feat(ci): dynamic test duration tracking for pytest-split

[vringar]

Summary

• Add pytest hooks in test/conftest.py to record per-test wall-clock durations, gated on GROUP env var (CI only)
• Update scripts/ci.sh to prefer cached durations over the committed .test_durations baseline
• Add cache restore/save steps and a consolidate-durations job to .github/workflows/run-tests.yaml that merges per-group artifacts into a single cached file

How it works

Per-run data flow:

Restore cache → ci.sh picks cached or committed file → pytest-split distributes tests
    → conftest hook records durations → upload per-group artifacts
    → consolidation job merges all groups → save to cache

Fallback chain (priority order):

1. Cached durations from the most recent run on the same branch
2. Cached durations from the most recent run on master
3. Committed .test_durations file (unchanged, serves as baseline)

Test plan

• Run pytest test/test_dataclass_validations.py -v with GROUP=1 — verified .test_durations_group_1 is created with correct node IDs matching .test_durations format
• Verify CI workflow: cache restore (miss on first run), artifact upload, consolidation, cache save
• Push again to verify cache hit and that splitting uses cached durations

[Copilot]

Pull request overview

This PR adds CI-driven feedback loops for pytest-split by collecting per-test durations during CI runs, consolidating them across split groups, and caching the merged result so future CI runs can split tests using the most recent timings.

Changes:

• Add pytest session hooks to record durations into per-group .test_durations_group_* files (CI-only via GROUP).
• Update scripts/ci.sh to prefer a cached durations file (.test_durations_cached) over the committed baseline (.test_durations).
• Extend the GitHub Actions workflow to restore/save the cache and add a consolidation job that merges per-group artifacts into the cached durations file.

Reviewed changes

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

File	Description
test/conftest.py	Adds pytest hooks to collect and write per-test duration JSON for the current CI split group.
scripts/ci.sh	Uses a cached durations file when available and passes it to pytest-split via --durations-path.
.github/workflows/run-tests.yaml	Restores durations cache for test jobs, uploads per-group duration artifacts, and consolidates/saves merged durations back to cache.

---

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

[Copilot]

pytest_runtest_logreport records durations for every local pytest run, even though the PR description says this should be gated on GROUP (CI-only). Consider checking os.environ.get("GROUP") (or a config flag) before updating _test_durations so local runs don't incur the hook overhead or generate unexpected in-memory state.

[Copilot]

The durations you persist are only for the call phase (report.when == "call"), but the PR description calls these "per-test wall-clock durations". For tests with expensive fixture setup/teardown (or skips/errors during setup), this will significantly undercount and can lead to poorer splitting; consider accumulating setup + call + teardown durations per nodeid (or otherwise clarifying/aligning what "wall-clock" means here).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 56.75%. Comparing base (12fae22) to head (4a6ef17).
⚠️ Report is 5 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1151   +/-   ##
=======================================
  Coverage   56.75%   56.75%           
=======================================
  Files          40       40           
  Lines        3908     3908           
=======================================
  Hits         2218     2218           
  Misses       1690     1690           

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.