Prototype matching issue

The prototype detection in SystemNormalizer is currently failing to match at least certain prototype structures. For example detecting the diamond prototype is failing. Most probably affects many other prototypes as well. The issue is also present in the current production system as many structures do not have section_prototype due to failed matching.

The issue boils down to information in nomadcore: python_common/common/python/nomadcore/structure_types.py. The normalized wyckoff positions listed there ("normalized_wysytax") seem to be wrong as compared to what MatID produces.

For now the easiest solution will be just to correct these entries by using the Wyckoff set normalization provided by MatID (which slightly differs from what is in AFLOW).

If I remember correctly @datom was working on this? Do you know if the file used to generate structure_types.py still exists? In the future, we will most probably want to update our information from AFLOW, and thus having this script within nomad-FAIR would be nice.