QMTP - HBI - resonance isomers #100
Description
When performing a QM calculation on a species with several resonance isomers, I assume they should all be the same. RDKit may estimate slightly different starting geometries depending on where it thinks there are double bonds or radicals, but the electronic structure calculation should be a mixture of all the resonance isomers so there is no point in calculating the same thing several times. (right?)
However, when doing a hybrid group/QM for radicals - eg. the hydrogen bond increment (HBI) method - the saturated structure that you end up with and run QM calculations for depends on which resonance isomer you saturate. Should we in fact be running QM calculations on all the possibilities (using each resonance isomer as a base), apply the corrections, and then choose the lowest (?) energy?
I suppose the chances are that the saturated species will already have been calculated (or will need to be at some point) so as long as we re-use existing calculations it might not be too expensive.
Anyway, this issue should remind us to implement or decide against this policy, which currently doesn't exist on the QMTP branch.
I have just noticed this is issue 100. I should have reserved that milestone for something more impressive. Ah well.