Several bugs in Lammps IO

[orionarcher]

I've encountered several bugs in the LAMMPS IO when using 'opls' dihedral type and 'cvff' improper style. I believe that I can fix them, but I am wondering if those fixs would be merged.

One of the errors is already addressed by PR #349, but that has been open for two years.

The other error occurs in the canonical_dihedral function in lammps_parser.py.

    def canonical_dihedral(self, params, dihedral, direction='into'):
        """Convert from the canonical form of this interaction. """

        if direction == 'into':
            converted_dihedral = dihedral  # Default
            if dihedral == ProperPeriodicDihedral:  # Proper dihedral
                convertfunc = convert_dihedral_from_proper_to_trig
                converted_dihedral = TrigDihedral
            elif dihedral == ImproperHarmonicDihedral:
                convertfunc = convert_nothing
            elif dihedral == RbDihedral:
                convertfunc = convert_dihedral_from_RB_to_trig
                converted_dihedral = TrigDihedral
            elif dihedral == FourierDihedral:
                convertfunc = convert_dihedral_from_fourier_to_trig
                converted_dihedral = TrigDihedral
                # Now actually convert the dihedral.
            params = convertfunc(params)

If dihedral == TrigDihedral at the beginning, then convertfunc will never be defined and an error will occur. An easy fix.

If I create a PR and fix these issues, is any developer available to review and merge them? I'd like to contribute but it appears the code has not been updated in some time.

[mrshirts]

Hi, Orion-

I have run out of time to support InterMol right now because of lack of funding/students and because of the future development of interchange. See https://github.com/ParmEd/ParmEd/issues/1204 for more discussion . . . I'd like to promise review, but because we never set up great testing on InterMol when we started in the olden days, I can't guarantee it won't break things, and I currently don't have time to do that sort of testing when I have grad students to mentor on their projects.

[yel21]

Dear Sirs,

If any chance to solve Intermol problems that related to dihedral angles for conversion from LAMMPS to GROMACS?

[orionarcher]

Hello @yel21,

Unfortunately this issue was never resolved. As alluded to above above, this is a planned feature for the not-yet-completed force field interchange library. Until that package is more mature, it is unlikely we'll have a solution to cross-application force field conversion.

In the meantime, the OpenFF Toolkit has some great tools for parameterizing forcefields.

[yel21]

Hello @yel21,

Unfortunately this issue was never resolved. As alluded to above above, this is a planned feature for the not-yet-completed force field interchange library. Until that package is more mature, it is unlikely we'll have a solution to cross-application force field conversion.

In the meantime, the OpenFF Toolkit has some great tools for parameterizing forcefields.

Hello @orioncohen, Thank you for answer.