Charlotte Hausman asked an interesting question: once we have adapted to the new product system, what parts of the current database schema will be going away? This page is meant to help us work that out, and one we have agreement, make a ticket to track it, so we can remove those bits once the new stuff is functional.
I think there are some guidelines we can use to come up with the list of what will go away:
- Access to any file goes through a filegroup
- Access to any filegroup is through the science products table or the ancillary products table
- Tying a science product to a project goes through the many to many relationship table
So here are some things I think will go away:
- execution_blocks.product_code
- execution_blocks.filegroup_id
- images.file_id
- calibrations.project_code
- calibrations.filegroup_id
- subscans.file_id
- scans.filegroup_id
- scans.filename
- image_products (these are now ancillary_products)
This isn't an exhaustive list, feel free to add to it.
Since we'll be linking from a science_product or ancillary_product to a filegroup and filegroups have files, I think we should decide on whether we need or want the recursive filegroup structure. I don't see a real need for it any more, this new stuff replaces what it did and does it better. The requirements for the product system don't definitively say one way or the other.