@steno-ai/engine 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/dist/adapters/cache.d.ts +9 -0
- package/dist/adapters/cache.d.ts.map +1 -0
- package/dist/adapters/cache.js +2 -0
- package/dist/adapters/cache.js.map +1 -0
- package/dist/adapters/embedding.d.ts +7 -0
- package/dist/adapters/embedding.d.ts.map +1 -0
- package/dist/adapters/embedding.js +2 -0
- package/dist/adapters/embedding.js.map +1 -0
- package/dist/adapters/gemini-embedding.d.ts +18 -0
- package/dist/adapters/gemini-embedding.d.ts.map +1 -0
- package/dist/adapters/gemini-embedding.js +53 -0
- package/dist/adapters/gemini-embedding.js.map +1 -0
- package/dist/adapters/index.d.ts +7 -0
- package/dist/adapters/index.d.ts.map +1 -0
- package/dist/adapters/index.js +7 -0
- package/dist/adapters/index.js.map +1 -0
- package/dist/adapters/llm.d.ts +19 -0
- package/dist/adapters/llm.d.ts.map +1 -0
- package/dist/adapters/llm.js +2 -0
- package/dist/adapters/llm.js.map +1 -0
- package/dist/adapters/perplexity-embedding.d.ts +24 -0
- package/dist/adapters/perplexity-embedding.d.ts.map +1 -0
- package/dist/adapters/perplexity-embedding.js +78 -0
- package/dist/adapters/perplexity-embedding.js.map +1 -0
- package/dist/adapters/storage.d.ts +172 -0
- package/dist/adapters/storage.d.ts.map +1 -0
- package/dist/adapters/storage.js +2 -0
- package/dist/adapters/storage.js.map +1 -0
- package/dist/auth/api-key.d.ts +8 -0
- package/dist/auth/api-key.d.ts.map +1 -0
- package/dist/auth/api-key.js +27 -0
- package/dist/auth/api-key.js.map +1 -0
- package/dist/auth/index.d.ts +2 -0
- package/dist/auth/index.d.ts.map +1 -0
- package/dist/auth/index.js +2 -0
- package/dist/auth/index.js.map +1 -0
- package/dist/config.d.ts +296 -0
- package/dist/config.d.ts.map +1 -0
- package/dist/config.js +92 -0
- package/dist/config.js.map +1 -0
- package/dist/extraction/contradiction.d.ts +15 -0
- package/dist/extraction/contradiction.d.ts.map +1 -0
- package/dist/extraction/contradiction.js +23 -0
- package/dist/extraction/contradiction.js.map +1 -0
- package/dist/extraction/dedup.d.ts +12 -0
- package/dist/extraction/dedup.d.ts.map +1 -0
- package/dist/extraction/dedup.js +93 -0
- package/dist/extraction/dedup.js.map +1 -0
- package/dist/extraction/entity-extractor.d.ts +30 -0
- package/dist/extraction/entity-extractor.d.ts.map +1 -0
- package/dist/extraction/entity-extractor.js +145 -0
- package/dist/extraction/entity-extractor.js.map +1 -0
- package/dist/extraction/hasher.d.ts +5 -0
- package/dist/extraction/hasher.d.ts.map +1 -0
- package/dist/extraction/hasher.js +8 -0
- package/dist/extraction/hasher.js.map +1 -0
- package/dist/extraction/heuristic.d.ts +3 -0
- package/dist/extraction/heuristic.d.ts.map +1 -0
- package/dist/extraction/heuristic.js +282 -0
- package/dist/extraction/heuristic.js.map +1 -0
- package/dist/extraction/index.d.ts +10 -0
- package/dist/extraction/index.d.ts.map +1 -0
- package/dist/extraction/index.js +10 -0
- package/dist/extraction/index.js.map +1 -0
- package/dist/extraction/llm-extractor.d.ts +23 -0
- package/dist/extraction/llm-extractor.d.ts.map +1 -0
- package/dist/extraction/llm-extractor.js +238 -0
- package/dist/extraction/llm-extractor.js.map +1 -0
- package/dist/extraction/pipeline.d.ts +30 -0
- package/dist/extraction/pipeline.d.ts.map +1 -0
- package/dist/extraction/pipeline.js +398 -0
- package/dist/extraction/pipeline.js.map +1 -0
- package/dist/extraction/prompts.d.ts +28 -0
- package/dist/extraction/prompts.d.ts.map +1 -0
- package/dist/extraction/prompts.js +196 -0
- package/dist/extraction/prompts.js.map +1 -0
- package/dist/extraction/sliding-window.d.ts +41 -0
- package/dist/extraction/sliding-window.d.ts.map +1 -0
- package/dist/extraction/sliding-window.js +84 -0
- package/dist/extraction/sliding-window.js.map +1 -0
- package/dist/extraction/types.d.ts +80 -0
- package/dist/extraction/types.d.ts.map +1 -0
- package/dist/extraction/types.js +2 -0
- package/dist/extraction/types.js.map +1 -0
- package/dist/feedback/index.d.ts +2 -0
- package/dist/feedback/index.d.ts.map +1 -0
- package/dist/feedback/index.js +2 -0
- package/dist/feedback/index.js.map +1 -0
- package/dist/feedback/tracker.d.ts +25 -0
- package/dist/feedback/tracker.d.ts.map +1 -0
- package/dist/feedback/tracker.js +90 -0
- package/dist/feedback/tracker.js.map +1 -0
- package/dist/index.d.ts +12 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +13 -0
- package/dist/index.js.map +1 -0
- package/dist/models/api-key.d.ts +54 -0
- package/dist/models/api-key.d.ts.map +1 -0
- package/dist/models/api-key.js +21 -0
- package/dist/models/api-key.js.map +1 -0
- package/dist/models/edge.d.ts +78 -0
- package/dist/models/edge.d.ts.map +1 -0
- package/dist/models/edge.js +29 -0
- package/dist/models/edge.js.map +1 -0
- package/dist/models/entity.d.ts +60 -0
- package/dist/models/entity.d.ts.map +1 -0
- package/dist/models/entity.js +22 -0
- package/dist/models/entity.js.map +1 -0
- package/dist/models/extraction.d.ts +111 -0
- package/dist/models/extraction.d.ts.map +1 -0
- package/dist/models/extraction.js +40 -0
- package/dist/models/extraction.js.map +1 -0
- package/dist/models/fact-entity.d.ts +33 -0
- package/dist/models/fact-entity.d.ts.map +1 -0
- package/dist/models/fact-entity.js +14 -0
- package/dist/models/fact-entity.js.map +1 -0
- package/dist/models/fact.d.ts +191 -0
- package/dist/models/fact.d.ts.map +1 -0
- package/dist/models/fact.js +72 -0
- package/dist/models/fact.js.map +1 -0
- package/dist/models/index.d.ts +13 -0
- package/dist/models/index.d.ts.map +1 -0
- package/dist/models/index.js +13 -0
- package/dist/models/index.js.map +1 -0
- package/dist/models/memory-access.d.ts +89 -0
- package/dist/models/memory-access.d.ts.map +1 -0
- package/dist/models/memory-access.js +33 -0
- package/dist/models/memory-access.js.map +1 -0
- package/dist/models/session.d.ts +60 -0
- package/dist/models/session.d.ts.map +1 -0
- package/dist/models/session.js +23 -0
- package/dist/models/session.js.map +1 -0
- package/dist/models/tenant.d.ts +448 -0
- package/dist/models/tenant.d.ts.map +1 -0
- package/dist/models/tenant.js +23 -0
- package/dist/models/tenant.js.map +1 -0
- package/dist/models/trigger.d.ts +87 -0
- package/dist/models/trigger.d.ts.map +1 -0
- package/dist/models/trigger.js +41 -0
- package/dist/models/trigger.js.map +1 -0
- package/dist/models/usage-record.d.ts +37 -0
- package/dist/models/usage-record.d.ts.map +1 -0
- package/dist/models/usage-record.js +14 -0
- package/dist/models/usage-record.js.map +1 -0
- package/dist/models/webhook.d.ts +50 -0
- package/dist/models/webhook.d.ts.map +1 -0
- package/dist/models/webhook.js +25 -0
- package/dist/models/webhook.js.map +1 -0
- package/dist/profiles/index.d.ts +3 -0
- package/dist/profiles/index.d.ts.map +1 -0
- package/dist/profiles/index.js +2 -0
- package/dist/profiles/index.js.map +1 -0
- package/dist/profiles/profile.d.ts +22 -0
- package/dist/profiles/profile.d.ts.map +1 -0
- package/dist/profiles/profile.js +59 -0
- package/dist/profiles/profile.js.map +1 -0
- package/dist/retrieval/compound-search.d.ts +13 -0
- package/dist/retrieval/compound-search.d.ts.map +1 -0
- package/dist/retrieval/compound-search.js +87 -0
- package/dist/retrieval/compound-search.js.map +1 -0
- package/dist/retrieval/contradiction-surfacer.d.ts +18 -0
- package/dist/retrieval/contradiction-surfacer.d.ts.map +1 -0
- package/dist/retrieval/contradiction-surfacer.js +64 -0
- package/dist/retrieval/contradiction-surfacer.js.map +1 -0
- package/dist/retrieval/embedding-cache.d.ts +17 -0
- package/dist/retrieval/embedding-cache.d.ts.map +1 -0
- package/dist/retrieval/embedding-cache.js +56 -0
- package/dist/retrieval/embedding-cache.js.map +1 -0
- package/dist/retrieval/fusion.d.ts +27 -0
- package/dist/retrieval/fusion.d.ts.map +1 -0
- package/dist/retrieval/fusion.js +87 -0
- package/dist/retrieval/fusion.js.map +1 -0
- package/dist/retrieval/graph-traversal.d.ts +29 -0
- package/dist/retrieval/graph-traversal.d.ts.map +1 -0
- package/dist/retrieval/graph-traversal.js +208 -0
- package/dist/retrieval/graph-traversal.js.map +1 -0
- package/dist/retrieval/index.d.ts +14 -0
- package/dist/retrieval/index.d.ts.map +1 -0
- package/dist/retrieval/index.js +13 -0
- package/dist/retrieval/index.js.map +1 -0
- package/dist/retrieval/keyword-search.d.ts +4 -0
- package/dist/retrieval/keyword-search.d.ts.map +1 -0
- package/dist/retrieval/keyword-search.js +27 -0
- package/dist/retrieval/keyword-search.js.map +1 -0
- package/dist/retrieval/query-expansion.d.ts +20 -0
- package/dist/retrieval/query-expansion.d.ts.map +1 -0
- package/dist/retrieval/query-expansion.js +76 -0
- package/dist/retrieval/query-expansion.js.map +1 -0
- package/dist/retrieval/reranker.d.ts +15 -0
- package/dist/retrieval/reranker.d.ts.map +1 -0
- package/dist/retrieval/reranker.js +47 -0
- package/dist/retrieval/reranker.js.map +1 -0
- package/dist/retrieval/salience-scorer.d.ts +15 -0
- package/dist/retrieval/salience-scorer.d.ts.map +1 -0
- package/dist/retrieval/salience-scorer.js +41 -0
- package/dist/retrieval/salience-scorer.js.map +1 -0
- package/dist/retrieval/search.d.ts +21 -0
- package/dist/retrieval/search.d.ts.map +1 -0
- package/dist/retrieval/search.js +228 -0
- package/dist/retrieval/search.js.map +1 -0
- package/dist/retrieval/temporal-scorer.d.ts +18 -0
- package/dist/retrieval/temporal-scorer.d.ts.map +1 -0
- package/dist/retrieval/temporal-scorer.js +106 -0
- package/dist/retrieval/temporal-scorer.js.map +1 -0
- package/dist/retrieval/trigger-matcher.d.ts +18 -0
- package/dist/retrieval/trigger-matcher.d.ts.map +1 -0
- package/dist/retrieval/trigger-matcher.js +134 -0
- package/dist/retrieval/trigger-matcher.js.map +1 -0
- package/dist/retrieval/types.d.ts +70 -0
- package/dist/retrieval/types.d.ts.map +1 -0
- package/dist/retrieval/types.js +9 -0
- package/dist/retrieval/types.js.map +1 -0
- package/dist/retrieval/vector-search.d.ts +5 -0
- package/dist/retrieval/vector-search.d.ts.map +1 -0
- package/dist/retrieval/vector-search.js +24 -0
- package/dist/retrieval/vector-search.js.map +1 -0
- package/dist/salience/decay.d.ts +9 -0
- package/dist/salience/decay.d.ts.map +1 -0
- package/dist/salience/decay.js +15 -0
- package/dist/salience/decay.js.map +1 -0
- package/dist/salience/index.d.ts +2 -0
- package/dist/salience/index.d.ts.map +1 -0
- package/dist/salience/index.js +2 -0
- package/dist/salience/index.js.map +1 -0
- package/dist/scratchpad/index.d.ts +2 -0
- package/dist/scratchpad/index.d.ts.map +1 -0
- package/dist/scratchpad/index.js +2 -0
- package/dist/scratchpad/index.js.map +1 -0
- package/dist/scratchpad/scratchpad.d.ts +23 -0
- package/dist/scratchpad/scratchpad.d.ts.map +1 -0
- package/dist/scratchpad/scratchpad.js +107 -0
- package/dist/scratchpad/scratchpad.js.map +1 -0
- package/dist/sessions/index.d.ts +2 -0
- package/dist/sessions/index.d.ts.map +1 -0
- package/dist/sessions/index.js +2 -0
- package/dist/sessions/index.js.map +1 -0
- package/dist/sessions/manager.d.ts +11 -0
- package/dist/sessions/manager.d.ts.map +1 -0
- package/dist/sessions/manager.js +63 -0
- package/dist/sessions/manager.js.map +1 -0
- package/package.json +38 -0
- package/src/adapters/cache.d.ts +9 -0
- package/src/adapters/cache.d.ts.map +1 -0
- package/src/adapters/cache.js.map +1 -0
- package/src/adapters/cache.ts +8 -0
- package/src/adapters/embedding.d.ts +7 -0
- package/src/adapters/embedding.d.ts.map +1 -0
- package/src/adapters/embedding.js.map +1 -0
- package/src/adapters/embedding.ts +6 -0
- package/src/adapters/gemini-embedding.ts +67 -0
- package/src/adapters/index.ts +6 -0
- package/src/adapters/llm.d.ts +19 -0
- package/src/adapters/llm.d.ts.map +1 -0
- package/src/adapters/llm.js.map +1 -0
- package/src/adapters/llm.ts +16 -0
- package/src/adapters/perplexity-embedding.d.ts +24 -0
- package/src/adapters/perplexity-embedding.d.ts.map +1 -0
- package/src/adapters/perplexity-embedding.js.map +1 -0
- package/src/adapters/perplexity-embedding.ts +98 -0
- package/src/adapters/storage.d.ts +172 -0
- package/src/adapters/storage.d.ts.map +1 -0
- package/src/adapters/storage.js.map +1 -0
- package/src/adapters/storage.ts +187 -0
- package/src/auth/api-key.ts +33 -0
- package/src/auth/index.ts +1 -0
- package/src/config.d.ts +86 -0
- package/src/config.d.ts.map +1 -0
- package/src/config.js.map +1 -0
- package/src/config.ts +131 -0
- package/src/extraction/contradiction.d.ts +15 -0
- package/src/extraction/contradiction.d.ts.map +1 -0
- package/src/extraction/contradiction.js.map +1 -0
- package/src/extraction/contradiction.ts +33 -0
- package/src/extraction/dedup.d.ts +12 -0
- package/src/extraction/dedup.d.ts.map +1 -0
- package/src/extraction/dedup.js.map +1 -0
- package/src/extraction/dedup.ts +133 -0
- package/src/extraction/entity-extractor.d.ts +30 -0
- package/src/extraction/entity-extractor.d.ts.map +1 -0
- package/src/extraction/entity-extractor.js.map +1 -0
- package/src/extraction/entity-extractor.ts +193 -0
- package/src/extraction/hasher.d.ts +5 -0
- package/src/extraction/hasher.d.ts.map +1 -0
- package/src/extraction/hasher.js.map +1 -0
- package/src/extraction/hasher.ts +7 -0
- package/src/extraction/heuristic.d.ts +3 -0
- package/src/extraction/heuristic.d.ts.map +1 -0
- package/src/extraction/heuristic.js.map +1 -0
- package/src/extraction/heuristic.ts +341 -0
- package/src/extraction/index.ts +9 -0
- package/src/extraction/llm-extractor.d.ts +21 -0
- package/src/extraction/llm-extractor.d.ts.map +1 -0
- package/src/extraction/llm-extractor.js.map +1 -0
- package/src/extraction/llm-extractor.ts +267 -0
- package/src/extraction/pipeline.d.ts +27 -0
- package/src/extraction/pipeline.d.ts.map +1 -0
- package/src/extraction/pipeline.js.map +1 -0
- package/src/extraction/pipeline.ts +515 -0
- package/src/extraction/prompts.js.map +1 -0
- package/src/extraction/prompts.ts +233 -0
- package/src/extraction/sliding-window.d.ts +41 -0
- package/src/extraction/sliding-window.d.ts.map +1 -0
- package/src/extraction/sliding-window.js.map +1 -0
- package/src/extraction/sliding-window.ts +121 -0
- package/src/extraction/types.d.ts +68 -0
- package/src/extraction/types.d.ts.map +1 -0
- package/src/extraction/types.js.map +1 -0
- package/src/extraction/types.ts +80 -0
- package/src/feedback/index.ts +1 -0
- package/src/feedback/tracker.d.ts +25 -0
- package/src/feedback/tracker.d.ts.map +1 -0
- package/src/feedback/tracker.js.map +1 -0
- package/src/feedback/tracker.ts +119 -0
- package/src/index.ts +12 -0
- package/src/models/api-key.d.ts +54 -0
- package/src/models/api-key.d.ts.map +1 -0
- package/src/models/api-key.js.map +1 -0
- package/src/models/api-key.ts +26 -0
- package/src/models/edge.d.ts +78 -0
- package/src/models/edge.d.ts.map +1 -0
- package/src/models/edge.js.map +1 -0
- package/src/models/edge.ts +34 -0
- package/src/models/entity.d.ts +60 -0
- package/src/models/entity.d.ts.map +1 -0
- package/src/models/entity.js.map +1 -0
- package/src/models/entity.ts +27 -0
- package/src/models/extraction.d.ts +111 -0
- package/src/models/extraction.d.ts.map +1 -0
- package/src/models/extraction.js.map +1 -0
- package/src/models/extraction.ts +45 -0
- package/src/models/fact-entity.d.ts +33 -0
- package/src/models/fact-entity.d.ts.map +1 -0
- package/src/models/fact-entity.js.map +1 -0
- package/src/models/fact-entity.ts +19 -0
- package/src/models/fact.ts +85 -0
- package/src/models/index.d.ts +13 -0
- package/src/models/index.d.ts.map +1 -0
- package/src/models/index.js.map +1 -0
- package/src/models/index.ts +12 -0
- package/src/models/memory-access.d.ts +89 -0
- package/src/models/memory-access.d.ts.map +1 -0
- package/src/models/memory-access.js.map +1 -0
- package/src/models/memory-access.ts +41 -0
- package/src/models/session.d.ts +60 -0
- package/src/models/session.d.ts.map +1 -0
- package/src/models/session.js.map +1 -0
- package/src/models/session.ts +28 -0
- package/src/models/tenant.d.ts +214 -0
- package/src/models/tenant.d.ts.map +1 -0
- package/src/models/tenant.js.map +1 -0
- package/src/models/tenant.ts +28 -0
- package/src/models/trigger.d.ts +87 -0
- package/src/models/trigger.d.ts.map +1 -0
- package/src/models/trigger.js.map +1 -0
- package/src/models/trigger.ts +58 -0
- package/src/models/usage-record.d.ts +37 -0
- package/src/models/usage-record.d.ts.map +1 -0
- package/src/models/usage-record.js.map +1 -0
- package/src/models/usage-record.ts +16 -0
- package/src/models/webhook.d.ts +50 -0
- package/src/models/webhook.d.ts.map +1 -0
- package/src/models/webhook.js.map +1 -0
- package/src/models/webhook.ts +30 -0
- package/src/profiles/index.ts +2 -0
- package/src/profiles/profile.ts +81 -0
- package/src/retrieval/compound-search.d.ts +13 -0
- package/src/retrieval/compound-search.d.ts.map +1 -0
- package/src/retrieval/compound-search.js.map +1 -0
- package/src/retrieval/compound-search.ts +104 -0
- package/src/retrieval/contradiction-surfacer.d.ts +18 -0
- package/src/retrieval/contradiction-surfacer.d.ts.map +1 -0
- package/src/retrieval/contradiction-surfacer.js.map +1 -0
- package/src/retrieval/contradiction-surfacer.ts +87 -0
- package/src/retrieval/embedding-cache.d.ts +17 -0
- package/src/retrieval/embedding-cache.d.ts.map +1 -0
- package/src/retrieval/embedding-cache.js.map +1 -0
- package/src/retrieval/embedding-cache.ts +63 -0
- package/src/retrieval/fusion.d.ts +26 -0
- package/src/retrieval/fusion.d.ts.map +1 -0
- package/src/retrieval/fusion.js.map +1 -0
- package/src/retrieval/fusion.ts +129 -0
- package/src/retrieval/graph-traversal.d.ts +28 -0
- package/src/retrieval/graph-traversal.d.ts.map +1 -0
- package/src/retrieval/graph-traversal.js.map +1 -0
- package/src/retrieval/graph-traversal.ts +235 -0
- package/src/retrieval/index.ts +13 -0
- package/src/retrieval/keyword-search.ts +39 -0
- package/src/retrieval/query-expansion.d.ts +20 -0
- package/src/retrieval/query-expansion.d.ts.map +1 -0
- package/src/retrieval/query-expansion.js.map +1 -0
- package/src/retrieval/query-expansion.ts +86 -0
- package/src/retrieval/reranker.d.ts +15 -0
- package/src/retrieval/reranker.d.ts.map +1 -0
- package/src/retrieval/reranker.js.map +1 -0
- package/src/retrieval/reranker.ts +56 -0
- package/src/retrieval/salience-scorer.d.ts +15 -0
- package/src/retrieval/salience-scorer.d.ts.map +1 -0
- package/src/retrieval/salience-scorer.js.map +1 -0
- package/src/retrieval/salience-scorer.ts +57 -0
- package/src/retrieval/search.d.ts +21 -0
- package/src/retrieval/search.d.ts.map +1 -0
- package/src/retrieval/search.js.map +1 -0
- package/src/retrieval/search.ts +271 -0
- package/src/retrieval/temporal-scorer.ts +111 -0
- package/src/retrieval/trigger-matcher.d.ts +18 -0
- package/src/retrieval/trigger-matcher.d.ts.map +1 -0
- package/src/retrieval/trigger-matcher.js.map +1 -0
- package/src/retrieval/trigger-matcher.ts +180 -0
- package/src/retrieval/types.d.ts +66 -0
- package/src/retrieval/types.d.ts.map +1 -0
- package/src/retrieval/types.js.map +1 -0
- package/src/retrieval/types.ts +82 -0
- package/src/retrieval/vector-search.d.ts +5 -0
- package/src/retrieval/vector-search.d.ts.map +1 -0
- package/src/retrieval/vector-search.js.map +1 -0
- package/src/retrieval/vector-search.ts +38 -0
- package/src/salience/decay.d.ts +9 -0
- package/src/salience/decay.d.ts.map +1 -0
- package/src/salience/decay.js.map +1 -0
- package/src/salience/decay.ts +25 -0
- package/src/salience/index.ts +1 -0
- package/src/scratchpad/index.ts +1 -0
- package/src/scratchpad/scratchpad.d.ts +23 -0
- package/src/scratchpad/scratchpad.d.ts.map +1 -0
- package/src/scratchpad/scratchpad.js.map +1 -0
- package/src/scratchpad/scratchpad.ts +140 -0
- package/src/sessions/index.ts +1 -0
- package/src/sessions/manager.ts +87 -0
|
@@ -0,0 +1,145 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Create or find all entities in the database.
|
|
3
|
+
* Does NOT link entities to any fact or create edges.
|
|
4
|
+
* Returns a map of canonicalName → entity.id for use in subsequent operations.
|
|
5
|
+
*/
|
|
6
|
+
export async function buildEntityIdMap(storage, embedding, tenantId, entities) {
|
|
7
|
+
const entityIdMap = new Map();
|
|
8
|
+
let entitiesCreated = 0;
|
|
9
|
+
// Deduplicate entities by canonical name within this batch
|
|
10
|
+
const uniqueEntities = new Map();
|
|
11
|
+
for (const entity of entities) {
|
|
12
|
+
if (!uniqueEntities.has(entity.canonicalName)) {
|
|
13
|
+
uniqueEntities.set(entity.canonicalName, entity);
|
|
14
|
+
}
|
|
15
|
+
}
|
|
16
|
+
// Create or find each entity
|
|
17
|
+
for (const entity of uniqueEntities.values()) {
|
|
18
|
+
const existing = await storage.findEntityByCanonicalName(tenantId, entity.canonicalName, entity.entityType);
|
|
19
|
+
if (existing) {
|
|
20
|
+
entityIdMap.set(entity.canonicalName, existing.id);
|
|
21
|
+
}
|
|
22
|
+
else {
|
|
23
|
+
const id = crypto.randomUUID();
|
|
24
|
+
const emb = await embedding.embed(entity.name);
|
|
25
|
+
await storage.createEntity({
|
|
26
|
+
...entity,
|
|
27
|
+
id,
|
|
28
|
+
tenantId,
|
|
29
|
+
embedding: emb,
|
|
30
|
+
embeddingModel: embedding.model,
|
|
31
|
+
embeddingDim: embedding.dimensions,
|
|
32
|
+
});
|
|
33
|
+
entityIdMap.set(entity.canonicalName, id);
|
|
34
|
+
entitiesCreated++;
|
|
35
|
+
}
|
|
36
|
+
}
|
|
37
|
+
return { entityIdMap, entitiesCreated };
|
|
38
|
+
}
|
|
39
|
+
/**
|
|
40
|
+
* Create edges between already-persisted entities.
|
|
41
|
+
* Uses the entityIdMap built by buildEntityIdMap.
|
|
42
|
+
*/
|
|
43
|
+
// Normalize relation names so "loves_deeply" and "loves" merge
|
|
44
|
+
const RELATION_SYNONYMS = {
|
|
45
|
+
loves_deeply: 'loves',
|
|
46
|
+
has_relationship: 'partner_of',
|
|
47
|
+
has_relationship_with: 'partner_of',
|
|
48
|
+
dating: 'partner_of',
|
|
49
|
+
in_relationship_with: 'partner_of',
|
|
50
|
+
romantic_partner: 'partner_of',
|
|
51
|
+
girlfriend_of: 'partner_of',
|
|
52
|
+
boyfriend_of: 'partner_of',
|
|
53
|
+
married_to: 'partner_of',
|
|
54
|
+
works_for: 'works_at',
|
|
55
|
+
employed_at: 'works_at',
|
|
56
|
+
employed_by: 'works_at',
|
|
57
|
+
resides_in: 'lives_in',
|
|
58
|
+
located_at: 'located_in',
|
|
59
|
+
friends_with: 'friend_of',
|
|
60
|
+
acquainted_with: 'knows',
|
|
61
|
+
knows_about: 'knows',
|
|
62
|
+
interested_in: 'prefers',
|
|
63
|
+
attracted_to: 'prefers',
|
|
64
|
+
likes: 'prefers',
|
|
65
|
+
replaces: 'updates',
|
|
66
|
+
supersedes: 'updates',
|
|
67
|
+
overrides: 'updates',
|
|
68
|
+
corrects: 'updates',
|
|
69
|
+
supplements: 'extends',
|
|
70
|
+
refines: 'extends',
|
|
71
|
+
elaborates: 'extends',
|
|
72
|
+
details: 'extends',
|
|
73
|
+
infers: 'derives',
|
|
74
|
+
implies: 'derives',
|
|
75
|
+
combines: 'derives',
|
|
76
|
+
};
|
|
77
|
+
function normalizeRelation(relation) {
|
|
78
|
+
const lower = relation.toLowerCase().trim();
|
|
79
|
+
return RELATION_SYNONYMS[lower] ?? lower;
|
|
80
|
+
}
|
|
81
|
+
export async function persistEdges(storage, tenantId, factId, edges, entityIdMap) {
|
|
82
|
+
// Deduplicate edges by (source, target, normalized_relation) within this batch
|
|
83
|
+
const seen = new Set();
|
|
84
|
+
const dedupedEdges = [];
|
|
85
|
+
for (const edge of edges) {
|
|
86
|
+
const normalizedRelation = normalizeRelation(edge.relation);
|
|
87
|
+
const key = `${edge.sourceName}|${normalizedRelation}|${edge.targetName}`;
|
|
88
|
+
if (!seen.has(key)) {
|
|
89
|
+
seen.add(key);
|
|
90
|
+
dedupedEdges.push({ ...edge, relation: normalizedRelation });
|
|
91
|
+
}
|
|
92
|
+
}
|
|
93
|
+
let edgesCreated = 0;
|
|
94
|
+
for (const edge of dedupedEdges) {
|
|
95
|
+
const sourceId = entityIdMap.get(edge.sourceName);
|
|
96
|
+
const targetId = entityIdMap.get(edge.targetName);
|
|
97
|
+
if (sourceId && targetId) {
|
|
98
|
+
try {
|
|
99
|
+
await storage.createEdge({
|
|
100
|
+
tenantId,
|
|
101
|
+
sourceId,
|
|
102
|
+
targetId,
|
|
103
|
+
relation: edge.relation,
|
|
104
|
+
edgeType: edge.edgeType,
|
|
105
|
+
confidence: edge.confidence,
|
|
106
|
+
weight: 1.0,
|
|
107
|
+
metadata: {},
|
|
108
|
+
factId,
|
|
109
|
+
id: crypto.randomUUID(),
|
|
110
|
+
});
|
|
111
|
+
edgesCreated++;
|
|
112
|
+
}
|
|
113
|
+
catch {
|
|
114
|
+
// Edge creation failed (e.g., duplicate) — continue
|
|
115
|
+
}
|
|
116
|
+
}
|
|
117
|
+
if (!sourceId || !targetId) {
|
|
118
|
+
console.warn(`[steno] Edge dropped: "${edge.sourceName}" → "${edge.relation}" → "${edge.targetName}" ` +
|
|
119
|
+
`(source=${sourceId ? 'found' : 'MISSING'}, target=${targetId ? 'found' : 'MISSING'}) ` +
|
|
120
|
+
`entityMap keys: [${[...entityIdMap.keys()].join(', ')}]`);
|
|
121
|
+
}
|
|
122
|
+
}
|
|
123
|
+
return edgesCreated;
|
|
124
|
+
}
|
|
125
|
+
/**
|
|
126
|
+
* Persist extracted entities and edges to the database.
|
|
127
|
+
* - Creates new entities if canonical name doesn't exist for this tenant
|
|
128
|
+
* - Reuses existing entities if canonical name matches
|
|
129
|
+
* - Links all entities to the fact via fact_entities junction
|
|
130
|
+
* - Creates edges between entities
|
|
131
|
+
*
|
|
132
|
+
* @deprecated Use buildEntityIdMap + persistEdges directly for better control.
|
|
133
|
+
* This function is kept for backward compatibility.
|
|
134
|
+
*/
|
|
135
|
+
export async function persistEntitiesAndEdges(storage, embedding, tenantId, factId, entities, edges) {
|
|
136
|
+
const { entityIdMap, entitiesCreated } = await buildEntityIdMap(storage, embedding, tenantId, entities);
|
|
137
|
+
// Link all entities to this fact
|
|
138
|
+
for (const [canonicalName, entityId] of entityIdMap) {
|
|
139
|
+
void canonicalName; // used as key only
|
|
140
|
+
await storage.linkFactEntity(factId, entityId, 'mentioned');
|
|
141
|
+
}
|
|
142
|
+
const edgesCreated = await persistEdges(storage, tenantId, factId, edges, entityIdMap);
|
|
143
|
+
return { entitiesCreated, edgesCreated, entityIdMap };
|
|
144
|
+
}
|
|
145
|
+
//# sourceMappingURL=entity-extractor.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"entity-extractor.js","sourceRoot":"","sources":["../../src/extraction/entity-extractor.ts"],"names":[],"mappings":"AAUA;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CACpC,OAAuB,EACvB,SAA2B,EAC3B,QAAgB,EAChB,QAA2B;IAE3B,MAAM,WAAW,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC9C,IAAI,eAAe,GAAG,CAAC,CAAC;IAExB,2DAA2D;IAC3D,MAAM,cAAc,GAAG,IAAI,GAAG,EAA2B,CAAC;IAC1D,KAAK,MAAM,MAAM,IAAI,QAAQ,EAAE,CAAC;QAC9B,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,aAAa,CAAC,EAAE,CAAC;YAC9C,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,aAAa,EAAE,MAAM,CAAC,CAAC;QACnD,CAAC;IACH,CAAC;IAED,6BAA6B;IAC7B,KAAK,MAAM,MAAM,IAAI,cAAc,CAAC,MAAM,EAAE,EAAE,CAAC;QAC7C,MAAM,QAAQ,GAAG,MAAM,OAAO,CAAC,yBAAyB,CACtD,QAAQ,EACR,MAAM,CAAC,aAAa,EACpB,MAAM,CAAC,UAAU,CAClB,CAAC;QAEF,IAAI,QAAQ,EAAE,CAAC;YACb,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,aAAa,EAAE,QAAQ,CAAC,EAAE,CAAC,CAAC;QACrD,CAAC;aAAM,CAAC;YACN,MAAM,EAAE,GAAG,MAAM,CAAC,UAAU,EAAE,CAAC;YAC/B,MAAM,GAAG,GAAG,MAAM,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;YAC/C,MAAM,OAAO,CAAC,YAAY,CAAC;gBACzB,GAAG,MAAM;gBACT,EAAE;gBACF,QAAQ;gBACR,SAAS,EAAE,GAAG;gBACd,cAAc,EAAE,SAAS,CAAC,KAAK;gBAC/B,YAAY,EAAE,SAAS,CAAC,UAAU;aACnC,CAAC,CAAC;YACH,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC;YAC1C,eAAe,EAAE,CAAC;QACpB,CAAC;IACH,CAAC;IAED,OAAO,EAAE,WAAW,EAAE,eAAe,EAAE,CAAC;AAC1C,CAAC;AAED;;;GAGG;AACH,+DAA+D;AAC/D,MAAM,iBAAiB,GAA2B;IAChD,YAAY,EAAE,OAAO;IACrB,gBAAgB,EAAE,YAAY;IAC9B,qBAAqB,EAAE,YAAY;IACnC,MAAM,EAAE,YAAY;IACpB,oBAAoB,EAAE,YAAY;IAClC,gBAAgB,EAAE,YAAY;IAC9B,aAAa,EAAE,YAAY;IAC3B,YAAY,EAAE,YAAY;IAC1B,UAAU,EAAE,YAAY;IACxB,SAAS,EAAE,UAAU;IACrB,WAAW,EAAE,UAAU;IACvB,WAAW,EAAE,UAAU;IACvB,UAAU,EAAE,UAAU;IACtB,UAAU,EAAE,YAAY;IACxB,YAAY,EAAE,WAAW;IACzB,eAAe,EAAE,OAAO;IACxB,WAAW,EAAE,OAAO;IACpB,aAAa,EAAE,SAAS;IACxB,YAAY,EAAE,SAAS;IACvB,KAAK,EAAE,SAAS;IAChB,QAAQ,EAAE,SAAS;IACnB,UAAU,EAAE,SAAS;IACrB,SAAS,EAAE,SAAS;IACpB,QAAQ,EAAE,SAAS;IACnB,WAAW,EAAE,SAAS;IACtB,OAAO,EAAE,SAAS;IAClB,UAAU,EAAE,SAAS;IACrB,OAAO,EAAE,SAAS;IAClB,MAAM,EAAE,SAAS;IACjB,OAAO,EAAE,SAAS;IAClB,QAAQ,EAAE,SAAS;CACpB,CAAC;AAEF,SAAS,iBAAiB,CAAC,QAAgB;IACzC,MAAM,KAAK,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;IAC5C,OAAO,iBAAiB,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC;AAC3C,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,OAAuB,EACvB,QAAgB,EAChB,MAAc,EACd,KAAsB,EACtB,WAAgC;IAEhC,+EAA+E;IAC/E,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,YAAY,GAAoB,EAAE,CAAC;IACzC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,kBAAkB,GAAG,iBAAiB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAC5D,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,UAAU,IAAI,kBAAkB,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;QAC1E,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACnB,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACd,YAAY,CAAC,IAAI,CAAC,EAAE,GAAG,IAAI,EAAE,QAAQ,EAAE,kBAAkB,EAAE,CAAC,CAAC;QAC/D,CAAC;IACH,CAAC;IAED,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;QAChC,MAAM,QAAQ,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;QAClD,MAAM,QAAQ,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;QAClD,IAAI,QAAQ,IAAI,QAAQ,EAAE,CAAC;YACzB,IAAI,CAAC;gBACH,MAAM,OAAO,CAAC,UAAU,CAAC;oBACvB,QAAQ;oBACR,QAAQ;oBACR,QAAQ;oBACR,QAAQ,EAAE,IAAI,CAAC,QAAQ;oBACvB,QAAQ,EAAE,IAAI,CAAC,QAAQ;oBACvB,UAAU,EAAE,IAAI,CAAC,UAAU;oBAC3B,MAAM,EAAE,GAAG;oBACX,QAAQ,EAAE,EAAE;oBACZ,MAAM;oBACN,EAAE,EAAE,MAAM,CAAC,UAAU,EAAE;iBACxB,CAAC,CAAC;gBACH,YAAY,EAAE,CAAC;YACjB,CAAC;YAAC,MAAM,CAAC;gBACP,oDAAoD;YACtD,CAAC;QACH,CAAC;QACD,IAAI,CAAC,QAAQ,IAAI,CAAC,QAAQ,EAAE,CAAC;YAC3B,OAAO,CAAC,IAAI,CACV,0BAA0B,IAAI,CAAC,UAAU,QAAQ,IAAI,CAAC,QAAQ,QAAQ,IAAI,CAAC,UAAU,IAAI;gBACzF,WAAW,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,YAAY,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,IAAI;gBACvF,oBAAoB,CAAC,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAC1D,CAAC;QACJ,CAAC;IACH,CAAC;IACD,OAAO,YAAY,CAAC;AACtB,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,CAAC,KAAK,UAAU,uBAAuB,CAC3C,OAAuB,EACvB,SAA2B,EAC3B,QAAgB,EAChB,MAAc,EACd,QAA2B,EAC3B,KAAsB;IAEtB,MAAM,EAAE,WAAW,EAAE,eAAe,EAAE,GAAG,MAAM,gBAAgB,CAC7D,OAAO,EACP,SAAS,EACT,QAAQ,EACR,QAAQ,CACT,CAAC;IAEF,iCAAiC;IACjC,KAAK,MAAM,CAAC,aAAa,EAAE,QAAQ,CAAC,IAAI,WAAW,EAAE,CAAC;QACpD,KAAK,aAAa,CAAC,CAAC,mBAAmB;QACvC,MAAM,OAAO,CAAC,cAAc,CAAC,MAAM,EAAE,QAAQ,EAAE,WAAW,CAAC,CAAC;IAC9D,CAAC;IAED,MAAM,YAAY,GAAG,MAAM,YAAY,CAAC,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,WAAW,CAAC,CAAC;IAEvF,OAAO,EAAE,eAAe,EAAE,YAAY,EAAE,WAAW,EAAE,CAAC;AACxD,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"hasher.d.ts","sourceRoot":"","sources":["../../src/extraction/hasher.ts"],"names":[],"mappings":"AAAA,wBAAsB,SAAS,CAAC,KAAK,EAAE;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,OAAO,CAAA;CAAE,GAAG,OAAO,CAAC,MAAM,CAAC,CAMvF"}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
export async function hashInput(input) {
|
|
2
|
+
const payload = JSON.stringify({ type: input.type, data: input.data });
|
|
3
|
+
const encoded = new TextEncoder().encode(payload);
|
|
4
|
+
const hashBuffer = await crypto.subtle.digest('SHA-256', encoded);
|
|
5
|
+
const hashArray = Array.from(new Uint8Array(hashBuffer));
|
|
6
|
+
return hashArray.map(b => b.toString(16).padStart(2, '0')).join('');
|
|
7
|
+
}
|
|
8
|
+
//# sourceMappingURL=hasher.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"hasher.js","sourceRoot":"","sources":["../../src/extraction/hasher.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,KAAsC;IACpE,MAAM,OAAO,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC;IACvE,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAClD,MAAM,UAAU,GAAG,MAAM,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IAClE,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC;IACzD,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACtE,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"heuristic.d.ts","sourceRoot":"","sources":["../../src/extraction/heuristic.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,gBAAgB,EAAkC,MAAM,YAAY,CAAC;AA8SnF,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,gBAAgB,CAqC/D"}
|
|
@@ -0,0 +1,282 @@
|
|
|
1
|
+
import nlp from 'compromise';
|
|
2
|
+
// ---------------------------------------------------------------------------
|
|
3
|
+
// Regex patterns
|
|
4
|
+
// ---------------------------------------------------------------------------
|
|
5
|
+
const REGEX = {
|
|
6
|
+
email: /\b[A-Za-z0-9._%+\-]+@[A-Za-z0-9.\-]+\.[A-Za-z]{2,}\b/g,
|
|
7
|
+
phone: /(?:\+?1[\s\-.]?)?\(?\d{3}\)?[\s\-.]?\d{3}[\s\-.]?\d{4}\b/g,
|
|
8
|
+
url: /https?:\/\/[^\s/$.?#].[^\s]*/g,
|
|
9
|
+
date: /\b(?:\d{1,2}[\/\-]\d{1,2}[\/\-]\d{2,4}|\d{4}[\/\-]\d{2}[\/\-]\d{2}|(?:Jan(?:uary)?|Feb(?:ruary)?|Mar(?:ch)?|Apr(?:il)?|May|Jun(?:e)?|Jul(?:y)?|Aug(?:ust)?|Sep(?:tember)?|Oct(?:ober)?|Nov(?:ember)?|Dec(?:ember)?)\s+\d{1,2}(?:st|nd|rd|th)?,?\s+\d{4}|\d{1,2}(?:st|nd|rd|th)?\s+(?:Jan(?:uary)?|Feb(?:ruary)?|Mar(?:ch)?|Apr(?:il)?|May|Jun(?:e)?|Jul(?:y)?|Aug(?:ust)?|Sep(?:tember)?|Oct(?:ober)?|Nov(?:ember)?|Dec(?:ember)?)\s+\d{4})\b/gi,
|
|
10
|
+
money: /\$\s?\d+(?:[.,]\d+)*(?:\.\d{2})?|\b\d+(?:[.,]\d+)*(?:\.\d{2})?\s?(?:dollars?|USD|EUR|GBP|euros?|pounds?)\b/gi,
|
|
11
|
+
};
|
|
12
|
+
const PATTERN_RULES = [
|
|
13
|
+
// Health / allergy — importance 0.95
|
|
14
|
+
{
|
|
15
|
+
pattern: /\bi(?:'m| am)\s+allergic\s+to\s+(.+?)(?:[.!?]|$)/i,
|
|
16
|
+
template: (m) => `User is allergic to ${m[1].trim()}`,
|
|
17
|
+
importance: 0.95,
|
|
18
|
+
tags: ['health', 'allergy'],
|
|
19
|
+
},
|
|
20
|
+
{
|
|
21
|
+
pattern: /\bi\s+have\s+(?:a\s+)?(?:allergy|allergies)\s+to\s+(.+?)(?:[.!?]|$)/i,
|
|
22
|
+
template: (m) => `User is allergic to ${m[1].trim()}`,
|
|
23
|
+
importance: 0.95,
|
|
24
|
+
tags: ['health', 'allergy'],
|
|
25
|
+
},
|
|
26
|
+
{
|
|
27
|
+
pattern: /\bi\s+(?:have|suffer from|was diagnosed with)\s+(.+?)(?:[.!?]|$)/i,
|
|
28
|
+
template: (m) => `User has ${m[1].trim()}`,
|
|
29
|
+
importance: 0.95,
|
|
30
|
+
tags: ['health'],
|
|
31
|
+
},
|
|
32
|
+
// Name — importance 0.9
|
|
33
|
+
{
|
|
34
|
+
pattern: /\bmy\s+name\s+is\s+([A-Za-z][a-zA-Z\s\-']{1,40}?)(?:[.!?,]|$)/i,
|
|
35
|
+
template: (m) => `User's name is ${m[1].trim()}`,
|
|
36
|
+
importance: 0.9,
|
|
37
|
+
tags: ['identity', 'name'],
|
|
38
|
+
},
|
|
39
|
+
{
|
|
40
|
+
pattern: /\bthey\s+call\s+me\s+([A-Za-z][a-zA-Z\s\-']{1,40}?)(?:[.!?,]|$)/i,
|
|
41
|
+
template: (m) => `User's name is ${m[1].trim()}`,
|
|
42
|
+
importance: 0.9,
|
|
43
|
+
tags: ['identity', 'name'],
|
|
44
|
+
},
|
|
45
|
+
// Identity — importance 0.85
|
|
46
|
+
{
|
|
47
|
+
pattern: /\bi(?:'m| am)\s+a(?:n)?\s+(.+?)(?:[.!?]|$)/i,
|
|
48
|
+
template: (m) => `User is a ${m[1].trim()}`,
|
|
49
|
+
importance: 0.85,
|
|
50
|
+
tags: ['identity'],
|
|
51
|
+
},
|
|
52
|
+
{
|
|
53
|
+
pattern: /\bi(?:'m| am)\s+([A-Z][a-z]+(?:\s+[A-Z][a-z]+)*)(?:[.!?,]|$)/,
|
|
54
|
+
template: (m) => `User is ${m[1].trim()}`,
|
|
55
|
+
importance: 0.85,
|
|
56
|
+
tags: ['identity'],
|
|
57
|
+
},
|
|
58
|
+
// Work / company — importance 0.8
|
|
59
|
+
{
|
|
60
|
+
pattern: /\bi\s+work\s+(?:at|for)\s+(.+?)(?:[.!?]|$)/i,
|
|
61
|
+
template: (m) => `User works at ${m[1].trim()}`,
|
|
62
|
+
importance: 0.8,
|
|
63
|
+
tags: ['work', 'company'],
|
|
64
|
+
},
|
|
65
|
+
{
|
|
66
|
+
pattern: /\bmy\s+(?:job|career|profession|occupation)\s+is\s+(.+?)(?:[.!?]|$)/i,
|
|
67
|
+
template: (m) => `User's job is ${m[1].trim()}`,
|
|
68
|
+
importance: 0.8,
|
|
69
|
+
tags: ['work'],
|
|
70
|
+
},
|
|
71
|
+
{
|
|
72
|
+
pattern: /\bi\s+(?:work|am employed)\s+as\s+(?:a(?:n)?\s+)?(.+?)(?:[.!?]|$)/i,
|
|
73
|
+
template: (m) => `User works as ${m[1].trim()}`,
|
|
74
|
+
importance: 0.8,
|
|
75
|
+
tags: ['work'],
|
|
76
|
+
},
|
|
77
|
+
// Location — importance 0.7
|
|
78
|
+
{
|
|
79
|
+
pattern: /\bi\s+live\s+in\s+(.+?)(?:[.!?]|$)/i,
|
|
80
|
+
template: (m) => `User lives in ${m[1].trim()}`,
|
|
81
|
+
importance: 0.7,
|
|
82
|
+
tags: ['location'],
|
|
83
|
+
},
|
|
84
|
+
{
|
|
85
|
+
pattern: /\bi(?:'m| am)\s+from\s+(.+?)(?:[.!?]|$)/i,
|
|
86
|
+
template: (m) => `User is from ${m[1].trim()}`,
|
|
87
|
+
importance: 0.7,
|
|
88
|
+
tags: ['location'],
|
|
89
|
+
},
|
|
90
|
+
{
|
|
91
|
+
pattern: /\bi\s+(?:moved|relocated)\s+to\s+(.+?)(?:[.!?]|$)/i,
|
|
92
|
+
template: (m) => `User moved to ${m[1].trim()}`,
|
|
93
|
+
importance: 0.7,
|
|
94
|
+
tags: ['location'],
|
|
95
|
+
},
|
|
96
|
+
// Preferences (like/love/enjoy) — importance 0.6
|
|
97
|
+
{
|
|
98
|
+
pattern: /\bi\s+(?:really\s+)?(?:like|love|enjoy|adore)\s+(.+?)(?:[.!?]|$)/i,
|
|
99
|
+
template: (m) => `User likes ${m[1].trim()}`,
|
|
100
|
+
importance: 0.6,
|
|
101
|
+
tags: ['preference', 'like'],
|
|
102
|
+
},
|
|
103
|
+
{
|
|
104
|
+
pattern: /\bmy\s+favorite\s+(?:\w+\s+)?is\s+(.+?)(?:[.!?]|$)/i,
|
|
105
|
+
template: (m) => `User's favorite is ${m[1].trim()}`,
|
|
106
|
+
importance: 0.6,
|
|
107
|
+
tags: ['preference', 'like'],
|
|
108
|
+
},
|
|
109
|
+
// Dislikes — importance 0.6
|
|
110
|
+
{
|
|
111
|
+
pattern: /\bi\s+(?:really\s+)?(?:hate|dislike|can'?t stand|despise|detest)\s+(.+?)(?:[.!?]|$)/i,
|
|
112
|
+
template: (m) => `User dislikes ${m[1].trim()}`,
|
|
113
|
+
importance: 0.6,
|
|
114
|
+
tags: ['preference', 'dislike'],
|
|
115
|
+
},
|
|
116
|
+
// Trivia / other — importance 0.3
|
|
117
|
+
{
|
|
118
|
+
pattern: /\bi\s+(?:think|believe|feel|guess|suppose)\s+(?:that\s+)?(.+?)(?:[.!?]|$)/i,
|
|
119
|
+
template: (m) => `User thinks ${m[1].trim()}`,
|
|
120
|
+
importance: 0.3,
|
|
121
|
+
tags: ['opinion'],
|
|
122
|
+
},
|
|
123
|
+
];
|
|
124
|
+
// ---------------------------------------------------------------------------
|
|
125
|
+
// Helpers
|
|
126
|
+
// ---------------------------------------------------------------------------
|
|
127
|
+
/** Strip a "role: " prefix like "user: " or "assistant: " from a line. */
|
|
128
|
+
function stripRolePrefix(line) {
|
|
129
|
+
return line.replace(/^[a-zA-Z_\-]+:\s*/, '');
|
|
130
|
+
}
|
|
131
|
+
function makeFact(content, originalContent, importance, confidence, tags, entityCanonicalNames = []) {
|
|
132
|
+
return {
|
|
133
|
+
content,
|
|
134
|
+
importance,
|
|
135
|
+
confidence,
|
|
136
|
+
sourceType: 'conversation',
|
|
137
|
+
modality: 'text',
|
|
138
|
+
tags,
|
|
139
|
+
originalContent,
|
|
140
|
+
operation: 'add',
|
|
141
|
+
entityCanonicalNames,
|
|
142
|
+
};
|
|
143
|
+
}
|
|
144
|
+
// ---------------------------------------------------------------------------
|
|
145
|
+
// Regex extraction (runs on the full text)
|
|
146
|
+
// ---------------------------------------------------------------------------
|
|
147
|
+
function extractRegex(text) {
|
|
148
|
+
const facts = [];
|
|
149
|
+
const seenContents = new Set();
|
|
150
|
+
function addFact(content, original, tags) {
|
|
151
|
+
if (!seenContents.has(content)) {
|
|
152
|
+
seenContents.add(content);
|
|
153
|
+
facts.push(makeFact(content, original, 0.8, 0.9, tags));
|
|
154
|
+
}
|
|
155
|
+
}
|
|
156
|
+
// Emails
|
|
157
|
+
for (const match of text.matchAll(REGEX.email)) {
|
|
158
|
+
const email = match[0];
|
|
159
|
+
addFact(`User's email is ${email}`, email, ['contact', 'email']);
|
|
160
|
+
}
|
|
161
|
+
// Phone numbers
|
|
162
|
+
for (const match of text.matchAll(REGEX.phone)) {
|
|
163
|
+
const phone = match[0];
|
|
164
|
+
addFact(`User's phone number is ${phone}`, phone, ['contact', 'phone']);
|
|
165
|
+
}
|
|
166
|
+
// URLs
|
|
167
|
+
for (const match of text.matchAll(REGEX.url)) {
|
|
168
|
+
const url = match[0];
|
|
169
|
+
addFact(`User mentioned URL: ${url}`, url, ['url']);
|
|
170
|
+
}
|
|
171
|
+
// Dates
|
|
172
|
+
for (const match of text.matchAll(REGEX.date)) {
|
|
173
|
+
const date = match[0];
|
|
174
|
+
addFact(`Mentioned date: ${date}`, date, ['date']);
|
|
175
|
+
}
|
|
176
|
+
// Money
|
|
177
|
+
for (const match of text.matchAll(REGEX.money)) {
|
|
178
|
+
const amount = match[0];
|
|
179
|
+
addFact(`Mentioned monetary amount: ${amount}`, amount, ['money']);
|
|
180
|
+
}
|
|
181
|
+
return facts;
|
|
182
|
+
}
|
|
183
|
+
// ---------------------------------------------------------------------------
|
|
184
|
+
// Pattern extraction (runs line by line, role prefix stripped)
|
|
185
|
+
// ---------------------------------------------------------------------------
|
|
186
|
+
function extractPatterns(lines) {
|
|
187
|
+
const facts = [];
|
|
188
|
+
const seenContents = new Set();
|
|
189
|
+
for (const rawLine of lines) {
|
|
190
|
+
const line = stripRolePrefix(rawLine).trim();
|
|
191
|
+
if (!line)
|
|
192
|
+
continue;
|
|
193
|
+
for (const rule of PATTERN_RULES) {
|
|
194
|
+
// Reset lastIndex for global regexes (these aren't global, but be safe)
|
|
195
|
+
const match = line.match(rule.pattern);
|
|
196
|
+
if (match) {
|
|
197
|
+
const content = rule.template(match);
|
|
198
|
+
if (!seenContents.has(content)) {
|
|
199
|
+
seenContents.add(content);
|
|
200
|
+
facts.push(makeFact(content, line, rule.importance, 0.7, rule.tags));
|
|
201
|
+
}
|
|
202
|
+
// Only use the first matching rule per line to avoid overlapping facts
|
|
203
|
+
break;
|
|
204
|
+
}
|
|
205
|
+
}
|
|
206
|
+
}
|
|
207
|
+
return facts;
|
|
208
|
+
}
|
|
209
|
+
// ---------------------------------------------------------------------------
|
|
210
|
+
// NER via compromise
|
|
211
|
+
// ---------------------------------------------------------------------------
|
|
212
|
+
function extractEntities(text) {
|
|
213
|
+
const doc = nlp(text);
|
|
214
|
+
const entities = [];
|
|
215
|
+
const seenCanonical = new Set();
|
|
216
|
+
function addEntity(name, entityType) {
|
|
217
|
+
const trimmed = name.trim();
|
|
218
|
+
if (!trimmed)
|
|
219
|
+
return;
|
|
220
|
+
const canonical = trimmed.toLowerCase();
|
|
221
|
+
if (!seenCanonical.has(`${entityType}:${canonical}`)) {
|
|
222
|
+
seenCanonical.add(`${entityType}:${canonical}`);
|
|
223
|
+
entities.push({
|
|
224
|
+
name: trimmed,
|
|
225
|
+
entityType,
|
|
226
|
+
canonicalName: canonical,
|
|
227
|
+
properties: {},
|
|
228
|
+
});
|
|
229
|
+
}
|
|
230
|
+
}
|
|
231
|
+
// People
|
|
232
|
+
const people = doc.people().out('array');
|
|
233
|
+
for (const name of people) {
|
|
234
|
+
addEntity(name, 'person');
|
|
235
|
+
}
|
|
236
|
+
// Organizations
|
|
237
|
+
const orgs = doc.organizations().out('array');
|
|
238
|
+
for (const org of orgs) {
|
|
239
|
+
addEntity(org, 'organization');
|
|
240
|
+
}
|
|
241
|
+
// Places
|
|
242
|
+
const places = doc.places().out('array');
|
|
243
|
+
for (const place of places) {
|
|
244
|
+
addEntity(place, 'location');
|
|
245
|
+
}
|
|
246
|
+
return entities;
|
|
247
|
+
}
|
|
248
|
+
// ---------------------------------------------------------------------------
|
|
249
|
+
// Main export
|
|
250
|
+
// ---------------------------------------------------------------------------
|
|
251
|
+
export function extractHeuristic(text) {
|
|
252
|
+
const lines = text.split('\n');
|
|
253
|
+
// 1. Regex extraction — run on full text
|
|
254
|
+
const regexFacts = extractRegex(text);
|
|
255
|
+
const regexContents = new Set(regexFacts.map((f) => f.content));
|
|
256
|
+
// 2. Pattern extraction — line by line, strip role prefix
|
|
257
|
+
const patternFacts = extractPatterns(lines);
|
|
258
|
+
// 3. De-duplicate: remove pattern facts whose content overlaps with regex facts
|
|
259
|
+
// (e.g., if regex grabbed the email and a pattern also mentions it)
|
|
260
|
+
const filteredPatternFacts = patternFacts.filter((f) => !regexContents.has(f.content));
|
|
261
|
+
// 4. NER entities via compromise
|
|
262
|
+
const entities = extractEntities(text);
|
|
263
|
+
const facts = [...regexFacts, ...filteredPatternFacts];
|
|
264
|
+
// Overall confidence: weighted average or fixed 0.9 for regex, 0.7 for pattern
|
|
265
|
+
// Use the maximum confidence across extracted facts, or 0.9 if only regex results exist
|
|
266
|
+
const confidence = facts.length > 0
|
|
267
|
+
? Math.max(...facts.map((f) => f.confidence))
|
|
268
|
+
: entities.length > 0
|
|
269
|
+
? 0.6
|
|
270
|
+
: 0.5;
|
|
271
|
+
return {
|
|
272
|
+
facts,
|
|
273
|
+
entities,
|
|
274
|
+
edges: [],
|
|
275
|
+
tier: 'heuristic',
|
|
276
|
+
confidence,
|
|
277
|
+
tokensInput: 0,
|
|
278
|
+
tokensOutput: 0,
|
|
279
|
+
model: null,
|
|
280
|
+
};
|
|
281
|
+
}
|
|
282
|
+
//# sourceMappingURL=heuristic.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"heuristic.js","sourceRoot":"","sources":["../../src/extraction/heuristic.ts"],"names":[],"mappings":"AAAA,OAAO,GAAG,MAAM,YAAY,CAAC;AAG7B,8EAA8E;AAC9E,iBAAiB;AACjB,8EAA8E;AAE9E,MAAM,KAAK,GAAG;IACZ,KAAK,EAAE,uDAAuD;IAC9D,KAAK,EAAE,2DAA2D;IAClE,GAAG,EAAE,+BAA+B;IACpC,IAAI,EAAE,ibAAib;IACvb,KAAK,EAAE,8GAA8G;CACtH,CAAC;AAaF,MAAM,aAAa,GAAkB;IACnC,qCAAqC;IACrC;QACE,OAAO,EAAE,mDAAmD;QAC5D,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,uBAAuB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QACtD,UAAU,EAAE,IAAI;QAChB,IAAI,EAAE,CAAC,QAAQ,EAAE,SAAS,CAAC;KAC5B;IACD;QACE,OAAO,EAAE,sEAAsE;QAC/E,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,uBAAuB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QACtD,UAAU,EAAE,IAAI;QAChB,IAAI,EAAE,CAAC,QAAQ,EAAE,SAAS,CAAC;KAC5B;IACD;QACE,OAAO,EAAE,mEAAmE;QAC5E,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAC3C,UAAU,EAAE,IAAI;QAChB,IAAI,EAAE,CAAC,QAAQ,CAAC;KACjB;IAED,wBAAwB;IACxB;QACE,OAAO,EAAE,gEAAgE;QACzE,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QACjD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,UAAU,EAAE,MAAM,CAAC;KAC3B;IACD;QACE,OAAO,EAAE,kEAAkE;QAC3E,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QACjD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,UAAU,EAAE,MAAM,CAAC;KAC3B;IAED,6BAA6B;IAC7B;QACE,OAAO,EAAE,6CAA6C;QACtD,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,aAAa,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAC5C,UAAU,EAAE,IAAI;QAChB,IAAI,EAAE,CAAC,UAAU,CAAC;KACnB;IACD;QACE,OAAO,EAAE,8DAA8D;QACvE,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAC1C,UAAU,EAAE,IAAI;QAChB,IAAI,EAAE,CAAC,UAAU,CAAC;KACnB;IAED,kCAAkC;IAClC;QACE,OAAO,EAAE,6CAA6C;QACtD,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAChD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;KAC1B;IACD;QACE,OAAO,EAAE,sEAAsE;QAC/E,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAChD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,MAAM,CAAC;KACf;IACD;QACE,OAAO,EAAE,oEAAoE;QAC7E,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAChD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,MAAM,CAAC;KACf;IAED,4BAA4B;IAC5B;QACE,OAAO,EAAE,qCAAqC;QAC9C,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAChD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,UAAU,CAAC;KACnB;IACD;QACE,OAAO,EAAE,0CAA0C;QACnD,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,gBAAgB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAC/C,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,UAAU,CAAC;KACnB;IACD;QACE,OAAO,EAAE,oDAAoD;QAC7D,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAChD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,UAAU,CAAC;KACnB;IAED,iDAAiD;IACjD;QACE,OAAO,EAAE,mEAAmE;QAC5E,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAC7C,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,YAAY,EAAE,MAAM,CAAC;KAC7B;IACD;QACE,OAAO,EAAE,qDAAqD;QAC9D,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QACrD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,YAAY,EAAE,MAAM,CAAC;KAC7B;IAED,4BAA4B;IAC5B;QACE,OAAO,EAAE,sFAAsF;QAC/F,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAChD,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,YAAY,EAAE,SAAS,CAAC;KAChC;IAED,kCAAkC;IAClC;QACE,OAAO,EAAE,4EAA4E;QACrF,QAAQ,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE;QAC9C,UAAU,EAAE,GAAG;QACf,IAAI,EAAE,CAAC,SAAS,CAAC;KAClB;CACF,CAAC;AAEF,8EAA8E;AAC9E,UAAU;AACV,8EAA8E;AAE9E,0EAA0E;AAC1E,SAAS,eAAe,CAAC,IAAY;IACnC,OAAO,IAAI,CAAC,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC;AAC/C,CAAC;AAED,SAAS,QAAQ,CACf,OAAe,EACf,eAAuB,EACvB,UAAkB,EAClB,UAAkB,EAClB,IAAc,EACd,uBAAiC,EAAE;IAEnC,OAAO;QACL,OAAO;QACP,UAAU;QACV,UAAU;QACV,UAAU,EAAE,cAAc;QAC1B,QAAQ,EAAE,MAAM;QAChB,IAAI;QACJ,eAAe;QACf,SAAS,EAAE,KAAK;QAChB,oBAAoB;KACrB,CAAC;AACJ,CAAC;AAED,8EAA8E;AAC9E,2CAA2C;AAC3C,8EAA8E;AAE9E,SAAS,YAAY,CAAC,IAAY;IAChC,MAAM,KAAK,GAAoB,EAAE,CAAC;IAClC,MAAM,YAAY,GAAG,IAAI,GAAG,EAAU,CAAC;IAEvC,SAAS,OAAO,CAAC,OAAe,EAAE,QAAgB,EAAE,IAAc;QAChE,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YAC/B,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;YAC1B,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,QAAQ,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,CAAC,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IAED,SAAS;IACT,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;QAC/C,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QACxB,OAAO,CAAC,mBAAmB,KAAK,EAAE,EAAE,KAAK,EAAE,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC,CAAC;IACnE,CAAC;IAED,gBAAgB;IAChB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;QAC/C,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QACxB,OAAO,CAAC,0BAA0B,KAAK,EAAE,EAAE,KAAK,EAAE,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC,CAAC;IAC1E,CAAC;IAED,OAAO;IACP,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QAC7C,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QACtB,OAAO,CAAC,uBAAuB,GAAG,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC;IACtD,CAAC;IAED,QAAQ;IACR,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QAC9C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QACvB,OAAO,CAAC,mBAAmB,IAAI,EAAE,EAAE,IAAI,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC;IACrD,CAAC;IAED,QAAQ;IACR,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;QAC/C,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;QACzB,OAAO,CAAC,8BAA8B,MAAM,EAAE,EAAE,MAAM,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;IACrE,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,8EAA8E;AAC9E,+DAA+D;AAC/D,8EAA8E;AAE9E,SAAS,eAAe,CAAC,KAAe;IACtC,MAAM,KAAK,GAAoB,EAAE,CAAC;IAClC,MAAM,YAAY,GAAG,IAAI,GAAG,EAAU,CAAC;IAEvC,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,eAAe,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;QAC7C,IAAI,CAAC,IAAI;YAAE,SAAS;QAEpB,KAAK,MAAM,IAAI,IAAI,aAAa,EAAE,CAAC;YACjC,wEAAwE;YACxE,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACvC,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,OAAO,GAAG,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;gBACrC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;oBAC/B,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;oBAC1B,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,UAAU,EAAE,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;gBACvE,CAAC;gBACD,uEAAuE;gBACvE,MAAM;YACR,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,8EAA8E;AAC9E,qBAAqB;AACrB,8EAA8E;AAE9E,SAAS,eAAe,CAAC,IAAY;IACnC,MAAM,GAAG,GAAG,GAAG,CAAC,IAAI,CAAC,CAAC;IACtB,MAAM,QAAQ,GAAsB,EAAE,CAAC;IACvC,MAAM,aAAa,GAAG,IAAI,GAAG,EAAU,CAAC;IAExC,SAAS,SAAS,CAAC,IAAY,EAAE,UAAkB;QACjD,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,OAAO;YAAE,OAAO;QACrB,MAAM,SAAS,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;QACxC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,GAAG,UAAU,IAAI,SAAS,EAAE,CAAC,EAAE,CAAC;YACrD,aAAa,CAAC,GAAG,CAAC,GAAG,UAAU,IAAI,SAAS,EAAE,CAAC,CAAC;YAChD,QAAQ,CAAC,IAAI,CAAC;gBACZ,IAAI,EAAE,OAAO;gBACb,UAAU;gBACV,aAAa,EAAE,SAAS;gBACxB,UAAU,EAAE,EAAE;aACf,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,SAAS;IACT,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,OAAO,CAAa,CAAC;IACrD,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;QAC1B,SAAS,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IAC5B,CAAC;IAED,gBAAgB;IAChB,MAAM,IAAI,GAAG,GAAG,CAAC,aAAa,EAAE,CAAC,GAAG,CAAC,OAAO,CAAa,CAAC;IAC1D,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,SAAS,CAAC,GAAG,EAAE,cAAc,CAAC,CAAC;IACjC,CAAC;IAED,SAAS;IACT,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,OAAO,CAAa,CAAC;IACrD,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,SAAS,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC;IAC/B,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,8EAA8E;AAC9E,cAAc;AACd,8EAA8E;AAE9E,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAE/B,yCAAyC;IACzC,MAAM,UAAU,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IACtC,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;IAEhE,0DAA0D;IAC1D,MAAM,YAAY,GAAG,eAAe,CAAC,KAAK,CAAC,CAAC;IAE5C,gFAAgF;IAChF,uEAAuE;IACvE,MAAM,oBAAoB,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;IAEvF,iCAAiC;IACjC,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,CAAC,CAAC;IAEvC,MAAM,KAAK,GAAG,CAAC,GAAG,UAAU,EAAE,GAAG,oBAAoB,CAAC,CAAC;IAEvD,+EAA+E;IAC/E,wFAAwF;IACxF,MAAM,UAAU,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC;QACjC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;QAC7C,CAAC,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC;YACnB,CAAC,CAAC,GAAG;YACL,CAAC,CAAC,GAAG,CAAC;IAEV,OAAO;QACL,KAAK;QACL,QAAQ;QACR,KAAK,EAAE,EAAE;QACT,IAAI,EAAE,WAAW;QACjB,UAAU;QACV,WAAW,EAAE,CAAC;QACd,YAAY,EAAE,CAAC;QACf,KAAK,EAAE,IAAI;KACZ,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
export * from './types.js';
|
|
2
|
+
export * from './hasher.js';
|
|
3
|
+
export * from './prompts.js';
|
|
4
|
+
export * from './heuristic.js';
|
|
5
|
+
export * from './contradiction.js';
|
|
6
|
+
export * from './entity-extractor.js';
|
|
7
|
+
export * from './llm-extractor.js';
|
|
8
|
+
export * from './dedup.js';
|
|
9
|
+
export * from './pipeline.js';
|
|
10
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/extraction/index.ts"],"names":[],"mappings":"AAAA,cAAc,YAAY,CAAC;AAC3B,cAAc,aAAa,CAAC;AAC5B,cAAc,cAAc,CAAC;AAC7B,cAAc,gBAAgB,CAAC;AAC/B,cAAc,oBAAoB,CAAC;AACnC,cAAc,uBAAuB,CAAC;AACtC,cAAc,oBAAoB,CAAC;AACnC,cAAc,YAAY,CAAC;AAC3B,cAAc,eAAe,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
export * from './types.js';
|
|
2
|
+
export * from './hasher.js';
|
|
3
|
+
export * from './prompts.js';
|
|
4
|
+
export * from './heuristic.js';
|
|
5
|
+
export * from './contradiction.js';
|
|
6
|
+
export * from './entity-extractor.js';
|
|
7
|
+
export * from './llm-extractor.js';
|
|
8
|
+
export * from './dedup.js';
|
|
9
|
+
export * from './pipeline.js';
|
|
10
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/extraction/index.ts"],"names":[],"mappings":"AAAA,cAAc,YAAY,CAAC;AAC3B,cAAc,aAAa,CAAC;AAC5B,cAAc,cAAc,CAAC;AAC7B,cAAc,gBAAgB,CAAC;AAC/B,cAAc,oBAAoB,CAAC;AACnC,cAAc,uBAAuB,CAAC;AACtC,cAAc,oBAAoB,CAAC;AACnC,cAAc,YAAY,CAAC;AAC3B,cAAc,eAAe,CAAC"}
|
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
import type { LLMAdapter } from '../adapters/llm.js';
|
|
2
|
+
import type { ExtractionResult } from './types.js';
|
|
3
|
+
import type { ExtractionTier, DomainEntityType } from '../config.js';
|
|
4
|
+
export interface LLMExtractorConfig {
|
|
5
|
+
llm: LLMAdapter;
|
|
6
|
+
tier: ExtractionTier;
|
|
7
|
+
entityTypes?: string[];
|
|
8
|
+
domainEntityTypes?: DomainEntityType[];
|
|
9
|
+
}
|
|
10
|
+
/**
|
|
11
|
+
* Two-pass extraction like Mem0:
|
|
12
|
+
* Pass 1: Extract facts as simple strings (focused, high quality)
|
|
13
|
+
* Pass 2: Extract entities + edges from the facts (separate concern)
|
|
14
|
+
*/
|
|
15
|
+
export declare function extractWithLLM(config: LLMExtractorConfig, input: string, existingFacts?: Array<{
|
|
16
|
+
lineageId: string;
|
|
17
|
+
content: string;
|
|
18
|
+
}>): Promise<ExtractionResult>;
|
|
19
|
+
/**
|
|
20
|
+
* Normalize an entity name to a clean canonical form.
|
|
21
|
+
*/
|
|
22
|
+
export declare function normalizeEntityName(raw: string): string;
|
|
23
|
+
//# sourceMappingURL=llm-extractor.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"llm-extractor.d.ts","sourceRoot":"","sources":["../../src/extraction/llm-extractor.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AACrD,OAAO,KAAK,EAAE,gBAAgB,EAAiD,MAAM,YAAY,CAAC;AAClG,OAAO,KAAK,EAAE,cAAc,EAAY,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAI/E,MAAM,WAAW,kBAAkB;IACjC,GAAG,EAAE,UAAU,CAAC;IAChB,IAAI,EAAE,cAAc,CAAC;IACrB,WAAW,CAAC,EAAE,MAAM,EAAE,CAAC;IACvB,iBAAiB,CAAC,EAAE,gBAAgB,EAAE,CAAC;CACxC;AAED;;;;GAIG;AACH,wBAAsB,cAAc,CAClC,MAAM,EAAE,kBAAkB,EAC1B,KAAK,EAAE,MAAM,EACb,aAAa,CAAC,EAAE,KAAK,CAAC;IAAE,SAAS,EAAE,MAAM,CAAC;IAAC,OAAO,EAAE,MAAM,CAAA;CAAE,CAAC,GAC5D,OAAO,CAAC,gBAAgB,CAAC,CA8M3B;AAsBD;;GAEG;AACH,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAavD"}
|