pulseed 0.5.3 → 0.6.0

package/dist/runtime/store/evidence-ledger.js

@@ -0,0 +1,1452 @@
+import { randomUUID } from "node:crypto";
+import * as fsp from "node:fs/promises";
+import * as path from "node:path";
+import { z } from "zod";
+import { createRuntimeStorePaths, ensureRuntimeStorePaths, } from "./runtime-paths.js";
+import { summarizeEvidenceMetricTrends } from "./metric-history.js";
+import { summarizeEvidenceEvaluatorResults, } from "./evaluator-results.js";
+import { summarizeEvidenceResearchMemos, } from "./research-evidence.js";
+import { summarizeEvidenceDreamCheckpoints, } from "./dream-checkpoints.js";
+import { summarizeArtifactRetention, RuntimeArtifactRetentionClassSchema, } from "./artifact-retention.js";
+export const RuntimeEvidenceOutcomeSchema = z.enum([
+    "improved",
+    "regressed",
+    "inconclusive",
+    "failed",
+    "blocked",
+    "continued",
+]);
+export const RuntimeEvidenceEntryKindSchema = z.enum([
+    "observation",
+    "strategy",
+    "task_generation",
+    "execution",
+    "verification",
+    "decision",
+    "metric",
+    "evaluator",
+    "research",
+    "dream_checkpoint",
+    "artifact",
+    "failure",
+    "other",
+]);
+export const RuntimeEvidenceArtifactRefSchema = z.object({
+    label: z.string().min(1),
+    path: z.string().min(1).optional(),
+    state_relative_path: z.string().min(1).optional(),
+    url: z.string().url().optional(),
+    kind: z.enum(["log", "metrics", "report", "diff", "url", "other"]).default("other"),
+    retention_class: RuntimeArtifactRetentionClassSchema.optional(),
+    size_bytes: z.number().int().nonnegative().optional(),
+    source: z.string().min(1).optional(),
+    dependency_refs: z.array(z.string().min(1)).optional(),
+}).strict();
+export const RuntimeEvidenceMetricSchema = z.object({
+    label: z.string().min(1),
+    value: z.union([z.string(), z.number(), z.boolean(), z.null()]).optional(),
+    unit: z.string().min(1).optional(),
+    direction: z.enum(["maximize", "minimize", "neutral"]).optional(),
+    confidence: z.number().min(0).max(1).optional(),
+    observed_at: z.string().datetime().optional(),
+    source: z.string().min(1).optional(),
+    summary: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceCandidateDispositionSchema = z.enum(["retained", "promoted", "retired"]);
+export const RuntimeEvidenceCandidateLineageSchema = z.object({
+    parent_candidate_id: z.string().min(1).optional(),
+    source_candidate_id: z.string().min(1).optional(),
+    source_strategy_id: z.string().min(1).optional(),
+    source_strategy: z.string().min(1).optional(),
+    strategy_family: z.string().min(1),
+    feature_lineage: z.array(z.string().min(1)).default([]),
+    model_lineage: z.array(z.string().min(1)).default([]),
+    config_lineage: z.array(z.string().min(1)).default([]),
+    seed_lineage: z.array(z.string().min(1)).default([]),
+    fold_lineage: z.array(z.string().min(1)).default([]),
+    postprocess_lineage: z.array(z.string().min(1)).default([]),
+    notes: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceCandidateSimilaritySchema = z.object({
+    candidate_id: z.string().min(1),
+    similarity: z.number().min(0).max(1),
+    signal: z.enum(["declared", "lineage", "metric_correlation", "artifact_overlap", "other"]).default("declared"),
+    summary: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceCandidateNearMissReasonSchema = z.enum([
+    "close_to_best",
+    "stability",
+    "novelty",
+    "weak_dimension_improvement",
+    "complementarity",
+    "ensemble_potential",
+]);
+export const RuntimeEvidenceCandidateNearMissSchema = z.object({
+    status: z.enum(["retained", "promoted", "rejected"]).default("retained"),
+    reason_to_keep: z.array(RuntimeEvidenceCandidateNearMissReasonSchema).min(1),
+    margin_to_best: z.number().min(0).optional(),
+    weak_dimensions: z.array(z.string().min(1)).default([]),
+    complementary_candidate_ids: z.array(z.string().min(1)).default([]),
+    follow_up: z.object({
+        title: z.string().min(1),
+        rationale: z.string().min(1),
+        target_dimensions: z.array(z.string().min(1)).default([]),
+        expected_evidence_gain: z.string().min(1).optional(),
+    }).strict().optional(),
+    evidence_refs: z.array(z.string().min(1)).default([]),
+    summary: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceCandidateRecordSchema = z.object({
+    candidate_id: z.string().min(1),
+    label: z.string().min(1).optional(),
+    lineage: RuntimeEvidenceCandidateLineageSchema,
+    metrics: z.array(RuntimeEvidenceMetricSchema).default([]),
+    artifacts: z.array(RuntimeEvidenceArtifactRefSchema).default([]),
+    similarity: z.array(RuntimeEvidenceCandidateSimilaritySchema).default([]),
+    robustness: z.object({
+        stability_score: z.number().min(0).max(1).optional(),
+        diversity_score: z.number().min(0).max(1).optional(),
+        risk_penalty: z.number().min(0).max(1).optional(),
+        robust_score: z.number().min(0).max(1).optional(),
+        evidence_confidence: z.number().min(0).max(1).optional(),
+        repeated_evaluations: z.number().int().nonnegative().optional(),
+        mean_score: z.number().optional(),
+        max_score: z.number().optional(),
+        score_stddev: z.number().min(0).optional(),
+        fold_score_range: z.number().min(0).optional(),
+        seed_score_range: z.number().min(0).optional(),
+        weak_dimensions: z.array(z.string().min(1)).default([]),
+        provenance_refs: z.array(z.string().min(1)).default([]),
+        summary: z.string().min(1).optional(),
+    }).strict().optional(),
+    near_miss: RuntimeEvidenceCandidateNearMissSchema.optional(),
+    disposition: RuntimeEvidenceCandidateDispositionSchema.default("retained"),
+    disposition_reason: z.string().min(1).optional(),
+    produced_at: z.string().datetime().optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorSignalSchema = z.enum(["local", "external"]);
+export const RuntimeEvidenceEvaluatorStatusSchema = z.enum([
+    "pending",
+    "ready",
+    "approval_required",
+    "submitted",
+    "passed",
+    "succeeded",
+    "completed",
+    "failed",
+    "regressed",
+    "blocked",
+    "unknown",
+]);
+export const RuntimeEvidenceEvaluatorPublishActionSchema = z.object({
+    id: z.string().min(1),
+    label: z.string().min(1),
+    tool_name: z.string().min(1).optional(),
+    payload_ref: z.string().min(1).optional(),
+    approval_required: z.literal(true).default(true),
+    status: z.enum(["approval_required", "approved", "submitted", "completed", "blocked"]).optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorValidationSchema = z.object({
+    status: z.enum(["pending", "passed", "failed", "blocked", "unknown"]).default("unknown"),
+    command: z.string().min(1).optional(),
+    summary: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorProvenanceSchema = z.object({
+    kind: z.enum(["local_command", "external_url", "ci", "benchmark", "human_review", "other"]).default("other"),
+    command: z.string().min(1).optional(),
+    url: z.string().url().optional(),
+    run_id: z.string().min(1).optional(),
+    external_id: z.string().min(1).optional(),
+    raw_ref: z.string().min(1).optional(),
+    retrieved_at: z.string().datetime().optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorBudgetSchema = z.object({
+    policy_id: z.string().min(1).optional(),
+    max_attempts: z.number().int().positive().optional(),
+    used_attempts: z.number().int().nonnegative().optional(),
+    remaining_attempts: z.number().int().nonnegative(),
+    approval_required: z.boolean().default(true),
+    deadline_at: z.string().datetime().optional(),
+    phase: z.enum(["exploration", "consolidation", "finalization", "other"]).optional(),
+    portfolio_policy: z.object({
+        diversified_portfolio_required: z.boolean().default(false),
+        reserve_for_finalization: z.boolean().default(false),
+        min_strategy_families: z.number().int().positive().optional(),
+    }).strict().optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorCandidateSnapshotSchema = z.object({
+    evidence_entry_id: z.string().min(1).optional(),
+    primary_metric_label: z.string().min(1).optional(),
+    local_metrics: z.array(RuntimeEvidenceMetricSchema).default([]),
+    robust_selection: z.object({
+        raw_rank: z.number().int().positive().optional(),
+        robust_score: z.number().min(0).max(1).optional(),
+        stability_score: z.number().min(0).max(1).optional(),
+        diversity_score: z.number().min(0).max(1).optional(),
+        risk_penalty: z.number().min(0).max(1).optional(),
+        portfolio_role: z.enum(["raw_best", "robust_best", "safe", "aggressive", "diverse", "near_miss", "other"]).optional(),
+    }).strict().optional(),
+    summary: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorCalibrationSchema = z.object({
+    mode: z.literal("calibration_only").default("calibration_only"),
+    use_for_selection: z.boolean().default(false),
+    direct_optimization_allowed: z.literal(false).default(false),
+    minimum_observations: z.number().int().positive().default(1),
+    conclusion: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceEvaluatorObservationSchema = z.object({
+    evaluator_id: z.string().min(1),
+    signal: RuntimeEvidenceEvaluatorSignalSchema,
+    source: z.string().min(1),
+    candidate_id: z.string().min(1),
+    candidate_label: z.string().min(1).optional(),
+    artifact_labels: z.array(z.string().min(1)).optional(),
+    status: RuntimeEvidenceEvaluatorStatusSchema.default("unknown"),
+    score: z.union([z.string(), z.number(), z.boolean(), z.null()]).optional(),
+    score_label: z.string().min(1).optional(),
+    direction: z.enum(["maximize", "minimize", "neutral"]).optional(),
+    observed_at: z.string().datetime().optional(),
+    expected_score: z.union([z.string(), z.number(), z.boolean(), z.null()]).optional(),
+    expected_status: RuntimeEvidenceEvaluatorStatusSchema.optional(),
+    expectation_source: z.string().min(1).optional(),
+    validation: RuntimeEvidenceEvaluatorValidationSchema.optional(),
+    publish_action: RuntimeEvidenceEvaluatorPublishActionSchema.optional(),
+    provenance: RuntimeEvidenceEvaluatorProvenanceSchema.optional(),
+    budget: RuntimeEvidenceEvaluatorBudgetSchema.optional(),
+    candidate_snapshot: RuntimeEvidenceEvaluatorCandidateSnapshotSchema.optional(),
+    calibration: RuntimeEvidenceEvaluatorCalibrationSchema.optional(),
+    summary: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceResearchSourceSchema = z.object({
+    url: z.string().url(),
+    title: z.string().min(1).optional(),
+    source_type: z.enum(["official_docs", "maintainer", "paper", "issue_thread", "example", "writeup", "other"]).default("other"),
+    provenance: z.enum(["quoted", "paraphrased", "summarized"]).default("summarized"),
+    relevance: z.string().min(1).optional(),
+}).strict();
+export const RuntimeEvidenceResearchFindingSchema = z.object({
+    finding: z.string().min(1),
+    source_urls: z.array(z.string().url()).min(1),
+    applicability: z.string().min(1),
+    risks_constraints: z.array(z.string().min(1)).default([]),
+    proposed_experiment: z.string().min(1),
+    expected_metric_impact: z.string().min(1),
+    fact_vs_adaptation: z.object({
+        facts: z.array(z.string().min(1)).default([]),
+        adaptation: z.string().min(1),
+    }).strict(),
+}).strict();
+export const RuntimeEvidenceResearchExternalActionSchema = z.object({
+    label: z.string().min(1),
+    reason: z.string().min(1),
+    approval_required: z.literal(true).default(true),
+}).strict();
+export const RuntimeEvidenceResearchMemoSchema = z.object({
+    trigger: z.enum(["plateau", "uncertainty", "knowledge_gap"]),
+    query: z.string().min(1),
+    summary: z.string().min(1),
+    sources: z.array(RuntimeEvidenceResearchSourceSchema).min(1),
+    findings: z.array(RuntimeEvidenceResearchFindingSchema).min(1),
+    candidate_playbook: z.object({
+        title: z.string().min(1),
+        steps: z.array(z.string().min(1)).default([]),
+        source_urls: z.array(z.string().url()).default([]),
+    }).strict().optional(),
+    untrusted_content_policy: z.literal("webpage_instructions_are_untrusted").default("webpage_instructions_are_untrusted"),
+    external_actions: z.array(RuntimeEvidenceResearchExternalActionSchema).default([]),
+    confidence: z.number().min(0).max(1).default(0.5),
+}).strict();
+export const RuntimeEvidenceDreamCheckpointTriggerSchema = z.enum([
+    "iteration",
+    "plateau",
+    "breakthrough",
+    "pre_finalization",
+]);
+export const RuntimeEvidenceDreamCheckpointMemoryRefSchema = z.object({
+    source_type: z.enum(["soil", "playbook", "runtime_evidence", "other"]),
+    ref: z.string().min(1).optional(),
+    summary: z.string().min(1),
+    authority: z.literal("advisory_only").default("advisory_only"),
+    relevance_score: z.number().min(0).max(1).optional(),
+    source_reliability: z.number().min(0).max(1).optional(),
+    recency_score: z.number().min(0).max(1).optional(),
+    prior_success_contribution: z.number().min(0).max(1).optional(),
+    retrieval: z.object({
+        kind: z.enum(["route_hit", "fallback_hit", "checkpoint", "manual", "unknown"]).default("unknown"),
+        score: z.number().min(0).max(1).optional(),
+        confidence: z.number().min(0).max(1).optional(),
+    }).strict().optional(),
+    ranking_trace: z.object({
+        score: z.number().min(0).max(1),
+        decision: z.enum(["admitted", "rejected"]),
+        reason: z.string().min(1),
+    }).strict().optional(),
+}).strict();
+export const RuntimeEvidenceDreamCheckpointStrategyCandidateSchema = z.object({
+    title: z.string().min(1),
+    rationale: z.string().min(1),
+    target_dimensions: z.array(z.string().min(1)).default([]),
+    expected_evidence_gain: z.string().min(1).optional(),
+    retry_reason: z.string().min(1).optional(),
+    failed_lineage_warning: z.object({
+        fingerprint: z.string().min(1),
+        count: z.number().int().positive(),
+        reason: z.string().min(1),
+    }).strict().optional(),
+}).strict();
+export const RuntimeEvidenceDreamCheckpointActiveHypothesisSchema = z.object({
+    hypothesis: z.string().min(1),
+    supporting_evidence_ref: z.string().min(1).optional(),
+    target_metric_or_dimension: z.string().min(1),
+    expected_next_observation: z.string().min(1),
+    status: z.enum(["active", "testing", "supported", "weakened"]).default("active"),
+}).strict();
+export const RuntimeEvidenceDreamCheckpointRejectedApproachSchema = z.object({
+    approach: z.string().min(1),
+    rejection_reason: z.string().min(1),
+    evidence_ref: z.string().min(1).optional(),
+    revisit_condition: z.string().min(1).optional(),
+    confidence: z.number().min(0).max(1).default(0.5),
+}).strict();
+export const RuntimeEvidenceDreamRunControlRecommendationSchema = z.object({
+    id: z.string().min(1).optional(),
+    action: z.enum([
+        "stay_current_mode",
+        "widen_exploration",
+        "consolidate_candidates",
+        "freeze_experiment_queue",
+        "enter_finalization",
+        "preserve_near_miss_candidates",
+        "retire_low_value_lineage",
+        "request_operator_approval",
+    ]),
+    rationale: z.string().min(1),
+    evidence: z.array(z.object({
+        kind: z.enum(["metric", "artifact", "lineage", "task_history", "deadline", "external_feedback", "memory", "runtime_state"]),
+        ref: z.string().min(1).optional(),
+        summary: z.string().min(1),
+    }).strict()).min(1),
+    target_mode: z.enum(["exploration", "consolidation", "finalization"]).optional(),
+    target_strategy_family: z.string().min(1).optional(),
+    candidate_refs: z.array(z.string().min(1)).default([]),
+    lineage_refs: z.array(z.string().min(1)).default([]),
+    approval_required: z.boolean().default(false),
+    risk: z.enum(["low", "medium", "high"]).default("medium"),
+    confidence: z.number().min(0).max(1).default(0.5),
+    policy_decision: z.object({
+        disposition: z.enum(["auto_apply", "approval_required", "advisory_only"]),
+        reason: z.string().min(1),
+    }).strict().optional(),
+}).strict();
+export const RuntimeEvidenceDreamCheckpointSchema = z.object({
+    trigger: RuntimeEvidenceDreamCheckpointTriggerSchema,
+    summary: z.string().min(1),
+    current_goal: z.string().min(1),
+    active_dimensions: z.array(z.string().min(1)).default([]),
+    best_evidence_so_far: z.string().min(1).optional(),
+    recent_strategy_families: z.array(z.string().min(1)).default([]),
+    exhausted: z.array(z.string().min(1)).default([]),
+    promising: z.array(z.string().min(1)).default([]),
+    relevant_memories: z.array(RuntimeEvidenceDreamCheckpointMemoryRefSchema).default([]),
+    active_hypotheses: z.array(RuntimeEvidenceDreamCheckpointActiveHypothesisSchema).default([]),
+    rejected_approaches: z.array(RuntimeEvidenceDreamCheckpointRejectedApproachSchema).default([]),
+    next_strategy_candidates: z.array(RuntimeEvidenceDreamCheckpointStrategyCandidateSchema).default([]),
+    run_control_recommendations: z.array(RuntimeEvidenceDreamRunControlRecommendationSchema).optional(),
+    guidance: z.string().min(1),
+    uncertainty: z.array(z.string().min(1)).default([]),
+    context_authority: z.literal("advisory_only").default("advisory_only"),
+    confidence: z.number().min(0).max(1).default(0.5),
+}).strict();
+export const RuntimeEvidenceDivergentHypothesisSchema = z.object({
+    strategy_id: z.string().min(1).optional(),
+    hypothesis: z.string().min(1),
+    strategy_family: z.string().min(1),
+    role: z.enum(["exploitation", "adjacent_exploration", "divergent_exploration"]),
+    novelty_score: z.number().min(0).max(1),
+    similarity_to_recent_failures: z.number().min(0).max(1).default(0),
+    expected_cost: z.enum(["low", "medium", "high"]),
+    relationship_to_lineage: z.enum([
+        "current_best",
+        "neighbor",
+        "failed_lineage",
+        "different_mechanism",
+        "different_assumption",
+        "unknown",
+    ]),
+    prior_evidence: z.string().min(1).optional(),
+    downrank_reason: z.string().min(1).optional(),
+    smoke_status: z.enum(["not_run", "promote", "defer", "retire"]).default("not_run"),
+    smoke_reason: z.string().min(1).optional(),
+    smoke_evidence_ref: z.string().min(1).optional(),
+    evidence_authority: z.literal("speculative_hypothesis").default("speculative_hypothesis"),
+}).strict();
+export const RuntimeEvidenceEntrySchema = z.object({
+    schema_version: z.literal("runtime-evidence-entry-v1"),
+    id: z.string().min(1),
+    occurred_at: z.string().datetime(),
+    kind: RuntimeEvidenceEntryKindSchema,
+    scope: z.object({
+        goal_id: z.string().min(1).optional(),
+        run_id: z.string().min(1).optional(),
+        task_id: z.string().min(1).optional(),
+        loop_index: z.number().int().nonnegative().optional(),
+        phase: z.string().min(1).optional(),
+    }).strict(),
+    hypothesis: z.string().min(1).optional(),
+    strategy: z.string().min(1).optional(),
+    task: z.object({
+        id: z.string().min(1).optional(),
+        description: z.string().min(1).optional(),
+        action: z.string().min(1).optional(),
+        primary_dimension: z.string().min(1).optional(),
+    }).strict().optional(),
+    verification: z.object({
+        command: z.string().min(1).optional(),
+        verdict: z.string().min(1).optional(),
+        confidence: z.number().min(0).max(1).optional(),
+        summary: z.string().min(1).optional(),
+    }).strict().optional(),
+    metrics: z.array(RuntimeEvidenceMetricSchema).default([]),
+    evaluators: z.array(RuntimeEvidenceEvaluatorObservationSchema).optional(),
+    research: z.array(RuntimeEvidenceResearchMemoSchema).optional(),
+    dream_checkpoints: z.array(RuntimeEvidenceDreamCheckpointSchema).optional(),
+    divergent_exploration: z.array(RuntimeEvidenceDivergentHypothesisSchema).optional(),
+    candidates: z.array(RuntimeEvidenceCandidateRecordSchema).optional(),
+    artifacts: z.array(RuntimeEvidenceArtifactRefSchema).default([]),
+    result: z.object({
+        status: z.string().min(1).optional(),
+        summary: z.string().min(1).optional(),
+        error: z.string().min(1).optional(),
+    }).strict().optional(),
+    outcome: RuntimeEvidenceOutcomeSchema.optional(),
+    decision_reason: z.string().min(1).optional(),
+    raw_refs: z.array(z.object({
+        kind: z.string().min(1),
+        id: z.string().min(1).optional(),
+        path: z.string().min(1).optional(),
+        state_relative_path: z.string().min(1).optional(),
+        url: z.string().url().optional(),
+    }).strict()).default([]),
+    summary: z.string().min(1).optional(),
+}).strict().refine((entry) => entry.scope.goal_id || entry.scope.run_id, {
+    message: "goal_id or run_id is required",
+    path: ["scope"],
+});
+export class RuntimeEvidenceLedger {
+    paths;
+    constructor(runtimeRootOrPaths) {
+        this.paths =
+            typeof runtimeRootOrPaths === "string"
+                ? createRuntimeStorePaths(runtimeRootOrPaths)
+                : runtimeRootOrPaths ?? createRuntimeStorePaths();
+    }
+    async ensureReady() {
+        await ensureRuntimeStorePaths(this.paths);
+    }
+    goalPath(goalId) {
+        return this.paths.evidenceGoalPath(goalId);
+    }
+    runPath(runId) {
+        return this.paths.evidenceRunPath(runId);
+    }
+    async append(input) {
+        const entry = RuntimeEvidenceEntrySchema.parse({
+            schema_version: "runtime-evidence-entry-v1",
+            id: input.id ?? randomUUID(),
+            occurred_at: input.occurred_at ?? new Date().toISOString(),
+            metrics: input.metrics ?? [],
+            evaluators: input.evaluators ?? [],
+            research: input.research ?? [],
+            dream_checkpoints: input.dream_checkpoints ?? [],
+            divergent_exploration: input.divergent_exploration ?? [],
+            candidates: input.candidates ?? [],
+            artifacts: input.artifacts ?? [],
+            raw_refs: input.raw_refs ?? [],
+            ...input,
+        });
+        await this.ensureReady();
+        const targets = new Set();
+        if (entry.scope.goal_id)
+            targets.add(this.paths.evidenceGoalPath(entry.scope.goal_id));
+        if (entry.scope.run_id)
+            targets.add(this.paths.evidenceRunPath(entry.scope.run_id));
+        await Promise.all([...targets].map(async (target) => {
+            await fsp.mkdir(path.dirname(target), { recursive: true });
+            await fsp.appendFile(target, `${JSON.stringify(entry)}\n`, "utf8");
+        }));
+        await Promise.all([...targets].map(async (target) => {
+            await rebuildSummaryIndex(target, this.paths);
+        }));
+        return [entry];
+    }
+    async readByGoal(goalId) {
+        return readEvidenceFile(this.paths.evidenceGoalPath(goalId));
+    }
+    async readByRun(runId) {
+        return readEvidenceFile(this.paths.evidenceRunPath(runId));
+    }
+    async summarizeGoal(goalId) {
+        const manifests = await readReproducibilityManifests(this.paths, { goal_id: goalId });
+        const indexed = manifests.length === 0
+            ? await readSummaryIndex(this.paths.evidenceGoalPath(goalId), { goal_id: goalId })
+            : null;
+        if (indexed)
+            return indexed.summary;
+        const read = await this.readByGoal(goalId);
+        const summary = summarizeEvidence({ goal_id: goalId }, read, manifests);
+        return summary;
+    }
+    async summarizeRun(runId) {
+        const manifests = await readReproducibilityManifests(this.paths, { run_id: runId });
+        const indexed = manifests.length === 0
+            ? await readSummaryIndex(this.paths.evidenceRunPath(runId), { run_id: runId })
+            : null;
+        if (indexed)
+            return indexed.summary;
+        const read = await this.readByRun(runId);
+        const summary = summarizeEvidence({ run_id: runId }, read, manifests);
+        return summary;
+    }
+    async rebuildSummaryIndexForGoal(goalId) {
+        return rebuildSummaryIndex(this.paths.evidenceGoalPath(goalId), this.paths);
+    }
+    async rebuildSummaryIndexForRun(runId) {
+        return rebuildSummaryIndex(this.paths.evidenceRunPath(runId), this.paths);
+    }
+}
+async function rebuildSummaryIndex(canonicalPath, paths) {
+    const scope = summaryScopeFromPath(canonicalPath);
+    const read = await readEvidenceFile(canonicalPath);
+    const manifests = await readReproducibilityManifests(paths, scope);
+    const summary = summarizeEvidence(scope, read, manifests);
+    await writeSummaryIndex(canonicalPath, summary);
+    return summary;
+}
+function summaryIndexPath(canonicalPath) {
+    return `${canonicalPath}.summary.json`;
+}
+function summaryScopeFromPath(canonicalPath) {
+    const basename = path.basename(canonicalPath, ".jsonl");
+    const decoded = decodeURIComponent(basename);
+    return canonicalPath.includes(`${path.sep}runs${path.sep}`) ? { run_id: decoded } : { goal_id: decoded };
+}
+async function readReproducibilityManifests(paths, scope) {
+    let fileNames;
+    try {
+        fileNames = await fsp.readdir(paths.reproducibilityManifestsDir);
+    }
+    catch (err) {
+        if (err.code === "ENOENT")
+            return [];
+        throw err;
+    }
+    const manifests = [];
+    for (const fileName of fileNames) {
+        if (!fileName.endsWith(".json"))
+            continue;
+        try {
+            const parsed = JSON.parse(await fsp.readFile(path.join(paths.reproducibilityManifestsDir, fileName), "utf8"));
+            if (parsed.schema_version !== "runtime-reproducibility-manifest-v1")
+                continue;
+            if (scope.goal_id && parsed.scope?.goal_id !== scope.goal_id)
+                continue;
+            if (scope.run_id && parsed.scope?.run_id !== scope.run_id)
+                continue;
+            manifests.push({
+                ...parsed,
+                artifacts: Array.isArray(parsed.artifacts) ? parsed.artifacts.filter(isManifestArtifactRef) : [],
+            });
+        }
+        catch {
+            continue;
+        }
+    }
+    return manifests;
+}
+function isManifestArtifactRef(value) {
+    return typeof value === "object"
+        && value !== null
+        && typeof value.label === "string";
+}
+async function readSummaryIndex(canonicalPath, expectedScope) {
+    let text;
+    try {
+        text = await fsp.readFile(summaryIndexPath(canonicalPath), "utf8");
+    }
+    catch (err) {
+        if (err.code === "ENOENT")
+            return null;
+        throw err;
+    }
+    try {
+        const parsed = JSON.parse(text);
+        if (parsed.schema_version !== "runtime-evidence-summary-index-v1")
+            return null;
+        if (parsed.summary.schema_version !== "runtime-evidence-summary-v1")
+            return null;
+        if (!isCurrentEvidenceSummaryShape(parsed.summary))
+            return null;
+        const stat = await fsp.stat(canonicalPath);
+        if (parsed.canonical_log_size !== stat.size)
+            return null;
+        if (parsed.canonical_log_mtime_ms !== stat.mtimeMs)
+            return null;
+        if (expectedScope.goal_id && parsed.summary.scope.goal_id !== expectedScope.goal_id)
+            return null;
+        if (expectedScope.run_id && parsed.summary.scope.run_id !== expectedScope.run_id)
+            return null;
+        return parsed;
+    }
+    catch {
+        return null;
+    }
+}
+function isCurrentEvidenceSummaryShape(summary) {
+    return Array.isArray(summary.candidate_lineages)
+        && Array.isArray(summary.recommended_candidate_portfolio)
+        && Array.isArray(summary.near_miss_candidates)
+        && typeof summary.artifact_retention === "object"
+        && summary.artifact_retention !== null
+        && Array.isArray(summary.evaluator_summary.budgets)
+        && Array.isArray(summary.evaluator_summary.calibration)
+        && typeof summary.candidate_selection_summary === "object"
+        && summary.candidate_selection_summary !== null;
+}
+async function writeSummaryIndex(canonicalPath, summary) {
+    const stat = await fsp.stat(canonicalPath);
+    const index = {
+        schema_version: "runtime-evidence-summary-index-v1",
+        generated_at: new Date().toISOString(),
+        canonical_log_path: canonicalPath,
+        canonical_log_size: stat.size,
+        canonical_log_mtime_ms: stat.mtimeMs,
+        summary,
+    };
+    await fsp.mkdir(path.dirname(canonicalPath), { recursive: true });
+    await fsp.writeFile(summaryIndexPath(canonicalPath), `${JSON.stringify(index)}\n`, "utf8");
+}
+async function readEvidenceFile(filePath) {
+    let text;
+    try {
+        text = await fsp.readFile(filePath, "utf8");
+    }
+    catch (err) {
+        if (err.code === "ENOENT") {
+            return { entries: [], warnings: [] };
+        }
+        throw err;
+    }
+    const entries = [];
+    const warnings = [];
+    const lines = text.split(/\r?\n/);
+    for (let index = 0; index < lines.length; index += 1) {
+        const line = lines[index];
+        if (!line?.trim())
+            continue;
+        try {
+            const parsed = RuntimeEvidenceEntrySchema.safeParse(JSON.parse(line));
+            if (parsed.success) {
+                entries.push(parsed.data);
+            }
+            else {
+                warnings.push({
+                    file: filePath,
+                    line: index + 1,
+                    message: parsed.error.issues.map((issue) => issue.message).join("; "),
+                });
+            }
+        }
+        catch (err) {
+            warnings.push({
+                file: filePath,
+                line: index + 1,
+                message: err instanceof Error ? err.message : String(err),
+            });
+        }
+    }
+    return { entries, warnings };
+}
+function summarizeEvidence(scope, read, manifests = []) {
+    const entries = [...read.entries].sort((a, b) => a.occurred_at.localeCompare(b.occurred_at));
+    const newestFirst = [...entries].reverse();
+    const evaluatorSummary = summarizeEvidenceEvaluatorResults(entries);
+    return {
+        schema_version: "runtime-evidence-summary-v1",
+        generated_at: new Date().toISOString(),
+        scope,
+        total_entries: entries.length,
+        latest_strategy: newestFirst.find((entry) => entry.kind === "strategy" || Boolean(entry.strategy) || Boolean(entry.decision_reason)) ?? null,
+        best_evidence: chooseBestEvidence(newestFirst),
+        metric_trends: summarizeEvidenceMetricTrends(entries),
+        evaluator_summary: evaluatorSummary,
+        research_memos: summarizeEvidenceResearchMemos(entries),
+        dream_checkpoints: summarizeEvidenceDreamCheckpoints(entries),
+        divergent_exploration: entries
+            .flatMap((entry) => entry.divergent_exploration ?? [])
+            .slice(-10)
+            .reverse(),
+        candidate_lineages: summarizeCandidateLineages(entries),
+        recommended_candidate_portfolio: selectDiversifiedCandidatePortfolio(entries),
+        candidate_selection_summary: summarizeCandidateSelection(entries, evaluatorSummary),
+        near_miss_candidates: summarizeNearMissCandidates(entries),
+        artifact_retention: summarizeArtifactRetention(entries, { manifests }),
+        recent_failed_attempts: newestFirst
+            .filter((entry) => entry.outcome === "failed"
+            || entry.outcome === "regressed"
+            || entry.kind === "failure"
+            || entry.result?.status === "failed"
+            || entry.verification?.verdict === "fail")
+            .slice(0, 5),
+        failed_lineages: summarizeFailedLineages(entries),
+        recent_entries: newestFirst.slice(0, 10),
+        warnings: read.warnings,
+    };
+}
+export function selectDiversifiedCandidatePortfolio(entriesOldestFirst, options = {}) {
+    const limit = options.limit ?? 3;
+    if (limit <= 0)
+        return [];
+    const nearDuplicateSimilarity = options.nearDuplicateSimilarity ?? 0.85;
+    const primaryMetric = resolvePrimaryCandidateMetricKey(entriesOldestFirst);
+    const candidates = extractCandidateEvidenceContexts(entriesOldestFirst, primaryMetric)
+        .filter((context) => context.candidate.disposition !== "retired")
+        .sort(compareCandidateEvidenceContexts);
+    const selected = [];
+    const skipped = [];
+    for (const candidate of candidates) {
+        if (selected.length >= limit)
+            break;
+        const duplicateSignal = mostSimilarSelectedCandidate(candidate, selected);
+        if (duplicateSignal && duplicateSignal.similarity >= nearDuplicateSimilarity) {
+            skipped.push({ ...candidate, similarity_to_selected: duplicateSignal });
+            continue;
+        }
+        selected.push({
+            ...candidate,
+            role: selected.length === 0 ? "top_metric" : "diverse_representative",
+            ...(duplicateSignal ? { similarity_to_selected: duplicateSignal } : {}),
+        });
+    }
+    for (const candidate of skipped) {
+        if (selected.length >= limit)
+            break;
+        selected.push({
+            ...candidate,
+            role: "lineage_representative",
+        });
+    }
+    return selected.map(toPortfolioSlot);
+}
+function summarizeCandidateSelection(entriesOldestFirst, evaluatorSummary) {
+    const primaryMetric = resolvePrimaryCandidateMetricKey(entriesOldestFirst);
+    const contexts = extractCandidateEvidenceContexts(entriesOldestFirst, primaryMetric)
+        .filter((context) => context.candidate.disposition !== "retired");
+    const rawRanked = [...contexts].sort(compareCandidateEvidenceContexts);
+    const scored = scoreCandidateSelectionContexts(rawRanked, evaluatorSummary?.calibration ?? []);
+    const ranked = [...scored].sort((a, b) => b.robust_score - a.robust_score || a.raw_rank - b.raw_rank);
+    const rawBest = scored.find((candidate) => candidate.raw_rank === 1) ?? null;
+    const robustBest = ranked[0] ?? null;
+    return {
+        primary_metric: primaryMetric,
+        raw_best: rawBest,
+        robust_best: robustBest,
+        ranked,
+        final_portfolio: {
+            safe: selectSafeCandidate(scored),
+            aggressive: rawBest,
+            diverse: selectDiverseCandidate(scored, robustBest),
+        },
+    };
+}
+function scoreCandidateSelectionContexts(rawRanked, calibration = []) {
+    const metricValues = rawRanked
+        .map((context) => context.metric?.value)
+        .filter((value) => typeof value === "number" && Number.isFinite(value));
+    const minMetric = metricValues.length > 0 ? Math.min(...metricValues) : 0;
+    const maxMetric = metricValues.length > 0 ? Math.max(...metricValues) : 0;
+    const rawBestFamily = rawRanked[0]?.candidate.lineage.strategy_family;
+    const allCandidates = rawRanked.map((context) => context.candidate);
+    return rawRanked.map((context, index) => {
+        const candidate = context.candidate;
+        const metricScore = normalizedMetricScore(context.metric, minMetric, maxMetric);
+        const stabilityScore = clamp01(candidate.robustness?.stability_score ?? context.metric?.confidence ?? 0.5);
+        const inferredDiversity = inferredDiversityScore(candidate, rawBestFamily, allCandidates);
+        const diversityScore = clamp01(candidate.robustness?.diversity_score === undefined
+            ? inferredDiversity
+            : Math.min(candidate.robustness.diversity_score, inferredDiversity));
+        const riskPenalty = clamp01(candidate.robustness?.risk_penalty ?? inferredCandidateRiskPenalty(candidate));
+        const evidenceConfidence = clamp01(candidate.robustness?.evidence_confidence ?? context.metric?.confidence ?? 0.5);
+        const calibrationAdjustment = evaluatorCalibrationAdjustment(candidate.candidate_id, calibration);
+        const robustScore = clamp01(candidate.robustness?.robust_score
+            ?? (metricScore * 0.45 + stabilityScore * 0.3 + diversityScore * 0.15 + evidenceConfidence * 0.1 - riskPenalty + calibrationAdjustment));
+        return {
+            candidate_id: candidate.candidate_id,
+            ...(candidate.label ? { label: candidate.label } : {}),
+            strategy_family: candidate.lineage.strategy_family,
+            evidence_entry_id: context.entry_id,
+            raw_rank: index + 1,
+            ...(context.metric ? { raw_metric: context.metric } : {}),
+            robust_score: roundScore(robustScore),
+            calibration_adjustment: roundScore(calibrationAdjustment),
+            metric_score: roundScore(metricScore),
+            stability_score: roundScore(stabilityScore),
+            diversity_score: roundScore(diversityScore),
+            risk_penalty: roundScore(riskPenalty),
+            evidence_confidence: roundScore(evidenceConfidence),
+            reasons: candidateSelectionReasons(candidate, {
+                metricScore,
+                stabilityScore,
+                diversityScore,
+                riskPenalty,
+                evidenceConfidence,
+                calibrationAdjustment,
+            }),
+        };
+    });
+}
+function evaluatorCalibrationAdjustment(candidateId, calibration) {
+    const relevant = calibration.filter((item) => item.candidate_id === candidateId
+        && item.use_for_selection
+        && item.direct_optimization_allowed === false);
+    if (relevant.length === 0)
+        return 0;
+    const average = relevant.reduce((sum, item) => sum + item.selection_adjustment, 0) / relevant.length;
+    if (relevant.length < Math.max(...relevant.map((item) => item.minimum_observations)))
+        return 0;
+    return Math.min(0.08, Math.max(-0.08, average));
+}
+function normalizedMetricScore(metric, minMetric, maxMetric) {
+    if (!metric)
+        return 0;
+    if (maxMetric === minMetric)
+        return 0.5;
+    const distance = metric.direction === "maximize"
+        ? (metric.value - minMetric) / (maxMetric - minMetric)
+        : (maxMetric - metric.value) / (maxMetric - minMetric);
+    return clamp01(distance);
+}
+function inferredDiversityScore(candidate, rawBestFamily, allCandidates) {
+    if (!rawBestFamily)
+        return 0.5;
+    const highestSimilarity = highestKnownSimilarity(candidate, allCandidates);
+    const lineageBase = candidate.lineage.strategy_family !== rawBestFamily ? 0.75 : 0.45;
+    if (highestSimilarity > 0)
+        return clamp01(Math.min(lineageBase, 1 - highestSimilarity));
+    return lineageBase;
+}
+function highestKnownSimilarity(candidate, allCandidates) {
+    let highest = candidate.similarity.reduce((max, similarity) => Math.max(max, similarity.similarity), 0);
+    for (const other of allCandidates) {
+        if (other.candidate_id === candidate.candidate_id)
+            continue;
+        for (const similarity of other.similarity) {
+            if (similarity.candidate_id === candidate.candidate_id) {
+                highest = Math.max(highest, similarity.similarity);
+            }
+        }
+    }
+    return highest;
+}
+function inferredCandidateRiskPenalty(candidate) {
+    const lineageText = [
+        ...candidate.lineage.config_lineage,
+        ...candidate.lineage.postprocess_lineage,
+    ].map(normalizeLineageText).join(" ");
+    if (lineageText.includes("manual") || lineageText.includes("threshold") || lineageText.includes("postprocess")) {
+        return 0.15;
+    }
+    if (lineageText.includes("calibration") || lineageText.includes("weight")) {
+        return 0.08;
+    }
+    return 0;
+}
+function candidateSelectionReasons(candidate, scores) {
+    const reasons = [];
+    if (scores.stabilityScore >= 0.8)
+        reasons.push("strong stability evidence");
+    if (scores.diversityScore >= 0.8)
+        reasons.push("diverse lineage");
+    if (scores.riskPenalty >= 0.15)
+        reasons.push("penalized for overfit-prone lineage or post-processing");
+    if (scores.metricScore >= 0.95)
+        reasons.push("top raw metric evidence");
+    if (scores.calibrationAdjustment > 0)
+        reasons.push("external feedback calibrates local validation upward");
+    if (scores.calibrationAdjustment < 0)
+        reasons.push("external feedback calibrates local validation downward");
+    if (candidate.robustness?.summary)
+        reasons.push(candidate.robustness.summary);
+    return reasons.length > 0 ? reasons : ["risk-adjusted candidate evidence"];
+}
+function selectSafeCandidate(candidates) {
+    return [...candidates].sort((a, b) => b.stability_score - a.stability_score
+        || a.risk_penalty - b.risk_penalty
+        || b.robust_score - a.robust_score
+        || a.raw_rank - b.raw_rank)[0] ?? null;
+}
+function selectDiverseCandidate(candidates, robustBest) {
+    const robustFamily = robustBest?.strategy_family;
+    return [...candidates]
+        .filter((candidate) => !robustFamily || candidate.strategy_family !== robustFamily)
+        .filter((candidate) => candidate.diversity_score >= 0.5)
+        .sort((a, b) => b.diversity_score - a.diversity_score
+        || b.robust_score - a.robust_score
+        || a.raw_rank - b.raw_rank)[0] ?? null;
+}
+function summarizeNearMissCandidates(entriesOldestFirst) {
+    const primaryMetric = resolvePrimaryCandidateMetricKey(entriesOldestFirst);
+    const contexts = extractCandidateEvidenceContexts(entriesOldestFirst, primaryMetric)
+        .filter((context) => context.candidate.disposition !== "retired");
+    const rawRanked = [...contexts].sort(compareCandidateEvidenceContexts);
+    const rawBest = rawRanked[0] ?? null;
+    const scoredByCandidateId = new Map(scoreCandidateSelectionContexts(rawRanked).map((candidate) => [candidate.candidate_id, candidate]));
+    const result = [];
+    for (const context of rawRanked) {
+        if (!rawBest || context.candidate.candidate_id === rawBest.candidate.candidate_id)
+            continue;
+        const scored = scoredByCandidateId.get(context.candidate.candidate_id);
+        const reasons = nearMissReasonsForCandidate(context, rawBest, scored);
+        if (reasons.length === 0)
+            continue;
+        const nearMiss = context.candidate.near_miss;
+        if (nearMiss?.status === "rejected")
+            continue;
+        result.push({
+            candidate_id: context.candidate.candidate_id,
+            ...(context.candidate.label ? { label: context.candidate.label } : {}),
+            strategy_family: context.candidate.lineage.strategy_family,
+            evidence_entry_id: context.entry_id,
+            occurred_at: context.occurred_at,
+            raw_rank: scored?.raw_rank ?? rawRanked.indexOf(context) + 1,
+            ...(context.metric ? { raw_metric: context.metric } : {}),
+            raw_best_candidate_id: rawBest.candidate.candidate_id,
+            ...nearMissMarginContext(context, rawBest, nearMiss),
+            reason_to_keep: reasons,
+            weak_dimensions: nearMiss?.weak_dimensions ?? context.candidate.robustness?.weak_dimensions ?? [],
+            complementary_candidate_ids: nearMiss?.complementary_candidate_ids ?? complementaryCandidateIds(context.candidate),
+            ...(nearMiss?.follow_up ? { follow_up: nearMiss.follow_up } : {}),
+            ...(context.candidate.disposition_reason ? { retained_reason: context.candidate.disposition_reason } : {}),
+            evidence_refs: nearMiss?.evidence_refs ?? context.candidate.robustness?.provenance_refs ?? [],
+            ...(nearMiss?.summary ?? context.candidate.robustness?.summary
+                ? { summary: nearMiss?.summary ?? context.candidate.robustness?.summary }
+                : {}),
+        });
+    }
+    return result.sort((a, b) => nearMissReasonRank(b.reason_to_keep) - nearMissReasonRank(a.reason_to_keep)
+        || (b.raw_metric?.confidence ?? 0) - (a.raw_metric?.confidence ?? 0)
+        || a.raw_rank - b.raw_rank).slice(0, 8);
+}
+function nearMissReasonsForCandidate(context, rawBest, scored) {
+    const explicit = context.candidate.near_miss?.reason_to_keep ?? [];
+    const reasons = new Set(explicit);
+    if (context.candidate.near_miss?.status === "retained" || context.candidate.near_miss?.status === "promoted") {
+        for (const reason of inferredNearMissReasons(context, rawBest, scored))
+            reasons.add(reason);
+    }
+    else if (context.candidate.near_miss) {
+        for (const reason of inferredNearMissReasons(context, rawBest, scored))
+            reasons.add(reason);
+    }
+    else if (isImplicitNearMiss(context, rawBest, scored)) {
+        for (const reason of inferredNearMissReasons(context, rawBest, scored))
+            reasons.add(reason);
+    }
+    return [...reasons];
+}
+function inferredNearMissReasons(context, rawBest, scored) {
+    const reasons = [];
+    const margin = candidateMetricMargin(context.metric, rawBest.metric);
+    if (margin !== null && isCloseToBestMargin(margin, rawBest.metric))
+        reasons.push("close_to_best");
+    if ((context.candidate.robustness?.stability_score ?? 0) >= 0.8)
+        reasons.push("stability");
+    if ((context.candidate.near_miss?.weak_dimensions.length ?? context.candidate.robustness?.weak_dimensions.length ?? 0) > 0) {
+        reasons.push("weak_dimension_improvement");
+    }
+    if (context.candidate.lineage.strategy_family !== rawBest.candidate.lineage.strategy_family
+        && (scored?.diversity_score ?? inferredDiversityScore(context.candidate, rawBest.candidate.lineage.strategy_family, [context.candidate, rawBest.candidate])) >= 0.5) {
+        reasons.push("novelty");
+    }
+    if (complementaryCandidateIds(context.candidate).length > 0 || highestKnownSimilarity(context.candidate, [context.candidate, rawBest.candidate]) <= 0.5) {
+        reasons.push("complementarity");
+    }
+    const lineageText = [
+        ...context.candidate.lineage.model_lineage,
+        ...context.candidate.lineage.config_lineage,
+        ...(context.candidate.near_miss?.summary ? [context.candidate.near_miss.summary] : []),
+    ].map(normalizeLineageText).join(" ");
+    if (lineageText.includes("stack") || lineageText.includes("ensemble") || lineageText.includes("blend")) {
+        reasons.push("ensemble_potential");
+    }
+    return reasons;
+}
+function isImplicitNearMiss(context, rawBest, scored) {
+    if (context.candidate.disposition !== "retained" && context.candidate.disposition !== "promoted")
+        return false;
+    const margin = candidateMetricMargin(context.metric, rawBest.metric);
+    const close = margin !== null && isCloseToBestMargin(margin, rawBest.metric);
+    const weakDimension = (context.candidate.robustness?.weak_dimensions.length ?? 0) > 0;
+    const distinctFamily = context.candidate.lineage.strategy_family !== rawBest.candidate.lineage.strategy_family
+        && (scored?.diversity_score ?? 0) >= 0.5;
+    return close || weakDimension || distinctFamily;
+}
+function nearMissMarginContext(context, rawBest, nearMiss) {
+    const margin = nearMiss?.margin_to_best ?? candidateMetricMargin(context.metric, rawBest.metric);
+    return margin === null || margin === undefined ? {} : { margin_to_raw_best: Math.round(margin * 1_000_000) / 1_000_000 };
+}
+function candidateMetricMargin(candidateMetric, bestMetric) {
+    if (!candidateMetric || !bestMetric || candidateMetric.direction !== bestMetric.direction)
+        return null;
+    const margin = candidateMetric.direction === "maximize"
+        ? bestMetric.value - candidateMetric.value
+        : candidateMetric.value - bestMetric.value;
+    return Number.isFinite(margin) ? Math.max(0, margin) : null;
+}
+function isCloseToBestMargin(margin, bestMetric) {
+    if (!bestMetric)
+        return false;
+    const tolerance = Math.max(Math.abs(bestMetric.value) * 0.005, 0.001);
+    return margin <= tolerance;
+}
+function complementaryCandidateIds(candidate) {
+    const explicit = candidate.near_miss?.complementary_candidate_ids ?? [];
+    if (explicit.length > 0)
+        return explicit;
+    return candidate.similarity
+        .filter((similarity) => similarity.signal === "metric_correlation" && similarity.similarity <= 0.5)
+        .map((similarity) => similarity.candidate_id);
+}
+function nearMissReasonRank(reasons) {
+    const weights = {
+        weak_dimension_improvement: 5,
+        novelty: 4,
+        complementarity: 4,
+        ensemble_potential: 3,
+        stability: 2,
+        close_to_best: 1,
+    };
+    return reasons.reduce((score, reason) => score + weights[reason], 0);
+}
+function summarizeCandidateLineages(entriesOldestFirst) {
+    const primaryMetric = resolvePrimaryCandidateMetricKey(entriesOldestFirst);
+    const byFamily = new Map();
+    for (const context of extractCandidateEvidenceContexts(entriesOldestFirst, primaryMetric)) {
+        const family = context.candidate.lineage.strategy_family;
+        byFamily.set(family, [...(byFamily.get(family) ?? []), context]);
+    }
+    return [...byFamily.entries()]
+        .map(([strategyFamily, contexts]) => {
+        const sorted = [...contexts].sort(compareCandidateEvidenceContexts);
+        const best = sorted[0];
+        const diversityNotes = new Set();
+        for (const context of contexts) {
+            for (const similarity of context.candidate.similarity) {
+                if (similarity.similarity >= 0.85) {
+                    diversityNotes.add(`${context.candidate.candidate_id} near-duplicate of ${similarity.candidate_id}`);
+                }
+            }
+        }
+        return {
+            strategy_family: strategyFamily,
+            candidate_ids: contexts.map((context) => context.candidate.candidate_id),
+            retained_representative_ids: sorted
+                .filter((context) => context.candidate.disposition === "retained" || context.candidate.disposition === "promoted")
+                .map((context) => context.candidate.candidate_id)
+                .slice(0, 3),
+            promoted_ids: contexts
+                .filter((context) => context.candidate.disposition === "promoted")
+                .map((context) => context.candidate.candidate_id),
+            retired_ids: contexts
+                .filter((context) => context.candidate.disposition === "retired")
+                .map((context) => context.candidate.candidate_id),
+            ...(best ? { best_candidate_id: best.candidate.candidate_id } : {}),
+            ...(best?.metric
+                ? {
+                    best_metric: {
+                        label: best.metric.label,
+                        value: best.metric.value,
+                        direction: best.metric.direction,
+                    },
+                }
+                : {}),
+            diversity_notes: [...diversityNotes].slice(0, 5),
+        };
+    })
+        .sort((a, b) => {
+        const aMetric = a.best_metric;
+        const bMetric = b.best_metric;
+        if (aMetric && bMetric && aMetric.direction === bMetric.direction) {
+            const delta = aMetric.direction === "maximize" ? bMetric.value - aMetric.value : aMetric.value - bMetric.value;
+            if (delta !== 0)
+                return delta;
+        }
+        if (aMetric && !bMetric)
+            return -1;
+        if (!aMetric && bMetric)
+            return 1;
+        return a.strategy_family.localeCompare(b.strategy_family);
+    });
+}
+function extractCandidateEvidenceContexts(entriesOldestFirst, primaryMetric) {
+    const contexts = [];
+    for (const entry of entriesOldestFirst) {
+        for (const candidate of entry.candidates ?? []) {
+            contexts.push({
+                entry_id: entry.id,
+                occurred_at: candidate.produced_at ?? entry.occurred_at,
+                candidate,
+                metric: candidateComparableMetric(candidate, primaryMetric),
+            });
+        }
+    }
+    return contexts;
+}
+function resolvePrimaryCandidateMetricKey(entriesOldestFirst) {
+    const candidates = entriesOldestFirst.flatMap((entry) => entry.candidates ?? []);
+    const byMetric = new Map();
+    candidates.forEach((candidate, candidateIndex) => {
+        const seenForCandidate = new Set();
+        candidate.metrics.forEach((metric, metricIndex) => {
+            if (typeof metric.value !== "number" || !Number.isFinite(metric.value))
+                return;
+            if (metric.direction !== "maximize" && metric.direction !== "minimize")
+                return;
+            const key = { label: metric.label, direction: metric.direction };
+            const mapKey = `${key.label}:${key.direction}`;
+            if (seenForCandidate.has(mapKey))
+                return;
+            seenForCandidate.add(mapKey);
+            const existing = byMetric.get(mapKey);
+            if (!existing) {
+                byMetric.set(mapKey, {
+                    key,
+                    candidate_count: 1,
+                    position_sum: metricIndex,
+                    latest_index: candidateIndex,
+                });
+                return;
+            }
+            existing.candidate_count += 1;
+            existing.position_sum += metricIndex;
+            existing.latest_index = candidateIndex;
+        });
+    });
+    return [...byMetric.values()].sort((a, b) => {
+        const coverageDelta = b.candidate_count - a.candidate_count;
+        if (coverageDelta !== 0)
+            return coverageDelta;
+        const localityDelta = Number(isLocalValidationMetricLabel(b.key.label)) - Number(isLocalValidationMetricLabel(a.key.label));
+        if (localityDelta !== 0)
+            return localityDelta;
+        const positionDelta = a.position_sum / a.candidate_count - b.position_sum / b.candidate_count;
+        if (positionDelta !== 0)
+            return positionDelta;
+        return b.latest_index - a.latest_index;
+    })[0]?.key ?? null;
+}
+function isLocalValidationMetricLabel(label) {
+    const normalized = normalizeLineageText(label);
+    const externalHints = ["public", "private", "leaderboard", "external", "lb", "submission"];
+    for (const hint of externalHints) {
+        if (normalized === hint || normalized.includes(hint)) {
+            return false;
+        }
+    }
+    return true;
+}
+function candidateComparableMetric(candidate, primaryMetric) {
+    for (const metric of candidate.metrics) {
+        if (primaryMetric && (metric.label !== primaryMetric.label || metric.direction !== primaryMetric.direction))
+            continue;
+        if (typeof metric.value !== "number" || !Number.isFinite(metric.value))
+            continue;
+        if (metric.direction !== "maximize" && metric.direction !== "minimize")
+            continue;
+        return {
+            label: metric.label,
+            value: metric.value,
+            direction: metric.direction,
+            confidence: metric.confidence ?? 1,
+        };
+    }
+    return null;
+}
+function compareCandidateEvidenceContexts(a, b) {
+    const metricDelta = compareCandidateMetrics(a.metric, b.metric);
+    if (metricDelta !== 0)
+        return metricDelta;
+    const dispositionDelta = dispositionRank(b.candidate.disposition) - dispositionRank(a.candidate.disposition);
+    if (dispositionDelta !== 0)
+        return dispositionDelta;
+    const confidenceDelta = (b.metric?.confidence ?? 0) - (a.metric?.confidence ?? 0);
+    if (confidenceDelta !== 0)
+        return confidenceDelta;
+    return b.occurred_at.localeCompare(a.occurred_at);
+}
+function compareCandidateMetrics(a, b) {
+    if (a && b && a.direction === b.direction) {
+        const valueDelta = a.direction === "maximize" ? b.value - a.value : a.value - b.value;
+        if (valueDelta !== 0)
+            return valueDelta;
+    }
+    if (a && !b)
+        return -1;
+    if (!a && b)
+        return 1;
+    return 0;
+}
+function dispositionRank(disposition) {
+    if (disposition === "promoted")
+        return 2;
+    if (disposition === "retained")
+        return 1;
+    return 0;
+}
+function mostSimilarSelectedCandidate(candidate, selected) {
+    let best;
+    for (const selectedCandidate of selected) {
+        const similarity = similarityBetweenCandidates(candidate.candidate, selectedCandidate.candidate);
+        if (!similarity)
+            continue;
+        if (!best || similarity.similarity > best.similarity)
+            best = similarity;
+    }
+    return best;
+}
+function similarityBetweenCandidates(candidate, selected) {
+    const direct = candidate.similarity.find((similarity) => similarity.candidate_id === selected.candidate_id);
+    const inverse = selected.similarity.find((similarity) => similarity.candidate_id === candidate.candidate_id);
+    if (direct && inverse)
+        return direct.similarity >= inverse.similarity ? direct : inverse;
+    if (direct)
+        return direct;
+    if (inverse) {
+        return {
+            ...inverse,
+            candidate_id: selected.candidate_id,
+        };
+    }
+    if (candidate.lineage.strategy_family === selected.lineage.strategy_family
+        && candidateLineageFingerprint(candidate) === candidateLineageFingerprint(selected)) {
+        return {
+            candidate_id: selected.candidate_id,
+            similarity: 0.9,
+            signal: "lineage",
+            summary: "candidate shares strategy family and lineage fingerprint",
+        };
+    }
+    return undefined;
+}
+function candidateLineageFingerprint(candidate) {
+    const lineage = candidate.lineage;
+    return [
+        lineage.strategy_family,
+        ...lineage.feature_lineage,
+        ...lineage.model_lineage,
+        ...lineage.config_lineage,
+        ...lineage.postprocess_lineage,
+    ].map(normalizeLineageText).filter(Boolean).join("|");
+}
+function toPortfolioSlot(context) {
+    const candidate = context.candidate;
+    return {
+        candidate_id: candidate.candidate_id,
+        ...(candidate.label ? { label: candidate.label } : {}),
+        strategy_family: candidate.lineage.strategy_family,
+        role: context.role,
+        evidence_entry_id: context.entry_id,
+        occurred_at: context.occurred_at,
+        ...(context.metric ? { metric: context.metric } : {}),
+        ...(candidate.lineage.parent_candidate_id ? { parent_candidate_id: candidate.lineage.parent_candidate_id } : {}),
+        ...(candidate.lineage.source_candidate_id ? { source_candidate_id: candidate.lineage.source_candidate_id } : {}),
+        ...(candidate.lineage.source_strategy_id ? { source_strategy_id: candidate.lineage.source_strategy_id } : {}),
+        disposition: candidate.disposition,
+        ...(candidate.disposition_reason ? { retained_reason: candidate.disposition_reason } : {}),
+        ...(context.similarity_to_selected ? { similarity_to_selected: context.similarity_to_selected } : {}),
+    };
+}
+function summarizeFailedLineages(entriesOldestFirst) {
+    const lineages = new Map();
+    for (const entry of entriesOldestFirst) {
+        if (!isFailedEvidenceEntry(entry))
+            continue;
+        const fingerprintInput = failedLineageFingerprintInput(entry);
+        const normalizedIdentityParts = [
+            normalizeLineageText(fingerprintInput.strategy_family),
+            normalizeLineageText(fingerprintInput.hypothesis),
+            normalizeLineageText(fingerprintInput.primary_dimension),
+            normalizeLineageText(fingerprintInput.task_action),
+        ].filter(Boolean);
+        const normalizedFallbackParts = [normalizeLineageText(fingerprintInput.failure_reason)].filter(Boolean);
+        const fingerprintParts = normalizedIdentityParts.length > 0 ? normalizedIdentityParts : normalizedFallbackParts;
+        if (fingerprintParts.length === 0)
+            continue;
+        const fingerprint = fingerprintParts.join("|");
+        const summary = entry.summary
+            ?? entry.result?.summary
+            ?? entry.verification?.summary
+            ?? entry.result?.error
+            ?? `${entry.kind} failed`;
+        const existing = lineages.get(fingerprint);
+        if (!existing) {
+            lineages.set(fingerprint, {
+                fingerprint,
+                count: 1,
+                first_seen_at: entry.occurred_at,
+                last_seen_at: entry.occurred_at,
+                ...(fingerprintInput.strategy_family ? { strategy_family: fingerprintInput.strategy_family } : {}),
+                ...(fingerprintInput.hypothesis ? { hypothesis: fingerprintInput.hypothesis } : {}),
+                ...(fingerprintInput.primary_dimension ? { primary_dimension: fingerprintInput.primary_dimension } : {}),
+                ...(fingerprintInput.task_action ? { task_action: fingerprintInput.task_action } : {}),
+                ...(fingerprintInput.failure_reason ? { failure_reason: fingerprintInput.failure_reason } : {}),
+                representative_entry_id: entry.id,
+                representative_summary: summary,
+                evidence_entry_ids: [entry.id],
+            });
+            continue;
+        }
+        existing.count += 1;
+        existing.last_seen_at = entry.occurred_at;
+        existing.representative_entry_id = entry.id;
+        existing.representative_summary = summary;
+        existing.evidence_entry_ids = [...existing.evidence_entry_ids, entry.id].slice(-5);
+    }
+    return [...lineages.values()]
+        .sort((a, b) => b.count - a.count || b.last_seen_at.localeCompare(a.last_seen_at))
+        .slice(0, 10);
+}
+function isFailedEvidenceEntry(entry) {
+    return entry.outcome === "failed"
+        || entry.outcome === "regressed"
+        || entry.kind === "failure"
+        || entry.result?.status === "failed"
+        || entry.verification?.verdict === "fail";
+}
+function failedLineageFingerprintInput(entry) {
+    const strategyFamily = entry.strategy ?? entry.task?.action;
+    return {
+        ...(strategyFamily ? { strategy_family: strategyFamily } : {}),
+        ...(entry.hypothesis ? { hypothesis: entry.hypothesis } : {}),
+        ...(entry.task?.primary_dimension ? { primary_dimension: entry.task.primary_dimension } : {}),
+        ...(entry.task?.action ? { task_action: entry.task.action } : {}),
+        ...(entry.result?.error || entry.result?.summary || entry.verification?.summary
+            ? { failure_reason: entry.result?.error ?? entry.result?.summary ?? entry.verification?.summary }
+            : {}),
+    };
+}
+function normalizeLineageText(value) {
+    return value?.normalize("NFKC").toLocaleLowerCase().replace(/[^\p{Letter}\p{Number}]+/gu, " ").trim() ?? "";
+}
+function clamp01(value) {
+    if (!Number.isFinite(value))
+        return 0;
+    return Math.min(1, Math.max(0, value));
+}
+function roundScore(value) {
+    return Math.round(value * 1_000_000) / 1_000_000;
+}
+function chooseBestEvidence(entriesNewestFirst) {
+    const metricBest = chooseBestMetricEvidence(entriesNewestFirst);
+    if (metricBest)
+        return metricBest;
+    return chooseBestFallbackEvidence(entriesNewestFirst);
+}
+function chooseBestFallbackEvidence(entriesNewestFirst) {
+    return entriesNewestFirst.find((entry) => entry.outcome === "improved")
+        ?? entriesNewestFirst.find((entry) => entry.verification?.verdict === "pass")
+        ?? entriesNewestFirst.find((entry) => entry.metrics.length > 0)
+        ?? entriesNewestFirst.find((entry) => entry.kind === "artifact")
+        ?? null;
+}
+function chooseBestMetricEvidence(entriesNewestFirst) {
+    const primaryMetric = resolvePrimaryMetricKey(entriesNewestFirst);
+    if (!primaryMetric)
+        return null;
+    const oldestFirst = [...entriesNewestFirst].reverse();
+    const baseline = findComparableMetric(oldestFirst, primaryMetric)?.value;
+    const candidates = entriesNewestFirst
+        .map((entry) => {
+        const metric = findComparableMetric([entry], primaryMetric);
+        if (!metric)
+            return null;
+        return {
+            entry,
+            metric: metric.metric,
+            value: metric.value,
+            direction: metric.direction,
+            primary_metric: primaryMetric,
+            improvement_strength: baseline === undefined
+                ? 0
+                : metric.direction === "maximize"
+                    ? metric.value - baseline
+                    : baseline - metric.value,
+            confidence: metric.metric.confidence ?? entry.verification?.confidence ?? 1,
+            has_pass_verification: entry.verification?.verdict === "pass",
+            has_artifact: entry.artifacts.length > 0 || entry.kind === "artifact",
+        };
+    })
+        .filter((candidate) => Boolean(candidate));
+    if (candidates.length === 0)
+        return null;
+    return candidates.sort(compareComparableEvidenceMetrics)[0]?.entry ?? null;
+}
+function resolvePrimaryMetricKey(entriesNewestFirst) {
+    for (const entry of entriesNewestFirst) {
+        const metric = findFirstDirectedNumericMetric(entry);
+        if (metric?.direction === "maximize" || metric?.direction === "minimize") {
+            return { label: metric.label, direction: metric.direction };
+        }
+    }
+    return null;
+}
+function findComparableMetric(entries, key) {
+    for (const entry of entries) {
+        for (const metric of entry.metrics) {
+            if (metric.label !== key.label || metric.direction !== key.direction)
+                continue;
+            const comparable = toComparableMetric(metric);
+            if (comparable)
+                return comparable;
+        }
+    }
+    return null;
+}
+function findFirstDirectedNumericMetric(entry) {
+    for (const metric of entry.metrics) {
+        if (toComparableMetric(metric))
+            return metric;
+    }
+    return null;
+}
+function toComparableMetric(metric) {
+    if (typeof metric.value !== "number" || !Number.isFinite(metric.value))
+        return null;
+    if (metric.direction !== "maximize" && metric.direction !== "minimize")
+        return null;
+    return {
+        metric,
+        value: metric.value,
+        direction: metric.direction,
+    };
+}
+function compareComparableEvidenceMetrics(a, b) {
+    const direction = a.direction;
+    const valueDelta = direction === "maximize" ? b.value - a.value : a.value - b.value;
+    if (valueDelta !== 0)
+        return valueDelta;
+    const passDelta = Number(b.has_pass_verification) - Number(a.has_pass_verification);
+    if (passDelta !== 0)
+        return passDelta;
+    const artifactDelta = Number(b.has_artifact) - Number(a.has_artifact);
+    if (artifactDelta !== 0)
+        return artifactDelta;
+    const confidenceDelta = b.confidence - a.confidence;
+    if (confidenceDelta !== 0)
+        return confidenceDelta;
+    const improvementDelta = b.improvement_strength - a.improvement_strength;
+    if (improvementDelta !== 0)
+        return improvementDelta;
+    return b.entry.occurred_at.localeCompare(a.entry.occurred_at);
+}
+//# sourceMappingURL=evidence-ledger.js.map