Bugs on handling interactions groups during hybrid topology construction?

Hi, there,

About the function _handle_interaction_groups() of the source code file openfe/protocols/openmm_rfe/_rfe_utils/relative.py, I have several questions:

1. According to the document on the function addInteractionGroup(), it has the following feature:
   If a particle pair appears in two different interaction groups, its interaction will be computed twice. This is sometimes useful, but be aware of it so you do not accidentally create unwanted duplicate interactions.
   If so, the interactions between unique_new_atoms and unique_new_atoms will be computed twice? So does unique_old_atoms. This point should be a bug?

2. Why not add the interactions between environment and environment, which should be added by sterics_custom_force.addInteractionGroup(environment_atoms, environment_atoms)?

3. This function only deals with the sterics interactions. Why not add these interaction groups for electrostatics interactions?

Looking forward to your comments. Thanks!

Pengfei

Start with _handle_interaction_groups() in openfe/protocols/openmm_rfe/_rfe_utils/relative.py and read the addInteractionGroup() documentation it references. Trace how unique_new_atoms, unique_old_atoms, and environment are passed to the sterics custom force, then compare the handling of sterics and electrostatics. Done means resolving the three interaction-group questions and documenting or filing a specific corrective change.