kordoc 3.0.1 → 3.2.0

package/dist/mcp.js

@@ -8,7 +8,7 @@ import {
   fillHwpx,
   markdownToHwpx,
   parse
-} from "./chunk-X7VQVMXQ.js";
+} from "./chunk-CYLOXVYY.js";
 import {
   detectFormat,
   detectZipFormat
@@ -19,7 +19,7 @@ import {
   blocksToMarkdown,
   sanitizeError,
   toArrayBuffer
-} from "./chunk-O5P6EG5L.js";
+} from "./chunk-ZWVTR2HQ.js";
 import "./chunk-MOL7MDBG.js";
 
 // src/mcp.ts
@@ -191,7 +191,7 @@ server.tool(
           break;
         case "pdf":
           try {
-            const { extractPdfMetadataOnly } = await import("./parser-3N6FZSKU.js");
+            const { extractPdfMetadataOnly } = await import("./parser-U6KRTEB7.js");
             metadata = await extractPdfMetadataOnly(buffer);
           } catch {
             metadata = void 0;

package/dist/page-range-LNMXJU6W.js

@@ -0,0 +1,7 @@
+import {
+  parsePageRange
+} from "./chunk-GE43BE46.js";
+export {
+  parsePageRange
+};
+//# sourceMappingURL=page-range-LNMXJU6W.js.map

package/dist/page-range-P7SDW6LR.cjs

@@ -0,0 +1,8 @@
+"use strict";Object.defineProperty(exports, "__esModule", {value: true});
+
+var _chunkDCZVOIEOcjs = require('./chunk-DCZVOIEO.cjs');
+require('./chunk-GS7T56RP.cjs');
+
+
+exports.parsePageRange = _chunkDCZVOIEOcjs.parsePageRange;
+//# sourceMappingURL=page-range-P7SDW6LR.cjs.map

package/dist/page-range-P7SDW6LR.cjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["/Users/chris_gomdori/workspace/kordoc/dist/page-range-P7SDW6LR.cjs"],"names":[],"mappings":"AAAA;AACE;AACF,wDAA6B;AAC7B,gCAA6B;AAC7B;AACE;AACF,0DAAC","file":"/Users/chris_gomdori/workspace/kordoc/dist/page-range-P7SDW6LR.cjs"}

package/dist/{parser-LZH7ZELV.js → parser-HAWIRFS4.js}

@@ -6,10 +6,10 @@ import {
   blocksToMarkdown,
   safeMax,
   safeMin
-} from "./chunk-X3SCCO5Q.js";
+} from "./chunk-7KFBGOXZ.js";
 import {
   parsePageRange
-} from "./chunk-SBVRCJFH.js";
+} from "./chunk-GE43BE46.js";
 
 // src/pdf/line-detector.ts
 import { OPS } from "pdfjs-dist/legacy/build/pdf.mjs";
@@ -1581,7 +1581,7 @@ async function parsePdfDocument(buffer, options) {
     if (totalChars / Math.max(parsedPageCount, 1) < 10) {
       if (options?.ocr) {
         try {
-          const { ocrPages } = await import("./provider-2SEHU2FM.js");
+          const { ocrPages } = await import("./provider-4ZJKV3DC.js");
           const ocrBlocks = await ocrPages(doc, options.ocr, pageFilter, effectivePageCount);
           if (ocrBlocks.length > 0) {
             const ocrMarkdown = ocrBlocks.map((b) => b.text || "").filter(Boolean).join("\n\n");
@@ -3017,7 +3017,7 @@ function mergeKoreanLines(text) {
   return result.join("\n");
 }
 async function applyFormulaOcr(buffer, blocks, pageFilter, effectivePageCount, warnings, _onProgress) {
-  const formulaMod = await import("./formula-3AQUUIRF.js");
+  const formulaMod = await import("./formula-RXVSQPXI.js");
   const { FormulaPipeline, ensureFormulaModels } = formulaMod;
   await ensureFormulaModels((p) => {
     if (p.phase === "download" && p.total) {
@@ -3143,4 +3143,4 @@ export {
   parsePdfDocument,
   removeHeaderFooterBlocks
 };
-//# sourceMappingURL=parser-LZH7ZELV.js.map
+//# sourceMappingURL=parser-HAWIRFS4.js.map

package/dist/{parser-5FZJVLQL.cjs → parser-LKIUKZJV.cjs}

@@ -6,10 +6,11 @@
 
 
 
-var _chunkNBJB6TJBcjs = require('./chunk-NBJB6TJB.cjs');
+var _chunkSWEBVX5Kcjs = require('./chunk-SWEBVX5K.cjs');
 
 
-var _chunkMUOQXDZ4cjs = require('./chunk-MUOQXDZ4.cjs');
+var _chunkDCZVOIEOcjs = require('./chunk-DCZVOIEO.cjs');
+require('./chunk-GS7T56RP.cjs');
 
 // src/pdf/line-detector.ts
 var _pdfmjs = require('pdfjs-dist/legacy/build/pdf.mjs');
@@ -1495,7 +1496,7 @@ async function loadPdfWithTimeout(buffer) {
       new Promise((_, reject) => {
         timer = setTimeout(() => {
           loadingTask.destroy();
-          reject(new (0, _chunkNBJB6TJBcjs.KordocError)("PDF \uB85C\uB529 \uD0C0\uC784\uC544\uC6C3 (30\uCD08 \uCD08\uACFC)"));
+          reject(new (0, _chunkSWEBVX5Kcjs.KordocError)("PDF \uB85C\uB529 \uD0C0\uC784\uC544\uC6C3 (30\uCD08 \uCD08\uACFC)"));
         }, PDF_LOAD_TIMEOUT_MS);
       })
     ]);
@@ -1508,7 +1509,7 @@ async function parsePdfDocument(buffer, options) {
   const doc = await loadPdfWithTimeout(buffer);
   try {
     const pageCount = doc.numPages;
-    if (pageCount === 0) throw new (0, _chunkNBJB6TJBcjs.KordocError)("PDF\uC5D0 \uD398\uC774\uC9C0\uAC00 \uC5C6\uC2B5\uB2C8\uB2E4.");
+    if (pageCount === 0) throw new (0, _chunkSWEBVX5Kcjs.KordocError)("PDF\uC5D0 \uD398\uC774\uC9C0\uAC00 \uC5C6\uC2B5\uB2C8\uB2E4.");
     const metadata = { pageCount };
     await extractPdfMetadata(doc, metadata);
     const blocks = [];
@@ -1517,7 +1518,7 @@ async function parsePdfDocument(buffer, options) {
     let totalChars = 0;
     let totalTextBytes = 0;
     const effectivePageCount = Math.min(pageCount, MAX_PAGES);
-    const pageFilter = _optionalChain([options, 'optionalAccess', _11 => _11.pages]) ? _chunkMUOQXDZ4cjs.parsePageRange.call(void 0, options.pages, effectivePageCount) : null;
+    const pageFilter = _optionalChain([options, 'optionalAccess', _11 => _11.pages]) ? _chunkDCZVOIEOcjs.parsePageRange.call(void 0, options.pages, effectivePageCount) : null;
     const totalTarget = pageFilter ? pageFilter.size : effectivePageCount;
     const fontSizeFreq = /* @__PURE__ */ new Map();
     const pageHeights = /* @__PURE__ */ new Map();
@@ -1568,11 +1569,11 @@ async function parsePdfDocument(buffer, options) {
           pageText += pageText ? "\n" + t : t;
         }
         pageQuality.push(computePageQuality(i, pageText));
-        if (totalTextBytes > MAX_TOTAL_TEXT) throw new (0, _chunkNBJB6TJBcjs.KordocError)("\uD14D\uC2A4\uD2B8 \uCD94\uCD9C \uD06C\uAE30 \uCD08\uACFC");
+        if (totalTextBytes > MAX_TOTAL_TEXT) throw new (0, _chunkSWEBVX5Kcjs.KordocError)("\uD14D\uC2A4\uD2B8 \uCD94\uCD9C \uD06C\uAE30 \uCD08\uACFC");
         parsedPages++;
         _optionalChain([options, 'optionalAccess', _12 => _12.onProgress, 'optionalCall', _13 => _13(parsedPages, totalTarget)]);
       } catch (pageErr) {
-        if (pageErr instanceof _chunkNBJB6TJBcjs.KordocError) throw pageErr;
+        if (pageErr instanceof _chunkSWEBVX5Kcjs.KordocError) throw pageErr;
         warnings.push({ page: i, message: `\uD398\uC774\uC9C0 ${i} \uD30C\uC2F1 \uC2E4\uD328: ${pageErr instanceof Error ? pageErr.message : "\uC54C \uC218 \uC5C6\uB294 \uC624\uB958"}`, code: "PARTIAL_PARSE" });
       }
     }
@@ -1581,7 +1582,7 @@ async function parsePdfDocument(buffer, options) {
     if (totalChars / Math.max(parsedPageCount, 1) < 10) {
       if (_optionalChain([options, 'optionalAccess', _14 => _14.ocr])) {
         try {
-          const { ocrPages } = await Promise.resolve().then(() => _interopRequireWildcard(require("./provider-SNONEZNW.cjs")));
+          const { ocrPages } = await Promise.resolve().then(() => _interopRequireWildcard(require("./provider-G4C2V2PD.cjs")));
           const ocrBlocks = await ocrPages(doc, options.ocr, pageFilter, effectivePageCount);
           if (ocrBlocks.length > 0) {
             const ocrMarkdown = ocrBlocks.map((b) => b.text || "").filter(Boolean).join("\n\n");
@@ -1640,7 +1641,7 @@ async function parsePdfDocument(buffer, options) {
     detectKoreanListBlocks(blocks);
     const outline = blocks.filter((b) => b.type === "heading" && b.level && b.text).map((b) => ({ level: b.level, text: b.text, pageNumber: b.pageNumber }));
     sanitizeBlockControlChars(blocks);
-    let markdown = cleanPdfText(_chunkNBJB6TJBcjs.blocksToMarkdown.call(void 0, blocks));
+    let markdown = cleanPdfText(_chunkSWEBVX5Kcjs.blocksToMarkdown.call(void 0, blocks));
     return {
       markdown,
       blocks,
@@ -1728,9 +1729,9 @@ function detectHeadings(blocks, medianFontSize) {
     if (/^\d+$/.test(text)) continue;
     const ratio = block.style.fontSize / medianFontSize;
     let level = 0;
-    if (ratio >= _chunkNBJB6TJBcjs.HEADING_RATIO_H1) level = 1;
-    else if (ratio >= _chunkNBJB6TJBcjs.HEADING_RATIO_H2) level = 2;
-    else if (ratio >= _chunkNBJB6TJBcjs.HEADING_RATIO_H3) level = 3;
+    if (ratio >= _chunkSWEBVX5Kcjs.HEADING_RATIO_H1) level = 1;
+    else if (ratio >= _chunkSWEBVX5Kcjs.HEADING_RATIO_H2) level = 2;
+    else if (ratio >= _chunkSWEBVX5Kcjs.HEADING_RATIO_H3) level = 3;
     if (level > 0) {
       block.type = "heading";
       block.level = level;
@@ -2135,7 +2136,7 @@ function extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals) {
     }
     if (remaining.length > 0) {
       const allY = remaining.map((i) => i.y);
-      const pageH = _chunkNBJB6TJBcjs.safeMax.call(void 0, allY) - _chunkNBJB6TJBcjs.safeMin.call(void 0, allY);
+      const pageH = _chunkSWEBVX5Kcjs.safeMax.call(void 0, allY) - _chunkSWEBVX5Kcjs.safeMin.call(void 0, allY);
       const groups = xyCutOrder(remaining, Math.max(15, pageH * 0.03));
       const textBlocks = [];
       for (const group of groups) {
@@ -2264,7 +2265,7 @@ function extractPageBlocksFallback(items, pageNum) {
       blocks.push({ type: "paragraph", text: tableText, pageNumber: pageNum, bbox, style: dominantStyle(items) });
     } else {
       const allY = items.map((i) => i.y);
-      const pageHeight = _chunkNBJB6TJBcjs.safeMax.call(void 0, allY) - _chunkNBJB6TJBcjs.safeMin.call(void 0, allY);
+      const pageHeight = _chunkSWEBVX5Kcjs.safeMax.call(void 0, allY) - _chunkSWEBVX5Kcjs.safeMin.call(void 0, allY);
       const gapThreshold = Math.max(15, pageHeight * 0.03);
       const orderedGroups = xyCutOrder(items, gapThreshold);
       for (const group of orderedGroups) {
@@ -2440,14 +2441,14 @@ function isProseSpread(items) {
   for (let i = 1; i < sorted.length; i++) {
     gaps.push(sorted[i].x - (sorted[i - 1].x + sorted[i - 1].w));
   }
-  const maxGap = _chunkNBJB6TJBcjs.safeMax.call(void 0, gaps);
+  const maxGap = _chunkSWEBVX5Kcjs.safeMax.call(void 0, gaps);
   const avgLen = items.reduce((s, i) => s + i.text.length, 0) / items.length;
   return maxGap < 40 && avgLen < 5;
 }
 function detectColumns(yLines) {
   const allItems = yLines.flat();
   if (allItems.length === 0) return null;
-  const pageWidth = _chunkNBJB6TJBcjs.safeMax.call(void 0, allItems.map((i) => i.x + i.w)) - _chunkNBJB6TJBcjs.safeMin.call(void 0, allItems.map((i) => i.x));
+  const pageWidth = _chunkSWEBVX5Kcjs.safeMax.call(void 0, allItems.map((i) => i.x + i.w)) - _chunkSWEBVX5Kcjs.safeMin.call(void 0, allItems.map((i) => i.x));
   if (pageWidth < 100) return null;
   let bigoLineIdx = -1;
   for (let i = 0; i < yLines.length; i++) {
@@ -3017,7 +3018,7 @@ function mergeKoreanLines(text) {
   return result.join("\n");
 }
 async function applyFormulaOcr(buffer, blocks, pageFilter, effectivePageCount, warnings, _onProgress) {
-  const formulaMod = await Promise.resolve().then(() => _interopRequireWildcard(require("./formula-XGG6ZP42.cjs")));
+  const formulaMod = await Promise.resolve().then(() => _interopRequireWildcard(require("./formula-5NKVS2LR.cjs")));
   const { FormulaPipeline, ensureFormulaModels } = formulaMod;
   await ensureFormulaModels((p) => {
     if (p.phase === "download" && p.total) {
@@ -3143,4 +3144,4 @@ function formatMb(bytes) {
 
 
 exports.cleanPdfText = cleanPdfText; exports.detectKoreanListBlocks = detectKoreanListBlocks; exports.detectTableCaptions = detectTableCaptions; exports.extractPdfMetadataOnly = extractPdfMetadataOnly; exports.mergeCrossPageTables = mergeCrossPageTables; exports.parsePdfDocument = parsePdfDocument; exports.removeHeaderFooterBlocks = removeHeaderFooterBlocks;
-//# sourceMappingURL=parser-5FZJVLQL.cjs.map
+//# sourceMappingURL=parser-LKIUKZJV.cjs.map