Make PROCESS compatible with easyVVUQ for Monte Carlo runs

[ym1906]

Some thought should be given to the best way to make PROCESS interact with easyVVUQ for generating large datasets of runs with the solver and optimiser (replicating functionality with previous Monte Carlo tools).

The problem is the way PROCESS and easyVVUQ interact. easyVVUQ needs to read in PROCESS output (mfiles). In principle, this just requires a custom decoder, however in the case that no MFILE is produced (ifail 6), the entire UQ run will be killed.

My current quick workaround for this is to add a new function to PROCESS post-processing. This function converts MFILES to json files with each entry in the dict being a PROCESS output parameter.

This is easily fed into easyVVUQ, and a template of this json can be added to the working folder for the case where no MFILE is produced.

Any thoughts/comments on this approach, or a better one, would be welcome.