Make a compound unique key constraint in mongo for biosamples

[aclum]

Is your feature request related to a problem? Please describe.
on biosample set a compound key of name + associated-studies should be unique

Screenshots
If applicable, add screenshots to help explain your problem.

Describe the solution you'd like
This is to prevent duplicate sample names with a given study, see microbiomedata/nmdc-schema#2878

Example user story template:
AS A {user|persona|system},
[INSTEAD OF {current condition}]
I WANT TO {action} [IN {mode} TIME | IN {differentiating performance units} TO {utility performance units}]
[SO THAT {value or justification}]
[NO LATER THAN {best by date}]

Acceptance Criteria

• Submissions to nmdc-runtime post endpoints that do not have a unique value for name + associated studies for biosample_set records return errors.
• this should be true for new submissions (via json:submit) and through endpoints which modify records like queries:run and changesheets.

Example scenario-based template:
Given (some given context or precondition), when (I take this action), then (this will be the specific outcome).

Describe alternatives you've considered
Not enforcing this may lead to duplicate sample names in mongo which then requires subsequent cleanup.

Additional context

• dependent on
  • 2901 Make slot name on class Biosample required nmdc-schema#2901

[eecavanna]

FYI: The migrator in nmdc-schema/pull/2901 just checks whether all biosamples have a name.

        if "name" not in biosample or not isinstance(biosample["name"], str):
            raise ValueError(f"Biosample {biosample['id']} lacks a string name")

[eecavanna]

I can create a migrator that checks whether any biosamples associated with the same study have the same name as one another. It does not exist yet, though.

[eecavanna]

The unique compound index is implemented in PR #1430 (ready for review).

I have a question: Is there something in nmdc-schema that says this combination must be unique (i.e. that no two biosamples associated with the same study can have the same name)? I checked the slot_usage for the name slot within the definition of the Biosample class (here) and didn't see anything about its uniqueness, other than a suggestion and a "should" statement in reference to NMDC:

we suggest that you make it [...] unique

For NMDC this should be unique within a study.

Given those alone, I consider this to be a suggestion and not a strict requirement. But the Runtime would be enforcing it as a strict requirement (no exceptions to the rule).