kordoc 4.21.9 → 4.21.11

package/README.md

@@ -12,7 +12,7 @@ HWP 3.x·5.x, HWPX, HWPML, PDF, XLS·XLSX, DOCX, PPTX, PNG·JPG·WebP를 Markdow
 [English](README-EN.md) · [상세 사용법](docs/usage.md) · [벤치마크](docs/benchmarks.md) · [변경 이력](CHANGELOG.md)
 
 - 📊 **PDF 공개 벤치 종합 0.963** — opendataloader-bench 200문서. 공개 12개 파서 비교 1위(2026-09-29 기록). [측정 조건](docs/benchmarks.md)
-- 🇰🇷 **한국 공문서 표 구조 100%** — 원본 HWPX 2,286문서의 보이는 표 **9,865개 전부 구조 일치**. [4.18.8 검증](docs/release-4.18.8.json)
+- 🇰🇷 **한국 공문서 표 구조 100%** — 원본 HWPX 2,424문서의 보이는 표 **10,342개 전부 구조 일치**. [4.21.11 검증](docs/benchmarks.md)
 
 [![Kordoc 활용하기 — 영상 보기](./docs/video-demo.jpg)](https://youtu.be/Q13GmgDcIw0)
 
@@ -98,28 +98,32 @@ await writeFile("보고서.hwpx", Buffer.from(hwpx))
 
 ## 검증 결과
 
-**4.18.8 · 2026-10-02 배포 검증**. 고정 코퍼스와 참조 기준으로 측정한 결과입니다.
+**4.21.11 · 2026-10-10 배포 검증**. 고정 코퍼스와 참조 기준으로 측정한 결과입니다.
 
 | 대상 | 규모 | 결과 |
 | --- | --- | --- |
-| HWPX | 2,286문서 · 9,865표 | 표 구조 일치 **9,865/9,865 · 100%** |
-| HWP 5.x ↔ HWPX | 1,120쌍 · 4,258표 | 짝 문서의 표 구조 일치 **100%** |
-| PDF 글 | 744쌍 | 글자 재현율 **99.83%** · 읽기 순서 **99.15%** |
-| PDF 표 | 708쌍 · 2,331표 | 표 탐지 **99.83%** · 구조 일치 **97.94%** |
+| HWPX | 2,424문서 · 10,342표 | 표 구조 일치 **10,342/10,342 · 100%** |
+| HWP 5.x ↔ HWPX | 1,130쌍 · 4,315표 | 짝 문서의 표 구조 일치 **100%** |
+| PDF 글 | 744쌍 | 글자 재현율 **99.83%** · 정확도 **99.64%** · 읽기 순서 **99.16%** · 어절 F1 **98.86%** |
+| PDF 표 | 708쌍 · 2,331표 | 표 탐지 **99.83%** · 구조 일치 **97.98%** · 칸 F1 **0.991** |
+| PDF 전체 | 1,911문서 중 1,729건 채점 | 글 커버리지 **99.80%** |
+| OCR | 53문서 · 102쪽 | CER **0.0412** · 글자 재현율 **99.03%** |
 
-표 구조 점수와 셀 내용·화면 재현은 별도 지표입니다. 채점 범위·제외 기준·OCR 결과는 [벤치마크 상세](docs/benchmarks.md), 실제 게시 검증은 [4.18.8 기록](docs/release-4.18.8.json)에 있습니다.
+표 구조 점수와 셀 내용·화면 재현은 별도 지표입니다. 채점 범위·제외 기준·전 벤치 문서별 비교는 [벤치마크 상세](docs/benchmarks.md)에 있습니다.
 
-외부 PDF 벤치 **opendataloader-bench 200문서**는 4.21.5 별도 측정에서 기본값 종합 **0.963**, OCR 끔 **0.939**입니다. [옵션별 결과와 재현 방법](docs/benchmarks.md#pdf--markdown--opendataloader-bench)
+외부 PDF 벤치 **opendataloader-bench 200문서**는 4.21.11 배포 측정에서 기본값 종합 **0.963**, OCR 끔 **0.939**입니다. [옵션별 결과와 재현 방법](docs/benchmarks.md#pdf--markdown--opendataloader-bench)
 
 ## 최근 업데이트
 
-최신 배포: **[4.21.9](https://github.com/chrisryugj/kordoc/releases/tag/v4.21.9)** · 2026-10-10
+최신 배포: **[4.21.11](https://github.com/chrisryugj/kordoc/releases/tag/v4.21.11)** · 2026-10-10
 
 | 버전 | 주요 변경 |
 | --- | --- |
-| 4.21.9 | PDF 줄 꺾임 이음 — 낱말 가운데 꺾임·겹조사·계사·접미사·명사+하다, 본문 위 첨자 행 흡수 (문서별 하락 0) |
-| 4.21.8 | PDF 표 묶음 — 돌려 찍은 가로 표 세우기·데이터 표의 차트 오판·좁은 숫자 열·칸 걸친 한 조각·칸 안 쌓인 금액 (문서별 하락 0) |
-| 4.21.7 | PDF 토막 괘선 칸·글 조각 클립·연락처 표 — 예산서 상세 칸(OCR CER 0.0512 → 0.0445), ezPDF·MS Print 글 조각 클립이 만든 가짜 표, 연락처 6열 (문서별 하락 0) |
+| 4.21.11 | PDF 목차·라벨탭 상자·칸 첨자 — 목차 리더 걷기, 제목 칩 상자를 글로, 목차 칸 제목·쪽 번호 짝짓기, 서식 칸 붙은 첨자 흡수 (PDF 글 정확도 99.58 → 99.64%·OCR CER 0.0421 → 0.0412, 문서별 하락 0 — 열화 OCR 은 표본 쪽 셋 이동) |
+| 4.21.10 | PDF 흐름도 표·괄호 풀이 — 선 격자 표를 품은 클러스터 후보 거름, 짧은 괄호 풀이 앞 줄 꺾임 이음 (PDF 어절 F1 98.84 → 98.85%·OCR CER 0.0443 → 0.0421, 문서별 하락 0) |
+| 4.21.9 | PDF 줄 꺾임 이음 — 낱말 가운데 꺾임·겹조사·계사·접미사·명사+하다, 본문 위 첨자 행 흡수 (PDF 어절 F1 98.81 → 98.84%·글 커버리지 99.79 → 99.80%, 문서별 하락 0) |
+| 4.21.8 | PDF 표 묶음 — 돌려 찍은 가로 표 세우기·데이터 표의 차트 오판·좁은 숫자 열·칸 걸친 한 조각·칸 안 쌓인 금액 (OCR CER 0.0445 → 0.0443·열화 OCR scan 0.0821 → 0.0805, 문서별 하락 0) |
+| 4.21.7 | PDF 토막 괘선 칸·글 조각 클립·연락처 표 — 예산서 상세 칸(OCR CER 0.0512 → 0.0445), ezPDF·MS Print 글 조각 클립이 만든 가짜 표, 연락처 6열 (열화 OCR scan 0.0954 → 0.0821·글 커버리지 99.78 → 99.79%, 문서별 하락 0) |
 | 4.21.6 | 연락처 표 부서별 병합·여러 사람 칸 분할(PDF 표 구조 일치 97.94→97.98%), XLSX·XLS 200열 밖 칸 유실 수리, HWP3 포함 그림 추출, 수식 OCR LaTeX 명령 쪼개기·영역 상한 수리, PPTX 서식 트랙·코퍼스 보강(HWP3 75·서식 58·rhwp 230) |
 | 4.21.5 | PDF 품질 묶음 — OCR 두 단 가짜 표(#141 부작용)·그림 OCR 라벨 묶음·한 자리 눈금, 주소 링크·벡터 차트 값 축·단 끝줄·각주 순서 (ODL 기본 0.958 → 0.963, 문서별 하락 0) · 개조식·보고서 요약 상자가 md→hwpx→md→hwpx 왕복에서 유지 |
 | 4.21.4 | MCP `fill_form` 값 파일 입력 `fields_file`(값이 대화에 안 남고 응답은 글자 수만), 전 벤치 묶음 러너·문서별 비교기 `bench/suite.mjs` |

package/dist/{-QXIJTZUP.js → -VONBNRVD.js}

@@ -54,17 +54,17 @@ import {
   unknownFontWarnings,
   usesGaejosikMunche,
   validateHwpx
-} from "./chunk-EBYYQO6A.js";
+} from "./chunk-S7J7VI4R.js";
 import {
   blocksToPages
-} from "./chunk-6EKJ6J47.js";
+} from "./chunk-2LFYBDX2.js";
 import {
   applySplices,
   buildParagraphSplices,
   buildRangeSplices,
   placeSealHwpx,
   scanSectionXml
-} from "./chunk-HDPQ7QBF.js";
+} from "./chunk-OEF4V654.js";
 import {
   DEFAULT_REDACT_RULES,
   redactMarkdown,
@@ -72,7 +72,7 @@ import {
 } from "./chunk-3WSNLH4B.js";
 import {
   blocksToChunks
-} from "./chunk-FRM6WYNC.js";
+} from "./chunk-GLZ4XPKX.js";
 import {
   SPACE_EM_FIXED,
   SPACE_EM_FONT,
@@ -97,20 +97,20 @@ import {
   renderSceneToHtml,
   simulateWrap,
   simulateWrapKeepWord
-} from "./chunk-ZK5L7VPR.js";
+} from "./chunk-NI3RMQTS.js";
 import "./chunk-7FCNOKOR.js";
-import "./chunk-EXCGDQPB.js";
+import "./chunk-SWV27B56.js";
 import "./chunk-VQFA55IC.js";
-import "./chunk-B6WEXUMV.js";
+import "./chunk-NMXMBYRP.js";
 import {
   blocksToMarkdown,
   flattenLayoutTables,
   hasStructuredCellContent
-} from "./chunk-UK4UHB23.js";
-import "./chunk-LMICG6MM.js";
+} from "./chunk-ZTNZ4UZM.js";
+import "./chunk-JDEWFE46.js";
 import {
   VERSION
-} from "./chunk-O53OIPMY.js";
+} from "./chunk-L652BNX6.js";
 export {
   AGENCY_COMMON_NAME,
   BUILTIN_TEMPLATES,
@@ -204,4 +204,4 @@ export {
   usesGaejosikMunche,
   validateHwpx
 };
-//# sourceMappingURL=-QXIJTZUP.js.map
+//# sourceMappingURL=-VONBNRVD.js.map

package/dist/{chunk-6EKJ6J47.js → chunk-2LFYBDX2.js}

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   blocksToMarkdown
-} from "./chunk-UK4UHB23.js";
+} from "./chunk-ZTNZ4UZM.js";
 
 // src/page-markdown.ts
 function blocksToPages(blocks, render = blocksToMarkdown) {
@@ -29,4 +29,4 @@ function blocksToPages(blocks, render = blocksToMarkdown) {
 export {
   blocksToPages
 };
-//# sourceMappingURL=chunk-6EKJ6J47.js.map
+//# sourceMappingURL=chunk-2LFYBDX2.js.map

package/dist/{chunk-TAPXIXAD.js → chunk-4U5ISDBU.js}

@@ -1,12 +1,12 @@
 #!/usr/bin/env node
 import {
   parse
-} from "./chunk-EBYYQO6A.js";
+} from "./chunk-S7J7VI4R.js";
 import {
   classifyError,
   sanitizeError,
   toArrayBuffer
-} from "./chunk-O53OIPMY.js";
+} from "./chunk-L652BNX6.js";
 
 // src/cli/worker-parse.ts
 import { readFileSync, statSync } from "fs";
@@ -26,4 +26,4 @@ async function parseWorkerFile(absPath, options) {
 export {
   parseWorkerFile
 };
-//# sourceMappingURL=chunk-TAPXIXAD.js.map
+//# sourceMappingURL=chunk-4U5ISDBU.js.map

package/dist/{chunk-QACOMAEX.js → chunk-4UMTGSRH.js}

@@ -10,10 +10,10 @@ import {
   extractPageBlocksWithLines,
   getOcrModelsDir,
   parseCharacterDict
-} from "./chunk-MWH3WQ2C.js";
+} from "./chunk-PUAKSOZD.js";
 import {
   OPTIONAL_DEP_INSTALL_HINT
-} from "./chunk-HE5XFKH2.js";
+} from "./chunk-LZGPS3FH.js";
 
 // src/ocr/ruling-lines.ts
 var INK_LUMA_MAX = 205;
@@ -2130,4 +2130,4 @@ export {
   groupOcrLines,
   ocrItemsToBlocks
 };
-//# sourceMappingURL=chunk-QACOMAEX.js.map
+//# sourceMappingURL=chunk-4UMTGSRH.js.map

package/dist/{chunk-LORBG6MG.cjs → chunk-5JRL2VQ6.cjs}

@@ -1,7 +1,7 @@
 "use strict";Object.defineProperty(exports, "__esModule", {value: true}); function _nullishCoalesce(lhs, rhsFn) { if (lhs != null) { return lhs; } else { return rhsFn(); } } function _optionalChain(ops) { let lastAccessLHS = undefined; let value = ops[0]; let i = 1; while (i < ops.length) { const op = ops[i]; const fn = ops[i + 1]; i += 2; if ((op === 'optionalAccess' || op === 'optionalCall') && value == null) { return undefined; } if (op === 'access' || op === 'optionalAccess') { lastAccessLHS = value; value = fn(value); } else if (op === 'call' || op === 'optionalCall') { value = fn((...args) => value.call(lastAccessLHS, ...args)); lastAccessLHS = undefined; } } return value; }
 
 
-var _chunkR6BSCIZRcjs = require('./chunk-R6BSCIZR.cjs');
+var _chunkWSSCBJSUcjs = require('./chunk-WSSCBJSU.cjs');
 
 // src/script-tags.ts
 var LITERAL_TAG = /<(?=\/?(?:u|sup|sub)>)/g;
@@ -804,8 +804,10 @@ function blocksToMarkdown(blocks, options) {
       continue;
     }
     if (block.type === "list" && block.text) {
-      const listText = sanitizeText(block.text);
+      let listText = sanitizeText(block.text);
       if (!listText) continue;
+      const num = _optionalChain([/^(\d{1,3}[.)])\s/, 'access', _28 => _28.exec, 'call', _29 => _29(block.text.trim()), 'optionalAccess', _30 => _30[1]]);
+      if (num && listText.startsWith(num) && !/^\s/.test(listText.slice(num.length))) listText = `${num} ${listText.slice(num.length)}`;
       const alreadyNumbered = block.listType === "ordered" && /^\d+\.\s/.test(listText);
       const alreadyBulleted = block.listType !== "ordered" && /^-\s/.test(listText);
       const prefix = alreadyNumbered || alreadyBulleted ? "" : block.listType === "ordered" ? "1. " : "- ";
@@ -824,7 +826,7 @@ function blocksToMarkdown(blocks, options) {
       if (lines.length && lines[lines.length - 1] !== "") lines.push("");
       if (/^\[별표\s*\d+/.test(text)) {
         const nextBlock = blocks[i + 1];
-        if (_optionalChain([nextBlock, 'optionalAccess', _28 => _28.type]) === "paragraph" && nextBlock.text && /관련\)?$/.test(nextBlock.text)) {
+        if (_optionalChain([nextBlock, 'optionalAccess', _31 => _31.type]) === "paragraph" && nextBlock.text && /관련\)?$/.test(nextBlock.text)) {
           lines.push("", `## ${escapeGfm(text)} ${escapeGfm(nextBlock.text)}`, "");
           i++;
         } else {
@@ -837,10 +839,10 @@ function blocksToMarkdown(blocks, options) {
         continue;
       }
       const listIndent = block.listDepth ? "  ".repeat(block.listDepth) : "";
-      if (_optionalChain([block, 'access', _29 => _29.spans, 'optionalAccess', _30 => _30.length])) {
+      if (_optionalChain([block, 'access', _32 => _32.spans, 'optionalAccess', _33 => _33.length])) {
         let rendered = spansToMarkdown(block.spans);
         if (block.href) {
-          const href = _chunkR6BSCIZRcjs.sanitizeHref.call(void 0, block.href);
+          const href = _chunkWSSCBJSUcjs.sanitizeHref.call(void 0, block.href);
           if (href) rendered = `[${rendered}](${href})`;
         }
         if (block.footnoteText) rendered += ` (\uC8FC: ${block.footnoteText})`;
@@ -848,7 +850,7 @@ function blocksToMarkdown(blocks, options) {
         continue;
       }
       if (block.href) {
-        const href = _chunkR6BSCIZRcjs.sanitizeHref.call(void 0, block.href);
+        const href = _chunkWSSCBJSUcjs.sanitizeHref.call(void 0, block.href);
         if (href) text = `[${text}](${href})`;
       }
       if (block.footnoteText) {
@@ -870,7 +872,7 @@ function blocksToMarkdown(blocks, options) {
   return lines.join("\n").trim();
 }
 function captionToMarkdown(table, gfm = false) {
-  if (_optionalChain([table, 'access', _31 => _31.captionBlocks, 'optionalAccess', _32 => _32.length])) {
+  if (_optionalChain([table, 'access', _34 => _34.captionBlocks, 'optionalAccess', _35 => _35.length])) {
     return table.captionBlocks.flatMap((b) => {
       if (b.type === "table" && b.table) {
         const md = tableToMarkdown(b.table, gfm);
@@ -885,7 +887,7 @@ function captionToMarkdown(table, gfm = false) {
   return caption ? [`**${escapeGfm(caption)}**`, ""] : [];
 }
 function captionToHtml(table) {
-  if (_optionalChain([table, 'access', _33 => _33.captionBlocks, 'optionalAccess', _34 => _34.length])) return cellInnerHtml({ text: "", colSpan: 1, rowSpan: 1, blocks: table.captionBlocks });
+  if (_optionalChain([table, 'access', _36 => _36.captionBlocks, 'optionalAccess', _37 => _37.length])) return cellInnerHtml({ text: "", colSpan: 1, rowSpan: 1, blocks: table.captionBlocks });
   const cap = table.caption ? sanitizeText(table.caption) : "";
   return cap ? escapeHtmlCellText(cap).replace(/\n/g, "<br>") : "";
 }
@@ -900,7 +902,7 @@ function hasMergedCells(table) {
 function hasNestedTables(table) {
   for (const row of table.cells) {
     for (const cell of row) {
-      if (_optionalChain([cell, 'access', _35 => _35.blocks, 'optionalAccess', _36 => _36.some, 'call', _37 => _37((b) => b.type === "table" && b.table)])) return true;
+      if (_optionalChain([cell, 'access', _38 => _38.blocks, 'optionalAccess', _39 => _39.some, 'call', _40 => _40((b) => b.type === "table" && b.table)])) return true;
     }
   }
   return false;
@@ -908,7 +910,7 @@ function hasNestedTables(table) {
 function hasStructuredCellContent(table) {
   for (const row of table.cells) {
     for (const cell of row) {
-      if (_optionalChain([cell, 'access', _38 => _38.blocks, 'optionalAccess', _39 => _39.some, 'call', _40 => _40((b) => b.type === "table" && b.table || b.type === "separator")])) return true;
+      if (_optionalChain([cell, 'access', _41 => _41.blocks, 'optionalAccess', _42 => _42.some, 'call', _43 => _43((b) => b.type === "table" && b.table || b.type === "separator")])) return true;
     }
   }
   return false;
@@ -917,19 +919,19 @@ function noteSuffix(b) {
   return b.footnoteText ? ` (\uC8FC: ${b.footnoteText})` : "";
 }
 function visibleText(b) {
-  return _optionalChain([b, 'access', _41 => _41.spans, 'optionalAccess', _42 => _42.some, 'call', _43 => _43((s) => s.placeholder)]) ? b.spans.filter((s) => !s.placeholder).map((s) => s.text).join("") : _nullishCoalesce(b.text, () => ( ""));
+  return _optionalChain([b, 'access', _44 => _44.spans, 'optionalAccess', _45 => _45.some, 'call', _46 => _46((s) => s.placeholder)]) ? b.spans.filter((s) => !s.placeholder).map((s) => s.text).join("") : _nullishCoalesce(b.text, () => ( ""));
 }
 function escapeHtmlCellText(text) {
-  return _chunkR6BSCIZRcjs.escapeHtml.call(void 0, text).replace(/(\\)?&lt;(\/?)(u|sup|sub)&gt;/g, (_, esc, sl, tag) => esc ? `&lt;${sl}${tag}&gt;` : `<${sl}${tag}>`);
+  return _chunkWSSCBJSUcjs.escapeHtml.call(void 0, text).replace(/(\\)?&lt;(\/?)(u|sup|sub)&gt;/g, (_, esc, sl, tag) => esc ? `&lt;${sl}${tag}&gt;` : `<${sl}${tag}>`);
 }
 function cellInnerHtml(cell) {
-  if (_optionalChain([cell, 'access', _44 => _44.blocks, 'optionalAccess', _45 => _45.length])) {
+  if (_optionalChain([cell, 'access', _47 => _47.blocks, 'optionalAccess', _48 => _48.length])) {
     return cell.blocks.map((b) => {
       if (b.type === "table" && b.table) {
         const cap = captionToHtml(b.table);
         return (cap ? cap + "<br>" : "") + tableToHtml(b.table);
       }
-      if (b.type === "image" && b.text) return `<img src="${_chunkR6BSCIZRcjs.escapeHtml.call(void 0, b.text, true)}" alt="image">`;
+      if (b.type === "image" && b.text) return `<img src="${_chunkWSSCBJSUcjs.escapeHtml.call(void 0, b.text, true)}" alt="image">`;
       const t = (sanitizeText(visibleText(b)) + noteSuffix(b)).trim();
       return t ? escapeHtmlCellText(t).replace(/\n/g, "<br>") : "";
     }).filter(Boolean).join("<br>");
@@ -945,7 +947,7 @@ function tableToHtml(table) {
     const rowHtml = [];
     for (let c = 0; c < numCols; c++) {
       if (skip.has(`${r},${c}`)) continue;
-      const cell = _optionalChain([cells, 'access', _46 => _46[r], 'optionalAccess', _47 => _47[c]]);
+      const cell = _optionalChain([cells, 'access', _49 => _49[r], 'optionalAccess', _50 => _50[c]]);
       if (!cell) continue;
       for (let dr = 0; dr < cell.rowSpan; dr++) {
         for (let dc = 0; dc < cell.colSpan; dc++) {
@@ -966,10 +968,10 @@ function tableToHtml(table) {
   return lines.join("\n");
 }
 function hasInlineCellBlocks(cell) {
-  return !!_optionalChain([cell, 'access', _48 => _48.blocks, 'optionalAccess', _49 => _49.some, 'call', _50 => _50((b) => _optionalChain([b, 'access', _51 => _51.spans, 'optionalAccess', _52 => _52.length]) || b.footnoteText || b.type === "image" && b.text)]);
+  return !!_optionalChain([cell, 'access', _51 => _51.blocks, 'optionalAccess', _52 => _52.some, 'call', _53 => _53((b) => _optionalChain([b, 'access', _54 => _54.spans, 'optionalAccess', _55 => _55.length]) || b.footnoteText || b.type === "image" && b.text)]);
 }
 function cellBlockToMarkdown(b) {
-  return b.type === "image" && b.text ? `![image](${b.text})` : _optionalChain([b, 'access', _53 => _53.spans, 'optionalAccess', _54 => _54.length]) ? spansToMarkdown(b.spans) + escapeGfm(noteSuffix(b)) : escapeGfm(sanitizeText(_nullishCoalesce(b.text, () => ( ""))) + noteSuffix(b));
+  return b.type === "image" && b.text ? `![image](${b.text})` : _optionalChain([b, 'access', _56 => _56.spans, 'optionalAccess', _57 => _57.length]) ? spansToMarkdown(b.spans) + escapeGfm(noteSuffix(b)) : escapeGfm(sanitizeText(_nullishCoalesce(b.text, () => ( ""))) + noteSuffix(b));
 }
 function cellToMarkdown(cell, separator) {
   if (!hasInlineCellBlocks(cell)) return escapeGfm(sanitizeText(cell.text));
@@ -988,7 +990,7 @@ function tableToGfm(table) {
   return out.join("\n");
 }
 function gfmCellText(cell, parentId, nested, markerless) {
-  if (!_optionalChain([cell, 'access', _55 => _55.blocks, 'optionalAccess', _56 => _56.some, 'call', _57 => _57((b) => b.type === "table" && b.table || b.type === "separator")])) return cellToMarkdown(cell, "<br>");
+  if (!_optionalChain([cell, 'access', _58 => _58.blocks, 'optionalAccess', _59 => _59.some, 'call', _60 => _60((b) => b.type === "table" && b.table || b.type === "separator")])) return cellToMarkdown(cell, "<br>");
   let out = "";
   let text = "";
   let hasMarker = false;
@@ -1051,7 +1053,7 @@ function pipeTable(table, cellText, rowSpanFill) {
   for (let r = 0; r < numRows; r++) {
     for (let c = 0; c < numCols; c++) {
       if (skip.has(`${r},${c}`)) continue;
-      const cell = _optionalChain([cells, 'access', _58 => _58[r], 'optionalAccess', _59 => _59[c]]);
+      const cell = _optionalChain([cells, 'access', _61 => _61[r], 'optionalAccess', _62 => _62[c]]);
       if (!cell) continue;
       const cellMd = (text2) => text2.replace(/\r\n|\r|\n/g, "<br>").replace(/(?<!\\)\|/g, "\\|");
       const text = cellText(cell);
@@ -1089,13 +1091,14 @@ function pipeTable(table, cellText, rowSpanFill) {
 var CELL_EDGES = /* @__PURE__ */ new WeakMap();
 var CELL_PAGES = /* @__PURE__ */ new WeakMap();
 var CONTENT_CELLS = /* @__PURE__ */ new WeakSet();
+var CELL_BASELINES = /* @__PURE__ */ new WeakMap();
 var FRACTION_MAX_CHARS = 40;
 function anchorsOf(t) {
   const out = [];
   const covered = /* @__PURE__ */ new Set();
   for (let r = 0; r < t.rows; r++) for (let c = 0; c < t.cols; c++) {
     if (covered.has(r * t.cols + c)) continue;
-    const cell = _optionalChain([t, 'access', _60 => _60.cells, 'access', _61 => _61[r], 'optionalAccess', _62 => _62[c]]);
+    const cell = _optionalChain([t, 'access', _63 => _63.cells, 'access', _64 => _64[r], 'optionalAccess', _65 => _65[c]]);
     if (!cell) continue;
     const rs = Math.max(1, Math.min(cell.rowSpan, t.rows - r)), cs = Math.max(1, Math.min(cell.colSpan, t.cols - c));
     for (let dr = 0; dr < rs; dr++) for (let dc = 0; dc < cs; dc++) covered.add((r + dr) * t.cols + c + dc);
@@ -1124,7 +1127,7 @@ function toTex(s) {
 }
 function fractionPart(cell) {
   const t = cell.text.trim();
-  return !!t && !t.includes("\n") && !_optionalChain([cell, 'access', _63 => _63.blocks, 'optionalAccess', _64 => _64.length]) && t.replace(/<\/?(?:u|sup|sub)>/g, "").length <= FRACTION_MAX_CHARS && !/(?<!\\)\$/.test(t);
+  return !!t && !t.includes("\n") && !_optionalChain([cell, 'access', _66 => _66.blocks, 'optionalAccess', _67 => _67.length]) && t.replace(/<\/?(?:u|sup|sub)>/g, "").length <= FRACTION_MAX_CHARS && !/(?<!\\)\$/.test(t);
 }
 function mergeFractions(t, anchors, H, V) {
   const at = /* @__PURE__ */ new Map();
@@ -1159,7 +1162,7 @@ function bandTable(t, anchors, V, H, r0, r1, keepEmptyCols) {
   let list = anchors.filter((a) => a.r >= r0 && a.r <= r1).map((a) => ({ ...a, rs: Math.min(a.rs, r1 - a.r + 1) }));
   const blank = (a) => {
     if (CONTENT_CELLS.has(a.cell)) return false;
-    if (!_optionalChain([a, 'access', _65 => _65.cell, 'access', _66 => _66.blocks, 'optionalAccess', _67 => _67.length])) return !a.cell.text.trim();
+    if (!_optionalChain([a, 'access', _68 => _68.cell, 'access', _69 => _69.blocks, 'optionalAccess', _70 => _70.length])) return !a.cell.text.trim();
     return a.cell.blocks.every((b) => (b.type === "paragraph" || b.type === "heading") && !visibleText(b).trim() && !b.footnoteText);
   };
   let lo = Infinity, hi = -Infinity;
@@ -1210,25 +1213,47 @@ function bandTable(t, anchors, V, H, r0, r1, keepEmptyCols) {
   }
   return { rows, cols, cells, hasHeader: t.hasHeader, ...emptyBand ? { renderAsTable: true } : {} };
 }
+function tocLines(cells) {
+  if (cells.length !== 2) return null;
+  const [titles, pages] = cells.map((c) => c.text.trim().split("\n").map((l) => l.trim()));
+  const [ty, py] = cells.map((c) => CELL_BASELINES.get(c));
+  if (!ty || !py || ty.length !== titles.length || py.length !== pages.length || titles.length < 2 || !pages.every((l) => /^\d{1,4}$/.test(l)) || titles.some((l) => !l || /^\d{1,4}$/.test(l))) return null;
+  const lines = [...titles], taken = /* @__PURE__ */ new Set();
+  for (let j = 0; j < pages.length; j++) {
+    const i = ty.findIndex((y) => Math.abs(y - py[j]) <= 3);
+    if (i < 0 || taken.has(i)) return null;
+    taken.add(i);
+    lines[i] += " " + pages[j];
+  }
+  return lines;
+}
 function textRow(row, out, pageNumber, keepEmptyCols) {
   const parts = [];
+  const cells = [];
   const flush = () => {
     if (!parts.length) return;
-    if (parts.length === 1) {
+    const toc = tocLines(cells);
+    if (toc) {
+      for (const line of toc) out.push({ type: "paragraph", text: line, pageNumber });
+    } else if (parts.length === 1) {
       for (const line of parts[0].split("\n")) if (line.trim()) out.push({ type: "paragraph", text: line.trim(), pageNumber });
     } else {
       const text = parts.map((p) => p.replace(/\s*\n\s*/g, " ").trim()).filter(Boolean).join(" ");
       if (text) out.push({ type: "paragraph", text, pageNumber });
     }
     parts.length = 0;
+    cells.length = 0;
   };
   for (const a of row) {
-    if (_optionalChain([a, 'access', _68 => _68.cell, 'access', _69 => _69.blocks, 'optionalAccess', _70 => _70.length])) {
+    if (_optionalChain([a, 'access', _71 => _71.cell, 'access', _72 => _72.blocks, 'optionalAccess', _73 => _73.length])) {
       flush();
       out.push(...unframeLayoutTables(a.cell.blocks, keepEmptyCols));
       continue;
     }
-    if (a.cell.text.trim()) parts.push(a.cell.text.trim());
+    if (a.cell.text.trim()) {
+      parts.push(a.cell.text.trim());
+      cells.push(a.cell);
+    }
   }
   flush();
 }
@@ -1257,7 +1282,7 @@ function unframeTable(t, pageNumber, keepEmptyCols) {
     }, pageNumber }];
   }
   const out = [];
-  if (_optionalChain([t, 'access', _71 => _71.captionBlocks, 'optionalAccess', _72 => _72.length])) out.push(...t.captionBlocks);
+  if (_optionalChain([t, 'access', _74 => _74.captionBlocks, 'optionalAccess', _75 => _75.length])) out.push(...t.captionBlocks);
   else if (t.caption) out.push({ type: "paragraph", text: t.caption, pageNumber });
   const rowPage = (r) => {
     const marks = anchors.filter((a) => a.r === r).map((a) => CELL_PAGES.get(a.cell)).filter((p) => p !== void 0);
@@ -1284,12 +1309,12 @@ function unframeLayoutTables(blocks, keepEmptyCols = false) {
       out.push(b);
       continue;
     }
-    if (_optionalChain([b, 'access', _73 => _73.table, 'access', _74 => _74.captionBlocks, 'optionalAccess', _75 => _75.length])) b.table.captionBlocks = unframeLayoutTables(b.table.captionBlocks, keepEmptyCols);
+    if (_optionalChain([b, 'access', _76 => _76.table, 'access', _77 => _77.captionBlocks, 'optionalAccess', _78 => _78.length])) b.table.captionBlocks = unframeLayoutTables(b.table.captionBlocks, keepEmptyCols);
     const flat = unframeTable(b.table, b.pageNumber, keepEmptyCols);
     const kept = _nullishCoalesce(flat, () => ( [b]));
     for (const k of kept) {
       if (k.type === "table" && k.table) {
-        for (const row of k.table.cells) for (const cell of row) if (_optionalChain([cell, 'access', _76 => _76.blocks, 'optionalAccess', _77 => _77.length])) cell.blocks = unframeLayoutTables(cell.blocks, keepEmptyCols);
+        for (const row of k.table.cells) for (const cell of row) if (_optionalChain([cell, 'access', _79 => _79.blocks, 'optionalAccess', _80 => _80.length])) cell.blocks = unframeLayoutTables(cell.blocks, keepEmptyCols);
       }
       out.push(k);
     }
@@ -1332,5 +1357,6 @@ var HEADING_RATIO_H3 = 1.15;
 
 
 
-exports.escapeLiteralTags = escapeLiteralTags; exports.wrapScript = wrapScript; exports.tidyScriptTags = tidyScriptTags; exports.plainScripts = plainScripts; exports.stripScriptTags = stripScriptTags; exports.WINGDINGS = WINGDINGS; exports.mapPuaText = mapPuaText; exports.MAX_COLS = MAX_COLS; exports.MAX_ROWS = MAX_ROWS; exports.MAX_TABLE_CELLS = MAX_TABLE_CELLS; exports.buildTable = buildTable; exports.convertTableToText = convertTableToText; exports.escapeGfm = escapeGfm; exports.escapeLiteralDollar = escapeLiteralDollar; exports.markNonLayoutTable = markNonLayoutTable; exports.flattenLayoutTables = flattenLayoutTables; exports.dedupeRunningHeaders = dedupeRunningHeaders; exports.assignTableIds = assignTableIds; exports.blocksToMarkdown = blocksToMarkdown; exports.hasStructuredCellContent = hasStructuredCellContent; exports.noteSuffix = noteSuffix; exports.escapeHtmlCellText = escapeHtmlCellText; exports.CELL_EDGES = CELL_EDGES; exports.CELL_PAGES = CELL_PAGES; exports.CONTENT_CELLS = CONTENT_CELLS; exports.unframeLayoutTables = unframeLayoutTables; exports.HEADING_RATIO_H1 = HEADING_RATIO_H1; exports.HEADING_RATIO_H2 = HEADING_RATIO_H2; exports.HEADING_RATIO_H3 = HEADING_RATIO_H3;
-//# sourceMappingURL=chunk-LORBG6MG.cjs.map
+
+exports.escapeLiteralTags = escapeLiteralTags; exports.wrapScript = wrapScript; exports.tidyScriptTags = tidyScriptTags; exports.plainScripts = plainScripts; exports.stripScriptTags = stripScriptTags; exports.WINGDINGS = WINGDINGS; exports.mapPuaText = mapPuaText; exports.MAX_COLS = MAX_COLS; exports.MAX_ROWS = MAX_ROWS; exports.MAX_TABLE_CELLS = MAX_TABLE_CELLS; exports.buildTable = buildTable; exports.convertTableToText = convertTableToText; exports.escapeGfm = escapeGfm; exports.escapeLiteralDollar = escapeLiteralDollar; exports.markNonLayoutTable = markNonLayoutTable; exports.flattenLayoutTables = flattenLayoutTables; exports.dedupeRunningHeaders = dedupeRunningHeaders; exports.assignTableIds = assignTableIds; exports.blocksToMarkdown = blocksToMarkdown; exports.hasStructuredCellContent = hasStructuredCellContent; exports.noteSuffix = noteSuffix; exports.escapeHtmlCellText = escapeHtmlCellText; exports.CELL_EDGES = CELL_EDGES; exports.CELL_PAGES = CELL_PAGES; exports.CONTENT_CELLS = CONTENT_CELLS; exports.CELL_BASELINES = CELL_BASELINES; exports.unframeLayoutTables = unframeLayoutTables; exports.HEADING_RATIO_H1 = HEADING_RATIO_H1; exports.HEADING_RATIO_H2 = HEADING_RATIO_H2; exports.HEADING_RATIO_H3 = HEADING_RATIO_H3;
+//# sourceMappingURL=chunk-5JRL2VQ6.cjs.map