Closed umesh-timalsina closed 1 month ago
I would rather avoid this, as it muddies the defined structure of the core object model; atom indices do not define interactions in force fields, other definitions associated with the typing procedures do. There may be situation in which keys needs to be generated on the fly, but it would be better for these keys to represent something in the force field, i.e. bond parameters having keys like "CT-CT"
for aliphatic carbons in OPLS-AA.
We've done okay smooshing together things into messy strings, which isn't the beset solution but works
Related to #114. Currently, I store the
Bond
,Angle
andDihedral
parameters based onatom_indices
. It might be nice to be able to support thePotentialKey
type to be a tuple of atom indices.