Open kyleredilla opened 2 years ago
Definitely include that HUC in validation efforts ^^ being on the southeastern border, also look at other edge cases.
This one is really, really hard. Unclear about the juice V squeeze: it can be extremely challenging to fully capture the range of interactions between our places (points, polys) vs. nodata vs. dataset extent. @kyleredilla what do you think we should do here?
I think we should consider doing at least a couple of simple tests for each endpoint, maybe even just one for each of point query, poly query, nodata query, and query beyond dataset extent. It just seems like we should have at least some sort of validation for the endpoints, which often are summarizing/reducing query results from Rasdaman.
We need to get some scripts or something for validating the endpoints that could be easily done before any deploy. Think this might be separate from any more precise unit testing that should be considered as well, but we need scripts that calculate what is expected from the endpoint using the raw base data (ie not queried from GS/Rasdaman) and compare with what endpoint returns.