Appion

Currently, Recon refinement run job generation can only be done once. I can only manually add further refinement cycles at the garibaldi and resubmit the job. This is not ideal since the job script on garabaldi-type of cluster that does not sync the job files automatically is not transferred back. It makes it possible to have inconsistent refinement parameters such as num_iter in ApRefineRunData. It would be better to make the process smoother and less-error prone.