kordoc 4.9.2 → 4.10.0

package/dist/index.d.cts

@@ -329,6 +329,8 @@ interface ExtractedImage {
     data: Uint8Array;
     /** MIME 타입 */
     mimeType: string;
+    /** 원본 컨테이너 내 항목명 (HWPX/DOCX ZIP 경로, HWP5 BinData 스토리지명 — PDF 등 합성 이미지는 없음, #70) */
+    source?: string;
 }
 interface ParseFailure extends ParseResultBase {
     success: false;

package/dist/index.d.ts

@@ -329,6 +329,8 @@ interface ExtractedImage {
     data: Uint8Array;
     /** MIME 타입 */
     mimeType: string;
+    /** 원본 컨테이너 내 항목명 (HWPX/DOCX ZIP 경로, HWP5 BinData 스토리지명 — PDF 등 합성 이미지는 없음, #70) */
+    source?: string;
 }
 interface ParseFailure extends ParseResultBase {
     success: false;

package/dist/index.js

@@ -8,7 +8,7 @@ import {
   flattenLayoutTables,
   inlineImagesIntoMarkdown,
   mapPuaText
-} from "./chunk-VYGNPB2X.js";
+} from "./chunk-VR3ECURZ.js";
 import {
   parsePageRange
 } from "./chunk-GE43BE46.js";
@@ -16,7 +16,7 @@ import {
   HEADING_RATIO_H1,
   HEADING_RATIO_H2,
   HEADING_RATIO_H3
-} from "./chunk-BN33O25L.js";
+} from "./chunk-ONQHGZTS.js";
 import {
   KordocError,
   VERSION,
@@ -29,7 +29,7 @@ import {
   sanitizeHref,
   stripDtd,
   toArrayBuffer
-} from "./chunk-OOT54JF2.js";
+} from "./chunk-7UUQXASL.js";
 
 // src/index.ts
 import { readFile } from "fs/promises";
@@ -2927,7 +2927,7 @@ function resolveImageBlocks(binDataMap, blocks, warnings, sweepUnreferenced) {
           const ext = mime.includes("jpeg") ? "jpg" : mime.includes("png") ? "png" : mime.includes("gif") ? "gif" : mime.includes("bmp") ? "bmp" : "bin";
           img = { filename: `image_${String(imageIndex).padStart(3, "0")}.${ext}`, data: new Uint8Array(bin.data), mime };
           resolved.set(storageId, img);
-          images.push({ filename: img.filename, data: img.data, mimeType: img.mime });
+          images.push({ filename: img.filename, data: img.data, mimeType: img.mime, source: bin.name });
           renamed.set(storageId, img.filename);
         }
       }
@@ -2951,7 +2951,7 @@ function resolveImageBlocks(binDataMap, blocks, warnings, sweepUnreferenced) {
       const ext = mime.includes("jpeg") ? "jpg" : mime.includes("png") ? "png" : mime.includes("gif") ? "gif" : mime.includes("bmp") ? "bmp" : "bin";
       const filename = `image_${String(imageIndex).padStart(3, "0")}.${ext}`;
       const data = new Uint8Array(bin.data);
-      images.push({ filename, data, mimeType: mime });
+      images.push({ filename, data, mimeType: mime, source: bin.name });
       blocks.push({ type: "image", text: filename, imageData: { data, mimeType: mime, filename: bin.name } });
     }
   }
@@ -3088,7 +3088,7 @@ async function extractImagesFromZip(zip, blocks, decompressed, warnings, sweepUn
           const mimeType = imageExtToMime(ext);
           imageIndex++;
           const filename2 = `image_${String(imageIndex).padStart(3, "0")}.${mimeToExt(mimeType)}`;
-          img = { filename: filename2, data, mimeType };
+          img = { filename: filename2, data, mimeType, source: path };
           images.push(img);
           usedPaths.add(path);
           break;
@@ -3127,7 +3127,7 @@ async function extractImagesFromZip(zip, blocks, decompressed, warnings, sweepUn
         }
         imageIndex++;
         const filename = `image_${String(imageIndex).padStart(3, "0")}.${mimeToExt(mimeType)}`;
-        const img = { filename, data, mimeType };
+        const img = { filename, data, mimeType, source: file.name };
         images.push(img);
         blocks.push({ type: "image", text: filename, imageData: { data, mimeType, filename: file.name } });
       } catch (err) {
@@ -19790,10 +19790,14 @@ async function buildImageMap(zip, rels, doc, warnings) {
         gif: "image/gif",
         bmp: "image/bmp",
         wmf: "image/wmf",
-        emf: "image/emf"
+        emf: "image/emf",
+        tif: "image/tiff",
+        tiff: "image/tiff",
+        svg: "image/svg+xml"
       };
       const filename = `image_${String(imgIdx).padStart(3, "0")}.${ext}`;
-      images.push({ filename, data, mimeType: mimeMap[ext] ?? "image/png" });
+      const mimeType = mimeMap[ext] ?? detectImageMime(data) ?? "application/octet-stream";
+      images.push({ filename, data, mimeType, source: imgPath });
       imageMap.set(embedId, filename);
     } catch (err) {
       warnings.push({
@@ -29530,7 +29534,7 @@ async function patchHwp(original, editedMarkdown, options) {
       for (let i = 0; i < scans.length; i++) {
         if (scans[i].repl.size === 0 && scans[i].inserts.size === 0) continue;
         const newStream = serializeRecords(scans[i].records, scans[i].repl, scans[i].inserts);
-        const content = compressed ? deflateRawSync2(newStream) : newStream;
+        const content = compressed ? compressWithTail(newStream) : newStream;
         out = replaceOleStream(out, sectionPaths[i], content);
       }
       data = new Uint8Array(out);
@@ -30013,6 +30017,26 @@ function rebuildCharShape(csData, coreStartUnit) {
   });
   return { buf, count: kept.length };
 }
+var CRC_TABLE2 = (() => {
+  const t = new Uint32Array(256);
+  for (let n = 0; n < 256; n++) {
+    let c = n;
+    for (let k = 0; k < 8; k++) c = c & 1 ? 3988292384 ^ c >>> 1 : c >>> 1;
+    t[n] = c >>> 0;
+  }
+  return t;
+})();
+function crc322(buf) {
+  let c = 4294967295;
+  for (let i = 0; i < buf.length; i++) c = CRC_TABLE2[(c ^ buf[i]) & 255] ^ c >>> 8;
+  return (c ^ 4294967295) >>> 0;
+}
+function compressWithTail(raw) {
+  const tail = Buffer.alloc(8);
+  tail.writeUInt32LE(crc322(raw), 0);
+  tail.writeUInt32LE(raw.length >>> 0, 4);
+  return Buffer.concat([deflateRawSync2(raw), tail]);
+}
 function synthesizeLineSegs(lineSegData, newRaw, startUnits) {
   if (lineSegData.length < 36) return null;
   const seg0 = lineSegData.subarray(0, 36);
@@ -30090,6 +30114,15 @@ function stageParaPatch(scan, para2, newPlain, skip) {
       scan.repl.set(para2.lineSegIdx, synth.buf);
       newHeader.writeUInt16LE(synth.count, 16);
     }
+  } else {
+    const lineSegData = records[para2.lineSegIdx].data;
+    const segCount = Math.floor(lineSegData.length / 36);
+    let keep = 1;
+    while (keep < segCount && lineSegData.readUInt32LE(keep * 36) < nChars) keep++;
+    if (keep < segCount) {
+      scan.repl.set(para2.lineSegIdx, Buffer.from(lineSegData.subarray(0, keep * 36)));
+      newHeader.writeUInt16LE(keep, 16);
+    }
   }
   scan.repl.set(para2.headerIdx, newHeader);
   return 1;
@@ -31959,7 +31992,7 @@ async function dispatch(format, buffer, opts) {
 }
 async function parseImage(buffer, options) {
   try {
-    const { parseImageDocument } = await import("./image-ocr-YZ56O63S.js");
+    const { parseImageDocument } = await import("./image-ocr-3YP5I2H4.js");
     const { blocks, warnings } = await parseImageDocument(buffer, options);
     return {
       success: true,
@@ -32018,7 +32051,7 @@ async function parseHwp(buffer, options) {
 async function parsePdf(buffer, options) {
   let parsePdfDocument;
   try {
-    const mod = await import("./parser-6UL2HLDH.js");
+    const mod = await import("./parser-6HKVNDZM.js");
     parsePdfDocument = mod.parsePdfDocument;
   } catch {
     return {