Inconsistency between specification and implementation: stimulus set CSV `filename` vs `stimulus_path_within_store`

[raj-magesh]

The BrainIO specification says that the CSV file corresponding to a stimulus set must have a column called filename which refers to the filepath of the image within the store (or just the filename, if the store is flat). However, in the reference implementation of the spec, I see code referencing stimulus_path_within_store. Which of the two is supposed to be used?

[ShreyaKapoor18]

Hi, I get the same error when I remove the cache. Any resolution?

[raj-magesh]

If you just need it to download and parse the datasets, I wrote some utilities to access BrainIO catalogs here:

https://github.com/BonnerLab/bonner-libraries/blob/main/src/bonner/brainio/_catalog.py

They used to work a couple of years ago; haven't tested them recently.