@namzu/sdk 62.12.0 → 62.14.0

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.
Files changed (248) hide show
  1. package/CHANGELOG.md +31 -0
  2. package/dist/agents/file-definitions.d.ts +7 -0
  3. package/dist/agents/file-definitions.d.ts.map +1 -1
  4. package/dist/agents/file-definitions.js +20 -0
  5. package/dist/agents/file-definitions.js.map +1 -1
  6. package/dist/agents/index.d.ts +1 -1
  7. package/dist/agents/index.d.ts.map +1 -1
  8. package/dist/agents/index.js +1 -1
  9. package/dist/agents/index.js.map +1 -1
  10. package/dist/bridge/acp/server.d.ts +8 -1
  11. package/dist/bridge/acp/server.d.ts.map +1 -1
  12. package/dist/bridge/acp/server.js +12 -0
  13. package/dist/bridge/acp/server.js.map +1 -1
  14. package/dist/connector/index.d.ts +1 -1
  15. package/dist/connector/index.d.ts.map +1 -1
  16. package/dist/connector/index.js +1 -1
  17. package/dist/connector/index.js.map +1 -1
  18. package/dist/connector/mcp/client.d.ts +96 -0
  19. package/dist/connector/mcp/client.d.ts.map +1 -1
  20. package/dist/connector/mcp/client.js +488 -13
  21. package/dist/connector/mcp/client.js.map +1 -1
  22. package/dist/connector/mcp/index.d.ts +1 -0
  23. package/dist/connector/mcp/index.d.ts.map +1 -1
  24. package/dist/connector/mcp/index.js +1 -0
  25. package/dist/connector/mcp/index.js.map +1 -1
  26. package/dist/connector/mcp/tasks.d.ts +40 -0
  27. package/dist/connector/mcp/tasks.d.ts.map +1 -0
  28. package/dist/connector/mcp/tasks.js +55 -0
  29. package/dist/connector/mcp/tasks.js.map +1 -0
  30. package/dist/constants/data-epoch.d.ts +18 -0
  31. package/dist/constants/data-epoch.d.ts.map +1 -0
  32. package/dist/constants/data-epoch.js +18 -0
  33. package/dist/constants/data-epoch.js.map +1 -0
  34. package/dist/constants/index.d.ts +1 -0
  35. package/dist/constants/index.d.ts.map +1 -1
  36. package/dist/constants/index.js +1 -0
  37. package/dist/constants/index.js.map +1 -1
  38. package/dist/constants/mcp/index.d.ts +1 -1
  39. package/dist/constants/mcp/index.d.ts.map +1 -1
  40. package/dist/constants/mcp/index.js +5 -0
  41. package/dist/constants/mcp/index.js.map +1 -1
  42. package/dist/constants/telemetry/index.d.ts +12 -0
  43. package/dist/constants/telemetry/index.d.ts.map +1 -1
  44. package/dist/constants/telemetry/index.js +12 -0
  45. package/dist/constants/telemetry/index.js.map +1 -1
  46. package/dist/eval/bootstrap.d.ts +122 -0
  47. package/dist/eval/bootstrap.d.ts.map +1 -0
  48. package/dist/eval/bootstrap.js +227 -0
  49. package/dist/eval/bootstrap.js.map +1 -0
  50. package/dist/eval/compare.d.ts +109 -0
  51. package/dist/eval/compare.d.ts.map +1 -0
  52. package/dist/eval/compare.js +324 -0
  53. package/dist/eval/compare.js.map +1 -0
  54. package/dist/eval/disclosure.d.ts +70 -0
  55. package/dist/eval/disclosure.d.ts.map +1 -0
  56. package/dist/eval/disclosure.js +64 -0
  57. package/dist/eval/disclosure.js.map +1 -0
  58. package/dist/eval/experiment.d.ts +23 -0
  59. package/dist/eval/experiment.d.ts.map +1 -1
  60. package/dist/eval/experiment.js +87 -15
  61. package/dist/eval/experiment.js.map +1 -1
  62. package/dist/eval/from-turn.d.ts.map +1 -1
  63. package/dist/eval/from-turn.js +5 -0
  64. package/dist/eval/from-turn.js.map +1 -1
  65. package/dist/eval/index.d.ts +13 -2
  66. package/dist/eval/index.d.ts.map +1 -1
  67. package/dist/eval/index.js +6 -1
  68. package/dist/eval/index.js.map +1 -1
  69. package/dist/eval/report-stats.d.ts +77 -0
  70. package/dist/eval/report-stats.d.ts.map +1 -0
  71. package/dist/eval/report-stats.js +184 -0
  72. package/dist/eval/report-stats.js.map +1 -0
  73. package/dist/eval/run-options.d.ts +30 -0
  74. package/dist/eval/run-options.d.ts.map +1 -0
  75. package/dist/eval/run-options.js +20 -0
  76. package/dist/eval/run-options.js.map +1 -0
  77. package/dist/eval/run-summary.d.ts +36 -0
  78. package/dist/eval/run-summary.d.ts.map +1 -0
  79. package/dist/eval/run-summary.js +57 -0
  80. package/dist/eval/run-summary.js.map +1 -0
  81. package/dist/eval/scorers.d.ts +32 -0
  82. package/dist/eval/scorers.d.ts.map +1 -1
  83. package/dist/eval/scorers.js +77 -0
  84. package/dist/eval/scorers.js.map +1 -1
  85. package/dist/eval/types.d.ts +74 -0
  86. package/dist/eval/types.d.ts.map +1 -1
  87. package/dist/manager/session/turn-recorder.d.ts +3 -0
  88. package/dist/manager/session/turn-recorder.d.ts.map +1 -1
  89. package/dist/manager/session/turn-recorder.js +18 -2
  90. package/dist/manager/session/turn-recorder.js.map +1 -1
  91. package/dist/plugin/index.d.ts +1 -1
  92. package/dist/plugin/index.d.ts.map +1 -1
  93. package/dist/plugin/index.js +1 -1
  94. package/dist/plugin/index.js.map +1 -1
  95. package/dist/plugin/loader.d.ts +11 -0
  96. package/dist/plugin/loader.d.ts.map +1 -1
  97. package/dist/plugin/loader.js +15 -2
  98. package/dist/plugin/loader.js.map +1 -1
  99. package/dist/public-runtime.d.ts +10 -6
  100. package/dist/public-runtime.d.ts.map +1 -1
  101. package/dist/public-runtime.js +13 -5
  102. package/dist/public-runtime.js.map +1 -1
  103. package/dist/public-types.d.ts +2 -0
  104. package/dist/public-types.d.ts.map +1 -1
  105. package/dist/registry/command/kernel-commands.d.ts +3 -0
  106. package/dist/registry/command/kernel-commands.d.ts.map +1 -1
  107. package/dist/registry/command/kernel-commands.js +76 -5
  108. package/dist/registry/command/kernel-commands.js.map +1 -1
  109. package/dist/runtime/query/cancelled-before-start.d.ts.map +1 -1
  110. package/dist/runtime/query/cancelled-before-start.js +2 -0
  111. package/dist/runtime/query/cancelled-before-start.js.map +1 -1
  112. package/dist/runtime/query/index.d.ts.map +1 -1
  113. package/dist/runtime/query/index.js +17 -0
  114. package/dist/runtime/query/index.js.map +1 -1
  115. package/dist/runtime/query/iteration/index.d.ts +18 -1
  116. package/dist/runtime/query/iteration/index.d.ts.map +1 -1
  117. package/dist/runtime/query/iteration/index.js +80 -1
  118. package/dist/runtime/query/iteration/index.js.map +1 -1
  119. package/dist/runtime/query/iteration/phases/tool-review.d.ts.map +1 -1
  120. package/dist/runtime/query/iteration/phases/tool-review.js +18 -1
  121. package/dist/runtime/query/iteration/phases/tool-review.js.map +1 -1
  122. package/dist/runtime/query/iteration/stream-turn.d.ts.map +1 -1
  123. package/dist/runtime/query/iteration/stream-turn.js +5 -0
  124. package/dist/runtime/query/iteration/stream-turn.js.map +1 -1
  125. package/dist/runtime/query/result.d.ts.map +1 -1
  126. package/dist/runtime/query/result.js +2 -0
  127. package/dist/runtime/query/result.js.map +1 -1
  128. package/dist/runtime/query/turn/answer-claims.d.ts +52 -0
  129. package/dist/runtime/query/turn/answer-claims.d.ts.map +1 -0
  130. package/dist/runtime/query/turn/answer-claims.js +348 -0
  131. package/dist/runtime/query/turn/answer-claims.js.map +1 -0
  132. package/dist/runtime/query/turn/outcome-ledger.d.ts +7 -0
  133. package/dist/runtime/query/turn/outcome-ledger.d.ts.map +1 -1
  134. package/dist/runtime/query/turn/outcome-ledger.js +13 -0
  135. package/dist/runtime/query/turn/outcome-ledger.js.map +1 -1
  136. package/dist/skills/index.d.ts +1 -1
  137. package/dist/skills/index.d.ts.map +1 -1
  138. package/dist/skills/index.js +1 -1
  139. package/dist/skills/index.js.map +1 -1
  140. package/dist/skills/loader.d.ts +15 -1
  141. package/dist/skills/loader.d.ts.map +1 -1
  142. package/dist/skills/loader.js +17 -3
  143. package/dist/skills/loader.js.map +1 -1
  144. package/dist/store/goal/index.d.ts +3 -1
  145. package/dist/store/goal/index.d.ts.map +1 -1
  146. package/dist/store/goal/index.js +6 -0
  147. package/dist/store/goal/index.js.map +1 -1
  148. package/dist/telemetry/genai.d.ts +98 -0
  149. package/dist/telemetry/genai.d.ts.map +1 -0
  150. package/dist/telemetry/genai.js +189 -0
  151. package/dist/telemetry/genai.js.map +1 -0
  152. package/dist/telemetry/runtime-accessors.d.ts +19 -0
  153. package/dist/telemetry/runtime-accessors.d.ts.map +1 -1
  154. package/dist/telemetry/runtime-accessors.js +43 -1
  155. package/dist/telemetry/runtime-accessors.js.map +1 -1
  156. package/dist/tools/goal/confirmation.d.ts +154 -0
  157. package/dist/tools/goal/confirmation.d.ts.map +1 -0
  158. package/dist/tools/goal/confirmation.js +290 -0
  159. package/dist/tools/goal/confirmation.js.map +1 -0
  160. package/dist/tools/goal/index.d.ts +4 -1
  161. package/dist/tools/goal/index.d.ts.map +1 -1
  162. package/dist/tools/goal/index.js +27 -2
  163. package/dist/tools/goal/index.js.map +1 -1
  164. package/dist/tools/hand-back/index.d.ts +48 -0
  165. package/dist/tools/hand-back/index.d.ts.map +1 -0
  166. package/dist/tools/hand-back/index.js +129 -0
  167. package/dist/tools/hand-back/index.js.map +1 -0
  168. package/dist/tools/schedules/schedule-tool.d.ts.map +1 -1
  169. package/dist/tools/schedules/schedule-tool.js +15 -1
  170. package/dist/tools/schedules/schedule-tool.js.map +1 -1
  171. package/dist/tools/schedules/types.d.ts +10 -0
  172. package/dist/tools/schedules/types.d.ts.map +1 -1
  173. package/dist/toolsets/manager.d.ts.map +1 -1
  174. package/dist/toolsets/manager.js +12 -0
  175. package/dist/toolsets/manager.js.map +1 -1
  176. package/dist/types/acp/index.d.ts +18 -0
  177. package/dist/types/acp/index.d.ts.map +1 -1
  178. package/dist/types/connector/mcp.d.ts +18 -0
  179. package/dist/types/connector/mcp.d.ts.map +1 -1
  180. package/dist/types/goal/index.d.ts +12 -0
  181. package/dist/types/goal/index.d.ts.map +1 -1
  182. package/dist/types/session/config.d.ts +8 -0
  183. package/dist/types/session/config.d.ts.map +1 -1
  184. package/dist/types/session/hand-back.d.ts +29 -0
  185. package/dist/types/session/hand-back.d.ts.map +1 -0
  186. package/dist/types/session/hand-back.js +11 -0
  187. package/dist/types/session/hand-back.js.map +1 -0
  188. package/dist/types/session/outcome.d.ts +37 -0
  189. package/dist/types/session/outcome.d.ts.map +1 -1
  190. package/dist/types/session/outcome.js.map +1 -1
  191. package/dist/types/session/turn.d.ts +19 -1
  192. package/dist/types/session/turn.d.ts.map +1 -1
  193. package/dist/types/session/turn.js.map +1 -1
  194. package/package.json +1 -1
  195. package/src/agents/file-definitions.ts +21 -0
  196. package/src/agents/index.ts +1 -0
  197. package/src/bridge/acp/server.ts +20 -0
  198. package/src/connector/index.ts +1 -0
  199. package/src/connector/mcp/client.ts +531 -14
  200. package/src/connector/mcp/index.ts +1 -0
  201. package/src/connector/mcp/tasks.ts +81 -0
  202. package/src/constants/data-epoch.ts +17 -0
  203. package/src/constants/index.ts +1 -0
  204. package/src/constants/mcp/index.ts +6 -1
  205. package/src/constants/telemetry/index.ts +13 -0
  206. package/src/eval/bootstrap.ts +314 -0
  207. package/src/eval/compare.ts +473 -0
  208. package/src/eval/disclosure.ts +136 -0
  209. package/src/eval/experiment.ts +127 -15
  210. package/src/eval/from-turn.ts +5 -0
  211. package/src/eval/index.ts +51 -0
  212. package/src/eval/report-stats.ts +257 -0
  213. package/src/eval/run-options.ts +49 -0
  214. package/src/eval/run-summary.ts +104 -0
  215. package/src/eval/scorers.ts +83 -0
  216. package/src/eval/types.ts +76 -0
  217. package/src/manager/session/turn-recorder.ts +23 -2
  218. package/src/plugin/index.ts +7 -1
  219. package/src/plugin/loader.ts +16 -2
  220. package/src/public-runtime.ts +70 -0
  221. package/src/public-types.ts +9 -0
  222. package/src/registry/command/kernel-commands.ts +94 -6
  223. package/src/runtime/query/cancelled-before-start.ts +2 -0
  224. package/src/runtime/query/index.ts +20 -0
  225. package/src/runtime/query/iteration/index.ts +85 -1
  226. package/src/runtime/query/iteration/phases/tool-review.ts +24 -1
  227. package/src/runtime/query/iteration/stream-turn.ts +10 -0
  228. package/src/runtime/query/result.ts +2 -0
  229. package/src/runtime/query/turn/answer-claims.ts +413 -0
  230. package/src/runtime/query/turn/outcome-ledger.ts +12 -0
  231. package/src/skills/index.ts +1 -0
  232. package/src/skills/loader.ts +21 -3
  233. package/src/store/goal/index.ts +11 -0
  234. package/src/telemetry/genai.ts +202 -0
  235. package/src/telemetry/runtime-accessors.ts +40 -1
  236. package/src/tools/goal/confirmation.ts +375 -0
  237. package/src/tools/goal/index.ts +37 -1
  238. package/src/tools/hand-back/index.ts +150 -0
  239. package/src/tools/schedules/schedule-tool.ts +19 -1
  240. package/src/tools/schedules/types.ts +10 -0
  241. package/src/toolsets/manager.ts +22 -0
  242. package/src/types/acp/index.ts +20 -0
  243. package/src/types/connector/mcp.ts +18 -0
  244. package/src/types/goal/index.ts +13 -0
  245. package/src/types/session/config.ts +8 -0
  246. package/src/types/session/hand-back.ts +31 -0
  247. package/src/types/session/outcome.ts +42 -0
  248. package/src/types/session/turn.ts +19 -1
@@ -1,4 +1,9 @@
1
+ import { EvalUsageError, validateSeed } from './bootstrap.js';
2
+ import { buildDisclosure, describeDisclosure } from './disclosure.js';
3
+ import { describeIntervals, describeSplits, duplicateMessage, duplicateName, groupRuns, statsOver, } from './report-stats.js';
4
+ import { currentEvalRunOptions } from './run-options.js';
1
5
  import { describeUncertainty, uncertaintyOf } from './uncertainty.js';
6
+ const MAX_REPEATS = 100;
2
7
  const MAX_CASE_TIMEOUT_MS = 2_147_483_647;
3
8
  /**
4
9
  * Run a dataset and score it.
@@ -17,25 +22,42 @@ export async function runExperiment(config) {
17
22
  const threshold = config.passThreshold ?? 1;
18
23
  const concurrency = Math.max(1, config.concurrency ?? 1);
19
24
  const timeoutMs = validateCaseTimeoutMs(config.timeoutMs);
20
- const results = new Array(config.cases.length);
25
+ const options = currentEvalRunOptions();
26
+ const repeats = validateRepeats(options.repeats ?? config.repeats);
27
+ const seed = validateSeed(options.seed ?? config.seed);
28
+ for (const evalCase of config.cases)
29
+ validateSplit(evalCase);
30
+ const alike = duplicateName(config.cases.map((c) => c.name));
31
+ if (alike !== undefined)
32
+ throw new Error(duplicateMessage(config.name, alike));
33
+ // Case-major, so a case's repeats sit together in the report.
34
+ const total = config.cases.length * repeats;
35
+ const results = new Array(total);
21
36
  let cursor = 0;
22
37
  const worker = async () => {
23
38
  for (;;) {
24
39
  const index = cursor++;
25
- const evalCase = config.cases[index];
26
- if (!evalCase)
40
+ if (index >= total)
27
41
  return;
42
+ const evalCase = config.cases[Math.floor(index / repeats)];
28
43
  const result = await evaluateCase(config, evalCase, threshold, timeoutMs);
44
+ result.repeat = (index % repeats) + 1;
45
+ if (evalCase.split)
46
+ result.split = evalCase.split;
29
47
  results[index] = result;
30
48
  config.onCaseFinish?.(result);
31
49
  }
32
50
  };
33
- await Promise.all(Array.from({ length: Math.min(concurrency, config.cases.length) }, worker));
51
+ await Promise.all(Array.from({ length: Math.min(concurrency, total) }, worker));
34
52
  const settled = results.filter(Boolean);
35
53
  // An inconclusive case has no mean to contribute; averaging its zero in
36
54
  // would report a suite-wide regression caused by the harness.
37
55
  const scored = settled.filter((r) => r.status !== 'inconclusive');
38
56
  const mean = scored.length === 0 ? 0 : scored.reduce((sum, r) => sum + r.mean, 0) / scored.length;
57
+ const stats = statsOver(groupRuns(settled), seed);
58
+ // One number per CASE, so a case's repeats are not counted as independent
59
+ // evidence about the suite.
60
+ const caseMeans = stats.caseSummaries.flatMap((c) => (c.mean === null ? [] : [c.mean]));
39
61
  return {
40
62
  name: config.name,
41
63
  cases: settled,
@@ -43,14 +65,35 @@ export async function runExperiment(config) {
43
65
  // Over the same cases the mean is over. Computing spread across a
44
66
  // different denominator than the average it qualifies would produce
45
67
  // an interval that does not belong to the number beside it.
46
- uncertainty: uncertaintyOf(scored.map((r) => r.mean)),
68
+ uncertainty: uncertaintyOf(caseMeans),
47
69
  passed: settled.filter((r) => r.status === 'passed').length,
48
70
  failed: settled.filter((r) => r.status === 'failed').length,
49
71
  inconclusive: settled.filter((r) => r.status === 'inconclusive').length,
50
72
  byScorer: meanByScorer(settled),
51
73
  durationMs: Date.now() - startedAt,
74
+ repeats,
75
+ seed,
76
+ ...(stats.passRate === null ? {} : { passRate: stats.passRate }),
77
+ intervals: stats.intervals,
78
+ ...(repeats > 1 ? { caseSummaries: stats.caseSummaries, reliability: stats.reliability } : {}),
79
+ splits: stats.splits,
80
+ disclosure: buildDisclosure(config.disclosure, { cases: config.cases.length, repeats, seed }, options, new Date().toISOString()),
52
81
  };
53
82
  }
83
+ function validateRepeats(value) {
84
+ if (value === undefined)
85
+ return 1;
86
+ if (!Number.isInteger(value) || value < 1 || value > MAX_REPEATS) {
87
+ throw new EvalUsageError(`repeats must be an integer from 1 to ${MAX_REPEATS}, or omitted; received ${String(value)}`);
88
+ }
89
+ return value;
90
+ }
91
+ function validateSplit(evalCase) {
92
+ const split = evalCase.split;
93
+ if (split !== undefined && split !== 'development' && split !== 'held-out') {
94
+ throw new RangeError(`case "${evalCase.name}" declares split ${JSON.stringify(split)}; a split is "development" or "held-out", or left out`);
95
+ }
96
+ }
54
97
  function validateCaseTimeoutMs(value) {
55
98
  if (value === undefined)
56
99
  return undefined;
@@ -263,17 +306,46 @@ function derivedUncertainty(report) {
263
306
  return uncertaintyOf(report.cases.filter((c) => c.status !== 'inconclusive').map((c) => c.mean));
264
307
  }
265
308
  export function formatReport(report) {
309
+ const repeats = report.repeats ?? 1;
310
+ const caseCount = repeats > 1 ? report.cases.length / repeats : report.cases.length;
266
311
  const lines = [
267
- `${report.name}: ${report.passed}/${report.cases.length} passed (mean ${report.mean.toFixed(2)}) in ${report.durationMs}ms`,
268
- // On its own line and always printed, including when the interval is
269
- // undefined. A mean printed alone is the thing that has been
270
- // over-read: two experiments three points apart look like a difference, and
271
- // at the n a hand-built suite has they are usually the same experiment
272
- // twice. Computing the interval and not showing it would leave the
273
- // reader exactly where they started.
274
- ` ${describeUncertainty(report.mean, report.uncertainty ?? derivedUncertainty(report))}`,
275
- '',
312
+ repeats > 1
313
+ ? `${report.name}: ${report.passed}/${report.cases.length} runs passed (${caseCount} cases x ${repeats} repeats, mean ${report.mean.toFixed(2)}) in ${report.durationMs}ms`
314
+ : `${report.name}: ${report.passed}/${report.cases.length} passed (mean ${report.mean.toFixed(2)}) in ${report.durationMs}ms`,
276
315
  ];
316
+ if (report.intervals) {
317
+ // The interval sits next to the number it belongs to. A mean printed
318
+ // alone is the thing that has been over-read: two experiments three
319
+ // points apart look like a difference, and at the n a hand-built
320
+ // suite has they are usually the same experiment twice.
321
+ for (const line of describeIntervals({
322
+ mean: report.cases.some((c) => c.status !== 'inconclusive') ? report.mean : null,
323
+ passRate: report.passRate ?? null,
324
+ intervals: report.intervals,
325
+ })) {
326
+ lines.push(` ${line}`);
327
+ }
328
+ }
329
+ else {
330
+ // A report from before bootstrap intervals, or one built by hand: the
331
+ // older interval, still never omitted.
332
+ lines.push(` ${describeUncertainty(report.mean, report.uncertainty ?? derivedUncertainty(report))}`);
333
+ }
334
+ if (report.reliability && repeats > 1) {
335
+ const r = report.reliability;
336
+ lines.push(` reliability across repeats: ${r.passedEveryRepeat} of ${r.cases} cases passed in every one of ${repeats} repeats${r.fraction === null ? '' : ` (${(r.fraction * 100).toFixed(1)}%)`}`);
337
+ for (const c of report.caseSummaries ?? []) {
338
+ if (!c.passedEveryRepeat)
339
+ lines.push(` ${c.case}: ${c.passed} of ${c.repeats} passed`);
340
+ }
341
+ }
342
+ if (report.splits) {
343
+ for (const line of describeSplits(report.splits))
344
+ lines.push(` ${line}`);
345
+ }
346
+ if (report.disclosure)
347
+ lines.push(` ${describeDisclosure(report.disclosure)}`);
348
+ lines.push('');
277
349
  // Surfaced above the failures, because an inconclusive case means the
278
350
  // harness is broken and every number below it is measured over less
279
351
  // evidence than it looks like.
@@ -287,7 +359,7 @@ export function formatReport(report) {
287
359
  if (failures.length > 0) {
288
360
  lines.push('', 'Failures:');
289
361
  for (const failure of failures) {
290
- lines.push(` ✗ ${failure.case} (${failure.mean.toFixed(2)})`);
362
+ lines.push(` ✗ ${failure.case}${repeats > 1 && failure.repeat ? ` [repeat ${failure.repeat} of ${repeats}]` : ''} (${failure.mean.toFixed(2)})`);
291
363
  if (failure.failedGates && failure.failedGates.length > 0) {
292
364
  // First line under the case, because it says WHY this failed.
293
365
  // A mean of 0.75 next to four scores leaves the reader
@@ -1 +1 @@
1
- {"version":3,"file":"experiment.js","sourceRoot":"","sources":["../../src/eval/experiment.ts"],"names":[],"mappings":"AAUA,OAAO,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAErE,MAAM,mBAAmB,GAAG,aAAa,CAAA;AAmCzC;;;;;;;;;;;GAWG;AACH,MAAM,CAAC,KAAK,UAAU,aAAa,CAClC,MAAgC;IAEhC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,MAAM,SAAS,GAAG,MAAM,CAAC,aAAa,IAAI,CAAC,CAAA;IAC3C,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,WAAW,IAAI,CAAC,CAAC,CAAA;IACxD,MAAM,SAAS,GAAG,qBAAqB,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IAEzD,MAAM,OAAO,GAAiB,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAA;IAC5D,IAAI,MAAM,GAAG,CAAC,CAAA;IAEd,MAAM,MAAM,GAAG,KAAK,IAAmB,EAAE;QACxC,SAAS,CAAC;YACT,MAAM,KAAK,GAAG,MAAM,EAAE,CAAA;YACtB,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;YACpC,IAAI,CAAC,QAAQ;gBAAE,OAAM;YAErB,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,SAAS,CAAC,CAAA;YACzE,OAAO,CAAC,KAAK,CAAC,GAAG,MAAM,CAAA;YACvB,MAAM,CAAC,YAAY,EAAE,CAAC,MAAM,CAAC,CAAA;QAC9B,CAAC;IACF,CAAC,CAAA;IAED,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC,CAAA;IAE7F,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;IACvC,wEAAwE;IACxE,8DAA8D;IAC9D,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAA;IACjE,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IAEjG,OAAO;QACN,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,KAAK,EAAE,OAAO;QACd,IAAI;QACJ,kEAAkE;QAClE,oEAAoE;QACpE,4DAA4D;QAC5D,WAAW,EAAE,aAAa,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QACrD,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM;QAC3D,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM;QAC3D,YAAY,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAC,MAAM;QACvE,QAAQ,EAAE,YAAY,CAAC,OAAO,CAAC;QAC/B,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;KAClC,CAAA;AACF,CAAC;AAED,SAAS,qBAAqB,CAAC,KAAyB;IACvD,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAA;IACzC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,GAAG,mBAAmB,EAAE,CAAC;QAC1E,MAAM,IAAI,UAAU,CACnB,0CAA0C,mBAAmB,0BAA0B,MAAM,CAAC,KAAK,CAAC,EAAE,CACtG,CAAA;IACF,CAAC;IACD,OAAO,KAAK,CAAA;AACb,CAAC;AAQD,6EAA6E;AAC7E,SAAS,gBAAgB,CAAC,SAA6B;IACtD,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAA;IACxC,IAAI,KAAgD,CAAA;IACpD,IAAI,MAAkC,CAAA;IACtC,IAAI,OAAiC,CAAA;IAErC,IAAI,SAAS,KAAK,SAAS,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;QAC9C,MAAM,GAAG,IAAI,OAAO,CAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE;YAChD,OAAO,GAAG,GAAG,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,MAAM,CAAC,CAAA;YAChD,UAAU,CAAC,MAAM,CAAC,gBAAgB,CAAC,OAAO,EAAE,OAAO,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAA;QACrE,CAAC,CAAC,CAAA;QACF,qEAAqE;QACrE,gEAAgE;QAChE,yDAAyD;QACzD,KAAK,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;QAC3B,KAAK,GAAG,UAAU,CACjB,GAAG,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,wBAAwB,SAAS,IAAI,CAAC,CAAC,EACxE,SAAS,CACT,CAAA;IACF,CAAC;IAED,OAAO;QACN,MAAM,EAAE,UAAU,CAAC,MAAM;QACzB,KAAK,CAAC,IAAI,CAAI,IAAgB;YAC7B,IAAI,CAAC,MAAM;gBAAE,OAAO,MAAM,IAAI,CAAA;YAC9B,qEAAqE;YACrE,mEAAmE;YACnE,KAAK,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;YACzB,OAAO,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAA;QAC1C,CAAC;QACD,OAAO;YACN,YAAY,CAAC,KAAK,CAAC,CAAA;YACnB,IAAI,OAAO;gBAAE,UAAU,CAAC,MAAM,CAAC,mBAAmB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAA;QACrE,CAAC;KACD,CAAA;AACF,CAAC;AAED,KAAK,UAAU,YAAY,CAC1B,MAAgC,EAChC,QAA0B,EAC1B,SAAiB,EACjB,SAA6B;IAE7B,MAAM,QAAQ,GAAG,gBAAgB,CAAC,SAAS,CAAC,CAAA;IAC5C,IAAI,CAAC;QACJ,MAAM,IAAI,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAA;QAC1D,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,IAAI,MAAM,CAAC,OAAO,CAAA;QAClD,MAAM,MAAM,GAA0B,EAAE,CAAA;QAExC,iEAAiE;QACjE,iEAAiE;QACjE,iDAAiD;QACjD,iEAAiE;QACjE,0DAA0D;QAC1D,6DAA6D;QAC7D,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;QAC9B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,KAAK,CACd,0BAA0B,MAAM,CAAC,IAAI,eAAe,QAAQ,CAAC,IAAI,wKAAwK,CACzO,CAAA;YACF,CAAC;YACD,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACrB,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAA;QACxE,CAAC;QAED,iEAAiE;QACjE,+DAA+D;QAC/D,2DAA2D;QAC3D,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC;aAClC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,IAAI,CAAC;aACrC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAA;QACrB,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;QAExF,iEAAiE;QACjE,8DAA8D;QAC9D,2DAA2D;QAC3D,gEAAgE;QAChE,2DAA2D;QAC3D,8DAA8D;QAC9D,8DAA8D;QAC9D,uBAAuB;QACvB,MAAM,WAAW,GAAG,OAAO;aACzB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC;aACpC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;YACb,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;YAC5B,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI;gBAAE,OAAO,KAAK,CAAA;YACtD,OAAO,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,SAAS,CAAC,CAAA;QAChD,CAAC,CAAC;aACD,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;QAEpB,MAAM,MAAM,GACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YACpD,CAAC,CAAC,cAAc;YAChB,CAAC,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC;gBACvB,CAAC,CAAC,QAAQ;gBACV,CAAC,CAAC,IAAI,IAAI,SAAS;oBAClB,CAAC,CAAC,QAAQ;oBACV,CAAC,CAAC,QAAQ,CAAA;QACd,OAAO;YACN,IAAI,EAAE,QAAQ,CAAC,IAAI;YACnB,IAAI;YACJ,MAAM;YACN,IAAI;YACJ,MAAM;YACN,MAAM,EAAE,MAAM,KAAK,QAAQ;YAC3B,8DAA8D;YAC9D,6DAA6D;YAC7D,GAAG,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SAClD,CAAA;IACF,CAAC;YAAS,CAAC;QACV,QAAQ,CAAC,OAAO,EAAE,CAAA;IACnB,CAAC;AACF,CAAC;AAED;;;GAGG;AACH,KAAK,UAAU,WAAW,CACzB,MAAgC,EAChC,QAA0B,EAC1B,QAAsB;IAEtB,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,IAAI,CAAC;QACJ,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAA;QAClE,OAAO,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACjC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACd,OAAO;YACN,MAAM,EAAE,IAAI;YACZ,KAAK,EAAE,EAAE;YACT,SAAS,EAAE,EAAE;YACb,WAAW,EAAE,CAAC;YACd,YAAY,EAAE,CAAC;YACf,4DAA4D;YAC5D,6DAA6D;YAC7D,sBAAsB;YACtB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;YAClC,KAAK,EAAE,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC;SACvD,CAAA;IACF,CAAC;AACF,CAAC;AAED,kEAAkE;AAClE,KAAK,UAAU,SAAS,CACvB,MAAc,EACd,IAAc,EACd,QAAkB,EAClB,QAAsB;IAEtB,sEAAsE;IACtE,qEAAqE;IACrE,2DAA2D;IAC3D,uEAAuE;IACvE,wEAAwE;IACxE,0EAA0E;IAC1E,qDAAqD;IACrD,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;QAC9B,OAAO,EAAE,KAAK,EAAE,CAAC,EAAE,MAAM,EAAE,gBAAgB,IAAI,CAAC,KAAK,EAAE,EAAE,OAAO,EAAE,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,EAAE,CAAA;IAC1F,CAAC;IAED,IAAI,CAAC;QACJ,QAAQ,CAAC,MAAM,CAAC,cAAc,EAAE,CAAA;QAChC,qEAAqE;QACrE,oEAAoE;QACpE,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,EAAE,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAA;QACxF,OAAO,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACjC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACd,oEAAoE;QACpE,6DAA6D;QAC7D,kEAAkE;QAClE,qEAAqE;QACrE,kEAAkE;QAClE,yCAAyC;QACzC,OAAO;YACN,KAAK,EAAE,CAAC;YACR,WAAW,EAAE,IAAI;YACjB,MAAM,EAAE,WAAW,MAAM,CAAC,IAAI,gCAC7B,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAChD,EAAE;SACF,CAAA;IACF,CAAC;AACF,CAAC;AAED,SAAS,YAAY,CAAC,OAA8B;IACnD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA4C,CAAA;IAChE,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;YAC3D,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI;gBAAE,SAAQ;YACxC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAA;YACtD,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAA;YAC1B,KAAK,CAAC,KAAK,EAAE,CAAA;YACb,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;QACtB,CAAC;IACF,CAAC;IACD,MAAM,GAAG,GAA2B,EAAE,CAAA;IACtC,oEAAoE;IACpE,sEAAsE;IACtE,qBAAqB;IACrB,KAAK,MAAM,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC;QAC7C,IAAI,KAAK,GAAG,CAAC;YAAE,GAAG,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,KAAK,CAAA;IACzC,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;GAKG;AACH;;;;;;;;;;;;GAYG;AACH,SAAS,kBAAkB,CAAC,MAAwB;IACnD,OAAO,aAAa,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;AACjG,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,MAAwB;IACpD,MAAM,KAAK,GAAa;QACvB,GAAG,MAAM,CAAC,IAAI,KAAK,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,iBAAiB,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,UAAU,IAAI;QAC3H,qEAAqE;QACrE,6DAA6D;QAC7D,4EAA4E;QAC5E,uEAAuE;QACvE,mEAAmE;QACnE,qCAAqC;QACrC,KAAK,mBAAmB,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,WAAW,IAAI,kBAAkB,CAAC,MAAM,CAAC,CAAC,EAAE;QACzF,EAAE;KACF,CAAA;IAED,sEAAsE;IACtE,oEAAoE;IACpE,+BAA+B;IAC/B,IAAI,MAAM,CAAC,YAAY,GAAG,CAAC,EAAE,CAAC;QAC7B,KAAK,CAAC,IAAI,CACT,KAAK,MAAM,CAAC,YAAY,QAAQ,MAAM,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,yDAAyD,EAC7H,EAAE,CACF,CAAA;IACF,CAAC;IAED,KAAK,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,KAAK,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;IAC5C,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAA;IAClE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,WAAW,CAAC,CAAA;QAC3B,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAChC,KAAK,CAAC,IAAI,CAAC,OAAO,OAAO,CAAC,IAAI,KAAK,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;YAC9D,IAAI,OAAO,CAAC,WAAW,IAAI,OAAO,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC3D,8DAA8D;gBAC9D,uDAAuD;gBACvD,+BAA+B;gBAC/B,KAAK,CAAC,IAAI,CAAC,sBAAsB,OAAO,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YACnE,CAAC;YACD,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC5D,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI,EAAE,CAAC;oBAChC,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,kBAAkB,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC1D,CAAC;qBAAM,IAAI,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC;oBAC5B,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC7C,CAAC;YACF,CAAC;YACD,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK;gBAAE,KAAK,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAA;QACzE,CAAC;IACF,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAA;IACxE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,aAAa,CAAC,CAAA;QAC7B,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,EAAE,CAAC,CAAA;YAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;gBACzD,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;YAC7C,CAAC;QACF,CAAC;IACF,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACxB,CAAC"}
1
+ {"version":3,"file":"experiment.js","sourceRoot":"","sources":["../../src/eval/experiment.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAA;AAC7D,OAAO,EAAmB,eAAe,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAA;AACtF,OAAO,EACN,iBAAiB,EACjB,cAAc,EACd,gBAAgB,EAChB,aAAa,EACb,SAAS,EACT,SAAS,GACT,MAAM,mBAAmB,CAAA;AAC1B,OAAO,EAAE,qBAAqB,EAAE,MAAM,kBAAkB,CAAA;AAWxD,OAAO,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAErE,MAAM,WAAW,GAAG,GAAG,CAAA;AAEvB,MAAM,mBAAmB,GAAG,aAAa,CAAA;AAyDzC;;;;;;;;;;;GAWG;AACH,MAAM,CAAC,KAAK,UAAU,aAAa,CAClC,MAAgC;IAEhC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,MAAM,SAAS,GAAG,MAAM,CAAC,aAAa,IAAI,CAAC,CAAA;IAC3C,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,WAAW,IAAI,CAAC,CAAC,CAAA;IACxD,MAAM,SAAS,GAAG,qBAAqB,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IACzD,MAAM,OAAO,GAAG,qBAAqB,EAAE,CAAA;IACvC,MAAM,OAAO,GAAG,eAAe,CAAC,OAAO,CAAC,OAAO,IAAI,MAAM,CAAC,OAAO,CAAC,CAAA;IAClE,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,CAAC,IAAI,IAAI,MAAM,CAAC,IAAI,CAAC,CAAA;IACtD,KAAK,MAAM,QAAQ,IAAI,MAAM,CAAC,KAAK;QAAE,aAAa,CAAC,QAAQ,CAAC,CAAA;IAC5D,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;IAC5D,IAAI,KAAK,KAAK,SAAS;QAAE,MAAM,IAAI,KAAK,CAAC,gBAAgB,CAAC,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAA;IAE9E,8DAA8D;IAC9D,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,GAAG,OAAO,CAAA;IAC3C,MAAM,OAAO,GAAiB,IAAI,KAAK,CAAC,KAAK,CAAC,CAAA;IAC9C,IAAI,MAAM,GAAG,CAAC,CAAA;IAEd,MAAM,MAAM,GAAG,KAAK,IAAmB,EAAE;QACxC,SAAS,CAAC;YACT,MAAM,KAAK,GAAG,MAAM,EAAE,CAAA;YACtB,IAAI,KAAK,IAAI,KAAK;gBAAE,OAAM;YAC1B,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,OAAO,CAAC,CAAqB,CAAA;YAE9E,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,SAAS,CAAC,CAAA;YACzE,MAAM,CAAC,MAAM,GAAG,CAAC,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,CAAA;YACrC,IAAI,QAAQ,CAAC,KAAK;gBAAE,MAAM,CAAC,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAA;YACjD,OAAO,CAAC,KAAK,CAAC,GAAG,MAAM,CAAA;YACvB,MAAM,CAAC,YAAY,EAAE,CAAC,MAAM,CAAC,CAAA;QAC9B,CAAC;IACF,CAAC,CAAA;IAED,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,KAAK,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC,CAAA;IAE/E,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;IACvC,wEAAwE;IACxE,8DAA8D;IAC9D,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAA;IACjE,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IAEjG,MAAM,KAAK,GAAG,SAAS,CAAC,SAAS,CAAC,OAAO,CAAC,EAAE,IAAI,CAAC,CAAA;IACjD,0EAA0E;IAC1E,4BAA4B;IAC5B,MAAM,SAAS,GAAG,KAAK,CAAC,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IAEvF,OAAO;QACN,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,KAAK,EAAE,OAAO;QACd,IAAI;QACJ,kEAAkE;QAClE,oEAAoE;QACpE,4DAA4D;QAC5D,WAAW,EAAE,aAAa,CAAC,SAAS,CAAC;QACrC,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM;QAC3D,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM;QAC3D,YAAY,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAC,MAAM;QACvE,QAAQ,EAAE,YAAY,CAAC,OAAO,CAAC;QAC/B,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;QAClC,OAAO;QACP,IAAI;QACJ,GAAG,CAAC,KAAK,CAAC,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,CAAC,QAAQ,EAAE,CAAC;QAChE,SAAS,EAAE,KAAK,CAAC,SAAS;QAC1B,GAAG,CAAC,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,KAAK,CAAC,aAAa,EAAE,WAAW,EAAE,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC9F,MAAM,EAAE,KAAK,CAAC,MAAM;QACpB,UAAU,EAAE,eAAe,CAC1B,MAAM,CAAC,UAAU,EACjB,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,EAC7C,OAAO,EACP,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CACxB;KACD,CAAA;AACF,CAAC;AAED,SAAS,eAAe,CAAC,KAAyB;IACjD,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,CAAC,CAAA;IACjC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,GAAG,WAAW,EAAE,CAAC;QAClE,MAAM,IAAI,cAAc,CACvB,wCAAwC,WAAW,0BAA0B,MAAM,CAAC,KAAK,CAAC,EAAE,CAC5F,CAAA;IACF,CAAC;IACD,OAAO,KAAK,CAAA;AACb,CAAC;AAED,SAAS,aAAa,CAAC,QAA2B;IACjD,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAgB,CAAA;IACvC,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,aAAa,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;QAC5E,MAAM,IAAI,UAAU,CACnB,SAAS,QAAQ,CAAC,IAAI,oBAAoB,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,uDAAuD,CACtH,CAAA;IACF,CAAC;AACF,CAAC;AAED,SAAS,qBAAqB,CAAC,KAAyB;IACvD,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAA;IACzC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,GAAG,mBAAmB,EAAE,CAAC;QAC1E,MAAM,IAAI,UAAU,CACnB,0CAA0C,mBAAmB,0BAA0B,MAAM,CAAC,KAAK,CAAC,EAAE,CACtG,CAAA;IACF,CAAC;IACD,OAAO,KAAK,CAAA;AACb,CAAC;AAQD,6EAA6E;AAC7E,SAAS,gBAAgB,CAAC,SAA6B;IACtD,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAA;IACxC,IAAI,KAAgD,CAAA;IACpD,IAAI,MAAkC,CAAA;IACtC,IAAI,OAAiC,CAAA;IAErC,IAAI,SAAS,KAAK,SAAS,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;QAC9C,MAAM,GAAG,IAAI,OAAO,CAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE;YAChD,OAAO,GAAG,GAAG,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,MAAM,CAAC,CAAA;YAChD,UAAU,CAAC,MAAM,CAAC,gBAAgB,CAAC,OAAO,EAAE,OAAO,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAA;QACrE,CAAC,CAAC,CAAA;QACF,qEAAqE;QACrE,gEAAgE;QAChE,yDAAyD;QACzD,KAAK,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;QAC3B,KAAK,GAAG,UAAU,CACjB,GAAG,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,wBAAwB,SAAS,IAAI,CAAC,CAAC,EACxE,SAAS,CACT,CAAA;IACF,CAAC;IAED,OAAO;QACN,MAAM,EAAE,UAAU,CAAC,MAAM;QACzB,KAAK,CAAC,IAAI,CAAI,IAAgB;YAC7B,IAAI,CAAC,MAAM;gBAAE,OAAO,MAAM,IAAI,CAAA;YAC9B,qEAAqE;YACrE,mEAAmE;YACnE,KAAK,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;YACzB,OAAO,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAA;QAC1C,CAAC;QACD,OAAO;YACN,YAAY,CAAC,KAAK,CAAC,CAAA;YACnB,IAAI,OAAO;gBAAE,UAAU,CAAC,MAAM,CAAC,mBAAmB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAA;QACrE,CAAC;KACD,CAAA;AACF,CAAC;AAED,KAAK,UAAU,YAAY,CAC1B,MAAgC,EAChC,QAA0B,EAC1B,SAAiB,EACjB,SAA6B;IAE7B,MAAM,QAAQ,GAAG,gBAAgB,CAAC,SAAS,CAAC,CAAA;IAC5C,IAAI,CAAC;QACJ,MAAM,IAAI,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAA;QAC1D,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,IAAI,MAAM,CAAC,OAAO,CAAA;QAClD,MAAM,MAAM,GAA0B,EAAE,CAAA;QAExC,iEAAiE;QACjE,iEAAiE;QACjE,iDAAiD;QACjD,iEAAiE;QACjE,0DAA0D;QAC1D,6DAA6D;QAC7D,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;QAC9B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,KAAK,CACd,0BAA0B,MAAM,CAAC,IAAI,eAAe,QAAQ,CAAC,IAAI,wKAAwK,CACzO,CAAA;YACF,CAAC;YACD,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACrB,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAA;QACxE,CAAC;QAED,iEAAiE;QACjE,+DAA+D;QAC/D,2DAA2D;QAC3D,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC;aAClC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,IAAI,CAAC;aACrC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAA;QACrB,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;QAExF,iEAAiE;QACjE,8DAA8D;QAC9D,2DAA2D;QAC3D,gEAAgE;QAChE,2DAA2D;QAC3D,8DAA8D;QAC9D,8DAA8D;QAC9D,uBAAuB;QACvB,MAAM,WAAW,GAAG,OAAO;aACzB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC;aACpC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;YACb,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;YAC5B,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI;gBAAE,OAAO,KAAK,CAAA;YACtD,OAAO,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,SAAS,CAAC,CAAA;QAChD,CAAC,CAAC;aACD,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;QAEpB,MAAM,MAAM,GACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YACpD,CAAC,CAAC,cAAc;YAChB,CAAC,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC;gBACvB,CAAC,CAAC,QAAQ;gBACV,CAAC,CAAC,IAAI,IAAI,SAAS;oBAClB,CAAC,CAAC,QAAQ;oBACV,CAAC,CAAC,QAAQ,CAAA;QACd,OAAO;YACN,IAAI,EAAE,QAAQ,CAAC,IAAI;YACnB,IAAI;YACJ,MAAM;YACN,IAAI;YACJ,MAAM;YACN,MAAM,EAAE,MAAM,KAAK,QAAQ;YAC3B,8DAA8D;YAC9D,6DAA6D;YAC7D,GAAG,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SAClD,CAAA;IACF,CAAC;YAAS,CAAC;QACV,QAAQ,CAAC,OAAO,EAAE,CAAA;IACnB,CAAC;AACF,CAAC;AAED;;;GAGG;AACH,KAAK,UAAU,WAAW,CACzB,MAAgC,EAChC,QAA0B,EAC1B,QAAsB;IAEtB,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,IAAI,CAAC;QACJ,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAA;QAClE,OAAO,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACjC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACd,OAAO;YACN,MAAM,EAAE,IAAI;YACZ,KAAK,EAAE,EAAE;YACT,SAAS,EAAE,EAAE;YACb,WAAW,EAAE,CAAC;YACd,YAAY,EAAE,CAAC;YACf,4DAA4D;YAC5D,6DAA6D;YAC7D,sBAAsB;YACtB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;YAClC,KAAK,EAAE,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC;SACvD,CAAA;IACF,CAAC;AACF,CAAC;AAED,kEAAkE;AAClE,KAAK,UAAU,SAAS,CACvB,MAAc,EACd,IAAc,EACd,QAAkB,EAClB,QAAsB;IAEtB,sEAAsE;IACtE,qEAAqE;IACrE,2DAA2D;IAC3D,uEAAuE;IACvE,wEAAwE;IACxE,0EAA0E;IAC1E,qDAAqD;IACrD,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;QAC9B,OAAO,EAAE,KAAK,EAAE,CAAC,EAAE,MAAM,EAAE,gBAAgB,IAAI,CAAC,KAAK,EAAE,EAAE,OAAO,EAAE,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,EAAE,CAAA;IAC1F,CAAC;IAED,IAAI,CAAC;QACJ,QAAQ,CAAC,MAAM,CAAC,cAAc,EAAE,CAAA;QAChC,qEAAqE;QACrE,oEAAoE;QACpE,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,EAAE,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAA;QACxF,OAAO,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACjC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACd,oEAAoE;QACpE,6DAA6D;QAC7D,kEAAkE;QAClE,qEAAqE;QACrE,kEAAkE;QAClE,yCAAyC;QACzC,OAAO;YACN,KAAK,EAAE,CAAC;YACR,WAAW,EAAE,IAAI;YACjB,MAAM,EAAE,WAAW,MAAM,CAAC,IAAI,gCAC7B,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAChD,EAAE;SACF,CAAA;IACF,CAAC;AACF,CAAC;AAED,SAAS,YAAY,CAAC,OAA8B;IACnD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA4C,CAAA;IAChE,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;YAC3D,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI;gBAAE,SAAQ;YACxC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAA;YACtD,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAA;YAC1B,KAAK,CAAC,KAAK,EAAE,CAAA;YACb,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;QACtB,CAAC;IACF,CAAC;IACD,MAAM,GAAG,GAA2B,EAAE,CAAA;IACtC,oEAAoE;IACpE,sEAAsE;IACtE,qBAAqB;IACrB,KAAK,MAAM,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC;QAC7C,IAAI,KAAK,GAAG,CAAC;YAAE,GAAG,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,KAAK,CAAA;IACzC,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;GAKG;AACH;;;;;;;;;;;;GAYG;AACH,SAAS,kBAAkB,CAAC,MAAwB;IACnD,OAAO,aAAa,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;AACjG,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,MAAwB;IACpD,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,IAAI,CAAC,CAAA;IACnC,MAAM,SAAS,GAAG,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,GAAG,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAA;IACnF,MAAM,KAAK,GAAa;QACvB,OAAO,GAAG,CAAC;YACV,CAAC,CAAC,GAAG,MAAM,CAAC,IAAI,KAAK,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,iBAAiB,SAAS,YAAY,OAAO,kBAAkB,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,UAAU,IAAI;YAC3K,CAAC,CAAC,GAAG,MAAM,CAAC,IAAI,KAAK,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,iBAAiB,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,UAAU,IAAI;KAC9H,CAAA;IACD,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;QACtB,qEAAqE;QACrE,oEAAoE;QACpE,iEAAiE;QACjE,wDAAwD;QACxD,KAAK,MAAM,IAAI,IAAI,iBAAiB,CAAC;YACpC,IAAI,EAAE,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI;YAChF,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,IAAI;YACjC,SAAS,EAAE,MAAM,CAAC,SAAS;SAC3B,CAAC,EAAE,CAAC;YACJ,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAA;QACxB,CAAC;IACF,CAAC;SAAM,CAAC;QACP,sEAAsE;QACtE,uCAAuC;QACvC,KAAK,CAAC,IAAI,CACT,KAAK,mBAAmB,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,WAAW,IAAI,kBAAkB,CAAC,MAAM,CAAC,CAAC,EAAE,CACzF,CAAA;IACF,CAAC;IACD,IAAI,MAAM,CAAC,WAAW,IAAI,OAAO,GAAG,CAAC,EAAE,CAAC;QACvC,MAAM,CAAC,GAAG,MAAM,CAAC,WAAW,CAAA;QAC5B,KAAK,CAAC,IAAI,CACT,iCAAiC,CAAC,CAAC,iBAAiB,OAAO,CAAC,CAAC,KAAK,iCAAiC,OAAO,WAAW,CAAC,CAAC,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,CACxL,CAAA;QACD,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,aAAa,IAAI,EAAE,EAAE,CAAC;YAC5C,IAAI,CAAC,CAAC,CAAC,iBAAiB;gBAAE,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,MAAM,OAAO,CAAC,CAAC,OAAO,SAAS,CAAC,CAAA;QAC1F,CAAC;IACF,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;QACnB,KAAK,MAAM,IAAI,IAAI,cAAc,CAAC,MAAM,CAAC,MAAM,CAAC;YAAE,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAA;IAC1E,CAAC;IACD,IAAI,MAAM,CAAC,UAAU;QAAE,KAAK,CAAC,IAAI,CAAC,KAAK,kBAAkB,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC,CAAA;IAC/E,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IAEd,sEAAsE;IACtE,oEAAoE;IACpE,+BAA+B;IAC/B,IAAI,MAAM,CAAC,YAAY,GAAG,CAAC,EAAE,CAAC;QAC7B,KAAK,CAAC,IAAI,CACT,KAAK,MAAM,CAAC,YAAY,QAAQ,MAAM,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,yDAAyD,EAC7H,EAAE,CACF,CAAA;IACF,CAAC;IAED,KAAK,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,KAAK,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;IAC5C,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAA;IAClE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,WAAW,CAAC,CAAA;QAC3B,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAChC,KAAK,CAAC,IAAI,CACT,OAAO,OAAO,CAAC,IAAI,GAAG,OAAO,GAAG,CAAC,IAAI,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,YAAY,OAAO,CAAC,MAAM,OAAO,OAAO,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CACrI,CAAA;YACD,IAAI,OAAO,CAAC,WAAW,IAAI,OAAO,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC3D,8DAA8D;gBAC9D,uDAAuD;gBACvD,+BAA+B;gBAC/B,KAAK,CAAC,IAAI,CAAC,sBAAsB,OAAO,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YACnE,CAAC;YACD,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC5D,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI,EAAE,CAAC;oBAChC,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,kBAAkB,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC1D,CAAC;qBAAM,IAAI,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC;oBAC5B,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC7C,CAAC;YACF,CAAC;YACD,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK;gBAAE,KAAK,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAA;QACzE,CAAC;IACF,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAA;IACxE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,aAAa,CAAC,CAAA;QAC7B,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,EAAE,CAAC,CAAA;YAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;gBACzD,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;YAC7C,CAAC;QACF,CAAC;IACF,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACxB,CAAC"}
@@ -1 +1 @@
1
- {"version":3,"file":"from-turn.d.ts","sourceRoot":"","sources":["../../src/eval/from-turn.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,0BAA0B,CAAA;AACpD,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,YAAY,CAAA;AAE1C;;;;;;;;GAQG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,GAAE;IAAE,UAAU,CAAC,EAAE,MAAM,CAAA;CAAO,GAAG,QAAQ,CAczF;AAED;;;;;;;;;;;;;GAaG;AACH,wBAAsB,iBAAiB,CACtC,SAAS,EAAE,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,GACtC,OAAO,CAAC,QAAQ,CAAC,CAKnB"}
1
+ {"version":3,"file":"from-turn.d.ts","sourceRoot":"","sources":["../../src/eval/from-turn.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,0BAA0B,CAAA;AACpD,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,YAAY,CAAA;AAE1C;;;;;;;;GAQG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,GAAE;IAAE,UAAU,CAAC,EAAE,MAAM,CAAA;CAAO,GAAG,QAAQ,CAkBzF;AAED;;;;;;;;;;;;;GAaG;AACH,wBAAsB,iBAAiB,CACtC,SAAS,EAAE,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,GACtC,OAAO,CAAC,QAAQ,CAAC,CAKnB"}
@@ -1,3 +1,4 @@
1
+ import { compareAnswerToOutcome } from '../runtime/query/turn/answer-claims.js';
1
2
  /**
2
3
  * Project a finished `Turn` into the shape scorers consume.
3
4
  *
@@ -9,6 +10,7 @@
9
10
  */
10
11
  export function evalTurnFromTurn(turn, opts = {}) {
11
12
  const steps = turn.steps ?? [];
13
+ const claimMismatch = compareAnswerToOutcome(turn.result, turn.outcome);
12
14
  return {
13
15
  output: turn.result ?? null,
14
16
  ...(turn.structuredOutput !== undefined ? { structuredOutput: turn.structuredOutput } : {}),
@@ -20,6 +22,9 @@ export function evalTurnFromTurn(turn, opts = {}) {
20
22
  totalCostUsd: turn.costInfo.totalCost,
21
23
  durationMs: opts.durationMs ?? (turn.endedAt ?? Date.now()) - turn.startedAt,
22
24
  ...(turn.lastError ? { error: turn.lastError } : {}),
25
+ ...(turn.outcome ? { outcome: turn.outcome } : {}),
26
+ ...(claimMismatch ? { claimMismatch } : {}),
27
+ ...(turn.handedBack ? { handedBack: turn.handedBack } : {}),
23
28
  };
24
29
  }
25
30
  /**
@@ -1 +1 @@
1
- {"version":3,"file":"from-turn.js","sourceRoot":"","sources":["../../src/eval/from-turn.ts"],"names":[],"mappings":"AAGA;;;;;;;;GAQG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAU,EAAE,OAAgC,EAAE;IAC9E,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,IAAI,EAAE,CAAA;IAC9B,OAAO;QACN,MAAM,EAAE,IAAI,CAAC,MAAM,IAAI,IAAI;QAC3B,GAAG,CAAC,IAAI,CAAC,gBAAgB,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,gBAAgB,EAAE,IAAI,CAAC,gBAAgB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3F,KAAK;QACL,yDAAyD;QACzD,SAAS,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QACxE,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3D,WAAW,EAAE,IAAI,CAAC,UAAU,CAAC,WAAW;QACxC,YAAY,EAAE,IAAI,CAAC,QAAQ,CAAC,SAAS;QACrC,UAAU,EAAE,IAAI,CAAC,UAAU,IAAI,CAAC,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS;QAC5E,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KACpD,CAAA;AACF,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACtC,SAAwC;IAExC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,IAAI,IAAI,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE,CAAA;IACjC,OAAO,CAAC,IAAI,CAAC,IAAI;QAAE,IAAI,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE,CAAA;IAChD,OAAO,gBAAgB,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,CAAC,CAAA;AAC5E,CAAC"}
1
+ {"version":3,"file":"from-turn.js","sourceRoot":"","sources":["../../src/eval/from-turn.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,sBAAsB,EAAE,MAAM,wCAAwC,CAAA;AAI/E;;;;;;;;GAQG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAU,EAAE,OAAgC,EAAE;IAC9E,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,IAAI,EAAE,CAAA;IAC9B,MAAM,aAAa,GAAG,sBAAsB,CAAC,IAAI,CAAC,MAAM,EAAE,IAAI,CAAC,OAAO,CAAC,CAAA;IACvE,OAAO;QACN,MAAM,EAAE,IAAI,CAAC,MAAM,IAAI,IAAI;QAC3B,GAAG,CAAC,IAAI,CAAC,gBAAgB,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,gBAAgB,EAAE,IAAI,CAAC,gBAAgB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3F,KAAK;QACL,yDAAyD;QACzD,SAAS,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QACxE,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3D,WAAW,EAAE,IAAI,CAAC,UAAU,CAAC,WAAW;QACxC,YAAY,EAAE,IAAI,CAAC,QAAQ,CAAC,SAAS;QACrC,UAAU,EAAE,IAAI,CAAC,UAAU,IAAI,CAAC,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS;QAC5E,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACpD,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClD,GAAG,CAAC,aAAa,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3C,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC3D,CAAA;AACF,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACtC,SAAwC;IAExC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,IAAI,IAAI,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE,CAAA;IACjC,OAAO,CAAC,IAAI,CAAC,IAAI;QAAE,IAAI,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE,CAAA;IAChD,OAAO,gBAAgB,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,CAAC,CAAA;AAC5E,CAAC"}
@@ -6,8 +6,19 @@ export { formatReport, runExperiment } from './experiment.js';
6
6
  export type { ExperimentConfig } from './experiment.js';
7
7
  export { judgeScorer } from './judge.js';
8
8
  export type { JudgeScorerConfig } from './judge.js';
9
- export { completionScorer, containsScorer, customScorer, stepBudgetScorer, trajectoryScorer, } from './scorers.js';
9
+ export { claimMismatchScorer, handedBackScorer, completionScorer, containsScorer, customScorer, outcomeScorer, stepBudgetScorer, trajectoryScorer, } from './scorers.js';
10
+ export { BOOTSTRAP_RESAMPLES, DEFAULT_EVAL_SEED, EvalUsageError, INTERVAL_LEVEL, MIN_CASES_FOR_INTERVAL, bootstrapCases, bootstrapPairedDifference, createSeededRandom, exactSignTestP, isEvalUsageError, } from './bootstrap.js';
11
+ export type { BootstrapMethod, CaseIntervals, CaseTally, DifferenceIntervals, Interval, PairedTally, } from './bootstrap.js';
12
+ export type { ReportDisclosure, RoleDisclosure, RoutingDisclosure, RunMode, SuiteFacts, } from './disclosure.js';
13
+ export { sdkVersion } from './disclosure.js';
14
+ export { currentEvalRunOptions, withEvalRunOptions } from './run-options.js';
15
+ export type { EvalRunOptions } from './run-options.js';
16
+ export { formatRunSummary, summarizeRun } from './run-summary.js';
17
+ export type { RunDisclosure, RunSummary } from './run-summary.js';
18
+ export type { RepeatReliability, SplitSummary } from './report-stats.js';
19
+ export { compareEvalReports, formatEvalComparison, loadEvalReportFile } from './compare.js';
20
+ export type { CaseRef, DisclosureComparison, DisclosureDifference, DisclosureDimension, DisclosureUnknown, EvalComparison, Flips, LoadedReportFile, LoadedSuite, SideNumbers, } from './compare.js';
10
21
  export { describeUncertainty, uncertaintyOf } from './uncertainty.js';
11
22
  export type { ScoreUncertainty } from './uncertainty.js';
12
- export type { CaseResult, CaseStatus, EvalCase, EvalTurn, ExperimentReport, Score, Scorer, } from './types.js';
23
+ export type { CaseRepeatSummary, CaseResult, EvalSplit, CaseStatus, EvalCase, EvalTurn, ExperimentReport, Score, Scorer, } from './types.js';
13
24
  //# sourceMappingURL=index.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/eval/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,gBAAgB,CAAA;AACpE,OAAO,EAAE,oBAAoB,EAAE,sBAAsB,EAAE,MAAM,2BAA2B,CAAA;AACxF,YAAY,EAAE,qBAAqB,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAA;AAC5F,YAAY,EACX,YAAY,EACZ,kBAAkB,EAClB,wBAAwB,EACxB,qBAAqB,EACrB,qBAAqB,EACrB,iBAAiB,EACjB,aAAa,GACb,MAAM,2BAA2B,CAAA;AAClC,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAA;AAC7D,YAAY,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAA;AACvD,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AACxC,YAAY,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAA;AACnD,OAAO,EACN,gBAAgB,EAChB,cAAc,EACd,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,GAChB,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA;AACrE,YAAY,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AACxD,YAAY,EACX,UAAU,EACV,UAAU,EACV,QAAQ,EACR,QAAQ,EACR,gBAAgB,EAChB,KAAK,EACL,MAAM,GACN,MAAM,YAAY,CAAA"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/eval/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,gBAAgB,CAAA;AACpE,OAAO,EAAE,oBAAoB,EAAE,sBAAsB,EAAE,MAAM,2BAA2B,CAAA;AACxF,YAAY,EAAE,qBAAqB,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAA;AAC5F,YAAY,EACX,YAAY,EACZ,kBAAkB,EAClB,wBAAwB,EACxB,qBAAqB,EACrB,qBAAqB,EACrB,iBAAiB,EACjB,aAAa,GACb,MAAM,2BAA2B,CAAA;AAClC,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAA;AAC7D,YAAY,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAA;AACvD,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AACxC,YAAY,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAA;AACnD,OAAO,EACN,mBAAmB,EACnB,gBAAgB,EAChB,gBAAgB,EAChB,cAAc,EACd,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,gBAAgB,GAChB,MAAM,cAAc,CAAA;AACrB,OAAO,EACN,mBAAmB,EACnB,iBAAiB,EACjB,cAAc,EACd,cAAc,EACd,sBAAsB,EACtB,cAAc,EACd,yBAAyB,EACzB,kBAAkB,EAClB,cAAc,EACd,gBAAgB,GAChB,MAAM,gBAAgB,CAAA;AACvB,YAAY,EACX,eAAe,EACf,aAAa,EACb,SAAS,EACT,mBAAmB,EACnB,QAAQ,EACR,WAAW,GACX,MAAM,gBAAgB,CAAA;AACvB,YAAY,EACX,gBAAgB,EAChB,cAAc,EACd,iBAAiB,EACjB,OAAO,EACP,UAAU,GACV,MAAM,iBAAiB,CAAA;AACxB,OAAO,EAAE,UAAU,EAAE,MAAM,iBAAiB,CAAA;AAC5C,OAAO,EAAE,qBAAqB,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAA;AAC5E,YAAY,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAA;AACtD,OAAO,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAA;AACjE,YAAY,EAAE,aAAa,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAA;AACjE,YAAY,EAAE,iBAAiB,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAA;AACxE,OAAO,EAAE,kBAAkB,EAAE,oBAAoB,EAAE,kBAAkB,EAAE,MAAM,cAAc,CAAA;AAC3F,YAAY,EACX,OAAO,EACP,oBAAoB,EACpB,oBAAoB,EACpB,mBAAmB,EACnB,iBAAiB,EACjB,cAAc,EACd,KAAK,EACL,gBAAgB,EAChB,WAAW,EACX,WAAW,GACX,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA;AACrE,YAAY,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AACxD,YAAY,EACX,iBAAiB,EACjB,UAAU,EACV,SAAS,EACT,UAAU,EACV,QAAQ,EACR,QAAQ,EACR,gBAAgB,EAChB,KAAK,EACL,MAAM,GACN,MAAM,YAAY,CAAA"}
@@ -2,6 +2,11 @@ export { evalTurnFromQuery, evalTurnFromTurn } from './from-turn.js';
2
2
  export { compareHarnessTrials, reviewHarnessCandidate } from './harness-verification.js';
3
3
  export { formatReport, runExperiment } from './experiment.js';
4
4
  export { judgeScorer } from './judge.js';
5
- export { completionScorer, containsScorer, customScorer, stepBudgetScorer, trajectoryScorer, } from './scorers.js';
5
+ export { claimMismatchScorer, handedBackScorer, completionScorer, containsScorer, customScorer, outcomeScorer, stepBudgetScorer, trajectoryScorer, } from './scorers.js';
6
+ export { BOOTSTRAP_RESAMPLES, DEFAULT_EVAL_SEED, EvalUsageError, INTERVAL_LEVEL, MIN_CASES_FOR_INTERVAL, bootstrapCases, bootstrapPairedDifference, createSeededRandom, exactSignTestP, isEvalUsageError, } from './bootstrap.js';
7
+ export { sdkVersion } from './disclosure.js';
8
+ export { currentEvalRunOptions, withEvalRunOptions } from './run-options.js';
9
+ export { formatRunSummary, summarizeRun } from './run-summary.js';
10
+ export { compareEvalReports, formatEvalComparison, loadEvalReportFile } from './compare.js';
6
11
  export { describeUncertainty, uncertaintyOf } from './uncertainty.js';
7
12
  //# sourceMappingURL=index.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/eval/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,gBAAgB,CAAA;AACpE,OAAO,EAAE,oBAAoB,EAAE,sBAAsB,EAAE,MAAM,2BAA2B,CAAA;AAWxF,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAA;AAE7D,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AAExC,OAAO,EACN,gBAAgB,EAChB,cAAc,EACd,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,GAChB,MAAM,cAAc,CAAA;AACrB,OAAO,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/eval/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,gBAAgB,CAAA;AACpE,OAAO,EAAE,oBAAoB,EAAE,sBAAsB,EAAE,MAAM,2BAA2B,CAAA;AAWxF,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAA;AAE7D,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AAExC,OAAO,EACN,mBAAmB,EACnB,gBAAgB,EAChB,gBAAgB,EAChB,cAAc,EACd,YAAY,EACZ,aAAa,EACb,gBAAgB,EAChB,gBAAgB,GAChB,MAAM,cAAc,CAAA;AACrB,OAAO,EACN,mBAAmB,EACnB,iBAAiB,EACjB,cAAc,EACd,cAAc,EACd,sBAAsB,EACtB,cAAc,EACd,yBAAyB,EACzB,kBAAkB,EAClB,cAAc,EACd,gBAAgB,GAChB,MAAM,gBAAgB,CAAA;AAgBvB,OAAO,EAAE,UAAU,EAAE,MAAM,iBAAiB,CAAA;AAC5C,OAAO,EAAE,qBAAqB,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAA;AAE5E,OAAO,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAA;AAGjE,OAAO,EAAE,kBAAkB,EAAE,oBAAoB,EAAE,kBAAkB,EAAE,MAAM,cAAc,CAAA;AAa3F,OAAO,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA"}
@@ -0,0 +1,77 @@
1
+ import { type CaseIntervals, type CaseTally } from './bootstrap.js';
2
+ import type { CaseRepeatSummary, CaseResult, EvalSplit } from './types.js';
3
+ /**
4
+ * The numbers a report carries about a set of runs, whether that set is one
5
+ * suite, one split of it, or every suite in a run.
6
+ *
7
+ * Cases are the unit. A case with several repeats is one case with several
8
+ * runs: the pass rate pools the runs, the intervals resample the cases, and
9
+ * reliability asks of each case whether it held up every time.
10
+ */
11
+ export interface RepeatReliability {
12
+ /** Cases that ran. */
13
+ cases: number;
14
+ /** Cases whose every repeat passed. */
15
+ passedEveryRepeat: number;
16
+ /** `passedEveryRepeat / cases`; null when no case ran. */
17
+ fraction: number | null;
18
+ }
19
+ export interface SplitSummary {
20
+ cases: number;
21
+ runs: number;
22
+ passed: number;
23
+ failed: number;
24
+ inconclusive: number;
25
+ passRate: number | null;
26
+ mean: number | null;
27
+ intervals: CaseIntervals;
28
+ }
29
+ /** The runs of one case, with the key that pairs it with the same case elsewhere. */
30
+ export interface CaseRuns {
31
+ readonly key: string;
32
+ readonly name: string;
33
+ readonly split?: EvalSplit;
34
+ readonly runs: readonly CaseResult[];
35
+ }
36
+ export interface RunStats {
37
+ cases: number;
38
+ runs: number;
39
+ passed: number;
40
+ failed: number;
41
+ inconclusive: number;
42
+ passRate: number | null;
43
+ mean: number | null;
44
+ intervals: CaseIntervals;
45
+ caseSummaries: CaseRepeatSummary[];
46
+ reliability: RepeatReliability;
47
+ splits: Partial<Record<EvalSplit | 'unassigned', SplitSummary>>;
48
+ }
49
+ export declare const SPLIT_ORDER: readonly ["development", "held-out", "unassigned"];
50
+ /** Group a flat list of runs by case name, in the order the cases first appear. */
51
+ export declare function groupRuns(results: readonly CaseResult[], keyOf?: (name: string) => string): CaseRuns[];
52
+ /** The first name that appears more than once, or undefined. */
53
+ export declare function duplicateName(names: readonly string[]): string | undefined;
54
+ export declare function duplicateMessage(suite: string, name: string): string;
55
+ /**
56
+ * The first case in a list of runs that appears twice for the same repeat: a
57
+ * second run of a case is a repeat only when it says it is a later one.
58
+ */
59
+ export declare function duplicateRun(runs: readonly Pick<CaseResult, 'case' | 'repeat'>[]): string | undefined;
60
+ /** Status of a run, tolerating a report old enough to carry only `passed`. */
61
+ export declare function statusOf(run: Pick<CaseResult, 'status' | 'passed'>): CaseResult['status'];
62
+ export declare function tallyOf(runs: readonly CaseResult[]): CaseTally | null;
63
+ /** True when the case has at least one judged run and every run passed. */
64
+ export declare function passedEveryRun(runs: readonly CaseResult[]): boolean;
65
+ export declare function statsOver(cases: readonly CaseRuns[], seed: number): RunStats;
66
+ /**
67
+ * The intervals, as lines a person reads next to the numbers. A number whose
68
+ * interval is not computed is printed with the reason, never alone.
69
+ */
70
+ export declare function describeIntervals(args: {
71
+ mean: number | null;
72
+ passRate: number | null;
73
+ intervals: CaseIntervals;
74
+ }): string[];
75
+ /** The line that says which splits a suite has, and says so plainly when none is held out. */
76
+ export declare function describeSplits(splits: RunStats['splits']): string[];
77
+ //# sourceMappingURL=report-stats.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"report-stats.d.ts","sourceRoot":"","sources":["../../src/eval/report-stats.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,aAAa,EAAE,KAAK,SAAS,EAAkB,MAAM,gBAAgB,CAAA;AACnF,OAAO,KAAK,EAAE,iBAAiB,EAAE,UAAU,EAAE,SAAS,EAAE,MAAM,YAAY,CAAA;AAE1E;;;;;;;GAOG;AAEH,MAAM,WAAW,iBAAiB;IACjC,sBAAsB;IACtB,KAAK,EAAE,MAAM,CAAA;IACb,uCAAuC;IACvC,iBAAiB,EAAE,MAAM,CAAA;IACzB,0DAA0D;IAC1D,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAA;CACvB;AAED,MAAM,WAAW,YAAY;IAC5B,KAAK,EAAE,MAAM,CAAA;IACb,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,EAAE,MAAM,CAAA;IACd,MAAM,EAAE,MAAM,CAAA;IACd,YAAY,EAAE,MAAM,CAAA;IACpB,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAA;IACvB,IAAI,EAAE,MAAM,GAAG,IAAI,CAAA;IACnB,SAAS,EAAE,aAAa,CAAA;CACxB;AAED,qFAAqF;AACrF,MAAM,WAAW,QAAQ;IACxB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;IACpB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;IACrB,QAAQ,CAAC,KAAK,CAAC,EAAE,SAAS,CAAA;IAC1B,QAAQ,CAAC,IAAI,EAAE,SAAS,UAAU,EAAE,CAAA;CACpC;AAED,MAAM,WAAW,QAAQ;IACxB,KAAK,EAAE,MAAM,CAAA;IACb,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,EAAE,MAAM,CAAA;IACd,MAAM,EAAE,MAAM,CAAA;IACd,YAAY,EAAE,MAAM,CAAA;IACpB,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAA;IACvB,IAAI,EAAE,MAAM,GAAG,IAAI,CAAA;IACnB,SAAS,EAAE,aAAa,CAAA;IACxB,aAAa,EAAE,iBAAiB,EAAE,CAAA;IAClC,WAAW,EAAE,iBAAiB,CAAA;IAC9B,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,SAAS,GAAG,YAAY,EAAE,YAAY,CAAC,CAAC,CAAA;CAC/D;AAED,eAAO,MAAM,WAAW,oDAAqD,CAAA;AAE7E,mFAAmF;AACnF,wBAAgB,SAAS,CACxB,OAAO,EAAE,SAAS,UAAU,EAAE,EAC9B,KAAK,GAAE,CAAC,IAAI,EAAE,MAAM,KAAK,MAAuB,GAC9C,QAAQ,EAAE,CAYZ;AAED,gEAAgE;AAChE,wBAAgB,aAAa,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,GAAG,SAAS,CAO1E;AAED,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,CAEpE;AAED;;;GAGG;AACH,wBAAgB,YAAY,CAC3B,IAAI,EAAE,SAAS,IAAI,CAAC,UAAU,EAAE,MAAM,GAAG,QAAQ,CAAC,EAAE,GAClD,MAAM,GAAG,SAAS,CAQpB;AAED,8EAA8E;AAC9E,wBAAgB,QAAQ,CAAC,GAAG,EAAE,IAAI,CAAC,UAAU,EAAE,QAAQ,GAAG,QAAQ,CAAC,GAAG,UAAU,CAAC,QAAQ,CAAC,CAKzF;AAED,wBAAgB,OAAO,CAAC,IAAI,EAAE,SAAS,UAAU,EAAE,GAAG,SAAS,GAAG,IAAI,CAYrE;AAED,2EAA2E;AAC3E,wBAAgB,cAAc,CAAC,IAAI,EAAE,SAAS,UAAU,EAAE,GAAG,OAAO,CAEnE;AAmCD,wBAAgB,SAAS,CAAC,KAAK,EAAE,SAAS,QAAQ,EAAE,EAAE,IAAI,EAAE,MAAM,GAAG,QAAQ,CAmC5E;AAKD;;;GAGG;AACH,wBAAgB,iBAAiB,CAAC,IAAI,EAAE;IACvC,IAAI,EAAE,MAAM,GAAG,IAAI,CAAA;IACnB,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAA;IACvB,SAAS,EAAE,aAAa,CAAA;CACxB,GAAG,MAAM,EAAE,CAsBX;AAED,8FAA8F;AAC9F,wBAAgB,cAAc,CAAC,MAAM,EAAE,QAAQ,CAAC,QAAQ,CAAC,GAAG,MAAM,EAAE,CAmBnE"}
@@ -0,0 +1,184 @@
1
+ import { bootstrapCases } from './bootstrap.js';
2
+ export const SPLIT_ORDER = ['development', 'held-out', 'unassigned'];
3
+ /** Group a flat list of runs by case name, in the order the cases first appear. */
4
+ export function groupRuns(results, keyOf = (name) => name) {
5
+ const groups = new Map();
6
+ for (const run of results) {
7
+ const key = keyOf(run.case);
8
+ let group = groups.get(key);
9
+ if (!group) {
10
+ group = { name: run.case, ...(run.split ? { split: run.split } : {}), runs: [] };
11
+ groups.set(key, group);
12
+ }
13
+ group.runs.push(run);
14
+ }
15
+ return [...groups].map(([key, g]) => ({ key, ...g }));
16
+ }
17
+ /** The first name that appears more than once, or undefined. */
18
+ export function duplicateName(names) {
19
+ const seen = new Set();
20
+ for (const name of names) {
21
+ if (seen.has(name))
22
+ return name;
23
+ seen.add(name);
24
+ }
25
+ return undefined;
26
+ }
27
+ export function duplicateMessage(suite, name) {
28
+ return `suite "${suite}" has the case "${name}" more than once; a case is named by its id, and two cases sharing one would be merged into one in every count and interval. Give each case its own name.`;
29
+ }
30
+ /**
31
+ * The first case in a list of runs that appears twice for the same repeat: a
32
+ * second run of a case is a repeat only when it says it is a later one.
33
+ */
34
+ export function duplicateRun(runs) {
35
+ const seen = new Set();
36
+ for (const run of runs) {
37
+ const key = `${run.case}\u0000${run.repeat ?? 1}`;
38
+ if (seen.has(key))
39
+ return run.case;
40
+ seen.add(key);
41
+ }
42
+ return undefined;
43
+ }
44
+ /** Status of a run, tolerating a report old enough to carry only `passed`. */
45
+ export function statusOf(run) {
46
+ if (run.status === 'passed' || run.status === 'failed' || run.status === 'inconclusive') {
47
+ return run.status;
48
+ }
49
+ return run.passed === true ? 'passed' : 'failed';
50
+ }
51
+ export function tallyOf(runs) {
52
+ let scoreSum = 0;
53
+ let passes = 0;
54
+ let judged = 0;
55
+ for (const run of runs) {
56
+ const status = statusOf(run);
57
+ if (status === 'inconclusive')
58
+ continue;
59
+ judged++;
60
+ scoreSum += run.mean;
61
+ if (status === 'passed')
62
+ passes++;
63
+ }
64
+ return judged === 0 ? null : { scoreSum, passes, judged };
65
+ }
66
+ /** True when the case has at least one judged run and every run passed. */
67
+ export function passedEveryRun(runs) {
68
+ return runs.length > 0 && runs.every((run) => statusOf(run) === 'passed');
69
+ }
70
+ function summaryOf(runs) {
71
+ let passed = 0;
72
+ let failed = 0;
73
+ let inconclusive = 0;
74
+ for (const run of runs) {
75
+ const status = statusOf(run);
76
+ if (status === 'passed')
77
+ passed++;
78
+ else if (status === 'failed')
79
+ failed++;
80
+ else
81
+ inconclusive++;
82
+ }
83
+ return { passed, failed, inconclusive };
84
+ }
85
+ function numbersOver(cases, seed) {
86
+ const tallies = [];
87
+ let runs = 0;
88
+ for (const c of cases) {
89
+ runs += c.runs.length;
90
+ const tally = tallyOf(c.runs);
91
+ if (tally)
92
+ tallies.push(tally);
93
+ }
94
+ const counts = summaryOf(cases.flatMap((c) => c.runs));
95
+ const judged = counts.passed + counts.failed;
96
+ const scoreSum = tallies.reduce((sum, t) => sum + t.scoreSum, 0);
97
+ return {
98
+ runs,
99
+ ...counts,
100
+ passRate: judged === 0 ? null : counts.passed / judged,
101
+ mean: judged === 0 ? null : scoreSum / judged,
102
+ intervals: bootstrapCases(tallies, seed),
103
+ };
104
+ }
105
+ export function statsOver(cases, seed) {
106
+ const whole = numbersOver(cases, seed);
107
+ const caseSummaries = cases.map((c) => {
108
+ const counts = summaryOf(c.runs);
109
+ const tally = tallyOf(c.runs);
110
+ return {
111
+ case: c.name,
112
+ ...(c.split ? { split: c.split } : {}),
113
+ repeats: c.runs.length,
114
+ ...counts,
115
+ mean: tally ? tally.scoreSum / tally.judged : null,
116
+ passedEveryRepeat: passedEveryRun(c.runs),
117
+ };
118
+ });
119
+ const everyRepeat = caseSummaries.filter((c) => c.passedEveryRepeat).length;
120
+ const splits = {};
121
+ for (const name of SPLIT_ORDER) {
122
+ const members = cases.filter((c) => (c.split ?? 'unassigned') === name);
123
+ if (members.length === 0)
124
+ continue;
125
+ const part = numbersOver(members, seed);
126
+ splits[name] = { cases: members.length, ...part };
127
+ }
128
+ return {
129
+ cases: cases.length,
130
+ ...whole,
131
+ caseSummaries,
132
+ reliability: {
133
+ cases: cases.length,
134
+ passedEveryRepeat: everyRepeat,
135
+ fraction: cases.length === 0 ? null : everyRepeat / cases.length,
136
+ },
137
+ splits,
138
+ };
139
+ }
140
+ const percent = (value) => `${(value * 100).toFixed(1)}%`;
141
+ const score = (value) => value.toFixed(3);
142
+ /**
143
+ * The intervals, as lines a person reads next to the numbers. A number whose
144
+ * interval is not computed is printed with the reason, never alone.
145
+ */
146
+ export function describeIntervals(args) {
147
+ const { intervals } = args;
148
+ const line = (label, estimate, interval, show) => {
149
+ if (estimate === null)
150
+ return `${label}: none, no run was judged`;
151
+ if (!interval.computed)
152
+ return `${label} ${show(estimate)}, no interval: ${interval.reason}`;
153
+ return `${label} ${show(estimate)}, 95% interval ${show(interval.low)} to ${show(interval.high)}`;
154
+ };
155
+ const lines = [
156
+ line('mean score', args.mean, intervals.mean, score),
157
+ line('pass rate', args.passRate, intervals.passRate, percent),
158
+ ];
159
+ if (intervals.mean.computed || intervals.passRate.computed) {
160
+ lines.push(`intervals: ${intervals.method.kind} (${intervals.method.cases} cases, ${intervals.method.resamples} resamples), seed ${intervals.method.seed}`);
161
+ }
162
+ return lines;
163
+ }
164
+ /** The line that says which splits a suite has, and says so plainly when none is held out. */
165
+ export function describeSplits(splits) {
166
+ const lines = [];
167
+ // A suite that declares no split has one bucket and nothing to break down.
168
+ const declared = SPLIT_ORDER.some((name) => name !== 'unassigned' && splits[name]);
169
+ for (const name of SPLIT_ORDER) {
170
+ const part = splits[name];
171
+ if (!part || !declared)
172
+ continue;
173
+ const rate = part.passRate === null ? 'no judged run' : `pass rate ${percent(part.passRate)}`;
174
+ const interval = part.intervals.passRate.computed
175
+ ? `, 95% interval ${percent(part.intervals.passRate.low)} to ${percent(part.intervals.passRate.high)}`
176
+ : '';
177
+ lines.push(`${name}: ${part.cases} case${part.cases === 1 ? '' : 's'}, ${rate}${interval}`);
178
+ }
179
+ if (!splits['held-out']) {
180
+ lines.push('no held-out cases: every number here is measured on cases the suite was developed against');
181
+ }
182
+ return lines;
183
+ }
184
+ //# sourceMappingURL=report-stats.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"report-stats.js","sourceRoot":"","sources":["../../src/eval/report-stats.ts"],"names":[],"mappings":"AAAA,OAAO,EAAsC,cAAc,EAAE,MAAM,gBAAgB,CAAA;AAsDnF,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,aAAa,EAAE,UAAU,EAAE,YAAY,CAAU,CAAA;AAE7E,mFAAmF;AACnF,MAAM,UAAU,SAAS,CACxB,OAA8B,EAC9B,QAAkC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI;IAEhD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAmE,CAAA;IACzF,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;QAC3B,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAC3B,IAAI,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QAC3B,IAAI,CAAC,KAAK,EAAE,CAAC;YACZ,KAAK,GAAG,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAA;YAChF,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAA;QACvB,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACrB,CAAC;IACD,OAAO,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC,CAAA;AACtD,CAAC;AAED,gEAAgE;AAChE,MAAM,UAAU,aAAa,CAAC,KAAwB;IACrD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC;YAAE,OAAO,IAAI,CAAA;QAC/B,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;IACf,CAAC;IACD,OAAO,SAAS,CAAA;AACjB,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,KAAa,EAAE,IAAY;IAC3D,OAAO,UAAU,KAAK,mBAAmB,IAAI,2JAA2J,CAAA;AACzM,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAC3B,IAAoD;IAEpD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACxB,MAAM,GAAG,GAAG,GAAG,GAAG,CAAC,IAAI,SAAS,GAAG,CAAC,MAAM,IAAI,CAAC,EAAE,CAAA;QACjD,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,OAAO,GAAG,CAAC,IAAI,CAAA;QAClC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;IACd,CAAC;IACD,OAAO,SAAS,CAAA;AACjB,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,QAAQ,CAAC,GAA0C;IAClE,IAAI,GAAG,CAAC,MAAM,KAAK,QAAQ,IAAI,GAAG,CAAC,MAAM,KAAK,QAAQ,IAAI,GAAG,CAAC,MAAM,KAAK,cAAc,EAAE,CAAC;QACzF,OAAO,GAAG,CAAC,MAAM,CAAA;IAClB,CAAC;IACD,OAAO,GAAG,CAAC,MAAM,KAAK,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAA;AACjD,CAAC;AAED,MAAM,UAAU,OAAO,CAAC,IAA2B;IAClD,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACxB,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAA;QAC5B,IAAI,MAAM,KAAK,cAAc;YAAE,SAAQ;QACvC,MAAM,EAAE,CAAA;QACR,QAAQ,IAAI,GAAG,CAAC,IAAI,CAAA;QACpB,IAAI,MAAM,KAAK,QAAQ;YAAE,MAAM,EAAE,CAAA;IAClC,CAAC;IACD,OAAO,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,CAAA;AAC1D,CAAC;AAED,2EAA2E;AAC3E,MAAM,UAAU,cAAc,CAAC,IAA2B;IACzD,OAAO,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,KAAK,QAAQ,CAAC,CAAA;AAC1E,CAAC;AAED,SAAS,SAAS,CAAC,IAA2B;IAC7C,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACxB,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAA;QAC5B,IAAI,MAAM,KAAK,QAAQ;YAAE,MAAM,EAAE,CAAA;aAC5B,IAAI,MAAM,KAAK,QAAQ;YAAE,MAAM,EAAE,CAAA;;YACjC,YAAY,EAAE,CAAA;IACpB,CAAC;IACD,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,YAAY,EAAE,CAAA;AACxC,CAAC;AAED,SAAS,WAAW,CAAC,KAA0B,EAAE,IAAY;IAC5D,MAAM,OAAO,GAAgB,EAAE,CAAA;IAC/B,IAAI,IAAI,GAAG,CAAC,CAAA;IACZ,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACvB,IAAI,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAA;QACrB,MAAM,KAAK,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;QAC7B,IAAI,KAAK;YAAE,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;IAC/B,CAAC;IACD,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;IACtD,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAA;IAC5C,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAA;IAChE,OAAO;QACN,IAAI;QACJ,GAAG,MAAM;QACT,QAAQ,EAAE,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,GAAG,MAAM;QACtD,IAAI,EAAE,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,GAAG,MAAM;QAC7C,SAAS,EAAE,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC;KACxC,CAAA;AACF,CAAC;AAED,MAAM,UAAU,SAAS,CAAC,KAA0B,EAAE,IAAY;IACjE,MAAM,KAAK,GAAG,WAAW,CAAC,KAAK,EAAE,IAAI,CAAC,CAAA;IACtC,MAAM,aAAa,GAAwB,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QAC1D,MAAM,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;QAChC,MAAM,KAAK,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;QAC7B,OAAO;YACN,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YACtC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM;YACtB,GAAG,MAAM;YACT,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI;YAClD,iBAAiB,EAAE,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC;SACzC,CAAA;IACF,CAAC,CAAC,CAAA;IACF,MAAM,WAAW,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,MAAM,CAAA;IAE3E,MAAM,MAAM,GAAuB,EAAE,CAAA;IACrC,KAAK,MAAM,IAAI,IAAI,WAAW,EAAE,CAAC;QAChC,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,YAAY,CAAC,KAAK,IAAI,CAAC,CAAA;QACvE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,SAAQ;QAClC,MAAM,IAAI,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,CAAA;QACvC,MAAM,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,EAAE,OAAO,CAAC,MAAM,EAAE,GAAG,IAAI,EAAE,CAAA;IAClD,CAAC;IAED,OAAO;QACN,KAAK,EAAE,KAAK,CAAC,MAAM;QACnB,GAAG,KAAK;QACR,aAAa;QACb,WAAW,EAAE;YACZ,KAAK,EAAE,KAAK,CAAC,MAAM;YACnB,iBAAiB,EAAE,WAAW;YAC9B,QAAQ,EAAE,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,WAAW,GAAG,KAAK,CAAC,MAAM;SAChE;QACD,MAAM;KACN,CAAA;AACF,CAAC;AAED,MAAM,OAAO,GAAG,CAAC,KAAa,EAAU,EAAE,CAAC,GAAG,CAAC,KAAK,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAA;AACzE,MAAM,KAAK,GAAG,CAAC,KAAa,EAAU,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAA;AAEzD;;;GAGG;AACH,MAAM,UAAU,iBAAiB,CAAC,IAIjC;IACA,MAAM,EAAE,SAAS,EAAE,GAAG,IAAI,CAAA;IAC1B,MAAM,IAAI,GAAG,CACZ,KAAa,EACb,QAAuB,EACvB,QAA+B,EAC/B,IAA2B,EAClB,EAAE;QACX,IAAI,QAAQ,KAAK,IAAI;YAAE,OAAO,GAAG,KAAK,2BAA2B,CAAA;QACjE,IAAI,CAAC,QAAQ,CAAC,QAAQ;YAAE,OAAO,GAAG,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,kBAAkB,QAAQ,CAAC,MAAM,EAAE,CAAA;QAC5F,OAAO,GAAG,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,kBAAkB,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAA;IAClG,CAAC,CAAA;IACD,MAAM,KAAK,GAAG;QACb,IAAI,CAAC,YAAY,EAAE,IAAI,CAAC,IAAI,EAAE,SAAS,CAAC,IAAI,EAAE,KAAK,CAAC;QACpD,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC,QAAQ,EAAE,SAAS,CAAC,QAAQ,EAAE,OAAO,CAAC;KAC7D,CAAA;IACD,IAAI,SAAS,CAAC,IAAI,CAAC,QAAQ,IAAI,SAAS,CAAC,QAAQ,CAAC,QAAQ,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CACT,cAAc,SAAS,CAAC,MAAM,CAAC,IAAI,KAAK,SAAS,CAAC,MAAM,CAAC,KAAK,WAAW,SAAS,CAAC,MAAM,CAAC,SAAS,qBAAqB,SAAS,CAAC,MAAM,CAAC,IAAI,EAAE,CAC/I,CAAA;IACF,CAAC;IACD,OAAO,KAAK,CAAA;AACb,CAAC;AAED,8FAA8F;AAC9F,MAAM,UAAU,cAAc,CAAC,MAA0B;IACxD,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,2EAA2E;IAC3E,MAAM,QAAQ,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,YAAY,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,CAAA;IAClF,KAAK,MAAM,IAAI,IAAI,WAAW,EAAE,CAAC;QAChC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,CAAA;QACzB,IAAI,CAAC,IAAI,IAAI,CAAC,QAAQ;YAAE,SAAQ;QAChC,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,aAAa,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAA;QAC7F,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,QAAQ;YAChD,CAAC,CAAC,kBAAkB,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,OAAO,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE;YACtG,CAAC,CAAC,EAAE,CAAA;QACL,KAAK,CAAC,IAAI,CAAC,GAAG,IAAI,KAAK,IAAI,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,IAAI,GAAG,QAAQ,EAAE,CAAC,CAAA;IAC5F,CAAC;IACD,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CACT,2FAA2F,CAC3F,CAAA;IACF,CAAC;IACD,OAAO,KAAK,CAAA;AACb,CAAC"}
@@ -0,0 +1,30 @@
1
+ /**
2
+ * Settings an operator gives a whole eval run, handed to every
3
+ * `runExperiment` that executes inside it.
4
+ *
5
+ * A suite file calls `runExperiment` itself and returns the report, so
6
+ * `namzu eval --repeat 5` has no argument to pass the five through. Rather
7
+ * than ask every suite to forward it, the runner opens a scope with
8
+ * {@link withEvalRunOptions} and `runExperiment` reads the scope. A suite
9
+ * that was written before the option existed repeats like any other.
10
+ *
11
+ * An option given here wins over the same option in the suite's own config:
12
+ * the person running the command asked for it.
13
+ */
14
+ export interface EvalRunOptions {
15
+ /** Times every case runs. */
16
+ readonly repeats?: number;
17
+ /** Seed for the bootstrap. */
18
+ readonly seed?: number;
19
+ /** Version of the program that is running the suites, for the disclosure. */
20
+ readonly harnessVersion?: string;
21
+ /** Source revision the suites came from, when known. */
22
+ readonly commit?: string;
23
+ /** Version of the suite being run, when known. */
24
+ readonly suiteVersion?: string;
25
+ }
26
+ /** Run `work` with `options` in force for every experiment it starts. */
27
+ export declare function withEvalRunOptions<T>(options: EvalRunOptions, work: () => T): T;
28
+ /** The options in force here, or an empty set outside any scope. */
29
+ export declare function currentEvalRunOptions(): EvalRunOptions;
30
+ //# sourceMappingURL=run-options.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"run-options.d.ts","sourceRoot":"","sources":["../../src/eval/run-options.ts"],"names":[],"mappings":"AAEA;;;;;;;;;;;;GAYG;AACH,MAAM,WAAW,cAAc;IAC9B,6BAA6B;IAC7B,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAA;IACzB,8BAA8B;IAC9B,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,CAAA;IACtB,6EAA6E;IAC7E,QAAQ,CAAC,cAAc,CAAC,EAAE,MAAM,CAAA;IAChC,wDAAwD;IACxD,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAA;IACxB,kDAAkD;IAClD,QAAQ,CAAC,YAAY,CAAC,EAAE,MAAM,CAAA;CAC9B;AAcD,yEAAyE;AACzE,wBAAgB,kBAAkB,CAAC,CAAC,EAAE,OAAO,EAAE,cAAc,EAAE,IAAI,EAAE,MAAM,CAAC,GAAG,CAAC,CAE/E;AAED,oEAAoE;AACpE,wBAAgB,qBAAqB,IAAI,cAAc,CAEtD"}