feat(event): score TDT mention links with precision and recall - #66
Conversation
Keep detected same-event pairs distinct from promoted instances and state transitions, and require computed precision, recall, and RMSE against known-truth pair sets.
|
Warning Review limit reachedNext included review available in 42 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (11)
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. Comment |
# Conflicts: # CHANGELOG.md # DOCUMENTATION.md # docs/TRACEABILITY.md
|
Reviewed and integrated current protected main without force-push. Current head: Verification at this exact head:
The link contract remains fail-closed: self-links and empty metric denominators reject, mention links cannot become instances or state transitions, and precision/recall are computed from normalized known-truth pairs. Please run protected current-head Checks and obtain the required independent approval before normal merge. |
|
Current-head review refresh for 633d41c:
|
|
Exact-head review receipt: current-main conflict repair was completed and pushed at |
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
Union of CHANGELOG/DOCUMENTATION rows; event_core error-variant and module/export union keeping link, intelligence, and first-story surfaces; ADR 0016 maturity consolidated to one statement; ADR index deduped to one 0016 row with combined wording; APA register keeps Allan et al. (1998) before Allen (1983); ledger keeps link precision/recall row and main's nested-ICC membership row.
| fn counted_rate(numerator: usize, denominator: usize) -> Result<f64, EventError> { | ||
| let numerator = u32::try_from(numerator).map_err(|_| EventError::InvalidWirePayload)?; | ||
| let denominator = u32::try_from(denominator).map_err(|_| EventError::InvalidWirePayload)?; | ||
| if denominator == 0 { | ||
| return Err(EventError::InvalidWirePayload); | ||
| } | ||
| Ok(f64::from(numerator) / f64::from(denominator)) | ||
| } |
There was a problem hiding this comment.
📝 Info: Empty-set precision/recall fails closed by design
event_link_precision errors on empty recovered set and event_link_recall errors on empty truth set, via the denominator==0 guard in counted_rate. This is a deliberate fail-closed choice matching the doctoring note and tests, not a defect.
Was this helpful? React with 👍 or 👎 to provide feedback.
| | evidence-bounded LLM interpretation | ADR 0010/0012; PRD | `tepp_api` router plus future `interpretation_gateway` | partial | | ||
| | evidence-bounded LLM interpretation | ADR 0010/0012; PRD | `tepp_api` router implemented-main plus future `interpretation_gateway` | partial | |
There was a problem hiding this comment.
📝 Info: Duplicate traceability rows for LLM interpretation
The change adds two near-identical 'evidence-bounded LLM interpretation' rows in docs/TRACEABILITY.md, differing only in wording ('router plus future' vs 'router implemented-main plus future'). The duplication looks unintentional and mirrors similar duplicate rows already present in the doc.
Was this helpful? React with 👍 or 👎 to provide feedback.
# Conflicts: # DOCUMENTATION.md # crates/event_core/src/error.rs # docs/TRACEABILITY.md # docs/adr/0016-tdt-chronos-event-intelligence-boundary.md # docs/research/standards-and-literature.md
# Conflicts: # CHANGELOG.md # DOCUMENTATION.md # crates/event_core/src/error.rs # docs/TRACEABILITY.md # docs/adr/0016-tdt-chronos-event-intelligence-boundary.md # docs/adr/README.md # docs/validation/temporal-event-foundation.md
There was a problem hiding this comment.
🔍 Pre-existing fused error-message tuple
The UnknownFirstStoryLabel entry is already fused with EventTrackIsNotEventInstance into a four-element tuple in the merge-base, the same defect as the newly introduced one. The test array was already structurally broken; fixing it in the same pass would restore both message assertions.
(Refers to this code)
Was this helpful? React with 👍 or 👎 to provide feedback.
| ( | ||
| EventError::UnknownEventLinkLabel, | ||
| "unknown event link label", |
There was a problem hiding this comment.
🔴 Link-label error test fused with first-story case
The added UnknownEventLinkLabel test entry never closes its tuple, so it absorbs the following FirstStoryIsNotEventInstance entry into one four-element tuple. The array now mixes tuple shapes and the message checks for both error variants no longer run.
| ( | |
| EventError::UnknownEventLinkLabel, | |
| "unknown event link label", | |
| ( | |
| EventError::UnknownEventLinkLabel, | |
| "unknown event link label", | |
| ), | |
| ( |
Was this helpful? React with 👍 or 👎 to provide feedback.
1 participant
Summary
#45 remains preferred merge (0007) but is only locally blocked: hourly re-drafted it; this cycle marked it ready again. Exact-head required checks were cancelled by the draft and are re-queuing. No qualifying independent APPROVE (CodeRabbit COMMENTED only). #38 must not leapfrog #45.
Next no-0008 gap: TDT link-detection precision/recall in existing
event_core(does not recreate TDT/CHRONOS crates, does not wait on #48/#49/#53/#65, and does not allocate 0008).refuse_event_link_as_instance/refuse_event_link_as_transitionnever promote a detected pairEventLinkPairrefuses self-links and normalizes pair orderevent_link_precisionandevent_link_recallfail closed on empty setsTest plan
link_detection_contractfailed withE0432/E0599cargo test -p event_core --offlinecargo clippy -p event_core --all-targets --offline -- -D warningscargo test --workspace --offlinepython3 scripts/validate_documentation.pyandcheck_docstrings.pyDo not self-approve or merge. Prefer merge #45 when exact-head required checks and a qualifying independent APPROVE exist. Do not allocate 0008.