@clazic/kordoc 2.4.17 → 2.4.19

package/dist/{auto-detect-2YGFYQCN.js → auto-detect-CBYICI6B.js}

@@ -2,14 +2,14 @@
 import {
   detectAvailableOcr,
   getAutoFallbackChain,
-  getTesseractFallbackMessage,
+  getNoCliMessage,
   validateOcrMode
-} from "./chunk-7NOZFYH6.js";
+} from "./chunk-CLK4PNZ7.js";
 import "./chunk-ZWE3DS7E.js";
 export {
   detectAvailableOcr,
   getAutoFallbackChain,
-  getTesseractFallbackMessage,
+  getNoCliMessage,
   validateOcrMode
 };
-//# sourceMappingURL=auto-detect-2YGFYQCN.js.map
+//# sourceMappingURL=auto-detect-CBYICI6B.js.map

package/dist/{chunk-WM3XI23V.js → chunk-463YQ2WL.js}

@@ -1,7 +1,4 @@
 #!/usr/bin/env node
-import {
-  markdownToBlocks
-} from "./chunk-YW5G6BCJ.js";
 import {
   KordocError,
   classifyError,
@@ -10,16 +7,16 @@ import {
   precheckZipSize,
   sanitizeHref,
   toArrayBuffer
-} from "./chunk-W2KDIKDF.js";
+} from "./chunk-MZN7PLTZ.js";
 import {
   parsePageRange
 } from "./chunk-MOL7MDBG.js";
-import {
-  createTesseractProvider
-} from "./chunk-7FMKAV4P.js";
 import {
   createCliOcrProvider
-} from "./chunk-34WIGIQC.js";
+} from "./chunk-Y4WFKJ5P.js";
+import {
+  markdownToBlocks
+} from "./chunk-YW5G6BCJ.js";
 import {
   createLoggerFromEnv,
   generateRunId
@@ -5688,7 +5685,7 @@ async function parsePdfDocument(buffer, options) {
       if (ocrMode === "off") {
         throw Object.assign(new KordocError(`\uC774\uBBF8\uC9C0 \uAE30\uBC18 PDF (${pageCount}\uD398\uC774\uC9C0, ${totalChars}\uC790)`), { isImageBased: true });
       }
-      const { resolveOcrProvider } = await import("./resolve-673XFZQ6.js");
+      const { resolveOcrProvider } = await import("./resolve-XWYJYKKH.js");
       const { ocrPages } = await import("./provider-T2D5XRTI.js");
       const tryProvider = async (provider, filter) => {
         try {
@@ -5709,7 +5706,7 @@ async function parsePdfDocument(buffer, options) {
       if (options?.ocr) {
         ocrBlocks = await tryProvider(options.ocr, pageFilter);
       } else if (ocrMode === "auto") {
-        const { getAutoFallbackChain } = await import("./auto-detect-2YGFYQCN.js");
+        const { getAutoFallbackChain } = await import("./auto-detect-CBYICI6B.js");
         const pendingPages = /* @__PURE__ */ new Set();
         for (let i = 1; i <= effectivePageCount; i++) {
           if (!pageFilter || pageFilter.has(i)) pendingPages.add(i);
@@ -9828,13 +9825,36 @@ async function markdownToXlsx(markdown, options) {
 import { performance } from "perf_hooks";
 import libre from "libreoffice-convert";
 var libreConvert = libre.convert;
-var PROOFREAD_PROMPT = [
-  "Perform non-destructive proofreading only on the Markdown below.",
-  "Rules:",
-  "- Do not add, remove, or infer any facts",
-  "- Do not change numbers, units, or proper nouns",
-  "- Correct only typos, spacing, line breaks, and Markdown structure",
-  "- Output the corrected Markdown body only"
+var OCR_PROMPT = [
+  "\uC774 PDF \uD398\uC774\uC9C0 \uC774\uBBF8\uC9C0\uC5D0\uC11C \uD14D\uC2A4\uD2B8\uC640 \uD45C\uB97C \uCD94\uCD9C\uD558\uC5EC Markdown\uC73C\uB85C \uBCC0\uD658\uD558\uACE0, OCR \uC624\uC778\uC2DD \uC624\uB958\uB97C \uC989\uC2DC \uAD50\uC815\uD558\uC5EC \uCD5C\uC885 \uACB0\uACFC\uBB3C\uC744 \uCD9C\uB825\uD558\uC138\uC694.",
+  "",
+  "\uCD94\uCD9C \uADDC\uCE59:",
+  "- \uD14D\uC2A4\uD2B8, \uD45C, \uC81C\uBAA9, \uB9AC\uC2A4\uD2B8\uB97C \uC6D0\uBB38 \uAD6C\uC870 \uADF8\uB300\uB85C Markdown\uC73C\uB85C \uBCC0\uD658",
+  "- \uD45C\uB294 Markdown \uD14C\uC774\uBE14 \uBB38\uBC95 \uC0AC\uC6A9 (| \uAD6C\uBD84, |---|---| \uD5E4\uB354 \uAD6C\uBD84\uC120 \uD3EC\uD568)",
+  "- \uD5E4\uB529\uC740 \uC2DC\uAC01\uC801 \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",
+  "",
+  "\uC808\uB300 \uAE08\uC9C0 \uC0AC\uD56D:",
+  "- \uBB38\uC7A5\xB7\uB2E8\uB77D\xB7\uD56D\uBAA9\uC744 \uCD94\uAC00\uD558\uAC70\uB098 \uC0AD\uC81C\uD558\uC9C0 \uB9D0 \uAC83",
+  "- \uC22B\uC790, \uD37C\uC13C\uD2B8, \uB0A0\uC9DC, \uB2E8\uC704, \uAE08\uC561\uC744 \uC808\uB300 \uBCC0\uACBD\uD558\uC9C0 \uB9D0 \uAC83",
+  "- \uACE0\uC720\uBA85\uC0AC, \uAE30\uAD00\uBA85, \uBC95\uB839\uBA85, \uC9C0\uBA85\uC744 \uBCC0\uACBD\uD558\uC9C0 \uB9D0 \uAC83",
+  "- \uD45C\uC758 \uC81C\uBAA9\uC744 \uBCC0\uACBD \uB610\uB294 \uC0AD\uC81C\uD558\uC9C0 \uB9D0 \uAC83",
+  "- \uD45C\uC758 \uD589\xB7\uC5F4 \uC218, \uC140 \uB0B4\uC6A9, \uD5E4\uB354\uB97C \uBCC0\uACBD\uD558\uC9C0 \uB9D0 \uAC83",
+  "- \uC81C\uBAA9 \uC218\uC900(#, ##, ### \uB4F1)\uC744 \uC784\uC758\uB85C \uBC14\uAFB8\uC9C0 \uB9D0 \uAC83",
+  "- \uC6D0\uBB38\uC5D0 \uC5C6\uB294 \uB0B4\uC6A9\uC744 \uC694\uC57D\xB7\uBCF4\uC644\xB7\uCD94\uB860\uD558\uC9C0 \uB9D0 \uAC83",
+  "- ` ``` `\uB85C \uAC10\uC2F8\uAC70\uB098 \uC124\uBA85 \uD14D\uC2A4\uD2B8\uB97C \uCD94\uAC00\uD558\uC9C0 \uB9D0 \uAC83",
+  "",
+  "\uD5C8\uC6A9\uB418\uB294 \uAD50\uC815 \uBC94\uC704 (OCR \uC624\uC778\uC2DD \uC218\uC815):",
+  "- \uBA85\uBC31\uD55C \uAE00\uC790 \uC624\uC778\uC2DD \uC218\uC815 (\uC608: '0' \u2192 'O', 'l' \u2192 '1' \uB4F1 \uB9E5\uB77D\uC0C1 \uBA85\uD655\uD55C \uACBD\uC6B0\uB9CC)",
+  "- \uB2E8\uC5B4 \uC911\uAC04\uC5D0 \uC798\uBABB \uC0BD\uC785\uB41C \uACF5\uBC31 \uC81C\uAC70",
+  "- \uC904\uBC14\uAFC8 \uC624\uB958\uB85C \uBD84\uB9AC\uB41C \uBB38\uC7A5 \uBCD1\uD569 (\uC758\uBBF8 \uB2E8\uC704 \uAE30\uC900)",
+  "- Markdown \uBB38\uBC95 \uC624\uB958 \uC218\uC815 (\uD45C \uAD6C\uBD84\uC120 \uB204\uB77D, \uB9AC\uC2A4\uD2B8 \uB4E4\uC5EC\uC4F0\uAE30 \uB4F1)",
+  "",
+  "\uCD9C\uB825 \uADDC\uCE59:",
+  "- \uBCC0\uD658\uB41C Markdown \uBCF8\uBB38\uB9CC \uCD9C\uB825\uD560 \uAC83 (\uC124\uBA85, \uC8FC\uC11D, \uBA54\uD0C0 \uD14D\uC2A4\uD2B8 \uC5C6\uC774)",
+  "- \uD655\uC2E4\uD558\uC9C0 \uC54A\uC73C\uBA74 \uC6D0\uBB38\uC744 \uADF8\uB300\uB85C \uC720\uC9C0\uD560 \uAC83"
 ].join("\n");
 
 // src/index.ts
@@ -9908,9 +9928,6 @@ async function parseImage(buffer, options) {
     if (ocrMode === "gemini" || ocrMode === "claude" || ocrMode === "codex" || ocrMode === "ollama") {
       ocrProvider = createCliOcrProvider(ocrMode);
       actualOcrMode = ocrMode;
-    } else if (ocrMode === "tesseract") {
-      ocrProvider = await createTesseractProvider();
-      actualOcrMode = ocrMode;
     } else if (ocrMode === "auto") {
       const modesToTry = ["gemini", "claude", "codex", "ollama"];
       for (const mode of modesToTry) {
@@ -9922,10 +9939,6 @@ async function parseImage(buffer, options) {
           console.warn(`[kordoc] OCR auto-detection: ${mode} CLI not available or failed. Trying next.`, e);
         }
       }
-      if (!ocrProvider) {
-        ocrProvider = await createTesseractProvider();
-        actualOcrMode = "tesseract";
-      }
     }
     if (!ocrProvider) {
       return { success: false, fileType: "image", error: "\uC0AC\uC6A9 \uAC00\uB2A5\uD55C OCR \uD504\uB85C\uBC14\uC774\uB354\uB97C \uCC3E\uC744 \uC218 \uC5C6\uC2B5\uB2C8\uB2E4.", code: "PARSE_ERROR" };
@@ -10193,4 +10206,4 @@ export {
 cfb/cfb.js:
   (*! crc32.js (C) 2014-present SheetJS -- http://sheetjs.com *)
 */
-//# sourceMappingURL=chunk-WM3XI23V.js.map
+//# sourceMappingURL=chunk-463YQ2WL.js.map