Write out ALL conformers of successful input molecules (I think the previous form of the workflow may have only written out the FIRST conformer, but since we want this to become the input for the QM step, it should write them all out)
Only test conformer 00 of each unique molecule
If a molecule PASSES coverage checking, write out ALL conformers to the output directory
If a molecule FAILS coverage checking, all its conformers should be output to error_mols
(maybe not super important) Support dataset "appending", where there are new unique molecules in the input directory, and only those get their coverage re-analyzed.
Newly-appeared errors from an "appending" run should get new error_mols indices (that is, if error_mol_{0...3} existing before the appending run, then any new error mols should start at index 4)
00
of each unique moleculeerror_mols
error_mols
indices (that is, iferror_mol_{0...3}
existing before the appending run, then any new error mols should start at index 4)