Repository navigation
Description
Task
T-074b· Milestone M6 — Analysis + risk · Size S · Priority P1 · Epic #6Implements FR-ANA-005
Branch
feat/t-074b-interaction-confirmation-orchestration-enqueue-runs— one task = one branch = one PR (doc 32 §6)
Summary
Application service runner/confirmations.py: on request (API :confirm-interaction) or automatically for top modes, enqueue planned scenarios as runs (phase=confirm, tag confirm, priority 10), wait for completion, call conclude, persist failure_modes.interaction.
Context / motivation
Confirmatory runs must be excluded from strategy comparison metrics (doc 09 §6) and must not disturb the strategy's history.
Scope
In scope
- Confirm runs excluded from
RunRepo.historyused by strategies. - Event
mode.discovered/campaign.updatedon update. - Optional auto-confirm of the top-N modes after final analysis (
analysis.interaction.auto_confirm_top_n, new, default 0 = off).
Implementation guidance
- Application service
roboprobe/runner/confirmations.py:class ConfirmationService: def request(self, campaign_id: UUID, mode_id: UUID) -> list[UUID](enqueue, returns run ids) anddef on_run_finished(self, run: RunRecord) -> None(called from the orchestrator tick; concludes when all confirm runs for a mode are terminal). - Enqueue via
RunQueue.enqueue(..., phase=RunPhase.CONFIRM, priority=settings.runner.confirm_priority)(new key, default 10 — same mechanism as manual runs). - Exclusion:
RunRepo.history(campaign_id, exclude_phases={CONFIRM, MANUAL})is what strategies and metrics consume — add the parameter if T-011a lacks it (coordinate via its owner, append-only change). - Persist
failure_modes.interactionJSON (verdict) and emitmode.discovered-stylecampaign.updatedevent through the outbox (T-011c). - Idempotency: a second
requestwhile confirmations are pending returns the existing run ids. - Budget: confirm runs do not count against the campaign budget (doc 09 §6) but do count in worker utilisation metrics.
- Standards: respect the layering contracts (doc 16 §2) and conventions in doc 34; no hard-coded tunables — thresholds/defaults come from typed settings (doc 27); no magic numbers (ruff
PLR2004); public APIs typed (mypy strict) and documented.
Configuration keys
From the configuration reference (doc 27). Keys marked (new) must be added to the typed settings models and appear in the generated reference (T-003c).
- runner.confirm_priority (new)
- analysis.interaction.auto_confirm_top_n (new)
Acceptance criteria
- Integration test on k2d: GT-A canonical mode confirmed as interaction ×3.
- Confirm runs excluded from strategy history and MET metrics (test with S-RANDOM campaign).
- Duplicate request is idempotent.
- Mode row updated and event emitted on conclusion.
- Tests listed below added and green in CI; coverage of new code meets the gate in doc 18 §3.
- Docs, config reference and CHANGELOG
Unreleasedupdated where behaviour, keys or interfaces changed. - Definition of Done (doc 32 §3) satisfied.
Tests required
tests/integration/analysis/test_confirmations.py.- Integration with Testcontainers Postgres + inline worker.
Dependencies
Blocked by: #133 (T-074a), #85 (T-041a)
Blocks: no other task (leaf)
Parallelization notes
Files / paths owned by this task — avoid editing other paths; shared files (pyproject.toml entry points, cli/main.py, api/routers/__init__.py, router.tsx, docker/compose.yaml, ci.yml) are append-only:
backend/src/roboprobe/runner/confirmations.py
Independent module.
Definition of Done
Per doc 32 §3 and AGENTS.md §5: lint, format, mypy strict and import-linter green; tests added and passing; docs updated; PR template completed with Closes #<this issue>; Conventional Commit messages referencing T-074b (doc 16 §6); commits authored only by the repository owner identity (AGENTS.md §4); CI green; reviewer checklist (doc 32 §5.1) satisfied.
References
Design docs: doc 09 §6 · doc 12 §2 · doc 11 §3 · doc 05 §9
Requirements: FR-ANA-005 (traceability: doc 28)
Metadata
Metadata
Assignees
Labels
Projects
- StatusShow more project fieldsBacklog
Activity