@ninebix/nmt-system 1.0.0
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 +70 -0
- package/README.md +465 -0
- package/dist/api/cli-server.d.ts +83 -0
- package/dist/api/cli-server.d.ts.map +1 -0
- package/dist/api/cli-server.js +597 -0
- package/dist/api/cli-server.js.map +1 -0
- package/dist/api/index.d.ts +6 -0
- package/dist/api/index.d.ts.map +1 -0
- package/dist/api/index.js +6 -0
- package/dist/api/index.js.map +1 -0
- package/dist/api/middleware/index.d.ts +12 -0
- package/dist/api/middleware/index.d.ts.map +1 -0
- package/dist/api/middleware/index.js +13 -0
- package/dist/api/middleware/index.js.map +1 -0
- package/dist/api/middleware/logger.d.ts +21 -0
- package/dist/api/middleware/logger.d.ts.map +1 -0
- package/dist/api/middleware/logger.js +134 -0
- package/dist/api/middleware/logger.js.map +1 -0
- package/dist/api/middleware/rate-limit.d.ts +26 -0
- package/dist/api/middleware/rate-limit.d.ts.map +1 -0
- package/dist/api/middleware/rate-limit.js +107 -0
- package/dist/api/middleware/rate-limit.js.map +1 -0
- package/dist/api/middleware/response.d.ts +61 -0
- package/dist/api/middleware/response.d.ts.map +1 -0
- package/dist/api/middleware/response.js +86 -0
- package/dist/api/middleware/response.js.map +1 -0
- package/dist/api/middleware/validation.d.ts +43 -0
- package/dist/api/middleware/validation.d.ts.map +1 -0
- package/dist/api/middleware/validation.js +257 -0
- package/dist/api/middleware/validation.js.map +1 -0
- package/dist/api/server.d.ts +79 -0
- package/dist/api/server.d.ts.map +1 -0
- package/dist/api/server.js +2011 -0
- package/dist/api/server.js.map +1 -0
- package/dist/bin/nmt.d.ts +9 -0
- package/dist/bin/nmt.d.ts.map +1 -0
- package/dist/bin/nmt.js +1142 -0
- package/dist/bin/nmt.js.map +1 -0
- package/dist/cli/commands/attractor.d.ts +6 -0
- package/dist/cli/commands/attractor.d.ts.map +1 -0
- package/dist/cli/commands/attractor.js +167 -0
- package/dist/cli/commands/attractor.js.map +1 -0
- package/dist/cli/commands/dimension.d.ts +6 -0
- package/dist/cli/commands/dimension.d.ts.map +1 -0
- package/dist/cli/commands/dimension.js +85 -0
- package/dist/cli/commands/dimension.js.map +1 -0
- package/dist/cli/commands/index.d.ts +11 -0
- package/dist/cli/commands/index.d.ts.map +1 -0
- package/dist/cli/commands/index.js +11 -0
- package/dist/cli/commands/index.js.map +1 -0
- package/dist/cli/commands/infer.d.ts +6 -0
- package/dist/cli/commands/infer.d.ts.map +1 -0
- package/dist/cli/commands/infer.js +139 -0
- package/dist/cli/commands/infer.js.map +1 -0
- package/dist/cli/commands/learn.d.ts +6 -0
- package/dist/cli/commands/learn.d.ts.map +1 -0
- package/dist/cli/commands/learn.js +87 -0
- package/dist/cli/commands/learn.js.map +1 -0
- package/dist/cli/commands/orchestrate.d.ts +6 -0
- package/dist/cli/commands/orchestrate.d.ts.map +1 -0
- package/dist/cli/commands/orchestrate.js +279 -0
- package/dist/cli/commands/orchestrate.js.map +1 -0
- package/dist/cli/commands/prob.d.ts +6 -0
- package/dist/cli/commands/prob.d.ts.map +1 -0
- package/dist/cli/commands/prob.js +256 -0
- package/dist/cli/commands/prob.js.map +1 -0
- package/dist/cli/commands/quantum.d.ts +6 -0
- package/dist/cli/commands/quantum.d.ts.map +1 -0
- package/dist/cli/commands/quantum.js +150 -0
- package/dist/cli/commands/quantum.js.map +1 -0
- package/dist/cli/commands/sync.d.ts +65 -0
- package/dist/cli/commands/sync.d.ts.map +1 -0
- package/dist/cli/commands/sync.js +338 -0
- package/dist/cli/commands/sync.js.map +1 -0
- package/dist/cli/index.d.ts +9 -0
- package/dist/cli/index.d.ts.map +1 -0
- package/dist/cli/index.js +9 -0
- package/dist/cli/index.js.map +1 -0
- package/dist/cli/probabilistic-commands.d.ts +39 -0
- package/dist/cli/probabilistic-commands.d.ts.map +1 -0
- package/dist/cli/probabilistic-commands.js +112 -0
- package/dist/cli/probabilistic-commands.js.map +1 -0
- package/dist/cli/types.d.ts +69 -0
- package/dist/cli/types.d.ts.map +1 -0
- package/dist/cli/types.js +5 -0
- package/dist/cli/types.js.map +1 -0
- package/dist/cli/utils/formatters.d.ts +51 -0
- package/dist/cli/utils/formatters.d.ts.map +1 -0
- package/dist/cli/utils/formatters.js +79 -0
- package/dist/cli/utils/formatters.js.map +1 -0
- package/dist/cli/utils/helpers.d.ts +21 -0
- package/dist/cli/utils/helpers.d.ts.map +1 -0
- package/dist/cli/utils/helpers.js +51 -0
- package/dist/cli/utils/helpers.js.map +1 -0
- package/dist/cli/utils/index.d.ts +7 -0
- package/dist/cli/utils/index.d.ts.map +1 -0
- package/dist/cli/utils/index.js +13 -0
- package/dist/cli/utils/index.js.map +1 -0
- package/dist/cli/utils/validators.d.ts +162 -0
- package/dist/cli/utils/validators.d.ts.map +1 -0
- package/dist/cli/utils/validators.js +351 -0
- package/dist/cli/utils/validators.js.map +1 -0
- package/dist/core/advanced-embedding.d.ts +154 -0
- package/dist/core/advanced-embedding.d.ts.map +1 -0
- package/dist/core/advanced-embedding.js +367 -0
- package/dist/core/advanced-embedding.js.map +1 -0
- package/dist/core/attractor-model.d.ts +381 -0
- package/dist/core/attractor-model.d.ts.map +1 -0
- package/dist/core/attractor-model.js +821 -0
- package/dist/core/attractor-model.js.map +1 -0
- package/dist/core/bidirectional-inference.d.ts +143 -0
- package/dist/core/bidirectional-inference.d.ts.map +1 -0
- package/dist/core/bidirectional-inference.js +501 -0
- package/dist/core/bidirectional-inference.js.map +1 -0
- package/dist/core/chunk-engine.d.ts +78 -0
- package/dist/core/chunk-engine.d.ts.map +1 -0
- package/dist/core/chunk-engine.js +192 -0
- package/dist/core/chunk-engine.js.map +1 -0
- package/dist/core/dynamic-embedding.d.ts +327 -0
- package/dist/core/dynamic-embedding.d.ts.map +1 -0
- package/dist/core/dynamic-embedding.js +527 -0
- package/dist/core/dynamic-embedding.js.map +1 -0
- package/dist/core/embedding-similarity.d.ts +68 -0
- package/dist/core/embedding-similarity.d.ts.map +1 -0
- package/dist/core/embedding-similarity.js +291 -0
- package/dist/core/embedding-similarity.js.map +1 -0
- package/dist/core/evolution-scheduler.d.ts +101 -0
- package/dist/core/evolution-scheduler.d.ts.map +1 -0
- package/dist/core/evolution-scheduler.js +235 -0
- package/dist/core/evolution-scheduler.js.map +1 -0
- package/dist/core/hierarchical-chunker.d.ts +108 -0
- package/dist/core/hierarchical-chunker.d.ts.map +1 -0
- package/dist/core/hierarchical-chunker.js +296 -0
- package/dist/core/hierarchical-chunker.js.map +1 -0
- package/dist/core/hnsw-index.d.ts +111 -0
- package/dist/core/hnsw-index.d.ts.map +1 -0
- package/dist/core/hnsw-index.js +466 -0
- package/dist/core/hnsw-index.js.map +1 -0
- package/dist/core/index.d.ts +23 -0
- package/dist/core/index.d.ts.map +1 -0
- package/dist/core/index.js +25 -0
- package/dist/core/index.js.map +1 -0
- package/dist/core/language-analyzers.d.ts +124 -0
- package/dist/core/language-analyzers.d.ts.map +1 -0
- package/dist/core/language-analyzers.js +365 -0
- package/dist/core/language-analyzers.js.map +1 -0
- package/dist/core/local-embedding.d.ts +109 -0
- package/dist/core/local-embedding.d.ts.map +1 -0
- package/dist/core/local-embedding.js +222 -0
- package/dist/core/local-embedding.js.map +1 -0
- package/dist/core/merkle-engine.d.ts +263 -0
- package/dist/core/merkle-engine.d.ts.map +1 -0
- package/dist/core/merkle-engine.js +528 -0
- package/dist/core/merkle-engine.js.map +1 -0
- package/dist/core/multi-layer-reasoning.d.ts +178 -0
- package/dist/core/multi-layer-reasoning.d.ts.map +1 -0
- package/dist/core/multi-layer-reasoning.js +607 -0
- package/dist/core/multi-layer-reasoning.js.map +1 -0
- package/dist/core/neuron-graph.d.ts +134 -0
- package/dist/core/neuron-graph.d.ts.map +1 -0
- package/dist/core/neuron-graph.js +436 -0
- package/dist/core/neuron-graph.js.map +1 -0
- package/dist/core/probabilistic-neuron.d.ts +251 -0
- package/dist/core/probabilistic-neuron.d.ts.map +1 -0
- package/dist/core/probabilistic-neuron.js +618 -0
- package/dist/core/probabilistic-neuron.js.map +1 -0
- package/dist/core/probabilistic-orchestrator.d.ts +408 -0
- package/dist/core/probabilistic-orchestrator.d.ts.map +1 -0
- package/dist/core/probabilistic-orchestrator.js +798 -0
- package/dist/core/probabilistic-orchestrator.js.map +1 -0
- package/dist/core/semantic-chunker.d.ts +117 -0
- package/dist/core/semantic-chunker.d.ts.map +1 -0
- package/dist/core/semantic-chunker.js +464 -0
- package/dist/core/semantic-chunker.js.map +1 -0
- package/dist/events/event-bus.d.ts +166 -0
- package/dist/events/event-bus.d.ts.map +1 -0
- package/dist/events/event-bus.js +228 -0
- package/dist/events/event-bus.js.map +1 -0
- package/dist/events/index.d.ts +7 -0
- package/dist/events/index.d.ts.map +1 -0
- package/dist/events/index.js +7 -0
- package/dist/events/index.js.map +1 -0
- package/dist/events/progress-tracker.d.ts +150 -0
- package/dist/events/progress-tracker.d.ts.map +1 -0
- package/dist/events/progress-tracker.js +290 -0
- package/dist/events/progress-tracker.js.map +1 -0
- package/dist/extensions/clustering/community-detection.d.ts +90 -0
- package/dist/extensions/clustering/community-detection.d.ts.map +1 -0
- package/dist/extensions/clustering/community-detection.js +470 -0
- package/dist/extensions/clustering/community-detection.js.map +1 -0
- package/dist/extensions/clustering/index.d.ts +114 -0
- package/dist/extensions/clustering/index.d.ts.map +1 -0
- package/dist/extensions/clustering/index.js +468 -0
- package/dist/extensions/clustering/index.js.map +1 -0
- package/dist/extensions/clustering/topic-modeling.d.ts +86 -0
- package/dist/extensions/clustering/topic-modeling.d.ts.map +1 -0
- package/dist/extensions/clustering/topic-modeling.js +355 -0
- package/dist/extensions/clustering/topic-modeling.js.map +1 -0
- package/dist/extensions/distributed/coordinator.d.ts +114 -0
- package/dist/extensions/distributed/coordinator.d.ts.map +1 -0
- package/dist/extensions/distributed/coordinator.js +319 -0
- package/dist/extensions/distributed/coordinator.js.map +1 -0
- package/dist/extensions/distributed/index.d.ts +10 -0
- package/dist/extensions/distributed/index.d.ts.map +1 -0
- package/dist/extensions/distributed/index.js +10 -0
- package/dist/extensions/distributed/index.js.map +1 -0
- package/dist/extensions/distributed/queue.d.ts +157 -0
- package/dist/extensions/distributed/queue.d.ts.map +1 -0
- package/dist/extensions/distributed/queue.js +326 -0
- package/dist/extensions/distributed/queue.js.map +1 -0
- package/dist/extensions/distributed/scheduler.d.ts +107 -0
- package/dist/extensions/distributed/scheduler.d.ts.map +1 -0
- package/dist/extensions/distributed/scheduler.js +301 -0
- package/dist/extensions/distributed/scheduler.js.map +1 -0
- package/dist/extensions/distributed/worker.d.ts +112 -0
- package/dist/extensions/distributed/worker.d.ts.map +1 -0
- package/dist/extensions/distributed/worker.js +260 -0
- package/dist/extensions/distributed/worker.js.map +1 -0
- package/dist/index.d.ts +14 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +20 -0
- package/dist/index.js.map +1 -0
- package/dist/mcp/server.d.ts +43 -0
- package/dist/mcp/server.d.ts.map +1 -0
- package/dist/mcp/server.js +494 -0
- package/dist/mcp/server.js.map +1 -0
- package/dist/services/adaptive-fallback.d.ts +140 -0
- package/dist/services/adaptive-fallback.d.ts.map +1 -0
- package/dist/services/adaptive-fallback.js +273 -0
- package/dist/services/adaptive-fallback.js.map +1 -0
- package/dist/services/answer-gate.d.ts +112 -0
- package/dist/services/answer-gate.d.ts.map +1 -0
- package/dist/services/answer-gate.js +299 -0
- package/dist/services/answer-gate.js.map +1 -0
- package/dist/services/auto-learning.d.ts +135 -0
- package/dist/services/auto-learning.d.ts.map +1 -0
- package/dist/services/auto-learning.js +413 -0
- package/dist/services/auto-learning.js.map +1 -0
- package/dist/services/context-compressor.d.ts +77 -0
- package/dist/services/context-compressor.d.ts.map +1 -0
- package/dist/services/context-compressor.js +234 -0
- package/dist/services/context-compressor.js.map +1 -0
- package/dist/services/efficient-rag.d.ts +140 -0
- package/dist/services/efficient-rag.d.ts.map +1 -0
- package/dist/services/efficient-rag.js +311 -0
- package/dist/services/efficient-rag.js.map +1 -0
- package/dist/services/embedding-provider.d.ts +72 -0
- package/dist/services/embedding-provider.d.ts.map +1 -0
- package/dist/services/embedding-provider.js +176 -0
- package/dist/services/embedding-provider.js.map +1 -0
- package/dist/services/file-ingestion.d.ts +72 -0
- package/dist/services/file-ingestion.d.ts.map +1 -0
- package/dist/services/file-ingestion.js +237 -0
- package/dist/services/file-ingestion.js.map +1 -0
- package/dist/services/four-stage-learning.d.ts +552 -0
- package/dist/services/four-stage-learning.d.ts.map +1 -0
- package/dist/services/four-stage-learning.js +1110 -0
- package/dist/services/four-stage-learning.js.map +1 -0
- package/dist/services/graph.d.ts +94 -0
- package/dist/services/graph.d.ts.map +1 -0
- package/dist/services/graph.js +292 -0
- package/dist/services/graph.js.map +1 -0
- package/dist/services/index.d.ts +15 -0
- package/dist/services/index.d.ts.map +1 -0
- package/dist/services/index.js +18 -0
- package/dist/services/index.js.map +1 -0
- package/dist/services/ingestion.d.ts +98 -0
- package/dist/services/ingestion.d.ts.map +1 -0
- package/dist/services/ingestion.js +259 -0
- package/dist/services/ingestion.js.map +1 -0
- package/dist/services/learning.d.ts +67 -0
- package/dist/services/learning.d.ts.map +1 -0
- package/dist/services/learning.js +262 -0
- package/dist/services/learning.js.map +1 -0
- package/dist/services/llm-router.d.ts +143 -0
- package/dist/services/llm-router.d.ts.map +1 -0
- package/dist/services/llm-router.js +284 -0
- package/dist/services/llm-router.js.map +1 -0
- package/dist/services/llm.d.ts +86 -0
- package/dist/services/llm.d.ts.map +1 -0
- package/dist/services/llm.js +283 -0
- package/dist/services/llm.js.map +1 -0
- package/dist/services/metrics-dashboard.d.ts +262 -0
- package/dist/services/metrics-dashboard.d.ts.map +1 -0
- package/dist/services/metrics-dashboard.js +417 -0
- package/dist/services/metrics-dashboard.js.map +1 -0
- package/dist/services/neuron-lifecycle.d.ts +137 -0
- package/dist/services/neuron-lifecycle.d.ts.map +1 -0
- package/dist/services/neuron-lifecycle.js +422 -0
- package/dist/services/neuron-lifecycle.js.map +1 -0
- package/dist/services/nmt-pipeline.d.ts +219 -0
- package/dist/services/nmt-pipeline.d.ts.map +1 -0
- package/dist/services/nmt-pipeline.js +449 -0
- package/dist/services/nmt-pipeline.js.map +1 -0
- package/dist/services/query-cache.d.ts +136 -0
- package/dist/services/query-cache.d.ts.map +1 -0
- package/dist/services/query-cache.js +255 -0
- package/dist/services/query-cache.js.map +1 -0
- package/dist/services/query-normalize.d.ts +107 -0
- package/dist/services/query-normalize.d.ts.map +1 -0
- package/dist/services/query-normalize.js +366 -0
- package/dist/services/query-normalize.js.map +1 -0
- package/dist/services/query.d.ts +102 -0
- package/dist/services/query.d.ts.map +1 -0
- package/dist/services/query.js +227 -0
- package/dist/services/query.js.map +1 -0
- package/dist/services/text-embedding.d.ts +183 -0
- package/dist/services/text-embedding.d.ts.map +1 -0
- package/dist/services/text-embedding.js +633 -0
- package/dist/services/text-embedding.js.map +1 -0
- package/dist/services/verification-gate.d.ts +147 -0
- package/dist/services/verification-gate.d.ts.map +1 -0
- package/dist/services/verification-gate.js +344 -0
- package/dist/services/verification-gate.js.map +1 -0
- package/dist/services/verify.d.ts +114 -0
- package/dist/services/verify.d.ts.map +1 -0
- package/dist/services/verify.js +237 -0
- package/dist/services/verify.js.map +1 -0
- package/dist/services/web-search.d.ts +145 -0
- package/dist/services/web-search.d.ts.map +1 -0
- package/dist/services/web-search.js +534 -0
- package/dist/services/web-search.js.map +1 -0
- package/dist/src/api/cli-server.d.ts +83 -0
- package/dist/src/api/cli-server.d.ts.map +1 -0
- package/dist/src/api/cli-server.js +597 -0
- package/dist/src/api/cli-server.js.map +1 -0
- package/dist/src/api/index.d.ts +6 -0
- package/dist/src/api/index.d.ts.map +1 -0
- package/dist/src/api/index.js +6 -0
- package/dist/src/api/index.js.map +1 -0
- package/dist/src/cli/commands/attractor.d.ts +6 -0
- package/dist/src/cli/commands/attractor.d.ts.map +1 -0
- package/dist/src/cli/commands/attractor.js +167 -0
- package/dist/src/cli/commands/attractor.js.map +1 -0
- package/dist/src/cli/commands/dimension.d.ts +6 -0
- package/dist/src/cli/commands/dimension.d.ts.map +1 -0
- package/dist/src/cli/commands/dimension.js +85 -0
- package/dist/src/cli/commands/dimension.js.map +1 -0
- package/dist/src/cli/commands/index.d.ts +11 -0
- package/dist/src/cli/commands/index.d.ts.map +1 -0
- package/dist/src/cli/commands/index.js +11 -0
- package/dist/src/cli/commands/index.js.map +1 -0
- package/dist/src/cli/commands/infer.d.ts +6 -0
- package/dist/src/cli/commands/infer.d.ts.map +1 -0
- package/dist/src/cli/commands/infer.js +139 -0
- package/dist/src/cli/commands/infer.js.map +1 -0
- package/dist/src/cli/commands/learn.d.ts +6 -0
- package/dist/src/cli/commands/learn.d.ts.map +1 -0
- package/dist/src/cli/commands/learn.js +87 -0
- package/dist/src/cli/commands/learn.js.map +1 -0
- package/dist/src/cli/commands/orchestrate.d.ts +6 -0
- package/dist/src/cli/commands/orchestrate.d.ts.map +1 -0
- package/dist/src/cli/commands/orchestrate.js +279 -0
- package/dist/src/cli/commands/orchestrate.js.map +1 -0
- package/dist/src/cli/commands/prob.d.ts +6 -0
- package/dist/src/cli/commands/prob.d.ts.map +1 -0
- package/dist/src/cli/commands/prob.js +256 -0
- package/dist/src/cli/commands/prob.js.map +1 -0
- package/dist/src/cli/commands/sync.d.ts +65 -0
- package/dist/src/cli/commands/sync.d.ts.map +1 -0
- package/dist/src/cli/commands/sync.js +338 -0
- package/dist/src/cli/commands/sync.js.map +1 -0
- package/dist/src/cli/index.d.ts +9 -0
- package/dist/src/cli/index.d.ts.map +1 -0
- package/dist/src/cli/index.js +9 -0
- package/dist/src/cli/index.js.map +1 -0
- package/dist/src/cli/probabilistic-commands.d.ts +39 -0
- package/dist/src/cli/probabilistic-commands.d.ts.map +1 -0
- package/dist/src/cli/probabilistic-commands.js +112 -0
- package/dist/src/cli/probabilistic-commands.js.map +1 -0
- package/dist/src/cli/types.d.ts +69 -0
- package/dist/src/cli/types.d.ts.map +1 -0
- package/dist/src/cli/types.js +5 -0
- package/dist/src/cli/types.js.map +1 -0
- package/dist/src/cli/utils/formatters.d.ts +51 -0
- package/dist/src/cli/utils/formatters.d.ts.map +1 -0
- package/dist/src/cli/utils/formatters.js +79 -0
- package/dist/src/cli/utils/formatters.js.map +1 -0
- package/dist/src/cli/utils/helpers.d.ts +21 -0
- package/dist/src/cli/utils/helpers.d.ts.map +1 -0
- package/dist/src/cli/utils/helpers.js +51 -0
- package/dist/src/cli/utils/helpers.js.map +1 -0
- package/dist/src/cli/utils/index.d.ts +7 -0
- package/dist/src/cli/utils/index.d.ts.map +1 -0
- package/dist/src/cli/utils/index.js +13 -0
- package/dist/src/cli/utils/index.js.map +1 -0
- package/dist/src/cli/utils/validators.d.ts +162 -0
- package/dist/src/cli/utils/validators.d.ts.map +1 -0
- package/dist/src/cli/utils/validators.js +351 -0
- package/dist/src/cli/utils/validators.js.map +1 -0
- package/dist/src/core/attractor-model.d.ts +381 -0
- package/dist/src/core/attractor-model.d.ts.map +1 -0
- package/dist/src/core/attractor-model.js +821 -0
- package/dist/src/core/attractor-model.js.map +1 -0
- package/dist/src/core/bidirectional-inference.d.ts +143 -0
- package/dist/src/core/bidirectional-inference.d.ts.map +1 -0
- package/dist/src/core/bidirectional-inference.js +501 -0
- package/dist/src/core/bidirectional-inference.js.map +1 -0
- package/dist/src/core/chunk-engine.d.ts +78 -0
- package/dist/src/core/chunk-engine.d.ts.map +1 -0
- package/dist/src/core/chunk-engine.js +192 -0
- package/dist/src/core/chunk-engine.js.map +1 -0
- package/dist/src/core/dynamic-embedding.d.ts +327 -0
- package/dist/src/core/dynamic-embedding.d.ts.map +1 -0
- package/dist/src/core/dynamic-embedding.js +527 -0
- package/dist/src/core/dynamic-embedding.js.map +1 -0
- package/dist/src/core/evolution-scheduler.d.ts +101 -0
- package/dist/src/core/evolution-scheduler.d.ts.map +1 -0
- package/dist/src/core/evolution-scheduler.js +235 -0
- package/dist/src/core/evolution-scheduler.js.map +1 -0
- package/dist/src/core/hnsw-index.d.ts +111 -0
- package/dist/src/core/hnsw-index.d.ts.map +1 -0
- package/dist/src/core/hnsw-index.js +466 -0
- package/dist/src/core/hnsw-index.js.map +1 -0
- package/dist/src/core/index.d.ts +23 -0
- package/dist/src/core/index.d.ts.map +1 -0
- package/dist/src/core/index.js +25 -0
- package/dist/src/core/index.js.map +1 -0
- package/dist/src/core/merkle-engine.d.ts +263 -0
- package/dist/src/core/merkle-engine.d.ts.map +1 -0
- package/dist/src/core/merkle-engine.js +528 -0
- package/dist/src/core/merkle-engine.js.map +1 -0
- package/dist/src/core/neuron-graph.d.ts +134 -0
- package/dist/src/core/neuron-graph.d.ts.map +1 -0
- package/dist/src/core/neuron-graph.js +436 -0
- package/dist/src/core/neuron-graph.js.map +1 -0
- package/dist/src/core/probabilistic-neuron.d.ts +251 -0
- package/dist/src/core/probabilistic-neuron.d.ts.map +1 -0
- package/dist/src/core/probabilistic-neuron.js +618 -0
- package/dist/src/core/probabilistic-neuron.js.map +1 -0
- package/dist/src/core/probabilistic-orchestrator.d.ts +408 -0
- package/dist/src/core/probabilistic-orchestrator.d.ts.map +1 -0
- package/dist/src/core/probabilistic-orchestrator.js +798 -0
- package/dist/src/core/probabilistic-orchestrator.js.map +1 -0
- package/dist/src/events/event-bus.d.ts +166 -0
- package/dist/src/events/event-bus.d.ts.map +1 -0
- package/dist/src/events/event-bus.js +228 -0
- package/dist/src/events/event-bus.js.map +1 -0
- package/dist/src/events/index.d.ts +7 -0
- package/dist/src/events/index.d.ts.map +1 -0
- package/dist/src/events/index.js +7 -0
- package/dist/src/events/index.js.map +1 -0
- package/dist/src/events/progress-tracker.d.ts +150 -0
- package/dist/src/events/progress-tracker.d.ts.map +1 -0
- package/dist/src/events/progress-tracker.js +290 -0
- package/dist/src/events/progress-tracker.js.map +1 -0
- package/dist/src/extensions/clustering/index.d.ts +114 -0
- package/dist/src/extensions/clustering/index.d.ts.map +1 -0
- package/dist/src/extensions/clustering/index.js +468 -0
- package/dist/src/extensions/clustering/index.js.map +1 -0
- package/dist/src/index.d.ts +14 -0
- package/dist/src/index.d.ts.map +1 -0
- package/dist/src/index.js +20 -0
- package/dist/src/index.js.map +1 -0
- package/dist/src/mcp/server.d.ts +43 -0
- package/dist/src/mcp/server.d.ts.map +1 -0
- package/dist/src/mcp/server.js +494 -0
- package/dist/src/mcp/server.js.map +1 -0
- package/dist/src/services/embedding-provider.d.ts +72 -0
- package/dist/src/services/embedding-provider.d.ts.map +1 -0
- package/dist/src/services/embedding-provider.js +176 -0
- package/dist/src/services/embedding-provider.js.map +1 -0
- package/dist/src/services/four-stage-learning.d.ts +552 -0
- package/dist/src/services/four-stage-learning.d.ts.map +1 -0
- package/dist/src/services/four-stage-learning.js +1110 -0
- package/dist/src/services/four-stage-learning.js.map +1 -0
- package/dist/src/services/index.d.ts +15 -0
- package/dist/src/services/index.d.ts.map +1 -0
- package/dist/src/services/index.js +18 -0
- package/dist/src/services/index.js.map +1 -0
- package/dist/src/services/ingestion.d.ts +98 -0
- package/dist/src/services/ingestion.d.ts.map +1 -0
- package/dist/src/services/ingestion.js +259 -0
- package/dist/src/services/ingestion.js.map +1 -0
- package/dist/src/services/query.d.ts +102 -0
- package/dist/src/services/query.d.ts.map +1 -0
- package/dist/src/services/query.js +227 -0
- package/dist/src/services/query.js.map +1 -0
- package/dist/src/services/text-embedding.d.ts +183 -0
- package/dist/src/services/text-embedding.d.ts.map +1 -0
- package/dist/src/services/text-embedding.js +633 -0
- package/dist/src/services/text-embedding.js.map +1 -0
- package/dist/src/services/verify.d.ts +114 -0
- package/dist/src/services/verify.d.ts.map +1 -0
- package/dist/src/services/verify.js +237 -0
- package/dist/src/services/verify.js.map +1 -0
- package/dist/src/storage/chunk-store.d.ts +107 -0
- package/dist/src/storage/chunk-store.d.ts.map +1 -0
- package/dist/src/storage/chunk-store.js +293 -0
- package/dist/src/storage/chunk-store.js.map +1 -0
- package/dist/src/storage/index-store.d.ts +126 -0
- package/dist/src/storage/index-store.d.ts.map +1 -0
- package/dist/src/storage/index-store.js +316 -0
- package/dist/src/storage/index-store.js.map +1 -0
- package/dist/src/storage/index.d.ts +45 -0
- package/dist/src/storage/index.d.ts.map +1 -0
- package/dist/src/storage/index.js +52 -0
- package/dist/src/storage/index.js.map +1 -0
- package/dist/src/storage/neuron-store.d.ts +121 -0
- package/dist/src/storage/neuron-store.d.ts.map +1 -0
- package/dist/src/storage/neuron-store.js +466 -0
- package/dist/src/storage/neuron-store.js.map +1 -0
- package/dist/src/storage/probabilistic-store.d.ts +104 -0
- package/dist/src/storage/probabilistic-store.d.ts.map +1 -0
- package/dist/src/storage/probabilistic-store.js +257 -0
- package/dist/src/storage/probabilistic-store.js.map +1 -0
- package/dist/src/sync/change-journal.d.ts +171 -0
- package/dist/src/sync/change-journal.d.ts.map +1 -0
- package/dist/src/sync/change-journal.js +362 -0
- package/dist/src/sync/change-journal.js.map +1 -0
- package/dist/src/sync/index.d.ts +8 -0
- package/dist/src/sync/index.d.ts.map +1 -0
- package/dist/src/sync/index.js +8 -0
- package/dist/src/sync/index.js.map +1 -0
- package/dist/src/sync/state-sync.d.ts +241 -0
- package/dist/src/sync/state-sync.d.ts.map +1 -0
- package/dist/src/sync/state-sync.js +396 -0
- package/dist/src/sync/state-sync.js.map +1 -0
- package/dist/src/sync/vector-clock.d.ts +144 -0
- package/dist/src/sync/vector-clock.d.ts.map +1 -0
- package/dist/src/sync/vector-clock.js +266 -0
- package/dist/src/sync/vector-clock.js.map +1 -0
- package/dist/src/types/index.d.ts +224 -0
- package/dist/src/types/index.d.ts.map +1 -0
- package/dist/src/types/index.js +24 -0
- package/dist/src/types/index.js.map +1 -0
- package/dist/src/utils/hash.d.ts +39 -0
- package/dist/src/utils/hash.d.ts.map +1 -0
- package/dist/src/utils/hash.js +56 -0
- package/dist/src/utils/hash.js.map +1 -0
- package/dist/src/utils/index.d.ts +26 -0
- package/dist/src/utils/index.d.ts.map +1 -0
- package/dist/src/utils/index.js +50 -0
- package/dist/src/utils/index.js.map +1 -0
- package/dist/src/utils/logger.d.ts +88 -0
- package/dist/src/utils/logger.d.ts.map +1 -0
- package/dist/src/utils/logger.js +157 -0
- package/dist/src/utils/logger.js.map +1 -0
- package/dist/src/utils/metrics.d.ts +232 -0
- package/dist/src/utils/metrics.d.ts.map +1 -0
- package/dist/src/utils/metrics.js +387 -0
- package/dist/src/utils/metrics.js.map +1 -0
- package/dist/src/utils/similarity.d.ts +64 -0
- package/dist/src/utils/similarity.d.ts.map +1 -0
- package/dist/src/utils/similarity.js +151 -0
- package/dist/src/utils/similarity.js.map +1 -0
- package/dist/src/utils/uuid.d.ts +23 -0
- package/dist/src/utils/uuid.d.ts.map +1 -0
- package/dist/src/utils/uuid.js +29 -0
- package/dist/src/utils/uuid.js.map +1 -0
- package/dist/storage/chunk-store.d.ts +107 -0
- package/dist/storage/chunk-store.d.ts.map +1 -0
- package/dist/storage/chunk-store.js +293 -0
- package/dist/storage/chunk-store.js.map +1 -0
- package/dist/storage/hybrid-adapters.d.ts +111 -0
- package/dist/storage/hybrid-adapters.d.ts.map +1 -0
- package/dist/storage/hybrid-adapters.js +223 -0
- package/dist/storage/hybrid-adapters.js.map +1 -0
- package/dist/storage/hybrid-store.d.ts +125 -0
- package/dist/storage/hybrid-store.d.ts.map +1 -0
- package/dist/storage/hybrid-store.js +655 -0
- package/dist/storage/hybrid-store.js.map +1 -0
- package/dist/storage/index-store.d.ts +126 -0
- package/dist/storage/index-store.d.ts.map +1 -0
- package/dist/storage/index-store.js +316 -0
- package/dist/storage/index-store.js.map +1 -0
- package/dist/storage/index.d.ts +45 -0
- package/dist/storage/index.d.ts.map +1 -0
- package/dist/storage/index.js +52 -0
- package/dist/storage/index.js.map +1 -0
- package/dist/storage/neuron-store.d.ts +121 -0
- package/dist/storage/neuron-store.d.ts.map +1 -0
- package/dist/storage/neuron-store.js +466 -0
- package/dist/storage/neuron-store.js.map +1 -0
- package/dist/storage/ontology-store.d.ts +132 -0
- package/dist/storage/ontology-store.d.ts.map +1 -0
- package/dist/storage/ontology-store.js +319 -0
- package/dist/storage/ontology-store.js.map +1 -0
- package/dist/storage/probabilistic-store.d.ts +104 -0
- package/dist/storage/probabilistic-store.d.ts.map +1 -0
- package/dist/storage/probabilistic-store.js +257 -0
- package/dist/storage/probabilistic-store.js.map +1 -0
- package/dist/storage/redis-adapters.d.ts +102 -0
- package/dist/storage/redis-adapters.d.ts.map +1 -0
- package/dist/storage/redis-adapters.js +205 -0
- package/dist/storage/redis-adapters.js.map +1 -0
- package/dist/storage/redis-ontology-store.d.ts +146 -0
- package/dist/storage/redis-ontology-store.d.ts.map +1 -0
- package/dist/storage/redis-ontology-store.js +384 -0
- package/dist/storage/redis-ontology-store.js.map +1 -0
- package/dist/storage/redis-store.d.ts +174 -0
- package/dist/storage/redis-store.d.ts.map +1 -0
- package/dist/storage/redis-store.js +506 -0
- package/dist/storage/redis-store.js.map +1 -0
- package/dist/sync/change-journal.d.ts +171 -0
- package/dist/sync/change-journal.d.ts.map +1 -0
- package/dist/sync/change-journal.js +362 -0
- package/dist/sync/change-journal.js.map +1 -0
- package/dist/sync/index.d.ts +8 -0
- package/dist/sync/index.d.ts.map +1 -0
- package/dist/sync/index.js +8 -0
- package/dist/sync/index.js.map +1 -0
- package/dist/sync/state-sync.d.ts +241 -0
- package/dist/sync/state-sync.d.ts.map +1 -0
- package/dist/sync/state-sync.js +396 -0
- package/dist/sync/state-sync.js.map +1 -0
- package/dist/sync/vector-clock.d.ts +144 -0
- package/dist/sync/vector-clock.d.ts.map +1 -0
- package/dist/sync/vector-clock.js +266 -0
- package/dist/sync/vector-clock.js.map +1 -0
- package/dist/types/index.d.ts +224 -0
- package/dist/types/index.d.ts.map +1 -0
- package/dist/types/index.js +24 -0
- package/dist/types/index.js.map +1 -0
- package/dist/utils/hash.d.ts +39 -0
- package/dist/utils/hash.d.ts.map +1 -0
- package/dist/utils/hash.js +56 -0
- package/dist/utils/hash.js.map +1 -0
- package/dist/utils/index.d.ts +26 -0
- package/dist/utils/index.d.ts.map +1 -0
- package/dist/utils/index.js +50 -0
- package/dist/utils/index.js.map +1 -0
- package/dist/utils/logger.d.ts +88 -0
- package/dist/utils/logger.d.ts.map +1 -0
- package/dist/utils/logger.js +157 -0
- package/dist/utils/logger.js.map +1 -0
- package/dist/utils/metrics.d.ts +232 -0
- package/dist/utils/metrics.d.ts.map +1 -0
- package/dist/utils/metrics.js +387 -0
- package/dist/utils/metrics.js.map +1 -0
- package/dist/utils/similarity.d.ts +64 -0
- package/dist/utils/similarity.d.ts.map +1 -0
- package/dist/utils/similarity.js +151 -0
- package/dist/utils/similarity.js.map +1 -0
- package/dist/utils/uuid.d.ts +23 -0
- package/dist/utils/uuid.d.ts.map +1 -0
- package/dist/utils/uuid.js +29 -0
- package/dist/utils/uuid.js.map +1 -0
- package/package.json +72 -0
|
@@ -0,0 +1,365 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Language-specific Analyzers
|
|
3
|
+
* 언어별 형태소 분석 및 청킹 최적화
|
|
4
|
+
*
|
|
5
|
+
* @module core/language-analyzers
|
|
6
|
+
*/
|
|
7
|
+
/**
|
|
8
|
+
* Korean Analyzer
|
|
9
|
+
* 한국어 형태소 분석기
|
|
10
|
+
*/
|
|
11
|
+
export class KoreanAnalyzer {
|
|
12
|
+
language = 'ko';
|
|
13
|
+
// 한국어 조사 목록
|
|
14
|
+
particles = new Set([
|
|
15
|
+
'은', '는', '이', '가', '을', '를', '에', '에서', '에게', '께',
|
|
16
|
+
'의', '로', '으로', '와', '과', '랑', '이랑', '도', '만', '부터',
|
|
17
|
+
'까지', '보다', '처럼', '같이', '마다', '밖에', '뿐', '조차'
|
|
18
|
+
]);
|
|
19
|
+
// 한국어 어미 목록
|
|
20
|
+
suffixes = new Set([
|
|
21
|
+
'다', '니다', '습니다', '세요', '요', '네', '죠', '지', '고',
|
|
22
|
+
'면', '어서', '아서', '니까', '는데', '던', '었', '았', '겠'
|
|
23
|
+
]);
|
|
24
|
+
// 문장 종결 패턴
|
|
25
|
+
sentenceEnders = /[.!?。!?]\s*/g;
|
|
26
|
+
tokenize(text) {
|
|
27
|
+
const tokens = [];
|
|
28
|
+
const words = text.split(/\s+/);
|
|
29
|
+
for (const word of words) {
|
|
30
|
+
if (!word)
|
|
31
|
+
continue;
|
|
32
|
+
// 조사 분리
|
|
33
|
+
let remaining = word;
|
|
34
|
+
let foundParticle = false;
|
|
35
|
+
for (const particle of this.particles) {
|
|
36
|
+
if (remaining.endsWith(particle) && remaining.length > particle.length) {
|
|
37
|
+
const stem = remaining.slice(0, -particle.length);
|
|
38
|
+
tokens.push({ text: stem, type: 'word' });
|
|
39
|
+
tokens.push({ text: particle, type: 'particle' });
|
|
40
|
+
foundParticle = true;
|
|
41
|
+
break;
|
|
42
|
+
}
|
|
43
|
+
}
|
|
44
|
+
if (!foundParticle) {
|
|
45
|
+
// 어미 분리 시도
|
|
46
|
+
let foundSuffix = false;
|
|
47
|
+
for (const suffix of this.suffixes) {
|
|
48
|
+
if (remaining.endsWith(suffix) && remaining.length > suffix.length) {
|
|
49
|
+
const stem = remaining.slice(0, -suffix.length);
|
|
50
|
+
tokens.push({ text: stem, type: 'word' });
|
|
51
|
+
tokens.push({ text: suffix, type: 'suffix' });
|
|
52
|
+
foundSuffix = true;
|
|
53
|
+
break;
|
|
54
|
+
}
|
|
55
|
+
}
|
|
56
|
+
if (!foundSuffix) {
|
|
57
|
+
tokens.push({ text: remaining, type: 'word' });
|
|
58
|
+
}
|
|
59
|
+
}
|
|
60
|
+
}
|
|
61
|
+
return tokens;
|
|
62
|
+
}
|
|
63
|
+
splitSentences(text) {
|
|
64
|
+
return text.split(this.sentenceEnders).filter(s => s.trim().length > 0);
|
|
65
|
+
}
|
|
66
|
+
splitPhrases(text) {
|
|
67
|
+
// 쉼표, 접속사로 분리
|
|
68
|
+
return text.split(/[,,、]|\s+(그리고|그러나|하지만|그래서|또한)\s+/g)
|
|
69
|
+
.filter(p => p && p.trim().length > 0);
|
|
70
|
+
}
|
|
71
|
+
splitWords(text) {
|
|
72
|
+
return text.split(/\s+/).filter(w => w.trim().length > 0);
|
|
73
|
+
}
|
|
74
|
+
splitMorphemes(text) {
|
|
75
|
+
return this.tokenize(text).map(t => t.text);
|
|
76
|
+
}
|
|
77
|
+
}
|
|
78
|
+
/**
|
|
79
|
+
* English Analyzer
|
|
80
|
+
* 영어 형태소 분석기
|
|
81
|
+
*/
|
|
82
|
+
export class EnglishAnalyzer {
|
|
83
|
+
language = 'en';
|
|
84
|
+
// Common English suffixes
|
|
85
|
+
suffixes = ['ing', 'ed', 'ly', 'tion', 'sion', 'ness', 'ment', 'able', 'ible'];
|
|
86
|
+
// Stop words
|
|
87
|
+
stopWords = new Set([
|
|
88
|
+
'the', 'a', 'an', 'is', 'are', 'was', 'were', 'be', 'been', 'being',
|
|
89
|
+
'have', 'has', 'had', 'do', 'does', 'did', 'will', 'would', 'could',
|
|
90
|
+
'should', 'may', 'might', 'must', 'shall', 'can', 'of', 'at', 'by',
|
|
91
|
+
'for', 'with', 'about', 'against', 'between', 'into', 'through',
|
|
92
|
+
'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up',
|
|
93
|
+
'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further',
|
|
94
|
+
'then', 'once', 'and', 'but', 'or', 'nor', 'so', 'yet'
|
|
95
|
+
]);
|
|
96
|
+
tokenize(text) {
|
|
97
|
+
const tokens = [];
|
|
98
|
+
const words = text.toLowerCase().split(/\s+/);
|
|
99
|
+
for (const word of words) {
|
|
100
|
+
if (!word)
|
|
101
|
+
continue;
|
|
102
|
+
// Remove punctuation
|
|
103
|
+
const cleanWord = word.replace(/[^\w]/g, '');
|
|
104
|
+
if (this.stopWords.has(cleanWord)) {
|
|
105
|
+
tokens.push({ text: cleanWord, type: 'particle' });
|
|
106
|
+
}
|
|
107
|
+
else {
|
|
108
|
+
// Try to find suffix
|
|
109
|
+
let foundSuffix = false;
|
|
110
|
+
for (const suffix of this.suffixes) {
|
|
111
|
+
if (cleanWord.endsWith(suffix) && cleanWord.length > suffix.length + 2) {
|
|
112
|
+
const stem = cleanWord.slice(0, -suffix.length);
|
|
113
|
+
tokens.push({ text: stem, type: 'word', lemma: stem });
|
|
114
|
+
tokens.push({ text: suffix, type: 'suffix' });
|
|
115
|
+
foundSuffix = true;
|
|
116
|
+
break;
|
|
117
|
+
}
|
|
118
|
+
}
|
|
119
|
+
if (!foundSuffix) {
|
|
120
|
+
tokens.push({ text: cleanWord, type: 'word' });
|
|
121
|
+
}
|
|
122
|
+
}
|
|
123
|
+
}
|
|
124
|
+
return tokens;
|
|
125
|
+
}
|
|
126
|
+
splitSentences(text) {
|
|
127
|
+
return text.split(/(?<=[.!?])\s+/).filter(s => s.trim().length > 0);
|
|
128
|
+
}
|
|
129
|
+
splitPhrases(text) {
|
|
130
|
+
return text.split(/[,;:]|\s+(and|but|or|because|although|however)\s+/gi)
|
|
131
|
+
.filter(p => p && p.trim().length > 0 && p.length > 3);
|
|
132
|
+
}
|
|
133
|
+
splitWords(text) {
|
|
134
|
+
return text.split(/\s+/).filter(w => w.trim().length > 0);
|
|
135
|
+
}
|
|
136
|
+
splitMorphemes(text) {
|
|
137
|
+
return this.tokenize(text).map(t => t.text);
|
|
138
|
+
}
|
|
139
|
+
}
|
|
140
|
+
/**
|
|
141
|
+
* Japanese Analyzer
|
|
142
|
+
* 일본어 형태소 분석기
|
|
143
|
+
*/
|
|
144
|
+
export class JapaneseAnalyzer {
|
|
145
|
+
language = 'ja';
|
|
146
|
+
// Unicode ranges
|
|
147
|
+
hiraganaRange = /[\u3040-\u309F]/;
|
|
148
|
+
katakanaRange = /[\u30A0-\u30FF]/;
|
|
149
|
+
kanjiRange = /[\u4E00-\u9FAF]/;
|
|
150
|
+
tokenize(text) {
|
|
151
|
+
const tokens = [];
|
|
152
|
+
let currentToken = '';
|
|
153
|
+
let currentType = 'word';
|
|
154
|
+
for (const char of text) {
|
|
155
|
+
let charType = 'word';
|
|
156
|
+
if (this.hiraganaRange.test(char)) {
|
|
157
|
+
charType = 'hiragana';
|
|
158
|
+
}
|
|
159
|
+
else if (this.katakanaRange.test(char)) {
|
|
160
|
+
charType = 'katakana';
|
|
161
|
+
}
|
|
162
|
+
else if (this.kanjiRange.test(char)) {
|
|
163
|
+
charType = 'kanji';
|
|
164
|
+
}
|
|
165
|
+
else if (/\s/.test(char)) {
|
|
166
|
+
charType = 'whitespace';
|
|
167
|
+
}
|
|
168
|
+
else if (/[。!?、]/.test(char)) {
|
|
169
|
+
charType = 'punctuation';
|
|
170
|
+
}
|
|
171
|
+
if (charType !== currentType && currentToken) {
|
|
172
|
+
tokens.push({ text: currentToken, type: currentType });
|
|
173
|
+
currentToken = '';
|
|
174
|
+
}
|
|
175
|
+
if (charType !== 'whitespace') {
|
|
176
|
+
currentToken += char;
|
|
177
|
+
}
|
|
178
|
+
currentType = charType;
|
|
179
|
+
}
|
|
180
|
+
if (currentToken) {
|
|
181
|
+
tokens.push({ text: currentToken, type: currentType });
|
|
182
|
+
}
|
|
183
|
+
return tokens;
|
|
184
|
+
}
|
|
185
|
+
splitSentences(text) {
|
|
186
|
+
return text.split(/[。!?]\s*/).filter(s => s.trim().length > 0);
|
|
187
|
+
}
|
|
188
|
+
splitPhrases(text) {
|
|
189
|
+
return text.split(/[、,]\s*/).filter(p => p.trim().length > 0);
|
|
190
|
+
}
|
|
191
|
+
splitWords(text) {
|
|
192
|
+
// Japanese doesn't use spaces, so we split by character type changes
|
|
193
|
+
const tokens = this.tokenize(text);
|
|
194
|
+
return tokens.filter(t => t.type !== 'punctuation').map(t => t.text);
|
|
195
|
+
}
|
|
196
|
+
splitMorphemes(text) {
|
|
197
|
+
return this.tokenize(text).map(t => t.text);
|
|
198
|
+
}
|
|
199
|
+
}
|
|
200
|
+
/**
|
|
201
|
+
* Chinese Analyzer
|
|
202
|
+
* 중국어 형태소 분석기
|
|
203
|
+
*/
|
|
204
|
+
export class ChineseAnalyzer {
|
|
205
|
+
language = 'zh';
|
|
206
|
+
// Common Chinese particles/function words
|
|
207
|
+
particles = new Set([
|
|
208
|
+
'的', '了', '是', '在', '有', '和', '与', '或', '但', '而',
|
|
209
|
+
'就', '也', '都', '还', '又', '才', '只', '已', '把', '被',
|
|
210
|
+
'让', '给', '向', '从', '到', '对', '为', '以', '因', '所'
|
|
211
|
+
]);
|
|
212
|
+
tokenize(text) {
|
|
213
|
+
const tokens = [];
|
|
214
|
+
for (const char of text) {
|
|
215
|
+
if (/\s/.test(char))
|
|
216
|
+
continue;
|
|
217
|
+
if (/[。!?,、]/.test(char)) {
|
|
218
|
+
tokens.push({ text: char, type: 'punctuation' });
|
|
219
|
+
}
|
|
220
|
+
else if (this.particles.has(char)) {
|
|
221
|
+
tokens.push({ text: char, type: 'particle' });
|
|
222
|
+
}
|
|
223
|
+
else {
|
|
224
|
+
tokens.push({ text: char, type: 'hanzi' });
|
|
225
|
+
}
|
|
226
|
+
}
|
|
227
|
+
return tokens;
|
|
228
|
+
}
|
|
229
|
+
splitSentences(text) {
|
|
230
|
+
return text.split(/[。!?]\s*/).filter(s => s.trim().length > 0);
|
|
231
|
+
}
|
|
232
|
+
splitPhrases(text) {
|
|
233
|
+
return text.split(/[,、;]\s*/).filter(p => p.trim().length > 0);
|
|
234
|
+
}
|
|
235
|
+
splitWords(text) {
|
|
236
|
+
// Basic bigram approach for Chinese
|
|
237
|
+
const chars = [...text].filter(c => !/\s/.test(c) && !/[。!?,、;]/.test(c));
|
|
238
|
+
const words = [];
|
|
239
|
+
for (let i = 0; i < chars.length; i += 2) {
|
|
240
|
+
if (i + 1 < chars.length) {
|
|
241
|
+
words.push(chars[i] + chars[i + 1]);
|
|
242
|
+
}
|
|
243
|
+
else {
|
|
244
|
+
words.push(chars[i]);
|
|
245
|
+
}
|
|
246
|
+
}
|
|
247
|
+
return words;
|
|
248
|
+
}
|
|
249
|
+
splitMorphemes(text) {
|
|
250
|
+
// Each character is a morpheme in Chinese
|
|
251
|
+
return [...text].filter(c => !/\s/.test(c));
|
|
252
|
+
}
|
|
253
|
+
}
|
|
254
|
+
/**
|
|
255
|
+
* Auto-detect language and return appropriate analyzer
|
|
256
|
+
*/
|
|
257
|
+
export function detectLanguage(text) {
|
|
258
|
+
// Count character types
|
|
259
|
+
let korean = 0, japanese = 0, chinese = 0, english = 0, total = 0;
|
|
260
|
+
for (const char of text) {
|
|
261
|
+
if (/\s/.test(char))
|
|
262
|
+
continue;
|
|
263
|
+
total++;
|
|
264
|
+
// Korean Hangul
|
|
265
|
+
if (/[\uAC00-\uD7AF\u1100-\u11FF\u3130-\u318F]/.test(char)) {
|
|
266
|
+
korean++;
|
|
267
|
+
}
|
|
268
|
+
// Japanese (Hiragana, Katakana)
|
|
269
|
+
else if (/[\u3040-\u309F\u30A0-\u30FF]/.test(char)) {
|
|
270
|
+
japanese++;
|
|
271
|
+
}
|
|
272
|
+
// CJK Unified Ideographs (shared by Chinese/Japanese/Korean)
|
|
273
|
+
else if (/[\u4E00-\u9FAF]/.test(char)) {
|
|
274
|
+
chinese++; // Assume Chinese if no kana
|
|
275
|
+
}
|
|
276
|
+
// Latin
|
|
277
|
+
else if (/[a-zA-Z]/.test(char)) {
|
|
278
|
+
english++;
|
|
279
|
+
}
|
|
280
|
+
}
|
|
281
|
+
if (total === 0)
|
|
282
|
+
return 'en';
|
|
283
|
+
const koRatio = korean / total;
|
|
284
|
+
const jaRatio = japanese / total;
|
|
285
|
+
const zhRatio = chinese / total;
|
|
286
|
+
const enRatio = english / total;
|
|
287
|
+
// Determine language
|
|
288
|
+
if (koRatio > 0.3)
|
|
289
|
+
return 'ko';
|
|
290
|
+
if (jaRatio > 0.1)
|
|
291
|
+
return 'ja'; // Even small amount of kana indicates Japanese
|
|
292
|
+
if (zhRatio > 0.5)
|
|
293
|
+
return 'zh';
|
|
294
|
+
if (enRatio > 0.5)
|
|
295
|
+
return 'en';
|
|
296
|
+
// Default fallback
|
|
297
|
+
if (koRatio > jaRatio && koRatio > zhRatio)
|
|
298
|
+
return 'ko';
|
|
299
|
+
if (jaRatio > zhRatio)
|
|
300
|
+
return 'ja';
|
|
301
|
+
if (zhRatio > 0)
|
|
302
|
+
return 'zh';
|
|
303
|
+
return 'en';
|
|
304
|
+
}
|
|
305
|
+
/**
|
|
306
|
+
* Get analyzer for language
|
|
307
|
+
*/
|
|
308
|
+
export function getAnalyzer(language, text) {
|
|
309
|
+
let lang = language;
|
|
310
|
+
if (language === 'auto' && text) {
|
|
311
|
+
lang = detectLanguage(text);
|
|
312
|
+
}
|
|
313
|
+
switch (lang) {
|
|
314
|
+
case 'ko':
|
|
315
|
+
return new KoreanAnalyzer();
|
|
316
|
+
case 'ja':
|
|
317
|
+
return new JapaneseAnalyzer();
|
|
318
|
+
case 'zh':
|
|
319
|
+
return new ChineseAnalyzer();
|
|
320
|
+
case 'en':
|
|
321
|
+
default:
|
|
322
|
+
return new EnglishAnalyzer();
|
|
323
|
+
}
|
|
324
|
+
}
|
|
325
|
+
/**
|
|
326
|
+
* Multi-language tokenizer
|
|
327
|
+
*/
|
|
328
|
+
export class MultiLanguageTokenizer {
|
|
329
|
+
analyzers = new Map();
|
|
330
|
+
constructor() {
|
|
331
|
+
this.analyzers.set('ko', new KoreanAnalyzer());
|
|
332
|
+
this.analyzers.set('en', new EnglishAnalyzer());
|
|
333
|
+
this.analyzers.set('ja', new JapaneseAnalyzer());
|
|
334
|
+
this.analyzers.set('zh', new ChineseAnalyzer());
|
|
335
|
+
}
|
|
336
|
+
/**
|
|
337
|
+
* Tokenize with auto language detection
|
|
338
|
+
*/
|
|
339
|
+
tokenize(text) {
|
|
340
|
+
const language = detectLanguage(text);
|
|
341
|
+
const analyzer = this.analyzers.get(language) || new EnglishAnalyzer();
|
|
342
|
+
return {
|
|
343
|
+
language,
|
|
344
|
+
tokens: analyzer.tokenize(text)
|
|
345
|
+
};
|
|
346
|
+
}
|
|
347
|
+
/**
|
|
348
|
+
* Split into all levels
|
|
349
|
+
*/
|
|
350
|
+
analyze(text) {
|
|
351
|
+
const language = detectLanguage(text);
|
|
352
|
+
const analyzer = this.analyzers.get(language) || new EnglishAnalyzer();
|
|
353
|
+
return {
|
|
354
|
+
language,
|
|
355
|
+
sentences: analyzer.splitSentences(text),
|
|
356
|
+
phrases: analyzer.splitPhrases(text),
|
|
357
|
+
words: analyzer.splitWords(text),
|
|
358
|
+
morphemes: analyzer.splitMorphemes(text),
|
|
359
|
+
tokens: analyzer.tokenize(text)
|
|
360
|
+
};
|
|
361
|
+
}
|
|
362
|
+
}
|
|
363
|
+
// Export singleton
|
|
364
|
+
export const multiLangTokenizer = new MultiLanguageTokenizer();
|
|
365
|
+
//# sourceMappingURL=language-analyzers.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"language-analyzers.js","sourceRoot":"","sources":["../../src/core/language-analyzers.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AA0CH;;;GAGG;AACH,MAAM,OAAO,cAAc;IACzB,QAAQ,GAAa,IAAI,CAAC;IAE1B,YAAY;IACJ,SAAS,GAAG,IAAI,GAAG,CAAC;QAC1B,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG;QAClD,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI;QACnD,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI;KAC9C,CAAC,CAAC;IAEH,YAAY;IACJ,QAAQ,GAAG,IAAI,GAAG,CAAC;QACzB,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;QAC/C,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;KAChD,CAAC,CAAC;IAEH,WAAW;IACH,cAAc,GAAG,cAAc,CAAC;IAExC,QAAQ,CAAC,IAAY;QACnB,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAEhC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,CAAC,IAAI;gBAAE,SAAS;YAEpB,QAAQ;YACR,IAAI,SAAS,GAAG,IAAI,CAAC;YACrB,IAAI,aAAa,GAAG,KAAK,CAAC;YAE1B,KAAK,MAAM,QAAQ,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;gBACtC,IAAI,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,SAAS,CAAC,MAAM,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC;oBACvE,MAAM,IAAI,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;oBAClD,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAC;oBAC1C,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;oBAClD,aAAa,GAAG,IAAI,CAAC;oBACrB,MAAM;gBACR,CAAC;YACH,CAAC;YAED,IAAI,CAAC,aAAa,EAAE,CAAC;gBACnB,WAAW;gBACX,IAAI,WAAW,GAAG,KAAK,CAAC;gBACxB,KAAK,MAAM,MAAM,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;oBACnC,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,SAAS,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;wBACnE,MAAM,IAAI,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;wBAChD,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAC;wBAC1C,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;wBAC9C,WAAW,GAAG,IAAI,CAAC;wBACnB,MAAM;oBACR,CAAC;gBACH,CAAC;gBAED,IAAI,CAAC,WAAW,EAAE,CAAC;oBACjB,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAC;gBACjD,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC1E,CAAC;IAED,YAAY,CAAC,IAAY;QACvB,cAAc;QACd,OAAO,IAAI,CAAC,KAAK,CAAC,mCAAmC,CAAC;aACnD,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC3C,CAAC;IAED,UAAU,CAAC,IAAY;QACrB,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5D,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IAC9C,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,OAAO,eAAe;IAC1B,QAAQ,GAAa,IAAI,CAAC;IAE1B,0BAA0B;IAClB,QAAQ,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;IAEvF,aAAa;IACL,SAAS,GAAG,IAAI,GAAG,CAAC;QAC1B,KAAK,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO;QACnE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;QACnE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;QAClE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS;QAC/D,QAAQ,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI;QACjE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,SAAS;QACrE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK;KACvD,CAAC,CAAC;IAEH,QAAQ,CAAC,IAAY;QACnB,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAE9C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,CAAC,IAAI;gBAAE,SAAS;YAEpB,qBAAqB;YACrB,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;YAE7C,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;gBAClC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;YACrD,CAAC;iBAAM,CAAC;gBACN,qBAAqB;gBACrB,IAAI,WAAW,GAAG,KAAK,CAAC;gBACxB,KAAK,MAAM,MAAM,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;oBACnC,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,SAAS,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBACvE,MAAM,IAAI,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;wBAChD,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC;wBACvD,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;wBAC9C,WAAW,GAAG,IAAI,CAAC;wBACnB,MAAM;oBACR,CAAC;gBACH,CAAC;gBAED,IAAI,CAAC,WAAW,EAAE,CAAC;oBACjB,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAC;gBACjD,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACtE,CAAC;IAED,YAAY,CAAC,IAAY;QACvB,OAAO,IAAI,CAAC,KAAK,CAAC,qDAAqD,CAAC;aACrE,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC3D,CAAC;IAED,UAAU,CAAC,IAAY;QACrB,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5D,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IAC9C,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,OAAO,gBAAgB;IAC3B,QAAQ,GAAa,IAAI,CAAC;IAE1B,iBAAiB;IACT,aAAa,GAAG,iBAAiB,CAAC;IAClC,aAAa,GAAG,iBAAiB,CAAC;IAClC,UAAU,GAAG,iBAAiB,CAAC;IAEvC,QAAQ,CAAC,IAAY;QACnB,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,IAAI,YAAY,GAAG,EAAE,CAAC;QACtB,IAAI,WAAW,GAAc,MAAM,CAAC;QAEpC,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACxB,IAAI,QAAQ,GAAc,MAAM,CAAC;YAEjC,IAAI,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAClC,QAAQ,GAAG,UAAU,CAAC;YACxB,CAAC;iBAAM,IAAI,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACzC,QAAQ,GAAG,UAAU,CAAC;YACxB,CAAC;iBAAM,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACtC,QAAQ,GAAG,OAAO,CAAC;YACrB,CAAC;iBAAM,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC3B,QAAQ,GAAG,YAAY,CAAC;YAC1B,CAAC;iBAAM,IAAI,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC/B,QAAQ,GAAG,aAAa,CAAC;YAC3B,CAAC;YAED,IAAI,QAAQ,KAAK,WAAW,IAAI,YAAY,EAAE,CAAC;gBAC7C,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,IAAI,EAAE,WAAW,EAAE,CAAC,CAAC;gBACvD,YAAY,GAAG,EAAE,CAAC;YACpB,CAAC;YAED,IAAI,QAAQ,KAAK,YAAY,EAAE,CAAC;gBAC9B,YAAY,IAAI,IAAI,CAAC;YACvB,CAAC;YACD,WAAW,GAAG,QAAQ,CAAC;QACzB,CAAC;QAED,IAAI,YAAY,EAAE,CAAC;YACjB,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,IAAI,EAAE,WAAW,EAAE,CAAC,CAAC;QACzD,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACjE,CAAC;IAED,YAAY,CAAC,IAAY;QACvB,OAAO,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAChE,CAAC;IAED,UAAU,CAAC,IAAY;QACrB,qEAAqE;QACrE,MAAM,MAAM,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QACnC,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,aAAa,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACvE,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IAC9C,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,OAAO,eAAe;IAC1B,QAAQ,GAAa,IAAI,CAAC;IAE1B,0CAA0C;IAClC,SAAS,GAAG,IAAI,GAAG,CAAC;QAC1B,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;QAChD,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;QAChD,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;KACjD,CAAC,CAAC;IAEH,QAAQ,CAAC,IAAY;QACnB,MAAM,MAAM,GAAY,EAAE,CAAC;QAE3B,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACxB,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC;gBAAE,SAAS;YAC9B,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACzB,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC,CAAC;YACnD,CAAC;iBAAM,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;gBACpC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;YAChD,CAAC;iBAAM,CAAC;gBACN,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,CAAC;YAC7C,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,OAAO,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACjE,CAAC;IAED,YAAY,CAAC,IAAY;QACvB,OAAO,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACjE,CAAC;IAED,UAAU,CAAC,IAAY;QACrB,oCAAoC;QACpC,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;QAC1E,MAAM,KAAK,GAAa,EAAE,CAAC;QAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACzC,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;gBACzB,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YACtC,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YACvB,CAAC;QACH,CAAC;QAED,OAAO,KAAK,CAAC;IACf,CAAC;IAED,cAAc,CAAC,IAAY;QACzB,0CAA0C;QAC1C,OAAO,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9C,CAAC;CACF;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAAC,IAAY;IACzC,wBAAwB;IACxB,IAAI,MAAM,GAAG,CAAC,EAAE,QAAQ,GAAG,CAAC,EAAE,OAAO,GAAG,CAAC,EAAE,OAAO,GAAG,CAAC,EAAE,KAAK,GAAG,CAAC,CAAC;IAElE,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;QACxB,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,SAAS;QAC9B,KAAK,EAAE,CAAC;QAER,gBAAgB;QAChB,IAAI,2CAA2C,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3D,MAAM,EAAE,CAAC;QACX,CAAC;QACD,gCAAgC;aAC3B,IAAI,8BAA8B,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACnD,QAAQ,EAAE,CAAC;QACb,CAAC;QACD,6DAA6D;aACxD,IAAI,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACtC,OAAO,EAAE,CAAC,CAAE,4BAA4B;QAC1C,CAAC;QACD,QAAQ;aACH,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC/B,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IAED,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAE7B,MAAM,OAAO,GAAG,MAAM,GAAG,KAAK,CAAC;IAC/B,MAAM,OAAO,GAAG,QAAQ,GAAG,KAAK,CAAC;IACjC,MAAM,OAAO,GAAG,OAAO,GAAG,KAAK,CAAC;IAChC,MAAM,OAAO,GAAG,OAAO,GAAG,KAAK,CAAC;IAEhC,qBAAqB;IACrB,IAAI,OAAO,GAAG,GAAG;QAAE,OAAO,IAAI,CAAC;IAC/B,IAAI,OAAO,GAAG,GAAG;QAAE,OAAO,IAAI,CAAC,CAAE,+CAA+C;IAChF,IAAI,OAAO,GAAG,GAAG;QAAE,OAAO,IAAI,CAAC;IAC/B,IAAI,OAAO,GAAG,GAAG;QAAE,OAAO,IAAI,CAAC;IAE/B,mBAAmB;IACnB,IAAI,OAAO,GAAG,OAAO,IAAI,OAAO,GAAG,OAAO;QAAE,OAAO,IAAI,CAAC;IACxD,IAAI,OAAO,GAAG,OAAO;QAAE,OAAO,IAAI,CAAC;IACnC,IAAI,OAAO,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAE7B,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,WAAW,CAAC,QAA2B,EAAE,IAAa;IACpE,IAAI,IAAI,GAAG,QAAQ,CAAC;IAEpB,IAAI,QAAQ,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC;QAChC,IAAI,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;IAC9B,CAAC;IAED,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,IAAI;YACP,OAAO,IAAI,cAAc,EAAE,CAAC;QAC9B,KAAK,IAAI;YACP,OAAO,IAAI,gBAAgB,EAAE,CAAC;QAChC,KAAK,IAAI;YACP,OAAO,IAAI,eAAe,EAAE,CAAC;QAC/B,KAAK,IAAI,CAAC;QACV;YACE,OAAO,IAAI,eAAe,EAAE,CAAC;IACjC,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,OAAO,sBAAsB;IACzB,SAAS,GAAoC,IAAI,GAAG,EAAE,CAAC;IAE/D;QACE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,cAAc,EAAE,CAAC,CAAC;QAC/C,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,eAAe,EAAE,CAAC,CAAC;QAChD,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,gBAAgB,EAAE,CAAC,CAAC;QACjD,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,eAAe,EAAE,CAAC,CAAC;IAClD,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,IAAY;QACnB,MAAM,QAAQ,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,IAAI,eAAe,EAAE,CAAC;QACvE,OAAO;YACL,QAAQ;YACR,MAAM,EAAE,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC;SAChC,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,OAAO,CAAC,IAAY;QAQlB,MAAM,QAAQ,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,IAAI,eAAe,EAAE,CAAC;QAEvE,OAAO;YACL,QAAQ;YACR,SAAS,EAAE,QAAQ,CAAC,cAAc,CAAC,IAAI,CAAC;YACxC,OAAO,EAAE,QAAQ,CAAC,YAAY,CAAC,IAAI,CAAC;YACpC,KAAK,EAAE,QAAQ,CAAC,UAAU,CAAC,IAAI,CAAC;YAChC,SAAS,EAAE,QAAQ,CAAC,cAAc,CAAC,IAAI,CAAC;YACxC,MAAM,EAAE,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC;SAChC,CAAC;IACJ,CAAC;CACF;AAED,mBAAmB;AACnB,MAAM,CAAC,MAAM,kBAAkB,GAAG,IAAI,sBAAsB,EAAE,CAAC"}
|
|
@@ -0,0 +1,109 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Local Embedding Service
|
|
3
|
+
* 로컬 초소형 모델을 사용한 임베딩 생성
|
|
4
|
+
*
|
|
5
|
+
* @module core/local-embedding
|
|
6
|
+
*/
|
|
7
|
+
/**
|
|
8
|
+
* 임베딩 결과
|
|
9
|
+
*/
|
|
10
|
+
interface EmbeddingResult {
|
|
11
|
+
text: string;
|
|
12
|
+
embedding: number[];
|
|
13
|
+
model: string;
|
|
14
|
+
processingTime: number;
|
|
15
|
+
}
|
|
16
|
+
/**
|
|
17
|
+
* 배치 임베딩 결과
|
|
18
|
+
*/
|
|
19
|
+
interface BatchEmbeddingResult {
|
|
20
|
+
results: EmbeddingResult[];
|
|
21
|
+
totalTime: number;
|
|
22
|
+
avgTime: number;
|
|
23
|
+
}
|
|
24
|
+
/**
|
|
25
|
+
* 유사도 결과
|
|
26
|
+
*/
|
|
27
|
+
interface SimilarityResult {
|
|
28
|
+
text1: string;
|
|
29
|
+
text2: string;
|
|
30
|
+
similarity: number;
|
|
31
|
+
method: 'cosine' | 'dot' | 'euclidean';
|
|
32
|
+
}
|
|
33
|
+
/**
|
|
34
|
+
* Local Embedding Service
|
|
35
|
+
*/
|
|
36
|
+
export declare class LocalEmbeddingService {
|
|
37
|
+
private extractor;
|
|
38
|
+
private modelName;
|
|
39
|
+
private isLoading;
|
|
40
|
+
private loadPromise;
|
|
41
|
+
private cache;
|
|
42
|
+
private cacheMaxSize;
|
|
43
|
+
private stats;
|
|
44
|
+
constructor(modelName?: string);
|
|
45
|
+
/**
|
|
46
|
+
* 모델 로드
|
|
47
|
+
*/
|
|
48
|
+
init(): Promise<void>;
|
|
49
|
+
/**
|
|
50
|
+
* 단일 텍스트 임베딩
|
|
51
|
+
*/
|
|
52
|
+
embed(text: string): Promise<EmbeddingResult>;
|
|
53
|
+
/**
|
|
54
|
+
* 배치 임베딩
|
|
55
|
+
*/
|
|
56
|
+
embedBatch(texts: string[]): Promise<BatchEmbeddingResult>;
|
|
57
|
+
/**
|
|
58
|
+
* 두 텍스트 간 유사도 계산
|
|
59
|
+
*/
|
|
60
|
+
similarity(text1: string, text2: string, method?: 'cosine' | 'dot' | 'euclidean'): Promise<SimilarityResult>;
|
|
61
|
+
/**
|
|
62
|
+
* 쿼리와 가장 유사한 텍스트 찾기
|
|
63
|
+
*/
|
|
64
|
+
findMostSimilar(query: string, candidates: string[], topK?: number): Promise<Array<{
|
|
65
|
+
text: string;
|
|
66
|
+
similarity: number;
|
|
67
|
+
}>>;
|
|
68
|
+
/**
|
|
69
|
+
* 코사인 유사도
|
|
70
|
+
*/
|
|
71
|
+
private cosineSimilarity;
|
|
72
|
+
/**
|
|
73
|
+
* 내적
|
|
74
|
+
*/
|
|
75
|
+
private dotProduct;
|
|
76
|
+
/**
|
|
77
|
+
* 유클리드 거리
|
|
78
|
+
*/
|
|
79
|
+
private euclideanDistance;
|
|
80
|
+
/**
|
|
81
|
+
* 캐시에 추가
|
|
82
|
+
*/
|
|
83
|
+
private addToCache;
|
|
84
|
+
/**
|
|
85
|
+
* 통계 반환
|
|
86
|
+
*/
|
|
87
|
+
getStats(): {
|
|
88
|
+
cacheSize: number;
|
|
89
|
+
cacheHitRate: number;
|
|
90
|
+
avgEmbeddingTime: number;
|
|
91
|
+
model: string;
|
|
92
|
+
isLoaded: boolean;
|
|
93
|
+
totalEmbeddings: number;
|
|
94
|
+
cacheHits: number;
|
|
95
|
+
totalTime: number;
|
|
96
|
+
modelLoadTime: number;
|
|
97
|
+
};
|
|
98
|
+
/**
|
|
99
|
+
* 캐시 클리어
|
|
100
|
+
*/
|
|
101
|
+
clearCache(): void;
|
|
102
|
+
/**
|
|
103
|
+
* 임베딩 차원 반환
|
|
104
|
+
*/
|
|
105
|
+
getDimension(): number;
|
|
106
|
+
}
|
|
107
|
+
export declare const localEmbedding: LocalEmbeddingService;
|
|
108
|
+
export {};
|
|
109
|
+
//# sourceMappingURL=local-embedding.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"local-embedding.d.ts","sourceRoot":"","sources":["../../src/core/local-embedding.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAYH;;GAEG;AACH,UAAU,eAAe;IACvB,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,EAAE,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,cAAc,EAAE,MAAM,CAAC;CACxB;AAED;;GAEG;AACH,UAAU,oBAAoB;IAC5B,OAAO,EAAE,eAAe,EAAE,CAAC;IAC3B,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED;;GAEG;AACH,UAAU,gBAAgB;IACxB,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE,QAAQ,GAAG,KAAK,GAAG,WAAW,CAAC;CACxC;AAED;;GAEG;AACH,qBAAa,qBAAqB;IAChC,OAAO,CAAC,SAAS,CAAiC;IAClD,OAAO,CAAC,SAAS,CAAS;IAC1B,OAAO,CAAC,SAAS,CAAkB;IACnC,OAAO,CAAC,WAAW,CAA8B;IAGjD,OAAO,CAAC,KAAK,CAAoC;IACjD,OAAO,CAAC,YAAY,CAAiB;IAGrC,OAAO,CAAC,KAAK,CAKX;gBAEU,SAAS,GAAE,MAAkC;IAIzD;;OAEG;IACG,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC;IA0B3B;;OAEG;IACG,KAAK,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,eAAe,CAAC;IAwCnD;;OAEG;IACG,UAAU,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,oBAAoB,CAAC;IAqBhE;;OAEG;IACG,UAAU,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,GAAE,QAAQ,GAAG,KAAK,GAAG,WAAsB,GAAG,OAAO,CAAC,gBAAgB,CAAC;IAuB5H;;OAEG;IACG,eAAe,CAAC,KAAK,EAAE,MAAM,EAAE,UAAU,EAAE,MAAM,EAAE,EAAE,IAAI,GAAE,MAAU,GAAG,OAAO,CAAC,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAclI;;OAEG;IACH,OAAO,CAAC,gBAAgB;IAcxB;;OAEG;IACH,OAAO,CAAC,UAAU;IAIlB;;OAEG;IACH,OAAO,CAAC,iBAAiB;IAIzB;;OAEG;IACH,OAAO,CAAC,UAAU;IASlB;;OAEG;IACH,QAAQ;;;;;;;;;;;IAeR;;OAEG;IACH,UAAU,IAAI,IAAI;IAIlB;;OAEG;IACH,YAAY,IAAI,MAAM;CAGvB;AAGD,eAAO,MAAM,cAAc,uBAA8B,CAAC"}
|