Create openeo UDP for generating worldcereal products

[jdegerickx]

We now have a preliminary version of a workflow generating a default worldcereal product for a custom AOI and year. For now, only crop/no-crop classification is supported, will be extended with crop type as well... Also, currently the generate_map function does not support supplying a custom model URL.

Demo can be found here --> https://github.com/WorldCereal/worldcereal-classification/blob/demo/notebooks/system_v1_inference.ipynb

Perhaps we can already convert this into a preliminary UDP for further testing (also in the Maps4GPP project)?

@jdries, feel free to assign this issue to someone else of course ;)

[jdries]

The UDP was created and added to APEx, use of it is documented here: https://worldcereal.github.io/worldcereal-documentation/usage/processing_api.html

I created it manually, better option is to do it via Python, to easily keep the latest version in sync.

[jdries]

@soxofaan for next steps, write script that parameterizes the datacube that is contstructed here: https://github.com/WorldCereal/worldcereal-classification/blob/main/src/worldcereal/job.py#L162

important parameters are:

• spatial extent
• temporal extent (with default settings for 2021)

[jdegerickx]

Update: the generate_map function has been extended and can also support crop type inference runs, even with custom models --> https://github.com/WorldCereal/worldcereal-classification/blob/main/src/worldcereal/job.py#L177

So we would need a UDP which can run this function based on:

• spatial extent
• temporal extent
• product type
• (optional) product parameters I would say that for output format, backend and tile size we can just use the defaults as specified in the generate_map function arguments.

See also here an example of how this function is used in a demo notebook --> https://github.com/WorldCereal/worldcereal-classification/blob/extra-demos/notebooks/system_v1_inference.ipynb