Noctua form refactor concealing GO-CAM model

[kltm]

Project link

https://github.com/orgs/geneontology/projects/148

Project description

This project aims to ease curators needing to reason about modeling and increase curator throughput in the Noctua form by:

• Refactoring form and flow to (better) hide the GO-CAM model from users doing "GAF/GPAD annotation"
• Simplify inputs so that only what is needed to represent GAF/GPAD is present
• Take into account curator feedback to increase throughput
• Consider other techniques to increase curator throughput (e.g. keyboard-only flow)
• (New) metadata to be displayed in search and new form when not compatible

As an add-on for this project, we'd be piloting automated/augmented testing (i.e. attempting some de-Kimberlification) for the user interface, using Kimberly's outage SOP as an initial basis. (noting deferred for the moment; possibly a different project)

While the above are the main deliverable targets, the following will be considered in engineering:

• The client code will be simplified, hopefully greatly, by removing the need for handling GO-CAM and edge cases
  • This will enable faster "small fixes" moving forward
  • Will enable more broad developer participation in the code base
    • Will include more detailed developer documentation
• The code product would be delivered as a "workbench"
• Some TBD engineering will need to be done to ensure that the newly simplified model of the form remains compatible with the current Minerva backend
  • Some form of semi-"bicameral" Minerva may need to be enabled by an additional model-level property for filtering (we wouldn't want this to affect the Minerva API beyond a parameter or so)
  • The client code could optionally more aggressively filter or collapse on its own (but that could defeat our simplification goals)

Outside the scope of this project would be:

• Adjustments to Minerva and data model beyond the addition of properties (for use in filtering)
• Changes in libraries or other client infrastructure
• Changes in the data upstreams to the client

Deferred topics are:

• Creation of new backend database
  • I.e. anything that smacks of https://docs.google.com/document/d/18Xc_9oKr9aETVZnFPCuAuMcnKCgLw-ih7WYWy5Ke-5M/edit#heading=h.yze1zl8h2s2l
  • Full stack code rewrites
• Larger changes client infrastructure (i.e. frontend code rewrite)

With this scope and restrictions, we hope to keep this as a "minimal" project, with larger changes, if any, put off to a TBD date.

PI

PaulT

Product owner (PO)

Kimberly

Technical lead (TL)

Seth

Other personnel (OP)

Tremayne, Patrick

Technical specs

Technically, if as above, there are no technical specs necessary as it would be the same scope as the current form.

Other comments

P2GO-like mockup from Kimberly: https://docs.google.com/drawings/d/1SoqT21fopJVWempRda-YCCgyRQDYw4zumAhD4fvxTxE/edit

Tremayne's initial repo for draft work: https://github.com/geneontology/noctua-standard-annotations