agentinputlab 0.1.0a1__py3-none-any.whl

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 (469) hide show
  1. agentinputlab/__init__.py +28 -0
  2. agentinputlab/algorithms/__init__.py +6 -0
  3. agentinputlab/algorithms/acquire/__init__.py +145 -0
  4. agentinputlab/algorithms/acquire/adaptive_rag/__init__.py +43 -0
  5. agentinputlab/algorithms/acquire/adaptive_rag/replay.py +367 -0
  6. agentinputlab/algorithms/acquire/adaptive_rag/runtime.py +1202 -0
  7. agentinputlab/algorithms/acquire/adaptive_rag_common_support.py +542 -0
  8. agentinputlab/algorithms/acquire/hyde.py +296 -0
  9. agentinputlab/algorithms/acquire/hyde_controlled.py +200 -0
  10. agentinputlab/algorithms/acquire/hyde_external.py +148 -0
  11. agentinputlab/algorithms/acquire/ircot.py +28 -0
  12. agentinputlab/algorithms/acquire/ircot_common_support.py +478 -0
  13. agentinputlab/algorithms/acquire/ircot_native.py +355 -0
  14. agentinputlab/algorithms/acquire/ircot_replay.py +327 -0
  15. agentinputlab/algorithms/acquire/qrag/__init__.py +79 -0
  16. agentinputlab/algorithms/acquire/qrag/bridge.py +828 -0
  17. agentinputlab/algorithms/acquire/qrag/common_support.py +891 -0
  18. agentinputlab/algorithms/acquire/qrag/native_axis.py +71 -0
  19. agentinputlab/algorithms/acquire/qrag/reader_adapter.py +139 -0
  20. agentinputlab/algorithms/acquire/r2ag/__init__.py +57 -0
  21. agentinputlab/algorithms/acquire/r2ag/bridge.py +1312 -0
  22. agentinputlab/algorithms/acquire/r2ag/common_support.py +1029 -0
  23. agentinputlab/algorithms/acquire/r2ag/socket_runtime.py +255 -0
  24. agentinputlab/algorithms/clarify/__init__.py +61 -0
  25. agentinputlab/algorithms/clarify/_structured.py +64 -0
  26. agentinputlab/algorithms/clarify/active_task_disambiguation.py +371 -0
  27. agentinputlab/algorithms/clarify/ask_or_assume/__init__.py +17 -0
  28. agentinputlab/algorithms/clarify/ask_or_assume/monitor.py +410 -0
  29. agentinputlab/algorithms/clarify/future_turns.py +379 -0
  30. agentinputlab/algorithms/clarify/udcs.py +255 -0
  31. agentinputlab/algorithms/clarify/udcs_deterministic.py +196 -0
  32. agentinputlab/algorithms/curate/__init__.py +65 -0
  33. agentinputlab/algorithms/curate/acon/__init__.py +27 -0
  34. agentinputlab/algorithms/curate/acon/common.py +284 -0
  35. agentinputlab/algorithms/curate/acon/common_support.py +136 -0
  36. agentinputlab/algorithms/curate/acon/history_bridge.py +309 -0
  37. agentinputlab/algorithms/curate/acon/observation_bridge.py +290 -0
  38. agentinputlab/algorithms/curate/llmlingua/__init__.py +30 -0
  39. agentinputlab/algorithms/curate/llmlingua/adapters.py +537 -0
  40. agentinputlab/algorithms/curate/llmlingua/budget.py +48 -0
  41. agentinputlab/algorithms/curate/llmlingua/controlled.py +175 -0
  42. agentinputlab/algorithms/curate/llmlingua/multilingual.py +105 -0
  43. agentinputlab/algorithms/curate/llmlingua/span_map.py +167 -0
  44. agentinputlab/algorithms/curate/recomp/__init__.py +47 -0
  45. agentinputlab/algorithms/curate/recomp/external.py +1129 -0
  46. agentinputlab/algorithms/curate/recomp/selective.py +304 -0
  47. agentinputlab/algorithms/decompose/__init__.py +107 -0
  48. agentinputlab/algorithms/decompose/decomp/__init__.py +53 -0
  49. agentinputlab/algorithms/decompose/decomp/common_support.py +478 -0
  50. agentinputlab/algorithms/decompose/decomp/planner_component.py +199 -0
  51. agentinputlab/algorithms/decompose/decomp/worker_bridge.py +416 -0
  52. agentinputlab/algorithms/decompose/llm_compiler/__init__.py +35 -0
  53. agentinputlab/algorithms/decompose/llm_compiler/bridge.py +520 -0
  54. agentinputlab/algorithms/decompose/llm_compiler/common_support.py +504 -0
  55. agentinputlab/algorithms/decompose/plan_and_budget/__init__.py +43 -0
  56. agentinputlab/algorithms/decompose/plan_and_budget/cached_soft_budget.py +362 -0
  57. agentinputlab/algorithms/decompose/plan_and_budget/decomposer_component.py +107 -0
  58. agentinputlab/algorithms/decompose/plan_and_budget/full.py +205 -0
  59. agentinputlab/algorithms/decompose/plan_and_budget/hard_cap.py +246 -0
  60. agentinputlab/algorithms/decompose/plan_and_budget/regenerated_soft_budget.py +448 -0
  61. agentinputlab/algorithms/decompose/plan_and_solve/__init__.py +33 -0
  62. agentinputlab/algorithms/decompose/plan_and_solve/plan_component.py +192 -0
  63. agentinputlab/algorithms/decompose/plan_and_solve/ps_prompt.py +127 -0
  64. agentinputlab/algorithms/decompose/self_ask/__init__.py +27 -0
  65. agentinputlab/algorithms/decompose/self_ask/loop.py +362 -0
  66. agentinputlab/algorithms/decompose/self_ask/planner_component.py +117 -0
  67. agentinputlab/algorithms/reformulate/__init__.py +85 -0
  68. agentinputlab/algorithms/reformulate/infocqr/__init__.py +40 -0
  69. agentinputlab/algorithms/reformulate/infocqr/cache_replay.py +362 -0
  70. agentinputlab/algorithms/reformulate/infocqr/external.py +675 -0
  71. agentinputlab/algorithms/reformulate/infocqr/student.py +189 -0
  72. agentinputlab/algorithms/reformulate/query2doc/__init__.py +42 -0
  73. agentinputlab/algorithms/reformulate/query2doc/controlled_live.py +190 -0
  74. agentinputlab/algorithms/reformulate/query2doc/models.py +95 -0
  75. agentinputlab/algorithms/reformulate/query2doc/official_cached.py +335 -0
  76. agentinputlab/algorithms/reformulate/query2doc/querygym_bridge.py +223 -0
  77. agentinputlab/algorithms/reformulate/query2doc/querygym_runtime.py +326 -0
  78. agentinputlab/algorithms/reformulate/reformer/__init__.py +23 -0
  79. agentinputlab/algorithms/reformulate/reformer/external.py +417 -0
  80. agentinputlab/algorithms/reformulate/rrr.py +696 -0
  81. agentinputlab/annotation/__init__.py +5 -0
  82. agentinputlab/annotation/independent.py +428 -0
  83. agentinputlab/annotation/server.py +233 -0
  84. agentinputlab/annotation/static/app.js +121 -0
  85. agentinputlab/annotation/static/index.html +60 -0
  86. agentinputlab/annotation/static/review.css +2 -0
  87. agentinputlab/annotation/static/review.html +10 -0
  88. agentinputlab/annotation/static/review.js +11 -0
  89. agentinputlab/annotation/static/styles.css +55 -0
  90. agentinputlab/annotation/store.py +444 -0
  91. agentinputlab/api.py +368 -0
  92. agentinputlab/artifacts/__init__.py +15 -0
  93. agentinputlab/artifacts/json_file.py +104 -0
  94. agentinputlab/artifacts/manifest.py +98 -0
  95. agentinputlab/artifacts/store.py +157 -0
  96. agentinputlab/backends/__init__.py +98 -0
  97. agentinputlab/backends/embedding/__init__.py +25 -0
  98. agentinputlab/backends/embedding/base.py +145 -0
  99. agentinputlab/backends/embedding/callable.py +75 -0
  100. agentinputlab/backends/embedding/replay.py +129 -0
  101. agentinputlab/backends/llm/__init__.py +31 -0
  102. agentinputlab/backends/llm/base.py +149 -0
  103. agentinputlab/backends/llm/callable.py +108 -0
  104. agentinputlab/backends/llm/decomp_completion.py +88 -0
  105. agentinputlab/backends/llm/final_content_stop.py +72 -0
  106. agentinputlab/backends/llm/gpt_oss_final_channel.py +84 -0
  107. agentinputlab/backends/llm/openai_compatible.py +332 -0
  108. agentinputlab/backends/llm/transformers.py +171 -0
  109. agentinputlab/backends/retrieval/__init__.py +35 -0
  110. agentinputlab/backends/retrieval/base.py +50 -0
  111. agentinputlab/backends/retrieval/dense.py +777 -0
  112. agentinputlab/backends/retrieval/memory.py +119 -0
  113. agentinputlab/backends/retrieval/sqlite_fts.py +155 -0
  114. agentinputlab/backends/user/__init__.py +15 -0
  115. agentinputlab/backends/user/base.py +66 -0
  116. agentinputlab/backends/user/callable.py +54 -0
  117. agentinputlab/backends/user/replay.py +89 -0
  118. agentinputlab/backends/vector.py +38 -0
  119. agentinputlab/benchmarks/__init__.py +246 -0
  120. agentinputlab/benchmarks/acquisition.py +106 -0
  121. agentinputlab/benchmarks/ale.py +540 -0
  122. agentinputlab/benchmarks/appworld_ul.py +669 -0
  123. agentinputlab/benchmarks/assets.py +226 -0
  124. agentinputlab/benchmarks/base.py +68 -0
  125. agentinputlab/benchmarks/beir_archive.py +235 -0
  126. agentinputlab/benchmarks/beir_sample.py +269 -0
  127. agentinputlab/benchmarks/bfcl_native.py +245 -0
  128. agentinputlab/benchmarks/bfcl_v3.py +607 -0
  129. agentinputlab/benchmarks/cache.py +218 -0
  130. agentinputlab/benchmarks/clariq.py +203 -0
  131. agentinputlab/benchmarks/data/registry.json +321 -0
  132. agentinputlab/benchmarks/data/smoke_assets.json +511 -0
  133. agentinputlab/benchmarks/data/statebench_v0_8_1_task_ids.json +475 -0
  134. agentinputlab/benchmarks/data/toolsandbox_165848b_scenario_names.json +1 -0
  135. agentinputlab/benchmarks/execution.py +1111 -0
  136. agentinputlab/benchmarks/interactive.py +257 -0
  137. agentinputlab/benchmarks/ir_query.py +204 -0
  138. agentinputlab/benchmarks/longbench.py +203 -0
  139. agentinputlab/benchmarks/paired_replay.py +479 -0
  140. agentinputlab/benchmarks/preprocess.py +2313 -0
  141. agentinputlab/benchmarks/production.py +236 -0
  142. agentinputlab/benchmarks/qa.py +963 -0
  143. agentinputlab/benchmarks/registry.py +110 -0
  144. agentinputlab/benchmarks/retrieval.py +459 -0
  145. agentinputlab/benchmarks/runtime.py +509 -0
  146. agentinputlab/benchmarks/statebench.py +938 -0
  147. agentinputlab/benchmarks/tau3.py +725 -0
  148. agentinputlab/benchmarks/toolsandbox.py +626 -0
  149. agentinputlab/benchmarks/trec_dl_sample.py +263 -0
  150. agentinputlab/campaign.py +654 -0
  151. agentinputlab/campaign_runner.py +953 -0
  152. agentinputlab/cli/__init__.py +6 -0
  153. agentinputlab/cli/main.py +933 -0
  154. agentinputlab/connectors/__init__.py +46 -0
  155. agentinputlab/connectors/knowledge/__init__.py +4 -0
  156. agentinputlab/connectors/tools/__init__.py +30 -0
  157. agentinputlab/connectors/tools/base.py +57 -0
  158. agentinputlab/connectors/tools/mcp.py +222 -0
  159. agentinputlab/connectors/tools/memory.py +41 -0
  160. agentinputlab/connectors/tools/openapi.py +695 -0
  161. agentinputlab/connectors/tools/schema.py +102 -0
  162. agentinputlab/contracts/__init__.py +91 -0
  163. agentinputlab/contracts/availability.py +24 -0
  164. agentinputlab/contracts/dependency.py +58 -0
  165. agentinputlab/contracts/enums.py +82 -0
  166. agentinputlab/contracts/errors.py +63 -0
  167. agentinputlab/contracts/evidence.py +43 -0
  168. agentinputlab/contracts/intent.py +110 -0
  169. agentinputlab/contracts/lifecycle.py +164 -0
  170. agentinputlab/contracts/patch.py +167 -0
  171. agentinputlab/contracts/serialization.py +93 -0
  172. agentinputlab/contracts/state.py +127 -0
  173. agentinputlab/contracts/trace.py +59 -0
  174. agentinputlab/contracts/worker.py +42 -0
  175. agentinputlab/core/__init__.py +7 -0
  176. agentinputlab/data/__init__.py +32 -0
  177. agentinputlab/data/checksums.py +9 -0
  178. agentinputlab/data/downloader.py +235 -0
  179. agentinputlab/data/http.py +171 -0
  180. agentinputlab/data/index.py +21 -0
  181. agentinputlab/data/overlap.py +28 -0
  182. agentinputlab/data/preprocess.py +32 -0
  183. agentinputlab/data/storage.py +57 -0
  184. agentinputlab/data/trec_sample.py +427 -0
  185. agentinputlab/evaluation/__init__.py +5 -0
  186. agentinputlab/evaluation/outcomes.py +32 -0
  187. agentinputlab/experiments/__init__.py +278 -0
  188. agentinputlab/experiments/analysis.py +981 -0
  189. agentinputlab/experiments/artifacts.py +687 -0
  190. agentinputlab/experiments/automatic_measurement.py +678 -0
  191. agentinputlab/experiments/auxiliary_matrix.py +203 -0
  192. agentinputlab/experiments/bindings.py +430 -0
  193. agentinputlab/experiments/blueprints.py +147 -0
  194. agentinputlab/experiments/builders/__init__.py +45 -0
  195. agentinputlab/experiments/builders/executable_external_payload.py +423 -0
  196. agentinputlab/experiments/builders/external_process_arena.py +1018 -0
  197. agentinputlab/experiments/builders/registry_intervention.py +463 -0
  198. agentinputlab/experiments/cli.py +158 -0
  199. agentinputlab/experiments/controls.py +340 -0
  200. agentinputlab/experiments/driver_guard.py +205 -0
  201. agentinputlab/experiments/drivers/__init__.py +89 -0
  202. agentinputlab/experiments/drivers/external_arena.py +2908 -0
  203. agentinputlab/experiments/drivers/method_process.py +662 -0
  204. agentinputlab/experiments/drivers/parity_process.py +1484 -0
  205. agentinputlab/experiments/drivers/static_production.py +1574 -0
  206. agentinputlab/experiments/executor.py +448 -0
  207. agentinputlab/experiments/loader.py +605 -0
  208. agentinputlab/experiments/measurement.py +675 -0
  209. agentinputlab/experiments/paper_analysis.py +1818 -0
  210. agentinputlab/experiments/parity_identity.py +605 -0
  211. agentinputlab/experiments/plan.py +277 -0
  212. agentinputlab/experiments/production_builders.py +3739 -0
  213. agentinputlab/experiments/readiness.py +432 -0
  214. agentinputlab/experiments/schema.py +514 -0
  215. agentinputlab/experiments/secrets.py +177 -0
  216. agentinputlab/experiments/stateful_analysis.py +1738 -0
  217. agentinputlab/experiments/transfer_analysis.py +1316 -0
  218. agentinputlab/experiments/unseen_analysis.py +1025 -0
  219. agentinputlab/integrations/__init__.py +1 -0
  220. agentinputlab/integrations/action_space/__init__.py +1 -0
  221. agentinputlab/integrations/external_workers/__init__.py +76 -0
  222. agentinputlab/integrations/external_workers/ace.py +431 -0
  223. agentinputlab/integrations/external_workers/base.py +181 -0
  224. agentinputlab/integrations/external_workers/decomp_live.py +249 -0
  225. agentinputlab/integrations/external_workers/decomp_planner.py +30 -0
  226. agentinputlab/integrations/external_workers/infocqr_student.py +280 -0
  227. agentinputlab/integrations/external_workers/ircot.py +576 -0
  228. agentinputlab/integrations/external_workers/llmcompiler_live.py +360 -0
  229. agentinputlab/integrations/external_workers/longllmlingua.py +162 -0
  230. agentinputlab/integrations/external_workers/method.py +314 -0
  231. agentinputlab/integrations/external_workers/process.py +678 -0
  232. agentinputlab/integrations/external_workers/supervisor.py +1071 -0
  233. agentinputlab/integrations/full_architecture/__init__.py +1 -0
  234. agentinputlab/integrations/full_architecture/hyde.py +212 -0
  235. agentinputlab/integrations/full_architecture/plan_and_budget.py +188 -0
  236. agentinputlab/integrations/live_history.py +222 -0
  237. agentinputlab/integrations/prompt_optimization/__init__.py +1 -0
  238. agentinputlab/integrations/training/__init__.py +1 -0
  239. agentinputlab/metrics/__init__.py +50 -0
  240. agentinputlab/metrics/catalog.py +313 -0
  241. agentinputlab/metrics/harm.py +95 -0
  242. agentinputlab/metrics/paired.py +40 -0
  243. agentinputlab/metrics/qa.py +76 -0
  244. agentinputlab/metrics/text.py +33 -0
  245. agentinputlab/metrics/third_party.py +161 -0
  246. agentinputlab/models/__init__.py +4 -0
  247. agentinputlab/observability/__init__.py +56 -0
  248. agentinputlab/observability/budget.py +316 -0
  249. agentinputlab/observability/instrumentation.py +754 -0
  250. agentinputlab/observability/models.py +403 -0
  251. agentinputlab/observability/monitor.py +380 -0
  252. agentinputlab/optimizers/__init__.py +1 -0
  253. agentinputlab/optimizers/clarify/__init__.py +12 -0
  254. agentinputlab/optimizers/context/__init__.py +22 -0
  255. agentinputlab/optimizers/decompose/__init__.py +64 -0
  256. agentinputlab/optimizers/evidence/__init__.py +22 -0
  257. agentinputlab/optimizers/rewrite/__init__.py +30 -0
  258. agentinputlab/parity/__init__.py +5 -0
  259. agentinputlab/parity/base.py +206 -0
  260. agentinputlab/production_campaign.py +763 -0
  261. agentinputlab/py.typed +1 -0
  262. agentinputlab/recipes/__init__.py +333 -0
  263. agentinputlab/recipes/ace.py +418 -0
  264. agentinputlab/recipes/ace_bfcl.py +273 -0
  265. agentinputlab/recipes/ace_common_support.py +169 -0
  266. agentinputlab/recipes/ace_native_plan.py +322 -0
  267. agentinputlab/recipes/ace_online.py +711 -0
  268. agentinputlab/recipes/ace_retrieval.py +330 -0
  269. agentinputlab/recipes/ace_statebench.py +382 -0
  270. agentinputlab/recipes/ace_tau.py +463 -0
  271. agentinputlab/recipes/ace_toolsandbox.py +432 -0
  272. agentinputlab/recipes/acon_guideline.py +22 -0
  273. agentinputlab/recipes/adaptive_rag.py +822 -0
  274. agentinputlab/recipes/llmlingua2_inference.py +146 -0
  275. agentinputlab/recipes/llmlingua2_training.py +796 -0
  276. agentinputlab/recipes/mce.py +1486 -0
  277. agentinputlab/recipes/mce_final.py +335 -0
  278. agentinputlab/recipes/mce_text_fit.py +499 -0
  279. agentinputlab/recipes/mce_text_only.py +298 -0
  280. agentinputlab/recipes/qrag.py +729 -0
  281. agentinputlab/recipes/qrag_inference.py +364 -0
  282. agentinputlab/recipes/r2ag_features.py +236 -0
  283. agentinputlab/recipes/r2ag_inference.py +503 -0
  284. agentinputlab/recipes/r2ag_training.py +948 -0
  285. agentinputlab/recipes/recomp_abstractive.py +592 -0
  286. agentinputlab/recipes/recomp_extractive.py +525 -0
  287. agentinputlab/recipes/reformer.py +17 -0
  288. agentinputlab/recipes/rqrag.py +1516 -0
  289. agentinputlab/recipes/rqrag_common_support.py +193 -0
  290. agentinputlab/recipes/rqrag_completion.py +147 -0
  291. agentinputlab/recipes/skillopt.py +462 -0
  292. agentinputlab/recipes/skillopt_common_support.py +154 -0
  293. agentinputlab/recipes/skillopt_environment.py +165 -0
  294. agentinputlab/recipes/skillopt_environment_cli.py +5 -0
  295. agentinputlab/recipes/skillopt_retrieval.py +188 -0
  296. agentinputlab/recipes/smartsearch.py +897 -0
  297. agentinputlab/recipes/smartsearch_curriculum.py +283 -0
  298. agentinputlab/recipes/smartsearch_grpo_contracts.py +58 -0
  299. agentinputlab/recipes/smartsearch_grpo_data.py +314 -0
  300. agentinputlab/recipes/smartsearch_grpo_meter.py +324 -0
  301. agentinputlab/recipes/smartsearch_grpo_prefix.py +144 -0
  302. agentinputlab/recipes/smartsearch_grpo_rollout.py +119 -0
  303. agentinputlab/recipes/smartsearch_grpo_tools.py +364 -0
  304. agentinputlab/recipes/smartsearch_grpo_training.py +305 -0
  305. agentinputlab/recipes/smartsearch_grpo_transport.py +248 -0
  306. agentinputlab/recipes/smartsearch_grpo_worker.py +364 -0
  307. agentinputlab/recipes/smartsearch_inference.py +524 -0
  308. agentinputlab/recipes/smartsearch_native_data.py +172 -0
  309. agentinputlab/recipes/smartsearch_policy.py +616 -0
  310. agentinputlab/recipes/smartsearch_policy_worker.py +9 -0
  311. agentinputlab/recipes/smartsearch_sampling.py +232 -0
  312. agentinputlab/recipes/smartsearch_teacher.py +203 -0
  313. agentinputlab/recipes/smartsearch_training.py +265 -0
  314. agentinputlab/registry/__init__.py +46 -0
  315. agentinputlab/registry/beta.py +727 -0
  316. agentinputlab/registry/data/algorithms.json +1959 -0
  317. agentinputlab/registry/data/auxiliary_method_profiles.json +34 -0
  318. agentinputlab/registry/loader.py +100 -0
  319. agentinputlab/registry/result_manifest.py +184 -0
  320. agentinputlab/registry/runtime.py +195 -0
  321. agentinputlab/registry/schema.py +449 -0
  322. agentinputlab/runner/__init__.py +30 -0
  323. agentinputlab/runner/balance.py +248 -0
  324. agentinputlab/runner/cache.py +135 -0
  325. agentinputlab/runner/composite.py +130 -0
  326. agentinputlab/runner/engine.py +285 -0
  327. agentinputlab/runner/runtime_guard.py +268 -0
  328. agentinputlab/runner/worker.py +80 -0
  329. agentinputlab/runtime_bootstrap.py +288 -0
  330. agentinputlab/selectors/__init__.py +196 -0
  331. agentinputlab/selectors/advanced.py +1514 -0
  332. agentinputlab/selectors/artifacts.py +1087 -0
  333. agentinputlab/selectors/base.py +112 -0
  334. agentinputlab/selectors/campaigns.py +1893 -0
  335. agentinputlab/selectors/catalog.py +86 -0
  336. agentinputlab/selectors/discovery.py +1860 -0
  337. agentinputlab/selectors/evaluation.py +1348 -0
  338. agentinputlab/selectors/features.py +383 -0
  339. agentinputlab/selectors/learned.py +502 -0
  340. agentinputlab/selectors/no_op.py +28 -0
  341. agentinputlab/selectors/outcome_table.py +475 -0
  342. agentinputlab/selectors/producer.py +853 -0
  343. agentinputlab/selectors/rules.py +109 -0
  344. agentinputlab/selectors/runtime.py +420 -0
  345. agentinputlab/studies/__init__.py +16 -0
  346. agentinputlab/studies/design.py +165 -0
  347. agentinputlab/studies/paired.py +322 -0
  348. agentinputlab/studio/__init__.py +6 -0
  349. agentinputlab/studio/campaigns.py +819 -0
  350. agentinputlab/studio/engine.py +1408 -0
  351. agentinputlab/studio/model_backends.py +425 -0
  352. agentinputlab/studio/prepared.py +410 -0
  353. agentinputlab/studio/server.py +613 -0
  354. agentinputlab/studio/static/app.js +397 -0
  355. agentinputlab/studio/static/campaigns.html +175 -0
  356. agentinputlab/studio/static/campaigns.js +426 -0
  357. agentinputlab/studio/static/index.html +192 -0
  358. agentinputlab/studio/static/result.html +103 -0
  359. agentinputlab/studio/static/result.js +273 -0
  360. agentinputlab/studio/static/styles.css +467 -0
  361. agentinputlab-0.1.0a1.data/data/share/agentinputlab/CITATION.cff +22 -0
  362. agentinputlab-0.1.0a1.dist-info/METADATA +392 -0
  363. agentinputlab-0.1.0a1.dist-info/RECORD +469 -0
  364. agentinputlab-0.1.0a1.dist-info/WHEEL +5 -0
  365. agentinputlab-0.1.0a1.dist-info/entry_points.txt +6 -0
  366. agentinputlab-0.1.0a1.dist-info/licenses/LICENSE +21 -0
  367. agentinputlab-0.1.0a1.dist-info/licenses/NOTICE +31 -0
  368. agentinputlab-0.1.0a1.dist-info/licenses/THIRD_PARTY_NOTICES.md +94 -0
  369. agentinputlab-0.1.0a1.dist-info/top_level.txt +2 -0
  370. workers/README.md +94 -0
  371. workers/ace/README.md +261 -0
  372. workers/ace/__init__.py +1 -0
  373. workers/ace/compatibility_manifest.json +39 -0
  374. workers/ace/driver.py +1682 -0
  375. workers/acon/README.md +215 -0
  376. workers/acon/__init__.py +5 -0
  377. workers/acon/compatibility_manifest.json +93 -0
  378. workers/acon/driver.py +1859 -0
  379. workers/acon/fitted.py +395 -0
  380. workers/acon/guideline.py +470 -0
  381. workers/acon/guideline_source_manifest.json +20 -0
  382. workers/adaptive_rag/README.md +96 -0
  383. workers/adaptive_rag/__init__.py +15 -0
  384. workers/adaptive_rag/compatibility_manifest.json +61 -0
  385. workers/adaptive_rag/driver.py +1470 -0
  386. workers/adaptive_rag/fixture_runtime.py +163 -0
  387. workers/adaptive_rag/runtime.py +252 -0
  388. workers/ask_or_assume/README.md +124 -0
  389. workers/ask_or_assume/__init__.py +1 -0
  390. workers/ask_or_assume/compatibility_manifest.json +36 -0
  391. workers/ask_or_assume/driver.py +1227 -0
  392. workers/decomp/README.md +133 -0
  393. workers/decomp/UPSTREAM_LICENSE.Apache-2.0.txt +201 -0
  394. workers/decomp/__init__.py +5 -0
  395. workers/decomp/compatibility_manifest.json +205 -0
  396. workers/decomp/driver.py +1015 -0
  397. workers/decomp/live.py +137 -0
  398. workers/decomp/runtime.py +231 -0
  399. workers/infocqr/README.md +106 -0
  400. workers/infocqr/__init__.py +1 -0
  401. workers/infocqr/compatibility_manifest.json +84 -0
  402. workers/infocqr/driver.py +1766 -0
  403. workers/ircot/README.md +167 -0
  404. workers/ircot/__init__.py +7 -0
  405. workers/ircot/compatibility_manifest.json +121 -0
  406. workers/ircot/controlled.py +457 -0
  407. workers/ircot/driver.py +1373 -0
  408. workers/ircot/fixture_bindings.json +39 -0
  409. workers/ircot/fixture_config.json +48 -0
  410. workers/ircot/fixture_prompt.txt +2 -0
  411. workers/ircot/native_qa.py +398 -0
  412. workers/ircot/qa_compatibility_manifest.json +57 -0
  413. workers/ircot/runtime.py +97 -0
  414. workers/llmcompiler/README.md +160 -0
  415. workers/llmcompiler/__init__.py +1 -0
  416. workers/llmcompiler/compatibility_manifest.json +45 -0
  417. workers/llmcompiler/driver.py +1514 -0
  418. workers/llmcompiler/export_prompts.py +74 -0
  419. workers/llmcompiler/live.py +174 -0
  420. workers/llmcompiler/patch_manifest.json +147 -0
  421. workers/mce/README.md +198 -0
  422. workers/mce/__init__.py +29 -0
  423. workers/mce/compatibility_manifest.json +115 -0
  424. workers/mce/driver.py +2585 -0
  425. workers/mce/final_eval.py +173 -0
  426. workers/official_arenas/README.md +25 -0
  427. workers/official_arenas/__init__.py +1 -0
  428. workers/official_arenas/capture_records.py +110 -0
  429. workers/official_arenas/driver.py +619 -0
  430. workers/official_arenas/protocol.py +334 -0
  431. workers/official_arenas/providers/__init__.py +1 -0
  432. workers/official_arenas/providers/_live_history.py +228 -0
  433. workers/official_arenas/providers/_message_compat.py +77 -0
  434. workers/official_arenas/providers/appworld_ul.py +196 -0
  435. workers/official_arenas/providers/fixture.py +112 -0
  436. workers/official_arenas/providers/statebench.py +960 -0
  437. workers/official_arenas/providers/tau3.py +1672 -0
  438. workers/official_arenas/providers/tau_retail.py +586 -0
  439. workers/official_arenas/providers/toolsandbox.py +1321 -0
  440. workers/official_arenas/providers/tooltalk.py +699 -0
  441. workers/official_arenas/retail_capture.py +188 -0
  442. workers/official_arenas/retail_driver.py +275 -0
  443. workers/qrag/README.md +172 -0
  444. workers/qrag/__init__.py +1 -0
  445. workers/qrag/compatibility_manifest.json +64 -0
  446. workers/qrag/driver.py +673 -0
  447. workers/r2ag/README.md +154 -0
  448. workers/r2ag/__init__.py +5 -0
  449. workers/r2ag/compatibility_manifest.json +55 -0
  450. workers/r2ag/driver.py +1287 -0
  451. workers/r2ag/feature_schema.json +8 -0
  452. workers/r2ag/runtime.py +589 -0
  453. workers/recomp/README.md +120 -0
  454. workers/recomp/__init__.py +1 -0
  455. workers/recomp/compatibility_manifest.json +43 -0
  456. workers/recomp/driver.py +1653 -0
  457. workers/reformer/README.md +125 -0
  458. workers/reformer/__init__.py +1 -0
  459. workers/reformer/compatibility_manifest.json +68 -0
  460. workers/reformer/driver.py +1459 -0
  461. workers/reformer/induction.py +457 -0
  462. workers/rrr/README.md +126 -0
  463. workers/rrr/__init__.py +1 -0
  464. workers/rrr/compatibility_manifest.json +57 -0
  465. workers/rrr/driver.py +1133 -0
  466. workers/smartsearch/README.md +331 -0
  467. workers/smartsearch/__init__.py +1 -0
  468. workers/smartsearch/compatibility_manifest.json +41 -0
  469. workers/smartsearch/driver.py +1643 -0
@@ -0,0 +1,410 @@
1
+ """External component bridge for Ask or Assume?'s intent monitor."""
2
+
3
+ from __future__ import annotations
4
+
5
+ import math
6
+ from collections.abc import Mapping
7
+ from dataclasses import asdict, dataclass
8
+
9
+ from agentinputlab.contracts import (
10
+ AlgorithmOutputError,
11
+ ExecutionTrace,
12
+ FrozenAgentState,
13
+ InputPatch,
14
+ InterventionContext,
15
+ PatchOp,
16
+ PatchOperation,
17
+ SemanticFamily,
18
+ TraceEvent,
19
+ TracePhase,
20
+ ValidationError,
21
+ )
22
+ from agentinputlab.integrations.external_workers import WorkerClient
23
+ from agentinputlab.integrations.external_workers.method import (
24
+ MethodWorkerRequest,
25
+ call_worker,
26
+ dependency_fingerprint,
27
+ mapping_sequence,
28
+ optional_text,
29
+ required_text,
30
+ resolve_worker,
31
+ string_tuple,
32
+ worker_fingerprint,
33
+ )
34
+ from agentinputlab.observability import external_usage_counters
35
+
36
+ ASK_OR_ASSUME_UPSTREAM_COMMIT = "728464a755f15fd448aff61a5cb32dce97fd9c1d"
37
+ ASK_OR_ASSUME_INTENT_PROMPT_SHA256 = (
38
+ "c0a150af1020090a0016930c0a319f1df0e19eff40c034d07d141c690b611190"
39
+ )
40
+ _TRANSITIONS = {"delegate", "verdict", "force", "await", "resume", "skip_first_turn"}
41
+
42
+
43
+ @dataclass(frozen=True, slots=True)
44
+ class AskOrAssumeControls:
45
+ skip_first_turn: bool = True
46
+ max_coding_turns: int = 100
47
+ max_user_turns: int = 3
48
+ max_model_calls: int = 2
49
+ max_tokens: int = 20_000
50
+ max_cost_usd: float = 10.0
51
+
52
+ def __post_init__(self) -> None:
53
+ if not isinstance(self.skip_first_turn, bool):
54
+ raise ValidationError("Ask-or-Assume skip_first_turn must be boolean")
55
+ for name in (
56
+ "max_coding_turns",
57
+ "max_user_turns",
58
+ "max_model_calls",
59
+ "max_tokens",
60
+ ):
61
+ value = getattr(self, name)
62
+ if isinstance(value, bool) or not isinstance(value, int) or value < 1:
63
+ raise ValidationError(f"Ask-or-Assume {name} must be positive")
64
+ if (
65
+ isinstance(self.max_cost_usd, bool)
66
+ or not isinstance(self.max_cost_usd, (int, float))
67
+ or not math.isfinite(float(self.max_cost_usd))
68
+ or self.max_cost_usd < 0
69
+ ):
70
+ raise ValidationError("Ask-or-Assume max_cost_usd must be non-negative")
71
+
72
+
73
+ @dataclass(frozen=True, slots=True)
74
+ class AskOrAssumeDecision:
75
+ needs_clarification: bool
76
+ reasons: tuple[str, ...]
77
+ clarification_question: str
78
+ clarification_questions: tuple[Mapping[str, object], ...]
79
+ intent_summary: str
80
+ message: str
81
+ decision_trace: tuple[Mapping[str, object], ...]
82
+ transition: str
83
+ delegated: bool
84
+ forced: bool
85
+ first_turn_skipped: bool
86
+
87
+ @classmethod
88
+ def from_mapping(cls, value: Mapping[str, object]) -> AskOrAssumeDecision:
89
+ needs_clarification = value.get("needs_clarification")
90
+ if not isinstance(needs_clarification, bool):
91
+ raise AlgorithmOutputError("Ask-or-Assume needs_clarification must be a strict boolean")
92
+ reasons = string_tuple(
93
+ value.get("reasons", ()),
94
+ label="Ask-or-Assume reasons",
95
+ unique=True,
96
+ )
97
+ if needs_clarification and not reasons:
98
+ raise AlgorithmOutputError(
99
+ "Ask-or-Assume clarification verdict requires IntentAgent reasons"
100
+ )
101
+ question = optional_text(
102
+ value,
103
+ "clarification_question",
104
+ label="Ask-or-Assume response",
105
+ )
106
+ if needs_clarification and not question.strip():
107
+ raise AlgorithmOutputError("Ask-or-Assume ask verdict requires a question")
108
+ if not needs_clarification and question.strip():
109
+ raise AlgorithmOutputError(
110
+ "Ask-or-Assume continue verdict cannot carry a clarification question"
111
+ )
112
+ transition = required_text(value, "transition", label="Ask-or-Assume response")
113
+ if transition not in _TRANSITIONS:
114
+ raise AlgorithmOutputError("Ask-or-Assume returned an invalid state transition")
115
+ booleans: dict[str, bool] = {}
116
+ for key in ("delegated", "forced", "first_turn_skipped"):
117
+ item = value.get(key, False)
118
+ if not isinstance(item, bool):
119
+ raise AlgorithmOutputError(f"Ask-or-Assume {key} must be boolean")
120
+ booleans[key] = item
121
+ return cls(
122
+ needs_clarification=needs_clarification,
123
+ reasons=reasons,
124
+ clarification_question=question,
125
+ clarification_questions=mapping_sequence(
126
+ value.get("clarification_questions", ()),
127
+ label="Ask-or-Assume clarification questions",
128
+ ),
129
+ intent_summary=optional_text(
130
+ value,
131
+ "intent_summary",
132
+ label="Ask-or-Assume response",
133
+ ),
134
+ message=optional_text(value, "message", label="Ask-or-Assume response"),
135
+ decision_trace=mapping_sequence(
136
+ value.get("decision_trace", ()),
137
+ label="Ask-or-Assume decision trace",
138
+ ),
139
+ transition=transition,
140
+ delegated=booleans["delegated"],
141
+ forced=booleans["forced"],
142
+ first_turn_skipped=booleans["first_turn_skipped"],
143
+ )
144
+
145
+
146
+ @dataclass(slots=True)
147
+ class AskOrAssumeIntentMonitor:
148
+ """Component-only uncertainty verdict; it does not impersonate OpenHands full."""
149
+
150
+ client: WorkerClient | None = None
151
+ controls: AskOrAssumeControls = AskOrAssumeControls()
152
+ intent_prompt_hash: str = ASK_OR_ASSUME_INTENT_PROMPT_SHA256
153
+ method_id: str = "ask_or_assume"
154
+ release_id: str = "clarify.ask_or_assume.intent_monitor_component"
155
+ family: SemanticFamily = SemanticFamily.CLARIFY
156
+
157
+ def __post_init__(self) -> None:
158
+ if self.intent_prompt_hash != ASK_OR_ASSUME_INTENT_PROMPT_SHA256:
159
+ raise ValidationError("Ask-or-Assume requires the pinned official IntentAgent prompt")
160
+
161
+ def propose(self, state: FrozenAgentState, context: InterventionContext) -> InputPatch:
162
+ turn_index_raw = state.metadata.get("turn_index", len(state.messages))
163
+ if isinstance(turn_index_raw, bool) or not isinstance(turn_index_raw, int):
164
+ raise ValidationError("Ask-or-Assume turn_index must be an integer")
165
+ if turn_index_raw < 0:
166
+ raise ValidationError("Ask-or-Assume turn_index must be non-negative")
167
+ coding_turns = state.metadata.get("ask_or_assume_coding_turns", turn_index_raw)
168
+ if isinstance(coding_turns, bool) or not isinstance(coding_turns, int):
169
+ raise ValidationError("ask_or_assume_coding_turns must be an integer")
170
+ if coding_turns < 0:
171
+ raise ValidationError("ask_or_assume_coding_turns must be non-negative")
172
+ clarification_turns = state.metadata.get(
173
+ "ask_or_assume_user_turns",
174
+ state.metadata.get("ask_or_assume_clarification_turns", 0),
175
+ )
176
+ if (
177
+ isinstance(clarification_turns, bool)
178
+ or not isinstance(clarification_turns, int)
179
+ or clarification_turns < 0
180
+ ):
181
+ raise ValidationError(
182
+ "ask_or_assume_clarification_turns must be a non-negative integer"
183
+ )
184
+ if self.controls.skip_first_turn and turn_index_raw == 0:
185
+ return self._first_turn_patch(state)
186
+ if coding_turns >= self.controls.max_coding_turns:
187
+ raise AlgorithmOutputError("Ask-or-Assume effective coding-turn budget is exhausted")
188
+
189
+ worker = resolve_worker(self.client, context, label="Ask-or-Assume")
190
+ coding_backend_fp = dependency_fingerprint(
191
+ context.backend,
192
+ label="Ask-or-Assume coding/clarification model",
193
+ required=True,
194
+ )
195
+ configured_intent_fp = context.run_config.get("ask_or_assume_intent_model_fingerprint")
196
+ if configured_intent_fp is None:
197
+ intent_backend_fp = coding_backend_fp
198
+ elif isinstance(configured_intent_fp, Mapping) and configured_intent_fp:
199
+ intent_backend_fp = dict(configured_intent_fp)
200
+ else:
201
+ raise ValidationError(
202
+ "ask_or_assume_intent_model_fingerprint must be a non-empty mapping"
203
+ )
204
+ request = MethodWorkerRequest(
205
+ method_id=self.method_id,
206
+ release_id=self.release_id,
207
+ upstream_commit=ASK_OR_ASSUME_UPSTREAM_COMMIT,
208
+ state_hash=state.state_hash,
209
+ payload={
210
+ "task_id": state.task_id,
211
+ "task_view": {
212
+ "query": state.query,
213
+ "instruction": state.instruction,
214
+ "context": state.context,
215
+ "tool_names": state.tool_names,
216
+ },
217
+ "history": state.messages,
218
+ "turn_index": coding_turns,
219
+ "clarification_turns": clarification_turns,
220
+ "controls": asdict(self.controls),
221
+ "intent_prompt_hash": self.intent_prompt_hash,
222
+ "seed": context.seed,
223
+ },
224
+ dependency_fingerprints={
225
+ "intent_model": intent_backend_fp,
226
+ "coding_model": coding_backend_fp,
227
+ },
228
+ minimum_reservation={"model_calls": 1},
229
+ protocol_version="agentinputlab.ask-or-assume.monitor.v1",
230
+ )
231
+ response = call_worker(worker, request)
232
+ decision = AskOrAssumeDecision.from_mapping(response.payload)
233
+ minimum_calls = 2 if decision.needs_clarification else 1
234
+ if response.usage.model_calls < minimum_calls:
235
+ raise AlgorithmOutputError(
236
+ "Ask-or-Assume usage omits the IntentAgent or Clarify-question model call"
237
+ )
238
+ if decision.first_turn_skipped:
239
+ raise AlgorithmOutputError(
240
+ "Ask-or-Assume worker cannot retroactively skip a delegated turn"
241
+ )
242
+ if not decision.delegated:
243
+ raise AlgorithmOutputError(
244
+ "Ask-or-Assume post-first-turn verdict must come from delegated IntentAgent"
245
+ )
246
+ if decision.forced and not decision.needs_clarification:
247
+ raise AlgorithmOutputError("Ask-or-Assume forced transition must request clarification")
248
+ if decision.needs_clarification and not decision.forced:
249
+ raise AlgorithmOutputError(
250
+ "Ask-or-Assume clarification must use the forced Clarify-only transition"
251
+ )
252
+ if decision.needs_clarification and decision.transition not in {"force", "await"}:
253
+ raise AlgorithmOutputError(
254
+ "Ask-or-Assume clarification requires a force/await controller transition"
255
+ )
256
+ if not decision.needs_clarification and decision.transition in {"force", "await"}:
257
+ raise AlgorithmOutputError(
258
+ "Ask-or-Assume continue verdict cannot enter a clarification transition"
259
+ )
260
+ if decision.needs_clarification and clarification_turns >= self.controls.max_user_turns:
261
+ raise AlgorithmOutputError(
262
+ "Ask-or-Assume worker requested clarification after max_user_turns"
263
+ )
264
+ if decision.needs_clarification:
265
+ operations = (
266
+ PatchOperation(
267
+ PatchOp.REQUEST_CLARIFICATION,
268
+ decision.clarification_question,
269
+ target="user",
270
+ rationale="Ask-or-Assume intent monitor requested missing task information",
271
+ ),
272
+ PatchOperation(
273
+ PatchOp.SET_METADATA,
274
+ {
275
+ "ask_or_assume_clarification_turns": clarification_turns + 1,
276
+ "ask_or_assume_user_turns": clarification_turns + 1,
277
+ },
278
+ target="metadata",
279
+ rationale="track the release-level clarification-turn safety limit",
280
+ ),
281
+ )
282
+ else:
283
+ operations = (
284
+ PatchOperation(
285
+ PatchOp.NO_OP,
286
+ rationale="Ask-or-Assume intent monitor allowed the coding agent to proceed",
287
+ ),
288
+ )
289
+ trace = ExecutionTrace(
290
+ method_id=self.method_id,
291
+ release_id=self.release_id,
292
+ events=(
293
+ TraceEvent(
294
+ TracePhase.PREPARE,
295
+ "intent_monitor_controls",
296
+ {
297
+ **asdict(self.controls),
298
+ "intent_prompt_hash": self.intent_prompt_hash,
299
+ "turn_index": turn_index_raw,
300
+ "effective_coding_turns": coding_turns,
301
+ "clarification_turns": clarification_turns,
302
+ },
303
+ ),
304
+ TraceEvent(
305
+ TracePhase.PROPOSE,
306
+ "intent_verdict",
307
+ {
308
+ "needs_clarification": decision.needs_clarification,
309
+ "reasons": decision.reasons,
310
+ "intent_summary": decision.intent_summary,
311
+ "message": decision.message,
312
+ "clarification_question": decision.clarification_question,
313
+ "clarification_questions": decision.clarification_questions,
314
+ "worker_decision_trace": decision.decision_trace,
315
+ },
316
+ ),
317
+ TraceEvent(
318
+ TracePhase.FINALIZE,
319
+ "controller_transition",
320
+ {
321
+ "transition": decision.transition,
322
+ "delegated": decision.delegated,
323
+ "forced": decision.forced,
324
+ "first_turn_skipped": decision.first_turn_skipped,
325
+ },
326
+ ),
327
+ ),
328
+ counters={
329
+ **external_usage_counters(response.usage),
330
+ "worker_calls": 1,
331
+ "clarification_requests": int(decision.needs_clarification),
332
+ },
333
+ )
334
+ return InputPatch(
335
+ method_id=self.method_id,
336
+ release_id=self.release_id,
337
+ source_state_hash=state.state_hash,
338
+ operations=operations,
339
+ trace=trace,
340
+ implementation_identity=(
341
+ f"official_component_adapter:nedwards99/ask-or-assume@"
342
+ f"{ASK_OR_ASSUME_UPSTREAM_COMMIT}"
343
+ ),
344
+ profile_kind="controlled_adapter",
345
+ metadata={
346
+ "implementation_origin": "official_component_adapter",
347
+ "component_only": True,
348
+ "openhands_full_system": False,
349
+ "needs_clarification": decision.needs_clarification,
350
+ "clarification_turns_before": clarification_turns,
351
+ "clarification_turns_after": (
352
+ clarification_turns + int(decision.needs_clarification)
353
+ ),
354
+ "reasons": decision.reasons,
355
+ "transition": decision.transition,
356
+ "worker_fingerprint": worker_fingerprint(worker, label="Ask-or-Assume"),
357
+ "intent_model_fingerprint": intent_backend_fp,
358
+ "coding_model_fingerprint": coding_backend_fp,
359
+ "effective_coding_turns": coding_turns,
360
+ "effective_user_turns": clarification_turns,
361
+ },
362
+ )
363
+
364
+ def _first_turn_patch(self, state: FrozenAgentState) -> InputPatch:
365
+ """Preserve the upstream rule: the coding agent acts once before delegation."""
366
+
367
+ trace = ExecutionTrace(
368
+ method_id=self.method_id,
369
+ release_id=self.release_id,
370
+ events=(
371
+ TraceEvent(
372
+ TracePhase.FINALIZE,
373
+ "skip_first_coding_turn",
374
+ {"transition": "skip_first_turn", "worker_called": False},
375
+ ),
376
+ ),
377
+ counters={"worker_calls": 0, "clarification_requests": 0},
378
+ )
379
+ return InputPatch(
380
+ method_id=self.method_id,
381
+ release_id=self.release_id,
382
+ source_state_hash=state.state_hash,
383
+ operations=(
384
+ PatchOperation(
385
+ PatchOp.NO_OP,
386
+ rationale="Ask-or-Assume delegates only after the first coding turn",
387
+ ),
388
+ ),
389
+ trace=trace,
390
+ implementation_identity=(
391
+ f"official_component_adapter:nedwards99/ask-or-assume@"
392
+ f"{ASK_OR_ASSUME_UPSTREAM_COMMIT}"
393
+ ),
394
+ profile_kind="controlled_adapter",
395
+ metadata={
396
+ "implementation_origin": "official_component_adapter",
397
+ "component_only": True,
398
+ "openhands_full_system": False,
399
+ "first_turn_skipped": True,
400
+ },
401
+ )
402
+
403
+
404
+ __all__ = [
405
+ "ASK_OR_ASSUME_INTENT_PROMPT_SHA256",
406
+ "ASK_OR_ASSUME_UPSTREAM_COMMIT",
407
+ "AskOrAssumeControls",
408
+ "AskOrAssumeDecision",
409
+ "AskOrAssumeIntentMonitor",
410
+ ]