kordoc 2.0.2 → 2.0.3

package/dist/chunk-25TXW6EP.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/utils.ts"],"sourcesContent":["/** kordoc 공용 유틸리티 */\r\n\r\n/** 빌드 타임에 tsup define으로 주입되는 버전 */\r\ndeclare const __KORDOC_VERSION__: string\r\nexport const VERSION: string = typeof __KORDOC_VERSION__ !== \"undefined\" ? __KORDOC_VERSION__ : \"0.0.0-dev\"\r\n\r\n/**\r\n * Node.js Buffer → ArrayBuffer 변환\r\n * pool Buffer의 공유 ArrayBuffer 문제를 안전하게 처리.\r\n * offset=0이고 전체 ArrayBuffer를 차지하면 복사 없이 직접 반환.\r\n */\r\nexport function toArrayBuffer(buf: Buffer): ArrayBuffer {\r\n  if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {\r\n    return buf.buffer as ArrayBuffer\r\n  }\r\n  return buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength) as ArrayBuffer\r\n}\r\n\r\n/**\r\n * kordoc 내부 에러 클래스 — 사용자에게 노출해도 안전한 메시지만 포함.\r\n * MCP 에러 정제에서 instanceof로 판별하여 allowlist 패턴 매칭 없이 안전하게 통과.\r\n */\r\nexport class KordocError extends Error {\r\n  constructor(message: string) {\r\n    super(message)\r\n    this.name = \"KordocError\"\r\n  }\r\n}\r\n\r\n/**\r\n * 에러 메시지 정제 — KordocError는 그대로, 나머지는 일반 메시지로 대체.\r\n * 파일시스템 경로, 스택 트레이스 등 내부 정보 노출 방지.\r\n */\r\nexport function sanitizeError(err: unknown): string {\r\n  if (err instanceof KordocError) return err.message\r\n  return \"문서 처리 중 오류가 발생했습니다\"\r\n}\r\n\r\n/**\r\n * ZIP 엔트리 경로의 경로 순회 여부 판별.\r\n * 백슬래시 정규화, .., 절대경로, Windows 드라이브 문자 모두 차단.\r\n */\r\nexport function isPathTraversal(name: string): boolean {\r\n  if (name.includes(\"\\x00\")) return true\r\n  const normalized = name.replace(/\\\\/g, \"/\")\r\n  return normalized.includes(\"..\") || normalized.startsWith(\"/\") || /^[A-Za-z]:/.test(normalized)\r\n}\r\n\r\n// ─── ZIP 안전 로딩 (ZIP bomb 방지) ────────────────────\r\n\r\n/**\r\n * ZIP bomb 사전 검사 — Central Directory에서 비압축 합계와 엔트리 수 확인.\r\n * HWPX/XLSX/DOCX 등 모든 ZIP 기반 포맷에서 공통 사용.\r\n */\r\nexport function precheckZipSize(\r\n  buffer: ArrayBuffer,\r\n  maxUncompressedSize = 100 * 1024 * 1024,\r\n  maxEntries = 500,\r\n): { totalUncompressed: number; entryCount: number } {\r\n  try {\r\n    const data = new DataView(buffer)\r\n    const len = buffer.byteLength\r\n    // EOCD 시그니처 역방향 스캔\r\n    let eocdOffset = -1\r\n    for (let i = len - 22; i >= Math.max(0, len - 65557); i--) {\r\n      if (data.getUint32(i, true) === 0x06054b50) { eocdOffset = i; break }\r\n    }\r\n    if (eocdOffset < 0) return { totalUncompressed: 0, entryCount: 0 }\r\n\r\n    const entryCount = data.getUint16(eocdOffset + 10, true)\r\n    if (entryCount > maxEntries) {\r\n      throw new KordocError(`ZIP 엔트리 수 초과: ${entryCount} (최대 ${maxEntries})`)\r\n    }\r\n\r\n    const cdSize = data.getUint32(eocdOffset + 12, true)\r\n    const cdOffset = data.getUint32(eocdOffset + 16, true)\r\n    if (cdOffset + cdSize > len) return { totalUncompressed: 0, entryCount }\r\n\r\n    let totalUncompressed = 0\r\n    let pos = cdOffset\r\n    for (let i = 0; i < entryCount && pos + 46 <= cdOffset + cdSize; i++) {\r\n      if (data.getUint32(pos, true) !== 0x02014b50) break\r\n      totalUncompressed += data.getUint32(pos + 24, true)\r\n      const nameLen = data.getUint16(pos + 28, true)\r\n      const extraLen = data.getUint16(pos + 30, true)\r\n      const commentLen = data.getUint16(pos + 32, true)\r\n      pos += 46 + nameLen + extraLen + commentLen\r\n    }\r\n\r\n    if (totalUncompressed > maxUncompressedSize) {\r\n      throw new KordocError(`ZIP 비압축 크기 초과: ${(totalUncompressed / 1024 / 1024).toFixed(1)}MB (최대 ${maxUncompressedSize / 1024 / 1024}MB)`)\r\n    }\r\n\r\n    return { totalUncompressed, entryCount }\r\n  } catch (err) {\r\n    if (err instanceof KordocError) throw err\r\n    return { totalUncompressed: 0, entryCount: 0 }\r\n  }\r\n}\r\n\r\n/** 하이퍼링크 URL 살균 — javascript: 등 XSS 위험 스킴 차단 */\r\nconst SAFE_HREF_RE = /^(?:https?:|mailto:|tel:|#)/i\r\nexport function sanitizeHref(href: string): string | null {\r\n  const trimmed = href.trim()\r\n  if (!trimmed || !SAFE_HREF_RE.test(trimmed)) return null\r\n  return trimmed\r\n}\r\n\r\n// ─── 에러 분류 ──────────────────────────────────────\r\n\r\nimport type { ErrorCode } from \"./types.js\"\r\n\r\n/** 에러를 구조화된 ErrorCode로 분류 — KordocError 메시지 패턴 매칭 */\r\nexport function classifyError(err: unknown): ErrorCode {\r\n  if (!(err instanceof Error)) return \"PARSE_ERROR\"\r\n  const msg = err.message\r\n  if (msg.includes(\"암호화\")) return \"ENCRYPTED\"\r\n  if (msg.includes(\"DRM\")) return \"DRM_PROTECTED\"\r\n  if (msg.includes(\"ZIP bomb\") || msg.includes(\"ZIP 비압축 크기 초과\") || msg.includes(\"ZIP 엔트리 수 초과\")) return \"ZIP_BOMB\"\r\n  if (msg.includes(\"bomb\") || msg.includes(\"크기 초과\") || msg.includes(\"압축 해제\")) return \"DECOMPRESSION_BOMB\"\r\n  if (msg.includes(\"이미지 기반\")) return \"IMAGE_BASED_PDF\"\r\n  if (msg.includes(\"섹션\") && (msg.includes(\"찾을 수 없\") || msg.includes(\"없음\"))) return \"NO_SECTIONS\"\r\n  if (msg.includes(\"시그니처\") || msg.includes(\"복구할 수 없\")) return \"CORRUPTED\"\r\n  return \"PARSE_ERROR\"\r\n}\r\n"],"mappings":";;;AAIO,IAAM,UAAkB,OAA4C,UAAqB;AAOzF,SAAS,cAAc,KAA0B;AACtD,MAAI,IAAI,eAAe,KAAK,IAAI,eAAe,IAAI,OAAO,YAAY;AACpE,WAAO,IAAI;AAAA,EACb;AACA,SAAO,IAAI,OAAO,MAAM,IAAI,YAAY,IAAI,aAAa,IAAI,UAAU;AACzE;AAMO,IAAM,cAAN,cAA0B,MAAM;AAAA,EACrC,YAAY,SAAiB;AAC3B,UAAM,OAAO;AACb,SAAK,OAAO;AAAA,EACd;AACF;AAMO,SAAS,cAAc,KAAsB;AAClD,MAAI,eAAe,YAAa,QAAO,IAAI;AAC3C,SAAO;AACT;AAMO,SAAS,gBAAgB,MAAuB;AACrD,MAAI,KAAK,SAAS,IAAM,EAAG,QAAO;AAClC,QAAM,aAAa,KAAK,QAAQ,OAAO,GAAG;AAC1C,SAAO,WAAW,SAAS,IAAI,KAAK,WAAW,WAAW,GAAG,KAAK,aAAa,KAAK,UAAU;AAChG;AAQO,SAAS,gBACd,QACA,sBAAsB,MAAM,OAAO,MACnC,aAAa,KACsC;AACnD,MAAI;AACF,UAAM,OAAO,IAAI,SAAS,MAAM;AAChC,UAAM,MAAM,OAAO;AAEnB,QAAI,aAAa;AACjB,aAAS,IAAI,MAAM,IAAI,KAAK,KAAK,IAAI,GAAG,MAAM,KAAK,GAAG,KAAK;AACzD,UAAI,KAAK,UAAU,GAAG,IAAI,MAAM,WAAY;AAAE,qBAAa;AAAG;AAAA,MAAM;AAAA,IACtE;AACA,QAAI,aAAa,EAAG,QAAO,EAAE,mBAAmB,GAAG,YAAY,EAAE;AAEjE,UAAM,aAAa,KAAK,UAAU,aAAa,IAAI,IAAI;AACvD,QAAI,aAAa,YAAY;AAC3B,YAAM,IAAI,YAAY,+CAAiB,UAAU,kBAAQ,UAAU,GAAG;AAAA,IACxE;AAEA,UAAM,SAAS,KAAK,UAAU,aAAa,IAAI,IAAI;AACnD,UAAM,WAAW,KAAK,UAAU,aAAa,IAAI,IAAI;AACrD,QAAI,WAAW,SAAS,IAAK,QAAO,EAAE,mBAAmB,GAAG,WAAW;AAEvE,QAAI,oBAAoB;AACxB,QAAI,MAAM;AACV,aAAS,IAAI,GAAG,IAAI,cAAc,MAAM,MAAM,WAAW,QAAQ,KAAK;AACpE,UAAI,KAAK,UAAU,KAAK,IAAI,MAAM,SAAY;AAC9C,2BAAqB,KAAK,UAAU,MAAM,IAAI,IAAI;AAClD,YAAM,UAAU,KAAK,UAAU,MAAM,IAAI,IAAI;AAC7C,YAAM,WAAW,KAAK,UAAU,MAAM,IAAI,IAAI;AAC9C,YAAM,aAAa,KAAK,UAAU,MAAM,IAAI,IAAI;AAChD,aAAO,KAAK,UAAU,WAAW;AAAA,IACnC;AAEA,QAAI,oBAAoB,qBAAqB;AAC3C,YAAM,IAAI,YAAY,sDAAmB,oBAAoB,OAAO,MAAM,QAAQ,CAAC,CAAC,oBAAU,sBAAsB,OAAO,IAAI,KAAK;AAAA,IACtI;AAEA,WAAO,EAAE,mBAAmB,WAAW;AAAA,EACzC,SAAS,KAAK;AACZ,QAAI,eAAe,YAAa,OAAM;AACtC,WAAO,EAAE,mBAAmB,GAAG,YAAY,EAAE;AAAA,EAC/C;AACF;AAGA,IAAM,eAAe;AACd,SAAS,aAAa,MAA6B;AACxD,QAAM,UAAU,KAAK,KAAK;AAC1B,MAAI,CAAC,WAAW,CAAC,aAAa,KAAK,OAAO,EAAG,QAAO;AACpD,SAAO;AACT;AAOO,SAAS,cAAc,KAAyB;AACrD,MAAI,EAAE,eAAe,OAAQ,QAAO;AACpC,QAAM,MAAM,IAAI;AAChB,MAAI,IAAI,SAAS,oBAAK,EAAG,QAAO;AAChC,MAAI,IAAI,SAAS,KAAK,EAAG,QAAO;AAChC,MAAI,IAAI,SAAS,UAAU,KAAK,IAAI,SAAS,kDAAe,KAAK,IAAI,SAAS,4CAAc,EAAG,QAAO;AACtG,MAAI,IAAI,SAAS,MAAM,KAAK,IAAI,SAAS,2BAAO,KAAK,IAAI,SAAS,2BAAO,EAAG,QAAO;AACnF,MAAI,IAAI,SAAS,iCAAQ,EAAG,QAAO;AACnC,MAAI,IAAI,SAAS,cAAI,MAAM,IAAI,SAAS,4BAAQ,KAAK,IAAI,SAAS,cAAI,GAAI,QAAO;AACjF,MAAI,IAAI,SAAS,0BAAM,KAAK,IAAI,SAAS,kCAAS,EAAG,QAAO;AAC5D,SAAO;AACT;","names":[]}

package/dist/{chunk-MOL7MDBG.js → chunk-3TBUDJDE.js}

@@ -32,4 +32,4 @@ function parsePageRange(spec, maxPages) {
 export {
   parsePageRange
 };
-//# sourceMappingURL=chunk-MOL7MDBG.js.map
+//# sourceMappingURL=chunk-3TBUDJDE.js.map

package/dist/chunk-3TBUDJDE.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/page-range.ts"],"sourcesContent":["/** 페이지/섹션 범위 파싱 유틸리티 */\r\n\r\n/**\r\n * 페이지 범위 지정을 1-based Set<number>로 변환.\r\n *\r\n * @param spec - [1,2,3] 또는 \"1-3\" 또는 \"1,3,5-7\"\r\n * @param maxPages - 최대 페이지 수 (클램핑 상한)\r\n * @returns 1-based 페이지 번호 Set\r\n */\r\nexport function parsePageRange(spec: number[] | string, maxPages: number): Set<number> {\r\n  const result = new Set<number>()\r\n  if (maxPages <= 0) return result\r\n\r\n  if (Array.isArray(spec)) {\r\n    for (const n of spec) {\r\n      const page = Math.round(n)\r\n      if (page >= 1 && page <= maxPages) result.add(page)\r\n    }\r\n    return result\r\n  }\r\n\r\n  if (typeof spec !== \"string\" || spec.trim() === \"\") return result\r\n\r\n  const parts = spec.split(\",\")\r\n  for (const part of parts) {\r\n    const trimmed = part.trim()\r\n    if (!trimmed) continue\r\n\r\n    const rangeMatch = trimmed.match(/^(\\d+)\\s*-\\s*(\\d+)$/)\r\n    if (rangeMatch) {\r\n      const start = Math.max(1, parseInt(rangeMatch[1], 10))\r\n      const end = Math.min(maxPages, parseInt(rangeMatch[2], 10))\r\n      for (let i = start; i <= end; i++) result.add(i)\r\n    } else {\r\n      const page = parseInt(trimmed, 10)\r\n      if (!isNaN(page) && page >= 1 && page <= maxPages) result.add(page)\r\n    }\r\n  }\r\n\r\n  return result\r\n}\r\n"],"mappings":";;;AASO,SAAS,eAAe,MAAyB,UAA+B;AACrF,QAAM,SAAS,oBAAI,IAAY;AAC/B,MAAI,YAAY,EAAG,QAAO;AAE1B,MAAI,MAAM,QAAQ,IAAI,GAAG;AACvB,eAAW,KAAK,MAAM;AACpB,YAAM,OAAO,KAAK,MAAM,CAAC;AACzB,UAAI,QAAQ,KAAK,QAAQ,SAAU,QAAO,IAAI,IAAI;AAAA,IACpD;AACA,WAAO;AAAA,EACT;AAEA,MAAI,OAAO,SAAS,YAAY,KAAK,KAAK,MAAM,GAAI,QAAO;AAE3D,QAAM,QAAQ,KAAK,MAAM,GAAG;AAC5B,aAAW,QAAQ,OAAO;AACxB,UAAM,UAAU,KAAK,KAAK;AAC1B,QAAI,CAAC,QAAS;AAEd,UAAM,aAAa,QAAQ,MAAM,qBAAqB;AACtD,QAAI,YAAY;AACd,YAAM,QAAQ,KAAK,IAAI,GAAG,SAAS,WAAW,CAAC,GAAG,EAAE,CAAC;AACrD,YAAM,MAAM,KAAK,IAAI,UAAU,SAAS,WAAW,CAAC,GAAG,EAAE,CAAC;AAC1D,eAAS,IAAI,OAAO,KAAK,KAAK,IAAK,QAAO,IAAI,CAAC;AAAA,IACjD,OAAO;AACL,YAAM,OAAO,SAAS,SAAS,EAAE;AACjC,UAAI,CAAC,MAAM,IAAI,KAAK,QAAQ,KAAK,QAAQ,SAAU,QAAO,IAAI,IAAI;AAAA,IACpE;AAAA,EACF;AAEA,SAAO;AACT;","names":[]}

package/dist/{chunk-XJYM2AUA.js → chunk-4UH6ABAY.js}

@@ -6,10 +6,10 @@ import {
   precheckZipSize,
   sanitizeHref,
   toArrayBuffer
-} from "./chunk-EVWOJ4T5.js";
+} from "./chunk-25TXW6EP.js";
 import {
   parsePageRange
-} from "./chunk-MOL7MDBG.js";
+} from "./chunk-3TBUDJDE.js";
 
 // src/detect.ts
 import JSZip from "jszip";
@@ -163,6 +163,47 @@ function sanitizeText(text) {
   }
   return result;
 }
+function flattenLayoutTables(blocks) {
+  const result = [];
+  for (const block of blocks) {
+    if (block.type !== "table" || !block.table) {
+      result.push(block);
+      continue;
+    }
+    const { rows: numRows, cols: numCols, cells } = block.table;
+    if (numRows === 1 && numCols === 1) {
+      result.push(block);
+      continue;
+    }
+    if (numRows <= 3) {
+      let totalNewlines = 0;
+      let totalTextLen = 0;
+      for (let r = 0; r < numRows; r++) {
+        for (let c = 0; c < numCols; c++) {
+          const t = cells[r]?.[c]?.text || "";
+          totalNewlines += (t.match(/\n/g) || []).length;
+          totalTextLen += t.length;
+        }
+      }
+      if (totalNewlines > 5 || numRows <= 2 && totalTextLen > 300) {
+        for (let r = 0; r < numRows; r++) {
+          for (let c = 0; c < numCols; c++) {
+            const cellText = cells[r]?.[c]?.text?.trim();
+            if (!cellText) continue;
+            for (const line of cellText.split("\n")) {
+              const trimmed = line.trim();
+              if (!trimmed) continue;
+              result.push({ type: "paragraph", text: trimmed, pageNumber: block.pageNumber });
+            }
+          }
+        }
+        continue;
+      }
+    }
+    result.push(block);
+  }
+  return result;
+}
 function blocksToMarkdown(blocks) {
   const lines = [];
   for (let i = 0; i < blocks.length; i++) {
@@ -1078,8 +1119,9 @@ var TAG_CHAR_SHAPE = 68;
 var TAG_CTRL_HEADER = 71;
 var TAG_LIST_HEADER = 72;
 var TAG_TABLE = 77;
-var TAG_DOC_CHAR_SHAPE = 55;
-var TAG_DOC_STYLE = 58;
+var TAG_DOC_CHAR_SHAPE = 21;
+var TAG_DOC_PARA_SHAPE = 25;
+var TAG_DOC_STYLE = 26;
 var CHAR_LINE = 0;
 var CHAR_SECTION_BREAK = 10;
 var CHAR_PARA = 13;
@@ -1135,8 +1177,14 @@ function parseFileHeader(data) {
 }
 function parseDocInfo(records) {
   const charShapes = [];
+  const paraShapes = [];
   const styles = [];
   for (const rec of records) {
+    if (rec.tagId === TAG_DOC_PARA_SHAPE && rec.data.length >= 4) {
+      const flags = rec.data.readUInt32LE(0);
+      const outlineLevel = flags >> 25 & 7;
+      paraShapes.push({ outlineLevel });
+    }
     if (rec.tagId === TAG_DOC_CHAR_SHAPE && rec.data.length >= 18) {
       if (rec.data.length >= 50) {
         const fontSize = rec.data.readUInt32LE(42);
@@ -1176,7 +1224,7 @@ function parseDocInfo(records) {
       }
     }
   }
-  return { charShapes, styles };
+  return { charShapes, paraShapes, styles };
 }
 function extractText(data) {
   let result = "";
@@ -2186,12 +2234,13 @@ function parseHwp5Document(buffer, options) {
     }
   }
   const images = cfb ? extractHwp5Images(cfb, blocks, compressed, warnings) : extractHwp5ImagesLenient(lenientCfb, blocks, compressed, warnings);
+  const flatBlocks = flattenLayoutTables(blocks);
   if (docInfo) {
-    detectHwp5Headings(blocks, docInfo);
+    detectHwp5Headings(flatBlocks, docInfo);
   }
-  const outline = blocks.filter((b) => b.type === "heading" && b.level && b.text).map((b) => ({ level: b.level, text: b.text, pageNumber: b.pageNumber }));
-  const markdown = blocksToMarkdown(blocks);
-  return { markdown, blocks, metadata, outline: outline.length > 0 ? outline : void 0, warnings: warnings.length > 0 ? warnings : void 0, images: images.length > 0 ? images : void 0 };
+  const outline = flatBlocks.filter((b) => b.type === "heading" && b.level && b.text).map((b) => ({ level: b.level, text: b.text, pageNumber: b.pageNumber }));
+  const markdown = blocksToMarkdown(flatBlocks);
+  return { markdown, blocks: flatBlocks, metadata, outline: outline.length > 0 ? outline : void 0, warnings: warnings.length > 0 ? warnings : void 0, images: images.length > 0 ? images : void 0 };
 }
 function parseDocInfoStream(cfb, compressed) {
   try {
@@ -2242,16 +2291,21 @@ function detectHwp5Headings(blocks, docInfo) {
   }
   if (baseFontSize <= 0) return;
   for (const block of blocks) {
-    if (block.type !== "paragraph" || !block.text || !block.style?.fontSize) continue;
+    if (block.type === "heading") continue;
+    if (block.type !== "paragraph" || !block.text) continue;
     const text = block.text.trim();
     if (text.length === 0 || text.length > 200) continue;
     if (/^\d+$/.test(text)) continue;
-    const ratio = block.style.fontSize / baseFontSize;
     let level = 0;
-    if (ratio >= HEADING_RATIO_H1) level = 1;
-    else if (ratio >= HEADING_RATIO_H2) level = 2;
-    else if (ratio >= HEADING_RATIO_H3) level = 3;
-    if (/^제\d+[조장절편]/.test(text) && text.length <= 50) {
+    if (block.style?.fontSize && baseFontSize > 0) {
+      const ratio = block.style.fontSize / baseFontSize;
+      if (ratio >= HEADING_RATIO_H1) level = 1;
+      else if (ratio >= HEADING_RATIO_H2) level = 2;
+      else if (ratio >= HEADING_RATIO_H3) level = 3;
+    }
+    if (/^제\d+[장절편]\s/.test(text) && text.length <= 50) {
+      if (level === 0) level = 2;
+    } else if (/^제\d+(조의?\d*)\s*[\(（]/.test(text) && text.length <= 80) {
       if (level === 0) level = 3;
     }
     if (level > 0) {
@@ -2497,13 +2551,20 @@ function parseSection(records, docInfo, warnings, sectionNum) {
   while (i < records.length) {
     const rec = records[i];
     if (rec.tagId === TAG_PARA_HEADER && rec.level === 0) {
-      const { paragraph, tables, nextIdx, charShapeIds } = parseParagraphWithTables(records, i);
+      const { paragraph, tables, nextIdx, charShapeIds, paraShapeId } = parseParagraphWithTables(records, i);
       if (paragraph) {
         const block = { type: "paragraph", text: paragraph, pageNumber: sectionNum };
         if (docInfo && charShapeIds.length > 0) {
           const style = resolveCharStyle(charShapeIds, docInfo);
           if (style) block.style = style;
         }
+        if (docInfo && paraShapeId >= 0 && paraShapeId < docInfo.paraShapes.length) {
+          const ol = docInfo.paraShapes[paraShapeId].outlineLevel;
+          if (ol >= 1 && ol <= 6) {
+            block.type = "heading";
+            block.level = ol;
+          }
+        }
         blocks.push(block);
       }
       for (const t of tables) blocks.push({ type: "table", table: t, pageNumber: sectionNum });
@@ -2623,6 +2684,8 @@ function parseParagraphWithTables(records, startIdx) {
   let text = "";
   const tables = [];
   const charShapeIds = [];
+  const paraHeaderData = records[startIdx].data;
+  const paraShapeId = paraHeaderData.length >= 10 ? paraHeaderData.readUInt16LE(8) : -1;
   let i = startIdx + 1;
   while (i < records.length) {
     const rec = records[i];
@@ -2647,7 +2710,7 @@ function parseParagraphWithTables(records, startIdx) {
     i++;
   }
   const trimmed = text.trim();
-  return { paragraph: trimmed || null, tables, nextIdx: i, charShapeIds };
+  return { paragraph: trimmed || null, tables, nextIdx: i, charShapeIds, paraShapeId };
 }
 function parseTableBlock(records, startIdx) {
   const tableLevel = records[startIdx].level;
@@ -3465,7 +3528,7 @@ async function parsePdfDocument(buffer, options) {
     if (totalChars / Math.max(parsedPageCount, 1) < 10) {
       if (options?.ocr) {
         try {
-          const { ocrPages } = await import("./provider-A4FHJSID.js");
+          const { ocrPages } = await import("./provider-EU3CG724.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");
@@ -4563,7 +4626,7 @@ async function parseXlsxDocument(buffer, options) {
   }
   let pageFilter = null;
   if (options?.pages) {
-    const { parsePageRange: parsePageRange2 } = await import("./page-range-737B4EZW.js");
+    const { parsePageRange: parsePageRange2 } = await import("./page-range-OF5I4PQY.js");
     pageFilter = parsePageRange2(options.pages, sheets.length);
   }
   const blocks = [];
@@ -5446,4 +5509,4 @@ export {
   extractFormFields,
   parse
 };
-//# sourceMappingURL=chunk-XJYM2AUA.js.map
+//# sourceMappingURL=chunk-4UH6ABAY.js.map