Missing WBO OpenFF Full Optimization Benchmark 1

[jthorton]

Some of the optimizations done as part of the OpenFF Full Optimization Benchmark 1 collection have missing WBO in the results despite it being requested, I think this is due to different qcengine versions on a manager which pulled down the job, meaning that for some molecules other connectivity checks are required.

Entry	first gradient id	wbo	qcengine version
CC(=O)O[C@H]1Cc2ccccc2N(c3c1cccc3)C(=O)N-0	6479626	True	v0.10.0
c1cc(c(cc1O)[C@@H]2CCN2)Br-0	3091838	False	v0.7.1

[mattwthompson]

Having WBO data in these records will make benchmarking/validation much easier, since it would allow us to avoid using more approximate methods and WBO comparisons seem like they'll be the best way to catch unexpected things happening during optimization (proton transfer, h-bonds, etc.)

[j-wags]

Interesting. I suspect we'll have the resources to simply redo these jobs once the current gen2 fitting datasets are complete. I'll see if we can do some queue planning for the next few weeks at Thursday's QCFractal call.