We need a way to be able to ask the database things like "give me all measurements of this object that are from DR1" or "that are from the current analysis version".
Provenances keep track of differnet code versions and parameter choices, but right now there's no way to decide what the "right" provenance is to use for any given purpose.
What I'm envisioning is something like a "tags" table. It would have (at least) two columns: a human readable tag name, and a provenance ID. One tag might be "current", another tag might be "DR1", etc. If you want to get all measurements from current, get all measurements who have a provenance that's in the tags table.
Perhaps three columns: tag, process, and provenance id. We could then enforce that the (tag, process) pair be unique together, so that we don't end up with redundant data in a single tag.
We need a way to be able to ask the database things like "give me all measurements of this object that are from DR1" or "that are from the current analysis version".
Provenances keep track of differnet code versions and parameter choices, but right now there's no way to decide what the "right" provenance is to use for any given purpose.
What I'm envisioning is something like a "tags" table. It would have (at least) two columns: a human readable tag name, and a provenance ID. One tag might be "current", another tag might be "DR1", etc. If you want to get all measurements from current, get all measurements who have a provenance that's in the tags table.
Perhaps three columns: tag, process, and provenance id. We could then enforce that the (tag, process) pair be unique together, so that we don't end up with redundant data in a single tag.