kordoc 4.19.1 → 4.19.2

package/README.md

@@ -111,10 +111,11 @@ await writeFile("보고서.hwpx", Buffer.from(hwpx))
 
 ## 최근 업데이트
 
-최신 배포: **[4.19.1](https://github.com/chrisryugj/kordoc/releases/tag/v4.19.1)** · 2026-10-07
+최신 배포: **[4.19.2](https://github.com/chrisryugj/kordoc/releases/tag/v4.19.2)** · 2026-10-07
 
 | 버전 | 주요 변경 |
 | --- | --- |
+| 4.19.2 | rhwp v0.8.7 파서 수정 반영 — HWP3 체크박스 □, HWP5 짝 없는 서로게이트 □, 손상 CFB 복구의 본문 경로 오인 수정, HWPX 압축 해제 상한을 엔트리 읽는 도중에 적용 |
 | 4.19.1 | PDF 쪽별 마크다운(`pages`)에서 쪽 넘김 표를 쪽마다 다시 갈라 뒤 쪽 행·상자가 앞 쪽에 붙거나 쪽 항목이 빠지지 않게 (#136) |
 | 4.19.0 | RAG 색인용 `tableFormat: "gfm"` — 병합·중첩 표도 HTML 없이 GFM 파이프 표로, 셀 안 표는 부모·자식 관계 표지와 함께 독립 표로 (#138) |
 | 4.18.12–13 | PDF 쪽 넘김 문단 잇기 판정을 HWPX·DOCX 원문 대조로 재조정 — 문장 끝·들여쓰기·나열 기호·탭 행 근거 (HWPX 쌍 오판 50 → 5) |

package/dist/{-LYUUVVQF.js → -3JIHF2AL.js}

@@ -51,17 +51,17 @@ import {
   unknownFontWarnings,
   usesGaejosikMunche,
   validateHwpx
-} from "./chunk-6KOQPPLX.js";
+} from "./chunk-NV46IWKF.js";
 import {
   blocksToPages
-} from "./chunk-X6ME7YYS.js";
+} from "./chunk-VO22VAR3.js";
 import {
   applySplices,
   buildParagraphSplices,
   buildRangeSplices,
   placeSealHwpx,
   scanSectionXml
-} from "./chunk-MSUGYK6A.js";
+} from "./chunk-PWQ2BVAT.js";
 import {
   DEFAULT_REDACT_RULES,
   redactMarkdown,
@@ -69,7 +69,7 @@ import {
 } from "./chunk-3WSNLH4B.js";
 import {
   blocksToChunks
-} from "./chunk-JNIJPP4B.js";
+} from "./chunk-TM4T4NYJ.js";
 import {
   SPACE_EM_FIXED,
   SPACE_EM_FONT,
@@ -94,20 +94,20 @@ import {
   renderSceneToHtml,
   simulateWrap,
   simulateWrapKeepWord
-} from "./chunk-5AZQEMMG.js";
+} from "./chunk-QZ63HXEG.js";
 import "./chunk-7FCNOKOR.js";
-import "./chunk-L5N5UQBP.js";
+import "./chunk-AOVMOCTK.js";
 import "./chunk-VQFA55IC.js";
-import "./chunk-UAAYIEIK.js";
+import "./chunk-QYJPU4FZ.js";
 import {
   blocksToMarkdown,
   flattenLayoutTables,
   hasStructuredCellContent
-} from "./chunk-TRUR4RUJ.js";
-import "./chunk-JA5B4NPF.js";
+} from "./chunk-B4R443MV.js";
+import "./chunk-6ZBORHFB.js";
 import {
   VERSION
-} from "./chunk-7L5GY54Q.js";
+} from "./chunk-OKHKEYTL.js";
 export {
   BUILTIN_TEMPLATES,
   DEFAULT_REDACT_RULES,
@@ -198,4 +198,4 @@ export {
   usesGaejosikMunche,
   validateHwpx
 };
-//# sourceMappingURL=-LYUUVVQF.js.map
+//# sourceMappingURL=-3JIHF2AL.js.map

package/dist/{chunk-KYFP5LDC.js → chunk-4CS7XE2Q.js}

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   KordocError
-} from "./chunk-7L5GY54Q.js";
+} from "./chunk-OKHKEYTL.js";
 
 // src/hwpx/profile-io.ts
 import { z } from "zod";
@@ -89,4 +89,4 @@ export {
   formatProfileSchema,
   parseFormatProfileJson
 };
-//# sourceMappingURL=chunk-KYFP5LDC.js.map
+//# sourceMappingURL=chunk-4CS7XE2Q.js.map

package/dist/{chunk-XFZMRCGJ.js → chunk-56ZWAREM.js}

@@ -3,7 +3,7 @@ import {
   ensureModelsIn,
   getModelStatusIn,
   getModelsDir
-} from "./chunk-BDTWXJNM.js";
+} from "./chunk-JK7N4R4K.js";
 
 // src/ocr/models.ts
 import { join } from "path";
@@ -93,4 +93,4 @@ export {
   parseCharacterDict,
   ocrModelPath
 };
-//# sourceMappingURL=chunk-XFZMRCGJ.js.map
+//# sourceMappingURL=chunk-56ZWAREM.js.map

package/dist/{chunk-FNX3ZAYB.cjs → chunk-5PGOZBEQ.cjs}

@@ -4,15 +4,15 @@
 
 
 
-var _chunkW5NXOMCDcjs = require('./chunk-W5NXOMCD.cjs');
+var _chunkORGWGR4Kcjs = require('./chunk-ORGWGR4K.cjs');
 
 
 
-var _chunkHMJXB7ZEcjs = require('./chunk-HMJXB7ZE.cjs');
+var _chunkWQAGDJIPcjs = require('./chunk-WQAGDJIP.cjs');
 
 
 
-var _chunkRHMIEEAZcjs = require('./chunk-RHMIEEAZ.cjs');
+var _chunkMZ6XNFZLcjs = require('./chunk-MZ6XNFZL.cjs');
 
 // src/ocr/models.ts
 var _path = require('path');
@@ -40,10 +40,10 @@ var OCR_REC_DICT = {
 };
 var ALL_OCR_MODELS = [OCR_DET_MODEL, OCR_REC_MODEL, OCR_REC_DICT];
 function getOcrModelsDir() {
-  return _chunkHMJXB7ZEcjs.getModelsDir.call(void 0, "ppocr");
+  return _chunkWQAGDJIPcjs.getModelsDir.call(void 0, "ppocr");
 }
 async function ensureOcrModels(onProgress) {
-  return _chunkHMJXB7ZEcjs.ensureModelsIn.call(void 0, getOcrModelsDir(), ALL_OCR_MODELS, onProgress);
+  return _chunkWQAGDJIPcjs.ensureModelsIn.call(void 0, getOcrModelsDir(), ALL_OCR_MODELS, onProgress);
 }
 async function ocrModelsCached() {
   const dir = getOcrModelsDir();
@@ -2091,7 +2091,7 @@ function tagScripts(text, lines) {
     const [s, e, kind] = marks[m];
     out = out.slice(0, s) + `<${kind}>` + out.slice(s, e) + `</${kind}>` + out.slice(e);
   }
-  return _chunkW5NXOMCDcjs.tidyScriptTags.call(void 0, out);
+  return _chunkORGWGR4Kcjs.tidyScriptTags.call(void 0, out);
 }
 
 // src/pdf/line-wrap.ts
@@ -3689,9 +3689,9 @@ function detectHeadings(blocks, medianFontSize) {
     if (/^\d+$/.test(text) && !isChapterNumber(block, blocks[bi + 1], medianFontSize)) continue;
     const ratio = block.style.fontSize / medianFontSize;
     let level = 0;
-    if (ratio >= _chunkW5NXOMCDcjs.HEADING_RATIO_H1) level = 1;
-    else if (ratio >= _chunkW5NXOMCDcjs.HEADING_RATIO_H2) level = 2;
-    else if (ratio >= _chunkW5NXOMCDcjs.HEADING_RATIO_H3) level = 3;
+    if (ratio >= _chunkORGWGR4Kcjs.HEADING_RATIO_H1) level = 1;
+    else if (ratio >= _chunkORGWGR4Kcjs.HEADING_RATIO_H2) level = 2;
+    else if (ratio >= _chunkORGWGR4Kcjs.HEADING_RATIO_H3) level = 3;
     if (level > 0) {
       block.type = "heading";
       block.level = level;
@@ -3704,7 +3704,7 @@ function isChapterNumber(block, next, medianFontSize) {
   const nextSize = _nullishCoalesce(_optionalChain([next, 'optionalAccess', _98 => _98.style, 'optionalAccess', _99 => _99.fontSize]), () => ( 0));
   const text = _nullishCoalesce(_optionalChain([next, 'optionalAccess', _100 => _100.text, 'optionalAccess', _101 => _101.trim, 'call', _102 => _102()]), () => ( ""));
   const a = block.bbox, b = _optionalChain([next, 'optionalAccess', _103 => _103.bbox]);
-  return /^\d{1,2}$/.test(_nullishCoalesce(_optionalChain([block, 'access', _104 => _104.text, 'optionalAccess', _105 => _105.trim, 'call', _106 => _106()]), () => ( ""))) && size >= medianFontSize * 2.5 && !!next && (next.type === "paragraph" || next.type === "heading") && next.pageNumber === block.pageNumber && nextSize >= medianFontSize * _chunkW5NXOMCDcjs.HEADING_RATIO_H1 && text.length > 0 && text.length <= 80 && !/^[\d\s.]+$/.test(text) && !!a && !!b && a.y > b.y && a.y - (b.y + b.height) <= size * 2;
+  return /^\d{1,2}$/.test(_nullishCoalesce(_optionalChain([block, 'access', _104 => _104.text, 'optionalAccess', _105 => _105.trim, 'call', _106 => _106()]), () => ( ""))) && size >= medianFontSize * 2.5 && !!next && (next.type === "paragraph" || next.type === "heading") && next.pageNumber === block.pageNumber && nextSize >= medianFontSize * _chunkORGWGR4Kcjs.HEADING_RATIO_H1 && text.length > 0 && text.length <= 80 && !/^[\d\s.]+$/.test(text) && !!a && !!b && a.y > b.y && a.y - (b.y + b.height) <= size * 2;
 }
 function mergeStackedHeadingLines(blocks, medianFontSize) {
   for (let i = 0; i < blocks.length - 1; ) {
@@ -6016,16 +6016,16 @@ 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 = _chunkRHMIEEAZcjs.safeMax.call(void 0, gaps);
+  const maxGap = _chunkMZ6XNFZLcjs.safeMax.call(void 0, gaps);
   const avgLen = items.reduce((s, i) => s + i.text.length, 0) / items.length;
   if (items.length >= 4 && maxGap < 40 && avgLen < 5) return true;
-  const minGap = _chunkRHMIEEAZcjs.safeMin.call(void 0, gaps);
+  const minGap = _chunkMZ6XNFZLcjs.safeMin.call(void 0, gaps);
   return minGap > 0 && maxGap - minGap <= Math.max(2, maxGap * 0.15) && items.filter((i) => /^\P{L}*\p{Ll}/u.test(i.text)).length * 2 >= items.length;
 }
 function detectColumns(yLines) {
   const allItems = yLines.flat();
   if (allItems.length === 0) return null;
-  const pageWidth = _chunkRHMIEEAZcjs.safeMax.call(void 0, allItems.map((i) => i.x + i.w)) - _chunkRHMIEEAZcjs.safeMin.call(void 0, allItems.map((i) => i.x));
+  const pageWidth = _chunkMZ6XNFZLcjs.safeMax.call(void 0, allItems.map((i) => i.x + i.w)) - _chunkMZ6XNFZLcjs.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++) {
@@ -6495,12 +6495,12 @@ function recordClipCellEdges(grids, horizontals, verticals, nonRules) {
 }
 function takeClipCellEdges(from, to) {
   const e = CLIP_CELL_EDGES.get(from);
-  if (e) _chunkW5NXOMCDcjs.CELL_EDGES.set(to, e);
+  if (e) _chunkORGWGR4Kcjs.CELL_EDGES.set(to, e);
 }
 function joinCellEdges(a, b) {
-  const u = _chunkW5NXOMCDcjs.CELL_EDGES.get(a), d = _chunkW5NXOMCDcjs.CELL_EDGES.get(b);
+  const u = _chunkORGWGR4Kcjs.CELL_EDGES.get(a), d = _chunkORGWGR4Kcjs.CELL_EDGES.get(b);
   if (!u || !d) return;
-  _chunkW5NXOMCDcjs.CELL_EDGES.set(a, { t: u.t, b: d.b, l: u.l || d.l, r: u.r || d.r });
+  _chunkORGWGR4Kcjs.CELL_EDGES.set(a, { t: u.t, b: d.b, l: u.l || d.l, r: u.r || d.r });
 }
 
 // src/pdf/table-unit-row.ts
@@ -6516,7 +6516,7 @@ function rebuildUnitLine(items, grid) {
 function prependUnitRow(cells, cols, line, usedItems) {
   const text = cleanCellText(line.map((item) => item.text).join(""));
   cells.unshift(Array.from({ length: cols }, (_, c) => ({ text: c === 0 ? text : "", colSpan: c === 0 ? cols : 1, rowSpan: 1 })));
-  for (const cell of cells[0]) _chunkW5NXOMCDcjs.CELL_EDGES.set(cell, NO_EDGES);
+  for (const cell of cells[0]) _chunkORGWGR4Kcjs.CELL_EDGES.set(cell, NO_EDGES);
   for (const item of line) usedItems.add(item);
 }
 function attachUnitRow(items, grid, cells, cols, usedItems) {
@@ -7046,7 +7046,7 @@ function recordFrameTitle(table, source, pageNum) {
   const [label, spacer, title] = cells;
   if (!/^(?:붙임|별첨|첨부|부록)\s*\d{1,3}$/.test(label.text.trim()) || spacer.text.trim() || !title.text.trim()) return;
   if (cells.some((c) => {
-    const e = _chunkW5NXOMCDcjs.CELL_EDGES.get(c);
+    const e = _chunkORGWGR4Kcjs.CELL_EDGES.get(c);
     return !e || e.t || e.b || e.l || e.r;
   })) return;
   const xs = TABLE_COLXS.get(table), style = dominantStyle(source), fs = _nullishCoalesce(_optionalChain([style, 'optionalAccess', _326 => _326.fontSize]), () => ( 0));
@@ -7958,7 +7958,7 @@ function extractBlocksWithGrids(items, pageNum, pageWidth, pageHeight, grids, ho
         groupSizes.push(1);
       }
       const allY = remaining.map((i) => i.y);
-      const pageH = _chunkRHMIEEAZcjs.safeMax.call(void 0, allY) - _chunkRHMIEEAZcjs.safeMin.call(void 0, allY);
+      const pageH = _chunkMZ6XNFZLcjs.safeMax.call(void 0, allY) - _chunkMZ6XNFZLcjs.safeMin.call(void 0, allY);
       const gapThreshold = Math.max(15, pageH * 0.03);
       const sides = frameColumnTextBands(remaining, spanning, gx);
       const textBlocks = [];
@@ -7974,7 +7974,7 @@ function extractBlocksWithGrids(items, pageNum, pageWidth, pageHeight, grids, ho
       finalTextBlocks.push(...detectListBlocks(textBlocks));
     } else {
       const allY = remaining.map((i) => i.y);
-      const pageH = _chunkRHMIEEAZcjs.safeMax.call(void 0, allY) - _chunkRHMIEEAZcjs.safeMin.call(void 0, allY);
+      const pageH = _chunkMZ6XNFZLcjs.safeMax.call(void 0, allY) - _chunkMZ6XNFZLcjs.safeMin.call(void 0, allY);
       const groups = xyCutOrder(remaining, Math.max(15, pageH * 0.03));
       const textBlocks = [];
       for (const group of groups) {
@@ -8299,7 +8299,7 @@ function extractPageBlocksFallback(items, pageNum, fullPage = false, detectTable
       blocks.push(...columnTextToBlocks(tableText, pageNum, bbox, dominantStyle(items)));
     } else {
       const allY = items.map((i) => i.y);
-      const pageHeight = _chunkRHMIEEAZcjs.safeMax.call(void 0, allY) - _chunkRHMIEEAZcjs.safeMin.call(void 0, allY);
+      const pageHeight = _chunkMZ6XNFZLcjs.safeMax.call(void 0, allY) - _chunkMZ6XNFZLcjs.safeMin.call(void 0, allY);
       const gapThreshold = Math.max(15, pageHeight * 0.03);
       const orderedGroups = proseCutX !== null ? splitTwoColumnProse(items, proseCutX) : xyCutOrder(items, gapThreshold);
       for (const group of orderedGroups) {
@@ -8378,4 +8378,4 @@ function extractPageBlocksFallback(items, pageNum, fullPage = false, detectTable
 
 
 exports.OCR_DET_MODEL = OCR_DET_MODEL; exports.OCR_REC_MODEL = OCR_REC_MODEL; exports.OCR_REC_DICT = OCR_REC_DICT; exports.getOcrModelsDir = getOcrModelsDir; exports.ensureOcrModels = ensureOcrModels; exports.ocrModelsCached = ocrModelsCached; exports.parseCharacterDict = parseCharacterDict; exports.extractLines = extractLines; exports.extractImageRegions = extractImageRegions; exports.filterHiddenText = filterHiddenText; exports.collapseEvenSpacing = collapseEvenSpacing; exports.computeBBox = computeBBox; exports.dominantStyle = dominantStyle; exports.normalizeItems = normalizeItems; exports.groupByY = groupByY; exports.mergeLineSimple = mergeLineSimple; exports.WrapLexicon = WrapLexicon; exports.startsNewItem = startsNewItem; exports.wrapJoiner = wrapJoiner; exports.joinPageBreakWraps = joinPageBreakWraps; exports.splitPageBreakWraps = splitPageBreakWraps; exports.latinSoftWrap = latinSoftWrap; exports.TOC_BLOCKS = TOC_BLOCKS; exports.computeMedianFontSizeFromFreq = computeMedianFontSizeFromFreq; exports.detectHeadings = detectHeadings; exports.mergeStackedHeadingLines = mergeStackedHeadingLines; exports.detectTypographyHeadings = detectTypographyHeadings; exports.detectDocumentStyleHeadings = detectDocumentStyleHeadings; exports.detectSiblingStyleHeadings = detectSiblingStyleHeadings; exports.detectRepeatedPageLabels = detectRepeatedPageLabels; exports.detectPageLeadHeadings = detectPageLeadHeadings; exports.refineDocumentStyleHeadings = refineDocumentStyleHeadings; exports.detectMarkerHeadings = detectMarkerHeadings; exports.detectTableCaptions = detectTableCaptions; exports.detectKoreanListBlocks = detectKoreanListBlocks; exports.removeHeaderFooterBlocks = removeHeaderFooterBlocks; exports.CLIP_TABLES = CLIP_TABLES; exports.FRAME_TITLE_BLOCKS = FRAME_TITLE_BLOCKS; exports.TABLE_COLXS = TABLE_COLXS; exports.TABLE_ROWYS = TABLE_ROWYS; exports.TABLE_TAIL = TABLE_TAIL; exports.PART_COLXS = PART_COLXS; exports.EMPTY_PARTS = EMPTY_PARTS; exports.IMAGE_CELLS = IMAGE_CELLS; exports.FILLER_CELLS = FILLER_CELLS; exports.CONT_PARTS = CONT_PARTS; exports.CELL_LINES = CELL_LINES; exports.ROW_RULES = ROW_RULES; exports.trimTrailingEmptyTableCols = trimTrailingEmptyTableCols; exports.NO_EDGES = NO_EDGES; exports.joinCellEdges = joinCellEdges; exports.FRAME_READING_UNITS = FRAME_READING_UNITS; exports.recordFrameReadingUnit = recordFrameReadingUnit; exports.restoreImageBullets = restoreImageBullets; exports.detectRightArrowRegions = detectRightArrowRegions; exports.removeSideTabs = removeSideTabs; exports.extractPageBlocksWithLines = extractPageBlocksWithLines;
-//# sourceMappingURL=chunk-FNX3ZAYB.cjs.map
+//# sourceMappingURL=chunk-5PGOZBEQ.cjs.map