I was looking into ways to connect an OpenFOAM model to my FEniCSx model and came across the experimental preCICE-adapter for FEniCSx. I understood it is not currently supporting mpi parallelization (nor vector-valued functions ?).
I was wondering if support for these features is planned at short to mid term as I would actually need both, to run on a supercomputer.
It is true that in its current state the FEniCSx adapter does not support MPI parallelization and also cannot handle vector-valued functions. In principle we have some idea on how both of these features can be implemented, but right now we do not have the resources to implement things ourselves. If you would really need these features and are well versed with FEniCSx, you could try contributing the features. We would gladly support you in making such a contribution.
Using a dedicated adapter would be probably the best solution, also in terms of reusability. However, preCICE also has excellent Python bindings. This might be an easier way to implement/couple or at least a good start for a proof of concept and some testing with FEniCSX.