This is a legacy from the very first commit of the DryRunDataCube where aggregate_spatial returns a AggregatePolygonResult instead of a DryRunDataCube. Raster_to_vector also returns an AggregatePolygonResult.
Normally we should be able to change this without breaking too many of the existing processes.
run_udf in the ProcessGraphDeserializer does contain the following check:
if dry_run_tracer and isinstance(data, AggregatePolygonResult):
return JSONResult({})
We should be able to just check if the current run is a dry_run instead.
Note that run_udf in the ProcessGraphDeserializer is only used for top level UDFs. It's mainly used when running UDFs on an entire vector cube.
There was also a discussion on if raster_to_vector should provide a weak_spatial_extent for the entire raster, because we don't know the geometries during the dry run.
This is a legacy from the very first commit of the DryRunDataCube where aggregate_spatial returns a AggregatePolygonResult instead of a DryRunDataCube. Raster_to_vector also returns an AggregatePolygonResult.
Normally we should be able to change this without breaking too many of the existing processes.
run_udf in the ProcessGraphDeserializer does contain the following check:
We should be able to just check if the current run is a dry_run instead.
Note that run_udf in the ProcessGraphDeserializer is only used for top level UDFs. It's mainly used when running UDFs on an entire vector cube.
There was also a discussion on if raster_to_vector should provide a weak_spatial_extent for the entire raster, because we don't know the geometries during the dry run.