Handling TENDL-2014 processing errors in ALARAJOYWrapper

[eitan-weinstein]

This PR introduces some minor updates for the PENDF production stage of ALARAJOYWrapper to allow for more robust handling in the case of a TENDL file missing necessary energy-dependent data in the unresolved region. The motivation behind this update is that I wanted to produce an ALARAJOY/TENDL-2014 library as well for my analysis, figuring that one possible solution to the question on FISPACT-II/ALARA comparisons that I posed in the discussion in #277 would be to compare ratio plots of FISPACT-II TENDL-2017/TENDL-2014 against ALARA/ALARAJOY TENDL-2017/TENDL-2014 to filter out the impacts of the fundamental differences in the two software's decay chain production. Additionally, while the primary motivation for building ALARAJOYWrapper is to make ALARA compatible specifically with FENDL3.2x (TENDL-2017 cross-sections), it would be undoubtedly valuable if this tool could be used for any TENDL release.

When running ALARAJOYWrapper with the TENDL-2014 distribution from IAEA, I ran into two issues. First, for a few files, UNRESR was failing with this NJOY error message:

***error in rdunf2***energy dependent data undefined
                      between  8.9001E+03 and  8.9001E+03 eV

This message indicates a deprecation of the data as the NJOY manual describes in section 5.6:

When using unresolved resonance formalisms with energy dependent parameters (e.g. the fission width, etc.), these data need to be defined over the entire unresolved resonance region. In rare cases, such as ENDF/BVII.0 Pu239, this is not the case, leading to NaN cross section values. This is an evaluation error.

While running UNRESR is generally preferable and part of our canonical methods for ALARAJOYWrapper, I think an evaluation error of this sort is not resolvable by us. Even without UNRESR, however, GASPR and GROUPR can still be run and the best possible data can be extracted from any given file with this issue.

A second issue in the TENDL-2014 data is that there are a number of files that are misnamed, such as the files for ²⁸⁵Nh and ²⁸⁶Nh, which for some reason have the symbol A3 instead. This update will catch this type of issue, identify the proper nuclide identifiers from the file itself and then rename it to match the correct nuclide data.

Finally, for a number of TENDL-2014 nuclides the RECONR error tolerance that was otherwise a fixed value was leading to NJOY stalling out and unable to resolve the reconstructions. I've included a looping mechanism to test out various timeout lengths before increasing this tolerance value to try to isolate the need to raise it only to cases where NJOY is truly unable to perform the calculation, rather than just doing so slowly.

[gonuke]

Can you summarize how many files and (if not too many) which files are affected by this UNRESR problem?

Also see my comment about handling poorly named files with additional workflow steps rather than while we are processing them.

[gonuke]

This seems intrusive to this code to handle a rare case. Can we just introduce something to the workflow that fixes the names of these files or something similar and run that before we process the data?

[eitan-weinstein]

Per my comment in #271 about moving away from relying on filenames for cross-section data, I'm wondering if it would make more sense to avoid this entirely both because of its intrusiveness but also because the actual data contained within the file is correct, just the naming is incorrect. Ultimately, I don't think we care that much about toying with the individual files in a TENDL or decay data release unless the data themselves need to be modified/reformatted.

[eitan-weinstein]

Rebased to follow #281, which should avoid the need for the special handling for misnamed files.

[gonuke]

I feel like this should not need a rebase again, but it looks like it does.