Expanding conformers in a past Dataset

[jthorton]

Users may want to add extra conformations to past datasets but not want to re-roll the entire dataset through the factory, as the dataset might be large and the order of the conformers may change. One way this can currently be done is the following

from qcsubmit.datasets import load_dataset

dataset = load_dataset("dataset.json")
# loop through the dataset and make new conformers
for entry in dataset.dataset.values():
    # get the molecule and gen conformers
    mol = entry.get_off_molecule()
    mol.generate_conformers()
    for i in mol.n_conformers:
        entry.initial_molecules.append(mol.to_qcschema(conformer=i))

here users just need to check that the same conformer is not entered twice into the entry, maybe we can add some functions to datasets to automatically do this for users?

[trevorgokey]

Thanks so much for detailing this! My initial reaction is that it "would be nice" if we could run this through the conformer component so we get the provenance and all the goodness from the exposed options (e.g. rms_cutoff). I'll take a look and see what comes of it.