feat(OMN-13366): add deterministic acceptance evidence

[jonahgabriel]

Summary

Implements OMN-13366 P1 deterministic acceptance for verifiable delegation classes.

• Declares score_source: deterministic_acceptance and deterministic acceptance_authority for code_generation, test, and validator_generation in task_class_contracts.v1.yaml.
• Extends quality-gate result evidence with score_source, acceptance_version, corpus_hash, validator_or_artifact_hash, acceptance_command, actual_score, pass, and failure_cases.
• Makes verifiable task classes with deterministic DoD use deterministic acceptance as the pass/fail authority; non-verifiable task classes keep existing heuristic authority.
• Extends node_generation_consumer corpus acceptance results with the same deterministic evidence bundle for validator-generation corpus gates.
• Adds focused OMN-13366 tests and updates corpus/DTO identity coverage.

Evidence-Ticket: OMN-13366
Evidence-Source: OCC#2831

Verification

Latest rebased head: 3b7316d53b5d260464802c0655c934be11892bd3 on origin/dev 2ce4b9672e30fa0647b87de5a042b31f2bf18a7c.

• uv run pytest tests/unit/delegation/test_quality_gate_deterministic_acceptance_omn13366.py tests/unit/nodes/node_generation_consumer/test_corpus_acceptance.py -q -> 17 passed.
• uv run pytest tests/unit/delegation tests/unit/nodes/node_generation_consumer -q -> 497 passed.
• uv run ruff check src/ tests/ && git diff --check -> passed.
• LC_ALL=C LANG=C uv run pytest -m "not integration and not kafka" -q -> 7554 passed, 35 skipped, 97 deselected.
• OMNIBASE_INFRA_PATH="$OMNI_HOME/omnibase_infra" LC_ALL=C LANG=C uv run pre-commit run --all-files -> passed after rerun; first run only normalized generated dep-health evidence, which was restored.

Earlier full uv run pytest -q on the pre-rebase head also ran: 7576 passed, 75 skipped, 4 failed. Failures were external/local-environment, not OMN-13366 paths:

• 3 Kafka tests could not use local Kafka/topic metadata on localhost:19092.
• 1 node_pr_polish unit assertion captured the local LC_ALL=C.UTF-8 warning from the fake uv subprocess; the broad unit/non-Kafka run passed with LC_ALL=C LANG=C.

Rollback

No rollback applied. Change is isolated to deterministic acceptance evidence/result semantics and task-class declarations.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds deterministic acceptance evidence to quality gate results: task class contracts gain score_source, acceptance_authority, and evidence_fields metadata; ModelQualityGateResult and ModelCorpusAcceptanceResult gain hash/command/score fields; handler_quality_gate.py computes and spreads evidence into all result paths; a compat shim redirects imports; and a new test suite validates the full deterministic/heuristic authority behavior.

Changes

Deterministic Acceptance Evidence

Layer / File(s)	Summary
Contract declarations and DTO evidence fields src/omnimarket/configs/task_class_contracts.v1.yaml, src/omnimarket/models/delegation/wire/model_quality_gate.py	YAML contracts add score_source, acceptance_authority, and evidence_fields to code_generation, test, and a new validator_generation task class. ModelQualityGateResult gains score_source, acceptance_version, hash/command fields, actual_score, pass_ (aliased to "pass"), and failure_cases, plus alias-aware model config.
Corpus acceptance evidence generation src/omnimarket/nodes/node_generation_consumer/corpus_acceptance.py	ModelCorpusAcceptanceResult gains the same evidence fields and alias config. New helpers compute stable SHA-256 corpus/handler hashes and the acceptance command string. A _result_with_evidence(...) helper fills all evidence fields and computes actual_score; both result-construction call sites delegate to it.
Quality gate handler: deterministic authority wiring src/omnimarket/nodes/node_delegation_quality_gate_reducer/handlers/handler_quality_gate.py	Adds module-level constants for acceptance version/source/verifiable task types and helper functions to compute hashes, derive deterministic scores, and decide authority applicability. delta() now computes acceptance_evidence before result construction and spreads it into all three result paths (deterministic-pass, heuristic-failure, final-pass). Legacy fail_category switches from enum members to string literals.
Compat import redirect src/omnimarket/nodes/node_delegation_quality_gate_reducer/models/model_quality_gate_result.py	Switches ModelQualityGateResult and EnumQualityGateCategory imports from omnibase_core.models.delegation.wire to omnimarket.models.delegation.wire.model_quality_gate, with an updated compatibility note.
Tests: compat assertions and deterministic acceptance behavior tests/test_delegation_dto_omnibase_compat_imports.py, tests/unit/delegation/test_quality_gate_deterministic_acceptance_omn13366.py, tests/unit/nodes/node_generation_consumer/test_corpus_acceptance.py	Compat tests add OMNIMARKET_OWNED_DTO_IMPORTS for ModelQualityGateResult and assert module ownership. New deterministic acceptance tests cover verifiable pass, fractional actual_score on failure, deterministic-over-heuristic precedence, heuristic fallback for non-verifiable contracts, and YAML authority field assertions. Corpus acceptance tests expand to assert evidence fields, aliased pass, and failure_cases.

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~60 minutes

Possibly related PRs

• OmniNode-ai/omnimarket#619: Modifies the same handler_quality_gate.py contract-driven deterministic/heuristic DoD handling and quality-gate evaluation logic that this PR extends with acceptance evidence generation.
• OmniNode-ai/omnimarket#675: Both PRs modify the same deterministic vs. heuristic check evaluation structure in handler_quality_gate.py, including float-equality handling in the deterministic check path.
• OmniNode-ai/omnimarket#1032: Directly overlaps with this PR's change of legacy fail_category assignment from EnumQualityGateCategory enum members to "pass"/"fail_heuristic" string literals in handler_quality_gate.py.

Poem

🐇 A bunny hops through hash and score,
Computing evidence, a deterministic floor.
SHA-256 for corpus, artifact in tow,
actual_score computed, pass_ set to glow!
No more heuristic guessing in the gate —
The rabbit stamps: deterministic. Accept the fate! 🎉

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 17.65% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'feat(OMN-13366): add deterministic acceptance evidence' directly and clearly summarizes the main change: adding deterministic acceptance evidence, matching the PR objectives and the substantial modifications across config, models, handlers, and tests.
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.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ 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 jonahgabriel/omn-13366-deterministic-acceptance

---

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

[github-actions]

✅ Architectural Review — PASSED

Errors: 0
Warnings: 0

What this checks

Rule	Description
ARCH-TOPIC-001	No hardcoded Kafka topic strings in handler code
ARCH-DI-001	No event_bus=None bypass
ARCH-DI-002	No direct Handler instantiation outside workflow_runner/adapters
ARCH-DI-003	No reinvented DI containers
ARCH-TOPIC-002	contract.yaml topics follow `onex.{cmd

---

No architectural violations found.

Powered by node_pr_review_bot

[github-actions]

✅ PR Review Bot — CLEAN

Findings: 0
Run ID: ``

Note: Reviewer models are running as stubs (OMN-7969–OMN-7972 pending). Architectural findings from Phase 1 (arch-lint) are the authoritative signal until model handlers land.

---

Verdict semantics

Verdict	Meaning
clean	No blocking findings
risks_noted	Findings noted; threads passed verification
blocking_issue	MAJOR/CRITICAL findings not resolved — do not merge

Powered by node_pr_review_bot — FSM pipeline with multi-model review fan-out

[github-actions]

⚠️ Hostile Reviewer — DEGRADED (informational)

Blocking findings (critical/error): 0
Total findings: 0
Models succeeded: none

Note: All reviewer models failed or were unavailable. Degraded results are informational during the pilot phase (OMN-8524) and do not block merge. Error: cli_review exited non-zero (all models failed or binary error)

---

Gate semantics (pilot phase)

Verdict	Meaning	Blocks merge?
passed	No critical/error findings	No
blocked	CRITICAL or ERROR findings found	Yes
degraded	All models unavailable (infra)	No (pilot)

Powered by omniintelligence.review_pairing.cli_review — node-based adversarial review via HandlerLlmCliSubprocess (OMN-8524)