Closed SimonBoothroyd closed 3 years ago
@SimonBoothroyd thanks for putting this together this is great and helps with the awkwardness of using a factory to make the CMILS objects. WRT to the naming let's stick with from_openff_molecule
for now I think only an openff molecule is safe to use as input due to missing information in the likes of qcelemental molecules and I like the idea of making it very clear that we are using a openff molecule as our goto structure which users might not expect when working with qcfractal.
Please feel free to merge tests are broken due to fragmenter and an rdkit update which I will work on separately.
Description
This PR adds a convenience method to create a CMILES containing
MoleculeAttributes
object from an OpenFF molecule. This allows the object to be created directly without first needing to create a dataset factory first.Questions
from_openff_molecule
to make it explicit that this method creates from an OpenFF molecule rather than say a QCElemental molecule, but it may be better to just have afrom_molecule
method and handle the molecule type in the method itself, maybe making private_from_openff_molecule
,_from_qcel_molecule
,_from_XXX_molecule
methods to handle the different molecule cases.Status