Skip to content

feat(event): score TDT tracks with pair precision and switch rate - #68

Draft
seonghobae wants to merge 1 commit into
mainfrom
agent/event-tracking-stability
Draft

feat(event): score TDT tracks with pair precision and switch rate#68
seonghobae wants to merge 1 commit into
mainfrom
agent/event-tracking-stability

Conversation

@seonghobae

Copy link
Copy Markdown
Contributor

Summary

#38 merged to protected main (4d287a2). Preferred merge #45 was draft+CONFLICTING after that merge. This cycle rebased #45 onto origin/main (kept mention SQL from #38 and 0007 retention), force-pushed ed1349f, and marked it ready. #45 is MERGEABLE BLOCKED: Checks queued, CodeRabbit COMMENTED only, no qualifying independent APPROVE. Do not self-approve or force-merge. Do not leapfrog #45 with #39.

Next no-0008 gap: TDT tracking pair precision/recall and identity-switch rate in existing event_core (does not recreate TDT/CHRONOS or model_selection crates, does not wait on #48/#49/#53/#65/#66/#67, and does not allocate 0008).

  • refuse_track_as_instance / refuse_track_as_transition never promote a hypothesized track
  • tracking_pair_precision and tracking_pair_recall fail closed on empty, mismatched, duplicate-mention, or pairless assignment streams
  • tracking_identity_switch_rate is computed only over consecutive mentions that share a truth track
  • Computed RMSE of known same-track pair targets is lower under calibrated scores than under an always-one-track detector

Test plan

  • RED: pair-precision fixture initially failed (0.5 vs 0.5 always-one-track); fixture corrected so calibrated precision exceeds the naive merge
  • cargo test -p event_core --offline
  • cargo clippy -p event_core --all-targets --offline -- -D warnings
  • cargo test --workspace --offline
  • python3 scripts/validate_documentation.py and check_docstrings.py
  • event_core line 144/144 and branch 40/40
  • exact-head required Checks (do not treat queued evidence as passed)

Do not self-approve or merge. Prefer merge #45 when exact-head required checks and a qualifying independent APPROVE exist. Do not allocate 0008.

Keep hypothesized track assignments in existing event_core as
measurement evidence. Refuse track-as-instance and track-as-transition
promotion, compute pair precision/recall and identity-switch rate from
known truth, and recover same-track targets with lower RMSE than an
always-one-track detector.
@coderabbitai

coderabbitai Bot commented Aug 13, 2026

Copy link
Copy Markdown

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 63bb261c-c696-44de-9a0e-da3fc2bc1dc5

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant