Better interface for run_udf without data parameter.

[EmileSonneveld]

Right now, a user needs to do something like this:

url = "https://raw.githubusercontent.com/Open-EO/openeo-geotrellis-kubernetes/master/openeo-geopyspark-k8s-custom-processes/src/openeo_geopyspark_k8s_custom_processes/cwl/dummy_stac.cwl"
cube = connection.datacube_from_process(
    "run_udf",
    data=None,
    udf=url,
    runtime="EOAP-CWL",
    context={}
)

But a nicer interface would be better

[soxofaan]

#845 is about replacing the somewhat datacube_from_process() with more succinct process() method.

But I'm not in favor of giving data a None value by default. That only saves on writing one simple line of code for an extreme niche use case, while making things more error prone for everybody else (when they forget to provide data to their UDF).