@ssmeo/mcp-local-rag 0.18.6
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +21 -0
- package/README.de.md +441 -0
- package/README.es.md +441 -0
- package/README.fr.md +441 -0
- package/README.md +530 -0
- package/README.pt-BR.md +441 -0
- package/README.zh-CN.md +441 -0
- package/dist/bin/install-skills.d.ts +19 -0
- package/dist/bin/install-skills.d.ts.map +1 -0
- package/dist/bin/install-skills.js +218 -0
- package/dist/bin/install-skills.js.map +1 -0
- package/dist/chunker/index.d.ts +24 -0
- package/dist/chunker/index.d.ts.map +1 -0
- package/dist/chunker/index.js +2 -0
- package/dist/chunker/index.js.map +1 -0
- package/dist/chunker/semantic-chunker.d.ts +96 -0
- package/dist/chunker/semantic-chunker.d.ts.map +1 -0
- package/dist/chunker/semantic-chunker.js +362 -0
- package/dist/chunker/semantic-chunker.js.map +1 -0
- package/dist/chunker/sentence-splitter.d.ts +32 -0
- package/dist/chunker/sentence-splitter.d.ts.map +1 -0
- package/dist/chunker/sentence-splitter.js +231 -0
- package/dist/chunker/sentence-splitter.js.map +1 -0
- package/dist/chunker/token-containment.d.ts +39 -0
- package/dist/chunker/token-containment.d.ts.map +1 -0
- package/dist/chunker/token-containment.js +83 -0
- package/dist/chunker/token-containment.js.map +1 -0
- package/dist/cli/common.d.ts +34 -0
- package/dist/cli/common.d.ts.map +1 -0
- package/dist/cli/common.js +121 -0
- package/dist/cli/common.js.map +1 -0
- package/dist/cli/delete.d.ts +3 -0
- package/dist/cli/delete.d.ts.map +1 -0
- package/dist/cli/delete.js +169 -0
- package/dist/cli/delete.js.map +1 -0
- package/dist/cli/file-collection.d.ts +2 -0
- package/dist/cli/file-collection.d.ts.map +1 -0
- package/dist/cli/file-collection.js +53 -0
- package/dist/cli/file-collection.js.map +1 -0
- package/dist/cli/ingest.d.ts +69 -0
- package/dist/cli/ingest.d.ts.map +1 -0
- package/dist/cli/ingest.js +323 -0
- package/dist/cli/ingest.js.map +1 -0
- package/dist/cli/list.d.ts +27 -0
- package/dist/cli/list.d.ts.map +1 -0
- package/dist/cli/list.js +207 -0
- package/dist/cli/list.js.map +1 -0
- package/dist/cli/options.d.ts +89 -0
- package/dist/cli/options.d.ts.map +1 -0
- package/dist/cli/options.js +244 -0
- package/dist/cli/options.js.map +1 -0
- package/dist/cli/query.d.ts +17 -0
- package/dist/cli/query.d.ts.map +1 -0
- package/dist/cli/query.js +194 -0
- package/dist/cli/query.js.map +1 -0
- package/dist/cli/read-neighbors.d.ts +3 -0
- package/dist/cli/read-neighbors.d.ts.map +1 -0
- package/dist/cli/read-neighbors.js +192 -0
- package/dist/cli/read-neighbors.js.map +1 -0
- package/dist/cli/status.d.ts +6 -0
- package/dist/cli/status.d.ts.map +1 -0
- package/dist/cli/status.js +78 -0
- package/dist/cli/status.js.map +1 -0
- package/dist/cli/sync.d.ts +6 -0
- package/dist/cli/sync.d.ts.map +1 -0
- package/dist/cli/sync.js +271 -0
- package/dist/cli/sync.js.map +1 -0
- package/dist/cli-main.d.ts +9 -0
- package/dist/cli-main.d.ts.map +1 -0
- package/dist/cli-main.js +60 -0
- package/dist/cli-main.js.map +1 -0
- package/dist/embedder/index.d.ts +143 -0
- package/dist/embedder/index.d.ts.map +1 -0
- package/dist/embedder/index.js +429 -0
- package/dist/embedder/index.js.map +1 -0
- package/dist/embedder/remote.d.ts +51 -0
- package/dist/embedder/remote.d.ts.map +1 -0
- package/dist/embedder/remote.js +182 -0
- package/dist/embedder/remote.js.map +1 -0
- package/dist/features/list.d.ts +37 -0
- package/dist/features/list.d.ts.map +1 -0
- package/dist/features/list.js +41 -0
- package/dist/features/list.js.map +1 -0
- package/dist/features/sync.d.ts +219 -0
- package/dist/features/sync.d.ts.map +1 -0
- package/dist/features/sync.js +461 -0
- package/dist/features/sync.js.map +1 -0
- package/dist/index.d.ts +3 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +56 -0
- package/dist/index.js.map +1 -0
- package/dist/ingest/compute.d.ts +63 -0
- package/dist/ingest/compute.d.ts.map +1 -0
- package/dist/ingest/compute.js +161 -0
- package/dist/ingest/compute.js.map +1 -0
- package/dist/ingest/file.d.ts +40 -0
- package/dist/ingest/file.d.ts.map +1 -0
- package/dist/ingest/file.js +74 -0
- package/dist/ingest/file.js.map +1 -0
- package/dist/ingest/visual.d.ts +52 -0
- package/dist/ingest/visual.d.ts.map +1 -0
- package/dist/ingest/visual.js +294 -0
- package/dist/ingest/visual.js.map +1 -0
- package/dist/parser/docx-parser.d.ts +12 -0
- package/dist/parser/docx-parser.d.ts.map +1 -0
- package/dist/parser/docx-parser.js +375 -0
- package/dist/parser/docx-parser.js.map +1 -0
- package/dist/parser/html-parser.d.ts +16 -0
- package/dist/parser/html-parser.d.ts.map +1 -0
- package/dist/parser/html-parser.js +119 -0
- package/dist/parser/html-parser.js.map +1 -0
- package/dist/parser/index.d.ts +125 -0
- package/dist/parser/index.d.ts.map +1 -0
- package/dist/parser/index.js +364 -0
- package/dist/parser/index.js.map +1 -0
- package/dist/parser/pdf-extract.d.ts +49 -0
- package/dist/parser/pdf-extract.d.ts.map +1 -0
- package/dist/parser/pdf-extract.js +114 -0
- package/dist/parser/pdf-extract.js.map +1 -0
- package/dist/parser/pdf-filter.d.ts +91 -0
- package/dist/parser/pdf-filter.d.ts.map +1 -0
- package/dist/parser/pdf-filter.js +522 -0
- package/dist/parser/pdf-filter.js.map +1 -0
- package/dist/parser/title-extractor.d.ts +28 -0
- package/dist/parser/title-extractor.d.ts.map +1 -0
- package/dist/parser/title-extractor.js +105 -0
- package/dist/parser/title-extractor.js.map +1 -0
- package/dist/pdf-visual/captioner.d.ts +11 -0
- package/dist/pdf-visual/captioner.d.ts.map +1 -0
- package/dist/pdf-visual/captioner.js +48 -0
- package/dist/pdf-visual/captioner.js.map +1 -0
- package/dist/pdf-visual/captioners/fast.d.ts +7 -0
- package/dist/pdf-visual/captioners/fast.d.ts.map +1 -0
- package/dist/pdf-visual/captioners/fast.js +108 -0
- package/dist/pdf-visual/captioners/fast.js.map +1 -0
- package/dist/pdf-visual/captioners/quality.d.ts +7 -0
- package/dist/pdf-visual/captioners/quality.d.ts.map +1 -0
- package/dist/pdf-visual/captioners/quality.js +101 -0
- package/dist/pdf-visual/captioners/quality.js.map +1 -0
- package/dist/pdf-visual/captioners/shared.d.ts +81 -0
- package/dist/pdf-visual/captioners/shared.d.ts.map +1 -0
- package/dist/pdf-visual/captioners/shared.js +156 -0
- package/dist/pdf-visual/captioners/shared.js.map +1 -0
- package/dist/pdf-visual/detector.d.ts +9 -0
- package/dist/pdf-visual/detector.d.ts.map +1 -0
- package/dist/pdf-visual/detector.js +269 -0
- package/dist/pdf-visual/detector.js.map +1 -0
- package/dist/pdf-visual/index.d.ts +13 -0
- package/dist/pdf-visual/index.d.ts.map +1 -0
- package/dist/pdf-visual/index.js +53 -0
- package/dist/pdf-visual/index.js.map +1 -0
- package/dist/pdf-visual/renderer.d.ts +9 -0
- package/dist/pdf-visual/renderer.d.ts.map +1 -0
- package/dist/pdf-visual/renderer.js +186 -0
- package/dist/pdf-visual/renderer.js.map +1 -0
- package/dist/pdf-visual/types.d.ts +53 -0
- package/dist/pdf-visual/types.d.ts.map +1 -0
- package/dist/pdf-visual/types.js +17 -0
- package/dist/pdf-visual/types.js.map +1 -0
- package/dist/server/error-utils.d.ts +63 -0
- package/dist/server/error-utils.d.ts.map +1 -0
- package/dist/server/error-utils.js +124 -0
- package/dist/server/error-utils.js.map +1 -0
- package/dist/server/index.d.ts +216 -0
- package/dist/server/index.d.ts.map +1 -0
- package/dist/server/index.js +1068 -0
- package/dist/server/index.js.map +1 -0
- package/dist/server/list-scanner.d.ts +30 -0
- package/dist/server/list-scanner.d.ts.map +1 -0
- package/dist/server/list-scanner.js +50 -0
- package/dist/server/list-scanner.js.map +1 -0
- package/dist/server/tool-definitions.d.ts +8 -0
- package/dist/server/tool-definitions.d.ts.map +1 -0
- package/dist/server/tool-definitions.js +181 -0
- package/dist/server/tool-definitions.js.map +1 -0
- package/dist/server/tool-input.d.ts +37 -0
- package/dist/server/tool-input.d.ts.map +1 -0
- package/dist/server/tool-input.js +211 -0
- package/dist/server/tool-input.js.map +1 -0
- package/dist/server/types.d.ts +307 -0
- package/dist/server/types.d.ts.map +1 -0
- package/dist/server/types.js +3 -0
- package/dist/server/types.js.map +1 -0
- package/dist/server-main.d.ts +40 -0
- package/dist/server-main.d.ts.map +1 -0
- package/dist/server-main.js +254 -0
- package/dist/server-main.js.map +1 -0
- package/dist/utils/base-dirs.d.ts +145 -0
- package/dist/utils/base-dirs.d.ts.map +1 -0
- package/dist/utils/base-dirs.js +344 -0
- package/dist/utils/base-dirs.js.map +1 -0
- package/dist/utils/errors.d.ts +32 -0
- package/dist/utils/errors.d.ts.map +1 -0
- package/dist/utils/errors.js +61 -0
- package/dist/utils/errors.js.map +1 -0
- package/dist/utils/limits.d.ts +26 -0
- package/dist/utils/limits.d.ts.map +1 -0
- package/dist/utils/limits.js +28 -0
- package/dist/utils/limits.js.map +1 -0
- package/dist/utils/list-sources.d.ts +38 -0
- package/dist/utils/list-sources.d.ts.map +1 -0
- package/dist/utils/list-sources.js +40 -0
- package/dist/utils/list-sources.js.map +1 -0
- package/dist/utils/raw-data-utils.d.ts +76 -0
- package/dist/utils/raw-data-utils.d.ts.map +1 -0
- package/dist/utils/raw-data-utils.js +202 -0
- package/dist/utils/raw-data-utils.js.map +1 -0
- package/dist/utils/scan.d.ts +114 -0
- package/dist/utils/scan.d.ts.map +1 -0
- package/dist/utils/scan.js +209 -0
- package/dist/utils/scan.js.map +1 -0
- package/dist/utils/scope-match.d.ts +38 -0
- package/dist/utils/scope-match.d.ts.map +1 -0
- package/dist/utils/scope-match.js +86 -0
- package/dist/utils/scope-match.js.map +1 -0
- package/dist/utils/sensitive-path.d.ts +9 -0
- package/dist/utils/sensitive-path.d.ts.map +1 -0
- package/dist/utils/sensitive-path.js +78 -0
- package/dist/utils/sensitive-path.js.map +1 -0
- package/dist/utils/sync-path-key.d.ts +14 -0
- package/dist/utils/sync-path-key.d.ts.map +1 -0
- package/dist/utils/sync-path-key.js +27 -0
- package/dist/utils/sync-path-key.js.map +1 -0
- package/dist/utils/type-guards.d.ts +18 -0
- package/dist/utils/type-guards.d.ts.map +1 -0
- package/dist/utils/type-guards.js +34 -0
- package/dist/utils/type-guards.js.map +1 -0
- package/dist/utils/visual-profile.d.ts +10 -0
- package/dist/utils/visual-profile.d.ts.map +1 -0
- package/dist/utils/visual-profile.js +15 -0
- package/dist/utils/visual-profile.js.map +1 -0
- package/dist/vectordb/index.d.ts +132 -0
- package/dist/vectordb/index.d.ts.map +1 -0
- package/dist/vectordb/index.js +617 -0
- package/dist/vectordb/index.js.map +1 -0
- package/dist/vectordb/search-filters.d.ts +15 -0
- package/dist/vectordb/search-filters.d.ts.map +1 -0
- package/dist/vectordb/search-filters.js +129 -0
- package/dist/vectordb/search-filters.js.map +1 -0
- package/dist/vectordb/types.d.ts +194 -0
- package/dist/vectordb/types.d.ts.map +1 -0
- package/dist/vectordb/types.js +237 -0
- package/dist/vectordb/types.js.map +1 -0
- package/package.json +108 -0
- package/skills/mcp-local-rag/SKILL.md +234 -0
- package/skills/mcp-local-rag/references/cli-reference.md +190 -0
- package/skills/mcp-local-rag/references/html-ingestion.md +78 -0
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"visual.js","sourceRoot":"","sources":["../../src/ingest/visual.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,wBAAwB,EAAE,sBAAsB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAA;AAoDjG,SAAS,iBAAiB,CAAC,IAAuB,EAAE,KAAwB;IAC1E,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;AACvF,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,kBAAkB,CACzB,SAA0C,EAC1C,MAA4B;IAE5B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,CAAC,CAAA;IACV,CAAC;IACD,IAAI,UAAU,GAAG,SAAS;SACvB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;SAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,iBAAiB,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAA;IAE1E,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5B,MAAM,YAAY,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACzC,MAAM,eAAe,GAAG,IAAI,CAAC,GAAG,CAC9B,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC,CAChF,CAAA;QACD,UAAU,GAAG,SAAS;aACnB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;aAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,KAAK,eAAe,CAAC,CAAA;IAClG,CAAC;IAED,MAAM,cAAc,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC3C,MAAM,SAAS,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,cAAc,CAAC,CAAA;IAC/F,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAA;IAC7D,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAA;AAC1D,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAuB;IAChD,IAAI,IAAI,CAAC,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5D,OAAO,IAAI,CAAC,aAAa,CAAA;IAC3B,CAAC;IACD,OAAO;QACL;YACE,OAAO,EAAE,IAAI,CAAC,OAAO;YACrB,YAAY,EAAE,CAAC;YACf,WAAW,EAAE,CAAC;YACd,eAAe,EAAE,CAAC;YAClB,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;YAClB,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,aAAa,EAAE,CAAC;YAChB,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM;SAC9B;KACF,CAAA;AACH,CAAC;AAQD,2EAA2E;AAC3E,SAAS,eAAe,CACtB,IAAuB,EACvB,aAA8C,EAC9C,gBAAkD;IAElD,OAAO,gBAAgB;SACpB,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,CAAC;SACnD,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,cAAc,EAAE,kBAAkB,CAAC,aAAa,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;SACxF,IAAI,CACH,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,IAAI,CAAC,cAAc,GAAG,KAAK,CAAC,cAAc;QAC1C,IAAI,CAAC,MAAM,CAAC,cAAc,GAAG,KAAK,CAAC,MAAM,CAAC,cAAc,CAC3D,CAAA;AACL,CAAC;AAED;;;GAGG;AACH,SAAS,sBAAsB,CAC7B,IAAuB,EACvB,aAA8C,EAC9C,UAAuC;IAEvC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAmC,CAAA;IACzD,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC9B,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,GAAG,CAAC,CAAC,CAAA;QACxD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;QACpD,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CACrB,CAAC,EACD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,WAAW,IAAI,SAAS,EAAE,aAAa,IAAI,CAAC,CAAC,CACpF,CAAA;QACD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAA;QACtC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QACvB,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;IAC3B,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AASD,kEAAkE;AAClE,SAAS,gBAAgB,CAAC,QAAgB;IACxC,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAA;IACvD,OAAO,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,CAAC,CAAA;AAC9C,CAAC;AAED;;;GAGG;AACH,SAAS,mBAAmB,CAC1B,QAAgB,EAChB,KAAuC,EACvC,OAKC;IAED,MAAM,EAAE,IAAI,EAAE,iBAAiB,EAAE,GAAG,IAAI,EAAE,GAAG,OAAO,CAAA;IACpD,IAAI,IAAI,GAAG,QAAQ,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,KAAK,MAAM,MAAM,IAAI,KAAK,EAAE,CAAC;QAC3B,IAAI,MAAM,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;YAC5B,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;YAC3C,SAAQ;QACV,CAAC;QACD,IAAI,YAAY,GAAG,CAAC,EAAE,CAAC;YACrB,IAAI,IAAI,MAAM,CAAA;QAChB,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAA;QACzB,IAAI,IAAI,2BAA2B,IAAI,CAAC,OAAO,YAAY,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,MAAM,CAAC,OAAO,GAAG,CAAA;QAC9G,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;QAC3D,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;QACrC,YAAY,IAAI,CAAC,CAAA;IACnB,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,sBAAsB,CAC7B,IAAuB,EACvB,eAAqD,EACrD,iBAAqD;IAErD,IAAI,QAAQ,GAAG,EAAE,CAAA;IACjB,IAAI,UAAU,GAAG,CAAC,CAAA;IAClB,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0C,CAAA;IACvE,MAAM,aAAa,GAAG,IAAI,GAAG,EAAiC,CAAA;IAE9D,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAC7F,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;QAC/C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;QAClE,IAAI,UAAU,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtC,QAAQ,IAAI,gBAAgB,CAAC,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QACjE,CAAC;QACD,QAAQ,GAAG,mBAAmB,CAAC,QAAQ,EAAE,KAAK,EAAE;YAC9C,IAAI;YACJ,iBAAiB;YACjB,aAAa;YACb,aAAa;SACd,CAAC,CAAA;QACF,IAAI,UAAU,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YAC5C,QAAQ,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAA;QACvD,CAAC;QACD,UAAU,GAAG,MAAM,CAAA;IACrB,CAAC;IACD,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAA;IAEvC,OAAO,EAAE,QAAQ,EAAE,aAAa,EAAE,aAAa,EAAE,CAAA;AACnD,CAAC;AAED,oEAAoE;AACpE,SAAS,eAAe,CACtB,MAA6B,EAC7B,QAAsB,EACtB,SAAiB,EACjB,WAAmB;IAEnB,MAAM,QAAQ,GAAG,QAAQ,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IACnD,MAAM,YAAY,GAAG,QAAQ;QAC3B,CAAC,CAAC,EAAE,KAAK,EAAE,SAAS,GAAG,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,SAAS,GAAG,QAAQ,CAAC,GAAG,EAAE;QACtE,CAAC,CAAC,SAAS,CAAA;IACb,MAAM,YAAY,GAAG,QAAQ,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IACvD,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;QAC/B,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAA;IACvD,CAAC;IACD,OAAO;QACL,GAAG,MAAM;QACT,WAAW;QACX,YAAY,EAAE,SAAS,GAAG,YAAY;QACtC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC1C,CAAA;AACH,CAAC;AAED,MAAM,UAAU,0BAA0B,CACxC,KAAmC,EACnC,gBAAkD;IAElD,IAAI,IAAI,GAAG,EAAE,CAAA;IACb,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,OAAO,GAA0B,EAAE,CAAA;IACzC,MAAM,eAAe,GAAG,IAAI,GAAG,EAAyB,CAAA;IAExD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAA;QAC7C,MAAM,UAAU,GAAG,eAAe,CAAC,IAAI,EAAE,aAAa,EAAE,gBAAgB,CAAC,CAAA;QACzE,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAC/B,UAAU,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE,WAAW,GAAG,KAAK,CAAU,CAAC,CAC7E,CAAA;QACD,MAAM,eAAe,GAAG,sBAAsB,CAAC,IAAI,EAAE,aAAa,EAAE,UAAU,CAAC,CAAA;QAC/E,MAAM,QAAQ,GAAG,sBAAsB,CAAC,IAAI,EAAE,eAAe,EAAE,iBAAiB,CAAC,CAAA;QAEjF,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAA;QACnF,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,CAAA;QACpD,IAAI,IAAI,aAAa,GAAG,QAAQ,CAAC,QAAQ,CAAA;QAEzC,KAAK,MAAM,gBAAgB,IAAI,UAAU,EAAE,CAAC;YAC1C,MAAM,MAAM,GAAG,eAAe,CAAC,gBAAgB,CAAC,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,WAAW,EAAE,CAAC,CAAA;YAC3F,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;gBACxB,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YACxC,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACpB,eAAe,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IAED,IAAI,eAAe,CAAC,IAAI,KAAK,gBAAgB,CAAC,MAAM,EAAE,CAAC;QACrD,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAA;IACvE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,CAAA;AACxC,CAAC;AAED,SAAS,uBAAuB,CAC9B,QAA+B,EAC/B,MAA4B,EAC5B,WAAwC;IAExC,MAAM,uBAAuB,GAAG,IAAI,GAAG,EAA8B,CAAA;IACrE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,uBAAuB,CAAA;IAChC,CAAC;IACD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CACjC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAU,CAAC,CACxE,CAAA;IACD,MAAM,iBAAiB,GAAG,IAAI,GAAG,EAAU,CAAA;IAE3C,KAAK,MAAM,UAAU,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,IAAI,CAC5C,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CACpD,EAAE,CAAC;QACF,IAAI,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YACjD,MAAM,IAAI,KAAK,CAAC,sCAAsC,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QAChF,CAAC;QACD,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC5C,MAAM,MAAM,GAAG,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC7D,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,MAAM,IAAI,KAAK,CAAC,4CAA4C,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QACtF,CAAC;QAED,MAAM,KAAK,GAAG,gBAAgB,CAAC,MAAM,EAAE,MAAM,CAAC,YAAY,CAAC,CAAA;QAC3D,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,MAAM,IAAI,KAAK,CAAC,UAAU,UAAU,CAAC,UAAU,sBAAsB,CAAC,CAAA;QACxE,CAAC;QACD,MAAM,OAAO,GAAG,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAA;QAC9D,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,CAAC,CAAA;IACnD,CAAC;IACD,OAAO,uBAAuB,CAAA;AAChC,CAAC;AAED,KAAK,UAAU,uBAAuB,CACpC,OAA+B,EAC/B,GAAiE;IAKjE,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;IAC1D,MAAM,SAAS,GAA4B,EAAE,CAAA;IAC7C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,kBAAkB,CAC7C,GAAG,EACH,MAAM,CAAC,OAAO,EACd,MAAM,CAAC,IAAI,EACX,MAAM,CAAC,QAAQ,CAChB,CAAA;YACD,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAA;QAChE,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;YACtB,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IACD,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAA;AACzC,CAAC;AAED;;;;GAIG;AACH,KAAK,UAAU,uBAAuB,CACpC,KAAqE,EACrE,GAAiE,EACjE,eAA4C,EAC5C,aAAsB;IAEtB,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAA;IAC5F,IAAI,eAAe,KAAK,SAAS,EAAE,CAAC;QAClC,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;QAC1D,OAAO,uBAAuB,CAAC,QAAQ,CAAC,mBAAmB,CAAC,QAAQ,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC,CAAA;IAClF,CAAC;IACD,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,wBAAwB,CAAC,CAAA;IACxD,MAAM,OAAO,GAAG,SAAS,CAAC,mBAAmB,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAA;IAC5D,MAAM,SAAS,GAAG,SAAS,CAAC,eAAe,CAAC,eAAe,CAAC,CAAA;IAC5D,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,MAAM,SAAS,CAAC,oBAAoB,CAAC,OAAO,EAAE,GAAG,EAAE;YACnE,SAAS;YACT,aAAa;SACd,CAAC,CAAA;QACF,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAC,EAAE,CAAA;IAC5C,CAAC;YAAS,CAAC;QACT,MAAM,SAAS,CAAC,OAAO,EAAE,CAAA;IAC3B,CAAC;AACH,CAAC;AAED,sFAAsF;AACtF,SAAS,kBAAkB,CAAC,OAAuC;IAIjE,MAAM,WAAW,GAAuB,EAAE,CAAA;IAC1C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;YACtB,SAAQ;QACV,CAAC;QACD,IAAI,CAAC;YACH,WAAW,CAAC,IAAI,CAAC,sBAAsB,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,SAAS,CAAC,CAAC,CAAA;QAChF,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;QACxB,CAAC;IACH,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,iBAAiB,EAAE,CAAA;AAC3C,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,QAAgB,EAChB,aAAwC,EACxC,OAAsC;IAEtC,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,aAAa,CAAA;IACnD,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,CAAA;IACzC,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAA;IAC5E,IAAI,CAAC;QACH,IAAI,iBAAiB,GAAG,CAAC,CAAA;QAEzB,MAAM,QAAQ,GAAG,MAAM,uBAAuB,CAAC,KAAK,EAAE,GAAG,EAAE,eAAe,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;QAC3F,MAAM,SAAS,GAAG,QAAQ,CAAC,SAAS,CAAA;QACpC,iBAAiB,IAAI,QAAQ,CAAC,iBAAiB,CAAA;QAE/C,MAAM,OAAO,GAAG,0BAA0B,CAAC,KAAK,EAAE,SAAS,CAAC,CAAA;QAC5D,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,MAAM,wBAAwB,CAC3D,OAAO,CAAC,IAAI,EACZ,OAAO,EACP,QAAQ,EACR,OAAO,CAAC,YAAY,CACrB,CAAA;QACD,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM;YAC9B,CAAC,CAAC,kBAAkB,CAAC,OAAO,CAAC,OAAO,CAAC;YACrC,CAAC,CAAC,EAAE,WAAW,EAAE,EAAE,EAAE,iBAAiB,EAAE,CAAC,EAAE,CAAA;QAC7C,MAAM,WAAW,GAAG,SAAS,CAAC,WAAW,CAAA;QACzC,iBAAiB,IAAI,SAAS,CAAC,iBAAiB,CAAA;QAChD,MAAM,iBAAiB,GAAG,uBAAuB,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,CAAA;QAC/E,OAAO;YACL,MAAM;YACN,UAAU;YACV,KAAK;YACL,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,iBAAiB;YACjB,iBAAiB;SAClB,CAAA;IACH,CAAC;YAAS,CAAC;QACT,IAAI,CAAC;YACH,GAAG,CAAC,OAAO,EAAE,CAAA;QACf,CAAC;QAAC,OAAO,YAAY,EAAE,CAAC;YACtB,MAAM,OAAO,GAAG,YAAY,YAAY,KAAK,CAAC,CAAC,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YAC3F,OAAO,CAAC,IAAI,CAAC,iDAAiD,OAAO,EAAE,CAAC,CAAA;QAC1E,CAAC;IACH,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { AtomicTextRange } from '../chunker/index.js';
|
|
2
|
+
export interface DocxBodyResult {
|
|
3
|
+
content: string;
|
|
4
|
+
atomicRanges: readonly AtomicTextRange[];
|
|
5
|
+
imageAnchors?: readonly {
|
|
6
|
+
offset: number;
|
|
7
|
+
imageIndex: number;
|
|
8
|
+
}[];
|
|
9
|
+
}
|
|
10
|
+
export declare function extractDocxCoreTitle(buffer: Buffer): Promise<string | undefined>;
|
|
11
|
+
export declare function convertDocxDocumentToText(document: Document): DocxBodyResult;
|
|
12
|
+
//# sourceMappingURL=docx-parser.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"docx-parser.d.ts","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;AAU1D,MAAM,WAAW,cAAc;IAC7B,OAAO,EAAE,MAAM,CAAA;IACf,YAAY,EAAE,SAAS,eAAe,EAAE,CAAA;IACxC,YAAY,CAAC,EAAE,SAAS;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,EAAE,CAAA;CACjE;AA6VD,wBAAsB,oBAAoB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,SAAS,CAAC,CAsBtF;AAsCD,wBAAgB,yBAAyB,CAAC,QAAQ,EAAE,QAAQ,GAAG,cAAc,CAmC5E"}
|
|
@@ -0,0 +1,375 @@
|
|
|
1
|
+
import { JSDOM } from 'jsdom';
|
|
2
|
+
import JSZip from 'jszip';
|
|
3
|
+
const CORE_TITLE_NAMESPACE = 'http://purl.org/dc/elements/1.1/';
|
|
4
|
+
const PROSE_BLOCK_TAGS = new Set(['H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'P', 'BLOCKQUOTE', 'PRE']);
|
|
5
|
+
const LIST_TAGS = new Set(['UL', 'OL']);
|
|
6
|
+
const SKIPPED_TAGS = new Set(['SCRIPT', 'STYLE', 'NOSCRIPT']);
|
|
7
|
+
const IMAGE_MARKER_PREFIX = '\u0000rag-image:';
|
|
8
|
+
const IMAGE_MARKER_SUFFIX = '\u0000';
|
|
9
|
+
const IMAGE_MARKER_PATTERN = new RegExp(`${IMAGE_MARKER_PREFIX}(\\d+)${IMAGE_MARKER_SUFFIX}`, 'gu');
|
|
10
|
+
/** DOM element check that narrows a `Node`, replacing a bare `nodeType` test. */
|
|
11
|
+
function isElement(node) {
|
|
12
|
+
return node.nodeType === 1;
|
|
13
|
+
}
|
|
14
|
+
function imageMarker(element) {
|
|
15
|
+
if (element.tagName !== 'IMG') {
|
|
16
|
+
return '';
|
|
17
|
+
}
|
|
18
|
+
const value = element.getAttribute('data-rag-image-index');
|
|
19
|
+
return value !== null && /^\d+$/.test(value)
|
|
20
|
+
? `${IMAGE_MARKER_PREFIX}${value}${IMAGE_MARKER_SUFFIX}`
|
|
21
|
+
: '';
|
|
22
|
+
}
|
|
23
|
+
function normalizeSingleLine(text) {
|
|
24
|
+
return text.replace(/\s+/g, ' ').trim();
|
|
25
|
+
}
|
|
26
|
+
/**
|
|
27
|
+
* Classify one node against the rules every DOCX serializer shares: text nodes
|
|
28
|
+
* emit their value, non-elements and skipped tags are dropped, an image
|
|
29
|
+
* placeholder emits its marker, and `<br>` is a line break the caller renders.
|
|
30
|
+
*/
|
|
31
|
+
function classifyNode(node) {
|
|
32
|
+
if (node.nodeType === 3) {
|
|
33
|
+
return { kind: 'emit', text: node.nodeValue ?? '' };
|
|
34
|
+
}
|
|
35
|
+
if (!isElement(node) || SKIPPED_TAGS.has(node.tagName)) {
|
|
36
|
+
return { kind: 'skip' };
|
|
37
|
+
}
|
|
38
|
+
const marker = imageMarker(node);
|
|
39
|
+
if (marker) {
|
|
40
|
+
return { kind: 'emit', text: marker };
|
|
41
|
+
}
|
|
42
|
+
if (node.tagName === 'BR') {
|
|
43
|
+
return { kind: 'break' };
|
|
44
|
+
}
|
|
45
|
+
return { kind: 'descend', element: node };
|
|
46
|
+
}
|
|
47
|
+
function normalizeTextContent(element) {
|
|
48
|
+
const parts = [];
|
|
49
|
+
const visit = (node) => {
|
|
50
|
+
const action = classifyNode(node);
|
|
51
|
+
if (action.kind === 'skip') {
|
|
52
|
+
return;
|
|
53
|
+
}
|
|
54
|
+
if (action.kind === 'emit') {
|
|
55
|
+
parts.push(action.text);
|
|
56
|
+
return;
|
|
57
|
+
}
|
|
58
|
+
if (action.kind === 'break') {
|
|
59
|
+
parts.push(' ');
|
|
60
|
+
return;
|
|
61
|
+
}
|
|
62
|
+
const child = action.element;
|
|
63
|
+
const ownsTextBoundary = PROSE_BLOCK_TAGS.has(child.tagName) || LIST_TAGS.has(child.tagName) || child.tagName === 'LI';
|
|
64
|
+
if (ownsTextBoundary) {
|
|
65
|
+
parts.push(' ');
|
|
66
|
+
}
|
|
67
|
+
for (const descendant of child.childNodes) {
|
|
68
|
+
visit(descendant);
|
|
69
|
+
}
|
|
70
|
+
if (ownsTextBoundary) {
|
|
71
|
+
parts.push(' ');
|
|
72
|
+
}
|
|
73
|
+
};
|
|
74
|
+
for (const child of element.childNodes) {
|
|
75
|
+
visit(child);
|
|
76
|
+
}
|
|
77
|
+
return normalizeSingleLine(parts.join(''));
|
|
78
|
+
}
|
|
79
|
+
function serializeInlineText(element) {
|
|
80
|
+
const parts = [];
|
|
81
|
+
const visit = (node) => {
|
|
82
|
+
const action = classifyNode(node);
|
|
83
|
+
if (action.kind === 'skip') {
|
|
84
|
+
return;
|
|
85
|
+
}
|
|
86
|
+
if (action.kind === 'emit') {
|
|
87
|
+
parts.push(action.text);
|
|
88
|
+
return;
|
|
89
|
+
}
|
|
90
|
+
if (action.kind === 'break') {
|
|
91
|
+
parts.push('\n');
|
|
92
|
+
return;
|
|
93
|
+
}
|
|
94
|
+
for (const descendant of action.element.childNodes) {
|
|
95
|
+
visit(descendant);
|
|
96
|
+
}
|
|
97
|
+
};
|
|
98
|
+
for (const child of element.childNodes) {
|
|
99
|
+
visit(child);
|
|
100
|
+
}
|
|
101
|
+
const lines = parts.join('').split('\n').map(normalizeSingleLine);
|
|
102
|
+
while (lines[0] === '') {
|
|
103
|
+
lines.shift();
|
|
104
|
+
}
|
|
105
|
+
while (lines.at(-1) === '') {
|
|
106
|
+
lines.pop();
|
|
107
|
+
}
|
|
108
|
+
return lines.join('\n');
|
|
109
|
+
}
|
|
110
|
+
function serializeList(list) {
|
|
111
|
+
const lines = [];
|
|
112
|
+
const serializeListElement = (currentList) => {
|
|
113
|
+
for (const child of currentList.children) {
|
|
114
|
+
if (child.tagName === 'LI') {
|
|
115
|
+
serializeListItem(child, lines, serializeListElement);
|
|
116
|
+
}
|
|
117
|
+
}
|
|
118
|
+
};
|
|
119
|
+
serializeListElement(list);
|
|
120
|
+
return lines.join('\n');
|
|
121
|
+
}
|
|
122
|
+
/**
|
|
123
|
+
* Append one `<li>`'s lines. A `<br>` or a nested list ends the current line,
|
|
124
|
+
* and the nested list is serialized in place through `serializeNested`.
|
|
125
|
+
*/
|
|
126
|
+
function serializeListItem(item, lines, serializeNested) {
|
|
127
|
+
let currentLine = '';
|
|
128
|
+
const flush = () => {
|
|
129
|
+
const normalized = normalizeSingleLine(currentLine);
|
|
130
|
+
if (normalized) {
|
|
131
|
+
lines.push(normalized);
|
|
132
|
+
}
|
|
133
|
+
currentLine = '';
|
|
134
|
+
};
|
|
135
|
+
const visitItemNode = (node) => {
|
|
136
|
+
const action = classifyNode(node);
|
|
137
|
+
if (action.kind === 'skip') {
|
|
138
|
+
return;
|
|
139
|
+
}
|
|
140
|
+
if (action.kind === 'emit') {
|
|
141
|
+
currentLine += action.text;
|
|
142
|
+
return;
|
|
143
|
+
}
|
|
144
|
+
if (action.kind === 'break') {
|
|
145
|
+
flush();
|
|
146
|
+
return;
|
|
147
|
+
}
|
|
148
|
+
if (LIST_TAGS.has(action.element.tagName)) {
|
|
149
|
+
flush();
|
|
150
|
+
serializeNested(action.element);
|
|
151
|
+
return;
|
|
152
|
+
}
|
|
153
|
+
for (const descendant of action.element.childNodes) {
|
|
154
|
+
visitItemNode(descendant);
|
|
155
|
+
}
|
|
156
|
+
};
|
|
157
|
+
for (const itemChild of item.childNodes) {
|
|
158
|
+
visitItemNode(itemChild);
|
|
159
|
+
}
|
|
160
|
+
flush();
|
|
161
|
+
}
|
|
162
|
+
function directTableRows(table) {
|
|
163
|
+
return Array.from(table.querySelectorAll('tr')).filter((row) => row.closest('table') === table);
|
|
164
|
+
}
|
|
165
|
+
function directRowCells(row) {
|
|
166
|
+
return Array.from(row.children).filter((cell) => cell.tagName === 'TH' || cell.tagName === 'TD');
|
|
167
|
+
}
|
|
168
|
+
function hasSpanningCell(cells) {
|
|
169
|
+
return cells.some((cell) => {
|
|
170
|
+
const rowSpan = Number.parseInt(cell.getAttribute('rowspan') ?? '1', 10);
|
|
171
|
+
const columnSpan = Number.parseInt(cell.getAttribute('colspan') ?? '1', 10);
|
|
172
|
+
return rowSpan > 1 || columnSpan > 1;
|
|
173
|
+
});
|
|
174
|
+
}
|
|
175
|
+
function isSupportedTable(table, rows) {
|
|
176
|
+
if (table.querySelector('table') !== null || rows.length < 2) {
|
|
177
|
+
return false;
|
|
178
|
+
}
|
|
179
|
+
const rowCells = rows.map(directRowCells);
|
|
180
|
+
const columnCount = rowCells[0]?.length ?? 0;
|
|
181
|
+
return (columnCount > 0 &&
|
|
182
|
+
rowCells.every((cells) => cells.length === columnCount && !hasSpanningCell(cells)));
|
|
183
|
+
}
|
|
184
|
+
function collectTextTokens(element) {
|
|
185
|
+
const tokens = [];
|
|
186
|
+
const visit = (node) => {
|
|
187
|
+
if (node.nodeType === 3) {
|
|
188
|
+
const token = normalizeSingleLine(node.nodeValue ?? '');
|
|
189
|
+
if (token) {
|
|
190
|
+
tokens.push(token);
|
|
191
|
+
}
|
|
192
|
+
return;
|
|
193
|
+
}
|
|
194
|
+
if (!isElement(node)) {
|
|
195
|
+
return;
|
|
196
|
+
}
|
|
197
|
+
const child = node;
|
|
198
|
+
if (SKIPPED_TAGS.has(child.tagName)) {
|
|
199
|
+
return;
|
|
200
|
+
}
|
|
201
|
+
const marker = imageMarker(child);
|
|
202
|
+
if (marker) {
|
|
203
|
+
tokens.push(marker);
|
|
204
|
+
return;
|
|
205
|
+
}
|
|
206
|
+
for (const descendant of child.childNodes) {
|
|
207
|
+
visit(descendant);
|
|
208
|
+
}
|
|
209
|
+
};
|
|
210
|
+
visit(element);
|
|
211
|
+
return tokens.join(' ');
|
|
212
|
+
}
|
|
213
|
+
function emitTable(table) {
|
|
214
|
+
const rows = directTableRows(table);
|
|
215
|
+
const headerRow = rows[0];
|
|
216
|
+
if (!isSupportedTable(table, rows) || headerRow === undefined) {
|
|
217
|
+
const text = collectTextTokens(table);
|
|
218
|
+
return text ? [{ text, atomic: false }] : [];
|
|
219
|
+
}
|
|
220
|
+
// Ordinary Word tables often contain only td cells, so the approved contract
|
|
221
|
+
// deliberately treats the first physical row as labels even without th markup.
|
|
222
|
+
const headerMarkers = [];
|
|
223
|
+
const headers = directRowCells(headerRow).map((cell, index) => {
|
|
224
|
+
const text = normalizeTextContent(cell);
|
|
225
|
+
headerMarkers.push(...(text.match(IMAGE_MARKER_PATTERN) ?? []));
|
|
226
|
+
return text.replace(IMAGE_MARKER_PATTERN, '') || `Column ${index + 1}`;
|
|
227
|
+
});
|
|
228
|
+
const rowBlocks = rows.slice(1).flatMap((row) => {
|
|
229
|
+
const values = directRowCells(row);
|
|
230
|
+
const text = values
|
|
231
|
+
.map((cell, index) => `${headers[index] ?? `Column ${index + 1}`}: ${normalizeTextContent(cell)}`)
|
|
232
|
+
.join('\n');
|
|
233
|
+
return text ? [{ text, atomic: true }] : [];
|
|
234
|
+
});
|
|
235
|
+
const first = rowBlocks[0];
|
|
236
|
+
if (first && headerMarkers.length > 0) {
|
|
237
|
+
first.text = headerMarkers.join('') + first.text;
|
|
238
|
+
}
|
|
239
|
+
return rowBlocks;
|
|
240
|
+
}
|
|
241
|
+
function hasRecognizedBlockDescendant(element) {
|
|
242
|
+
return Array.from(element.querySelectorAll('*')).some((descendant) => PROSE_BLOCK_TAGS.has(descendant.tagName) ||
|
|
243
|
+
LIST_TAGS.has(descendant.tagName) ||
|
|
244
|
+
descendant.tagName === 'TABLE');
|
|
245
|
+
}
|
|
246
|
+
/** How a recognized block element turns into text, or `null` when it is not one. */
|
|
247
|
+
function blockSerializerFor(element) {
|
|
248
|
+
if (PROSE_BLOCK_TAGS.has(element.tagName)) {
|
|
249
|
+
return serializeInlineText;
|
|
250
|
+
}
|
|
251
|
+
if (LIST_TAGS.has(element.tagName)) {
|
|
252
|
+
return serializeList;
|
|
253
|
+
}
|
|
254
|
+
return null;
|
|
255
|
+
}
|
|
256
|
+
function emitDocumentBlocks(document) {
|
|
257
|
+
const blocks = [];
|
|
258
|
+
/** Push one non-empty prose block. */
|
|
259
|
+
const pushText = (text) => {
|
|
260
|
+
if (text) {
|
|
261
|
+
blocks.push({ text, atomic: false });
|
|
262
|
+
}
|
|
263
|
+
};
|
|
264
|
+
const emitElement = (element) => {
|
|
265
|
+
if (SKIPPED_TAGS.has(element.tagName)) {
|
|
266
|
+
return;
|
|
267
|
+
}
|
|
268
|
+
const serialize = blockSerializerFor(element);
|
|
269
|
+
if (serialize !== null) {
|
|
270
|
+
pushText(serialize(element));
|
|
271
|
+
return;
|
|
272
|
+
}
|
|
273
|
+
if (element.tagName === 'TABLE') {
|
|
274
|
+
blocks.push(...emitTable(element));
|
|
275
|
+
return;
|
|
276
|
+
}
|
|
277
|
+
if (!hasRecognizedBlockDescendant(element)) {
|
|
278
|
+
pushText(normalizeTextContent(element));
|
|
279
|
+
return;
|
|
280
|
+
}
|
|
281
|
+
for (const child of element.childNodes) {
|
|
282
|
+
emitNode(child);
|
|
283
|
+
}
|
|
284
|
+
};
|
|
285
|
+
const emitNode = (node) => {
|
|
286
|
+
if (node.nodeType === 3) {
|
|
287
|
+
pushText(normalizeSingleLine(node.nodeValue ?? ''));
|
|
288
|
+
return;
|
|
289
|
+
}
|
|
290
|
+
if (isElement(node)) {
|
|
291
|
+
emitElement(node);
|
|
292
|
+
}
|
|
293
|
+
};
|
|
294
|
+
for (const node of document.body.childNodes) {
|
|
295
|
+
emitNode(node);
|
|
296
|
+
}
|
|
297
|
+
return blocks;
|
|
298
|
+
}
|
|
299
|
+
export async function extractDocxCoreTitle(buffer) {
|
|
300
|
+
const zip = await JSZip.loadAsync(buffer);
|
|
301
|
+
const coreProperties = zip.file('docProps/core.xml');
|
|
302
|
+
if (!coreProperties) {
|
|
303
|
+
return undefined;
|
|
304
|
+
}
|
|
305
|
+
const xml = await coreProperties.async('string');
|
|
306
|
+
try {
|
|
307
|
+
const document = new JSDOM(xml, { contentType: 'text/xml' }).window.document;
|
|
308
|
+
if (document.querySelector('parsererror')) {
|
|
309
|
+
return undefined;
|
|
310
|
+
}
|
|
311
|
+
const title = document.getElementsByTagNameNS(CORE_TITLE_NAMESPACE, 'title').item(0);
|
|
312
|
+
const normalized = normalizeSingleLine(title?.textContent ?? '');
|
|
313
|
+
return normalized || undefined;
|
|
314
|
+
}
|
|
315
|
+
catch {
|
|
316
|
+
// Core properties are optional display metadata; malformed XML falls
|
|
317
|
+
// through to the heading/filename chain without rejecting readable body text.
|
|
318
|
+
return undefined;
|
|
319
|
+
}
|
|
320
|
+
}
|
|
321
|
+
/** Marker-free text, with each line normalized as the body text is. */
|
|
322
|
+
function stripImageMarkers(text) {
|
|
323
|
+
return text.replace(IMAGE_MARKER_PATTERN, '').split('\n').map(normalizeSingleLine).join('\n');
|
|
324
|
+
}
|
|
325
|
+
/** Where each of a block's image markers lands once markers are stripped. */
|
|
326
|
+
function anchorsWithinBlock(blockText, matches, start) {
|
|
327
|
+
return matches.map((match) => ({
|
|
328
|
+
offset: start + stripImageMarkers(blockText.slice(0, match.index ?? 0)).length,
|
|
329
|
+
imageIndex: Number(match[1]),
|
|
330
|
+
}));
|
|
331
|
+
}
|
|
332
|
+
/** Keep the first anchor per image index, in image order. */
|
|
333
|
+
function dedupeAnchors(imageAnchors) {
|
|
334
|
+
const seen = new Set();
|
|
335
|
+
return [...imageAnchors]
|
|
336
|
+
.sort((left, right) => left.imageIndex - right.imageIndex)
|
|
337
|
+
.filter((anchor) => {
|
|
338
|
+
if (seen.has(anchor.imageIndex)) {
|
|
339
|
+
return false;
|
|
340
|
+
}
|
|
341
|
+
seen.add(anchor.imageIndex);
|
|
342
|
+
return true;
|
|
343
|
+
});
|
|
344
|
+
}
|
|
345
|
+
export function convertDocxDocumentToText(document) {
|
|
346
|
+
let content = '';
|
|
347
|
+
const atomicRanges = [];
|
|
348
|
+
const imageAnchors = [];
|
|
349
|
+
for (const block of emitDocumentBlocks(document)) {
|
|
350
|
+
const matches = [...block.text.matchAll(IMAGE_MARKER_PATTERN)];
|
|
351
|
+
const blockText = matches.length === 0 ? block.text : stripImageMarkers(block.text);
|
|
352
|
+
// A block that is nothing but markers contributes no text: its images
|
|
353
|
+
// anchor at the current end of the document instead.
|
|
354
|
+
if (matches.length > 0 && !blockText) {
|
|
355
|
+
imageAnchors.push(...matches.map((match) => ({ offset: content.length, imageIndex: Number(match[1]) })));
|
|
356
|
+
continue;
|
|
357
|
+
}
|
|
358
|
+
if (content) {
|
|
359
|
+
content += '\n\n';
|
|
360
|
+
}
|
|
361
|
+
const start = content.length;
|
|
362
|
+
imageAnchors.push(...anchorsWithinBlock(block.text, matches, start));
|
|
363
|
+
content += blockText;
|
|
364
|
+
if (block.atomic) {
|
|
365
|
+
atomicRanges.push({ start, end: content.length });
|
|
366
|
+
}
|
|
367
|
+
}
|
|
368
|
+
const uniqueImageAnchors = dedupeAnchors(imageAnchors);
|
|
369
|
+
return {
|
|
370
|
+
content,
|
|
371
|
+
atomicRanges,
|
|
372
|
+
...(uniqueImageAnchors.length === 0 ? {} : { imageAnchors: uniqueImageAnchors }),
|
|
373
|
+
};
|
|
374
|
+
}
|
|
375
|
+
//# sourceMappingURL=docx-parser.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"docx-parser.js","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,KAAK,MAAM,OAAO,CAAA;AAGzB,MAAM,oBAAoB,GAAG,kCAAkC,CAAA;AAC/D,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,YAAY,EAAE,KAAK,CAAC,CAAC,CAAA;AAChG,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAA;AACvC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,UAAU,CAAC,CAAC,CAAA;AAC7D,MAAM,mBAAmB,GAAG,kBAAkB,CAAA;AAC9C,MAAM,mBAAmB,GAAG,QAAQ,CAAA;AACpC,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,mBAAmB,SAAS,mBAAmB,EAAE,EAAE,IAAI,CAAC,CAAA;AAanG,iFAAiF;AACjF,SAAS,SAAS,CAAC,IAAU;IAC3B,OAAO,IAAI,CAAC,QAAQ,KAAK,CAAC,CAAA;AAC5B,CAAC;AAED,SAAS,WAAW,CAAC,OAAgB;IACnC,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,EAAE,CAAC;QAC9B,OAAO,EAAE,CAAA;IACX,CAAC;IACD,MAAM,KAAK,GAAG,OAAO,CAAC,YAAY,CAAC,sBAAsB,CAAC,CAAA;IAC1D,OAAO,KAAK,KAAK,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC;QAC1C,CAAC,CAAC,GAAG,mBAAmB,GAAG,KAAK,GAAG,mBAAmB,EAAE;QACxD,CAAC,CAAC,EAAE,CAAA;AACR,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AACzC,CAAC;AASD;;;;GAIG;AACH,SAAS,YAAY,CAAC,IAAU;IAC9B,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,IAAI,EAAE,EAAE,CAAA;IACrD,CAAC;IACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QACvD,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;IACzB,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,IAAI,CAAC,CAAA;IAChC,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;IACvC,CAAC;IACD,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;QAC1B,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,CAAA;IAC1B,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAA;AAC3C,CAAC;AAED,SAAS,oBAAoB,CAAC,OAAgB;IAC5C,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACvB,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACf,OAAM;QACR,CAAC;QAED,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAA;QAC5B,MAAM,gBAAgB,GACpB,gBAAgB,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAA;QAC/F,IAAI,gBAAgB,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACjB,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YAC1C,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;QACD,IAAI,gBAAgB,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACjB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvC,KAAK,CAAC,KAAK,CAAC,CAAA;IACd,CAAC;IACD,OAAO,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAA;AAC5C,CAAC;AAED,SAAS,mBAAmB,CAAC,OAAgB;IAC3C,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACvB,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAChB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YACnD,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvC,KAAK,CAAC,KAAK,CAAC,CAAA;IACd,CAAC;IAED,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAA;IACjE,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QACvB,KAAK,CAAC,KAAK,EAAE,CAAA;IACf,CAAC;IACD,OAAO,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QAC3B,KAAK,CAAC,GAAG,EAAE,CAAA;IACb,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,aAAa,CAAC,IAAa;IAClC,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,oBAAoB,GAAG,CAAC,WAAoB,EAAQ,EAAE;QAC1D,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,QAAQ,EAAE,CAAC;YACzC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3B,iBAAiB,CAAC,KAAK,EAAE,KAAK,EAAE,oBAAoB,CAAC,CAAA;YACvD,CAAC;QACH,CAAC;IACH,CAAC,CAAA;IAED,oBAAoB,CAAC,IAAI,CAAC,CAAA;IAC1B,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED;;;GAGG;AACH,SAAS,iBAAiB,CACxB,IAAa,EACb,KAAe,EACf,eAAwC;IAExC,IAAI,WAAW,GAAG,EAAE,CAAA;IACpB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,MAAM,UAAU,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAA;QACnD,IAAI,UAAU,EAAE,CAAC;YACf,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,CAAC;QACD,WAAW,GAAG,EAAE,CAAA;IAClB,CAAC,CAAA;IAED,MAAM,aAAa,GAAG,CAAC,IAAU,EAAQ,EAAE;QACzC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,WAAW,IAAI,MAAM,CAAC,IAAI,CAAA;YAC1B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,EAAE,CAAA;YACP,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YAC1C,KAAK,EAAE,CAAA;YACP,eAAe,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAC/B,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YACnD,aAAa,CAAC,UAAU,CAAC,CAAA;QAC3B,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;QACxC,aAAa,CAAC,SAAS,CAAC,CAAA;IAC1B,CAAC;IACD,KAAK,EAAE,CAAA;AACT,CAAC;AAED,SAAS,eAAe,CAAC,KAAc;IACrC,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,CAAA;AACjG,CAAC;AAED,SAAS,cAAc,CAAC,GAAY;IAClC,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;AAClG,CAAC;AAED,SAAS,eAAe,CAAC,KAAyB;IAChD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE;QACzB,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QACxE,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QAC3E,OAAO,OAAO,GAAG,CAAC,IAAI,UAAU,GAAG,CAAC,CAAA;IACtC,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAc,EAAE,IAAwB;IAChE,IAAI,KAAK,CAAC,aAAa,CAAC,OAAO,CAAC,KAAK,IAAI,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7D,OAAO,KAAK,CAAA;IACd,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;IACzC,MAAM,WAAW,GAAG,QAAQ,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAA;IAC5C,OAAO,CACL,WAAW,GAAG,CAAC;QACf,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,WAAW,IAAI,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC,CACnF,CAAA;AACH,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAgB;IACzC,MAAM,MAAM,GAAa,EAAE,CAAA;IAC3B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,KAAK,GAAG,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YACvD,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACpB,CAAC;YACD,OAAM;QACR,CAAC;QACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,OAAM;QACR,CAAC;QAED,MAAM,KAAK,GAAG,IAAI,CAAA;QAClB,IAAI,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;YACpC,OAAM;QACR,CAAC;QACD,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,MAAM,EAAE,CAAC;YACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACnB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YAC1C,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IACD,KAAK,CAAC,OAAO,CAAC,CAAA;IACd,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,SAAS,CAAC,KAAc;IAC/B,MAAM,IAAI,GAAG,eAAe,CAAC,KAAK,CAAC,CAAA;IACnC,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;IACzB,IAAI,CAAC,gBAAgB,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC9D,MAAM,IAAI,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAA;QACrC,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC9C,CAAC;IAED,6EAA6E;IAC7E,+EAA+E;IAC/E,MAAM,aAAa,GAAa,EAAE,CAAA;IAClC,MAAM,OAAO,GAAG,cAAc,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5D,MAAM,IAAI,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAA;QACvC,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;QAC/D,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,CAAA;IACxE,CAAC,CAAC,CAAA;IAEF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC9C,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,CAAA;QAClC,MAAM,IAAI,GAAG,MAAM;aAChB,GAAG,CACF,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,KAAK,oBAAoB,CAAC,IAAI,CAAC,EAAE,CAC7F;aACA,IAAI,CAAC,IAAI,CAAC,CAAA;QACb,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC7C,CAAC,CAAC,CAAA;IACF,MAAM,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,CAAA;IAC1B,IAAI,KAAK,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtC,KAAK,CAAC,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAA;IAClD,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,SAAS,4BAA4B,CAAC,OAAgB;IACpD,OAAO,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CACnD,CAAC,UAAU,EAAE,EAAE,CACb,gBAAgB,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACxC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACjC,UAAU,CAAC,OAAO,KAAK,OAAO,CACjC,CAAA;AACH,CAAC;AAED,oFAAoF;AACpF,SAAS,kBAAkB,CAAC,OAAgB;IAC1C,IAAI,gBAAgB,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1C,OAAO,mBAAmB,CAAA;IAC5B,CAAC;IACD,IAAI,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACnC,OAAO,aAAa,CAAA;IACtB,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,kBAAkB,CAAC,QAAkB;IAC5C,MAAM,MAAM,GAAmB,EAAE,CAAA;IAEjC,sCAAsC;IACtC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAQ,EAAE;QACtC,IAAI,IAAI,EAAE,CAAC;YACT,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;QACtC,CAAC;IACH,CAAC,CAAA;IAED,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAQ,EAAE;QAC7C,IAAI,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YACtC,OAAM;QACR,CAAC;QACD,MAAM,SAAS,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAA;QAC7C,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACvB,QAAQ,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAA;YAC5B,OAAM;QACR,CAAC;QACD,IAAI,OAAO,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;YAChC,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAA;YAClC,OAAM;QACR,CAAC;QACD,IAAI,CAAC,4BAA4B,CAAC,OAAO,CAAC,EAAE,CAAC;YAC3C,QAAQ,CAAC,oBAAoB,CAAC,OAAO,CAAC,CAAC,CAAA;YACvC,OAAM;QACR,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;YACvC,QAAQ,CAAC,KAAK,CAAC,CAAA;QACjB,CAAC;IACH,CAAC,CAAA;IAED,MAAM,QAAQ,GAAG,CAAC,IAAU,EAAQ,EAAE;QACpC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,QAAQ,CAAC,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,CAAA;YACnD,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACpB,WAAW,CAAC,IAAI,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;QAC5C,QAAQ,CAAC,IAAI,CAAC,CAAA;IAChB,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,oBAAoB,CAAC,MAAc;IACvD,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,cAAc,GAAG,GAAG,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAA;IACpD,IAAI,CAAC,cAAc,EAAE,CAAC;QACpB,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,cAAc,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAChD,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,IAAI,KAAK,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,UAAU,EAAE,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAA;QAC5E,IAAI,QAAQ,CAAC,aAAa,CAAC,aAAa,CAAC,EAAE,CAAC;YAC1C,OAAO,SAAS,CAAA;QAClB,CAAC;QAED,MAAM,KAAK,GAAG,QAAQ,CAAC,sBAAsB,CAAC,oBAAoB,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACpF,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,EAAE,WAAW,IAAI,EAAE,CAAC,CAAA;QAChE,OAAO,UAAU,IAAI,SAAS,CAAA;IAChC,CAAC;IAAC,MAAM,CAAC;QACP,qEAAqE;QACrE,8EAA8E;QAC9E,OAAO,SAAS,CAAA;IAClB,CAAC;AACH,CAAC;AAED,uEAAuE;AACvE,SAAS,iBAAiB,CAAC,IAAY;IACrC,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AAC/F,CAAC;AAED,6EAA6E;AAC7E,SAAS,kBAAkB,CACzB,SAAiB,EACjB,OAAmC,EACnC,KAAa;IAEb,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7B,MAAM,EAAE,KAAK,GAAG,iBAAiB,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM;QAC9E,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KAC7B,CAAC,CAAC,CAAA;AACL,CAAC;AAED,6DAA6D;AAC7D,SAAS,aAAa,CAAC,YAAoC;IACzD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,OAAO,CAAC,GAAG,YAAY,CAAC;SACrB,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CAAC;SACzD,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE;QACjB,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;YAChC,OAAO,KAAK,CAAA;QACd,CAAC;QACD,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;QAC3B,OAAO,IAAI,CAAA;IACb,CAAC,CAAC,CAAA;AACN,CAAC;AAOD,MAAM,UAAU,yBAAyB,CAAC,QAAkB;IAC1D,IAAI,OAAO,GAAG,EAAE,CAAA;IAChB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,YAAY,GAAkB,EAAE,CAAA;IAEtC,KAAK,MAAM,KAAK,IAAI,kBAAkB,CAAC,QAAQ,CAAC,EAAE,CAAC;QACjD,MAAM,OAAO,GAAG,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC,CAAA;QAC9D,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,iBAAiB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAEnF,sEAAsE;QACtE,qDAAqD;QACrD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;YACrC,YAAY,CAAC,IAAI,CACf,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CACtF,CAAA;YACD,SAAQ;QACV,CAAC;QAED,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,IAAI,MAAM,CAAA;QACnB,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;QAC5B,YAAY,CAAC,IAAI,CAAC,GAAG,kBAAkB,CAAC,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAA;QACpE,OAAO,IAAI,SAAS,CAAA;QACpB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;YACjB,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;QACnD,CAAC;IACH,CAAC;IAED,MAAM,kBAAkB,GAAG,aAAa,CAAC,YAAY,CAAC,CAAA;IACtD,OAAO;QACL,OAAO;QACP,YAAY;QACZ,GAAG,CAAC,kBAAkB,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,kBAAkB,EAAE,CAAC;KACjF,CAAA;AACH,CAAC"}
|
|
@@ -0,0 +1,16 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Parse HTML content and extract main content as Markdown
|
|
3
|
+
*
|
|
4
|
+
* Flow:
|
|
5
|
+
* 1. HTML string → JSDOM (DOM creation)
|
|
6
|
+
* 2. JSDOM → Readability (main content extraction, noise removal)
|
|
7
|
+
* 3. Readability result → Turndown (Markdown conversion)
|
|
8
|
+
* 4. Title extracted separately via extractHtmlTitle (NOT prepended to content)
|
|
9
|
+
*
|
|
10
|
+
* @param url - Source URL (used for resolving relative links)
|
|
11
|
+
*/
|
|
12
|
+
export declare function parseHtml(html: string, url: string): Promise<{
|
|
13
|
+
content: string;
|
|
14
|
+
title: string;
|
|
15
|
+
}>;
|
|
16
|
+
//# sourceMappingURL=html-parser.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"html-parser.d.ts","sourceRoot":"","sources":["../../src/parser/html-parser.ts"],"names":[],"mappings":"AA4EA;;;;;;;;;;GAUG;AACH,wBAAsB,SAAS,CAC7B,IAAI,EAAE,MAAM,EACZ,GAAG,EAAE,MAAM,GACV,OAAO,CAAC;IAAE,OAAO,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CA0D7C"}
|
|
@@ -0,0 +1,119 @@
|
|
|
1
|
+
// HTML Parser using Readability and Turndown
|
|
2
|
+
// Extracts main content from HTML and converts to Markdown
|
|
3
|
+
import { Readability } from '@mozilla/readability';
|
|
4
|
+
import { JSDOM } from 'jsdom';
|
|
5
|
+
import TurndownService from 'turndown';
|
|
6
|
+
import { isRecord } from '../utils/type-guards.js';
|
|
7
|
+
import { extractHtmlTitle } from './title-extractor.js';
|
|
8
|
+
/** DOM element check that narrows a `Node`, replacing a bare `nodeType` test. */
|
|
9
|
+
function isElement(node) {
|
|
10
|
+
return node.nodeType === 1;
|
|
11
|
+
}
|
|
12
|
+
/** Readability returns `null` when it cannot extract an article. */
|
|
13
|
+
function toReadabilityResult(parsed) {
|
|
14
|
+
if (!isRecord(parsed)) {
|
|
15
|
+
return null;
|
|
16
|
+
}
|
|
17
|
+
const { title, content } = parsed;
|
|
18
|
+
if (typeof title !== 'string' || typeof content !== 'string') {
|
|
19
|
+
return null;
|
|
20
|
+
}
|
|
21
|
+
return { title, content };
|
|
22
|
+
}
|
|
23
|
+
// ============================================
|
|
24
|
+
// Turndown Service Configuration
|
|
25
|
+
// ============================================
|
|
26
|
+
/**
|
|
27
|
+
* Create and configure Turndown service for HTML to Markdown conversion
|
|
28
|
+
*/
|
|
29
|
+
function createTurndownService() {
|
|
30
|
+
const turndownService = new TurndownService({
|
|
31
|
+
headingStyle: 'atx', // Use # style headings
|
|
32
|
+
codeBlockStyle: 'fenced', // Use ``` for code blocks
|
|
33
|
+
bulletListMarker: '-', // Use - for bullet lists
|
|
34
|
+
emDelimiter: '_', // Use _ for emphasis
|
|
35
|
+
strongDelimiter: '**', // Use ** for bold
|
|
36
|
+
});
|
|
37
|
+
// Keep code blocks intact
|
|
38
|
+
turndownService.addRule('codeBlocks', {
|
|
39
|
+
filter: ['pre'],
|
|
40
|
+
replacement: (_content, node) => {
|
|
41
|
+
if (!isElement(node)) {
|
|
42
|
+
return '';
|
|
43
|
+
}
|
|
44
|
+
const element = node;
|
|
45
|
+
const codeElement = element.querySelector('code');
|
|
46
|
+
const code = codeElement ? codeElement.textContent : element.textContent;
|
|
47
|
+
const language = codeElement?.className?.replace('language-', '') || '';
|
|
48
|
+
return `\n\`\`\`${language}\n${code?.trim() || ''}\n\`\`\`\n`;
|
|
49
|
+
},
|
|
50
|
+
});
|
|
51
|
+
return turndownService;
|
|
52
|
+
}
|
|
53
|
+
// ============================================
|
|
54
|
+
// HTML Parser
|
|
55
|
+
// ============================================
|
|
56
|
+
/**
|
|
57
|
+
* Parse HTML content and extract main content as Markdown
|
|
58
|
+
*
|
|
59
|
+
* Flow:
|
|
60
|
+
* 1. HTML string → JSDOM (DOM creation)
|
|
61
|
+
* 2. JSDOM → Readability (main content extraction, noise removal)
|
|
62
|
+
* 3. Readability result → Turndown (Markdown conversion)
|
|
63
|
+
* 4. Title extracted separately via extractHtmlTitle (NOT prepended to content)
|
|
64
|
+
*
|
|
65
|
+
* @param url - Source URL (used for resolving relative links)
|
|
66
|
+
*/
|
|
67
|
+
export async function parseHtml(html, url) {
|
|
68
|
+
// Handle empty or whitespace-only HTML
|
|
69
|
+
if (!html || html.trim().length === 0) {
|
|
70
|
+
return { content: '', title: '' };
|
|
71
|
+
}
|
|
72
|
+
try {
|
|
73
|
+
// Create DOM from HTML string
|
|
74
|
+
const dom = new JSDOM(html, {
|
|
75
|
+
url,
|
|
76
|
+
// Enable features needed for Readability
|
|
77
|
+
runScripts: 'outside-only',
|
|
78
|
+
});
|
|
79
|
+
const document = dom.window.document;
|
|
80
|
+
// Use Readability to extract main content
|
|
81
|
+
const reader = new Readability(document, {
|
|
82
|
+
keepClasses: false,
|
|
83
|
+
debug: false,
|
|
84
|
+
});
|
|
85
|
+
const article = toReadabilityResult(reader.parse());
|
|
86
|
+
// If Readability couldn't extract content, fall back to body text
|
|
87
|
+
if (!article?.content) {
|
|
88
|
+
// Try to get body content directly
|
|
89
|
+
const bodyContent = document.body?.innerHTML || '';
|
|
90
|
+
if (!bodyContent.trim()) {
|
|
91
|
+
return { content: '', title: '' };
|
|
92
|
+
}
|
|
93
|
+
// Convert raw body HTML to Markdown
|
|
94
|
+
const turndownService = createTurndownService();
|
|
95
|
+
return { content: turndownService.turndown(bodyContent).trim(), title: '' };
|
|
96
|
+
}
|
|
97
|
+
// Convert extracted HTML content to Markdown
|
|
98
|
+
const turndownService = createTurndownService();
|
|
99
|
+
const markdown = turndownService.turndown(article.content);
|
|
100
|
+
// Extract title separately (NOT prepended to markdown content)
|
|
101
|
+
// Use URL-derived filename as fallback when Readability has no title
|
|
102
|
+
let urlFileName = '';
|
|
103
|
+
try {
|
|
104
|
+
urlFileName = new URL(url).pathname.split('/').filter(Boolean).pop() || '';
|
|
105
|
+
}
|
|
106
|
+
catch {
|
|
107
|
+
// Non-URL string, empty fallback
|
|
108
|
+
}
|
|
109
|
+
const titleResult = extractHtmlTitle(article.title || '', urlFileName);
|
|
110
|
+
const title = titleResult.title;
|
|
111
|
+
return { content: markdown.trim(), title };
|
|
112
|
+
}
|
|
113
|
+
catch (error) {
|
|
114
|
+
// Log error but don't throw - return empty values for graceful degradation
|
|
115
|
+
console.error('Failed to parse HTML:', error);
|
|
116
|
+
return { content: '', title: '' };
|
|
117
|
+
}
|
|
118
|
+
}
|
|
119
|
+
//# sourceMappingURL=html-parser.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"html-parser.js","sourceRoot":"","sources":["../../src/parser/html-parser.ts"],"names":[],"mappings":"AAAA,6CAA6C;AAC7C,2DAA2D;AAE3D,OAAO,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAA;AAClD,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,eAAe,MAAM,UAAU,CAAA;AACtC,OAAO,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAA;AAClD,OAAO,EAAE,gBAAgB,EAAE,MAAM,sBAAsB,CAAA;AAcvD,iFAAiF;AACjF,SAAS,SAAS,CAAC,IAAU;IAC3B,OAAO,IAAI,CAAC,QAAQ,KAAK,CAAC,CAAA;AAC5B,CAAC;AAED,oEAAoE;AACpE,SAAS,mBAAmB,CAAC,MAAe;IAC1C,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;QACtB,OAAO,IAAI,CAAA;IACb,CAAC;IACD,MAAM,EAAE,KAAK,EAAE,OAAO,EAAE,GAAG,MAAM,CAAA;IACjC,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,OAAO,OAAO,KAAK,QAAQ,EAAE,CAAC;QAC7D,OAAO,IAAI,CAAA;IACb,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,CAAA;AAC3B,CAAC;AAED,+CAA+C;AAC/C,iCAAiC;AACjC,+CAA+C;AAE/C;;GAEG;AACH,SAAS,qBAAqB;IAC5B,MAAM,eAAe,GAAG,IAAI,eAAe,CAAC;QAC1C,YAAY,EAAE,KAAK,EAAE,uBAAuB;QAC5C,cAAc,EAAE,QAAQ,EAAE,0BAA0B;QACpD,gBAAgB,EAAE,GAAG,EAAE,yBAAyB;QAChD,WAAW,EAAE,GAAG,EAAE,qBAAqB;QACvC,eAAe,EAAE,IAAI,EAAE,kBAAkB;KAC1C,CAAC,CAAA;IAEF,0BAA0B;IAC1B,eAAe,CAAC,OAAO,CAAC,YAAY,EAAE;QACpC,MAAM,EAAE,CAAC,KAAK,CAAC;QACf,WAAW,EAAE,CAAC,QAAQ,EAAE,IAAI,EAAE,EAAE;YAC9B,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;gBACrB,OAAO,EAAE,CAAA;YACX,CAAC;YACD,MAAM,OAAO,GAAG,IAAI,CAAA;YACpB,MAAM,WAAW,GAAG,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,CAAA;YACjD,MAAM,IAAI,GAAG,WAAW,CAAC,CAAC,CAAC,WAAW,CAAC,WAAW,CAAC,CAAC,CAAC,OAAO,CAAC,WAAW,CAAA;YACxE,MAAM,QAAQ,GAAG,WAAW,EAAE,SAAS,EAAE,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,IAAI,EAAE,CAAA;YACvE,OAAO,WAAW,QAAQ,KAAK,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,YAAY,CAAA;QAC/D,CAAC;KACF,CAAC,CAAA;IAEF,OAAO,eAAe,CAAA;AACxB,CAAC;AAED,+CAA+C;AAC/C,cAAc;AACd,+CAA+C;AAE/C;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAC7B,IAAY,EACZ,GAAW;IAEX,uCAAuC;IACvC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACtC,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;IACnC,CAAC;IAED,IAAI,CAAC;QACH,8BAA8B;QAC9B,MAAM,GAAG,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE;YAC1B,GAAG;YACH,yCAAyC;YACzC,UAAU,EAAE,cAAc;SAC3B,CAAC,CAAA;QAEF,MAAM,QAAQ,GAAG,GAAG,CAAC,MAAM,CAAC,QAAQ,CAAA;QAEpC,0CAA0C;QAC1C,MAAM,MAAM,GAAG,IAAI,WAAW,CAAC,QAAQ,EAAE;YACvC,WAAW,EAAE,KAAK;YAClB,KAAK,EAAE,KAAK;SACb,CAAC,CAAA;QAEF,MAAM,OAAO,GAAG,mBAAmB,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC,CAAA;QAEnD,kEAAkE;QAClE,IAAI,CAAC,OAAO,EAAE,OAAO,EAAE,CAAC;YACtB,mCAAmC;YACnC,MAAM,WAAW,GAAG,QAAQ,CAAC,IAAI,EAAE,SAAS,IAAI,EAAE,CAAA;YAClD,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,EAAE,CAAC;gBACxB,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;YACnC,CAAC;YAED,oCAAoC;YACpC,MAAM,eAAe,GAAG,qBAAqB,EAAE,CAAA;YAC/C,OAAO,EAAE,OAAO,EAAE,eAAe,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;QAC7E,CAAC;QAED,6CAA6C;QAC7C,MAAM,eAAe,GAAG,qBAAqB,EAAE,CAAA;QAC/C,MAAM,QAAQ,GAAG,eAAe,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,CAAA;QAE1D,+DAA+D;QAC/D,qEAAqE;QACrE,IAAI,WAAW,GAAG,EAAE,CAAA;QACpB,IAAI,CAAC;YACH,WAAW,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAA;QAC5E,CAAC;QAAC,MAAM,CAAC;YACP,iCAAiC;QACnC,CAAC;QACD,MAAM,WAAW,GAAG,gBAAgB,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE,EAAE,WAAW,CAAC,CAAA;QACtE,MAAM,KAAK,GAAG,WAAW,CAAC,KAAK,CAAA;QAE/B,OAAO,EAAE,OAAO,EAAE,QAAQ,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,CAAA;IAC5C,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,2EAA2E;QAC3E,OAAO,CAAC,KAAK,CAAC,uBAAuB,EAAE,KAAK,CAAC,CAAA;QAC7C,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;IACnC,CAAC;AACH,CAAC"}
|