[ENH] more specific bids entities and fixes to be bids compatible

[36000]

This is an attempt to address some of the issues raised in #1013 , and to make pyAFQ outputs more bids compatible

[36000]

produces names like this:

[arokem]

I think that should keep "space" (rather than "coordsys") as that's already an existing BIDS entity.

…

On Thu, Sep 21, 2023 at 11:13 AM John Kruper ***@***.***> wrote: produces names like this: [image: image] <https://user-images.githubusercontent.com/9562159/269715362-ed9086a2-340f-44f1-84b0-74425b69d835.png> — Reply to this email directly, view it on GitHub <#1035 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAA46NTICEMREF4PX24EIHLX3R7ULANCNFSM6AAAAAA5B6L6M4> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[36000]

But space is a different concept. You could use RASMM in template space or in subject space. I think if we use too many existing BIDS entities, we will just create more conflicts in the file name. It is better to have entities that are specific, no?

[36000]

drop coordsys, its implied

[36000]

This is going to make it so pyAFQ doesn't recognize old datasets. how do we handle this? do we need a function that updates the names? This could get tricky to maintain, especially as the BIDS spec keeps getting updated

[36000]

@arokem the docs are failing with this error that I do not understand:

Handler <function run_autoapi at 0x7f394a2d2a60> for event 'builder-inited' threw an exception (exception: 'Module' object has no attribute 'doc')

[arokem]

Yeah - we're seeing this across all our CI, so this is not specific to this PR. I wonder if it's related to a recent (Sep 13th) release of Sphinx. Would you mind pinning sphinx to the 7.2.5 and seeing whether we still get this?

…

On Thu, Sep 28, 2023 at 1:55 PM John Kruper ***@***.***> wrote: @arokem <https://github.com/arokem> the docs are failing with this error that I do not understand: Handler <function run_autoapi at 0x7f394a2d2a60> for event 'builder-inited' threw an exception (exception: 'Module' object has no attribute 'doc') — Reply to this email directly, view it on GitHub <#1035 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAA46NVYEBESBFUCZJBC5K3X4XP5XANCNFSM6AAAAAA5B6L6M4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[arokem]

OK - back to this one, now that we have the CI issue under control 🙌

I think that the issue of support for old datasets is of real concern, but I am not sure how to handle this. I think that maintaining converters between datasets from different versions of pyAFQ would become a real hassle to keep track of, and I do think that we want to keep improving here. My inclination would be to address issues with old and less compliant naming schemes as users bring this up. I wonder whether we'd need to add a documentation page that keeps track of the naming schemes as they evolve, so that at least there is one place that we could refer to for guidance about this.