Closed ajkerr0 closed 8 years ago
Adding atomtype definitions carefully should fix the atomtype issue mentioned above. This means that bondtype problem will need to be dealt with explicitly, which is probably better because hydrogenation wouldn't likely take care of it. Going to close this issue.
To be clear: adding NEW atomtypes is not what we're doing, just making sure the sp carbons get assigned to aromatic.
This eliminates problems with atomtype and bondtype assignment.
With atomtype, there are no interactions between 'CA' and 'CZ'. Hydrogenation would make every carbon atom 'CA'. With bondtype, finding valence states for the large structures with the mixed # of neighbors is a big O problem. Hydrogenation should mean the first suggested valence state should work, but this might change with the chains being added.
Defining my own atomtype (and carefully defining the interactions) could eliminate the first problem. That would leave me solving the bondtype problem using the method I had developed before, in which the user can hardcode valences for a subset of the atoms and the bondtype program only searches for the valence state which fixes the leftover subset.