There are currently three different extraction pipelines:
Patch extractions
Point extractions
Patch-to-point extractions
For the first two, worldcereal_preprocessed_inputs can be used. This makes sure that the preprocessing is always aligned between these two pipelines.
However, in its current form, worldcereal_preprocessed_inputs cannot be used for the patch-to-point extractions, since Sentinel-1 and Sentinel-2 collections are loaded from STAC instead of from collection and METEO is loaded from Terrascope instead of CDSE.
We need to either unify the preprocessing, or install a safeguard to make sure that the data preprocessing between the three pipelines is always aligned.
There are currently three different extraction pipelines:
For the first two,
worldcereal_preprocessed_inputscan be used. This makes sure that the preprocessing is always aligned between these two pipelines.However, in its current form,
worldcereal_preprocessed_inputscannot be used for the patch-to-point extractions, since Sentinel-1 and Sentinel-2 collections are loaded from STAC instead of from collection and METEO is loaded from Terrascope instead of CDSE.We need to either unify the preprocessing, or install a safeguard to make sure that the data preprocessing between the three pipelines is always aligned.