Drop stale _runtime.tmp.linked_instances.* collections

[dwinston]

On this branch, I add a fastapi.BackgroundTasks parameter to the path operation function for the GET /nmdcschema/linked_instances endpoint, and schedule a task -- to be performed after the endpoint returns -- that drops any temporary linked-instances collections that were generated earlier than one day prior to the task being run.

Related issue(s)

Fixes #1240

Related subsystem(s)

• Runtime API (except the Minter)

Testing

• I tested these changes (explain below)

I tested these changes by inspection. Calling the endpoint locally dropped local stale temporary collections.

Documentation

• Other (explain below)

N/A

Maintainability

• Every Python function I defined includes a docstring (test functions are exempt from this)
• Every Python function parameter I introduced includes a type hint (e.g. study_id: str)
• All "to do" or "fix me" Python comments I added begin with either # TODO or # FIXME
• I used black to format all the Python files I created/modified
• The PR title is in the imperative mood (e.g. "Do X") and not the declarative mood (e.g. "Does X" or "Did X")

[eecavanna]

This looks good to me, and I understand how it works based on our discussion earlier "today" (thanks for introducing me to this FastAPI feature).

I'm OK with this being merged in as is.

If time permits (note: I can add this via a future PR if it is not added to this one), I'd prefer there be a logging.info(f"Dropping temporary collection: {collection_name}") statement in the cleanup function, since AFAIK it's the first use of the FastAPI BackgroundTask in the codebase and I think having a logging statement there will help us (maintainers of the deployments) internalize its behavior.

[eecavanna]

Edit: I approve of this being merged in once the tests pass (I just noticed some failed). No need for subsequent review by me after updating tests.

[eecavanna]

Thanks for implementing this "clean-up" functionality. This task remaining unresolved beyond this quarter, is not something I'm worried about. The changes made sense to me when I reviewed the branch at commit a8d1945. I'd be OK taking over trying to get the tests to pass (starting next week).