@clazic/kordoc 2.6.1 → 2.7.1

package/dist/index.d.cts

@@ -549,12 +549,14 @@ interface UnifiedOcrProgressEvent {
     pageNumber?: number;
     workerCount?: number;
 }
+/** unified-ocr에서 onEvent로 전달되는 보조 이벤트 — LibreOffice 설치/검증/에러 */
+type UnifiedOcrAuxEvent = ConvertInstallEvent | ConvertValidateEvent | ConvertErrorEvent;
 interface UnifiedOcrOptions {
     workspaceDir?: string;
     outputPath?: string;
     dpi?: number;
     baseUrl?: string;
-    onEvent?: (event: UnifiedOcrProgressEvent) => void;
+    onEvent?: (event: UnifiedOcrProgressEvent | UnifiedOcrAuxEvent) => void;
     modelCandidates?: string[];
     modelMaxTokens?: Record<string, number>;
     stageWeights?: Partial<Record<UnifiedStage, number>>;
@@ -564,6 +566,10 @@ interface UnifiedOcrOptions {
     logger?: Logger;
     runId?: string;
     concurrencyPerKey?: number;
+    /** LibreOffice 자동 설치 허용 (기본 false) — convert 단계에서 install/validate 이벤트가 onEvent로 전달됨 */
+    autoInstallLibreOffice?: boolean;
+    /** LibreOffice 바이너리 직접 지정 (선택) — 지정 시 자동 탐색/설치 건너뜀 */
+    sofficePath?: string;
 }
 interface UnifiedOcrResult {
     outputPath: string;
@@ -602,4 +608,4 @@ declare function parseXlsx(buffer: ArrayBuffer, options?: ParseOptions, zip?: JS
 /** DOCX 파일을 Markdown으로 변환 */
 declare function parseDocx(buffer: ArrayBuffer, options?: ParseOptions, zip?: JSZip): Promise<ParseResult>;
 
-export { type AcquireResult, AllKeysCoolingDownError, type ApiKeyRotationOptions, ApiKeyRotationPool, type BlockDiff, type BoundingBox, type CellContext, type CellDiff, type ConvertToPdfFailure, type ConvertToPdfOptions, type ConvertToPdfResult, type ConvertToPdfSuccess, type DiffChangeType, type DiffResult, type DocumentMetadata, type ErrorCode, type ExtractedImage, type FileType, type FormField, type FormResult, type IRBlock, type IRBlockType, type IRCell, type IRTable, type ImageData, type InlineStyle, type KeyHealthSnapshot, type MarkdownToXlsxOptions, type OcrMode, type OcrProvider, type OutlineItem, type ParseFailure, type ParseOptions, type ParseResult, type ParseSuccess, type ParseWarning, type StructuredOcrResult, type UnifiedOcrOptions, type UnifiedOcrProgressEvent, type UnifiedOcrResult, type UnifiedStage, VERSION, type WarningCode, type WatchOptions, blocksToMarkdown, compare, convertHwpToPdf, convertHwpxToPdf, convertToPdf, detectFormat, detectZipFormat, diffBlocks, extractFormFields, isHwpxFile, isOldHwpFile, isPdfFile, isZipFile, markdownToHwpx, markdownToXlsx, parse, parseDocx, parseHwp, parseHwpx, parsePdf, parseXlsx, runUnifiedOcrPipeline };
+export { type AcquireResult, AllKeysCoolingDownError, type ApiKeyRotationOptions, ApiKeyRotationPool, type BlockDiff, type BoundingBox, type CellContext, type CellDiff, type ConvertErrorEvent, type ConvertInstallEvent, type ConvertToPdfFailure, type ConvertToPdfOptions, type ConvertToPdfResult, type ConvertToPdfSuccess, type ConvertValidateEvent, type DiffChangeType, type DiffResult, type DocumentMetadata, type ErrorCode, type ExtractedImage, type FileType, type FormField, type FormResult, type IRBlock, type IRBlockType, type IRCell, type IRTable, type ImageData, type InlineStyle, type KeyHealthSnapshot, type MarkdownToXlsxOptions, type OcrMode, type OcrProvider, type OutlineItem, type ParseFailure, type ParseOptions, type ParseResult, type ParseSuccess, type ParseWarning, type StructuredOcrResult, type UnifiedOcrAuxEvent, type UnifiedOcrOptions, type UnifiedOcrProgressEvent, type UnifiedOcrResult, type UnifiedStage, VERSION, type WarningCode, type WatchOptions, blocksToMarkdown, compare, convertHwpToPdf, convertHwpxToPdf, convertToPdf, detectFormat, detectZipFormat, diffBlocks, extractFormFields, isHwpxFile, isOldHwpFile, isPdfFile, isZipFile, markdownToHwpx, markdownToXlsx, parse, parseDocx, parseHwp, parseHwpx, parsePdf, parseXlsx, runUnifiedOcrPipeline };

package/dist/index.d.ts

@@ -549,12 +549,14 @@ interface UnifiedOcrProgressEvent {
     pageNumber?: number;
     workerCount?: number;
 }
+/** unified-ocr에서 onEvent로 전달되는 보조 이벤트 — LibreOffice 설치/검증/에러 */
+type UnifiedOcrAuxEvent = ConvertInstallEvent | ConvertValidateEvent | ConvertErrorEvent;
 interface UnifiedOcrOptions {
     workspaceDir?: string;
     outputPath?: string;
     dpi?: number;
     baseUrl?: string;
-    onEvent?: (event: UnifiedOcrProgressEvent) => void;
+    onEvent?: (event: UnifiedOcrProgressEvent | UnifiedOcrAuxEvent) => void;
     modelCandidates?: string[];
     modelMaxTokens?: Record<string, number>;
     stageWeights?: Partial<Record<UnifiedStage, number>>;
@@ -564,6 +566,10 @@ interface UnifiedOcrOptions {
     logger?: Logger;
     runId?: string;
     concurrencyPerKey?: number;
+    /** LibreOffice 자동 설치 허용 (기본 false) — convert 단계에서 install/validate 이벤트가 onEvent로 전달됨 */
+    autoInstallLibreOffice?: boolean;
+    /** LibreOffice 바이너리 직접 지정 (선택) — 지정 시 자동 탐색/설치 건너뜀 */
+    sofficePath?: string;
 }
 interface UnifiedOcrResult {
     outputPath: string;
@@ -602,4 +608,4 @@ declare function parseXlsx(buffer: ArrayBuffer, options?: ParseOptions, zip?: JS
 /** DOCX 파일을 Markdown으로 변환 */
 declare function parseDocx(buffer: ArrayBuffer, options?: ParseOptions, zip?: JSZip): Promise<ParseResult>;
 
-export { type AcquireResult, AllKeysCoolingDownError, type ApiKeyRotationOptions, ApiKeyRotationPool, type BlockDiff, type BoundingBox, type CellContext, type CellDiff, type ConvertToPdfFailure, type ConvertToPdfOptions, type ConvertToPdfResult, type ConvertToPdfSuccess, type DiffChangeType, type DiffResult, type DocumentMetadata, type ErrorCode, type ExtractedImage, type FileType, type FormField, type FormResult, type IRBlock, type IRBlockType, type IRCell, type IRTable, type ImageData, type InlineStyle, type KeyHealthSnapshot, type MarkdownToXlsxOptions, type OcrMode, type OcrProvider, type OutlineItem, type ParseFailure, type ParseOptions, type ParseResult, type ParseSuccess, type ParseWarning, type StructuredOcrResult, type UnifiedOcrOptions, type UnifiedOcrProgressEvent, type UnifiedOcrResult, type UnifiedStage, VERSION, type WarningCode, type WatchOptions, blocksToMarkdown, compare, convertHwpToPdf, convertHwpxToPdf, convertToPdf, detectFormat, detectZipFormat, diffBlocks, extractFormFields, isHwpxFile, isOldHwpFile, isPdfFile, isZipFile, markdownToHwpx, markdownToXlsx, parse, parseDocx, parseHwp, parseHwpx, parsePdf, parseXlsx, runUnifiedOcrPipeline };
+export { type AcquireResult, AllKeysCoolingDownError, type ApiKeyRotationOptions, ApiKeyRotationPool, type BlockDiff, type BoundingBox, type CellContext, type CellDiff, type ConvertErrorEvent, type ConvertInstallEvent, type ConvertToPdfFailure, type ConvertToPdfOptions, type ConvertToPdfResult, type ConvertToPdfSuccess, type ConvertValidateEvent, type DiffChangeType, type DiffResult, type DocumentMetadata, type ErrorCode, type ExtractedImage, type FileType, type FormField, type FormResult, type IRBlock, type IRBlockType, type IRCell, type IRTable, type ImageData, type InlineStyle, type KeyHealthSnapshot, type MarkdownToXlsxOptions, type OcrMode, type OcrProvider, type OutlineItem, type ParseFailure, type ParseOptions, type ParseResult, type ParseSuccess, type ParseWarning, type StructuredOcrResult, type UnifiedOcrAuxEvent, type UnifiedOcrOptions, type UnifiedOcrProgressEvent, type UnifiedOcrResult, type UnifiedStage, VERSION, type WarningCode, type WatchOptions, blocksToMarkdown, compare, convertHwpToPdf, convertHwpxToPdf, convertToPdf, detectFormat, detectZipFormat, diffBlocks, extractFormFields, isHwpxFile, isOldHwpFile, isPdfFile, isZipFile, markdownToHwpx, markdownToXlsx, parse, parseDocx, parseHwp, parseHwpx, parsePdf, parseXlsx, runUnifiedOcrPipeline };

package/dist/index.js

@@ -37,118 +37,6 @@ var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__ge
   mod
 ));
 
-// src/utils.ts
-var utils_exports = {};
-__export(utils_exports, {
-  KordocError: () => KordocError,
-  VERSION: () => VERSION,
-  classifyError: () => classifyError,
-  isPathTraversal: () => isPathTraversal,
-  normalizeKordocError: () => normalizeKordocError,
-  precheckZipSize: () => precheckZipSize,
-  sanitizeError: () => sanitizeError,
-  sanitizeHref: () => sanitizeHref,
-  toArrayBuffer: () => toArrayBuffer
-});
-function toArrayBuffer(buf) {
-  if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {
-    return buf.buffer;
-  }
-  return buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength);
-}
-function sanitizeError(err) {
-  if (err instanceof KordocError) return err.message;
-  return "\uBB38\uC11C \uCC98\uB9AC \uC911 \uC624\uB958\uAC00 \uBC1C\uC0DD\uD588\uC2B5\uB2C8\uB2E4";
-}
-function isPathTraversal(name) {
-  if (name.includes("\0")) return true;
-  const normalized = name.replace(/\\/g, "/");
-  return normalized.includes("..") || normalized.startsWith("/") || /^[A-Za-z]:/.test(normalized);
-}
-function precheckZipSize(buffer, maxUncompressedSize = 100 * 1024 * 1024, maxEntries = 500) {
-  try {
-    const data = new DataView(buffer);
-    const len = buffer.byteLength;
-    let eocdOffset = -1;
-    for (let i = len - 22; i >= Math.max(0, len - 65557); i--) {
-      if (data.getUint32(i, true) === 101010256) {
-        eocdOffset = i;
-        break;
-      }
-    }
-    if (eocdOffset < 0) return { totalUncompressed: 0, entryCount: 0 };
-    const entryCount = data.getUint16(eocdOffset + 10, true);
-    if (entryCount > maxEntries) {
-      throw new KordocError(`ZIP \uC5D4\uD2B8\uB9AC \uC218 \uCD08\uACFC: ${entryCount} (\uCD5C\uB300 ${maxEntries})`);
-    }
-    const cdSize = data.getUint32(eocdOffset + 12, true);
-    const cdOffset = data.getUint32(eocdOffset + 16, true);
-    if (cdOffset + cdSize > len) return { totalUncompressed: 0, entryCount };
-    let totalUncompressed = 0;
-    let pos = cdOffset;
-    for (let i = 0; i < entryCount && pos + 46 <= cdOffset + cdSize; i++) {
-      if (data.getUint32(pos, true) !== 33639248) break;
-      totalUncompressed += data.getUint32(pos + 24, true);
-      const nameLen = data.getUint16(pos + 28, true);
-      const extraLen = data.getUint16(pos + 30, true);
-      const commentLen = data.getUint16(pos + 32, true);
-      pos += 46 + nameLen + extraLen + commentLen;
-    }
-    if (totalUncompressed > maxUncompressedSize) {
-      throw new KordocError(`ZIP \uBE44\uC555\uCD95 \uD06C\uAE30 \uCD08\uACFC: ${(totalUncompressed / 1024 / 1024).toFixed(1)}MB (\uCD5C\uB300 ${maxUncompressedSize / 1024 / 1024}MB)`);
-    }
-    return { totalUncompressed, entryCount };
-  } catch (err) {
-    if (err instanceof KordocError) throw err;
-    return { totalUncompressed: 0, entryCount: 0 };
-  }
-}
-function sanitizeHref(href) {
-  const trimmed = href.trim();
-  if (!trimmed || !SAFE_HREF_RE.test(trimmed)) return null;
-  return trimmed;
-}
-function classifyError(err) {
-  if (!(err instanceof Error)) return "PARSE_ERROR";
-  const msg = err.message;
-  if (msg.includes("\uC554\uD638\uD654")) return "ENCRYPTED";
-  if (msg.includes("DRM")) return "DRM_PROTECTED";
-  if (msg.includes("ZIP bomb") || msg.includes("ZIP \uBE44\uC555\uCD95 \uD06C\uAE30 \uCD08\uACFC") || msg.includes("ZIP \uC5D4\uD2B8\uB9AC \uC218 \uCD08\uACFC")) return "ZIP_BOMB";
-  if (msg.includes("bomb") || msg.includes("\uD06C\uAE30 \uCD08\uACFC") || msg.includes("\uC555\uCD95 \uD574\uC81C")) return "DECOMPRESSION_BOMB";
-  if (msg.includes("\uC774\uBBF8\uC9C0 \uAE30\uBC18")) return "IMAGE_BASED_PDF";
-  if (msg.includes("\uC139\uC158") && (msg.includes("\uCC3E\uC744 \uC218 \uC5C6") || msg.includes("\uC5C6\uC74C"))) return "NO_SECTIONS";
-  if (msg.includes("\uC2DC\uADF8\uB2C8\uCC98") || msg.includes("\uBCF5\uAD6C\uD560 \uC218 \uC5C6")) return "CORRUPTED";
-  return "PARSE_ERROR";
-}
-function normalizeKordocError(err, fallbackMessage, stage = "unknown", fallbackCode = "PARSE_ERROR") {
-  if (err instanceof KordocError) {
-    if (!err.stage) err.stage = stage;
-    if (!err.code) err.code = fallbackCode;
-    return err;
-  }
-  const message = err instanceof Error ? err.message : fallbackMessage;
-  const code = err instanceof Error ? classifyError(err) : fallbackCode;
-  return new KordocError(message || fallbackMessage, { code, stage });
-}
-var VERSION, KordocError, SAFE_HREF_RE;
-var init_utils = __esm({
-  "src/utils.ts"() {
-    "use strict";
-    VERSION = true ? "2.6.0" : "0.0.0-dev";
-    KordocError = class extends Error {
-      code;
-      stage;
-      constructor(message, opts = {}) {
-        super(message);
-        this.name = "KordocError";
-        this.code = opts.code;
-        this.stage = opts.stage;
-      }
-    };
-    SAFE_HREF_RE = /^(?:https?:|mailto:|tel:|#)/i;
-  }
-});
-
 // src/page-range.ts
 var page_range_exports = {};
 __export(page_range_exports, {
@@ -3223,8 +3111,97 @@ async function detectZipFormat(buffer) {
 import JSZip2 from "jszip";
 import { DOMParser } from "@xmldom/xmldom";
 
+// src/utils.ts
+var VERSION = true ? "2.7.1" : "0.0.0-dev";
+function toArrayBuffer(buf) {
+  if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {
+    return buf.buffer;
+  }
+  return buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength);
+}
+var KordocError = class extends Error {
+  code;
+  stage;
+  constructor(message, opts = {}) {
+    super(message);
+    this.name = "KordocError";
+    this.code = opts.code;
+    this.stage = opts.stage;
+  }
+};
+function isPathTraversal(name) {
+  if (name.includes("\0")) return true;
+  const normalized = name.replace(/\\/g, "/");
+  return normalized.includes("..") || normalized.startsWith("/") || /^[A-Za-z]:/.test(normalized);
+}
+function precheckZipSize(buffer, maxUncompressedSize = 100 * 1024 * 1024, maxEntries = 500) {
+  try {
+    const data = new DataView(buffer);
+    const len = buffer.byteLength;
+    let eocdOffset = -1;
+    for (let i = len - 22; i >= Math.max(0, len - 65557); i--) {
+      if (data.getUint32(i, true) === 101010256) {
+        eocdOffset = i;
+        break;
+      }
+    }
+    if (eocdOffset < 0) return { totalUncompressed: 0, entryCount: 0 };
+    const entryCount = data.getUint16(eocdOffset + 10, true);
+    if (entryCount > maxEntries) {
+      throw new KordocError(`ZIP \uC5D4\uD2B8\uB9AC \uC218 \uCD08\uACFC: ${entryCount} (\uCD5C\uB300 ${maxEntries})`);
+    }
+    const cdSize = data.getUint32(eocdOffset + 12, true);
+    const cdOffset = data.getUint32(eocdOffset + 16, true);
+    if (cdOffset + cdSize > len) return { totalUncompressed: 0, entryCount };
+    let totalUncompressed = 0;
+    let pos = cdOffset;
+    for (let i = 0; i < entryCount && pos + 46 <= cdOffset + cdSize; i++) {
+      if (data.getUint32(pos, true) !== 33639248) break;
+      totalUncompressed += data.getUint32(pos + 24, true);
+      const nameLen = data.getUint16(pos + 28, true);
+      const extraLen = data.getUint16(pos + 30, true);
+      const commentLen = data.getUint16(pos + 32, true);
+      pos += 46 + nameLen + extraLen + commentLen;
+    }
+    if (totalUncompressed > maxUncompressedSize) {
+      throw new KordocError(`ZIP \uBE44\uC555\uCD95 \uD06C\uAE30 \uCD08\uACFC: ${(totalUncompressed / 1024 / 1024).toFixed(1)}MB (\uCD5C\uB300 ${maxUncompressedSize / 1024 / 1024}MB)`);
+    }
+    return { totalUncompressed, entryCount };
+  } catch (err) {
+    if (err instanceof KordocError) throw err;
+    return { totalUncompressed: 0, entryCount: 0 };
+  }
+}
+var SAFE_HREF_RE = /^(?:https?:|mailto:|tel:|#)/i;
+function sanitizeHref(href) {
+  const trimmed = href.trim();
+  if (!trimmed || !SAFE_HREF_RE.test(trimmed)) return null;
+  return trimmed;
+}
+function classifyError(err) {
+  if (!(err instanceof Error)) return "PARSE_ERROR";
+  const msg = err.message;
+  if (msg.includes("\uC554\uD638\uD654")) return "ENCRYPTED";
+  if (msg.includes("DRM")) return "DRM_PROTECTED";
+  if (msg.includes("ZIP bomb") || msg.includes("ZIP \uBE44\uC555\uCD95 \uD06C\uAE30 \uCD08\uACFC") || msg.includes("ZIP \uC5D4\uD2B8\uB9AC \uC218 \uCD08\uACFC")) return "ZIP_BOMB";
+  if (msg.includes("bomb") || msg.includes("\uD06C\uAE30 \uCD08\uACFC") || msg.includes("\uC555\uCD95 \uD574\uC81C")) return "DECOMPRESSION_BOMB";
+  if (msg.includes("\uC774\uBBF8\uC9C0 \uAE30\uBC18")) return "IMAGE_BASED_PDF";
+  if (msg.includes("\uC139\uC158") && (msg.includes("\uCC3E\uC744 \uC218 \uC5C6") || msg.includes("\uC5C6\uC74C"))) return "NO_SECTIONS";
+  if (msg.includes("\uC2DC\uADF8\uB2C8\uCC98") || msg.includes("\uBCF5\uAD6C\uD560 \uC218 \uC5C6")) return "CORRUPTED";
+  return "PARSE_ERROR";
+}
+function normalizeKordocError(err, fallbackMessage, stage = "unknown", fallbackCode = "PARSE_ERROR") {
+  if (err instanceof KordocError) {
+    if (!err.stage) err.stage = stage;
+    if (!err.code) err.code = fallbackCode;
+    return err;
+  }
+  const message = err instanceof Error ? err.message : fallbackMessage;
+  const code = err instanceof Error ? classifyError(err) : fallbackCode;
+  return new KordocError(message || fallbackMessage, { code, stage });
+}
+
 // src/table/builder.ts
-init_utils();
 var MAX_COLS = 200;
 var MAX_ROWS = 1e4;
 function buildTable(rows) {
@@ -3484,8 +3461,6 @@ var HEADING_RATIO_H2 = 1.3;
 var HEADING_RATIO_H3 = 1.15;
 
 // src/hwpx/parser.ts
-init_utils();
-init_utils();
 init_page_range();
 init_logger();
 var MAX_DECOMPRESS_SIZE = 500 * 1024 * 1024;
@@ -4327,7 +4302,6 @@ function extractTextFromNode(node) {
 }
 
 // src/hwp5/record.ts
-init_utils();
 import { inflateRawSync, inflateSync } from "zlib";
 var TAG_PARA_HEADER = 66;
 var TAG_PARA_TEXT = 67;
@@ -5378,7 +5352,6 @@ function parseLenientCfb(data) {
 }
 
 // src/hwp5/parser.ts
-init_utils();
 init_page_range();
 init_logger();
 var CFB = __toESM(require_cfb(), 1);
@@ -6034,7 +6007,6 @@ function arrangeCells(rows, cols, cells) {
 }
 
 // src/pdf/parser.ts
-init_utils();
 init_page_range();
 import { createRequire } from "module";
 import { dirname as dirname2, join as join3, resolve as resolve2 } from "path";
@@ -7926,7 +7898,6 @@ function mergeKoreanLines(text) {
 }
 
 // src/xlsx/parser.ts
-init_utils();
 import JSZip3 from "jszip";
 import { DOMParser as DOMParser2 } from "@xmldom/xmldom";
 init_logger();
@@ -8255,7 +8226,6 @@ async function parseXlsxDocument(buffer, options, existingZip) {
 }
 
 // src/docx/parser.ts
-init_utils();
 import JSZip4 from "jszip";
 import { DOMParser as DOMParser3 } from "@xmldom/xmldom";
 init_logger();
@@ -8737,7 +8707,6 @@ async function parseDocxDocument(buffer, options, existingZip) {
 }
 
 // src/index.ts
-init_utils();
 init_cli_provider();
 init_markdown_to_blocks();
 init_logger();
@@ -11241,7 +11210,6 @@ async function markdownToXlsx(markdown, options) {
 
 // src/convert/index.ts
 import { readFile } from "fs/promises";
-init_utils();
 
 // src/convert/libreoffice.ts
 import libre from "libreoffice-convert";
@@ -11261,6 +11229,7 @@ import { join as join4, delimiter } from "path";
 import { mkdir, access, symlink, rm } from "fs/promises";
 import { createWriteStream } from "fs";
 import { spawn as spawn2 } from "child_process";
+var installInFlight = null;
 var CACHE_DIR = join4(homedir(), ".cache", "kordoc", "libreoffice");
 var VERSION_FILE = join4(CACHE_DIR, "version");
 var PACKAGES = {
@@ -11281,13 +11250,11 @@ var PACKAGES = {
   }
 };
 async function findInPath() {
-  try {
-    const { runCommand: runCommand2 } = await Promise.resolve().then(() => (init_utils(), utils_exports));
-    await runCommand2("soffice", ["--version"]);
-    return "soffice";
-  } catch {
-    return null;
-  }
+  return new Promise((resolve4) => {
+    const child = spawn2("soffice", ["--version"], { stdio: "ignore" });
+    child.on("close", (code) => resolve4(code === 0 ? "soffice" : null));
+    child.on("error", () => resolve4(null));
+  });
 }
 async function findInCache() {
   const cachedBin = join4(CACHE_DIR, "bin", "soffice");
@@ -11298,6 +11265,38 @@ async function findInCache() {
     return null;
   }
 }
+async function findInDefaultPaths() {
+  const platform = process.platform;
+  const paths = [];
+  if (platform === "darwin") {
+    paths.push(
+      "/Applications/LibreOffice.app/Contents/MacOS/soffice",
+      "/opt/homebrew/bin/soffice",
+      "/usr/local/bin/soffice"
+    );
+  } else if (platform === "linux") {
+    paths.push(
+      "/usr/bin/soffice",
+      "/usr/lib/libreoffice/program/soffice"
+    );
+  } else if (platform === "win32") {
+    const pf = process.env["ProgramFiles"] ?? "C:\\Program Files";
+    const pf86 = process.env["ProgramFiles(x86)"] ?? "C:\\Program Files (x86)";
+    paths.push(
+      join4(pf, "LibreOffice", "program", "soffice.exe"),
+      join4(pf86, "LibreOffice", "program", "soffice.exe")
+    );
+  }
+  for (const p of paths) {
+    try {
+      await access(p);
+      return p;
+    } catch {
+      continue;
+    }
+  }
+  return null;
+}
 async function downloadWithProgress(url, dest, totalBytes, onProgress) {
   const response = await fetch(url);
   if (!response.body) throw new Error("\uB2E4\uC6B4\uB85C\uB4DC \uC2E4\uD328: response body \uC5C6\uC74C");
@@ -11308,13 +11307,17 @@ async function downloadWithProgress(url, dest, totalBytes, onProgress) {
     while (true) {
       const { done, value } = await reader.read();
       if (done) break;
-      file.write(value);
+      if (!file.write(value)) {
+        await new Promise((resolve4) => file.once("drain", resolve4));
+      }
       downloaded += value.length;
       onProgress?.(downloaded, totalBytes);
     }
   } finally {
-    file.end();
     reader.releaseLock();
+    await new Promise((resolve4, reject) => {
+      file.end((err) => err ? reject(err) : resolve4());
+    });
   }
 }
 async function installForPlatform(pkg, onProgress) {
@@ -11425,6 +11428,11 @@ async function resolveSoffice(emitter, autoInstall = true) {
     emitter.validate("soffice_found", "\uCE90\uC2DC\uB41C LibreOffice \uBC1C\uACAC", { sofficePath: inCache });
     return inCache;
   }
+  const inDefault = await findInDefaultPaths();
+  if (inDefault) {
+    emitter.validate("soffice_found", "\uAE30\uBCF8 \uACBD\uB85C\uC5D0\uC11C LibreOffice \uBC1C\uACAC", { sofficePath: inDefault });
+    return inDefault;
+  }
   if (!autoInstall) {
     emitter.error(
       "validate",
@@ -11434,38 +11442,35 @@ async function resolveSoffice(emitter, autoInstall = true) {
     );
     throw new ConvertError("SOFFICE_NOT_FOUND", "LibreOffice\uAC00 \uC124\uCE58\uB418\uC9C0 \uC54A\uC558\uC2B5\uB2C8\uB2E4");
   }
+  if (installInFlight) {
+    return installInFlight;
+  }
   emitter.install("install_start", "LibreOffice \uC790\uB3D9 \uC124\uCE58\uB97C \uC2DC\uC791\uD569\uB2C8\uB2E4...");
-  try {
-    const installed = await installLibreOffice((downloaded, total) => {
-      const percent = Math.round(downloaded / total * 100);
-      emitter.install("download_progress", `\uB2E4\uC6B4\uB85C\uB4DC \uC911... ${percent}%`, {
-        percent,
-        downloadedBytes: downloaded,
-        totalBytes: total
+  installInFlight = (async () => {
+    try {
+      const installed = await installLibreOffice((downloaded, total) => {
+        const percent = Math.round(downloaded / total * 100);
+        emitter.install("download_progress", `\uB2E4\uC6B4\uB85C\uB4DC \uC911... ${percent}%`, {
+          percent,
+          downloadedBytes: downloaded,
+          totalBytes: total
+        });
       });
-    });
-    emitter.install("install_complete", "\uC124\uCE58 \uC644\uB8CC", { installedPath: installed });
-    return installed;
-  } catch (err) {
-    const errorMsg = err instanceof Error ? err.message : String(err);
-    emitter.install("install_failed", "\uC124\uCE58 \uC2E4\uD328", { error: errorMsg });
-    throw err;
-  }
+      emitter.install("install_complete", "\uC124\uCE58 \uC644\uB8CC", { installedPath: installed });
+      return installed;
+    } catch (err) {
+      const errorMsg = err instanceof Error ? err.message : String(err);
+      emitter.install("install_failed", "\uC124\uCE58 \uC2E4\uD328", { error: errorMsg });
+      throw err;
+    } finally {
+      installInFlight = null;
+    }
+  })();
+  return installInFlight;
 }
 
 // src/convert/libreoffice.ts
 var libreConvert = libre.convert;
-async function assertSofficeAvailable() {
-  const { runCommand: runCommand2 } = await Promise.resolve().then(() => (init_utils(), utils_exports));
-  try {
-    await runCommand2("soffice", ["--version"]);
-  } catch {
-    throw new ConvertError(
-      "SOFFICE_NOT_FOUND",
-      "soffice\uB97C \uCC3E\uC744 \uC218 \uC5C6\uC2B5\uB2C8\uB2E4. LibreOffice\uB97C \uC124\uCE58\uD574 \uC8FC\uC138\uC694."
-    );
-  }
-}
 async function convertBuffer(buffer, targetExt, timeoutMs = 6e4) {
   return new Promise((resolve4, reject) => {
     const timer = setTimeout(() => {
@@ -11711,9 +11716,6 @@ async function convertHwpxToPdf(input, options) {
   return result;
 }
 
-// src/index.ts
-init_utils();
-
 // src/ocr/api-key-rotation.ts
 var AllKeysCoolingDownError = class extends Error {
   waitMs;
@@ -11809,7 +11811,7 @@ var ApiKeyRotationPool = class _ApiKeyRotationPool {
 
 // src/pipeline/unified-ocr.ts
 import { mkdir as mkdir2, readdir, readFile as readFile2, stat, writeFile as writeFile2 } from "fs/promises";
-import { basename as basename2, dirname as dirname3, extname, join as join5, resolve as resolve3 } from "path";
+import { basename as basename2, delimiter as delimiter2, dirname as dirname3, extname, join as join5, resolve as resolve3 } from "path";
 import { spawn as spawn3 } from "child_process";
 import { performance } from "perf_hooks";
 init_logger();
@@ -11985,7 +11987,25 @@ async function runUnifiedOcrPipeline(inputPath, options = {}) {
     markStageStart("convert", "\uBB38\uC11C\uB97C PDF\uB85C \uBCC0\uD658 \uC911");
     logStage("info", "convert", "start", "\uBB38\uC11C\uB97C PDF\uB85C \uBCC0\uD658 \uC2DC\uC791", { input: absInput });
     if (extname(absInput).toLowerCase() !== ".pdf") {
-      await assertSofficeAvailable();
+      const convertEmitter = new ConvertEventEmitter();
+      if (options.onEvent) {
+        convertEmitter.setListener((evt) => {
+          if (evt.type === "install" || evt.type === "validate" || evt.type === "error") {
+            try {
+              ;
+              options.onEvent(evt);
+            } catch {
+            }
+          }
+        });
+      }
+      if (options.sofficePath) {
+        const sofficeDir = dirname3(options.sofficePath);
+        process.env.PATH = `${sofficeDir}${delimiter2}${process.env.PATH ?? ""}`;
+        convertEmitter.validate("soffice_found", "\uC9C1\uC811 \uC9C0\uC815\uB41C LibreOffice \uACBD\uB85C \uC0AC\uC6A9", { sofficePath: options.sofficePath });
+      } else {
+        await resolveSoffice(convertEmitter, options.autoInstallLibreOffice ?? false);
+      }
       workingPdfPath = join5(workspaceDir, `${stem}.pdf`);
       const inputBuffer = await readFile2(absInput);
       const out = await convertBuffer(inputBuffer, ".pdf");
@@ -12558,6 +12578,16 @@ function ensureSupportedInput(path) {
 }
 function normalizePipelineError(err, stage) {
   if (err instanceof UnifiedOcrError) return err;
+  if (err instanceof ConvertError) {
+    const codeMap = {
+      SOFFICE_NOT_FOUND: "SOFFICE_NOT_FOUND",
+      CONVERT_FAILED: "CONVERT_FAILED",
+      TIMEOUT: "CONVERT_FAILED",
+      UNSUPPORTED_PLATFORM: "CONVERT_FAILED",
+      UNSUPPORTED_FORMAT: "UNSUPPORTED_INPUT"
+    };
+    return new UnifiedOcrError(codeMap[err.code] ?? "CONVERT_FAILED", stage, err.message);
+  }
   const message = err instanceof Error ? err.message : String(err);
   const codeByStage = {
     convert: "CONVERT_FAILED",