chemical_formula_reduced is not actually reduced

@himanel1 We based results.material definition on optimade. results.material.chemical_formula_reduced and metadata.optimade.checmical_formula_reduces seem to use the same normalizer function optimade_chemical_formula_reduced to produce its values. When fixing our optimade problems during the workshop, I changed optimade_chemical_formula_reduced, because the optimade spec wants 'C2H2' to appear as CH and not as C2H2 (as asserted in our normalizer tests). But now this breaks our normaliser tests for results.material. #60268c85

What should I do? Change also our meaning of results.material.chemical_formula_reduced? This would effect many formulas in our database. Are there different meanings for a reduced formula in MS?