fix(polymarket): admit trailing completion proof inputs

[proerror77]

Change contract

For explicit-market Polymarket evidence only, admit authenticated trailing reference segments that contain redundant metadata plus exact completion proofs but no trade or settlement history, while keeping every segment ordered/non-overlapping and every data-bearing segment consecutive by UTC hour.

Out of scope

Collector or deployment changes; evidence byte or digest formulas; snapshot sampling or factor logic; Binance inputs; evaluator/MCTS; Paper or Live.

Dependency or merge order

Base: main@23af4258bf452a94c8c630ad1f8d8d5479f6b954. Depends on the event-local sparse-reference producer contract merged in #278 and verifier fixes through #294. The unrelated frontend audit blocker was split into #299/#300 and is merged. Merge this PR before retrying #232 with an exact-main research image.

Focused validation

• Red proof before implementation: the explicit 17:00 data plus 20:00 completion-proof shape failed with validated reference inputs must be consecutive UTC hours.
• Focused positive/negative tests: explicit 17 -> 20 accepted; the same non-explicit shape rejected; early, reversed, ordered-interleaved, malformed, and data-gap inputs rejected.
• cargo test -p ploy-market-data - 60 passed.
• cargo test -p ploy-research --test mission_search_policies -- polymarket_missions_pin_current_search_policy_snapshot - passed for BTC 5m and SOL 5m governed missions.
• cargo fmt -p ploy-market-data -- --check - passed.
• cargo clippy -p ploy-market-data --all-targets --no-deps -- -D warnings - passed.
• git diff --check - passed.
• Matt two-axis read-only code review - Standards: APPROVE; Spec: APPROVE, including the required policy-pin update.

Rollout or rollback impact

Rollout publishes a new immutable exact-main research image, then reruns the unchanged #232 evidence in ACK. Rollback is a one-commit revert; no existing artifact or collector state is mutated.

Root cause

PR #278 intentionally allowed explicit event-local evidence normalization to consume nondecreasing sparse reference hours, but the downstream typed verifier retained PR #154's earlier consecutive-hour rule. This producer/verifier contract drift rejected a valid later completion-proof segment.

Why no PRD

#297 defines one already-approved verifier predicate and its counterexamples, so Matt TDD was the narrowest workflow.

Tracks #297. Cloud acceptance remains post-merge and will be recorded on the issue before it is closed.

[coderabbitai]

📝 Walkthrough

Walkthrough

Validation now uses explicit market selection to gate quote-collection failures and completion-only references. Reference ordering, UTC-hour constraints, event composition, test helpers, and completion-only scenarios were updated.

Changes

Explicit reference validation

Layer / File(s)	Summary
Explicit-selection event gating rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs	validate_inputs uses explicit selection to determine whether quote_collection_failure is an allowed event type.
Completion-only reference rules and coverage rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs	Reference validation adds completion-only detection, UTC-hour ordering and sequencing checks, explicit manifest rewriting in test utilities, and coverage for valid, sparse, early, interleaved, and malformed references.

Estimated code review effort: 4 (Complex) | ~45 minutes

Possibly related issues

• Allow nonconsecutive completion-only Polymarket evidence inputs #297 — Directly covers completion-only reference validation for explicit manifests.
• Prediction evidence: admit only bounded recovered quote gaps per explicit market #283 — Addresses explicit-market validation for constrained quote-collection failures.
• Polymarket sealer: admit explicit evidence inputs that disclose quote collection failures #284 — Covers allowing quote_collection_failure for explicitly selected manifests.

Possibly related PRs

• proerror77/monday#154 — Extends validation of ordered and consecutive UTC-hour reference segments.
• proerror77/monday#183 — Binds evidence validation and manifest handling to explicit market selection.
• proerror77/monday#211 — Modifies reference UTC-hour sequencing and trade-completion timestamp validation.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title accurately summarizes the main change: allowing trailing completion-proof Polymarket inputs.
Description check	✅ Passed	The description covers the required contract, out-of-scope, dependencies, validation, and rollout details.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/polymarket-completion-input-hours-297

---

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

• X
• Mastodon
• Reddit
• LinkedIn

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

[coderabbitai]

🧹 Nitpick comments (2)

rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs (2)

695-753: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick win

Completion-only sequencing logic verified correct.

Traced this against all new/existing tests (later-only OK, sparse/non-explicit rejected, early/interleaved rejected, reversed rejected, same-hour overlap rejected): the all-references ordering check (701-711) plus the data_references-only consecutive-hour check (735-743) plus the has_prior_metadata/data_references.is_empty() guard (716-719) and the "must follow all evidence data" guard (726-733) together correctly implement "ordered & non-overlapping for all segments, consecutive UTC hours for data-bearing segments only, trailing-only completion proofs." No correctness issue found.

One coverage gap: no test exercises a chain of multiple consecutive completion-only references after the data segments (the logic appears to permit this, consistent with "trailing reference segments" being plural in the PR summary), so a regression here wouldn't be caught.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In
`@rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs`
around lines 695 - 753, ​​Add a regression test covering multiple consecutive
completion-only references after the data-bearing segments in the validation
flow around the ordering and sequencing checks. Ensure the test confirms the
trailing completion-only chain is accepted while preserving the existing
requirements that completion-only references follow evidence data and remain
trailing.

---

1419-1549: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick win

Good coverage of new sequencing rules; consider adding a chained completion-only test.

These four tests correctly pin down: later completion-only accepted, sparse completion-only rejected when the manifest isn't explicit, early/interleaved completion-only rejected with distinct messages, and malformed completion counts still rejected. Consider adding a case with two (or more) consecutive completion-only references trailing the data segments, since the implementation in validate_inputs (695-753) appears to permit that but it's currently untested.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In
`@rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs`
around lines 1419 - 1549, Add a test alongside the existing completion-only
sequencing tests that appends at least two consecutive valid completion-only
references after the final evidence-data reference. Use the existing input
builders and sealing flow, preserve chronological ordering and valid completion
metadata, and assert that validate_inputs via seal_polymarket_evidence_triplet
succeeds.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In
`@rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs`:
- Around line 695-753: ​​Add a regression test covering multiple consecutive
completion-only references after the data-bearing segments in the validation
flow around the ordering and sequencing checks. Ensure the test confirms the
trailing completion-only chain is accepted while preserving the existing
requirements that completion-only references follow evidence data and remain
trailing.
- Around line 1419-1549: Add a test alongside the existing completion-only
sequencing tests that appends at least two consecutive valid completion-only
references after the final evidence-data reference. Use the existing input
builders and sealing flow, preserve chronological ordering and valid completion
metadata, and assert that validate_inputs via seal_polymarket_evidence_triplet
succeeds.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 72137ead-9e44-413e-8319-116916d64605

📥 Commits

Reviewing files that changed from the base of the PR and between 000efe1 and d965da9.

📒 Files selected for processing (1)

• rust_hft/prediction-markets/crates/ploy-market-data/src/polymarket_evidence/artifact.rs