Separate Download Bool for images and workspace
Might want to mount one bucket but download from the other (SOURCE_BUCKET vs WORKSPACE_BUCKET). (e.g. reading images off of the public bucket Cell Painting Gallery, mount is good but workspace files are likely in another bucket)
We currently only support: Same bucket - Download = True for both Same bucket - Download = False for both
I think we want to support: Different bucket - Download = True for both Different bucket - Download = False for source, Download = True for workspace
I think we don't need to support: Different bucket - Download = False for both (i.e. mounting two buckets)
I'm indifferent to supporting: Different bucket - Download = True for source, Download = False for workspace
I'm leaning toward:
- we have mounting for just the SOURCE_BUCKET (where the images are) and default to download workspace files regardless of download bool. Pro, doesn't expand the config, easily adds on support for the two additional options I think we should support. Con, might be situations with large workspace files that we really would want to mount if we're getting more into DL stuff?? If so, could likely workaround by making custom docker as likely large workspace files would be consistent between batches.
Possible that instead: 2) we separate the bool for mount of different buckets. Pro, more configurable. Con, would mean we would need to support mounting two buckets as I really don't want a situation where someone could set something in the config that we don't support
@bethac07 do you have an opinion?
I think your preferred option (second bucket is always download, current boolean is only for the source_bucket) makes a lot of sense to me!