@tangle-network/agent-eval 0.216.0 → 0.216.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +39 -0
- package/dist/adapters/http.d.ts +3 -3
- package/dist/{agent-profile-CIA2Atnt.d.ts → agent-profile-VFcQncbW.d.ts} +5 -5
- package/dist/{agent-profile-CIA2Atnt.d.ts.map → agent-profile-VFcQncbW.d.ts.map} +1 -1
- package/dist/{agent-profile-cell-BYc0bxsL.js → agent-profile-cell-BK8mlqlq.js} +3 -3
- package/dist/{agent-profile-cell-BYc0bxsL.js.map → agent-profile-cell-BK8mlqlq.js.map} +1 -1
- package/dist/{agreement-irr-DsTURxdr.js → agreement-irr-BEWLIF3_.js} +3 -3
- package/dist/{agreement-irr-DsTURxdr.js.map → agreement-irr-BEWLIF3_.js.map} +1 -1
- package/dist/analyst/index.d.ts +14 -14
- package/dist/analyst/index.d.ts.map +1 -1
- package/dist/analyst/index.js +15 -15
- package/dist/{analyze-runs-3xpby3p2.js → analyze-runs-C84irm4i.js} +608 -604
- package/dist/analyze-runs-C84irm4i.js.map +1 -0
- package/dist/{analyze-runs-CSDAh8A7.d.ts → analyze-runs-D68fd-su.d.ts} +17 -15
- package/dist/analyze-runs-D68fd-su.d.ts.map +1 -0
- package/dist/{baseline-Dm8966du.js → baseline-Udi_PLuT.js} +5 -5
- package/dist/{baseline-Dm8966du.js.map → baseline-Udi_PLuT.js.map} +1 -1
- package/dist/{benchmark-BywPeAUo.js → benchmark-CXmVVHH2.js} +2 -2
- package/dist/{benchmark-BywPeAUo.js.map → benchmark-CXmVVHH2.js.map} +1 -1
- package/dist/{benchmark-command-CByl9dxG.js → benchmark-command-By0vh8q7.js} +62 -45
- package/dist/benchmark-command-By0vh8q7.js.map +1 -0
- package/dist/benchmarks/index.d.ts +5 -5
- package/dist/benchmarks/index.js +3 -3
- package/dist/builder-eval/index.js +2 -2
- package/dist/campaign/index.d.ts +7 -7
- package/dist/campaign/index.js +13 -13
- package/dist/{campaign-BQ-Jujuj.js → campaign-BudZ70zo.js} +245 -247
- package/dist/campaign-BudZ70zo.js.map +1 -0
- package/dist/{campaign-evidence-OcUciD-N.js → campaign-evidence-Clrn4hBV.js} +8 -8
- package/dist/campaign-evidence-Clrn4hBV.js.map +1 -0
- package/dist/{canonical-BsZVgBhW.js → canonical-B2RqLofv.js} +2 -2
- package/dist/{canonical-BsZVgBhW.js.map → canonical-B2RqLofv.js.map} +1 -1
- package/dist/{chat-client-DthPOYGd.js → chat-client-l-MVDlDk.js} +2 -2
- package/dist/{chat-client-DthPOYGd.js.map → chat-client-l-MVDlDk.js.map} +1 -1
- package/dist/{chat-json-call-BrKsyspg.js → chat-json-call-puMArmtW.js} +2 -2
- package/dist/{chat-json-call-BrKsyspg.js.map → chat-json-call-puMArmtW.js.map} +1 -1
- package/dist/{claim-BXcZw-jo.js → claim-q4HQhK1G.js} +4 -4
- package/dist/claim-q4HQhK1G.js.map +1 -0
- package/dist/cli.js +9 -9
- package/dist/cli.js.map +1 -1
- package/dist/{cluster-bootstrap-B3JywQQn.js → cluster-bootstrap-qcUWYFl-.js} +3 -3
- package/dist/{cluster-bootstrap-B3JywQQn.js.map → cluster-bootstrap-qcUWYFl-.js.map} +1 -1
- package/dist/contract/index.d.ts +11 -11
- package/dist/contract/index.js +1022 -1078
- package/dist/contract/index.js.map +1 -1
- package/dist/{cost-ledger-D2FnJPBF.js → cost-ledger-BP8V4kUQ.js} +835 -816
- package/dist/cost-ledger-BP8V4kUQ.js.map +1 -0
- package/dist/{cost-ledger-CO7UyEcQ.d.ts → cost-ledger-Jk62OBfF.d.ts} +12 -47
- package/dist/cost-ledger-Jk62OBfF.d.ts.map +1 -0
- package/dist/{counterfactual-DtMWvn7D.js → counterfactual-Dn3veueY.js} +2 -2
- package/dist/{counterfactual-DtMWvn7D.js.map → counterfactual-Dn3veueY.js.map} +1 -1
- package/dist/{default-registry-Cn1LQ2Ve.js → default-registry-COhk9g36.js} +5 -5
- package/dist/{default-registry-Cn1LQ2Ve.js.map → default-registry-COhk9g36.js.map} +1 -1
- package/dist/{default-registry-BrWh62zf.d.ts → default-registry-DzHiGruW.d.ts} +6 -6
- package/dist/{default-registry-BrWh62zf.d.ts.map → default-registry-DzHiGruW.d.ts.map} +1 -1
- package/dist/{define-agent-eval-Ch3MWX6k.js → define-agent-eval-CNTPzIkp.js} +10 -10
- package/dist/define-agent-eval-CNTPzIkp.js.map +1 -0
- package/dist/{define-agent-eval-CEWkX5wZ.d.ts → define-agent-eval-WSck0CZJ.d.ts} +8 -8
- package/dist/{define-agent-eval-CEWkX5wZ.d.ts.map → define-agent-eval-WSck0CZJ.d.ts.map} +1 -1
- package/dist/{descriptive-CnU8Yq29.js → descriptive-DboxPZrG.js} +2 -2
- package/dist/{descriptive-CnU8Yq29.js.map → descriptive-DboxPZrG.js.map} +1 -1
- package/dist/diagnosis/index.d.ts +3 -3
- package/dist/diagnosis/index.js +4 -4
- package/dist/{dspy-rlm-engine-D-KX_e64.js → dspy-rlm-engine-LRCu61m5.js} +3 -3
- package/dist/{dspy-rlm-engine-D-KX_e64.js.map → dspy-rlm-engine-LRCu61m5.js.map} +1 -1
- package/dist/{effect-sizes-DiH8MGOH.js → effect-sizes-BhzEpXMc.js} +2 -2
- package/dist/{effect-sizes-DiH8MGOH.js.map → effect-sizes-BhzEpXMc.js.map} +1 -1
- package/dist/{engine-C_zNWfM8.d.ts → engine-3A_4jOlI.d.ts} +5 -5
- package/dist/{engine-C_zNWfM8.d.ts.map → engine-3A_4jOlI.d.ts.map} +1 -1
- package/dist/{equal-terms-Blx9Zk5I.js → equal-terms-BtdXayCJ.js} +2 -2
- package/dist/{equal-terms-Blx9Zk5I.js.map → equal-terms-BtdXayCJ.js.map} +1 -1
- package/dist/{errors-Dngq5h35.js → errors-DVZiqS7K.js} +6 -2
- package/dist/{errors-Dngq5h35.js.map → errors-DVZiqS7K.js.map} +1 -1
- package/dist/{eval-campaign-4byTtqrf.js → eval-campaign-90jeaCAI.js} +6 -6
- package/dist/{eval-campaign-4byTtqrf.js.map → eval-campaign-90jeaCAI.js.map} +1 -1
- package/dist/{evaluation-Cn2dRZIz.d.ts → evaluation-6Yoi74gM.d.ts} +4 -4
- package/dist/{evaluation-Cn2dRZIz.d.ts.map → evaluation-6Yoi74gM.d.ts.map} +1 -1
- package/dist/evaluation.d.ts +1 -1
- package/dist/evaluation.js +1 -1
- package/dist/{exact-types-CD0lf5Qd.d.ts → exact-types-ClugMJfI.d.ts} +2 -2
- package/dist/{exact-types-CD0lf5Qd.d.ts.map → exact-types-ClugMJfI.d.ts.map} +1 -1
- package/dist/experiment/index.d.ts +404 -3
- package/dist/experiment/index.d.ts.map +1 -1
- package/dist/experiment/index.js +313 -19
- package/dist/experiment/index.js.map +1 -1
- package/dist/{external-optimizer-process-nnsY5D_4.js → external-optimizer-process-BjMGUuds.js} +3 -3
- package/dist/{external-optimizer-process-nnsY5D_4.js.map → external-optimizer-process-BjMGUuds.js.map} +1 -1
- package/dist/{external-optimizer-subprocess-DcerlZSG.js → external-optimizer-subprocess-BqW0hHcI.js} +548 -548
- package/dist/external-optimizer-subprocess-BqW0hHcI.js.map +1 -0
- package/dist/{feedback-trajectory-DkaDj0dj.d.ts → feedback-trajectory-C4EG0nNH.d.ts} +5 -8
- package/dist/feedback-trajectory-C4EG0nNH.d.ts.map +1 -0
- package/dist/fuzz.d.ts +1 -1
- package/dist/fuzz.js +3 -3
- package/dist/hosted/index.js +1 -1
- package/dist/{index-CPJJr98o.d.ts → index-BFRO6chs.d.ts} +3 -2
- package/dist/index-BFRO6chs.d.ts.map +1 -0
- package/dist/{index-Ch9VxCpK.d.ts → index-BO41eGj5.d.ts} +2 -2
- package/dist/{index-Ch9VxCpK.d.ts.map → index-BO41eGj5.d.ts.map} +1 -1
- package/dist/{index-NvGBiO8H.d.ts → index-CyszSg4O.d.ts} +165 -162
- package/dist/index-CyszSg4O.d.ts.map +1 -0
- package/dist/{index-DlH0Y9gI.d.ts → index-DPcngtwt.d.ts} +30 -17
- package/dist/index-DPcngtwt.d.ts.map +1 -0
- package/dist/{index-oQ4aHziK.d.ts → index-DV-jSzxF.d.ts} +8 -8
- package/dist/{index-oQ4aHziK.d.ts.map → index-DV-jSzxF.d.ts.map} +1 -1
- package/dist/index.d.ts +1969 -1871
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +161 -201
- package/dist/index.js.map +1 -1
- package/dist/{integrity-BvnVARVl.js → integrity-BXOsAb3i.js} +2 -2
- package/dist/{integrity-BvnVARVl.js.map → integrity-BXOsAb3i.js.map} +1 -1
- package/dist/{integrity-CCAxGVK6.d.ts → integrity-DAMEDFaC.d.ts} +2 -2
- package/dist/{integrity-CCAxGVK6.d.ts.map → integrity-DAMEDFaC.d.ts.map} +1 -1
- package/dist/{integrity-BV1ZDf42.js → integrity-dLiDe2kD.js} +2 -5
- package/dist/integrity-dLiDe2kD.js.map +1 -0
- package/dist/{internal-CRHwS9iz.js → internal-Epl-YBZW.js} +3 -3
- package/dist/{internal-CRHwS9iz.js.map → internal-Epl-YBZW.js.map} +1 -1
- package/dist/interval-CELpye1n.js +219 -0
- package/dist/interval-CELpye1n.js.map +1 -0
- package/dist/interval-ohuOJurX.d.ts +47 -0
- package/dist/interval-ohuOJurX.d.ts.map +1 -0
- package/dist/{journal-DvKCvjty.js → journal-DfRz9kps.js} +2 -2
- package/dist/{journal-DvKCvjty.js.map → journal-DfRz9kps.js.map} +1 -1
- package/dist/{judge-calibration-N44V04Se.js → judge-calibration-BpCy5XQl.js} +3 -3
- package/dist/{judge-calibration-N44V04Se.js.map → judge-calibration-BpCy5XQl.js.map} +1 -1
- package/dist/{judge-gate-DNcUrRFA.d.ts → judge-gate-CE7CQAgt.d.ts} +3 -3
- package/dist/{judge-gate-DNcUrRFA.d.ts.map → judge-gate-CE7CQAgt.d.ts.map} +1 -1
- package/dist/{judge-gate-B1qYnkpi.js → judge-gate-iY9s4L44.js} +3 -3
- package/dist/judge-gate-iY9s4L44.js.map +1 -0
- package/dist/{kind-factory-CbZ-6tvu.js → kind-factory-BobI5Jy3.js} +4 -4
- package/dist/{kind-factory-CbZ-6tvu.js.map → kind-factory-BobI5Jy3.js.map} +1 -1
- package/dist/ledger-core/index.js +3 -3
- package/dist/{ledger-core-B2hahzrt.js → ledger-core-Gq7yp8lP.js} +4 -4
- package/dist/{ledger-core-B2hahzrt.js.map → ledger-core-Gq7yp8lP.js.map} +1 -1
- package/dist/{llm-client-DQveoxSM.js → llm-client-CyBAoYiX.js} +192 -219
- package/dist/llm-client-CyBAoYiX.js.map +1 -0
- package/dist/{llm-judge-PXV7Xljl.js → llm-judge-CG12SQJq.js} +499 -921
- package/dist/llm-judge-CG12SQJq.js.map +1 -0
- package/dist/matrix/index.d.ts +2 -2
- package/dist/matrix/index.js +1 -1
- package/dist/{matrix-CHMiebFx.d.ts → matrix-DSrSxjjy.d.ts} +4 -4
- package/dist/{matrix-CHMiebFx.d.ts.map → matrix-DSrSxjjy.d.ts.map} +1 -1
- package/dist/{matrix-_BV9-Lk5.js → matrix-Dm0BQW7U.js} +7 -11
- package/dist/matrix-Dm0BQW7U.js.map +1 -0
- package/dist/{paired-tests-jbwPdSz5.js → mean-DHoMiudD.js} +11 -5
- package/dist/mean-DHoMiudD.js.map +1 -0
- package/dist/meta-eval/index.d.ts +3 -3
- package/dist/meta-eval/index.js +9 -9
- package/dist/meta-eval/index.js.map +1 -1
- package/dist/mint-DNU8QPG4.d.ts +81 -0
- package/dist/mint-DNU8QPG4.d.ts.map +1 -0
- package/dist/{mint-Csor1BLJ.js → mint-QaD5cA0u.js} +3 -3
- package/dist/{mint-Csor1BLJ.js.map → mint-QaD5cA0u.js.map} +1 -1
- package/dist/{multiplicity-D9cjmStq.js → multiplicity-DQUFcbW4.js} +2 -2
- package/dist/{multiplicity-D9cjmStq.js.map → multiplicity-DQUFcbW4.js.map} +1 -1
- package/dist/multishot/golden/index.d.ts +1 -1
- package/dist/multishot/golden/index.d.ts.map +1 -1
- package/dist/multishot/golden/index.js +3 -6
- package/dist/multishot/golden/index.js.map +1 -1
- package/dist/multishot/index.d.ts +2 -2
- package/dist/multishot/index.js +1 -1
- package/dist/nonnegative-number-DvdMxg_f.js +9 -0
- package/dist/nonnegative-number-DvdMxg_f.js.map +1 -0
- package/dist/openapi.json +1 -1
- package/dist/{otel-spans-C8YIJWy1.js → otel-spans-B9tUXRHF.js} +3 -3
- package/dist/{otel-spans-C8YIJWy1.js.map → otel-spans-B9tUXRHF.js.map} +1 -1
- package/dist/{paired-arms-DzKn0TfY.js → paired-arms-CwMGJsrS.js} +5 -5
- package/dist/{paired-arms-DzKn0TfY.js.map → paired-arms-CwMGJsrS.js.map} +1 -1
- package/dist/{paired-binary-Q36bhn2R.js → paired-binary-CvpFHfKZ.js} +2 -2
- package/dist/{paired-binary-Q36bhn2R.js.map → paired-binary-CvpFHfKZ.js.map} +1 -1
- package/dist/{paired-promotion-decision-BN6IQTiu.js → paired-promotion-decision-DsAPYGCm.js} +3 -3
- package/dist/{paired-promotion-decision-BN6IQTiu.js.map → paired-promotion-decision-DsAPYGCm.js.map} +1 -1
- package/dist/{search-kernel-BztlERLv.js → pareto-Cm6ndrWN.js} +883 -861
- package/dist/pareto-Cm6ndrWN.js.map +1 -0
- package/dist/pipelines/index.d.ts +2 -2
- package/dist/pipelines/index.js +3 -3
- package/dist/{power-and-mde-fK5Qocwo.js → power-and-mde-Hj2ubb4B.js} +2 -2
- package/dist/{power-and-mde-fK5Qocwo.js.map → power-and-mde-Hj2ubb4B.js.map} +1 -1
- package/dist/predicate-FdKmElA_.js +38 -0
- package/dist/predicate-FdKmElA_.js.map +1 -0
- package/dist/{prime-protocol-DcAK9pC8.js → prime-protocol-CHKgA3a6.js} +2 -2
- package/dist/{prime-protocol-DcAK9pC8.js.map → prime-protocol-CHKgA3a6.js.map} +1 -1
- package/dist/{prime-protocol-CEE4BR9N.d.ts → prime-protocol-HiB9HX7E.d.ts} +3 -3
- package/dist/{prime-protocol-CEE4BR9N.d.ts.map → prime-protocol-HiB9HX7E.d.ts.map} +1 -1
- package/dist/{produced-state-ZCtDBxAE.js → produced-state-BsDmnwxc.js} +8 -8
- package/dist/{produced-state-ZCtDBxAE.js.map → produced-state-BsDmnwxc.js.map} +1 -1
- package/dist/profile-cell.js +1 -1
- package/dist/{release-confidence-BK2QCB5M.d.ts → promotion-gate-BawrdPYr.d.ts} +102 -102
- package/dist/promotion-gate-BawrdPYr.d.ts.map +1 -0
- package/dist/{promotion-policy-CWW35UWq.js → promotion-policy-KO57gbyV.js} +4 -4
- package/dist/{promotion-policy-CWW35UWq.js.map → promotion-policy-KO57gbyV.js.map} +1 -1
- package/dist/{proposal-findings-BZiMTzAF.js → proposal-findings-CY1xKn_a.js} +2 -2
- package/dist/{proposal-findings-BZiMTzAF.js.map → proposal-findings-CY1xKn_a.js.map} +1 -1
- package/dist/{query-B1D1qp2Z.js → query-BUCZCK7F.js} +3 -3
- package/dist/{query-B1D1qp2Z.js.map → query-BUCZCK7F.js.map} +1 -1
- package/dist/{random-Dn5fPWkt.js → random-ByUSnV9w.js} +2 -2
- package/dist/{random-Dn5fPWkt.js.map → random-ByUSnV9w.js.map} +1 -1
- package/dist/{raw-provider-sink-rB2DYxr6.js → raw-provider-sink-DoHz4JNG.js} +2 -2
- package/dist/{raw-provider-sink-rB2DYxr6.js.map → raw-provider-sink-DoHz4JNG.js.map} +1 -1
- package/dist/{redact-Dk1ZdJEm.js → redact-B5xXKqQ3.js} +97 -98
- package/dist/redact-B5xXKqQ3.js.map +1 -0
- package/dist/{redact-Cveoq6fJ.d.ts → redact-mtCCcT8z.d.ts} +4 -20
- package/dist/redact-mtCCcT8z.d.ts.map +1 -0
- package/dist/{registry-JYfPxcVz.d.ts → registry-D7_hm2fK.d.ts} +5 -5
- package/dist/{registry-JYfPxcVz.d.ts.map → registry-D7_hm2fK.d.ts.map} +1 -1
- package/dist/{release-confidence-Dbe9AS_X.js → release-confidence-CpuP65S5.js} +5 -5
- package/dist/{release-confidence-Dbe9AS_X.js.map → release-confidence-CpuP65S5.js.map} +1 -1
- package/dist/{report-command-y8H6RSX8.js → report-command-SC9nqb4M.js} +106 -171
- package/dist/report-command-SC9nqb4M.js.map +1 -0
- package/dist/reporting-html-Kvlc8oMt.js +100 -0
- package/dist/reporting-html-Kvlc8oMt.js.map +1 -0
- package/dist/reporting.d.ts +3 -3
- package/dist/reporting.js +6 -6
- package/dist/{researcher-f6mZhh58.d.ts → researcher-CEcfku-T.d.ts} +5 -5
- package/dist/{researcher-f6mZhh58.d.ts.map → researcher-CEcfku-T.d.ts.map} +1 -1
- package/dist/reward-CpCmAM8L.d.ts +78 -0
- package/dist/reward-CpCmAM8L.d.ts.map +1 -0
- package/dist/{reward-hacking-BFcQgkux.js → reward-hacking-E0d-WVcP.js} +2 -2
- package/dist/{reward-hacking-BFcQgkux.js.map → reward-hacking-E0d-WVcP.js.map} +1 -1
- package/dist/{reward-hacking-DHA7c1l8.d.ts → reward-hacking-RVdZ6Kkt.d.ts} +2 -2
- package/dist/{reward-hacking-DHA7c1l8.d.ts.map → reward-hacking-RVdZ6Kkt.d.ts.map} +1 -1
- package/dist/rl.d.ts +8 -8
- package/dist/rl.js +14 -14
- package/dist/rollout/index.d.ts +3 -2
- package/dist/rollout/index.js +2 -2
- package/dist/{rollout-D9bKKBnZ.js → rollout-DZV1t2Wg.js} +2 -2
- package/dist/{rollout-D9bKKBnZ.js.map → rollout-DZV1t2Wg.js.map} +1 -1
- package/dist/round-decimal-COrKglMh.js +10 -0
- package/dist/round-decimal-COrKglMh.js.map +1 -0
- package/dist/{rubric-predictive-validity-2SgHxoa5.d.ts → rubric-predictive-validity-C03EC7ZE.d.ts} +2 -2
- package/dist/{rubric-predictive-validity-2SgHxoa5.d.ts.map → rubric-predictive-validity-C03EC7ZE.d.ts.map} +1 -1
- package/dist/{rubric-predictive-validity-BsgkBfT0.js → rubric-predictive-validity-DaFW70y9.js} +3 -3
- package/dist/{rubric-predictive-validity-BsgkBfT0.js.map → rubric-predictive-validity-DaFW70y9.js.map} +1 -1
- package/dist/{run-record-DlTZ_lqe.js → run-record-Boc5IUTp.js} +4 -4
- package/dist/{run-record-DlTZ_lqe.js.map → run-record-Boc5IUTp.js.map} +1 -1
- package/dist/{run-record-D1Hw-Jml.d.ts → run-record-C1E24G2A.d.ts} +2 -2
- package/dist/{run-record-D1Hw-Jml.d.ts.map → run-record-C1E24G2A.d.ts.map} +1 -1
- package/dist/{run-record-ES07wfz9.js → run-record-C5e7FjhJ.js} +3 -5
- package/dist/run-record-C5e7FjhJ.js.map +1 -0
- package/dist/search/index.d.ts +46 -32
- package/dist/search/index.d.ts.map +1 -1
- package/dist/search/index.js +3 -3
- package/dist/{search-kernel-CpcKH5Ih.d.ts → search-kernel-DyNEa6IU.d.ts} +14 -10
- package/dist/search-kernel-DyNEa6IU.d.ts.map +1 -0
- package/dist/{search-ledger-d7V8o9Lk.js → search-ledger-CVUGLDZp.js} +189 -199
- package/dist/search-ledger-CVUGLDZp.js.map +1 -0
- package/dist/{search-ledger-file-BqOo-bVo.js → search-ledger-file-DDQUgvCg.js} +4 -4
- package/dist/{search-ledger-file-BqOo-bVo.js.map → search-ledger-file-DDQUgvCg.js.map} +1 -1
- package/dist/search-ledger-types-p2xKfiDJ.d.ts.map +1 -1
- package/dist/{search-shipper-er4pK5_Y.js → search-shipper-BOSD46qv.js} +3 -3
- package/dist/{search-shipper-er4pK5_Y.js.map → search-shipper-BOSD46qv.js.map} +1 -1
- package/dist/{search-summary-CBvz3u_S.js → search-summary-BQCSWf1g.js} +5 -5
- package/dist/{search-summary-CBvz3u_S.js.map → search-summary-BQCSWf1g.js.map} +1 -1
- package/dist/{semantic-concept-judge-zQBG9hHU.js → semantic-concept-judge-BQF8te0T.js} +12 -14
- package/dist/semantic-concept-judge-BQF8te0T.js.map +1 -0
- package/dist/{sequential-D-1PPCmM.js → sequential-Bf_pbE50.js} +5 -5
- package/dist/{sequential-D-1PPCmM.js.map → sequential-Bf_pbE50.js.map} +1 -1
- package/dist/{sequential-eprocess-D1jKoihe.js → sequential-eprocess-DM_qENCl.js} +2 -2
- package/dist/{sequential-eprocess-D1jKoihe.js.map → sequential-eprocess-DM_qENCl.js.map} +1 -1
- package/dist/{server-DbcC7xMl.js → server-e9oWO_bN.js} +4 -4
- package/dist/{server-DbcC7xMl.js.map → server-e9oWO_bN.js.map} +1 -1
- package/dist/{single-run-lock-fVu3MZWH.js → single-run-lock-CV9Eyd_q.js} +2 -2
- package/dist/{single-run-lock-fVu3MZWH.js.map → single-run-lock-CV9Eyd_q.js.map} +1 -1
- package/dist/{skillopt-optimization-method-CCtzralq.js → skillopt-optimization-method-BiwA8nfX.js} +427 -12
- package/dist/skillopt-optimization-method-BiwA8nfX.js.map +1 -0
- package/dist/{spans-D9_ehFuk.d.ts → spans-BHAUXMGN.d.ts} +2 -2
- package/dist/{spans-D9_ehFuk.d.ts.map → spans-BHAUXMGN.d.ts.map} +1 -1
- package/dist/{statistical-heldout-1Gbmex7v.js → statistical-heldout-42NlCNDe.js} +3 -3
- package/dist/{statistical-heldout-1Gbmex7v.js.map → statistical-heldout-42NlCNDe.js.map} +1 -1
- package/dist/{statistical-heldout-BeR4bOdg.d.ts → statistical-heldout-CYpGaoDH.d.ts} +3 -3
- package/dist/{statistical-heldout-BeR4bOdg.d.ts.map → statistical-heldout-CYpGaoDH.d.ts.map} +1 -1
- package/dist/{store-otlp-DtB4Slsr.js → store-otlp-DJ9dtA5x.js} +135 -132
- package/dist/store-otlp-DJ9dtA5x.js.map +1 -0
- package/dist/{store-tool-spans-fZIigrpP.js → store-tool-spans-CIX5arAG.js} +7 -7
- package/dist/{store-tool-spans-fZIigrpP.js.map → store-tool-spans-CIX5arAG.js.map} +1 -1
- package/dist/{store-tool-spans-DR1SrCLK.d.ts → store-tool-spans-DdCSPa7C.d.ts} +162 -161
- package/dist/store-tool-spans-DdCSPa7C.d.ts.map +1 -0
- package/dist/storyboard/index.d.ts.map +1 -1
- package/dist/storyboard/index.js +20 -3
- package/dist/storyboard/index.js.map +1 -1
- package/dist/{student-t-BQeKJx2N.js → student-t-DVPW7CdL.js} +2 -2
- package/dist/{student-t-BQeKJx2N.js.map → student-t-DVPW7CdL.js.map} +1 -1
- package/dist/{summary-report-DtdRucM0.js → summary-report-V_JACR4Y.js} +526 -614
- package/dist/summary-report-V_JACR4Y.js.map +1 -0
- package/dist/{summary-report-C-CtItCN.d.ts → summary-report-bdFDXq0b.d.ts} +10 -5
- package/dist/summary-report-bdFDXq0b.d.ts.map +1 -0
- package/dist/supervisor-run/index.d.ts +153 -149
- package/dist/supervisor-run/index.d.ts.map +1 -1
- package/dist/supervisor-run/index.js +65 -3
- package/dist/supervisor-run/index.js.map +1 -1
- package/dist/{systemone-CpTLAlqb.d.ts → systemone-BomhHr2c.d.ts} +6 -6
- package/dist/{systemone-CpTLAlqb.d.ts.map → systemone-BomhHr2c.d.ts.map} +1 -1
- package/dist/systemone-protocol.js +1 -1
- package/dist/{systemone-review-B_yo1xsd.d.ts → systemone-review-CUVcQJu6.d.ts} +2 -2
- package/dist/{systemone-review-B_yo1xsd.d.ts.map → systemone-review-CUVcQJu6.d.ts.map} +1 -1
- package/dist/{systemone-review-BZWfrtq1.js → systemone-review-D7sXwsqq.js} +2 -2
- package/dist/{systemone-review-BZWfrtq1.js.map → systemone-review-D7sXwsqq.js.map} +1 -1
- package/dist/systemone.d.ts +3 -3
- package/dist/systemone.js +4 -4
- package/dist/{task-failure-attributes-DngaLdE4.js → task-failure-attributes-hEmASjix.js} +2 -2
- package/dist/{task-failure-attributes-DngaLdE4.js.map → task-failure-attributes-hEmASjix.js.map} +1 -1
- package/dist/{terminal-record-BHocwFw1.js → terminal-record-DZyQZ3pz.js} +14 -10
- package/dist/terminal-record-DZyQZ3pz.js.map +1 -0
- package/dist/{tool-groups-mcxOXVlb.d.ts → tool-groups-C3NEcdCM.d.ts} +3 -3
- package/dist/tool-groups-C3NEcdCM.d.ts.map +1 -0
- package/dist/{tool-waste-Wi4g6Ekb.d.ts → tool-waste-BKgvd-TP.d.ts} +30 -30
- package/dist/tool-waste-BKgvd-TP.d.ts.map +1 -0
- package/dist/{tool-waste-D52PqU9Q.js → tool-waste-BlesqTqW.js} +4 -4
- package/dist/{tool-waste-D52PqU9Q.js.map → tool-waste-BlesqTqW.js.map} +1 -1
- package/dist/{trace-questions-DG7nqLQ7.d.ts → trace-questions-CM2CAocU.d.ts} +4 -4
- package/dist/{trace-questions-DG7nqLQ7.d.ts.map → trace-questions-CM2CAocU.d.ts.map} +1 -1
- package/dist/trace-repair/index.d.ts +3 -3
- package/dist/trace-repair/index.d.ts.map +1 -1
- package/dist/trace-repair/index.js +7 -13
- package/dist/trace-repair/index.js.map +1 -1
- package/dist/traces.d.ts +76 -58
- package/dist/traces.d.ts.map +1 -1
- package/dist/traces.js +9 -9
- package/dist/trajectory-replay/index.js +3 -3
- package/dist/{tree-rXLW3M_4.js → tree-C1XYeXvy.js} +2277 -2339
- package/dist/tree-C1XYeXvy.js.map +1 -0
- package/dist/{types-DBZ7keha.d.ts → types-C02qtnG8.d.ts} +4 -46
- package/dist/types-C02qtnG8.d.ts.map +1 -0
- package/dist/{types-oKRhQAde.d.ts → types-C1tnkEgM.d.ts} +4 -4
- package/dist/{types-oKRhQAde.d.ts.map → types-C1tnkEgM.d.ts.map} +1 -1
- package/dist/{types-CfZMIzB_.d.ts → types-Cbe2fWBa.d.ts} +3 -3
- package/dist/{types-CfZMIzB_.d.ts.map → types-Cbe2fWBa.d.ts.map} +1 -1
- package/dist/{types-B4C5nFw6.d.ts → types-Cq8mSbq8.d.ts} +78 -72
- package/dist/types-Cq8mSbq8.d.ts.map +1 -0
- package/dist/{verdict-DhPoXApk.js → verdict-Dl0tiT1X.js} +2 -2
- package/dist/{verdict-DhPoXApk.js.map → verdict-Dl0tiT1X.js.map} +1 -1
- package/dist/{verdict-cache-BQEsVy9n.js → verdict-cache-Bm8EKL2s.js} +2 -2
- package/dist/{verdict-cache-BQEsVy9n.js.map → verdict-cache-Bm8EKL2s.js.map} +1 -1
- package/dist/wire/index.d.ts +3 -3
- package/dist/wire/index.js +1 -1
- package/docs/design/mlbenchmarks-book-review.md +1 -1
- package/docs/design/mlbenchmarks-review/probes.mts +23 -31
- package/docs/experiment.md +1 -1
- package/package.json +1 -1
- package/dist/analyze-runs-3xpby3p2.js.map +0 -1
- package/dist/analyze-runs-CSDAh8A7.d.ts.map +0 -1
- package/dist/ast-DJ6Bc7ab.js +0 -558
- package/dist/ast-DJ6Bc7ab.js.map +0 -1
- package/dist/ast-hI-vjW6J.d.ts +0 -457
- package/dist/ast-hI-vjW6J.d.ts.map +0 -1
- package/dist/benchmark-command-CByl9dxG.js.map +0 -1
- package/dist/campaign-BQ-Jujuj.js.map +0 -1
- package/dist/campaign-evidence-OcUciD-N.js.map +0 -1
- package/dist/claim-BXcZw-jo.js.map +0 -1
- package/dist/cost-ledger-CO7UyEcQ.d.ts.map +0 -1
- package/dist/cost-ledger-D2FnJPBF.js.map +0 -1
- package/dist/define-agent-eval-Ch3MWX6k.js.map +0 -1
- package/dist/external-optimizer-subprocess-DcerlZSG.js.map +0 -1
- package/dist/feedback-trajectory-DkaDj0dj.d.ts.map +0 -1
- package/dist/index-CPJJr98o.d.ts.map +0 -1
- package/dist/index-DlH0Y9gI.d.ts.map +0 -1
- package/dist/index-NvGBiO8H.d.ts.map +0 -1
- package/dist/integrity-BV1ZDf42.js.map +0 -1
- package/dist/judge-gate-B1qYnkpi.js.map +0 -1
- package/dist/llm-client-DQveoxSM.js.map +0 -1
- package/dist/llm-judge-PXV7Xljl.js.map +0 -1
- package/dist/matrix-_BV9-Lk5.js.map +0 -1
- package/dist/mint-DCGoBYhB.d.ts +0 -155
- package/dist/mint-DCGoBYhB.d.ts.map +0 -1
- package/dist/paired-tests-jbwPdSz5.js.map +0 -1
- package/dist/redact-Cveoq6fJ.d.ts.map +0 -1
- package/dist/redact-Dk1ZdJEm.js.map +0 -1
- package/dist/release-confidence-BK2QCB5M.d.ts.map +0 -1
- package/dist/report-command-y8H6RSX8.js.map +0 -1
- package/dist/run-record-ES07wfz9.js.map +0 -1
- package/dist/search-kernel-BztlERLv.js.map +0 -1
- package/dist/search-kernel-CpcKH5Ih.d.ts.map +0 -1
- package/dist/search-ledger-d7V8o9Lk.js.map +0 -1
- package/dist/semantic-concept-judge-zQBG9hHU.js.map +0 -1
- package/dist/skillopt-optimization-method-CCtzralq.js.map +0 -1
- package/dist/store-otlp-DtB4Slsr.js.map +0 -1
- package/dist/store-tool-spans-DR1SrCLK.d.ts.map +0 -1
- package/dist/summary-report-C-CtItCN.d.ts.map +0 -1
- package/dist/summary-report-DtdRucM0.js.map +0 -1
- package/dist/terminal-record-BHocwFw1.js.map +0 -1
- package/dist/tool-groups-mcxOXVlb.d.ts.map +0 -1
- package/dist/tool-waste-Wi4g6Ekb.d.ts.map +0 -1
- package/dist/tree-rXLW3M_4.js.map +0 -1
- package/dist/types-B4C5nFw6.d.ts.map +0 -1
- package/dist/types-DBZ7keha.d.ts.map +0 -1
|
@@ -1,5 +1,5 @@
|
|
|
1
|
+
import { a as CostLedgerHandle, c as CostLedgerSummary, p as CustomTokenPricing } from "./cost-ledger-Jk62OBfF.js";
|
|
1
2
|
import { t as AgentEvalError } from "./errors-DEE6u6ot.js";
|
|
2
|
-
import { b as CustomTokenPricing, c as CostLedgerHandle, f as CostLedgerSummary, g as CostReceiptInput, x as MaximumCharge } from "./cost-ledger-CO7UyEcQ.js";
|
|
3
3
|
//#region src/judge-families.d.ts
|
|
4
4
|
/**
|
|
5
5
|
* Judge model-family classification + cross-family enforcement.
|
|
@@ -286,7 +286,7 @@ declare class FileSystemRawProviderSink implements RawProviderSink {
|
|
|
286
286
|
*/
|
|
287
287
|
declare function providerFromBaseUrl(baseUrl: string): string;
|
|
288
288
|
//#endregion
|
|
289
|
-
//#region src/llm-client.d.ts
|
|
289
|
+
//#region src/llm-client/contract.d.ts
|
|
290
290
|
interface LlmMessage {
|
|
291
291
|
role: 'system' | 'user' | 'assistant' | 'tool';
|
|
292
292
|
/**
|
|
@@ -374,7 +374,6 @@ interface LlmChargeBounds {
|
|
|
374
374
|
/** Token rates used when the provider omits cost or package pricing does not cover the model. */
|
|
375
375
|
customTokenPricing?: CustomTokenPricing;
|
|
376
376
|
}
|
|
377
|
-
declare function maximumChargeForLlmRequest(request: Pick<LlmCallRequest, 'model' | 'messages' | 'jsonSchema' | 'tools' | 'toolChoice' | 'maxTokens' | 'thinking'>, options?: LlmChargeBounds): MaximumCharge | undefined;
|
|
378
377
|
interface LlmUsage {
|
|
379
378
|
promptTokens: number;
|
|
380
379
|
completionTokens: number;
|
|
@@ -458,29 +457,6 @@ interface LlmCallResult {
|
|
|
458
457
|
raw: Record<string, unknown>;
|
|
459
458
|
}
|
|
460
459
|
type LlmCallMetadata = Pick<LlmCallResult, 'usage' | 'costUsd' | 'model' | 'durationMs'>;
|
|
461
|
-
/** Convert a provider result into the canonical paid-call receipt input.
|
|
462
|
-
* The receipt is JSON-clean: absent optional fields are omitted, never carried
|
|
463
|
-
* as explicit-undefined keys. Receipts cross JSON boundaries (the external
|
|
464
|
-
* optimizer's loopback proxy validates them with `assertJsonValue`), where an
|
|
465
|
-
* explicit-undefined value is rejected as non-serializable. */
|
|
466
|
-
declare function costReceiptFromLlm(result: LlmCallResult, customTokenPricing?: CustomTokenPricing): CostReceiptInput;
|
|
467
|
-
/** Structured-response failures retain their completed provider receipt. */
|
|
468
|
-
declare function costReceiptFromLlmError(error: Error, customTokenPricing?: CustomTokenPricing): CostReceiptInput | undefined;
|
|
469
|
-
declare class LlmCallError extends AgentEvalError {
|
|
470
|
-
readonly status: number;
|
|
471
|
-
readonly body: string;
|
|
472
|
-
readonly model: string;
|
|
473
|
-
constructor(message: string, status: number, body: string, model: string);
|
|
474
|
-
}
|
|
475
|
-
/** A provider response completed and incurred measurable usage, but its content
|
|
476
|
-
* could not satisfy the caller's response contract. The response envelope is
|
|
477
|
-
* retained so accounting can commit the receipt before the error propagates. */
|
|
478
|
-
declare class LlmResponseError extends AgentEvalError {
|
|
479
|
-
readonly result: LlmCallResult;
|
|
480
|
-
constructor(message: string, result: LlmCallResult, options?: {
|
|
481
|
-
cause?: unknown;
|
|
482
|
-
});
|
|
483
|
-
}
|
|
484
460
|
interface LlmClientOptions extends LlmChargeBounds {
|
|
485
461
|
/** Base URL (without trailing slash), ending at the `/v1` prefix. Required: there is no default endpoint. */
|
|
486
462
|
baseUrl?: string;
|
|
@@ -557,24 +533,6 @@ interface LlmClientOptions extends LlmChargeBounds {
|
|
|
557
533
|
*/
|
|
558
534
|
assertServedModel?: boolean | AssertServedModelOptions;
|
|
559
535
|
}
|
|
560
|
-
/**
|
|
561
|
-
* True when an error is a transient transport/network fault worth retrying,
|
|
562
|
-
* as opposed to a deterministic failure (4xx schema reject, JSON parse) that
|
|
563
|
-
* a retry cannot fix. Inspects `LlmCallError.status`, then the error's
|
|
564
|
-
* name/message/code, then recurses into `error.cause` — undici nests the
|
|
565
|
-
* real socket fault one or more levels under `.cause`.
|
|
566
|
-
*
|
|
567
|
-
* This is the retry classifier for the package: `callLlm` and
|
|
568
|
-
* `withJudgeRetry` both route through it, so connection failures are treated
|
|
569
|
-
* consistently across transports.
|
|
570
|
-
*/
|
|
571
|
-
declare function isTransientLlmError(err: unknown): boolean;
|
|
572
|
-
/**
|
|
573
|
-
* Strip a ```json / ``` code fence if the model emitted one.
|
|
574
|
-
* Idempotent for naked JSON. Some models (claude-code via router, certain
|
|
575
|
-
* deepseek models) wrap output even under json_object.
|
|
576
|
-
*/
|
|
577
|
-
declare function stripFencedJson(raw: string): string;
|
|
578
536
|
//#endregion
|
|
579
537
|
//#region src/analyst/chat-client.d.ts
|
|
580
538
|
/**
|
|
@@ -868,5 +826,5 @@ interface CheckResult {
|
|
|
868
826
|
actual: string;
|
|
869
827
|
}
|
|
870
828
|
//#endregion
|
|
871
|
-
export {
|
|
872
|
-
//# sourceMappingURL=types-
|
|
829
|
+
export { providerFromBaseUrl as $, createChatClient as A, LlmToolDefinition as B, ChatResponse as C, MockTransportOpts as D, CustomTransportOpts as E, LlmMessage as F, InMemoryRawProviderSinkOptions as G, FileSystemRawProviderSink as H, LlmThinkingMode as I, RawProviderDirection as J, NoopRawProviderSink as K, LlmTokenLogprob as L, LlmCallRequest as M, LlmCallResult as N, OpenAiCompatibleTransportOpts as O, LlmChargeBounds as P, defaultProviderRedactor as Q, LlmToolCall as R, ChatRequest as S, CreateChatClientOpts as T, FileSystemRawProviderSinkOptions as U, LlmUsage as V, InMemoryRawProviderSink as W, RawProviderSink as X, RawProviderEvent as Y, RawProviderSinkFilter as Z, ScenarioResult as _, CompletionCriterion as a, ServedModelPolicy as at, ChatCallOpts as b, JudgeInput as c, assertServedModel as ct, PersonaConfig as d, servedModelAcceptable as dt, AssertCrossFamilyServedOptions as et, ProductClientConfig as f, AssertCrossFamilyOptions as ft, ScenarioFile as g, judgeFamily as gt, Scenario as h, assertCrossFamily as ht, CollectedArtifacts as i, ServedModelCheck as it, LlmCallMetadata as j, SandboxSdkTransportOpts as k, JudgeRubric as l, assertServedModels as lt, RubricDimension as m, JudgeFamily as mt, ArtifactResult as n, ModelSubstitutionError as nt, DriverState as o, ServedModelVerdict as ot, RouteMap as p, CrossFamilyError as pt, ProviderRedactor as q, CheckResult as r, ServedCrossFamilyError as rt, JudgeFn as s, assertCrossFamilyServed as st, ArtifactCheck as t, AssertServedModelOptions as tt, JudgeScore as u, checkServedModel as ut, Turn as v, ChatTransport as w, ChatClient as x, TurnResult as y, LlmToolChoice as z };
|
|
830
|
+
//# sourceMappingURL=types-C02qtnG8.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"types-C02qtnG8.d.ts","names":[],"sources":["../src/judge-families.ts","../src/integrity/served-model.ts","../src/trace/raw-provider-sink.ts","../src/llm-client/contract.ts","../src/analyst/chat-client.ts","../src/types.ts"],"mappings":";;;;;;;;;;;;;KAWY;;;;;;;iBAsEI,YAAY,kBAAkB;UAc7B;;EAEf;;;;EAIA;;cAGW,yBAAyB;WAGlB,UAAU;WACV;EAHlB,YACE,iBACgB,UAAU,eACV;;;;;;;;;;;;;;;iBAoBJ,kBACd,kBACA,OAAM,2BACL;;;;KCvFS;;;;;;;;;;;UAYK;;EAEf;;EAEA;EACA,iBAAiB;;EAEjB,cAAc;EACd,SAAS;;EAET;;;;;;;;;;;;iBA4Cc,iBACd,mBACA,oCACC;cA4CU,+BAA+B;WAGxB,QAAQ,cAAc;EAFxC,YACE,iBACgB,QAAQ,cAAc;;;;;;;;;KAc9B;UAWK;;;;;;EAMf;;;;;EAKA;;EAEA;;;;;;;iBAQc,sBACd,OAAO,kBACP,OAAM;;;;;;iBA6BQ,kBACd,mBACA,mCACA,OAAM,2BACL;;;;;iBAea,mBACd,OAAO;EAAgB;EAAmB;IAC1C,OAAM,2BACL;UAec,uCAAuC;;EAEtD;;EAEA;;cAGW,+BAA+B;WAGxB,UAAU;WACV,QAAQ,cAAc;EAHxC,YACE,iBACgB,UAAU,eACV,QAAQ,cAAc;;;;;;;;;;;iBAgB1B,wBACd,OAAO;EAAgB;EAAmB;IAC1C,OAAM,iCACL;;;;;;;;;;;;;;;;;;;;;;;;;;;;;KCjRS;UAEK;;EAEf;;EAEA;EACA;;;;;;EAMA;EACA;;EAEA;;EAEA;;EAEA;EACA,WAAW;;EAEX;;EAEA;EACA;EACA,iBAAiB;EACjB;EACA,kBAAkB;EAClB;;EAEA;;EAEA;;UAGe;EACf;EACA;EACA,YAAY;EACZ;;UAGe;EACf,OAAO,OAAO,mBAAmB;;EAEjC,MAAM,SAAS,wBAAwB,QAAQ;;EAE/C,UAAU;;KAGA,oBAAoB,OAAO,qBAAqB;;;;;iBAM5C,wBAAwB,OAAO,mBAAmB;UAqBjD;EACf,WAAW;;cAGA,mCAAmC;UACtC;UACA;EAER,YAAY,OAAM;EAIZ,OAAO,OAAO,mBAAmB;EAIjC,KAAK,SAAQ,wBAA6B,QAAQ;EAUxD;;cAKW,+BAA+B;EACpC,UAAU;;;;;;;EASV,QAAQ,QAAQ;;UAOP;;EAEf;;EAEA;;EAEA;EACA,WAAW;;cAGA,qCAAqC;UACxC;UACA;UACA;UACA;UACA;UACA;UACA;EAER,YAAY,MAAM;UAOJ;UAON;EAKF,OAAO,OAAO,mBAAmB;EAYjC,KAAK,SAAQ,wBAA6B,QAAQ;;;;;;iBAkC1C,oBAAoB;;;UC5OnB;EACf;;;;;EAKA,kBAEI;IACM;IAAc;;IACd;IAAmB;MAAa;MAAa;;;;EAGvD,YAAY;;EAEZ;;KAGU;;UAGK;EACf;EACA;IACE;IACA;IACA,YAAY;;;;KAKJ;EAIN;EAAkB;IAAY;;;;UAGnB;EACf;EACA;;EAEA;;UAGe;EACf;EACA,UAAU;;EAEV;;EAEA;IAAe;IAAc,QAAQ;;;EAErC,QAAQ;;EAER,aAAa;EACb;EACA;;;;;EAKA;IAAa;;;EAEb,WAAW;;EAEX;;;;;;UAOe;;EAEf;;EAEA;;EAEA,WAAW;;EAEX,qBAAqB;;UAGN;EACf;EACA;EACA;;EAEA;;EAEA;;EAEA;;;;UAKe;EACf;EACA;EACA,KAAK;IAAgB;IAAe;;;UAGrB;;EAEf;;EAEA,YAAY;EACZ,OAAO;;;;;EAKP;;;;;;;;;EASA;;;;;;;;;;;;EAYA;;EAEA;;;;;;;;;EASA;;;;;;;EAOA;;;;;;;;EAQA,WAAW,cAAc;;EAEzB,KAAK;;KAGK,kBAAkB,KAAK;UAElB,yBAAyB;;EAExC;;EAEA;EACA;;EAEA;IAAe;IAAc;;;EAE7B;;EAEA;;;;;;;EAOA;;;;;;;EAOA,SAAS;;;;;;;;EAQT;;;;;;EAMA;;EAEA,eAAe;;;;;;;;EAQf,UAAU;;;;;EAKV;;EAEA;IAAiB;IAAgB;;;EAEjC,WAAW;;;;;;;;;;EAUX,8BAA8B;;;;;;;UCzMf;;WAEN,WAAW;;WAEX;;WAEA;;EAGT,KAAK,KAAK,aAAa,OAAO,eAAe,QAAQ;;KAG3C;UAMK,oBAAoB,KAAK;;EAExC;;KAGU,eAAe;UAEV;;EAEf,SAAS;;EAET;;EAEA;;EAEA;;KAKU,uBACR,0BACA,sBACA,gCACA;UAEM;EACR;;EAEA;;;;;;UAOe,gCAAgC;EAC/C;EACA,OAAO,KAAK,aAAa,OAAO,iBAAiB,QAAQ;;;UAI1C,4BAA4B;EAC3C;EACA,OAAO,KAAK,aAAa,OAAO,iBAAiB,QAAQ;;;;;;;;;;;;;;;;;;;;UAqB1C,sCACP,mBACN,KACE;EAcJ;;EAEA;;EAEA;;EAEA;;EAEA;IAAe;IAAc;;;;;;;UAOd,0BAA0B;EACzC;EACA,UAAU,KAAK,aAAa,OAAO,iBAAiB,QAAQ;;;;;;iBAO9C,iBAAiB,MAAM,uBAAuB;;;UChK7C;EACf;EACA;EACA;EACA;EACA;EACA,OAAO;EACP,gBAAgB;EAChB;;UAGe;EACf;EACA;EACA;EACA;;UAKe;EACf;EAQA;EACA;EACA;EACA;;UAKe;EACf;EACA;EACA,YAAY;;UAGG;EACf;EACA;EACA;EACA;EACA;;UAKe;EACf;EACA;EACA,OAAO;EACP,iBAAiB;EACjB,aAAa;EACb;EACA;EACA;EACA,WAAW;;EAEX,OAAO;;UAGQ;EACf;EACA;EACA;EACA;EACA;IAAmB;IAAc;;EACjC;EACA;;UAGe;EACf,OAAO;EACP;EACA;;UAGe;EACf;EACA;EACA;EACA;EACA;;UAGe;EACf;IAAc;IAAc;;EAC5B;IAAmB;IAAc,QAAQ;;EACzC;IAAc;IAAkB;;EAChC;;UAKe;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;GACC;;UAGc;EACf;EACA,QAAQ;;EAER;;UAKe;EACf;EACA;EACA;EACA;EACA;EACA;EACA;IACE;MACE;MACA;MACA;;;EAGJ,OAAO;EACP,gBAAgB;;UAKD;EACf;EACA,QAAQ,OAAO;EACf,YAAY,OAAO;;;;;;;;;;KAWT;UAEK;EACf;EACA;EACA;EACA,oBAAoB;EACpB;;EAEA,QAAQ;;;;;;;EAOR;;;;;;;EAOA;;;;;;EAMA;;UAGe;EACf;EACA;EACA;IAAa;IAAiB;IAAkB;;EAChD;EACA;EACA;;UAGe;EACf,UAAU;EACV,OAAO;EACP,WAAW;;EAEX,aAAa;EACb;EACA,WAAW;EACX,SAAS;;KAGC,WAAW,MAAM,YAAY,OAAO,eAAe,QAAQ;UAYtD;EACf;EACA;EACA;EACA"}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
|
-
import {
|
|
2
|
-
import { a as RunRecord, d as RunTokenUsage } from "./run-record-
|
|
3
|
-
import { c as JudgeInput, x as ChatClient } from "./types-
|
|
1
|
+
import { a as CostLedgerHandle, l as CostProvenance } from "./cost-ledger-Jk62OBfF.js";
|
|
2
|
+
import { a as RunRecord, d as RunTokenUsage } from "./run-record-C1E24G2A.js";
|
|
3
|
+
import { c as JudgeInput, x as ChatClient } from "./types-C02qtnG8.js";
|
|
4
4
|
import { z } from "zod";
|
|
5
5
|
import { SpanKind } from "@tangle-network/agent-trace-contract";
|
|
6
6
|
import { RE2JS } from "re2js";
|
|
@@ -971,4 +971,4 @@ type AnalystRunEvent = {
|
|
|
971
971
|
};
|
|
972
972
|
//#endregion
|
|
973
973
|
export { DatasetOverview as $, FindingMeasure as A, lineageRoots as B, makeFinding as C, FindingCalibration as D, FINDING_LAYERS as E, FindingScopeReading as F, BoundedTraceAnalysisStoreOptions as G, parseAnalystFinding as H, ProposalFindingSchema as I, SpanSourceReader as J, ReadSpanSourceInput as K, UNKNOWN_LINEAGE as L, FindingParseOutcome as M, FindingReaderLabel as N, FindingEvidenceUnit as O, FindingScope as P, DEFAULT_TRACE_ANALYST_BUDGETS as Q, findingLayerRank as R, computeFindingId as S, AnalystFindingSchema as T, readFindingScope as U, mergeEvidenceUnits as V, createBoundedTraceAnalysisStore as W, TraceAnalysisStore as X, TRACE_ANALYSIS_LIMITS as Y, TraceAnalysisStoreContext as Z, ExecutionProbeOutcome as _, AnalystFinding as a, TRACE_ANALYST_TRUNCATION_MARKER_PREFIX as at, ProposalFinding as b, AnalystRequirements as c, TraceAnalystSpan as ct, AnalystRunResult as d, TraceAnalystTraceSummary as dt, ErrorCluster as et, AnalystRunSummary as f, ViewSpansResult as ft, ExecutionProbe as g, EvidenceRef as h, AnalystCost as i, SpanMatchRecord as it, FindingMethod as j, FindingLayer as k, AnalystRunEvent as l, TraceAnalystSpanKind as lt, AnalystUsageReceipt as m, ViewTraceResult as mt, Analyst as n, SearchSpanResult as nt, AnalystFindingSchemaVersion as o, TraceAnalystByteBudgets as ot, AnalystSeverity as p, ViewTraceOversized as pt, ReadSpanSourceResult as q, AnalystContext as r, SearchTraceResult as rt, AnalystInputKind as s, TraceAnalystFilters as st, ANALYST_FINDING_SCHEMA_VERSIONS as t, QueryTracesPage as tt, AnalystRunInputs as u, TraceAnalystSpanStatus as ut, ExecutionProbeRequest as v, makeProposalFinding as w, ProposalFindingOrigin as x, FINDING_SCOPE_FIELDS as y, independentUnitCount as z };
|
|
974
|
-
//# sourceMappingURL=types-
|
|
974
|
+
//# sourceMappingURL=types-C1tnkEgM.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"types-
|
|
1
|
+
{"version":3,"file":"types-C1tnkEgM.d.ts","names":[],"sources":["../src/trace-analyst/types.ts","../src/trace-analyst/store-contract.ts","../src/trace-analyst/store-boundary.ts","../src/analyst/finding-scope.ts","../src/analyst/types.ts"],"mappings":";;;;;;;;;;;;KAsBY,uBAAuB;KAEvB;;;;UAKK;EACf;EACA;EACA;EACA;EACA,MAAM;EACN;EACA;EACA;EACA,QAAQ;EACR;EACA;EACA;EACA;EACA;;;EAGA,YAAY;;UAGG;EACf;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;;UAGe;;EAEf;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;EAEA;;;EAGA;;;;;;;UAQe;;EAEf;;EAEA;;EAEA;;EAEA;EACA;EACA;;EAEA;;EAEA;;EAEA;;UAGe;EACf;EACA;EACA;EACA;EACA;EACA;;EAEA;EACA;IAAU;IAAqB;;;;;;EAK/B,gBAAgB;EAChB;IAAc;IAAkB;;;UAGjB;EACf,QAAQ;EACR;EACA;;;;;UAMe;EACf;EACA,QAAQ;EACR,YAAY;;UAGG;EACf;;EAEA,gBAAgB;;EAEhB;EACA;;UAGe;EACf;EACA,OAAO;;EAEP;;EAEA;;EAEA;;EAEA;;UAGe;EACf;EACA;EACA;EACA,WAAW;;;EAGX;EACA;EACA;EACA;EACA;;UAGe;EACf;EACA,MAAM;EACN;;UAGe;EACf;EACA;EACA,MAAM;EACN;;;UAIe;;;EAGf;;;EAGA;;;EAGA;;;EAGA;;cAGW,+BAA+B;;;cAS/B;;;cC3MA;WACX;WACA;WACA;WACA;WACA;WACA;WACA;WACA;;UAGe;EACf,SAAS;;;UAIM;EACf;EACA;EACA;EACA;EACA;EACA;;KAGU;EAEN;EACA;EACA;EACA;EACA;EACA;;EAGA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;EACA;IACE;IACA;IACA;IACA;IACA;;;;KAKI,oBACV,OAAO,qBACP,UAAU,8BACP,QAAQ;;;;;;;;UASI;;EAEf,iBAAiB;EAEjB,SAAS,kBAAkB,UAAU,4BAA4B;EAEjE,SACE;IAAS;IAAkB;KAC3B,UAAU,4BACT;EAEH,YACE,UAAU,qBACV,UAAU,4BACT,QAAQ;EAEX,YACE;IAAS,UAAU;IAAqB;IAAe;KACvD,UAAU,4BACT,QAAQ;EAEX,YAAY,UAAU,qBAAqB,UAAU,4BAA4B;EAEjF,UACE;IACE;IACA;KAEF,UAAU,4BACT,QAAQ;EAEX,UACE;IACE;IACA;IACA;KAEF,UAAU,4BACT,QAAQ;EAEX,YACE;IACE;IACA;IACA;KAEF,UAAU,4BACT,QAAQ;EAEX,WACE;IACE;IACA;IACA;IACA;KAEF,UAAU,4BACT,QAAQ;;UAGI;EACf,UAAU,QAAQ;;;;;iBC1GJ,gCACd,QAAQ,oBACR,UAAS,mCACR;;;;;;;;cCCU;KAaD,uBAAuB;;iBAGnB,iBAAiB,OAAO;UAIvB;;;;;;;;EAQf;;EAEA;;EAEA;IAAU;IAAc;;;EAExB;;UAGe;;EAEf;;EAEA;;EAEA;;UAGe;EACf;EACA;EACA;EACA,cAAc;;UAGC;EACf;EACA;EACA;EACA;;EAEA;;;;;;;;UASe;EACf;EACA,OAAO;EACP;;;cAoKW,sBAAoB,EAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAA0C,EAAA,KAAA;;cAG9D,uBAAqB,EAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAGN,EAAA,KAAA;KAQhB;EACN;EAAiB,OAAO;;EACxB;EAAkB;IAAS;IAAc;;;;iBAG/B,oBAAoB,iBAAiB;;;;;UAiBpC;EACf,OAAO;EACP;;EAEA;EACA;EACA,QAAQ;;;;;;;;KASE;iBAEI,iBAAiB,SAAS,iBAAiB;;cA2B9C;;;;;;iBAOG,aAAa,OAAO,cAAc,uBAAuB;;iBA0BzD,qBAAqB,OAAO,cAAc;;iBAK1C,mBAAmB,gBAAgB,wBAAwB;;;;;;;;;;;;;UC5V1D;EACf,gBAAgB;;;;;;;EAOhB;EACA;EACA;EACA,UAAU;;;;;;;EAOV;EACA;EACA;EACA,eAAe;EACf;EACA;;EAEA;;;;;;EAMA;;;;EAIA;;EAEA,WAAW;;EAEX,QAAQ;;EAER,QAAQ;;EAER,SAAS;;EAET,UAAU;;EAEV;;EAEA;;;;;EAKA;;;;;EAKA,iBAAiB;;;cAIN;KAED,sCAAsC;;;cAIrC;KAED,0BAA0B;;KAG1B;;KAGA,kBAAkB;WACnB,iBAAiB;;UAGX;;;;;;;EAOf;EACA;EACA;;;;;;;;KAWU;UAOK;;EAEf;;EAEA;;EAEA;;EAEA;;UAGe;;EAEf;;EAEA;;;;;;UAOe;EACf,aAAa;EACb;EACA,YAAY;EACZ,aAAa;;EAEb,SAAS;;UAGM;EACf;;EAEA;;EAEA;;EAEA;;EAEA,aAAa;;EAEb;;;;;;;EAOA,OAAO;;;;;;;;;;EAUP,gBAAgB,cAAc;;;;;;;EAO9B,mBAAmB,cAAc;;;;;EAKjC,eAAe,SAAS;;EAExB,OAAO;;EAEP,OAAO,aAAa,SAAS;;EAE7B,SAAS;;;;;;;;EAQT,QAAQ;;;UAMO;EACf;;EAEA;;EAEA;;EAEA;EACA,SAAS;;;;;;;KAQC;EAEN;EACA;EACA;EACA;EACA;;EAEA;;EAEA;EAAkB;IAAS;IAAe;;;;;;;;UAO/B;;WAEN;EACT,QAAQ,SAAS,wBAAwB,QAAQ;;;;;;;;UASlC,QAAQ;;WAEd;;WAEA;WACA,WAAW;WACX,MAAM;WACN,WAAW;;WAEX;EACT,QAAQ,OAAO,QAAQ,KAAK,iBAAiB,QAAQ;;;UAItC;;EAEf;;EAEA,QAAQ;;EAER,MAAM;;EAEN;;;;;;;;EAQA;IAAkB;IAAsB;;;;;;;;EAOxC;;;;;;;;;;iBAac,iBAAiB;EAC/B;EACA;EACA;EACA;;EAEA;;;cAsBW;;;;;iBAeG,YACd,MAAM,KAAK;EACT;EACA;IAED;;iBAmBa,oBACd,MAAM,KAAK;EACT;EACA;IAED;UAOc;EACf;EACA;;EAEA;EACA;EACA;;EAEA,OAAO;;EAEP;IAAU;IAAe;;;UAGV;EACf;EACA;EACA;EACA;EACA,UAAU;EACV,aAAa;;EAEb;;;;;EAKA,wBAAwB;;;;;;;;;;;;;;;KAkBd;EAEN;EACA;EACA;EACA;;EAEA,aAAa;;EAGb;EACA,SAAS;;EAGT;EACA;EACA;;EAGA;;EAEA,SAAS;EACT,UAAU,cAAc;;EAGxB;EACA,QAAQ"}
|
|
@@ -1,5 +1,5 @@
|
|
|
1
|
-
import { c as RunSplitTag } from "./run-record-
|
|
2
|
-
import { R as Scenario, d as DispatchContext } from "./types-
|
|
1
|
+
import { c as RunSplitTag } from "./run-record-C1E24G2A.js";
|
|
2
|
+
import { R as Scenario, d as DispatchContext } from "./types-Cq8mSbq8.js";
|
|
3
3
|
//#region src/benchmarks/types.d.ts
|
|
4
4
|
type BenchmarkTaskKind = 'retrieval' | 'rag-answer' | 'hallucination' | 'kb-improvement' | 'routing' | 'custom';
|
|
5
5
|
type BenchmarkFamily = 'beir' | 'mteb-retrieval' | 'msmarco' | 'trec-dl' | 'miracl' | 'lotte' | 'bright' | 'crag' | 'hotpotqa' | 'kilt' | 'ragtruth' | 'faithbench' | 'first-party' | 'custom';
|
|
@@ -99,4 +99,4 @@ declare const BENCHMARK_SPLIT_SEED = "agent-eval-v1";
|
|
|
99
99
|
declare function deterministicSplit(itemId: string, seed?: string): RunSplitTag;
|
|
100
100
|
//#endregion
|
|
101
101
|
export { BenchmarkFamily as a, BenchmarkSource as c, BenchmarkEvaluation as i, BenchmarkTaskKind as l, BenchmarkAdapter as n, BenchmarkResponder as o, BenchmarkDatasetItem as r, BenchmarkScenario as s, BENCHMARK_SPLIT_SEED as t, deterministicSplit as u };
|
|
102
|
-
//# sourceMappingURL=types-
|
|
102
|
+
//# sourceMappingURL=types-Cbe2fWBa.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"types-
|
|
1
|
+
{"version":3,"file":"types-Cbe2fWBa.d.ts","names":[],"sources":["../src/benchmarks/types.ts"],"mappings":";;;KASY;KAQA;UAgBK,qBAAqB;;;EAGpC;;EAEA,SAAS;;EAET,QAAQ;;EAER,SAAS;;EAET,WAAW;;EAEX;;EAEA,SAAS;EACT,WAAW;;UAGI;;;;EAIf;;EAEA;;EAEA,aAAa;;;EAGb,KAAK;EACL;;UAGe;EACf;EACA;EACA;;EAEA;EACA;EACA;;EAEA;;;;;EAKA;;;UAQe,iBAAiB,kBAAkB,oBAAoB;;EAEtE;EACA,SAAS;EACT,WAAW;EACX;EACA,SAAS;EACT;;;;;EAKA,YAAY,OAAO,cAAc,QAAQ,qBAAqB;;EAE9D,SAAS,MAAM,qBAAqB,WAAW,UAAU,YAAY,QAAQ;;;;EAI7E,YAAY,iBAAiB;;UAGd,kBAAkB,4BAA4B;EAC7D;EACA;EACA,QAAQ;EACR,UAAU;EACV,UAAU;EACV,MAAM,qBAAqB;;KAGjB,mBAAmB,oBAAoB,uBAAuB;EACxE,UAAU,kBAAkB;EAC5B,MAAM,qBAAqB;EAC3B,SAAS;MACL,QAAQ,aAAa;;;cAoBd;;;;;;;;;iBAUG,mBACd,gBACA,gBACC"}
|
|
@@ -1,18 +1,20 @@
|
|
|
1
|
-
import {
|
|
2
|
-
import { d as RunTokenUsage, s as RunSearchCoordinates } from "./run-record-
|
|
3
|
-
import { b as ProposalFinding } from "./types-
|
|
4
|
-
import {
|
|
1
|
+
import { a as CostLedgerHandle, c as CostLedgerSummary, h as PaidCallResult, l as CostProvenance, r as CostChannel, v as RunPaidCallInput } from "./cost-ledger-Jk62OBfF.js";
|
|
2
|
+
import { d as RunTokenUsage, s as RunSearchCoordinates } from "./run-record-C1E24G2A.js";
|
|
3
|
+
import { b as ProposalFinding } from "./types-C1tnkEgM.js";
|
|
4
|
+
import { j as LlmCallMetadata } from "./types-C02qtnG8.js";
|
|
5
5
|
import { t as SeriesDistribution } from "./descriptive-B2iPaT9J.js";
|
|
6
6
|
import { A as SearchLedgerHash, G as SearchSplit, I as SearchNodeStatus, L as SearchOpenedEvent, N as SearchNodeDecision, O as SearchLedgerEntry, P as SearchNodeRef, Q as SearchTaskOutcome, R as SearchOperationKind, U as SearchReservation, Z as SearchTask, b as SearchClosedEvent, d as SearchCellCancelledEvent, i as SearchArtifactRef, l as SearchCandidateSurface, o as SearchAudit, p as SearchCellStage, t as NodeEstimate, w as SearchEdgeRecordedEvent, z as SearchOperationRecordedEvent } from "./search-ledger-types-p2xKfiDJ.js";
|
|
7
7
|
import { w as LedgerProjector } from "./index-DHKEAv3s.js";
|
|
8
|
-
import { i as RolloutSearchLineage } from "./mint-
|
|
9
|
-
//#region src/campaign/search-state.d.ts
|
|
8
|
+
import { i as RolloutSearchLineage } from "./mint-DNU8QPG4.js";
|
|
9
|
+
//#region src/campaign/search-state-identity.d.ts
|
|
10
10
|
/** Deterministic node id: identical content in one search is one node. */
|
|
11
11
|
declare function searchNodeId(searchId: string, artifactDigest: string): string;
|
|
12
12
|
/** Deterministic cell id: one node on one task in one split at one repeat. */
|
|
13
13
|
declare function searchCellId(searchId: string, nodeId: string, taskId: string, split: SearchSplit, rep: number): string;
|
|
14
14
|
/** Digest a split's task list in taskId order. */
|
|
15
15
|
declare function searchTaskSetDigest(tasks: readonly SearchTask[]): SearchLedgerHash;
|
|
16
|
+
//#endregion
|
|
17
|
+
//#region src/campaign/search-state-types.d.ts
|
|
16
18
|
interface SearchDecisionRecord {
|
|
17
19
|
sequence: number;
|
|
18
20
|
decision: SearchNodeDecision;
|
|
@@ -42,13 +44,6 @@ interface SearchScoredCell {
|
|
|
42
44
|
attempt: number;
|
|
43
45
|
score: number;
|
|
44
46
|
}
|
|
45
|
-
/**
|
|
46
|
-
* Per-unit means of scored cells, in unitId order. Each unit sums its cells in
|
|
47
|
-
* cellId order, so the result depends on the set of cells alone, never on the
|
|
48
|
-
* order they settled in: a verifier that reads the same cells from a database
|
|
49
|
-
* gets the same bits.
|
|
50
|
-
*/
|
|
51
|
-
declare function searchUnitScores(cells: readonly SearchScoredCell[]): SearchUnitScore[];
|
|
52
47
|
interface SearchNode {
|
|
53
48
|
nodeId: string;
|
|
54
49
|
/** Registration order; the root is 0. */
|
|
@@ -125,6 +120,74 @@ interface SearchCompletion {
|
|
|
125
120
|
complete: boolean;
|
|
126
121
|
reasons: string[];
|
|
127
122
|
}
|
|
123
|
+
//#endregion
|
|
124
|
+
//#region src/campaign/search-state-view.d.ts
|
|
125
|
+
/**
|
|
126
|
+
* Per-unit means of scored cells, in unitId order. Each unit sums its cells in
|
|
127
|
+
* cellId order, so the result depends on the set of cells alone, never on the
|
|
128
|
+
* order they settled in: a verifier that reads the same cells from a database
|
|
129
|
+
* gets the same bits.
|
|
130
|
+
*/
|
|
131
|
+
declare function searchUnitScores(cells: readonly SearchScoredCell[]): SearchUnitScore[];
|
|
132
|
+
/**
|
|
133
|
+
* A read of `SearchState` at one ledger position. Header, head, audit and
|
|
134
|
+
* claim are values. Node, edge, cell and score reads go to the live indexes and
|
|
135
|
+
* throw once the state has applied a later entry.
|
|
136
|
+
*/
|
|
137
|
+
declare class SearchStateView {
|
|
138
|
+
readonly searchId: string;
|
|
139
|
+
readonly header: SearchOpenedEvent | null;
|
|
140
|
+
readonly closed: SearchClosedEvent | null;
|
|
141
|
+
readonly head: {
|
|
142
|
+
sequence: number;
|
|
143
|
+
entryHash: SearchLedgerHash;
|
|
144
|
+
} | null;
|
|
145
|
+
/** `occurredAt` of the head entry; a writer stamps later events at or after it. */
|
|
146
|
+
readonly lastOccurredAt: string | null;
|
|
147
|
+
readonly audit: SearchAudit;
|
|
148
|
+
readonly completion: SearchCompletion;
|
|
149
|
+
/**
|
|
150
|
+
* The admission rule's room: `headroomUsd` is the largest reservation a new
|
|
151
|
+
* non-claim cell or operation may hold (committed spend, open holds and the
|
|
152
|
+
* unspent claim reserve all count against the cap), or null when the search
|
|
153
|
+
* declares no cap. `claimReserveUsd` is what claim cells may still draw.
|
|
154
|
+
*/
|
|
155
|
+
readonly budget: {
|
|
156
|
+
headroomUsd: number | null;
|
|
157
|
+
claimReserveUsd: number;
|
|
158
|
+
};
|
|
159
|
+
private readonly state;
|
|
160
|
+
private readonly version;
|
|
161
|
+
constructor(state: SearchState, version: number);
|
|
162
|
+
/** The node the search started from: the first registered. */
|
|
163
|
+
get rootNodeId(): string | null;
|
|
164
|
+
node(nodeId: string): SearchNode | undefined;
|
|
165
|
+
hasNode(nodeId: string): boolean;
|
|
166
|
+
/** Node ids in registration order. */
|
|
167
|
+
nodeIds(): string[];
|
|
168
|
+
/** Nodes in registration order. */
|
|
169
|
+
nodes(): SearchNode[];
|
|
170
|
+
nodeIdForDigest(artifactDigest: string): string | undefined;
|
|
171
|
+
edge(edgeId: string): SearchEdgeRecordedEvent | undefined;
|
|
172
|
+
/** Edges in ledger order. */
|
|
173
|
+
edges(): SearchEdgeRecordedEvent[];
|
|
174
|
+
cell(cellId: string): SearchCell | undefined;
|
|
175
|
+
/** Cells in allocation order, optionally of one node. */
|
|
176
|
+
cells(filter?: {
|
|
177
|
+
nodeId?: string;
|
|
178
|
+
}): SearchCell[];
|
|
179
|
+
operation(operationId: string): SearchOperation | undefined;
|
|
180
|
+
/** A node's scored cells on one split, in allocation order. Unscored cells
|
|
181
|
+
* (errored, cancelled or in flight) are absent, never zero. */
|
|
182
|
+
scoredCells(nodeId: string, split: SearchSplit): SearchScoredCell[];
|
|
183
|
+
/** Per-unit score means of a node on one split, in unitId order. */
|
|
184
|
+
unitScores(nodeId: string, split: SearchSplit): SearchUnitScore[];
|
|
185
|
+
/** The lineage a RunRecord mint needs for one cell attempt. */
|
|
186
|
+
lineage(coordinates: RunSearchCoordinates): RolloutSearchLineage;
|
|
187
|
+
private live;
|
|
188
|
+
}
|
|
189
|
+
//#endregion
|
|
190
|
+
//#region src/campaign/search-state-projection.d.ts
|
|
128
191
|
/**
|
|
129
192
|
* The pure, incremental projection of one search ledger. `apply` refuses any
|
|
130
193
|
* entry that breaks an invariant; the journal then discards this instance and
|
|
@@ -232,63 +295,6 @@ declare class SearchState implements LedgerProjector<SearchLedgerEntry, SearchSt
|
|
|
232
295
|
/** @internal */
|
|
233
296
|
readScoredCells(nodeId: string, split: SearchSplit): SearchScoredCell[];
|
|
234
297
|
}
|
|
235
|
-
/**
|
|
236
|
-
* A read of `SearchState` at one ledger position. Header, head, audit and
|
|
237
|
-
* claim are values. Node, edge, cell and score reads go to the live indexes and
|
|
238
|
-
* throw once the state has applied a later entry.
|
|
239
|
-
*/
|
|
240
|
-
declare class SearchStateView {
|
|
241
|
-
readonly searchId: string;
|
|
242
|
-
readonly header: SearchOpenedEvent | null;
|
|
243
|
-
readonly closed: SearchClosedEvent | null;
|
|
244
|
-
readonly head: {
|
|
245
|
-
sequence: number;
|
|
246
|
-
entryHash: SearchLedgerHash;
|
|
247
|
-
} | null;
|
|
248
|
-
/** `occurredAt` of the head entry; a writer stamps later events at or after it. */
|
|
249
|
-
readonly lastOccurredAt: string | null;
|
|
250
|
-
readonly audit: SearchAudit;
|
|
251
|
-
readonly completion: SearchCompletion;
|
|
252
|
-
/**
|
|
253
|
-
* The admission rule's room: `headroomUsd` is the largest reservation a new
|
|
254
|
-
* non-claim cell or operation may hold (committed spend, open holds and the
|
|
255
|
-
* unspent claim reserve all count against the cap), or null when the search
|
|
256
|
-
* declares no cap. `claimReserveUsd` is what claim cells may still draw.
|
|
257
|
-
*/
|
|
258
|
-
readonly budget: {
|
|
259
|
-
headroomUsd: number | null;
|
|
260
|
-
claimReserveUsd: number;
|
|
261
|
-
};
|
|
262
|
-
private readonly state;
|
|
263
|
-
private readonly version;
|
|
264
|
-
constructor(state: SearchState, version: number);
|
|
265
|
-
/** The node the search started from: the first registered. */
|
|
266
|
-
get rootNodeId(): string | null;
|
|
267
|
-
node(nodeId: string): SearchNode | undefined;
|
|
268
|
-
hasNode(nodeId: string): boolean;
|
|
269
|
-
/** Node ids in registration order. */
|
|
270
|
-
nodeIds(): string[];
|
|
271
|
-
/** Nodes in registration order. */
|
|
272
|
-
nodes(): SearchNode[];
|
|
273
|
-
nodeIdForDigest(artifactDigest: string): string | undefined;
|
|
274
|
-
edge(edgeId: string): SearchEdgeRecordedEvent | undefined;
|
|
275
|
-
/** Edges in ledger order. */
|
|
276
|
-
edges(): SearchEdgeRecordedEvent[];
|
|
277
|
-
cell(cellId: string): SearchCell | undefined;
|
|
278
|
-
/** Cells in allocation order, optionally of one node. */
|
|
279
|
-
cells(filter?: {
|
|
280
|
-
nodeId?: string;
|
|
281
|
-
}): SearchCell[];
|
|
282
|
-
operation(operationId: string): SearchOperation | undefined;
|
|
283
|
-
/** A node's scored cells on one split, in allocation order. Unscored cells
|
|
284
|
-
* (errored, cancelled or in flight) are absent, never zero. */
|
|
285
|
-
scoredCells(nodeId: string, split: SearchSplit): SearchScoredCell[];
|
|
286
|
-
/** Per-unit score means of a node on one split, in unitId order. */
|
|
287
|
-
unitScores(nodeId: string, split: SearchSplit): SearchUnitScore[];
|
|
288
|
-
/** The lineage a RunRecord mint needs for one cell attempt. */
|
|
289
|
-
lineage(coordinates: RunSearchCoordinates): RolloutSearchLineage;
|
|
290
|
-
private live;
|
|
291
|
-
}
|
|
292
298
|
//#endregion
|
|
293
299
|
//#region src/campaign/search-summary.d.ts
|
|
294
300
|
/** The proposer's own read of a search: its node's cells on the train split,
|
|
@@ -1014,5 +1020,5 @@ interface CampaignResult<TArtifact = unknown, TScenario extends Scenario = Scena
|
|
|
1014
1020
|
scenarios: Array<CampaignScenarioIdentity & Pick<TScenario, 'id' | 'kind'>>;
|
|
1015
1021
|
}
|
|
1016
1022
|
//#endregion
|
|
1017
|
-
export {
|
|
1018
|
-
//# sourceMappingURL=types-
|
|
1023
|
+
export { SearchCell as $, LabeledScenarioWrite as A, ScoredSurfaceOutcome as B, JudgeDimension as C, LabeledScenarioSampleArgs as D, LabeledScenarioRecord as E, ProposeContext as F, labelTrustRank as G, SurfaceProposer as H, ProposedCandidate as I, renderSearchSummary as J, SearchProposerView as K, RedactionStatus as L, OptimizerConfig as M, ParetoParent as N, LabeledScenarioSource as O, ProposalTrackContext as P, searchUnitScores as Q, Scenario as R, JudgeConfig as S, LabelTrust as T, TraceSpan as U, SessionScript as V, isProposedCandidate as W, SearchState as X, searchProposerView as Y, SearchStateView as Z, GateDecision as _, CampaignResult as a, SearchSpend as at, GenerationRecord as b, CampaignTraceWriter as c, searchNodeId as ct, DispatchContext as d, SearchCompletion as et, DispatchFn as f, GateContribution as g, GateContext as h, CampaignCostMeter as i, SearchScoredCell as it, MutableSurface as j, LabeledScenarioStore as k, CodeSurface as l, searchTaskSetDigest as lt, GateCheckStatus as m, CampaignArtifactWriter as n, SearchNode as nt, CampaignScenarioIdentity as o, SearchUnitScore as ot, Gate as p, SearchSummaryOptions as q, CampaignCellResult as r, SearchOperation as rt, CampaignTokenUsage as s, searchCellId as st, CampaignAggregates as t, SearchDecisionRecord as tt, ComponentSurface as u, GateResult as v, JudgeScore as w, JudgeAggregate as x, GenerationCandidate as y, ScenarioAggregate as z };
|
|
1024
|
+
//# sourceMappingURL=types-Cq8mSbq8.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"types-Cq8mSbq8.d.ts","names":[],"sources":["../src/campaign/search-state-identity.ts","../src/campaign/search-state-types.ts","../src/campaign/search-state-view.ts","../src/campaign/search-state-projection.ts","../src/campaign/search-summary.ts","../src/campaign/types.ts"],"mappings":";;;;;;;;;;iBAOgB,aAAa,kBAAkB;;iBAK/B,aACd,kBACA,gBACA,gBACA,OAAO,aACP;;iBAMc,oBAAoB,gBAAgB,eAAe;;;UCHlD;EACf;EACA,UAAU;EACV,OAAO;EACP;EACA;;UAGe;EACf;EACA;EACA;EACA;;;;UAKe;EACf;EACA;EACA;EACA;;;UAIe;EACf;EACA;;EAEA;EACA;;UAGe;EACf;;EAEA;EACA,gBAAgB;EAChB,UAAU;EACV,mBAAmB;;;EAGnB;;EAEA;;;;EAIA,kBAAkB;EAClB;EACA;EACA,QAAQ;EACR;;EAEA;EACA,oBAAoB;EACpB;EACA,OAAO;EACP;EACA;;UAGe;EACf;EACA;EACA;EACA;EACA,OAAO;EACP;EACA,OAAO;EACP;EACA,aAAa;EACb;EACA;EACA,SAAS;;EAET;;EAEA;EACA,WAAW;;EAEX;;EAEA;EACA;EACA;EACA;;UAGe;EACf;EACA,eAAe;EACf,aAAa;EACb;;EAEA,SAAS;;;EAGT,WAAW;;;EAGX,oBAAoB;;EAEpB;;EAEA;;UAGe;EACf;EACA;;;;;;;;;;iBCnGc,iBAAiB,gBAAgB,qBAAqB;;;;;;cAyBzD;WACF;WACA,QAAQ;WACR,QAAQ;WACR;IAAQ;IAAkB,WAAW;;;WAErC;WACA,OAAO;WACP,YAAY;;;;;;;WAOZ;IAAU;IAA4B;;mBAC9B;mBACA;EAEjB,YAAY,OAAO,aAAa;;MAgB5B;EAIJ,KAAK,iBAAiB;EAItB,QAAQ;;EAKR;;EAKA,SAAS;EAKT,gBAAgB;EAIhB,KAAK,iBAAiB;;EAKtB,SAAS;EAKT,KAAK,iBAAiB;;EAKtB,MAAM;IAAU;MAAyB;EAKzC,UAAU,sBAAsB;;;EAMhC,YAAY,gBAAgB,OAAO,cAAc;;EAKjD,WAAW,gBAAgB,OAAO,cAAc;;EAKhD,QAAQ,aAAa,uBAAuB;UAmCpC;;;;;;;;;cCzIG,uBAAuB,gBAAgB,mBAAmB;WAC5D;UACD;UACA;UACA;UACA;UACA;UACA;mBACS;mBACA;mBACA;mBACA;mBACA;mBACA;mBACA;mBACA;UACT;UACA;;mBAES;;UAET;mBACS;EAEjB,YAAY;EAgCZ,MAAM,OAAO,mBAAmB;EA6BhC,YAAY;;EAKZ,KAAK;UAIG;UAgCA;UA6CA;UAsBA;UA8BA;UA8CA;UAsHA;UAmEA;UA4EA;UAoBA;UAmDA;UAaA;;;UA2EA;;EASR;;;UAmBQ;;;UAkBA;;;;;;;;;;;UAoBA;;;EAsBR;IAAgB;IAA4B;;;EAa5C,cAAc;;EAKd,cAAc;;EAKd;IAAc;IAAkB,WAAW;;;EAK3C;;EAKA,aAAa;;EAcb,SAAS,iBAAiB;;EAM1B;;EAKA,YAAY;;EAKZ,oBAAoB;;EAKpB,SAAS,iBAAiB;;EAK1B;;EAKA,SAAS,iBAAiB;;EAM1B,YAAY;;EAMZ,cAAc,sBAAsB;;EAMpC,gBAAgB,gBAAgB,OAAO,cAAc;;;;;;UCj4BtC;WACN;WACA;;;EAGT,YAAY,0BAA0B;;EAEtC,WAAW,0BAA0B;;;iBAIvB,mBAAmB,OAAO,kBAAkB;UAW3C;;;;;;;;EAQf,OAAO;;EAEP;;;;;;;;;;;iBAYc,oBAAoB,OAAO,iBAAiB,SAAS;;;;;UCtCpD;EACf;EACA;EACA;;;;;EAKA;;;UAIe,iCAAiC,KAAK;EACrD;;;;;UAMe;EACf;;EAEA;EACA;EACA;EACA;EACA,QAAQ;EACR,OAAO;EACP,WAAW;EACX,MAAM;;EAEN;;EAEA;;;;;;;;EAQA;;;;KAKU,WAAW,kBAAkB,UAAU,cACjD,UAAU,WACV,KAAK,oBACF,QAAQ;;;;UAOI,cAAc,WAAW;EACxC;EACA;EACA;;EAEA;;;EAGA,sBAAsB,UAAU,WAAW,sBAAsB,UAAU,cAAc;;UAK1E;;EAEf;;EAEA;;;;;;;;;UAUe,YAAY,WAAW,kBAAkB,WAAW;EACnE;EACA,YAAY;;;;EAIZ;;;EAGA,MAAM;IACJ,UAAU;IACV,UAAU;IACV,QAAQ;;IAER,aAAa;IACb;IACA,WAAW;MACT,aAAa,QAAQ;EACzB,aAAa,UAAU;;;;;;;;;;;UAYR;EACf,YAAY;EACZ;EACA;;EAEA,UAAU;;;;;;;EAOV;;;;EAIA,eAAe,eAAe;IAAQ;IAAe;;;;;EAIrD;;;EAGA;;EAEA;;EAEA,WAAW,eAAe;;;;;;;UAUX;WACN;;;WAGA;;WAEA;;WAEA;;WAEA;;WAEA;;WAEA;;;WAGA;aACE;aACA;aACA;;;WAGF;;;UAIM;WACN;WACA,YAAY,SAAS;;;;;;;;;;;KAYpB,0BAA0B,mBAAmB;;;;;;;UAQxC;EACf,SAAS;;EAET;;;;EAIA;;;;;;EAMA,cAAc,SAAS;;;;iBAKT,oBACd,OAAO,iBAAiB,oBACvB,SAAS;;;;;;;;;UAkBK;EACf,SAAS;EACT;;;EAGA,YAAY;;;EAGZ;;EAEA;EACA;EACA;;;UAIe;EACf;EACA;EACA;EACA;EACA;;;;;UAMe;;;EAGf;;EAEA;EACA;EACA;EACA,YAAY;;;;;EAKZ,WAAW;IAAQ;IAAoB;IAAmB;IAAgB;;EAC1E;IACE;IACA;;;;;UAMa,eAAe,YAAY;;;;WAIjC,gBAAgB;;;;;WAKhB;;;;WAIA,UAAU;IAAgB;IAAgB,UAAU;;;;WAGpD,QAAQ;;;;WAIR;;;WAGA,SAAS,cAAc;WACvB,UAAU,cAAc;;WAExB;WACA;WACA,QAAQ;;WAER,QAAQ;;;WAGR,kBAAkB;;;WAGlB,mBAAmB;;;;WAInB,gBAAgB;;;;WAIhB;;;;;;;WAOA,gBAAgB,cAAc;;WAE9B,aAAa;WACb;;;;;;;;;;;;;;UAeM,gBAAgB,YAAY;EAC3C;;;;;EAKA,QAAQ,KAAK,eAAe,aAAa,QAAQ,MAAM,iBAAiB;;;EAGxE,QAAQ;IAAQ,SAAS,cAAc;;IAAwB;IAAe;;;UAG/D;EACf;EACA;EACA,mBAAmB,qBAAqB;;UAGzB,wBAAwB;EACvC,UAAU;;;KAMA;;KAGA;UAEK;EACf;EACA,QAAQ;EACR;;UAGe,YAAY,WAAW,kBAAkB;EACxD,oBAAoB,YAAY;EAChC,oBAAoB,YAAY;;EAEhC,aAAa,YAAY,eAAe;;;;;EAKxC,sBAAsB,YAAY,eAAe;;;;;;;EAOjD,yBAAyB,YAAY,eAAe;;;EAGpD,uBAAuB,YAAY;EACnC,WAAW;EACX;IAAQ;IAAmB;;;EAE3B,aAAa;EACb;EACA,QAAQ;;UAGO;EACf,UAAU;EACV;EACA,mBAAmB;EACnB;;;UAIe,KAAK,qBAAqB,kBAAkB,WAAW;EACtE;EACA,OAAO,KAAK,YAAY,WAAW,aAAa,QAAQ;;;;UAOzC;EACf,KAAK,cAAc,aAAa,0BAA0B;EAC1D,SAAS;;UAGM;EACf,IAAI,aAAa;EACjB,aAAa,aAAa;;;;;UAMX;EACf,MAAM,cAAc,kBAAkB,aAAa;EACnD,UAAU,cAAc,iBAAiB;;;;;;KAO/B,qBAAqB;;;;;UAMhB;;EAEf,YAAY,GACV,OAAO,KAAK,iBAAiB;IAC3B,UAAU;MAEX,QAAQ,eAAe;;;;;KAQhB;KAOA;;;;;;;;;;;;;;;KAgBA;;iBASI,eAAe,OAAO;;;;UAOrB,qBAAqB,kBAAkB,WAAW,UAAU;EAC3E,UAAU;EACV,UAAU;EACV,aAAa,eAAe;EAC5B,QAAQ;EACR;EACA;EACA,iBAAiB;;;;;EAKjB,aAAa;;EAEb;;UAGe,sBAAsB,kBAAkB,WAAW,UAAU,6BACpE,qBAAqB,WAAW;;EAExC;;;EAGA;;UAGe;EACf;;EAEA;;;;EAIA;EACA;IACE;IACA,SAAS,wBAAwB;IACjC;IACA;;;;;IAKA,WAAW;;;UAIE;EACf,QAAQ,OAAO,uBAAuB;EACtC,OAAO,MAAM,4BAA4B,QAAQ;EACjD,QAAQ;IACN;IACA;IACA,UAAU;;;IAGV,SAAS,OAAO;;;UAMH,mBAAmB;;;EAGlC;EACA;;EAEA;IAAY;IAAsB;;EAClC;EACA;EACA;EACA,UAAU;EACV,aAAa,eAAe;;;EAG5B;;EAEA,gBAAgB;;EAEhB;;;EAGA,YAAY;;;EAGZ;;;EAGA;EACA;EACA;EACA;;;;EAIA;;EAEA;;EAEA;EACA;;UAGe;EACf;EACA;EACA;EACA;;;;;;EAMA,cAAc;;UAGC;EACf;EACA;EACA;;;EAGA,cAAc;;UAGC;EACf;EACA,YAAY;EACZ;;;;;;UAOe;EACf;;;;EAIA;;;;EAIA;;;;EAIA;IACE;IACA;IACA,iBAAiB;MAAQ;MAAgB;;;;;EAI3C,YAAY;;;;;;;;;;EAUZ,WAAW;IAAQ;IAAoB;IAAmB;IAAgB;;;;EAG1E;;;;EAIA;;;;EAIA,cAAc,SAAS;;UAGR;EACf,SAAS,eAAe;EACxB,YAAY,eAAe;;EAE3B,MAAM;;EAEN;EACA;EACA;;EAEA;;EAEA;;EAEA;;EAEA;;UAGe,eAAe,qBAAqB,kBAAkB,WAAW;;EAEhF;;EAEA;EACA;;EAEA;EACA;EACA;EACA;EACA,OAAO,MAAM,mBAAmB;EAChC,YAAY;EACZ;IACE,aAAa;IACb;;EAEF,OAAO;EACP;EACA;EACA,iBAAiB;;;EAGjB,WAAW,MAAM,2BAA2B,KAAK"}
|
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
import { n as contentHash } from "./verdict-cache-
|
|
1
|
+
import { n as contentHash } from "./verdict-cache-Bm8EKL2s.js";
|
|
2
2
|
//#region src/verdict.ts
|
|
3
3
|
/**
|
|
4
4
|
* Validator-output verdict — substrate primitive for "did this output pass,
|
|
@@ -58,4 +58,4 @@ function equivalenceVerdict(record) {
|
|
|
58
58
|
//#endregion
|
|
59
59
|
export { equivalenceVerdict as n, certificationEvidenceDigest as t };
|
|
60
60
|
|
|
61
|
-
//# sourceMappingURL=verdict-
|
|
61
|
+
//# sourceMappingURL=verdict-Dl0tiT1X.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"verdict-
|
|
1
|
+
{"version":3,"file":"verdict-Dl0tiT1X.js","names":[],"sources":["../src/verdict.ts"],"sourcesContent":["/**\n * Validator-output verdict — substrate primitive for \"did this output pass,\n * and how well?\"\n *\n * Used by:\n * - `@tangle-network/agent-eval/matrix` — verdict per cell in the cartesian.\n * - `@tangle-network/agent-runtime` — Validator<Output, Verdict = DefaultVerdict>.\n * Runtime keeps `Validator` because it's coupled to runtime-shaped\n * `ValidationCtx` (iteration, signal, traceEmitter); the verdict TYPE\n * itself is a substrate concept and lives here.\n *\n * Repo layering: agent-eval is the substrate (no upward deps). Both\n * agent-runtime and agent-knowledge consume this type FROM agent-eval —\n * never the other way around. See CLAUDE.md \"Repo layering\" for the rule.\n */\n\nimport { contentHash } from './verdict-cache'\nimport type {\n CheckerIdentity,\n EquivalenceRecord,\n VerificationStrategySource,\n} from './verification-strategy'\n\n/**\n * What certified a verdict — the epistemics a bare `valid` + `score` pair\n * cannot carry. A kernel-checked proof and an LLM judge can produce the\n * same `{ valid: true, score: 1 }`; this record is what tells them apart.\n *\n * Every field is required on purpose. A certification without a checker\n * identity or an evidence digest is exactly the unverifiable claim this\n * record exists to kill.\n */\nexport interface VerdictCertification {\n /**\n * Which verification-strategy member produced the certificate. Each\n * member's failure mode is documented on the family type and in\n * `VERIFICATION_STRATEGIES` (src/verification-strategy.ts).\n */\n strategy: VerificationStrategySource\n /**\n * Exact checker identity, e.g. `{ name: 'lean4', version: '4.33.0',\n * pins: { mathlib: 'db584cd6d46c' } }` for a kernel, or a judge model\n * id + snapshot for a judge.\n */\n checker: CheckerIdentity\n /**\n * Every step the certificate rests on that the checker did NOT verify,\n * e.g. 'diagonal embedding argued in prose' or 'agreement arms share\n * training-corpus blind spots'. An empty array is the producer's\n * explicit, auditable claim that there are none — it is not a default.\n */\n assumptions: string[]\n /** Digest (e.g. sha256) of the evidence artifact the checker consumed or produced. */\n evidenceDigest: string\n}\n\n/**\n * Minimal verdict shape — `valid` + `score` are required; `scores` +\n * `notes` are optional surface. Validators that need richer shapes\n * parameterise `Validator<Output, MyVerdict>` with their own type.\n *\n * Need structured extras? Extend DefaultVerdict with typed fields — never\n * serialize extras into `notes`.\n */\nexport interface DefaultVerdict {\n /** Whether the output meets the validator's pass criteria. */\n valid: boolean\n /** Aggregate score in [0, 1]. Drivers use this for winner selection. */\n score: number\n /** Per-dimension scores. Free-form; weighted into `score` by the validator. */\n scores?: Record<string, number>\n /** Human-readable rationale; surfaces in trace + final-result `winner.verdict`. */\n notes?: string\n /**\n * What certified this verdict, when anything did. Absent = uncertified:\n * scored, but no strategy vouches for it. A consumer that cares reads\n * `certification.strategy` to tell a kernel-checked verdict from a\n * judge-scored one; a consumer that does not read it sees the exact\n * verdict it always saw.\n */\n certification?: VerdictCertification\n}\n\n/**\n * Digest a certification's evidence artifact: sha-256 over the canonical\n * JSON of the evidence AS SERIALIZED. The JSON round-trip drops\n * `undefined`-valued fields on purpose — certified verdicts persist as\n * JSON, so the digest must cover exactly what a reader of the persisted\n * artifact can re-digest, nothing the wire cannot carry. Throws when the\n * evidence has no JSON form at all.\n */\nexport function certificationEvidenceDigest(evidence: unknown): string {\n const serialized = JSON.stringify(evidence)\n if (serialized === undefined) {\n throw new Error(\n 'certificationEvidenceDigest: evidence has no JSON form — a certificate over nothing is the unverifiable claim this record exists to kill',\n )\n }\n return contentHash(JSON.parse(serialized))\n}\n\n/**\n * Land a finished statement-equivalence record in the shared verdict\n * spine. `valid` answers the protocol's question — are the two blind\n * statements equivalent — so a refutation is `valid: false` even though a\n * separating witness in hand is a successful protocol outcome.\n *\n * Certification is present exactly when the obligation was discharged\n * (proved or refuted): the checker vouches for the discharge either way.\n * An unresolved obligation certifies nothing and keeps its reason in\n * `notes`.\n */\nexport function equivalenceVerdict(record: EquivalenceRecord): DefaultVerdict {\n const { obligation, spec, arms } = record\n const discharged = obligation.status !== 'unresolved'\n const notes =\n obligation.status === 'proved'\n ? `statements equivalent: ${spec.artifact}`\n : obligation.status === 'refuted-with-separating-witness'\n ? `formalization gap witnessed: ${obligation.separatingWitness}`\n : `obligation undischarged: ${obligation.unresolvedReason}`\n return {\n valid: obligation.status === 'proved',\n score: obligation.status === 'proved' ? 1 : 0,\n notes,\n ...(discharged && obligation.evidenceDigest !== undefined\n ? {\n certification: {\n strategy: spec.source,\n checker: obligation.checker,\n // The arms' blindness declarations are attested by the arms,\n // never verified by the checker — the one step every\n // equivalence certificate rests on.\n assumptions: [\n `arm '${arms[0].armId}' and arm '${arms[1].armId}' self-declare blind derivation`,\n ],\n evidenceDigest: obligation.evidenceDigest,\n } satisfies VerdictCertification,\n }\n : {}),\n }\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;AA2FA,SAAgB,4BAA4B,UAA2B;CACrE,MAAM,aAAa,KAAK,UAAU,QAAQ;CAC1C,IAAI,eAAe,KAAA,GACjB,MAAM,IAAI,MACR,0IACF;CAEF,OAAO,YAAY,KAAK,MAAM,UAAU,CAAC;AAC3C;;;;;;;;;;;;AAaA,SAAgB,mBAAmB,QAA2C;CAC5E,MAAM,EAAE,YAAY,MAAM,SAAS;CACnC,MAAM,aAAa,WAAW,WAAW;CACzC,MAAM,QACJ,WAAW,WAAW,WAClB,0BAA0B,KAAK,aAC/B,WAAW,WAAW,oCACpB,gCAAgC,WAAW,sBAC3C,4BAA4B,WAAW;CAC/C,OAAO;EACL,OAAO,WAAW,WAAW;EAC7B,OAAO,WAAW,WAAW,WAAW,IAAI;EAC5C;EACA,GAAI,cAAc,WAAW,mBAAmB,KAAA,IAC5C,EACE,eAAe;GACb,UAAU,KAAK;GACf,SAAS,WAAW;GAIpB,aAAa,CACX,QAAQ,KAAK,EAAE,CAAC,MAAM,aAAa,KAAK,EAAE,CAAC,MAAM,gCACnD;GACA,gBAAgB,WAAW;EAC7B,EACF,IACA,CAAC;CACP;AACF"}
|
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
import { a as hashCanonical, r as canonicalString } from "./canonical-
|
|
1
|
+
import { a as hashCanonical, r as canonicalString } from "./canonical-B2RqLofv.js";
|
|
2
2
|
import { appendFileSync, existsSync, readFileSync } from "node:fs";
|
|
3
3
|
//#region src/verdict-cache.ts
|
|
4
4
|
/**
|
|
@@ -85,4 +85,4 @@ function fileVerdictCache(path) {
|
|
|
85
85
|
//#endregion
|
|
86
86
|
export { contentHash as n, fileVerdictCache as r, canonicalJson as t };
|
|
87
87
|
|
|
88
|
-
//# sourceMappingURL=verdict-cache-
|
|
88
|
+
//# sourceMappingURL=verdict-cache-Bm8EKL2s.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"verdict-cache-
|
|
1
|
+
{"version":3,"file":"verdict-cache-Bm8EKL2s.js","names":[],"sources":["../src/verdict-cache.ts"],"sourcesContent":["/**\n * Content-addressed judge-verdict caching.\n *\n * LAW: cache JUDGE VERDICTS only — judging the same artifact with the same\n * judge+rubric is pure. NEVER cache agent rollouts. (A router that cached\n * identical fanout prompts silently destroyed best-of-N diversity; rollout\n * caching reintroduces that failure class. Judging has no diversity to\n * destroy — same artifact + same rubric ⇒ same verdict is the desired\n * property, not a bug.)\n *\n * The cache key is `v2:` + a sha-256 over the canonical JSON of everything\n * that can change a verdict: the artifact content, the scenario id, the judge\n * name, the full dimension list (key + description — the description IS the\n * rubric text shown to the judge), and a caller-supplied `judgeVersion`.\n * `judgeVersion` is REQUIRED: a judge whose prompt/model/ensemble changes\n * without a version bump would otherwise silently serve stale verdicts.\n * The `v2:` prefix names the key scheme; a store written under another scheme\n * misses once and repopulates.\n *\n * Canonical JSON is RFC 8785 from `ledger-core/canonical` — the package's one\n * identity scheme. It refuses `undefined`, functions, symbols, non-finite\n * numbers, class instances (including `Date`), `Map`, `Set`, cycles, and\n * unpaired surrogates: an artifact that cannot be unambiguously serialized\n * cannot be content-addressed, and coercing it would let two different\n * artifacts collide on one key. Pass timestamps as ISO strings.\n */\n\nimport { appendFileSync, existsSync, readFileSync } from 'node:fs'\nimport type { JudgeConfig, JudgeScore, Scenario } from './campaign/types'\nimport { canonicalString, hashCanonical } from './ledger-core/canonical'\n\n// ── canonical JSON + content hash ─────────────────────────────────────────\n\n/** Key scheme of every entry `cachedJudge` writes. */\nexport const VERDICT_CACHE_KEY_SCHEME = 'v2'\n\n/**\n * Canonical JSON (RFC 8785): object keys sorted recursively, so two\n * semantically equal values produce byte-identical output regardless of key\n * insertion order. Throws `LedgerCanonicalizationError` on any value with no\n * faithful JSON form — see `ledger-core/canonical`.\n */\nexport function canonicalJson(value: unknown): string {\n return canonicalString(value)\n}\n\n/** Hex sha-256 over `canonicalJson(value)`. The content address used by the\n * verdict cache and report attestation. Equals `hashCanonical(value)` with\n * the `sha256:` prefix removed. */\nexport function contentHash(value: unknown): string {\n return hashCanonical(value).slice('sha256:'.length)\n}\n\n// ── store contract ─────────────────────────────────────────────────────────\n\n/** Pluggable verdict store. Sync or async on both legs — `cachedJudge`\n * awaits the results either way. */\nexport interface VerdictCacheStore {\n get(key: string): Promise<JudgeScore | undefined> | JudgeScore | undefined\n set(key: string, score: JudgeScore): Promise<void> | void\n}\n\n/** Process-local Map-backed store. */\nexport function inMemoryVerdictCache(): VerdictCacheStore {\n const entries = new Map<string, JudgeScore>()\n return {\n get: (key) => entries.get(key),\n set: (key, score) => {\n entries.set(key, score)\n },\n }\n}\n\ninterface VerdictCacheLine {\n key: string\n score: JudgeScore\n}\n\nfunction parseCacheLine(line: string, path: string, lineNo: number): VerdictCacheLine {\n let parsed: unknown\n try {\n parsed = JSON.parse(line)\n } catch (err) {\n throw new Error(\n `fileVerdictCache: corrupt JSONL at ${path}:${lineNo} — ${err instanceof Error ? err.message : String(err)}`,\n )\n }\n const rec = parsed as Partial<VerdictCacheLine>\n if (\n typeof rec !== 'object' ||\n rec === null ||\n typeof rec.key !== 'string' ||\n typeof rec.score !== 'object' ||\n rec.score === null ||\n typeof rec.score.composite !== 'number' ||\n typeof rec.score.dimensions !== 'object'\n ) {\n throw new Error(\n `fileVerdictCache: invalid record shape at ${path}:${lineNo} — expected {key, score:{dimensions, composite, notes}}`,\n )\n }\n return rec as VerdictCacheLine\n}\n\n/**\n * JSONL-file-backed store: the full file is loaded into an in-memory index at\n * construction; every `set` appends one line synchronously (durable before\n * the verdict is returned). A corrupt or malformed line throws at load with\n * file:line — a skipped line would silently re-judge (cost) or, worse, mask\n * a half-written file that needs operator attention.\n */\nexport function fileVerdictCache(path: string): VerdictCacheStore {\n const entries = new Map<string, JudgeScore>()\n if (existsSync(path)) {\n const lines = readFileSync(path, 'utf8').split('\\n')\n for (let i = 0; i < lines.length; i++) {\n const line = lines[i]\n if (line === undefined || line.trim() === '') continue\n const rec = parseCacheLine(line, path, i + 1)\n entries.set(rec.key, rec.score)\n }\n }\n return {\n get: (key) => entries.get(key),\n set: (key, score) => {\n appendFileSync(path, `${JSON.stringify({ key, score })}\\n`, 'utf8')\n entries.set(key, score)\n },\n }\n}\n\n// ── cached judge wrapper ───────────────────────────────────────────────────\n\nexport interface VerdictCacheStats {\n hits: number\n misses: number\n}\n\nexport interface CachedJudgeOptions {\n /** REQUIRED — part of the cache key. Bump on any change to the judge's\n * prompt, model, ensemble, or scoring logic; silent judge upgrades must\n * never serve stale verdicts. */\n judgeVersion: string\n}\n\n/** The wrapped judge: same `JudgeConfig` seam, plus hit/miss observability. */\nexport type CachedJudge<TArtifact, TScenario extends Scenario = Scenario> = JudgeConfig<\n TArtifact,\n TScenario\n> & {\n stats(): VerdictCacheStats\n}\n\n/**\n * Wrap a `JudgeConfig` so repeat judgments of the same artifact are served\n * from the store instead of re-invoking `score()`. The wrapper is generic\n * over the judge's own type parameters and preserves `appliesTo` — it is a\n * drop-in replacement anywhere a `JudgeConfig` is accepted.\n *\n * A judge that throws is NOT cached: the error propagates and the next\n * attempt re-judges (caching a failure would pin a transient outage forever).\n */\nexport function cachedJudge<TArtifact, TScenario extends Scenario = Scenario>(\n judge: JudgeConfig<TArtifact, TScenario>,\n store: VerdictCacheStore,\n options: CachedJudgeOptions,\n): CachedJudge<TArtifact, TScenario> {\n if (typeof options.judgeVersion !== 'string' || options.judgeVersion.trim() === '') {\n throw new Error('cachedJudge: judgeVersion is required and must be a non-empty string')\n }\n const stats: VerdictCacheStats = { hits: 0, misses: 0 }\n const wrapped: CachedJudge<TArtifact, TScenario> = {\n name: judge.name,\n dimensions: judge.dimensions,\n judgeVersion: options.judgeVersion,\n async score(input) {\n const key = `${VERDICT_CACHE_KEY_SCHEME}:${contentHash({\n artifact: canonicalJson(input.artifact),\n scenarioId: input.scenario.id,\n judgeName: judge.name,\n dimensions: judge.dimensions,\n judgeVersion: options.judgeVersion,\n })}`\n const cached = await store.get(key)\n if (cached !== undefined) {\n stats.hits += 1\n return cached\n }\n const score = await judge.score(input)\n await store.set(key, score)\n stats.misses += 1\n return score\n },\n stats: () => ({ ...stats }),\n }\n if (judge.appliesTo) wrapped.appliesTo = judge.appliesTo\n return wrapped\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA0CA,SAAgB,cAAc,OAAwB;CACpD,OAAO,gBAAgB,KAAK;AAC9B;;;;AAKA,SAAgB,YAAY,OAAwB;CAClD,OAAO,cAAc,KAAK,CAAC,CAAC,MAAM,CAAgB;AACpD;AA2BA,SAAS,eAAe,MAAc,MAAc,QAAkC;CACpF,IAAI;CACJ,IAAI;EACF,SAAS,KAAK,MAAM,IAAI;CAC1B,SAAS,KAAK;EACZ,MAAM,IAAI,MACR,sCAAsC,KAAK,GAAG,OAAO,KAAK,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG,GAC3G;CACF;CACA,MAAM,MAAM;CACZ,IACE,OAAO,QAAQ,YACf,QAAQ,QACR,OAAO,IAAI,QAAQ,YACnB,OAAO,IAAI,UAAU,YACrB,IAAI,UAAU,QACd,OAAO,IAAI,MAAM,cAAc,YAC/B,OAAO,IAAI,MAAM,eAAe,UAEhC,MAAM,IAAI,MACR,6CAA6C,KAAK,GAAG,OAAO,wDAC9D;CAEF,OAAO;AACT;;;;;;;;AASA,SAAgB,iBAAiB,MAAiC;CAChE,MAAM,0BAAU,IAAI,IAAwB;CAC5C,IAAI,WAAW,IAAI,GAAG;EACpB,MAAM,QAAQ,aAAa,MAAM,MAAM,CAAC,CAAC,MAAM,IAAI;EACnD,KAAK,IAAI,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK;GACrC,MAAM,OAAO,MAAM;GACnB,IAAI,SAAS,KAAA,KAAa,KAAK,KAAK,MAAM,IAAI;GAC9C,MAAM,MAAM,eAAe,MAAM,MAAM,IAAI,CAAC;GAC5C,QAAQ,IAAI,IAAI,KAAK,IAAI,KAAK;EAChC;CACF;CACA,OAAO;EACL,MAAM,QAAQ,QAAQ,IAAI,GAAG;EAC7B,MAAM,KAAK,UAAU;GACnB,eAAe,MAAM,GAAG,KAAK,UAAU;IAAE;IAAK;GAAM,CAAC,EAAE,KAAK,MAAM;GAClE,QAAQ,IAAI,KAAK,KAAK;EACxB;CACF;AACF"}
|
package/dist/wire/index.d.ts
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
|
-
import {
|
|
2
|
-
import { x as ChatClient } from "../types-
|
|
1
|
+
import { a as CostLedgerHandle, p as CustomTokenPricing } from "../cost-ledger-Jk62OBfF.js";
|
|
2
|
+
import { x as ChatClient } from "../types-C02qtnG8.js";
|
|
3
3
|
import { s as TraceStore } from "../store-C4avFjIi.js";
|
|
4
|
-
import { m as FeedbackTrajectoryStore } from "../feedback-trajectory-
|
|
4
|
+
import { m as FeedbackTrajectoryStore } from "../feedback-trajectory-C4EG0nNH.js";
|
|
5
5
|
import { z } from "zod";
|
|
6
6
|
import { ServerType } from "@hono/node-server";
|
|
7
7
|
import { Hono } from "hono";
|
package/dist/wire/index.js
CHANGED
|
@@ -1,2 +1,2 @@
|
|
|
1
|
-
import { A as WIRE_VERSION, C as RUBRIC_VERSION_SCHEME, D as TracesIngestRequestSchema, E as TraceEventSchema, O as TracesIngestResponseSchema, S as ListRubricsResponseSchema, T as RubricSchema, _ as FeedbackIngestResponseSchema, a as runRpcBatch, b as JudgeRequestSchema, c as WireError, d as handleListRubrics, f as handleTracesIngest, g as ErrorResponseSchema, h as listBuiltinRubrics, i as dispatchRpc, j as hashRubric, k as VersionResponseSchema, l as handleFeedbackIngest, m as getBuiltinRubric, n as startServer, o as runRpcOnce, p as handleVersion, r as startServerAsync, s as buildOpenApi, t as createApp, u as handleJudge, v as FeedbackTrajectorySchema, w as RubricDimensionSchema, x as JudgeResultSchema, y as HealthResponseSchema } from "../server-
|
|
1
|
+
import { A as WIRE_VERSION, C as RUBRIC_VERSION_SCHEME, D as TracesIngestRequestSchema, E as TraceEventSchema, O as TracesIngestResponseSchema, S as ListRubricsResponseSchema, T as RubricSchema, _ as FeedbackIngestResponseSchema, a as runRpcBatch, b as JudgeRequestSchema, c as WireError, d as handleListRubrics, f as handleTracesIngest, g as ErrorResponseSchema, h as listBuiltinRubrics, i as dispatchRpc, j as hashRubric, k as VersionResponseSchema, l as handleFeedbackIngest, m as getBuiltinRubric, n as startServer, o as runRpcOnce, p as handleVersion, r as startServerAsync, s as buildOpenApi, t as createApp, u as handleJudge, v as FeedbackTrajectorySchema, w as RubricDimensionSchema, x as JudgeResultSchema, y as HealthResponseSchema } from "../server-e9oWO_bN.js";
|
|
2
2
|
export { ErrorResponseSchema, FeedbackIngestResponseSchema, FeedbackTrajectorySchema, HealthResponseSchema, JudgeRequestSchema, JudgeResultSchema, ListRubricsResponseSchema, RUBRIC_VERSION_SCHEME, RubricDimensionSchema, RubricSchema, TraceEventSchema, TracesIngestRequestSchema, TracesIngestResponseSchema, VersionResponseSchema, WIRE_VERSION, WireError, buildOpenApi, createApp, dispatchRpc, getBuiltinRubric, handleFeedbackIngest, handleJudge, handleListRubrics, handleTracesIngest, handleVersion, hashRubric, listBuiltinRubrics, runRpcBatch, runRpcOnce, startServer, startServerAsync };
|
|
@@ -111,7 +111,7 @@ These are implementation findings, not an assessment of adoption in every consum
|
|
|
111
111
|
| Concern | Checked implementation | Assessment and remaining gap |
|
|
112
112
|
| --- | --- | --- |
|
|
113
113
|
| Statistical comparisons | [statistics](../../src/statistics/index.ts), [paired decisions](../../src/paired-promotion-decision.ts), [heldout pairing](../../src/campaign/gates/statistical-heldout.ts) | Present. Paired tests, uncertainty, exact binary methods, multiplicity, and power do not need replacement. Claim scope and independent sampling units need stronger binding. |
|
|
114
|
-
| Cluster-aware design | [power](../../src/experiment/power.ts), [registered rule AST](../../src/experiment/ast
|
|
114
|
+
| Cluster-aware design | [power](../../src/experiment/power.ts), [registered rule AST](../../src/experiment/ast/) | Present. The high-level campaign path does not automatically select these methods from a declared generalization target. |
|
|
115
115
|
| Executable preregistration | [define/seal/open](../../src/experiment/define.ts), [acceptance tests](../../tests/experiment/preregistration-acceptance.test.ts) | Present. Extend this rule representation instead of creating a second experiment language. |
|
|
116
116
|
| Sequential testing | [sequential gate](../../src/campaign/gates/sequential.ts), [e-process](../../src/statistics/sequential-eprocess.ts) | Present. Optional stopping within a valid stream differs from adapting hypotheses across streams or counting dependent replicas as independent evidence. |
|
|
117
117
|
| Search/final separation | [selfImprove](../../src/contract/self-improve.ts), [method comparison](../../src/campaign/presets/compare-optimization-methods.ts) | Present within calls. Final cases are withheld from method inputs. A persistent cross-campaign exposure policy is missing from these paths. |
|