@wanshi-kg/wanshi 0.1.0

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 (443) hide show
  1. package/LICENSE +21 -0
  2. package/README.md +458 -0
  3. package/dist/__tests__/helpers.js +27 -0
  4. package/dist/__tests__/helpers.js.map +1 -0
  5. package/dist/cli/commands/export.command.js +99 -0
  6. package/dist/cli/commands/export.command.js.map +1 -0
  7. package/dist/cli/commands/index.js +22 -0
  8. package/dist/cli/commands/index.js.map +1 -0
  9. package/dist/cli/commands/inspectMerges.command.js +84 -0
  10. package/dist/cli/commands/inspectMerges.command.js.map +1 -0
  11. package/dist/cli/commands/metrics.command.js +196 -0
  12. package/dist/cli/commands/metrics.command.js.map +1 -0
  13. package/dist/cli/commands/process.command.js +82 -0
  14. package/dist/cli/commands/process.command.js.map +1 -0
  15. package/dist/cli/commands/watch.command.js +91 -0
  16. package/dist/cli/commands/watch.command.js.map +1 -0
  17. package/dist/cli/index.js +269 -0
  18. package/dist/cli/index.js.map +1 -0
  19. package/dist/cli/optionsToConfig.js +160 -0
  20. package/dist/cli/optionsToConfig.js.map +1 -0
  21. package/dist/config/index.js +59 -0
  22. package/dist/config/index.js.map +1 -0
  23. package/dist/config/legacyHints.js +113 -0
  24. package/dist/config/legacyHints.js.map +1 -0
  25. package/dist/config/schema.js +803 -0
  26. package/dist/config/schema.js.map +1 -0
  27. package/dist/config/ui.js +221 -0
  28. package/dist/config/ui.js.map +1 -0
  29. package/dist/core/DirectoryProcessor.js +725 -0
  30. package/dist/core/DirectoryProcessor.js.map +1 -0
  31. package/dist/core/adapters/IStructuredAdapter.js +3 -0
  32. package/dist/core/adapters/IStructuredAdapter.js.map +1 -0
  33. package/dist/core/adapters/SqliteAdapter.js +267 -0
  34. package/dist/core/adapters/SqliteAdapter.js.map +1 -0
  35. package/dist/core/adapters/StructuredAdapterRegistry.js +31 -0
  36. package/dist/core/adapters/StructuredAdapterRegistry.js.map +1 -0
  37. package/dist/core/adapters/index.js +20 -0
  38. package/dist/core/adapters/index.js.map +1 -0
  39. package/dist/core/checkpoint/CheckpointService.js +188 -0
  40. package/dist/core/checkpoint/CheckpointService.js.map +1 -0
  41. package/dist/core/checkpoint/index.js +18 -0
  42. package/dist/core/checkpoint/index.js.map +1 -0
  43. package/dist/core/corpus/CorpusAnalyzer.js +266 -0
  44. package/dist/core/corpus/CorpusAnalyzer.js.map +1 -0
  45. package/dist/core/corpus/CorpusProfileStore.js +92 -0
  46. package/dist/core/corpus/CorpusProfileStore.js.map +1 -0
  47. package/dist/core/corpus/index.js +21 -0
  48. package/dist/core/corpus/index.js.map +1 -0
  49. package/dist/core/corpus/normalizeGlossary.js +60 -0
  50. package/dist/core/corpus/normalizeGlossary.js.map +1 -0
  51. package/dist/core/corpus/relPath.js +52 -0
  52. package/dist/core/corpus/relPath.js.map +1 -0
  53. package/dist/core/corpus/termFrequency.js +86 -0
  54. package/dist/core/corpus/termFrequency.js.map +1 -0
  55. package/dist/core/cost/CostMeter.js +235 -0
  56. package/dist/core/cost/CostMeter.js.map +1 -0
  57. package/dist/core/cost/index.js +19 -0
  58. package/dist/core/cost/index.js.map +1 -0
  59. package/dist/core/cost/prices.js +38 -0
  60. package/dist/core/cost/prices.js.map +1 -0
  61. package/dist/core/cv/ObjectDetectionService.js +119 -0
  62. package/dist/core/cv/ObjectDetectionService.js.map +1 -0
  63. package/dist/core/di/ContainerFactory.js +670 -0
  64. package/dist/core/di/ContainerFactory.js.map +1 -0
  65. package/dist/core/di/DIContainer.js +103 -0
  66. package/dist/core/di/DIContainer.js.map +1 -0
  67. package/dist/core/di/index.js +19 -0
  68. package/dist/core/di/index.js.map +1 -0
  69. package/dist/core/errors/CustomErrors.js +342 -0
  70. package/dist/core/errors/CustomErrors.js.map +1 -0
  71. package/dist/core/errors/index.js +18 -0
  72. package/dist/core/errors/index.js.map +1 -0
  73. package/dist/core/export/KnowledgeGraphExportService.js +56 -0
  74. package/dist/core/export/KnowledgeGraphExportService.js.map +1 -0
  75. package/dist/core/export/index.js +19 -0
  76. package/dist/core/export/index.js.map +1 -0
  77. package/dist/core/export/strategies/GraphitiExportStrategy.js +115 -0
  78. package/dist/core/export/strategies/GraphitiExportStrategy.js.map +1 -0
  79. package/dist/core/export/strategies/GraphvizDotExportStrategy.js +331 -0
  80. package/dist/core/export/strategies/GraphvizDotExportStrategy.js.map +1 -0
  81. package/dist/core/export/strategies/IExportStrategy.js +3 -0
  82. package/dist/core/export/strategies/IExportStrategy.js.map +1 -0
  83. package/dist/core/export/strategies/JsonExportStrategy.js +19 -0
  84. package/dist/core/export/strategies/JsonExportStrategy.js.map +1 -0
  85. package/dist/core/export/strategies/JsonlExportStrategy.js +69 -0
  86. package/dist/core/export/strategies/JsonlExportStrategy.js.map +1 -0
  87. package/dist/core/export/strategies/KblamExportStrategy.js +36 -0
  88. package/dist/core/export/strategies/KblamExportStrategy.js.map +1 -0
  89. package/dist/core/export/strategies/LoraExportStrategy.js +46 -0
  90. package/dist/core/export/strategies/LoraExportStrategy.js.map +1 -0
  91. package/dist/core/export/strategies/McpExportStrategy.js +67 -0
  92. package/dist/core/export/strategies/McpExportStrategy.js.map +1 -0
  93. package/dist/core/export/strategies/index.js +25 -0
  94. package/dist/core/export/strategies/index.js.map +1 -0
  95. package/dist/core/export/strategies/kbTriples.js +60 -0
  96. package/dist/core/export/strategies/kbTriples.js.map +1 -0
  97. package/dist/core/index.js +22 -0
  98. package/dist/core/index.js.map +1 -0
  99. package/dist/core/knowledge/KnowledgeGraphBuilder.js +627 -0
  100. package/dist/core/knowledge/KnowledgeGraphBuilder.js.map +1 -0
  101. package/dist/core/knowledge/MergeRecord.js +3 -0
  102. package/dist/core/knowledge/MergeRecord.js.map +1 -0
  103. package/dist/core/knowledge/canon/Canonicalizer.js +414 -0
  104. package/dist/core/knowledge/canon/Canonicalizer.js.map +1 -0
  105. package/dist/core/knowledge/canon/index.js +18 -0
  106. package/dist/core/knowledge/canon/index.js.map +1 -0
  107. package/dist/core/knowledge/contradiction/HeuristicContradictionChecker.js +92 -0
  108. package/dist/core/knowledge/contradiction/HeuristicContradictionChecker.js.map +1 -0
  109. package/dist/core/knowledge/contradiction/LlmContradictionChecker.js +52 -0
  110. package/dist/core/knowledge/contradiction/LlmContradictionChecker.js.map +1 -0
  111. package/dist/core/knowledge/contradiction/index.js +19 -0
  112. package/dist/core/knowledge/contradiction/index.js.map +1 -0
  113. package/dist/core/knowledge/grounding/KeywordGroundingChecker.js +33 -0
  114. package/dist/core/knowledge/grounding/KeywordGroundingChecker.js.map +1 -0
  115. package/dist/core/knowledge/grounding/MiniCheckGroundingChecker.js +82 -0
  116. package/dist/core/knowledge/grounding/MiniCheckGroundingChecker.js.map +1 -0
  117. package/dist/core/knowledge/grounding/index.js +20 -0
  118. package/dist/core/knowledge/grounding/index.js.map +1 -0
  119. package/dist/core/knowledge/grounding/verbalize.js +38 -0
  120. package/dist/core/knowledge/grounding/verbalize.js.map +1 -0
  121. package/dist/core/knowledge/images/imageMetaGraph.js +136 -0
  122. package/dist/core/knowledge/images/imageMetaGraph.js.map +1 -0
  123. package/dist/core/knowledge/index.js +20 -0
  124. package/dist/core/knowledge/index.js.map +1 -0
  125. package/dist/core/knowledge/merging/KnowledgeMerger.js +624 -0
  126. package/dist/core/knowledge/merging/KnowledgeMerger.js.map +1 -0
  127. package/dist/core/knowledge/references/ReferenceResolver.js +184 -0
  128. package/dist/core/knowledge/references/ReferenceResolver.js.map +1 -0
  129. package/dist/core/knowledge/references/citations/CitationEvidenceProcessor.js +401 -0
  130. package/dist/core/knowledge/references/citations/CitationEvidenceProcessor.js.map +1 -0
  131. package/dist/core/knowledge/references/citations/CitationResolver.js +95 -0
  132. package/dist/core/knowledge/references/citations/CitationResolver.js.map +1 -0
  133. package/dist/core/knowledge/references/citations/GrobidClient.js +143 -0
  134. package/dist/core/knowledge/references/citations/GrobidClient.js.map +1 -0
  135. package/dist/core/knowledge/references/citations/TitleIdResolver.js +101 -0
  136. package/dist/core/knowledge/references/citations/TitleIdResolver.js.map +1 -0
  137. package/dist/core/knowledge/references/web/FetchCacheService.js +114 -0
  138. package/dist/core/knowledge/references/web/FetchCacheService.js.map +1 -0
  139. package/dist/core/knowledge/references/web/GatedFetcher.js +228 -0
  140. package/dist/core/knowledge/references/web/GatedFetcher.js.map +1 -0
  141. package/dist/core/knowledge/references/web/WebReferenceProcessor.js +164 -0
  142. package/dist/core/knowledge/references/web/WebReferenceProcessor.js.map +1 -0
  143. package/dist/core/knowledge/search/KnowledgeGraphSearch.js +261 -0
  144. package/dist/core/knowledge/search/KnowledgeGraphSearch.js.map +1 -0
  145. package/dist/core/knowledge/vocabulary.js +162 -0
  146. package/dist/core/knowledge/vocabulary.js.map +1 -0
  147. package/dist/core/llm/EmbeddingService.js +113 -0
  148. package/dist/core/llm/EmbeddingService.js.map +1 -0
  149. package/dist/core/llm/OllamaService.js +146 -0
  150. package/dist/core/llm/OllamaService.js.map +1 -0
  151. package/dist/core/llm/OpenAICompatibleService.js +190 -0
  152. package/dist/core/llm/OpenAICompatibleService.js.map +1 -0
  153. package/dist/core/llm/OpenAIEmbeddingService.js +129 -0
  154. package/dist/core/llm/OpenAIEmbeddingService.js.map +1 -0
  155. package/dist/core/llm/embeddingUtils.js +25 -0
  156. package/dist/core/llm/embeddingUtils.js.map +1 -0
  157. package/dist/core/llm/index.js +23 -0
  158. package/dist/core/llm/index.js.map +1 -0
  159. package/dist/core/llm/prompts/PromptManager.js +388 -0
  160. package/dist/core/llm/prompts/PromptManager.js.map +1 -0
  161. package/dist/core/llm/prompts/PromptTemplateEngine.js +257 -0
  162. package/dist/core/llm/prompts/PromptTemplateEngine.js.map +1 -0
  163. package/dist/core/llm/prompts/templates/partials/examples/EXAMPLE_STYLE_GUIDE.md +84 -0
  164. package/dist/core/llm/prompts/templates/partials/examples/article.md +187 -0
  165. package/dist/core/llm/prompts/templates/partials/examples/code.md +229 -0
  166. package/dist/core/llm/prompts/templates/partials/examples/communication.md +205 -0
  167. package/dist/core/llm/prompts/templates/partials/examples/documentation.md +262 -0
  168. package/dist/core/llm/prompts/templates/partials/examples/financial.md +157 -0
  169. package/dist/core/llm/prompts/templates/partials/examples/legal.md +153 -0
  170. package/dist/core/llm/prompts/templates/partials/examples/logs.md +127 -0
  171. package/dist/core/llm/prompts/templates/partials/examples/medical.md +218 -0
  172. package/dist/core/llm/prompts/templates/partials/examples/notes.md +201 -0
  173. package/dist/core/llm/prompts/templates/partials/examples/research.md +208 -0
  174. package/dist/core/llm/prompts/templates/partials/examples/tabular.md +178 -0
  175. package/dist/core/llm/prompts/templates/partials/examples/transcript.md +204 -0
  176. package/dist/core/llm/prompts/templates/partials/retrieved-context.hbs +18 -0
  177. package/dist/core/llm/prompts/templates/v1/system.hbs +371 -0
  178. package/dist/core/llm/prompts/templates/v1/user.hbs +20 -0
  179. package/dist/core/llm/prompts/templates/v2/system.hbs +573 -0
  180. package/dist/core/llm/prompts/templates/v2/user.hbs +20 -0
  181. package/dist/core/llm/prompts/templates/v3/system.hbs +861 -0
  182. package/dist/core/llm/prompts/templates/v3/user.hbs +16 -0
  183. package/dist/core/llm/prompts/templates/v4/system.hbs +800 -0
  184. package/dist/core/llm/prompts/templates/v4/user.hbs +40 -0
  185. package/dist/core/llm/prompts/templates/v4.5/system.hbs +71 -0
  186. package/dist/core/llm/prompts/templates/v4.5/user.hbs +46 -0
  187. package/dist/core/llm/prompts/templates/v5/glossary/system.hbs +40 -0
  188. package/dist/core/llm/prompts/templates/v5/glossary/user.hbs +11 -0
  189. package/dist/core/llm/prompts/templates/v5/system.hbs +163 -0
  190. package/dist/core/llm/prompts/templates/v5/user.hbs +55 -0
  191. package/dist/core/pipeline/GroundingTransform.js +52 -0
  192. package/dist/core/pipeline/GroundingTransform.js.map +1 -0
  193. package/dist/core/pipeline/PipelineRunner.js +51 -0
  194. package/dist/core/pipeline/PipelineRunner.js.map +1 -0
  195. package/dist/core/pipeline/RelationFilterTransform.js +72 -0
  196. package/dist/core/pipeline/RelationFilterTransform.js.map +1 -0
  197. package/dist/core/pipeline/index.js +20 -0
  198. package/dist/core/pipeline/index.js.map +1 -0
  199. package/dist/core/processor/FileProcessor.js +184 -0
  200. package/dist/core/processor/FileProcessor.js.map +1 -0
  201. package/dist/core/processor/ProcessedRegistry.js +38 -0
  202. package/dist/core/processor/ProcessedRegistry.js.map +1 -0
  203. package/dist/core/processor/ast/AstSeedService.js +0 -0
  204. package/dist/core/processor/ast/AstSeedService.js.map +1 -0
  205. package/dist/core/processor/ast/AstSymbolStore.js +110 -0
  206. package/dist/core/processor/ast/AstSymbolStore.js.map +1 -0
  207. package/dist/core/processor/ast/index.js +19 -0
  208. package/dist/core/processor/ast/index.js.map +1 -0
  209. package/dist/core/processor/chunking/TextChunker.js +98 -0
  210. package/dist/core/processor/chunking/TextChunker.js.map +1 -0
  211. package/dist/core/processor/chunking/index.js +18 -0
  212. package/dist/core/processor/chunking/index.js.map +1 -0
  213. package/dist/core/processor/classifier/CONTENT_CLASSES.js +294 -0
  214. package/dist/core/processor/classifier/CONTENT_CLASSES.js.map +1 -0
  215. package/dist/core/processor/classifier/CascadeContentClassifier.js +107 -0
  216. package/dist/core/processor/classifier/CascadeContentClassifier.js.map +1 -0
  217. package/dist/core/processor/classifier/HeuristicContentClassifier.js +113 -0
  218. package/dist/core/processor/classifier/HeuristicContentClassifier.js.map +1 -0
  219. package/dist/core/processor/classifier/IContentTypeClassifier.js +3 -0
  220. package/dist/core/processor/classifier/IContentTypeClassifier.js.map +1 -0
  221. package/dist/core/processor/classifier/LlmContentClassifier.js +107 -0
  222. package/dist/core/processor/classifier/LlmContentClassifier.js.map +1 -0
  223. package/dist/core/processor/classifier/NER_DOMAIN_EXAMPLES.js +498 -0
  224. package/dist/core/processor/classifier/NER_DOMAIN_EXAMPLES.js.map +1 -0
  225. package/dist/core/processor/classifier/index.js +21 -0
  226. package/dist/core/processor/classifier/index.js.map +1 -0
  227. package/dist/core/processor/classifier/mergeClassifications.js +32 -0
  228. package/dist/core/processor/classifier/mergeClassifications.js.map +1 -0
  229. package/dist/core/processor/index.js +20 -0
  230. package/dist/core/processor/index.js.map +1 -0
  231. package/dist/core/processor/readers/AudioReader.js +462 -0
  232. package/dist/core/processor/readers/AudioReader.js.map +1 -0
  233. package/dist/core/processor/readers/BinaryReader.js +90 -0
  234. package/dist/core/processor/readers/BinaryReader.js.map +1 -0
  235. package/dist/core/processor/readers/ChandraPdfReader.js +187 -0
  236. package/dist/core/processor/readers/ChandraPdfReader.js.map +1 -0
  237. package/dist/core/processor/readers/ChatExportReader.js +365 -0
  238. package/dist/core/processor/readers/ChatExportReader.js.map +1 -0
  239. package/dist/core/processor/readers/DoclingReader.js +445 -0
  240. package/dist/core/processor/readers/DoclingReader.js.map +1 -0
  241. package/dist/core/processor/readers/EmailReader.js +259 -0
  242. package/dist/core/processor/readers/EmailReader.js.map +1 -0
  243. package/dist/core/processor/readers/EpubReader.js +175 -0
  244. package/dist/core/processor/readers/EpubReader.js.map +1 -0
  245. package/dist/core/processor/readers/FileReader.js +90 -0
  246. package/dist/core/processor/readers/FileReader.js.map +1 -0
  247. package/dist/core/processor/readers/FileReaderFactory.js +49 -0
  248. package/dist/core/processor/readers/FileReaderFactory.js.map +1 -0
  249. package/dist/core/processor/readers/HtmlReader.js +371 -0
  250. package/dist/core/processor/readers/HtmlReader.js.map +1 -0
  251. package/dist/core/processor/readers/ImageReader.js +162 -0
  252. package/dist/core/processor/readers/ImageReader.js.map +1 -0
  253. package/dist/core/processor/readers/JsonFileReader.js +232 -0
  254. package/dist/core/processor/readers/JsonFileReader.js.map +1 -0
  255. package/dist/core/processor/readers/JupyterReader.js +178 -0
  256. package/dist/core/processor/readers/JupyterReader.js.map +1 -0
  257. package/dist/core/processor/readers/LatexReader.js +176 -0
  258. package/dist/core/processor/readers/LatexReader.js.map +1 -0
  259. package/dist/core/processor/readers/MarkdownReader.js +289 -0
  260. package/dist/core/processor/readers/MarkdownReader.js.map +1 -0
  261. package/dist/core/processor/readers/MarkerPdfReader.js +193 -0
  262. package/dist/core/processor/readers/MarkerPdfReader.js.map +1 -0
  263. package/dist/core/processor/readers/MistralOcrReader.js +198 -0
  264. package/dist/core/processor/readers/MistralOcrReader.js.map +1 -0
  265. package/dist/core/processor/readers/OfficeReader.js +174 -0
  266. package/dist/core/processor/readers/OfficeReader.js.map +1 -0
  267. package/dist/core/processor/readers/PdfReader.js +116 -0
  268. package/dist/core/processor/readers/PdfReader.js.map +1 -0
  269. package/dist/core/processor/readers/RtfReader.js +107 -0
  270. package/dist/core/processor/readers/RtfReader.js.map +1 -0
  271. package/dist/core/processor/readers/SubtitleReader.js +145 -0
  272. package/dist/core/processor/readers/SubtitleReader.js.map +1 -0
  273. package/dist/core/processor/readers/TesseractPdfReader.js +183 -0
  274. package/dist/core/processor/readers/TesseractPdfReader.js.map +1 -0
  275. package/dist/core/processor/readers/TextReader.js +129 -0
  276. package/dist/core/processor/readers/TextReader.js.map +1 -0
  277. package/dist/core/processor/readers/TranscriptReader.js +234 -0
  278. package/dist/core/processor/readers/TranscriptReader.js.map +1 -0
  279. package/dist/core/processor/readers/image/imageMetadata.js +155 -0
  280. package/dist/core/processor/readers/image/imageMetadata.js.map +1 -0
  281. package/dist/core/processor/readers/index.js +41 -0
  282. package/dist/core/processor/readers/index.js.map +1 -0
  283. package/dist/core/processor/readers/referenceExtraction.js +198 -0
  284. package/dist/core/processor/readers/referenceExtraction.js.map +1 -0
  285. package/dist/core/processor/readers/stripReferences.js +59 -0
  286. package/dist/core/processor/readers/stripReferences.js.map +1 -0
  287. package/dist/core/processor/readers/transcript/turnPacking.js +81 -0
  288. package/dist/core/processor/readers/transcript/turnPacking.js.map +1 -0
  289. package/dist/core/progress/NdjsonProgressEmitter.js +30 -0
  290. package/dist/core/progress/NdjsonProgressEmitter.js.map +1 -0
  291. package/dist/core/progress/NoopProgressEmitter.js +15 -0
  292. package/dist/core/progress/NoopProgressEmitter.js.map +1 -0
  293. package/dist/core/progress/index.js +19 -0
  294. package/dist/core/progress/index.js.map +1 -0
  295. package/dist/core/trace/TraceWriter.js +100 -0
  296. package/dist/core/trace/TraceWriter.js.map +1 -0
  297. package/dist/core/trace/events.js +13 -0
  298. package/dist/core/trace/events.js.map +1 -0
  299. package/dist/core/trace/index.js +20 -0
  300. package/dist/core/trace/index.js.map +1 -0
  301. package/dist/core/trace/lineage.js +97 -0
  302. package/dist/core/trace/lineage.js.map +1 -0
  303. package/dist/evaluation/BenchmarkRunner.js +171 -0
  304. package/dist/evaluation/BenchmarkRunner.js.map +1 -0
  305. package/dist/evaluation/classifier/ClassifierAccuracy.js +185 -0
  306. package/dist/evaluation/classifier/ClassifierAccuracy.js.map +1 -0
  307. package/dist/evaluation/classifier/labeledSamples.js +379 -0
  308. package/dist/evaluation/classifier/labeledSamples.js.map +1 -0
  309. package/dist/evaluation/compare/goldCompare.js +126 -0
  310. package/dist/evaluation/compare/goldCompare.js.map +1 -0
  311. package/dist/evaluation/crossre/compareScoring.js +30 -0
  312. package/dist/evaluation/crossre/compareScoring.js.map +1 -0
  313. package/dist/evaluation/datasets/CrossREDataset.js +170 -0
  314. package/dist/evaluation/datasets/CrossREDataset.js.map +1 -0
  315. package/dist/evaluation/datasets/IDataset.js +3 -0
  316. package/dist/evaluation/datasets/IDataset.js.map +1 -0
  317. package/dist/evaluation/datasets/RebelDataset.js +117 -0
  318. package/dist/evaluation/datasets/RebelDataset.js.map +1 -0
  319. package/dist/evaluation/datasets/RedocredDataset.js +218 -0
  320. package/dist/evaluation/datasets/RedocredDataset.js.map +1 -0
  321. package/dist/evaluation/datasets/SemEval2010Dataset.js +150 -0
  322. package/dist/evaluation/datasets/SemEval2010Dataset.js.map +1 -0
  323. package/dist/evaluation/index.js +33 -0
  324. package/dist/evaluation/index.js.map +1 -0
  325. package/dist/evaluation/matching/ExactMatcher.js +75 -0
  326. package/dist/evaluation/matching/ExactMatcher.js.map +1 -0
  327. package/dist/evaluation/matching/SemanticMatcher.js +143 -0
  328. package/dist/evaluation/matching/SemanticMatcher.js.map +1 -0
  329. package/dist/evaluation/metrics/TripleMetrics.js +64 -0
  330. package/dist/evaluation/metrics/TripleMetrics.js.map +1 -0
  331. package/dist/evaluation/mine/MineCheckpoint.js +114 -0
  332. package/dist/evaluation/mine/MineCheckpoint.js.map +1 -0
  333. package/dist/evaluation/mine/MineDataset.js +208 -0
  334. package/dist/evaluation/mine/MineDataset.js.map +1 -0
  335. package/dist/evaluation/mine/MineReporter.js +98 -0
  336. package/dist/evaluation/mine/MineReporter.js.map +1 -0
  337. package/dist/evaluation/mine/MineRunner.js +148 -0
  338. package/dist/evaluation/mine/MineRunner.js.map +1 -0
  339. package/dist/evaluation/mine/MineScorer.js +127 -0
  340. package/dist/evaluation/mine/MineScorer.js.map +1 -0
  341. package/dist/evaluation/mine/types.js +12 -0
  342. package/dist/evaluation/mine/types.js.map +1 -0
  343. package/dist/evaluation/reporters/ConsoleReporter.js +55 -0
  344. package/dist/evaluation/reporters/ConsoleReporter.js.map +1 -0
  345. package/dist/evaluation/reporters/JsonReporter.js +50 -0
  346. package/dist/evaluation/reporters/JsonReporter.js.map +1 -0
  347. package/dist/index.js +28 -0
  348. package/dist/index.js.map +1 -0
  349. package/dist/quality/CompositeScore.js +61 -0
  350. package/dist/quality/CompositeScore.js.map +1 -0
  351. package/dist/quality/ConsistencyMetrics.js +70 -0
  352. package/dist/quality/ConsistencyMetrics.js.map +1 -0
  353. package/dist/quality/FactualMetrics.js +76 -0
  354. package/dist/quality/FactualMetrics.js.map +1 -0
  355. package/dist/quality/GraphHealthMetrics.js +68 -0
  356. package/dist/quality/GraphHealthMetrics.js.map +1 -0
  357. package/dist/quality/SemanticMetrics.js +102 -0
  358. package/dist/quality/SemanticMetrics.js.map +1 -0
  359. package/dist/quality/StructuralMetrics.js +60 -0
  360. package/dist/quality/StructuralMetrics.js.map +1 -0
  361. package/dist/quality/index.js +23 -0
  362. package/dist/quality/index.js.map +1 -0
  363. package/dist/shared/index.js +20 -0
  364. package/dist/shared/index.js.map +1 -0
  365. package/dist/shared/logger/Logger.js +3 -0
  366. package/dist/shared/logger/Logger.js.map +1 -0
  367. package/dist/shared/logger/LoggerFactory.js +75 -0
  368. package/dist/shared/logger/LoggerFactory.js.map +1 -0
  369. package/dist/shared/logger/index.js +19 -0
  370. package/dist/shared/logger/index.js.map +1 -0
  371. package/dist/shared/shutdown.js +30 -0
  372. package/dist/shared/shutdown.js.map +1 -0
  373. package/dist/shared/utils/agglomerativeCluster.js +269 -0
  374. package/dist/shared/utils/agglomerativeCluster.js.map +1 -0
  375. package/dist/shared/utils/astSymbols.js +69 -0
  376. package/dist/shared/utils/astSymbols.js.map +1 -0
  377. package/dist/shared/utils/cosineSimilarity.js +18 -0
  378. package/dist/shared/utils/cosineSimilarity.js.map +1 -0
  379. package/dist/shared/utils/directoryTree.js +184 -0
  380. package/dist/shared/utils/directoryTree.js.map +1 -0
  381. package/dist/shared/utils/documentOutline.js +74 -0
  382. package/dist/shared/utils/documentOutline.js.map +1 -0
  383. package/dist/shared/utils/index.js +24 -0
  384. package/dist/shared/utils/index.js.map +1 -0
  385. package/dist/shared/utils/jaroWinklerSimilarity.js +60 -0
  386. package/dist/shared/utils/jaroWinklerSimilarity.js.map +1 -0
  387. package/dist/shared/utils/parseJsonLenient.js +27 -0
  388. package/dist/shared/utils/parseJsonLenient.js.map +1 -0
  389. package/dist/shared/utils/readConfig.js +42 -0
  390. package/dist/shared/utils/readConfig.js.map +1 -0
  391. package/dist/shared/utils/readRtf.js +216 -0
  392. package/dist/shared/utils/readRtf.js.map +1 -0
  393. package/dist/shared/utils/softmax.js +26 -0
  394. package/dist/shared/utils/softmax.js.map +1 -0
  395. package/dist/types/ContentClass.js +3 -0
  396. package/dist/types/ContentClass.js.map +1 -0
  397. package/dist/types/CorpusProfile.js +3 -0
  398. package/dist/types/CorpusProfile.js.map +1 -0
  399. package/dist/types/IContradictionChecker.js +3 -0
  400. package/dist/types/IContradictionChecker.js.map +1 -0
  401. package/dist/types/ICorpusAnalyzer.js +3 -0
  402. package/dist/types/ICorpusAnalyzer.js.map +1 -0
  403. package/dist/types/IDirectoryProcessor.js +3 -0
  404. package/dist/types/IDirectoryProcessor.js.map +1 -0
  405. package/dist/types/IEmbeddingProvider.js +3 -0
  406. package/dist/types/IEmbeddingProvider.js.map +1 -0
  407. package/dist/types/IEmbeddingService.js +6 -0
  408. package/dist/types/IEmbeddingService.js.map +1 -0
  409. package/dist/types/IFileProcessor.js +3 -0
  410. package/dist/types/IFileProcessor.js.map +1 -0
  411. package/dist/types/IGroundingChecker.js +3 -0
  412. package/dist/types/IGroundingChecker.js.map +1 -0
  413. package/dist/types/IKnowledgeGraphBuilder.js +3 -0
  414. package/dist/types/IKnowledgeGraphBuilder.js.map +1 -0
  415. package/dist/types/IKnowledgeGraphExporter.js +3 -0
  416. package/dist/types/IKnowledgeGraphExporter.js.map +1 -0
  417. package/dist/types/IKnowledgeGraphMerger.js +3 -0
  418. package/dist/types/IKnowledgeGraphMerger.js.map +1 -0
  419. package/dist/types/IKnowledgeGraphSearch.js +3 -0
  420. package/dist/types/IKnowledgeGraphSearch.js.map +1 -0
  421. package/dist/types/ILLMProvider.js +3 -0
  422. package/dist/types/ILLMProvider.js.map +1 -0
  423. package/dist/types/ILLMService.js +3 -0
  424. package/dist/types/ILLMService.js.map +1 -0
  425. package/dist/types/IObjectDetector.js +3 -0
  426. package/dist/types/IObjectDetector.js.map +1 -0
  427. package/dist/types/IProcessingService.js +3 -0
  428. package/dist/types/IProcessingService.js.map +1 -0
  429. package/dist/types/IProgressEmitter.js +3 -0
  430. package/dist/types/IProgressEmitter.js.map +1 -0
  431. package/dist/types/IPromptManager.js +3 -0
  432. package/dist/types/IPromptManager.js.map +1 -0
  433. package/dist/types/KnowledgeGraph.js +3 -0
  434. package/dist/types/KnowledgeGraph.js.map +1 -0
  435. package/dist/types/MCPKnowledgeGraph.js +3 -0
  436. package/dist/types/MCPKnowledgeGraph.js.map +1 -0
  437. package/dist/types/Observation.js +21 -0
  438. package/dist/types/Observation.js.map +1 -0
  439. package/dist/types/ProcessingOptions.js +3 -0
  440. package/dist/types/ProcessingOptions.js.map +1 -0
  441. package/dist/types/index.js +40 -0
  442. package/dist/types/index.js.map +1 -0
  443. package/package.json +122 -0
@@ -0,0 +1,23 @@
1
+ "use strict";
2
+ var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
+ if (k2 === undefined) k2 = k;
4
+ var desc = Object.getOwnPropertyDescriptor(m, k);
5
+ if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
+ desc = { enumerable: true, get: function() { return m[k]; } };
7
+ }
8
+ Object.defineProperty(o, k2, desc);
9
+ }) : (function(o, m, k, k2) {
10
+ if (k2 === undefined) k2 = k;
11
+ o[k2] = m[k];
12
+ }));
13
+ var __exportStar = (this && this.__exportStar) || function(m, exports) {
14
+ for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
15
+ };
16
+ Object.defineProperty(exports, "__esModule", { value: true });
17
+ __exportStar(require("./StructuralMetrics"), exports);
18
+ __exportStar(require("./GraphHealthMetrics"), exports);
19
+ __exportStar(require("./SemanticMetrics"), exports);
20
+ __exportStar(require("./FactualMetrics"), exports);
21
+ __exportStar(require("./ConsistencyMetrics"), exports);
22
+ __exportStar(require("./CompositeScore"), exports);
23
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/quality/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,sDAAoC;AACpC,uDAAqC;AACrC,oDAAkC;AAClC,mDAAiC;AACjC,uDAAqC;AACrC,mDAAiC"}
@@ -0,0 +1,20 @@
1
+ "use strict";
2
+ var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
+ if (k2 === undefined) k2 = k;
4
+ var desc = Object.getOwnPropertyDescriptor(m, k);
5
+ if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
+ desc = { enumerable: true, get: function() { return m[k]; } };
7
+ }
8
+ Object.defineProperty(o, k2, desc);
9
+ }) : (function(o, m, k, k2) {
10
+ if (k2 === undefined) k2 = k;
11
+ o[k2] = m[k];
12
+ }));
13
+ var __exportStar = (this && this.__exportStar) || function(m, exports) {
14
+ for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
15
+ };
16
+ Object.defineProperty(exports, "__esModule", { value: true });
17
+ __exportStar(require("./logger"), exports);
18
+ __exportStar(require("./utils"), exports);
19
+ __exportStar(require("./shutdown"), exports);
20
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/shared/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,2CAAyB;AACzB,0CAAwB;AACxB,6CAA2B"}
@@ -0,0 +1,3 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ //# sourceMappingURL=Logger.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"Logger.js","sourceRoot":"","sources":["../../../src/shared/logger/Logger.ts"],"names":[],"mappings":""}
@@ -0,0 +1,75 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.LoggerFactory = void 0;
4
+ const fs_1 = require("fs");
5
+ const tslog_1 = require("tslog");
6
+ class LoggerFactory {
7
+ static createLogger(options) {
8
+ var _a;
9
+ const { file: logFile, level: logLevel, silent, debug, progressNdjson, } = (_a = options.logging) !== null && _a !== void 0 ? _a : {};
10
+ const logger = new tslog_1.Logger({
11
+ name: "wanshi",
12
+ // In NDJSON progress mode, suppress tslog's built-in pretty output so it
13
+ // can't pollute the structured stdout stream; we re-emit logs ourselves as
14
+ // `channel: "log"` lines below.
15
+ type: progressNdjson ? "hidden" : "pretty",
16
+ // tslog scale: silly=0, trace=1, debug=2, info=3, warn=4, error=5, fatal=6.
17
+ // The old mapping was two notches loose (info→1 = "trace and up"), and
18
+ // `silent` at 4 still printed warn/error. (KG-19)
19
+ minLevel: silent
20
+ ? 7 // above fatal → nothing prints
21
+ : debug
22
+ ? 2
23
+ : logLevel === "debug"
24
+ ? 2
25
+ : logLevel === "info"
26
+ ? 3
27
+ : logLevel === "warning"
28
+ ? 4
29
+ : logLevel === "error"
30
+ ? 5
31
+ : 3,
32
+ });
33
+ if (progressNdjson) {
34
+ // Bridge logs onto the same NDJSON stdout stream the progress emitter uses,
35
+ // so a parent process gets a live log tail demuxed by `channel`.
36
+ logger.attachTransport((logObj) => {
37
+ var _a, _b;
38
+ try {
39
+ process.stdout.write(JSON.stringify({
40
+ channel: "log",
41
+ ts: Date.now(),
42
+ level: String((_b = (_a = logObj === null || logObj === void 0 ? void 0 : logObj._meta) === null || _a === void 0 ? void 0 : _a.logLevelName) !== null && _b !== void 0 ? _b : "INFO").toLowerCase(),
43
+ message: stringifyLogArgs(logObj),
44
+ }) + "\n");
45
+ }
46
+ catch (_c) {
47
+ // side-channel only — swallow write errors (e.g. EPIPE)
48
+ }
49
+ });
50
+ }
51
+ if (logFile) {
52
+ logger.attachTransport((logObj) => {
53
+ (0, fs_1.appendFileSync)(logFile, JSON.stringify(logObj) + "\n");
54
+ });
55
+ }
56
+ return logger;
57
+ }
58
+ }
59
+ exports.LoggerFactory = LoggerFactory;
60
+ /**
61
+ * tslog passes logged arguments as numeric keys (`0`, `1`, …) on the log object
62
+ * alongside the `_meta` block. Join the string parts into a single message,
63
+ * JSON-stringifying any non-string argument.
64
+ */
65
+ function stringifyLogArgs(logObj) {
66
+ const parts = [];
67
+ for (const key of Object.keys(logObj)) {
68
+ if (key === "_meta")
69
+ continue;
70
+ const value = logObj[key];
71
+ parts.push(typeof value === "string" ? value : JSON.stringify(value));
72
+ }
73
+ return parts.join(" ");
74
+ }
75
+ //# sourceMappingURL=LoggerFactory.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"LoggerFactory.js","sourceRoot":"","sources":["../../../src/shared/logger/LoggerFactory.ts"],"names":[],"mappings":";;;AAAA,2BAAoC;AACpC,iCAA2C;AAI3C,MAAa,aAAa;IACxB,MAAM,CAAC,YAAY,CACjB,OAA4D;;QAE5D,MAAM,EACJ,IAAI,EAAE,OAAO,EACb,KAAK,EAAE,QAAQ,EACf,MAAM,EACN,KAAK,EACL,cAAc,GACf,GAAG,MAAA,OAAO,CAAC,OAAO,mCAAI,EAAE,CAAC;QAE1B,MAAM,MAAM,GAAG,IAAI,cAAQ,CAAM;YAC/B,IAAI,EAAE,QAAQ;YACd,yEAAyE;YACzE,2EAA2E;YAC3E,gCAAgC;YAChC,IAAI,EAAE,cAAc,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ;YAC1C,4EAA4E;YAC5E,uEAAuE;YACvE,kDAAkD;YAClD,QAAQ,EAAE,MAAM;gBACd,CAAC,CAAC,CAAC,CAAC,+BAA+B;gBACnC,CAAC,CAAC,KAAK;oBACP,CAAC,CAAC,CAAC;oBACH,CAAC,CAAC,QAAQ,KAAK,OAAO;wBACtB,CAAC,CAAC,CAAC;wBACH,CAAC,CAAC,QAAQ,KAAK,MAAM;4BACrB,CAAC,CAAC,CAAC;4BACH,CAAC,CAAC,QAAQ,KAAK,SAAS;gCACxB,CAAC,CAAC,CAAC;gCACH,CAAC,CAAC,QAAQ,KAAK,OAAO;oCACtB,CAAC,CAAC,CAAC;oCACH,CAAC,CAAC,CAAC;SACN,CAAC,CAAC;QAEH,IAAI,cAAc,EAAE,CAAC;YACnB,4EAA4E;YAC5E,iEAAiE;YACjE,MAAM,CAAC,eAAe,CAAC,CAAC,MAAM,EAAE,EAAE;;gBAChC,IAAI,CAAC;oBACH,OAAO,CAAC,MAAM,CAAC,KAAK,CAClB,IAAI,CAAC,SAAS,CAAC;wBACb,OAAO,EAAE,KAAK;wBACd,EAAE,EAAE,IAAI,CAAC,GAAG,EAAE;wBACd,KAAK,EAAE,MAAM,CAAC,MAAA,MAAA,MAAM,aAAN,MAAM,uBAAN,MAAM,CAAE,KAAK,0CAAE,YAAY,mCAAI,MAAM,CAAC,CAAC,WAAW,EAAE;wBAClE,OAAO,EAAE,gBAAgB,CAAC,MAAM,CAAC;qBAClC,CAAC,GAAG,IAAI,CACV,CAAC;gBACJ,CAAC;gBAAC,WAAM,CAAC;oBACP,wDAAwD;gBAC1D,CAAC;YACH,CAAC,CAAC,CAAC;QACL,CAAC;QAED,IAAI,OAAO,EAAE,CAAC;YACZ,MAAM,CAAC,eAAe,CAAC,CAAC,MAAM,EAAE,EAAE;gBAChC,IAAA,mBAAc,EAAC,OAAO,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC;YACzD,CAAC,CAAC,CAAC;QACL,CAAC;QAED,OAAO,MAAgB,CAAC;IAC1B,CAAC;CACF;AA/DD,sCA+DC;AAED;;;;GAIG;AACH,SAAS,gBAAgB,CAAC,MAA+B;IACvD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QACtC,IAAI,GAAG,KAAK,OAAO;YAAE,SAAS;QAC9B,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QAC1B,KAAK,CAAC,IAAI,CAAC,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC;IACxE,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC"}
@@ -0,0 +1,19 @@
1
+ "use strict";
2
+ var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
+ if (k2 === undefined) k2 = k;
4
+ var desc = Object.getOwnPropertyDescriptor(m, k);
5
+ if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
+ desc = { enumerable: true, get: function() { return m[k]; } };
7
+ }
8
+ Object.defineProperty(o, k2, desc);
9
+ }) : (function(o, m, k, k2) {
10
+ if (k2 === undefined) k2 = k;
11
+ o[k2] = m[k];
12
+ }));
13
+ var __exportStar = (this && this.__exportStar) || function(m, exports) {
14
+ for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
15
+ };
16
+ Object.defineProperty(exports, "__esModule", { value: true });
17
+ __exportStar(require("./Logger"), exports);
18
+ __exportStar(require("./LoggerFactory"), exports);
19
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/shared/logger/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,2CAAyB;AACzB,kDAAgC"}
@@ -0,0 +1,30 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.shutdown = void 0;
4
+ /**
5
+ * Process-wide cooperative shutdown flag for graceful interrupts (Ctrl+C / Ctrl+D).
6
+ *
7
+ * The long-running processing loops poll `isRequested()` between units of work
8
+ * (between files in DirectoryProcessor, between chunks in KnowledgeGraphBuilder)
9
+ * and stop cleanly — finishing the in-flight chunk, checkpointing it, and flushing
10
+ * a partial graph — instead of being killed mid-chunk and wasting a paid LLM call.
11
+ *
12
+ * A module singleton keeps this low-ceremony for a single-run CLI: the signal
13
+ * handlers in the CLI flip the flag, the loops read it. No DI plumbing.
14
+ */
15
+ class ShutdownController {
16
+ constructor() {
17
+ this.requested = false;
18
+ }
19
+ isRequested() {
20
+ return this.requested;
21
+ }
22
+ request() {
23
+ this.requested = true;
24
+ }
25
+ reset() {
26
+ this.requested = false;
27
+ }
28
+ }
29
+ exports.shutdown = new ShutdownController();
30
+ //# sourceMappingURL=shutdown.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"shutdown.js","sourceRoot":"","sources":["../../src/shared/shutdown.ts"],"names":[],"mappings":";;;AAAA;;;;;;;;;;GAUG;AACH,MAAM,kBAAkB;IAAxB;QACU,cAAS,GAAG,KAAK,CAAC;IAa5B,CAAC;IAXC,WAAW;QACT,OAAO,IAAI,CAAC,SAAS,CAAC;IACxB,CAAC;IAED,OAAO;QACL,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC;IACxB,CAAC;IAED,KAAK;QACH,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC;IACzB,CAAC;CACF;AAEY,QAAA,QAAQ,GAAG,IAAI,kBAAkB,EAAE,CAAC"}
@@ -0,0 +1,269 @@
1
+ "use strict";
2
+ var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
3
+ function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
4
+ return new (P || (P = Promise))(function (resolve, reject) {
5
+ function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
6
+ function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
7
+ function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
8
+ step((generator = generator.apply(thisArg, _arguments || [])).next());
9
+ });
10
+ };
11
+ Object.defineProperty(exports, "__esModule", { value: true });
12
+ exports.agglomerativeClusters = agglomerativeClusters;
13
+ exports.clusterByEmbedding = clusterByEmbedding;
14
+ const cosineSimilarity_1 = require("./cosineSimilarity");
15
+ /** Minimal union-find (path compression + union by size). */
16
+ class UnionFind {
17
+ constructor(n) {
18
+ this.parent = Array.from({ length: n }, (_, i) => i);
19
+ this.size = new Array(n).fill(1);
20
+ }
21
+ find(x) {
22
+ while (this.parent[x] !== x) {
23
+ this.parent[x] = this.parent[this.parent[x]];
24
+ x = this.parent[x];
25
+ }
26
+ return x;
27
+ }
28
+ union(a, b) {
29
+ let ra = this.find(a);
30
+ let rb = this.find(b);
31
+ if (ra === rb)
32
+ return;
33
+ if (this.size[ra] < this.size[rb])
34
+ [ra, rb] = [rb, ra];
35
+ this.parent[rb] = ra;
36
+ this.size[ra] += this.size[rb];
37
+ }
38
+ connected(a, b) {
39
+ return this.find(a) === this.find(b);
40
+ }
41
+ }
42
+ /**
43
+ * Symmetric top-N eligibility matrix: `m[i][j]` is true iff j is among i's N most
44
+ * similar neighbours OR vice versa. Returns null when blocking is off (topN ≤ 0),
45
+ * meaning "all pairs eligible". Pure.
46
+ */
47
+ function blockingEligibility(sim, n, topN) {
48
+ if (!topN || topN <= 0 || topN >= n - 1)
49
+ return null;
50
+ const m = Array.from({ length: n }, () => new Array(n).fill(false));
51
+ for (let i = 0; i < n; i++) {
52
+ const neighbours = [...Array(n).keys()]
53
+ .filter((j) => j !== i)
54
+ .sort((a, b) => sim[i][b] - sim[i][a])
55
+ .slice(0, topN);
56
+ for (const j of neighbours) {
57
+ m[i][j] = true;
58
+ m[j][i] = true; // symmetric: eligible if in EITHER endpoint's top-N
59
+ }
60
+ }
61
+ return m;
62
+ }
63
+ /** Group item indices by their union-find root, returned as id clusters. */
64
+ function clustersFromUF(items, uf) {
65
+ const groups = new Map();
66
+ for (let i = 0; i < items.length; i++) {
67
+ const root = uf.find(i);
68
+ const g = groups.get(root);
69
+ if (g)
70
+ g.push(items[i].id);
71
+ else
72
+ groups.set(root, [items[i].id]);
73
+ }
74
+ return Array.from(groups.values());
75
+ }
76
+ /**
77
+ * Single-linkage agglomerative clustering: connected components of the graph
78
+ * where an edge joins two items whose cosine similarity ≥ `threshold`. Pure and
79
+ * synchronous; singletons are returned as one-element clusters. This is the
80
+ * embeddings-method workhorse for canonicalization.
81
+ */
82
+ function agglomerativeClusters(items, threshold) {
83
+ const uf = new UnionFind(items.length);
84
+ for (let i = 0; i < items.length; i++) {
85
+ for (let j = i + 1; j < items.length; j++) {
86
+ if ((0, cosineSimilarity_1.cosineSimilarity)(items[i].embedding, items[j].embedding) >= threshold) {
87
+ uf.union(i, j);
88
+ }
89
+ }
90
+ }
91
+ return clustersFromUF(items, uf);
92
+ }
93
+ /**
94
+ * Clustering with a pluggable per-pair policy, shared by all canonicalization
95
+ * methods:
96
+ * - embeddings: decide = sim ≥ threshold ? "merge" : "reject" (no escalation)
97
+ * - llm/hybrid: decide = sim ≥ high ? "merge" : sim ≥ low ? "escalate" : "reject",
98
+ * with the band escalated to the LLM adjudicator.
99
+ *
100
+ * "merge" pairs are unioned first; "escalate" pairs are then adjudicated highest
101
+ * similarity first (and skipped when already transitively merged). Pairs whose
102
+ * similarity falls in `band` are recorded as borderline with their final
103
+ * merged/not-merged outcome — the signal the merge-log viewer surfaces.
104
+ */
105
+ function clusterByEmbedding(items, opts) {
106
+ return __awaiter(this, void 0, void 0, function* () {
107
+ var _a;
108
+ if (((_a = opts.linkage) !== null && _a !== void 0 ? _a : "single") === "complete") {
109
+ return completeLinkageCluster(items, opts);
110
+ }
111
+ const uf = new UnionFind(items.length);
112
+ const escalate = [];
113
+ const bandPairs = [];
114
+ for (let i = 0; i < items.length; i++) {
115
+ for (let j = i + 1; j < items.length; j++) {
116
+ const sim = (0, cosineSimilarity_1.cosineSimilarity)(items[i].embedding, items[j].embedding);
117
+ if (opts.band && sim >= opts.band[0] && sim < opts.band[1]) {
118
+ bandPairs.push({ i, j, sim });
119
+ }
120
+ const d = opts.decide(sim, items[i].id, items[j].id);
121
+ if (d === "merge")
122
+ uf.union(i, j);
123
+ else if (d === "escalate")
124
+ escalate.push({ i, j, sim });
125
+ }
126
+ }
127
+ // Adjudicate escalated pairs, most-similar first; skip pairs already merged
128
+ // transitively (a no-op the LLM shouldn't be billed for).
129
+ escalate.sort((a, b) => b.sim - a.sim);
130
+ for (const { i, j, sim } of escalate) {
131
+ if (uf.connected(i, j))
132
+ continue;
133
+ if (opts.adjudicate && (yield opts.adjudicate(items[i].id, items[j].id, sim))) {
134
+ uf.union(i, j);
135
+ }
136
+ }
137
+ // Borderline = band pairs ∪ escalated pairs, with their final outcome.
138
+ const seen = new Set();
139
+ const borderline = [];
140
+ for (const { i, j, sim } of [...bandPairs, ...escalate]) {
141
+ const key = i + "-" + j;
142
+ if (seen.has(key))
143
+ continue;
144
+ seen.add(key);
145
+ borderline.push({
146
+ a: items[i].id,
147
+ b: items[j].id,
148
+ sim,
149
+ merged: uf.connected(i, j),
150
+ });
151
+ }
152
+ return { clusters: clustersFromUF(items, uf), borderline };
153
+ });
154
+ }
155
+ /**
156
+ * Complete-linkage agglomerative clustering over the same pluggable pairwise policy.
157
+ * A cluster is valid only when EVERY internal pair is "joinable" (decide()==="merge",
158
+ * or an "escalate" pair the adjudicator approves); a single "reject"/veto cross-pair
159
+ * blocks the whole merge. Clusters grow greedily by highest min-cross-similarity
160
+ * (tightest first), so — unlike single-linkage — one borderline bridge between two
161
+ * otherwise-dissimilar siblings can't chain a family into one node (CESI, WWW 2018;
162
+ * the 8-member Epicure fusion). Adjudication is lazy: an escalate pair is sent to the
163
+ * LLM only when it is the binding cross-pair of a candidate merge (rejects short-
164
+ * circuit first), which keeps call volume bounded.
165
+ */
166
+ function completeLinkageCluster(items, opts) {
167
+ return __awaiter(this, void 0, void 0, function* () {
168
+ const n = items.length;
169
+ const sim = Array.from({ length: n }, () => new Array(n).fill(0));
170
+ const decision = Array.from({ length: n }, () => new Array(n).fill("reject"));
171
+ const bandPairs = [];
172
+ const escalatePairs = [];
173
+ for (let i = 0; i < n; i++) {
174
+ for (let j = i + 1; j < n; j++) {
175
+ sim[i][j] = sim[j][i] = (0, cosineSimilarity_1.cosineSimilarity)(items[i].embedding, items[j].embedding);
176
+ }
177
+ }
178
+ // Blocking: keep only each item's top-N neighbours as merge-eligible; the rest are
179
+ // forced to "reject" so decide/adjudicate is never spent on them.
180
+ const eligible = blockingEligibility(sim, n, opts.blockTopN);
181
+ for (let i = 0; i < n; i++) {
182
+ for (let j = i + 1; j < n; j++) {
183
+ const s = sim[i][j];
184
+ if (eligible && !eligible[i][j]) {
185
+ decision[i][j] = decision[j][i] = "reject";
186
+ continue;
187
+ }
188
+ if (opts.band && s >= opts.band[0] && s < opts.band[1])
189
+ bandPairs.push({ i, j, sim: s });
190
+ const d = opts.decide(s, items[i].id, items[j].id);
191
+ decision[i][j] = decision[j][i] = d;
192
+ if (d === "escalate")
193
+ escalatePairs.push({ i, j, sim: s });
194
+ }
195
+ }
196
+ // Lazy, cached per-pair joinability (only escalate pairs hit the adjudicator).
197
+ const adjCache = new Map();
198
+ const joinable = (i, j) => __awaiter(this, void 0, void 0, function* () {
199
+ const d = decision[i][j];
200
+ if (d === "merge")
201
+ return true;
202
+ if (d === "reject")
203
+ return false;
204
+ const key = (i < j ? i : j) * n + (i < j ? j : i);
205
+ const cached = adjCache.get(key);
206
+ if (cached !== undefined)
207
+ return cached;
208
+ const ok = opts.adjudicate ? yield opts.adjudicate(items[i].id, items[j].id, sim[i][j]) : false;
209
+ adjCache.set(key, ok);
210
+ return ok;
211
+ });
212
+ // Every cross-pair joinable? Cheap rejects first, then adjudicate the rest.
213
+ const allJoinable = (A, B) => __awaiter(this, void 0, void 0, function* () {
214
+ for (const x of A)
215
+ for (const y of B)
216
+ if (decision[x][y] === "reject")
217
+ return false;
218
+ for (const x of A)
219
+ for (const y of B)
220
+ if (!(yield joinable(x, y)))
221
+ return false;
222
+ return true;
223
+ });
224
+ let clusters = items.map((_, i) => [i]);
225
+ // Greedy: merge the tightest fully-joinable cluster pair until none qualify.
226
+ for (;;) {
227
+ const candidates = [];
228
+ for (let a = 0; a < clusters.length; a++) {
229
+ for (let b = a + 1; b < clusters.length; b++) {
230
+ let mn = Infinity;
231
+ for (const x of clusters[a])
232
+ for (const y of clusters[b])
233
+ mn = Math.min(mn, sim[x][y]);
234
+ candidates.push({ a, b, min: mn });
235
+ }
236
+ }
237
+ candidates.sort((p, q) => q.min - p.min);
238
+ let mergedAny = false;
239
+ for (const c of candidates) {
240
+ if (yield allJoinable(clusters[c.a], clusters[c.b])) {
241
+ clusters[c.a] = clusters[c.a].concat(clusters[c.b]);
242
+ clusters.splice(c.b, 1); // c.b > c.a, so c.a index is unaffected
243
+ mergedAny = true;
244
+ break; // recompute candidates from scratch after each merge
245
+ }
246
+ }
247
+ if (!mergedAny)
248
+ break;
249
+ }
250
+ const clusterOf = new Map();
251
+ clusters.forEach((c, ci) => c.forEach((idx) => clusterOf.set(idx, ci)));
252
+ const seen = new Set();
253
+ const borderline = [];
254
+ for (const { i, j, sim: s } of [...bandPairs, ...escalatePairs]) {
255
+ const key = i + "-" + j;
256
+ if (seen.has(key))
257
+ continue;
258
+ seen.add(key);
259
+ borderline.push({
260
+ a: items[i].id,
261
+ b: items[j].id,
262
+ sim: s,
263
+ merged: clusterOf.get(i) === clusterOf.get(j),
264
+ });
265
+ }
266
+ return { clusters: clusters.map((c) => c.map((idx) => items[idx].id)), borderline };
267
+ });
268
+ }
269
+ //# sourceMappingURL=agglomerativeCluster.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"agglomerativeCluster.js","sourceRoot":"","sources":["../../../src/shared/utils/agglomerativeCluster.ts"],"names":[],"mappings":";;;;;;;;;;;AA+EA,sDAUC;AAoDD,gDAiDC;AA9LD,yDAAsD;AAQtD,6DAA6D;AAC7D,MAAM,SAAS;IAGb,YAAY,CAAS;QACnB,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;QACrD,IAAI,CAAC,IAAI,GAAG,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACnC,CAAC;IACD,IAAI,CAAC,CAAS;QACZ,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC;YAC5B,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;YAC7C,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QACrB,CAAC;QACD,OAAO,CAAC,CAAC;IACX,CAAC;IACD,KAAK,CAAC,CAAS,EAAE,CAAS;QACxB,IAAI,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACtB,IAAI,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACtB,IAAI,EAAE,KAAK,EAAE;YAAE,OAAO;QACtB,IAAI,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAAE,CAAC,EAAE,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;QACvD,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;QACrB,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACjC,CAAC;IACD,SAAS,CAAC,CAAS,EAAE,CAAS;QAC5B,OAAO,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACvC,CAAC;CACF;AAED;;;;GAIG;AACH,SAAS,mBAAmB,CAC1B,GAAe,EACf,CAAS,EACT,IAAa;IAEb,IAAI,CAAC,IAAI,IAAI,IAAI,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACrD,MAAM,CAAC,GAAgB,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,IAAI,KAAK,CAAU,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAC1F,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3B,MAAM,UAAU,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;aACpC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC;aACtB,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;aACrC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QAClB,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;YAC3B,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC;YACf,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,oDAAoD;QACtE,CAAC;IACH,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,4EAA4E;AAC5E,SAAS,cAAc,CAAC,KAAiB,EAAE,EAAa;IACtD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAoB,CAAC;IAC3C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,IAAI,GAAG,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACxB,MAAM,CAAC,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAC3B,IAAI,CAAC;YAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;;YACtB,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IACvC,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;AACrC,CAAC;AAED;;;;;GAKG;AACH,SAAgB,qBAAqB,CAAC,KAAiB,EAAE,SAAiB;IACxE,MAAM,EAAE,GAAG,IAAI,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACvC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC1C,IAAI,IAAA,mCAAgB,EAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,IAAI,SAAS,EAAE,CAAC;gBAC1E,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YACjB,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,cAAc,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;AACnC,CAAC;AAwCD;;;;;;;;;;;GAWG;AACH,SAAsB,kBAAkB,CACtC,KAAiB,EACjB,IAA+B;;;QAE/B,IAAI,CAAC,MAAA,IAAI,CAAC,OAAO,mCAAI,QAAQ,CAAC,KAAK,UAAU,EAAE,CAAC;YAC9C,OAAO,sBAAsB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAC7C,CAAC;QACD,MAAM,EAAE,GAAG,IAAI,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QACvC,MAAM,QAAQ,GAAiD,EAAE,CAAC;QAClE,MAAM,SAAS,GAAiD,EAAE,CAAC;QAEnE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC1C,MAAM,GAAG,GAAG,IAAA,mCAAgB,EAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;gBACrE,IAAI,IAAI,CAAC,IAAI,IAAI,GAAG,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;oBAC3D,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC;gBAChC,CAAC;gBACD,MAAM,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBACrD,IAAI,CAAC,KAAK,OAAO;oBAAE,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;qBAC7B,IAAI,CAAC,KAAK,UAAU;oBAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC;YAC1D,CAAC;QACH,CAAC;QAED,4EAA4E;QAC5E,0DAA0D;QAC1D,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;QACvC,KAAK,MAAM,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,IAAI,QAAQ,EAAE,CAAC;YACrC,IAAI,EAAE,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC;gBAAE,SAAS;YACjC,IAAI,IAAI,CAAC,UAAU,IAAI,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;gBAC9E,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YACjB,CAAC;QACH,CAAC;QAED,uEAAuE;QACvE,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,MAAM,UAAU,GAAqB,EAAE,CAAC;QACxC,KAAK,MAAM,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,SAAS,EAAE,GAAG,QAAQ,CAAC,EAAE,CAAC;YACxD,MAAM,GAAG,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC;YACxB,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC5B,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACd,UAAU,CAAC,IAAI,CAAC;gBACd,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE;gBACd,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE;gBACd,GAAG;gBACH,MAAM,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC;aAC3B,CAAC,CAAC;QACL,CAAC;QAED,OAAO,EAAE,QAAQ,EAAE,cAAc,CAAC,KAAK,EAAE,EAAE,CAAC,EAAE,UAAU,EAAE,CAAC;IAC7D,CAAC;CAAA;AAED;;;;;;;;;;GAUG;AACH,SAAe,sBAAsB,CACnC,KAAiB,EACjB,IAA+B;;QAE/B,MAAM,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;QACvB,MAAM,GAAG,GAAe,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,IAAI,KAAK,CAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;QACtF,MAAM,QAAQ,GAAsB,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,EAAE,GAAG,EAAE,CACjE,IAAI,KAAK,CAAgB,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAC3C,CAAC;QACF,MAAM,SAAS,GAAiD,EAAE,CAAC;QACnE,MAAM,aAAa,GAAiD,EAAE,CAAC;QAEvE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC/B,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,IAAA,mCAAgB,EAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;YACnF,CAAC;QACH,CAAC;QAED,mFAAmF;QACnF,kEAAkE;QAClE,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,EAAE,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC;QAE7D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC/B,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACpB,IAAI,QAAQ,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;oBAChC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC;oBAC3C,SAAS;gBACX,CAAC;gBACD,IAAI,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;oBAAE,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC;gBACzF,MAAM,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBACnD,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;gBACpC,IAAI,CAAC,KAAK,UAAU;oBAAE,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC;YAC7D,CAAC;QACH,CAAC;QAED,+EAA+E;QAC/E,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAmB,CAAC;QAC5C,MAAM,QAAQ,GAAG,CAAO,CAAS,EAAE,CAAS,EAAoB,EAAE;YAChE,MAAM,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACzB,IAAI,CAAC,KAAK,OAAO;gBAAE,OAAO,IAAI,CAAC;YAC/B,IAAI,CAAC,KAAK,QAAQ;gBAAE,OAAO,KAAK,CAAC;YACjC,MAAM,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAClD,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACjC,IAAI,MAAM,KAAK,SAAS;gBAAE,OAAO,MAAM,CAAC;YACxC,MAAM,EAAE,GAAG,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;YAChG,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;YACtB,OAAO,EAAE,CAAC;QACZ,CAAC,CAAA,CAAC;QAEF,4EAA4E;QAC5E,MAAM,WAAW,GAAG,CAAO,CAAW,EAAE,CAAW,EAAoB,EAAE;YACvE,KAAK,MAAM,CAAC,IAAI,CAAC;gBAAE,KAAK,MAAM,CAAC,IAAI,CAAC;oBAAE,IAAI,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,QAAQ;wBAAE,OAAO,KAAK,CAAC;YACpF,KAAK,MAAM,CAAC,IAAI,CAAC;gBAAE,KAAK,MAAM,CAAC,IAAI,CAAC;oBAAE,IAAI,CAAC,CAAC,MAAM,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;wBAAE,OAAO,KAAK,CAAC;YAChF,OAAO,IAAI,CAAC;QACd,CAAC,CAAA,CAAC;QAEF,IAAI,QAAQ,GAAe,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAEpD,6EAA6E;QAC7E,SAAS,CAAC;YACR,MAAM,UAAU,GAAiD,EAAE,CAAC;YACpE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACzC,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC7C,IAAI,EAAE,GAAG,QAAQ,CAAC;oBAClB,KAAK,MAAM,CAAC,IAAI,QAAQ,CAAC,CAAC,CAAC;wBAAE,KAAK,MAAM,CAAC,IAAI,QAAQ,CAAC,CAAC,CAAC;4BAAE,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;oBACvF,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,CAAC;gBACrC,CAAC;YACH,CAAC;YACD,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;YAEzC,IAAI,SAAS,GAAG,KAAK,CAAC;YACtB,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;gBAC3B,IAAI,MAAM,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;oBACpD,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;oBACpD,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,wCAAwC;oBACjE,SAAS,GAAG,IAAI,CAAC;oBACjB,MAAM,CAAC,qDAAqD;gBAC9D,CAAC;YACH,CAAC;YACD,IAAI,CAAC,SAAS;gBAAE,MAAM;QACxB,CAAC;QAED,MAAM,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;QAC5C,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;QAExE,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,MAAM,UAAU,GAAqB,EAAE,CAAC;QACxC,KAAK,MAAM,EAAE,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,IAAI,CAAC,GAAG,SAAS,EAAE,GAAG,aAAa,CAAC,EAAE,CAAC;YAChE,MAAM,GAAG,GAAG,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC;YACxB,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC5B,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACd,UAAU,CAAC,IAAI,CAAC;gBACd,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE;gBACd,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE;gBACd,GAAG,EAAE,CAAC;gBACN,MAAM,EAAE,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC;aAC9C,CAAC,CAAC;QACL,CAAC;QAED,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC;IACtF,CAAC;CAAA"}
@@ -0,0 +1,69 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.MODULE_MARKER = exports.REFERENCE_KIND_TO_PREDICATE = exports.SYMBOL_KIND_TO_ENTITY_TYPE = exports.MEMBER_KINDS = exports.DEFINITION_KINDS = exports.hashContent = void 0;
4
+ exports.shouldSeedSymbol = shouldSeedSymbol;
5
+ const outlion_1 = require("@wanshi-kg/outlion");
6
+ Object.defineProperty(exports, "hashContent", { enumerable: true, get: function () { return outlion_1.hashContent; } });
7
+ /**
8
+ * Mapping tables + filters for the AST symbol seed (Phase 8). Per the locked
9
+ * decision, every target type/predicate is an **existing** Phase-2 vocabulary
10
+ * member (`src/core/knowledge/vocabulary.ts`) — no `vocabulary.ts` / `system.hbs`
11
+ * extension — so seeded nodes/edges validate against the closed enum.
12
+ */
13
+ /** Structural definitions — always seeded as entities. */
14
+ exports.DEFINITION_KINDS = new Set([
15
+ "module",
16
+ "namespace",
17
+ "class",
18
+ "struct",
19
+ "interface",
20
+ "trait",
21
+ "enum",
22
+ "type_alias",
23
+ "function",
24
+ "method",
25
+ "constructor",
26
+ ]);
27
+ /** Fine-grained members — seeded only when `exported` (public surface; locals are noise). */
28
+ exports.MEMBER_KINDS = new Set([
29
+ "field",
30
+ "property",
31
+ "variable",
32
+ "constant",
33
+ "enum_member",
34
+ ]);
35
+ /** `SymbolKind` → an existing `BASE_ENTITY_TYPE`. */
36
+ exports.SYMBOL_KIND_TO_ENTITY_TYPE = {
37
+ module: "module",
38
+ namespace: "module",
39
+ class: "class",
40
+ struct: "class",
41
+ interface: "interface",
42
+ trait: "interface",
43
+ enum: "data_structure",
44
+ type_alias: "data_structure",
45
+ function: "function",
46
+ method: "function",
47
+ constructor: "function",
48
+ field: "term",
49
+ property: "term",
50
+ variable: "term",
51
+ constant: "term",
52
+ enum_member: "term",
53
+ };
54
+ /** `SymbolReference.kind` → an existing `BASE_RELATION_TYPE` predicate. */
55
+ exports.REFERENCE_KIND_TO_PREDICATE = {
56
+ calls: "calls",
57
+ imports: "depends_on",
58
+ };
59
+ /** The marker the Symbol API uses for a reference `from` at file/module top level. */
60
+ exports.MODULE_MARKER = "<module>";
61
+ /** Seed definitions always; members only when exported (covers all 16 kinds). */
62
+ function shouldSeedSymbol(symbol) {
63
+ if (exports.DEFINITION_KINDS.has(symbol.kind))
64
+ return true;
65
+ if (exports.MEMBER_KINDS.has(symbol.kind))
66
+ return symbol.exported;
67
+ return false;
68
+ }
69
+ //# sourceMappingURL=astSymbols.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"astSymbols.js","sourceRoot":"","sources":["../../../src/shared/utils/astSymbols.ts"],"names":[],"mappings":";;;AAwEA,4CAIC;AA5ED,gDAM4B;AAEnB,4FAHP,qBAAW,OAGO;AAGpB;;;;;GAKG;AAEH,0DAA0D;AAC7C,QAAA,gBAAgB,GAA4B,IAAI,GAAG,CAAa;IAC3E,QAAQ;IACR,WAAW;IACX,OAAO;IACP,QAAQ;IACR,WAAW;IACX,OAAO;IACP,MAAM;IACN,YAAY;IACZ,UAAU;IACV,QAAQ;IACR,aAAa;CACd,CAAC,CAAC;AAEH,6FAA6F;AAChF,QAAA,YAAY,GAA4B,IAAI,GAAG,CAAa;IACvE,OAAO;IACP,UAAU;IACV,UAAU;IACV,UAAU;IACV,aAAa;CACd,CAAC,CAAC;AAEH,qDAAqD;AACxC,QAAA,0BAA0B,GAA+B;IACpE,MAAM,EAAE,QAAQ;IAChB,SAAS,EAAE,QAAQ;IACnB,KAAK,EAAE,OAAO;IACd,MAAM,EAAE,OAAO;IACf,SAAS,EAAE,WAAW;IACtB,KAAK,EAAE,WAAW;IAClB,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,gBAAgB;IAC5B,QAAQ,EAAE,UAAU;IACpB,MAAM,EAAE,UAAU;IAClB,WAAW,EAAE,UAAU;IACvB,KAAK,EAAE,MAAM;IACb,QAAQ,EAAE,MAAM;IAChB,QAAQ,EAAE,MAAM;IAChB,QAAQ,EAAE,MAAM;IAChB,WAAW,EAAE,MAAM;CACpB,CAAC;AAEF,2EAA2E;AAC9D,QAAA,2BAA2B,GAA4C;IAClF,KAAK,EAAE,OAAO;IACd,OAAO,EAAE,YAAY;CACtB,CAAC;AAEF,sFAAsF;AACzE,QAAA,aAAa,GAAG,UAAU,CAAC;AAExC,iFAAiF;AACjF,SAAgB,gBAAgB,CAAC,MAAmB;IAClD,IAAI,wBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC;IACnD,IAAI,oBAAY,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC;QAAE,OAAO,MAAM,CAAC,QAAQ,CAAC;IAC1D,OAAO,KAAK,CAAC;AACf,CAAC"}
@@ -0,0 +1,18 @@
1
+ "use strict";
2
+ // Calculate cosine similarity between two vectors
3
+ Object.defineProperty(exports, "__esModule", { value: true });
4
+ exports.cosineSimilarity = cosineSimilarity;
5
+ function cosineSimilarity(a, b) {
6
+ if (a.length !== b.length)
7
+ return 0;
8
+ let dotProduct = 0;
9
+ let normA = 0;
10
+ let normB = 0;
11
+ for (let i = 0; i < a.length; i++) {
12
+ dotProduct += a[i] * b[i];
13
+ normA += a[i] * a[i];
14
+ normB += b[i] * b[i];
15
+ }
16
+ return dotProduct / (Math.sqrt(normA) * Math.sqrt(normB));
17
+ }
18
+ //# sourceMappingURL=cosineSimilarity.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"cosineSimilarity.js","sourceRoot":"","sources":["../../../src/shared/utils/cosineSimilarity.ts"],"names":[],"mappings":";AAAA,kDAAkD;;AAElD,4CAcC;AAdD,SAAgB,gBAAgB,CAAC,CAAW,EAAE,CAAW;IACvD,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM;QAAE,OAAO,CAAC,CAAC;IAEpC,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClC,UAAU,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAC1B,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACrB,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;IAED,OAAO,UAAU,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;AAC5D,CAAC"}