@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 +1,150 @@
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 __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
14
+ Object.defineProperty(o, "default", { enumerable: true, value: v });
15
+ }) : function(o, v) {
16
+ o["default"] = v;
17
+ });
18
+ var __importStar = (this && this.__importStar) || (function () {
19
+ var ownKeys = function(o) {
20
+ ownKeys = Object.getOwnPropertyNames || function (o) {
21
+ var ar = [];
22
+ for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
23
+ return ar;
24
+ };
25
+ return ownKeys(o);
26
+ };
27
+ return function (mod) {
28
+ if (mod && mod.__esModule) return mod;
29
+ var result = {};
30
+ if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
31
+ __setModuleDefault(result, mod);
32
+ return result;
33
+ };
34
+ })();
35
+ var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
36
+ function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
37
+ return new (P || (P = Promise))(function (resolve, reject) {
38
+ function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
39
+ function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
40
+ function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
41
+ step((generator = generator.apply(thisArg, _arguments || [])).next());
42
+ });
43
+ };
44
+ var __asyncValues = (this && this.__asyncValues) || function (o) {
45
+ if (!Symbol.asyncIterator) throw new TypeError("Symbol.asyncIterator is not defined.");
46
+ var m = o[Symbol.asyncIterator], i;
47
+ return m ? m.call(o) : (o = typeof __values === "function" ? __values(o) : o[Symbol.iterator](), i = {}, verb("next"), verb("throw"), verb("return"), i[Symbol.asyncIterator] = function () { return this; }, i);
48
+ function verb(n) { i[n] = o[n] && function (v) { return new Promise(function (resolve, reject) { v = o[n](v), settle(resolve, reject, v.done, v.value); }); }; }
49
+ function settle(resolve, reject, d, v) { Promise.resolve(v).then(function(v) { resolve({ value: v, done: d }); }, reject); }
50
+ };
51
+ Object.defineProperty(exports, "__esModule", { value: true });
52
+ exports.SemEval2010Dataset = exports.SEMEVAL_LABELS = void 0;
53
+ const fs = __importStar(require("fs"));
54
+ const readline = __importStar(require("readline"));
55
+ /** The 19 ClassLabel names, in index order (HF schema). Used to resolve an
56
+ * integer `relation` to its string label; canonical and fixed. */
57
+ exports.SEMEVAL_LABELS = [
58
+ 'Cause-Effect(e1,e2)', 'Cause-Effect(e2,e1)',
59
+ 'Component-Whole(e1,e2)', 'Component-Whole(e2,e1)',
60
+ 'Content-Container(e1,e2)', 'Content-Container(e2,e1)',
61
+ 'Entity-Destination(e1,e2)', 'Entity-Destination(e2,e1)',
62
+ 'Entity-Origin(e1,e2)', 'Entity-Origin(e2,e1)',
63
+ 'Instrument-Agency(e1,e2)', 'Instrument-Agency(e2,e1)',
64
+ 'Member-Collection(e1,e2)', 'Member-Collection(e2,e1)',
65
+ 'Message-Topic(e1,e2)', 'Message-Topic(e2,e1)',
66
+ 'Product-Producer(e1,e2)', 'Product-Producer(e2,e1)',
67
+ 'Other',
68
+ ];
69
+ class SemEval2010Dataset {
70
+ load(dataPath, limit, domain) {
71
+ return __awaiter(this, void 0, void 0, function* () {
72
+ var _a, e_1, _b, _c;
73
+ if (!fs.existsSync(dataPath)) {
74
+ throw new Error(`SemEval-2010 Task 8 data not found at: ${dataPath}\n` +
75
+ `Fetch it first: npx ts-node scripts/fetch-semeval.ts\n` +
76
+ `(downloads SemEvalWorkshop/sem_eval_2010_task_8 → data/semeval/{train,test}.jsonl)`);
77
+ }
78
+ const samples = [];
79
+ const stream = fs.createReadStream(dataPath, { encoding: 'utf-8' });
80
+ const rl = readline.createInterface({ input: stream, crlfDelay: Infinity });
81
+ try {
82
+ for (var _d = true, rl_1 = __asyncValues(rl), rl_1_1; rl_1_1 = yield rl_1.next(), _a = rl_1_1.done, !_a; _d = true) {
83
+ _c = rl_1_1.value;
84
+ _d = false;
85
+ const line = _c;
86
+ if (samples.length >= limit)
87
+ break;
88
+ const trimmed = line.trim();
89
+ if (!trimmed)
90
+ continue;
91
+ let raw;
92
+ try {
93
+ raw = JSON.parse(trimmed);
94
+ }
95
+ catch (_e) {
96
+ continue;
97
+ }
98
+ const sample = SemEval2010Dataset.parseInstance(raw, samples.length);
99
+ if (sample)
100
+ samples.push(sample);
101
+ }
102
+ }
103
+ catch (e_1_1) { e_1 = { error: e_1_1 }; }
104
+ finally {
105
+ try {
106
+ if (!_d && !_a && (_b = rl_1.return)) yield _b.call(rl_1);
107
+ }
108
+ finally { if (e_1) throw e_1.error; }
109
+ }
110
+ return samples;
111
+ });
112
+ }
113
+ /** Parse one raw line into a BenchmarkSample (or null when malformed). Static
114
+ * + exported-via-class so the unit test can exercise it without a file. */
115
+ static parseInstance(raw, index) {
116
+ var _a, _b, _c, _d;
117
+ const sentence = raw.sentence;
118
+ if (!sentence)
119
+ return null;
120
+ const e1 = (_b = (_a = sentence.match(/<e1>([\s\S]*?)<\/e1>/)) === null || _a === void 0 ? void 0 : _a[1]) === null || _b === void 0 ? void 0 : _b.trim();
121
+ const e2 = (_d = (_c = sentence.match(/<e2>([\s\S]*?)<\/e2>/)) === null || _c === void 0 ? void 0 : _c[1]) === null || _d === void 0 ? void 0 : _d.trim();
122
+ if (!e1 || !e2)
123
+ return null;
124
+ const label = SemEval2010Dataset.resolveLabel(raw.relation);
125
+ if (!label)
126
+ return null;
127
+ // "Cause-Effect(e2,e1)" → base "Cause-Effect", reversed = true.
128
+ const base = label.replace(/\(e[12],e[12]\)\s*$/, '');
129
+ const reversed = /\(e2,e1\)\s*$/.test(label);
130
+ const [subject, object] = reversed ? [e2, e1] : [e1, e2];
131
+ // Strip the entity markup (the model must not see annotation tags) and any
132
+ // wrapping quotes the original format left on the sentence.
133
+ const text = sentence
134
+ .replace(/<\/?e[12]>/g, '')
135
+ .replace(/^\s*"|"\s*$/g, '')
136
+ .trim();
137
+ const groundTruth = [{ subject, predicate: base, object }];
138
+ return { id: `semeval-${index}`, text, groundTruth, domain: 'semeval' };
139
+ }
140
+ static resolveLabel(relation) {
141
+ var _a;
142
+ if (typeof relation === 'number')
143
+ return (_a = exports.SEMEVAL_LABELS[relation]) !== null && _a !== void 0 ? _a : null;
144
+ if (typeof relation === 'string' && relation.length > 0)
145
+ return relation;
146
+ return null;
147
+ }
148
+ }
149
+ exports.SemEval2010Dataset = SemEval2010Dataset;
150
+ //# sourceMappingURL=SemEval2010Dataset.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"SemEval2010Dataset.js","sourceRoot":"","sources":["../../../src/evaluation/datasets/SemEval2010Dataset.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,uCAAyB;AACzB,mDAAqC;AAqBrC;mEACmE;AACtD,QAAA,cAAc,GAAG;IAC5B,qBAAqB,EAAM,qBAAqB;IAChD,wBAAwB,EAAG,wBAAwB;IACnD,0BAA0B,EAAC,0BAA0B;IACrD,2BAA2B,EAAC,2BAA2B;IACvD,sBAAsB,EAAK,sBAAsB;IACjD,0BAA0B,EAAC,0BAA0B;IACrD,0BAA0B,EAAC,0BAA0B;IACrD,sBAAsB,EAAK,sBAAsB;IACjD,yBAAyB,EAAE,yBAAyB;IACpD,OAAO;CACR,CAAC;AAEF,MAAa,kBAAkB;IACvB,IAAI,CAAC,QAAgB,EAAE,KAAa,EAAE,MAAe;;;YACzD,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;gBAC7B,MAAM,IAAI,KAAK,CACb,0CAA0C,QAAQ,IAAI;oBACtD,yDAAyD;oBACzD,oFAAoF,CACrF,CAAC;YACJ,CAAC;YAED,MAAM,OAAO,GAAsB,EAAE,CAAC;YACtC,MAAM,MAAM,GAAG,EAAE,CAAC,gBAAgB,CAAC,QAAQ,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC,CAAC;YACpE,MAAM,EAAE,GAAG,QAAQ,CAAC,eAAe,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC,CAAC;;gBAE5E,KAAyB,eAAA,OAAA,cAAA,EAAE,CAAA,QAAA,gEAAE,CAAC;oBAAL,kBAAE;oBAAF,WAAE;oBAAhB,MAAM,IAAI,KAAA,CAAA;oBACnB,IAAI,OAAO,CAAC,MAAM,IAAI,KAAK;wBAAE,MAAM;oBACnC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;oBAC5B,IAAI,CAAC,OAAO;wBAAE,SAAS;oBAEvB,IAAI,GAAmB,CAAC;oBACxB,IAAI,CAAC;wBACH,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;oBAC5B,CAAC;oBAAC,WAAM,CAAC;wBACP,SAAS;oBACX,CAAC;oBAED,MAAM,MAAM,GAAG,kBAAkB,CAAC,aAAa,CAAC,GAAG,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;oBACrE,IAAI,MAAM;wBAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBACnC,CAAC;;;;;;;;;YAED,OAAO,OAAO,CAAC;QACjB,CAAC;KAAA;IAED;gFAC4E;IAC5E,MAAM,CAAC,aAAa,CAAC,GAAmB,EAAE,KAAa;;QACrD,MAAM,QAAQ,GAAG,GAAG,CAAC,QAAQ,CAAC;QAC9B,IAAI,CAAC,QAAQ;YAAE,OAAO,IAAI,CAAC;QAE3B,MAAM,EAAE,GAAG,MAAA,MAAA,QAAQ,CAAC,KAAK,CAAC,sBAAsB,CAAC,0CAAG,CAAC,CAAC,0CAAE,IAAI,EAAE,CAAC;QAC/D,MAAM,EAAE,GAAG,MAAA,MAAA,QAAQ,CAAC,KAAK,CAAC,sBAAsB,CAAC,0CAAG,CAAC,CAAC,0CAAE,IAAI,EAAE,CAAC;QAC/D,IAAI,CAAC,EAAE,IAAI,CAAC,EAAE;YAAE,OAAO,IAAI,CAAC;QAE5B,MAAM,KAAK,GAAG,kBAAkB,CAAC,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC5D,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,CAAC;QAExB,gEAAgE;QAChE,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,qBAAqB,EAAE,EAAE,CAAC,CAAC;QACtD,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC7C,MAAM,CAAC,OAAO,EAAE,MAAM,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;QAEzD,2EAA2E;QAC3E,4DAA4D;QAC5D,MAAM,IAAI,GAAG,QAAQ;aAClB,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC;aAC1B,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC;aAC3B,IAAI,EAAE,CAAC;QAEV,MAAM,WAAW,GAAc,CAAC,EAAE,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAC;QACtE,OAAO,EAAE,EAAE,EAAE,WAAW,KAAK,EAAE,EAAE,IAAI,EAAE,WAAW,EAAE,MAAM,EAAE,SAAS,EAAE,CAAC;IAC1E,CAAC;IAEO,MAAM,CAAC,YAAY,CAAC,QAAqC;;QAC/D,IAAI,OAAO,QAAQ,KAAK,QAAQ;YAAE,OAAO,MAAA,sBAAc,CAAC,QAAQ,CAAC,mCAAI,IAAI,CAAC;QAC1E,IAAI,OAAO,QAAQ,KAAK,QAAQ,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,QAAQ,CAAC;QACzE,OAAO,IAAI,CAAC;IACd,CAAC;CACF;AAnED,gDAmEC"}
@@ -0,0 +1,33 @@
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("./datasets/IDataset"), exports);
18
+ __exportStar(require("./datasets/RebelDataset"), exports);
19
+ __exportStar(require("./datasets/CrossREDataset"), exports);
20
+ __exportStar(require("./datasets/RedocredDataset"), exports);
21
+ __exportStar(require("./datasets/SemEval2010Dataset"), exports);
22
+ __exportStar(require("./matching/ExactMatcher"), exports);
23
+ __exportStar(require("./matching/SemanticMatcher"), exports);
24
+ __exportStar(require("./metrics/TripleMetrics"), exports);
25
+ __exportStar(require("./reporters/ConsoleReporter"), exports);
26
+ __exportStar(require("./reporters/JsonReporter"), exports);
27
+ __exportStar(require("./BenchmarkRunner"), exports);
28
+ __exportStar(require("./mine/types"), exports);
29
+ __exportStar(require("./mine/MineDataset"), exports);
30
+ __exportStar(require("./mine/MineScorer"), exports);
31
+ __exportStar(require("./mine/MineRunner"), exports);
32
+ __exportStar(require("./mine/MineReporter"), exports);
33
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/evaluation/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,sDAAoC;AACpC,0DAAwC;AACxC,4DAA0C;AAC1C,6DAA2C;AAC3C,gEAA8C;AAC9C,0DAAwC;AACxC,6DAA2C;AAC3C,0DAAwC;AACxC,8DAA4C;AAC5C,2DAAyC;AACzC,oDAAkC;AAClC,+CAA6B;AAC7B,qDAAmC;AACnC,oDAAkC;AAClC,oDAAkC;AAClC,sDAAoC"}
@@ -0,0 +1,75 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.ExactMatcher = void 0;
4
+ class ExactMatcher {
5
+ normalize(s) {
6
+ return s.toLowerCase().replace(/[^a-z0-9\s]/g, '').trim();
7
+ }
8
+ /**
9
+ * Returns the set of extracted triplets that have at least one match
10
+ * in groundTruth (true positives), and counts FP/FN.
11
+ */
12
+ matchTriplets(extracted, groundTruth) {
13
+ const gtNorm = groundTruth.map(t => this.normalizeTriplet(t));
14
+ const exNorm = extracted.map(t => this.normalizeTriplet(t));
15
+ let tp = 0;
16
+ const matchedGT = new Set();
17
+ for (const ex of exNorm) {
18
+ const idx = gtNorm.findIndex((gt, i) => !matchedGT.has(i) && this.tripletsEqual(ex, gt));
19
+ if (idx !== -1) {
20
+ tp++;
21
+ matchedGT.add(idx);
22
+ }
23
+ }
24
+ return { tp, fp: extracted.length - tp, fn: groundTruth.length - tp };
25
+ }
26
+ /**
27
+ * Matches entity sets (subjects ∪ objects), ignoring predicates.
28
+ */
29
+ matchEntities(extracted, groundTruth) {
30
+ const gtEntities = this.extractEntitySet(groundTruth).map(e => this.normalize(e));
31
+ const exEntities = this.extractEntitySet(extracted).map(e => this.normalize(e));
32
+ return this.matchSets(exEntities, gtEntities);
33
+ }
34
+ /**
35
+ * Matches predicate sets only.
36
+ */
37
+ matchRelations(extracted, groundTruth) {
38
+ const gtRels = groundTruth.map(t => this.normalize(t.predicate));
39
+ const exRels = extracted.map(t => this.normalize(t.predicate));
40
+ return this.matchSets(exRels, gtRels);
41
+ }
42
+ // ─── Private helpers ──────────────────────────────────────────────────────
43
+ normalizeTriplet(t) {
44
+ return {
45
+ subject: this.normalize(t.subject),
46
+ predicate: this.normalize(t.predicate),
47
+ object: this.normalize(t.object),
48
+ };
49
+ }
50
+ tripletsEqual(a, b) {
51
+ return a.subject === b.subject && a.predicate === b.predicate && a.object === b.object;
52
+ }
53
+ extractEntitySet(triplets) {
54
+ const set = new Set();
55
+ for (const t of triplets) {
56
+ set.add(t.subject);
57
+ set.add(t.object);
58
+ }
59
+ return Array.from(set);
60
+ }
61
+ matchSets(extracted, groundTruth) {
62
+ let tp = 0;
63
+ const matchedGT = new Set();
64
+ for (const ex of extracted) {
65
+ const idx = groundTruth.findIndex((gt, i) => !matchedGT.has(i) && gt === ex);
66
+ if (idx !== -1) {
67
+ tp++;
68
+ matchedGT.add(idx);
69
+ }
70
+ }
71
+ return { tp, fp: extracted.length - tp, fn: groundTruth.length - tp };
72
+ }
73
+ }
74
+ exports.ExactMatcher = ExactMatcher;
75
+ //# sourceMappingURL=ExactMatcher.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"ExactMatcher.js","sourceRoot":"","sources":["../../../src/evaluation/matching/ExactMatcher.ts"],"names":[],"mappings":";;;AAEA,MAAa,YAAY;IACvB,SAAS,CAAC,CAAS;QACjB,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAC5D,CAAC;IAED;;;OAGG;IACH,aAAa,CAAC,SAAoB,EAAE,WAAsB;QAGxD,MAAM,MAAM,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9D,MAAM,MAAM,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC,CAAC,CAAC;QAE5D,IAAI,EAAE,GAAG,CAAC,CAAC;QACX,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;QAEpC,KAAK,MAAM,EAAE,IAAI,MAAM,EAAE,CAAC;YACxB,MAAM,GAAG,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,aAAa,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC;YACzF,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;gBACf,EAAE,EAAE,CAAC;gBACL,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACrB,CAAC;QACH,CAAC;QAED,OAAO,EAAE,EAAE,EAAE,EAAE,EAAE,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,EAAE,EAAE,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;IACxE,CAAC;IAED;;OAEG;IACH,aAAa,CAAC,SAAoB,EAAE,WAAsB;QAGxD,MAAM,UAAU,GAAG,IAAI,CAAC,gBAAgB,CAAC,WAAW,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;QAClF,MAAM,UAAU,GAAG,IAAI,CAAC,gBAAgB,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;QAChF,OAAO,IAAI,CAAC,SAAS,CAAC,UAAU,EAAE,UAAU,CAAC,CAAC;IAChD,CAAC;IAED;;OAEG;IACH,cAAc,CAAC,SAAoB,EAAE,WAAsB;QAGzD,MAAM,MAAM,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC;QACjE,MAAM,MAAM,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC;QAC/D,OAAO,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACxC,CAAC;IAED,6EAA6E;IAErE,gBAAgB,CAAC,CAAU;QACjC,OAAO;YACL,OAAO,EAAI,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC;YACpC,SAAS,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC;YACtC,MAAM,EAAK,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;SACpC,CAAC;IACJ,CAAC;IAEO,aAAa,CAAC,CAAU,EAAE,CAAU;QAC1C,OAAO,CAAC,CAAC,OAAO,KAAK,CAAC,CAAC,OAAO,IAAI,CAAC,CAAC,SAAS,KAAK,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM,CAAC;IACzF,CAAC;IAEO,gBAAgB,CAAC,QAAmB;QAC1C,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;QAC9B,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;YACzB,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YACnB,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QACpB,CAAC;QACD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACzB,CAAC;IAEO,SAAS,CAAC,SAAmB,EAAE,WAAqB;QAG1D,IAAI,EAAE,GAAG,CAAC,CAAC;QACX,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;QAEpC,KAAK,MAAM,EAAE,IAAI,SAAS,EAAE,CAAC;YAC3B,MAAM,GAAG,GAAG,WAAW,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;YAC7E,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;gBACf,EAAE,EAAE,CAAC;gBACL,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACrB,CAAC;QACH,CAAC;QAED,OAAO,EAAE,EAAE,EAAE,EAAE,EAAE,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,EAAE,EAAE,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;IACxE,CAAC;CACF;AA1FD,oCA0FC"}
@@ -0,0 +1,143 @@
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.SemanticMatcher = void 0;
13
+ const shared_1 = require("../../shared");
14
+ const ExactMatcher_1 = require("./ExactMatcher");
15
+ class SemanticMatcher {
16
+ // Depends only on the provider interface (embed/embedBatch), so the metrics
17
+ // command can pass whichever provider the container built (Ollama | OpenAI).
18
+ constructor(embeddingService, threshold = 0.80) {
19
+ this.embeddingService = embeddingService;
20
+ this.threshold = threshold;
21
+ this.exactMatcher = new ExactMatcher_1.ExactMatcher();
22
+ this.embeddingCache = new Map();
23
+ }
24
+ /**
25
+ * Pre-warm the embedding cache for a batch of strings.
26
+ * Call this before running matchTriplets on many samples to batch
27
+ * the embedding calls efficiently.
28
+ */
29
+ warmCache(strings) {
30
+ return __awaiter(this, void 0, void 0, function* () {
31
+ const unique = Array.from(new Set(strings)).filter(s => !this.embeddingCache.has(s));
32
+ if (unique.length === 0)
33
+ return;
34
+ const embeddings = yield this.embeddingService.embedBatch(unique);
35
+ for (let i = 0; i < unique.length; i++) {
36
+ this.embeddingCache.set(unique[i], embeddings[i]);
37
+ }
38
+ });
39
+ }
40
+ matchTriplets(extracted, groundTruth) {
41
+ return __awaiter(this, void 0, void 0, function* () {
42
+ // Warm cache for all strings — subjects, predicates, and objects
43
+ const allStrings = [
44
+ ...extracted.flatMap(t => [t.subject, t.predicate, t.object]),
45
+ ...groundTruth.flatMap(t => [t.subject, t.predicate, t.object]),
46
+ ];
47
+ yield this.warmCache(allStrings);
48
+ let tp = 0;
49
+ const matchedGT = new Set();
50
+ for (const ex of extracted) {
51
+ const idx = yield this.findMatch(ex, groundTruth, matchedGT);
52
+ if (idx !== -1) {
53
+ tp++;
54
+ matchedGT.add(idx);
55
+ }
56
+ }
57
+ return { tp, fp: extracted.length - tp, fn: groundTruth.length - tp };
58
+ });
59
+ }
60
+ matchEntities(extracted, groundTruth) {
61
+ return __awaiter(this, void 0, void 0, function* () {
62
+ const exEntities = this.uniqueEntities(extracted);
63
+ const gtEntities = this.uniqueEntities(groundTruth);
64
+ yield this.warmCache([...exEntities, ...gtEntities]);
65
+ let tp = 0;
66
+ const matchedGT = new Set();
67
+ for (const ex of exEntities) {
68
+ const idx = gtEntities.findIndex((gt, i) => {
69
+ if (matchedGT.has(i))
70
+ return false;
71
+ return this.stringSimilar(ex, gt);
72
+ });
73
+ if (idx !== -1) {
74
+ tp++;
75
+ matchedGT.add(idx);
76
+ }
77
+ }
78
+ return { tp, fp: exEntities.length - tp, fn: gtEntities.length - tp };
79
+ });
80
+ }
81
+ /**
82
+ * Semantically match predicate sets. Uses embedding similarity rather than
83
+ * exact string matching — critical when the extractor uses free-form labels
84
+ * (e.g. "defeated") while ground truth uses a fixed taxonomy ("win-defeat").
85
+ */
86
+ matchRelations(extracted, groundTruth) {
87
+ return __awaiter(this, void 0, void 0, function* () {
88
+ const exPreds = extracted.map(t => t.predicate);
89
+ const gtPreds = groundTruth.map(t => t.predicate);
90
+ yield this.warmCache([...exPreds, ...gtPreds]);
91
+ let tp = 0;
92
+ const matchedGT = new Set();
93
+ for (const ex of exPreds) {
94
+ const idx = gtPreds.findIndex((gt, i) => {
95
+ if (matchedGT.has(i))
96
+ return false;
97
+ return this.stringSimilar(ex, gt);
98
+ });
99
+ if (idx !== -1) {
100
+ tp++;
101
+ matchedGT.add(idx);
102
+ }
103
+ }
104
+ return { tp, fp: exPreds.length - tp, fn: gtPreds.length - tp };
105
+ });
106
+ }
107
+ // ─── Private helpers ──────────────────────────────────────────────────────
108
+ findMatch(ex, groundTruth, matchedGT) {
109
+ return __awaiter(this, void 0, void 0, function* () {
110
+ for (let i = 0; i < groundTruth.length; i++) {
111
+ if (matchedGT.has(i))
112
+ continue;
113
+ const gt = groundTruth[i];
114
+ // All three components: semantic similarity
115
+ if (this.stringSimilar(ex.subject, gt.subject) &&
116
+ this.stringSimilar(ex.predicate, gt.predicate) &&
117
+ this.stringSimilar(ex.object, gt.object)) {
118
+ return i;
119
+ }
120
+ }
121
+ return -1;
122
+ });
123
+ }
124
+ stringSimilar(a, b) {
125
+ const aVec = this.embeddingCache.get(a);
126
+ const bVec = this.embeddingCache.get(b);
127
+ if (!aVec || !bVec) {
128
+ // Fall back to exact match if embeddings not available
129
+ return this.exactMatcher.normalize(a) === this.exactMatcher.normalize(b);
130
+ }
131
+ return (0, shared_1.cosineSimilarity)(aVec, bVec) >= this.threshold;
132
+ }
133
+ uniqueEntities(triplets) {
134
+ const set = new Set();
135
+ for (const t of triplets) {
136
+ set.add(t.subject);
137
+ set.add(t.object);
138
+ }
139
+ return Array.from(set);
140
+ }
141
+ }
142
+ exports.SemanticMatcher = SemanticMatcher;
143
+ //# sourceMappingURL=SemanticMatcher.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"SemanticMatcher.js","sourceRoot":"","sources":["../../../src/evaluation/matching/SemanticMatcher.ts"],"names":[],"mappings":";;;;;;;;;;;;AAAA,yCAAgD;AAEhD,iDAA8C;AAG9C,MAAa,eAAe;IAI1B,4EAA4E;IAC5E,6EAA6E;IAC7E,YACU,gBAAoC,EACpC,YAAoB,IAAI;QADxB,qBAAgB,GAAhB,gBAAgB,CAAoB;QACpC,cAAS,GAAT,SAAS,CAAe;QAP1B,iBAAY,GAAG,IAAI,2BAAY,EAAE,CAAC;QAClC,mBAAc,GAA0B,IAAI,GAAG,EAAE,CAAC;IAOvD,CAAC;IAEJ;;;;OAIG;IACG,SAAS,CAAC,OAAiB;;YAC/B,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;YACrF,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;gBAAE,OAAO;YAChC,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,gBAAgB,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;YAClE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACvC,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;YACpD,CAAC;QACH,CAAC;KAAA;IAEK,aAAa,CAAC,SAAoB,EAAE,WAAsB;;YAG9D,iEAAiE;YACjE,MAAM,UAAU,GAAG;gBACjB,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC;gBAC7D,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC;aAChE,CAAC;YACF,MAAM,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC;YAEjC,IAAI,EAAE,GAAG,CAAC,CAAC;YACX,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;YAEpC,KAAK,MAAM,EAAE,IAAI,SAAS,EAAE,CAAC;gBAC3B,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,EAAE,EAAE,WAAW,EAAE,SAAS,CAAC,CAAC;gBAC7D,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;oBACf,EAAE,EAAE,CAAC;oBACL,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;gBACrB,CAAC;YACH,CAAC;YAED,OAAO,EAAE,EAAE,EAAE,EAAE,EAAE,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,EAAE,EAAE,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QACxE,CAAC;KAAA;IAEK,aAAa,CAAC,SAAoB,EAAE,WAAsB;;YAG9D,MAAM,UAAU,GAAG,IAAI,CAAC,cAAc,CAAC,SAAS,CAAC,CAAC;YAClD,MAAM,UAAU,GAAG,IAAI,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC;YACpD,MAAM,IAAI,CAAC,SAAS,CAAC,CAAC,GAAG,UAAU,EAAE,GAAG,UAAU,CAAC,CAAC,CAAC;YAErD,IAAI,EAAE,GAAG,CAAC,CAAC;YACX,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;YAEpC,KAAK,MAAM,EAAE,IAAI,UAAU,EAAE,CAAC;gBAC5B,MAAM,GAAG,GAAG,UAAU,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE;oBACzC,IAAI,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC;wBAAE,OAAO,KAAK,CAAC;oBACnC,OAAO,IAAI,CAAC,aAAa,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;gBACpC,CAAC,CAAC,CAAC;gBACH,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;oBACf,EAAE,EAAE,CAAC;oBACL,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;gBACrB,CAAC;YACH,CAAC;YAED,OAAO,EAAE,EAAE,EAAE,EAAE,EAAE,UAAU,CAAC,MAAM,GAAG,EAAE,EAAE,EAAE,EAAE,UAAU,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QACxE,CAAC;KAAA;IAED;;;;OAIG;IACG,cAAc,CAAC,SAAoB,EAAE,WAAsB;;YAG/D,MAAM,OAAO,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;YAChD,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;YAClD,MAAM,IAAI,CAAC,SAAS,CAAC,CAAC,GAAG,OAAO,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC;YAE/C,IAAI,EAAE,GAAG,CAAC,CAAC;YACX,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;YAEpC,KAAK,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;gBACzB,MAAM,GAAG,GAAG,OAAO,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE;oBACtC,IAAI,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC;wBAAE,OAAO,KAAK,CAAC;oBACnC,OAAO,IAAI,CAAC,aAAa,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;gBACpC,CAAC,CAAC,CAAC;gBACH,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;oBACf,EAAE,EAAE,CAAC;oBACL,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;gBACrB,CAAC;YACH,CAAC;YAED,OAAO,EAAE,EAAE,EAAE,EAAE,EAAE,OAAO,CAAC,MAAM,GAAG,EAAE,EAAE,EAAE,EAAE,OAAO,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QAClE,CAAC;KAAA;IAED,6EAA6E;IAE/D,SAAS,CAAC,EAAW,EAAE,WAAsB,EAAE,SAAsB;;YACjF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC5C,IAAI,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC;oBAAE,SAAS;gBAC/B,MAAM,EAAE,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;gBAE1B,4CAA4C;gBAC5C,IACE,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC,OAAO,EAAI,EAAE,CAAC,OAAO,CAAC;oBAC5C,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC;oBAC9C,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC,MAAM,EAAK,EAAE,CAAC,MAAM,CAAC,EAC3C,CAAC;oBACD,OAAO,CAAC,CAAC;gBACX,CAAC;YACH,CAAC;YACD,OAAO,CAAC,CAAC,CAAC;QACZ,CAAC;KAAA;IAEO,aAAa,CAAC,CAAS,EAAE,CAAS;QACxC,MAAM,IAAI,GAAG,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QACxC,MAAM,IAAI,GAAG,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QACxC,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;YACnB,uDAAuD;YACvD,OAAO,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;QAC3E,CAAC;QACD,OAAO,IAAA,yBAAgB,EAAC,IAAI,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,SAAS,CAAC;IACxD,CAAC;IAEO,cAAc,CAAC,QAAmB;QACxC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;QAC9B,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;YACzB,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YACnB,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QACpB,CAAC;QACD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACzB,CAAC;CACF;AA3ID,0CA2IC"}
@@ -0,0 +1,64 @@
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.computeMetrics = computeMetrics;
13
+ exports.computeExactMetrics = computeExactMetrics;
14
+ exports.computeSemanticMetrics = computeSemanticMetrics;
15
+ exports.microAverage = microAverage;
16
+ function computeMetrics(tp, fp, fn) {
17
+ const precision = tp + fp > 0 ? tp / (tp + fp) : 0;
18
+ const recall = tp + fn > 0 ? tp / (tp + fn) : 0;
19
+ const f1 = precision + recall > 0 ? (2 * precision * recall) / (precision + recall) : 0;
20
+ return { precision, recall, f1, tp, fp, fn };
21
+ }
22
+ /**
23
+ * Compute exact-match P/R/F1 at entity, relation, and triple level.
24
+ */
25
+ function computeExactMetrics(extracted, groundTruth, matcher) {
26
+ const entity = computeMetrics(...Object.values(matcher.matchEntities(extracted, groundTruth)));
27
+ const relation = computeMetrics(...Object.values(matcher.matchRelations(extracted, groundTruth)));
28
+ const triple = computeMetrics(...Object.values(matcher.matchTriplets(extracted, groundTruth)));
29
+ return { entity, relation, triple };
30
+ }
31
+ /**
32
+ * Compute semantic P/R/F1 at entity, relation, and triple level.
33
+ */
34
+ function computeSemanticMetrics(extracted, groundTruth, matcher) {
35
+ return __awaiter(this, void 0, void 0, function* () {
36
+ const entityRaw = yield matcher.matchEntities(extracted, groundTruth);
37
+ const relationRaw = yield matcher.matchRelations(extracted, groundTruth);
38
+ const tripleRaw = yield matcher.matchTriplets(extracted, groundTruth);
39
+ return {
40
+ entity: computeMetrics(entityRaw.tp, entityRaw.fp, entityRaw.fn),
41
+ relation: computeMetrics(relationRaw.tp, relationRaw.fp, relationRaw.fn),
42
+ triple: computeMetrics(tripleRaw.tp, tripleRaw.fp, tripleRaw.fn),
43
+ };
44
+ });
45
+ }
46
+ /**
47
+ * Micro-average LevelMetrics across multiple samples.
48
+ */
49
+ function microAverage(results) {
50
+ if (results.length === 0) {
51
+ const zero = { precision: 0, recall: 0, f1: 0, tp: 0, fp: 0, fn: 0 };
52
+ return { entity: zero, relation: zero, triple: zero };
53
+ }
54
+ const levels = ['entity', 'relation', 'triple'];
55
+ const averaged = {};
56
+ for (const level of levels) {
57
+ const totalTP = results.reduce((s, r) => s + r[level].tp, 0);
58
+ const totalFP = results.reduce((s, r) => s + r[level].fp, 0);
59
+ const totalFN = results.reduce((s, r) => s + r[level].fn, 0);
60
+ averaged[level] = computeMetrics(totalTP, totalFP, totalFN);
61
+ }
62
+ return averaged;
63
+ }
64
+ //# sourceMappingURL=TripleMetrics.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"TripleMetrics.js","sourceRoot":"","sources":["../../../src/evaluation/metrics/TripleMetrics.ts"],"names":[],"mappings":";;;;;;;;;;;AAIA,wCAKC;AAKD,kDASC;AAKD,wDAcC;AAKD,oCAiBC;AA5DD,SAAgB,cAAc,CAAC,EAAU,EAAE,EAAU,EAAE,EAAU;IAC/D,MAAM,SAAS,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACnD,MAAM,MAAM,GAAM,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACnD,MAAM,EAAE,GAAG,SAAS,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,SAAS,GAAG,MAAM,CAAC,GAAG,CAAC,SAAS,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACxF,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC;AAC/C,CAAC;AAED;;GAEG;AACH,SAAgB,mBAAmB,CACjC,SAAoB,EACpB,WAAsB,EACtB,OAAqB;IAErB,MAAM,MAAM,GAAK,cAAc,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,SAAS,EAAE,WAAW,CAAC,CAA6B,CAAC,CAAC;IAC7H,MAAM,QAAQ,GAAG,cAAc,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,cAAc,CAAC,SAAS,EAAE,WAAW,CAAC,CAA6B,CAAC,CAAC;IAC9H,MAAM,MAAM,GAAK,cAAc,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,SAAS,EAAE,WAAW,CAAC,CAA6B,CAAC,CAAC;IAC7H,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC;AACtC,CAAC;AAED;;GAEG;AACH,SAAsB,sBAAsB,CAC1C,SAAoB,EACpB,WAAsB,EACtB,OAAwB;;QAExB,MAAM,SAAS,GAAK,MAAM,OAAO,CAAC,aAAa,CAAC,SAAS,EAAE,WAAW,CAAC,CAAC;QACxE,MAAM,WAAW,GAAG,MAAM,OAAO,CAAC,cAAc,CAAC,SAAS,EAAE,WAAW,CAAC,CAAC;QACzE,MAAM,SAAS,GAAK,MAAM,OAAO,CAAC,aAAa,CAAC,SAAS,EAAE,WAAW,CAAC,CAAC;QAExE,OAAO;YACL,MAAM,EAAI,cAAc,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,EAAE,CAAC;YAClE,QAAQ,EAAE,cAAc,CAAC,WAAW,CAAC,EAAE,EAAE,WAAW,CAAC,EAAE,EAAE,WAAW,CAAC,EAAE,CAAC;YACxE,MAAM,EAAI,cAAc,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,EAAE,CAAC;SACnE,CAAC;IACJ,CAAC;CAAA;AAED;;GAEG;AACH,SAAgB,YAAY,CAAC,OAAuB;IAClD,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACzB,MAAM,IAAI,GAAgB,EAAE,SAAS,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC;QAClF,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;IACxD,CAAC;IAED,MAAM,MAAM,GAA2B,CAAC,QAAQ,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;IACxE,MAAM,QAAQ,GAA0B,EAAE,CAAC;IAE3C,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;QAC7D,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;QAC7D,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;QAC7D,QAAQ,CAAC,KAAK,CAAC,GAAG,cAAc,CAAC,OAAO,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;IAC9D,CAAC;IAED,OAAO,QAAwB,CAAC;AAClC,CAAC"}
@@ -0,0 +1,114 @@
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 __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
14
+ Object.defineProperty(o, "default", { enumerable: true, value: v });
15
+ }) : function(o, v) {
16
+ o["default"] = v;
17
+ });
18
+ var __importStar = (this && this.__importStar) || (function () {
19
+ var ownKeys = function(o) {
20
+ ownKeys = Object.getOwnPropertyNames || function (o) {
21
+ var ar = [];
22
+ for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
23
+ return ar;
24
+ };
25
+ return ownKeys(o);
26
+ };
27
+ return function (mod) {
28
+ if (mod && mod.__esModule) return mod;
29
+ var result = {};
30
+ if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
31
+ __setModuleDefault(result, mod);
32
+ return result;
33
+ };
34
+ })();
35
+ var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
36
+ function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
37
+ return new (P || (P = Promise))(function (resolve, reject) {
38
+ function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
39
+ function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
40
+ function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
41
+ step((generator = generator.apply(thisArg, _arguments || [])).next());
42
+ });
43
+ };
44
+ Object.defineProperty(exports, "__esModule", { value: true });
45
+ exports.MineCheckpoint = void 0;
46
+ const fs = __importStar(require("fs"));
47
+ const path = __importStar(require("path"));
48
+ /**
49
+ * Append-only JSONL sidecar so a crash mid-sweep doesn't lose completed articles
50
+ * (the overnight run was wiped by an OS reboot). Mirrors FetchCacheService /
51
+ * CheckpointService: tolerant load (drops a truncated final line), parent-dir
52
+ * ensured on first append, in-memory index keyed by the run+article signature.
53
+ */
54
+ class MineCheckpoint {
55
+ constructor(checkpointPath, logger) {
56
+ this.records = new Map();
57
+ this.loaded = false;
58
+ this.dirEnsured = false;
59
+ this.path = checkpointPath;
60
+ this.logger = logger;
61
+ }
62
+ /** Load any existing checkpoint. Tolerant of a truncated final line. */
63
+ load() {
64
+ return __awaiter(this, void 0, void 0, function* () {
65
+ if (this.loaded)
66
+ return this.records.size;
67
+ this.loaded = true;
68
+ if (!fs.existsSync(this.path))
69
+ return 0;
70
+ const content = yield fs.promises.readFile(this.path, 'utf-8');
71
+ let loaded = 0;
72
+ let skipped = 0;
73
+ for (const line of content.split('\n')) {
74
+ const trimmed = line.trim();
75
+ if (!trimmed)
76
+ continue;
77
+ try {
78
+ const record = JSON.parse(trimmed);
79
+ if (record.key && record.result) {
80
+ this.records.set(record.key, record.result);
81
+ loaded++;
82
+ }
83
+ }
84
+ catch (_a) {
85
+ skipped++; // truncated/corrupt final line — drop it, it'll be recomputed
86
+ }
87
+ }
88
+ if (loaded > 0) {
89
+ this.logger.info(`Resuming MINE from ${loaded} checkpointed article(s) at ${this.path}` +
90
+ (skipped ? ` (skipped ${skipped} corrupt line(s))` : ''));
91
+ }
92
+ return loaded;
93
+ });
94
+ }
95
+ has(key) {
96
+ return this.records.has(key);
97
+ }
98
+ get(key) {
99
+ return this.records.get(key);
100
+ }
101
+ /** Persist one article's result and keep the in-memory index in sync. */
102
+ append(key, result) {
103
+ return __awaiter(this, void 0, void 0, function* () {
104
+ this.records.set(key, result);
105
+ if (!this.dirEnsured) {
106
+ yield fs.promises.mkdir(path.dirname(this.path), { recursive: true });
107
+ this.dirEnsured = true;
108
+ }
109
+ yield fs.promises.appendFile(this.path, JSON.stringify({ key, result }) + '\n');
110
+ });
111
+ }
112
+ }
113
+ exports.MineCheckpoint = MineCheckpoint;
114
+ //# sourceMappingURL=MineCheckpoint.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"MineCheckpoint.js","sourceRoot":"","sources":["../../../src/evaluation/mine/MineCheckpoint.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,uCAAyB;AACzB,2CAA6B;AAe7B;;;;;GAKG;AACH,MAAa,cAAc;IAOzB,YAAY,cAAsB,EAAE,MAAc;QAJ1C,YAAO,GAAmC,IAAI,GAAG,EAAE,CAAC;QACpD,WAAM,GAAG,KAAK,CAAC;QACf,eAAU,GAAG,KAAK,CAAC;QAGzB,IAAI,CAAC,IAAI,GAAG,cAAc,CAAC;QAC3B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED,wEAAwE;IAClE,IAAI;;YACR,IAAI,IAAI,CAAC,MAAM;gBAAE,OAAO,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC;YAC1C,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;YACnB,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC;gBAAE,OAAO,CAAC,CAAC;YAExC,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;YAC/D,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,IAAI,OAAO,GAAG,CAAC,CAAC;YAChB,KAAK,MAAM,IAAI,IAAI,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;gBACvC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;gBAC5B,IAAI,CAAC,OAAO;oBAAE,SAAS;gBACvB,IAAI,CAAC;oBACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAyB,CAAC;oBAC3D,IAAI,MAAM,CAAC,GAAG,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;wBAChC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC;wBAC5C,MAAM,EAAE,CAAC;oBACX,CAAC;gBACH,CAAC;gBAAC,WAAM,CAAC;oBACP,OAAO,EAAE,CAAC,CAAC,8DAA8D;gBAC3E,CAAC;YACH,CAAC;YACD,IAAI,MAAM,GAAG,CAAC,EAAE,CAAC;gBACf,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,sBAAsB,MAAM,+BAA+B,IAAI,CAAC,IAAI,EAAE;oBACpE,CAAC,OAAO,CAAC,CAAC,CAAC,aAAa,OAAO,mBAAmB,CAAC,CAAC,CAAC,EAAE,CAAC,CAC3D,CAAC;YACJ,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;KAAA;IAED,GAAG,CAAC,GAAW;QACb,OAAO,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC/B,CAAC;IAED,GAAG,CAAC,GAAW;QACb,OAAO,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC/B,CAAC;IAED,yEAAyE;IACnE,MAAM,CAAC,GAAW,EAAE,MAAyB;;YACjD,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;YAC9B,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;gBACrB,MAAM,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;gBACtE,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC;YACzB,CAAC;YACD,MAAM,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,EAAE,MAAM,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC;QAClF,CAAC;KAAA;CACF;AA5DD,wCA4DC"}