Versions Compared

Old Version 1

changes.mady.by.user David Dotson

Saved on

compared with

New Version Current

changes.mady.by.user David Dotson

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Participants

Goals

  • QCFractal release process, server deployment

    • GHA for building OpenFF worker Docker image?

  • Psi4 / QCA support for PCM-based implicit solvent?

  • Update manager prod envs with MM tooling.

  • Livecoding session for INCOMPLETE restarts - see how far we can get

    • want to get current datasets unstuck if we can, even if the fix is a bit dirty

  • New datasets

    • First direct submission from David Cerutti

  • What gaps in our current pipeline do we want to address next?

    • Submission is getting easier, but what about end-user use for:

      • fitting

      • benchmarking

      • scientific investigations

  • Trawl automation issues for additional functionality:

    • Github link macro
      linkhttps://github.com/openforcefield/qca-dataset-submission/issues/53

    • Github link macro
      linkhttps://github.com/openforcefield/openforcefield/issues/516

Discussion topics

Time

Item

Presenter

Notes

QCFractal releases

Ben

  • A couple of large migrations mid-October targeting

    • setting up for upcoming items

  • Another release Dec. or Jan. with deduplication rework

  • Targeting September for an INCOMPLETE remediation pathway

    • would require some logic in server to create new task for an INCOMPLETE optimization record

Optimizations running

Trevor

  • Seeing optimizations rolling through his manager

    • David: not sure where these are coming from; no datasets error cycling at this time

    • They are tail-ends of torsiondrives; new tasks will spawn periodically until the TD finishes

OpenFF worker docker image (MM tooling)

David

  • Decided we will move the QCFractal openff conda env files to an openff repo, with Dockerhub image builds, etc., since these change as the openff tooling changes and aren’t really coupled to fractal

  • Ben will merge Fractal#609, push new conda env files, Docker image for our managers (last time; David will get new repo in place next week)

Disaccharide submission

David

  • Trevor: Actually we should have enough information from the JSON submissions to generate CMILES; will take a stab and see if he hits any showstoppers

  • David: This would be huge!

Gaps analysis

David

  • What gaps in usability can we identify?

    • Trevor: fitting team doing their thing, benchmarking doing theirs

      • given two or more datasets, give me all data belonging to this CMILES - this is really valuable

    • David: we’ll want the vital pathways for e.g. benchmarking to go into openff-benchmark, but this doesn’t have to happen immediately if Trevor can move faster in his own repos

Psi4 - PCM support

Ben, Trevor

  • David: this will likely be coming down the pipe as an official request soon, so we should identify where changes would need to be made to support it

  • PCM support should only require changes in QCElemental, perhaps also QCSchema

Action items

  •  Ben Pritchard will merge QCFractal#609, push new conda env files, Docker image for use on our deployed managers
  •  Trevor Gokey will take a stab at extracting CMILES from the inputs provided by David Cerutti (Deactivated) in qca-dataset-submission#124.
  •  David Dotson will migrate conda env for production workers to an OpenFF repo (perhaps qca-dataset-submission), with automatic push to Anaconda cloud and build/push to DockerHub via QCA
  •  Trevor Gokey will investigate the optimizations running on his manager that we don’t know the dataset for

Decisions