existing client implementations, such as those described here.
We have now rolled back this CMR-STAC release, restoring functionality to the previous baseline.

Thanks for your patience and we apologize for the inconvenience.

Re: LPCLOUD STAC search fails

Posted: Thu Feb 23, 2023 3:24 pm America/New_York
by waltersdan
Thanks so much for your help and quick replies!
Cheers