Allow custom forcefield to be used in planckton-flow

[gwenwhite]

Right now planckton-flow does not allow custom forcefields to be made. It would be nice to change this for new molecules we create in mbuild.

[jennyfothergill]

Hi @gwenwhite, just to clarify this issue a little further, this may be two separate problems:

1. Gaff does not parameterize all of the nonbonded/bond/angle/dihedral types in our molecules, and we want to fix this by adding the unparameterized types to gaff. (related to debug forcefield issues: unparameterized dihedrals planckton#30) I know you are working on this right now :) Once this is finished we can submit a PR to gaff-foyer or antechamber to update the gaff repo, and update it in planckton.
2. As for a custom forcefield (i.e., the user provides the xml), this functionality can be added, but I am not sure it is ideal... in general it is better to move away from single-use custom forcefields in favor of a well-documented one under version control. (related to Plan to move away from gaff-custom forcefield and *_typed.mol2 compounds planckton#59)

[chrisjonesBSU]

" know you are working on this right now :) Once this is finished we can submit a PR to gaff-foyer or antechamber to update the gaff repo, and update it in planckton."

Wouldn't this require some testing via running simulations with the new parameters to understand if the new parameters make sense? As written, planckton will accept anything that is a foyer FF object, so planckton allows the user to use an edited/customized version of gaff. If this feature is already in planckton then it would be nice to be able to access it via planckton flow as planckton-flow essentially serves as an access point to planckton.

And I think moving away from single-use custom forcefields in favor of version controlled forcefields is kind of subjective. I think some people prefer working with a trimmed down focefield file that doesn't contain any more infoormation than needed for whatever project they are working on.

[chrisjonesBSU]

Also, I think cmelab/planckton#59 is more about removing the extra logic/steps of checking a file name, and having the gaff_custom file essentially hard coded in more-so than about moving away from custom focefields in general.