kernel: deterministic static_check stage (§B-§5) + critique-outcome producer (§B-§6)
Two reviewer-reliability tracks, both deterministic and unit-verified (no model/network). §B-§5 — static_check stage seam. The StaticAnalysisRunner + StaticFindingsRecordedEvent + reviewer-context filter already existed; what was missing was a stage that runs the tool and emits the event. Added: - StaticCheckStageExecutor: reads stage metadata (static_tool/static_argv), runs the configured command via StaticAnalysisRunner, returns a StaticFindingsRecordedEvent. No-op (empty findings) when no runner is wired or no command is configured — safe to carry unconfigured. - A deterministic-stage seam in DefaultSessionOrchestrator.enterStage: any stage with metadata["stage_type"] == "static_check" is run by the executor instead of the LLM subagent, then advances on its (unconditional) exit edge. - TomlWorkflowLoader: stage_type/static_tool/static_argv fields → StageConfig.metadata. - role_pipeline.toml: a static_check stage between implementer and reviewer (no-op until static_argv + a CommandRunner are set; activation is live-QA-gated, see StaticAnalysisRunner doc). §B-§6 — critique-outcome producer. The CritiqueFinding type + CriticCalibrationProjection existed but nothing fed them. Added: - CritiqueFindingsRecordedEvent (+ CritiqueVerdict): the producing side — a critic's findings + verdict for one review iteration, carrying modelHash for per-model calibration. - CritiqueOutcomeCorrelator: pure loop-resolution logic deciding UPHELD (fixed between rounds) / DISMISSED (persisted into an approved final) / INCONCLUSIVE (open at a non-approved terminal), per critic (role + modelHash) and per finding id. - A hook in completeWorkflow/failWorkflow that correlates recorded findings into CritiqueOutcomeCorrelatedEvents at loop resolution — no-op when none recorded, idempotent. (LLM-side finding emission stays a separate model-gated activation.) Tests: StaticCheckStageExecutorTest (4), StaticCheckStageTest integration (1, fake runner → event + transition), CritiqueOutcomeCorrelatorTest (8), CritiqueCalibrationWiringTest integration (1, seeded findings → outcomes at completion), updated RolePipelineWorkflowTest. Full suites for core:events/kernel/critique, infrastructure:workflow, testing:integration green; detekt clean. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
@@ -0,0 +1,161 @@
|
||||
import com.correx.core.approvals.ApprovalProjector
|
||||
import com.correx.core.approvals.DefaultApprovalReducer
|
||||
import com.correx.core.approvals.DefaultApprovalRepository
|
||||
import com.correx.core.approvals.domain.DefaultApprovalEngine
|
||||
import com.correx.core.artifacts.DefaultArtifactReducer
|
||||
import com.correx.core.events.events.CritiqueFinding
|
||||
import com.correx.core.events.events.CritiqueFindingsRecordedEvent
|
||||
import com.correx.core.events.events.CritiqueOutcome
|
||||
import com.correx.core.events.events.CritiqueOutcomeCorrelatedEvent
|
||||
import com.correx.core.events.events.CritiqueRole
|
||||
import com.correx.core.events.events.CritiqueSeverity
|
||||
import com.correx.core.events.events.CritiqueVerdict
|
||||
import com.correx.core.events.events.EventMetadata
|
||||
import com.correx.core.events.events.NewEvent
|
||||
import com.correx.core.events.types.EventId
|
||||
import com.correx.core.events.execution.RetryPolicy
|
||||
import com.correx.core.events.types.SessionId
|
||||
import com.correx.core.events.types.StageId
|
||||
import com.correx.core.events.types.TransitionId
|
||||
import com.correx.core.inference.InferenceRepository
|
||||
import com.correx.core.inference.InferenceState
|
||||
import com.correx.core.journal.DecisionJournalProjector
|
||||
import com.correx.core.journal.DefaultDecisionJournalReducer
|
||||
import com.correx.core.journal.DefaultDecisionJournalRepository
|
||||
import com.correx.core.kernel.orchestration.DefaultOrchestrationReducer
|
||||
import com.correx.core.kernel.orchestration.DefaultSessionOrchestrator
|
||||
import com.correx.core.kernel.orchestration.OrchestrationConfig
|
||||
import com.correx.core.kernel.orchestration.OrchestrationProjector
|
||||
import com.correx.core.kernel.orchestration.OrchestrationRepository
|
||||
import com.correx.core.kernel.orchestration.OrchestratorEngines
|
||||
import com.correx.core.kernel.orchestration.OrchestratorRepositories
|
||||
import com.correx.core.kernel.retry.DefaultRetryCoordinator
|
||||
import com.correx.core.risk.DefaultRiskAssessor
|
||||
import com.correx.core.sessions.DefaultSessionRepository
|
||||
import com.correx.core.sessions.projections.replay.DefaultEventReplayer
|
||||
import com.correx.core.sessions.projections.replay.EventReplayer
|
||||
import com.correx.core.transitions.graph.StageConfig
|
||||
import com.correx.core.transitions.graph.TransitionEdge
|
||||
import com.correx.core.transitions.graph.WorkflowGraph
|
||||
import com.correx.core.validation.pipeline.ValidationPipeline
|
||||
import com.correx.infrastructure.persistence.InMemoryEventStore
|
||||
import com.correx.infrastructure.persistence.artifact.LiveArtifactRepository
|
||||
import com.correx.testing.contracts.fixtures.artifactstore.NoopArtifactStore
|
||||
import com.correx.testing.fixtures.InferenceFixtures
|
||||
import com.correx.testing.fixtures.context.ContextFixtures
|
||||
import com.correx.testing.fixtures.cyclePolicyMissingValidator
|
||||
import com.correx.testing.fixtures.transitions.TransitionFixtures
|
||||
import com.correx.testing.kernel.MockSessionEventReplayer
|
||||
import kotlinx.coroutines.runBlocking
|
||||
import kotlinx.datetime.Clock
|
||||
import org.junit.jupiter.api.Assertions.assertEquals
|
||||
import org.junit.jupiter.api.Assertions.assertTrue
|
||||
import org.junit.jupiter.api.Test
|
||||
|
||||
/**
|
||||
* End-to-end proof that the reviewer-loop runtime correlates recorded critique findings into
|
||||
* outcome events at loop resolution (BACKLOG §B-§6), and that those events feed the
|
||||
* `:core:critique` calibration projection. The LLM-side emission of findings is mocked here by
|
||||
* seeding [CritiqueFindingsRecordedEvent]s directly (that half is a separate model-gated step).
|
||||
*/
|
||||
class CritiqueCalibrationWiringTest {
|
||||
|
||||
private val eventStore = InMemoryEventStore()
|
||||
private val sessionReplayer = MockSessionEventReplayer()
|
||||
private val sessionRepository = DefaultSessionRepository(sessionReplayer)
|
||||
private val orchestrationRepository = OrchestrationRepository(
|
||||
DefaultEventReplayer(eventStore, OrchestrationProjector(DefaultOrchestrationReducer())),
|
||||
)
|
||||
private val inferenceRepository = InferenceRepository(
|
||||
object : EventReplayer<InferenceState> {
|
||||
override fun rebuild(sessionId: SessionId) = InferenceState()
|
||||
},
|
||||
)
|
||||
private val approvalRepository = DefaultApprovalRepository(
|
||||
DefaultEventReplayer(eventStore, ApprovalProjector(DefaultApprovalReducer())),
|
||||
)
|
||||
private val decisionJournalRepository = DefaultDecisionJournalRepository(
|
||||
DefaultEventReplayer(eventStore, DecisionJournalProjector(DefaultDecisionJournalReducer())),
|
||||
)
|
||||
private val retryCoordinator = DefaultRetryCoordinator(eventStore)
|
||||
|
||||
private val repositories = OrchestratorRepositories(
|
||||
eventStore = eventStore,
|
||||
inferenceRepository = inferenceRepository,
|
||||
orchestrationRepository = orchestrationRepository,
|
||||
sessionRepository = sessionRepository,
|
||||
artifactRepository = LiveArtifactRepository(eventStore, DefaultArtifactReducer()),
|
||||
approvalRepository = approvalRepository,
|
||||
)
|
||||
|
||||
private val orchestrator = DefaultSessionOrchestrator(
|
||||
repositories = repositories,
|
||||
engines = OrchestratorEngines(
|
||||
transitionResolver = TransitionFixtures.simpleResolver(),
|
||||
contextPackBuilder = ContextFixtures.simpleBuilder(),
|
||||
inferenceRouter = InferenceFixtures.fixedRouter(),
|
||||
validationPipeline = ValidationPipeline(validators = listOf(cyclePolicyMissingValidator())),
|
||||
approvalEngine = DefaultApprovalEngine(),
|
||||
riskAssessor = DefaultRiskAssessor(),
|
||||
),
|
||||
retryCoordinator = retryCoordinator,
|
||||
artifactStore = NoopArtifactStore(),
|
||||
decisionJournalRepository = decisionJournalRepository,
|
||||
)
|
||||
|
||||
private val defaultConfig = OrchestrationConfig(retryPolicy = RetryPolicy(maxAttempts = 1, backoffMs = 0))
|
||||
|
||||
private fun finding(id: String) = CritiqueFinding(
|
||||
id = id,
|
||||
role = CritiqueRole.CODE_REVIEWER,
|
||||
severity = CritiqueSeverity.MAJOR,
|
||||
category = "correctness",
|
||||
message = "msg-$id",
|
||||
)
|
||||
|
||||
private suspend fun seedReview(sessionId: SessionId, iteration: Int, verdict: CritiqueVerdict, findings: List<CritiqueFinding>) {
|
||||
eventStore.append(
|
||||
NewEvent(
|
||||
EventMetadata(EventId("rev-$iteration"), sessionId, Clock.System.now(), 1, null, null),
|
||||
CritiqueFindingsRecordedEvent(
|
||||
sessionId = sessionId,
|
||||
stageId = StageId("reviewer"),
|
||||
role = CritiqueRole.CODE_REVIEWER,
|
||||
modelHash = "model-A",
|
||||
iteration = iteration,
|
||||
verdict = verdict,
|
||||
findings = findings,
|
||||
),
|
||||
),
|
||||
)
|
||||
}
|
||||
|
||||
@Test
|
||||
fun `recorded findings are correlated at completion and feed the calibration projection`(): Unit = runBlocking {
|
||||
val sessionId = SessionId("calib-1")
|
||||
|
||||
// Two review rounds: f1 fixed between them (→ UPHELD); f2 persists to an approved final (→ DISMISSED).
|
||||
seedReview(sessionId, 1, CritiqueVerdict.CHANGES_REQUESTED, listOf(finding("f1"), finding("f2")))
|
||||
seedReview(sessionId, 2, CritiqueVerdict.APPROVED, listOf(finding("f2")))
|
||||
|
||||
val stageA = StageId("A")
|
||||
val done = StageId("done")
|
||||
val graph = WorkflowGraph(
|
||||
id = "calib-wiring",
|
||||
stages = mapOf(stageA to StageConfig(), done to StageConfig()),
|
||||
transitions = setOf(TransitionEdge(id = TransitionId("t1"), from = stageA, to = done, condition = { true })),
|
||||
start = stageA,
|
||||
)
|
||||
|
||||
orchestrator.run(sessionId, graph, defaultConfig)
|
||||
|
||||
// The orchestrator emits one outcome per recorded finding at completion — the events the
|
||||
// :core:critique calibration projection consumes (its folding is covered by that module's
|
||||
// DefaultCriticCalibrationReducerTest).
|
||||
val outcomes = eventStore.read(sessionId).mapNotNull { it.payload as? CritiqueOutcomeCorrelatedEvent }
|
||||
assertEquals(2, outcomes.size, "one outcome per recorded finding")
|
||||
assertEquals(CritiqueOutcome.UPHELD, outcomes.single { it.findingId == "f1" }.outcome)
|
||||
assertEquals(CritiqueOutcome.DISMISSED, outcomes.single { it.findingId == "f2" }.outcome)
|
||||
assertTrue(outcomes.all { it.modelHash == "model-A" && it.role == CritiqueRole.CODE_REVIEWER })
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,158 @@
|
||||
import com.correx.core.approvals.ApprovalProjector
|
||||
import com.correx.core.approvals.DefaultApprovalReducer
|
||||
import com.correx.core.approvals.DefaultApprovalRepository
|
||||
import com.correx.core.approvals.domain.DefaultApprovalEngine
|
||||
import com.correx.core.artifacts.DefaultArtifactReducer
|
||||
import com.correx.core.events.events.StaticFindingsRecordedEvent
|
||||
import com.correx.core.events.events.WorkflowCompletedEvent
|
||||
import com.correx.core.events.execution.RetryPolicy
|
||||
import com.correx.core.events.types.SessionId
|
||||
import com.correx.core.events.types.StageId
|
||||
import com.correx.core.events.types.TransitionId
|
||||
import com.correx.core.inference.InferenceRepository
|
||||
import com.correx.core.inference.InferenceState
|
||||
import com.correx.core.journal.DecisionJournalProjector
|
||||
import com.correx.core.journal.DefaultDecisionJournalReducer
|
||||
import com.correx.core.journal.DefaultDecisionJournalRepository
|
||||
import com.correx.core.kernel.orchestration.CommandOutput
|
||||
import com.correx.core.kernel.orchestration.CommandRunner
|
||||
import com.correx.core.kernel.orchestration.DefaultOrchestrationReducer
|
||||
import com.correx.core.kernel.orchestration.DefaultSessionOrchestrator
|
||||
import com.correx.core.kernel.orchestration.OrchestrationConfig
|
||||
import com.correx.core.kernel.orchestration.OrchestrationProjector
|
||||
import com.correx.core.kernel.orchestration.OrchestrationRepository
|
||||
import com.correx.core.kernel.orchestration.OrchestratorEngines
|
||||
import com.correx.core.kernel.orchestration.OrchestratorRepositories
|
||||
import com.correx.core.kernel.retry.DefaultRetryCoordinator
|
||||
import com.correx.core.risk.DefaultRiskAssessor
|
||||
import com.correx.core.sessions.DefaultSessionRepository
|
||||
import com.correx.core.sessions.projections.replay.DefaultEventReplayer
|
||||
import com.correx.core.sessions.projections.replay.EventReplayer
|
||||
import com.correx.core.transitions.graph.StageConfig
|
||||
import com.correx.core.transitions.graph.TransitionEdge
|
||||
import com.correx.core.transitions.graph.WorkflowGraph
|
||||
import com.correx.core.validation.pipeline.ValidationPipeline
|
||||
import com.correx.infrastructure.persistence.InMemoryEventStore
|
||||
import com.correx.infrastructure.persistence.artifact.LiveArtifactRepository
|
||||
import com.correx.testing.contracts.fixtures.artifactstore.NoopArtifactStore
|
||||
import com.correx.testing.fixtures.InferenceFixtures
|
||||
import com.correx.testing.fixtures.context.ContextFixtures
|
||||
import com.correx.testing.fixtures.cyclePolicyMissingValidator
|
||||
import com.correx.testing.fixtures.transitions.TransitionFixtures
|
||||
import com.correx.testing.kernel.MockSessionEventReplayer
|
||||
import kotlinx.coroutines.runBlocking
|
||||
import org.junit.jupiter.api.Assertions.assertEquals
|
||||
import org.junit.jupiter.api.Assertions.assertNotNull
|
||||
import org.junit.jupiter.api.Test
|
||||
|
||||
/**
|
||||
* End-to-end proof of the deterministic static_check stage seam (BACKLOG §B-§5): a stage marked
|
||||
* `stage_type = "static_check"` runs the configured tool via an injected [CommandRunner] (no LLM
|
||||
* subagent), records a [StaticFindingsRecordedEvent] with the parsed findings, and the workflow
|
||||
* advances on the unconditional exit edge.
|
||||
*/
|
||||
class StaticCheckStageTest {
|
||||
|
||||
private val eventStore = InMemoryEventStore()
|
||||
private val sessionReplayer = MockSessionEventReplayer()
|
||||
private val sessionRepository = DefaultSessionRepository(sessionReplayer)
|
||||
private val orchestrationRepository = OrchestrationRepository(
|
||||
DefaultEventReplayer(eventStore, OrchestrationProjector(DefaultOrchestrationReducer())),
|
||||
)
|
||||
private val inferenceRepository = InferenceRepository(
|
||||
object : EventReplayer<InferenceState> {
|
||||
override fun rebuild(sessionId: SessionId) = InferenceState()
|
||||
},
|
||||
)
|
||||
private val approvalRepository = DefaultApprovalRepository(
|
||||
DefaultEventReplayer(eventStore, ApprovalProjector(DefaultApprovalReducer())),
|
||||
)
|
||||
private val decisionJournalRepository = DefaultDecisionJournalRepository(
|
||||
DefaultEventReplayer(eventStore, DecisionJournalProjector(DefaultDecisionJournalReducer())),
|
||||
)
|
||||
private val retryCoordinator = DefaultRetryCoordinator(eventStore)
|
||||
|
||||
private val repositories = OrchestratorRepositories(
|
||||
eventStore = eventStore,
|
||||
inferenceRepository = inferenceRepository,
|
||||
orchestrationRepository = orchestrationRepository,
|
||||
sessionRepository = sessionRepository,
|
||||
artifactRepository = LiveArtifactRepository(eventStore, DefaultArtifactReducer()),
|
||||
approvalRepository = approvalRepository,
|
||||
)
|
||||
|
||||
// Canned tool output: one stderr error + one stdout detekt finding (+ noise that must be ignored).
|
||||
private val fakeRunner = object : CommandRunner {
|
||||
var lastArgv: List<String>? = null
|
||||
override suspend fun run(argv: List<String>): CommandOutput {
|
||||
lastArgv = argv
|
||||
return CommandOutput(
|
||||
exitCode = 1,
|
||||
stdout = "src/Foo.kt:42:9: Magic number found [MagicNumber]\nBUILD FAILED",
|
||||
stderr = "src/Bar.kt:5:9: error: type mismatch",
|
||||
)
|
||||
}
|
||||
}
|
||||
|
||||
private val orchestrator = DefaultSessionOrchestrator(
|
||||
repositories = repositories,
|
||||
engines = OrchestratorEngines(
|
||||
transitionResolver = TransitionFixtures.simpleResolver(),
|
||||
contextPackBuilder = ContextFixtures.simpleBuilder(),
|
||||
inferenceRouter = InferenceFixtures.fixedRouter(),
|
||||
validationPipeline = ValidationPipeline(validators = listOf(cyclePolicyMissingValidator())),
|
||||
approvalEngine = DefaultApprovalEngine(),
|
||||
riskAssessor = DefaultRiskAssessor(),
|
||||
),
|
||||
retryCoordinator = retryCoordinator,
|
||||
artifactStore = NoopArtifactStore(),
|
||||
decisionJournalRepository = decisionJournalRepository,
|
||||
commandRunner = fakeRunner,
|
||||
)
|
||||
|
||||
private val defaultConfig = OrchestrationConfig(
|
||||
retryPolicy = RetryPolicy(maxAttempts = 1, backoffMs = 0),
|
||||
)
|
||||
|
||||
@Test
|
||||
fun `static_check stage runs the tool deterministically and records findings`(): Unit = runBlocking {
|
||||
val sessionId = SessionId("static-check-1")
|
||||
val staticCheck = StageId("static_check")
|
||||
val done = StageId("done")
|
||||
|
||||
val graph = WorkflowGraph(
|
||||
id = "static-check-seam",
|
||||
stages = mapOf(
|
||||
staticCheck to StageConfig(
|
||||
metadata = mapOf(
|
||||
"stage_type" to "static_check",
|
||||
"static_tool" to "detekt",
|
||||
"static_argv" to "./gradlew detekt --console=plain",
|
||||
),
|
||||
),
|
||||
done to StageConfig(),
|
||||
),
|
||||
transitions = setOf(
|
||||
TransitionEdge(id = TransitionId("t1"), from = staticCheck, to = done, condition = { true }),
|
||||
),
|
||||
start = staticCheck,
|
||||
)
|
||||
|
||||
orchestrator.run(sessionId, graph, defaultConfig)
|
||||
|
||||
val events = eventStore.read(sessionId)
|
||||
val recorded = events.mapNotNull { it.payload as? StaticFindingsRecordedEvent }
|
||||
assertEquals(1, recorded.size, "the static_check stage must record exactly one findings event")
|
||||
val event = recorded.single()
|
||||
assertEquals("detekt", event.tool)
|
||||
assertEquals(2, event.findings.size, "both the stdout detekt finding and the stderr error parse")
|
||||
assertEquals("MagicNumber", event.findings[0].ruleId)
|
||||
assertEquals("src/Bar.kt", event.findings[1].file)
|
||||
assertEquals(listOf("./gradlew", "detekt", "--console=plain"), fakeRunner.lastArgv)
|
||||
|
||||
assertNotNull(
|
||||
events.find { it.payload is WorkflowCompletedEvent },
|
||||
"workflow advances past static_check on the unconditional exit edge and completes",
|
||||
)
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user