kordoc 3.4.1 → 3.5.1

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.
Files changed (85) hide show
  1. package/LICENSE +21 -21
  2. package/README.md +590 -584
  3. package/dist/{-YOQGQ22S.js → -BNJMPTQ2.js} +9 -5
  4. package/dist/{chunk-X34YWRL5.cjs → chunk-DCZVOIEO.cjs} +1 -1
  5. package/dist/chunk-DCZVOIEO.cjs.map +1 -0
  6. package/dist/{chunk-AIQ3ISQU.js → chunk-GE43BE46.js} +1 -1
  7. package/dist/chunk-GE43BE46.js.map +1 -0
  8. package/dist/{chunk-CGHPBI3X.js → chunk-GMPDZD7Q.js} +206 -42
  9. package/dist/chunk-GMPDZD7Q.js.map +1 -0
  10. package/dist/chunk-GS7T56RP.cjs.map +1 -1
  11. package/dist/{chunk-FPFB7AFX.js → chunk-IKAXUVSU.js} +16 -2
  12. package/dist/chunk-IKAXUVSU.js.map +1 -0
  13. package/dist/{chunk-CFQLK3LM.js → chunk-MEPHGCPQ.js} +1 -1
  14. package/dist/chunk-MEPHGCPQ.js.map +1 -0
  15. package/dist/{chunk-3TBUDJDE.js → chunk-MOL7MDBG.js} +1 -1
  16. package/dist/chunk-MOL7MDBG.js.map +1 -0
  17. package/dist/{chunk-SE3KEL4G.js → chunk-OXOPDVNV.js} +16 -2
  18. package/dist/chunk-OXOPDVNV.js.map +1 -0
  19. package/dist/{chunk-W7JHJQ2O.cjs → chunk-RWSGBXSA.cjs} +25 -11
  20. package/dist/chunk-RWSGBXSA.cjs.map +1 -0
  21. package/dist/cli.js +10 -27
  22. package/dist/cli.js.map +1 -1
  23. package/dist/{detect-NTR2FUEG.js → detect-RI2MQ33K.js} +2 -2
  24. package/dist/{formula-ZYUDQHD3.cjs → formula-5NKVS2LR.cjs} +1 -1
  25. package/dist/formula-5NKVS2LR.cjs.map +1 -0
  26. package/dist/{formula-KOQV353G.js → formula-JCNF43NE.js} +1 -1
  27. package/dist/formula-JCNF43NE.js.map +1 -0
  28. package/dist/{formula-MEDGYQXS.js → formula-RXVSQPXI.js} +1 -1
  29. package/dist/formula-RXVSQPXI.js.map +1 -0
  30. package/dist/index.cjs +377 -213
  31. package/dist/index.cjs.map +1 -1
  32. package/dist/index.d.cts +9 -3
  33. package/dist/index.d.ts +9 -3
  34. package/dist/index.js +204 -40
  35. package/dist/index.js.map +1 -1
  36. package/dist/mcp.js +42 -6
  37. package/dist/mcp.js.map +1 -1
  38. package/dist/page-range-737B4EZW.js +8 -0
  39. package/dist/page-range-LNMXJU6W.js +7 -0
  40. package/dist/page-range-P7SDW6LR.cjs +8 -0
  41. package/dist/page-range-P7SDW6LR.cjs.map +1 -0
  42. package/dist/{parser-5HDZZWDH.js → parser-GHZMDUGG.js} +5 -5
  43. package/dist/parser-GHZMDUGG.js.map +1 -0
  44. package/dist/{parser-ET64VEEX.js → parser-XXQAMYGV.js} +5 -5
  45. package/dist/parser-XXQAMYGV.js.map +1 -0
  46. package/dist/{parser-JAZCZJI6.cjs → parser-ZLFOUVR4.cjs} +18 -18
  47. package/dist/parser-ZLFOUVR4.cjs.map +1 -0
  48. package/dist/{provider-H4WWSPOV.js → provider-4ZJKV3DC.js} +1 -1
  49. package/dist/provider-4ZJKV3DC.js.map +1 -0
  50. package/dist/{provider-7H4CPZYS.js → provider-AKROB7WQ.js} +1 -1
  51. package/dist/provider-AKROB7WQ.js.map +1 -0
  52. package/dist/{provider-5K7O3Z2O.cjs → provider-G4C2V2PD.cjs} +1 -1
  53. package/dist/provider-G4C2V2PD.cjs.map +1 -0
  54. package/dist/{setup-QSJ2INNS.js → setup-57FB3LSP.js} +1 -1
  55. package/dist/setup-57FB3LSP.js.map +1 -0
  56. package/dist/{watch-56GNC2YN.js → watch-IN3YN72W.js} +5 -5
  57. package/dist/watch-IN3YN72W.js.map +1 -0
  58. package/package.json +98 -98
  59. package/dist/chunk-3TBUDJDE.js.map +0 -1
  60. package/dist/chunk-AIQ3ISQU.js.map +0 -1
  61. package/dist/chunk-CFQLK3LM.js.map +0 -1
  62. package/dist/chunk-CGHPBI3X.js.map +0 -1
  63. package/dist/chunk-FPFB7AFX.js.map +0 -1
  64. package/dist/chunk-SE3KEL4G.js.map +0 -1
  65. package/dist/chunk-W7JHJQ2O.cjs.map +0 -1
  66. package/dist/chunk-X34YWRL5.cjs.map +0 -1
  67. package/dist/formula-KOQV353G.js.map +0 -1
  68. package/dist/formula-MEDGYQXS.js.map +0 -1
  69. package/dist/formula-ZYUDQHD3.cjs.map +0 -1
  70. package/dist/page-range-CIRRJPXJ.js +0 -7
  71. package/dist/page-range-OF5I4PQY.js +0 -8
  72. package/dist/page-range-TPIRSA3J.cjs +0 -8
  73. package/dist/page-range-TPIRSA3J.cjs.map +0 -1
  74. package/dist/parser-5HDZZWDH.js.map +0 -1
  75. package/dist/parser-ET64VEEX.js.map +0 -1
  76. package/dist/parser-JAZCZJI6.cjs.map +0 -1
  77. package/dist/provider-5K7O3Z2O.cjs.map +0 -1
  78. package/dist/provider-7H4CPZYS.js.map +0 -1
  79. package/dist/provider-H4WWSPOV.js.map +0 -1
  80. package/dist/setup-QSJ2INNS.js.map +0 -1
  81. package/dist/watch-56GNC2YN.js.map +0 -1
  82. /package/dist/{-YOQGQ22S.js.map → -BNJMPTQ2.js.map} +0 -0
  83. /package/dist/{detect-NTR2FUEG.js.map → detect-RI2MQ33K.js.map} +0 -0
  84. /package/dist/{page-range-CIRRJPXJ.js.map → page-range-737B4EZW.js.map} +0 -0
  85. /package/dist/{page-range-OF5I4PQY.js.map → page-range-LNMXJU6W.js.map} +0 -0
@@ -34,4 +34,4 @@ async function renderPageToPng(page) {
34
34
  export {
35
35
  ocrPages
36
36
  };
37
- //# sourceMappingURL=provider-H4WWSPOV.js.map
37
+ //# sourceMappingURL=provider-4ZJKV3DC.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"sources":["../src/ocr/provider.ts"],"sourcesContent":["/**\n * OCR 프로바이더 브릿지 — PDF 페이지를 이미지로 렌더링하여 OCR 호출\n *\n * kordoc은 OCR 라이브러리를 번들하지 않음.\n * 사용자가 OcrProvider 함수를 제공하면 이미지 기반 PDF도 텍스트 추출 가능.\n *\n * @example\n * ```ts\n * import { parse } from \"kordoc\"\n *\n * const result = await parse(buffer, {\n * ocr: async (pageImage, pageNumber, mimeType) => {\n * // Tesseract, Claude Vision, Google Vision 등 사용\n * return await myOcrService.recognize(pageImage)\n * }\n * })\n * ```\n */\n\nimport type { OcrProvider, IRBlock } from \"../types.js\"\n\n/**\n * 이미지 기반 PDF 페이지에 OCR을 적용하여 IRBlock[] 반환.\n *\n * pdfjs page 객체에서 viewport + render를 통해 PNG 생성 후\n * 사용자 제공 OcrProvider 호출.\n *\n * canvas 미설치 시 pdfjs render 불가하므로 에러 반환.\n */\nexport async function ocrPages(\n doc: { numPages: number; getPage(n: number): Promise<PdfPageProxy> },\n provider: OcrProvider,\n pageFilter: Set<number> | null,\n effectivePageCount: number\n): Promise<IRBlock[]> {\n const blocks: IRBlock[] = []\n\n for (let i = 1; i <= effectivePageCount; i++) {\n if (pageFilter && !pageFilter.has(i)) continue\n const page = await doc.getPage(i)\n try {\n const imageData = await renderPageToPng(page)\n const text = await provider(imageData, i, \"image/png\")\n if (text.trim()) {\n blocks.push({ type: \"paragraph\", text: text.trim(), pageNumber: i })\n }\n } catch {\n blocks.push({ type: \"paragraph\" as const, text: `[OCR 실패: 페이지 ${i}]` })\n }\n }\n\n return blocks\n}\n\ninterface PdfPageProxy {\n getViewport(params: { scale: number }): { width: number; height: number }\n render(params: { canvasContext: unknown; viewport: unknown }): { promise: Promise<void> }\n}\n\n/**\n * PDF 페이지를 PNG로 렌더링.\n * node-canvas가 설치되어 있어야 동작.\n * 미설치 시 에러 throw → 호출측에서 catch.\n */\nasync function renderPageToPng(page: PdfPageProxy): Promise<Uint8Array> {\n // node-canvas 동적 로드 (선택적 의존성)\n let createCanvas: (w: number, h: number) => { getContext(t: string): unknown; toBuffer(t: string): Buffer }\n try {\n const canvasModule = await import(\"canvas\")\n createCanvas = canvasModule.createCanvas\n } catch {\n throw new Error(\"OCR을 사용하려면 'canvas' 패키지를 설치하세요: npm install canvas\")\n }\n\n const scale = 2.0 // 300 DPI 근사\n const viewport = page.getViewport({ scale })\n const canvas = createCanvas(Math.floor(viewport.width), Math.floor(viewport.height))\n const ctx = canvas.getContext(\"2d\")\n\n await page.render({ canvasContext: ctx, viewport }).promise\n return new Uint8Array(canvas.toBuffer(\"image/png\"))\n}\n"],"mappings":";AA6BA,eAAsB,SACpB,KACA,UACA,YACA,oBACoB;AACpB,QAAM,SAAoB,CAAC;AAE3B,WAAS,IAAI,GAAG,KAAK,oBAAoB,KAAK;AAC5C,QAAI,cAAc,CAAC,WAAW,IAAI,CAAC,EAAG;AACtC,UAAM,OAAO,MAAM,IAAI,QAAQ,CAAC;AAChC,QAAI;AACF,YAAM,YAAY,MAAM,gBAAgB,IAAI;AAC5C,YAAM,OAAO,MAAM,SAAS,WAAW,GAAG,WAAW;AACrD,UAAI,KAAK,KAAK,GAAG;AACf,eAAO,KAAK,EAAE,MAAM,aAAa,MAAM,KAAK,KAAK,GAAG,YAAY,EAAE,CAAC;AAAA,MACrE;AAAA,IACF,QAAQ;AACN,aAAO,KAAK,EAAE,MAAM,aAAsB,MAAM,yCAAgB,CAAC,IAAI,CAAC;AAAA,IACxE;AAAA,EACF;AAEA,SAAO;AACT;AAYA,eAAe,gBAAgB,MAAyC;AAEtE,MAAI;AACJ,MAAI;AACF,UAAM,eAAe,MAAM,OAAO,QAAQ;AAC1C,mBAAe,aAAa;AAAA,EAC9B,QAAQ;AACN,UAAM,IAAI,MAAM,+HAAoD;AAAA,EACtE;AAEA,QAAM,QAAQ;AACd,QAAM,WAAW,KAAK,YAAY,EAAE,MAAM,CAAC;AAC3C,QAAM,SAAS,aAAa,KAAK,MAAM,SAAS,KAAK,GAAG,KAAK,MAAM,SAAS,MAAM,CAAC;AACnF,QAAM,MAAM,OAAO,WAAW,IAAI;AAElC,QAAM,KAAK,OAAO,EAAE,eAAe,KAAK,SAAS,CAAC,EAAE;AACpD,SAAO,IAAI,WAAW,OAAO,SAAS,WAAW,CAAC;AACpD;","names":[]}
@@ -36,4 +36,4 @@ async function renderPageToPng(page) {
36
36
  export {
37
37
  ocrPages
38
38
  };
39
- //# sourceMappingURL=provider-7H4CPZYS.js.map
39
+ //# sourceMappingURL=provider-AKROB7WQ.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"sources":["../src/ocr/provider.ts"],"sourcesContent":["/**\n * OCR 프로바이더 브릿지 — PDF 페이지를 이미지로 렌더링하여 OCR 호출\n *\n * kordoc은 OCR 라이브러리를 번들하지 않음.\n * 사용자가 OcrProvider 함수를 제공하면 이미지 기반 PDF도 텍스트 추출 가능.\n *\n * @example\n * ```ts\n * import { parse } from \"kordoc\"\n *\n * const result = await parse(buffer, {\n * ocr: async (pageImage, pageNumber, mimeType) => {\n * // Tesseract, Claude Vision, Google Vision 등 사용\n * return await myOcrService.recognize(pageImage)\n * }\n * })\n * ```\n */\n\nimport type { OcrProvider, IRBlock } from \"../types.js\"\n\n/**\n * 이미지 기반 PDF 페이지에 OCR을 적용하여 IRBlock[] 반환.\n *\n * pdfjs page 객체에서 viewport + render를 통해 PNG 생성 후\n * 사용자 제공 OcrProvider 호출.\n *\n * canvas 미설치 시 pdfjs render 불가하므로 에러 반환.\n */\nexport async function ocrPages(\n doc: { numPages: number; getPage(n: number): Promise<PdfPageProxy> },\n provider: OcrProvider,\n pageFilter: Set<number> | null,\n effectivePageCount: number\n): Promise<IRBlock[]> {\n const blocks: IRBlock[] = []\n\n for (let i = 1; i <= effectivePageCount; i++) {\n if (pageFilter && !pageFilter.has(i)) continue\n const page = await doc.getPage(i)\n try {\n const imageData = await renderPageToPng(page)\n const text = await provider(imageData, i, \"image/png\")\n if (text.trim()) {\n blocks.push({ type: \"paragraph\", text: text.trim(), pageNumber: i })\n }\n } catch {\n blocks.push({ type: \"paragraph\" as const, text: `[OCR 실패: 페이지 ${i}]` })\n }\n }\n\n return blocks\n}\n\ninterface PdfPageProxy {\n getViewport(params: { scale: number }): { width: number; height: number }\n render(params: { canvasContext: unknown; viewport: unknown }): { promise: Promise<void> }\n}\n\n/**\n * PDF 페이지를 PNG로 렌더링.\n * node-canvas가 설치되어 있어야 동작.\n * 미설치 시 에러 throw → 호출측에서 catch.\n */\nasync function renderPageToPng(page: PdfPageProxy): Promise<Uint8Array> {\n // node-canvas 동적 로드 (선택적 의존성)\n let createCanvas: (w: number, h: number) => { getContext(t: string): unknown; toBuffer(t: string): Buffer }\n try {\n const canvasModule = await import(\"canvas\")\n createCanvas = canvasModule.createCanvas\n } catch {\n throw new Error(\"OCR을 사용하려면 'canvas' 패키지를 설치하세요: npm install canvas\")\n }\n\n const scale = 2.0 // 300 DPI 근사\n const viewport = page.getViewport({ scale })\n const canvas = createCanvas(Math.floor(viewport.width), Math.floor(viewport.height))\n const ctx = canvas.getContext(\"2d\")\n\n await page.render({ canvasContext: ctx, viewport }).promise\n return new Uint8Array(canvas.toBuffer(\"image/png\"))\n}\n"],"mappings":";;;AA6BA,eAAsB,SACpB,KACA,UACA,YACA,oBACoB;AACpB,QAAM,SAAoB,CAAC;AAE3B,WAAS,IAAI,GAAG,KAAK,oBAAoB,KAAK;AAC5C,QAAI,cAAc,CAAC,WAAW,IAAI,CAAC,EAAG;AACtC,UAAM,OAAO,MAAM,IAAI,QAAQ,CAAC;AAChC,QAAI;AACF,YAAM,YAAY,MAAM,gBAAgB,IAAI;AAC5C,YAAM,OAAO,MAAM,SAAS,WAAW,GAAG,WAAW;AACrD,UAAI,KAAK,KAAK,GAAG;AACf,eAAO,KAAK,EAAE,MAAM,aAAa,MAAM,KAAK,KAAK,GAAG,YAAY,EAAE,CAAC;AAAA,MACrE;AAAA,IACF,QAAQ;AACN,aAAO,KAAK,EAAE,MAAM,aAAsB,MAAM,yCAAgB,CAAC,IAAI,CAAC;AAAA,IACxE;AAAA,EACF;AAEA,SAAO;AACT;AAYA,eAAe,gBAAgB,MAAyC;AAEtE,MAAI;AACJ,MAAI;AACF,UAAM,eAAe,MAAM,OAAO,QAAQ;AAC1C,mBAAe,aAAa;AAAA,EAC9B,QAAQ;AACN,UAAM,IAAI,MAAM,+HAAoD;AAAA,EACtE;AAEA,QAAM,QAAQ;AACd,QAAM,WAAW,KAAK,YAAY,EAAE,MAAM,CAAC;AAC3C,QAAM,SAAS,aAAa,KAAK,MAAM,SAAS,KAAK,GAAG,KAAK,MAAM,SAAS,MAAM,CAAC;AACnF,QAAM,MAAM,OAAO,WAAW,IAAI;AAElC,QAAM,KAAK,OAAO,EAAE,eAAe,KAAK,SAAS,CAAC,EAAE;AACpD,SAAO,IAAI,WAAW,OAAO,SAAS,WAAW,CAAC;AACpD;","names":[]}
@@ -36,4 +36,4 @@ async function renderPageToPng(page) {
36
36
 
37
37
 
38
38
  exports.ocrPages = ocrPages;
39
- //# sourceMappingURL=provider-5K7O3Z2O.cjs.map
39
+ //# sourceMappingURL=provider-G4C2V2PD.cjs.map
@@ -0,0 +1 @@
1
+ {"version":3,"sources":["/Users/chris_gomdori/workspace/kordoc/dist/provider-G4C2V2PD.cjs","../src/ocr/provider.ts"],"names":[],"mappings":"AAAA,2XAA6B;AAC7B;AACA;AC2BA,MAAA,SAAsB,QAAA,CACpB,GAAA,EACA,QAAA,EACA,UAAA,EACA,kBAAA,EACoB;AACpB,EAAA,MAAM,OAAA,EAAoB,CAAC,CAAA;AAE3B,EAAA,IAAA,CAAA,IAAS,EAAA,EAAI,CAAA,EAAG,EAAA,GAAK,kBAAA,EAAoB,CAAA,EAAA,EAAK;AAC5C,IAAA,GAAA,CAAI,WAAA,GAAc,CAAC,UAAA,CAAW,GAAA,CAAI,CAAC,CAAA,EAAG,QAAA;AACtC,IAAA,MAAM,KAAA,EAAO,MAAM,GAAA,CAAI,OAAA,CAAQ,CAAC,CAAA;AAChC,IAAA,IAAI;AACF,MAAA,MAAM,UAAA,EAAY,MAAM,eAAA,CAAgB,IAAI,CAAA;AAC5C,MAAA,MAAM,KAAA,EAAO,MAAM,QAAA,CAAS,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AACrD,MAAA,GAAA,CAAI,IAAA,CAAK,IAAA,CAAK,CAAA,EAAG;AACf,QAAA,MAAA,CAAO,IAAA,CAAK,EAAE,IAAA,EAAM,WAAA,EAAa,IAAA,EAAM,IAAA,CAAK,IAAA,CAAK,CAAA,EAAG,UAAA,EAAY,EAAE,CAAC,CAAA;AAAA,MACrE;AAAA,IACF,EAAA,UAAQ;AACN,MAAA,MAAA,CAAO,IAAA,CAAK,EAAE,IAAA,EAAM,WAAA,EAAsB,IAAA,EAAM,CAAA,sCAAA,EAAgB,CAAC,CAAA,CAAA,EAAI,CAAC,CAAA;AAAA,IACxE;AAAA,EACF;AAEA,EAAA,OAAO,MAAA;AACT;AAYA,MAAA,SAAe,eAAA,CAAgB,IAAA,EAAyC;AAEtE,EAAA,IAAI,YAAA;AACJ,EAAA,IAAI;AACF,IAAA,MAAM,aAAA,EAAe,MAAM,4DAAA,CAAO,QAAQ,GAAA;AAC1C,IAAA,aAAA,EAAe,YAAA,CAAa,YAAA;AAAA,EAC9B,EAAA,WAAQ;AACN,IAAA,MAAM,IAAI,KAAA,CAAM,+HAAoD,CAAA;AAAA,EACtE;AAEA,EAAA,MAAM,MAAA,EAAQ,CAAA;AACd,EAAA,MAAM,SAAA,EAAW,IAAA,CAAK,WAAA,CAAY,EAAE,MAAM,CAAC,CAAA;AAC3C,EAAA,MAAM,OAAA,EAAS,YAAA,CAAa,IAAA,CAAK,KAAA,CAAM,QAAA,CAAS,KAAK,CAAA,EAAG,IAAA,CAAK,KAAA,CAAM,QAAA,CAAS,MAAM,CAAC,CAAA;AACnF,EAAA,MAAM,IAAA,EAAM,MAAA,CAAO,UAAA,CAAW,IAAI,CAAA;AAElC,EAAA,MAAM,IAAA,CAAK,MAAA,CAAO,EAAE,aAAA,EAAe,GAAA,EAAK,SAAS,CAAC,CAAA,CAAE,OAAA;AACpD,EAAA,OAAO,IAAI,UAAA,CAAW,MAAA,CAAO,QAAA,CAAS,WAAW,CAAC,CAAA;AACpD;AD9CA;AACE;AACF,4BAAC","file":"/Users/chris_gomdori/workspace/kordoc/dist/provider-G4C2V2PD.cjs","sourcesContent":[null,"/**\n * OCR 프로바이더 브릿지 — PDF 페이지를 이미지로 렌더링하여 OCR 호출\n *\n * kordoc은 OCR 라이브러리를 번들하지 않음.\n * 사용자가 OcrProvider 함수를 제공하면 이미지 기반 PDF도 텍스트 추출 가능.\n *\n * @example\n * ```ts\n * import { parse } from \"kordoc\"\n *\n * const result = await parse(buffer, {\n * ocr: async (pageImage, pageNumber, mimeType) => {\n * // Tesseract, Claude Vision, Google Vision 등 사용\n * return await myOcrService.recognize(pageImage)\n * }\n * })\n * ```\n */\n\nimport type { OcrProvider, IRBlock } from \"../types.js\"\n\n/**\n * 이미지 기반 PDF 페이지에 OCR을 적용하여 IRBlock[] 반환.\n *\n * pdfjs page 객체에서 viewport + render를 통해 PNG 생성 후\n * 사용자 제공 OcrProvider 호출.\n *\n * canvas 미설치 시 pdfjs render 불가하므로 에러 반환.\n */\nexport async function ocrPages(\n doc: { numPages: number; getPage(n: number): Promise<PdfPageProxy> },\n provider: OcrProvider,\n pageFilter: Set<number> | null,\n effectivePageCount: number\n): Promise<IRBlock[]> {\n const blocks: IRBlock[] = []\n\n for (let i = 1; i <= effectivePageCount; i++) {\n if (pageFilter && !pageFilter.has(i)) continue\n const page = await doc.getPage(i)\n try {\n const imageData = await renderPageToPng(page)\n const text = await provider(imageData, i, \"image/png\")\n if (text.trim()) {\n blocks.push({ type: \"paragraph\", text: text.trim(), pageNumber: i })\n }\n } catch {\n blocks.push({ type: \"paragraph\" as const, text: `[OCR 실패: 페이지 ${i}]` })\n }\n }\n\n return blocks\n}\n\ninterface PdfPageProxy {\n getViewport(params: { scale: number }): { width: number; height: number }\n render(params: { canvasContext: unknown; viewport: unknown }): { promise: Promise<void> }\n}\n\n/**\n * PDF 페이지를 PNG로 렌더링.\n * node-canvas가 설치되어 있어야 동작.\n * 미설치 시 에러 throw → 호출측에서 catch.\n */\nasync function renderPageToPng(page: PdfPageProxy): Promise<Uint8Array> {\n // node-canvas 동적 로드 (선택적 의존성)\n let createCanvas: (w: number, h: number) => { getContext(t: string): unknown; toBuffer(t: string): Buffer }\n try {\n const canvasModule = await import(\"canvas\")\n createCanvas = canvasModule.createCanvas\n } catch {\n throw new Error(\"OCR을 사용하려면 'canvas' 패키지를 설치하세요: npm install canvas\")\n }\n\n const scale = 2.0 // 300 DPI 근사\n const viewport = page.getViewport({ scale })\n const canvas = createCanvas(Math.floor(viewport.width), Math.floor(viewport.height))\n const ctx = canvas.getContext(\"2d\")\n\n await page.render({ canvasContext: ctx, viewport }).promise\n return new Uint8Array(canvas.toBuffer(\"image/png\"))\n}\n"]}
@@ -198,4 +198,4 @@ function printManualConfig() {
198
198
  export {
199
199
  runSetup
200
200
  };
201
- //# sourceMappingURL=setup-QSJ2INNS.js.map
201
+ //# sourceMappingURL=setup-57FB3LSP.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"sources":["../src/setup.ts"],"sourcesContent":["/**\n * 대화형 설치 마법사 — kordoc MCP\n *\n * `npx kordoc setup` 으로 실행.\n * 선택한 AI 클라이언트 설정 파일에 kordoc-mcp 서버를 자동 등록합니다.\n * API 키 불필요, macOS / Linux / Windows 공용.\n */\n\nimport { createInterface } from \"node:readline/promises\"\nimport { readFile, writeFile, mkdir } from \"node:fs/promises\"\nimport { existsSync } from \"node:fs\"\nimport { resolve, dirname } from \"node:path\"\nimport { homedir, platform } from \"node:os\"\nimport { stdin, stdout } from \"node:process\"\n\ninterface ClientConfig {\n readonly name: string\n readonly configPath: string\n readonly format: \"mcpServers\" | \"servers\" | \"context_servers\"\n}\n\nfunction detectClients(): readonly ClientConfig[] {\n const home = homedir()\n const os = platform()\n const clients: ClientConfig[] = []\n\n const claudePaths: Record<string, string> = {\n darwin: resolve(home, \"Library/Application Support/Claude/claude_desktop_config.json\"),\n win32: resolve(process.env[\"APPDATA\"] ?? resolve(home, \"AppData/Roaming\"), \"Claude/claude_desktop_config.json\"),\n linux: resolve(home, \".config/Claude/claude_desktop_config.json\"),\n }\n const claudePath = claudePaths[os]\n if (claudePath) clients.push({ name: \"Claude Desktop\", configPath: claudePath, format: \"mcpServers\" })\n\n clients.push({ name: \"Claude Code (현재 디렉토리)\", configPath: resolve(process.cwd(), \".mcp.json\"), format: \"mcpServers\" })\n clients.push({ name: \"Cursor\", configPath: resolve(home, \".cursor/mcp.json\"), format: \"mcpServers\" })\n clients.push({ name: \"VS Code (현재 디렉토리)\", configPath: resolve(process.cwd(), \".vscode/mcp.json\"), format: \"servers\" })\n clients.push({ name: \"Windsurf\", configPath: resolve(home, \".codeium/windsurf/mcp_config.json\"), format: \"mcpServers\" })\n clients.push({ name: \"Gemini CLI\", configPath: resolve(home, \".gemini/settings.json\"), format: \"mcpServers\" })\n\n const zedPaths: Record<string, string> = {\n darwin: resolve(home, \".zed/settings.json\"),\n linux: resolve(home, \".config/zed/settings.json\"),\n win32: resolve(home, \".zed/settings.json\"),\n }\n const zedPath = zedPaths[os]\n if (zedPath) clients.push({ name: \"Zed\", configPath: zedPath, format: \"context_servers\" })\n\n clients.push({ name: \"Antigravity\", configPath: resolve(home, \".gemini/antigravity/mcp_config.json\"), format: \"mcpServers\" })\n\n return clients\n}\n\nasync function readJsonFile(path: string): Promise<Record<string, unknown>> {\n if (!existsSync(path)) return {}\n const raw = await readFile(path, \"utf-8\")\n return JSON.parse(raw) as Record<string, unknown>\n}\n\nasync function writeJsonFile(path: string, data: Record<string, unknown>): Promise<void> {\n const dir = dirname(path)\n if (!existsSync(dir)) await mkdir(dir, { recursive: true })\n await writeFile(path, JSON.stringify(data, null, 2) + \"\\n\", \"utf-8\")\n}\n\n/**\n * Windows 에선 npx.cmd 를 Claude Desktop 이 해석 못 해서 `cmd /c` 래핑이 필요하다.\n * `kordoc` 은 `npx kordoc mcp` 로 MCP 서버 실행, 또는 `kordoc-mcp` bin 직접 실행.\n * npx 래핑이 범용적이므로 `npx -y kordoc mcp` 로 통일.\n */\nfunction buildServerEntry(): Record<string, unknown> {\n if (platform() === \"win32\") {\n return { command: \"cmd\", args: [\"/c\", \"npx\", \"-y\", \"kordoc\", \"mcp\"] }\n }\n return { command: \"npx\", args: [\"-y\", \"kordoc\", \"mcp\"] }\n}\n\nfunction buildZedEntry(): Record<string, unknown> {\n const base = platform() === \"win32\"\n ? { path: \"cmd\", args: [\"/c\", \"npx\", \"-y\", \"kordoc\", \"mcp\"] }\n : { path: \"npx\", args: [\"-y\", \"kordoc\", \"mcp\"] }\n return { command: base }\n}\n\n// ─── ANSI ─────────────────────────────────────────────────────────────\nconst ESC = \"\\x1b[\"\nconst c = {\n reset: `${ESC}0m`, bold: `${ESC}1m`, dim: `${ESC}2m`,\n cyan: `${ESC}36m`, green: `${ESC}32m`, yellow: `${ESC}33m`,\n red: `${ESC}31m`, white: `${ESC}37m`,\n} as const\n\nfunction rgb(r: number, g: number, b: number): string {\n return `${ESC}38;2;${r};${g};${b}m`\n}\nfunction sleep(ms: number): Promise<void> {\n return new Promise((res) => setTimeout(res, ms))\n}\nasync function typewrite(text: string, delay = 15): Promise<void> {\n for (const ch of text) { process.stdout.write(ch); await sleep(delay) }\n console.log()\n}\n\nasync function printBanner(): Promise<void> {\n const gradients = [\n rgb(255, 120, 80), rgb(255, 140, 80), rgb(255, 160, 80),\n rgb(240, 180, 80), rgb(220, 200, 80), rgb(200, 220, 80),\n ]\n const logo = [\n \" _ _ \",\n \" | | _____ _ __ __| | ___ ___ \",\n \" | |/ / _ \\\\| '__/ _` |/ _ \\\\ / __|\",\n \" | < (_) | | | (_| | (_) | (__ \",\n \" |_|\\\\_\\\\___/|_| \\\\__,_|\\\\___/ \\\\___|\",\n ]\n console.log()\n for (let i = 0; i < logo.length; i++) {\n console.log(`${gradients[i % gradients.length]}${c.bold}${logo[i]}${c.reset}`)\n await sleep(60)\n }\n console.log()\n await typewrite(`${c.dim} 모두 파싱해버리겠다 ━━ HWP · HWPX · PDF · XLSX · DOCX → Markdown${c.reset}`, 10)\n console.log()\n console.log(`${c.cyan} ${\"━\".repeat(60)}${c.reset}`)\n console.log()\n}\n\nfunction stepHeader(step: number, total: number, title: string): void {\n const dots = `${c.dim}${\"·\".repeat(Math.max(0, 40 - title.length))}${c.reset}`\n console.log(` ${c.cyan}${c.bold}[${step}/${total}]${c.reset} ${c.white}${c.bold}${title}${c.reset} ${dots}`)\n console.log()\n}\n\nfunction successLine(label: string, detail: string): void {\n console.log(` ${c.green}${c.bold}+${c.reset} ${c.white}${label}${c.reset}${c.dim} ${detail}${c.reset}`)\n}\nfunction failLine(label: string, detail: string): void {\n console.log(` ${c.red}${c.bold}x${c.reset} ${c.white}${label}${c.reset}${c.dim} ${detail}${c.reset}`)\n}\n\nasync function printComplete(): Promise<void> {\n console.log()\n const box = [\n ` ${c.green}${c.bold}╔${\"═\".repeat(50)}╗${c.reset}`,\n ` ${c.green}${c.bold}║${c.reset}${\" \".repeat(14)}${c.green}${c.bold}Setup Complete!${c.reset}${\" \".repeat(22)}${c.green}${c.bold}║${c.reset}`,\n ` ${c.green}${c.bold}╚${\"═\".repeat(50)}╝${c.reset}`,\n ]\n for (const line of box) { console.log(line); await sleep(40) }\n console.log()\n console.log(` ${c.dim}클라이언트를 재시작하면 'kordoc' MCP 서버가 활성화됩니다.${c.reset}`)\n console.log(` ${c.dim}8개 도구: parse_document / parse_metadata / parse_pages / parse_table${c.reset}`)\n console.log(` ${c.dim} detect_format / compare_documents / parse_form / fill_form${c.reset}`)\n console.log()\n}\n\nexport async function runSetup(): Promise<void> {\n const rl = createInterface({ input: stdin, output: stdout })\n\n try {\n await printBanner()\n\n stepHeader(1, 2, \"MCP 클라이언트 선택\")\n const clients = detectClients()\n clients.forEach((cl, i) => {\n const exists = existsSync(cl.configPath)\n const badge = exists ? `${c.green} [감지됨]${c.reset}` : \"\"\n const num = `${c.cyan}${String(i + 1).padStart(2)}${c.reset}`\n console.log(` ${num}) ${c.white}${cl.name}${c.reset}${badge}`)\n })\n console.log()\n const clientInput = (await rl.question(` ${c.cyan}>${c.reset} 번호 (예: 1,3): `)).trim()\n\n if (!clientInput) {\n console.log(`\\n ${c.yellow}선택 없음${c.reset} — 수동 설정 안내:`)\n printManualConfig()\n return\n }\n\n const indices = clientInput\n .split(\",\")\n .map((s) => parseInt(s.trim(), 10) - 1)\n .filter((i) => i >= 0 && i < clients.length)\n\n if (indices.length === 0) {\n console.log(`\\n ${c.yellow}유효한 선택 없음${c.reset} — 수동 설정 안내:`)\n printManualConfig()\n return\n }\n\n console.log()\n stepHeader(2, 2, \"설정 파일 업데이트\")\n const entry = buildServerEntry()\n\n for (const idx of indices) {\n const client = clients[idx]\n await sleep(150)\n try {\n const config = await readJsonFile(client.configPath)\n const key = client.format\n const serverEntry = key === \"context_servers\" ? buildZedEntry() : entry\n const servers = (config[key] ?? {}) as Record<string, unknown>\n servers[\"kordoc\"] = serverEntry\n config[key] = servers\n await writeJsonFile(client.configPath, config)\n successLine(client.name, client.configPath)\n } catch (err) {\n const msg = err instanceof Error ? err.message : String(err)\n failLine(client.name, msg)\n }\n }\n\n await printComplete()\n } catch (err) {\n if ((err as NodeJS.ErrnoException)?.code === \"ERR_USE_AFTER_CLOSE\") return\n throw err\n } finally {\n rl.close()\n }\n}\n\nfunction printManualConfig(): void {\n const entry = buildServerEntry()\n console.log()\n console.log(` ${c.dim}아래 JSON을 설정 파일의 mcpServers에 추가하세요:${c.reset}`)\n console.log()\n console.log(` ${c.cyan}\"kordoc\"${c.reset}: ${JSON.stringify(entry, null, 4)}`)\n console.log()\n}\n"],"mappings":";;;AAQA,SAAS,uBAAuB;AAChC,SAAS,UAAU,WAAW,aAAa;AAC3C,SAAS,kBAAkB;AAC3B,SAAS,SAAS,eAAe;AACjC,SAAS,SAAS,gBAAgB;AAClC,SAAS,OAAO,cAAc;AAQ9B,SAAS,gBAAyC;AAChD,QAAM,OAAO,QAAQ;AACrB,QAAM,KAAK,SAAS;AACpB,QAAM,UAA0B,CAAC;AAEjC,QAAM,cAAsC;AAAA,IAC1C,QAAQ,QAAQ,MAAM,+DAA+D;AAAA,IACrF,OAAO,QAAQ,QAAQ,IAAI,SAAS,KAAK,QAAQ,MAAM,iBAAiB,GAAG,mCAAmC;AAAA,IAC9G,OAAO,QAAQ,MAAM,2CAA2C;AAAA,EAClE;AACA,QAAM,aAAa,YAAY,EAAE;AACjC,MAAI,WAAY,SAAQ,KAAK,EAAE,MAAM,kBAAkB,YAAY,YAAY,QAAQ,aAAa,CAAC;AAErG,UAAQ,KAAK,EAAE,MAAM,uDAAyB,YAAY,QAAQ,QAAQ,IAAI,GAAG,WAAW,GAAG,QAAQ,aAAa,CAAC;AACrH,UAAQ,KAAK,EAAE,MAAM,UAAU,YAAY,QAAQ,MAAM,kBAAkB,GAAG,QAAQ,aAAa,CAAC;AACpG,UAAQ,KAAK,EAAE,MAAM,mDAAqB,YAAY,QAAQ,QAAQ,IAAI,GAAG,kBAAkB,GAAG,QAAQ,UAAU,CAAC;AACrH,UAAQ,KAAK,EAAE,MAAM,YAAY,YAAY,QAAQ,MAAM,mCAAmC,GAAG,QAAQ,aAAa,CAAC;AACvH,UAAQ,KAAK,EAAE,MAAM,cAAc,YAAY,QAAQ,MAAM,uBAAuB,GAAG,QAAQ,aAAa,CAAC;AAE7G,QAAM,WAAmC;AAAA,IACvC,QAAQ,QAAQ,MAAM,oBAAoB;AAAA,IAC1C,OAAO,QAAQ,MAAM,2BAA2B;AAAA,IAChD,OAAO,QAAQ,MAAM,oBAAoB;AAAA,EAC3C;AACA,QAAM,UAAU,SAAS,EAAE;AAC3B,MAAI,QAAS,SAAQ,KAAK,EAAE,MAAM,OAAO,YAAY,SAAS,QAAQ,kBAAkB,CAAC;AAEzF,UAAQ,KAAK,EAAE,MAAM,eAAe,YAAY,QAAQ,MAAM,qCAAqC,GAAG,QAAQ,aAAa,CAAC;AAE5H,SAAO;AACT;AAEA,eAAe,aAAa,MAAgD;AAC1E,MAAI,CAAC,WAAW,IAAI,EAAG,QAAO,CAAC;AAC/B,QAAM,MAAM,MAAM,SAAS,MAAM,OAAO;AACxC,SAAO,KAAK,MAAM,GAAG;AACvB;AAEA,eAAe,cAAc,MAAc,MAA8C;AACvF,QAAM,MAAM,QAAQ,IAAI;AACxB,MAAI,CAAC,WAAW,GAAG,EAAG,OAAM,MAAM,KAAK,EAAE,WAAW,KAAK,CAAC;AAC1D,QAAM,UAAU,MAAM,KAAK,UAAU,MAAM,MAAM,CAAC,IAAI,MAAM,OAAO;AACrE;AAOA,SAAS,mBAA4C;AACnD,MAAI,SAAS,MAAM,SAAS;AAC1B,WAAO,EAAE,SAAS,OAAO,MAAM,CAAC,MAAM,OAAO,MAAM,UAAU,KAAK,EAAE;AAAA,EACtE;AACA,SAAO,EAAE,SAAS,OAAO,MAAM,CAAC,MAAM,UAAU,KAAK,EAAE;AACzD;AAEA,SAAS,gBAAyC;AAChD,QAAM,OAAO,SAAS,MAAM,UACxB,EAAE,MAAM,OAAO,MAAM,CAAC,MAAM,OAAO,MAAM,UAAU,KAAK,EAAE,IAC1D,EAAE,MAAM,OAAO,MAAM,CAAC,MAAM,UAAU,KAAK,EAAE;AACjD,SAAO,EAAE,SAAS,KAAK;AACzB;AAGA,IAAM,MAAM;AACZ,IAAM,IAAI;AAAA,EACR,OAAO,GAAG,GAAG;AAAA,EAAM,MAAM,GAAG,GAAG;AAAA,EAAM,KAAK,GAAG,GAAG;AAAA,EAChD,MAAM,GAAG,GAAG;AAAA,EAAO,OAAO,GAAG,GAAG;AAAA,EAAO,QAAQ,GAAG,GAAG;AAAA,EACrD,KAAK,GAAG,GAAG;AAAA,EAAO,OAAO,GAAG,GAAG;AACjC;AAEA,SAAS,IAAI,GAAW,GAAW,GAAmB;AACpD,SAAO,GAAG,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;AAClC;AACA,SAAS,MAAM,IAA2B;AACxC,SAAO,IAAI,QAAQ,CAAC,QAAQ,WAAW,KAAK,EAAE,CAAC;AACjD;AACA,eAAe,UAAU,MAAc,QAAQ,IAAmB;AAChE,aAAW,MAAM,MAAM;AAAE,YAAQ,OAAO,MAAM,EAAE;AAAG,UAAM,MAAM,KAAK;AAAA,EAAE;AACtE,UAAQ,IAAI;AACd;AAEA,eAAe,cAA6B;AAC1C,QAAM,YAAY;AAAA,IAChB,IAAI,KAAK,KAAK,EAAE;AAAA,IAAG,IAAI,KAAK,KAAK,EAAE;AAAA,IAAG,IAAI,KAAK,KAAK,EAAE;AAAA,IACtD,IAAI,KAAK,KAAK,EAAE;AAAA,IAAG,IAAI,KAAK,KAAK,EAAE;AAAA,IAAG,IAAI,KAAK,KAAK,EAAE;AAAA,EACxD;AACA,QAAM,OAAO;AAAA,IACX;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,EACF;AACA,UAAQ,IAAI;AACZ,WAAS,IAAI,GAAG,IAAI,KAAK,QAAQ,KAAK;AACpC,YAAQ,IAAI,GAAG,UAAU,IAAI,UAAU,MAAM,CAAC,GAAG,EAAE,IAAI,GAAG,KAAK,CAAC,CAAC,GAAG,EAAE,KAAK,EAAE;AAC7E,UAAM,MAAM,EAAE;AAAA,EAChB;AACA,UAAQ,IAAI;AACZ,QAAM,UAAU,GAAG,EAAE,GAAG,sIAA8D,EAAE,KAAK,IAAI,EAAE;AACnG,UAAQ,IAAI;AACZ,UAAQ,IAAI,GAAG,EAAE,IAAI,KAAK,SAAI,OAAO,EAAE,CAAC,GAAG,EAAE,KAAK,EAAE;AACpD,UAAQ,IAAI;AACd;AAEA,SAAS,WAAW,MAAc,OAAe,OAAqB;AACpE,QAAM,OAAO,GAAG,EAAE,GAAG,GAAG,OAAI,OAAO,KAAK,IAAI,GAAG,KAAK,MAAM,MAAM,CAAC,CAAC,GAAG,EAAE,KAAK;AAC5E,UAAQ,IAAI,KAAK,EAAE,IAAI,GAAG,EAAE,IAAI,IAAI,IAAI,IAAI,KAAK,IAAI,EAAE,KAAK,IAAI,EAAE,KAAK,GAAG,EAAE,IAAI,GAAG,KAAK,GAAG,EAAE,KAAK,IAAI,IAAI,EAAE;AAC5G,UAAQ,IAAI;AACd;AAEA,SAAS,YAAY,OAAe,QAAsB;AACxD,UAAQ,IAAI,KAAK,EAAE,KAAK,GAAG,EAAE,IAAI,IAAI,EAAE,KAAK,IAAI,EAAE,KAAK,GAAG,KAAK,GAAG,EAAE,KAAK,GAAG,EAAE,GAAG,IAAI,MAAM,GAAG,EAAE,KAAK,EAAE;AACzG;AACA,SAAS,SAAS,OAAe,QAAsB;AACrD,UAAQ,IAAI,KAAK,EAAE,GAAG,GAAG,EAAE,IAAI,IAAI,EAAE,KAAK,IAAI,EAAE,KAAK,GAAG,KAAK,GAAG,EAAE,KAAK,GAAG,EAAE,GAAG,IAAI,MAAM,GAAG,EAAE,KAAK,EAAE;AACvG;AAEA,eAAe,gBAA+B;AAC5C,UAAQ,IAAI;AACZ,QAAM,MAAM;AAAA,IACV,KAAK,EAAE,KAAK,GAAG,EAAE,IAAI,SAAI,SAAI,OAAO,EAAE,CAAC,SAAI,EAAE,KAAK;AAAA,IAClD,KAAK,EAAE,KAAK,GAAG,EAAE,IAAI,SAAI,EAAE,KAAK,GAAG,IAAI,OAAO,EAAE,CAAC,GAAG,EAAE,KAAK,GAAG,EAAE,IAAI,kBAAkB,EAAE,KAAK,GAAG,IAAI,OAAO,EAAE,CAAC,GAAG,EAAE,KAAK,GAAG,EAAE,IAAI,SAAI,EAAE,KAAK;AAAA,IAC5I,KAAK,EAAE,KAAK,GAAG,EAAE,IAAI,SAAI,SAAI,OAAO,EAAE,CAAC,SAAI,EAAE,KAAK;AAAA,EACpD;AACA,aAAW,QAAQ,KAAK;AAAE,YAAQ,IAAI,IAAI;AAAG,UAAM,MAAM,EAAE;AAAA,EAAE;AAC7D,UAAQ,IAAI;AACZ,UAAQ,IAAI,KAAK,EAAE,GAAG,4IAAwC,EAAE,KAAK,EAAE;AACvE,UAAQ,IAAI,KAAK,EAAE,GAAG,oFAAqE,EAAE,KAAK,EAAE;AACpG,UAAQ,IAAI,KAAK,EAAE,GAAG,sEAAsE,EAAE,KAAK,EAAE;AACrG,UAAQ,IAAI;AACd;AAEA,eAAsB,WAA0B;AAC9C,QAAM,KAAK,gBAAgB,EAAE,OAAO,OAAO,QAAQ,OAAO,CAAC;AAE3D,MAAI;AACF,UAAM,YAAY;AAElB,eAAW,GAAG,GAAG,iDAAc;AAC/B,UAAM,UAAU,cAAc;AAC9B,YAAQ,QAAQ,CAAC,IAAI,MAAM;AACzB,YAAM,SAAS,WAAW,GAAG,UAAU;AACvC,YAAM,QAAQ,SAAS,GAAG,EAAE,KAAK,wBAAS,EAAE,KAAK,KAAK;AACtD,YAAM,MAAM,GAAG,EAAE,IAAI,GAAG,OAAO,IAAI,CAAC,EAAE,SAAS,CAAC,CAAC,GAAG,EAAE,KAAK;AAC3D,cAAQ,IAAI,KAAK,GAAG,KAAK,EAAE,KAAK,GAAG,GAAG,IAAI,GAAG,EAAE,KAAK,GAAG,KAAK,EAAE;AAAA,IAChE,CAAC;AACD,YAAQ,IAAI;AACZ,UAAM,eAAe,MAAM,GAAG,SAAS,KAAK,EAAE,IAAI,IAAI,EAAE,KAAK,+BAAgB,GAAG,KAAK;AAErF,QAAI,CAAC,aAAa;AAChB,cAAQ,IAAI;AAAA,IAAO,EAAE,MAAM,4BAAQ,EAAE,KAAK,iDAAc;AACxD,wBAAkB;AAClB;AAAA,IACF;AAEA,UAAM,UAAU,YACb,MAAM,GAAG,EACT,IAAI,CAAC,MAAM,SAAS,EAAE,KAAK,GAAG,EAAE,IAAI,CAAC,EACrC,OAAO,CAAC,MAAM,KAAK,KAAK,IAAI,QAAQ,MAAM;AAE7C,QAAI,QAAQ,WAAW,GAAG;AACxB,cAAQ,IAAI;AAAA,IAAO,EAAE,MAAM,+CAAY,EAAE,KAAK,iDAAc;AAC5D,wBAAkB;AAClB;AAAA,IACF;AAEA,YAAQ,IAAI;AACZ,eAAW,GAAG,GAAG,oDAAY;AAC7B,UAAM,QAAQ,iBAAiB;AAE/B,eAAW,OAAO,SAAS;AACzB,YAAM,SAAS,QAAQ,GAAG;AAC1B,YAAM,MAAM,GAAG;AACf,UAAI;AACF,cAAM,SAAS,MAAM,aAAa,OAAO,UAAU;AACnD,cAAM,MAAM,OAAO;AACnB,cAAM,cAAc,QAAQ,oBAAoB,cAAc,IAAI;AAClE,cAAM,UAAW,OAAO,GAAG,KAAK,CAAC;AACjC,gBAAQ,QAAQ,IAAI;AACpB,eAAO,GAAG,IAAI;AACd,cAAM,cAAc,OAAO,YAAY,MAAM;AAC7C,oBAAY,OAAO,MAAM,OAAO,UAAU;AAAA,MAC5C,SAAS,KAAK;AACZ,cAAM,MAAM,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC3D,iBAAS,OAAO,MAAM,GAAG;AAAA,MAC3B;AAAA,IACF;AAEA,UAAM,cAAc;AAAA,EACtB,SAAS,KAAK;AACZ,QAAK,KAA+B,SAAS,sBAAuB;AACpE,UAAM;AAAA,EACR,UAAE;AACA,OAAG,MAAM;AAAA,EACX;AACF;AAEA,SAAS,oBAA0B;AACjC,QAAM,QAAQ,iBAAiB;AAC/B,UAAQ,IAAI;AACZ,UAAQ,IAAI,KAAK,EAAE,GAAG,2GAAqC,EAAE,KAAK,EAAE;AACpE,UAAQ,IAAI;AACZ,UAAQ,IAAI,KAAK,EAAE,IAAI,WAAW,EAAE,KAAK,KAAK,KAAK,UAAU,OAAO,MAAM,CAAC,CAAC,EAAE;AAC9E,UAAQ,IAAI;AACd;","names":[]}
@@ -1,14 +1,14 @@
1
1
  #!/usr/bin/env node
2
2
  import {
3
3
  parse
4
- } from "./chunk-CGHPBI3X.js";
4
+ } from "./chunk-GMPDZD7Q.js";
5
5
  import {
6
6
  detectFormat
7
- } from "./chunk-CFQLK3LM.js";
7
+ } from "./chunk-MEPHGCPQ.js";
8
8
  import {
9
9
  toArrayBuffer
10
- } from "./chunk-FPFB7AFX.js";
11
- import "./chunk-3TBUDJDE.js";
10
+ } from "./chunk-IKAXUVSU.js";
11
+ import "./chunk-MOL7MDBG.js";
12
12
 
13
13
  // src/watch.ts
14
14
  import { watch, readFileSync, writeFileSync, mkdirSync, statSync, existsSync, realpathSync } from "fs";
@@ -147,4 +147,4 @@ async function sendWebhook(url, payload) {
147
147
  export {
148
148
  watchDirectory
149
149
  };
150
- //# sourceMappingURL=watch-56GNC2YN.js.map
150
+ //# sourceMappingURL=watch-IN3YN72W.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"sources":["../src/watch.ts"],"sourcesContent":["/** 디렉토리 감시 모드 — 새 문서 자동 변환 + Webhook 알림 */\n\nimport { watch, readFileSync, writeFileSync, mkdirSync, statSync, existsSync, realpathSync } from \"fs\"\nimport { basename, resolve, extname, sep } from \"path\"\nimport { parse, detectFormat } from \"./index.js\"\nimport { toArrayBuffer } from \"./utils.js\"\nimport type { WatchOptions } from \"./types.js\"\n\nconst SUPPORTED_EXTENSIONS = new Set([\".hwp\", \".hwpx\", \".pdf\", \".xlsx\", \".docx\"])\nconst DEBOUNCE_MS = 1000\n/** 파일 쓰기 완료 판정: 연속 2회 동일 크기 확인 간격 */\nconst STABLE_CHECK_MS = 300\nconst MAX_FILE_SIZE = 500 * 1024 * 1024\n\n/**\n * 디렉토리를 감시하여 새 문서 파일을 자동 변환.\n *\n * @example\n * ```bash\n * kordoc watch ./incoming -d ./output --webhook https://api.example.com/docs\n * ```\n */\nexport async function watchDirectory(options: WatchOptions): Promise<void> {\n const { dir, outDir, webhook, format = \"markdown\", pages, silent } = options\n\n if (!existsSync(dir)) throw new Error(`디렉토리를 찾을 수 없습니다: ${dir}`)\n if (webhook) validateWebhookUrl(webhook)\n if (outDir) mkdirSync(outDir, { recursive: true })\n\n const log = silent ? () => {} : (msg: string) => process.stderr.write(msg + \"\\n\")\n log(`[kordoc watch] 감시 시작: ${resolve(dir)}`)\n if (outDir) log(`[kordoc watch] 출력: ${resolve(outDir)}`)\n if (webhook) log(`[kordoc watch] 웹훅: ${webhook}`)\n\n // 디바운스 맵\n const pending = new Map<string, ReturnType<typeof setTimeout>>()\n // 동시 처리 제한 — 메모리 폭주 방지\n const MAX_CONCURRENT = 3\n let activeCount = 0\n const inProgress = new Set<string>()\n\n /** 파일 크기가 안정화될 때까지 대기 (쓰기 완료 감지) */\n const waitForStableSize = async (absPath: string): Promise<number> => {\n let prevSize = statSync(absPath).size\n await new Promise(r => setTimeout(r, STABLE_CHECK_MS))\n if (!existsSync(absPath)) return 0\n const currSize = statSync(absPath).size\n if (currSize !== prevSize) {\n // 크기가 변했으면 한 번 더 대기\n await new Promise(r => setTimeout(r, STABLE_CHECK_MS))\n if (!existsSync(absPath)) return 0\n return statSync(absPath).size\n }\n return currSize\n }\n\n const processFile = async (filePath: string) => {\n const ext = extname(filePath).toLowerCase()\n if (!SUPPORTED_EXTENSIONS.has(ext)) return\n // 동일 파일 동시 처리 방지 + 동시 처리 수 제한\n if (inProgress.has(filePath) || activeCount >= MAX_CONCURRENT) return\n inProgress.add(filePath)\n activeCount++\n\n const fileName = basename(filePath)\n try {\n const rawPath = resolve(dir, filePath)\n if (!existsSync(rawPath)) return\n // 심볼릭 링크 해석 후 감시 디렉토리 외부 파일 차단\n let absPath: string\n try { absPath = realpathSync(rawPath) } catch { return }\n const realDir = realpathSync(resolve(dir))\n if (!absPath.startsWith(realDir + sep) && absPath !== realDir) return\n\n const fileSize = await waitForStableSize(absPath)\n if (fileSize > MAX_FILE_SIZE || fileSize === 0) return\n\n log(`[kordoc watch] 변환 중: ${fileName}`)\n\n const buffer = readFileSync(absPath)\n const arrayBuffer = toArrayBuffer(buffer)\n const parseOptions = pages ? { pages } : undefined\n const result = await parse(arrayBuffer, parseOptions)\n\n if (!result.success) {\n log(`[kordoc watch] 실패: ${fileName} — ${result.error}`)\n await sendWebhook(webhook, { file: fileName, format: detectFormat(arrayBuffer), success: false, error: result.error })\n return\n }\n\n const output = format === \"json\" ? JSON.stringify(result, null, 2) : result.markdown\n\n if (outDir) {\n const outExt = format === \"json\" ? \".json\" : \".md\"\n const outPath = resolve(outDir, fileName.replace(/\\.[^.]+$/, outExt))\n writeFileSync(outPath, output, \"utf-8\")\n log(`[kordoc watch] 완료: ${fileName} → ${basename(outPath)}`)\n } else {\n process.stdout.write(output + \"\\n\")\n }\n\n await sendWebhook(webhook, {\n file: fileName,\n format: result.fileType,\n success: true,\n markdown: format === \"markdown\" ? output.substring(0, 1000) : undefined,\n })\n } catch (err) {\n log(`[kordoc watch] 에러: ${fileName} — ${err instanceof Error ? err.message : err}`)\n } finally {\n inProgress.delete(filePath)\n activeCount--\n }\n }\n\n // fs.watch recursive (Node 18+ Windows/macOS, Node 19+ Linux)\n watch(dir, { recursive: true }, (event, filename) => {\n if (!filename) return\n const filePath = filename.toString()\n\n // 디바운스\n const existing = pending.get(filePath)\n if (existing) clearTimeout(existing)\n pending.set(filePath, setTimeout(() => {\n pending.delete(filePath)\n processFile(filePath).catch((err) => {\n process.stderr.write(`[kordoc watch] 처리 실패: ${filePath} — ${err instanceof Error ? err.message : String(err)}\\n`)\n })\n }, DEBOUNCE_MS))\n })\n\n // 프로세스 종료 방지 (Ctrl+C로 종료)\n return new Promise(() => {})\n}\n\n/** Webhook URL 검증 — SSRF 방지: http/https만 허용, localhost/private IP 차단 */\nfunction validateWebhookUrl(url: string): void {\n let parsed: URL\n try {\n parsed = new URL(url)\n } catch {\n throw new Error(`유효하지 않은 webhook URL: ${url}`)\n }\n if (parsed.protocol !== \"http:\" && parsed.protocol !== \"https:\") {\n throw new Error(`허용되지 않는 webhook 프로토콜: ${parsed.protocol}`)\n }\n const hostname = parsed.hostname.toLowerCase()\n if (\n hostname === \"localhost\" ||\n hostname === \"[::1]\" ||\n hostname.startsWith(\"127.\") ||\n hostname.startsWith(\"10.\") ||\n hostname.startsWith(\"192.168.\") ||\n /^172\\.(1[6-9]|2\\d|3[01])\\./.test(hostname) ||\n hostname === \"0.0.0.0\" ||\n hostname.startsWith(\"169.254.\") ||\n hostname.endsWith(\".local\") ||\n // IPv6 사설 대역\n hostname.startsWith(\"[fc\") ||\n hostname.startsWith(\"[fd\") ||\n hostname.startsWith(\"[fe80:\") ||\n hostname === \"[::0]\" ||\n hostname === \"[::]\" ||\n // 클라우드 메타데이터 엔드포인트\n hostname === \"metadata.google.internal\" ||\n hostname === \"metadata.google\" ||\n // 16진수/8진수/10진수 정수 IP 인코딩 우회 방지\n /^0x[0-9a-f]+$/i.test(hostname) ||\n /^0[0-7]+$/.test(hostname) ||\n /^\\d+$/.test(hostname)\n ) {\n throw new Error(`내부 네트워크 대상 webhook은 허용되지 않습니다: ${hostname}`)\n }\n}\n\nasync function sendWebhook(url: string | undefined, payload: Record<string, unknown>): Promise<void> {\n if (!url) return\n try {\n validateWebhookUrl(url)\n await fetch(url, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({ ...payload, timestamp: new Date().toISOString() }),\n redirect: \"error\",\n })\n } catch (err) {\n process.stderr.write(`[kordoc watch] webhook 전송 실패: ${err instanceof Error ? err.message : String(err)}\\n`)\n }\n}\n"],"mappings":";;;;;;;;;;;;;AAEA,SAAS,OAAO,cAAc,eAAe,WAAW,UAAU,YAAY,oBAAoB;AAClG,SAAS,UAAU,SAAS,SAAS,WAAW;AAKhD,IAAM,uBAAuB,oBAAI,IAAI,CAAC,QAAQ,SAAS,QAAQ,SAAS,OAAO,CAAC;AAChF,IAAM,cAAc;AAEpB,IAAM,kBAAkB;AACxB,IAAM,gBAAgB,MAAM,OAAO;AAUnC,eAAsB,eAAe,SAAsC;AACzE,QAAM,EAAE,KAAK,QAAQ,SAAS,SAAS,YAAY,OAAO,OAAO,IAAI;AAErE,MAAI,CAAC,WAAW,GAAG,EAAG,OAAM,IAAI,MAAM,gFAAoB,GAAG,EAAE;AAC/D,MAAI,QAAS,oBAAmB,OAAO;AACvC,MAAI,OAAQ,WAAU,QAAQ,EAAE,WAAW,KAAK,CAAC;AAEjD,QAAM,MAAM,SAAS,MAAM;AAAA,EAAC,IAAI,CAAC,QAAgB,QAAQ,OAAO,MAAM,MAAM,IAAI;AAChF,MAAI,6CAAyB,QAAQ,GAAG,CAAC,EAAE;AAC3C,MAAI,OAAQ,KAAI,gCAAsB,QAAQ,MAAM,CAAC,EAAE;AACvD,MAAI,QAAS,KAAI,gCAAsB,OAAO,EAAE;AAGhD,QAAM,UAAU,oBAAI,IAA2C;AAE/D,QAAM,iBAAiB;AACvB,MAAI,cAAc;AAClB,QAAM,aAAa,oBAAI,IAAY;AAGnC,QAAM,oBAAoB,OAAO,YAAqC;AACpE,QAAI,WAAW,SAAS,OAAO,EAAE;AACjC,UAAM,IAAI,QAAQ,OAAK,WAAW,GAAG,eAAe,CAAC;AACrD,QAAI,CAAC,WAAW,OAAO,EAAG,QAAO;AACjC,UAAM,WAAW,SAAS,OAAO,EAAE;AACnC,QAAI,aAAa,UAAU;AAEzB,YAAM,IAAI,QAAQ,OAAK,WAAW,GAAG,eAAe,CAAC;AACrD,UAAI,CAAC,WAAW,OAAO,EAAG,QAAO;AACjC,aAAO,SAAS,OAAO,EAAE;AAAA,IAC3B;AACA,WAAO;AAAA,EACT;AAEA,QAAM,cAAc,OAAO,aAAqB;AAC9C,UAAM,MAAM,QAAQ,QAAQ,EAAE,YAAY;AAC1C,QAAI,CAAC,qBAAqB,IAAI,GAAG,EAAG;AAEpC,QAAI,WAAW,IAAI,QAAQ,KAAK,eAAe,eAAgB;AAC/D,eAAW,IAAI,QAAQ;AACvB;AAEA,UAAM,WAAW,SAAS,QAAQ;AAClC,QAAI;AACF,YAAM,UAAU,QAAQ,KAAK,QAAQ;AACrC,UAAI,CAAC,WAAW,OAAO,EAAG;AAE1B,UAAI;AACJ,UAAI;AAAE,kBAAU,aAAa,OAAO;AAAA,MAAE,QAAQ;AAAE;AAAA,MAAO;AACvD,YAAM,UAAU,aAAa,QAAQ,GAAG,CAAC;AACzC,UAAI,CAAC,QAAQ,WAAW,UAAU,GAAG,KAAK,YAAY,QAAS;AAE/D,YAAM,WAAW,MAAM,kBAAkB,OAAO;AAChD,UAAI,WAAW,iBAAiB,aAAa,EAAG;AAEhD,UAAI,uCAAwB,QAAQ,EAAE;AAEtC,YAAM,SAAS,aAAa,OAAO;AACnC,YAAM,cAAc,cAAc,MAAM;AACxC,YAAM,eAAe,QAAQ,EAAE,MAAM,IAAI;AACzC,YAAM,SAAS,MAAM,MAAM,aAAa,YAAY;AAEpD,UAAI,CAAC,OAAO,SAAS;AACnB,YAAI,gCAAsB,QAAQ,WAAM,OAAO,KAAK,EAAE;AACtD,cAAM,YAAY,SAAS,EAAE,MAAM,UAAU,QAAQ,aAAa,WAAW,GAAG,SAAS,OAAO,OAAO,OAAO,MAAM,CAAC;AACrH;AAAA,MACF;AAEA,YAAM,SAAS,WAAW,SAAS,KAAK,UAAU,QAAQ,MAAM,CAAC,IAAI,OAAO;AAE5E,UAAI,QAAQ;AACV,cAAM,SAAS,WAAW,SAAS,UAAU;AAC7C,cAAM,UAAU,QAAQ,QAAQ,SAAS,QAAQ,YAAY,MAAM,CAAC;AACpE,sBAAc,SAAS,QAAQ,OAAO;AACtC,YAAI,gCAAsB,QAAQ,WAAM,SAAS,OAAO,CAAC,EAAE;AAAA,MAC7D,OAAO;AACL,gBAAQ,OAAO,MAAM,SAAS,IAAI;AAAA,MACpC;AAEA,YAAM,YAAY,SAAS;AAAA,QACzB,MAAM;AAAA,QACN,QAAQ,OAAO;AAAA,QACf,SAAS;AAAA,QACT,UAAU,WAAW,aAAa,OAAO,UAAU,GAAG,GAAI,IAAI;AAAA,MAChE,CAAC;AAAA,IACH,SAAS,KAAK;AACZ,UAAI,gCAAsB,QAAQ,WAAM,eAAe,QAAQ,IAAI,UAAU,GAAG,EAAE;AAAA,IACpF,UAAE;AACA,iBAAW,OAAO,QAAQ;AAC1B;AAAA,IACF;AAAA,EACF;AAGA,QAAM,KAAK,EAAE,WAAW,KAAK,GAAG,CAAC,OAAO,aAAa;AACnD,QAAI,CAAC,SAAU;AACf,UAAM,WAAW,SAAS,SAAS;AAGnC,UAAM,WAAW,QAAQ,IAAI,QAAQ;AACrC,QAAI,SAAU,cAAa,QAAQ;AACnC,YAAQ,IAAI,UAAU,WAAW,MAAM;AACrC,cAAQ,OAAO,QAAQ;AACvB,kBAAY,QAAQ,EAAE,MAAM,CAAC,QAAQ;AACnC,gBAAQ,OAAO,MAAM,6CAAyB,QAAQ,WAAM,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG,CAAC;AAAA,CAAI;AAAA,MAClH,CAAC;AAAA,IACH,GAAG,WAAW,CAAC;AAAA,EACjB,CAAC;AAGD,SAAO,IAAI,QAAQ,MAAM;AAAA,EAAC,CAAC;AAC7B;AAGA,SAAS,mBAAmB,KAAmB;AAC7C,MAAI;AACJ,MAAI;AACF,aAAS,IAAI,IAAI,GAAG;AAAA,EACtB,QAAQ;AACN,UAAM,IAAI,MAAM,sDAAwB,GAAG,EAAE;AAAA,EAC/C;AACA,MAAI,OAAO,aAAa,WAAW,OAAO,aAAa,UAAU;AAC/D,UAAM,IAAI,MAAM,2EAAyB,OAAO,QAAQ,EAAE;AAAA,EAC5D;AACA,QAAM,WAAW,OAAO,SAAS,YAAY;AAC7C,MACE,aAAa,eACb,aAAa,WACb,SAAS,WAAW,MAAM,KAC1B,SAAS,WAAW,KAAK,KACzB,SAAS,WAAW,UAAU,KAC9B,6BAA6B,KAAK,QAAQ,KAC1C,aAAa,aACb,SAAS,WAAW,UAAU,KAC9B,SAAS,SAAS,QAAQ;AAAA,EAE1B,SAAS,WAAW,KAAK,KACzB,SAAS,WAAW,KAAK,KACzB,SAAS,WAAW,QAAQ,KAC5B,aAAa,WACb,aAAa;AAAA,EAEb,aAAa,8BACb,aAAa;AAAA,EAEb,iBAAiB,KAAK,QAAQ,KAC9B,YAAY,KAAK,QAAQ,KACzB,QAAQ,KAAK,QAAQ,GACrB;AACA,UAAM,IAAI,MAAM,uHAAkC,QAAQ,EAAE;AAAA,EAC9D;AACF;AAEA,eAAe,YAAY,KAAyB,SAAiD;AACnG,MAAI,CAAC,IAAK;AACV,MAAI;AACF,uBAAmB,GAAG;AACtB,UAAM,MAAM,KAAK;AAAA,MACf,QAAQ;AAAA,MACR,SAAS,EAAE,gBAAgB,mBAAmB;AAAA,MAC9C,MAAM,KAAK,UAAU,EAAE,GAAG,SAAS,YAAW,oBAAI,KAAK,GAAE,YAAY,EAAE,CAAC;AAAA,MACxE,UAAU;AAAA,IACZ,CAAC;AAAA,EACH,SAAS,KAAK;AACZ,YAAQ,OAAO,MAAM,qDAAiC,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG,CAAC;AAAA,CAAI;AAAA,EAC5G;AACF;","names":[]}
package/package.json CHANGED
@@ -1,98 +1,98 @@
1
- {
2
- "name": "kordoc",
3
- "version": "3.4.1",
4
- "description": "Parse Korean documents (HWP3, HWP, HWPX, HWPML, PDF, XLS, XLSX, DOCX) to Markdown + Print Renderer",
5
- "type": "module",
6
- "exports": {
7
- ".": {
8
- "types": "./dist/index.d.ts",
9
- "import": "./dist/index.js",
10
- "require": "./dist/index.cjs"
11
- }
12
- },
13
- "main": "./dist/index.cjs",
14
- "module": "./dist/index.js",
15
- "types": "./dist/index.d.ts",
16
- "bin": {
17
- "kordoc": "./dist/cli.js",
18
- "kordoc-mcp": "./dist/mcp.js"
19
- },
20
- "files": [
21
- "dist"
22
- ],
23
- "scripts": {
24
- "build": "tsup",
25
- "dev": "tsup --watch",
26
- "test": "node --import tsx --test tests/*.test.ts",
27
- "prepublishOnly": "npm run test && npm run build"
28
- },
29
- "keywords": [
30
- "hwp",
31
- "hwpx",
32
- "hancom",
33
- "hangul",
34
- "korean",
35
- "document",
36
- "parser",
37
- "pdf",
38
- "xls",
39
- "xlsx",
40
- "biff8",
41
- "docx",
42
- "markdown",
43
- "converter",
44
- "mcp",
45
- "model-context-protocol",
46
- "print",
47
- "puppeteer"
48
- ],
49
- "author": "chrisryugj",
50
- "license": "MIT",
51
- "repository": {
52
- "type": "git",
53
- "url": "https://github.com/chrisryugj/kordoc.git"
54
- },
55
- "engines": {
56
- "node": ">=18"
57
- },
58
- "dependencies": {
59
- "@modelcontextprotocol/sdk": "^1.28.0",
60
- "@xmldom/xmldom": "^0.9.8",
61
- "cfb": "1.2.2",
62
- "commander": "^13.0.0",
63
- "jszip": "^3.10.1",
64
- "markdown-it": "^14.1.1",
65
- "zod": "^3.23.0"
66
- },
67
- "peerDependencies": {
68
- "pdfjs-dist": ">=4.0.0",
69
- "puppeteer-core": ">=22.0.0"
70
- },
71
- "peerDependenciesMeta": {
72
- "pdfjs-dist": {
73
- "optional": true
74
- },
75
- "puppeteer-core": {
76
- "optional": true
77
- }
78
- },
79
- "optionalDependencies": {
80
- "@huggingface/transformers": "^4.1.0",
81
- "@hyzyla/pdfium": "^2.1.0",
82
- "onnxruntime-node": "^1.24.0",
83
- "sharp": "^0.34.0"
84
- },
85
- "devDependencies": {
86
- "@huggingface/transformers": "^4.1.0",
87
- "@hyzyla/pdfium": "^2.1.0",
88
- "@types/markdown-it": "^14.1.2",
89
- "@types/node": "^18.19.130",
90
- "onnxruntime-node": "^1.24.0",
91
- "pdfjs-dist": "^4.10.38",
92
- "puppeteer-core": "^24.42.0",
93
- "sharp": "^0.34.0",
94
- "tsup": "^8.4.0",
95
- "tsx": "^4.21.0",
96
- "typescript": "^5.9.0"
97
- }
98
- }
1
+ {
2
+ "name": "kordoc",
3
+ "version": "3.5.1",
4
+ "description": "Parse Korean documents (HWP3, HWP, HWPX, HWPML, PDF, XLS, XLSX, DOCX) to Markdown + Print Renderer",
5
+ "type": "module",
6
+ "exports": {
7
+ ".": {
8
+ "types": "./dist/index.d.ts",
9
+ "import": "./dist/index.js",
10
+ "require": "./dist/index.cjs"
11
+ }
12
+ },
13
+ "main": "./dist/index.cjs",
14
+ "module": "./dist/index.js",
15
+ "types": "./dist/index.d.ts",
16
+ "bin": {
17
+ "kordoc": "./dist/cli.js",
18
+ "kordoc-mcp": "./dist/mcp.js"
19
+ },
20
+ "files": [
21
+ "dist"
22
+ ],
23
+ "scripts": {
24
+ "build": "tsup",
25
+ "dev": "tsup --watch",
26
+ "test": "node --import tsx --test tests/*.test.ts",
27
+ "prepublishOnly": "npm run test && npm run build"
28
+ },
29
+ "keywords": [
30
+ "hwp",
31
+ "hwpx",
32
+ "hancom",
33
+ "hangul",
34
+ "korean",
35
+ "document",
36
+ "parser",
37
+ "pdf",
38
+ "xls",
39
+ "xlsx",
40
+ "biff8",
41
+ "docx",
42
+ "markdown",
43
+ "converter",
44
+ "mcp",
45
+ "model-context-protocol",
46
+ "print",
47
+ "puppeteer"
48
+ ],
49
+ "author": "chrisryugj",
50
+ "license": "MIT",
51
+ "repository": {
52
+ "type": "git",
53
+ "url": "https://github.com/chrisryugj/kordoc.git"
54
+ },
55
+ "engines": {
56
+ "node": ">=18"
57
+ },
58
+ "dependencies": {
59
+ "@modelcontextprotocol/sdk": "^1.28.0",
60
+ "@xmldom/xmldom": "^0.9.8",
61
+ "cfb": "1.2.2",
62
+ "commander": "^13.0.0",
63
+ "jszip": "^3.10.1",
64
+ "markdown-it": "^14.1.1",
65
+ "zod": "^3.23.0"
66
+ },
67
+ "peerDependencies": {
68
+ "pdfjs-dist": ">=4.0.0",
69
+ "puppeteer-core": ">=22.0.0"
70
+ },
71
+ "peerDependenciesMeta": {
72
+ "pdfjs-dist": {
73
+ "optional": true
74
+ },
75
+ "puppeteer-core": {
76
+ "optional": true
77
+ }
78
+ },
79
+ "optionalDependencies": {
80
+ "@huggingface/transformers": "^4.1.0",
81
+ "@hyzyla/pdfium": "^2.1.0",
82
+ "onnxruntime-node": "^1.24.0",
83
+ "sharp": "^0.34.0"
84
+ },
85
+ "devDependencies": {
86
+ "@huggingface/transformers": "^4.1.0",
87
+ "@hyzyla/pdfium": "^2.1.0",
88
+ "@types/markdown-it": "^14.1.2",
89
+ "@types/node": "^18.19.130",
90
+ "onnxruntime-node": "^1.24.0",
91
+ "pdfjs-dist": "^4.10.38",
92
+ "puppeteer-core": "^24.42.0",
93
+ "sharp": "^0.34.0",
94
+ "tsup": "^8.4.0",
95
+ "tsx": "^4.21.0",
96
+ "typescript": "^5.9.0"
97
+ }
98
+ }
@@ -1 +0,0 @@
1
- {"version":3,"sources":["../src/page-range.ts"],"sourcesContent":["/** 페이지/섹션 범위 파싱 유틸리티 */\r\n\r\n/**\r\n * 페이지 범위 지정을 1-based Set<number>로 변환.\r\n *\r\n * @param spec - [1,2,3] 또는 \"1-3\" 또는 \"1,3,5-7\"\r\n * @param maxPages - 최대 페이지 수 (클램핑 상한)\r\n * @returns 1-based 페이지 번호 Set\r\n */\r\nexport function parsePageRange(spec: number[] | string, maxPages: number): Set<number> {\r\n const result = new Set<number>()\r\n if (maxPages <= 0) return result\r\n\r\n if (Array.isArray(spec)) {\r\n for (const n of spec) {\r\n const page = Math.round(n)\r\n if (page >= 1 && page <= maxPages) result.add(page)\r\n }\r\n return result\r\n }\r\n\r\n if (typeof spec !== \"string\" || spec.trim() === \"\") return result\r\n\r\n const parts = spec.split(\",\")\r\n for (const part of parts) {\r\n const trimmed = part.trim()\r\n if (!trimmed) continue\r\n\r\n const rangeMatch = trimmed.match(/^(\\d+)\\s*-\\s*(\\d+)$/)\r\n if (rangeMatch) {\r\n const start = Math.max(1, parseInt(rangeMatch[1], 10))\r\n const end = Math.min(maxPages, parseInt(rangeMatch[2], 10))\r\n for (let i = start; i <= end; i++) result.add(i)\r\n } else {\r\n const page = parseInt(trimmed, 10)\r\n if (!isNaN(page) && page >= 1 && page <= maxPages) result.add(page)\r\n }\r\n }\r\n\r\n return result\r\n}\r\n"],"mappings":";;;AASO,SAAS,eAAe,MAAyB,UAA+B;AACrF,QAAM,SAAS,oBAAI,IAAY;AAC/B,MAAI,YAAY,EAAG,QAAO;AAE1B,MAAI,MAAM,QAAQ,IAAI,GAAG;AACvB,eAAW,KAAK,MAAM;AACpB,YAAM,OAAO,KAAK,MAAM,CAAC;AACzB,UAAI,QAAQ,KAAK,QAAQ,SAAU,QAAO,IAAI,IAAI;AAAA,IACpD;AACA,WAAO;AAAA,EACT;AAEA,MAAI,OAAO,SAAS,YAAY,KAAK,KAAK,MAAM,GAAI,QAAO;AAE3D,QAAM,QAAQ,KAAK,MAAM,GAAG;AAC5B,aAAW,QAAQ,OAAO;AACxB,UAAM,UAAU,KAAK,KAAK;AAC1B,QAAI,CAAC,QAAS;AAEd,UAAM,aAAa,QAAQ,MAAM,qBAAqB;AACtD,QAAI,YAAY;AACd,YAAM,QAAQ,KAAK,IAAI,GAAG,SAAS,WAAW,CAAC,GAAG,EAAE,CAAC;AACrD,YAAM,MAAM,KAAK,IAAI,UAAU,SAAS,WAAW,CAAC,GAAG,EAAE,CAAC;AAC1D,eAAS,IAAI,OAAO,KAAK,KAAK,IAAK,QAAO,IAAI,CAAC;AAAA,IACjD,OAAO;AACL,YAAM,OAAO,SAAS,SAAS,EAAE;AACjC,UAAI,CAAC,MAAM,IAAI,KAAK,QAAQ,KAAK,QAAQ,SAAU,QAAO,IAAI,IAAI;AAAA,IACpE;AAAA,EACF;AAEA,SAAO;AACT;","names":[]}
@@ -1 +0,0 @@
1
- {"version":3,"sources":["../src/page-range.ts"],"sourcesContent":["/** 페이지/섹션 범위 파싱 유틸리티 */\r\n\r\n/**\r\n * 페이지 범위 지정을 1-based Set<number>로 변환.\r\n *\r\n * @param spec - [1,2,3] 또는 \"1-3\" 또는 \"1,3,5-7\"\r\n * @param maxPages - 최대 페이지 수 (클램핑 상한)\r\n * @returns 1-based 페이지 번호 Set\r\n */\r\nexport function parsePageRange(spec: number[] | string, maxPages: number): Set<number> {\r\n const result = new Set<number>()\r\n if (maxPages <= 0) return result\r\n\r\n if (Array.isArray(spec)) {\r\n for (const n of spec) {\r\n const page = Math.round(n)\r\n if (page >= 1 && page <= maxPages) result.add(page)\r\n }\r\n return result\r\n }\r\n\r\n if (typeof spec !== \"string\" || spec.trim() === \"\") return result\r\n\r\n const parts = spec.split(\",\")\r\n for (const part of parts) {\r\n const trimmed = part.trim()\r\n if (!trimmed) continue\r\n\r\n const rangeMatch = trimmed.match(/^(\\d+)\\s*-\\s*(\\d+)$/)\r\n if (rangeMatch) {\r\n const start = Math.max(1, parseInt(rangeMatch[1], 10))\r\n const end = Math.min(maxPages, parseInt(rangeMatch[2], 10))\r\n for (let i = start; i <= end; i++) result.add(i)\r\n } else {\r\n const page = parseInt(trimmed, 10)\r\n if (!isNaN(page) && page >= 1 && page <= maxPages) result.add(page)\r\n }\r\n }\r\n\r\n return result\r\n}\r\n"],"mappings":";AASO,SAAS,eAAe,MAAyB,UAA+B;AACrF,QAAM,SAAS,oBAAI,IAAY;AAC/B,MAAI,YAAY,EAAG,QAAO;AAE1B,MAAI,MAAM,QAAQ,IAAI,GAAG;AACvB,eAAW,KAAK,MAAM;AACpB,YAAM,OAAO,KAAK,MAAM,CAAC;AACzB,UAAI,QAAQ,KAAK,QAAQ,SAAU,QAAO,IAAI,IAAI;AAAA,IACpD;AACA,WAAO;AAAA,EACT;AAEA,MAAI,OAAO,SAAS,YAAY,KAAK,KAAK,MAAM,GAAI,QAAO;AAE3D,QAAM,QAAQ,KAAK,MAAM,GAAG;AAC5B,aAAW,QAAQ,OAAO;AACxB,UAAM,UAAU,KAAK,KAAK;AAC1B,QAAI,CAAC,QAAS;AAEd,UAAM,aAAa,QAAQ,MAAM,qBAAqB;AACtD,QAAI,YAAY;AACd,YAAM,QAAQ,KAAK,IAAI,GAAG,SAAS,WAAW,CAAC,GAAG,EAAE,CAAC;AACrD,YAAM,MAAM,KAAK,IAAI,UAAU,SAAS,WAAW,CAAC,GAAG,EAAE,CAAC;AAC1D,eAAS,IAAI,OAAO,KAAK,KAAK,IAAK,QAAO,IAAI,CAAC;AAAA,IACjD,OAAO;AACL,YAAM,OAAO,SAAS,SAAS,EAAE;AACjC,UAAI,CAAC,MAAM,IAAI,KAAK,QAAQ,KAAK,QAAQ,SAAU,QAAO,IAAI,IAAI;AAAA,IACpE;AAAA,EACF;AAEA,SAAO;AACT;","names":[]}
@@ -1 +0,0 @@
1
- {"version":3,"sources":["../src/detect.ts","../src/hwp5/cfb-lenient.ts"],"sourcesContent":["/** 매직 바이트 기반 파일 포맷 감지 */\r\n\r\nimport JSZip from \"jszip\"\r\nimport type { FileType } from \"./types.js\"\r\nimport { parseLenientCfb } from \"./hwp5/cfb-lenient.js\"\r\n\r\n/** 매직 바이트 뷰 생성 (복사 없이 view) */\r\nfunction magicBytes(buffer: ArrayBuffer): Uint8Array {\r\n return new Uint8Array(buffer, 0, Math.min(4, buffer.byteLength))\r\n}\r\n\r\n/** ZIP 파일 여부: PK\\x03\\x04 */\r\nexport function isZipFile(buffer: ArrayBuffer): boolean {\r\n const b = magicBytes(buffer)\r\n return b[0] === 0x50 && b[1] === 0x4b && b[2] === 0x03 && b[3] === 0x04\r\n}\r\n\r\n/** HWPX (ZIP 기반 한컴 문서): PK\\x03\\x04 — 하위 호환용 */\r\nexport function isHwpxFile(buffer: ArrayBuffer): boolean {\r\n return isZipFile(buffer)\r\n}\r\n\r\n/** HWP 5.x (OLE2 바이너리 한컴 문서): \\xD0\\xCF\\x11\\xE0 */\r\nexport function isOldHwpFile(buffer: ArrayBuffer): boolean {\r\n const b = magicBytes(buffer)\r\n return b[0] === 0xd0 && b[1] === 0xcf && b[2] === 0x11 && b[3] === 0xe0\r\n}\r\n\r\n/**\r\n * HWP 3.x (한글 워드프로세서 3.0): \"HWP Document File V3.00 \\x1A\\x01\\x02\\x03\\x04\\x05\" 30 byte.\r\n * CFB(OLE2) 컨테이너 아닌 단일 binary stream — isOldHwpFile 과 magic 이 다르다.\r\n */\r\nconst HWP3_PREFIX = new TextEncoder().encode(\"HWP Document File V3.00\")\r\nexport function isHwp3File(buffer: ArrayBuffer): boolean {\r\n if (buffer.byteLength < HWP3_PREFIX.length) return false\r\n const head = new Uint8Array(buffer, 0, HWP3_PREFIX.length)\r\n for (let i = 0; i < HWP3_PREFIX.length; i++) {\r\n if (head[i] !== HWP3_PREFIX[i]) return false\r\n }\r\n return true\r\n}\r\n\r\n/** PDF 문서: %PDF */\r\nexport function isPdfFile(buffer: ArrayBuffer): boolean {\r\n const b = magicBytes(buffer)\r\n return b[0] === 0x25 && b[1] === 0x50 && b[2] === 0x44 && b[3] === 0x46\r\n}\r\n\r\n/** HWPML (XML 기반 한컴 문서): <?xml ... <HWPML */\r\nexport function isHwpmlFile(buffer: ArrayBuffer): boolean {\r\n const bytes = new Uint8Array(buffer, 0, Math.min(512, buffer.byteLength))\r\n const head = new TextDecoder(\"utf-8\", { fatal: false }).decode(bytes).replace(/^\\uFEFF/, \"\")\r\n return head.trimStart().startsWith(\"<?xml\") && head.includes(\"<HWPML\")\r\n}\r\n\r\n/** 동기 포맷 감지 — ZIP은 모두 \"hwpx\"로 반환 (하위 호환) */\r\nexport function detectFormat(buffer: ArrayBuffer): FileType {\r\n if (buffer.byteLength < 4) return \"unknown\"\r\n if (isHwp3File(buffer)) return \"hwp3\"\r\n if (isZipFile(buffer)) return \"hwpx\"\r\n if (isOldHwpFile(buffer)) return \"hwp\"\r\n if (isPdfFile(buffer)) return \"pdf\"\r\n if (isHwpmlFile(buffer)) return \"hwpml\"\r\n return \"unknown\"\r\n}\r\n\r\n/**\r\n * OLE2 컨테이너 내부 스트림 기반 포맷 세분화.\r\n * HWP 5.x, XLS 모두 OLE2이므로 스트림 이름으로 구분.\r\n * - \"Workbook\" 또는 \"Book\" → 'xls'\r\n * - 그 외 (FileHeader 등) → 'hwp'\r\n */\r\nexport function detectOle2Format(buffer: ArrayBuffer): \"hwp\" | \"xls\" | \"unknown\" {\r\n try {\r\n const cfb = parseLenientCfb(Buffer.from(buffer))\r\n const names = cfb.entries().map(e => e.name)\r\n if (names.includes(\"Workbook\") || names.includes(\"Book\")) return \"xls\"\r\n if (names.includes(\"FileHeader\")) return \"hwp\"\r\n // FileHeader 없어도 BodyText/DocInfo 있으면 hwp\r\n if (names.some(n => n === \"DocInfo\" || n.startsWith(\"Section\"))) return \"hwp\"\r\n return \"unknown\"\r\n } catch {\r\n return \"unknown\"\r\n }\r\n}\r\n\r\n/**\r\n * ZIP 내부 구조 기반 포맷 세분화.\r\n * HWPX, XLSX, DOCX 모두 ZIP이므로 내부 파일로 구분.\r\n */\r\nexport async function detectZipFormat(buffer: ArrayBuffer): Promise<\"hwpx\" | \"xlsx\" | \"docx\" | \"unknown\"> {\r\n try {\r\n const zip = await JSZip.loadAsync(buffer)\r\n // XLSX: xl/workbook.xml\r\n if (zip.file(\"xl/workbook.xml\")) return \"xlsx\"\r\n // DOCX: word/document.xml\r\n if (zip.file(\"word/document.xml\")) return \"docx\"\r\n // HWPX: Contents/ 또는 content.hpf 또는 mimetype\r\n if (zip.file(\"Contents/content.hpf\") || zip.file(\"mimetype\")) return \"hwpx\"\r\n // 기타 ZIP 내에 section 파일이 있으면 HWPX로 추정\r\n const hasSection = Object.keys(zip.files).some(f => f.startsWith(\"Contents/\"))\r\n if (hasSection) return \"hwpx\"\r\n return \"unknown\"\r\n } catch {\r\n return \"unknown\"\r\n }\r\n}\r\n","/**\r\n * Lenient CFB (Compound File Binary / OLE2) 파서.\r\n *\r\n * 표준 cfb 모듈이 FAT 검증 실패로 거부하는 손상된 HWP 파일을 열기 위한 폴백.\r\n * 직접 헤더/FAT/디렉토리를 파싱하여 스트림 데이터를 추출.\r\n *\r\n * 참조: rhwp (MIT) src/parser/cfb_reader.rs (LenientCfbReader)\r\n * 참조: MS-CFB spec (https://learn.microsoft.com/en-us/openspecs/windows_protocols/ms-cfb)\r\n */\r\n\r\nimport { decompressStream } from \"./record.js\"\r\n\r\n// ── 상수 ──\r\n\r\nconst CFB_MAGIC = Buffer.from([0xd0, 0xcf, 0x11, 0xe0, 0xa1, 0xb1, 0x1a, 0xe1])\r\nconst END_OF_CHAIN = 0xfffffffe\r\nconst FREE_SECT = 0xffffffff\r\n\r\n/** 순환 감지용 최대 체인 길이 */\r\nconst MAX_CHAIN_LENGTH = 1_000_000\r\n/** 최대 디렉토리 엔트리 수 */\r\nconst MAX_DIR_ENTRIES = 100_000\r\n/** 최대 스트림 크기 (100MB) */\r\nconst MAX_STREAM_SIZE = 100 * 1024 * 1024\r\n\r\n// ── 디렉토리 엔트리 ──\r\n\r\ninterface DirEntry {\r\n name: string\r\n type: number // 0=unknown, 1=storage, 2=stream, 5=root\r\n startSector: number\r\n size: number\r\n}\r\n\r\n// ── CFB 컨테이너 ──\r\n\r\nexport interface LenientCfbContainer {\r\n /** 이름 기반 스트림 탐색 */\r\n findStream(path: string): Buffer | null\r\n /** 디렉토리 엔트리 목록 */\r\n entries(): DirEntry[]\r\n}\r\n\r\n// ── 구현 ──\r\n\r\nexport function parseLenientCfb(data: Buffer): LenientCfbContainer {\r\n if (data.length < 512) throw new Error(\"CFB 파일이 너무 짧습니다 (최소 512바이트)\")\r\n if (!data.subarray(0, 8).equals(CFB_MAGIC)) throw new Error(\"CFB 매직 바이트 불일치\")\r\n\r\n // ── 헤더 파싱 ──\r\n\r\n const sectorSizeShift = data.readUInt16LE(30)\r\n if (sectorSizeShift < 7 || sectorSizeShift > 16) throw new Error(\"유효하지 않은 섹터 크기 시프트: \" + sectorSizeShift)\r\n const sectorSize = 1 << sectorSizeShift // 보통 512\r\n const miniSectorSizeShift = data.readUInt16LE(32)\r\n if (miniSectorSizeShift > 16) throw new Error(\"유효하지 않은 미니 섹터 크기 시프트: \" + miniSectorSizeShift)\r\n const miniSectorSize = 1 << miniSectorSizeShift // 보통 64\r\n\r\n const fatSectorCount = data.readUInt32LE(44)\r\n if (fatSectorCount > 10000) throw new Error(\"FAT 섹터 수가 너무 많습니다: \" + fatSectorCount)\r\n const firstDirSector = data.readUInt32LE(48)\r\n const miniStreamCutoff = data.readUInt32LE(56) // 보통 4096\r\n const firstMiniFatSector = data.readUInt32LE(60)\r\n const miniFatSectorCount = data.readUInt32LE(64)\r\n const firstDifatSector = data.readUInt32LE(68)\r\n const difatSectorCount = data.readUInt32LE(72)\r\n\r\n // ── 유틸 ──\r\n\r\n function sectorOffset(id: number): number {\r\n return 512 + id * sectorSize\r\n }\r\n\r\n function readSectorData(id: number): Buffer {\r\n const off = sectorOffset(id)\r\n if (off + sectorSize > data.length) return Buffer.alloc(0)\r\n return data.subarray(off, off + sectorSize)\r\n }\r\n\r\n // ── DIFAT → FAT 섹터 목록 ──\r\n\r\n const fatSectors: number[] = []\r\n\r\n // 헤더 내 DIFAT (최대 109개)\r\n for (let i = 0; i < 109 && fatSectors.length < fatSectorCount; i++) {\r\n const sid = data.readUInt32LE(76 + i * 4)\r\n if (sid === FREE_SECT || sid === END_OF_CHAIN) break\r\n fatSectors.push(sid)\r\n }\r\n\r\n // 추가 DIFAT 섹터 체인\r\n let difatSector = firstDifatSector\r\n const visitedDifat = new Set<number>()\r\n for (let d = 0; d < difatSectorCount && difatSector !== END_OF_CHAIN && difatSector !== FREE_SECT; d++) {\r\n if (visitedDifat.has(difatSector)) break\r\n visitedDifat.add(difatSector)\r\n\r\n const buf = readSectorData(difatSector)\r\n const entriesPerSector = (sectorSize / 4) - 1 // 마지막 4바이트는 다음 DIFAT 포인터\r\n for (let i = 0; i < entriesPerSector && fatSectors.length < fatSectorCount; i++) {\r\n const sid = buf.readUInt32LE(i * 4)\r\n if (sid === FREE_SECT || sid === END_OF_CHAIN) continue\r\n fatSectors.push(sid)\r\n }\r\n difatSector = buf.readUInt32LE(entriesPerSector * 4)\r\n }\r\n\r\n // ── FAT 테이블 구축 ──\r\n\r\n const entriesPerFatSector = sectorSize / 4\r\n const fatTable = new Uint32Array(fatSectors.length * entriesPerFatSector)\r\n\r\n for (let fi = 0; fi < fatSectors.length; fi++) {\r\n const buf = readSectorData(fatSectors[fi])\r\n for (let i = 0; i < entriesPerFatSector; i++) {\r\n fatTable[fi * entriesPerFatSector + i] = i * 4 + 3 < buf.length\r\n ? buf.readUInt32LE(i * 4)\r\n : FREE_SECT\r\n }\r\n }\r\n\r\n // ── 체인 리더 (순환 방지) ──\r\n\r\n function readChain(startSector: number, maxBytes: number): Buffer {\r\n if (startSector === END_OF_CHAIN || startSector === FREE_SECT) return Buffer.alloc(0)\r\n if (maxBytes > MAX_STREAM_SIZE) throw new Error(\"스트림이 너무 큽니다\")\r\n\r\n const chunks: Buffer[] = []\r\n let current = startSector\r\n let totalRead = 0\r\n const visited = new Set<number>()\r\n\r\n while (current !== END_OF_CHAIN && current !== FREE_SECT && totalRead < maxBytes) {\r\n if (visited.has(current)) break // 순환 감지\r\n if (visited.size > MAX_CHAIN_LENGTH) break\r\n visited.add(current)\r\n\r\n const buf = readSectorData(current)\r\n const remaining = maxBytes - totalRead\r\n chunks.push(remaining < sectorSize ? buf.subarray(0, remaining) : buf)\r\n totalRead += Math.min(buf.length, remaining)\r\n\r\n current = current < fatTable.length ? fatTable[current] : END_OF_CHAIN\r\n }\r\n\r\n return Buffer.concat(chunks)\r\n }\r\n\r\n // ── Mini-FAT 테이블 ──\r\n\r\n let miniFatTable: Uint32Array | null = null\r\n\r\n function getMiniFatTable(): Uint32Array {\r\n if (miniFatTable) return miniFatTable\r\n\r\n if (miniFatSectorCount === 0 || firstMiniFatSector === END_OF_CHAIN) {\r\n miniFatTable = new Uint32Array(0)\r\n return miniFatTable\r\n }\r\n\r\n const miniFatData = readChain(firstMiniFatSector, miniFatSectorCount * sectorSize)\r\n const entries = miniFatData.length / 4\r\n miniFatTable = new Uint32Array(entries)\r\n for (let i = 0; i < entries; i++) {\r\n miniFatTable[i] = miniFatData.readUInt32LE(i * 4)\r\n }\r\n return miniFatTable\r\n }\r\n\r\n // ── 디렉토리 엔트리 파싱 ──\r\n\r\n const dirData = readChain(firstDirSector, MAX_DIR_ENTRIES * 128)\r\n const dirEntries: DirEntry[] = []\r\n\r\n for (let offset = 0; offset + 128 <= dirData.length && dirEntries.length < MAX_DIR_ENTRIES; offset += 128) {\r\n const nameLen = dirData.readUInt16LE(offset + 64) // 바이트 수 (null 포함)\r\n if (nameLen <= 0 || nameLen > 64) {\r\n dirEntries.push({ name: \"\", type: 0, startSector: 0, size: 0 })\r\n continue\r\n }\r\n\r\n const nameBytes = nameLen - 2 // null terminator 제외\r\n const name = nameBytes > 0\r\n ? dirData.subarray(offset, offset + nameBytes).toString(\"utf16le\")\r\n : \"\"\r\n\r\n const type = dirData[offset + 66]\r\n const startSector = dirData.readUInt32LE(offset + 116)\r\n // CFBv3에서는 size가 u32 (offset 120), v4에서는 u64\r\n const size = dirData.readUInt32LE(offset + 120)\r\n\r\n dirEntries.push({ name, type, startSector, size })\r\n }\r\n\r\n // ── Root 엔트리에서 미니 스트림 추출 ──\r\n\r\n let miniStreamData: Buffer | null = null\r\n\r\n function getMiniStream(): Buffer {\r\n if (miniStreamData) return miniStreamData\r\n const root = dirEntries[0]\r\n if (!root || root.type !== 5) {\r\n miniStreamData = Buffer.alloc(0)\r\n return miniStreamData\r\n }\r\n miniStreamData = readChain(root.startSector, root.size || MAX_STREAM_SIZE)\r\n return miniStreamData\r\n }\r\n\r\n // ── 미니 스트림에서 읽기 ──\r\n\r\n function readMiniStream(startSector: number, size: number): Buffer {\r\n const mft = getMiniFatTable()\r\n const ms = getMiniStream()\r\n if (mft.length === 0 || ms.length === 0) return Buffer.alloc(0)\r\n\r\n const chunks: Buffer[] = []\r\n let current = startSector\r\n let totalRead = 0\r\n const visited = new Set<number>()\r\n\r\n while (current !== END_OF_CHAIN && current !== FREE_SECT && totalRead < size) {\r\n if (visited.has(current)) break\r\n if (visited.size > MAX_CHAIN_LENGTH) break\r\n visited.add(current)\r\n\r\n const off = current * miniSectorSize\r\n const remaining = size - totalRead\r\n const chunkSize = Math.min(miniSectorSize, remaining)\r\n if (off + chunkSize <= ms.length) {\r\n chunks.push(ms.subarray(off, off + chunkSize))\r\n }\r\n totalRead += chunkSize\r\n\r\n current = current < mft.length ? mft[current] : END_OF_CHAIN\r\n }\r\n\r\n return Buffer.concat(chunks)\r\n }\r\n\r\n // ── 스트림 읽기 (일반/미니 자동 분기) ──\r\n\r\n function readStreamData(entry: DirEntry): Buffer {\r\n if (entry.size === 0) return Buffer.alloc(0)\r\n if (entry.size < miniStreamCutoff) {\r\n const miniResult = readMiniStream(entry.startSector, entry.size)\r\n // 미니스트림이 비어있으면 일반 체인으로 폴백 (lenient)\r\n if (miniResult.length > 0) return miniResult\r\n }\r\n return readChain(entry.startSector, entry.size)\r\n }\r\n\r\n // ── 경로 기반 탐색 ──\r\n\r\n // 전체 경로 맵 구축 (간이: 이름 기반 flat lookup)\r\n // HWP 파일의 디렉토리 구조는 보통 1~2 depth이므로 이름 매칭으로 충분\r\n function findEntryByPath(path: string): DirEntry | null {\r\n // \"/FileHeader\" → \"FileHeader\"\r\n // \"/BodyText/Section0\" → path component matching\r\n const parts = path.replace(/^\\//, \"\").split(\"/\")\r\n\r\n if (parts.length === 1) {\r\n // 단일 이름 매칭\r\n return dirEntries.find(e => e.name === parts[0] && e.type === 2) ?? null\r\n }\r\n\r\n // 2-depth: storage/stream\r\n // HWP 구조: Root/BodyText/Section0, Root/DocInfo, Root/BinData/BIN0001 등\r\n const storageName = parts[0]\r\n const streamName = parts.slice(1).join(\"/\")\r\n\r\n // 디렉토리 구조 대신 이름 패턴으로 찾기 (lenient)\r\n for (const e of dirEntries) {\r\n if (e.type === 2 && e.name === streamName) {\r\n // 부모 확인은 생략 (lenient) — 중복 이름 시 첫 번째 반환\r\n return e\r\n }\r\n }\r\n\r\n // 정확한 이름이 아닌 경우 (ViewText/Section0 등)\r\n const lastPart = parts[parts.length - 1]\r\n return dirEntries.find(e => e.type === 2 && e.name === lastPart) ?? null\r\n }\r\n\r\n // ── 공개 API ──\r\n\r\n return {\r\n findStream(path: string): Buffer | null {\r\n // \\005 prefix 처리 (SummaryInformation)\r\n const normalized = path.replace(/^\\//, \"\")\r\n const entry = findEntryByPath(normalized)\r\n if (!entry || entry.type !== 2) return null\r\n const stream = readStreamData(entry)\r\n return stream.length > 0 ? stream : null\r\n },\r\n\r\n entries(): DirEntry[] {\r\n return dirEntries.filter(e => e.type === 2) // stream만\r\n },\r\n }\r\n}\r\n"],"mappings":";;;AAEA,OAAO,WAAW;;;ACYlB,IAAM,YAAY,OAAO,KAAK,CAAC,KAAM,KAAM,IAAM,KAAM,KAAM,KAAM,IAAM,GAAI,CAAC;AAC9E,IAAM,eAAe;AACrB,IAAM,YAAY;AAGlB,IAAM,mBAAmB;AAEzB,IAAM,kBAAkB;AAExB,IAAM,kBAAkB,MAAM,OAAO;AAsB9B,SAAS,gBAAgB,MAAmC;AACjE,MAAI,KAAK,SAAS,IAAK,OAAM,IAAI,MAAM,mGAA6B;AACpE,MAAI,CAAC,KAAK,SAAS,GAAG,CAAC,EAAE,OAAO,SAAS,EAAG,OAAM,IAAI,MAAM,wDAAgB;AAI5E,QAAM,kBAAkB,KAAK,aAAa,EAAE;AAC5C,MAAI,kBAAkB,KAAK,kBAAkB,GAAI,OAAM,IAAI,MAAM,yFAAwB,eAAe;AACxG,QAAM,aAAa,KAAK;AACxB,QAAM,sBAAsB,KAAK,aAAa,EAAE;AAChD,MAAI,sBAAsB,GAAI,OAAM,IAAI,MAAM,sGAA2B,mBAAmB;AAC5F,QAAM,iBAAiB,KAAK;AAE5B,QAAM,iBAAiB,KAAK,aAAa,EAAE;AAC3C,MAAI,iBAAiB,IAAO,OAAM,IAAI,MAAM,0EAAwB,cAAc;AAClF,QAAM,iBAAiB,KAAK,aAAa,EAAE;AAC3C,QAAM,mBAAmB,KAAK,aAAa,EAAE;AAC7C,QAAM,qBAAqB,KAAK,aAAa,EAAE;AAC/C,QAAM,qBAAqB,KAAK,aAAa,EAAE;AAC/C,QAAM,mBAAmB,KAAK,aAAa,EAAE;AAC7C,QAAM,mBAAmB,KAAK,aAAa,EAAE;AAI7C,WAAS,aAAa,IAAoB;AACxC,WAAO,MAAM,KAAK;AAAA,EACpB;AAEA,WAAS,eAAe,IAAoB;AAC1C,UAAM,MAAM,aAAa,EAAE;AAC3B,QAAI,MAAM,aAAa,KAAK,OAAQ,QAAO,OAAO,MAAM,CAAC;AACzD,WAAO,KAAK,SAAS,KAAK,MAAM,UAAU;AAAA,EAC5C;AAIA,QAAM,aAAuB,CAAC;AAG9B,WAAS,IAAI,GAAG,IAAI,OAAO,WAAW,SAAS,gBAAgB,KAAK;AAClE,UAAM,MAAM,KAAK,aAAa,KAAK,IAAI,CAAC;AACxC,QAAI,QAAQ,aAAa,QAAQ,aAAc;AAC/C,eAAW,KAAK,GAAG;AAAA,EACrB;AAGA,MAAI,cAAc;AAClB,QAAM,eAAe,oBAAI,IAAY;AACrC,WAAS,IAAI,GAAG,IAAI,oBAAoB,gBAAgB,gBAAgB,gBAAgB,WAAW,KAAK;AACtG,QAAI,aAAa,IAAI,WAAW,EAAG;AACnC,iBAAa,IAAI,WAAW;AAE5B,UAAM,MAAM,eAAe,WAAW;AACtC,UAAM,mBAAoB,aAAa,IAAK;AAC5C,aAAS,IAAI,GAAG,IAAI,oBAAoB,WAAW,SAAS,gBAAgB,KAAK;AAC/E,YAAM,MAAM,IAAI,aAAa,IAAI,CAAC;AAClC,UAAI,QAAQ,aAAa,QAAQ,aAAc;AAC/C,iBAAW,KAAK,GAAG;AAAA,IACrB;AACA,kBAAc,IAAI,aAAa,mBAAmB,CAAC;AAAA,EACrD;AAIA,QAAM,sBAAsB,aAAa;AACzC,QAAM,WAAW,IAAI,YAAY,WAAW,SAAS,mBAAmB;AAExE,WAAS,KAAK,GAAG,KAAK,WAAW,QAAQ,MAAM;AAC7C,UAAM,MAAM,eAAe,WAAW,EAAE,CAAC;AACzC,aAAS,IAAI,GAAG,IAAI,qBAAqB,KAAK;AAC5C,eAAS,KAAK,sBAAsB,CAAC,IAAI,IAAI,IAAI,IAAI,IAAI,SACrD,IAAI,aAAa,IAAI,CAAC,IACtB;AAAA,IACN;AAAA,EACF;AAIA,WAAS,UAAU,aAAqB,UAA0B;AAChE,QAAI,gBAAgB,gBAAgB,gBAAgB,UAAW,QAAO,OAAO,MAAM,CAAC;AACpF,QAAI,WAAW,gBAAiB,OAAM,IAAI,MAAM,0DAAa;AAE7D,UAAM,SAAmB,CAAC;AAC1B,QAAI,UAAU;AACd,QAAI,YAAY;AAChB,UAAM,UAAU,oBAAI,IAAY;AAEhC,WAAO,YAAY,gBAAgB,YAAY,aAAa,YAAY,UAAU;AAChF,UAAI,QAAQ,IAAI,OAAO,EAAG;AAC1B,UAAI,QAAQ,OAAO,iBAAkB;AACrC,cAAQ,IAAI,OAAO;AAEnB,YAAM,MAAM,eAAe,OAAO;AAClC,YAAM,YAAY,WAAW;AAC7B,aAAO,KAAK,YAAY,aAAa,IAAI,SAAS,GAAG,SAAS,IAAI,GAAG;AACrE,mBAAa,KAAK,IAAI,IAAI,QAAQ,SAAS;AAE3C,gBAAU,UAAU,SAAS,SAAS,SAAS,OAAO,IAAI;AAAA,IAC5D;AAEA,WAAO,OAAO,OAAO,MAAM;AAAA,EAC7B;AAIA,MAAI,eAAmC;AAEvC,WAAS,kBAA+B;AACtC,QAAI,aAAc,QAAO;AAEzB,QAAI,uBAAuB,KAAK,uBAAuB,cAAc;AACnE,qBAAe,IAAI,YAAY,CAAC;AAChC,aAAO;AAAA,IACT;AAEA,UAAM,cAAc,UAAU,oBAAoB,qBAAqB,UAAU;AACjF,UAAM,UAAU,YAAY,SAAS;AACrC,mBAAe,IAAI,YAAY,OAAO;AACtC,aAAS,IAAI,GAAG,IAAI,SAAS,KAAK;AAChC,mBAAa,CAAC,IAAI,YAAY,aAAa,IAAI,CAAC;AAAA,IAClD;AACA,WAAO;AAAA,EACT;AAIA,QAAM,UAAU,UAAU,gBAAgB,kBAAkB,GAAG;AAC/D,QAAM,aAAyB,CAAC;AAEhC,WAAS,SAAS,GAAG,SAAS,OAAO,QAAQ,UAAU,WAAW,SAAS,iBAAiB,UAAU,KAAK;AACzG,UAAM,UAAU,QAAQ,aAAa,SAAS,EAAE;AAChD,QAAI,WAAW,KAAK,UAAU,IAAI;AAChC,iBAAW,KAAK,EAAE,MAAM,IAAI,MAAM,GAAG,aAAa,GAAG,MAAM,EAAE,CAAC;AAC9D;AAAA,IACF;AAEA,UAAM,YAAY,UAAU;AAC5B,UAAM,OAAO,YAAY,IACrB,QAAQ,SAAS,QAAQ,SAAS,SAAS,EAAE,SAAS,SAAS,IAC/D;AAEJ,UAAM,OAAO,QAAQ,SAAS,EAAE;AAChC,UAAM,cAAc,QAAQ,aAAa,SAAS,GAAG;AAErD,UAAM,OAAO,QAAQ,aAAa,SAAS,GAAG;AAE9C,eAAW,KAAK,EAAE,MAAM,MAAM,aAAa,KAAK,CAAC;AAAA,EACnD;AAIA,MAAI,iBAAgC;AAEpC,WAAS,gBAAwB;AAC/B,QAAI,eAAgB,QAAO;AAC3B,UAAM,OAAO,WAAW,CAAC;AACzB,QAAI,CAAC,QAAQ,KAAK,SAAS,GAAG;AAC5B,uBAAiB,OAAO,MAAM,CAAC;AAC/B,aAAO;AAAA,IACT;AACA,qBAAiB,UAAU,KAAK,aAAa,KAAK,QAAQ,eAAe;AACzE,WAAO;AAAA,EACT;AAIA,WAAS,eAAe,aAAqB,MAAsB;AACjE,UAAM,MAAM,gBAAgB;AAC5B,UAAM,KAAK,cAAc;AACzB,QAAI,IAAI,WAAW,KAAK,GAAG,WAAW,EAAG,QAAO,OAAO,MAAM,CAAC;AAE9D,UAAM,SAAmB,CAAC;AAC1B,QAAI,UAAU;AACd,QAAI,YAAY;AAChB,UAAM,UAAU,oBAAI,IAAY;AAEhC,WAAO,YAAY,gBAAgB,YAAY,aAAa,YAAY,MAAM;AAC5E,UAAI,QAAQ,IAAI,OAAO,EAAG;AAC1B,UAAI,QAAQ,OAAO,iBAAkB;AACrC,cAAQ,IAAI,OAAO;AAEnB,YAAM,MAAM,UAAU;AACtB,YAAM,YAAY,OAAO;AACzB,YAAM,YAAY,KAAK,IAAI,gBAAgB,SAAS;AACpD,UAAI,MAAM,aAAa,GAAG,QAAQ;AAChC,eAAO,KAAK,GAAG,SAAS,KAAK,MAAM,SAAS,CAAC;AAAA,MAC/C;AACA,mBAAa;AAEb,gBAAU,UAAU,IAAI,SAAS,IAAI,OAAO,IAAI;AAAA,IAClD;AAEA,WAAO,OAAO,OAAO,MAAM;AAAA,EAC7B;AAIA,WAAS,eAAe,OAAyB;AAC/C,QAAI,MAAM,SAAS,EAAG,QAAO,OAAO,MAAM,CAAC;AAC3C,QAAI,MAAM,OAAO,kBAAkB;AACjC,YAAM,aAAa,eAAe,MAAM,aAAa,MAAM,IAAI;AAE/D,UAAI,WAAW,SAAS,EAAG,QAAO;AAAA,IACpC;AACA,WAAO,UAAU,MAAM,aAAa,MAAM,IAAI;AAAA,EAChD;AAMA,WAAS,gBAAgB,MAA+B;AAGtD,UAAM,QAAQ,KAAK,QAAQ,OAAO,EAAE,EAAE,MAAM,GAAG;AAE/C,QAAI,MAAM,WAAW,GAAG;AAEtB,aAAO,WAAW,KAAK,OAAK,EAAE,SAAS,MAAM,CAAC,KAAK,EAAE,SAAS,CAAC,KAAK;AAAA,IACtE;AAIA,UAAM,cAAc,MAAM,CAAC;AAC3B,UAAM,aAAa,MAAM,MAAM,CAAC,EAAE,KAAK,GAAG;AAG1C,eAAW,KAAK,YAAY;AAC1B,UAAI,EAAE,SAAS,KAAK,EAAE,SAAS,YAAY;AAEzC,eAAO;AAAA,MACT;AAAA,IACF;AAGA,UAAM,WAAW,MAAM,MAAM,SAAS,CAAC;AACvC,WAAO,WAAW,KAAK,OAAK,EAAE,SAAS,KAAK,EAAE,SAAS,QAAQ,KAAK;AAAA,EACtE;AAIA,SAAO;AAAA,IACL,WAAW,MAA6B;AAEtC,YAAM,aAAa,KAAK,QAAQ,OAAO,EAAE;AACzC,YAAM,QAAQ,gBAAgB,UAAU;AACxC,UAAI,CAAC,SAAS,MAAM,SAAS,EAAG,QAAO;AACvC,YAAM,SAAS,eAAe,KAAK;AACnC,aAAO,OAAO,SAAS,IAAI,SAAS;AAAA,IACtC;AAAA,IAEA,UAAsB;AACpB,aAAO,WAAW,OAAO,OAAK,EAAE,SAAS,CAAC;AAAA,IAC5C;AAAA,EACF;AACF;;;ADrSA,SAAS,WAAW,QAAiC;AACnD,SAAO,IAAI,WAAW,QAAQ,GAAG,KAAK,IAAI,GAAG,OAAO,UAAU,CAAC;AACjE;AAGO,SAAS,UAAU,QAA8B;AACtD,QAAM,IAAI,WAAW,MAAM;AAC3B,SAAO,EAAE,CAAC,MAAM,MAAQ,EAAE,CAAC,MAAM,MAAQ,EAAE,CAAC,MAAM,KAAQ,EAAE,CAAC,MAAM;AACrE;AAGO,SAAS,WAAW,QAA8B;AACvD,SAAO,UAAU,MAAM;AACzB;AAGO,SAAS,aAAa,QAA8B;AACzD,QAAM,IAAI,WAAW,MAAM;AAC3B,SAAO,EAAE,CAAC,MAAM,OAAQ,EAAE,CAAC,MAAM,OAAQ,EAAE,CAAC,MAAM,MAAQ,EAAE,CAAC,MAAM;AACrE;AAMA,IAAM,cAAc,IAAI,YAAY,EAAE,OAAO,yBAAyB;AAC/D,SAAS,WAAW,QAA8B;AACvD,MAAI,OAAO,aAAa,YAAY,OAAQ,QAAO;AACnD,QAAM,OAAO,IAAI,WAAW,QAAQ,GAAG,YAAY,MAAM;AACzD,WAAS,IAAI,GAAG,IAAI,YAAY,QAAQ,KAAK;AAC3C,QAAI,KAAK,CAAC,MAAM,YAAY,CAAC,EAAG,QAAO;AAAA,EACzC;AACA,SAAO;AACT;AAGO,SAAS,UAAU,QAA8B;AACtD,QAAM,IAAI,WAAW,MAAM;AAC3B,SAAO,EAAE,CAAC,MAAM,MAAQ,EAAE,CAAC,MAAM,MAAQ,EAAE,CAAC,MAAM,MAAQ,EAAE,CAAC,MAAM;AACrE;AAGO,SAAS,YAAY,QAA8B;AACxD,QAAM,QAAQ,IAAI,WAAW,QAAQ,GAAG,KAAK,IAAI,KAAK,OAAO,UAAU,CAAC;AACxE,QAAM,OAAO,IAAI,YAAY,SAAS,EAAE,OAAO,MAAM,CAAC,EAAE,OAAO,KAAK,EAAE,QAAQ,WAAW,EAAE;AAC3F,SAAO,KAAK,UAAU,EAAE,WAAW,OAAO,KAAK,KAAK,SAAS,QAAQ;AACvE;AAGO,SAAS,aAAa,QAA+B;AAC1D,MAAI,OAAO,aAAa,EAAG,QAAO;AAClC,MAAI,WAAW,MAAM,EAAG,QAAO;AAC/B,MAAI,UAAU,MAAM,EAAG,QAAO;AAC9B,MAAI,aAAa,MAAM,EAAG,QAAO;AACjC,MAAI,UAAU,MAAM,EAAG,QAAO;AAC9B,MAAI,YAAY,MAAM,EAAG,QAAO;AAChC,SAAO;AACT;AAQO,SAAS,iBAAiB,QAAgD;AAC/E,MAAI;AACF,UAAM,MAAM,gBAAgB,OAAO,KAAK,MAAM,CAAC;AAC/C,UAAM,QAAQ,IAAI,QAAQ,EAAE,IAAI,OAAK,EAAE,IAAI;AAC3C,QAAI,MAAM,SAAS,UAAU,KAAK,MAAM,SAAS,MAAM,EAAG,QAAO;AACjE,QAAI,MAAM,SAAS,YAAY,EAAG,QAAO;AAEzC,QAAI,MAAM,KAAK,OAAK,MAAM,aAAa,EAAE,WAAW,SAAS,CAAC,EAAG,QAAO;AACxE,WAAO;AAAA,EACT,QAAQ;AACN,WAAO;AAAA,EACT;AACF;AAMA,eAAsB,gBAAgB,QAAoE;AACxG,MAAI;AACF,UAAM,MAAM,MAAM,MAAM,UAAU,MAAM;AAExC,QAAI,IAAI,KAAK,iBAAiB,EAAG,QAAO;AAExC,QAAI,IAAI,KAAK,mBAAmB,EAAG,QAAO;AAE1C,QAAI,IAAI,KAAK,sBAAsB,KAAK,IAAI,KAAK,UAAU,EAAG,QAAO;AAErE,UAAM,aAAa,OAAO,KAAK,IAAI,KAAK,EAAE,KAAK,OAAK,EAAE,WAAW,WAAW,CAAC;AAC7E,QAAI,WAAY,QAAO;AACvB,WAAO;AAAA,EACT,QAAQ;AACN,WAAO;AAAA,EACT;AACF;","names":[]}