search_data results for OPERA_L2_CSLC-S1_V1 are much larger than expected

[scottstanie]

I was checking how hard it woudl be to use earthaccess to download some OPERA CSLC products (taking the small search from here as an example)

import earthaccess
earthaccess.login()

results = earthaccess.search_data(
    short_name='OPERA_L2_CSLC-S1_V1',
    cloud_hosted=True,
    bounding_box=(-117.33, 35.541, -117.880, 35.991),
    temporal=("2023-10-01", "2023-10-15"),
    count=500
)
Granules found: 2737

This is a small bounding and only 15 days, so I'd expect just a few results.

Entering this into ASF (or using asf_search) yields the expected 12 granules

I'm not sure if this is something odd or unexpected with the OPERA data compared to others, or how earthaccess would match this many granules.

[scottstanie]

Shoot I'm sorry, I just copied those parameters to test, but I didn't notice that the bounding box from that issue was incorrect and had the order (right, bottom, left, top).

Using bounding_box=(-117.33, 35.541, -117.880, 35.991) returns 10 results.

I actually saw this because the asf_search library printed the helpful warning

WARNING:asf_search:WKT REPAIR/VALIDATION: The following repairs were performed on the provided AOI:
["'type': 'REVERSE': 'report': Reversed polygon winding order"]

Not sure if that's worth adding here or not, but it helped me pick out that problem.

[mfisher87]

I really appreciate you sharing the problem solving process! It may help someone else some day :) ❤️

[jhkennedy]

@scottstanie it might be worth opening a feature request for python-cmr to do the same validation of geometries as asf_search.