2020-08-21 QCA Submission Meeting notes

Disacharrides

Josh

• Have everything we need from David Cerutti for now; will work on the tarball next week

ANI2X additional compute

Josh

• add an ignore_error flag for submissions to allow forcing, but requires explicit usage

MM/ANI2x submission from Dominic Rufa

David

• Josh: if it’s any of the JACS molecules, should already be in there

• David: not sure if it’s an existing dataset/molecules, or a new set; will find out today

Theory Benchmarking set

David

• David: planning to make a v2 that combines these three submissions into one with multiple specs according to our STANDARDS

DANCE datasets

Trevor

• David: pinging Jessica if these are still of interest

• Trevor: original submission had wrong indices on torsions; needs a rework of the torsion indices at the very least, version bump, run through QCSubmit

Prod environments

David

• David: could I get a review on this PR? Just added pytorch, torchani to main env

• Trevor: I’ll review, test each one

  • Warning: you have pip-installed dependencies in your environment file, but you do not list pip itself as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging you.

  • ANI env

    • ResolvePackageNotFound:

      • pytorch=1.6.0

      • pulls pytorch-nightly/linux-64::pytorch-1.7.0.dev20200821-py3.7_cuda10.2.89_cudnn7.6.5_0

      • EnvironmentSectionNotValid:

        • channel_priority

        • ERROR: Invalid requirement: 'torchani =2.2' (from line 1 of /home/tgokey/staging/condaenv.haxwix3v.requirements.txt)
          Hint: = is not a valid operator. Did you mean == ?

STANDARDS expansion

Trevor

• Trevor: adding additional required fields for this declarative policy document

  • also added a major version for the standard itself, which can be used to add meaning to dataset versions

• Added a statement on torsion drives that we won’t accept (rings of 3,4,5,6 atoms)

• David: can you put this as a PR so we can discuss and get it merged? Will be very much inform how we evolve infrastructure to support future datasets

PCM and ESPs

Josh

• Josh: will give PCM a try with QCEngine to see if it works with Psi4 as expected; we’ll see if the parameters get passed all the way through

• Waiting on Hyesu for info on ESPs usable from wavefunction