2022-09-12 Core Developers meeting notes

General updates

• DD offline/largely unavailable for 9 days starting Thurs. JW will run F@H meetings.

Individua

l updates

• JW –

  • Followup workshop planning - Tricky because my constraints are 1) open invite with no zoom bombers 2) I don’t want to run an email campaign and get into conversations with all attendees 3) record how I did it this time so anyone else can do it in the future. See 2022 Follow-up workshops

  • I took too long to get biopolymer support out so OpenFE is now committed to a first-iteration prototype for protien loading/representation that doesn’t use the Toolkit. I’m going to get in contact with them to rein in their second iteration to be more integrated with the toolkit.

  • Bespoke deployment support for Arjun. JHorton helped a lot too. Had great feedback! I need to update install instructions/methods and other parts of the docs.

• DN

  • Been think about how we’ve been working since the annual meeting - Would be good if we could prepare the product to be given to future generations. I think it’s common to not understand the role of PM - One of my big goals is to make sure that progress can continue after there are different people in all of our roles. But one thing that I’d like built is a WBS dictionary. It’s hard for outsiders to understand what each item is - and this means that new team members and outsiders won’t understand this either. So the goal would be that each item in the WBS would have a short description in a document, with one paragraph as an absolute max. This would be a huge plus, so if this would be helpful for you I’d encourage you to try it out.

    • CC – Are you envisioning a google doc or confluence page or something else?

      • DN – Confluence would be a little better. GDocs is more natural for the WBSes since they should undergo constant editing/updating, but they can link to the confluence pages.

      • DD – It may make more sense to have these dictionary entries be in the notes of the gdocs slides - otherwise we may have trouble keeping everything together.

      • DN – We could do that, but I’d still like the WBSes to be findable from Confluence.

      • JW – Agree with DD - Should go in the notes slides - Let’s start refreshing the WBSes at our check in on wednesday.

• MT

  • from_openmm in Interchange example “working”

    • Very early example:

  • Updating openmmforcefields CI

  • Updating science-facing software packages

    • New Evaluator release expected this week

    • Updating Recharge - delayed for the time being

• DD

  • Protein-ligand-benchmarks

    • got AlchemicalNetwork roundtripping working in neo4j storage system for fah-alchemy

      • same approach works for fetching other GufeTokenizables out of the database

    • working this week on FAHScheduler server (distributed.core.Server subclass), with dask-scheduler+dask-worker

    • also building out FAHAlchemyExecutor (server with fastapi RESTful API) and FAHAlchemyClient

    • coordinated effort with OpenFE, Chodera Lab on key components they are building out

  • QCArchive

    • at the moment just keeping workers running; will need to sync up with Pavan on the state of our submission needs to make sure compute is appropriate

• PB

  • Out last week, no update this week

• CC

  • Updated Protein FF Project Plan based on feedback from Biopolymer FF meeting

  • Working on parameter fits targeting protein QC datasets based on workflow from PB. Resolved an issue with assigning library charges correctly last week, continuing setting up ForceBalance runs this week

  • Capped 3-mer backbones QC dataset here has stopped making progress. This dataset is not blocking my current work but anticipates future needs. Any blockers from QCFractal workers?