niceeval 0.18.0 → 0.19.0

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 (679) hide show
  1. package/INDEX.md +4 -2
  2. package/dist/adapter-usage.cjs.map +1 -1
  3. package/dist/adapter-usage.d.cts +1 -1
  4. package/dist/adapter-usage.d.mts +1 -1
  5. package/dist/adapter-usage.d.ts +1 -1
  6. package/dist/adapter.cjs +8 -7
  7. package/dist/adapter.cjs.map +1 -1
  8. package/dist/adapter.d.cts +21 -21
  9. package/dist/adapter.d.mts +21 -21
  10. package/dist/adapter.d.ts +21 -21
  11. package/dist/agents/{effect-runtime.cjs → agent-effects.cjs} +12 -18
  12. package/dist/agents/agent-effects.cjs.map +1 -0
  13. package/dist/agents/agent-effects.d.cts +23 -0
  14. package/dist/agents/agent-effects.d.mts +23 -0
  15. package/dist/agents/agent-effects.d.ts +23 -0
  16. package/dist/agents/agent-effects.js +16 -0
  17. package/dist/agents/agent-effects.js.map +1 -0
  18. package/dist/agents/ai-sdk.cjs +1 -1
  19. package/dist/agents/ai-sdk.cjs.map +1 -1
  20. package/dist/agents/bub.cjs +31 -18
  21. package/dist/agents/bub.cjs.map +1 -1
  22. package/dist/agents/bub.d.cts +9 -12
  23. package/dist/agents/bub.d.mts +9 -12
  24. package/dist/agents/bub.d.ts +9 -12
  25. package/dist/agents/claude-code-native.cjs +2 -2
  26. package/dist/agents/claude-code-native.cjs.map +1 -1
  27. package/dist/agents/claude-code.cjs +17 -11
  28. package/dist/agents/claude-code.cjs.map +1 -1
  29. package/dist/agents/claude-code.d.cts +11 -11
  30. package/dist/agents/claude-code.d.mts +11 -11
  31. package/dist/agents/claude-code.d.ts +11 -11
  32. package/dist/agents/codex-app-server.cjs +2 -2
  33. package/dist/agents/codex-app-server.cjs.map +1 -1
  34. package/dist/agents/codex.cjs +18 -12
  35. package/dist/agents/codex.cjs.map +1 -1
  36. package/dist/agents/codex.d.cts +12 -12
  37. package/dist/agents/codex.d.mts +12 -12
  38. package/dist/agents/codex.d.ts +12 -12
  39. package/dist/agents/deepseek-harness.cjs +2 -2
  40. package/dist/agents/deepseek-harness.cjs.map +1 -1
  41. package/dist/agents/hermes.cjs +3 -3
  42. package/dist/agents/hermes.cjs.map +1 -1
  43. package/dist/agents/index.cjs +3 -3
  44. package/dist/agents/index.cjs.map +1 -1
  45. package/dist/agents/index.d.cts +3 -3
  46. package/dist/agents/index.d.mts +3 -3
  47. package/dist/agents/index.d.ts +3 -3
  48. package/dist/agents/index.js +2 -2
  49. package/dist/agents/index.mjs +2 -2
  50. package/dist/agents/mcp.cjs.map +1 -1
  51. package/dist/agents/mcp.d.cts +2 -2
  52. package/dist/agents/mcp.d.mts +2 -2
  53. package/dist/agents/mcp.d.ts +2 -2
  54. package/dist/agents/omp.cjs +2 -2
  55. package/dist/agents/omp.cjs.map +1 -1
  56. package/dist/agents/openai-compat.cjs +2 -2
  57. package/dist/agents/openai-compat.cjs.map +1 -1
  58. package/dist/agents/openclaw.cjs +3 -3
  59. package/dist/agents/openclaw.cjs.map +1 -1
  60. package/dist/agents/opencode.cjs +16 -4
  61. package/dist/agents/opencode.cjs.map +1 -1
  62. package/dist/agents/plugin-extensions.cjs +71 -0
  63. package/dist/agents/plugin-extensions.cjs.map +1 -0
  64. package/dist/agents/plugin-extensions.d.cts +26 -0
  65. package/dist/agents/plugin-extensions.d.mts +26 -0
  66. package/dist/agents/plugin-extensions.d.ts +26 -0
  67. package/dist/agents/plugin-extensions.js +8 -0
  68. package/dist/agents/plugin-extensions.js.map +1 -0
  69. package/dist/agents/post-setup.cjs +14 -3
  70. package/dist/agents/post-setup.cjs.map +1 -1
  71. package/dist/agents/post-setup.d.cts +4 -2
  72. package/dist/agents/post-setup.d.mts +4 -2
  73. package/dist/agents/post-setup.d.ts +4 -2
  74. package/dist/agents/post-setup.js +8 -6
  75. package/dist/agents/sdk-streams.cjs +3 -3
  76. package/dist/agents/sdk-streams.cjs.map +1 -1
  77. package/dist/agents/{session-slot.cjs → session-key.cjs} +4 -4
  78. package/dist/agents/session-key.cjs.map +1 -0
  79. package/dist/agents/session-key.d.cts +12 -0
  80. package/dist/agents/session-key.d.mts +12 -0
  81. package/dist/agents/session-key.d.ts +12 -0
  82. package/dist/agents/session-key.js +6 -0
  83. package/dist/agents/session-key.js.map +1 -0
  84. package/dist/agents/shared.cjs +2 -2
  85. package/dist/agents/shared.cjs.map +1 -1
  86. package/dist/agents/shared.d.cts +4 -4
  87. package/dist/agents/shared.d.mts +4 -4
  88. package/dist/agents/shared.d.ts +4 -4
  89. package/dist/agents/skills.cjs.map +1 -1
  90. package/dist/agents/skills.d.cts +3 -3
  91. package/dist/agents/skills.d.mts +3 -3
  92. package/dist/agents/skills.d.ts +3 -3
  93. package/dist/agents/types.cjs.map +1 -1
  94. package/dist/agents/types.d.cts +55 -54
  95. package/dist/agents/types.d.mts +55 -54
  96. package/dist/agents/types.d.ts +55 -54
  97. package/dist/agents/ui-message-stream.cjs +4 -4
  98. package/dist/agents/ui-message-stream.cjs.map +1 -1
  99. package/dist/agents/unrecognized-events.cjs +32 -0
  100. package/dist/agents/unrecognized-events.cjs.map +1 -0
  101. package/dist/agents/unrecognized-events.d.cts +8 -0
  102. package/dist/agents/unrecognized-events.d.mts +8 -0
  103. package/dist/agents/unrecognized-events.d.ts +8 -0
  104. package/dist/agents/unrecognized-events.js +6 -0
  105. package/dist/agents/unrecognized-events.js.map +1 -0
  106. package/dist/assertions/api.cjs.map +1 -1
  107. package/dist/assertions/api.d.cts +3 -0
  108. package/dist/assertions/api.d.mts +3 -0
  109. package/dist/assertions/api.d.ts +3 -0
  110. package/dist/assertions/collection.cjs +1 -1
  111. package/dist/assertions/collection.cjs.map +1 -1
  112. package/dist/assertions/judge.cjs +1 -1
  113. package/dist/assertions/judge.cjs.map +1 -1
  114. package/dist/assertions/match.cjs +52 -21
  115. package/dist/assertions/match.cjs.map +1 -1
  116. package/dist/assertions/match.d.cts +16 -5
  117. package/dist/assertions/match.d.mts +16 -5
  118. package/dist/assertions/match.d.ts +16 -5
  119. package/dist/assertions/match.js +96 -94
  120. package/dist/assertions/measurements.cjs +22 -6
  121. package/dist/assertions/measurements.cjs.map +1 -1
  122. package/dist/assertions/measurements.d.cts +8 -1
  123. package/dist/assertions/measurements.d.mts +8 -1
  124. package/dist/assertions/measurements.d.ts +8 -1
  125. package/dist/assertions/numeric.cjs +2 -0
  126. package/dist/assertions/numeric.cjs.map +1 -1
  127. package/dist/assertions/numeric.d.cts +1 -1
  128. package/dist/assertions/numeric.d.mts +1 -1
  129. package/dist/assertions/numeric.d.ts +1 -1
  130. package/dist/assertions/quantity.cjs +42 -0
  131. package/dist/assertions/quantity.cjs.map +1 -0
  132. package/dist/assertions/quantity.d.cts +18 -0
  133. package/dist/assertions/quantity.d.mts +18 -0
  134. package/dist/assertions/quantity.d.ts +18 -0
  135. package/dist/assertions/quantity.js +22 -0
  136. package/dist/assertions/quantity.js.map +1 -0
  137. package/dist/assertions/record/attachment.cjs +2 -0
  138. package/dist/assertions/record/attachment.cjs.map +1 -1
  139. package/dist/assertions/record/codec.cjs +5 -1
  140. package/dist/assertions/record/codec.cjs.map +1 -1
  141. package/dist/assertions/record/model.cjs.map +1 -1
  142. package/dist/assertions/record/model.d.cts +3 -1
  143. package/dist/assertions/record/model.d.mts +3 -1
  144. package/dist/assertions/record/model.d.ts +3 -1
  145. package/dist/assertions/runtime.cjs +11 -6
  146. package/dist/assertions/runtime.cjs.map +1 -1
  147. package/dist/cli/program.cjs +3 -0
  148. package/dist/cli/program.cjs.map +1 -1
  149. package/dist/context/assert-first.cjs +6 -4
  150. package/dist/context/assert-first.cjs.map +1 -1
  151. package/dist/context/assert-first.d.cts +9 -3
  152. package/dist/context/assert-first.d.mts +9 -3
  153. package/dist/context/assert-first.d.ts +9 -3
  154. package/dist/context/notes.cjs +47 -0
  155. package/dist/context/notes.cjs.map +1 -0
  156. package/dist/context/notes.d.cts +10 -0
  157. package/dist/context/notes.d.mts +10 -0
  158. package/dist/context/notes.d.ts +10 -0
  159. package/dist/context/notes.js +6 -0
  160. package/dist/context/notes.js.map +1 -0
  161. package/dist/context/send-failures.cjs +1 -7
  162. package/dist/context/send-failures.cjs.map +1 -1
  163. package/dist/context/send-failures.d.cts +2 -3
  164. package/dist/context/send-failures.d.mts +2 -3
  165. package/dist/context/send-failures.d.ts +2 -3
  166. package/dist/context/send-retry.cjs +0 -1
  167. package/dist/context/send-retry.cjs.map +1 -1
  168. package/dist/context/send-retry.d.cts +1 -2
  169. package/dist/context/send-retry.d.mts +1 -2
  170. package/dist/context/send-retry.d.ts +1 -2
  171. package/dist/context/session.cjs +10 -17
  172. package/dist/context/session.cjs.map +1 -1
  173. package/dist/context/session.d.cts +4 -10
  174. package/dist/context/session.d.mts +4 -10
  175. package/dist/context/session.d.ts +4 -10
  176. package/dist/define.cjs +88 -65
  177. package/dist/define.cjs.map +1 -1
  178. package/dist/define.d.cts +3 -3
  179. package/dist/define.d.mts +3 -3
  180. package/dist/define.d.ts +3 -3
  181. package/dist/eval/cli/contribution.cjs +1 -1
  182. package/dist/eval/cli/contribution.cjs.map +1 -1
  183. package/dist/eval/host/index.cjs +1 -1
  184. package/dist/eval/host/index.cjs.map +1 -1
  185. package/dist/eval/record/evaluation.cjs +1 -1
  186. package/dist/eval/record/evaluation.cjs.map +1 -1
  187. package/dist/eval/record/membership-provenance.d.cts +3 -3
  188. package/dist/eval/record/membership-provenance.d.mts +3 -3
  189. package/dist/eval/record/membership-provenance.d.ts +3 -3
  190. package/dist/eval/record/score.cjs +2 -2
  191. package/dist/eval/record/score.cjs.map +1 -1
  192. package/dist/eval/record/sealed-assertion.cjs +1 -1
  193. package/dist/eval/record/sealed-assertion.cjs.map +1 -1
  194. package/dist/eval/record/verdict.cjs +2 -2
  195. package/dist/eval/record/verdict.cjs.map +1 -1
  196. package/dist/expect/index.cjs +10 -1
  197. package/dist/expect/index.cjs.map +1 -1
  198. package/dist/expect/index.d.cts +3 -1
  199. package/dist/expect/index.d.mts +3 -1
  200. package/dist/expect/index.d.ts +3 -1
  201. package/dist/expect/index.js +92 -76
  202. package/dist/expect/index.mjs +92 -76
  203. package/dist/experiment/host/cli/contribution.cjs +52 -39
  204. package/dist/experiment/host/cli/contribution.cjs.map +1 -1
  205. package/dist/experiment/host/cli/contribution.d.cts +18 -4
  206. package/dist/experiment/host/cli/contribution.d.mts +18 -4
  207. package/dist/experiment/host/cli/contribution.d.ts +18 -4
  208. package/dist/experiment/host/cli/contribution.js +2 -2
  209. package/dist/experiment/host/cli/debug-protocol.cjs +2 -2
  210. package/dist/experiment/host/cli/debug-protocol.cjs.map +1 -1
  211. package/dist/experiment/host/cli/debug-protocol.d.cts +2 -2
  212. package/dist/experiment/host/cli/debug-protocol.d.mts +2 -2
  213. package/dist/experiment/host/cli/debug-protocol.d.ts +2 -2
  214. package/dist/experiment/host/cli/output-protocol.cjs +1 -1
  215. package/dist/experiment/host/cli/output-protocol.cjs.map +1 -1
  216. package/dist/experiment/host/cli/output-protocol.d.cts +2 -2
  217. package/dist/experiment/host/cli/output-protocol.d.mts +2 -2
  218. package/dist/experiment/host/cli/output-protocol.d.ts +2 -2
  219. package/dist/experiment/host/current.cjs +2 -2
  220. package/dist/experiment/host/current.cjs.map +1 -1
  221. package/dist/experiment/host/index.cjs +1 -1
  222. package/dist/experiment/host/index.cjs.map +1 -1
  223. package/dist/experiment/host/operations.cjs +21 -22
  224. package/dist/experiment/host/operations.cjs.map +1 -1
  225. package/dist/experiment/host/runtime.cjs +8 -9
  226. package/dist/experiment/host/runtime.cjs.map +1 -1
  227. package/dist/experiment/host/runtime.d.cts +3 -3
  228. package/dist/experiment/host/runtime.d.mts +3 -3
  229. package/dist/experiment/host/runtime.d.ts +3 -3
  230. package/dist/experiment/host/teardown.cjs +3 -3
  231. package/dist/experiment/host/teardown.cjs.map +1 -1
  232. package/dist/experiment/host/teardown.d.cts +3 -3
  233. package/dist/experiment/host/teardown.d.mts +3 -3
  234. package/dist/experiment/host/teardown.d.ts +3 -3
  235. package/dist/experiment/host/types.cjs.map +1 -1
  236. package/dist/experiment/host/types.d.cts +19 -19
  237. package/dist/experiment/host/types.d.mts +19 -19
  238. package/dist/experiment/host/types.d.ts +19 -19
  239. package/dist/index.cjs +3 -3
  240. package/dist/index.cjs.map +1 -1
  241. package/dist/index.d.cts +3 -3
  242. package/dist/index.d.mts +3 -3
  243. package/dist/index.d.ts +3 -3
  244. package/dist/index.js +2 -2
  245. package/dist/index.mjs +2 -2
  246. package/dist/inspection/assertions.cjs +1 -1
  247. package/dist/inspection/assertions.cjs.map +1 -1
  248. package/dist/inspection/cli/contribution.cjs +8 -7
  249. package/dist/inspection/cli/contribution.cjs.map +1 -1
  250. package/dist/inspection/facts.cjs +4 -4
  251. package/dist/inspection/facts.cjs.map +1 -1
  252. package/dist/inspection/index.cjs +1 -5
  253. package/dist/inspection/index.cjs.map +1 -1
  254. package/dist/inspection/index.d.cts +1 -2
  255. package/dist/inspection/index.d.mts +1 -2
  256. package/dist/inspection/index.d.ts +1 -2
  257. package/dist/inspection/index.js +20 -26
  258. package/dist/inspection/lifecycle.cjs +7 -0
  259. package/dist/inspection/lifecycle.cjs.map +1 -0
  260. package/dist/inspection/lifecycle.d.cts +4 -0
  261. package/dist/inspection/lifecycle.d.mts +4 -0
  262. package/dist/inspection/lifecycle.d.ts +4 -0
  263. package/dist/inspection/lifecycle.js +6 -0
  264. package/dist/inspection/lifecycle.js.map +1 -0
  265. package/dist/inspection/notes.cjs +40 -0
  266. package/dist/inspection/notes.cjs.map +1 -0
  267. package/dist/inspection/notes.d.cts +12 -0
  268. package/dist/inspection/notes.d.mts +12 -0
  269. package/dist/inspection/notes.d.ts +12 -0
  270. package/dist/inspection/notes.js +8 -0
  271. package/dist/inspection/notes.js.map +1 -0
  272. package/dist/inspection/overview.cjs +12 -15
  273. package/dist/inspection/overview.cjs.map +1 -1
  274. package/dist/inspection/overview.d.cts +8 -6
  275. package/dist/inspection/overview.d.mts +8 -6
  276. package/dist/inspection/overview.d.ts +8 -6
  277. package/dist/inspection/planned-attempt.cjs +45 -0
  278. package/dist/inspection/planned-attempt.cjs.map +1 -0
  279. package/dist/inspection/planned-attempt.d.cts +109 -0
  280. package/dist/inspection/planned-attempt.d.mts +109 -0
  281. package/dist/inspection/planned-attempt.d.ts +109 -0
  282. package/dist/inspection/planned-attempt.js +24 -0
  283. package/dist/inspection/planned-attempt.js.map +1 -0
  284. package/dist/inspection/project-result.cjs +5 -5
  285. package/dist/inspection/project-result.cjs.map +1 -1
  286. package/dist/inspection/project-result.d.cts +58 -57
  287. package/dist/inspection/project-result.d.mts +58 -57
  288. package/dist/inspection/project-result.d.ts +58 -57
  289. package/dist/inspection/project-result.js +4 -4
  290. package/dist/inspection/project.cjs +3 -3
  291. package/dist/inspection/project.cjs.map +1 -1
  292. package/dist/inspection/protocol-values.cjs +1 -1
  293. package/dist/inspection/protocol-values.cjs.map +1 -1
  294. package/dist/inspection/protocol-values.d.cts +1 -1
  295. package/dist/inspection/protocol-values.d.mts +1 -1
  296. package/dist/inspection/protocol-values.d.ts +1 -1
  297. package/dist/inspection/protocol.cjs +8 -8
  298. package/dist/inspection/protocol.cjs.map +1 -1
  299. package/dist/inspection/protocol.d.cts +237 -219
  300. package/dist/inspection/protocol.d.mts +237 -219
  301. package/dist/inspection/protocol.d.ts +237 -219
  302. package/dist/inspection/public.cjs +1 -5
  303. package/dist/inspection/public.cjs.map +1 -1
  304. package/dist/inspection/public.d.cts +1 -2
  305. package/dist/inspection/public.d.mts +1 -2
  306. package/dist/inspection/public.d.ts +1 -2
  307. package/dist/inspection/public.js +14 -20
  308. package/dist/inspection/public.mjs +166 -129
  309. package/dist/inspection/query.cjs +1 -1
  310. package/dist/inspection/query.cjs.map +1 -1
  311. package/dist/inspection/results.cjs +61 -52
  312. package/dist/inspection/results.cjs.map +1 -1
  313. package/dist/inspection/results.d.cts +157 -140
  314. package/dist/inspection/results.d.mts +157 -140
  315. package/dist/inspection/results.d.ts +157 -140
  316. package/dist/inspection/select.cjs +33 -28
  317. package/dist/inspection/select.cjs.map +1 -1
  318. package/dist/inspection/source.cjs +1 -1
  319. package/dist/inspection/source.cjs.map +1 -1
  320. package/dist/inspection/sources.cjs +2 -2
  321. package/dist/inspection/sources.cjs.map +1 -1
  322. package/dist/inspection/trace.cjs +9 -8
  323. package/dist/inspection/trace.cjs.map +1 -1
  324. package/dist/inspection/trace.d.cts +4 -4
  325. package/dist/inspection/trace.d.mts +4 -4
  326. package/dist/inspection/trace.d.ts +4 -4
  327. package/dist/loaders/index.cjs +7 -279
  328. package/dist/loaders/index.cjs.map +1 -1
  329. package/dist/loaders/index.d.cts +2 -95
  330. package/dist/loaders/index.d.mts +2 -95
  331. package/dist/loaders/index.d.ts +2 -95
  332. package/dist/loaders/index.js +10 -12
  333. package/dist/loaders/index.mjs +10 -12
  334. package/dist/loaders/runtime.cjs +283 -0
  335. package/dist/loaders/runtime.cjs.map +1 -0
  336. package/dist/loaders/runtime.d.cts +95 -0
  337. package/dist/loaders/runtime.d.mts +95 -0
  338. package/dist/loaders/runtime.d.ts +95 -0
  339. package/dist/loaders/runtime.js +16 -0
  340. package/dist/loaders/runtime.js.map +1 -0
  341. package/dist/migrations/judge-provider.md +6 -6
  342. package/dist/{model-slots.cjs → model-roles.cjs} +12 -12
  343. package/dist/{model-slots.cjs.map → model-roles.cjs.map} +1 -1
  344. package/dist/{model-slots.d.cts → model-roles.d.cts} +10 -10
  345. package/dist/{model-slots.d.mts → model-roles.d.mts} +10 -10
  346. package/dist/{model-slots.d.ts → model-roles.d.ts} +10 -10
  347. package/dist/model-roles.js +14 -0
  348. package/dist/model-roles.js.map +1 -0
  349. package/dist/o11y/adapter-usage-projection.cjs +9 -8
  350. package/dist/o11y/adapter-usage-projection.cjs.map +1 -1
  351. package/dist/o11y/adapter-usage-projection.d.cts +23 -15
  352. package/dist/o11y/adapter-usage-projection.d.mts +23 -15
  353. package/dist/o11y/adapter-usage-projection.d.ts +23 -15
  354. package/dist/o11y/cache-read.cjs +14 -14
  355. package/dist/o11y/cache-read.cjs.map +1 -1
  356. package/dist/o11y/cache-read.d.cts +5 -5
  357. package/dist/o11y/cache-read.d.mts +5 -5
  358. package/dist/o11y/cache-read.d.ts +5 -5
  359. package/dist/o11y/derive.cjs.map +1 -1
  360. package/dist/o11y/derive.d.cts +2 -2
  361. package/dist/o11y/derive.d.mts +2 -2
  362. package/dist/o11y/derive.d.ts +2 -2
  363. package/dist/o11y/eval-usage.cjs +9 -9
  364. package/dist/o11y/eval-usage.cjs.map +1 -1
  365. package/dist/o11y/eval-usage.d.cts +7 -7
  366. package/dist/o11y/eval-usage.d.mts +7 -7
  367. package/dist/o11y/eval-usage.d.ts +7 -7
  368. package/dist/o11y/parsers/bub.cjs +57 -81
  369. package/dist/o11y/parsers/bub.cjs.map +1 -1
  370. package/dist/o11y/parsers/index.cjs.map +1 -1
  371. package/dist/o11y/parsers/index.d.cts +2 -0
  372. package/dist/o11y/parsers/index.d.mts +2 -0
  373. package/dist/o11y/parsers/index.d.ts +2 -0
  374. package/dist/o11y/parsers/opencode.cjs +129 -217
  375. package/dist/o11y/parsers/opencode.cjs.map +1 -1
  376. package/dist/o11y/parsers/opencode.d.cts +2 -5
  377. package/dist/o11y/parsers/opencode.d.mts +2 -5
  378. package/dist/o11y/parsers/opencode.d.ts +2 -5
  379. package/dist/o11y/types.cjs.map +1 -1
  380. package/dist/o11y/types.d.cts +1 -1
  381. package/dist/o11y/types.d.mts +1 -1
  382. package/dist/o11y/types.d.ts +1 -1
  383. package/dist/plugin/contracts.cjs +170 -120
  384. package/dist/plugin/contracts.cjs.map +1 -1
  385. package/dist/plugin/contracts.d.cts +102 -90
  386. package/dist/plugin/contracts.d.mts +102 -90
  387. package/dist/plugin/contracts.d.ts +102 -90
  388. package/dist/plugin/contracts.js +16 -12
  389. package/dist/plugin/index.cjs.map +1 -1
  390. package/dist/plugin/index.d.cts +1 -1
  391. package/dist/plugin/index.d.mts +1 -1
  392. package/dist/plugin/index.d.ts +1 -1
  393. package/dist/plugin/link.cjs +69 -64
  394. package/dist/plugin/link.cjs.map +1 -1
  395. package/dist/plugin/link.d.cts +5 -3
  396. package/dist/plugin/link.d.mts +5 -3
  397. package/dist/plugin/link.d.ts +5 -3
  398. package/dist/plugin/link.js +12 -8
  399. package/dist/record/codec/seal-manifest.d.cts +3 -3
  400. package/dist/record/codec/seal-manifest.d.mts +3 -3
  401. package/dist/record/codec/seal-manifest.d.ts +3 -3
  402. package/dist/record/family/attempt-note/definition.cjs +12 -0
  403. package/dist/record/family/attempt-note/definition.cjs.map +1 -0
  404. package/dist/record/family/attempt-note/definition.d.cts +6 -0
  405. package/dist/record/family/attempt-note/definition.d.mts +6 -0
  406. package/dist/record/family/attempt-note/definition.d.ts +6 -0
  407. package/dist/record/family/attempt-note/definition.js +6 -0
  408. package/dist/record/family/attempt-note/definition.js.map +1 -0
  409. package/dist/record/family/attempt-note/persistence.cjs +10 -0
  410. package/dist/record/family/attempt-note/persistence.cjs.map +1 -0
  411. package/dist/record/family/attempt-note/persistence.d.cts +6 -0
  412. package/dist/record/family/attempt-note/persistence.d.mts +6 -0
  413. package/dist/record/family/attempt-note/persistence.d.ts +6 -0
  414. package/dist/record/family/attempt-note/persistence.js +6 -0
  415. package/dist/record/family/attempt-note/persistence.js.map +1 -0
  416. package/dist/record/family/attempt-note/schema.cjs +20 -0
  417. package/dist/record/family/attempt-note/schema.cjs.map +1 -0
  418. package/dist/record/family/attempt-note/schema.d.cts +21 -0
  419. package/dist/record/family/attempt-note/schema.d.mts +21 -0
  420. package/dist/record/family/attempt-note/schema.d.ts +21 -0
  421. package/dist/record/family/attempt-note/schema.js +10 -0
  422. package/dist/record/family/attempt-note/schema.js.map +1 -0
  423. package/dist/record/family/catalog.cjs +32 -29
  424. package/dist/record/family/catalog.cjs.map +1 -1
  425. package/dist/record/family/catalog.d.cts +7 -2
  426. package/dist/record/family/catalog.d.mts +7 -2
  427. package/dist/record/family/catalog.d.ts +7 -2
  428. package/dist/record/family/current.cjs +48 -45
  429. package/dist/record/family/current.cjs.map +1 -1
  430. package/dist/record/family/current.d.cts +15 -0
  431. package/dist/record/family/current.d.mts +15 -0
  432. package/dist/record/family/current.d.ts +15 -0
  433. package/dist/record/model/index.js +66 -64
  434. package/dist/record/model/run-context.cjs +9 -3
  435. package/dist/record/model/run-context.cjs.map +1 -1
  436. package/dist/record/model/run-context.d.cts +9 -2
  437. package/dist/record/model/run-context.d.mts +9 -2
  438. package/dist/record/model/run-context.d.ts +9 -2
  439. package/dist/record/model/run-context.js +8 -6
  440. package/dist/record/model/seal-manifest.d.cts +1 -1
  441. package/dist/record/model/seal-manifest.d.mts +1 -1
  442. package/dist/record/model/seal-manifest.d.ts +1 -1
  443. package/dist/run/cli/contribution.cjs +6 -32
  444. package/dist/run/cli/contribution.cjs.map +1 -1
  445. package/dist/run/host/runtime.cjs +1 -1
  446. package/dist/run/host/runtime.cjs.map +1 -1
  447. package/dist/run/storage/sqlite.cjs +7 -7
  448. package/dist/run/storage/sqlite.cjs.map +1 -1
  449. package/dist/runner/adapter-usage.cjs +5 -3
  450. package/dist/runner/adapter-usage.cjs.map +1 -1
  451. package/dist/runner/adapter-usage.d.cts +2 -2
  452. package/dist/runner/adapter-usage.d.mts +2 -2
  453. package/dist/runner/adapter-usage.d.ts +2 -2
  454. package/dist/runner/adoption.cjs +9 -12
  455. package/dist/runner/adoption.cjs.map +1 -1
  456. package/dist/runner/attempt-notes.cjs +8 -0
  457. package/dist/runner/attempt-notes.cjs.map +1 -0
  458. package/dist/runner/attempt-notes.d.cts +4 -0
  459. package/dist/runner/attempt-notes.d.mts +4 -0
  460. package/dist/runner/attempt-notes.d.ts +4 -0
  461. package/dist/runner/attempt-notes.js +8 -0
  462. package/dist/runner/attempt-notes.js.map +1 -0
  463. package/dist/runner/attempt.cjs +61 -73
  464. package/dist/runner/attempt.cjs.map +1 -1
  465. package/dist/runner/attempt.d.cts +3 -3
  466. package/dist/runner/attempt.d.mts +3 -3
  467. package/dist/runner/attempt.d.ts +3 -3
  468. package/dist/runner/command-plan.cjs +21 -26
  469. package/dist/runner/command-plan.cjs.map +1 -1
  470. package/dist/runner/command-plan.d.cts +4 -4
  471. package/dist/runner/command-plan.d.mts +4 -4
  472. package/dist/runner/command-plan.d.ts +4 -4
  473. package/dist/runner/command-plan.js +4 -4
  474. package/dist/runner/config-identity.cjs.map +1 -1
  475. package/dist/runner/config-identity.d.cts +2 -2
  476. package/dist/runner/config-identity.d.mts +2 -2
  477. package/dist/runner/config-identity.d.ts +2 -2
  478. package/dist/runner/discover.cjs +6 -6
  479. package/dist/runner/discover.cjs.map +1 -1
  480. package/dist/runner/eval-selection.cjs +2 -2
  481. package/dist/runner/eval-selection.cjs.map +1 -1
  482. package/dist/runner/eval-selection.d.cts +2 -2
  483. package/dist/runner/eval-selection.d.mts +2 -2
  484. package/dist/runner/eval-selection.d.ts +2 -2
  485. package/dist/runner/feedback/human.cjs +1 -1
  486. package/dist/runner/feedback/human.cjs.map +1 -1
  487. package/dist/runner/feedback/json.cjs +4 -2
  488. package/dist/runner/feedback/json.cjs.map +1 -1
  489. package/dist/runner/feedback/json.d.cts +1 -1
  490. package/dist/runner/feedback/json.d.mts +1 -1
  491. package/dist/runner/feedback/json.d.ts +1 -1
  492. package/dist/runner/feedback/sink.cjs +1 -1
  493. package/dist/runner/feedback/sink.cjs.map +1 -1
  494. package/dist/runner/feedback/sink.d.cts +1 -1
  495. package/dist/runner/feedback/sink.d.mts +1 -1
  496. package/dist/runner/feedback/sink.d.ts +1 -1
  497. package/dist/runner/fingerprint.cjs +1 -1
  498. package/dist/runner/fingerprint.cjs.map +1 -1
  499. package/dist/runner/project-target.cjs.map +1 -1
  500. package/dist/runner/project-target.d.cts +2 -2
  501. package/dist/runner/project-target.d.mts +2 -2
  502. package/dist/runner/project-target.d.ts +2 -2
  503. package/dist/runner/record/attachments.cjs +2 -2
  504. package/dist/runner/record/attachments.cjs.map +1 -1
  505. package/dist/runner/record/planning.cjs +8 -2
  506. package/dist/runner/record/planning.cjs.map +1 -1
  507. package/dist/runner/record.cjs +2 -0
  508. package/dist/runner/record.cjs.map +1 -1
  509. package/dist/runner/report.cjs +3 -3
  510. package/dist/runner/report.cjs.map +1 -1
  511. package/dist/runner/report.d.cts +1 -1
  512. package/dist/runner/report.d.mts +1 -1
  513. package/dist/runner/report.d.ts +1 -1
  514. package/dist/runner/reuse-plan.cjs +2 -2
  515. package/dist/runner/reuse-plan.cjs.map +1 -1
  516. package/dist/runner/run.cjs +10 -45
  517. package/dist/runner/run.cjs.map +1 -1
  518. package/dist/runner/sandbox-pool.cjs +1 -2
  519. package/dist/runner/sandbox-pool.cjs.map +1 -1
  520. package/dist/runner/setup-prefix-preparation.cjs +1 -1
  521. package/dist/runner/setup-prefix-preparation.cjs.map +1 -1
  522. package/dist/runner/teardown-registry.cjs +2 -2
  523. package/dist/runner/teardown-registry.cjs.map +1 -1
  524. package/dist/runner/teardown-registry.d.cts +1 -1
  525. package/dist/runner/teardown-registry.d.mts +1 -1
  526. package/dist/runner/teardown-registry.d.ts +1 -1
  527. package/dist/runner/types.cjs +4 -4
  528. package/dist/runner/types.cjs.map +1 -1
  529. package/dist/runner/types.d.cts +64 -93
  530. package/dist/runner/types.d.mts +64 -93
  531. package/dist/runner/types.d.ts +64 -93
  532. package/dist/runner/types.js +2 -2
  533. package/dist/sandbox/action.cjs +1 -1
  534. package/dist/sandbox/action.cjs.map +1 -1
  535. package/dist/sandbox/cli/contribution.cjs +17 -14
  536. package/dist/sandbox/cli/contribution.cjs.map +1 -1
  537. package/dist/sandbox/cli/contribution.d.cts +1 -1
  538. package/dist/sandbox/cli/contribution.d.mts +1 -1
  539. package/dist/sandbox/cli/contribution.d.ts +1 -1
  540. package/dist/sandbox/commands.cjs +1 -1
  541. package/dist/sandbox/commands.cjs.map +1 -1
  542. package/dist/sandbox/commands.d.cts +1 -1
  543. package/dist/sandbox/commands.d.mts +1 -1
  544. package/dist/sandbox/commands.d.ts +1 -1
  545. package/dist/sandbox/compose.cjs +2 -2
  546. package/dist/sandbox/compose.cjs.map +1 -1
  547. package/dist/sandbox/docker-profile/registry.cjs +5 -9
  548. package/dist/sandbox/docker-profile/registry.cjs.map +1 -1
  549. package/dist/sandbox/docker-profile/registry.d.cts +0 -3
  550. package/dist/sandbox/docker-profile/registry.d.mts +0 -3
  551. package/dist/sandbox/docker-profile/registry.d.ts +0 -3
  552. package/dist/sandbox/docker-profile/registry.js +6 -12
  553. package/dist/sandbox/docker-profile/runtime.cjs +1 -1
  554. package/dist/sandbox/docker-profile/runtime.cjs.map +1 -1
  555. package/dist/sandbox/docker-profile/schema.cjs +12 -16
  556. package/dist/sandbox/docker-profile/schema.cjs.map +1 -1
  557. package/dist/sandbox/docker-profile/schema.d.cts +2 -6
  558. package/dist/sandbox/docker-profile/schema.d.mts +2 -6
  559. package/dist/sandbox/docker-profile/schema.d.ts +2 -6
  560. package/dist/sandbox/docker-profile/schema.js +24 -32
  561. package/dist/sandbox/incus.cjs +2 -2
  562. package/dist/sandbox/incus.cjs.map +1 -1
  563. package/dist/sandbox/index.cjs +4 -35
  564. package/dist/sandbox/index.cjs.map +1 -1
  565. package/dist/sandbox/index.d.cts +4 -8
  566. package/dist/sandbox/index.d.mts +4 -8
  567. package/dist/sandbox/index.d.ts +4 -8
  568. package/dist/sandbox/index.js +116 -172
  569. package/dist/sandbox/index.mjs +116 -172
  570. package/dist/sandbox/layer.cjs +3 -3
  571. package/dist/sandbox/layer.cjs.map +1 -1
  572. package/dist/sandbox/link.cjs +5 -0
  573. package/dist/sandbox/link.cjs.map +1 -1
  574. package/dist/sandbox/link.d.cts +3 -0
  575. package/dist/sandbox/link.d.mts +3 -0
  576. package/dist/sandbox/link.d.ts +3 -0
  577. package/dist/sandbox/runtime.cjs +9 -4
  578. package/dist/sandbox/runtime.cjs.map +1 -1
  579. package/dist/shared/failure-class.cjs +5 -5
  580. package/dist/shared/failure-class.cjs.map +1 -1
  581. package/dist/shared/failure-class.d.cts +4 -10
  582. package/dist/shared/failure-class.d.mts +4 -10
  583. package/dist/shared/failure-class.d.ts +4 -10
  584. package/dist/shared/types.cjs.map +1 -1
  585. package/dist/shared/types.d.cts +1 -1
  586. package/dist/shared/types.d.mts +1 -1
  587. package/dist/shared/types.d.ts +1 -1
  588. package/dist/show/contribution.cjs +17 -2
  589. package/dist/show/contribution.cjs.map +1 -1
  590. package/dist/show/contribution.d.cts +14 -0
  591. package/dist/show/contribution.d.mts +14 -0
  592. package/dist/show/contribution.d.ts +14 -0
  593. package/dist/show/model.cjs +8 -5
  594. package/dist/show/model.cjs.map +1 -1
  595. package/dist/show/model.d.cts +11 -8
  596. package/dist/show/model.d.mts +11 -8
  597. package/dist/show/model.d.ts +11 -8
  598. package/dist/show/render.cjs +34 -22
  599. package/dist/show/render.cjs.map +1 -1
  600. package/dist/sources/projector.cjs +1 -1
  601. package/dist/sources/projector.cjs.map +1 -1
  602. package/dist/types.cjs +1 -1
  603. package/dist/types.cjs.map +1 -1
  604. package/dist/types.d.cts +1 -1
  605. package/dist/types.d.mts +1 -1
  606. package/dist/types.d.ts +1 -1
  607. package/dist/types.js +2 -2
  608. package/dist/view/app-dist/.vite/manifest.json +2 -2
  609. package/dist/view/app-dist/assets/{index-B3ijZRph.js → index-CxrOUvHC.js} +17 -17
  610. package/dist/view/app-dist/assets/{quality-cost-scatter-B--DfJ9s.js → quality-cost-scatter-B6Aiss7x.js} +1 -1
  611. package/dist/view/app-dist/index.html +1 -1
  612. package/dist/view/cli/contribution.cjs +1 -1
  613. package/dist/view/cli/contribution.cjs.map +1 -1
  614. package/dist/view/function-runtime.mjs +318 -121
  615. package/dist/view/http-protocol.d.cts +1 -1
  616. package/dist/view/http-protocol.d.mts +1 -1
  617. package/dist/view/http-protocol.d.ts +1 -1
  618. package/dist/view/inspection-host.cjs +2 -2
  619. package/dist/view/inspection-host.cjs.map +1 -1
  620. package/dist/view/inspection-host.d.cts +349 -331
  621. package/dist/view/inspection-host.d.mts +349 -331
  622. package/dist/view/inspection-host.d.ts +349 -331
  623. package/docs-site/zh/examples/integrations/ai-sdk-v7.mdx +6 -6
  624. package/docs-site/zh/examples/integrations/claude-sdk.mdx +5 -5
  625. package/docs-site/zh/examples/integrations/codex-sdk.mdx +4 -4
  626. package/docs-site/zh/examples/integrations/langgraph.mdx +4 -4
  627. package/docs-site/zh/examples/integrations/pi-sdk.mdx +5 -5
  628. package/docs-site/zh/explanation/adapter.mdx +6 -6
  629. package/docs-site/zh/explanation/assert.mdx +7 -7
  630. package/docs-site/zh/explanation/drive.mdx +4 -4
  631. package/docs-site/zh/explanation/hitl.mdx +2 -2
  632. package/docs-site/zh/explanation/judge.mdx +2 -2
  633. package/docs-site/zh/introduction.mdx +2 -2
  634. package/docs-site/zh/reference/builtin-agents.mdx +46 -49
  635. package/docs-site/zh/reference/cli.mdx +16 -14
  636. package/docs-site/zh/reference/define-agent.mdx +19 -19
  637. package/docs-site/zh/reference/define-config.mdx +7 -15
  638. package/docs-site/zh/reference/define-eval.mdx +20 -14
  639. package/docs-site/zh/reference/define-experiment.mdx +18 -39
  640. package/docs-site/zh/reference/events.mdx +1 -1
  641. package/docs-site/zh/reference/expect.mdx +21 -0
  642. package/docs-site/zh/reference/official-adapters.mdx +12 -5
  643. package/docs-site/zh/troubleshooting/debugging.mdx +2 -2
  644. package/docs-site/zh/troubleshooting/recover-after-kill.mdx +1 -1
  645. package/docs-site/zh/tutorials/agent-onboarding.mdx +2 -2
  646. package/docs-site/zh/tutorials/authoring.mdx +2 -2
  647. package/docs-site/zh/tutorials/compare-runs.mdx +1 -1
  648. package/docs-site/zh/tutorials/concurrency.mdx +8 -8
  649. package/docs-site/zh/tutorials/configuration.mdx +5 -5
  650. package/docs-site/zh/tutorials/control-run-cost.mdx +19 -19
  651. package/docs-site/zh/tutorials/eval-groups.mdx +1 -1
  652. package/docs-site/zh/tutorials/evaluation-kinds.mdx +5 -5
  653. package/docs-site/zh/tutorials/experiment-services.mdx +32 -32
  654. package/docs-site/zh/tutorials/handle-execution-failures.mdx +2 -28
  655. package/docs-site/zh/tutorials/plugins.mdx +49 -143
  656. package/docs-site/zh/tutorials/sandbox-agent.mdx +1 -1
  657. package/docs-site/zh/tutorials/sandbox-reuse.mdx +7 -7
  658. package/docs-site/zh/tutorials/upgrade-0-18.mdx +1 -1
  659. package/docs-site/zh/tutorials/upgrade-0-19.mdx +121 -0
  660. package/docs-site/zh/tutorials/upgrade-0-20.mdx +12 -0
  661. package/docs-site/zh/tutorials/verify-judge.mdx +2 -2
  662. package/docs-site/zh/tutorials/viewing-results.mdx +2 -2
  663. package/docs-site/zh/tutorials/write-experiment.mdx +1 -1
  664. package/docs-site/zh/tutorials/write-send.mdx +13 -13
  665. package/package.json +1 -1
  666. package/dist/agents/effect-runtime.cjs.map +0 -1
  667. package/dist/agents/effect-runtime.d.cts +0 -29
  668. package/dist/agents/effect-runtime.d.mts +0 -29
  669. package/dist/agents/effect-runtime.d.ts +0 -29
  670. package/dist/agents/effect-runtime.js +0 -16
  671. package/dist/agents/effect-runtime.js.map +0 -1
  672. package/dist/agents/session-slot.cjs.map +0 -1
  673. package/dist/agents/session-slot.d.cts +0 -12
  674. package/dist/agents/session-slot.d.mts +0 -12
  675. package/dist/agents/session-slot.d.ts +0 -12
  676. package/dist/agents/session-slot.js +0 -6
  677. package/dist/agents/session-slot.js.map +0 -1
  678. package/dist/model-slots.js +0 -14
  679. package/dist/model-slots.js.map +0 -1
@@ -74,7 +74,7 @@ ai-sdk-v7/
74
74
  <div className="gd-file">
75
75
  <div className="gd-head"><span className="gd-name">{"niceeval.config.ts"}</span><span className="gd-stats"><span className="gd-plus">{"+8"}</span></span></div>
76
76
  <div className="gd-body">
77
- <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineConfig } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { OpenAIProvider } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/judge\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineConfig"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" name: { "}</span><span className="gdt2">{"\"zh-CN\""}</span><span className="gdt0">{": "}</span><span className="gdt2">{"\"AI SDK v7 HTTP 无侵入示例\""}</span><span className="gdt0">{", en: "}</span><span className="gdt2">{"\"AI SDK v7 HTTP non-invasive example\""}</span><span className="gdt0">{" },"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" judgeRuntime: "}</span><span className="gdt5">{"OpenAIProvider"}</span><span className="gdt0">{"({ model: "}</span><span className="gdt2">{"\"gpt-5.4\""}</span><span className="gdt0">{" }),"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" timeoutMs: "}</span><span className="gdt1">{"60_000"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
77
+ <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineConfig } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { OpenAIProvider } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/judge\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineConfig"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" name: { "}</span><span className="gdt2">{"\"zh-CN\""}</span><span className="gdt0">{": "}</span><span className="gdt2">{"\"AI SDK v7 HTTP 无侵入示例\""}</span><span className="gdt0">{", en: "}</span><span className="gdt2">{"\"AI SDK v7 HTTP non-invasive example\""}</span><span className="gdt0">{" },"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" judgeProvider: "}</span><span className="gdt5">{"OpenAIProvider"}</span><span className="gdt0">{"({ model: "}</span><span className="gdt2">{"\"gpt-5.4\""}</span><span className="gdt0">{" }),"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" timeoutMs: "}</span><span className="gdt1">{"60_000"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
78
78
  </div>
79
79
  </div>
80
80
 
@@ -88,7 +88,7 @@ ai-sdk-v7/
88
88
  <div className="gd-file">
89
89
  <div className="gd-head"><span className="gd-name">{"evals/basic-qa.eval.ts"}</span><span className="gd-stats"><span className="gd-plus">{"+25"}</span></span></div>
90
90
  <div className="gd-body">
91
- <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineEval } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineJudge } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/expect\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"answerQuality"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineJudge"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" name: "}</span><span className="gdt2">{"\"answer-quality\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" rubric: "}</span><span className="gdt2">{"\"助手是否用一两句话正常介绍了自己,而不是报错或答非所问?\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 这条 eval 验证 agent 能正常问答且不瞎调工具。断言依据全部来自 UI Message Stream 协议帧"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// (uiMessageStreamAgent 直构);协议帧里没有 usage,所以这里不做用量断言(OTel span 只进瀑布图)。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineEval"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"12"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"测试 agent 能正常问答且不瞎调工具\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"13"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"14"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"async"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"test"}</span><span className="gdt0">{"("}</span><span className="gdt7">{"t"}</span><span className="gdt0">{") {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"15"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"turn"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"send"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"用一句话介绍一下你自己,这轮不用查天气也不用算数。\""}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"16"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" turn."}</span><span className="gdt5">{"succeeded"}</span><span className="gdt0">{"()."}</span><span className="gdt5">{"orStop"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"17"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"18"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"group"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"正常收发、没有多余工具调用\""}</span><span className="gdt0">{", () "}</span><span className="gdt4">{"=>"}</span><span className="gdt0">{" {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"19"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"succeeded"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"20"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"usedNoTools"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"21"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" });"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"22"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"23"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"judge"}</span><span className="gdt0">{"({ prompt: turn.input, answer: turn.message }, answerQuality)."}</span><span className="gdt5">{"gate"}</span><span className="gdt0">{"("}</span><span className="gdt1">{"0.6"}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"24"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" },"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"25"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
91
+ <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineEval } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineJudge } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/expect\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"answerQuality"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineJudge"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" name: "}</span><span className="gdt2">{"\"answer-quality\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" rubric: "}</span><span className="gdt2">{"\"助手是否用一两句话正常介绍了自己,而不是报错或答非所问?\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 这条 eval 验证 agent 能正常问答且不瞎调工具。断言依据全部来自 UI Message Stream 协议帧"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// (uiMessageStreamAgent 直构);协议帧里没有 usage,所以这里不做用量断言(OTel span 只进瀑布图)。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineEval"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"12"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"测试 agent 能正常问答且不瞎调工具\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"13"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"14"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"async"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"test"}</span><span className="gdt0">{"("}</span><span className="gdt7">{"t"}</span><span className="gdt0">{") {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"15"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"turn"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"send"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"用一句话介绍一下你自己,这轮不用查天气也不用算数。\""}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"16"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" turn."}</span><span className="gdt5">{"succeeded"}</span><span className="gdt0">{"()."}</span><span className="gdt5">{"orStop"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"17"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"18"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"group"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"正常收发、没有多余工具调用\""}</span><span className="gdt0">{", () "}</span><span className="gdt4">{"=>"}</span><span className="gdt0">{" {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"19"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"succeeded"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"20"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"usedNoTools"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"21"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" });"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"22"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"23"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"judge"}</span><span className="gdt0">{"({ prompt: turn.input, answer: turn.reply }, answerQuality)."}</span><span className="gdt5">{"gate"}</span><span className="gdt0">{"("}</span><span className="gdt1">{"0.6"}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"24"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" },"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"25"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
92
92
  </div>
93
93
  </div>
94
94
 
@@ -116,27 +116,27 @@ ai-sdk-v7/
116
116
  <div className="gd-file">
117
117
  <div className="gd-head"><span className="gd-name">{"evals/weather-tool.eval.ts"}</span><span className="gd-stats"><span className="gd-plus">{"+26"}</span></span></div>
118
118
  <div className="gd-body">
119
- <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineEval } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineJudge } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/expect\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { jsonMatch, pattern, toolMatch } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/expect\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"answerQuality"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineJudge"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" name: "}</span><span className="gdt2">{"\"answer-quality\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" rubric: "}</span><span className="gdt2">{"\"助手是否给出了具体的天气数据(温度或天气状况),而不是拒绝回答或含糊其辞?\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 这条 eval 验证 agent 遇到实时天气问题时会调 get_weather,而不是直接编一个答案。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 工具断言完全来自 UI Message Stream 协议帧直构(uiMessageStreamAgent 内置),adapter 没有为它写一行帧映射。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"12"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineEval"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"13"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"测试 agent 在天气问题中正确调用 get_weather 并基于结果作答\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"14"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"15"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"async"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"test"}</span><span className="gdt0">{"("}</span><span className="gdt7">{"t"}</span><span className="gdt0">{") {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"16"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"turn"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"send"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"北京今天天气怎么样?\""}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"17"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" turn."}</span><span className="gdt5">{"succeeded"}</span><span className="gdt0">{"()."}</span><span className="gdt5">{"orStop"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"18"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"19"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"group"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"调用 get_weather 且城市正确\""}</span><span className="gdt0">{", () "}</span><span className="gdt4">{"=>"}</span><span className="gdt0">{" {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"20"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"calledTool"}</span><span className="gdt0">{"("}</span><span className="gdt5">{"toolMatch"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"get_weather\""}</span><span className="gdt0">{", { input: "}</span><span className="gdt5">{"jsonMatch"}</span><span className="gdt0">{"({ city: "}</span><span className="gdt2">{"\"北京\""}</span><span className="gdt0">{" }) }));"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"21"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"check"}</span><span className="gdt0">{"(turn.message, "}</span><span className="gdt5">{"pattern"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"/"}</span><span className="gdt8">{"°C"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"气温"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"天气"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"晴"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"多云"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"雨"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"阴"}</span><span className="gdt2">{"/"}</span><span className="gdt0">{"));"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"22"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" });"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"23"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"24"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"judge"}</span><span className="gdt0">{"({ prompt: turn.input, answer: turn.message }, answerQuality)."}</span><span className="gdt5">{"gate"}</span><span className="gdt0">{"("}</span><span className="gdt1">{"0.7"}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"25"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" },"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"26"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
119
+ <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineEval } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineJudge } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/expect\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { jsonMatch, pattern, toolMatch } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval/expect\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"answerQuality"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineJudge"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" name: "}</span><span className="gdt2">{"\"answer-quality\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" rubric: "}</span><span className="gdt2">{"\"助手是否给出了具体的天气数据(温度或天气状况),而不是拒绝回答或含糊其辞?\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 这条 eval 验证 agent 遇到实时天气问题时会调 get_weather,而不是直接编一个答案。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 工具断言完全来自 UI Message Stream 协议帧直构(uiMessageStreamAgent 内置),adapter 没有为它写一行帧映射。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"12"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineEval"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"13"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"测试 agent 在天气问题中正确调用 get_weather 并基于结果作答\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"14"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"15"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"async"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"test"}</span><span className="gdt0">{"("}</span><span className="gdt7">{"t"}</span><span className="gdt0">{") {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"16"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"const"}</span><span className="gdt0">{" "}</span><span className="gdt1">{"turn"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"="}</span><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"send"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"北京今天天气怎么样?\""}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"17"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" turn."}</span><span className="gdt5">{"succeeded"}</span><span className="gdt0">{"()."}</span><span className="gdt5">{"orStop"}</span><span className="gdt0">{"();"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"18"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"19"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" "}</span><span className="gdt4">{"await"}</span><span className="gdt0">{" t."}</span><span className="gdt5">{"group"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"调用 get_weather 且城市正确\""}</span><span className="gdt0">{", () "}</span><span className="gdt4">{"=>"}</span><span className="gdt0">{" {"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"20"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"calledTool"}</span><span className="gdt0">{"("}</span><span className="gdt5">{"toolMatch"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"\"get_weather\""}</span><span className="gdt0">{", { input: "}</span><span className="gdt5">{"jsonMatch"}</span><span className="gdt0">{"({ city: "}</span><span className="gdt2">{"\"北京\""}</span><span className="gdt0">{" }) }));"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"21"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"check"}</span><span className="gdt0">{"(turn.reply, "}</span><span className="gdt5">{"pattern"}</span><span className="gdt0">{"("}</span><span className="gdt2">{"/"}</span><span className="gdt8">{"°C"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"气温"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"天气"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"晴"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"多云"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"雨"}</span><span className="gdt4">{"|"}</span><span className="gdt8">{"阴"}</span><span className="gdt2">{"/"}</span><span className="gdt0">{"));"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"22"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" });"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"23"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"24"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" t."}</span><span className="gdt5">{"judge"}</span><span className="gdt0">{"({ prompt: turn.input, answer: turn.reply }, answerQuality)."}</span><span className="gdt5">{"gate"}</span><span className="gdt0">{"("}</span><span className="gdt1">{"0.7"}</span><span className="gdt0">{");"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"25"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" },"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"26"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
120
120
  </div>
121
121
  </div>
122
122
 
123
123
  <div className="gd-file">
124
124
  <div className="gd-head"><span className="gd-name">{"experiments/assistant.ts"}</span><span className="gd-stats"><span className="gd-plus">{"+9"}</span></span></div>
125
125
  <div className="gd-body">
126
- <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineExperiment } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" agent "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"../agents/ai-sdk-v7.ts\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 单配置基线:不比较模型,用请求体不传 model 时应用自己的默认值。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineExperiment"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"ai-sdk-v7:真实 AI SDK v7 后端(默认模型)\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" agent,"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" attempts: "}</span><span className="gdt1">{"1"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
126
+ <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineExperiment } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" agent "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"../agents/ai-sdk-v7.ts\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 单配置基线:不比较模型,用请求体不传 model 时应用自己的默认值。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineExperiment"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"ai-sdk-v7:真实 AI SDK v7 后端(默认模型)\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" adapter: agent,"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" attempts: "}</span><span className="gdt1">{"1"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
127
127
  </div>
128
128
  </div>
129
129
 
130
130
  <div className="gd-file">
131
131
  <div className="gd-head"><span className="gd-name">{"experiments/compare-models/deepseek-v4-flash.ts"}</span><span className="gd-stats"><span className="gd-plus">{"+12"}</span></span></div>
132
132
  <div className="gd-body">
133
- <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineExperiment } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" agent "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"../../agents/ai-sdk-v7.ts\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// compare-models 组的一格:deepseek-v4-flash。一文件一配置(单 model),model 经 ctx.model"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 走请求体传给应用,同一个 server 实例服务所有 model,不用重启进程。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineExperiment"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"deepseek-v4-flash: 对比模型\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" agent,"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" model: "}</span><span className="gdt2">{"\"deepseek-v4-flash\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" attempts: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{", "}</span><span className="gdt6">{"// 跑满 2 次,才能比较 model 间的通过率"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" budget: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"12"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
133
+ <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineExperiment } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" agent "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"../../agents/ai-sdk-v7.ts\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// compare-models 组的一格:deepseek-v4-flash。一文件一配置(单 model),model 经 ctx.model"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// 走请求体传给应用,同一个 server 实例服务所有 model,不用重启进程。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineExperiment"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"deepseek-v4-flash: 对比模型\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" adapter: agent,"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" model: "}</span><span className="gdt2">{"\"deepseek-v4-flash\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" attempts: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{", "}</span><span className="gdt6">{"// 跑满 2 次,才能比较 model 间的通过率"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" budget: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"12"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
134
134
  </div>
135
135
  </div>
136
136
 
137
137
  <div className="gd-file">
138
138
  <div className="gd-head"><span className="gd-name">{"experiments/compare-models/deepseek-v4-pro.ts"}</span><span className="gd-stats"><span className="gd-plus">{"+11"}</span></span></div>
139
139
  <div className="gd-body">
140
- <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineExperiment } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" agent "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"../../agents/ai-sdk-v7.ts\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// compare-models 组的一格:deepseek-v4-pro。与 deepseek-v4-flash.ts 钉住一切、只差 model。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineExperiment"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"deepseek-v4-pro: 对比模型\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" agent,"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" model: "}</span><span className="gdt2">{"\"deepseek-v4-pro\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" attempts: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{", "}</span><span className="gdt6">{"// 跑满 2 次,才能比较 model 间的通过率"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" budget: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
140
+ <table className="gd-table"><tbody><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"1"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" { defineExperiment } "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"niceeval\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"2"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"import"}</span><span className="gdt0">{" agent "}</span><span className="gdt4">{"from"}</span><span className="gdt0">{" "}</span><span className="gdt2">{"\"../../agents/ai-sdk-v7.ts\""}</span><span className="gdt0">{";"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"3"}</td><td className="gd-sign">{"+"}</td><td className="gd-code">{" "}</td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"4"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt6">{"// compare-models 组的一格:deepseek-v4-pro。与 deepseek-v4-flash.ts 钉住一切、只差 model。"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"5"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt4">{"export"}</span><span className="gdt0">{" "}</span><span className="gdt4">{"default"}</span><span className="gdt0">{" "}</span><span className="gdt5">{"defineExperiment"}</span><span className="gdt0">{"({"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"6"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" description: "}</span><span className="gdt2">{"\"deepseek-v4-pro: 对比模型\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"7"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" adapter: agent,"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"8"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" model: "}</span><span className="gdt2">{"\"deepseek-v4-pro\""}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"9"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" attempts: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{", "}</span><span className="gdt6">{"// 跑满 2 次,才能比较 model 间的通过率"}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"10"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{" budget: "}</span><span className="gdt1">{"2"}</span><span className="gdt0">{","}</span></td></tr><tr className="gd-add"><td className="gd-ln"></td><td className="gd-ln">{"11"}</td><td className="gd-sign">{"+"}</td><td className="gd-code"><span className="gdt0">{"});"}</span></td></tr></tbody></table>
141
141
  </div>
142
142
  </div>