[Compilation]: Future Improvements & Ideas

[CodyCBakerPhD]

This is a meta-issue compiling many individual ones; with the current round of work on the GUIDE coming to an end, the individual issues will be closed for bookkeeping but may be re-opened when/if they are tackled in the future

These are organized according to an estimate of 'how much time' each enhancement would take

1. Simple

These are relatively minor, could still take a day or two for experience developers to thoroughly test.

• 1.a https://github.com/NeurodataWithoutBorders/nwb-guide/issues/173

  • A button on the frontend to enable easy copy/paste of existing pipelines; this is technically possible ATM by manually playing with the folders in the GUIDE home folder, but would need a set of structured interactions with various validation assurances

• 1.b https://github.com/NeurodataWithoutBorders/nwb-guide/issues/191

  • An initial implementation was added in https://github.com/NeurodataWithoutBorders/nwb-guide/pull/625 and could theoretically be restored to some effect; was deemed too low of a priority to finish polishing and test all edge cases. Otherwise, feedback from some users was that it could reduce the amount of clutter that occurs when all accordions are expanded

• 1.c https://github.com/NeurodataWithoutBorders/nwb-guide/issues/682

• 1.d https://github.com/NeurodataWithoutBorders/nwb-guide/issues/853

  • Not a huge priority since that feature is not expected to be used super often, and as mentioned in the issue, still performs the functionality just not the appearence

• 1.e https://github.com/NeurodataWithoutBorders/nwb-guide/pull/833

  • When going through a guided mode pipeline (reliant on NeuroCon interfaces) there are a handful of NWB Inspector checks that the user has no control over so we shouldn't warn about it on the report page (they should be ignored). The proposal was to make two separate endpoints for guided mode vs. standalone page (which should still display such things generally)

• 1.f Expose number of jobs control to multi-session conversion (currently parallelizes over 2)

• 1.g https://github.com/NeurodataWithoutBorders/nwb-guide/issues/905

  • Assets have metadata that indicate how they are generated and what software is associated with the files. It would be great to add NWB GUIDE, which would help us track the usage of this app and provide better provenance for those assets.
  • Should be a pretty easy injection to sent this info along the DANDI API on creation, or sent as a separate request

2. Complex

These have multiple interaction points requiring extensive integration and testing. Probably a week or more of effort.

• 2.a https://github.com/NeurodataWithoutBorders/nwb-guide/issues/584
  • Icon and popup interaction should be simple, most of work would be documenting all the information.
• 2.b https://github.com/NeurodataWithoutBorders/nwb-guide/issues/592
• 2.c https://github.com/NeurodataWithoutBorders/nwb-guide/issues/684
  • May just boil down to better names for certain pages
• 2.d https://github.com/NeurodataWithoutBorders/nwb-guide/issues/761
• 2.e https://github.com/NeurodataWithoutBorders/nwb-guide/issues/805
  • Might just be a special NWB Inspector check that compares all the different fields of the file
• 2.f https://github.com/NeurodataWithoutBorders/nwb-guide/issues/806
  • better keyboard shortcuts (hard to perfect cross-platform, very hard to test)
• 2.g https://github.com/NeurodataWithoutBorders/nwb-guide/issues/852
  • This was seen in many user tests on the tutorial as well; confusion between the SpikeGLX Converter (multi-stream) and SpikeGLX Recording (single-stream); the tutorial only uses a single stream, but one must read the instructions carefully and most people rush it and choose the wrong one
• 2.h Export NeuroConv and/or PyNWB code that a GUIDE pipeline is equivalent to

3. Unknown

These could be easy or could be difficult. Can't know without a deep dive.

• 3.a https://github.com/NeurodataWithoutBorders/nwb-guide/issues/153
  • first of several issues related to improved copy/paste from the current HandsOnTable used on the subject page and a few metadata sections
• 3.b https://github.com/NeurodataWithoutBorders/nwb-guide/issues/181
• 3.c https://github.com/NeurodataWithoutBorders/nwb-guide/issues/214
  • Could also work in conjunction with better minimal window sizing
• 3.d https://github.com/NeurodataWithoutBorders/nwb-guide/issues/504
  • A pretty subtle distinction of behavior as discussed in the original issue
• 3.e https://github.com/NeurodataWithoutBorders/nwb-guide/issues/517
  • Like the copy/paste but for dragging; a feature common to find in places like Google Sheets or Microsoft Excel
• 3.f https://github.com/NeurodataWithoutBorders/nwb-guide/issues/660
• 3.g https://github.com/NeurodataWithoutBorders/nwb-guide/issues/747
  • General feature for any button that triggers a backend endpoint; grey out and disable the button until the Flask server signals that it is running
• 3.h https://github.com/NeurodataWithoutBorders/nwb-guide/issues/807
  • it might be simpler to not distinguish warnings vs. errors to reduce the cognitive burden on users
• 3.i https://github.com/NeurodataWithoutBorders/nwb-guide/issues/850
• 3.j Drag-and-drop for file/folder selector components does not work on Windows production (possibly dev as well)
• 3.k Explore adding the GIN testing data repositories as git submodules https://github.com/NeurodataWithoutBorders/nwb-guide/pull/910#discussion_r1716188117

[bendichter]

Another idea:

On the path expansion page, it would be nice to see the session and subject id values that were extracted from the paths. It may not be what you expect if the values are poorly specified or the paths don't follow a strict structure.