use OSF to store the data (and the OSF API to access it)?

[jrosen48]

Hi all, in #5, @datalorax you mentioned the possibility of using OSF to store the data (and using the OSF API via {osfr} to access it).

Just as a proof of concept, I checked whether it is possible to download a file in a public OSF repository without adding an OSF API key, and it does appear to be so. So, if this worked and was not less efficient than the current route (with the data stored in a separate repository), then it would remove the step of adding a GITHUB PAT (or any kind of key/token)... which may be cool?

@ivelasq is it worth possibly exploring in a branch? I don't think it would take too long to try out (just to gauge whether or not it may be worth it with our actual data.

library(osfr)
f <- osfr::osf_retrieve_file('63d7f')
osfr::osf_download(f)

[datalorax]

That makes sense to me! Removing the requirement for the GITHUB PAT and not having to have an OSF API Key would be great for the user-end of things and would lead to one less thing to manage on the maintainer side. Also, the data storage is probably more stable there than in an individual repo (as it is now). You could probably even remove the piggyback dependency too (although that part is kind of neat, it's always nice to remove dependencies). So this is an all-around positive move from my perspective!