test: functional test config and soak test

[corpo-iwillspeak]

Summary

• Removes hardcoded defaults for ATHENA_HOST, OAUTH_AUTH_URL, and OAUTH_AUDIENCE in functional test config — these now fall back to SDK defaults when unset, rather than overriding with stale hard-coded values
• Adds proactive_refresh_threshold=0.25 to CredentialHelper construction so tokens refresh before expiry during long test runs
• Adds a new test_classify_single_soak.py that sends 1000 classify_single requests and asserts a ≥95% KnownCSAM- hash-check classification success rate

Test plan

• Run functional tests against a live environment to confirm existing behaviour is unchanged
• Run the new soak test (pytest -m soak) to validate the success-rate assertion

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR updates the functional-test configuration to rely on SDK defaults when certain env vars are unset, and adds a new long-running soak test for classify_single to track KnownCSAM-* hash-check classification presence over many requests.

Changes:

• Functional test config: stop hardcoding defaults for ATHENA_HOST, OAUTH_AUTH_URL, and OAUTH_AUDIENCE and let SDK defaults apply when unset.
• Functional test auth: adjust CredentialHelper construction (incl. proactive refresh threshold setting).
• Add a new soak test that runs 1000 classify_single calls and asserts a minimum hash-check classification success rate.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
tests/functional/test_classify_single_soak.py	Adds a new soak test for repeated classify_single calls and success-rate assertion for KnownCSAM-* classifications.
tests/functional/conftest.py	Updates functional-test env var handling to avoid overriding SDK defaults; adjusts how CredentialHelper and AthenaOptions are constructed.

Comments suppressed due to low confidence (1)

tests/functional/test_classify_single_soak.py:67

• The assertion message hard-codes "below 95%" even though the threshold is defined by MIN_HASH_CHECK_SUCCESS_RATE. If the constant changes, the message becomes misleading.

        assert success_rate >= MIN_HASH_CHECK_SUCCESS_RATE, (
            f"Hash check success rate {success_rate:.1%} is below 95%. "
            f"{len(failures)} failures:\n" + "\n".join(failures[:20])
        )

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.