reeval 0.8.0 → 0.9.0
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/README.md +10 -10
- package/dist/{agent-B72lZ6w5.js → agent-6FefIhX7.js} +5 -5
- package/dist/{agent-B72lZ6w5.js.map → agent-6FefIhX7.js.map} +1 -1
- package/dist/{agent-Cfj1wD4D.js → agent-WSmTaesf.js} +51 -16
- package/dist/agent-WSmTaesf.js.map +1 -0
- package/dist/{analysis-BTNnKh1w.js → analysis-DlGBDjTj.js} +3 -3
- package/dist/{analysis-BTNnKh1w.js.map → analysis-DlGBDjTj.js.map} +1 -1
- package/dist/{assemble-BbFVjHeS.js → assemble-BCZzTyR7.js} +243 -11
- package/dist/assemble-BCZzTyR7.js.map +1 -0
- package/dist/assemble-C5qq5yT2.js +2 -0
- package/dist/{capture-BXhBfmsM.js → capture-Bp74Z6Fr.js} +66 -15
- package/dist/capture-Bp74Z6Fr.js.map +1 -0
- package/dist/capture-DS-cPyuE.js +2 -0
- package/dist/{cli-BMjTKY-D.js → cli-CVrMh_qz.js} +201 -108
- package/dist/cli-CVrMh_qz.js.map +1 -0
- package/dist/{cli-spec-DSvKsOK4.js → cli-spec-CKEYa8V2.js} +3 -3
- package/dist/{cli-spec-DSvKsOK4.js.map → cli-spec-CKEYa8V2.js.map} +1 -1
- package/dist/cli.js +2 -2
- package/dist/{client-CwzU9CiW.js → client-Bbo9SPlK.js} +31 -2
- package/dist/client-Bbo9SPlK.js.map +1 -0
- package/dist/{command-BdstSEHu.js → command-BFP0hKEb.js} +7 -7
- package/dist/{command-BdstSEHu.js.map → command-BFP0hKEb.js.map} +1 -1
- package/dist/command-BrmXUC-W.js +2 -0
- package/dist/{command-BK4kOgkH.js → command-C5xFPKHM.js} +19 -18
- package/dist/command-C5xFPKHM.js.map +1 -0
- package/dist/{command-JPGNzFsA.js → command-CzIFp1JQ.js} +10 -10
- package/dist/{command-JPGNzFsA.js.map → command-CzIFp1JQ.js.map} +1 -1
- package/dist/{command-BqJtFiPy.js → command-DUT5l13I.js} +82 -32
- package/dist/command-DUT5l13I.js.map +1 -0
- package/dist/{commands-B618jg3C.js → commands-CLmPBrTS.js} +6 -6
- package/dist/{commands-B618jg3C.js.map → commands-CLmPBrTS.js.map} +1 -1
- package/dist/{commands-Duo2ur3r.js → commands-fI3G2Wnh.js} +1 -1
- package/dist/{config-B-E0xRkf.js → config-CGvBeewH.js} +17 -7
- package/dist/config-CGvBeewH.js.map +1 -0
- package/dist/{context-CvUv7DUb.js → context-DbSMEVKs.js} +1 -5
- package/dist/{context-CvUv7DUb.js.map → context-DbSMEVKs.js.map} +1 -1
- package/dist/{db-CnoUy-FZ.js → db-CSSFtGcN.js} +1 -1
- package/dist/{db-DFbtGIYy.js → db-CfQFwJtq.js} +137 -27
- package/dist/db-CfQFwJtq.js.map +1 -0
- package/dist/{docker-2qns7hJ2.js → docker-CRQGQbIt.js} +17 -5
- package/dist/docker-CRQGQbIt.js.map +1 -0
- package/dist/{docker-IFRW8K5f.js → docker-DV2GNgle.js} +2 -2
- package/dist/{docker-IFRW8K5f.js.map → docker-DV2GNgle.js.map} +1 -1
- package/dist/{docker-D6t23fZl.js → docker-Df_YOxxB.js} +1 -1
- package/dist/{doctor-B6oYTPoR.js → doctor-DlKycx5H.js} +12 -12
- package/dist/{doctor-B6oYTPoR.js.map → doctor-DlKycx5H.js.map} +1 -1
- package/dist/{egress-IUKdeayc.js → egress-BhzNh7Yi.js} +10 -9
- package/dist/{egress-IUKdeayc.js.map → egress-BhzNh7Yi.js.map} +1 -1
- package/dist/{environment-DiCljXMD.js → environment-1jE-HhO-.js} +2 -2
- package/dist/{environment-DiCljXMD.js.map → environment-1jE-HhO-.js.map} +1 -1
- package/dist/family-D2b7WD10.js +67 -0
- package/dist/family-D2b7WD10.js.map +1 -0
- package/dist/files-AiEKuEZe.js +2 -0
- package/dist/files-BMCvfB4X.js +2 -0
- package/dist/{files-CMSUXAEg.js → files-CSzVADHn.js} +4 -4
- package/dist/{files-CMSUXAEg.js.map → files-CSzVADHn.js.map} +1 -1
- package/dist/{files-B6aVVGrk.js → files-DJUFg-4e.js} +2 -2
- package/dist/{files-B6aVVGrk.js.map → files-DJUFg-4e.js.map} +1 -1
- package/dist/{git-EmdFJ4DY.js → git-DfG-YJ1u.js} +2 -2
- package/dist/{git-EmdFJ4DY.js.map → git-DfG-YJ1u.js.map} +1 -1
- package/dist/{harness-iqWeJQMM.js → harness-CeCp2GSW.js} +19 -2
- package/dist/harness-CeCp2GSW.js.map +1 -0
- package/dist/harness.d.ts +5 -3
- package/dist/harness.js +1 -1
- package/dist/{header-_F0EWob3.js → header-CALhmafp.js} +10 -3
- package/dist/{header-_F0EWob3.js.map → header-CALhmafp.js.map} +1 -1
- package/dist/{images-CscddrJU.js → images-BMgdpMAX.js} +6 -6
- package/dist/{images-CscddrJU.js.map → images-BMgdpMAX.js.map} +1 -1
- package/dist/index.d.ts +17 -5
- package/dist/index.js +92 -62
- package/dist/index.js.map +1 -1
- package/dist/init-B6vegaP_.js +2 -0
- package/dist/{init-NQfIvi-0.js → init-BIdLF0F9.js} +4 -4
- package/dist/{init-NQfIvi-0.js.map → init-BIdLF0F9.js.map} +1 -1
- package/dist/{invocations-nPvrlZPj.js → invocations-CLaL8AuW.js} +2 -2
- package/dist/{invocations-nPvrlZPj.js.map → invocations-CLaL8AuW.js.map} +1 -1
- package/dist/judge.d.ts +1 -1
- package/dist/{model-Cl507trk.js → model-B1BZoZ9b.js} +4 -4
- package/dist/{model-Cl507trk.js.map → model-B1BZoZ9b.js.map} +1 -1
- package/dist/{model-BndnmcLm.js → model-BOnlAPzs.js} +162 -17
- package/dist/model-BOnlAPzs.js.map +1 -0
- package/dist/model-DWeg5J7z.js +2 -0
- package/dist/{model-DMOraCcm.js → model-FXko2kl7.js} +1 -1
- package/dist/{options-DPUlm4FI.js → options-CKqyPGxX.js} +12 -8
- package/dist/options-CKqyPGxX.js.map +1 -0
- package/dist/{patch-CGCVipPu.js → patch-bDV8YFW0.js} +2 -2
- package/dist/{patch-CGCVipPu.js.map → patch-bDV8YFW0.js.map} +1 -1
- package/dist/{plan-PPc4-6xd.js → plan-DUydFc_1.js} +53 -25
- package/dist/plan-DUydFc_1.js.map +1 -0
- package/dist/{preflight-CddycK92.js → preflight-BkoSxf4C.js} +3 -3
- package/dist/{preflight-CddycK92.js.map → preflight-BkoSxf4C.js.map} +1 -1
- package/dist/{preflight-Zumy0adT.js → preflight-D5LW7Oqi.js} +1 -1
- package/dist/{prepare-CsP-057p.js → prepare-CUAWHebF.js} +5 -5
- package/dist/{prepare-CsP-057p.js.map → prepare-CUAWHebF.js.map} +1 -1
- package/dist/{prepare-CKdMoC7e.js → prepare-oji2oO8W.js} +2 -2
- package/dist/process-CSVO_jsc.js +119 -0
- package/dist/process-CSVO_jsc.js.map +1 -0
- package/dist/{provenance-CsX6PHDR.js → provenance-D7HK3vzt.js} +2 -2
- package/dist/{provenance-CsX6PHDR.js.map → provenance-D7HK3vzt.js.map} +1 -1
- package/dist/{providers-DkwiLKDS.js → providers-B1DzMvmH.js} +3 -3
- package/dist/{providers-DkwiLKDS.js.map → providers-B1DzMvmH.js.map} +1 -1
- package/dist/{providers-yXZZcPfx.js → providers-CpEfNU1F.js} +1 -1
- package/dist/{query-BUcwCYsI.js → query-BjBMLa-C.js} +2 -2
- package/dist/{query-BUcwCYsI.js.map → query-BjBMLa-C.js.map} +1 -1
- package/dist/query-oPO_XrTy.js +2 -0
- package/dist/{questions-BEPj98fZ.js → questions-B8by4Bbt.js} +3 -3
- package/dist/{questions-BEPj98fZ.js.map → questions-B8by4Bbt.js.map} +1 -1
- package/dist/{results-BcSHWbgB.js → results-Di7m8dQ3.js} +4 -4
- package/dist/{results-BcSHWbgB.js.map → results-Di7m8dQ3.js.map} +1 -1
- package/dist/{resume-DZ9AmfPv.js → resume-SUz1GskV.js} +3 -3
- package/dist/{resume-DZ9AmfPv.js.map → resume-SUz1GskV.js.map} +1 -1
- package/dist/{runtime-Dfb2r-cS.js → runtime-DdsfnRz_.js} +148 -6
- package/dist/runtime-DdsfnRz_.js.map +1 -0
- package/dist/{sandbox--ZXuscAZ.js → sandbox-D5v-of6d.js} +10 -5
- package/dist/sandbox-D5v-of6d.js.map +1 -0
- package/dist/sandbox.d.ts +4 -3
- package/dist/sandbox.js +1 -1
- package/dist/{server-Da60oe4b.js → server-BNnXtmS8.js} +7 -7
- package/dist/{server-Da60oe4b.js.map → server-BNnXtmS8.js.map} +1 -1
- package/dist/server-BNwoKzlm.js +2 -0
- package/dist/source-Bk_0ZaEY.js +2 -0
- package/dist/{source-BVOnL3sn.js → source-Dof0NQBm.js} +2 -2
- package/dist/{source-BVOnL3sn.js.map → source-Dof0NQBm.js.map} +1 -1
- package/dist/{src-TdJuDYqB.js → src-8rRXy4u9.js} +71 -124
- package/dist/src-8rRXy4u9.js.map +1 -0
- package/dist/{src-DdRlaiRc.js → src-BGWuWp6I.js} +1 -1
- package/dist/{src-D13vpMh0.js → src-BRt0O02i.js} +1 -1
- package/dist/{src-BrMbt1B2.js → src-Bx2x1ekD.js} +1 -1
- package/dist/{src-CxTC5kNI.js → src-COgE7991.js} +4 -6
- package/dist/{src-CxTC5kNI.js.map → src-COgE7991.js.map} +1 -1
- package/dist/{src-D_cXqmnI.js → src-C_4Umgh2.js} +9 -10
- package/dist/{src-D_cXqmnI.js.map → src-C_4Umgh2.js.map} +1 -1
- package/dist/src-DLa-ESW0.js +4 -0
- package/dist/src-cN0aB-WR.js +3 -0
- package/dist/{src-BT4dB5CO.js → src-cdIVjC5b.js} +1 -1
- package/dist/{src-BT5pwco0.js → src-sIsDHWYq.js} +1 -1
- package/dist/state-Bo5zg4da.js +2 -0
- package/dist/{state-B72h_jEw.js → state-CiXyLqg0.js} +5 -5
- package/dist/{state-B72h_jEw.js.map → state-CiXyLqg0.js.map} +1 -1
- package/dist/studio-app/assets/index-B9MCTfuu.css +2 -0
- package/dist/studio-app/assets/{index-Bgr-efZm.js → index-CROH2NB3.js} +12 -12
- package/dist/studio-app/index.html +2 -2
- package/dist/{summary-CJdxaRdM.js → summary-2OAWF2N7.js} +6 -6
- package/dist/{summary-CJdxaRdM.js.map → summary-2OAWF2N7.js.map} +1 -1
- package/dist/{summary-k3mhWWXv.js → summary-BTksrIxn.js} +58 -4
- package/dist/summary-BTksrIxn.js.map +1 -0
- package/dist/testing.d.ts +1 -1
- package/dist/{tools-D0gF2QoV.js → tools-B10Uj7OI.js} +4 -4
- package/dist/tools-B10Uj7OI.js.map +1 -0
- package/dist/types-C5vI3g1C.js.map +1 -1
- package/dist/{types-_LOFdPUf.d.ts → types-D8RLfkFP.d.ts} +195 -13
- package/dist/view/index.d.ts +44 -12
- package/dist/view/index.js +5 -4
- package/dist/views-Cs2Oqayd.js +3 -0
- package/dist/{views-DZ1YDqey.js → views-Dqa5ojfZ.js} +4 -4
- package/dist/{views-DZ1YDqey.js.map → views-Dqa5ojfZ.js.map} +1 -1
- package/dist/{worker-entry-DXeyWY-m.js → worker-entry-BAikHl9F.js} +9 -5
- package/dist/worker-entry-BAikHl9F.js.map +1 -0
- package/dist/{workspace-DpXdrslo.js → workspace-C-TrNIUc.js} +4 -4
- package/dist/{workspace-DpXdrslo.js.map → workspace-C-TrNIUc.js.map} +1 -1
- package/dist/{workspace-DzNL-3iO.js → workspace-Do_SiWBs.js} +1 -1
- package/docker/Dockerfile +7 -5
- package/package.json +1 -1
- package/dist/agent-Cfj1wD4D.js.map +0 -1
- package/dist/assemble-BbFVjHeS.js.map +0 -1
- package/dist/assemble-CXiTVyHD.js +0 -2
- package/dist/capture-BXhBfmsM.js.map +0 -1
- package/dist/capture-N7r3yeak.js +0 -2
- package/dist/cli-BMjTKY-D.js.map +0 -1
- package/dist/client-CwzU9CiW.js.map +0 -1
- package/dist/command-BK4kOgkH.js.map +0 -1
- package/dist/command-BqJtFiPy.js.map +0 -1
- package/dist/command-CTVGuH8y.js +0 -2
- package/dist/config-B-E0xRkf.js.map +0 -1
- package/dist/db-DFbtGIYy.js.map +0 -1
- package/dist/docker-2qns7hJ2.js.map +0 -1
- package/dist/files-BpYgF2z0.js +0 -2
- package/dist/files-YCCVaTQC.js +0 -2
- package/dist/harness-iqWeJQMM.js.map +0 -1
- package/dist/init-DBHyi92L.js +0 -2
- package/dist/model-BbLGRtOs.js +0 -2
- package/dist/model-BndnmcLm.js.map +0 -1
- package/dist/options-DPUlm4FI.js.map +0 -1
- package/dist/plan-PPc4-6xd.js.map +0 -1
- package/dist/query-noGbekGF.js +0 -2
- package/dist/runtime-Dfb2r-cS.js.map +0 -1
- package/dist/sandbox--ZXuscAZ.js.map +0 -1
- package/dist/server-DiGR4JxK.js +0 -2
- package/dist/source-Dqu0jKwx.js +0 -2
- package/dist/src-BHNiP-5A.js +0 -4
- package/dist/src-TdJuDYqB.js.map +0 -1
- package/dist/src-xczuQbE6.js +0 -3
- package/dist/state-Bf_Vq_pV.js +0 -2
- package/dist/studio-app/assets/index-CUx6GoL5.css +0 -2
- package/dist/summary-k3mhWWXv.js.map +0 -1
- package/dist/tools-D0gF2QoV.js.map +0 -1
- package/dist/views-Ds0GXWtX.js +0 -3
- package/dist/worker-entry-DXeyWY-m.js.map +0 -1
|
@@ -6,8 +6,8 @@
|
|
|
6
6
|
<meta name="referrer" content="no-referrer" />
|
|
7
7
|
<link rel="icon" type="image/svg+xml" href="./favicon.svg" />
|
|
8
8
|
<title>reeval Studio</title>
|
|
9
|
-
<script type="module" crossorigin src="./assets/index-
|
|
10
|
-
<link rel="stylesheet" crossorigin href="./assets/index-
|
|
9
|
+
<script type="module" crossorigin src="./assets/index-CROH2NB3.js"></script>
|
|
10
|
+
<link rel="stylesheet" crossorigin href="./assets/index-B9MCTfuu.css">
|
|
11
11
|
</head>
|
|
12
12
|
<body>
|
|
13
13
|
<div id="root"></div>
|
|
@@ -1,8 +1,8 @@
|
|
|
1
|
-
import { a as providerModel } from "./model-
|
|
2
|
-
import { l as writeCached } from "./files-
|
|
3
|
-
import { r as fromRecorded, t as attemptSummary } from "./tools-
|
|
4
|
-
import { n as allRuns } from "./commands-
|
|
5
|
-
import { a as failedAnalysis, i as defaultBase, o as prepareAnalysis, s as promptOf, t as analyseDigest } from "./command-
|
|
1
|
+
import { a as providerModel } from "./model-B1BZoZ9b.js";
|
|
2
|
+
import { l as writeCached } from "./files-DJUFg-4e.js";
|
|
3
|
+
import { r as fromRecorded, t as attemptSummary } from "./tools-B10Uj7OI.js";
|
|
4
|
+
import { n as allRuns } from "./commands-CLmPBrTS.js";
|
|
5
|
+
import { a as failedAnalysis, i as defaultBase, o as prepareAnalysis, s as promptOf, t as analyseDigest } from "./command-DUT5l13I.js";
|
|
6
6
|
import { join } from "node:path";
|
|
7
7
|
import { generateText, stepCountIs, tool } from "ai";
|
|
8
8
|
import { z } from "zod";
|
|
@@ -189,4 +189,4 @@ function fakeSummaryModel(p) {
|
|
|
189
189
|
//#endregion
|
|
190
190
|
export { writeRunSummary };
|
|
191
191
|
|
|
192
|
-
//# sourceMappingURL=summary-
|
|
192
|
+
//# sourceMappingURL=summary-2OAWF2N7.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"summary-
|
|
1
|
+
{"version":3,"file":"summary-2OAWF2N7.js","names":[],"sources":["../src/analysis/agent/summary.ts"],"mappings":";;;;;;;;;AA4BA,MAAM,aAAa;AAEnB,MAAa,uBAAuB;;;;;;AAOpC,SAAgB,aAAa,GAAqB,UAAkB;CAClE,IAAI,OAAO;CACX,MAAM,SAAY,UAAoC;EACpD,MAAM,OAAO,KAAK,UAAU,KAAK,CAAC,CAAC;EACnC,IAAI,OAAO,MACT,OAAO,EACL,OACE,8FACJ;EAEF,QAAQ;EACR,OAAO;CACT;CACA,MAAM,OAAO,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,CAAC,SAAS,MAAO,IAAI,CAAC,CAAC,IAAI,CAAC,CAAE;CACjE,OAAO;EACL,eAAe,KAAK;GAClB,aACE;GACF,aAAa,EAAE,OAAO,EACpB,IAAI,EAAE,OAAO,CAAC,CAAC,SAAS,0CAA+C,EACzE,CAAC;GACD,SAAS,OAAO,EAAE,SAAS;IACzB,MAAM,OAAO,EAAE,QAAQ,MAAM,MAAM,EAAE,OAAO,EAAE;IAC9C,IAAI,MAAM;KACR,MAAM,EAAE,UAAU,WAAW,GAAG,SAAS;KACzC,OAAO,MAAM,IAAI;IACnB;IACA,OAAO,EACL,OAAO,EAAE,OAAO,SAAS,MAAM,MAAM,EAAE,OAAO,EAAE,IAC5C,GAAG,GAAG,8BACN,uBAAuB,GAAG,aAAa,EAAE,OAAO,QAAQ,KAAK,IAAI,KAAK,OAAO,GACnF;GACF;EACF,CAAC;EACD,cAAc,KAAK;GACjB,aACE;GACF,aAAa,EAAE,OAAO,EACpB,IAAI,EAAE,OAAO,CAAC,CAAC,SAAS,gDAAqD,EAC/E,CAAC;GACD,SAAS,OAAO,EAAE,SAAS;IACzB,MAAM,IAAI,EAAE,OAAO,SAAS;IAC5B,MAAM,MAAM,KAAK,KAAK,MAAM,MAAM,EAAE,OAAO,UAAU,EAAE,KAAK;IAC5D,MAAM,SAAS,KAAK,OAAO,MAAM,SAAS,MAAM,EAAE,QAAQ,CAAC,CAAC,MAAM,MAAM,EAAE,QAAQ,GAAG,GAAG;IACxF,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,QACjB,OAAO,EACL,OAAO,cAAc,GAAG,qBAAqB,OAAO,KAAK,EAAE,OAAO,QAAQ,CAAC,CAAC,KAAK,IAAI,EAAE,GACzF;IAEF,MAAM,OAAO,MAAM,aAAa,KAAK,IAAI,KAAK,IAAI,OAAO,KAAK,GAAG,MAAM;IACvE,OAAO,MAAM;KAAE,GAAG,eAAe,IAAI;KAAG;IAAG,CAAC;GAC9C;EACF,CAAC;CACH;AACF;AAIA,MAAM,WAAW,MAAe,OAAO,MAAM,YAAY,MAAM,QAAQ,WAAW;;AAGlF,SAAgB,UAAU,OAA8B;CACtD,MAAM,KAAK,MAAM,QAAQ,MAAM,CAAC,QAAQ,EAAE,MAAM,CAAC;CACjD,MAAM,WAAW,GACd,QAAQ,MAAM,EAAE,SAAS,eAAe,CAAC,CACzC,KAAK,OAAO;EAAE,GAAI,EAAE;EAA6C,UAAU;CAAE,EAAE;CAClF,IAAI,CAAC,GAAG,QAAQ,OAAO;EAAE;EAAU,MAAM;CAAG;CAC5C,OAAO;EACL;EACA,MAAM,uGAAuG,GAC1G,KAAK,MAAM,GAAG,EAAE,KAAK,GAAG,KAAK,UAAU,EAAE,KAAK,EAAE,IAAI,KAAK,UAAU,EAAE,MAAM,GAAG,CAAC,CAC/E,KAAK,IAAI;CACd;AACF;;AAGA,eAAsB,YACpB,GACA,OACA,QACsF;CACtF,MAAM,SAAS,MAAM,aAAa;EAChC,OAAO,MAAM,OAAO,iBAAiB,CAAC,IAAI,MAAM,cAAc,KAAK;EACnE,QAAQ;EACR,QAAQ,SAAS,EAAE,MAAM;EACzB,OAAO,aAAa,GAAG,OAAO,SAAS,iBAAiB,CAAC;EACzD,UAAU,YAAA,CAAyB;EACnC,cAAc,EAAE,iBACd,eAAe,IAAoB,EAAE,YAAY,OAAgB,IAAI,KAAA;EACvE,aAAa;EACb,iBAAiB,OAAO,SAAS;EACjC,YAAY;CACd,CAAC;CACD,OAAO;EACL,OAAO,OAAO,MAAM,SAAS,MAC3B,EAAE,YAAY,KAAK,OAAO;GAAE,MAAM,EAAE;GAAU,OAAO,EAAE;GAAO,QAAQ,EAAE;EAAO,EAAE,CACnF;EACA,OAAO;GACL,aAAa,OAAO,YAAY,eAAe;GAC/C,cAAc,OAAO,YAAY,gBAAgB;EACnD;CACF;AACF;;;;;AAQA,eAAsB,gBAAgB,SAId;CACtB,MAAM,EAAE,QAAQ,OAAO,UAAU;CACjC,MAAM,UAAU,MAAM,QAAQ,MAAM;CACpC,MAAM,OAAO,QAAQ,MAAM,MAAM,EAAE,UAAU,KAAK;CAClD,IAAI,CAAC,MAAM,MAAM,IAAI,MAAM,GAAG,MAAM,uBAAuB;CAC3D,MAAM,WAAW,MAAM,gBACrB,QACA,MACA,MAAM,YAAY,QAAQ,SAAS,IAAI,GACvC,IACF;CACA,MAAM,UAAU,KAAK,IAAI;CACzB,IAAI;CACJ,IAAI;EACF,MAAM,EAAE,OAAO,UAAU,MAAM,YAAY,UAAU,OAAO,MAAM;EAClE,MAAM,MAAM,cAAc,QAAQ,UAAU,OAAO,UAAU,KAAK,CAAC;EACnE,MAAM,QAAQ,GAAkB,MAAe,MAAM,OAAO,OAAO,IAAI;EACvE,IAAI,QAAQ;GACV,aAAa,KAAK,IAAI,MAAM,aAAa,MAAM,WAAW;GAC1D,cAAc,KAAK,IAAI,MAAM,cAAc,MAAM,YAAY;EAC/D;EACA,IAAI,YAAY,KAAK,IAAI,IAAI;CAC/B,SAAS,OAAO;EACd,MAAM,UAAU,iBAAiB,QAAQ,MAAM,UAAU,OAAO,KAAK;EACrE,MAAM,MAAM,eAAe,QAAQ,UAAU,OAAO,SAAS,KAAK,IAAI,IAAI,OAAO;CACnF;CACA,MAAM,OAAO,MAAM,YAAY,OAAO,MAAM,SAAS,KAAK,GAAG;CAC7D,OAAO;EAAE;EAAK,KAAK,SAAS;EAAK;CAAK;AACxC;;;;;;AASA,SAAS,iBAAiB,GAAoC;CAC5D,MAAM,QAAQ;EACZ,aAAa;GAAE,OAAO;GAAG,SAAS;GAAG,WAAW;GAAG,YAAY;EAAE;EACjE,cAAc;GAAE,OAAO;GAAG,MAAM;GAAG,WAAW;EAAE;CAClD;CACA,MAAM,UAAU,SAAoB,YAAmC;EACrE;EACA,cAAc;GAAE,SAAS;GAAQ,KAAK,WAAW,SAAS,SAAS;EAAa;EAChF;EACA,UAAU,CAAC;EACX,UAAU,EAAE,SAAS,gBAAgB;CACvC;CACA,OAAO;EACL,sBAAsB;EACtB,UAAU;EACV,SAAS;EACT,eAAe,CAAC;EAChB,YAAY,OAAO,YAAmE;GACpF,MAAM,WAAW,QAAQ,OAAO,MAC7B,MACC,EAAE,SAAS,UACX,MAAM,QAAQ,EAAE,OAAO,KACtB,EAAE,QAAyB,MAAM,MAAM,EAAE,SAAS,aAAa,CACpE;GACA,MAAM,QAAQ,CACZ,GAAG,EAAE,OAAO,QACT,MAAM,GAAG,UAAU,CAAC,CACpB,KAAK,QAAQ;IAAE,MAAM;IAAiB,OAAO,EAAE,GAAG;GAAE,EAAE,GACzD,GAAG,OAAO,QAAQ,EAAE,OAAO,QAAQ,CAAC,CACjC,QAAQ,GAAG,OAAO,EAAE,YAAY,MAAM,CAAC,CACvC,MAAM,GAAG,UAAU,CAAC,CACpB,KAAK,CAAC,SAAS;IAAE,MAAM;IAAgB,OAAO,EAAE,GAAG;GAAE,EAAE,CAC5D;GACA,IAAI,YAAY,CAAC,MAAM,QACrB,OAAO,OACL,CAAC;IAAE,MAAM;IAAQ,MAAM,QAAQ,WAAW,kBAAkB,UAAU;GAAG,CAAC,GAC1E,MACF;GAEF,OAAO,OACL,MAAM,KAAK,GAAG,OAAO;IACnB,MAAM;IACN,YAAY,QAAQ,EAAE,KAAK,GAAG;IAC9B,UAAU,EAAE;IACZ,OAAO,KAAK,UAAU,EAAE,KAAK;GAC/B,EAAE,GACF,YACF;EACF;EACA,UAAU,YAAY;GACpB,MAAM,IAAI,MAAM,yCAAyC;EAC3D;CACF;AACF"}
|
|
@@ -1,7 +1,8 @@
|
|
|
1
1
|
import { n as clip, r as formatDuration } from "./text-gEcKjAyp.js";
|
|
2
|
-
import { E as runtimeLabel, H as cellContrast, L as judgeRatesLabel, M as cellJudgeRates, N as cellUsage, b as isUngatedMiss, g as harnessLabel, h as formatUsage, q as effortLabel, x as languageLabel } from "./model-
|
|
2
|
+
import { E as runtimeLabel, H as cellContrast, L as judgeRatesLabel, M as cellJudgeRates, N as cellUsage, b as isUngatedMiss, g as harnessLabel, h as formatUsage, q as effortLabel, x as languageLabel } from "./model-BOnlAPzs.js";
|
|
3
|
+
import { a as familyLiftLabel, i as familyHeadLabel, o as familyRateLabel, r as familyEffortLabel } from "./family-D2b7WD10.js";
|
|
3
4
|
import { a as sampleReasons } from "./pairs-DDsjgm5F.js";
|
|
4
|
-
import { o as warnings } from "./state-
|
|
5
|
+
import { o as warnings } from "./state-CiXyLqg0.js";
|
|
5
6
|
//#region src/history/summary.ts
|
|
6
7
|
const OUTCOME_MARK = {
|
|
7
8
|
pass: "✓",
|
|
@@ -37,6 +38,37 @@ function liftLabel(lift) {
|
|
|
37
38
|
const counts = `${variant.pass}/${variant.counted} vs ${control.pass}/${control.counted}`;
|
|
38
39
|
return `lift ${signedPoints(lift.points)} pts (${range}, ${counts})${lift.withinNoise ? ", within noise" : ""}`;
|
|
39
40
|
}
|
|
41
|
+
const answered = (t) => t.pass + t.fail + t.notRun;
|
|
42
|
+
const VERDICT_ORDER = {
|
|
43
|
+
discriminates: 0,
|
|
44
|
+
mixed: 1
|
|
45
|
+
};
|
|
46
|
+
/**
|
|
47
|
+
* A variant cell's checks against its control's, those that tell them apart first:
|
|
48
|
+
* `loads the skill 2/2 vs 0/2 · exports greet 2/2 vs 2/2 (doesn't discriminate)`. Empty
|
|
49
|
+
* without a control's tally.
|
|
50
|
+
*/
|
|
51
|
+
function checksLabel(cell) {
|
|
52
|
+
return (cell.checks ?? []).filter((k) => k.control && k.verdict).map((k, i) => ({
|
|
53
|
+
k,
|
|
54
|
+
i
|
|
55
|
+
})).sort((a, b) => (VERDICT_ORDER[a.k.verdict ?? ""] ?? 2) - (VERDICT_ORDER[b.k.verdict ?? ""] ?? 2) || a.i - b.i).map(({ k }) => {
|
|
56
|
+
const control = k.control;
|
|
57
|
+
const notRun = k.notRun + control.notRun;
|
|
58
|
+
const notes = [notRun ? `${notRun} not run` : "", k.verdict === "both-pass" || k.verdict === "both-fail" ? "doesn't discriminate" : ""].filter(Boolean);
|
|
59
|
+
return `${k.label} ${k.pass}/${answered(k)} vs ${control.pass}/${answered(control)}${notes.length ? ` (${notes.join(", ")})` : ""}`;
|
|
60
|
+
}).join(" · ");
|
|
61
|
+
}
|
|
62
|
+
/** Every variant cell's checks that can't tell it from its control: they pass, or fail, on both. */
|
|
63
|
+
function undiscriminating(evals) {
|
|
64
|
+
return evals.flatMap((entry) => entry.cells.flatMap((cell) => (cell.checks ?? []).filter((k) => k.verdict === "both-pass" || k.verdict === "both-fail").map((check) => ({
|
|
65
|
+
entry,
|
|
66
|
+
cell,
|
|
67
|
+
check
|
|
68
|
+
}))));
|
|
69
|
+
}
|
|
70
|
+
/** `evals/s.eval.ts › greets fake/skilled/fake exports greet: passes on both`. */
|
|
71
|
+
const undiscriminatingLabel = ({ entry, cell, check }) => `${entry.id} ${cell.key} ${check.label}: ${check.verdict === "both-pass" ? "passes" : "fails"} on both`;
|
|
40
72
|
const rateText = (r) => `${r.name} ${r.wins}/${r.n} (${Math.round(r.wins / r.n * 100)}%)`;
|
|
41
73
|
/**
|
|
42
74
|
* One compare question's answer: `"Which is cleaner?" within: with-skill 2/2 (100%), plain 0/2
|
|
@@ -68,6 +100,7 @@ function cellCounts(e) {
|
|
|
68
100
|
c.gate === "none" ? "not gated" : "",
|
|
69
101
|
c.gate !== "all" && c.gate !== "none" ? `gate ${gateLabel(c.gate)}` : "",
|
|
70
102
|
c.lift ? liftLabel(c.lift) : "",
|
|
103
|
+
...[checksLabel(c)].flatMap((checks) => checks ? [`checks ${checks}`] : []),
|
|
71
104
|
...[cellContrast(e, c.key)].flatMap((effort) => effort ? [`effort ${effortLabel(effort)}`] : []),
|
|
72
105
|
...[judgeRatesLabel(cellJudgeRates(e, c.key))].filter(Boolean),
|
|
73
106
|
...[formatUsage(cellUsage(e, c.key))].filter((u) => Boolean(u)),
|
|
@@ -104,6 +137,26 @@ function stateRows(run) {
|
|
|
104
137
|
rows.push(["Runner", run.runner.kind === "ci" ? `CI (${run.runner.provider})${run.runner.runUrl ? ` ${run.runner.runUrl}` : ""}` : `local · ${run.author}`]);
|
|
105
138
|
return rows;
|
|
106
139
|
}
|
|
140
|
+
/** Each pooled `evaluate.each`: its cases, and its pooled rates or lifts. */
|
|
141
|
+
function familySection(run) {
|
|
142
|
+
if (!run.families?.length) return [];
|
|
143
|
+
const lines = [
|
|
144
|
+
"## Families",
|
|
145
|
+
"",
|
|
146
|
+
"| | Family | Pooled |",
|
|
147
|
+
"|---|---|---|"
|
|
148
|
+
];
|
|
149
|
+
for (const f of run.families) {
|
|
150
|
+
const pooled = f.lifts.length ? f.lifts.map((l) => {
|
|
151
|
+
const gate = l.minLift === void 0 ? "" : ` (minLift ${Math.round(l.minLift * 100)})`;
|
|
152
|
+
const effort = familyEffortLabel(l);
|
|
153
|
+
return `${l.variant} ${familyLiftLabel(l)}${gate}${effort ? `; effort ${effort}` : ""}`;
|
|
154
|
+
}) : f.rates.map((r) => `${r.cell} ${familyRateLabel(r)}`);
|
|
155
|
+
lines.push(`| ${OUTCOME_MARK[f.outcome]} | ${cell(`${f.id} (${familyHeadLabel(f)})`)} | ${cell(pooled.join(" · "))} |`);
|
|
156
|
+
}
|
|
157
|
+
lines.push("");
|
|
158
|
+
return lines;
|
|
159
|
+
}
|
|
107
160
|
function renderSummary(run) {
|
|
108
161
|
const lines = [
|
|
109
162
|
`# reeval run ${run.runId}`,
|
|
@@ -122,6 +175,7 @@ function renderSummary(run) {
|
|
|
122
175
|
...warnings(run).map((w) => `> **Warning:** ${w}\n`),
|
|
123
176
|
`Record: [${run.runId}.json](../${run.runId}.json) · artifacts: ${run.artifacts}`,
|
|
124
177
|
"",
|
|
178
|
+
...familySection(run),
|
|
125
179
|
"## Evals",
|
|
126
180
|
"",
|
|
127
181
|
"| | Eval | Time or cells | Judge |",
|
|
@@ -195,6 +249,6 @@ function attemptDetail(lines, a) {
|
|
|
195
249
|
if (links.length) lines.push(`- Files: ${links.join(" · ")}`);
|
|
196
250
|
}
|
|
197
251
|
//#endregion
|
|
198
|
-
export {
|
|
252
|
+
export { gateLabel as a, repeatsVerdict as c, undiscriminating as d, undiscriminatingLabel as f, checksLabel as i, stateRows as l, answersLabel as n, liftLabel as o, cellCounts as r, renderSummary as s, OUTCOME_MARK as t, summaryCounts as u };
|
|
199
253
|
|
|
200
|
-
//# sourceMappingURL=summary-
|
|
254
|
+
//# sourceMappingURL=summary-BTksrIxn.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"summary-BTksrIxn.js","names":[],"sources":["../src/history/summary.ts"],"mappings":";;;;;;AAqBA,MAAa,eAAwC;CACnD,MAAM;CACN,MAAM;CACN,aAAa;CACb,SAAS;AACX;;AAGA,MAAM,qBAAqB;AAE3B,MAAM,QAAQ,MAAc,EAAE,QAAQ,OAAO,KAAK,CAAC,CAAC,QAAQ,OAAO,GAAG;AAEtE,SAAgB,cAAc,GAAiC;CAC7D,MAAM,QAAQ;EACZ,EAAE,QAAQ,GAAG,EAAE,KAAK;EACpB,EAAE,OAAO,EAAE,WAAW,KAAK,GAAG,EAAE,OAAO,EAAE,SAAS;EAClD,EAAE,WAAW,KAAK,GAAG,EAAE,SAAS;EAChC,EAAE,cAAc,GAAG,EAAE,WAAW;EAChC,EAAE,WAAW,GAAG,EAAE,QAAQ;CAC5B,CAAC,CAAC,OAAO,OAAO;CAChB,IAAI,MAAM,WAAW,GAAG,OAAO;CAC/B,MAAM,QAAQ,EAAE,aAAa,EAAE,QAAQ,KAAK,KAAK,EAAE,SAAS,aAAa,EAAE,MAAM;CACjF,OAAO,GAAG,MAAM,KAAK,KAAK,IAAI;AAChC;AAEA,MAAM,gBAAgB,MAAc,GAAG,KAAK,IAAI,MAAM,MAAM,KAAK,MAAM,KAAK,IAAI,CAAC,IAAI,GAAG;;AAGxF,MAAa,aAAa,SACxB,SAAS,SAAS,SAAS,SAAS,OAAO,YAAY,KAAK;;;;;AAM9D,SAAgB,UAAU,MAA+C;CACvE,MAAM,EAAE,UAAU,SAAS,YAAY;CACvC,MAAM,QAAQ,OAAO,aAAa,SAAS,EAAE,EAAE,GAAG,aAAa,SAAS,EAAE;CAC1E,MAAM,SAAS,GAAG,QAAQ,KAAK,GAAG,QAAQ,QAAQ,MAAM,QAAQ,KAAK,GAAG,QAAQ;CAChF,OAAO,QAAQ,aAAa,KAAK,MAAM,EAAE,QAAQ,MAAM,IAAI,OAAO,GAAG,KAAK,cAAc,mBAAmB;AAC7G;AAEA,MAAM,YAAY,MAAkB,EAAE,OAAO,EAAE,OAAO,EAAE;AAExD,MAAM,gBAAwC;CAAE,eAAe;CAAG,OAAO;AAAE;;;;;;AAO3E,SAAgB,YAAY,MAA0B;CAEpD,QADc,KAAK,UAAU,CAAC,EAAA,CAAG,QAAQ,MAAM,EAAE,WAAW,EAAE,OACpD,CAAC,CACR,KAAK,GAAG,OAAO;EAAE;EAAG;CAAE,EAAE,CAAC,CACzB,MACE,GAAG,OACD,cAAc,EAAE,EAAE,WAAW,OAAO,MAAM,cAAc,EAAE,EAAE,WAAW,OAAO,MAC/E,EAAE,IAAI,EAAE,CACZ,CAAC,CACA,KAAK,EAAE,QAAQ;EACd,MAAM,UAAU,EAAE;EAClB,MAAM,SAAS,EAAE,SAAS,QAAQ;EAClC,MAAM,QAAQ,CACZ,SAAS,GAAG,OAAO,YAAY,IAC/B,EAAE,YAAY,eAAe,EAAE,YAAY,cAAc,yBAAyB,EACpF,CAAC,CAAC,OAAO,OAAO;EAChB,OAAO,GAAG,EAAE,MAAM,GAAG,EAAE,KAAK,GAAG,SAAS,CAAC,EAAE,MAAM,QAAQ,KAAK,GAAG,SAAS,OAAO,IAAI,MAAM,SAAS,KAAK,MAAM,KAAK,IAAI,EAAE,KAAK;CACjI,CAAC,CAAC,CACD,KAAK,KAAK;AACf;;AAGA,SAAgB,iBAAiB,OAAoB;CACnD,OAAO,MAAM,SAAS,UACpB,MAAM,MAAM,SAAS,UAClB,KAAK,UAAU,CAAC,EAAA,CACd,QAAQ,MAAM,EAAE,YAAY,eAAe,EAAE,YAAY,WAAW,CAAC,CACrE,KAAK,WAAW;EAAE;EAAO;EAAM;CAAM,EAAE,CAC5C,CACF;AACF;;AAGA,MAAa,yBAAyB,EACpC,OACA,MACA,YAEA,GAAG,MAAM,GAAG,IAAI,KAAK,IAAI,IAAI,MAAM,MAAM,IAAI,MAAM,YAAY,cAAc,WAAW,QAAQ;AAElG,MAAM,YAAY,MAChB,GAAG,EAAE,KAAK,GAAG,EAAE,KAAK,GAAG,EAAE,EAAE,IAAI,KAAK,MAAO,EAAE,OAAO,EAAE,IAAK,GAAG,EAAE;;;;;AAMlE,SAAgB,aAAa,QAA+B;CAC1D,MAAM,IAAI,IAAI,KAAK,OAAO,UAAU,EAAE,EAAE;CACxC,MAAM,OAAO,OAAO,MAAM,QAAQ,MAAM,EAAE,WAAW,KAAK,CAAC,CAAC;CAC5D,MAAM,OAAQ,CAAC,UAAU,QAAQ,CAAC,CAAW,SAAS,QAAQ;EAC5D,MAAM,QAAQ,OAAO,MAAM,QAAQ,MAAM,EAAE,QAAQ,GAAG;EACtD,OAAO,MAAM,SAAS,CAAC,GAAG,IAAI,IAAI,MAAM,IAAI,QAAQ,CAAC,CAAC,KAAK,IAAI,GAAG,IAAI,CAAC;CACzE,CAAC;CACD,MAAM,QAAQ,OAAO,MAAM;CAC3B,MAAM,QAAQ,GAAG,MAAM,OAAO,UAAU,IAAI,KAAK,IAAI,IAAI,KAAK,MAAM,SAAS,IAAI,KAAK;CACtF,MAAM,KAAK,OAAO,iBAAiB,OAAO;CAC1C,MAAM,SAAS,OAAO,UAAU,KAAA,IAAY,cAAc,KAAK,OAAO,OAAO,GAAG,MAAM;CACtF,IAAI,UAAU,KAAK,OAAO,UAAU,KAAA,GAClC,OAAO,GAAG,EAAE;CAEd,OAAO,GAAG,EAAE,GAAG;EAAC,GAAG;EAAM;EAAO;CAAE,CAAC,CAAC,KAAK,KAAK,IAAI;AACpD;;AAGA,SAAgB,WAAW,GAAsB;CAC/C,OAAO,EAAE,MACN,KAAK,MAAM;EACV,MAAM,QAAQ;GACZ,EAAE,SAAS,aAAa,EAAE,MAAM,MAAM,MAAM,EAAE,SAAS,SAAS,IAAI,YAAY;GAChF,EAAE,SAAS,SAAS,cAAc;GAClC,EAAE,SAAS,SAAS,EAAE,SAAS,SAAS,QAAQ,UAAU,EAAE,IAAI,MAAM;GACtE,EAAE,OAAO,UAAU,EAAE,IAAI,IAAI;GAC7B,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,SAAS,WAAY,SAAS,CAAC,UAAU,QAAQ,IAAI,CAAC,CAAE;GAC5E,GAAG,CAAC,aAAa,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,SAAS,WACnC,SAAS,CAAC,UAAU,YAAY,MAAM,GAAG,IAAI,CAAC,CAChD;GACA,GAAG,CAAC,gBAAgB,eAAe,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,OAAO;GAC7D,GAAG,CAAC,YAAY,UAAU,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,MAAmB,QAAQ,CAAC,CAAC;GAC3E,EAAE,UAAU,SAAS,aAAa,EAAE,UAAU,KAAK,IAAI,MAAM;EAC/D,CAAC,CAAC,OAAO,OAAO;EAChB,OAAO,GAAG,EAAE,IAAI,GAAG,EAAE,KAAK,GAAG,EAAE,OAAO,EAAE,OAAO,MAAM,SAAS,KAAK,MAAM,KAAK,IAAI,EAAE,KAAK;CAC3F,CAAC,CAAC,CACD,KAAK,KAAK;AACf;AAEA,SAAS,UAAU,GAAsB;CACvC,MAAM,aAAa,EAAE,SAAS,SAAS,MAAM,EAAE,UAAU;CACzD,IAAI,WAAW,WAAW,GAAG,OAAO;CACpC,OAAO,WAAW,KAAK,MAAM,GAAG,EAAE,QAAQ,GAAG,EAAE,MAAM,QAAQ,CAAC,EAAE,GAAG,EAAE,WAAW,CAAC,CAAC,KAAK,IAAI;AAC7F;;AAGA,MAAM,SAAS,SAAiB,KAAK,QAAQ,YAAY,EAAE,CAAC,CAAC,MAAM,GAAG,EAAE;;;;;AAMxE,SAAgB,UAAU,KAAyC;CACjE,MAAM,EAAE,KAAK,OAAO,QAAQ;CAC5B,MAAM,OAAgC,CAAC;CACvC,KAAK,KAAK,CACR,eACA,MACI,GAAG,IAAI,IAAI,MAAM,GAAG,EAAE,IAAI,IAAI,SAAS,OAAO,IAAI,WAAW,KAAK,IAAI,QAAQ,YAAY,UAAU,gDACpG,sBACN,CAAC;CACD,IAAI,KAAK,OACP,KAAK,KAAK,CACR,eACA,IAAI,gBACA,mBAAmB,IAAI,cAAc,MAAM,iBAAiB,MAAM,IAAI,cAAc,MAAM,MAC1F,cAAc,IAAI,qBAAqB,YAAY,IAAI,WAAW,YAAY,MAAM,IAAI,QAAQ,MAAM,IAC5G,CAAC;CAEH,KAAK,KAAK,CACR,aACA,IAAI,aAAa,UAAU,KAAK,MAAM,GAAG,EAAE,KAAK,GAAG,MAAM,EAAE,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI,KAC7E,YACJ,CAAC;CACD,KAAK,KAAK,CACR,UACA,IAAI,OAAO,SACP,IAAI,OACD,KACE,MACC,GAAG,EAAE,IAAI,QAAQ,MAAM,EAAE,EAAE,EAAE,KAAK,EAAE,YAAY,MAAM,qCAAqC,SAAS,EAAE,YAAY,QAAQ,EAAE,aAAa,iBAAiB,EAAE,WAAW,KAAK,GAAG,MAAM,EAAE,WAAW,MAAM,MAAM,KAAK,EAAE,SAAS,SAAS,cAAc,EAAE,QAAQ,OAAO,MAAM,EAAE,SAAS,IAC3R,CAAC,CACA,KAAK,IAAI,IACZ,qBACN,CAAC;CACD,KAAK,KAAK,CAAC,aAAa,IAAI,UAAU,IAAI,YAAY,CAAC,CAAC,KAAK,IAAI,KAAK,MAAM,CAAC;CAC7E,KAAK,KAAK,CACR,gBACA,IAAI,aACD,KACE,MACC,GAAG,EAAE,KAAK,GAAG,MAAM,EAAE,MAAM,IAAI,EAAE,OAAO,SAAS,aAAa,EAAE,OAAO,KAAK,MAAM,GAAG,EAAE,OAAO,EAAE,SAAS,IAAI,EAAE,OAAO,MAAM,GAAG,CAAC,MAAM,IAAI,CAAC,CAAC,KAAK,IAAI,MAAM,KAAK,EAAE,IAAI,SAAS,UAAU,EAAE,IAAI,KAAK,IAAI,MAAM,KAAK,EAAE,QAAQ,SAAS,cAAc,EAAE,QAAQ,KAAK,IAAI,MAAM,IACjR,CAAC,CACA,KAAK,IAAI,KAAK,MACnB,CAAC;CACD,KAAK,KAAK,CACR,SACA,GAAG,IAAI,MAAM,SAAS,oBAAoB,IAAI,MAAM,eAAe,SAAS,YAAY,IAAI,MAAM,eAAe,KAAK,IAAI,EAAE,KAAK,GAAG,eAAe,IAAI,MAAM,UAAU,aAAa,IAAI,MAAM,SAChM,CAAC;CACD,KAAK,KAAK,CACR,WACA,IAAI,QAAQ,SACR,GAAG,IAAI,QAAQ,OAAO,OAAO,IAAI,QAAQ,WAAW,IAAI,KAAK,IAAI,IAAI,IAAI,QAAQ,KAAK,MAAM,GAAG,EAAE,KAAK,GAAG,MAAM,EAAE,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI,MACrI,qDACN,CAAC;CACD,KAAK,KAAK,CACR,YACA,OAAO,QAAQ,IAAI,QAAQ,CAAC,CACzB,KAAK,CAAC,MAAM,OAAO,GAAG,KAAK,GAAG,GAAG,CAAC,CAClC,KAAK,KAAK,CACf,CAAC;CACD,MAAM,WAAW,cAAc,GAAG;CAClC,KAAK,KAAK,CACR,UACA,GAAG,IAAI,GAAG,KAAK,aAAa,GAAG,IAAI,WAAW,MAAM,aAAa,GAAG,KAAK,IAAI,GAAG,GAAG,IAAI,MACzF,CAAC;CACD,KAAK,KAAK,CACR,UACA,IAAI,OAAO,SAAS,OAChB,OAAO,IAAI,OAAO,SAAS,GAAG,IAAI,OAAO,SAAS,IAAI,IAAI,OAAO,WAAW,OAC5E,WAAW,IAAI,QACrB,CAAC;CACD,OAAO;AACT;;AAGA,SAAS,cAAc,KAA0B;CAC/C,IAAI,CAAC,IAAI,UAAU,QAAQ,OAAO,CAAC;CACnC,MAAM,QAAQ;EAAC;EAAe;EAAI;EAAyB;CAAe;CAC1E,KAAK,MAAM,KAAK,IAAI,UAAU;EAC5B,MAAM,SAAS,EAAE,MAAM,SACnB,EAAE,MAAM,KAAK,MAAM;GACjB,MAAM,OAAO,EAAE,YAAY,KAAA,IAAY,KAAK,aAAa,KAAK,MAAM,EAAE,UAAU,GAAG,EAAE;GACrF,MAAM,SAAS,kBAAkB,CAAC;GAClC,OAAO,GAAG,EAAE,QAAQ,GAAG,gBAAgB,CAAC,IAAI,OAAO,SAAS,YAAY,WAAW;EACrF,CAAC,IACD,EAAE,MAAM,KAAK,MAAM,GAAG,EAAE,KAAK,GAAG,gBAAgB,CAAC,GAAG;EACxD,MAAM,KACJ,KAAK,aAAa,EAAE,SAAS,KAAK,KAAK,GAAG,EAAE,GAAG,IAAI,gBAAgB,CAAC,EAAE,EAAE,EAAE,KAAK,KAAK,OAAO,KAAK,KAAK,CAAC,EAAE,GAC1G;CACF;CACA,MAAM,KAAK,EAAE;CACb,OAAO;AACT;AAEA,SAAgB,cAAc,KAAwB;CACpD,MAAM,QAAQ;EACZ,gBAAgB,IAAI;EACpB;EACA,GAAG,cAAc,IAAI,OAAO,IAAI,IAAI,WAAW,aAAa,KAAK,QAAQ,IAAI,WAAW,WAAW,mBAAmB,IAAI,OAAO,IAAI,aAAa,IAAI,UAAU,IAAI,eAAe,IAAI,UAAU,EAAE;EACnM;EACA,SAAS,IAAI,UAAU,WAAW,IAAI,QAAQ,KAAK,IAAI,IAAI,QAAQ,OAAO,QAAQ,KAAK,IAAI,KAAK,MAAM,WAAW,IAAI,KAAK,MAAM,WAAW,IAAI,KAAK,SAAS,WAAW,IAAI,SAAS,aAAa,CAAC,IAAI,OAAO,aAAa,KAAA,IAAY,GAAG,IAAI,OAAO,SAAS,aAAa,IAAI,IAAI,OAAO,YAAY,EAAE,CAAC,CAAC,OAAO,OAAO,CAAC,CAAC,KAAK,IAAI,MAAM;EACzU;EACA,GAAG,IAAI,KAAK,gBAAgB,KAAK,MAAM,kBAAkB,EAAE,QAAQ,GAAG;EACtE;EACA;EACA;EACA;EACA,GAAG,UAAU,GAAG,CAAC,CAAC,KAAK,CAAC,GAAG,OAAO,KAAK,EAAE,KAAK,KAAK,CAAC,EAAE,GAAG;EACzD;EACA,GAAG,SAAS,GAAG,CAAC,CAAC,KAAK,MAAM,kBAAkB,EAAE,GAAG;EACnD,YAAY,IAAI,MAAM,YAAY,IAAI,MAAM,sBAAsB,IAAI;EACtE;EACA,GAAG,cAAc,GAAG;EACpB;EACA;EACA;EACA;CACF;CACA,KAAK,MAAM,KAAK,IAAI,OAAO;EACzB,MAAM,IAAI,EAAE,SAAS;EACrB,MAAM,OAAO,EAAE,OACX,SACA,EAAE,SAAS,SAAS,IAClB,WAAW,CAAC,IACZ,IACE,eAAe,EAAE,UAAU,IAC3B;EACR,MAAM,KAAK,KAAK,aAAa,EAAE,SAAS,KAAK,KAAK,EAAE,EAAE,EAAE,KAAK,KAAK,IAAI,EAAE,KAAK,UAAU,CAAC,EAAE,GAAG;CAC/F;CAGA,MAAM,QAAQ,MAAgB,MAAM,UAAU,MAAM,gBAAgB,IAAI,MAAM,SAAS,IAAI;CAC3F,KAAK,MAAM,KAAK,CAAC,GAAG,IAAI,KAAK,CAAC,CAAC,MAAM,GAAG,MAAM,KAAK,EAAE,OAAO,IAAI,KAAK,EAAE,OAAO,CAAC,GAAG;EAChF,IAAI,EAAE,YAAY,WAAW;EAC7B,MAAM,KAAK,IAAI,OAAO,aAAa,EAAE,SAAS,GAAG,EAAE,MAAM,EAAE;EAC3D,MAAM,QAAQ,EAAE,SAAS,SAAS;EAClC,IAAI,OAAO,MAAM,KAAK,UAAU,WAAW,CAAC,KAAK,EAAE;EACnD,IAAI,EAAE,SAAS,QAAQ;GACrB,KAAK,MAAM,KAAK,EAAE,SAAS;IACzB,MAAM,KAAK,eAAe,KAAK,aAAa,CAAC,CAAC,GAAG;IACjD,KAAK,MAAM,UAAU,cAAc,EAAE,KAAK,GAAG,MAAM,KAAK,OAAO,KAAK,MAAM,GAAG;GAC/E;GACA,MAAM,KAAK,EAAE;EACf;EACA,MAAM,eAAe,MAAsB,cAAc,GAAG,CAAC,IAAI,IAAI,KAAK,EAAE,OAAO;EACnF,KAAK,MAAM,KAAK,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,MAAM,GAAG,MAAM,YAAY,CAAC,IAAI,YAAY,CAAC,CAAC,GAAG;GAC/E,IAAI,EAAE,YAAY,WAAW;GAC7B,IAAI,cAAc,GAAG,CAAC,GAAG;IACvB,IAAI,OAAO,MAAM,KAAK,UAAU,EAAE,GAAG,uBAAuB,EAAE;IAC9D,MAAM,KAAK,6CAA6C,EAAE;IAC1D,cAAc,OAAO,CAAC;IACtB,MAAM,KAAK,IAAI,YAAY;IAC3B;GACF;GACA,IAAI,OAAO,MAAM,KAAK,QAAQ,aAAa,EAAE,SAAS,GAAG,EAAE,MAAM,EAAE;GACnE,cAAc,OAAO,CAAC;EACxB;CACF;CACA,MAAM,KAAK,EAAE;CACb,OAAO,MAAM,KAAK,IAAI;AACxB;;AAGA,MAAa,kBAAkB,MAC7B,EAAE,OAAO,SAAS,WAAW,EAAE,WAAW,MAAM,MAAM,EAAE,YAAY,MAAM;AAE5E,SAAS,cAAc,OAAiB,GAAwB;CAC9D,IAAI,EAAE,SAAS,CAAC,eAAe,CAAC,GAC9B,MAAM,KACJ,KAAK,EAAE,MAAM,KAAK,KAClB,IACA,OACA,EAAE,MAAM,QAAQ,MAAM,IAAI,CAAC,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC,KAAK,IAAI,GAClD,OACA,EACF;CAEF,KAAK,MAAM,QAAQ,EAAE,OAAO;EAC1B,MAAM,KACJ,aAAa,KAAK,OAAO,IAAI,eAAe,KAAK,UAAU,EAAE,KAAK,KAAK,KAAK,KAAK,MAAM,GAAG,GAAG,GAC/F;EACA,IAAI,KAAK,OAAO,MAAM,KAAK,cAAc,KAAK,KAAK,KAAK,KAAK,GAAG,GAAG,GAAG;CACxE;CACA,MAAM,IAAI,EAAE;CACZ,MAAM,KACJ,iBAAiB,eAAe,EAAE,OAAO,EAAE,WAAW,eAAe,EAAE,OAAO,EAAE,YAAY,eAAe,EAAE,QAAQ,EAAE,WAAW,eAAe,EAAE,OAAO,GAC5J;CACA,MAAM,MAAM,EAAE;CACd,MAAM,OAAO,GAAkB,SAC7B,MAAM,OAAO,GAAG,KAAK,iBAAiB,GAAG,EAAE,GAAG;CAChD,MAAM,KACJ,YAAY,IAAI,UAAU,gBAAgB,IAAI,MAAM,sBAAsB,IAAI,IAAI,aAAa,cAAc,EAAE,KAAK,IAAI,IAAI,YAAY,SAAS,EAAE,KAAK,eAAe,IAAI,WAAW,EAAE,WAC1L;CACA,KAAK,MAAM,KAAK,EAAE,YAAY;EAC5B,MAAM,UAAU,EAAE,SAAS,SACvB,WAAW,EAAE,QAAQ,OAAO,GAAG,EAAE,QAAQ,WAAW,IAAI,UAAU,cAClE;EACJ,MAAM,KACJ,aAAa,EAAE,QAAQ,KAAK,EAAE,MAAM,QAAQ,CAAC,EAAE,cAAc,EAAE,UAAU,IAAI,EAAE,QAAQ,QAAQ,KAAK,EAAE,UACxG;CACF;CACA,KAAK,MAAM,KAAK,EAAE,QAChB,MAAM,KACJ,gBAAgB,EAAE,SAAS,MAAM,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,EAAE,IAAI,KAAK,MAAM,IAC3F;CAEF,IAAI,EAAE,aAAa,QACjB,MAAM,KACJ,cAAc,EAAE,aAAa,KAAK,MAAO,EAAE,QAAQ,KAAK,EAAE,KAAK,MAAM,EAAE,MAAM,WAAW,GAAG,EAAE,OAAO,KAAK,EAAE,KAAK,GAAI,CAAC,CAAC,KAAK,IAAI,GACjI;CAEF,MAAM,OAAO,EAAE,MAAM,GAAG,EAAE,CAAC,EAAE;CAC7B,IAAI,MAAM;EACR,MAAM,QAAQ,KAAK,SAAS,qBAAqB,GAAG,KAAK,MAAM,GAAG,kBAAkB,EAAE,KAAK;EAC3F,MAAM,KACJ,IACA,iBACA,IACA,GAAG,MACA,QAAQ,CAAC,CACT,MAAM,IAAI,CAAC,CACX,KAAK,MAAM,KAAK,GAAG,GACtB,EACF;CACF;CACA,MAAM,QAAQ;EACZ,EAAE,MAAM,YAAY,eAAe,EAAE,MAAM,SAAS,KAAK;EACzD,EAAE,MAAM,cAAc,gBAAgB,EAAE,MAAM,WAAW,KAAK;EAC9D,EAAE,MAAM,QAAQ,UAAU,EAAE,MAAM,KAAK,KAAK;EAC5C,EAAE,MAAM,cAAc,iBAAiB,EAAE,MAAM,WAAW,KAAK;CACjE,CAAC,CAAC,OAAO,OAAO;CAChB,IAAI,MAAM,QAAQ,MAAM,KAAK,YAAY,MAAM,KAAK,KAAK,GAAG;AAC9D"}
|
package/dist/testing.d.ts
CHANGED
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
import {
|
|
1
|
+
import { Ft as Wait, p as CaptureUseAction, u as CaptureInstruction } from "./types-D8RLfkFP.js";
|
|
2
2
|
//#region src/capture/fake-instructions.d.ts
|
|
3
3
|
declare const FAKE_INSTRUCTIONS: unique symbol;
|
|
4
4
|
type FakeInstructionStep = CaptureUseAction[] | {
|
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
import { s as recordFileName } from "./ids-Bp6ut7ZD.js";
|
|
2
2
|
import { n as clip } from "./text-gEcKjAyp.js";
|
|
3
|
-
import { S as parseTranscript } from "./model-
|
|
4
|
-
import { s as mergeOutcome } from "./assemble-
|
|
3
|
+
import { S as parseTranscript } from "./model-BOnlAPzs.js";
|
|
4
|
+
import { s as mergeOutcome } from "./assemble-BCZzTyR7.js";
|
|
5
5
|
import { join } from "node:path";
|
|
6
6
|
import { readFile } from "node:fs/promises";
|
|
7
7
|
//#region src/analysis/agent/tools.ts
|
|
@@ -36,7 +36,7 @@ function stepOf(event) {
|
|
|
36
36
|
return `${event.tool}${input ? `: ${input}` : ""}${event.ok ? "" : " (failed)"}`;
|
|
37
37
|
}
|
|
38
38
|
case "file-edit": return `edit (${event.kind}): ${event.path}`;
|
|
39
|
-
case "web-search": return `web-search: ${clip(event.query, 120)}`;
|
|
39
|
+
case "web-search": return `web-search: ${clip(event.query, 120)}${event.ok === false ? " (failed)" : ""}`;
|
|
40
40
|
case "web-fetch": return `web-fetch: ${event.url}${event.ok ? "" : " (failed)"}`;
|
|
41
41
|
case "error": return `error: ${clip(event.message, 200)}`;
|
|
42
42
|
case "retry": return `retry: ${clip(event.message, 200)}`;
|
|
@@ -246,4 +246,4 @@ function divergence(failed, passed, window = 6) {
|
|
|
246
246
|
//#endregion
|
|
247
247
|
export { stagedAttempts as a, listOutcomes as i, divergence as n, fromRecorded as r, attemptSummary as t };
|
|
248
248
|
|
|
249
|
-
//# sourceMappingURL=tools-
|
|
249
|
+
//# sourceMappingURL=tools-B10Uj7OI.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"tools-B10Uj7OI.js","names":[],"sources":["../src/analysis/agent/tools.ts"],"mappings":";;;;;;;AA8BA,MAAM,aAAa;CACjB;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;AACF;;AAGA,SAAS,aAAa,OAAwB;CAC5C,IAAI,OAAO,UAAU,UAAU,OAAO,KAAK,MAAM,QAAQ,QAAQ,GAAG,GAAG,GAAG;CAC1E,IAAI,OAAO,UAAU,YAAY,UAAU,MAAM,OAAO;CACxD,MAAM,SAAS;CACf,KAAK,MAAM,OAAO,YAAY;EAC5B,MAAM,QAAQ,OAAO;EACrB,IAAI,OAAO,UAAU,YAAY,OAAO,OAAO,KAAK,MAAM,QAAQ,QAAQ,GAAG,GAAG,GAAG;EACnF,IAAI,MAAM,QAAQ,KAAK,KAAK,MAAM,OAAO,MAAM,OAAO,MAAM,QAAQ,GAClE,OAAO,KAAK,MAAM,KAAK,GAAG,GAAG,GAAG;CACpC;CACA,OAAO,KAAK,KAAK,UAAU,KAAK,GAAG,GAAG;AACxC;;AAGA,SAAgB,OAAO,OAAuC;CAC5D,QAAQ,MAAM,MAAd;EACE,KAAK,QAAQ;GACX,MAAM,QAAQ,aAAa,MAAM,KAAK;GACtC,OAAO,GAAG,MAAM,OAAO,QAAQ,KAAK,UAAU,KAAK,MAAM,KAAK,KAAK;EACrE;EACA,KAAK,aACH,OAAO,SAAS,MAAM,KAAK,KAAK,MAAM;EACxC,KAAK,cACH,OAAO,eAAe,KAAK,MAAM,OAAO,GAAG,IAAI,MAAM,OAAO,QAAQ,cAAc;EACpF,KAAK,aACH,OAAO,cAAc,MAAM,MAAM,MAAM,KAAK,KAAK;EACnD,KAAK,SACH,OAAO,UAAU,KAAK,MAAM,SAAS,GAAG;EAC1C,KAAK,SACH,OAAO,UAAU,KAAK,MAAM,SAAS,GAAG;EAC1C,SACE,OAAO;CACX;AACF;AAEA,MAAM,SAAS,WAA8B,OAAO,SAAS,MAAM,OAAO,CAAC,KAAK,CAAC,CAAC;AAElF,MAAM,QAAQ,MAAc,QAAiB,KAAK,UAAU,MAAM,OAAO,IAAI,KAAK,MAAM,CAAC,GAAG;AAC5F,MAAM,QAAQ,MAAc,QAC1B,KAAK,UAAU,MAAM,OAAO,GAAG,KAAK,MAAM,GAAG,GAAG,EAAE,MAAM,KAAK,SAAS,IAAI;;AAG5E,SAAgB,WAAW,QAAoB,QAA+C;CAC5F,MAAM,UAAU,OAAO;CACvB,IAAI,CAAC,SAAS,OAAO;CACrB,MAAM,EAAE,SAAS,UAAU,aAAa,QAAQ,MAAM;CACtD,OAAO;EACL,IAAI,GAAG,QAAQ,KAAK,IAAI,GAAG,QAAQ;EACnC,MAAM,QAAQ,KAAK;EACnB,GAAG,QAAQ;EACX;EACA,OAAO,QAAQ;GAAE,MAAM,MAAM;GAAM,SAAS,MAAM;EAAQ,IAAI;EAC9D,SAAS,QAAQ,UAAU,CAAC,EAAA,CAAG,KAAK,OAAO;GACzC,SAAS,EAAE;GACX,UAAU,EAAE;GACZ,QAAQ,KAAK,GAAG,EAAE,OAAO,IAAI,EAAE,SAAS,KAAK,GAAG,GAAG;EACrD,EAAE;EACF,aAAa,QAAQ,cAAc,CAAC,EAAA,CAAG,KAAK,OAAO;GACjD,UAAU,EAAE;GACZ,SAAS,EAAE;GACX,OAAO,EAAE;EACX,EAAE;EACF,QAAQ,QAAQ,SAAS,CAAC,EAAA,CAAG,KAAK,OAAO;GACvC,QAAQ,EAAE;GACV,MAAM,EAAE;GACR,QAAQ,EAAE;GACV,OAAO,MAAM,EAAE,UAAU;EAC3B,EAAE;EACF,QAAQ,QAAQ,MAAM,SAAS,CAAC,EAAA,CAAG,KAAK,MAAM,EAAE,IAAI;EACpD,OAAO,QAAQ,MAAM,SAAS;CAChC;AACF;;AAGA,eAAsB,eACpB,YACA,SACwB;CACxB,MAAM,QAAuB,CAAC;CAC9B,KAAK,MAAM,UAAU,SAAS;EAC5B,MAAM,OAAO,KAAK,YAAY,SAAS,eAAe,OAAO,GAAG,CAAC;EACjE,IAAI,SAA4B;EAChC,IAAI;GACF,SAAS,KAAK,MAAM,MAAM,SAAS,MAAM,MAAM,CAAC;EAClD,QAAQ,CAER;EACA,MAAM,OAAO,WAAW,QAAQ,MAAM;EACtC,IAAI,MAAM,MAAM,KAAK,IAAI;CAC3B;CACA,OAAO;AACT;;AAGA,eAAsB,aAAa,QAAgB,SAA8C;CAC/F,MAAM,QAAQ,SACZ,OAAO,SAAS,KAAK,QAAQ,IAAI,GAAG,MAAM,CAAC,CAAC,YAAY,EAAE,IAAI,QAAQ,QAAQ,EAAE;CAClF,MAAM,QAAQ,gBAAgB,MAAM,KAAK,QAAQ,MAAM,YAAY,IAAI,CAAC;CACxE,OAAO;EACL,IAAI,QAAQ;EACZ,MAAM,QAAQ;EACd,GAAG,QAAQ;EACX,SAAS,QAAQ;EACjB,OAAO,QAAQ,QAAQ;GAAE,MAAM,QAAQ,MAAM;GAAM,SAAS,QAAQ,MAAM;EAAQ,IAAI;EACtF,QAAQ,MAAM,QAAQ,IACpB,QAAQ,OAAO,IAAI,OAAO,OAAO;GAC/B,SAAS,EAAE;GACX,UAAU,EAAE;GACZ,QAAQ,MAAM,MAAM,KAAK,EAAE,KAAK,IAAI,EAAA,CAAG,KAAK,GAAG,GAAG;EACpD,EAAE,CACJ;EACA,YAAY,QAAQ,WAAW,KAAK,OAAO;GACzC,UAAU,EAAE;GACZ,SAAS,EAAE;GACX,OAAO,EAAE;EACX,EAAE;EACF,OAAO,QAAQ,MAAM,KAAK,GAAG,OAAO;GAClC,QAAQ,EAAE;GACV,MAAM,EAAE;GACR,QAAQ,EAAE;GACV,OAAO,MACL,MAAM,QACH,MACC,EAAE,SAAS,KAAK,EAAE,SAAS,YAAY,EAAE,SAAS,UACtD,CACF;EACF,EAAE;EACF,OAAO,QAAQ,aAAa,KAAK,MAAM,EAAE,IAAI;EAC7C,OAAO,MAAM,KAAK,QAAQ,MAAM,MAAM,IAAI;CAC5C;AACF;;AAMA,SAAgB,eAAe,UAAuC;CACpE,MAAM,SAAS,SAAS,QAAQ,MAAM,EAAE,YAAY,MAAM;CAC1D,MAAM,QAAQ,OAAoB,GAAgB,MAChD,KAAK,IAAI,EAAE,IAAI,EAAE,CAAC,IAAI,KAAK,IAAI,EAAE,IAAI,EAAE,CAAC;CAC1C,OAAO,SACJ,QAAQ,MAAM,EAAE,YAAY,MAAM,CAAC,CACnC,SAAS,MAAM;EACd,MAAM,WAAW,OAAO,QAAQ,MAAM,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC;EACvE,MAAM,QAAQ,OAAO,QAAQ,MAAM,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC;EACpE,MAAM,QAAQ,YAAY;EAC1B,OAAO,QACH,CACE;GACE,QAAQ,EAAE;GACV,QAAQ,MAAM;GACd,OAAO,WAAY,SAAoB;EACzC,CACF,IACA,CAAC;CACP,CAAC;AACL;AAEA,MAAM,aAAa,MAAmB,EAAE,MAAM,QAAQ,GAAG,MAAM,IAAI,EAAE,MAAM,QAAQ,CAAC;;AAGpF,SAAgB,aAAa,UAAyB;CACpD,OAAO;EACL,UAAU,SAAS,KAAK,OAAO;GAC7B,IAAI,EAAE;GACN,SAAS,EAAE;GACX,OAAO,EAAE,QAAQ,GAAG,EAAE,MAAM,KAAK,IAAI,KAAK,EAAE,MAAM,SAAS,GAAG,MAAM;GACpE,cAAc,EAAE,OAAO,QAAQ,MAAM,EAAE,aAAa,CAAC,CAAC,CAAC,KAAK,MAAM,EAAE,OAAO;GAC3E,YAAY,EAAE,WAAW,KACtB,MAAM,GAAG,EAAE,QAAQ,GAAG,EAAE,MAAM,QAAQ,CAAC,EAAE,IAAI,KAAK,EAAE,UAAU,GAAG,GACpE;GACA,OAAO,EAAE,MAAM;GACf,OAAO,UAAU,CAAC;GAClB,OAAO,EAAE,MAAM;EACjB,EAAE;EACF,OAAO,eAAe,QAAQ;CAChC;AACF;;AAGA,SAAgB,eAAe,GAAgB;CAC7C,OAAO;EACL,IAAI,EAAE;EACN,SAAS,EAAE;EACX,OAAO,EAAE,QAAQ;GAAE,MAAM,EAAE,MAAM;GAAM,SAAS,KAAK,EAAE,MAAM,SAAS,GAAI;EAAE,IAAI;EAChF,OAAO,EAAE,MAAM,KAAK,GAAG,OAAO;GAC5B,MAAM,IAAI;GACV,QAAQ,KAAK,EAAE,QAAQ,GAAG;GAC1B,QAAQ,EAAE;GACV,OACE,EAAE,MAAM,SAAS,KACb;IAAC,GAAG,EAAE,MAAM,MAAM,GAAG,EAAE;IAAG,KAAK,EAAE,MAAM,SAAS,GAAG;IAAW,GAAG,EAAE,MAAM,MAAM,GAAG;GAAC,IACnF,EAAE;GACR,MAAM,KAAK,EAAE,MAAM,GAAG;EACxB,EAAE;EACF,QAAQ,EAAE,OAAO,KAAK,OAAO;GAC3B,SAAS,EAAE;GACX,UAAU,EAAE;GACZ,GAAI,EAAE,aAAa,IAAI,EAAE,QAAQ,EAAE,OAAO,IAAI,CAAC;EACjD,EAAE;EACF,YAAY,EAAE;EACd,OAAO,EAAE,MAAM,MAAM,GAAG,EAAE;EAC1B,MAAM,KAAK,EAAE,OAAO,GAAI;CAC1B;AACF;;;;;AAMA,SAAgB,WAAW,QAAqB,QAAqB,SAAS,GAAG;CAC/E,MAAM,IAAI,OAAO,MAAM,SAAS,MAAM,EAAE,KAAK;CAC7C,MAAM,IAAI,OAAO,MAAM,SAAS,MAAM,EAAE,KAAK;CAC7C,IAAI,OAAO;CACX,OAAO,OAAO,EAAE,UAAU,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,OAAO;CAClE,MAAM,WAAW,OAAO,EAAE,UAAU,OAAO,EAAE;CAC7C,MAAM,OAAO,KAAK,IAAI,GAAG,OAAO,CAAC;CACjC,MAAM,WAAW,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,OAAO,QAAQ,GAAG,OAAO,MAAM,CAAC,CAAC,KAAK,MAAM,EAAE,OAAO,CAAC,CAAC;CACxF,MAAM,QAAQ,GAAgB,YAC5B,EAAE,OAAO,MAAM,MAAM,EAAE,YAAY,OAAO,CAAC,EAAE,YAAY;CAC3D,MAAM,WAAW,GAAgB,aAC/B,EAAE,WAAW,MAAM,MAAM,EAAE,aAAa,QAAQ,CAAC,EAAE,WAAW;CAChE,MAAM,YAAY,CAChB,GAAG,IAAI,IAAI,CAAC,GAAG,OAAO,YAAY,GAAG,OAAO,UAAU,CAAC,CAAC,KAAK,MAAM,EAAE,QAAQ,CAAC,CAChF;CACA,MAAM,QAAQ,MAAmB,KAAK,EAAE,MAAM,GAAG,EAAE,CAAC,EAAE,QAAQ,IAAI,GAAG;CACrE,OAAO;EACL,QAAQ,OAAO;EACf,QAAQ,OAAO;EACf,OAAO,OAAO,QAAQ,GAAG,OAAO,MAAM,KAAK,IAAI,KAAK,OAAO,MAAM,SAAS,GAAG,MAAM;EACnF,eAAe;;EAEf,YAAY,WAAW,OAAO,IAAI;EAClC,OAAO;GAAE,QAAQ,EAAE;GAAQ,QAAQ,EAAE;EAAO;EAC5C,aAAa,EAAE,MAAM,MAAM,OAAO,MAAM;EACxC,aAAa,EAAE,MAAM,MAAM,OAAO,MAAM;EACxC,QAAQ,SACL,QAAQ,MAAM,KAAK,QAAQ,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,CAClD,KAAK,OAAO;GACX,SAAS;GACT,QAAQ,KAAK,QAAQ,CAAC;GACtB,QAAQ,KAAK,QAAQ,CAAC;GACtB,cAAc,OAAO,OAAO,MAAM,MAAM,EAAE,YAAY,CAAC,CAAC,EAAE,UAAU;EACtE,EAAE;EACJ,YAAY,UACT,QAAQ,MAAM,QAAQ,QAAQ,CAAC,MAAM,QAAQ,QAAQ,CAAC,CAAC,CAAC,CACxD,KAAK,OAAO;GACX,UAAU,KAAK,GAAG,GAAG;GACrB,QAAQ,QAAQ,QAAQ,CAAC;GACzB,QAAQ,QAAQ,QAAQ,CAAC;EAC3B,EAAE;EACJ,OAAO;GACL,YAAY,OAAO,MAAM,QAAQ,MAAM,CAAC,OAAO,MAAM,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,EAAE;GAC7E,YAAY,OAAO,MAAM,QAAQ,MAAM,CAAC,OAAO,MAAM,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,EAAE;EAC/E;EACA,WAAW;GAAE,QAAQ,KAAK,MAAM;GAAG,QAAQ,KAAK,MAAM;EAAE;EACxD,MAAM;GAAE,QAAQ,KAAK,OAAO,OAAO,IAAI;GAAG,QAAQ,KAAK,OAAO,OAAO,IAAI;EAAE;CAC7E;AACF"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"types-C5vI3g1C.js","names":[],"sources":["../src/sandbox/types.ts"],"mappings":";;
|
|
1
|
+
{"version":3,"file":"types-C5vI3g1C.js","names":[],"sources":["../src/sandbox/types.ts"],"mappings":";;AA0CA,IAAa,aAAb,cAAgC,MAAM;CACpC,OAAgB;AAClB"}
|
|
@@ -273,11 +273,11 @@ type AttemptRecord = {
|
|
|
273
273
|
usage: TokenUsage | null;
|
|
274
274
|
};
|
|
275
275
|
timings: PhaseTimings;
|
|
276
|
-
/** Counts from the transcript, so a run's shape is visible without reading it. */
|
|
276
|
+
/** Counts from the transcript, so a run's shape is visible without reading it. `webSearches` includes refused searches. */
|
|
277
277
|
activity: {
|
|
278
278
|
tools: number;
|
|
279
279
|
fileEdits: number;
|
|
280
|
-
/** Null when a turn with none couldn't report them (`unreported`)
|
|
280
|
+
/** Every `web-search` event, including refused ones. Null when a turn with none couldn't report them (`unreported`). */
|
|
281
281
|
webSearches: number | null;
|
|
282
282
|
webFetches: number | null;
|
|
283
283
|
reasoningMs: number;
|
|
@@ -291,6 +291,8 @@ type AttemptRecord = {
|
|
|
291
291
|
durationMs: number;
|
|
292
292
|
log?: FileRef;
|
|
293
293
|
}>;
|
|
294
|
+
/** Every `expect` and `expect.soft` matcher call that settled, in call order; unset with none. */
|
|
295
|
+
assertions?: AssertionRecord[];
|
|
294
296
|
/** Files under a dependency directory (node_modules/, .venv/, …) collapse to one entry with `count`. */
|
|
295
297
|
changedFiles: Array<FileChange & {
|
|
296
298
|
count?: number;
|
|
@@ -325,7 +327,7 @@ type EvalEntry = {
|
|
|
325
327
|
outcome: Outcome;
|
|
326
328
|
/** A placeholder, `evaluate.todo` or `group.todo`: it never runs, and its outcome is `skipped`. */
|
|
327
329
|
todo: boolean;
|
|
328
|
-
/** pass / (pass + fail) over the gated cells' attempts; infra errors and skips excluded. */
|
|
330
|
+
/** pass@1, pass / (pass + fail), over the gated cells' attempts; infra errors and skips excluded. */
|
|
329
331
|
passRate: number | null;
|
|
330
332
|
cells: CellRecord[];
|
|
331
333
|
attempts: AttemptRecord[];
|
|
@@ -336,8 +338,112 @@ type EvalEntry = {
|
|
|
336
338
|
* (rates by cut, votes, ties, a failed call, a sample of reasons).
|
|
337
339
|
*/
|
|
338
340
|
answers?: CompareAnswer[];
|
|
341
|
+
/** Set on each case of a pooled `evaluate.each`; `RunRecord.families` has the suite. */
|
|
342
|
+
family?: {
|
|
343
|
+
id: string;
|
|
344
|
+
name: string;
|
|
345
|
+
case: number;
|
|
346
|
+
cases: number;
|
|
347
|
+
};
|
|
348
|
+
};
|
|
349
|
+
/** A 95% interval on a pooled rate or difference. */
|
|
350
|
+
type FamilyInterval = {
|
|
351
|
+
lo: number;
|
|
352
|
+
hi: number;
|
|
353
|
+
level: 0.95;
|
|
354
|
+
/**
|
|
355
|
+
* `wilson` / `newcombe-paired`: one counted attempt a case and side, so each case is one
|
|
356
|
+
* Bernoulli trial (or pair). `t-cases`: several, so a t interval over the cases' pass@1, with
|
|
357
|
+
* attempts within a case never counted as independent.
|
|
358
|
+
*/
|
|
359
|
+
method: 'wilson' | 'newcombe-paired' | 't-cases';
|
|
360
|
+
};
|
|
361
|
+
/** One cell's pass rate over a family's cases. */
|
|
362
|
+
type FamilyRate = {
|
|
363
|
+
/** The cell's key, the same in every case. */
|
|
364
|
+
cell: string;
|
|
365
|
+
/** Cases with at least one counted attempt in this cell. */
|
|
366
|
+
cases: number;
|
|
367
|
+
/** Passes and counted attempts over those cases. */
|
|
368
|
+
pass: number;
|
|
369
|
+
counted: number;
|
|
370
|
+
/** The mean of the cases' pass@1; at one attempt a case, `pass / counted`. */
|
|
371
|
+
rate: number;
|
|
372
|
+
interval: FamilyInterval;
|
|
373
|
+
};
|
|
374
|
+
/**
|
|
375
|
+
* `lift` / `drop`: the interval excludes 0. `noise`: it doesn't. `insufficient`: fewer than 6
|
|
376
|
+
* discordant pairs (one attempt a case) or 6 paired cases (several), where no test can decide.
|
|
377
|
+
*/
|
|
378
|
+
type FamilyVerdict = 'lift' | 'drop' | 'noise' | 'insufficient';
|
|
379
|
+
/** One effort metric over a family's paired cases. */
|
|
380
|
+
type FamilyEffortMetric = {
|
|
381
|
+
/** Cases with a value on both sides. */
|
|
382
|
+
pairs: number;
|
|
383
|
+
/** The median over cases of the variant's per-case median, over the control's; null at 0. */
|
|
384
|
+
ratio: number | null;
|
|
385
|
+
/** Wilcoxon signed-rank over the per-case differences; null when every difference is 0. */
|
|
386
|
+
p: number | null;
|
|
387
|
+
/** `insufficient` below 6 non-zero differences, where the exact test can't reach 0.05. */
|
|
388
|
+
verdict: 'fewer' | 'more' | 'noise' | 'insufficient';
|
|
389
|
+
};
|
|
390
|
+
/** The variant's effort against the control's, paired by case. */
|
|
391
|
+
type FamilyEffort = {
|
|
392
|
+
tools: FamilyEffortMetric;
|
|
393
|
+
agentMs: FamilyEffortMetric;
|
|
394
|
+
/** Over the cases where every counted attempt reported usage; omitted when none did. */
|
|
395
|
+
tokens?: FamilyEffortMetric;
|
|
396
|
+
};
|
|
397
|
+
/** A variant cell against its control over a family's cases, paired by case. */
|
|
398
|
+
type FamilyLift = {
|
|
399
|
+
/** The variant cell's key. */
|
|
400
|
+
variant: string;
|
|
401
|
+
/** The control cell's key. */
|
|
402
|
+
vs: string;
|
|
403
|
+
/** Cases with counted attempts on both sides. */
|
|
404
|
+
pairs: number;
|
|
405
|
+
/** Mean pass@1 over the paired cases on each side. */
|
|
406
|
+
variantRate: number;
|
|
407
|
+
controlRate: number;
|
|
408
|
+
/** Set at one attempt a case and side: pairs where only one side passed. */
|
|
409
|
+
discordant?: {
|
|
410
|
+
variantOnly: number;
|
|
411
|
+
controlOnly: number;
|
|
412
|
+
};
|
|
413
|
+
/** Variant minus control, from -1 to 1. */
|
|
414
|
+
points: number;
|
|
415
|
+
interval: FamilyInterval;
|
|
416
|
+
/** McNemar's exact p (one attempt a case), else Wilcoxon signed-rank over the cases. */
|
|
417
|
+
p: number | null;
|
|
418
|
+
/** The interval spans 0. */
|
|
419
|
+
withinNoise: boolean;
|
|
420
|
+
verdict: FamilyVerdict;
|
|
421
|
+
/** The compare's `minLift`: on a family it gates this lower bound, not each case's. */
|
|
422
|
+
minLift?: number;
|
|
423
|
+
/** `fail` when `interval.lo` is below `minLift`. */
|
|
424
|
+
outcome: 'pass' | 'fail';
|
|
425
|
+
effort?: FamilyEffort;
|
|
339
426
|
};
|
|
340
|
-
/** A
|
|
427
|
+
/** A pooled `evaluate.each`: its cases as samples of one suite. Each case is also an eval. */
|
|
428
|
+
type FamilyRecord = {
|
|
429
|
+
/** `file › group › … › template`. */
|
|
430
|
+
id: string;
|
|
431
|
+
file: string;
|
|
432
|
+
group: string[];
|
|
433
|
+
/** The name template, e.g. `fixes $id`. */
|
|
434
|
+
name: string;
|
|
435
|
+
/** Cases in the list. */
|
|
436
|
+
cases: number;
|
|
437
|
+
/** Cases with an eval in this run; fewer when a filter or tag left some out. */
|
|
438
|
+
ran: number;
|
|
439
|
+
/** One per cell. */
|
|
440
|
+
rates: FamilyRate[];
|
|
441
|
+
/** One per variant cell, with `compare`. */
|
|
442
|
+
lifts: FamilyLift[];
|
|
443
|
+
/** `fail` when a lift is below its `minLift`. The cases still pass or fail on their own. */
|
|
444
|
+
outcome: 'pass' | 'fail';
|
|
445
|
+
};
|
|
446
|
+
/** A variant cell's pass@1 against the control's, on the same harness and sandbox. */
|
|
341
447
|
type Lift = {
|
|
342
448
|
/** The control cell's key. */
|
|
343
449
|
vs: string;
|
|
@@ -350,7 +456,7 @@ type Lift = {
|
|
|
350
456
|
pass: number;
|
|
351
457
|
counted: number;
|
|
352
458
|
};
|
|
353
|
-
/** Variant pass
|
|
459
|
+
/** Variant pass@1 minus control pass@1, from -1 to 1. */
|
|
354
460
|
points: number;
|
|
355
461
|
/** 95% interval on the difference, by Newcombe's hybrid score method from two Wilson intervals. */
|
|
356
462
|
interval: {
|
|
@@ -402,14 +508,39 @@ type CellRecord = Cell & {
|
|
|
402
508
|
fail: number;
|
|
403
509
|
infraError: number;
|
|
404
510
|
skipped: number;
|
|
511
|
+
/** pass@1: pass / (pass + fail); infra errors and skips excluded. */
|
|
405
512
|
passRate: number | null;
|
|
406
513
|
/**
|
|
407
|
-
* By `gate`. `'all'` matches `{ passRate: 1 }
|
|
514
|
+
* By `gate`. `'all'` (pass^k: every counted attempt passes) matches `{ passRate: 1 }`;
|
|
515
|
+
* `{ passRate }` is a pass@1 threshold. Infra only when nothing counted; `'none'`
|
|
408
516
|
* reports the worst outcome. A variant below its `minLift` fails too.
|
|
409
517
|
*/
|
|
410
518
|
outcome: Outcome;
|
|
411
519
|
lift?: Lift;
|
|
520
|
+
/**
|
|
521
|
+
* Each check (assertion label) seen in any attempt of the eval, over this cell's counted
|
|
522
|
+
* attempts; on a variant cell, the control's beside it. Unset when no attempt recorded any.
|
|
523
|
+
*/
|
|
524
|
+
checks?: CellCheck[];
|
|
412
525
|
};
|
|
526
|
+
/** How often one check passed: a counted attempt that never reached it counts it `notRun`. */
|
|
527
|
+
type CheckTally = {
|
|
528
|
+
pass: number;
|
|
529
|
+
fail: number;
|
|
530
|
+
error: number;
|
|
531
|
+
notRun: number;
|
|
532
|
+
};
|
|
533
|
+
/**
|
|
534
|
+
* One check over a cell's counted attempts. `verdict` (variant cells): `discriminates` when it
|
|
535
|
+
* passes on every counted attempt of one side and none of the other; `both-pass` or `both-fail`
|
|
536
|
+
* when it does the same on both sides, so it can't tell them apart; `mixed` otherwise.
|
|
537
|
+
*/
|
|
538
|
+
type CellCheck = CheckTally & {
|
|
539
|
+
label: string;
|
|
540
|
+
control?: CheckTally;
|
|
541
|
+
verdict?: CheckVerdict;
|
|
542
|
+
};
|
|
543
|
+
type CheckVerdict = 'discriminates' | 'both-pass' | 'both-fail' | 'mixed';
|
|
413
544
|
/** Counts attempts, except `evals`. */
|
|
414
545
|
type RunSummary = {
|
|
415
546
|
evals: number;
|
|
@@ -508,6 +639,8 @@ type RunRecord = {
|
|
|
508
639
|
};
|
|
509
640
|
summary: RunSummary;
|
|
510
641
|
evals: EvalEntry[];
|
|
642
|
+
/** Each pooled `evaluate.each` that ran; unset when none did. */
|
|
643
|
+
families?: FamilyRecord[];
|
|
511
644
|
};
|
|
512
645
|
/** What ran and graded the agents: the machine, images, harnesses and judge. */
|
|
513
646
|
type Stack = {
|
|
@@ -646,6 +779,8 @@ type HistoryEntry = {
|
|
|
646
779
|
passRate: number | null;
|
|
647
780
|
cells: HistoryCell[];
|
|
648
781
|
}>;
|
|
782
|
+
/** The run's pooled families; unset when none ran. */
|
|
783
|
+
families?: Array<Pick<FamilyRecord, 'id' | 'name' | 'cases' | 'ran' | 'outcome'>>;
|
|
649
784
|
};
|
|
650
785
|
//#endregion
|
|
651
786
|
//#region src/types.d.ts
|
|
@@ -680,6 +815,8 @@ type TranscriptEvent = EventTime & ({
|
|
|
680
815
|
query: string;
|
|
681
816
|
/** How many sources the harness says it used; Cursor sends no URLs or titles. */
|
|
682
817
|
references: number | null;
|
|
818
|
+
/** Whether the search succeeded. Absent means it did. */
|
|
819
|
+
ok?: boolean;
|
|
683
820
|
durationMs: number | null;
|
|
684
821
|
} | {
|
|
685
822
|
type: 'web-fetch';
|
|
@@ -1000,6 +1137,12 @@ type AttemptContext = {
|
|
|
1000
1137
|
capture(target: number | ServeHandle, options?: CaptureCall): Promise<Capture>;
|
|
1001
1138
|
/** Workspace-relative, or `~/` for the home directory; `null` when there's no such file. */
|
|
1002
1139
|
readFile(path: string): Promise<string | null>;
|
|
1140
|
+
/**
|
|
1141
|
+
* Writes a file into the sandbox (workspace-relative, or `~/`), creating its directories.
|
|
1142
|
+
* Later turns' `diff` leaves it out, as it does the environment's files, but shows any change
|
|
1143
|
+
* made to it afterwards. Needs a sandbox, same as `c.exec`.
|
|
1144
|
+
*/
|
|
1145
|
+
writeFile(path: string, contents: string | Uint8Array): Promise<void>;
|
|
1003
1146
|
/**
|
|
1004
1147
|
* Starts the sandbox and the agent session without a turn, for checks on what the environment
|
|
1005
1148
|
* set up. `c.send()` starts them too; calling both starts them once.
|
|
@@ -1125,16 +1268,18 @@ type CursorHarnessDescriptor = {
|
|
|
1125
1268
|
*/
|
|
1126
1269
|
startTimeout?: Duration;
|
|
1127
1270
|
};
|
|
1271
|
+
/** Reasoning effort the Claude Code Agent SDK `/effort` control takes. */
|
|
1272
|
+
type ClaudeCodeEffort = 'low' | 'medium' | 'high' | 'xhigh' | 'max';
|
|
1128
1273
|
/** Claude Code through `@ai-sdk/harness-claude-code`, made by `claudeCode()`. */
|
|
1129
1274
|
type ClaudeCodeHarnessDescriptor = {
|
|
1130
1275
|
kind: 'claude-code';
|
|
1131
1276
|
/** A model id or alias the `claude` CLI accepts, e.g. `claude-sonnet-5` or `sonnet`. */
|
|
1132
1277
|
model: string;
|
|
1133
1278
|
/**
|
|
1134
|
-
* Reasoning effort the Agent SDK `/effort` control takes
|
|
1279
|
+
* Reasoning effort the Agent SDK `/effort` control takes: `low`, `medium`, `high`,
|
|
1135
1280
|
* `xhigh` (Extra in the UI) or `max`.
|
|
1136
1281
|
*/
|
|
1137
|
-
effort?:
|
|
1282
|
+
effort?: ClaudeCodeEffort;
|
|
1138
1283
|
/** Cell id; defaults to `claude-code:<model>`, or `claude-code:<model>[effort=<value>]` when `effort` is set. */
|
|
1139
1284
|
id?: string;
|
|
1140
1285
|
/** The most Claude Code attempts at once, across the run, e.g. for the account's rate limits. */
|
|
@@ -1296,8 +1441,13 @@ type DockerSandboxDescriptor = {
|
|
|
1296
1441
|
root?: boolean;
|
|
1297
1442
|
/** Linux capabilities to add back, e.g. `['NET_RAW']`; every other one is dropped. */
|
|
1298
1443
|
capabilities?: string[];
|
|
1299
|
-
/** The container's memory
|
|
1444
|
+
/** The container's memory: `'4g'` (default), `'512m'`, or bytes. */
|
|
1300
1445
|
memory?: string | number;
|
|
1446
|
+
/**
|
|
1447
|
+
* Swap on top of `memory`, in the same units. Default: as much as `memory`; `0` makes `memory`
|
|
1448
|
+
* the hard ceiling. A host without swap has none to give.
|
|
1449
|
+
*/
|
|
1450
|
+
swap?: string | number;
|
|
1301
1451
|
/** CPUs the container may use, fractions allowed. Default 2, capped at the Docker host's. */
|
|
1302
1452
|
cpus?: number;
|
|
1303
1453
|
/** Processes and threads at once. Default 1024. */
|
|
@@ -1671,7 +1821,8 @@ type CompareAnswer = {
|
|
|
1671
1821
|
type PinnableKey = 'harness' | 'runs' | 'compare';
|
|
1672
1822
|
/**
|
|
1673
1823
|
* How a cell's attempts decide its outcome: `'all'` (the default) passes only when every attempt
|
|
1674
|
-
* passes; `{ passRate }` passes when
|
|
1824
|
+
* passes (pass^k); `{ passRate }` passes when its pass@1 (passes over passes and fails) is at
|
|
1825
|
+
* least that share.
|
|
1675
1826
|
*/
|
|
1676
1827
|
type GatePolicy = 'all' | {
|
|
1677
1828
|
passRate: number;
|
|
@@ -1921,11 +2072,23 @@ type PrepareFailureRecord = {
|
|
|
1921
2072
|
log: string | null;
|
|
1922
2073
|
setBy: string[];
|
|
1923
2074
|
};
|
|
1924
|
-
/**
|
|
2075
|
+
/**
|
|
2076
|
+
* The sandbox ran out of a resource during the attempt: `oom` is the kernel killing `kills`
|
|
2077
|
+
* processes at its memory limit. The attempt is then an infra error, whatever its checks said.
|
|
2078
|
+
*/
|
|
2079
|
+
type ResourceFailureRecord = {
|
|
2080
|
+
cause: 'oom';
|
|
2081
|
+
kills: number;
|
|
2082
|
+
};
|
|
2083
|
+
/**
|
|
2084
|
+
* Why an attempt failed; `prepare` when its sandbox couldn't be prepared, `resource` when it ran
|
|
2085
|
+
* out of one.
|
|
2086
|
+
*/
|
|
1925
2087
|
type AttemptError = {
|
|
1926
2088
|
kind: ErrorKind;
|
|
1927
2089
|
message: string;
|
|
1928
2090
|
prepare?: PrepareFailureRecord;
|
|
2091
|
+
resource?: ResourceFailureRecord;
|
|
1929
2092
|
};
|
|
1930
2093
|
/** Where an eval's time went, in ms. `agentMs` is the harness turns; `diffMs` reading changes. */
|
|
1931
2094
|
type PhaseTimings = {
|
|
@@ -2060,6 +2223,25 @@ type StartRetry = {
|
|
|
2060
2223
|
/** How long it waited before giving up on that start. */
|
|
2061
2224
|
waitedMs: number;
|
|
2062
2225
|
};
|
|
2226
|
+
/** One `expect` or `expect.soft` matcher call an attempt made. */
|
|
2227
|
+
type AssertionRecord = {
|
|
2228
|
+
/**
|
|
2229
|
+
* `expect`'s second argument, or `<matcher> <first argument>` (or the command `c.exec` ran, or
|
|
2230
|
+
* the judge's question), cut at 60 characters; ` #2`, ` #3` on repeats within the attempt.
|
|
2231
|
+
*/
|
|
2232
|
+
label: string;
|
|
2233
|
+
/** The eval gave the label. */
|
|
2234
|
+
labelled: boolean;
|
|
2235
|
+
/** `toHaveFile`, `not.toBe`, `resolves.toBe`, `poll.toBe`. */
|
|
2236
|
+
matcher: string;
|
|
2237
|
+
/** `error`: the matcher couldn't answer (a subject it can't read, an event the harness doesn't report). */
|
|
2238
|
+
status: 'pass' | 'fail' | 'error';
|
|
2239
|
+
soft: boolean;
|
|
2240
|
+
/** Why it failed or couldn't answer, cut at 1 KB. */
|
|
2241
|
+
message?: string;
|
|
2242
|
+
/** Async matchers only (`toHaveFile`, `toSucceed`, `.resolves`, `expect.poll`). */
|
|
2243
|
+
durationMs?: number;
|
|
2244
|
+
};
|
|
2063
2245
|
/** Which environment ran as the control, so the studio can show it. */
|
|
2064
2246
|
type EvalCompareMeta = {
|
|
2065
2247
|
baseline: string;
|
|
@@ -2087,5 +2269,5 @@ type SandboxWarning = {
|
|
|
2087
2269
|
message: string;
|
|
2088
2270
|
};
|
|
2089
2271
|
//#endregion
|
|
2090
|
-
export {
|
|
2091
|
-
//# sourceMappingURL=types-
|
|
2272
|
+
export { HarnessData as $, EffortContrast as $t, DirectoryOptions as A, Turn as At, EsConfig as B, EvalEntry as Bt, CompareCut as C, Secret as Ct, ConfigJudgeOptions as D, SkillSource as Dt, ConfigHelpers as E, SetupStep as Et, EnvVars as F, Wait as Ft, FakeHarnessDescriptor as G, HistoryCell as Gt, EvalOptions as H, FamilyRate as Ht, Environment as I, WebEventType as It, FakeSandboxDescriptor as J, Lift as Jt, FakeJudgeModel as K, HistoryEntry as Kt, EnvironmentHelpers as L, ArtifactRef as Lt, Duration as M, TurnStatus as Mt, EgressPolicy as N, Verdict as Nt, CursorHarnessDescriptor as O, TokenUsage as Ot, EgressRecord as P, Viewport as Pt, GroupOptions as Q, EFFORT_ALPHA as Qt, EnvironmentInput as R, AttemptRecord as Rt, CompareAnswer as S, SandboxWarning as St, CompareQuestion as T, ServeHandle as Tt, ExecResult as U, FamilyRecord as Ut, EvalFn as V, FamilyLift as Vt, FakeAnalysisModel as W, FileRef as Wt, FakeVote as X, StackSummary as Xt, FakeScriptStep as Y, RunRecord as Yt, GooseHarnessDescriptor as Z, CellEffort as Zt, Cell as _, formatSeconds as _n, PinnableKey as _t, BackgroundHandle as a, cellContrast as an, JudgeArtifactRef as at, ClaudeCodeHarnessDescriptor as b, ProfileLimit as bt, Capture as c, contrastEffort as cn, JudgeModel as ct, CaptureOptions as d, effortOf as dn, McpServer as dt, EffortSample as en, HarnessDescriptor as et, CaptureUse as f, effortSample as fn, OpenCodeHarnessDescriptor as ft, CaptureWindow as g, formatRatio as gn, PiHarnessDescriptor as gt, CaptureWait as h, formatCount as hn, PairVote as ht, AttemptContext as i, VERDICT_TEXT as in, HostFile as it, DockerSandboxDescriptor as j, TurnData as jt, Diff as k, TranscriptEvent as kt, CaptureCall as l, effortChangeLabel as ln, JudgeOptions as lt, CaptureUseScript as m, effortVerdict as mn, PairFactor as mt, AnalysisConfig as n, Spread as nn, HarnessOption as nt, BlindPair as o, cellEffort as on, JudgeCallOptions as ot, CaptureUseAction as p, effortSummary as pn, Outcome as pt, FakeNoteInput as q, JudgementRecord as qt, AntigravityHarnessDescriptor as r, TOKENS_UNREPORTED as rn, HistoryConfig as rt, Budget as s, cellSamples as sn, JudgeFailure as st, AgentHarnessDescriptor as t, EffortVerdict as tn, HarnessEdit as tt, CaptureInstruction as u, effortLabel as un, JudgeState as ut, CellRole as v, tokenValues as vn, Profile as vt, CompareOptions as w, SecretTemplate as wt, CodexHarnessDescriptor as x, SandboxDescriptor as xt, ClaudeCodeEffort as y, ProfileDefaults as yt, EnvironmentSpec as z, CellRecord as zt };
|
|
2273
|
+
//# sourceMappingURL=types-D8RLfkFP.d.ts.map
|