Rethink how env vars are mocked in tests

[rmanaem]

• How do we test an os.env read?
• How do we test that changing the env var actually results in the functions in the test having the new value?

For testing the time at which the env vars are read is at import time as opposed to at app start time/test runtime.

In the n-API (and soon f-API: neurobagel/federation-api#90) we load ENV variables in a utility module like so

api/app/api/utility.py

Lines 16 to 18 in 84c0e99

	ALLOWED_ORIGINS = EnvVar(
	"NB_API_ALLOWED_ORIGINS", os.environ.get("NB_API_ALLOWED_ORIGINS", "")
	)

Because the loading happens at the top level of the module, the ENV variable gets read in when the module is imported. In other words: we have no control over when the ENV variable is read in (e.g. during or after app startup).

This mainly matters during testing where we have to make sure we set ENV variables before the test run starts at all using something like https://pypi.org/project/pytest-env/ because by the time we have launched the TestApp client or imported the utility module any further changes to ENV variables will have no effect (because the reading has already happened).

This has two immediate consequences:

• it's tricky / hard to actually test the reading of ENV variables (e.g. testing for several ENV values)
• we need hard-to-fully-understand workarounds like monkeypatching the internal variable that the ENV value is read into

Possible solutions

• Using a pydantic model for environment variables / app settings: https://fastapi.tiangolo.com/advanced/settings/#pydantic-settings
  • This should (?) allow us to dynamically replace variable values
• Setting all variables before tests run using https://pypi.org/project/pytest-env/

[surchs]

Transferred this to n-API because we use env vars here more so far.

[github-actions]

We want to keep our issues up to date and active. This issue hasn't seen any activity in the last 75 days.
We have applied the _flag:stale label to indicate that this issue should be reviewed again.
When you review, please reread the spec and then apply one of these three options:

• prioritize: apply the flag:schedule label to suggest moving this issue into the backlog now
• close: if the issue is no longer relevant, explain why (give others a chance to reply) and then close.
• archive: sometimes an issue has important information or ideas but we won't work on it soon. In this case
  apply the someday label to show that this won't be prioritized. The stalebot will ignore issues with this
  label in the future. Use sparingly!

[alyssadai]

See also pytest-dev/pytest#10027 (comment)