Issues in v2ff Branch: Orca QM Module & Gromacs Export

[mmagithub]

Hi Selim,

I explored the v2ff branch ahead of the full release and have encountered a couple of issues:

1. Orca QM Module:

Problem:
The module fails to extract point charges regardless of whether ESP or Hirshfeld methods are used.

Troubleshooting:
I attempted modifications in both qm.py and orca.py with no success.

Affected Versions:
This issue appears consistently with both Orca 6 and 5.04.

2. Format Exports

A. OpenMM XML Export
Observation:
Saving to OpenMM XML works fine, though note that all charges default to xtb (likely related to the Orca issue).

B. Gromacs Export
There are two specific issues with Gromacs format saving:

i. Forcefield Settings Syntax

Current Code:

class Gromacs(ForcefieldSettings): always_on_terms = ['bond', 'angle']

Proposed Revision:

class Gromacs(ForcefieldSettings): _always_on_terms = { 'bond': ('harmonic', ['harmonic']), 'angle': ('harmonic', ['harmonic']), }

This change should better represent the intended mapping between terms and their parameters.

ii. Dihedral Parameters (RB Formalism)

Problem:
The dihedrals saved according to the RB formalism are inconsistent; it seems necessary to pad these with zeros (or apply another suitable approach) to ensure correct indexing.

I have attached a modified gromacs.py (gromacs_updated_py.txt) module for your review. Please let me know if these modifications make sense or if you have any further suggestions.

Thanks for all your efforts on this project. This is by far the best parametrization tool I've encountered, and I’m excited to test the full version as soon as it’s available. I’m also happy to try out any work-in-progress or pre-release versions if needed.

Best regards,
Marawan

gromacs_updated_py.txt