@sprinterai/runtime 0.4.1 → 0.6.0

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 (389) hide show
  1. package/dist/adapters/a2a-adapter.d.ts +37 -9
  2. package/dist/adapters/a2a-adapter.d.ts.map +1 -1
  3. package/dist/adapters/a2a-adapter.js +132 -9
  4. package/dist/adapters/a2a-adapter.js.map +1 -1
  5. package/dist/adapters/a2a-adapter.test.d.ts +2 -0
  6. package/dist/adapters/a2a-adapter.test.d.ts.map +1 -0
  7. package/dist/adapters/a2a-adapter.test.js +295 -0
  8. package/dist/adapters/a2a-adapter.test.js.map +1 -0
  9. package/dist/adapters/http-agent-adapter.d.ts +19 -9
  10. package/dist/adapters/http-agent-adapter.d.ts.map +1 -1
  11. package/dist/adapters/http-agent-adapter.js +55 -9
  12. package/dist/adapters/http-agent-adapter.js.map +1 -1
  13. package/dist/adapters/http-agent-adapter.test.d.ts +2 -0
  14. package/dist/adapters/http-agent-adapter.test.d.ts.map +1 -0
  15. package/dist/adapters/http-agent-adapter.test.js +164 -0
  16. package/dist/adapters/http-agent-adapter.test.js.map +1 -0
  17. package/dist/adapters/index.d.ts +8 -8
  18. package/dist/adapters/index.d.ts.map +1 -1
  19. package/dist/adapters/index.js +4 -4
  20. package/dist/adapters/index.js.map +1 -1
  21. package/dist/adapters/mcp-adapter.d.ts +29 -9
  22. package/dist/adapters/mcp-adapter.d.ts.map +1 -1
  23. package/dist/adapters/mcp-adapter.js +43 -8
  24. package/dist/adapters/mcp-adapter.js.map +1 -1
  25. package/dist/adapters/mcp-adapter.test.d.ts +2 -0
  26. package/dist/adapters/mcp-adapter.test.d.ts.map +1 -0
  27. package/dist/adapters/mcp-adapter.test.js +118 -0
  28. package/dist/adapters/mcp-adapter.test.js.map +1 -0
  29. package/dist/adapters/openclaw-adapter.d.ts +34 -8
  30. package/dist/adapters/openclaw-adapter.d.ts.map +1 -1
  31. package/dist/adapters/openclaw-adapter.js +38 -8
  32. package/dist/adapters/openclaw-adapter.js.map +1 -1
  33. package/dist/adapters/openclaw-adapter.test.d.ts +2 -0
  34. package/dist/adapters/openclaw-adapter.test.d.ts.map +1 -0
  35. package/dist/adapters/openclaw-adapter.test.js +77 -0
  36. package/dist/adapters/openclaw-adapter.test.js.map +1 -0
  37. package/dist/agent/agent-registry.test.js +1 -1
  38. package/dist/agent/agent-resolver.d.ts +8 -2
  39. package/dist/agent/agent-resolver.d.ts.map +1 -1
  40. package/dist/agent/agent-resolver.js +7 -1
  41. package/dist/agent/agent-resolver.js.map +1 -1
  42. package/dist/agent/agent-resolver.test.js +21 -3
  43. package/dist/agent/agent-resolver.test.js.map +1 -1
  44. package/dist/agent/delegate.test.js +1 -1
  45. package/dist/agent/execute-agent.d.ts +32 -11
  46. package/dist/agent/execute-agent.d.ts.map +1 -1
  47. package/dist/agent/execute-agent.js +39 -3
  48. package/dist/agent/execute-agent.js.map +1 -1
  49. package/dist/agent/index.d.ts +9 -9
  50. package/dist/agent/index.d.ts.map +1 -1
  51. package/dist/agent/index.js +5 -5
  52. package/dist/agent/index.js.map +1 -1
  53. package/dist/agent/prompt-builder.test.js +1 -1
  54. package/dist/approval/approval-manager.d.ts +19 -0
  55. package/dist/approval/approval-manager.d.ts.map +1 -0
  56. package/dist/approval/approval-manager.js +36 -0
  57. package/dist/approval/approval-manager.js.map +1 -0
  58. package/dist/approval/approval-manager.test.d.ts +2 -0
  59. package/dist/approval/approval-manager.test.d.ts.map +1 -0
  60. package/dist/approval/approval-manager.test.js +239 -0
  61. package/dist/approval/approval-manager.test.js.map +1 -0
  62. package/dist/approval/index.d.ts +3 -0
  63. package/dist/approval/index.d.ts.map +1 -0
  64. package/dist/approval/index.js +2 -0
  65. package/dist/approval/index.js.map +1 -0
  66. package/dist/chat/chat-handler.d.ts +16 -7
  67. package/dist/chat/chat-handler.d.ts.map +1 -1
  68. package/dist/chat/chat-handler.js +68 -20
  69. package/dist/chat/chat-handler.js.map +1 -1
  70. package/dist/chat/chat-handler.test.js +101 -11
  71. package/dist/chat/chat-handler.test.js.map +1 -1
  72. package/dist/chat/index.d.ts +3 -3
  73. package/dist/chat/index.js +2 -2
  74. package/dist/chat/message-utils.d.ts +15 -7
  75. package/dist/chat/message-utils.d.ts.map +1 -1
  76. package/dist/chat/message-utils.js +94 -22
  77. package/dist/chat/message-utils.js.map +1 -1
  78. package/dist/chat/message-utils.test.js +71 -1
  79. package/dist/chat/message-utils.test.js.map +1 -1
  80. package/dist/document/chunk-generator.d.ts +6 -0
  81. package/dist/document/chunk-generator.d.ts.map +1 -0
  82. package/dist/document/chunk-generator.js +107 -0
  83. package/dist/document/chunk-generator.js.map +1 -0
  84. package/dist/document/chunk-generator.test.d.ts +2 -0
  85. package/dist/document/chunk-generator.test.d.ts.map +1 -0
  86. package/dist/document/chunk-generator.test.js +166 -0
  87. package/dist/document/chunk-generator.test.js.map +1 -0
  88. package/dist/document/document-processor.d.ts +27 -0
  89. package/dist/document/document-processor.d.ts.map +1 -0
  90. package/dist/document/document-processor.js +44 -0
  91. package/dist/document/document-processor.js.map +1 -0
  92. package/dist/document/document-processor.test.d.ts +2 -0
  93. package/dist/document/document-processor.test.d.ts.map +1 -0
  94. package/dist/document/document-processor.test.js +197 -0
  95. package/dist/document/document-processor.test.js.map +1 -0
  96. package/dist/document/index.d.ts +5 -0
  97. package/dist/document/index.d.ts.map +1 -0
  98. package/dist/document/index.js +4 -0
  99. package/dist/document/index.js.map +1 -0
  100. package/dist/document/parsers/index.d.ts +2 -0
  101. package/dist/document/parsers/index.d.ts.map +1 -0
  102. package/dist/document/parsers/index.js +2 -0
  103. package/dist/document/parsers/index.js.map +1 -0
  104. package/dist/document/parsers/text-parser.d.ts +4 -0
  105. package/dist/document/parsers/text-parser.d.ts.map +1 -0
  106. package/dist/document/parsers/text-parser.js +23 -0
  107. package/dist/document/parsers/text-parser.js.map +1 -0
  108. package/dist/document/parsers/text-parser.test.d.ts +2 -0
  109. package/dist/document/parsers/text-parser.test.d.ts.map +1 -0
  110. package/dist/document/parsers/text-parser.test.js +64 -0
  111. package/dist/document/parsers/text-parser.test.js.map +1 -0
  112. package/dist/eval/eval-runner.test.js +2 -2
  113. package/dist/eval/index.d.ts +3 -3
  114. package/dist/eval/index.js +2 -2
  115. package/dist/eval/scorers.test.js +1 -1
  116. package/dist/events/event-bus.d.ts +12 -0
  117. package/dist/events/event-bus.d.ts.map +1 -0
  118. package/dist/events/event-bus.js +77 -0
  119. package/dist/events/event-bus.js.map +1 -0
  120. package/dist/events/event-bus.test.d.ts +2 -0
  121. package/dist/events/event-bus.test.d.ts.map +1 -0
  122. package/dist/events/event-bus.test.js +155 -0
  123. package/dist/events/event-bus.test.js.map +1 -0
  124. package/dist/events/index.d.ts +2 -0
  125. package/dist/events/index.d.ts.map +1 -0
  126. package/dist/events/index.js +2 -0
  127. package/dist/events/index.js.map +1 -0
  128. package/dist/guardrail/guardrail-pipeline.test.js +1 -1
  129. package/dist/guardrail/index.d.ts +2 -2
  130. package/dist/guardrail/index.js +1 -1
  131. package/dist/index.d.ts +81 -45
  132. package/dist/index.d.ts.map +1 -1
  133. package/dist/index.js +55 -27
  134. package/dist/index.js.map +1 -1
  135. package/dist/jobs/in-memory-job-queue.d.ts +20 -0
  136. package/dist/jobs/in-memory-job-queue.d.ts.map +1 -0
  137. package/dist/jobs/in-memory-job-queue.js +120 -0
  138. package/dist/jobs/in-memory-job-queue.js.map +1 -0
  139. package/dist/jobs/in-memory-job-queue.test.d.ts +2 -0
  140. package/dist/jobs/in-memory-job-queue.test.d.ts.map +1 -0
  141. package/dist/jobs/in-memory-job-queue.test.js +146 -0
  142. package/dist/jobs/in-memory-job-queue.test.js.map +1 -0
  143. package/dist/jobs/index.d.ts +4 -0
  144. package/dist/jobs/index.d.ts.map +1 -0
  145. package/dist/jobs/index.js +3 -0
  146. package/dist/jobs/index.js.map +1 -0
  147. package/dist/jobs/job-runner.d.ts +42 -0
  148. package/dist/jobs/job-runner.d.ts.map +1 -0
  149. package/dist/jobs/job-runner.js +119 -0
  150. package/dist/jobs/job-runner.js.map +1 -0
  151. package/dist/jobs/job-runner.test.d.ts +2 -0
  152. package/dist/jobs/job-runner.test.d.ts.map +1 -0
  153. package/dist/jobs/job-runner.test.js +190 -0
  154. package/dist/jobs/job-runner.test.js.map +1 -0
  155. package/dist/memory/index.d.ts +3 -3
  156. package/dist/memory/index.js +2 -2
  157. package/dist/memory/memory-prompt.d.ts +1 -1
  158. package/dist/module/create-runtime.d.ts +21 -6
  159. package/dist/module/create-runtime.d.ts.map +1 -1
  160. package/dist/module/create-runtime.js +7 -7
  161. package/dist/module/create-runtime.js.map +1 -1
  162. package/dist/module/create-runtime.test.js +8 -8
  163. package/dist/module/index.d.ts +5 -4
  164. package/dist/module/index.d.ts.map +1 -1
  165. package/dist/module/index.js +3 -2
  166. package/dist/module/index.js.map +1 -1
  167. package/dist/module/map-supabase-stores.d.ts +33 -0
  168. package/dist/module/map-supabase-stores.d.ts.map +1 -0
  169. package/dist/module/map-supabase-stores.js +28 -0
  170. package/dist/module/map-supabase-stores.js.map +1 -0
  171. package/dist/module/module-loader.d.ts +9 -1
  172. package/dist/module/module-loader.d.ts.map +1 -1
  173. package/dist/module/module-loader.js +40 -1
  174. package/dist/module/module-loader.js.map +1 -1
  175. package/dist/module/module-loader.test.js +38 -1
  176. package/dist/module/module-loader.test.js.map +1 -1
  177. package/dist/notification/digest-scheduler.d.ts +18 -0
  178. package/dist/notification/digest-scheduler.d.ts.map +1 -0
  179. package/dist/notification/digest-scheduler.js +44 -0
  180. package/dist/notification/digest-scheduler.js.map +1 -0
  181. package/dist/notification/digest-scheduler.test.d.ts +2 -0
  182. package/dist/notification/digest-scheduler.test.d.ts.map +1 -0
  183. package/dist/notification/digest-scheduler.test.js +306 -0
  184. package/dist/notification/digest-scheduler.test.js.map +1 -0
  185. package/dist/notification/index.d.ts +5 -0
  186. package/dist/notification/index.d.ts.map +1 -0
  187. package/dist/notification/index.js +3 -0
  188. package/dist/notification/index.js.map +1 -0
  189. package/dist/notification/notification-engine.d.ts +20 -0
  190. package/dist/notification/notification-engine.d.ts.map +1 -0
  191. package/dist/notification/notification-engine.js +78 -0
  192. package/dist/notification/notification-engine.js.map +1 -0
  193. package/dist/notification/notification-engine.test.d.ts +2 -0
  194. package/dist/notification/notification-engine.test.d.ts.map +1 -0
  195. package/dist/notification/notification-engine.test.js +364 -0
  196. package/dist/notification/notification-engine.test.js.map +1 -0
  197. package/dist/providers/index.d.ts +5 -0
  198. package/dist/providers/index.d.ts.map +1 -0
  199. package/dist/providers/index.js +3 -0
  200. package/dist/providers/index.js.map +1 -0
  201. package/dist/providers/model-resolver.d.ts +93 -0
  202. package/dist/providers/model-resolver.d.ts.map +1 -0
  203. package/dist/providers/model-resolver.js +152 -0
  204. package/dist/providers/model-resolver.js.map +1 -0
  205. package/dist/providers/model-resolver.test.d.ts +2 -0
  206. package/dist/providers/model-resolver.test.d.ts.map +1 -0
  207. package/dist/providers/model-resolver.test.js +199 -0
  208. package/dist/providers/model-resolver.test.js.map +1 -0
  209. package/dist/providers/noop-providers.d.ts +6 -0
  210. package/dist/providers/noop-providers.d.ts.map +1 -0
  211. package/dist/providers/noop-providers.js +15 -0
  212. package/dist/providers/noop-providers.js.map +1 -0
  213. package/dist/providers/noop-providers.test.d.ts +2 -0
  214. package/dist/providers/noop-providers.test.d.ts.map +1 -0
  215. package/dist/providers/noop-providers.test.js +63 -0
  216. package/dist/providers/noop-providers.test.js.map +1 -0
  217. package/dist/providers/provider-interfaces.d.ts +65 -0
  218. package/dist/providers/provider-interfaces.d.ts.map +1 -0
  219. package/dist/providers/provider-interfaces.js +2 -0
  220. package/dist/providers/provider-interfaces.js.map +1 -0
  221. package/dist/realtime/index.d.ts +3 -0
  222. package/dist/realtime/index.d.ts.map +1 -0
  223. package/dist/realtime/index.js +2 -0
  224. package/dist/realtime/index.js.map +1 -0
  225. package/dist/realtime/realtime-manager.d.ts +20 -0
  226. package/dist/realtime/realtime-manager.d.ts.map +1 -0
  227. package/dist/realtime/realtime-manager.js +110 -0
  228. package/dist/realtime/realtime-manager.js.map +1 -0
  229. package/dist/realtime/realtime-manager.test.d.ts +2 -0
  230. package/dist/realtime/realtime-manager.test.d.ts.map +1 -0
  231. package/dist/realtime/realtime-manager.test.js +273 -0
  232. package/dist/realtime/realtime-manager.test.js.map +1 -0
  233. package/dist/scoring/compute-score.test.js +1 -1
  234. package/dist/scoring/index.d.ts +2 -2
  235. package/dist/scoring/index.js +1 -1
  236. package/dist/search/cosine-similarity.d.ts +8 -0
  237. package/dist/search/cosine-similarity.d.ts.map +1 -0
  238. package/dist/search/cosine-similarity.js +28 -0
  239. package/dist/search/cosine-similarity.js.map +1 -0
  240. package/dist/search/cosine-similarity.test.d.ts +2 -0
  241. package/dist/search/cosine-similarity.test.d.ts.map +1 -0
  242. package/dist/search/cosine-similarity.test.js +49 -0
  243. package/dist/search/cosine-similarity.test.js.map +1 -0
  244. package/dist/search/hybrid-search.d.ts +47 -0
  245. package/dist/search/hybrid-search.d.ts.map +1 -0
  246. package/dist/search/hybrid-search.js +111 -0
  247. package/dist/search/hybrid-search.js.map +1 -0
  248. package/dist/search/hybrid-search.test.d.ts +2 -0
  249. package/dist/search/hybrid-search.test.d.ts.map +1 -0
  250. package/dist/search/hybrid-search.test.js +238 -0
  251. package/dist/search/hybrid-search.test.js.map +1 -0
  252. package/dist/search/in-memory-search.d.ts +17 -0
  253. package/dist/search/in-memory-search.d.ts.map +1 -0
  254. package/dist/search/in-memory-search.js +59 -0
  255. package/dist/search/in-memory-search.js.map +1 -0
  256. package/dist/search/in-memory-search.test.d.ts +2 -0
  257. package/dist/search/in-memory-search.test.d.ts.map +1 -0
  258. package/dist/search/in-memory-search.test.js +169 -0
  259. package/dist/search/in-memory-search.test.js.map +1 -0
  260. package/dist/search/index.d.ts +6 -0
  261. package/dist/search/index.d.ts.map +1 -0
  262. package/dist/search/index.js +4 -0
  263. package/dist/search/index.js.map +1 -0
  264. package/dist/testing/in-memory-agent-store.d.ts +5 -1
  265. package/dist/testing/in-memory-agent-store.d.ts.map +1 -1
  266. package/dist/testing/in-memory-agent-store.js +19 -0
  267. package/dist/testing/in-memory-agent-store.js.map +1 -1
  268. package/dist/testing/in-memory-agent-store.test.js +1 -1
  269. package/dist/testing/in-memory-chat-store.d.ts +4 -3
  270. package/dist/testing/in-memory-chat-store.d.ts.map +1 -1
  271. package/dist/testing/in-memory-chat-store.js +3 -2
  272. package/dist/testing/in-memory-chat-store.js.map +1 -1
  273. package/dist/testing/in-memory-entity-store.d.ts +5 -1
  274. package/dist/testing/in-memory-entity-store.d.ts.map +1 -1
  275. package/dist/testing/in-memory-entity-store.js +69 -4
  276. package/dist/testing/in-memory-entity-store.js.map +1 -1
  277. package/dist/testing/in-memory-entity-store.test.js +1 -1
  278. package/dist/testing/in-memory-memory-store.d.ts +3 -1
  279. package/dist/testing/in-memory-memory-store.d.ts.map +1 -1
  280. package/dist/testing/in-memory-memory-store.js +20 -0
  281. package/dist/testing/in-memory-memory-store.js.map +1 -1
  282. package/dist/testing/in-memory-tool-store.d.ts +13 -2
  283. package/dist/testing/in-memory-tool-store.d.ts.map +1 -1
  284. package/dist/testing/in-memory-tool-store.js +51 -0
  285. package/dist/testing/in-memory-tool-store.js.map +1 -1
  286. package/dist/testing/index.d.ts +7 -7
  287. package/dist/testing/index.js +7 -7
  288. package/dist/tool/ai-bridge.d.ts +1 -0
  289. package/dist/tool/ai-bridge.d.ts.map +1 -1
  290. package/dist/tool/ai-bridge.js +1 -0
  291. package/dist/tool/ai-bridge.js.map +1 -1
  292. package/dist/tool/ai-bridge.test.js +1 -1
  293. package/dist/tool/catalog.d.ts +19 -0
  294. package/dist/tool/catalog.d.ts.map +1 -0
  295. package/dist/tool/catalog.js +88 -0
  296. package/dist/tool/catalog.js.map +1 -0
  297. package/dist/tool/catalog.test.d.ts +2 -0
  298. package/dist/tool/catalog.test.d.ts.map +1 -0
  299. package/dist/tool/catalog.test.js +129 -0
  300. package/dist/tool/catalog.test.js.map +1 -0
  301. package/dist/tool/entity-tools-factory.d.ts.map +1 -1
  302. package/dist/tool/entity-tools-factory.js +80 -0
  303. package/dist/tool/entity-tools-factory.js.map +1 -1
  304. package/dist/tool/entity-tools-factory.test.js +70 -4
  305. package/dist/tool/entity-tools-factory.test.js.map +1 -1
  306. package/dist/tool/execute-registered-tool.d.ts +17 -0
  307. package/dist/tool/execute-registered-tool.d.ts.map +1 -0
  308. package/dist/tool/execute-registered-tool.js +24 -0
  309. package/dist/tool/execute-registered-tool.js.map +1 -0
  310. package/dist/tool/execute-registered-tool.test.d.ts +2 -0
  311. package/dist/tool/execute-registered-tool.test.d.ts.map +1 -0
  312. package/dist/tool/execute-registered-tool.test.js +73 -0
  313. package/dist/tool/execute-registered-tool.test.js.map +1 -0
  314. package/dist/tool/index.d.ts +11 -7
  315. package/dist/tool/index.d.ts.map +1 -1
  316. package/dist/tool/index.js +7 -5
  317. package/dist/tool/index.js.map +1 -1
  318. package/dist/tool/resolve-tools.d.ts +3 -1
  319. package/dist/tool/resolve-tools.d.ts.map +1 -1
  320. package/dist/tool/resolve-tools.js +1 -1
  321. package/dist/tool/resolve-tools.js.map +1 -1
  322. package/dist/tool/resolve-tools.test.js +1 -1
  323. package/dist/tool/tool-executor.d.ts +3 -2
  324. package/dist/tool/tool-executor.d.ts.map +1 -1
  325. package/dist/tool/tool-executor.js +4 -2
  326. package/dist/tool/tool-executor.js.map +1 -1
  327. package/dist/tool/tool-executor.test.js +2 -2
  328. package/dist/tool/tool-registry.test.js +1 -1
  329. package/dist/tool/zod-to-json-schema.d.ts +7 -0
  330. package/dist/tool/zod-to-json-schema.d.ts.map +1 -0
  331. package/dist/tool/zod-to-json-schema.js +12 -0
  332. package/dist/tool/zod-to-json-schema.js.map +1 -0
  333. package/dist/tool/zod-to-json-schema.test.d.ts +2 -0
  334. package/dist/tool/zod-to-json-schema.test.d.ts.map +1 -0
  335. package/dist/tool/zod-to-json-schema.test.js +127 -0
  336. package/dist/tool/zod-to-json-schema.test.js.map +1 -0
  337. package/dist/webhook/index.d.ts +4 -0
  338. package/dist/webhook/index.d.ts.map +1 -0
  339. package/dist/webhook/index.js +3 -0
  340. package/dist/webhook/index.js.map +1 -0
  341. package/dist/webhook/webhook-delivery.d.ts +12 -0
  342. package/dist/webhook/webhook-delivery.d.ts.map +1 -0
  343. package/dist/webhook/webhook-delivery.js +102 -0
  344. package/dist/webhook/webhook-delivery.js.map +1 -0
  345. package/dist/webhook/webhook-delivery.test.d.ts +2 -0
  346. package/dist/webhook/webhook-delivery.test.d.ts.map +1 -0
  347. package/dist/webhook/webhook-delivery.test.js +284 -0
  348. package/dist/webhook/webhook-delivery.test.js.map +1 -0
  349. package/dist/webhook/webhook-signer.d.ts +14 -0
  350. package/dist/webhook/webhook-signer.d.ts.map +1 -0
  351. package/dist/webhook/webhook-signer.js +31 -0
  352. package/dist/webhook/webhook-signer.js.map +1 -0
  353. package/dist/webhook/webhook-signer.test.d.ts +2 -0
  354. package/dist/webhook/webhook-signer.test.d.ts.map +1 -0
  355. package/dist/webhook/webhook-signer.test.js +74 -0
  356. package/dist/webhook/webhook-signer.test.js.map +1 -0
  357. package/dist/workflow/compile.js +4 -1
  358. package/dist/workflow/compile.js.map +1 -1
  359. package/dist/workflow/compile.test.js +1 -1
  360. package/dist/workflow/evaluate-nodes.d.ts +24 -0
  361. package/dist/workflow/evaluate-nodes.d.ts.map +1 -0
  362. package/dist/workflow/evaluate-nodes.js +126 -0
  363. package/dist/workflow/evaluate-nodes.js.map +1 -0
  364. package/dist/workflow/evaluate-nodes.test.d.ts +2 -0
  365. package/dist/workflow/evaluate-nodes.test.d.ts.map +1 -0
  366. package/dist/workflow/evaluate-nodes.test.js +363 -0
  367. package/dist/workflow/evaluate-nodes.test.js.map +1 -0
  368. package/dist/workflow/index.d.ts +8 -3
  369. package/dist/workflow/index.d.ts.map +1 -1
  370. package/dist/workflow/index.js +4 -2
  371. package/dist/workflow/index.js.map +1 -1
  372. package/dist/workflow/node-executor.d.ts +40 -0
  373. package/dist/workflow/node-executor.d.ts.map +1 -0
  374. package/dist/workflow/node-executor.js +2 -0
  375. package/dist/workflow/node-executor.js.map +1 -0
  376. package/dist/workflow/node-executor.test.d.ts +2 -0
  377. package/dist/workflow/node-executor.test.d.ts.map +1 -0
  378. package/dist/workflow/node-executor.test.js +91 -0
  379. package/dist/workflow/node-executor.test.js.map +1 -0
  380. package/dist/workflow/status.test.js +1 -1
  381. package/dist/workflow/workflow-runner.d.ts +57 -0
  382. package/dist/workflow/workflow-runner.d.ts.map +1 -0
  383. package/dist/workflow/workflow-runner.js +263 -0
  384. package/dist/workflow/workflow-runner.js.map +1 -0
  385. package/dist/workflow/workflow-runner.test.d.ts +2 -0
  386. package/dist/workflow/workflow-runner.test.d.ts.map +1 -0
  387. package/dist/workflow/workflow-runner.test.js +657 -0
  388. package/dist/workflow/workflow-runner.test.js.map +1 -0
  389. package/package.json +9 -4
@@ -0,0 +1,107 @@
1
+ const DEFAULT_CHUNK_SIZE = 1000;
2
+ const DEFAULT_OVERLAP = 200;
3
+ /** Ordered list of separators for recursive text splitting. */
4
+ const SEPARATORS = ['\n\n', '\n', '. ', ' '];
5
+ /**
6
+ * Recursively split text using the separator hierarchy.
7
+ * Tries the first separator; if a segment is still too large, recurses with the next.
8
+ */
9
+ function recursiveSplit(text, chunkSize, separatorIndex) {
10
+ // Base case: text fits in a single chunk
11
+ if (text.length <= chunkSize) {
12
+ return text.length > 0 ? [text] : [];
13
+ }
14
+ // If we've exhausted all separators, hard-split at chunkSize
15
+ if (separatorIndex >= SEPARATORS.length) {
16
+ const chunks = [];
17
+ for (let i = 0; i < text.length; i += chunkSize) {
18
+ chunks.push(text.slice(i, i + chunkSize));
19
+ }
20
+ return chunks;
21
+ }
22
+ const separator = SEPARATORS[separatorIndex];
23
+ const parts = text.split(separator);
24
+ // If the separator doesn't split anything useful, try the next separator
25
+ if (parts.length <= 1) {
26
+ return recursiveSplit(text, chunkSize, separatorIndex + 1);
27
+ }
28
+ const chunks = [];
29
+ let currentChunk = '';
30
+ for (const part of parts) {
31
+ const candidate = currentChunk.length === 0 ? part : currentChunk + separator + part;
32
+ if (candidate.length <= chunkSize) {
33
+ currentChunk = candidate;
34
+ }
35
+ else {
36
+ // Flush current chunk if it has content
37
+ if (currentChunk.length > 0) {
38
+ chunks.push(currentChunk);
39
+ }
40
+ // If the part itself is too large, recursively split it
41
+ if (part.length > chunkSize) {
42
+ const subChunks = recursiveSplit(part, chunkSize, separatorIndex + 1);
43
+ // Add all sub-chunks except the last as completed chunks
44
+ for (let i = 0; i < subChunks.length - 1; i++) {
45
+ chunks.push(subChunks[i]);
46
+ }
47
+ // The last sub-chunk becomes the new current chunk
48
+ currentChunk = subChunks.length > 0 ? subChunks[subChunks.length - 1] : '';
49
+ }
50
+ else {
51
+ currentChunk = part;
52
+ }
53
+ }
54
+ }
55
+ if (currentChunk.length > 0) {
56
+ chunks.push(currentChunk);
57
+ }
58
+ return chunks;
59
+ }
60
+ /**
61
+ * Apply overlap: each chunk (except the first) includes the last `overlap`
62
+ * characters from the previous chunk prepended to it.
63
+ */
64
+ function applyOverlap(chunks, overlap) {
65
+ if (chunks.length <= 1 || overlap <= 0) {
66
+ return chunks;
67
+ }
68
+ const result = [chunks[0]];
69
+ for (let i = 1; i < chunks.length; i++) {
70
+ const prev = chunks[i - 1];
71
+ const overlapText = prev.slice(Math.max(0, prev.length - overlap));
72
+ result.push(overlapText + chunks[i]);
73
+ }
74
+ return result;
75
+ }
76
+ /** Split text into chunks with overlap. */
77
+ export function chunkText(content, options) {
78
+ const chunkSize = options?.chunkSize ?? DEFAULT_CHUNK_SIZE;
79
+ const overlap = options?.overlap ?? DEFAULT_OVERLAP;
80
+ if (!content || content.length === 0) {
81
+ return [];
82
+ }
83
+ const rawChunks = recursiveSplit(content, chunkSize, 0);
84
+ return applyOverlap(rawChunks, overlap);
85
+ }
86
+ /** Split parsed pages into page-aware chunks. */
87
+ export function chunkPages(pages, options) {
88
+ if (!pages || pages.length === 0) {
89
+ return [];
90
+ }
91
+ const chunks = [];
92
+ let globalIndex = 0;
93
+ for (const page of pages) {
94
+ const textChunks = chunkText(page.content, options);
95
+ for (const text of textChunks) {
96
+ chunks.push({
97
+ content: text,
98
+ chunkIndex: globalIndex,
99
+ pageNumber: page.pageNumber,
100
+ metadata: page.metadata,
101
+ });
102
+ globalIndex++;
103
+ }
104
+ }
105
+ return chunks;
106
+ }
107
+ //# sourceMappingURL=chunk-generator.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"chunk-generator.js","sourceRoot":"","sources":["../../src/document/chunk-generator.ts"],"names":[],"mappings":"AAEA,MAAM,kBAAkB,GAAG,IAAI,CAAC;AAChC,MAAM,eAAe,GAAG,GAAG,CAAC;AAE5B,+DAA+D;AAC/D,MAAM,UAAU,GAAG,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC;AAE7C;;;GAGG;AACH,SAAS,cAAc,CAAC,IAAY,EAAE,SAAiB,EAAE,cAAsB;IAC7E,yCAAyC;IACzC,IAAI,IAAI,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACvC,CAAC;IAED,6DAA6D;IAC7D,IAAI,cAAc,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC;QACxC,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;YAChD,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC;QAC5C,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,MAAM,SAAS,GAAG,UAAU,CAAC,cAAc,CAAC,CAAC;IAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IAEpC,yEAAyE;IACzE,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,cAAc,CAAC,IAAI,EAAE,SAAS,EAAE,cAAc,GAAG,CAAC,CAAC,CAAC;IAC7D,CAAC;IAED,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,YAAY,GAAG,EAAE,CAAC;IAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,SAAS,GAAG,YAAY,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,YAAY,GAAG,SAAS,GAAG,IAAI,CAAC;QAErF,IAAI,SAAS,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;YAClC,YAAY,GAAG,SAAS,CAAC;QAC3B,CAAC;aAAM,CAAC;YACN,wCAAwC;YACxC,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5B,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;YAC5B,CAAC;YAED,wDAAwD;YACxD,IAAI,IAAI,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;gBAC5B,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,EAAE,SAAS,EAAE,cAAc,GAAG,CAAC,CAAC,CAAC;gBACtE,yDAAyD;gBACzD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC9C,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;gBAC5B,CAAC;gBACD,mDAAmD;gBACnD,YAAY,GAAG,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YAC7E,CAAC;iBAAM,CAAC;gBACN,YAAY,GAAG,IAAI,CAAC;YACtB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;IAC5B,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;GAGG;AACH,SAAS,YAAY,CAAC,MAAgB,EAAE,OAAe;IACrD,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,IAAI,OAAO,IAAI,CAAC,EAAE,CAAC;QACvC,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,MAAM,MAAM,GAAa,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IACrC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAC3B,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,CAAC,CAAC;QACnE,MAAM,CAAC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IACvC,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,2CAA2C;AAC3C,MAAM,UAAU,SAAS,CAAC,OAAe,EAAE,OAAsB;IAC/D,MAAM,SAAS,GAAG,OAAO,EAAE,SAAS,IAAI,kBAAkB,CAAC;IAC3D,MAAM,OAAO,GAAG,OAAO,EAAE,OAAO,IAAI,eAAe,CAAC;IAEpD,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACrC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,SAAS,GAAG,cAAc,CAAC,OAAO,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;IACxD,OAAO,YAAY,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;AAC1C,CAAC;AAED,iDAAiD;AACjD,MAAM,UAAU,UAAU,CAAC,KAAmB,EAAE,OAAsB;IACpE,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,MAAM,GAAoB,EAAE,CAAC;IACnC,IAAI,WAAW,GAAG,CAAC,CAAC;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,UAAU,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QACpD,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,CAAC,IAAI,CAAC;gBACV,OAAO,EAAE,IAAI;gBACb,UAAU,EAAE,WAAW;gBACvB,UAAU,EAAE,IAAI,CAAC,UAAU;gBAC3B,QAAQ,EAAE,IAAI,CAAC,QAAQ;aACxB,CAAC,CAAC;YACH,WAAW,EAAE,CAAC;QAChB,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
@@ -0,0 +1,2 @@
1
+ export {};
2
+ //# sourceMappingURL=chunk-generator.test.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"chunk-generator.test.d.ts","sourceRoot":"","sources":["../../src/document/chunk-generator.test.ts"],"names":[],"mappings":""}
@@ -0,0 +1,166 @@
1
+ import { describe, it, expect } from 'vitest';
2
+ import { chunkText, chunkPages } from './chunk-generator.js';
3
+ describe('chunkText', () => {
4
+ it('returns empty array for empty input', () => {
5
+ expect(chunkText('')).toEqual([]);
6
+ });
7
+ it('returns empty array for undefined-like empty string', () => {
8
+ expect(chunkText('', { chunkSize: 100, overlap: 0 })).toEqual([]);
9
+ });
10
+ it('returns single chunk for short text', () => {
11
+ const result = chunkText('Hello, world!', { chunkSize: 100, overlap: 0 });
12
+ expect(result).toHaveLength(1);
13
+ expect(result[0]).toBe('Hello, world!');
14
+ });
15
+ it('splits at paragraph boundaries (double newline)', () => {
16
+ const text = 'Paragraph one.\n\nParagraph two.\n\nParagraph three.';
17
+ const result = chunkText(text, { chunkSize: 25, overlap: 0 });
18
+ // Each paragraph should be its own chunk
19
+ expect(result.length).toBeGreaterThanOrEqual(2);
20
+ expect(result[0]).toContain('Paragraph one.');
21
+ });
22
+ it('splits at single newline boundaries', () => {
23
+ const text = 'Line one.\nLine two.\nLine three.\nLine four.\nLine five.';
24
+ const result = chunkText(text, { chunkSize: 25, overlap: 0 });
25
+ expect(result.length).toBeGreaterThan(1);
26
+ // All original content should be preserved across chunks
27
+ const combined = result.join('');
28
+ expect(combined).toContain('Line one.');
29
+ expect(combined).toContain('Line five.');
30
+ });
31
+ it('splits at period boundaries', () => {
32
+ const text = 'First sentence. Second sentence. Third sentence. Fourth sentence.';
33
+ const result = chunkText(text, { chunkSize: 40, overlap: 0 });
34
+ expect(result.length).toBeGreaterThan(1);
35
+ });
36
+ it('splits at space boundaries', () => {
37
+ const text = 'word1 word2 word3 word4 word5 word6 word7 word8 word9 word10';
38
+ const result = chunkText(text, { chunkSize: 20, overlap: 0 });
39
+ expect(result.length).toBeGreaterThan(1);
40
+ for (const chunk of result) {
41
+ // Chunks should not exceed the limit by too much (spaces may cause slight variation)
42
+ expect(chunk.length).toBeLessThanOrEqual(25);
43
+ }
44
+ });
45
+ it('hard splits when no natural boundaries exist', () => {
46
+ const text = 'a'.repeat(50);
47
+ const result = chunkText(text, { chunkSize: 20, overlap: 0 });
48
+ expect(result.length).toBe(3); // 20 + 20 + 10
49
+ expect(result[0]).toBe('a'.repeat(20));
50
+ expect(result[1]).toBe('a'.repeat(20));
51
+ expect(result[2]).toBe('a'.repeat(10));
52
+ });
53
+ it('applies overlap correctly', () => {
54
+ const text = 'AAAAAAAAAA\n\nBBBBBBBBBB\n\nCCCCCCCCCC';
55
+ const result = chunkText(text, { chunkSize: 15, overlap: 5 });
56
+ expect(result.length).toBeGreaterThanOrEqual(2);
57
+ // Second chunk should start with last 5 chars of first chunk
58
+ if (result.length >= 2) {
59
+ const firstChunkEnd = result[0].slice(-5);
60
+ expect(result[1].startsWith(firstChunkEnd)).toBe(true);
61
+ }
62
+ });
63
+ it('uses default chunk size and overlap', () => {
64
+ // Create text that's just over 1000 chars (default chunkSize)
65
+ const text = 'word '.repeat(250); // 1250 chars
66
+ const result = chunkText(text);
67
+ expect(result.length).toBeGreaterThan(1);
68
+ });
69
+ it('handles text exactly at chunk size', () => {
70
+ const text = 'x'.repeat(100);
71
+ const result = chunkText(text, { chunkSize: 100, overlap: 0 });
72
+ expect(result).toHaveLength(1);
73
+ expect(result[0]).toBe(text);
74
+ });
75
+ it('preserves content integrity without overlap', () => {
76
+ const paragraphs = ['Alpha paragraph.', 'Beta paragraph.', 'Gamma paragraph.'];
77
+ const text = paragraphs.join('\n\n');
78
+ const result = chunkText(text, { chunkSize: 20, overlap: 0 });
79
+ // Every paragraph should appear in at least one chunk
80
+ for (const para of paragraphs) {
81
+ const found = result.some((chunk) => chunk.includes(para));
82
+ expect(found).toBe(true);
83
+ }
84
+ });
85
+ it('handles zero overlap', () => {
86
+ const text = 'Chunk A.\n\nChunk B.\n\nChunk C.';
87
+ const result = chunkText(text, { chunkSize: 15, overlap: 0 });
88
+ expect(result.length).toBeGreaterThan(1);
89
+ // With no overlap, second chunk should not start with content from first
90
+ });
91
+ });
92
+ describe('chunkPages', () => {
93
+ it('returns empty array for empty pages', () => {
94
+ expect(chunkPages([])).toEqual([]);
95
+ });
96
+ it('preserves page numbers in chunks', () => {
97
+ const pages = [
98
+ { pageNumber: 1, content: 'Page one content' },
99
+ { pageNumber: 2, content: 'Page two content' },
100
+ ];
101
+ const result = chunkPages(pages, { chunkSize: 1000, overlap: 0 });
102
+ expect(result).toHaveLength(2);
103
+ expect(result[0].pageNumber).toBe(1);
104
+ expect(result[0].content).toBe('Page one content');
105
+ expect(result[1].pageNumber).toBe(2);
106
+ expect(result[1].content).toBe('Page two content');
107
+ });
108
+ it('assigns sequential chunk indexes across pages', () => {
109
+ const pages = [
110
+ { pageNumber: 1, content: 'Short content' },
111
+ { pageNumber: 2, content: 'More content' },
112
+ { pageNumber: 3, content: 'Final content' },
113
+ ];
114
+ const result = chunkPages(pages, { chunkSize: 1000, overlap: 0 });
115
+ expect(result.map((c) => c.chunkIndex)).toEqual([0, 1, 2]);
116
+ });
117
+ it('splits long pages into multiple chunks', () => {
118
+ const pages = [
119
+ { pageNumber: 1, content: 'word '.repeat(300) }, // 1500 chars
120
+ ];
121
+ const result = chunkPages(pages, { chunkSize: 500, overlap: 0 });
122
+ expect(result.length).toBeGreaterThan(1);
123
+ // All chunks should reference page 1
124
+ for (const chunk of result) {
125
+ expect(chunk.pageNumber).toBe(1);
126
+ }
127
+ });
128
+ it('handles empty page content', () => {
129
+ const pages = [
130
+ { pageNumber: 1, content: '' },
131
+ { pageNumber: 2, content: 'Has content' },
132
+ ];
133
+ const result = chunkPages(pages, { chunkSize: 1000, overlap: 0 });
134
+ // Empty page produces no chunks
135
+ expect(result).toHaveLength(1);
136
+ expect(result[0].pageNumber).toBe(2);
137
+ expect(result[0].chunkIndex).toBe(0);
138
+ });
139
+ it('propagates page metadata to chunks', () => {
140
+ const pages = [
141
+ { pageNumber: 1, content: 'Content', metadata: { source: 'pdf', font: 'Arial' } },
142
+ ];
143
+ const result = chunkPages(pages, { chunkSize: 1000, overlap: 0 });
144
+ expect(result[0].metadata).toEqual({ source: 'pdf', font: 'Arial' });
145
+ });
146
+ it('handles multiple pages with varying sizes', () => {
147
+ const pages = [
148
+ { pageNumber: 1, content: 'Short' },
149
+ { pageNumber: 2, content: 'A much longer page. '.repeat(100) }, // 2000 chars
150
+ { pageNumber: 3, content: 'Also short' },
151
+ ];
152
+ const result = chunkPages(pages, { chunkSize: 500, overlap: 0 });
153
+ // Page 1: 1 chunk, page 2: multiple chunks, page 3: 1 chunk
154
+ const page1Chunks = result.filter((c) => c.pageNumber === 1);
155
+ const page2Chunks = result.filter((c) => c.pageNumber === 2);
156
+ const page3Chunks = result.filter((c) => c.pageNumber === 3);
157
+ expect(page1Chunks).toHaveLength(1);
158
+ expect(page2Chunks.length).toBeGreaterThan(1);
159
+ expect(page3Chunks).toHaveLength(1);
160
+ // Chunk indexes should be sequential across all pages
161
+ for (let i = 0; i < result.length; i++) {
162
+ expect(result[i].chunkIndex).toBe(i);
163
+ }
164
+ });
165
+ });
166
+ //# sourceMappingURL=chunk-generator.test.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"chunk-generator.test.js","sourceRoot":"","sources":["../../src/document/chunk-generator.test.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAE9C,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAE1D,QAAQ,CAAC,WAAW,EAAE,GAAG,EAAE;IACzB,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;QAC7C,MAAM,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACpC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,qDAAqD,EAAE,GAAG,EAAE;QAC7D,MAAM,CAAC,SAAS,CAAC,EAAE,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACpE,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;QAC7C,MAAM,MAAM,GAAG,SAAS,CAAC,eAAe,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAC1E,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;IAC1C,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,iDAAiD,EAAE,GAAG,EAAE;QACzD,MAAM,IAAI,GAAG,sDAAsD,CAAC;QACpE,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,yCAAyC;QACzC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAC;QAChD,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,gBAAgB,CAAC,CAAC;IAChD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;QAC7C,MAAM,IAAI,GAAG,2DAA2D,CAAC;QACzE,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QACzC,yDAAyD;QACzD,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACjC,MAAM,CAAC,QAAQ,CAAC,CAAC,SAAS,CAAC,WAAW,CAAC,CAAC;QACxC,MAAM,CAAC,QAAQ,CAAC,CAAC,SAAS,CAAC,YAAY,CAAC,CAAC;IAC3C,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,6BAA6B,EAAE,GAAG,EAAE;QACrC,MAAM,IAAI,GAAG,mEAAmE,CAAC;QACjF,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;IAC3C,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,4BAA4B,EAAE,GAAG,EAAE;QACpC,MAAM,IAAI,GAAG,8DAA8D,CAAC;QAC5E,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QACzC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,qFAAqF;YACrF,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,mBAAmB,CAAC,EAAE,CAAC,CAAC;QAC/C,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,8CAA8C,EAAE,GAAG,EAAE;QACtD,MAAM,IAAI,GAAG,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;QAC5B,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,eAAe;QAC9C,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC;QACvC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC;QACvC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC;IACzC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,2BAA2B,EAAE,GAAG,EAAE;QACnC,MAAM,IAAI,GAAG,wCAAwC,CAAC;QACtD,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAC;QAChD,6DAA6D;QAC7D,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;YACvB,MAAM,aAAa,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YAC1C,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACzD,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;QAC7C,8DAA8D;QAC9D,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,aAAa;QAC/C,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;QAE/B,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;IAC3C,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,oCAAoC,EAAE,GAAG,EAAE;QAC5C,MAAM,IAAI,GAAG,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAC7B,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE/D,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/B,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,6CAA6C,EAAE,GAAG,EAAE;QACrD,MAAM,UAAU,GAAG,CAAC,kBAAkB,EAAE,iBAAiB,EAAE,kBAAkB,CAAC,CAAC;QAC/E,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrC,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,sDAAsD;QACtD,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;YAC3D,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,sBAAsB,EAAE,GAAG,EAAE;QAC9B,MAAM,IAAI,GAAG,kCAAkC,CAAC;QAChD,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAE9D,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QACzC,yEAAyE;IAC3E,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEH,QAAQ,CAAC,YAAY,EAAE,GAAG,EAAE;IAC1B,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;QAC7C,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACrC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,kCAAkC,EAAE,GAAG,EAAE;QAC1C,MAAM,KAAK,GAAiB;YAC1B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,kBAAkB,EAAE;YAC9C,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,kBAAkB,EAAE;SAC/C,CAAC;QACF,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACrC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC;QACnD,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACrC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC;IACrD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,+CAA+C,EAAE,GAAG,EAAE;QACvD,MAAM,KAAK,GAAiB;YAC1B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,eAAe,EAAE;YAC3C,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,cAAc,EAAE;YAC1C,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,eAAe,EAAE;SAC5C,CAAC;QACF,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAC7D,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,wCAAwC,EAAE,GAAG,EAAE;QAChD,MAAM,KAAK,GAAiB;YAC1B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE,aAAa;SAC/D,CAAC;QACF,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAEjE,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QACzC,qCAAqC;QACrC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACnC,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,4BAA4B,EAAE,GAAG,EAAE;QACpC,MAAM,KAAK,GAAiB;YAC1B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,EAAE,EAAE;YAC9B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,aAAa,EAAE;SAC1C,CAAC;QACF,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAElE,gCAAgC;QAChC,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACrC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACvC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,oCAAoC,EAAE,GAAG,EAAE;QAC5C,MAAM,KAAK,GAAiB;YAC1B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,EAAE;SAClF,CAAC;QACF,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,OAAO,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,CAAC;IACvE,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,2CAA2C,EAAE,GAAG,EAAE;QACnD,MAAM,KAAK,GAAiB;YAC1B,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,OAAO,EAAE;YACnC,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,sBAAsB,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE,aAAa;YAC7E,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,YAAY,EAAE;SACzC,CAAC;QACF,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;QAEjE,4DAA4D;QAC5D,MAAM,WAAW,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,KAAK,CAAC,CAAC,CAAC;QAC7D,MAAM,WAAW,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,KAAK,CAAC,CAAC,CAAC;QAC7D,MAAM,WAAW,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,KAAK,CAAC,CAAC,CAAC;QAE7D,MAAM,CAAC,WAAW,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;QACpC,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QAC9C,MAAM,CAAC,WAAW,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;QAEpC,sDAAsD;QACtD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACvC,CAAC;IACH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
@@ -0,0 +1,27 @@
1
+ import type { ChunkOptions, DocumentChunk, DocumentParser, EmbeddingProvider, ParsedDocument, ParsedPage } from '@sprinterai/core';
2
+ /** Options for processing a document. */
3
+ export interface ProcessDocumentOptions {
4
+ documentId: string;
5
+ tenantId: string;
6
+ fileName: string;
7
+ mimeType: string;
8
+ fileData: Buffer | Uint8Array;
9
+ chunkOptions?: ChunkOptions;
10
+ embeddingProvider?: EmbeddingProvider;
11
+ }
12
+ /** Result of processing a document through the pipeline. */
13
+ export interface ProcessDocumentResult {
14
+ pages: ParsedPage[];
15
+ chunks: DocumentChunk[];
16
+ embeddings?: number[][];
17
+ metadata: ParsedDocument['metadata'];
18
+ }
19
+ /**
20
+ * Create a document processor with a registry of parsers.
21
+ * Falls back to the text parser for unrecognized MIME types.
22
+ */
23
+ export declare function createDocumentProcessor(parsers?: DocumentParser[]): {
24
+ process(options: ProcessDocumentOptions): Promise<ProcessDocumentResult>;
25
+ getParser(mimeType: string): DocumentParser | null;
26
+ };
27
+ //# sourceMappingURL=document-processor.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"document-processor.d.ts","sourceRoot":"","sources":["../../src/document/document-processor.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,YAAY,EACZ,aAAa,EACb,cAAc,EACd,iBAAiB,EACjB,cAAc,EACd,UAAU,EACX,MAAM,kBAAkB,CAAC;AAK1B,yCAAyC;AACzC,MAAM,WAAW,sBAAsB;IACrC,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,GAAG,UAAU,CAAC;IAC9B,YAAY,CAAC,EAAE,YAAY,CAAC;IAC5B,iBAAiB,CAAC,EAAE,iBAAiB,CAAC;CACvC;AAED,4DAA4D;AAC5D,MAAM,WAAW,qBAAqB;IACpC,KAAK,EAAE,UAAU,EAAE,CAAC;IACpB,MAAM,EAAE,aAAa,EAAE,CAAC;IACxB,UAAU,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;IACxB,QAAQ,EAAE,cAAc,CAAC,UAAU,CAAC,CAAC;CACtC;AAED;;;GAGG;AACH,wBAAgB,uBAAuB,CAAC,OAAO,CAAC,EAAE,cAAc,EAAE,GAAG;IACnE,OAAO,CAAC,OAAO,EAAE,sBAAsB,GAAG,OAAO,CAAC,qBAAqB,CAAC,CAAC;IACzE,SAAS,CAAC,QAAQ,EAAE,MAAM,GAAG,cAAc,GAAG,IAAI,CAAC;CACpD,CA6CA"}
@@ -0,0 +1,44 @@
1
+ import { chunkPages } from './chunk-generator.js';
2
+ import { textParser } from './parsers/index.js';
3
+ /**
4
+ * Create a document processor with a registry of parsers.
5
+ * Falls back to the text parser for unrecognized MIME types.
6
+ */
7
+ export function createDocumentProcessor(parsers) {
8
+ // Build MIME type -> parser lookup from registered parsers + built-in text parser
9
+ const allParsers = [...(parsers ?? []), textParser];
10
+ function getParser(mimeType) {
11
+ const normalized = mimeType.toLowerCase().split(';')[0].trim();
12
+ for (const parser of allParsers) {
13
+ if (parser.mimeTypes.includes(normalized)) {
14
+ return parser;
15
+ }
16
+ }
17
+ return null;
18
+ }
19
+ async function process(options) {
20
+ const { mimeType, fileData, chunkOptions, embeddingProvider } = options;
21
+ // 1. Parse: find a parser or fall back to text parser
22
+ const parser = getParser(mimeType) ?? textParser;
23
+ const parsed = await parser.parse(fileData instanceof Buffer ? fileData : Buffer.from(fileData));
24
+ // 2. Chunk: split parsed pages into chunks
25
+ const chunks = chunkPages(parsed.pages, chunkOptions);
26
+ // 3. Embed (optional): generate embeddings if provider is available
27
+ let embeddings;
28
+ if (embeddingProvider && chunks.length > 0) {
29
+ const texts = chunks.map((c) => c.content);
30
+ const result = await embeddingProvider.generateEmbeddings(texts);
31
+ if (result !== null) {
32
+ embeddings = result;
33
+ }
34
+ }
35
+ return {
36
+ pages: parsed.pages,
37
+ chunks,
38
+ embeddings,
39
+ metadata: parsed.metadata,
40
+ };
41
+ }
42
+ return { process, getParser };
43
+ }
44
+ //# sourceMappingURL=document-processor.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"document-processor.js","sourceRoot":"","sources":["../../src/document/document-processor.ts"],"names":[],"mappings":"AASA,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAC/C,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAqBvC;;;GAGG;AACH,MAAM,UAAU,uBAAuB,CAAC,OAA0B;IAIhE,kFAAkF;IAClF,MAAM,UAAU,GAAqB,CAAC,GAAG,CAAC,OAAO,IAAI,EAAE,CAAC,EAAE,UAAU,CAAC,CAAC;IAEtE,SAAS,SAAS,CAAC,QAAgB;QACjC,MAAM,UAAU,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC/D,KAAK,MAAM,MAAM,IAAI,UAAU,EAAE,CAAC;YAChC,IAAI,MAAM,CAAC,SAAS,CAAC,QAAQ,CAAC,UAAU,CAAC,EAAE,CAAC;gBAC1C,OAAO,MAAM,CAAC;YAChB,CAAC;QACH,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,KAAK,UAAU,OAAO,CAAC,OAA+B;QACpD,MAAM,EAAE,QAAQ,EAAE,QAAQ,EAAE,YAAY,EAAE,iBAAiB,EAAE,GAAG,OAAO,CAAC;QAExE,sDAAsD;QACtD,MAAM,MAAM,GAAG,SAAS,CAAC,QAAQ,CAAC,IAAI,UAAU,CAAC;QACjD,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,KAAK,CAC/B,QAAQ,YAAY,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAC9D,CAAC;QAEF,2CAA2C;QAC3C,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,CAAC,KAAK,EAAE,YAAY,CAAC,CAAC;QAEtD,oEAAoE;QACpE,IAAI,UAAkC,CAAC;QACvC,IAAI,iBAAiB,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC3C,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YAC3C,MAAM,MAAM,GAAG,MAAM,iBAAiB,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAC;YACjE,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;gBACpB,UAAU,GAAG,MAAM,CAAC;YACtB,CAAC;QACH,CAAC;QAED,OAAO;YACL,KAAK,EAAE,MAAM,CAAC,KAAK;YACnB,MAAM;YACN,UAAU;YACV,QAAQ,EAAE,MAAM,CAAC,QAAQ;SAC1B,CAAC;IACJ,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;AAChC,CAAC"}
@@ -0,0 +1,2 @@
1
+ export {};
2
+ //# sourceMappingURL=document-processor.test.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"document-processor.test.d.ts","sourceRoot":"","sources":["../../src/document/document-processor.test.ts"],"names":[],"mappings":""}
@@ -0,0 +1,197 @@
1
+ import { describe, it, expect, vi } from 'vitest';
2
+ import { createDocumentProcessor } from './document-processor.js';
3
+ describe('createDocumentProcessor', () => {
4
+ const processor = createDocumentProcessor();
5
+ describe('getParser', () => {
6
+ it('returns text parser for text/plain', () => {
7
+ const parser = processor.getParser('text/plain');
8
+ expect(parser).not.toBeNull();
9
+ expect(parser.mimeTypes).toContain('text/plain');
10
+ });
11
+ it('returns text parser for text/markdown', () => {
12
+ const parser = processor.getParser('text/markdown');
13
+ expect(parser).not.toBeNull();
14
+ });
15
+ it('returns text parser for application/json', () => {
16
+ const parser = processor.getParser('application/json');
17
+ expect(parser).not.toBeNull();
18
+ });
19
+ it('returns null for unsupported MIME types', () => {
20
+ const parser = processor.getParser('application/pdf');
21
+ expect(parser).toBeNull();
22
+ });
23
+ it('handles MIME type with charset parameter', () => {
24
+ const parser = processor.getParser('text/plain; charset=utf-8');
25
+ expect(parser).not.toBeNull();
26
+ });
27
+ it('is case-insensitive', () => {
28
+ const parser = processor.getParser('TEXT/PLAIN');
29
+ expect(parser).not.toBeNull();
30
+ });
31
+ });
32
+ describe('process', () => {
33
+ it('processes a text document end-to-end', async () => {
34
+ const result = await processor.process({
35
+ documentId: 'doc-1',
36
+ tenantId: 'tenant-1',
37
+ fileName: 'test.txt',
38
+ mimeType: 'text/plain',
39
+ fileData: Buffer.from('Hello, world!'),
40
+ });
41
+ expect(result.pages).toHaveLength(1);
42
+ expect(result.pages[0].content).toBe('Hello, world!');
43
+ expect(result.chunks.length).toBeGreaterThanOrEqual(1);
44
+ expect(result.chunks[0].content).toContain('Hello, world!');
45
+ expect(result.metadata.pageCount).toBe(1);
46
+ expect(result.embeddings).toBeUndefined();
47
+ });
48
+ it('processes Uint8Array input', async () => {
49
+ const encoder = new TextEncoder();
50
+ const result = await processor.process({
51
+ documentId: 'doc-2',
52
+ tenantId: 'tenant-1',
53
+ fileName: 'test.txt',
54
+ mimeType: 'text/plain',
55
+ fileData: encoder.encode('From Uint8Array'),
56
+ });
57
+ expect(result.pages[0].content).toBe('From Uint8Array');
58
+ });
59
+ it('falls back to text parser for unknown MIME types', async () => {
60
+ const result = await processor.process({
61
+ documentId: 'doc-3',
62
+ tenantId: 'tenant-1',
63
+ fileName: 'unknown.xyz',
64
+ mimeType: 'application/x-unknown',
65
+ fileData: Buffer.from('Unknown format but readable text'),
66
+ });
67
+ expect(result.pages).toHaveLength(1);
68
+ expect(result.pages[0].content).toBe('Unknown format but readable text');
69
+ });
70
+ it('reports correct page and chunk counts', async () => {
71
+ const longText = 'This is a sentence. '.repeat(100); // ~2000 chars
72
+ const result = await processor.process({
73
+ documentId: 'doc-4',
74
+ tenantId: 'tenant-1',
75
+ fileName: 'long.txt',
76
+ mimeType: 'text/plain',
77
+ fileData: Buffer.from(longText),
78
+ chunkOptions: { chunkSize: 500, overlap: 0 },
79
+ });
80
+ expect(result.pages).toHaveLength(1);
81
+ expect(result.chunks.length).toBeGreaterThan(1);
82
+ expect(result.metadata.pageCount).toBe(1);
83
+ });
84
+ it('generates embeddings when provider is given', async () => {
85
+ const mockProvider = {
86
+ dimensions: 384,
87
+ generateEmbeddings: vi.fn().mockResolvedValue([
88
+ [0.1, 0.2, 0.3],
89
+ [0.4, 0.5, 0.6],
90
+ ]),
91
+ };
92
+ const text = 'First paragraph.\n\nSecond paragraph.';
93
+ const result = await processor.process({
94
+ documentId: 'doc-5',
95
+ tenantId: 'tenant-1',
96
+ fileName: 'test.txt',
97
+ mimeType: 'text/plain',
98
+ fileData: Buffer.from(text),
99
+ chunkOptions: { chunkSize: 20, overlap: 0 },
100
+ embeddingProvider: mockProvider,
101
+ });
102
+ expect(result.embeddings).toBeDefined();
103
+ expect(mockProvider.generateEmbeddings).toHaveBeenCalledTimes(1);
104
+ });
105
+ it('returns undefined embeddings when provider returns null', async () => {
106
+ const mockProvider = {
107
+ dimensions: 384,
108
+ generateEmbeddings: vi.fn().mockResolvedValue(null),
109
+ };
110
+ const result = await processor.process({
111
+ documentId: 'doc-6',
112
+ tenantId: 'tenant-1',
113
+ fileName: 'test.txt',
114
+ mimeType: 'text/plain',
115
+ fileData: Buffer.from('Some text'),
116
+ embeddingProvider: mockProvider,
117
+ });
118
+ expect(result.embeddings).toBeUndefined();
119
+ });
120
+ it('skips embedding for empty content', async () => {
121
+ const mockProvider = {
122
+ dimensions: 384,
123
+ generateEmbeddings: vi.fn().mockResolvedValue([]),
124
+ };
125
+ const result = await processor.process({
126
+ documentId: 'doc-7',
127
+ tenantId: 'tenant-1',
128
+ fileName: 'empty.txt',
129
+ mimeType: 'text/plain',
130
+ fileData: Buffer.from(''),
131
+ embeddingProvider: mockProvider,
132
+ });
133
+ // Empty content produces no chunks, so embedding should not be called
134
+ expect(mockProvider.generateEmbeddings).not.toHaveBeenCalled();
135
+ expect(result.embeddings).toBeUndefined();
136
+ });
137
+ it('respects chunk options', async () => {
138
+ const text = 'word '.repeat(200); // 1000 chars
139
+ const result = await processor.process({
140
+ documentId: 'doc-8',
141
+ tenantId: 'tenant-1',
142
+ fileName: 'test.txt',
143
+ mimeType: 'text/plain',
144
+ fileData: Buffer.from(text),
145
+ chunkOptions: { chunkSize: 200, overlap: 50 },
146
+ });
147
+ expect(result.chunks.length).toBeGreaterThan(1);
148
+ });
149
+ });
150
+ describe('custom parsers', () => {
151
+ it('uses custom parser when registered', async () => {
152
+ const customParser = {
153
+ mimeTypes: ['application/pdf'],
154
+ parse: vi.fn().mockResolvedValue({
155
+ pages: [
156
+ { pageNumber: 1, content: 'PDF page 1' },
157
+ { pageNumber: 2, content: 'PDF page 2' },
158
+ ],
159
+ fullText: 'PDF page 1\nPDF page 2',
160
+ metadata: { pageCount: 2, title: 'Test PDF' },
161
+ }),
162
+ };
163
+ const customProcessor = createDocumentProcessor([customParser]);
164
+ const result = await customProcessor.process({
165
+ documentId: 'doc-pdf',
166
+ tenantId: 'tenant-1',
167
+ fileName: 'test.pdf',
168
+ mimeType: 'application/pdf',
169
+ fileData: Buffer.from('fake pdf data'),
170
+ });
171
+ expect(customParser.parse).toHaveBeenCalledTimes(1);
172
+ expect(result.pages).toHaveLength(2);
173
+ expect(result.metadata.pageCount).toBe(2);
174
+ expect(result.metadata.title).toBe('Test PDF');
175
+ });
176
+ it('custom parser takes precedence over built-in', async () => {
177
+ const customTextParser = {
178
+ mimeTypes: ['text/plain'],
179
+ parse: vi.fn().mockResolvedValue({
180
+ pages: [{ pageNumber: 1, content: 'CUSTOM' }],
181
+ fullText: 'CUSTOM',
182
+ metadata: { pageCount: 1 },
183
+ }),
184
+ };
185
+ const customProcessor = createDocumentProcessor([customTextParser]);
186
+ const result = await customProcessor.process({
187
+ documentId: 'doc-custom',
188
+ tenantId: 'tenant-1',
189
+ fileName: 'test.txt',
190
+ mimeType: 'text/plain',
191
+ fileData: Buffer.from('original'),
192
+ });
193
+ expect(result.pages[0].content).toBe('CUSTOM');
194
+ });
195
+ });
196
+ });
197
+ //# sourceMappingURL=document-processor.test.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"document-processor.test.js","sourceRoot":"","sources":["../../src/document/document-processor.test.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,QAAQ,CAAC;AAElD,OAAO,EAAE,uBAAuB,EAAE,MAAM,sBAAsB,CAAC;AAE/D,QAAQ,CAAC,yBAAyB,EAAE,GAAG,EAAE;IACvC,MAAM,SAAS,GAAG,uBAAuB,EAAE,CAAC;IAE5C,QAAQ,CAAC,WAAW,EAAE,GAAG,EAAE;QACzB,EAAE,CAAC,oCAAoC,EAAE,GAAG,EAAE;YAC5C,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,YAAY,CAAC,CAAC;YACjD,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC;YAC9B,MAAM,CAAC,MAAO,CAAC,SAAS,CAAC,CAAC,SAAS,CAAC,YAAY,CAAC,CAAC;QACpD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,uCAAuC,EAAE,GAAG,EAAE;YAC/C,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,eAAe,CAAC,CAAC;YACpD,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC;QAChC,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,0CAA0C,EAAE,GAAG,EAAE;YAClD,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,kBAAkB,CAAC,CAAC;YACvD,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC;QAChC,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,yCAAyC,EAAE,GAAG,EAAE;YACjD,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,iBAAiB,CAAC,CAAC;YACtD,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,CAAC;QAC5B,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,0CAA0C,EAAE,GAAG,EAAE;YAClD,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,2BAA2B,CAAC,CAAC;YAChE,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC;QAChC,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,qBAAqB,EAAE,GAAG,EAAE;YAC7B,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,YAAY,CAAC,CAAC;YACjD,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC;QAChC,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,SAAS,EAAE,GAAG,EAAE;QACvB,EAAE,CAAC,sCAAsC,EAAE,KAAK,IAAI,EAAE;YACpD,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,eAAe,CAAC;aACvC,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;YACtD,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAC;YACvD,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,SAAS,CAAC,eAAe,CAAC,CAAC;YAC5D,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC1C,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,aAAa,EAAE,CAAC;QAC5C,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,4BAA4B,EAAE,KAAK,IAAI,EAAE;YAC1C,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAC;YAClC,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,OAAO,CAAC,MAAM,CAAC,iBAAiB,CAAC;aAC5C,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,iBAAiB,CAAC,CAAC;QAC1D,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,kDAAkD,EAAE,KAAK,IAAI,EAAE;YAChE,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,aAAa;gBACvB,QAAQ,EAAE,uBAAuB;gBACjC,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,kCAAkC,CAAC;aAC1D,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,kCAAkC,CAAC,CAAC;QAC3E,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,uCAAuC,EAAE,KAAK,IAAI,EAAE;YACrD,MAAM,QAAQ,GAAG,sBAAsB,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,cAAc;YACnE,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC;gBAC/B,YAAY,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,EAAE;aAC7C,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;YAChD,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC5C,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,6CAA6C,EAAE,KAAK,IAAI,EAAE;YAC3D,MAAM,YAAY,GAAsB;gBACtC,UAAU,EAAE,GAAG;gBACf,kBAAkB,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC;oBAC5C,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC;oBACf,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC;iBAChB,CAAC;aACH,CAAC;YAEF,MAAM,IAAI,GAAG,uCAAuC,CAAC;YACrD,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC;gBAC3B,YAAY,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE;gBAC3C,iBAAiB,EAAE,YAAY;aAChC,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,WAAW,EAAE,CAAC;YACxC,MAAM,CAAC,YAAY,CAAC,kBAAkB,CAAC,CAAC,qBAAqB,CAAC,CAAC,CAAC,CAAC;QACnE,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,yDAAyD,EAAE,KAAK,IAAI,EAAE;YACvE,MAAM,YAAY,GAAsB;gBACtC,UAAU,EAAE,GAAG;gBACf,kBAAkB,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC;aACpD,CAAC;YAEF,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC;gBAClC,iBAAiB,EAAE,YAAY;aAChC,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,aAAa,EAAE,CAAC;QAC5C,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,mCAAmC,EAAE,KAAK,IAAI,EAAE;YACjD,MAAM,YAAY,GAAsB;gBACtC,UAAU,EAAE,GAAG;gBACf,kBAAkB,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,EAAE,CAAC;aAClD,CAAC;YAEF,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,WAAW;gBACrB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;gBACzB,iBAAiB,EAAE,YAAY;aAChC,CAAC,CAAC;YAEH,sEAAsE;YACtE,MAAM,CAAC,YAAY,CAAC,kBAAkB,CAAC,CAAC,GAAG,CAAC,gBAAgB,EAAE,CAAC;YAC/D,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,aAAa,EAAE,CAAC;QAC5C,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,wBAAwB,EAAE,KAAK,IAAI,EAAE;YACtC,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,aAAa;YAC/C,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC;gBACrC,UAAU,EAAE,OAAO;gBACnB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC;gBAC3B,YAAY,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,EAAE,EAAE,EAAE;aAC9C,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QAClD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,gBAAgB,EAAE,GAAG,EAAE;QAC9B,EAAE,CAAC,oCAAoC,EAAE,KAAK,IAAI,EAAE;YAClD,MAAM,YAAY,GAAmB;gBACnC,SAAS,EAAE,CAAC,iBAAiB,CAAC;gBAC9B,KAAK,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC;oBAC/B,KAAK,EAAE;wBACL,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,YAAY,EAAE;wBACxC,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,YAAY,EAAE;qBACzC;oBACD,QAAQ,EAAE,wBAAwB;oBAClC,QAAQ,EAAE,EAAE,SAAS,EAAE,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE;iBAC9C,CAAC;aACH,CAAC;YAEF,MAAM,eAAe,GAAG,uBAAuB,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC;YAEhE,MAAM,MAAM,GAAG,MAAM,eAAe,CAAC,OAAO,CAAC;gBAC3C,UAAU,EAAE,SAAS;gBACrB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,iBAAiB;gBAC3B,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,eAAe,CAAC;aACvC,CAAC,CAAC;YAEH,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,qBAAqB,CAAC,CAAC,CAAC,CAAC;YACpD,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC1C,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;QACjD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,8CAA8C,EAAE,KAAK,IAAI,EAAE;YAC5D,MAAM,gBAAgB,GAAmB;gBACvC,SAAS,EAAE,CAAC,YAAY,CAAC;gBACzB,KAAK,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC;oBAC/B,KAAK,EAAE,CAAC,EAAE,UAAU,EAAE,CAAC,EAAE,OAAO,EAAE,QAAQ,EAAE,CAAC;oBAC7C,QAAQ,EAAE,QAAQ;oBAClB,QAAQ,EAAE,EAAE,SAAS,EAAE,CAAC,EAAE;iBAC3B,CAAC;aACH,CAAC;YAEF,MAAM,eAAe,GAAG,uBAAuB,CAAC,CAAC,gBAAgB,CAAC,CAAC,CAAC;YAEpE,MAAM,MAAM,GAAG,MAAM,eAAe,CAAC,OAAO,CAAC;gBAC3C,UAAU,EAAE,YAAY;gBACxB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,UAAU;gBACpB,QAAQ,EAAE,YAAY;gBACtB,QAAQ,EAAE,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC;aAClC,CAAC,CAAC;YAEH,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACjD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
@@ -0,0 +1,5 @@
1
+ export { chunkText, chunkPages } from './chunk-generator.js';
2
+ export { createDocumentProcessor } from './document-processor.js';
3
+ export type { ProcessDocumentOptions, ProcessDocumentResult } from './document-processor.js';
4
+ export { textParser } from './parsers/index.js';
5
+ //# sourceMappingURL=index.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/document/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAC1D,OAAO,EAAE,uBAAuB,EAAE,MAAM,sBAAsB,CAAC;AAC/D,YAAY,EAAE,sBAAsB,EAAE,qBAAqB,EAAE,MAAM,sBAAsB,CAAC;AAC1F,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC"}
@@ -0,0 +1,4 @@
1
+ export { chunkText, chunkPages } from './chunk-generator.js';
2
+ export { createDocumentProcessor } from './document-processor.js';
3
+ export { textParser } from './parsers/index.js';
4
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/document/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAC1D,OAAO,EAAE,uBAAuB,EAAE,MAAM,sBAAsB,CAAC;AAE/D,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC"}
@@ -0,0 +1,2 @@
1
+ export { textParser } from './text-parser.js';
2
+ //# sourceMappingURL=index.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/document/parsers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,eAAe,CAAC"}
@@ -0,0 +1,2 @@
1
+ export { textParser } from './text-parser.js';
2
+ //# sourceMappingURL=index.js.map