kordoc 4.2.7 → 4.2.9

package/dist/index.d.cts

@@ -92,6 +92,12 @@ interface IRTable {
     hasHeader: boolean;
     /** 표 캡션 (예: "표 1. 부서별 예산") — v3.0 */
     caption?: string;
+    /**
+     * 캡션 내부 블록 콘텐츠 — v4.2.8 (#55).
+     * 캡션 안 중첩표·문단을 구조 그대로, 원문 순서대로 보존한다 (IRCell.blocks와 같은 계약).
+     * 표가 있을 때만 채워지며, caption은 하위 호환용 평탄화 문자열로 계속 제공된다.
+     */
+    captionBlocks?: IRBlock[];
 }
 interface IRCell {
     text: string;

package/dist/index.d.ts

@@ -92,6 +92,12 @@ interface IRTable {
     hasHeader: boolean;
     /** 표 캡션 (예: "표 1. 부서별 예산") — v3.0 */
     caption?: string;
+    /**
+     * 캡션 내부 블록 콘텐츠 — v4.2.8 (#55).
+     * 캡션 안 중첩표·문단을 구조 그대로, 원문 순서대로 보존한다 (IRCell.blocks와 같은 계약).
+     * 표가 있을 때만 채워지며, caption은 하위 호환용 평탄화 문자열로 계속 제공된다.
+     */
+    captionBlocks?: IRBlock[];
 }
 interface IRCell {
     text: string;

package/dist/index.js

@@ -8,7 +8,7 @@ import {
   flattenLayoutTables,
   inlineImagesIntoMarkdown,
   mapPuaText
-} from "./chunk-YAQRGQUJ.js";
+} from "./chunk-STYN45MS.js";
 import {
   parsePageRange
 } from "./chunk-GE43BE46.js";
@@ -27,7 +27,7 @@ import {
   sanitizeHref,
   stripDtd,
   toArrayBuffer
-} from "./chunk-IVX3YLQH.js";
+} from "./chunk-7ZARSBQ3.js";
 
 // src/index.ts
 import { readFile } from "fs/promises";
@@ -1914,6 +1914,7 @@ function resolveParaHeading(paraEl, ctx) {
 function buildTableWithCellMeta(state, keepAnchoredEmptyCols) {
   const table2 = buildTable(state.rows, { keepAnchoredEmptyCols });
   if (state.caption) table2.caption = state.caption;
+  if (state.captionBlocks?.length) table2.captionBlocks = state.captionBlocks;
   const anchors = [];
   {
     const covered = /* @__PURE__ */ new Set();
@@ -2061,12 +2062,13 @@ function walkSection(node, blocks, tableCtx, tableStack, ctx, depth = 0) {
       }
       // 표/도표 캡션 — IRTable.caption으로 보존 (v3.0, 기존 무음 드롭 수정)
       case "caption": {
-        const capText = collectSubListText(el, ctx);
-        if (capText) {
+        const cap = collectSubListContent(el, ctx);
+        if (cap.text) {
           if (tableCtx) {
-            tableCtx.caption = (tableCtx.caption ? tableCtx.caption + "\n" : "") + capText;
+            tableCtx.caption = (tableCtx.caption ? tableCtx.caption + "\n" : "") + cap.text;
+            if (cap.hasStructure) (tableCtx.captionBlocks ??= []).push(...cap.blocks);
           } else {
-            blocks.push({ type: "paragraph", text: capText, pageNumber: ctx.sectionNum });
+            blocks.push({ type: "paragraph", text: cap.text, pageNumber: ctx.sectionNum });
           }
         }
         break;
@@ -2316,40 +2318,60 @@ function hasPageAutoNum(el, depth = 0) {
   return false;
 }
 function collectSubListText(el, ctx, depth = 0) {
-  if (depth > 10) return "";
+  return collectSubListContent(el, ctx, depth).text;
+}
+function collectSubListContent(el, ctx, depth = 0) {
+  const out = { text: "", blocks: [], hasStructure: false };
+  if (depth > 10) return out;
   const parts = [];
   const children = el.childNodes;
-  if (!children) return "";
+  if (!children) return out;
   for (let i = 0; i < children.length; i++) {
     const ch = children[i];
     if (ch.nodeType !== 1) continue;
     const tag = (ch.tagName || ch.localName || "").replace(/^[^:]+:/, "");
     if (tag === "p" || tag === "para") {
       const t = extractParagraphInfo(ch, ctx.styleMap, ctx).text;
-      if (t) parts.push(t);
+      if (t) {
+        parts.push(t);
+        out.blocks.push({ type: "paragraph", text: t, pageNumber: ctx.sectionNum });
+      }
       const tbls = [];
       findTopLevelTbls(ch, tbls);
       for (const tbl2 of tbls) {
-        const flat = flattenSubListTable(tbl2, ctx, depth);
-        if (flat) parts.push(flat);
+        const built = buildSubListTable(tbl2, ctx, depth);
+        if (built.text) parts.push(built.text);
+        if (built.block) {
+          out.blocks.push(built.block);
+          out.hasStructure = true;
+        }
       }
     } else if (tag === "tbl") {
-      const flat = flattenSubListTable(ch, ctx, depth);
-      if (flat) parts.push(flat);
+      const built = buildSubListTable(ch, ctx, depth);
+      if (built.text) parts.push(built.text);
+      if (built.block) {
+        out.blocks.push(built.block);
+        out.hasStructure = true;
+      }
     } else {
-      const t = collectSubListText(ch, ctx, depth + 1);
-      if (t) parts.push(t);
+      const sub = collectSubListContent(ch, ctx, depth + 1);
+      if (sub.text) parts.push(sub.text);
+      out.blocks.push(...sub.blocks);
+      if (sub.hasStructure) out.hasStructure = true;
     }
   }
-  return parts.join("\n").trim();
+  out.text = parts.join("\n").trim();
+  return out;
 }
-function flattenSubListTable(el, ctx, depth) {
+function buildSubListTable(el, ctx, depth) {
   const sink = [];
   const st = { rows: [], currentRow: [], cell: null };
   walkSection(el, sink, st, [], ctx, depth + 1);
   let flat = convertTableToText(st.rows);
   if (st.caption) flat = st.caption + (flat ? "\n" + flat : "");
-  return flat;
+  const built = [];
+  completeTable(st, [], built, ctx);
+  return { text: flat, block: built.find((b) => b.type === "table") ?? null };
 }
 function findTopLevelTbls(el, out, depth = 0) {
   if (depth > MAX_XML_DEPTH) return;
@@ -2395,11 +2417,13 @@ function walkParagraphChildren(node, blocks, tableCtx, tableStack, ctx, depth =
         walkSection(el, blocks, newTable, tableStack, ctx, d + 1);
         tableCtx = completeTable(newTable, tableStack, blocks, ctx);
       } else if (localTag === "caption" && !inShape) {
-        const capText = collectSubListText(el, ctx);
-        if (capText) {
-          if (tableCtx?.cell) mergeBlocksIntoCell(tableCtx.cell, [{ type: "paragraph", text: capText, pageNumber: ctx.sectionNum }]);
-          else if (tableCtx) tableCtx.caption = (tableCtx.caption ? tableCtx.caption + "\n" : "") + capText;
-          else blocks.push({ type: "paragraph", text: capText, pageNumber: ctx.sectionNum });
+        const cap = collectSubListContent(el, ctx);
+        if (cap.text) {
+          if (tableCtx?.cell) mergeBlocksIntoCell(tableCtx.cell, [{ type: "paragraph", text: cap.text, pageNumber: ctx.sectionNum }]);
+          else if (tableCtx) {
+            tableCtx.caption = (tableCtx.caption ? tableCtx.caption + "\n" : "") + cap.text;
+            if (cap.hasStructure) (tableCtx.captionBlocks ??= []).push(...cap.blocks);
+          } else blocks.push({ type: "paragraph", text: cap.text, pageNumber: ctx.sectionNum });
         }
       } else if (localTag === "pic" || localTag === "shape" || localTag === "drawingObject") {
         if (tableCtx?.cell) {
@@ -17544,9 +17568,8 @@ function decodeJohab(ch) {
     const jongIdx = ch & 31;
     const cho = CHO_MAP[choIdx];
     const jung = JUNG_MAP[jungIdx];
-    let jong = JONG_MAP[jongIdx];
-    if (cho !== -1 && jung !== -1) {
-      if (jong === -1) jong = 0;
+    const jong = JONG_MAP[jongIdx];
+    if (cho !== -1 && jung !== -1 && jong !== -1) {
       return 44032 + cho * 588 + jung * 28 + jong;
     }
     const hit = lookupSymbol(ch);
@@ -17715,8 +17738,6 @@ var LINE_INFO_SIZE = 14;
 var INLINE_CHAR_SHAPE_SIZE = 31;
 var SIMPLE_CTRL = /* @__PURE__ */ new Map([
   [9, { extraBytes: 6, extraHchar: 3, emit: "	" }],
-  [7, { extraBytes: 6, extraHchar: 3, emit: "\uFFFC" }],
-  [8, { extraBytes: 6, extraHchar: 3, emit: "\uFFFC" }],
   [18, { extraBytes: 6, extraHchar: 3, emit: " " }],
   // AutoNumber → 공백 (HWP5 패턴)
   [19, { extraBytes: 6, extraHchar: 3, emit: "\uFFFC" }],
@@ -17725,7 +17746,8 @@ var SIMPLE_CTRL = /* @__PURE__ */ new Map([
   [22, { extraBytes: 22, extraHchar: 11, emit: "\uFFFC" }],
   [23, { extraBytes: 8, extraHchar: 4, emit: "\uFFFC" }],
   [24, { extraBytes: 4, extraHchar: 2, emit: "-" }],
-  [25, { extraBytes: 4, extraHchar: 2, emit: "-" }],
+  [25, { extraBytes: 4, extraHchar: 2, emit: null }],
+  // 차례 표식 — 비가시 (rhwp #2765)
   [26, { extraBytes: 244, extraHchar: 122, emit: "\uFFFC" }],
   [28, { extraBytes: 62, extraHchar: 31, emit: "\uFFFC" }],
   [30, { extraBytes: 2, extraHchar: 1, emit: "\xA0" }],
@@ -17869,9 +17891,6 @@ function parseCharStream(reader, charCount, ctx) {
       case 11:
         parsePicture(reader, ctx);
         break;
-      case 12:
-        reader.skip(84);
-        break;
       case 14:
         reader.skip(84);
         break;
@@ -17896,6 +17915,12 @@ function parseCharStream(reader, charCount, ctx) {
       case 6:
         reader.skip(34);
         break;
+      case 7:
+        reader.skip(76);
+        break;
+      case 8:
+        reader.skip(88);
+        break;
       case 29:
         if (headerVal1 < 1e6) reader.skip(headerVal1);
         break;
@@ -29815,7 +29840,7 @@ function findFirst(el, name, depth = 0) {
   return null;
 }
 function escapeXml3(s) {
-  return s.replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;").replace(/"/g, "&quot;");
+  return s.replace(/[\x00-\x08\x0B\x0C\x0E-\x1F]/g, "").replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;").replace(/"/g, "&quot;");
 }
 var OBJ_TAGS = /* @__PURE__ */ new Set(["tbl", "pic", "container", "equation", "rect", "ellipse", "polygon", "curv", "line", "arc", "ole", "textart"]);
 var pt = (u) => String(Math.round(u) / 100);
@@ -30574,7 +30599,7 @@ async function parse(input, options) {
 }
 async function parseImage(buffer, options) {
   try {
-    const { parseImageDocument } = await import("./image-ocr-OS4ETBOL.js");
+    const { parseImageDocument } = await import("./image-ocr-PS7XCV6K.js");
     const { blocks, warnings } = await parseImageDocument(buffer, options);
     return {
       success: true,
@@ -30632,7 +30657,7 @@ async function parseHwp(buffer, options) {
 async function parsePdf(buffer, options) {
   let parsePdfDocument;
   try {
-    const mod = await import("./parser-2JAC5WMM.js");
+    const mod = await import("./parser-A34NAYKA.js");
     parsePdfDocument = mod.parsePdfDocument;
   } catch {
     return {