@clazic/kordoc 2.5.1 → 2.5.2

package/dist/mcp.js

@@ -10,15 +10,15 @@ import {
   markdownToHwpx,
   markdownToXlsx,
   parse
-} from "./chunk-QG6BYZMR.js";
+} from "./chunk-25ZYYLVP.js";
 import {
   KordocError,
   VERSION,
   sanitizeError,
   toArrayBuffer
-} from "./chunk-IJGNPAK2.js";
+} from "./chunk-5CILZHRW.js";
 import "./chunk-MOL7MDBG.js";
-import "./chunk-Y4WFKJ5P.js";
+import "./chunk-S7BHLD2V.js";
 import "./chunk-YW5G6BCJ.js";
 import {
   createLoggerFromEnv,

package/dist/{resolve-XWYJYKKH.js → resolve-ZSUEJK3E.js}

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   createCliOcrProvider
-} from "./chunk-Y4WFKJ5P.js";
+} from "./chunk-S7BHLD2V.js";
 import {
   detectAvailableOcr,
   validateOcrMode
@@ -22,7 +22,7 @@ async function resolveOcrProvider(mode, warnings, concurrency, batchSize) {
   if (mode !== "auto") {
     validateOcrMode(mode);
     if (mode === "gemini" || mode === "claude" || mode === "codex") {
-      const { createBatchCliProvider, DEFAULT_BATCH_SIZES } = await import("./batch-provider-5BFJRKAZ.js");
+      const { createBatchCliProvider, DEFAULT_BATCH_SIZES } = await import("./batch-provider-XRF6F26E.js");
       const effectiveBatch = batchSize ?? DEFAULT_BATCH_SIZES[mode];
       if (effectiveBatch > 1) {
         logger.log({ level: "info", event: "done", message: "Batch CLI provider \uC120\uD0DD", meta: { mode, batchSize: effectiveBatch } });
@@ -48,7 +48,7 @@ async function resolveOcrProvider(mode, warnings, concurrency, batchSize) {
     });
   }
   if (detected === "gemini" || detected === "codex" || detected === "claude") {
-    const { createBatchCliProvider, DEFAULT_BATCH_SIZES } = await import("./batch-provider-5BFJRKAZ.js");
+    const { createBatchCliProvider, DEFAULT_BATCH_SIZES } = await import("./batch-provider-XRF6F26E.js");
     const effectiveBatch = batchSize ?? DEFAULT_BATCH_SIZES[detected];
     if (effectiveBatch > 1) {
       logger.log({ level: "info", event: "done", message: "AUTO: Batch CLI provider \uC120\uD0DD", meta: { mode: detected, batchSize: effectiveBatch } });
@@ -63,4 +63,4 @@ async function resolveOcrProvider(mode, warnings, concurrency, batchSize) {
 export {
   resolveOcrProvider
 };
-//# sourceMappingURL=resolve-XWYJYKKH.js.map
+//# sourceMappingURL=resolve-ZSUEJK3E.js.map

package/dist/{utils-RBXHHCLI.js → utils-H2BL5GNR.js}

@@ -9,7 +9,7 @@ import {
   sanitizeError,
   sanitizeHref,
   toArrayBuffer
-} from "./chunk-IJGNPAK2.js";
+} from "./chunk-5CILZHRW.js";
 import "./chunk-ZWE3DS7E.js";
 export {
   KordocError,
@@ -22,4 +22,4 @@ export {
   sanitizeHref,
   toArrayBuffer
 };
-//# sourceMappingURL=utils-RBXHHCLI.js.map
+//# sourceMappingURL=utils-H2BL5GNR.js.map

package/dist/{watch-5CCMTZ7F.js → watch-D6ODQLPJ.js}

@@ -2,12 +2,12 @@
 import {
   detectFormat,
   parse
-} from "./chunk-QG6BYZMR.js";
+} from "./chunk-25ZYYLVP.js";
 import {
   toArrayBuffer
-} from "./chunk-IJGNPAK2.js";
+} from "./chunk-5CILZHRW.js";
 import "./chunk-MOL7MDBG.js";
-import "./chunk-Y4WFKJ5P.js";
+import "./chunk-S7BHLD2V.js";
 import "./chunk-YW5G6BCJ.js";
 import "./chunk-I6YC6ZGK.js";
 import "./chunk-ZWE3DS7E.js";
@@ -136,4 +136,4 @@ async function sendWebhook(url, payload) {
 export {
   watchDirectory
 };
-//# sourceMappingURL=watch-5CCMTZ7F.js.map
+//# sourceMappingURL=watch-D6ODQLPJ.js.map

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@clazic/kordoc",
-  "version": "2.5.1",
+  "version": "2.5.2",
   "description": "Parse Korean documents (HWP, HWPX, PDF, XLSX, DOCX) to Markdown",
   "type": "module",
   "exports": {

package/dist/batch-provider-5BFJRKAZ.js

@@ -1,190 +0,0 @@
-#!/usr/bin/env node
-import "./chunk-ZWE3DS7E.js";
-
-// src/ocr/batch-provider.ts
-import { spawn, execSync } from "child_process";
-import { writeFileSync, readFileSync, unlinkSync, mkdirSync } from "fs";
-import { join } from "path";
-import { tmpdir } from "os";
-var BATCH_OCR_PROMPT = "\uB2E4\uC74C \uBB38\uC11C \uD398\uC774\uC9C0 \uC774\uBBF8\uC9C0\uB4E4\uC744 OCR\uD558\uC5EC \uC21C\uC218 Markdown\uC73C\uB85C \uBCC0\uD658\uD558\uC138\uC694.\n\n\uADDC\uCE59:\n- \uAC01 \uD398\uC774\uC9C0 \uACB0\uACFC \uC0AC\uC774\uC5D0 \uBC18\uB4DC\uC2DC \uC774 \uAD6C\uBD84\uC790\uB97C \uC0BD\uC785: <!-- PAGE_BREAK -->\n- \uD14C\uC774\uBE14\uC740 Markdown \uD14C\uC774\uBE14 \uBB38\uBC95 \uC0AC\uC6A9 (| \uAD6C\uBD84, |---|---| \uD5E4\uB354 \uAD6C\uBD84\uC120 \uD3EC\uD568)\n- \uBCD1\uD569\uB41C \uC140\uC740 \uD574\uB2F9 \uC704\uCE58\uC5D0 \uB0B4\uC6A9 \uAE30\uC7AC\n- \uD5E4\uB529\uC740 \uAE00\uC790 \uD06C\uAE30\uC5D0 \uB530\uB77C ## ~ ###### \uC0AC\uC6A9\n- \uB9AC\uC2A4\uD2B8\uB294 - \uB610\uB294 1. \uC0AC\uC6A9\n- \uC774\uBBF8\uC9C0, \uB3C4\uD615 \uB4F1 \uBE44\uD14D\uC2A4\uD2B8 \uC694\uC18C\uB294 \uBB34\uC2DC\n- \uC6D0\uBB38\uC758 \uC77D\uAE30 \uC21C\uC11C\uC640 \uAD6C\uC870\uB97C \uC720\uC9C0\n- ```\uB85C \uAC10\uC2F8\uC9C0 \uB9D0\uACE0 \uC21C\uC218 Markdown\uB9CC \uCD9C\uB825";
-var DEFAULT_BATCH_SIZES = {
-  gemini: 5,
-  claude: 5,
-  codex: 10
-};
-var _batchTempDir = null;
-function getBatchTempDir() {
-  if (!_batchTempDir) {
-    _batchTempDir = join(process.cwd(), ".kordoc_ocr_tmp");
-    mkdirSync(_batchTempDir, { recursive: true });
-    if (process.platform === "win32") {
-      try {
-        execSync(`attrib +h "${_batchTempDir}"`, { stdio: "ignore" });
-      } catch {
-      }
-    }
-  }
-  return _batchTempDir;
-}
-function createBatchCliProvider(mode, batchSize) {
-  return {
-    __batch: true,
-    batchSize,
-    async processBatch(pages) {
-      const results = /* @__PURE__ */ new Map();
-      const tempDir = getBatchTempDir();
-      const tempFiles = [];
-      try {
-        for (const { image, pageNum } of pages) {
-          const path = join(tempDir, `batch-p${pageNum}.png`);
-          writeFileSync(path, image);
-          tempFiles.push(path);
-        }
-        let output;
-        if (mode === "codex") {
-          output = await callBatchCodexCli(tempFiles);
-        } else {
-          output = await callBatchCli(mode, tempFiles);
-        }
-        const cleaned = stripCodeFence(output.trim());
-        const parts = cleaned.split(/<!--\s*PAGE_BREAK\s*-->/).map((p) => p.trim()).filter((p) => p.length > 0);
-        for (let i = 0; i < pages.length; i++) {
-          const pageNum = pages[i].pageNum;
-          if (i < parts.length) {
-            results.set(pageNum, { markdown: parts[i] });
-          }
-        }
-      } finally {
-        for (const f of tempFiles) {
-          try {
-            unlinkSync(f);
-          } catch {
-          }
-        }
-      }
-      return results;
-    }
-  };
-}
-function spawnAsync(cmd, args, opts) {
-  return new Promise((resolve, reject) => {
-    const child = spawn(cmd, args, {
-      cwd: opts.cwd,
-      env: process.env,
-      stdio: ["pipe", "pipe", "pipe"],
-      shell: process.platform === "win32"
-    });
-    let stdout = "";
-    let stderr = "";
-    let killed = false;
-    child.stdout.setEncoding("utf-8");
-    child.stderr.setEncoding("utf-8");
-    child.stdout.on("data", (d) => {
-      stdout += d;
-    });
-    child.stderr.on("data", (d) => {
-      stderr += d;
-    });
-    const timer = setTimeout(() => {
-      killed = true;
-      if (process.platform === "win32") {
-        child.kill();
-      } else {
-        child.kill("SIGTERM");
-      }
-    }, opts.timeoutMs);
-    if (opts.stdin !== void 0) {
-      child.stdin.end(opts.stdin);
-    } else {
-      child.stdin.end();
-    }
-    child.on("close", (code) => {
-      clearTimeout(timer);
-      if (killed) {
-        reject(new Error(`\uD0C0\uC784\uC544\uC6C3 (${Math.round(opts.timeoutMs / 1e3)}\uCD08)`));
-      } else {
-        resolve({ stdout, stderr, exitCode: code ?? 1 });
-      }
-    });
-    child.on("error", (err) => {
-      clearTimeout(timer);
-      reject(err);
-    });
-  });
-}
-async function callBatchCli(mode, imagePaths) {
-  const fileRefs = imagePaths.map((p) => `@${p.replace(/\\/g, "/")}`).join("\n");
-  const prompt = `${BATCH_OCR_PROMPT}
-
-${fileRefs}`;
-  let args;
-  if (mode === "gemini") {
-    const model = process.env.KORDOC_GEMINI_MODEL ?? "gemini-2.5-flash";
-    args = ["--prompt", prompt, "--yolo", "--model", model];
-  } else {
-    args = ["--print", prompt];
-    const model = process.env.KORDOC_CLAUDE_MODEL;
-    if (model) args.push("--model", model);
-  }
-  const timeoutMs = 6e4 + imagePaths.length * 2e4;
-  const result = await spawnAsync(mode, args, {
-    timeoutMs,
-    ...mode === "claude" ? { cwd: tmpdir() } : {}
-  });
-  if (result.exitCode !== 0) {
-    const errMsg = result.stderr?.trim() || `exit code ${result.exitCode}`;
-    throw new Error(`${mode} \uBC30\uCE58 OCR \uC2E4\uD328: ${errMsg}`);
-  }
-  const output = result.stdout || "";
-  checkForLimitError(output, mode);
-  return output;
-}
-async function callBatchCodexCli(imagePaths) {
-  const outPath = join(tmpdir(), `kordoc-codex-batch-${Date.now()}-${Math.random().toString(36).slice(2)}.txt`);
-  try {
-    const args = ["exec", BATCH_OCR_PROMPT];
-    for (const p of imagePaths) {
-      args.push("--image", p);
-    }
-    args.push("--output-last-message", outPath);
-    const model = process.env.KORDOC_CODEX_MODEL;
-    if (model) args.push("--model", model);
-    const timeoutMs = 6e4 + imagePaths.length * 2e4;
-    const result = await spawnAsync("codex", args, {
-      timeoutMs,
-      stdin: ""
-    });
-    if (result.exitCode !== 0) {
-      const errMsg = result.stderr?.trim() || `exit code ${result.exitCode}`;
-      throw new Error(`codex \uBC30\uCE58 OCR \uC2E4\uD328: ${errMsg}`);
-    }
-    let text;
-    try {
-      text = readFileSync(outPath, "utf-8");
-    } catch {
-      text = result.stdout || "";
-    }
-    checkForLimitError(text, "codex");
-    return text;
-  } finally {
-    try {
-      unlinkSync(outPath);
-    } catch {
-    }
-  }
-}
-function checkForLimitError(output, mode) {
-  const lower = output.toLowerCase();
-  if (lower.includes("usage limit") || lower.includes("rate limit")) {
-    throw new Error(`${mode} \uC0AC\uC6A9\uB7C9/\uC18D\uB3C4 \uC81C\uD55C: ${output.trim().slice(0, 200)}`);
-  }
-}
-function stripCodeFence(text) {
-  const match = text.match(/^```(?:markdown|md)?\s*\n([\s\S]*?)\n```\s*$/m);
-  return match ? match[1].trim() : text;
-}
-export {
-  DEFAULT_BATCH_SIZES,
-  createBatchCliProvider
-};
-//# sourceMappingURL=batch-provider-5BFJRKAZ.js.map

package/dist/batch-provider-5BFJRKAZ.js.map

@@ -1 +0,0 @@
-{"version":3,"sources":["../src/ocr/batch-provider.ts"],"sourcesContent":["/**\n * CLI 배치 OCR 프로바이더\n *\n * 여러 페이지 이미지를 단일 CLI 호출로 처리하여 API 호출 수를 대폭 감소.\n * gemini/claude: @file 멀티 참조, codex: --image 멀티 플래그\n *\n * 299페이지 기준:\n * - 기존: CLI 299회 호출 (~30분)\n * - 배치: CLI 3~6회 호출 (~3분)\n */\n\nimport { spawn, execSync } from \"child_process\"\nimport { writeFileSync, readFileSync, unlinkSync, mkdirSync } from \"fs\"\nimport { join } from \"path\"\nimport { tmpdir } from \"os\"\nimport type { StructuredOcrResult, BatchOcrProvider } from \"../types.js\"\n\n/** 배치 OCR 프롬프트 */\nconst BATCH_OCR_PROMPT =\n  \"다음 문서 페이지 이미지들을 OCR하여 순수 Markdown으로 변환하세요.\\n\\n\" +\n  \"규칙:\\n\" +\n  \"- 각 페이지 결과 사이에 반드시 이 구분자를 삽입: <!-- PAGE_BREAK -->\\n\" +\n  \"- 테이블은 Markdown 테이블 문법 사용 (| 구분, |---|---| 헤더 구분선 포함)\\n\" +\n  \"- 병합된 셀은 해당 위치에 내용 기재\\n\" +\n  \"- 헤딩은 글자 크기에 따라 ## ~ ###### 사용\\n\" +\n  \"- 리스트는 - 또는 1. 사용\\n\" +\n  \"- 이미지, 도형 등 비텍스트 요소는 무시\\n\" +\n  \"- 원문의 읽기 순서와 구조를 유지\\n\" +\n  \"- ```로 감싸지 말고 순수 Markdown만 출력\"\n\n/** 모드별 기본 배치 크기 (CLI 내부 타임아웃 + 실측 기반)\n *\n * gemini CLI: 10장 이상에서 AbortError 발생 (내부 타임아웃).\n * 5장 배치가 안정적으로 동작 확인 (35초/배치).\n * 299페이지 = 60배치 = 기존 299회 대비 80% 감소.\n */\nexport const DEFAULT_BATCH_SIZES: Record<string, number> = {\n  gemini: 5,\n  claude: 5,\n  codex: 10,\n}\n\n/**\n * 임시 디렉토리 — gemini CLI는 cwd 하위 + gitignore 밖만 @참조 가능\n *\n * 숨김 처리:\n * - macOS/Linux: '.' 접두사로 기본 숨김 (ls -a 로만 표시)\n * - Windows: '.' 접두사 + attrib +h 로 숨김 속성 부여\n */\nlet _batchTempDir: string | null = null\nfunction getBatchTempDir(): string {\n  if (!_batchTempDir) {\n    _batchTempDir = join(process.cwd(), \".kordoc_ocr_tmp\")\n    mkdirSync(_batchTempDir, { recursive: true })\n    // Windows: dot-prefix만으로 숨김 처리 불충분 → attrib +h 추가\n    if (process.platform === \"win32\") {\n      try { execSync(`attrib +h \"${_batchTempDir}\"`, { stdio: \"ignore\" }) } catch { /* ignore */ }\n    }\n  }\n  return _batchTempDir\n}\n\n/**\n * 배치 CLI 프로바이더 생성\n */\nexport function createBatchCliProvider(\n  mode: \"gemini\" | \"claude\" | \"codex\",\n  batchSize: number\n): BatchOcrProvider {\n  return {\n    __batch: true as const,\n    batchSize,\n    async processBatch(pages) {\n      const results = new Map<number, StructuredOcrResult>()\n      const tempDir = getBatchTempDir()\n      const tempFiles: string[] = []\n\n      try {\n        // 1. Write all page images to temp files\n        for (const { image, pageNum } of pages) {\n          const path = join(tempDir, `batch-p${pageNum}.png`)\n          writeFileSync(path, image)\n          tempFiles.push(path)\n        }\n\n        // 2. Call CLI with all file references (비동기 — 병렬 배치 실행 가능)\n        let output: string\n        if (mode === \"codex\") {\n          output = await callBatchCodexCli(tempFiles)\n        } else {\n          output = await callBatchCli(mode, tempFiles)\n        }\n\n        // 3. Parse response by PAGE_BREAK separator\n        const cleaned = stripCodeFence(output.trim())\n        const parts = cleaned.split(/<!--\\s*PAGE_BREAK\\s*-->/)\n          .map(p => p.trim())\n          .filter(p => p.length > 0)\n\n        // 4. Map results to page numbers (best-effort if count mismatch)\n        for (let i = 0; i < pages.length; i++) {\n          const pageNum = pages[i].pageNum\n          if (i < parts.length) {\n            results.set(pageNum, { markdown: parts[i] })\n          }\n          // If fewer parts than pages, remaining pages get no result\n        }\n      } finally {\n        // 5. Clean up temp files\n        for (const f of tempFiles) {\n          try { unlinkSync(f) } catch { /* ignore */ }\n        }\n      }\n\n      return results\n    },\n  }\n}\n\n/**\n * 비동기 CLI 실행 헬퍼 — spawn + Promise 래핑.\n * spawnSync는 이벤트 루프를 차단하여 병렬 배치 실행 불가.\n */\nfunction spawnAsync(\n  cmd: string,\n  args: string[],\n  opts: { timeoutMs: number; cwd?: string; stdin?: string }\n): Promise<{ stdout: string; stderr: string; exitCode: number }> {\n  return new Promise((resolve, reject) => {\n    const child = spawn(cmd, args, {\n      cwd: opts.cwd,\n      env: process.env,\n      stdio: [\"pipe\", \"pipe\", \"pipe\"],\n      shell: process.platform === \"win32\",\n    })\n\n    let stdout = \"\"\n    let stderr = \"\"\n    let killed = false\n\n    child.stdout.setEncoding(\"utf-8\")\n    child.stderr.setEncoding(\"utf-8\")\n    child.stdout.on(\"data\", (d: string) => { stdout += d })\n    child.stderr.on(\"data\", (d: string) => { stderr += d })\n\n    const timer = setTimeout(() => {\n      killed = true\n      if (process.platform === \"win32\") {\n        child.kill()\n      } else {\n        child.kill(\"SIGTERM\")\n      }\n    }, opts.timeoutMs)\n\n    if (opts.stdin !== undefined) {\n      child.stdin.end(opts.stdin)\n    } else {\n      child.stdin.end()\n    }\n\n    child.on(\"close\", (code) => {\n      clearTimeout(timer)\n      if (killed) {\n        reject(new Error(`타임아웃 (${Math.round(opts.timeoutMs / 1000)}초)`))\n      } else {\n        resolve({ stdout, stderr, exitCode: code ?? 1 })\n      }\n    })\n    child.on(\"error\", (err) => {\n      clearTimeout(timer)\n      reject(err)\n    })\n  })\n}\n\n/** gemini/claude 배치 호출 (비동기) */\nasync function callBatchCli(mode: \"gemini\" | \"claude\", imagePaths: string[]): Promise<string> {\n  const fileRefs = imagePaths.map(p => `@${p.replace(/\\\\/g, \"/\")}`).join(\"\\n\")\n  const prompt = `${BATCH_OCR_PROMPT}\\n\\n${fileRefs}`\n\n  let args: string[]\n  if (mode === \"gemini\") {\n    const model = process.env.KORDOC_GEMINI_MODEL ?? \"gemini-2.5-flash\"\n    args = [\"--prompt\", prompt, \"--yolo\", \"--model\", model]\n  } else {\n    args = [\"--print\", prompt]\n    const model = process.env.KORDOC_CLAUDE_MODEL\n    if (model) args.push(\"--model\", model)\n  }\n\n  const timeoutMs = 60_000 + imagePaths.length * 20_000\n  const result = await spawnAsync(mode, args, {\n    timeoutMs,\n    ...(mode === \"claude\" ? { cwd: tmpdir() } : {}),\n  })\n\n  if (result.exitCode !== 0) {\n    const errMsg = result.stderr?.trim() || `exit code ${result.exitCode}`\n    throw new Error(`${mode} 배치 OCR 실패: ${errMsg}`)\n  }\n\n  const output = result.stdout || \"\"\n  checkForLimitError(output, mode)\n  return output\n}\n\n/** codex 배치 호출 (비동기) — --image를 여러 번 지정 */\nasync function callBatchCodexCli(imagePaths: string[]): Promise<string> {\n  const outPath = join(tmpdir(), `kordoc-codex-batch-${Date.now()}-${Math.random().toString(36).slice(2)}.txt`)\n  try {\n    const args = [\"exec\", BATCH_OCR_PROMPT]\n    for (const p of imagePaths) {\n      args.push(\"--image\", p)\n    }\n    args.push(\"--output-last-message\", outPath)\n    const model = process.env.KORDOC_CODEX_MODEL\n    if (model) args.push(\"--model\", model)\n\n    const timeoutMs = 60_000 + imagePaths.length * 20_000\n    const result = await spawnAsync(\"codex\", args, {\n      timeoutMs,\n      stdin: \"\",\n    })\n\n    if (result.exitCode !== 0) {\n      const errMsg = result.stderr?.trim() || `exit code ${result.exitCode}`\n      throw new Error(`codex 배치 OCR 실패: ${errMsg}`)\n    }\n\n    let text: string\n    try {\n      text = readFileSync(outPath, \"utf-8\")\n    } catch {\n      text = result.stdout || \"\"\n    }\n    checkForLimitError(text, \"codex\")\n    return text\n  } finally {\n    try { unlinkSync(outPath) } catch { /* ignore */ }\n  }\n}\n\n/**\n * 출력 텍스트에서 사용량·속도 제한 에러 감지.\n * 해당 메시지가 포함된 경우 throw하여 다음 엔진으로 fallback 트리거.\n */\nfunction checkForLimitError(output: string, mode: string): void {\n  const lower = output.toLowerCase()\n  if (lower.includes(\"usage limit\") || lower.includes(\"rate limit\")) {\n    throw new Error(`${mode} 사용량/속도 제한: ${output.trim().slice(0, 200)}`)\n  }\n}\n\n/** LLM 출력에서 코드 펜스 제거 (cli-provider.ts와 동일 로직) */\nfunction stripCodeFence(text: string): string {\n  const match = text.match(/^```(?:markdown|md)?\\s*\\n([\\s\\S]*?)\\n```\\s*$/m)\n  return match ? match[1].trim() : text\n}\n"],"mappings":";;;;AAWA,SAAS,OAAO,gBAAgB;AAChC,SAAS,eAAe,cAAc,YAAY,iBAAiB;AACnE,SAAS,YAAY;AACrB,SAAS,cAAc;AAIvB,IAAM,mBACJ;AAiBK,IAAM,sBAA8C;AAAA,EACzD,QAAQ;AAAA,EACR,QAAQ;AAAA,EACR,OAAO;AACT;AASA,IAAI,gBAA+B;AACnC,SAAS,kBAA0B;AACjC,MAAI,CAAC,eAAe;AAClB,oBAAgB,KAAK,QAAQ,IAAI,GAAG,iBAAiB;AACrD,cAAU,eAAe,EAAE,WAAW,KAAK,CAAC;AAE5C,QAAI,QAAQ,aAAa,SAAS;AAChC,UAAI;AAAE,iBAAS,cAAc,aAAa,KAAK,EAAE,OAAO,SAAS,CAAC;AAAA,MAAE,QAAQ;AAAA,MAAe;AAAA,IAC7F;AAAA,EACF;AACA,SAAO;AACT;AAKO,SAAS,uBACd,MACA,WACkB;AAClB,SAAO;AAAA,IACL,SAAS;AAAA,IACT;AAAA,IACA,MAAM,aAAa,OAAO;AACxB,YAAM,UAAU,oBAAI,IAAiC;AACrD,YAAM,UAAU,gBAAgB;AAChC,YAAM,YAAsB,CAAC;AAE7B,UAAI;AAEF,mBAAW,EAAE,OAAO,QAAQ,KAAK,OAAO;AACtC,gBAAM,OAAO,KAAK,SAAS,UAAU,OAAO,MAAM;AAClD,wBAAc,MAAM,KAAK;AACzB,oBAAU,KAAK,IAAI;AAAA,QACrB;AAGA,YAAI;AACJ,YAAI,SAAS,SAAS;AACpB,mBAAS,MAAM,kBAAkB,SAAS;AAAA,QAC5C,OAAO;AACL,mBAAS,MAAM,aAAa,MAAM,SAAS;AAAA,QAC7C;AAGA,cAAM,UAAU,eAAe,OAAO,KAAK,CAAC;AAC5C,cAAM,QAAQ,QAAQ,MAAM,yBAAyB,EAClD,IAAI,OAAK,EAAE,KAAK,CAAC,EACjB,OAAO,OAAK,EAAE,SAAS,CAAC;AAG3B,iBAAS,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK;AACrC,gBAAM,UAAU,MAAM,CAAC,EAAE;AACzB,cAAI,IAAI,MAAM,QAAQ;AACpB,oBAAQ,IAAI,SAAS,EAAE,UAAU,MAAM,CAAC,EAAE,CAAC;AAAA,UAC7C;AAAA,QAEF;AAAA,MACF,UAAE;AAEA,mBAAW,KAAK,WAAW;AACzB,cAAI;AAAE,uBAAW,CAAC;AAAA,UAAE,QAAQ;AAAA,UAAe;AAAA,QAC7C;AAAA,MACF;AAEA,aAAO;AAAA,IACT;AAAA,EACF;AACF;AAMA,SAAS,WACP,KACA,MACA,MAC+D;AAC/D,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,QAAQ,MAAM,KAAK,MAAM;AAAA,MAC7B,KAAK,KAAK;AAAA,MACV,KAAK,QAAQ;AAAA,MACb,OAAO,CAAC,QAAQ,QAAQ,MAAM;AAAA,MAC9B,OAAO,QAAQ,aAAa;AAAA,IAC9B,CAAC;AAED,QAAI,SAAS;AACb,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,UAAM,OAAO,YAAY,OAAO;AAChC,UAAM,OAAO,YAAY,OAAO;AAChC,UAAM,OAAO,GAAG,QAAQ,CAAC,MAAc;AAAE,gBAAU;AAAA,IAAE,CAAC;AACtD,UAAM,OAAO,GAAG,QAAQ,CAAC,MAAc;AAAE,gBAAU;AAAA,IAAE,CAAC;AAEtD,UAAM,QAAQ,WAAW,MAAM;AAC7B,eAAS;AACT,UAAI,QAAQ,aAAa,SAAS;AAChC,cAAM,KAAK;AAAA,MACb,OAAO;AACL,cAAM,KAAK,SAAS;AAAA,MACtB;AAAA,IACF,GAAG,KAAK,SAAS;AAEjB,QAAI,KAAK,UAAU,QAAW;AAC5B,YAAM,MAAM,IAAI,KAAK,KAAK;AAAA,IAC5B,OAAO;AACL,YAAM,MAAM,IAAI;AAAA,IAClB;AAEA,UAAM,GAAG,SAAS,CAAC,SAAS;AAC1B,mBAAa,KAAK;AAClB,UAAI,QAAQ;AACV,eAAO,IAAI,MAAM,6BAAS,KAAK,MAAM,KAAK,YAAY,GAAI,CAAC,SAAI,CAAC;AAAA,MAClE,OAAO;AACL,gBAAQ,EAAE,QAAQ,QAAQ,UAAU,QAAQ,EAAE,CAAC;AAAA,MACjD;AAAA,IACF,CAAC;AACD,UAAM,GAAG,SAAS,CAAC,QAAQ;AACzB,mBAAa,KAAK;AAClB,aAAO,GAAG;AAAA,IACZ,CAAC;AAAA,EACH,CAAC;AACH;AAGA,eAAe,aAAa,MAA2B,YAAuC;AAC5F,QAAM,WAAW,WAAW,IAAI,OAAK,IAAI,EAAE,QAAQ,OAAO,GAAG,CAAC,EAAE,EAAE,KAAK,IAAI;AAC3E,QAAM,SAAS,GAAG,gBAAgB;AAAA;AAAA,EAAO,QAAQ;AAEjD,MAAI;AACJ,MAAI,SAAS,UAAU;AACrB,UAAM,QAAQ,QAAQ,IAAI,uBAAuB;AACjD,WAAO,CAAC,YAAY,QAAQ,UAAU,WAAW,KAAK;AAAA,EACxD,OAAO;AACL,WAAO,CAAC,WAAW,MAAM;AACzB,UAAM,QAAQ,QAAQ,IAAI;AAC1B,QAAI,MAAO,MAAK,KAAK,WAAW,KAAK;AAAA,EACvC;AAEA,QAAM,YAAY,MAAS,WAAW,SAAS;AAC/C,QAAM,SAAS,MAAM,WAAW,MAAM,MAAM;AAAA,IAC1C;AAAA,IACA,GAAI,SAAS,WAAW,EAAE,KAAK,OAAO,EAAE,IAAI,CAAC;AAAA,EAC/C,CAAC;AAED,MAAI,OAAO,aAAa,GAAG;AACzB,UAAM,SAAS,OAAO,QAAQ,KAAK,KAAK,aAAa,OAAO,QAAQ;AACpE,UAAM,IAAI,MAAM,GAAG,IAAI,mCAAe,MAAM,EAAE;AAAA,EAChD;AAEA,QAAM,SAAS,OAAO,UAAU;AAChC,qBAAmB,QAAQ,IAAI;AAC/B,SAAO;AACT;AAGA,eAAe,kBAAkB,YAAuC;AACtE,QAAM,UAAU,KAAK,OAAO,GAAG,sBAAsB,KAAK,IAAI,CAAC,IAAI,KAAK,OAAO,EAAE,SAAS,EAAE,EAAE,MAAM,CAAC,CAAC,MAAM;AAC5G,MAAI;AACF,UAAM,OAAO,CAAC,QAAQ,gBAAgB;AACtC,eAAW,KAAK,YAAY;AAC1B,WAAK,KAAK,WAAW,CAAC;AAAA,IACxB;AACA,SAAK,KAAK,yBAAyB,OAAO;AAC1C,UAAM,QAAQ,QAAQ,IAAI;AAC1B,QAAI,MAAO,MAAK,KAAK,WAAW,KAAK;AAErC,UAAM,YAAY,MAAS,WAAW,SAAS;AAC/C,UAAM,SAAS,MAAM,WAAW,SAAS,MAAM;AAAA,MAC7C;AAAA,MACA,OAAO;AAAA,IACT,CAAC;AAED,QAAI,OAAO,aAAa,GAAG;AACzB,YAAM,SAAS,OAAO,QAAQ,KAAK,KAAK,aAAa,OAAO,QAAQ;AACpE,YAAM,IAAI,MAAM,wCAAoB,MAAM,EAAE;AAAA,IAC9C;AAEA,QAAI;AACJ,QAAI;AACF,aAAO,aAAa,SAAS,OAAO;AAAA,IACtC,QAAQ;AACN,aAAO,OAAO,UAAU;AAAA,IAC1B;AACA,uBAAmB,MAAM,OAAO;AAChC,WAAO;AAAA,EACT,UAAE;AACA,QAAI;AAAE,iBAAW,OAAO;AAAA,IAAE,QAAQ;AAAA,IAAe;AAAA,EACnD;AACF;AAMA,SAAS,mBAAmB,QAAgB,MAAoB;AAC9D,QAAM,QAAQ,OAAO,YAAY;AACjC,MAAI,MAAM,SAAS,aAAa,KAAK,MAAM,SAAS,YAAY,GAAG;AACjE,UAAM,IAAI,MAAM,GAAG,IAAI,kDAAe,OAAO,KAAK,EAAE,MAAM,GAAG,GAAG,CAAC,EAAE;AAAA,EACrE;AACF;AAGA,SAAS,eAAe,MAAsB;AAC5C,QAAM,QAAQ,KAAK,MAAM,+CAA+C;AACxE,SAAO,QAAQ,MAAM,CAAC,EAAE,KAAK,IAAI;AACnC;","names":[]}

package/dist/chunk-Y4WFKJ5P.js

@@ -1,167 +0,0 @@
-#!/usr/bin/env node
-
-// src/ocr/cli-provider.ts
-import { spawnSync } from "child_process";
-import { writeFileSync, readFileSync, unlinkSync, mkdirSync } from "fs";
-import { join } from "path";
-import { tmpdir } from "os";
-var OCR_PROMPT = `\uC774 PDF \uD398\uC774\uC9C0 \uC774\uBBF8\uC9C0\uC5D0\uC11C \uD14D\uC2A4\uD2B8\uC640 \uD14C\uC774\uBE14\uC744 \uCD94\uCD9C\uD558\uC5EC \uC21C\uC218 Markdown\uC73C\uB85C \uBCC0\uD658\uD558\uC138\uC694.
-\uADDC\uCE59:
-- \uD14C\uC774\uBE14\uC740 Markdown \uD14C\uC774\uBE14 \uBB38\uBC95 \uC0AC\uC6A9 (| \uAD6C\uBD84, |---|---| \uD5E4\uB354 \uAD6C\uBD84\uC120 \uD3EC\uD568)
-- \uBCD1\uD569\uB41C \uC140\uC740 \uD574\uB2F9 \uC704\uCE58\uC5D0 \uB0B4\uC6A9 \uAE30\uC7AC
-- \uD5E4\uB529\uC740 \uAE00\uC790 \uD06C\uAE30\uC5D0 \uB530\uB77C ## ~ ###### \uC0AC\uC6A9
-- \uB9AC\uC2A4\uD2B8\uB294 - \uB610\uB294 1. \uC0AC\uC6A9
-- \uC774\uBBF8\uC9C0, \uB3C4\uD615 \uB4F1 \uBE44\uD14D\uC2A4\uD2B8 \uC694\uC18C\uB294 \uBB34\uC2DC
-- \uC6D0\uBB38\uC758 \uC77D\uAE30 \uC21C\uC11C\uC640 \uAD6C\uC870\uB97C \uC720\uC9C0
-- \`\`\`\uB85C \uAC10\uC2F8\uC9C0 \uB9D0\uACE0 \uC21C\uC218 Markdown\uB9CC \uCD9C\uB825`;
-var _tempDir = null;
-function getTempDir() {
-  if (!_tempDir) {
-    _tempDir = join(process.cwd(), ".kordoc_ocr_tmp");
-    mkdirSync(_tempDir, { recursive: true });
-  }
-  return _tempDir;
-}
-function createCliOcrProvider(mode) {
-  return async (pageImage, pageNumber) => {
-    const tempPath = join(getTempDir(), `page-${pageNumber}.png`);
-    try {
-      writeFileSync(tempPath, pageImage);
-      let output;
-      if (mode === "ollama") {
-        output = await callOllamaApi(tempPath);
-      } else {
-        output = callCli(mode, tempPath);
-      }
-      return { markdown: stripCodeFence(output.trim()) };
-    } finally {
-      try {
-        unlinkSync(tempPath);
-      } catch {
-      }
-    }
-  };
-}
-function checkForLimitError(output, mode) {
-  const lower = output.toLowerCase();
-  if (lower.includes("usage limit") || lower.includes("rate limit")) {
-    throw new Error(`${mode} \uC0AC\uC6A9\uB7C9/\uC18D\uB3C4 \uC81C\uD55C: ${output.trim().slice(0, 200)}`);
-  }
-}
-function callCli(mode, imagePath) {
-  if (mode === "codex") {
-    return callCodexCli(imagePath);
-  }
-  const args = buildCliArgs(mode, imagePath);
-  const result = spawnSync(mode, args, {
-    encoding: "utf-8",
-    timeout: 6e5,
-    maxBuffer: 10 * 1024 * 1024,
-    shell: process.platform === "win32",
-    // claude: /tmp에서 실행하여 프로젝트 CLAUDE.md의 규칙 간섭 방지
-    ...mode === "claude" ? { cwd: tmpdir() } : {}
-  });
-  if (result.error) {
-    throw new Error(`${mode} CLI \uC2E4\uD589 \uC2E4\uD328: ${result.error.message}`);
-  }
-  if (result.status !== 0) {
-    const errMsg = result.stderr?.trim() || `exit code ${result.status}`;
-    throw new Error(`${mode} OCR \uC2E4\uD328: ${errMsg}`);
-  }
-  const output = result.stdout || "";
-  checkForLimitError(output, mode);
-  return output;
-}
-function callCodexCli(imagePath) {
-  const outPath = join(tmpdir(), `kordoc-codex-out-${Date.now()}.txt`);
-  try {
-    const args = ["exec", OCR_PROMPT, "--image", imagePath, "--output-last-message", outPath];
-    const model = process.env.KORDOC_CODEX_MODEL;
-    if (model) args.push("--model", model);
-    const result = spawnSync("codex", args, {
-      encoding: "utf-8",
-      timeout: 18e4,
-      maxBuffer: 10 * 1024 * 1024,
-      input: "",
-      // stdin EOF 즉시 전달 (대화형 입력 차단)
-      shell: process.platform === "win32"
-    });
-    if (result.error) {
-      throw new Error(`codex CLI \uC2E4\uD589 \uC2E4\uD328: ${result.error.message}`);
-    }
-    if (result.status !== 0) {
-      const errMsg = result.stderr?.trim() || `exit code ${result.status}`;
-      throw new Error(`codex OCR \uC2E4\uD328: ${errMsg}`);
-    }
-    let text;
-    try {
-      text = readFileSync(outPath, "utf-8");
-    } catch {
-      text = result.stdout || "";
-    }
-    checkForLimitError(text, "codex");
-    return text;
-  } finally {
-    try {
-      unlinkSync(outPath);
-    } catch {
-    }
-  }
-}
-function buildCliArgs(mode, imagePath) {
-  const normalizedPath = imagePath.replace(/\\/g, "/");
-  const promptWithImage = `${OCR_PROMPT}
-
-\uC774\uBBF8\uC9C0: @${normalizedPath}`;
-  switch (mode) {
-    case "gemini": {
-      const args = ["--prompt", promptWithImage, "--yolo"];
-      const model = process.env.KORDOC_GEMINI_MODEL;
-      if (model) args.push("--model", model);
-      return args;
-    }
-    case "claude": {
-      const args = ["--print", promptWithImage];
-      const model = process.env.KORDOC_CLAUDE_MODEL;
-      if (model) args.push("--model", model);
-      return args;
-    }
-    default:
-      throw new Error(`\uC9C0\uC6D0\uD558\uC9C0 \uC54A\uB294 CLI: ${mode}`);
-  }
-}
-async function callOllamaApi(imagePath) {
-  const { readFileSync: readFileSync2 } = await import("fs");
-  const imageBase64 = readFileSync2(imagePath).toString("base64");
-  const model = process.env.KORDOC_OLLAMA_MODEL || "qwen3-vl:8b";
-  const host = process.env.KORDOC_OLLAMA_HOST || "http://localhost:11434";
-  const timeoutMs = Number(process.env.KORDOC_OLLAMA_TIMEOUT) || 12e4;
-  const response = await fetch(`${host}/api/chat`, {
-    method: "POST",
-    headers: { "Content-Type": "application/json" },
-    body: JSON.stringify({
-      model,
-      messages: [{
-        role: "user",
-        content: OCR_PROMPT,
-        images: [imageBase64]
-      }],
-      stream: false
-    }),
-    signal: AbortSignal.timeout(timeoutMs)
-  });
-  if (!response.ok) {
-    throw new Error(`Ollama API \uC624\uB958: ${response.status} ${response.statusText}`);
-  }
-  const data = await response.json();
-  return data.message?.content || "";
-}
-function stripCodeFence(text) {
-  const match = text.match(/^```(?:markdown|md)?\s*([\s\S]*?)```\s*$/m);
-  return match ? match[1].trim() : text;
-}
-
-export {
-  createCliOcrProvider
-};
-//# sourceMappingURL=chunk-Y4WFKJ5P.js.map