ErinWeisbart
commented
10 months ago I'd provide the counter argument that the cellpainting-gallery should provide data from raw images through profiles for all datasets we host. I think it's confusing to have some datasets have profiles on S3 and some where you need to instead go to a GitHub repository. Doing so decreases the utility of our resource.
I think we should make a public note that profiles are the most likely data to change and anyone using them should check linked external repositories for updates.
Shantanu noted here:
For example, s3://cellpainting-gallery/cpg0001-cellpainting-protocol/workspace/profiles is redundant with https://github.com/jump-cellpainting/pilot-data-public/tree/main/profiles.
Eventually, we'd want to get rid of that redundancy (in favor of the latter)