Closed soxofaan closed 1 year ago
Yes, that's on the todo list for sure.
Allowing aggregate_spatial
to get vector-cube as input is necessary to proceed with the random_forest pipeline. In the meanwhile I will modify the process locally for some tests.
Closely related: #343 and #346
I'm not sure I'd add vector-cube as input for load_result and load_collection. These are compromises of simplicity and utility. You may then just add it to filter_spatial as a dedicated process that can be used to filter based on vector cubes.
Processes
aggregate_spatial
,filter_spatial
,mask_polygon
have ageometries
(ormask
) parameter that currently is of typegeojson:object
. Likewise,load_collection
andload_result
have aspatial_extent
parameter that can begeojson:object
(among other options).Once we improve support for loading/processing vector cubes (e.g. #322) these processes should also allow
vector-cube
as data type for these parameters, so that a user can pass the geometries with a file/url instead of embedding geojson in the process graph