Matt Thompson

Ah, I forgot we had another issue open for that, thanks!

Does the above behavior hold true for other molecules in that list?

Would it be possible to track this molecule back to another source, which may provide more information about whether or not this hydrogen should be there? `openff/toolkit/data/molecules/MiniDrugBank_ff.mol2` has no charges...

That molecule provides a nice test case; there are no explicit hydrogens to cause the ambiguity of the molecule above (a mix of explicit hydrogens and some hydrogens that could...

Agreed, another solution would be to convert the wrappers for the user. This would be backwards-compatible at a high-level but would introduce some behavior changes, i.e. a specific exception that...

Here's a look as some behaviors in this implementation: ```python3 >>> from openff.toolkit.typing.engines.smirnoff.parameters import vdWHandler >>> from openff.toolkit import ForceField >>> ForceField("openff-2.0.0.offxml")['vdW'].long_range_dispersion 'isotropic' >>> vdWHandler(version=0.3).long_range_dispersion 'isotropic' >>> vdWHandler(version=0.3).version >>> vdWHandler(version=0.4).long_range_dispersion...

Closing as this has stalled, but it is linked in the EP for posterity.

It would be useful to update this unit test (and the typos in the comments 😅) in the process of tackling this issue: https://github.com/openforcefield/openforcefield/blob/8149b36e73b57cab83c958b1da4ca6d775093c34/openforcefield/tests/test_toolkits.py#L2456-L2473

Apologies for the wall of text, I just got really interested in this for some reason - the short of it is that I think the original issue is resolved,...

I have a basic singleton implemented at the below commit. Making it an actual class has the benefit of enforcing its contents at instantiation but using `GLOBAL_TOOLKIT_REGISTRY` as a variable...