The main fix is to update the failing test_sample_data_point test caused by an apparent change in the EE server error handling. Previously, attempting to generate random points from a point geometry would raise an error, and sankee would catch that and rethrow. Now, we catch more specifically when an empty feature collection is sampled.
Some other minor changes:
Added the more specific SamplingError exception for all errors related to data sampling.
Fixed a pandas deprecation warning related to using the builtin sum in a groupby.
Simplified the CI by removing the unneeded Python script.
Added linting to the CI.
Switched from black to ruff for formatting.
Fixed auth errors in the CI by updating the repo secret to include the full contents of the credentials file rather than just the refresh_token.
The main fix is to update the failing
test_sample_data_pointtest caused by an apparent change in the EE server error handling. Previously, attempting to generate random points from a point geometry would raise an error, andsankeewould catch that and rethrow. Now, we catch more specifically when an empty feature collection is sampled.Some other minor changes:
SamplingErrorexception for all errors related to data sampling.sumin a groupby.credentialsfile rather than just therefresh_token.