Problem Statement

This document revision applies to Q4 2021 - Q3 2022.

The QCArchive (QCA) resource is the central data system for all quantum chemistry calculations required by the Open Force Field Initiative. Over the past two years, many high-quality datasets have been generated and stored in QCA, with subsequent use as the basis for forcefield fitting and benchmarking.

Our aim in fulfilling the items within scope is to enable the continued usefulness of QCA for OpenFF fitting, benchmarking, and exploratory science. All effort we expend on improving the speed and ease of QCA use is tied directly to this aim.

Where possible, OpenFF will work to develop as much of the functionality required into QCArchive components directly.

Scope

High priority:

• STANDARDSv3 implementation

  • will require components specific to OpenFF usage in openff-qcsubmit

  • will almost certainly require functionality in QCFractal to support persistent state needs of STANDARDSv3 that can’t currently be represented in data models

• In-server, policy-based error cycling

• 2D torsiondrive support (ensure it works as advertised; effort available to address)

Nice to have:

• Frozen QCArchive datasets with DOI minting

• Substructure search middleware / webserver.

  • New infrastructure to cache/store search results for combining entries by molecule

  • Maybe combine efforts with the benchmarking dashboard for traversing data efficiently

  • Support common cheminformatics queries

• Help port Psi4 to conda-forge

  • We have reached out, LB has enumerated some of the issues:

    • 

• Collection that supports chained operations

  • cheap optimization, expensive optimization

  • QM then MM

  • QM then Hessian point calculation

  • etc.

• Officially-supported way to build the ESP from the wavefunction in psi4 that can then be used for e.g. optimizations as the initial guess

  • need support for passing in and out of psi4

Low priority:

Uncategorized:

• Ability to run QM/MM hybrid models in QCEngine and QCArchive

• Create new versions of old datasets as demanded by users compliant with STANDARDSv3

Milestones and deadlines

Plot out milestones and deadlines using the table below. Put each milestone on its own row, @mention milestone owners, and type /date to quickly add deadlines.

Reference materials

Add links to relevant research and any other key documents

Development notes

1. Things we already know we want from experience,
   current operational needs

   • STANDARDSv3

   • in-server, policy-based error cycling

   • 2D torsiondrive support
     (ensure it works as advertised;
     effort available to address)

2. Aspirational needs, future needs

   • check that ESP needs are being met

     • Trevor Gokey , Pavan Behara , Tobias Huefner

   • ask Simon Boothroyd what needs may exist for Rosemary

3. Grant aims

   • can leave AI/ML needs to MolSSI to fulfill

   • person we hire should plan to do software engineering in service to openff-qcsubmit