kordoc 3.11.0 → 3.12.0

package/dist/index.js

@@ -19,7 +19,7 @@ import {
   sanitizeHref,
   stripDtd,
   toArrayBuffer
-} from "./chunk-D4JKKSSI.js";
+} from "./chunk-NGKIAZ6G.js";
 import {
   parsePageRange
 } from "./chunk-GE43BE46.js";
@@ -25548,7 +25548,7 @@ async function parseHwp(buffer, options) {
 async function parsePdf(buffer, options) {
   let parsePdfDocument;
   try {
-    const mod = await import("./parser-STGI5I5K.js");
+    const mod = await import("./parser-AWEGMHHV.js");
     parsePdfDocument = mod.parsePdfDocument;
   } catch {
     return {

package/dist/mcp.js

@@ -10,22 +10,22 @@ import {
   parse,
   patchHwp,
   patchHwpx
-} from "./chunk-TUBD3EEF.js";
+} from "./chunk-RUYSNNYZ.js";
 import {
   detectFormat,
   detectZipFormat
 } from "./chunk-MEPHGCPQ.js";
-import "./chunk-GETZFABK.js";
+import "./chunk-ILR7GF3W.js";
 import {
   blocksToMarkdown
-} from "./chunk-O27B5IRL.js";
+} from "./chunk-BUDM6ULT.js";
 import "./chunk-MOL7MDBG.js";
 import {
   KordocError,
   VERSION,
   sanitizeError,
   toArrayBuffer
-} from "./chunk-SNSAZ7FA.js";
+} from "./chunk-7Z43WEML.js";
 
 // src/mcp.ts
 import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
@@ -196,7 +196,7 @@ server.tool(
           break;
         case "pdf":
           try {
-            const { extractPdfMetadataOnly } = await import("./parser-3EARWV2H.js");
+            const { extractPdfMetadataOnly } = await import("./parser-5FSEFR3R.js");
             metadata = await extractPdfMetadataOnly(buffer);
           } catch {
             metadata = void 0;

package/dist/{parser-3EARWV2H.js → parser-5FSEFR3R.js}

@@ -4,7 +4,7 @@ import {
   HEADING_RATIO_H2,
   HEADING_RATIO_H3,
   blocksToMarkdown
-} from "./chunk-O27B5IRL.js";
+} from "./chunk-BUDM6ULT.js";
 import {
   parsePageRange
 } from "./chunk-MOL7MDBG.js";
@@ -12,7 +12,7 @@ import {
   KordocError,
   safeMax,
   safeMin
-} from "./chunk-SNSAZ7FA.js";
+} from "./chunk-7Z43WEML.js";
 
 // src/pdf/line-extract.ts
 import { OPS } from "pdfjs-dist/legacy/build/pdf.mjs";
@@ -240,10 +240,40 @@ var EDGE_MIN_SPAN = 12;
 var EDGE_INSET = 15;
 var EDGE_NEAR = 10;
 var EDGE_CONNECT_TOL = 5;
+var CHAIN_Y_TOL = 1.5;
+var CHAIN_GAP = 3;
+function chainCollinearRules(horizontals) {
+  if (horizontals.length <= 1) return horizontals;
+  const sorted = [...horizontals].sort((a, b) => a.y1 - b.y1 || a.x1 - b.x1);
+  const rules = [];
+  let bandStart = 0;
+  const flushBand = (end) => {
+    const band = sorted.slice(bandStart, end).sort((a, b) => a.x1 - b.x1);
+    let cur = { ...band[0] };
+    for (let i = 1; i < band.length; i++) {
+      const seg = band[i];
+      if (seg.x1 - cur.x2 <= CHAIN_GAP) {
+        if (seg.x2 > cur.x2) cur.x2 = seg.x2;
+        if (seg.lineWidth > cur.lineWidth) cur.lineWidth = seg.lineWidth;
+      } else {
+        rules.push(cur);
+        cur = { ...seg };
+      }
+    }
+    rules.push(cur);
+  };
+  for (let i = 1; i <= sorted.length; i++) {
+    if (i === sorted.length || sorted[i].y1 - sorted[bandStart].y1 > CHAIN_Y_TOL) {
+      flushBand(i);
+      bandStart = i;
+    }
+  }
+  return rules;
+}
 function closeOpenTableEdges(horizontals, verticals) {
   if (horizontals.length < EDGE_MIN_RULES) return verticals;
   const groups = [];
-  for (const hl of horizontals) {
+  for (const hl of chainCollinearRules(horizontals)) {
     let placed = false;
     for (const g2 of groups) {
       if (Math.abs(g2[0].x1 - hl.x1) <= EDGE_ALIGN_TOL && Math.abs(g2[0].x2 - hl.x2) <= EDGE_ALIGN_TOL) {
@@ -2298,6 +2328,10 @@ function detectHeadings(blocks, medianFontSize) {
 function shouldDemoteTable(table) {
   const allCells = table.cells.flatMap((row) => row.map((c) => c.text.trim())).filter(Boolean);
   const allText = allCells.join(" ");
+  if (table.rows >= 2 && table.cols >= 2 && table.cells[0].every((c) => {
+    const t = c.text.trim();
+    return t.length > 0 && t.length <= 12 && !/[□■◆○●▶ㅇ<>]/.test(t);
+  }) && table.cells.slice(1).some((row) => row.some((c) => c.text.trim() !== ""))) return false;
   if (table.rows <= 3 && table.cols <= 3) {
     const totalCells2 = table.rows * table.cols;
     const emptyCells2 = totalCells2 - allCells.length;
@@ -3440,4 +3474,4 @@ export {
   parsePdfDocument,
   removeHeaderFooterBlocks
 };
-//# sourceMappingURL=parser-3EARWV2H.js.map
+//# sourceMappingURL=parser-5FSEFR3R.js.map