@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.
- package/LICENSE +21 -0
- package/README.md +458 -0
- package/dist/__tests__/helpers.js +27 -0
- package/dist/__tests__/helpers.js.map +1 -0
- package/dist/cli/commands/export.command.js +99 -0
- package/dist/cli/commands/export.command.js.map +1 -0
- package/dist/cli/commands/index.js +22 -0
- package/dist/cli/commands/index.js.map +1 -0
- package/dist/cli/commands/inspectMerges.command.js +84 -0
- package/dist/cli/commands/inspectMerges.command.js.map +1 -0
- package/dist/cli/commands/metrics.command.js +196 -0
- package/dist/cli/commands/metrics.command.js.map +1 -0
- package/dist/cli/commands/process.command.js +82 -0
- package/dist/cli/commands/process.command.js.map +1 -0
- package/dist/cli/commands/watch.command.js +91 -0
- package/dist/cli/commands/watch.command.js.map +1 -0
- package/dist/cli/index.js +269 -0
- package/dist/cli/index.js.map +1 -0
- package/dist/cli/optionsToConfig.js +160 -0
- package/dist/cli/optionsToConfig.js.map +1 -0
- package/dist/config/index.js +59 -0
- package/dist/config/index.js.map +1 -0
- package/dist/config/legacyHints.js +113 -0
- package/dist/config/legacyHints.js.map +1 -0
- package/dist/config/schema.js +803 -0
- package/dist/config/schema.js.map +1 -0
- package/dist/config/ui.js +221 -0
- package/dist/config/ui.js.map +1 -0
- package/dist/core/DirectoryProcessor.js +725 -0
- package/dist/core/DirectoryProcessor.js.map +1 -0
- package/dist/core/adapters/IStructuredAdapter.js +3 -0
- package/dist/core/adapters/IStructuredAdapter.js.map +1 -0
- package/dist/core/adapters/SqliteAdapter.js +267 -0
- package/dist/core/adapters/SqliteAdapter.js.map +1 -0
- package/dist/core/adapters/StructuredAdapterRegistry.js +31 -0
- package/dist/core/adapters/StructuredAdapterRegistry.js.map +1 -0
- package/dist/core/adapters/index.js +20 -0
- package/dist/core/adapters/index.js.map +1 -0
- package/dist/core/checkpoint/CheckpointService.js +188 -0
- package/dist/core/checkpoint/CheckpointService.js.map +1 -0
- package/dist/core/checkpoint/index.js +18 -0
- package/dist/core/checkpoint/index.js.map +1 -0
- package/dist/core/corpus/CorpusAnalyzer.js +266 -0
- package/dist/core/corpus/CorpusAnalyzer.js.map +1 -0
- package/dist/core/corpus/CorpusProfileStore.js +92 -0
- package/dist/core/corpus/CorpusProfileStore.js.map +1 -0
- package/dist/core/corpus/index.js +21 -0
- package/dist/core/corpus/index.js.map +1 -0
- package/dist/core/corpus/normalizeGlossary.js +60 -0
- package/dist/core/corpus/normalizeGlossary.js.map +1 -0
- package/dist/core/corpus/relPath.js +52 -0
- package/dist/core/corpus/relPath.js.map +1 -0
- package/dist/core/corpus/termFrequency.js +86 -0
- package/dist/core/corpus/termFrequency.js.map +1 -0
- package/dist/core/cost/CostMeter.js +235 -0
- package/dist/core/cost/CostMeter.js.map +1 -0
- package/dist/core/cost/index.js +19 -0
- package/dist/core/cost/index.js.map +1 -0
- package/dist/core/cost/prices.js +38 -0
- package/dist/core/cost/prices.js.map +1 -0
- package/dist/core/cv/ObjectDetectionService.js +119 -0
- package/dist/core/cv/ObjectDetectionService.js.map +1 -0
- package/dist/core/di/ContainerFactory.js +670 -0
- package/dist/core/di/ContainerFactory.js.map +1 -0
- package/dist/core/di/DIContainer.js +103 -0
- package/dist/core/di/DIContainer.js.map +1 -0
- package/dist/core/di/index.js +19 -0
- package/dist/core/di/index.js.map +1 -0
- package/dist/core/errors/CustomErrors.js +342 -0
- package/dist/core/errors/CustomErrors.js.map +1 -0
- package/dist/core/errors/index.js +18 -0
- package/dist/core/errors/index.js.map +1 -0
- package/dist/core/export/KnowledgeGraphExportService.js +56 -0
- package/dist/core/export/KnowledgeGraphExportService.js.map +1 -0
- package/dist/core/export/index.js +19 -0
- package/dist/core/export/index.js.map +1 -0
- package/dist/core/export/strategies/GraphitiExportStrategy.js +115 -0
- package/dist/core/export/strategies/GraphitiExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/GraphvizDotExportStrategy.js +331 -0
- package/dist/core/export/strategies/GraphvizDotExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/IExportStrategy.js +3 -0
- package/dist/core/export/strategies/IExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/JsonExportStrategy.js +19 -0
- package/dist/core/export/strategies/JsonExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/JsonlExportStrategy.js +69 -0
- package/dist/core/export/strategies/JsonlExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/KblamExportStrategy.js +36 -0
- package/dist/core/export/strategies/KblamExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/LoraExportStrategy.js +46 -0
- package/dist/core/export/strategies/LoraExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/McpExportStrategy.js +67 -0
- package/dist/core/export/strategies/McpExportStrategy.js.map +1 -0
- package/dist/core/export/strategies/index.js +25 -0
- package/dist/core/export/strategies/index.js.map +1 -0
- package/dist/core/export/strategies/kbTriples.js +60 -0
- package/dist/core/export/strategies/kbTriples.js.map +1 -0
- package/dist/core/index.js +22 -0
- package/dist/core/index.js.map +1 -0
- package/dist/core/knowledge/KnowledgeGraphBuilder.js +627 -0
- package/dist/core/knowledge/KnowledgeGraphBuilder.js.map +1 -0
- package/dist/core/knowledge/MergeRecord.js +3 -0
- package/dist/core/knowledge/MergeRecord.js.map +1 -0
- package/dist/core/knowledge/canon/Canonicalizer.js +414 -0
- package/dist/core/knowledge/canon/Canonicalizer.js.map +1 -0
- package/dist/core/knowledge/canon/index.js +18 -0
- package/dist/core/knowledge/canon/index.js.map +1 -0
- package/dist/core/knowledge/contradiction/HeuristicContradictionChecker.js +92 -0
- package/dist/core/knowledge/contradiction/HeuristicContradictionChecker.js.map +1 -0
- package/dist/core/knowledge/contradiction/LlmContradictionChecker.js +52 -0
- package/dist/core/knowledge/contradiction/LlmContradictionChecker.js.map +1 -0
- package/dist/core/knowledge/contradiction/index.js +19 -0
- package/dist/core/knowledge/contradiction/index.js.map +1 -0
- package/dist/core/knowledge/grounding/KeywordGroundingChecker.js +33 -0
- package/dist/core/knowledge/grounding/KeywordGroundingChecker.js.map +1 -0
- package/dist/core/knowledge/grounding/MiniCheckGroundingChecker.js +82 -0
- package/dist/core/knowledge/grounding/MiniCheckGroundingChecker.js.map +1 -0
- package/dist/core/knowledge/grounding/index.js +20 -0
- package/dist/core/knowledge/grounding/index.js.map +1 -0
- package/dist/core/knowledge/grounding/verbalize.js +38 -0
- package/dist/core/knowledge/grounding/verbalize.js.map +1 -0
- package/dist/core/knowledge/images/imageMetaGraph.js +136 -0
- package/dist/core/knowledge/images/imageMetaGraph.js.map +1 -0
- package/dist/core/knowledge/index.js +20 -0
- package/dist/core/knowledge/index.js.map +1 -0
- package/dist/core/knowledge/merging/KnowledgeMerger.js +624 -0
- package/dist/core/knowledge/merging/KnowledgeMerger.js.map +1 -0
- package/dist/core/knowledge/references/ReferenceResolver.js +184 -0
- package/dist/core/knowledge/references/ReferenceResolver.js.map +1 -0
- package/dist/core/knowledge/references/citations/CitationEvidenceProcessor.js +401 -0
- package/dist/core/knowledge/references/citations/CitationEvidenceProcessor.js.map +1 -0
- package/dist/core/knowledge/references/citations/CitationResolver.js +95 -0
- package/dist/core/knowledge/references/citations/CitationResolver.js.map +1 -0
- package/dist/core/knowledge/references/citations/GrobidClient.js +143 -0
- package/dist/core/knowledge/references/citations/GrobidClient.js.map +1 -0
- package/dist/core/knowledge/references/citations/TitleIdResolver.js +101 -0
- package/dist/core/knowledge/references/citations/TitleIdResolver.js.map +1 -0
- package/dist/core/knowledge/references/web/FetchCacheService.js +114 -0
- package/dist/core/knowledge/references/web/FetchCacheService.js.map +1 -0
- package/dist/core/knowledge/references/web/GatedFetcher.js +228 -0
- package/dist/core/knowledge/references/web/GatedFetcher.js.map +1 -0
- package/dist/core/knowledge/references/web/WebReferenceProcessor.js +164 -0
- package/dist/core/knowledge/references/web/WebReferenceProcessor.js.map +1 -0
- package/dist/core/knowledge/search/KnowledgeGraphSearch.js +261 -0
- package/dist/core/knowledge/search/KnowledgeGraphSearch.js.map +1 -0
- package/dist/core/knowledge/vocabulary.js +162 -0
- package/dist/core/knowledge/vocabulary.js.map +1 -0
- package/dist/core/llm/EmbeddingService.js +113 -0
- package/dist/core/llm/EmbeddingService.js.map +1 -0
- package/dist/core/llm/OllamaService.js +146 -0
- package/dist/core/llm/OllamaService.js.map +1 -0
- package/dist/core/llm/OpenAICompatibleService.js +190 -0
- package/dist/core/llm/OpenAICompatibleService.js.map +1 -0
- package/dist/core/llm/OpenAIEmbeddingService.js +129 -0
- package/dist/core/llm/OpenAIEmbeddingService.js.map +1 -0
- package/dist/core/llm/embeddingUtils.js +25 -0
- package/dist/core/llm/embeddingUtils.js.map +1 -0
- package/dist/core/llm/index.js +23 -0
- package/dist/core/llm/index.js.map +1 -0
- package/dist/core/llm/prompts/PromptManager.js +388 -0
- package/dist/core/llm/prompts/PromptManager.js.map +1 -0
- package/dist/core/llm/prompts/PromptTemplateEngine.js +257 -0
- package/dist/core/llm/prompts/PromptTemplateEngine.js.map +1 -0
- package/dist/core/llm/prompts/templates/partials/examples/EXAMPLE_STYLE_GUIDE.md +84 -0
- package/dist/core/llm/prompts/templates/partials/examples/article.md +187 -0
- package/dist/core/llm/prompts/templates/partials/examples/code.md +229 -0
- package/dist/core/llm/prompts/templates/partials/examples/communication.md +205 -0
- package/dist/core/llm/prompts/templates/partials/examples/documentation.md +262 -0
- package/dist/core/llm/prompts/templates/partials/examples/financial.md +157 -0
- package/dist/core/llm/prompts/templates/partials/examples/legal.md +153 -0
- package/dist/core/llm/prompts/templates/partials/examples/logs.md +127 -0
- package/dist/core/llm/prompts/templates/partials/examples/medical.md +218 -0
- package/dist/core/llm/prompts/templates/partials/examples/notes.md +201 -0
- package/dist/core/llm/prompts/templates/partials/examples/research.md +208 -0
- package/dist/core/llm/prompts/templates/partials/examples/tabular.md +178 -0
- package/dist/core/llm/prompts/templates/partials/examples/transcript.md +204 -0
- package/dist/core/llm/prompts/templates/partials/retrieved-context.hbs +18 -0
- package/dist/core/llm/prompts/templates/v1/system.hbs +371 -0
- package/dist/core/llm/prompts/templates/v1/user.hbs +20 -0
- package/dist/core/llm/prompts/templates/v2/system.hbs +573 -0
- package/dist/core/llm/prompts/templates/v2/user.hbs +20 -0
- package/dist/core/llm/prompts/templates/v3/system.hbs +861 -0
- package/dist/core/llm/prompts/templates/v3/user.hbs +16 -0
- package/dist/core/llm/prompts/templates/v4/system.hbs +800 -0
- package/dist/core/llm/prompts/templates/v4/user.hbs +40 -0
- package/dist/core/llm/prompts/templates/v4.5/system.hbs +71 -0
- package/dist/core/llm/prompts/templates/v4.5/user.hbs +46 -0
- package/dist/core/llm/prompts/templates/v5/glossary/system.hbs +40 -0
- package/dist/core/llm/prompts/templates/v5/glossary/user.hbs +11 -0
- package/dist/core/llm/prompts/templates/v5/system.hbs +163 -0
- package/dist/core/llm/prompts/templates/v5/user.hbs +55 -0
- package/dist/core/pipeline/GroundingTransform.js +52 -0
- package/dist/core/pipeline/GroundingTransform.js.map +1 -0
- package/dist/core/pipeline/PipelineRunner.js +51 -0
- package/dist/core/pipeline/PipelineRunner.js.map +1 -0
- package/dist/core/pipeline/RelationFilterTransform.js +72 -0
- package/dist/core/pipeline/RelationFilterTransform.js.map +1 -0
- package/dist/core/pipeline/index.js +20 -0
- package/dist/core/pipeline/index.js.map +1 -0
- package/dist/core/processor/FileProcessor.js +184 -0
- package/dist/core/processor/FileProcessor.js.map +1 -0
- package/dist/core/processor/ProcessedRegistry.js +38 -0
- package/dist/core/processor/ProcessedRegistry.js.map +1 -0
- package/dist/core/processor/ast/AstSeedService.js +0 -0
- package/dist/core/processor/ast/AstSeedService.js.map +1 -0
- package/dist/core/processor/ast/AstSymbolStore.js +110 -0
- package/dist/core/processor/ast/AstSymbolStore.js.map +1 -0
- package/dist/core/processor/ast/index.js +19 -0
- package/dist/core/processor/ast/index.js.map +1 -0
- package/dist/core/processor/chunking/TextChunker.js +98 -0
- package/dist/core/processor/chunking/TextChunker.js.map +1 -0
- package/dist/core/processor/chunking/index.js +18 -0
- package/dist/core/processor/chunking/index.js.map +1 -0
- package/dist/core/processor/classifier/CONTENT_CLASSES.js +294 -0
- package/dist/core/processor/classifier/CONTENT_CLASSES.js.map +1 -0
- package/dist/core/processor/classifier/CascadeContentClassifier.js +107 -0
- package/dist/core/processor/classifier/CascadeContentClassifier.js.map +1 -0
- package/dist/core/processor/classifier/HeuristicContentClassifier.js +113 -0
- package/dist/core/processor/classifier/HeuristicContentClassifier.js.map +1 -0
- package/dist/core/processor/classifier/IContentTypeClassifier.js +3 -0
- package/dist/core/processor/classifier/IContentTypeClassifier.js.map +1 -0
- package/dist/core/processor/classifier/LlmContentClassifier.js +107 -0
- package/dist/core/processor/classifier/LlmContentClassifier.js.map +1 -0
- package/dist/core/processor/classifier/NER_DOMAIN_EXAMPLES.js +498 -0
- package/dist/core/processor/classifier/NER_DOMAIN_EXAMPLES.js.map +1 -0
- package/dist/core/processor/classifier/index.js +21 -0
- package/dist/core/processor/classifier/index.js.map +1 -0
- package/dist/core/processor/classifier/mergeClassifications.js +32 -0
- package/dist/core/processor/classifier/mergeClassifications.js.map +1 -0
- package/dist/core/processor/index.js +20 -0
- package/dist/core/processor/index.js.map +1 -0
- package/dist/core/processor/readers/AudioReader.js +462 -0
- package/dist/core/processor/readers/AudioReader.js.map +1 -0
- package/dist/core/processor/readers/BinaryReader.js +90 -0
- package/dist/core/processor/readers/BinaryReader.js.map +1 -0
- package/dist/core/processor/readers/ChandraPdfReader.js +187 -0
- package/dist/core/processor/readers/ChandraPdfReader.js.map +1 -0
- package/dist/core/processor/readers/ChatExportReader.js +365 -0
- package/dist/core/processor/readers/ChatExportReader.js.map +1 -0
- package/dist/core/processor/readers/DoclingReader.js +445 -0
- package/dist/core/processor/readers/DoclingReader.js.map +1 -0
- package/dist/core/processor/readers/EmailReader.js +259 -0
- package/dist/core/processor/readers/EmailReader.js.map +1 -0
- package/dist/core/processor/readers/EpubReader.js +175 -0
- package/dist/core/processor/readers/EpubReader.js.map +1 -0
- package/dist/core/processor/readers/FileReader.js +90 -0
- package/dist/core/processor/readers/FileReader.js.map +1 -0
- package/dist/core/processor/readers/FileReaderFactory.js +49 -0
- package/dist/core/processor/readers/FileReaderFactory.js.map +1 -0
- package/dist/core/processor/readers/HtmlReader.js +371 -0
- package/dist/core/processor/readers/HtmlReader.js.map +1 -0
- package/dist/core/processor/readers/ImageReader.js +162 -0
- package/dist/core/processor/readers/ImageReader.js.map +1 -0
- package/dist/core/processor/readers/JsonFileReader.js +232 -0
- package/dist/core/processor/readers/JsonFileReader.js.map +1 -0
- package/dist/core/processor/readers/JupyterReader.js +178 -0
- package/dist/core/processor/readers/JupyterReader.js.map +1 -0
- package/dist/core/processor/readers/LatexReader.js +176 -0
- package/dist/core/processor/readers/LatexReader.js.map +1 -0
- package/dist/core/processor/readers/MarkdownReader.js +289 -0
- package/dist/core/processor/readers/MarkdownReader.js.map +1 -0
- package/dist/core/processor/readers/MarkerPdfReader.js +193 -0
- package/dist/core/processor/readers/MarkerPdfReader.js.map +1 -0
- package/dist/core/processor/readers/MistralOcrReader.js +198 -0
- package/dist/core/processor/readers/MistralOcrReader.js.map +1 -0
- package/dist/core/processor/readers/OfficeReader.js +174 -0
- package/dist/core/processor/readers/OfficeReader.js.map +1 -0
- package/dist/core/processor/readers/PdfReader.js +116 -0
- package/dist/core/processor/readers/PdfReader.js.map +1 -0
- package/dist/core/processor/readers/RtfReader.js +107 -0
- package/dist/core/processor/readers/RtfReader.js.map +1 -0
- package/dist/core/processor/readers/SubtitleReader.js +145 -0
- package/dist/core/processor/readers/SubtitleReader.js.map +1 -0
- package/dist/core/processor/readers/TesseractPdfReader.js +183 -0
- package/dist/core/processor/readers/TesseractPdfReader.js.map +1 -0
- package/dist/core/processor/readers/TextReader.js +129 -0
- package/dist/core/processor/readers/TextReader.js.map +1 -0
- package/dist/core/processor/readers/TranscriptReader.js +234 -0
- package/dist/core/processor/readers/TranscriptReader.js.map +1 -0
- package/dist/core/processor/readers/image/imageMetadata.js +155 -0
- package/dist/core/processor/readers/image/imageMetadata.js.map +1 -0
- package/dist/core/processor/readers/index.js +41 -0
- package/dist/core/processor/readers/index.js.map +1 -0
- package/dist/core/processor/readers/referenceExtraction.js +198 -0
- package/dist/core/processor/readers/referenceExtraction.js.map +1 -0
- package/dist/core/processor/readers/stripReferences.js +59 -0
- package/dist/core/processor/readers/stripReferences.js.map +1 -0
- package/dist/core/processor/readers/transcript/turnPacking.js +81 -0
- package/dist/core/processor/readers/transcript/turnPacking.js.map +1 -0
- package/dist/core/progress/NdjsonProgressEmitter.js +30 -0
- package/dist/core/progress/NdjsonProgressEmitter.js.map +1 -0
- package/dist/core/progress/NoopProgressEmitter.js +15 -0
- package/dist/core/progress/NoopProgressEmitter.js.map +1 -0
- package/dist/core/progress/index.js +19 -0
- package/dist/core/progress/index.js.map +1 -0
- package/dist/core/trace/TraceWriter.js +100 -0
- package/dist/core/trace/TraceWriter.js.map +1 -0
- package/dist/core/trace/events.js +13 -0
- package/dist/core/trace/events.js.map +1 -0
- package/dist/core/trace/index.js +20 -0
- package/dist/core/trace/index.js.map +1 -0
- package/dist/core/trace/lineage.js +97 -0
- package/dist/core/trace/lineage.js.map +1 -0
- package/dist/evaluation/BenchmarkRunner.js +171 -0
- package/dist/evaluation/BenchmarkRunner.js.map +1 -0
- package/dist/evaluation/classifier/ClassifierAccuracy.js +185 -0
- package/dist/evaluation/classifier/ClassifierAccuracy.js.map +1 -0
- package/dist/evaluation/classifier/labeledSamples.js +379 -0
- package/dist/evaluation/classifier/labeledSamples.js.map +1 -0
- package/dist/evaluation/compare/goldCompare.js +126 -0
- package/dist/evaluation/compare/goldCompare.js.map +1 -0
- package/dist/evaluation/crossre/compareScoring.js +30 -0
- package/dist/evaluation/crossre/compareScoring.js.map +1 -0
- package/dist/evaluation/datasets/CrossREDataset.js +170 -0
- package/dist/evaluation/datasets/CrossREDataset.js.map +1 -0
- package/dist/evaluation/datasets/IDataset.js +3 -0
- package/dist/evaluation/datasets/IDataset.js.map +1 -0
- package/dist/evaluation/datasets/RebelDataset.js +117 -0
- package/dist/evaluation/datasets/RebelDataset.js.map +1 -0
- package/dist/evaluation/datasets/RedocredDataset.js +218 -0
- package/dist/evaluation/datasets/RedocredDataset.js.map +1 -0
- package/dist/evaluation/datasets/SemEval2010Dataset.js +150 -0
- package/dist/evaluation/datasets/SemEval2010Dataset.js.map +1 -0
- package/dist/evaluation/index.js +33 -0
- package/dist/evaluation/index.js.map +1 -0
- package/dist/evaluation/matching/ExactMatcher.js +75 -0
- package/dist/evaluation/matching/ExactMatcher.js.map +1 -0
- package/dist/evaluation/matching/SemanticMatcher.js +143 -0
- package/dist/evaluation/matching/SemanticMatcher.js.map +1 -0
- package/dist/evaluation/metrics/TripleMetrics.js +64 -0
- package/dist/evaluation/metrics/TripleMetrics.js.map +1 -0
- package/dist/evaluation/mine/MineCheckpoint.js +114 -0
- package/dist/evaluation/mine/MineCheckpoint.js.map +1 -0
- package/dist/evaluation/mine/MineDataset.js +208 -0
- package/dist/evaluation/mine/MineDataset.js.map +1 -0
- package/dist/evaluation/mine/MineReporter.js +98 -0
- package/dist/evaluation/mine/MineReporter.js.map +1 -0
- package/dist/evaluation/mine/MineRunner.js +148 -0
- package/dist/evaluation/mine/MineRunner.js.map +1 -0
- package/dist/evaluation/mine/MineScorer.js +127 -0
- package/dist/evaluation/mine/MineScorer.js.map +1 -0
- package/dist/evaluation/mine/types.js +12 -0
- package/dist/evaluation/mine/types.js.map +1 -0
- package/dist/evaluation/reporters/ConsoleReporter.js +55 -0
- package/dist/evaluation/reporters/ConsoleReporter.js.map +1 -0
- package/dist/evaluation/reporters/JsonReporter.js +50 -0
- package/dist/evaluation/reporters/JsonReporter.js.map +1 -0
- package/dist/index.js +28 -0
- package/dist/index.js.map +1 -0
- package/dist/quality/CompositeScore.js +61 -0
- package/dist/quality/CompositeScore.js.map +1 -0
- package/dist/quality/ConsistencyMetrics.js +70 -0
- package/dist/quality/ConsistencyMetrics.js.map +1 -0
- package/dist/quality/FactualMetrics.js +76 -0
- package/dist/quality/FactualMetrics.js.map +1 -0
- package/dist/quality/GraphHealthMetrics.js +68 -0
- package/dist/quality/GraphHealthMetrics.js.map +1 -0
- package/dist/quality/SemanticMetrics.js +102 -0
- package/dist/quality/SemanticMetrics.js.map +1 -0
- package/dist/quality/StructuralMetrics.js +60 -0
- package/dist/quality/StructuralMetrics.js.map +1 -0
- package/dist/quality/index.js +23 -0
- package/dist/quality/index.js.map +1 -0
- package/dist/shared/index.js +20 -0
- package/dist/shared/index.js.map +1 -0
- package/dist/shared/logger/Logger.js +3 -0
- package/dist/shared/logger/Logger.js.map +1 -0
- package/dist/shared/logger/LoggerFactory.js +75 -0
- package/dist/shared/logger/LoggerFactory.js.map +1 -0
- package/dist/shared/logger/index.js +19 -0
- package/dist/shared/logger/index.js.map +1 -0
- package/dist/shared/shutdown.js +30 -0
- package/dist/shared/shutdown.js.map +1 -0
- package/dist/shared/utils/agglomerativeCluster.js +269 -0
- package/dist/shared/utils/agglomerativeCluster.js.map +1 -0
- package/dist/shared/utils/astSymbols.js +69 -0
- package/dist/shared/utils/astSymbols.js.map +1 -0
- package/dist/shared/utils/cosineSimilarity.js +18 -0
- package/dist/shared/utils/cosineSimilarity.js.map +1 -0
- package/dist/shared/utils/directoryTree.js +184 -0
- package/dist/shared/utils/directoryTree.js.map +1 -0
- package/dist/shared/utils/documentOutline.js +74 -0
- package/dist/shared/utils/documentOutline.js.map +1 -0
- package/dist/shared/utils/index.js +24 -0
- package/dist/shared/utils/index.js.map +1 -0
- package/dist/shared/utils/jaroWinklerSimilarity.js +60 -0
- package/dist/shared/utils/jaroWinklerSimilarity.js.map +1 -0
- package/dist/shared/utils/parseJsonLenient.js +27 -0
- package/dist/shared/utils/parseJsonLenient.js.map +1 -0
- package/dist/shared/utils/readConfig.js +42 -0
- package/dist/shared/utils/readConfig.js.map +1 -0
- package/dist/shared/utils/readRtf.js +216 -0
- package/dist/shared/utils/readRtf.js.map +1 -0
- package/dist/shared/utils/softmax.js +26 -0
- package/dist/shared/utils/softmax.js.map +1 -0
- package/dist/types/ContentClass.js +3 -0
- package/dist/types/ContentClass.js.map +1 -0
- package/dist/types/CorpusProfile.js +3 -0
- package/dist/types/CorpusProfile.js.map +1 -0
- package/dist/types/IContradictionChecker.js +3 -0
- package/dist/types/IContradictionChecker.js.map +1 -0
- package/dist/types/ICorpusAnalyzer.js +3 -0
- package/dist/types/ICorpusAnalyzer.js.map +1 -0
- package/dist/types/IDirectoryProcessor.js +3 -0
- package/dist/types/IDirectoryProcessor.js.map +1 -0
- package/dist/types/IEmbeddingProvider.js +3 -0
- package/dist/types/IEmbeddingProvider.js.map +1 -0
- package/dist/types/IEmbeddingService.js +6 -0
- package/dist/types/IEmbeddingService.js.map +1 -0
- package/dist/types/IFileProcessor.js +3 -0
- package/dist/types/IFileProcessor.js.map +1 -0
- package/dist/types/IGroundingChecker.js +3 -0
- package/dist/types/IGroundingChecker.js.map +1 -0
- package/dist/types/IKnowledgeGraphBuilder.js +3 -0
- package/dist/types/IKnowledgeGraphBuilder.js.map +1 -0
- package/dist/types/IKnowledgeGraphExporter.js +3 -0
- package/dist/types/IKnowledgeGraphExporter.js.map +1 -0
- package/dist/types/IKnowledgeGraphMerger.js +3 -0
- package/dist/types/IKnowledgeGraphMerger.js.map +1 -0
- package/dist/types/IKnowledgeGraphSearch.js +3 -0
- package/dist/types/IKnowledgeGraphSearch.js.map +1 -0
- package/dist/types/ILLMProvider.js +3 -0
- package/dist/types/ILLMProvider.js.map +1 -0
- package/dist/types/ILLMService.js +3 -0
- package/dist/types/ILLMService.js.map +1 -0
- package/dist/types/IObjectDetector.js +3 -0
- package/dist/types/IObjectDetector.js.map +1 -0
- package/dist/types/IProcessingService.js +3 -0
- package/dist/types/IProcessingService.js.map +1 -0
- package/dist/types/IProgressEmitter.js +3 -0
- package/dist/types/IProgressEmitter.js.map +1 -0
- package/dist/types/IPromptManager.js +3 -0
- package/dist/types/IPromptManager.js.map +1 -0
- package/dist/types/KnowledgeGraph.js +3 -0
- package/dist/types/KnowledgeGraph.js.map +1 -0
- package/dist/types/MCPKnowledgeGraph.js +3 -0
- package/dist/types/MCPKnowledgeGraph.js.map +1 -0
- package/dist/types/Observation.js +21 -0
- package/dist/types/Observation.js.map +1 -0
- package/dist/types/ProcessingOptions.js +3 -0
- package/dist/types/ProcessingOptions.js.map +1 -0
- package/dist/types/index.js +40 -0
- package/dist/types/index.js.map +1 -0
- package/package.json +122 -0
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"CitationEvidenceProcessor.js","sourceRoot":"","sources":["../../../../../src/core/knowledge/references/citations/CitationEvidenceProcessor.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAiEA,kDA4BC;AASD,sCAmBC;AAzHD,uCAAyB;AAGzB,oDAAmF;AAEnF,4DAA8E;AAyC9E,MAAM,KAAK,GAAG,uBAAuB,CAAC;AAEtC;qEACqE;AACrE,MAAM,gBAAgB,GAAG,wFAAwF,CAAC;AAElH;;;;;;;;;;;;GAYG;AACH,SAAgB,mBAAmB,CAAC,MAAgB;IAClD,6EAA6E;IAC7E,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IACvD,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,KAAK,GAAG,GAAG,CAAC;QAClB,GAAG,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;QACxB,IAAI,KAAK,GAAG,KAAK,GAAG,GAAG;YAAE,SAAS,CAAC,mDAAmD;QACtF,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACpC,MAAM,CAAC,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAClE,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACX,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YACjD,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YAChC,IAAI,IAAI;gBAAE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC1B,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC;QACrC,CAAC;IACH,CAAC;IACD,2EAA2E;IAC3E,MAAM,UAAU,GAAG,CAAC,CAAS,EAAW,EAAE;QACxC,MAAM,KAAK,GAAG,CAAC,EAAU,EAAE,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,GAAG,IAAI,EAAE,CAAC,CAAC,CAAC;QACrG,MAAM,OAAO,GACX,KAAK,CAAC,qBAAqB,CAAC;YAC5B,KAAK,CAAC,cAAc,EAAE,CAAC,CAAC;YACxB,KAAK,CAAC,wIAAwI,EAAE,CAAC,CAAC,CAAC;QACrJ,OAAO,OAAO,IAAI,CAAC,CAAC;IACtB,CAAC,CAAC;IACF,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;IAClD,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC;AACrC,CAAC;AAED;;;;;;GAMG;AACH,SAAgB,aAAa,CAAC,IAAY,EAAE,MAAM,GAAG,GAAG;IACtD,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/C,IAAI,KAAK,CAAC,MAAM,IAAI,MAAM;QAAE,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACxD,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,MAAM,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,eAAe,CAAC,EAAE,CAAC;QAC7C,IAAI,GAAG,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC,GAAG,MAAM,EAAE,CAAC;YAC9C,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACd,GAAG,GAAG,CAAC,CAAC;QACV,CAAC;aAAM,CAAC;YACN,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QAChC,CAAC;QACD,OAAO,GAAG,CAAC,MAAM,GAAG,MAAM,GAAG,GAAG,EAAE,CAAC;YACjC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,6CAA6C;YAC7E,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QAC1B,CAAC;IACH,CAAC;IACD,IAAI,GAAG,CAAC,IAAI,EAAE;QAAE,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC9B,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAa,yBAAyB;IAKpC,YACmB,OAAqB,EACrB,KAAwB,EACxB,QAA0B,EAC1B,OAA0B,EAC1B,UAA8B,EAC9B,MAAc,EAC/B,OAA6B,EAAE;;QANd,YAAO,GAAP,OAAO,CAAc;QACrB,UAAK,GAAL,KAAK,CAAmB;QACxB,aAAQ,GAAR,QAAQ,CAAkB;QAC1B,YAAO,GAAP,OAAO,CAAmB;QAC1B,eAAU,GAAV,UAAU,CAAoB;QAC9B,WAAM,GAAN,MAAM,CAAQ;QAG/B,IAAI,CAAC,MAAM,GAAG,MAAA,IAAI,CAAC,MAAM,mCAAI,IAAI,CAAC;QAClC,IAAI,CAAC,YAAY,GAAG,MAAA,IAAI,CAAC,YAAY,mCAAI,IAAI,CAAC;QAC9C,IAAI,CAAC,IAAI,GAAG,MAAA,IAAI,CAAC,aAAa,mCAAI,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IACjD,CAAC;IAEK,OAAO,CACX,SAAiB,EACjB,cAAsB,EACtB,cAA6B;;YAE7B,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,cAAc,EAAE,cAAc,CAAC,CAAC;YAC1E,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;gBAAE,OAAO,IAAI,CAAC;YAEvC,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;YAC3C,MAAM,SAAS,GAAe,EAAE,CAAC;YACjC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;YAC/B,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,eAA8B,EAAE,EAAE,EAAE;gBACnE,MAAM,CAAC,GAAG,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;gBAC7B,IAAI,CAAC,EAAE,CAAC;oBACN,IAAI,YAAY,CAAC,MAAM;wBAAE,CAAC,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC;oBAC9D,OAAO;gBACT,CAAC;gBACD,QAAQ,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,EAAE,KAAK,EAAE,EAAE,EAAE,YAAY,EAAE,CAAC,CAAC;YAChF,CAAC,CAAC;YACF,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,oCAAoC;YAE1D,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;gBAC3B,MAAM,IAAI,GAAG,IAAA,oCAAgB,kBAAG,GAAG,EAAE,GAAG,CAAC,GAAG,IAAK,GAAG,CAAC,GAAG,EAAG,CAAC;gBAC5D,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC;oBAAE,SAAS;gBACnD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;gBAEf,MAAM,YAAY,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC;gBACjE,KAAK,MAAM,CAAC,IAAI,YAAY,CAAC,QAAQ;oBAAE,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,CAAC;gBACzE,SAAS,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,SAAS,CAAC,CAAC;YAC5C,CAAC;YAED,OAAO,SAAS,CAAC,MAAM,IAAI,QAAQ,CAAC,IAAI,GAAG,CAAC;gBAC1C,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,SAAS,EAAE;gBACxD,CAAC,CAAC,IAAI,CAAC;QACX,CAAC;KAAA;IAED,gFAAgF;IAClE,UAAU,CACtB,SAAiB,EACjB,QAAgB,EAChB,GAAoB;;YAEpB,MAAM,SAAS,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YAC3C,MAAM,SAAS,GAAG,IAAA,wCAAoB,kBAAG,GAAG,EAAE,GAAG,CAAC,GAAG,IAAK,GAAG,CAAC,GAAG,GAAI,SAAS,EAAE,SAAS,CAAC,CAAC;YAE3F,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACpD,IAAI,CAAC,MAAM,EAAE,CAAC;gBACZ,sEAAsE;gBACtE,OAAO,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC;YACzD,CAAC;YAED,8EAA8E;YAC9E,6EAA6E;YAC7E,6EAA6E;YAC7E,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAC1C,IAAI,MAAM,aAAN,MAAM,uBAAN,MAAM,CAAE,KAAK,EAAE,CAAC;gBAClB,OAAO,IAAI,CAAC,SAAS,CAAC,SAAS,EAAE,QAAQ,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,EAAE,SAAS,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC;YACvF,CAAC;YAED,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;YACnD,IAAI,KAAqB,CAAC;YAC1B,IAAI,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC;gBAC7B,IAAI,MAAM,GAAa,EAAE,CAAC;gBAC1B,IAAI,MAAM,GAAqB,EAAE,CAAC;gBAClC,IAAI,CAAC;oBACH,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;gBACpE,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,oCAAoC,MAAM,CAAC,GAAG,KAAK,GAAG,EAAE,CAAC,CAAC;gBAC7E,CAAC;wBAAS,CAAC;oBACT,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;gBACxD,CAAC;gBACD,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;gBAC5D,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;gBAC5C,MAAM,SAAS,GACb,CAAC,OAAO,IAAI,GAAG,CAAC,WAAW,IAAI,GAAG,CAAC,YAAY,KAAK,KAAK;oBACvD,CAAC,CAAC,uDAAuD;oBACzD,CAAC,CAAC,SAAS,CAAC;gBAChB,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,SAAS,EAAE,QAAQ,EAAE,MAAM,CAAC,GAAG,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,SAAS,EAAE,SAAS,CAAC,CAAC;YACrH,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAC,MAAM,MAAM,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;gBACvE,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,SAAS,EAAE,QAAQ,EAAE,MAAM,CAAC,GAAG,EAAE,SAAS,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,SAAS,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC;YACrH,CAAC;YAED,MAAM,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;gBACtB,GAAG,EAAE,MAAM,CAAC,GAAG;gBACf,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC,QAAQ;gBACtB,MAAM,EAAE,CAAC,CAAC,MAAM;gBAChB,MAAM,EAAE,CAAC,CAAC,MAAM;gBAChB,WAAW,EAAE,CAAC,CAAC,WAAW;gBAC1B,SAAS,EAAE,SAAS;gBACpB,KAAK;aACN,CAAC,CAAC;YACH,OAAO,KAAK,CAAC;QACf,CAAC;KAAA;IAED;2EACuE;IAC/D,YAAY,CAClB,SAAiB,EACjB,QAAgB,EAChB,GAAW,EACX,SAAwB,EACxB,aAA+B,EAC/B,IAA4C,EAC5C,OAAmF,EACnF,SAA6B,EAC7B,SAAiB,EACjB,QAAQ,GAAG,IAAI;QAEf,MAAM,QAAQ,GAAa,EAAE,CAAC;QAC9B,MAAM,GAAG,GAAkB,CAAC,GAAG,SAAS,CAAC,CAAC;QAC1C,IAAI,QAAQ;YAAE,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,6BAA6B,GAAG,EAAE,EAAE,MAAM,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC,CAAC;QAC7F,IAAI,OAAO,EAAE,CAAC;YACZ,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,GAAG,KAAK,KAAK,OAAO,CAAC,KAAK,KAAK,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,EAAE,MAAM,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC,CAAC;QACzG,CAAC;aAAM,IAAI,SAAS,EAAE,CAAC;YACrB,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,GAAG,KAAK,KAAK,SAAS,EAAE,EAAE,MAAM,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC,CAAC;QACvE,CAAC;QACD,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,UAAU,EAAE,UAAU,EAAE,KAAK,EAAE,EAAE,EAAE,YAAY,EAAE,GAAG,EAAE,CAAC,CAAC;QAExF,6EAA6E;QAC7E,KAAK,MAAM,CAAC,IAAI,aAAa,EAAE,CAAC;YAC9B,KAAK,MAAM,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC;gBAC3B,QAAQ,CAAC,IAAI,iCAAM,CAAC,KAAE,YAAY,EAAE,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,iCAAM,CAAC,KAAE,MAAM,EAAE,GAAG,IAAG,CAAC,IAAG,CAAC;YAC5F,CAAC;QACH,CAAC;QACD,MAAM,SAAS,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;QAE5D,MAAM,IAAI,GAAa,EAAE,IAAI,EAAE,SAAS,EAAE,EAAE,EAAE,QAAQ,EAAE,YAAY,EAAE,CAAC,OAAO,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC;QAC/G,IAAI,OAAO,IAAI,IAAI,EAAE,CAAC;YACpB,IAAI,CAAC,YAAY,GAAG,OAAO,CAAC,KAAK,CAAC;YAClC,IAAI,CAAC,iBAAiB,GAAG,OAAO,CAAC,KAAK,CAAC;YACvC,IAAI,CAAC,cAAc,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QACjD,CAAC;QACD,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACrB,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;IACjC,CAAC;IAEO,UAAU,CAAC,SAAiB,EAAE,QAAgB,EAAE,SAAwB;QAC9E,OAAO;YACL,QAAQ,EAAE,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,UAAU,EAAE,UAAU,EAAE,KAAK,EAAE,EAAE,EAAE,YAAY,EAAE,SAAS,EAAE,CAAC;YAC1F,SAAS,EAAE,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,EAAE,EAAE,QAAQ,EAAE,YAAY,EAAE,CAAC,OAAO,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC;SAC5G,CAAC;IACJ,CAAC;IAED,gFAAgF;IACxE,SAAS,CACf,SAAiB,EACjB,QAAgB,EAChB,GAAoB,EACpB,GAAW,EACX,SAAwB,EACxB,WAA2B;;QAE3B,MAAM,QAAQ,GAAa;YACzB,EAAE,IAAI,EAAE,QAAQ,EAAE,UAAU,EAAE,UAAU,EAAE,KAAK,EAAE,EAAE,EAAE,YAAY,EAAE,SAAS,EAAE;SAC/E,CAAC;QACF,KAAK,MAAM,CAAC,IAAI,WAAW,CAAC,QAAQ,EAAE,CAAC;YACrC,IAAI,CAAC,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBACxB,uEAAuE;gBACvE,QAAQ,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC;YACnF,CAAC;iBAAM,CAAC;gBACN,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;QACD,MAAM,UAAU,GAAG,WAAW,CAAC,SAAS,CAAC,IAAI,CAC3C,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,QAAQ,IAAI,CAAC,CAAC,YAAY,CAAC,QAAQ,CAAC,OAAO,CAAC,CAC7D,CAAC;QACF,MAAM,QAAQ,GAAG,MAAA,UAAU,aAAV,UAAU,uBAAV,UAAU,CAAE,QAAQ,mCAAI,IAAI,CAAC;QAC9C,MAAM,IAAI,GAAa,EAAE,IAAI,EAAE,SAAS,EAAE,EAAE,EAAE,QAAQ,EAAE,YAAY,EAAE,CAAC,OAAO,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC;QAC/G,mFAAmF;QACnF,IAAI,CAAA,UAAU,aAAV,UAAU,uBAAV,UAAU,CAAE,YAAY,KAAI,UAAU,CAAC,MAAM,KAAK,SAAS,EAAE,CAAC;YAChE,IAAI,CAAC,YAAY,GAAG,UAAU,CAAC,YAAY,CAAC;YAC5C,IAAI,CAAC,iBAAiB,GAAG,UAAU,CAAC,iBAAiB,CAAC;YACtD,IAAI,CAAC,cAAc,GAAG,UAAU,CAAC,cAAc,CAAC;QAClD,CAAC;QACD,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;IACzC,CAAC;IAED,oFAAoF;IACtE,aAAa,CAAC,QAAgB,EAAE,KAAoB;;YAChE,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;gBAChB,IAAI,CAAC;oBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;oBAC9C,IAAI,CAAC,CAAC,MAAM;wBAAE,OAAO,CAAC,CAAC;oBACvB,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,uEAAuE,CAAC,CAAC;gBAC5F,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,uBAAuB,GAAG,oCAAoC,CAAC,CAAC;gBACnF,CAAC;YACH,CAAC;YACD,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;gBACvB,GAAG,EAAE,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE;gBACrE,GAAG,EAAE,CAAC,CAAC,GAAG;aACX,CAAC,CAAC,CAAC;QACN,CAAC;KAAA;IAED;uFACmF;IACrE,UAAU,CACtB,KAAyB,EACzB,SAAmB;;YAEnB,IAAI,CAAC,KAAK,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;gBAAE,OAAO,IAAI,CAAC;YAClD,4EAA4E;YAC5E,4EAA4E;YAC5E,MAAM,MAAM,GAAG,mBAAmB,CAAC,SAAS,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC;YAC/E,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;gBAAE,OAAO,IAAI,CAAC;YACrC,MAAM,IAAI,GAAG,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;YACxE,MAAM,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;YACvB,IAAI,EAAE,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;gBACpB,MAAM,MAAM,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;gBAC/B,MAAM,GAAG,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC;gBAC3D,IAAI,GAAG;oBAAE,OAAO,EAAE,IAAI,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;YAC1C,CAAC;YACD,IAAI,CAAC;gBACH,MAAM,CAAC,EAAE,EAAE,GAAG,GAAG,CAAC,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,KAAK,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC;gBAC1E,IAAI,IAAI,GAAG,CAAC,CAAC,CAAC;gBACd,IAAI,EAAE,GAAG,CAAC,CAAC;gBACX,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;oBACnB,MAAM,CAAC,GAAG,IAAA,wBAAgB,EAAC,EAAE,EAAE,CAAC,CAAC,CAAC;oBAClC,IAAI,CAAC,GAAG,IAAI,EAAE,CAAC;wBACb,IAAI,GAAG,CAAC,CAAC;wBACT,EAAE,GAAG,CAAC,CAAC;oBACT,CAAC;gBACH,CAAC,CAAC,CAAC;gBACH,OAAO,EAAE,IAAI,EAAE,MAAM,CAAC,EAAE,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;YAC3C,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,0BAA0B,GAAG,wBAAwB,CAAC,CAAC;gBACxE,IAAI,IAAI,GAAG,CAAC,CAAC,CAAC;gBACd,IAAI,EAAE,GAAG,CAAC,CAAC;gBACX,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE;oBACvB,MAAM,CAAC,GAAG,IAAA,6BAAqB,EAAC,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;oBAC1E,IAAI,CAAC,GAAG,IAAI,EAAE,CAAC;wBACb,IAAI,GAAG,CAAC,CAAC;wBACT,EAAE,GAAG,CAAC,CAAC;oBACT,CAAC;gBACH,CAAC,CAAC,CAAC;gBACH,OAAO,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,EAAE,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;YAC7D,CAAC;QACH,CAAC;KAAA;IAED;;;;oCAIgC;IAClB,KAAK,CACjB,GAAoB,EACpB,IAA4C;;YAE5C,IAAI,CAAC,IAAI,CAAC,YAAY,IAAI,CAAC,GAAG,CAAC,WAAW,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACjE,IAAI,GAAG,CAAC,YAAY,KAAK,KAAK;gBAAE,OAAO,IAAI,CAAC,CAAC,8BAA8B;YAC3E,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,KAAK,CAAC,GAAG,CAAC,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;gBACpE,OAAO,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC;YACxD,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,8BAA8B,GAAG,EAAE,CAAC,CAAC;gBACtD,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;KAAA;IAEO,KAAK,CAAC,KAAa;QACzB,MAAM,CAAC,EAAE,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC;QAC3B,IAAI,KAAK,IAAI,EAAE;YAAE,OAAO,WAAW,CAAC;QACpC,IAAI,KAAK,IAAI,EAAE;YAAE,OAAO,aAAa,CAAC;QACtC,OAAO,WAAW,CAAC;IACrB,CAAC;CACF;AA7RD,8DA6RC"}
|
|
@@ -0,0 +1,95 @@
|
|
|
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.CitationResolver = void 0;
|
|
13
|
+
class CitationResolver {
|
|
14
|
+
constructor(opts, logger, titleResolver = null, fetchFn = globalThis.fetch) {
|
|
15
|
+
this.opts = opts;
|
|
16
|
+
this.logger = logger;
|
|
17
|
+
this.titleResolver = titleResolver;
|
|
18
|
+
this.fetchFn = fetchFn;
|
|
19
|
+
}
|
|
20
|
+
resolve(c_1) {
|
|
21
|
+
return __awaiter(this, arguments, void 0, function* (c, _depth = 0) {
|
|
22
|
+
if (c.arxivId)
|
|
23
|
+
return this.target(`https://arxiv.org/pdf/${c.arxivId}`);
|
|
24
|
+
if (c.doi) {
|
|
25
|
+
const url = yield this.unpaywall(c.doi);
|
|
26
|
+
return url ? this.target(url) : null;
|
|
27
|
+
}
|
|
28
|
+
if (c.pmid) {
|
|
29
|
+
const url = yield this.pmc(c.pmid);
|
|
30
|
+
return url ? this.target(url) : null;
|
|
31
|
+
}
|
|
32
|
+
// id-less: try the title resolver once (Phase 2d), then resolve the hard id.
|
|
33
|
+
if (c.title && this.titleResolver && _depth === 0) {
|
|
34
|
+
try {
|
|
35
|
+
const ids = yield this.titleResolver.resolve(c.title);
|
|
36
|
+
if (ids && (ids.doi || ids.arxivId))
|
|
37
|
+
return this.resolve(Object.assign({}, ids), _depth + 1);
|
|
38
|
+
}
|
|
39
|
+
catch (err) {
|
|
40
|
+
this.logger.warn(`Title resolution failed for "${c.title.slice(0, 60)}": ${err}`);
|
|
41
|
+
}
|
|
42
|
+
}
|
|
43
|
+
return null;
|
|
44
|
+
});
|
|
45
|
+
}
|
|
46
|
+
target(url) {
|
|
47
|
+
try {
|
|
48
|
+
return { url, host: new URL(url).hostname };
|
|
49
|
+
}
|
|
50
|
+
catch (_a) {
|
|
51
|
+
return null;
|
|
52
|
+
}
|
|
53
|
+
}
|
|
54
|
+
/** DOI → Unpaywall best OA PDF (or landing) URL; null when closed or no email. */
|
|
55
|
+
unpaywall(doi) {
|
|
56
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
57
|
+
if (!this.opts.unpaywallEmail) {
|
|
58
|
+
this.logger.warn("Unpaywall email unset (references.citations.fetch.unpaywallEmail) — skipping DOI resolution");
|
|
59
|
+
return null;
|
|
60
|
+
}
|
|
61
|
+
try {
|
|
62
|
+
const r = yield this.fetchFn(`https://api.unpaywall.org/v2/${encodeURIComponent(doi)}?email=${encodeURIComponent(this.opts.unpaywallEmail)}`);
|
|
63
|
+
if (!r.ok)
|
|
64
|
+
return null;
|
|
65
|
+
const j = (yield r.json());
|
|
66
|
+
const loc = j.best_oa_location;
|
|
67
|
+
return (loc === null || loc === void 0 ? void 0 : loc.url_for_pdf) || (loc === null || loc === void 0 ? void 0 : loc.url) || null;
|
|
68
|
+
}
|
|
69
|
+
catch (err) {
|
|
70
|
+
this.logger.warn(`Unpaywall lookup failed for ${doi}: ${err}`);
|
|
71
|
+
return null;
|
|
72
|
+
}
|
|
73
|
+
});
|
|
74
|
+
}
|
|
75
|
+
/** PMID → PMCID (id-converter) → PMC article PDF URL; null when not in PMC. */
|
|
76
|
+
pmc(pmid) {
|
|
77
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
78
|
+
var _a, _b;
|
|
79
|
+
try {
|
|
80
|
+
const r = yield this.fetchFn(`https://www.ncbi.nlm.nih.gov/pmc/utils/idconv/v1.0/?ids=${encodeURIComponent(pmid)}&format=json`);
|
|
81
|
+
if (!r.ok)
|
|
82
|
+
return null;
|
|
83
|
+
const j = (yield r.json());
|
|
84
|
+
const pmcid = (_b = (_a = j.records) === null || _a === void 0 ? void 0 : _a[0]) === null || _b === void 0 ? void 0 : _b.pmcid;
|
|
85
|
+
return pmcid ? `https://www.ncbi.nlm.nih.gov/pmc/articles/${pmcid}/pdf/` : null;
|
|
86
|
+
}
|
|
87
|
+
catch (err) {
|
|
88
|
+
this.logger.warn(`PMC id-convert failed for PMID ${pmid}: ${err}`);
|
|
89
|
+
return null;
|
|
90
|
+
}
|
|
91
|
+
});
|
|
92
|
+
}
|
|
93
|
+
}
|
|
94
|
+
exports.CitationResolver = CitationResolver;
|
|
95
|
+
//# sourceMappingURL=CitationResolver.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"CitationResolver.js","sourceRoot":"","sources":["../../../../../src/core/knowledge/references/citations/CitationResolver.ts"],"names":[],"mappings":";;;;;;;;;;;;AAsCA,MAAa,gBAAgB;IAC3B,YACmB,IAA6B,EAC7B,MAAc,EACd,gBAAyC,IAAI,EAC7C,UAAoB,UAAkB,CAAC,KAAK;QAH5C,SAAI,GAAJ,IAAI,CAAyB;QAC7B,WAAM,GAAN,MAAM,CAAQ;QACd,kBAAa,GAAb,aAAa,CAAgC;QAC7C,YAAO,GAAP,OAAO,CAAqC;IAC5D,CAAC;IAEE,OAAO;6DAAC,CAAqB,EAAE,MAAM,GAAG,CAAC;YAC7C,IAAI,CAAC,CAAC,OAAO;gBAAE,OAAO,IAAI,CAAC,MAAM,CAAC,yBAAyB,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;YACxE,IAAI,CAAC,CAAC,GAAG,EAAE,CAAC;gBACV,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;gBACxC,OAAO,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YACvC,CAAC;YACD,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;gBACX,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;gBACnC,OAAO,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YACvC,CAAC;YACD,6EAA6E;YAC7E,IAAI,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,aAAa,IAAI,MAAM,KAAK,CAAC,EAAE,CAAC;gBAClD,IAAI,CAAC;oBACH,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;oBACtD,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC;wBAAE,OAAO,IAAI,CAAC,OAAO,mBAAM,GAAG,GAAI,MAAM,GAAG,CAAC,CAAC,CAAC;gBACnF,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,gCAAgC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;gBACpF,CAAC;YACH,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;KAAA;IAEO,MAAM,CAAC,GAAW;QACxB,IAAI,CAAC;YACH,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,CAAC;QAC9C,CAAC;QAAC,WAAM,CAAC;YACP,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IAED,kFAAkF;IACpE,SAAS,CAAC,GAAW;;YACjC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,CAAC;gBAC9B,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,6FAA6F,CAAC,CAAC;gBAChH,OAAO,IAAI,CAAC;YACd,CAAC;YACD,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAC1B,gCAAgC,kBAAkB,CAAC,GAAG,CAAC,UAAU,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,EAAE,CAChH,CAAC;gBACF,IAAI,CAAC,CAAC,CAAC,EAAE;oBAAE,OAAO,IAAI,CAAC;gBACvB,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAExB,CAAC;gBACF,MAAM,GAAG,GAAG,CAAC,CAAC,gBAAgB,CAAC;gBAC/B,OAAO,CAAA,GAAG,aAAH,GAAG,uBAAH,GAAG,CAAE,WAAW,MAAI,GAAG,aAAH,GAAG,uBAAH,GAAG,CAAE,GAAG,CAAA,IAAI,IAAI,CAAC;YAC9C,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,+BAA+B,GAAG,KAAK,GAAG,EAAE,CAAC,CAAC;gBAC/D,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;KAAA;IAED,+EAA+E;IACjE,GAAG,CAAC,IAAY;;;YAC5B,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAC1B,2DAA2D,kBAAkB,CAAC,IAAI,CAAC,cAAc,CAClG,CAAC;gBACF,IAAI,CAAC,CAAC,CAAC,EAAE;oBAAE,OAAO,IAAI,CAAC;gBACvB,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAA4C,CAAC;gBACtE,MAAM,KAAK,GAAG,MAAA,MAAA,CAAC,CAAC,OAAO,0CAAG,CAAC,CAAC,0CAAE,KAAK,CAAC;gBACpC,OAAO,KAAK,CAAC,CAAC,CAAC,6CAA6C,KAAK,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC;YAClF,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,kCAAkC,IAAI,KAAK,GAAG,EAAE,CAAC,CAAC;gBACnE,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;KAAA;CACF;AA3ED,4CA2EC"}
|
|
@@ -0,0 +1,143 @@
|
|
|
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.GrobidClient = void 0;
|
|
46
|
+
const fs = __importStar(require("fs"));
|
|
47
|
+
const path = __importStar(require("path"));
|
|
48
|
+
const cheerio = __importStar(require("cheerio"));
|
|
49
|
+
const ARXIV_RE = /(\d{4}\.\d{4,5}(?:v\d+)?)/;
|
|
50
|
+
const DOI_RE = /10\.\d{4,9}\/[-._;()/:a-z0-9]+/i;
|
|
51
|
+
class GrobidClient {
|
|
52
|
+
constructor(url, logger, fetchFn = globalThis.fetch) {
|
|
53
|
+
this.url = url;
|
|
54
|
+
this.logger = logger;
|
|
55
|
+
this.fetchFn = fetchFn;
|
|
56
|
+
}
|
|
57
|
+
/** Liveness probe so wiring can warn early; the processor degrades regardless. */
|
|
58
|
+
isAlive() {
|
|
59
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
60
|
+
try {
|
|
61
|
+
const r = yield this.fetchFn(`${this.url}/api/isalive`);
|
|
62
|
+
return r.ok;
|
|
63
|
+
}
|
|
64
|
+
catch (_a) {
|
|
65
|
+
return false;
|
|
66
|
+
}
|
|
67
|
+
});
|
|
68
|
+
}
|
|
69
|
+
process(filePath) {
|
|
70
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
71
|
+
const buf = yield fs.promises.readFile(filePath);
|
|
72
|
+
const form = new FormData();
|
|
73
|
+
form.append("input", new Blob([buf]), path.basename(filePath) || "doc.pdf");
|
|
74
|
+
form.append("consolidateCitations", "1");
|
|
75
|
+
form.append("includeRawCitations", "1");
|
|
76
|
+
form.append("segmentSentences", "1"); // emit <s> so the citing CLAIM is a sentence, not the whole paragraph
|
|
77
|
+
const res = yield this.fetchFn(`${this.url}/api/processFulltextDocument`, {
|
|
78
|
+
method: "POST",
|
|
79
|
+
body: form,
|
|
80
|
+
});
|
|
81
|
+
if (!res.ok)
|
|
82
|
+
throw new Error(`GROBID ${res.status}`);
|
|
83
|
+
return this.parseTei(yield res.text());
|
|
84
|
+
});
|
|
85
|
+
}
|
|
86
|
+
/** TEI → contexts: join each reference's ids/title with the sentence that cites it. */
|
|
87
|
+
parseTei(xml) {
|
|
88
|
+
const $ = cheerio.load(xml, { xmlMode: true });
|
|
89
|
+
// 1) bibliography: xml:id → {ids, title, raw}
|
|
90
|
+
const refs = new Map();
|
|
91
|
+
$("listBibl biblStruct").each((_, el) => {
|
|
92
|
+
const $el = $(el);
|
|
93
|
+
const id = $el.attr("xml:id");
|
|
94
|
+
if (!id)
|
|
95
|
+
return;
|
|
96
|
+
let doi;
|
|
97
|
+
let arxivId;
|
|
98
|
+
$el.find("idno").each((__, idno) => {
|
|
99
|
+
var _a, _b, _c, _d;
|
|
100
|
+
const type = ($(idno).attr("type") || "").toLowerCase();
|
|
101
|
+
const txt = $(idno).text().trim();
|
|
102
|
+
if (type === "doi" || DOI_RE.test(txt))
|
|
103
|
+
doi = doi || ((_b = (_a = DOI_RE.exec(txt)) === null || _a === void 0 ? void 0 : _a[0]) !== null && _b !== void 0 ? _b : txt);
|
|
104
|
+
if (type === "arxiv")
|
|
105
|
+
arxivId = arxivId || ((_d = (_c = ARXIV_RE.exec(txt)) === null || _c === void 0 ? void 0 : _c[1]) !== null && _d !== void 0 ? _d : txt);
|
|
106
|
+
});
|
|
107
|
+
const title = $el.find('title[level="a"]').first().text().trim() ||
|
|
108
|
+
$el.find('title[level="m"]').first().text().trim() ||
|
|
109
|
+
$el.find("title").first().text().trim() ||
|
|
110
|
+
undefined;
|
|
111
|
+
const raw = $el.find("note[type='raw_reference']").first().text().trim() || title || id;
|
|
112
|
+
refs.set(id, { ids: { arxivId, doi, title }, raw });
|
|
113
|
+
});
|
|
114
|
+
// 2) in-text callouts: target #bN → the enclosing sentence/paragraph text.
|
|
115
|
+
// Also record whether the citing sentence cites ONLY this work (soleReferent):
|
|
116
|
+
// a sentence with several markers makes a collective claim not attributable to
|
|
117
|
+
// one work, so Phase-2c won't assert a faithfulness label for it.
|
|
118
|
+
$('ref[type="bibr"]').each((_, el) => {
|
|
119
|
+
const target = ($(el).attr("target") || "").replace(/^#/, "");
|
|
120
|
+
const ctx = refs.get(target);
|
|
121
|
+
if (!ctx || ctx.citingClaim)
|
|
122
|
+
return; // first claim wins; keep it cheap
|
|
123
|
+
const $el = $(el);
|
|
124
|
+
const host = $el.closest("s").length ? $el.closest("s") : $el.closest("p");
|
|
125
|
+
const claim = host.text().replace(/\s+/g, " ").trim();
|
|
126
|
+
if (!claim)
|
|
127
|
+
return;
|
|
128
|
+
const coCited = new Set();
|
|
129
|
+
host.find('ref[type="bibr"]').each((__, r) => {
|
|
130
|
+
const t = ($(r).attr("target") || "").replace(/^#/, "");
|
|
131
|
+
if (t)
|
|
132
|
+
coCited.add(t);
|
|
133
|
+
});
|
|
134
|
+
ctx.citingClaim = claim;
|
|
135
|
+
ctx.soleReferent = coCited.size <= 1;
|
|
136
|
+
});
|
|
137
|
+
const out = Array.from(refs.values()).filter((c) => c.ids.arxivId || c.ids.doi || c.ids.title);
|
|
138
|
+
this.logger.info(`GROBID parsed ${out.length} reference(s), ${out.filter((c) => c.citingClaim).length} with a citing claim`);
|
|
139
|
+
return out;
|
|
140
|
+
}
|
|
141
|
+
}
|
|
142
|
+
exports.GrobidClient = GrobidClient;
|
|
143
|
+
//# sourceMappingURL=GrobidClient.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"GrobidClient.js","sourceRoot":"","sources":["../../../../../src/core/knowledge/references/citations/GrobidClient.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,uCAAyB;AACzB,2CAA6B;AAC7B,iDAAmC;AAuBnC,MAAM,QAAQ,GAAG,2BAA2B,CAAC;AAC7C,MAAM,MAAM,GAAG,iCAAiC,CAAC;AAEjD,MAAa,YAAY;IACvB,YACmB,GAAW,EACX,MAAc,EACd,UAAoB,UAAkB,CAAC,KAAK;QAF5C,QAAG,GAAH,GAAG,CAAQ;QACX,WAAM,GAAN,MAAM,CAAQ;QACd,YAAO,GAAP,OAAO,CAAqC;IAC5D,CAAC;IAEJ,kFAAkF;IAC5E,OAAO;;YACX,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,GAAG,IAAI,CAAC,GAAG,cAAc,CAAC,CAAC;gBACxD,OAAO,CAAC,CAAC,EAAE,CAAC;YACd,CAAC;YAAC,WAAM,CAAC;gBACP,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC;KAAA;IAEK,OAAO,CAAC,QAAgB;;YAC5B,MAAM,GAAG,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YACjD,MAAM,IAAI,GAAG,IAAI,QAAQ,EAAE,CAAC;YAC5B,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,SAAS,CAAC,CAAC;YAC5E,IAAI,CAAC,MAAM,CAAC,sBAAsB,EAAE,GAAG,CAAC,CAAC;YACzC,IAAI,CAAC,MAAM,CAAC,qBAAqB,EAAE,GAAG,CAAC,CAAC;YACxC,IAAI,CAAC,MAAM,CAAC,kBAAkB,EAAE,GAAG,CAAC,CAAC,CAAC,sEAAsE;YAE5G,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,GAAG,IAAI,CAAC,GAAG,8BAA8B,EAAE;gBACxE,MAAM,EAAE,MAAM;gBACd,IAAI,EAAE,IAAI;aACX,CAAC,CAAC;YACH,IAAI,CAAC,GAAG,CAAC,EAAE;gBAAE,MAAM,IAAI,KAAK,CAAC,UAAU,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC;YACrD,OAAO,IAAI,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;QACzC,CAAC;KAAA;IAED,uFAAuF;IACvF,QAAQ,CAAC,GAAW;QAClB,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;QAE/C,8CAA8C;QAC9C,MAAM,IAAI,GAAG,IAAI,GAAG,EAA2B,CAAC;QAChD,CAAC,CAAC,qBAAqB,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;YACtC,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC;YAClB,MAAM,EAAE,GAAG,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YAC9B,IAAI,CAAC,EAAE;gBAAE,OAAO;YAEhB,IAAI,GAAuB,CAAC;YAC5B,IAAI,OAA2B,CAAC;YAChC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,EAAE;;gBACjC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;gBACxD,MAAM,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;gBAClC,IAAI,IAAI,KAAK,KAAK,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC;oBAAE,GAAG,GAAG,GAAG,IAAI,CAAC,MAAA,MAAA,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,0CAAG,CAAC,CAAC,mCAAI,GAAG,CAAC,CAAC;gBACpF,IAAI,IAAI,KAAK,OAAO;oBAAE,OAAO,GAAG,OAAO,IAAI,CAAC,MAAA,MAAA,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,0CAAG,CAAC,CAAC,mCAAI,GAAG,CAAC,CAAC;YAC9E,CAAC,CAAC,CAAC;YACH,MAAM,KAAK,GACT,GAAG,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE;gBAClD,GAAG,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE;gBAClD,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE;gBACvC,SAAS,CAAC;YACZ,MAAM,GAAG,GAAG,GAAG,CAAC,IAAI,CAAC,4BAA4B,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,IAAI,KAAK,IAAI,EAAE,CAAC;YAExF,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,OAAO,EAAE,GAAG,EAAE,KAAK,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;QACtD,CAAC,CAAC,CAAC;QAEH,2EAA2E;QAC3E,+EAA+E;QAC/E,+EAA+E;QAC/E,kEAAkE;QAClE,CAAC,CAAC,kBAAkB,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;YACnC,MAAM,MAAM,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YAC9D,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YAC7B,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC,WAAW;gBAAE,OAAO,CAAC,kCAAkC;YACvE,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC;YAClB,MAAM,IAAI,GAAG,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;YAC3E,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;YACtD,IAAI,CAAC,KAAK;gBAAE,OAAO;YACnB,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;YAClC,IAAI,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE;gBAC3C,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACxD,IAAI,CAAC;oBAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YACxB,CAAC,CAAC,CAAC;YACH,GAAG,CAAC,WAAW,GAAG,KAAK,CAAC;YACxB,GAAG,CAAC,YAAY,GAAG,OAAO,CAAC,IAAI,IAAI,CAAC,CAAC;QACvC,CAAC,CAAC,CAAC;QAEH,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,MAAM,CAC1C,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,IAAI,CAAC,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,KAAK,CACjD,CAAC;QACF,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,iBAAiB,GAAG,CAAC,MAAM,kBAAkB,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,MAAM,sBAAsB,CAC3G,CAAC;QACF,OAAO,GAAG,CAAC;IACb,CAAC;CACF;AA3FD,oCA2FC"}
|
|
@@ -0,0 +1,101 @@
|
|
|
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.TitleIdResolver = void 0;
|
|
13
|
+
const utils_1 = require("../../../../shared/utils");
|
|
14
|
+
const norm = (s) => s.toLowerCase().replace(/\s+/g, " ").trim();
|
|
15
|
+
class TitleIdResolver {
|
|
16
|
+
constructor(opts, logger, fetchFn = globalThis.fetch) {
|
|
17
|
+
this.opts = opts;
|
|
18
|
+
this.logger = logger;
|
|
19
|
+
this.fetchFn = fetchFn;
|
|
20
|
+
}
|
|
21
|
+
resolve(title) {
|
|
22
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
23
|
+
var _a, _b;
|
|
24
|
+
const q = title.trim();
|
|
25
|
+
if (q.length < 8)
|
|
26
|
+
return null; // too short to match reliably
|
|
27
|
+
return ((_b = (_a = (yield this.crossref(q))) !== null && _a !== void 0 ? _a : (yield this.semanticScholar(q))) !== null && _b !== void 0 ? _b : (yield this.openAlex(q)));
|
|
28
|
+
});
|
|
29
|
+
}
|
|
30
|
+
similar(a, b) {
|
|
31
|
+
return !!b && (0, utils_1.jaroWinklerSimilarity)(norm(a), norm(b)) >= this.opts.minTitleSimilarity;
|
|
32
|
+
}
|
|
33
|
+
mailtoParam() {
|
|
34
|
+
return this.opts.mailto ? `&mailto=${encodeURIComponent(this.opts.mailto)}` : "";
|
|
35
|
+
}
|
|
36
|
+
crossref(title) {
|
|
37
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
38
|
+
var _a, _b, _c;
|
|
39
|
+
try {
|
|
40
|
+
const r = yield this.fetchFn(`https://api.crossref.org/works?rows=3&query.bibliographic=${encodeURIComponent(title)}${this.mailtoParam()}`);
|
|
41
|
+
if (!r.ok)
|
|
42
|
+
return null;
|
|
43
|
+
const j = (yield r.json());
|
|
44
|
+
for (const it of (_b = (_a = j.message) === null || _a === void 0 ? void 0 : _a.items) !== null && _b !== void 0 ? _b : []) {
|
|
45
|
+
if (it.DOI && this.similar(title, (_c = it.title) === null || _c === void 0 ? void 0 : _c[0]))
|
|
46
|
+
return { doi: it.DOI };
|
|
47
|
+
}
|
|
48
|
+
}
|
|
49
|
+
catch (err) {
|
|
50
|
+
this.logger.warn(`Crossref title lookup failed: ${err}`);
|
|
51
|
+
}
|
|
52
|
+
return null;
|
|
53
|
+
});
|
|
54
|
+
}
|
|
55
|
+
semanticScholar(title) {
|
|
56
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
57
|
+
var _a, _b, _c;
|
|
58
|
+
try {
|
|
59
|
+
const r = yield this.fetchFn(`https://api.semanticscholar.org/graph/v1/paper/search?limit=3&fields=title,externalIds&query=${encodeURIComponent(title)}`, this.opts.semanticScholarKey ? { headers: { "x-api-key": this.opts.semanticScholarKey } } : undefined);
|
|
60
|
+
if (!r.ok)
|
|
61
|
+
return null;
|
|
62
|
+
const j = (yield r.json());
|
|
63
|
+
for (const it of (_a = j.data) !== null && _a !== void 0 ? _a : []) {
|
|
64
|
+
if (!this.similar(title, it.title))
|
|
65
|
+
continue;
|
|
66
|
+
if ((_b = it.externalIds) === null || _b === void 0 ? void 0 : _b.ArXiv)
|
|
67
|
+
return { arxivId: it.externalIds.ArXiv };
|
|
68
|
+
if ((_c = it.externalIds) === null || _c === void 0 ? void 0 : _c.DOI)
|
|
69
|
+
return { doi: it.externalIds.DOI };
|
|
70
|
+
}
|
|
71
|
+
}
|
|
72
|
+
catch (err) {
|
|
73
|
+
this.logger.warn(`Semantic Scholar title lookup failed: ${err}`);
|
|
74
|
+
}
|
|
75
|
+
return null;
|
|
76
|
+
});
|
|
77
|
+
}
|
|
78
|
+
openAlex(title) {
|
|
79
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
80
|
+
var _a;
|
|
81
|
+
try {
|
|
82
|
+
const key = this.opts.openAlexKey ? `&api_key=${encodeURIComponent(this.opts.openAlexKey)}` : this.mailtoParam();
|
|
83
|
+
const r = yield this.fetchFn(`https://api.openalex.org/works?per_page=3&filter=title.search:${encodeURIComponent(title)}${key}`);
|
|
84
|
+
if (!r.ok)
|
|
85
|
+
return null;
|
|
86
|
+
const j = (yield r.json());
|
|
87
|
+
for (const it of (_a = j.results) !== null && _a !== void 0 ? _a : []) {
|
|
88
|
+
if (it.doi && this.similar(title, it.title)) {
|
|
89
|
+
return { doi: it.doi.replace(/^https?:\/\/(dx\.)?doi\.org\//i, "") };
|
|
90
|
+
}
|
|
91
|
+
}
|
|
92
|
+
}
|
|
93
|
+
catch (err) {
|
|
94
|
+
this.logger.warn(`OpenAlex title lookup failed: ${err}`);
|
|
95
|
+
}
|
|
96
|
+
return null;
|
|
97
|
+
});
|
|
98
|
+
}
|
|
99
|
+
}
|
|
100
|
+
exports.TitleIdResolver = TitleIdResolver;
|
|
101
|
+
//# sourceMappingURL=TitleIdResolver.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"TitleIdResolver.js","sourceRoot":"","sources":["../../../../../src/core/knowledge/references/citations/TitleIdResolver.ts"],"names":[],"mappings":";;;;;;;;;;;;AACA,oDAAiE;AAuBjE,MAAM,IAAI,GAAG,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAExE,MAAa,eAAe;IAC1B,YACmB,IAA4B,EAC5B,MAAc,EACd,UAAoB,UAAkB,CAAC,KAAK;QAF5C,SAAI,GAAJ,IAAI,CAAwB;QAC5B,WAAM,GAAN,MAAM,CAAQ;QACd,YAAO,GAAP,OAAO,CAAqC;IAC5D,CAAC;IAEE,OAAO,CAAC,KAAa;;;YACzB,MAAM,CAAC,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;YACvB,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC;gBAAE,OAAO,IAAI,CAAC,CAAC,8BAA8B;YAC7D,OAAO,CACL,MAAA,MAAA,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,mCAAI,CAAC,MAAM,IAAI,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC,mCAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CACxF,CAAC;QACJ,CAAC;KAAA;IAEO,OAAO,CAAC,CAAS,EAAE,CAAqB;QAC9C,OAAO,CAAC,CAAC,CAAC,IAAI,IAAA,6BAAqB,EAAC,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,kBAAkB,CAAC;IACxF,CAAC;IAEO,WAAW;QACjB,OAAO,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,WAAW,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;IACnF,CAAC;IAEa,QAAQ,CAAC,KAAa;;;YAClC,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAC1B,6DAA6D,kBAAkB,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,WAAW,EAAE,EAAE,CAC9G,CAAC;gBACF,IAAI,CAAC,CAAC,CAAC,EAAE;oBAAE,OAAO,IAAI,CAAC;gBACvB,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAwE,CAAC;gBAClG,KAAK,MAAM,EAAE,IAAI,MAAA,MAAA,CAAC,CAAC,OAAO,0CAAE,KAAK,mCAAI,EAAE,EAAE,CAAC;oBACxC,IAAI,EAAE,CAAC,GAAG,IAAI,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,MAAA,EAAE,CAAC,KAAK,0CAAG,CAAC,CAAC,CAAC;wBAAE,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,EAAE,CAAC;gBAC3E,CAAC;YACH,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,iCAAiC,GAAG,EAAE,CAAC,CAAC;YAC3D,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;KAAA;IAEa,eAAe,CAAC,KAAa;;;YACzC,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAC1B,gGAAgG,kBAAkB,CAAC,KAAK,CAAC,EAAE,EAC3H,IAAI,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CACtG,CAAC;gBACF,IAAI,CAAC,CAAC,CAAC,EAAE;oBAAE,OAAO,IAAI,CAAC;gBACvB,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAExB,CAAC;gBACF,KAAK,MAAM,EAAE,IAAI,MAAA,CAAC,CAAC,IAAI,mCAAI,EAAE,EAAE,CAAC;oBAC9B,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC;wBAAE,SAAS;oBAC7C,IAAI,MAAA,EAAE,CAAC,WAAW,0CAAE,KAAK;wBAAE,OAAO,EAAE,OAAO,EAAE,EAAE,CAAC,WAAW,CAAC,KAAK,EAAE,CAAC;oBACpE,IAAI,MAAA,EAAE,CAAC,WAAW,0CAAE,GAAG;wBAAE,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE,CAAC;gBAC9D,CAAC;YACH,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,yCAAyC,GAAG,EAAE,CAAC,CAAC;YACnE,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;KAAA;IAEa,QAAQ,CAAC,KAAa;;;YAClC,IAAI,CAAC;gBACH,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,YAAY,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;gBACjH,MAAM,CAAC,GAAG,MAAM,IAAI,CAAC,OAAO,CAC1B,iEAAiE,kBAAkB,CAAC,KAAK,CAAC,GAAG,GAAG,EAAE,CACnG,CAAC;gBACF,IAAI,CAAC,CAAC,CAAC,EAAE;oBAAE,OAAO,IAAI,CAAC;gBACvB,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAA0D,CAAC;gBACpF,KAAK,MAAM,EAAE,IAAI,MAAA,CAAC,CAAC,OAAO,mCAAI,EAAE,EAAE,CAAC;oBACjC,IAAI,EAAE,CAAC,GAAG,IAAI,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,EAAE,CAAC;wBAC5C,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,gCAAgC,EAAE,EAAE,CAAC,EAAE,CAAC;oBACvE,CAAC;gBACH,CAAC;YACH,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,iCAAiC,GAAG,EAAE,CAAC,CAAC;YAC3D,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;KAAA;CACF;AA9ED,0CA8EC"}
|
|
@@ -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.FetchCacheService = void 0;
|
|
46
|
+
const fs = __importStar(require("fs"));
|
|
47
|
+
const path = __importStar(require("path"));
|
|
48
|
+
/**
|
|
49
|
+
* Append-only JSONL sidecar for the gated web fetcher — the "never refetch"
|
|
50
|
+
* guard. Mirrors `CheckpointService`: tolerant load (drops a truncated final
|
|
51
|
+
* line), parent-dir ensured on first append, in-memory index keyed by URL.
|
|
52
|
+
*/
|
|
53
|
+
class FetchCacheService {
|
|
54
|
+
constructor(cachePath, logger) {
|
|
55
|
+
this.records = new Map();
|
|
56
|
+
this.loaded = false;
|
|
57
|
+
this.dirEnsured = false;
|
|
58
|
+
this.path = cachePath;
|
|
59
|
+
this.logger = logger;
|
|
60
|
+
}
|
|
61
|
+
getPath() {
|
|
62
|
+
return this.path;
|
|
63
|
+
}
|
|
64
|
+
/** Load any existing cache. Tolerant of a truncated final line. */
|
|
65
|
+
load() {
|
|
66
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
67
|
+
if (this.loaded)
|
|
68
|
+
return this.records.size;
|
|
69
|
+
this.loaded = true;
|
|
70
|
+
if (!fs.existsSync(this.path))
|
|
71
|
+
return 0;
|
|
72
|
+
const content = yield fs.promises.readFile(this.path, "utf-8");
|
|
73
|
+
let loaded = 0;
|
|
74
|
+
let skipped = 0;
|
|
75
|
+
for (const line of content.split("\n")) {
|
|
76
|
+
const trimmed = line.trim();
|
|
77
|
+
if (!trimmed)
|
|
78
|
+
continue;
|
|
79
|
+
try {
|
|
80
|
+
const record = JSON.parse(trimmed);
|
|
81
|
+
if (record.url) {
|
|
82
|
+
this.records.set(record.url, record);
|
|
83
|
+
loaded++;
|
|
84
|
+
}
|
|
85
|
+
}
|
|
86
|
+
catch (_a) {
|
|
87
|
+
skipped++; // truncated/corrupt line — drop it, it'll be regenerated
|
|
88
|
+
}
|
|
89
|
+
}
|
|
90
|
+
this.logger.info(`Loaded ${loaded} cached web fetch(es) from ${this.path}` +
|
|
91
|
+
(skipped ? ` (skipped ${skipped} corrupt line(s))` : ""));
|
|
92
|
+
return loaded;
|
|
93
|
+
});
|
|
94
|
+
}
|
|
95
|
+
has(url) {
|
|
96
|
+
return this.records.has(url);
|
|
97
|
+
}
|
|
98
|
+
get(url) {
|
|
99
|
+
return this.records.get(url);
|
|
100
|
+
}
|
|
101
|
+
/** Persist a fetch outcome and keep the in-memory index in sync. */
|
|
102
|
+
append(record) {
|
|
103
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
104
|
+
this.records.set(record.url, record);
|
|
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(record) + "\n");
|
|
110
|
+
});
|
|
111
|
+
}
|
|
112
|
+
}
|
|
113
|
+
exports.FetchCacheService = FetchCacheService;
|
|
114
|
+
//# sourceMappingURL=FetchCacheService.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"FetchCacheService.js","sourceRoot":"","sources":["../../../../../src/core/knowledge/references/web/FetchCacheService.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,uCAAyB;AACzB,2CAA6B;AAoB7B;;;;GAIG;AACH,MAAa,iBAAiB;IAO5B,YAAY,SAAiB,EAAE,MAAc;QAJrC,YAAO,GAAkC,IAAI,GAAG,EAAE,CAAC;QACnD,WAAM,GAAG,KAAK,CAAC;QACf,eAAU,GAAG,KAAK,CAAC;QAGzB,IAAI,CAAC,IAAI,GAAG,SAAS,CAAC;QACtB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED,OAAO;QACL,OAAO,IAAI,CAAC,IAAI,CAAC;IACnB,CAAC;IAED,mEAAmE;IAC7D,IAAI;;YACR,IAAI,IAAI,CAAC,MAAM;gBAAE,OAAO,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC;YAC1C,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;YAEnB,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,CAAqB,CAAC;oBACvD,IAAI,MAAM,CAAC,GAAG,EAAE,CAAC;wBACf,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;wBACrC,MAAM,EAAE,CAAC;oBACX,CAAC;gBACH,CAAC;gBAAC,WAAM,CAAC;oBACP,OAAO,EAAE,CAAC,CAAC,yDAAyD;gBACtE,CAAC;YACH,CAAC;YACD,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,UAAU,MAAM,8BAA8B,IAAI,CAAC,IAAI,EAAE;gBACvD,CAAC,OAAO,CAAC,CAAC,CAAC,aAAa,OAAO,mBAAmB,CAAC,CAAC,CAAC,EAAE,CAAC,CAC3D,CAAC;YACF,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,oEAAoE;IAC9D,MAAM,CAAC,MAAwB;;YACnC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;YACrC,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,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC;QACzE,CAAC;KAAA;CACF;AA/DD,8CA+DC"}
|