Core functionality of /data_objects/study/{study_id} endpoint is not being tested

[eecavanna]

The only test that targets the /data_objects/study/{study_id} endpoint, targets a corner case where the specified study does not exist.

nmdc-runtime/tests/test_api/test_endpoints.py

Lines 349 to 364 in bcdf8f3

	def test_find_data_objects_for_nonexistent_study(api_site_client):
	r"""
	Confirms the endpoint returns an unsuccessful status code when no `Study` having the specified `id` exists.
	Reference: https://docs.pytest.org/en/stable/reference/reference.html#pytest.raises
	Note: The `api_site_client` fixture's `request` method will raise an exception if the server responds with
	an unsuccessful status code.
	TODO: Add tests focused on the situation where the `Study` _does_ exist.
	"""
	ensure_schema_collections_and_alldocs()
	with pytest.raises(requests.exceptions.HTTPError):
	api_site_client.request(
	"GET",
	"/data_objects/study/nmdc:sty-11-hdd4bf83",
	)

There are no tests that target the core functionality (a.k.a. "happy path") of the endpoint.

nmdc-runtime/tests/test_api/test_endpoints.py

Lines 36 to 38 in bcdf8f3

	# FIXME: Seed the database with documents that would be included in an `alldocs` collection,
	# such that the `/data_objects/study/{study_id}` endpoint (which uses that collection)
	# would return some data. Currently, we are practically _not testing_ that endpoint.

Previously, there was a test that targeted the core functionality; but it was removed because it relied on a static database dump (which conformed to a specific schema) residing on the NERSC filesystem.

Related (bug report involving this endpoint): #723

[aclum]

I am advocating for this be escalated, this is the second time we've caught problem in production with this endpoint. cc @shreddd

[eecavanna]

I'll create a branch that contains the skeleton of a test, and then work with @sujaypatil96 to flesh it out.

[eecavanna]

I created a branch (in a draft PR) in which I implemented the following two new tests:

1. test_find_data_objects_for_study_having_none: the Study exists, but doesn't have anything associated with it (e.g. no Biosamples, no DataGenerations, no DataObjects). In this case, the endpoint returns no DataObjects.
2. test_find_data_objects_for_study_having_one: the Study exists, a Biosample is associated with that Study, a MassSpectrometry (which inherits from DataObject) has that Biosample as an input, and a DataObject is an output of that MassSpectrometry. In this case, the endpoint returns one DataObject.

Here's a link to the PR: #849

I will invite others to contribute to it there.