Support UDP parameters in run_udf calls

We recently had a couple of use cases where a user wanted to access UDP parameters from UDF functions (e.g. apply_datacube) through the context argument.

Currently this has to be done with a bit of boilerplate code like for example

    fapar_masked = S2_bands_mask.reduce_dimension(dimension="bands", reducer=PGNode(
        process_id="run_udf",
        data={"from_parameter": "data"},
        udf=udf,
        runtime="Python",
        context={"biopar": {"from_parameter": "biopar"}}
    ))

This context/parameter handling should be supported directly by reduce_dimension() and the like, or better: through a helper for easily building a run_udf node

Start by tracing reduce_dimension() and related process-building entry points, then compare them with the run_udf node example using context and UDP parameters. Define how parameters should reach apply_datacube through context, and consider whether direct support or a run_udf helper gives a complete, consistent API.