kordoc 4.18.2 → 4.18.3

package/dist/{parser-KB4BYQBZ.js → parser-NKP2NCSV.js}

@@ -40,15 +40,15 @@ import {
   startsNewItem,
   trimTrailingEmptyTableCols,
   wrapJoiner
-} from "./chunk-OGAYGUJF.js";
+} from "./chunk-J2A46VCA.js";
 import {
   ocrModelsCached
-} from "./chunk-NZM2SSI4.js";
-import "./chunk-4B2EXELH.js";
-import "./chunk-6OKBTY72.js";
+} from "./chunk-5X7XWYR2.js";
+import "./chunk-RDBTOE5X.js";
+import "./chunk-AZQWESRH.js";
 import {
   blocksToPages
-} from "./chunk-GM7XIZF2.js";
+} from "./chunk-6GLTGOP6.js";
 import {
   encodePng
 } from "./chunk-7FCNOKOR.js";
@@ -58,7 +58,7 @@ import {
   escapeLiteralTags,
   stripScriptTags,
   unframeLayoutTables
-} from "./chunk-F5HING4Y.js";
+} from "./chunk-UVUKT3IQ.js";
 import {
   hasRequestedPagesAfter,
   parsePageRange
@@ -67,11 +67,11 @@ import {
   WINGDINGS,
   blocksToMarkdown,
   escapeLiteralDollar
-} from "./chunk-IA3UKDHA.js";
+} from "./chunk-W3RADOLO.js";
 import {
   KordocError,
   sanitizeHref
-} from "./chunk-OBILAVWC.js";
+} from "./chunk-VGDTONTB.js";
 
 // src/pdf/ocr-region-merge.ts
 function mergeOcrImageRegions(blocks, page, regions, ocrBlocks) {
@@ -924,6 +924,17 @@ var CARRIED_MIN_COL_W = 4;
 var LABEL_ROOM = 1.5;
 var HANGING_MIN = 0.5;
 var HANGING_TOL = 1;
+var KO_ITEMS = "\uAC00\uB098\uB2E4\uB77C\uB9C8\uBC14\uC0AC\uC544\uC790\uCC28\uCE74\uD0C0\uD30C\uD558\uAC70\uB108\uB354\uB7EC\uBA38\uBC84\uC11C\uC5B4\uC800\uCC98\uCEE4\uD130\uD37C\uD5C8\uACE0\uB178\uB3C4\uB85C\uBAA8\uBCF4\uC18C\uC624\uC870\uCD08\uCF54\uD1A0\uD3EC\uD638\uAD6C\uB204\uB450\uB8E8\uBB34\uBD80\uC218\uC6B0\uC8FC\uCD94\uCFE0\uD22C\uD478\uD6C4";
+var ITEM_MARK = /^\s*(?:([가-힣])\.|(\d{1,3})[.)]|\(([가-힣\d]{1,3})\))(?=\s)/;
+function nextItemHead(prev, next) {
+  const a = ITEM_MARK.exec(prev), b = ITEM_MARK.exec(next);
+  if (!a || !b) return false;
+  const succ = (x, y) => /^\d+$/.test(x) ? /^\d+$/.test(y) && Number(y) === Number(x) + 1 : KO_ITEMS.indexOf(x) >= 0 && KO_ITEMS.indexOf(y) === KO_ITEMS.indexOf(x) + 1;
+  if (a[1] && b[1]) return succ(a[1], b[1]);
+  if (a[2] && b[2]) return succ(a[2], b[2]) && prev.trimStart()[a[2].length] === next.trimStart()[b[2].length];
+  if (a[3] && b[3]) return succ(a[3], b[3]);
+  return false;
+}
 function continuesAcross(u, d, x1, x2) {
   const U = CELL_LINES.get(u), D = CELL_LINES.get(d);
   if (!U?.length || !D?.length) return false;
@@ -1055,6 +1066,8 @@ function mergeSplitRow(table, owner, first, colXs, textOnly = false, cut = NO_CU
   if (!carriedSplit) {
     const norm = (c) => flowText(c.text).replace(/\s+/g, "");
     if (pairs.some(([u, d]) => hasContent(d.cell) && norm(cell(d)) !== norm(cell(u)) && lineEnded(cell(u), colXs[u.c], colXs[u.c + u.cs], cell(d)))) return false;
+    const oneLine = (c) => CELL_LINES.get(c)?.length === 1;
+    if (pairs.some(([u, d]) => nextItemHead(cell(u).text, cell(d).text)) && pairs.some(([u, d]) => hasContent(u.cell) && hasContent(d.cell) && oneLine(cell(u)) && oneLine(cell(d)) && norm(cell(u)) !== norm(cell(d)))) return false;
     const flows = ([u, d]) => continuesAcross(cell(u), cell(d), colXs[u.c], colXs[u.c + u.cs]) || clauseContinues(cell(u), cell(d));
     if (!(cut.open && !newCell)) {
       if (pairs.some((p) => p[1].rs > 1 && !flows(p))) return false;
@@ -2686,7 +2699,7 @@ function mergeLinkRuns(markdown) {
 
 // src/pdf/formula-ocr.ts
 async function applyFormulaOcr(buffer, blocks, pageFilter, effectivePageCount, warnings, _onProgress) {
-  const formulaMod = await import("./formula-E3N7KDDG.js");
+  const formulaMod = await import("./formula-QZAOKPBX.js");
   const { FormulaPipeline, ensureFormulaModels } = formulaMod;
   await ensureFormulaModels((p) => {
     if (p.phase === "download" && p.total) {
@@ -3070,7 +3083,7 @@ async function parsePdfDocument(buffer, options) {
       }
       if (targets.size > 0) {
         try {
-          const { runPdfOcr } = await import("./pdf-ocr-YEBFWU2K.js");
+          const { runPdfOcr } = await import("./pdf-ocr-VJL7ON2Q.js");
           const mode = typeof options?.ocr === "function" ? options.ocr : "builtin";
           const regionPages = new Map([...ocrRegions].filter(([p]) => options?.ocr !== "force" && !isImageBased && !pageQuality.find((q) => q.page === p)?.needsOcr));
           const ocrPageBlocks = await runPdfOcr(ocrBuffer, targets, mode, warnings, options?.onProgress, options?.tables !== false, vectorPageOps, regionPages);
@@ -3302,4 +3315,4 @@ export {
   parsePdfDocument,
   removeHeaderFooterBlocks
 };
-//# sourceMappingURL=parser-KB4BYQBZ.js.map
+//# sourceMappingURL=parser-NKP2NCSV.js.map