@tangle-network/agent-eval 0.150.2 → 0.161.0

package/dist/contract/index.js

@@ -1,21 +1,22 @@
 import { s as ValidationError } from "../errors-Dngq5h35.js";
+import { i as hashCanonical } from "../canonical-IL-Bu-14.js";
 import { r as pairedBootstrap } from "../paired-tests-BHIhYVdu.js";
-import { a as summarizeExecution, i as analyzeRuns, n as SelfImproveRunError, r as selfImprove, t as defineAgentEval } from "../define-agent-eval-C8V8sMqP.js";
-import { a as parseRunRecordSafe } from "../run-record-D2lDdSAz.js";
-import { $ as runEval, et as runCampaign, f as runImprovementLoop, q as defaultProductionGate, t as llmJudge, ut as campaignSplitDigest, x as compareOptimizationMethods } from "../llm-judge-BqqMS8t7.js";
+import { a as summarizeExecution, i as analyzeRuns, n as SelfImproveRunError, r as selfImprove, t as defineAgentEval } from "../define-agent-eval-h-s-sI-v.js";
+import { a as parseRunRecordSafe } from "../run-record-BC0ebuRP.js";
+import { E as compareOptimizationMethods, Y as defaultProductionGate, ft as campaignSplitDigest, i as transientDispatchFailure, m as runImprovementLoop, nt as runCampaign, t as llmJudge, tt as runEval } from "../llm-judge-BhasIPFT.js";
 import { i as isModelPriced, n as estimateCost } from "../metrics-Qv-cpptD.js";
 import { i as CostLedger } from "../cost-ledger-B1qx30B4.js";
-import { d as mapConcurrentRange } from "../ledger-core-DTae9rv_.js";
-import { S as inMemoryCampaignStorage, x as fsCampaignStorage } from "../external-optimizer-subprocess-CKNb42oM.js";
+import { d as mapConcurrentRange } from "../ledger-core-BOzlRygb.js";
+import { S as inMemoryCampaignStorage, x as fsCampaignStorage } from "../external-optimizer-subprocess-BIWbHpgD.js";
 import { a as heldoutSignificance, s as decidePairedPromotion } from "../power-preflight-DEw-uC7q.js";
 import { r as makeProposalFinding } from "../types-BI4fT3HN.js";
-import { $ as classifyOtlpSpanRole, et as isOtlpModelCall } from "../kind-factory-DmAa0h3K.js";
-import { n as buildDefaultAnalystRegistry, t as createChatClient } from "../chat-client-Bvmxedyv.js";
+import { G as isOtlpModelCall, W as classifyOtlpSpanRole } from "../kind-factory-DY8FdoXf.js";
+import { n as buildDefaultAnalystRegistry, t as createChatClient } from "../chat-client-DlMlAeYI.js";
 import { LLM_MODEL_ATTR_KEYS, SPAN_KIND_ATTR_KEYS } from "../trace-attributes.js";
 import { t as extractUsage } from "../extract-usage-BrQ8mCLX.js";
 import { n as InMemoryOutcomeStore, t as FileSystemOutcomeStore } from "../outcome-store-ChBKlTd_.js";
-import { i as summarizeTraceErrors, n as recordAggregateMeasurements, r as summarizeExecutionMeasurements, t as readTaskFailureLabels } from "../task-failure-attributes--ZTP3tYO.js";
-import { a as externalTextOptimizationMethod, c as REFERENCE_EQUIVALENCE_INPUT_LIMITS, d as runReferenceEquivalenceJudge, i as composeGate, l as REFERENCE_EQUIVALENCE_JUDGE_VERSION, n as gepaOptimizationMethod, r as heldOutGate, t as skillOptOptimizationMethod, u as createReferenceEquivalenceJudge } from "../skillopt-optimization-method-CPBlTcj5.js";
+import { i as summarizeTraceErrors, n as recordAggregateMeasurements, r as summarizeExecutionMeasurements, t as readTaskFailureLabels } from "../task-failure-attributes-DTl-7-Kw.js";
+import { c as externalTextOptimizationMethod, d as REFERENCE_EQUIVALENCE_INPUT_LIMITS, f as REFERENCE_EQUIVALENCE_JUDGE_VERSION, m as runReferenceEquivalenceJudge, n as gepaOptimizationMethod, o as heldOutGate, p as createReferenceEquivalenceJudge, s as composeGate, t as skillOptOptimizationMethod } from "../skillopt-optimization-method-DbaekMcn.js";
 import { n as paretoPolicy, r as paretoSignificanceGate, t as buildEvidenceVector } from "../promotion-policy-xzA40Evo.js";
 import { createHash } from "node:crypto";
 import { agentCandidateBenchmarkSuiteSchema, agentCandidateBenchmarkTaskSchema, agentCandidateBundleSchema, agentCandidateEvaluationPolicySchema, agentCandidateExperimentSchema, agentImprovementMeasuredComparisonSchema, agentProfileImprovementExperimentSchema, agentProfileImprovementMeasuredComparisonSchema, agentProfileImprovementRunCellSchema, agentProfileImprovementRunReceiptSchema, agentProfileImprovementSuiteInputsSchema, agentProfileImprovementSuiteSchema, agentProfileImprovementTaskSchema, candidateExecutionEvidenceSchema, canonicalCandidateDigest, canonicalCandidateJson, numbersApproximatelyEqual, omitTopLevelDigest } from "@tangle-network/agent-interface";
@@ -900,9 +901,6 @@ function sealAgentProfileImprovementExperiment(material) {
 function verifyAgentProfileImprovementTask(input) {
 	return agentProfileImprovementTaskSchema.parse(input);
 }
-function verifyAgentProfileImprovementSuiteInputs(input) {
-	return agentProfileImprovementSuiteInputsSchema.parse(input);
-}
 function verifyAgentProfileImprovementExperiment(input) {
 	return agentProfileImprovementExperimentSchema.parse(input);
 }
@@ -2309,7 +2307,6 @@ function fromKimiCodeSession(options) {
 function fromPiSession(options) {
 	return fromCodeAgentSession("pi", options);
 }
-const fromPigraphSession = fromPiSession;
 function fromCodeAgentSession(source, options) {
 	const entries = options.entries.filter(isRecord);
 	if (entries.length === 0) return {
@@ -2355,8 +2352,8 @@ function fromCodeAgentSession(source, options) {
 	const configHash = options.configHash ?? hashJson({
 		source,
 		model,
-		sourcePath: options.sourcePath,
-		cwd: cwdFromEntries(entries),
+		sourcePath: options.sourcePath ?? null,
+		cwd: cwdFromEntries(entries) ?? null,
 		entryCount: entries.length
 	});
 	const explicitTerminal = observation.terminal.explicit;
@@ -3034,7 +3031,7 @@ function withSnapshot(model) {
 	return `${trimmed}@observed-local`;
 }
 function hashJson(value) {
-	return hashString(JSON.stringify(value));
+	return hashCanonical(value);
 }
 function hashString(value) {
 	return `sha256:${createHash("sha256").update(value).digest("hex")}`;
@@ -3421,6 +3418,6 @@ function collectNumericAttrs(spans) {
 	return raw;
 }
 //#endregion
-export { FileSystemOutcomeStore, InMemoryOutcomeStore, REFERENCE_EQUIVALENCE_INPUT_LIMITS, REFERENCE_EQUIVALENCE_JUDGE_VERSION, SelfImproveRunError, analyzeRuns, buildDefaultAnalystRegistry, buildEvidenceVector, campaignSplitDigest, compareOptimizationMethods, composeGate, createChatClient, createReferenceEquivalenceJudge, defaultProductionGate, defineAgentEval, diffGenerations, diffRunBaselineToWinner, diffRuns, evalReportingSuite, evaluatePairedMeasurements, externalTextOptimizationMethod, fromClaudeCodeSession, fromCodexSession, fromFeedbackTable, fromKimiCodeSession, fromOpenCodeSession, fromOtelSpans, fromPiSession, fromPigraphSession, fromRunRecordDir, fsCampaignStorage, gepaOptimizationMethod, heldOutGate, inMemoryCampaignStorage, llmJudge, makeProposalFinding, measuredComparisonFromAgentProfileImprovementExperiment, measuredComparisonFromCandidateExperiment, observeCodeAgentSession, paretoPolicy, paretoSignificanceGate, parseAgentTrace, parseCodeAgentJsonl, parseCodeAgentJsonlFile, partitionRunsByAuthoringModel, runAgentProfileImprovementExperiment, runCampaign, runCandidateExperiment, runEval, runImprovementLoop, runReferenceEquivalenceJudge, sealAgentProfileImprovementExperiment, sealAgentProfileImprovementSuite, sealAgentProfileImprovementTask, sealCandidateBenchmarkSuite, sealCandidateBenchmarkTask, sealCandidateExperiment, selfImprove, skillOptOptimizationMethod, streamCodeAgentJsonlFile, summarizeExecution, verifyAgentProfileImprovementExperiment, verifyAgentProfileImprovementExperimentComparison, verifyAgentProfileImprovementSuiteInputs, verifyAgentProfileImprovementTask, verifyCandidateBenchmarkSuite, verifyCandidateBenchmarkSuiteInputs, verifyCandidateBenchmarkTask, verifyCandidateExperiment, verifyCandidateExperimentComparison };
+export { FileSystemOutcomeStore, InMemoryOutcomeStore, REFERENCE_EQUIVALENCE_INPUT_LIMITS, REFERENCE_EQUIVALENCE_JUDGE_VERSION, SelfImproveRunError, analyzeRuns, buildDefaultAnalystRegistry, buildEvidenceVector, campaignSplitDigest, compareOptimizationMethods, composeGate, createChatClient, createReferenceEquivalenceJudge, defaultProductionGate, defineAgentEval, diffGenerations, diffRunBaselineToWinner, diffRuns, evalReportingSuite, evaluatePairedMeasurements, externalTextOptimizationMethod, fromClaudeCodeSession, fromCodexSession, fromFeedbackTable, fromKimiCodeSession, fromOpenCodeSession, fromOtelSpans, fromPiSession, fromRunRecordDir, fsCampaignStorage, gepaOptimizationMethod, heldOutGate, inMemoryCampaignStorage, llmJudge, makeProposalFinding, measuredComparisonFromAgentProfileImprovementExperiment, measuredComparisonFromCandidateExperiment, observeCodeAgentSession, paretoPolicy, paretoSignificanceGate, parseAgentTrace, parseCodeAgentJsonl, parseCodeAgentJsonlFile, partitionRunsByAuthoringModel, runAgentProfileImprovementExperiment, runCampaign, runCandidateExperiment, runEval, runImprovementLoop, runReferenceEquivalenceJudge, sealAgentProfileImprovementExperiment, sealAgentProfileImprovementSuite, sealAgentProfileImprovementTask, sealCandidateBenchmarkSuite, sealCandidateBenchmarkTask, sealCandidateExperiment, selfImprove, skillOptOptimizationMethod, streamCodeAgentJsonlFile, summarizeExecution, transientDispatchFailure, verifyAgentProfileImprovementExperimentComparison, verifyCandidateExperiment, verifyCandidateExperimentComparison };
 
 //# sourceMappingURL=index.js.map