Allow processing of all JEF(F) and ENDF evaluations in parallel mode #7
Description
So far, parallelization of PyNjoy processing is only effective for sampled nuclear data (for Total Monte-Carlo applications). A generalization of the parallelization to the (usual) nuclear data best estimates is desirable, dealing with both european evaluations (from JEF-2.2 to JEFF-4) and american evaluations (from ENDF-B/VI to ENDF/BVIII).
Only PyNjoy2012 does produce Draglib for these evaluations, so far. This development would allow to deprecate PyNjoy2012. Moreover, this would make disappear the huge code duplication constituted by the PyNjoy.py and PyNjoy_mp.py scripts.
Such a development would most probably require to rewrite PyNjoy_mp.py, because calling it in a dataset is too complex, showing too much of what should be internal to PyNjoy.
Also, it would be possible to simplify the management of the production of all the evaluations, for the different energy meshes, by mutualizing (in a separate file) the options related to each evaluation.
Rigorous non-regression testing (compared to PyNjoy2012) is highly required for that task.