@tangle-network/agent-eval 0.53.0 → 0.55.0

package/dist/contract/index.d.ts

@@ -1,20 +1,21 @@
-import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, I as ImprovementDriver, G as Gate } from '../types-Dbj5gu8n.js';
-export { f as CampaignAggregates, g as CampaignArtifactWriter, h as CampaignCellResult, i as CampaignCostMeter, j as CampaignResult, k as CampaignTraceWriter, C as CodeSurface, D as Dispatch, l as GateContext, m as GateDecision, n as GateResult, o as GenerationCandidate, p as GenerationRecord, r as JudgeDimension, J as JudgeScore, t as Mutator, O as OptimizerConfig, v as SessionScript } from '../types-Dbj5gu8n.js';
-import { C as CampaignStorage, e as RunImprovementLoopResult } from '../run-improvement-loop-Cc7oZlRP.js';
-export { D as DefaultProductionGateOptions, E as EvolutionaryDriverOptions, a as GepaDriverOptions, H as HeldOutGateOptions, R as RunCampaignOptions, c as RunEvalOptions, d as RunImprovementLoopOptions, h as composeGate, j as defaultProductionGate, k as evolutionaryDriver, m as fsCampaignStorage, n as gepaDriver, o as heldOutGate, p as inMemoryCampaignStorage, r as runCampaign, s as runEval, t as runImprovementLoop } from '../run-improvement-loop-Cc7oZlRP.js';
+import { S as Scenario, M as MutableSurface, b as DispatchContext, a as JudgeConfig, I as ImprovementDriver, G as Gate } from '../types-BgrxOJSf.js';
+export { g as CampaignAggregates, h as CampaignArtifactWriter, i as CampaignCellResult, j as CampaignCostMeter, k as CampaignResult, l as CampaignTraceWriter, C as CodeSurface, D as Dispatch, m as GateContext, n as GateDecision, o as GateResult, p as GenerationCandidate, q as GenerationRecord, s as JudgeDimension, J as JudgeScore, u as Mutator, O as OptimizerConfig, w as SessionScript } from '../types-BgrxOJSf.js';
+import { C as CampaignStorage, e as RunImprovementLoopResult } from '../run-improvement-loop-BhfdjrMY.js';
+export { D as DefaultProductionGateOptions, E as EvolutionaryDriverOptions, a as GepaDriverOptions, H as HeldOutGateOptions, R as RunCampaignOptions, c as RunEvalOptions, d as RunImprovementLoopOptions, h as composeGate, j as defaultProductionGate, k as evolutionaryDriver, m as fsCampaignStorage, n as gepaDriver, o as heldOutGate, p as inMemoryCampaignStorage, r as runCampaign, s as runEval, t as runImprovementLoop } from '../run-improvement-loop-BhfdjrMY.js';
 export { D as DeploymentOutcome, F as FileSystemOutcomeStore, b as FileSystemOutcomeStoreOptions, I as InMemoryOutcomeStore, O as OutcomeStore } from '../outcome-store-D6KWmYvj.js';
-import { a as HostedTenant, I as InsightReport, T as TraceSpanEvent } from '../index-C7RhhEME.js';
-export { F as FailureClusterInsight, b as InterRaterInsight, J as JudgeInsight, L as LiftInsight, O as OutcomeCorrelationInsight, R as Recommendation, c as ReleaseSummary, S as ScalarDistribution } from '../index-C7RhhEME.js';
-import { A as AnalystRegistry } from '../registry-8KAs18kY.js';
+import { a as HostedTenant, I as InsightReport, T as TraceSpanEvent } from '../index-D2nT6_KT.js';
+export { F as FailureClusterInsight, b as InterRaterInsight, J as JudgeInsight, L as LiftInsight, O as OutcomeCorrelationInsight, R as Recommendation, c as ReleaseSummary, S as ScalarDistribution } from '../index-D2nT6_KT.js';
+import { A as AnalystRegistry } from '../registry-BSWy0rvH.js';
 import { a as DatasetScenario } from '../dataset-BlwAtYYf.js';
-import { R as RunRecord, a as RunSplitTag } from '../run-record-BGY6bHRh.js';
+import { R as RunRecord, a as RunSplitTag } from '../run-record-etiCMsUq.js';
 import '../llm-client-BXVRUZyX.js';
 import '../errors-mje_cKOs.js';
 import '../raw-provider-sink-C46HDghv.js';
-import '../red-team-30II1T4o.js';
-import '../store-Db2Bv8Cf.js';
-import '../summary-report-B7gNRX-r.js';
-import '../failure-cluster-Cw65_5FY.js';
+import '../red-team-CrC5MZYd.js';
+import '../store-CKUAgsJz.js';
+import '../schema-m0gsnbt3.js';
+import '../summary-report-DLxh4yWk.js';
+import '../failure-cluster-CL7IVgkJ.js';
 import '../judge-calibration-DilmB3Ml.js';
 import '../store-CJbzDxZ2.js';
 import '../types-DhqpAi_z.js';

package/dist/contract/index.js

@@ -75,6 +75,7 @@ async function analyzeRuns(opts) {
   const interRater = opts.raterScores ? computeInterRater(opts.raterScores) : void 0;
   const lift = computeLift(runs, opts.baselineCandidateId, opts.candidateCandidateId, split);
   const failureClusters = opts.analyst ? await computeFailureClusters(runs, opts.analyst, split) : void 0;
+  const failureModes = computeFailureModes(runs);
   const contamination = opts.canaryScenarios ? computeContamination(runs, opts.canaryScenarios) : void 0;
   const outcomeCorrelation = opts.outcomeSignal ? computeOutcomeCorrelation(runs, opts.outcomeSignal, split) : void 0;
   const release = buildReleaseScorecard(composite, lift, contamination);
@@ -85,6 +86,7 @@ async function analyzeRuns(opts) {
     interRater,
     lift,
     failureClusters,
+    failureModes,
     contamination,
     outcomeCorrelation,
     priorPeriodComparison,
@@ -102,10 +104,21 @@ async function analyzeRuns(opts) {
     contamination,
     outcomeCorrelation,
     release,
+    ...failureModes ? { failureModes } : {},
     ...priorPeriodComparison ? { priorPeriodComparison } : {},
     recommendations
   };
 }
+function computeFailureModes(runs) {
+  const counts = /* @__PURE__ */ new Map();
+  for (const r of runs) {
+    const key = r.failureClass ?? r.failureMode;
+    if (key) counts.set(key, (counts.get(key) ?? 0) + 1);
+  }
+  if (counts.size === 0) return void 0;
+  const n = runs.length;
+  return [...counts.entries()].map(([mode, count]) => ({ mode, count, share: n > 0 ? count / n : 0 })).sort((a, b) => b.count - a.count || a.mode.localeCompare(b.mode));
+}
 function computePriorPeriodComparison(current, baseline, split, windowLabel) {
   if (current.length === 0 || baseline.length === 0) return void 0;
   const metrics = {};
@@ -661,6 +674,18 @@ function buildRecommendations(ctx) {
       });
     }
   }
+  if (ctx.failureModes && ctx.failureModes.length > 0) {
+    const top = ctx.failureModes[0];
+    if (top.count >= 3 && top.share >= 0.15) {
+      out.push({
+        priority: top.share >= 0.25 ? "high" : "medium",
+        kind: "investigate",
+        title: `'${top.mode}' is the dominant failure mode \u2014 ${top.count} runs (${(top.share * 100).toFixed(0)}% of the corpus)`,
+        detail: `The mean composite can look acceptable while one named failure dominates the lower tail. ${top.count} of ${ctx.composite.n} runs failed with '${top.mode}'${ctx.failureModes.length > 1 ? ` (next: '${ctx.failureModes[1].mode}' \xD7${ctx.failureModes[1].count})` : ""}. Fix this cause first.`,
+        evidencePath: "failureModes"
+      });
+    }
+  }
   if (Object.keys(ctx.judges).length === 0 && ctx.composite.n > 0) {
     out.push({
       priority: "medium",