ocr-provenance-mcp 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.
Potentially problematic release.
This version of ocr-provenance-mcp might be problematic. Click here for more details.
- package/.env.example +55 -0
- package/LICENSE +78 -0
- package/README.md +1154 -0
- package/dist/bin-http.d.ts +24 -0
- package/dist/bin-http.d.ts.map +1 -0
- package/dist/bin-http.js +275 -0
- package/dist/bin-http.js.map +1 -0
- package/dist/bin-setup.d.ts +11 -0
- package/dist/bin-setup.d.ts.map +1 -0
- package/dist/bin-setup.js +610 -0
- package/dist/bin-setup.js.map +1 -0
- package/dist/bin.d.ts +16 -0
- package/dist/bin.d.ts.map +1 -0
- package/dist/bin.js +16 -0
- package/dist/bin.js.map +1 -0
- package/dist/index.d.ts +13 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +90 -0
- package/dist/index.js.map +1 -0
- package/dist/models/chunk.d.ts +136 -0
- package/dist/models/chunk.d.ts.map +1 -0
- package/dist/models/chunk.js +27 -0
- package/dist/models/chunk.js.map +1 -0
- package/dist/models/cluster.d.ts +79 -0
- package/dist/models/cluster.d.ts.map +1 -0
- package/dist/models/cluster.js +10 -0
- package/dist/models/cluster.js.map +1 -0
- package/dist/models/comparison.d.ts +62 -0
- package/dist/models/comparison.d.ts.map +1 -0
- package/dist/models/comparison.js +8 -0
- package/dist/models/comparison.js.map +1 -0
- package/dist/models/document.d.ts +104 -0
- package/dist/models/document.d.ts.map +1 -0
- package/dist/models/document.js +15 -0
- package/dist/models/document.js.map +1 -0
- package/dist/models/embedding.d.ts +87 -0
- package/dist/models/embedding.d.ts.map +1 -0
- package/dist/models/embedding.js +23 -0
- package/dist/models/embedding.js.map +1 -0
- package/dist/models/extraction.d.ts +15 -0
- package/dist/models/extraction.d.ts.map +1 -0
- package/dist/models/extraction.js +2 -0
- package/dist/models/extraction.js.map +1 -0
- package/dist/models/form-fill.d.ts +23 -0
- package/dist/models/form-fill.d.ts.map +1 -0
- package/dist/models/form-fill.js +2 -0
- package/dist/models/form-fill.js.map +1 -0
- package/dist/models/image.d.ts +177 -0
- package/dist/models/image.d.ts.map +1 -0
- package/dist/models/image.js +8 -0
- package/dist/models/image.js.map +1 -0
- package/dist/models/index.d.ts +14 -0
- package/dist/models/index.d.ts.map +1 -0
- package/dist/models/index.js +22 -0
- package/dist/models/index.js.map +1 -0
- package/dist/models/provenance.d.ts +174 -0
- package/dist/models/provenance.d.ts.map +1 -0
- package/dist/models/provenance.js +53 -0
- package/dist/models/provenance.js.map +1 -0
- package/dist/models/uploaded-file.d.ts +20 -0
- package/dist/models/uploaded-file.d.ts.map +1 -0
- package/dist/models/uploaded-file.js +2 -0
- package/dist/models/uploaded-file.js.map +1 -0
- package/dist/server/errors.d.ts +93 -0
- package/dist/server/errors.d.ts.map +1 -0
- package/dist/server/errors.js +256 -0
- package/dist/server/errors.js.map +1 -0
- package/dist/server/events.d.ts +36 -0
- package/dist/server/events.d.ts.map +1 -0
- package/dist/server/events.js +48 -0
- package/dist/server/events.js.map +1 -0
- package/dist/server/permissions.d.ts +26 -0
- package/dist/server/permissions.d.ts.map +1 -0
- package/dist/server/permissions.js +194 -0
- package/dist/server/permissions.js.map +1 -0
- package/dist/server/register-tools.d.ts +25 -0
- package/dist/server/register-tools.d.ts.map +1 -0
- package/dist/server/register-tools.js +102 -0
- package/dist/server/register-tools.js.map +1 -0
- package/dist/server/startup.d.ts +16 -0
- package/dist/server/startup.d.ts.map +1 -0
- package/dist/server/startup.js +37 -0
- package/dist/server/startup.js.map +1 -0
- package/dist/server/state.d.ts +166 -0
- package/dist/server/state.d.ts.map +1 -0
- package/dist/server/state.js +424 -0
- package/dist/server/state.js.map +1 -0
- package/dist/server/transports/http-transport.d.ts +37 -0
- package/dist/server/transports/http-transport.d.ts.map +1 -0
- package/dist/server/transports/http-transport.js +204 -0
- package/dist/server/transports/http-transport.js.map +1 -0
- package/dist/server/transports/index.d.ts +9 -0
- package/dist/server/transports/index.d.ts.map +1 -0
- package/dist/server/transports/index.js +9 -0
- package/dist/server/transports/index.js.map +1 -0
- package/dist/server/transports/session-manager.d.ts +40 -0
- package/dist/server/transports/session-manager.d.ts.map +1 -0
- package/dist/server/transports/session-manager.js +74 -0
- package/dist/server/transports/session-manager.js.map +1 -0
- package/dist/server/types.d.ts +82 -0
- package/dist/server/types.d.ts.map +1 -0
- package/dist/server/types.js +14 -0
- package/dist/server/types.js.map +1 -0
- package/dist/services/audit.d.ts +26 -0
- package/dist/services/audit.d.ts.map +1 -0
- package/dist/services/audit.js +43 -0
- package/dist/services/audit.js.map +1 -0
- package/dist/services/chunking/chunk-deduplicator.d.ts +33 -0
- package/dist/services/chunking/chunk-deduplicator.d.ts.map +1 -0
- package/dist/services/chunking/chunk-deduplicator.js +46 -0
- package/dist/services/chunking/chunk-deduplicator.js.map +1 -0
- package/dist/services/chunking/chunk-merger.d.ts +26 -0
- package/dist/services/chunking/chunk-merger.d.ts.map +1 -0
- package/dist/services/chunking/chunk-merger.js +94 -0
- package/dist/services/chunking/chunk-merger.js.map +1 -0
- package/dist/services/chunking/chunker.d.ts +62 -0
- package/dist/services/chunking/chunker.d.ts.map +1 -0
- package/dist/services/chunking/chunker.js +566 -0
- package/dist/services/chunking/chunker.js.map +1 -0
- package/dist/services/chunking/heading-normalizer.d.ts +33 -0
- package/dist/services/chunking/heading-normalizer.d.ts.map +1 -0
- package/dist/services/chunking/heading-normalizer.js +101 -0
- package/dist/services/chunking/heading-normalizer.js.map +1 -0
- package/dist/services/chunking/json-block-analyzer.d.ts +163 -0
- package/dist/services/chunking/json-block-analyzer.d.ts.map +1 -0
- package/dist/services/chunking/json-block-analyzer.js +1033 -0
- package/dist/services/chunking/json-block-analyzer.js.map +1 -0
- package/dist/services/chunking/markdown-parser.d.ts +75 -0
- package/dist/services/chunking/markdown-parser.d.ts.map +1 -0
- package/dist/services/chunking/markdown-parser.js +428 -0
- package/dist/services/chunking/markdown-parser.js.map +1 -0
- package/dist/services/chunking/text-normalizer.d.ts +20 -0
- package/dist/services/chunking/text-normalizer.d.ts.map +1 -0
- package/dist/services/chunking/text-normalizer.js +36 -0
- package/dist/services/chunking/text-normalizer.js.map +1 -0
- package/dist/services/clm/contract-schemas.d.ts +36 -0
- package/dist/services/clm/contract-schemas.d.ts.map +1 -0
- package/dist/services/clm/contract-schemas.js +92 -0
- package/dist/services/clm/contract-schemas.js.map +1 -0
- package/dist/services/clm/summarization.d.ts +46 -0
- package/dist/services/clm/summarization.d.ts.map +1 -0
- package/dist/services/clm/summarization.js +61 -0
- package/dist/services/clm/summarization.js.map +1 -0
- package/dist/services/clustering/clustering-service.d.ts +58 -0
- package/dist/services/clustering/clustering-service.d.ts.map +1 -0
- package/dist/services/clustering/clustering-service.js +467 -0
- package/dist/services/clustering/clustering-service.js.map +1 -0
- package/dist/services/comparison/diff-service.d.ts +41 -0
- package/dist/services/comparison/diff-service.d.ts.map +1 -0
- package/dist/services/comparison/diff-service.js +120 -0
- package/dist/services/comparison/diff-service.js.map +1 -0
- package/dist/services/embedding/embedder.d.ts +55 -0
- package/dist/services/embedding/embedder.d.ts.map +1 -0
- package/dist/services/embedding/embedder.js +202 -0
- package/dist/services/embedding/embedder.js.map +1 -0
- package/dist/services/embedding/nomic.d.ts +67 -0
- package/dist/services/embedding/nomic.d.ts.map +1 -0
- package/dist/services/embedding/nomic.js +280 -0
- package/dist/services/embedding/nomic.js.map +1 -0
- package/dist/services/gemini/circuit-breaker.d.ts +106 -0
- package/dist/services/gemini/circuit-breaker.d.ts.map +1 -0
- package/dist/services/gemini/circuit-breaker.js +237 -0
- package/dist/services/gemini/circuit-breaker.js.map +1 -0
- package/dist/services/gemini/client.d.ts +173 -0
- package/dist/services/gemini/client.d.ts.map +1 -0
- package/dist/services/gemini/client.js +483 -0
- package/dist/services/gemini/client.js.map +1 -0
- package/dist/services/gemini/config.d.ts +116 -0
- package/dist/services/gemini/config.d.ts.map +1 -0
- package/dist/services/gemini/config.js +118 -0
- package/dist/services/gemini/config.js.map +1 -0
- package/dist/services/gemini/index.d.ts +9 -0
- package/dist/services/gemini/index.d.ts.map +1 -0
- package/dist/services/gemini/index.js +13 -0
- package/dist/services/gemini/index.js.map +1 -0
- package/dist/services/gemini/rate-limiter.d.ts +62 -0
- package/dist/services/gemini/rate-limiter.d.ts.map +1 -0
- package/dist/services/gemini/rate-limiter.js +120 -0
- package/dist/services/gemini/rate-limiter.js.map +1 -0
- package/dist/services/images/extractor.d.ts +88 -0
- package/dist/services/images/extractor.d.ts.map +1 -0
- package/dist/services/images/extractor.js +340 -0
- package/dist/services/images/extractor.js.map +1 -0
- package/dist/services/images/optimizer.d.ts +130 -0
- package/dist/services/images/optimizer.d.ts.map +1 -0
- package/dist/services/images/optimizer.js +228 -0
- package/dist/services/images/optimizer.js.map +1 -0
- package/dist/services/ocr/datalab.d.ts +64 -0
- package/dist/services/ocr/datalab.d.ts.map +1 -0
- package/dist/services/ocr/datalab.js +425 -0
- package/dist/services/ocr/datalab.js.map +1 -0
- package/dist/services/ocr/errors.d.ts +38 -0
- package/dist/services/ocr/errors.d.ts.map +1 -0
- package/dist/services/ocr/errors.js +83 -0
- package/dist/services/ocr/errors.js.map +1 -0
- package/dist/services/ocr/file-manager.d.ts +76 -0
- package/dist/services/ocr/file-manager.d.ts.map +1 -0
- package/dist/services/ocr/file-manager.js +238 -0
- package/dist/services/ocr/file-manager.js.map +1 -0
- package/dist/services/ocr/form-fill.d.ts +48 -0
- package/dist/services/ocr/form-fill.d.ts.map +1 -0
- package/dist/services/ocr/form-fill.js +213 -0
- package/dist/services/ocr/form-fill.js.map +1 -0
- package/dist/services/ocr/processor.d.ts +95 -0
- package/dist/services/ocr/processor.d.ts.map +1 -0
- package/dist/services/ocr/processor.js +259 -0
- package/dist/services/ocr/processor.js.map +1 -0
- package/dist/services/provenance/agent-metadata.d.ts +82 -0
- package/dist/services/provenance/agent-metadata.d.ts.map +1 -0
- package/dist/services/provenance/agent-metadata.js +106 -0
- package/dist/services/provenance/agent-metadata.js.map +1 -0
- package/dist/services/provenance/chain-hash.d.ts +57 -0
- package/dist/services/provenance/chain-hash.d.ts.map +1 -0
- package/dist/services/provenance/chain-hash.js +131 -0
- package/dist/services/provenance/chain-hash.js.map +1 -0
- package/dist/services/provenance/exporter.d.ts +202 -0
- package/dist/services/provenance/exporter.d.ts.map +1 -0
- package/dist/services/provenance/exporter.js +457 -0
- package/dist/services/provenance/exporter.js.map +1 -0
- package/dist/services/provenance/index.d.ts +15 -0
- package/dist/services/provenance/index.d.ts.map +1 -0
- package/dist/services/provenance/index.js +17 -0
- package/dist/services/provenance/index.js.map +1 -0
- package/dist/services/provenance/tracker.d.ts +138 -0
- package/dist/services/provenance/tracker.d.ts.map +1 -0
- package/dist/services/provenance/tracker.js +293 -0
- package/dist/services/provenance/tracker.js.map +1 -0
- package/dist/services/provenance/verifier.d.ts +153 -0
- package/dist/services/provenance/verifier.d.ts.map +1 -0
- package/dist/services/provenance/verifier.js +536 -0
- package/dist/services/provenance/verifier.js.map +1 -0
- package/dist/services/python-pool.d.ts +70 -0
- package/dist/services/python-pool.d.ts.map +1 -0
- package/dist/services/python-pool.js +265 -0
- package/dist/services/python-pool.js.map +1 -0
- package/dist/services/search/bm25.d.ts +180 -0
- package/dist/services/search/bm25.d.ts.map +1 -0
- package/dist/services/search/bm25.js +656 -0
- package/dist/services/search/bm25.js.map +1 -0
- package/dist/services/search/fusion.d.ts +103 -0
- package/dist/services/search/fusion.d.ts.map +1 -0
- package/dist/services/search/fusion.js +122 -0
- package/dist/services/search/fusion.js.map +1 -0
- package/dist/services/search/local-reranker.d.ts +30 -0
- package/dist/services/search/local-reranker.d.ts.map +1 -0
- package/dist/services/search/local-reranker.js +123 -0
- package/dist/services/search/local-reranker.js.map +1 -0
- package/dist/services/search/quality.d.ts +11 -0
- package/dist/services/search/quality.d.ts.map +1 -0
- package/dist/services/search/quality.js +17 -0
- package/dist/services/search/quality.js.map +1 -0
- package/dist/services/search/query-classifier.d.ts +34 -0
- package/dist/services/search/query-classifier.d.ts.map +1 -0
- package/dist/services/search/query-classifier.js +114 -0
- package/dist/services/search/query-classifier.js.map +1 -0
- package/dist/services/search/query-expander.d.ts +73 -0
- package/dist/services/search/query-expander.d.ts.map +1 -0
- package/dist/services/search/query-expander.js +281 -0
- package/dist/services/search/query-expander.js.map +1 -0
- package/dist/services/search/reranker.d.ts +44 -0
- package/dist/services/search/reranker.d.ts.map +1 -0
- package/dist/services/search/reranker.js +101 -0
- package/dist/services/search/reranker.js.map +1 -0
- package/dist/services/storage/database/annotation-operations.d.ts +113 -0
- package/dist/services/storage/database/annotation-operations.d.ts.map +1 -0
- package/dist/services/storage/database/annotation-operations.js +177 -0
- package/dist/services/storage/database/annotation-operations.js.map +1 -0
- package/dist/services/storage/database/approval-operations.d.ts +132 -0
- package/dist/services/storage/database/approval-operations.d.ts.map +1 -0
- package/dist/services/storage/database/approval-operations.js +206 -0
- package/dist/services/storage/database/approval-operations.js.map +1 -0
- package/dist/services/storage/database/chunk-operations.d.ts +132 -0
- package/dist/services/storage/database/chunk-operations.d.ts.map +1 -0
- package/dist/services/storage/database/chunk-operations.js +306 -0
- package/dist/services/storage/database/chunk-operations.js.map +1 -0
- package/dist/services/storage/database/cluster-operations.d.ts +97 -0
- package/dist/services/storage/database/cluster-operations.d.ts.map +1 -0
- package/dist/services/storage/database/cluster-operations.js +258 -0
- package/dist/services/storage/database/cluster-operations.js.map +1 -0
- package/dist/services/storage/database/comparison-operations.d.ts +41 -0
- package/dist/services/storage/database/comparison-operations.d.ts.map +1 -0
- package/dist/services/storage/database/comparison-operations.js +65 -0
- package/dist/services/storage/database/comparison-operations.js.map +1 -0
- package/dist/services/storage/database/converters.d.ts +36 -0
- package/dist/services/storage/database/converters.d.ts.map +1 -0
- package/dist/services/storage/database/converters.js +244 -0
- package/dist/services/storage/database/converters.js.map +1 -0
- package/dist/services/storage/database/document-operations.d.ts +145 -0
- package/dist/services/storage/database/document-operations.d.ts.map +1 -0
- package/dist/services/storage/database/document-operations.js +498 -0
- package/dist/services/storage/database/document-operations.js.map +1 -0
- package/dist/services/storage/database/embedding-operations.d.ts +130 -0
- package/dist/services/storage/database/embedding-operations.d.ts.map +1 -0
- package/dist/services/storage/database/embedding-operations.js +315 -0
- package/dist/services/storage/database/embedding-operations.js.map +1 -0
- package/dist/services/storage/database/extraction-operations.d.ts +47 -0
- package/dist/services/storage/database/extraction-operations.d.ts.map +1 -0
- package/dist/services/storage/database/extraction-operations.js +85 -0
- package/dist/services/storage/database/extraction-operations.js.map +1 -0
- package/dist/services/storage/database/form-fill-operations.d.ts +58 -0
- package/dist/services/storage/database/form-fill-operations.d.ts.map +1 -0
- package/dist/services/storage/database/form-fill-operations.js +116 -0
- package/dist/services/storage/database/form-fill-operations.js.map +1 -0
- package/dist/services/storage/database/helpers.d.ts +29 -0
- package/dist/services/storage/database/helpers.d.ts.map +1 -0
- package/dist/services/storage/database/helpers.js +55 -0
- package/dist/services/storage/database/helpers.js.map +1 -0
- package/dist/services/storage/database/image-operations.d.ts +202 -0
- package/dist/services/storage/database/image-operations.d.ts.map +1 -0
- package/dist/services/storage/database/image-operations.js +484 -0
- package/dist/services/storage/database/image-operations.js.map +1 -0
- package/dist/services/storage/database/index.d.ts +13 -0
- package/dist/services/storage/database/index.d.ts.map +1 -0
- package/dist/services/storage/database/index.js +16 -0
- package/dist/services/storage/database/index.js.map +1 -0
- package/dist/services/storage/database/lock-operations.d.ts +59 -0
- package/dist/services/storage/database/lock-operations.d.ts.map +1 -0
- package/dist/services/storage/database/lock-operations.js +89 -0
- package/dist/services/storage/database/lock-operations.js.map +1 -0
- package/dist/services/storage/database/obligation-operations.d.ts +88 -0
- package/dist/services/storage/database/obligation-operations.d.ts.map +1 -0
- package/dist/services/storage/database/obligation-operations.js +206 -0
- package/dist/services/storage/database/obligation-operations.js.map +1 -0
- package/dist/services/storage/database/ocr-operations.d.ts +33 -0
- package/dist/services/storage/database/ocr-operations.d.ts.map +1 -0
- package/dist/services/storage/database/ocr-operations.js +70 -0
- package/dist/services/storage/database/ocr-operations.js.map +1 -0
- package/dist/services/storage/database/playbook-operations.d.ts +72 -0
- package/dist/services/storage/database/playbook-operations.d.ts.map +1 -0
- package/dist/services/storage/database/playbook-operations.js +247 -0
- package/dist/services/storage/database/playbook-operations.js.map +1 -0
- package/dist/services/storage/database/provenance-operations.d.ts +112 -0
- package/dist/services/storage/database/provenance-operations.d.ts.map +1 -0
- package/dist/services/storage/database/provenance-operations.js +251 -0
- package/dist/services/storage/database/provenance-operations.js.map +1 -0
- package/dist/services/storage/database/service.d.ts +142 -0
- package/dist/services/storage/database/service.d.ts.map +1 -0
- package/dist/services/storage/database/service.js +310 -0
- package/dist/services/storage/database/service.js.map +1 -0
- package/dist/services/storage/database/static-operations.d.ts +30 -0
- package/dist/services/storage/database/static-operations.d.ts.map +1 -0
- package/dist/services/storage/database/static-operations.js +218 -0
- package/dist/services/storage/database/static-operations.js.map +1 -0
- package/dist/services/storage/database/stats-operations.d.ts +101 -0
- package/dist/services/storage/database/stats-operations.d.ts.map +1 -0
- package/dist/services/storage/database/stats-operations.js +394 -0
- package/dist/services/storage/database/stats-operations.js.map +1 -0
- package/dist/services/storage/database/tag-operations.d.ts +76 -0
- package/dist/services/storage/database/tag-operations.d.ts.map +1 -0
- package/dist/services/storage/database/tag-operations.js +178 -0
- package/dist/services/storage/database/tag-operations.js.map +1 -0
- package/dist/services/storage/database/types.d.ts +286 -0
- package/dist/services/storage/database/types.d.ts.map +1 -0
- package/dist/services/storage/database/types.js +39 -0
- package/dist/services/storage/database/types.js.map +1 -0
- package/dist/services/storage/database/upload-operations.d.ts +71 -0
- package/dist/services/storage/database/upload-operations.d.ts.map +1 -0
- package/dist/services/storage/database/upload-operations.js +124 -0
- package/dist/services/storage/database/upload-operations.js.map +1 -0
- package/dist/services/storage/database/user-operations.d.ts +102 -0
- package/dist/services/storage/database/user-operations.d.ts.map +1 -0
- package/dist/services/storage/database/user-operations.js +151 -0
- package/dist/services/storage/database/user-operations.js.map +1 -0
- package/dist/services/storage/database/workflow-operations.d.ts +98 -0
- package/dist/services/storage/database/workflow-operations.d.ts.map +1 -0
- package/dist/services/storage/database/workflow-operations.js +157 -0
- package/dist/services/storage/database/workflow-operations.js.map +1 -0
- package/dist/services/storage/database.d.ts +16 -0
- package/dist/services/storage/database.d.ts.map +1 -0
- package/dist/services/storage/database.js +15 -0
- package/dist/services/storage/database.js.map +1 -0
- package/dist/services/storage/index.d.ts +10 -0
- package/dist/services/storage/index.d.ts.map +1 -0
- package/dist/services/storage/index.js +10 -0
- package/dist/services/storage/index.js.map +1 -0
- package/dist/services/storage/migrations/index.d.ts +16 -0
- package/dist/services/storage/migrations/index.d.ts.map +1 -0
- package/dist/services/storage/migrations/index.js +20 -0
- package/dist/services/storage/migrations/index.js.map +1 -0
- package/dist/services/storage/migrations/operations.d.ts +40 -0
- package/dist/services/storage/migrations/operations.d.ts.map +1 -0
- package/dist/services/storage/migrations/operations.js +2910 -0
- package/dist/services/storage/migrations/operations.js.map +1 -0
- package/dist/services/storage/migrations/schema-definitions.d.ts +306 -0
- package/dist/services/storage/migrations/schema-definitions.d.ts.map +1 -0
- package/dist/services/storage/migrations/schema-definitions.js +1006 -0
- package/dist/services/storage/migrations/schema-definitions.js.map +1 -0
- package/dist/services/storage/migrations/schema-helpers.d.ts +50 -0
- package/dist/services/storage/migrations/schema-helpers.d.ts.map +1 -0
- package/dist/services/storage/migrations/schema-helpers.js +176 -0
- package/dist/services/storage/migrations/schema-helpers.js.map +1 -0
- package/dist/services/storage/migrations/types.d.ts +15 -0
- package/dist/services/storage/migrations/types.d.ts.map +1 -0
- package/dist/services/storage/migrations/types.js +21 -0
- package/dist/services/storage/migrations/types.js.map +1 -0
- package/dist/services/storage/migrations/verification.d.ts +20 -0
- package/dist/services/storage/migrations/verification.d.ts.map +1 -0
- package/dist/services/storage/migrations/verification.js +78 -0
- package/dist/services/storage/migrations/verification.js.map +1 -0
- package/dist/services/storage/migrations.d.ts +16 -0
- package/dist/services/storage/migrations.d.ts.map +1 -0
- package/dist/services/storage/migrations.js +17 -0
- package/dist/services/storage/migrations.js.map +1 -0
- package/dist/services/storage/types.d.ts +12 -0
- package/dist/services/storage/types.d.ts.map +1 -0
- package/dist/services/storage/types.js +5 -0
- package/dist/services/storage/types.js.map +1 -0
- package/dist/services/storage/vector.d.ts +208 -0
- package/dist/services/storage/vector.d.ts.map +1 -0
- package/dist/services/storage/vector.js +526 -0
- package/dist/services/storage/vector.js.map +1 -0
- package/dist/services/vlm/pipeline.d.ts +194 -0
- package/dist/services/vlm/pipeline.d.ts.map +1 -0
- package/dist/services/vlm/pipeline.js +800 -0
- package/dist/services/vlm/pipeline.js.map +1 -0
- package/dist/services/vlm/prompts.d.ts +171 -0
- package/dist/services/vlm/prompts.d.ts.map +1 -0
- package/dist/services/vlm/prompts.js +229 -0
- package/dist/services/vlm/prompts.js.map +1 -0
- package/dist/services/vlm/service.d.ts +174 -0
- package/dist/services/vlm/service.d.ts.map +1 -0
- package/dist/services/vlm/service.js +256 -0
- package/dist/services/vlm/service.js.map +1 -0
- package/dist/services/webhook-delivery.d.ts +4 -0
- package/dist/services/webhook-delivery.d.ts.map +1 -0
- package/dist/services/webhook-delivery.js +140 -0
- package/dist/services/webhook-delivery.js.map +1 -0
- package/dist/tools/chunks.d.ts +19 -0
- package/dist/tools/chunks.d.ts.map +1 -0
- package/dist/tools/chunks.js +392 -0
- package/dist/tools/chunks.js.map +1 -0
- package/dist/tools/clm.d.ts +16 -0
- package/dist/tools/clm.d.ts.map +1 -0
- package/dist/tools/clm.js +668 -0
- package/dist/tools/clm.js.map +1 -0
- package/dist/tools/clustering.d.ts +13 -0
- package/dist/tools/clustering.d.ts.map +1 -0
- package/dist/tools/clustering.js +498 -0
- package/dist/tools/clustering.js.map +1 -0
- package/dist/tools/collaboration.d.ts +15 -0
- package/dist/tools/collaboration.d.ts.map +1 -0
- package/dist/tools/collaboration.js +516 -0
- package/dist/tools/collaboration.js.map +1 -0
- package/dist/tools/comparison.d.ts +13 -0
- package/dist/tools/comparison.d.ts.map +1 -0
- package/dist/tools/comparison.js +735 -0
- package/dist/tools/comparison.js.map +1 -0
- package/dist/tools/compliance.d.ts +15 -0
- package/dist/tools/compliance.d.ts.map +1 -0
- package/dist/tools/compliance.js +640 -0
- package/dist/tools/compliance.js.map +1 -0
- package/dist/tools/config.d.ts +19 -0
- package/dist/tools/config.d.ts.map +1 -0
- package/dist/tools/config.js +213 -0
- package/dist/tools/config.js.map +1 -0
- package/dist/tools/database.d.ts +62 -0
- package/dist/tools/database.d.ts.map +1 -0
- package/dist/tools/database.js +288 -0
- package/dist/tools/database.js.map +1 -0
- package/dist/tools/documents.d.ts +61 -0
- package/dist/tools/documents.d.ts.map +1 -0
- package/dist/tools/documents.js +1624 -0
- package/dist/tools/documents.js.map +1 -0
- package/dist/tools/embeddings.d.ts +14 -0
- package/dist/tools/embeddings.d.ts.map +1 -0
- package/dist/tools/embeddings.js +626 -0
- package/dist/tools/embeddings.js.map +1 -0
- package/dist/tools/evaluation.d.ts +25 -0
- package/dist/tools/evaluation.d.ts.map +1 -0
- package/dist/tools/evaluation.js +523 -0
- package/dist/tools/evaluation.js.map +1 -0
- package/dist/tools/events.d.ts +16 -0
- package/dist/tools/events.d.ts.map +1 -0
- package/dist/tools/events.js +493 -0
- package/dist/tools/events.js.map +1 -0
- package/dist/tools/extraction-structured.d.ts +13 -0
- package/dist/tools/extraction-structured.d.ts.map +1 -0
- package/dist/tools/extraction-structured.js +390 -0
- package/dist/tools/extraction-structured.js.map +1 -0
- package/dist/tools/extraction.d.ts +24 -0
- package/dist/tools/extraction.d.ts.map +1 -0
- package/dist/tools/extraction.js +424 -0
- package/dist/tools/extraction.js.map +1 -0
- package/dist/tools/file-management.d.ts +14 -0
- package/dist/tools/file-management.d.ts.map +1 -0
- package/dist/tools/file-management.js +523 -0
- package/dist/tools/file-management.js.map +1 -0
- package/dist/tools/form-fill.d.ts +13 -0
- package/dist/tools/form-fill.d.ts.map +1 -0
- package/dist/tools/form-fill.js +250 -0
- package/dist/tools/form-fill.js.map +1 -0
- package/dist/tools/health.d.ts +19 -0
- package/dist/tools/health.d.ts.map +1 -0
- package/dist/tools/health.js +229 -0
- package/dist/tools/health.js.map +1 -0
- package/dist/tools/images.d.ts +54 -0
- package/dist/tools/images.d.ts.map +1 -0
- package/dist/tools/images.js +787 -0
- package/dist/tools/images.js.map +1 -0
- package/dist/tools/ingestion.d.ts +94 -0
- package/dist/tools/ingestion.d.ts.map +1 -0
- package/dist/tools/ingestion.js +1659 -0
- package/dist/tools/ingestion.js.map +1 -0
- package/dist/tools/intelligence.d.ts +18 -0
- package/dist/tools/intelligence.d.ts.map +1 -0
- package/dist/tools/intelligence.js +1039 -0
- package/dist/tools/intelligence.js.map +1 -0
- package/dist/tools/provenance.d.ts +51 -0
- package/dist/tools/provenance.d.ts.map +1 -0
- package/dist/tools/provenance.js +691 -0
- package/dist/tools/provenance.js.map +1 -0
- package/dist/tools/reports.d.ts +41 -0
- package/dist/tools/reports.d.ts.map +1 -0
- package/dist/tools/reports.js +1394 -0
- package/dist/tools/reports.js.map +1 -0
- package/dist/tools/search.d.ts +35 -0
- package/dist/tools/search.d.ts.map +1 -0
- package/dist/tools/search.js +2528 -0
- package/dist/tools/search.js.map +1 -0
- package/dist/tools/shared.d.ts +52 -0
- package/dist/tools/shared.d.ts.map +1 -0
- package/dist/tools/shared.js +54 -0
- package/dist/tools/shared.js.map +1 -0
- package/dist/tools/tags.d.ts +15 -0
- package/dist/tools/tags.d.ts.map +1 -0
- package/dist/tools/tags.js +287 -0
- package/dist/tools/tags.js.map +1 -0
- package/dist/tools/timeline.d.ts +15 -0
- package/dist/tools/timeline.d.ts.map +1 -0
- package/dist/tools/timeline.js +14 -0
- package/dist/tools/timeline.js.map +1 -0
- package/dist/tools/users.d.ts +14 -0
- package/dist/tools/users.d.ts.map +1 -0
- package/dist/tools/users.js +257 -0
- package/dist/tools/users.js.map +1 -0
- package/dist/tools/vlm.d.ts +40 -0
- package/dist/tools/vlm.d.ts.map +1 -0
- package/dist/tools/vlm.js +475 -0
- package/dist/tools/vlm.js.map +1 -0
- package/dist/tools/workflow.d.ts +16 -0
- package/dist/tools/workflow.d.ts.map +1 -0
- package/dist/tools/workflow.js +495 -0
- package/dist/tools/workflow.js.map +1 -0
- package/dist/utils/backoff.d.ts +53 -0
- package/dist/utils/backoff.d.ts.map +1 -0
- package/dist/utils/backoff.js +78 -0
- package/dist/utils/backoff.js.map +1 -0
- package/dist/utils/config-persistence.d.ts +33 -0
- package/dist/utils/config-persistence.d.ts.map +1 -0
- package/dist/utils/config-persistence.js +61 -0
- package/dist/utils/config-persistence.js.map +1 -0
- package/dist/utils/hash.d.ts +65 -0
- package/dist/utils/hash.d.ts.map +1 -0
- package/dist/utils/hash.js +146 -0
- package/dist/utils/hash.js.map +1 -0
- package/dist/utils/math.d.ts +21 -0
- package/dist/utils/math.d.ts.map +1 -0
- package/dist/utils/math.js +39 -0
- package/dist/utils/math.js.map +1 -0
- package/dist/utils/validation.d.ts +697 -0
- package/dist/utils/validation.d.ts.map +1 -0
- package/dist/utils/validation.js +529 -0
- package/dist/utils/validation.js.map +1 -0
- package/package.json +96 -0
- package/python/.gitkeep +0 -0
- package/python/__init__.py +104 -0
- package/python/clustering_worker.py +440 -0
- package/python/docx_image_extractor.py +524 -0
- package/python/embedding_worker.py +552 -0
- package/python/file_manager_worker.py +564 -0
- package/python/form_fill_worker.py +399 -0
- package/python/gpu_utils.py +582 -0
- package/python/image_extractor.py +317 -0
- package/python/image_optimizer.py +444 -0
- package/python/ocr_worker.py +712 -0
- package/python/pyproject.toml +76 -0
- package/python/requirements.txt +51 -0
- package/python/reranker_worker.py +87 -0
|
@@ -0,0 +1,340 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Image Extractor Service
|
|
3
|
+
*
|
|
4
|
+
* TypeScript wrapper for Python image extraction scripts.
|
|
5
|
+
* Extracts images from PDF and DOCX documents for VLM analysis.
|
|
6
|
+
*/
|
|
7
|
+
import { spawn } from 'child_process';
|
|
8
|
+
import * as path from 'path';
|
|
9
|
+
import * as fs from 'fs';
|
|
10
|
+
import { fileURLToPath } from 'url';
|
|
11
|
+
const __dirname = path.dirname(fileURLToPath(import.meta.url));
|
|
12
|
+
/** Max stderr accumulation: 10KB (matches nomic.ts pattern) */
|
|
13
|
+
const MAX_STDERR_LENGTH = 10_240;
|
|
14
|
+
/** Supported file types for image extraction */
|
|
15
|
+
const SUPPORTED_EXTRACTION_TYPES = new Set(['.pdf', '.docx']);
|
|
16
|
+
const DEFAULT_CONFIG = {
|
|
17
|
+
pythonPath: process.platform === 'win32' ? 'python' : 'python3',
|
|
18
|
+
timeout: 120000,
|
|
19
|
+
};
|
|
20
|
+
/**
|
|
21
|
+
* Service for extracting images from PDF documents
|
|
22
|
+
*/
|
|
23
|
+
export class ImageExtractor {
|
|
24
|
+
config;
|
|
25
|
+
pdfScriptPath;
|
|
26
|
+
docxScriptPath;
|
|
27
|
+
constructor(config = {}) {
|
|
28
|
+
this.config = { ...DEFAULT_CONFIG, ...config };
|
|
29
|
+
this.pdfScriptPath =
|
|
30
|
+
this.config.scriptPath || path.resolve(__dirname, '../../../python/image_extractor.py');
|
|
31
|
+
this.docxScriptPath = path.resolve(__dirname, '../../../python/docx_image_extractor.py');
|
|
32
|
+
}
|
|
33
|
+
/**
|
|
34
|
+
* Extract images from any supported document type.
|
|
35
|
+
* Routes to the correct extractor based on file extension.
|
|
36
|
+
*
|
|
37
|
+
* @param filePath - Path to the document file
|
|
38
|
+
* @param options - Extraction options
|
|
39
|
+
* @returns Promise<ExtractedImage[]> - Array of extracted images
|
|
40
|
+
* @throws Error if file type is unsupported or extraction fails
|
|
41
|
+
*/
|
|
42
|
+
async extractImages(filePath, options) {
|
|
43
|
+
const ext = path.extname(filePath).toLowerCase();
|
|
44
|
+
switch (ext) {
|
|
45
|
+
case '.pdf':
|
|
46
|
+
return this.extractFromPDF(filePath, options);
|
|
47
|
+
case '.docx':
|
|
48
|
+
return this.extractFromDOCX(filePath, options);
|
|
49
|
+
default:
|
|
50
|
+
throw new Error(`Unsupported file type for image extraction: '${ext}'. ` +
|
|
51
|
+
`Supported types: ${[...SUPPORTED_EXTRACTION_TYPES].join(', ')}`);
|
|
52
|
+
}
|
|
53
|
+
}
|
|
54
|
+
/**
|
|
55
|
+
* Check if a file type is supported for image extraction.
|
|
56
|
+
*/
|
|
57
|
+
static isSupported(filePath) {
|
|
58
|
+
return SUPPORTED_EXTRACTION_TYPES.has(path.extname(filePath).toLowerCase());
|
|
59
|
+
}
|
|
60
|
+
/**
|
|
61
|
+
* Extract images from a PDF document
|
|
62
|
+
*
|
|
63
|
+
* @param pdfPath - Path to the PDF file
|
|
64
|
+
* @param options - Extraction options
|
|
65
|
+
* @returns Promise<ExtractedImage[]> - Array of extracted images
|
|
66
|
+
* @throws Error if extraction fails
|
|
67
|
+
*/
|
|
68
|
+
async extractFromPDF(pdfPath, options) {
|
|
69
|
+
// Validate PDF exists
|
|
70
|
+
if (!fs.existsSync(pdfPath)) {
|
|
71
|
+
throw new Error(`PDF file not found: ${pdfPath}`);
|
|
72
|
+
}
|
|
73
|
+
// Validate script exists
|
|
74
|
+
if (!fs.existsSync(this.pdfScriptPath)) {
|
|
75
|
+
throw new Error(`Image extractor script not found: ${this.pdfScriptPath}. ` +
|
|
76
|
+
`Ensure python/image_extractor.py exists.`);
|
|
77
|
+
}
|
|
78
|
+
// Create output directory if it doesn't exist
|
|
79
|
+
if (!fs.existsSync(options.outputDir)) {
|
|
80
|
+
fs.mkdirSync(options.outputDir, { recursive: true });
|
|
81
|
+
}
|
|
82
|
+
const result = await this.runPythonExtractorScript(this.pdfScriptPath, pdfPath, options);
|
|
83
|
+
if (!result.success) {
|
|
84
|
+
throw new Error(`Image extraction failed: ${result.error}`);
|
|
85
|
+
}
|
|
86
|
+
if (result.warnings && result.warnings.length > 0) {
|
|
87
|
+
console.error(`[WARN] [ImageExtractor] Warnings during extraction: ${result.warnings.join('; ')}`);
|
|
88
|
+
}
|
|
89
|
+
return result.images;
|
|
90
|
+
}
|
|
91
|
+
/**
|
|
92
|
+
* Extract images from a DOCX document
|
|
93
|
+
*
|
|
94
|
+
* @param docxPath - Path to the DOCX file
|
|
95
|
+
* @param options - Extraction options
|
|
96
|
+
* @returns Promise<ExtractedImage[]> - Array of extracted images
|
|
97
|
+
* @throws Error if extraction fails
|
|
98
|
+
*/
|
|
99
|
+
async extractFromDOCX(docxPath, options) {
|
|
100
|
+
if (!fs.existsSync(docxPath)) {
|
|
101
|
+
throw new Error(`DOCX file not found: ${docxPath}`);
|
|
102
|
+
}
|
|
103
|
+
if (!fs.existsSync(this.docxScriptPath)) {
|
|
104
|
+
throw new Error(`DOCX image extractor script not found: ${this.docxScriptPath}. ` +
|
|
105
|
+
`Ensure python/docx_image_extractor.py exists.`);
|
|
106
|
+
}
|
|
107
|
+
if (!fs.existsSync(options.outputDir)) {
|
|
108
|
+
fs.mkdirSync(options.outputDir, { recursive: true });
|
|
109
|
+
}
|
|
110
|
+
const result = await this.runPythonExtractorScript(this.docxScriptPath, docxPath, options);
|
|
111
|
+
if (!result.success) {
|
|
112
|
+
throw new Error(`DOCX image extraction failed: ${result.error}`);
|
|
113
|
+
}
|
|
114
|
+
if (result.warnings && result.warnings.length > 0) {
|
|
115
|
+
console.error(`[WARN] [ImageExtractor] DOCX extraction warnings: ${result.warnings.join('; ')}`);
|
|
116
|
+
}
|
|
117
|
+
return result.images;
|
|
118
|
+
}
|
|
119
|
+
/**
|
|
120
|
+
* Check if the Python environment is properly configured
|
|
121
|
+
*
|
|
122
|
+
* @returns Promise<boolean> - True if Python and dependencies are available
|
|
123
|
+
*/
|
|
124
|
+
async checkEnvironment() {
|
|
125
|
+
const missingDeps = [];
|
|
126
|
+
// Check Python version
|
|
127
|
+
let pythonVersion;
|
|
128
|
+
try {
|
|
129
|
+
pythonVersion = await this.runCommand(this.config.pythonPath, ['--version']);
|
|
130
|
+
}
|
|
131
|
+
catch (error) {
|
|
132
|
+
console.error('[ImageExtractor] Python version check failed:', error instanceof Error ? error.message : String(error));
|
|
133
|
+
return {
|
|
134
|
+
available: false,
|
|
135
|
+
missingDependencies: ['python'],
|
|
136
|
+
};
|
|
137
|
+
}
|
|
138
|
+
// Check PyMuPDF
|
|
139
|
+
try {
|
|
140
|
+
await this.runCommand(this.config.pythonPath, ['-c', 'import fitz; print(fitz.version)']);
|
|
141
|
+
}
|
|
142
|
+
catch (error) {
|
|
143
|
+
console.error('[ImageExtractor] PyMuPDF import check failed:', error instanceof Error ? error.message : String(error));
|
|
144
|
+
missingDeps.push('PyMuPDF');
|
|
145
|
+
}
|
|
146
|
+
// Check Pillow
|
|
147
|
+
try {
|
|
148
|
+
await this.runCommand(this.config.pythonPath, ['-c', 'from PIL import Image; print("OK")']);
|
|
149
|
+
}
|
|
150
|
+
catch (error) {
|
|
151
|
+
console.error('[ImageExtractor] Pillow import check failed:', error instanceof Error ? error.message : String(error));
|
|
152
|
+
missingDeps.push('Pillow');
|
|
153
|
+
}
|
|
154
|
+
return {
|
|
155
|
+
available: missingDeps.length === 0,
|
|
156
|
+
pythonVersion: pythonVersion?.trim(),
|
|
157
|
+
missingDependencies: missingDeps,
|
|
158
|
+
};
|
|
159
|
+
}
|
|
160
|
+
/**
|
|
161
|
+
* Run a Python extraction script (works for both PDF and DOCX extractors)
|
|
162
|
+
*/
|
|
163
|
+
runPythonExtractorScript(scriptPath, inputPath, options) {
|
|
164
|
+
return new Promise((resolve, reject) => {
|
|
165
|
+
const args = [
|
|
166
|
+
scriptPath,
|
|
167
|
+
'--input',
|
|
168
|
+
inputPath,
|
|
169
|
+
'--output',
|
|
170
|
+
options.outputDir,
|
|
171
|
+
'--min-size',
|
|
172
|
+
String(options.minSize ?? 50),
|
|
173
|
+
'--max-images',
|
|
174
|
+
String(options.maxImages ?? 100),
|
|
175
|
+
];
|
|
176
|
+
const timeout = this.config.timeout ?? 120000;
|
|
177
|
+
const proc = spawn(this.config.pythonPath, args, {
|
|
178
|
+
timeout,
|
|
179
|
+
});
|
|
180
|
+
const stdoutChunks = [];
|
|
181
|
+
let stderr = '';
|
|
182
|
+
let settled = false;
|
|
183
|
+
let sigkillTimer = null;
|
|
184
|
+
const cleanup = () => {
|
|
185
|
+
if (sigkillTimer) {
|
|
186
|
+
clearTimeout(sigkillTimer);
|
|
187
|
+
sigkillTimer = null;
|
|
188
|
+
}
|
|
189
|
+
};
|
|
190
|
+
proc.stdout.on('data', (data) => {
|
|
191
|
+
stdoutChunks.push(data);
|
|
192
|
+
});
|
|
193
|
+
// Cap stderr accumulation to prevent unbounded memory growth
|
|
194
|
+
proc.stderr.on('data', (data) => {
|
|
195
|
+
if (stderr.length < MAX_STDERR_LENGTH) {
|
|
196
|
+
stderr += data.toString();
|
|
197
|
+
}
|
|
198
|
+
});
|
|
199
|
+
proc.on('error', (err) => {
|
|
200
|
+
cleanup();
|
|
201
|
+
if (settled)
|
|
202
|
+
return;
|
|
203
|
+
settled = true;
|
|
204
|
+
reject(new Error(`Failed to start Python process: ${err.message}`));
|
|
205
|
+
});
|
|
206
|
+
proc.on('close', (code, signal) => {
|
|
207
|
+
cleanup();
|
|
208
|
+
if (settled)
|
|
209
|
+
return;
|
|
210
|
+
settled = true;
|
|
211
|
+
if (stderr) {
|
|
212
|
+
console.error(`[WARN] [ImageExtractor] stderr: ${stderr.substring(0, 2000)}`);
|
|
213
|
+
}
|
|
214
|
+
if (signal === 'SIGTERM' || signal === 'SIGKILL') {
|
|
215
|
+
reject(new Error(`Python process killed by ${signal} (timeout: ${timeout}ms)`));
|
|
216
|
+
return;
|
|
217
|
+
}
|
|
218
|
+
const stdout = Buffer.concat(stdoutChunks).toString('utf-8');
|
|
219
|
+
try {
|
|
220
|
+
// Python may output debug/warning lines before the JSON. Find the last
|
|
221
|
+
// valid JSON line (starts with '{' or '[') to handle multi-line output.
|
|
222
|
+
const lines = stdout.trim().split('\n');
|
|
223
|
+
let parsed;
|
|
224
|
+
for (let i = lines.length - 1; i >= 0; i--) {
|
|
225
|
+
const line = lines[i].trim();
|
|
226
|
+
if (line.startsWith('{') || line.startsWith('[')) {
|
|
227
|
+
try {
|
|
228
|
+
parsed = JSON.parse(line);
|
|
229
|
+
break;
|
|
230
|
+
}
|
|
231
|
+
catch (error) {
|
|
232
|
+
console.error('[ImageExtractor] JSON parse failed for output line, trying previous:', error instanceof Error ? error.message : String(error));
|
|
233
|
+
/* not valid JSON, try previous line */
|
|
234
|
+
}
|
|
235
|
+
}
|
|
236
|
+
}
|
|
237
|
+
if (parsed === undefined) {
|
|
238
|
+
// Fallback: try parsing the entire stdout as one JSON blob
|
|
239
|
+
parsed = JSON.parse(stdout);
|
|
240
|
+
}
|
|
241
|
+
resolve(parsed);
|
|
242
|
+
}
|
|
243
|
+
catch (parseError) {
|
|
244
|
+
if (code !== 0) {
|
|
245
|
+
reject(new Error(`Python script exited with code ${code}: ${(stderr || stdout).substring(0, 2000)}`));
|
|
246
|
+
}
|
|
247
|
+
else {
|
|
248
|
+
reject(new Error(`Failed to parse extraction result: ${parseError}`));
|
|
249
|
+
}
|
|
250
|
+
}
|
|
251
|
+
});
|
|
252
|
+
// F-INTEG-12: SIGKILL escalation if SIGTERM doesn't exit within 5s.
|
|
253
|
+
// Also settles the promise to prevent zombie hangs if close event never fires.
|
|
254
|
+
if (timeout > 0) {
|
|
255
|
+
sigkillTimer = setTimeout(() => {
|
|
256
|
+
if (!proc.killed) {
|
|
257
|
+
console.error(`[ImageExtractor] Process did not exit after SIGTERM, sending SIGKILL (pid: ${proc.pid})`);
|
|
258
|
+
try {
|
|
259
|
+
proc.kill('SIGKILL');
|
|
260
|
+
}
|
|
261
|
+
catch (error) {
|
|
262
|
+
console.error('[ImageExtractor] Failed to SIGKILL process (may already be gone):', error instanceof Error ? error.message : String(error));
|
|
263
|
+
// Process may already be gone
|
|
264
|
+
}
|
|
265
|
+
}
|
|
266
|
+
// Settle the promise if close event hasn't fired yet (zombie prevention)
|
|
267
|
+
if (!settled) {
|
|
268
|
+
settled = true;
|
|
269
|
+
reject(new Error(`Python process killed by SIGKILL after timeout (${timeout}ms + 5s grace)`));
|
|
270
|
+
}
|
|
271
|
+
}, timeout + 5000);
|
|
272
|
+
}
|
|
273
|
+
});
|
|
274
|
+
}
|
|
275
|
+
/**
|
|
276
|
+
* Run a simple command and return output
|
|
277
|
+
*/
|
|
278
|
+
runCommand(cmd, args) {
|
|
279
|
+
const timeout = 10000;
|
|
280
|
+
return new Promise((resolve, reject) => {
|
|
281
|
+
const proc = spawn(cmd, args, { timeout });
|
|
282
|
+
let stdout = '';
|
|
283
|
+
let stderr = '';
|
|
284
|
+
let settled = false;
|
|
285
|
+
let sigkillTimer = null;
|
|
286
|
+
const cleanup = () => {
|
|
287
|
+
if (sigkillTimer) {
|
|
288
|
+
clearTimeout(sigkillTimer);
|
|
289
|
+
sigkillTimer = null;
|
|
290
|
+
}
|
|
291
|
+
};
|
|
292
|
+
proc.stdout.on('data', (d) => { if (stdout.length < 65536)
|
|
293
|
+
stdout += d; });
|
|
294
|
+
proc.stderr.on('data', (d) => { if (stderr.length < 10240)
|
|
295
|
+
stderr += d; });
|
|
296
|
+
proc.on('error', (err) => {
|
|
297
|
+
cleanup();
|
|
298
|
+
if (settled)
|
|
299
|
+
return;
|
|
300
|
+
settled = true;
|
|
301
|
+
reject(err);
|
|
302
|
+
});
|
|
303
|
+
proc.on('close', (code, signal) => {
|
|
304
|
+
cleanup();
|
|
305
|
+
if (settled)
|
|
306
|
+
return;
|
|
307
|
+
settled = true;
|
|
308
|
+
if (signal === 'SIGTERM' || signal === 'SIGKILL') {
|
|
309
|
+
reject(new Error(`Process killed by ${signal} (timeout: ${timeout}ms)`));
|
|
310
|
+
return;
|
|
311
|
+
}
|
|
312
|
+
if (code === 0) {
|
|
313
|
+
resolve(stdout);
|
|
314
|
+
}
|
|
315
|
+
else {
|
|
316
|
+
reject(new Error(stderr || `Exit code ${code}`));
|
|
317
|
+
}
|
|
318
|
+
});
|
|
319
|
+
// SIGKILL escalation if SIGTERM doesn't exit within 5s
|
|
320
|
+
if (timeout > 0) {
|
|
321
|
+
sigkillTimer = setTimeout(() => {
|
|
322
|
+
if (!proc.killed) {
|
|
323
|
+
console.error(`[ImageExtractor] runCommand process did not exit after SIGTERM, sending SIGKILL (pid: ${proc.pid})`);
|
|
324
|
+
try {
|
|
325
|
+
proc.kill('SIGKILL');
|
|
326
|
+
}
|
|
327
|
+
catch (error) {
|
|
328
|
+
console.error('[ImageExtractor] Failed to SIGKILL runCommand process:', error instanceof Error ? error.message : String(error));
|
|
329
|
+
}
|
|
330
|
+
}
|
|
331
|
+
if (!settled) {
|
|
332
|
+
settled = true;
|
|
333
|
+
reject(new Error(`Process killed by SIGKILL after timeout (${timeout}ms + 5s grace)`));
|
|
334
|
+
}
|
|
335
|
+
}, timeout + 5000);
|
|
336
|
+
}
|
|
337
|
+
});
|
|
338
|
+
}
|
|
339
|
+
}
|
|
340
|
+
//# sourceMappingURL=extractor.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"extractor.js","sourceRoot":"","sources":["../../../src/services/images/extractor.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,KAAK,EAAE,MAAM,eAAe,CAAC;AACtC,OAAO,KAAK,IAAI,MAAM,MAAM,CAAC;AAC7B,OAAO,KAAK,EAAE,MAAM,IAAI,CAAC;AACzB,OAAO,EAAE,aAAa,EAAE,MAAM,KAAK,CAAC;AAGpC,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAyB/D,+DAA+D;AAC/D,MAAM,iBAAiB,GAAG,MAAM,CAAC;AAEjC,gDAAgD;AAChD,MAAM,0BAA0B,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE9D,MAAM,cAAc,GAAoB;IACtC,UAAU,EAAE,OAAO,CAAC,QAAQ,KAAK,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;IAC/D,OAAO,EAAE,MAAM;CAChB,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,cAAc;IACR,MAAM,CAAkB;IACxB,aAAa,CAAS;IACtB,cAAc,CAAS;IAExC,YAAY,SAAmC,EAAE;QAC/C,IAAI,CAAC,MAAM,GAAG,EAAE,GAAG,cAAc,EAAE,GAAG,MAAM,EAAE,CAAC;QAC/C,IAAI,CAAC,aAAa;YAChB,IAAI,CAAC,MAAM,CAAC,UAAU,IAAI,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,oCAAoC,CAAC,CAAC;QAC1F,IAAI,CAAC,cAAc,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,yCAAyC,CAAC,CAAC;IAC3F,CAAC;IAED;;;;;;;;OAQG;IACH,KAAK,CAAC,aAAa,CACjB,QAAgB,EAChB,OAA+B;QAE/B,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;QAEjD,QAAQ,GAAG,EAAE,CAAC;YACZ,KAAK,MAAM;gBACT,OAAO,IAAI,CAAC,cAAc,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;YAChD,KAAK,OAAO;gBACV,OAAO,IAAI,CAAC,eAAe,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;YACjD;gBACE,MAAM,IAAI,KAAK,CACb,gDAAgD,GAAG,KAAK;oBACtD,oBAAoB,CAAC,GAAG,0BAA0B,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACnE,CAAC;QACN,CAAC;IACH,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,WAAW,CAAC,QAAgB;QACjC,OAAO,0BAA0B,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IAC9E,CAAC;IAED;;;;;;;OAOG;IACH,KAAK,CAAC,cAAc,CAClB,OAAe,EACf,OAA+B;QAE/B,sBAAsB;QACtB,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;YAC5B,MAAM,IAAI,KAAK,CAAC,uBAAuB,OAAO,EAAE,CAAC,CAAC;QACpD,CAAC;QAED,yBAAyB;QACzB,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC;YACvC,MAAM,IAAI,KAAK,CACb,qCAAqC,IAAI,CAAC,aAAa,IAAI;gBACzD,0CAA0C,CAC7C,CAAC;QACJ,CAAC;QAED,8CAA8C;QAC9C,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;YACtC,EAAE,CAAC,SAAS,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QACvD,CAAC;QAED,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,wBAAwB,CAAC,IAAI,CAAC,aAAa,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;QAEzF,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,4BAA4B,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;QAC9D,CAAC;QAED,IAAI,MAAM,CAAC,QAAQ,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAClD,OAAO,CAAC,KAAK,CACX,uDAAuD,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACpF,CAAC;QACJ,CAAC;QAED,OAAO,MAAM,CAAC,MAAM,CAAC;IACvB,CAAC;IAED;;;;;;;OAOG;IACH,KAAK,CAAC,eAAe,CACnB,QAAgB,EAChB,OAA+B;QAE/B,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC7B,MAAM,IAAI,KAAK,CAAC,wBAAwB,QAAQ,EAAE,CAAC,CAAC;QACtD,CAAC;QAED,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,cAAc,CAAC,EAAE,CAAC;YACxC,MAAM,IAAI,KAAK,CACb,0CAA0C,IAAI,CAAC,cAAc,IAAI;gBAC/D,+CAA+C,CAClD,CAAC;QACJ,CAAC;QAED,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;YACtC,EAAE,CAAC,SAAS,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QACvD,CAAC;QAED,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,wBAAwB,CAAC,IAAI,CAAC,cAAc,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;QAE3F,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,iCAAiC,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;QACnE,CAAC;QAED,IAAI,MAAM,CAAC,QAAQ,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAClD,OAAO,CAAC,KAAK,CACX,qDAAqD,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAClF,CAAC;QACJ,CAAC;QAED,OAAO,MAAM,CAAC,MAAM,CAAC;IACvB,CAAC;IAED;;;;OAIG;IACH,KAAK,CAAC,gBAAgB;QAKpB,MAAM,WAAW,GAAa,EAAE,CAAC;QAEjC,uBAAuB;QACvB,IAAI,aAAiC,CAAC;QACtC,IAAI,CAAC;YACH,aAAa,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,UAAU,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC;QAC/E,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,CAAC,KAAK,CACX,+CAA+C,EAC/C,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;YACF,OAAO;gBACL,SAAS,EAAE,KAAK;gBAChB,mBAAmB,EAAE,CAAC,QAAQ,CAAC;aAChC,CAAC;QACJ,CAAC;QAED,gBAAgB;QAChB,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,kCAAkC,CAAC,CAAC,CAAC;QAC5F,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,CAAC,KAAK,CACX,+CAA+C,EAC/C,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;YACF,WAAW,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAC9B,CAAC;QAED,eAAe;QACf,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,oCAAoC,CAAC,CAAC,CAAC;QAC9F,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,CAAC,KAAK,CACX,8CAA8C,EAC9C,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;YACF,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAC7B,CAAC;QAED,OAAO;YACL,SAAS,EAAE,WAAW,CAAC,MAAM,KAAK,CAAC;YACnC,aAAa,EAAE,aAAa,EAAE,IAAI,EAAE;YACpC,mBAAmB,EAAE,WAAW;SACjC,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,wBAAwB,CAC9B,UAAkB,EAClB,SAAiB,EACjB,OAA+B;QAE/B,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACrC,MAAM,IAAI,GAAG;gBACX,UAAU;gBACV,SAAS;gBACT,SAAS;gBACT,UAAU;gBACV,OAAO,CAAC,SAAS;gBACjB,YAAY;gBACZ,MAAM,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC;gBAC7B,cAAc;gBACd,MAAM,CAAC,OAAO,CAAC,SAAS,IAAI,GAAG,CAAC;aACjC,CAAC;YAEF,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,IAAI,MAAM,CAAC;YAC9C,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,UAAU,EAAE,IAAI,EAAE;gBAC/C,OAAO;aACR,CAAC,CAAC;YAEH,MAAM,YAAY,GAAa,EAAE,CAAC;YAClC,IAAI,MAAM,GAAG,EAAE,CAAC;YAChB,IAAI,OAAO,GAAG,KAAK,CAAC;YACpB,IAAI,YAAY,GAAyC,IAAI,CAAC;YAE9D,MAAM,OAAO,GAAG,GAAG,EAAE;gBACnB,IAAI,YAAY,EAAE,CAAC;oBACjB,YAAY,CAAC,YAAY,CAAC,CAAC;oBAC3B,YAAY,GAAG,IAAI,CAAC;gBACtB,CAAC;YACH,CAAC,CAAC;YAEF,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,IAAY,EAAE,EAAE;gBACtC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC1B,CAAC,CAAC,CAAC;YAEH,6DAA6D;YAC7D,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,IAAI,EAAE,EAAE;gBAC9B,IAAI,MAAM,CAAC,MAAM,GAAG,iBAAiB,EAAE,CAAC;oBACtC,MAAM,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;gBAC5B,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE;gBACvB,OAAO,EAAE,CAAC;gBACV,IAAI,OAAO;oBAAE,OAAO;gBACpB,OAAO,GAAG,IAAI,CAAC;gBACf,MAAM,CAAC,IAAI,KAAK,CAAC,mCAAmC,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;YACtE,CAAC,CAAC,CAAC;YAEH,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE;gBAChC,OAAO,EAAE,CAAC;gBACV,IAAI,OAAO;oBAAE,OAAO;gBACpB,OAAO,GAAG,IAAI,CAAC;gBAEf,IAAI,MAAM,EAAE,CAAC;oBACX,OAAO,CAAC,KAAK,CAAC,mCAAmC,MAAM,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;gBAChF,CAAC;gBAED,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;oBACjD,MAAM,CAAC,IAAI,KAAK,CAAC,4BAA4B,MAAM,cAAc,OAAO,KAAK,CAAC,CAAC,CAAC;oBAChF,OAAO;gBACT,CAAC;gBAED,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,YAAY,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;gBAC7D,IAAI,CAAC;oBACH,uEAAuE;oBACvE,wEAAwE;oBACxE,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;oBACxC,IAAI,MAAoC,CAAC;oBACzC,KAAK,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;wBAC3C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;wBAC7B,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;4BACjD,IAAI,CAAC;gCACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAqB,CAAC;gCAC9C,MAAM;4BACR,CAAC;4BAAC,OAAO,KAAK,EAAE,CAAC;gCACf,OAAO,CAAC,KAAK,CACX,sEAAsE,EACtE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;gCACF,uCAAuC;4BACzC,CAAC;wBACH,CAAC;oBACH,CAAC;oBACD,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;wBACzB,2DAA2D;wBAC3D,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAqB,CAAC;oBAClD,CAAC;oBACD,OAAO,CAAC,MAAM,CAAC,CAAC;gBAClB,CAAC;gBAAC,OAAO,UAAU,EAAE,CAAC;oBACpB,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC;wBACf,MAAM,CACJ,IAAI,KAAK,CACP,kCAAkC,IAAI,KAAK,CAAC,MAAM,IAAI,MAAM,CAAC,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,CACnF,CACF,CAAC;oBACJ,CAAC;yBAAM,CAAC;wBACN,MAAM,CAAC,IAAI,KAAK,CAAC,sCAAsC,UAAU,EAAE,CAAC,CAAC,CAAC;oBACxE,CAAC;gBACH,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,oEAAoE;YACpE,+EAA+E;YAC/E,IAAI,OAAO,GAAG,CAAC,EAAE,CAAC;gBAChB,YAAY,GAAG,UAAU,CAAC,GAAG,EAAE;oBAC7B,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;wBACjB,OAAO,CAAC,KAAK,CACX,8EAA8E,IAAI,CAAC,GAAG,GAAG,CAC1F,CAAC;wBACF,IAAI,CAAC;4BACH,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;wBACvB,CAAC;wBAAC,OAAO,KAAK,EAAE,CAAC;4BACf,OAAO,CAAC,KAAK,CACX,mEAAmE,EACnE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;4BACF,8BAA8B;wBAChC,CAAC;oBACH,CAAC;oBACD,yEAAyE;oBACzE,IAAI,CAAC,OAAO,EAAE,CAAC;wBACb,OAAO,GAAG,IAAI,CAAC;wBACf,MAAM,CACJ,IAAI,KAAK,CAAC,mDAAmD,OAAO,gBAAgB,CAAC,CACtF,CAAC;oBACJ,CAAC;gBACH,CAAC,EAAE,OAAO,GAAG,IAAI,CAAC,CAAC;YACrB,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED;;OAEG;IACK,UAAU,CAAC,GAAW,EAAE,IAAc;QAC5C,MAAM,OAAO,GAAG,KAAK,CAAC;QACtB,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACrC,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,OAAO,EAAE,CAAC,CAAC;YAC3C,IAAI,MAAM,GAAG,EAAE,CAAC;YAChB,IAAI,MAAM,GAAG,EAAE,CAAC;YAChB,IAAI,OAAO,GAAG,KAAK,CAAC;YACpB,IAAI,YAAY,GAAyC,IAAI,CAAC;YAE9D,MAAM,OAAO,GAAG,GAAG,EAAE;gBACnB,IAAI,YAAY,EAAE,CAAC;oBACjB,YAAY,CAAC,YAAY,CAAC,CAAC;oBAC3B,YAAY,GAAG,IAAI,CAAC;gBACtB,CAAC;YACH,CAAC,CAAC;YAEF,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,MAAM,CAAC,MAAM,GAAG,KAAK;gBAAE,MAAM,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3E,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,MAAM,CAAC,MAAM,GAAG,KAAK;gBAAE,MAAM,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAE3E,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE;gBACvB,OAAO,EAAE,CAAC;gBACV,IAAI,OAAO;oBAAE,OAAO;gBACpB,OAAO,GAAG,IAAI,CAAC;gBACf,MAAM,CAAC,GAAG,CAAC,CAAC;YACd,CAAC,CAAC,CAAC;YAEH,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE;gBAChC,OAAO,EAAE,CAAC;gBACV,IAAI,OAAO;oBAAE,OAAO;gBACpB,OAAO,GAAG,IAAI,CAAC;gBAEf,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;oBACjD,MAAM,CAAC,IAAI,KAAK,CAAC,qBAAqB,MAAM,cAAc,OAAO,KAAK,CAAC,CAAC,CAAC;oBACzE,OAAO;gBACT,CAAC;gBAED,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC;oBACf,OAAO,CAAC,MAAM,CAAC,CAAC;gBAClB,CAAC;qBAAM,CAAC;oBACN,MAAM,CAAC,IAAI,KAAK,CAAC,MAAM,IAAI,aAAa,IAAI,EAAE,CAAC,CAAC,CAAC;gBACnD,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,uDAAuD;YACvD,IAAI,OAAO,GAAG,CAAC,EAAE,CAAC;gBAChB,YAAY,GAAG,UAAU,CAAC,GAAG,EAAE;oBAC7B,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;wBACjB,OAAO,CAAC,KAAK,CACX,yFAAyF,IAAI,CAAC,GAAG,GAAG,CACrG,CAAC;wBACF,IAAI,CAAC;4BACH,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;wBACvB,CAAC;wBAAC,OAAO,KAAK,EAAE,CAAC;4BACf,OAAO,CAAC,KAAK,CACX,wDAAwD,EACxD,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;wBACJ,CAAC;oBACH,CAAC;oBACD,IAAI,CAAC,OAAO,EAAE,CAAC;wBACb,OAAO,GAAG,IAAI,CAAC;wBACf,MAAM,CACJ,IAAI,KAAK,CAAC,4CAA4C,OAAO,gBAAgB,CAAC,CAC/E,CAAC;oBACJ,CAAC;gBACH,CAAC,EAAE,OAAO,GAAG,IAAI,CAAC,CAAC;YACrB,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;CACF"}
|
|
@@ -0,0 +1,130 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Image Optimizer Service
|
|
3
|
+
*
|
|
4
|
+
* TypeScript wrapper for Python image optimizer providing:
|
|
5
|
+
* 1. Resize for OCR (max 4800px width for Datalab API)
|
|
6
|
+
* 2. Resize for VLM (optimize token usage, max 2048px)
|
|
7
|
+
* 3. Relevance analysis to filter logos, icons, and decorative elements
|
|
8
|
+
*
|
|
9
|
+
* The relevance analysis uses multi-layer heuristics:
|
|
10
|
+
* - Size filtering (tiny images are likely icons)
|
|
11
|
+
* - Aspect ratio analysis (extreme ratios = banners/logos)
|
|
12
|
+
* - Color diversity (low color count = likely logo/icon)
|
|
13
|
+
*/
|
|
14
|
+
/**
|
|
15
|
+
* Image category classification
|
|
16
|
+
*/
|
|
17
|
+
type ImageCategory = 'photo' | 'chart' | 'document' | 'logo' | 'icon' | 'decorative' | 'unknown';
|
|
18
|
+
/**
|
|
19
|
+
* Result of image relevance analysis
|
|
20
|
+
*/
|
|
21
|
+
interface ImageAnalysisResult {
|
|
22
|
+
success: true;
|
|
23
|
+
path: string;
|
|
24
|
+
width: number;
|
|
25
|
+
height: number;
|
|
26
|
+
aspect_ratio: number;
|
|
27
|
+
unique_colors: number;
|
|
28
|
+
color_diversity_score: number;
|
|
29
|
+
size_score: number;
|
|
30
|
+
aspect_score: number;
|
|
31
|
+
overall_relevance: number;
|
|
32
|
+
predicted_category: ImageCategory;
|
|
33
|
+
should_vlm: boolean;
|
|
34
|
+
skip_reason?: string;
|
|
35
|
+
}
|
|
36
|
+
/**
|
|
37
|
+
* Result of resize operation
|
|
38
|
+
*/
|
|
39
|
+
interface ResizeResult {
|
|
40
|
+
success: true;
|
|
41
|
+
resized: boolean;
|
|
42
|
+
original_width: number;
|
|
43
|
+
original_height: number;
|
|
44
|
+
output_width: number;
|
|
45
|
+
output_height: number;
|
|
46
|
+
scale_factor?: number;
|
|
47
|
+
output_path: string;
|
|
48
|
+
}
|
|
49
|
+
/**
|
|
50
|
+
* Result when image is skipped (too small)
|
|
51
|
+
*/
|
|
52
|
+
interface SkipResult {
|
|
53
|
+
success: true;
|
|
54
|
+
skipped: true;
|
|
55
|
+
skip_reason: string;
|
|
56
|
+
original_width: number;
|
|
57
|
+
original_height: number;
|
|
58
|
+
}
|
|
59
|
+
/**
|
|
60
|
+
* Error result from Python script
|
|
61
|
+
*/
|
|
62
|
+
interface ErrorResult {
|
|
63
|
+
success: false;
|
|
64
|
+
error: string;
|
|
65
|
+
}
|
|
66
|
+
/**
|
|
67
|
+
* Configuration for the image optimizer
|
|
68
|
+
*/
|
|
69
|
+
interface ImageOptimizerConfig {
|
|
70
|
+
/** Path to Python executable */
|
|
71
|
+
pythonPath: string;
|
|
72
|
+
/** Timeout in milliseconds */
|
|
73
|
+
timeout: number;
|
|
74
|
+
/** Maximum dimension for VLM resize (default: 2048) */
|
|
75
|
+
vlmMaxDimension: number;
|
|
76
|
+
/** Minimum size to skip for VLM (default: 50) */
|
|
77
|
+
vlmSkipBelowSize: number;
|
|
78
|
+
/** Minimum relevance score for VLM (default: 0.3) */
|
|
79
|
+
minRelevanceScore: number;
|
|
80
|
+
}
|
|
81
|
+
/**
|
|
82
|
+
* Service for optimizing images for OCR and VLM processing
|
|
83
|
+
*/
|
|
84
|
+
export declare class ImageOptimizer {
|
|
85
|
+
private readonly config;
|
|
86
|
+
private readonly scriptPath;
|
|
87
|
+
constructor(config?: Partial<ImageOptimizerConfig>);
|
|
88
|
+
/**
|
|
89
|
+
* Analyze an image to determine if it should be processed by VLM.
|
|
90
|
+
*
|
|
91
|
+
* Uses multi-layer heuristics:
|
|
92
|
+
* 1. Size (tiny images = skip)
|
|
93
|
+
* 2. Aspect ratio (extreme ratios = skip)
|
|
94
|
+
* 3. Color diversity (low = likely logo/icon)
|
|
95
|
+
* 4. Category prediction
|
|
96
|
+
*
|
|
97
|
+
* @param imagePath - Path to the image file
|
|
98
|
+
* @returns Analysis result with should_vlm recommendation
|
|
99
|
+
*/
|
|
100
|
+
analyzeImage(imagePath: string): Promise<ImageAnalysisResult | ErrorResult>;
|
|
101
|
+
/**
|
|
102
|
+
* Resize an image for VLM processing (Gemini).
|
|
103
|
+
*
|
|
104
|
+
* @param inputPath - Path to input image
|
|
105
|
+
* @param outputPath - Path for output (optional, creates temp file if not provided)
|
|
106
|
+
* @returns Resize result or skip result if too small
|
|
107
|
+
*/
|
|
108
|
+
resizeForVLM(inputPath: string, outputPath?: string): Promise<ResizeResult | SkipResult | ErrorResult>;
|
|
109
|
+
/**
|
|
110
|
+
* Check if the Python optimizer script exists.
|
|
111
|
+
*/
|
|
112
|
+
isAvailable(): boolean;
|
|
113
|
+
/**
|
|
114
|
+
* Create a temporary file path for resized output.
|
|
115
|
+
*/
|
|
116
|
+
private createTempPath;
|
|
117
|
+
/**
|
|
118
|
+
* Run the Python optimizer script.
|
|
119
|
+
* M-14: Sends SIGKILL after 5s if SIGTERM from Node.js timeout didn't terminate the process.
|
|
120
|
+
*/
|
|
121
|
+
private runPython;
|
|
122
|
+
}
|
|
123
|
+
/**
|
|
124
|
+
* Get an optimizer instance.
|
|
125
|
+
* When config is provided, always creates a new instance and updates the cache.
|
|
126
|
+
* When no config is provided, returns the cached instance (creating one if needed).
|
|
127
|
+
*/
|
|
128
|
+
export declare function getImageOptimizer(config?: Partial<ImageOptimizerConfig>): ImageOptimizer;
|
|
129
|
+
export {};
|
|
130
|
+
//# sourceMappingURL=optimizer.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"optimizer.d.ts","sourceRoot":"","sources":["../../../src/services/images/optimizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAUH;;GAEG;AACH,KAAK,aAAa,GAAG,OAAO,GAAG,OAAO,GAAG,UAAU,GAAG,MAAM,GAAG,MAAM,GAAG,YAAY,GAAG,SAAS,CAAC;AAEjG;;GAEG;AACH,UAAU,mBAAmB;IAC3B,OAAO,EAAE,IAAI,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;IACb,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,EAAE,MAAM,CAAC;IACrB,aAAa,EAAE,MAAM,CAAC;IACtB,qBAAqB,EAAE,MAAM,CAAC;IAC9B,UAAU,EAAE,MAAM,CAAC;IACnB,YAAY,EAAE,MAAM,CAAC;IACrB,iBAAiB,EAAE,MAAM,CAAC;IAC1B,kBAAkB,EAAE,aAAa,CAAC;IAClC,UAAU,EAAE,OAAO,CAAC;IACpB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED;;GAEG;AACH,UAAU,YAAY;IACpB,OAAO,EAAE,IAAI,CAAC;IACd,OAAO,EAAE,OAAO,CAAC;IACjB,cAAc,EAAE,MAAM,CAAC;IACvB,eAAe,EAAE,MAAM,CAAC;IACxB,YAAY,EAAE,MAAM,CAAC;IACrB,aAAa,EAAE,MAAM,CAAC;IACtB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,WAAW,EAAE,MAAM,CAAC;CACrB;AAED;;GAEG;AACH,UAAU,UAAU;IAClB,OAAO,EAAE,IAAI,CAAC;IACd,OAAO,EAAE,IAAI,CAAC;IACd,WAAW,EAAE,MAAM,CAAC;IACpB,cAAc,EAAE,MAAM,CAAC;IACvB,eAAe,EAAE,MAAM,CAAC;CACzB;AAED;;GAEG;AACH,UAAU,WAAW;IACnB,OAAO,EAAE,KAAK,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;CACf;AAED;;GAEG;AACH,UAAU,oBAAoB;IAC5B,gCAAgC;IAChC,UAAU,EAAE,MAAM,CAAC;IACnB,8BAA8B;IAC9B,OAAO,EAAE,MAAM,CAAC;IAChB,uDAAuD;IACvD,eAAe,EAAE,MAAM,CAAC;IACxB,iDAAiD;IACjD,gBAAgB,EAAE,MAAM,CAAC;IACzB,qDAAqD;IACrD,iBAAiB,EAAE,MAAM,CAAC;CAC3B;AAaD;;GAEG;AACH,qBAAa,cAAc;IACzB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAuB;IAC9C,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAS;gBAExB,MAAM,GAAE,OAAO,CAAC,oBAAoB,CAAM;IAKtD;;;;;;;;;;;OAWG;IACG,YAAY,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAAC,mBAAmB,GAAG,WAAW,CAAC;IAIjF;;;;;;OAMG;IACG,YAAY,CAChB,SAAS,EAAE,MAAM,EACjB,UAAU,CAAC,EAAE,MAAM,GAClB,OAAO,CAAC,YAAY,GAAG,UAAU,GAAG,WAAW,CAAC;IAYnD;;OAEG;IACH,WAAW,IAAI,OAAO;IAItB;;OAEG;IACH,OAAO,CAAC,cAAc;IAOtB;;;OAGG;IACH,OAAO,CAAC,SAAS;CAoIlB;AAOD;;;;GAIG;AACH,wBAAgB,iBAAiB,CAAC,MAAM,CAAC,EAAE,OAAO,CAAC,oBAAoB,CAAC,GAAG,cAAc,CASxF"}
|