fix: restore utcnow() in test_detections.py

[Acruve15]

Summary

Seven test functions in src/tests/endpoints/test_detections.py still call datetime.utcnow(), but the datetime import was removed in #574 when those call sites were missed during the cleanup. This breaks pytest, pytest-client, and precommit-hooks on every PR against main.

This PR restores the calls by using the already-imported utcnow() helper from app.core.time (same fix that was applied elsewhere in #574).

Test plan

• pytest tests/endpoints/test_detections.py -v (67 / 67 pass)
• ruff 0.11.9 check clean

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.75%. Comparing base (f86f703) to head (e66ae87).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #578      +/-   ##
==========================================
- Coverage   90.40%   88.75%   -1.65%     
==========================================
  Files           3       52      +49     
  Lines         125     2196    +2071     
==========================================
+ Hits          113     1949    +1836     
- Misses         12      247     +235     

Flag	Coverage Δ
backend	88.65% <ø> (?)
client	90.40% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Acruve15]

@fe51 @MateoLostanlen Heads up on the codecov check: the failure here is a measurement artifact, not a real coverage drop.

What's happening?

Our tests.yml workflow uploads coverage from two separate jobs (pytest for backend, pytest-client for client). Since #574 merged, the pytest job has been failing on main because of leftover datetime.utcnow() calls in test_detections.py. That means every push to main since then only uploaded the client flag (3 files).

Codecov's flag carryforward kept the old backend numbers around for a while, but the cache eventually expired. Today main's stored coverage is just the 3 client files.

Why this PR looks like a regression

main (base):  3 files,   125 lines, 90.40%   (client only)
this PR:     52 files, 2225 lines, 88.75%   (full backend + client)

Codecov subtracts these two and reports -1.65%, but they're not measuring the same thing.

Path forward

I split the datetime.utcnow() cleanup into #578 so it can land on its own. Once that merges, main's pytest job will be green again and the next push to main will upload a complete backend report. Codecov's base will then match this PR's scope and the project check should pass automatically (no force-push needed on this PR).

#578 itself will also fail codecov/project for the same reason. It would need a one-time admin override to merge. After that the system is self-healing and we don't need a codecov.yml workaround.

[MateoLostanlen]

Good catch thanks