kordoc 2.2.3 → 2.2.5

package/dist/parser-CYBX5MP4.cjs

@@ -0,0 +1,2278 @@
+"use strict";Object.defineProperty(exports, "__esModule", {value: true}); function _interopRequireWildcard(obj) { if (obj && obj.__esModule) { return obj; } else { var newObj = {}; if (obj != null) { for (var key in obj) { if (Object.prototype.hasOwnProperty.call(obj, key)) { newObj[key] = obj[key]; } } } newObj.default = obj; return newObj; } } function _optionalChain(ops) { let lastAccessLHS = undefined; let value = ops[0]; let i = 1; while (i < ops.length) { const op = ops[i]; const fn = ops[i + 1]; i += 2; if ((op === 'optionalAccess' || op === 'optionalCall') && value == null) { return undefined; } if (op === 'access' || op === 'optionalAccess') { lastAccessLHS = value; value = fn(value); } else if (op === 'call' || op === 'optionalCall') { value = fn((...args) => value.call(lastAccessLHS, ...args)); lastAccessLHS = undefined; } } return value; } var _class;
+
+
+
+
+
+
+
+var _chunkOJ4QR33Vcjs = require('./chunk-OJ4QR33V.cjs');
+
+
+var _chunkMUOQXDZ4cjs = require('./chunk-MUOQXDZ4.cjs');
+
+// src/pdf/line-detector.ts
+var _pdfmjs = require('pdfjs-dist/legacy/build/pdf.mjs');
+var ORIENTATION_TOL = 2;
+var MIN_LINE_LENGTH = 15;
+var MAX_LINE_WIDTH = 5;
+var CONNECT_TOL = 5;
+var CELL_PADDING = 2;
+var MIN_COL_WIDTH = 15;
+var MIN_ROW_HEIGHT = 6;
+var VERTEX_MERGE_FACTOR = 4;
+var MIN_COORD_MERGE_TOL = 8;
+function extractLines(fnArray, argsArray) {
+  const horizontals = [];
+  const verticals = [];
+  let lineWidth = 1;
+  let currentPath = [];
+  let pathStartX = 0, pathStartY = 0;
+  let curX = 0, curY = 0;
+  function pushRectangle(path, rx, ry, rw, rh) {
+    if (Math.abs(rh) < ORIENTATION_TOL * 2) {
+      path.push({ x1: rx, y1: ry + rh / 2, x2: rx + rw, y2: ry + rh / 2 });
+    } else if (Math.abs(rw) < ORIENTATION_TOL * 2) {
+      path.push({ x1: rx + rw / 2, y1: ry, x2: rx + rw / 2, y2: ry + rh });
+    } else {
+      path.push(
+        { x1: rx, y1: ry, x2: rx + rw, y2: ry },
+        { x1: rx + rw, y1: ry, x2: rx + rw, y2: ry + rh },
+        { x1: rx + rw, y1: ry + rh, x2: rx, y2: ry + rh },
+        { x1: rx, y1: ry + rh, x2: rx, y2: ry }
+      );
+    }
+  }
+  function flushPath(isStroke) {
+    if (!isStroke) {
+      currentPath = [];
+      return;
+    }
+    for (const seg of currentPath) {
+      classifyAndAdd(seg, lineWidth, horizontals, verticals);
+    }
+    currentPath = [];
+  }
+  for (let i = 0; i < fnArray.length; i++) {
+    const op = fnArray[i];
+    const args = argsArray[i];
+    switch (op) {
+      case _pdfmjs.OPS.setLineWidth:
+        lineWidth = args[0] || 1;
+        break;
+      case _pdfmjs.OPS.constructPath: {
+        const arg0 = args[0];
+        if (Array.isArray(arg0)) {
+          const subOps = arg0;
+          const coords = args[1];
+          let ci = 0;
+          for (const subOp of subOps) {
+            if (subOp === _pdfmjs.OPS.moveTo) {
+              curX = coords[ci++];
+              curY = coords[ci++];
+              pathStartX = curX;
+              pathStartY = curY;
+            } else if (subOp === _pdfmjs.OPS.lineTo) {
+              const x2 = coords[ci++], y2 = coords[ci++];
+              currentPath.push({ x1: curX, y1: curY, x2, y2 });
+              curX = x2;
+              curY = y2;
+            } else if (subOp === _pdfmjs.OPS.rectangle) {
+              const rx = coords[ci++], ry = coords[ci++];
+              const rw = coords[ci++], rh = coords[ci++];
+              pushRectangle(currentPath, rx, ry, rw, rh);
+            } else if (subOp === _pdfmjs.OPS.closePath) {
+              if (curX !== pathStartX || curY !== pathStartY) {
+                currentPath.push({ x1: curX, y1: curY, x2: pathStartX, y2: pathStartY });
+              }
+              curX = pathStartX;
+              curY = pathStartY;
+            } else if (subOp === _pdfmjs.OPS.curveTo) {
+              ci += 6;
+            } else if (subOp === _pdfmjs.OPS.curveTo2 || subOp === _pdfmjs.OPS.curveTo3) {
+              ci += 4;
+            }
+          }
+        } else {
+          const afterOp = arg0;
+          const dataArr = args[1];
+          const pathData = _optionalChain([dataArr, 'optionalAccess', _3 => _3[0]]);
+          if (pathData && typeof pathData === "object") {
+            const len = Object.keys(pathData).length;
+            let di = 0;
+            while (di < len) {
+              const drawOp = pathData[di++];
+              if (drawOp === 0 /* moveTo */) {
+                curX = pathData[di++];
+                curY = pathData[di++];
+                pathStartX = curX;
+                pathStartY = curY;
+              } else if (drawOp === 1 /* lineTo */) {
+                const x2 = pathData[di++], y2 = pathData[di++];
+                currentPath.push({ x1: curX, y1: curY, x2, y2 });
+                curX = x2;
+                curY = y2;
+              } else if (drawOp === 2 /* curveTo */) {
+                di += 6;
+              } else if (drawOp === 3 /* quadraticCurveTo */) {
+                di += 4;
+              } else if (drawOp === 4 /* closePath */) {
+                if (curX !== pathStartX || curY !== pathStartY) {
+                  currentPath.push({ x1: curX, y1: curY, x2: pathStartX, y2: pathStartY });
+                }
+                curX = pathStartX;
+                curY = pathStartY;
+              } else {
+                break;
+              }
+            }
+          }
+          if (afterOp === _pdfmjs.OPS.stroke || afterOp === _pdfmjs.OPS.closeStroke) {
+            flushPath(true);
+          } else if (afterOp === _pdfmjs.OPS.fill || afterOp === _pdfmjs.OPS.eoFill || afterOp === _pdfmjs.OPS.fillStroke || afterOp === _pdfmjs.OPS.eoFillStroke || afterOp === _pdfmjs.OPS.closeFillStroke || afterOp === _pdfmjs.OPS.closeEOFillStroke) {
+            flushPath(true);
+          } else if (afterOp === _pdfmjs.OPS.endPath) {
+            flushPath(false);
+          }
+        }
+        break;
+      }
+      case _pdfmjs.OPS.stroke:
+      case _pdfmjs.OPS.closeStroke:
+        flushPath(true);
+        break;
+      case _pdfmjs.OPS.fill:
+      case _pdfmjs.OPS.eoFill:
+      case _pdfmjs.OPS.fillStroke:
+      case _pdfmjs.OPS.eoFillStroke:
+      case _pdfmjs.OPS.closeFillStroke:
+      case _pdfmjs.OPS.closeEOFillStroke:
+        flushPath(true);
+        break;
+      case _pdfmjs.OPS.endPath:
+        flushPath(false);
+        break;
+    }
+  }
+  return { horizontals, verticals };
+}
+function classifyAndAdd(seg, lineWidth, horizontals, verticals) {
+  const dx = Math.abs(seg.x2 - seg.x1);
+  const dy = Math.abs(seg.y2 - seg.y1);
+  const length = Math.sqrt(dx * dx + dy * dy);
+  if (length < MIN_LINE_LENGTH) return;
+  if (dy <= ORIENTATION_TOL) {
+    const y = (seg.y1 + seg.y2) / 2;
+    const x1 = Math.min(seg.x1, seg.x2);
+    const x2 = Math.max(seg.x1, seg.x2);
+    horizontals.push({ x1, y1: y, x2, y2: y, lineWidth });
+  } else if (dx <= ORIENTATION_TOL) {
+    const x = (seg.x1 + seg.x2) / 2;
+    const y1 = Math.min(seg.y1, seg.y2);
+    const y2 = Math.max(seg.y1, seg.y2);
+    verticals.push({ x1: x, y1, x2: x, y2, lineWidth });
+  }
+}
+function preprocessLines(horizontals, verticals) {
+  let h = horizontals.filter((l) => l.lineWidth <= MAX_LINE_WIDTH);
+  let v = verticals.filter((l) => l.lineWidth <= MAX_LINE_WIDTH);
+  h = mergeParallelLines(h, "h");
+  v = mergeParallelLines(v, "v");
+  return { horizontals: h, verticals: v };
+}
+function mergeParallelLines(lines, dir) {
+  if (lines.length <= 1) return lines;
+  const sorted = [...lines].sort((a, b) => {
+    const posA = dir === "h" ? a.y1 : a.x1;
+    const posB = dir === "h" ? b.y1 : b.x1;
+    if (Math.abs(posA - posB) > 0.1) return posA - posB;
+    return dir === "h" ? a.x1 - b.x1 : a.y1 - b.y1;
+  });
+  const MERGE_TOL = 3;
+  const result = [sorted[0]];
+  for (let i = 1; i < sorted.length; i++) {
+    const prev = result[result.length - 1];
+    const curr = sorted[i];
+    const prevPos = dir === "h" ? prev.y1 : prev.x1;
+    const currPos = dir === "h" ? curr.y1 : curr.x1;
+    if (Math.abs(prevPos - currPos) <= MERGE_TOL) {
+      const prevStart = dir === "h" ? prev.x1 : prev.y1;
+      const prevEnd = dir === "h" ? prev.x2 : prev.y2;
+      const currStart = dir === "h" ? curr.x1 : curr.y1;
+      const currEnd = dir === "h" ? curr.x2 : curr.y2;
+      const overlap = Math.min(prevEnd, currEnd) - Math.max(prevStart, currStart);
+      const minLen = Math.min(prevEnd - prevStart, currEnd - currStart);
+      if (overlap > minLen * 0.3) {
+        if (dir === "h") {
+          prev.x1 = Math.min(prev.x1, curr.x1);
+          prev.x2 = Math.max(prev.x2, curr.x2);
+          prev.y1 = (prev.y1 + curr.y1) / 2;
+          prev.y2 = prev.y1;
+        } else {
+          prev.y1 = Math.min(prev.y1, curr.y1);
+          prev.y2 = Math.max(prev.y2, curr.y2);
+          prev.x1 = (prev.x1 + curr.x1) / 2;
+          prev.x2 = prev.x1;
+        }
+        prev.lineWidth = Math.max(prev.lineWidth, curr.lineWidth);
+        continue;
+      }
+    }
+    result.push(curr);
+  }
+  return result;
+}
+function filterPageBorderLines(horizontals, verticals, pageWidth, pageHeight) {
+  const margin = 5;
+  return {
+    horizontals: horizontals.filter(
+      (l) => !(Math.abs(l.y1) < margin || Math.abs(l.y1 - pageHeight) < margin) || l.x2 - l.x1 < pageWidth * 0.9
+    ),
+    verticals: verticals.filter(
+      (l) => !(Math.abs(l.x1) < margin || Math.abs(l.x1 - pageWidth) < margin) || l.y2 - l.y1 < pageHeight * 0.9
+    )
+  };
+}
+function buildVertices(horizontals, verticals) {
+  const vertices = [];
+  const tol = CONNECT_TOL;
+  for (const h of horizontals) {
+    for (const v of verticals) {
+      if (v.x1 >= h.x1 - tol && v.x1 <= h.x2 + tol && h.y1 >= v.y1 - tol && h.y1 <= v.y2 + tol) {
+        const radius = Math.max(h.lineWidth, v.lineWidth, 1);
+        vertices.push({ x: v.x1, y: h.y1, radius });
+      }
+    }
+  }
+  return vertices;
+}
+function mergeVertices(vertices) {
+  if (vertices.length <= 1) return vertices;
+  const merged = [];
+  const used = new Array(vertices.length).fill(false);
+  for (let i = 0; i < vertices.length; i++) {
+    if (used[i]) continue;
+    let sumX = vertices[i].x, sumY = vertices[i].y;
+    let maxRadius = vertices[i].radius;
+    let count = 1;
+    for (let j = i + 1; j < vertices.length; j++) {
+      if (used[j]) continue;
+      const mergeTol = VERTEX_MERGE_FACTOR * Math.max(maxRadius, vertices[j].radius);
+      if (Math.abs(vertices[i].x - vertices[j].x) <= mergeTol && Math.abs(vertices[i].y - vertices[j].y) <= mergeTol) {
+        sumX += vertices[j].x;
+        sumY += vertices[j].y;
+        maxRadius = Math.max(maxRadius, vertices[j].radius);
+        count++;
+        used[j] = true;
+      }
+    }
+    merged.push({ x: sumX / count, y: sumY / count, radius: maxRadius });
+  }
+  return merged;
+}
+function buildTableGrids(horizontals, verticals) {
+  if (horizontals.length < 2 || verticals.length < 2) return [];
+  const allVertices = buildVertices(horizontals, verticals);
+  const vertices = mergeVertices(allVertices);
+  if (vertices.length < 4) return [];
+  const globalRadius = vertices.reduce((max, v) => Math.max(max, v.radius), 1);
+  const allLines = [
+    ...horizontals.map((l, i) => ({ ...l, type: "h", id: i })),
+    ...verticals.map((l, i) => ({ ...l, type: "v", id: i + horizontals.length }))
+  ];
+  const groups = groupConnectedLines(allLines);
+  const grids = [];
+  for (const group of groups) {
+    const hLines = group.filter((l) => l.type === "h");
+    const vLines = group.filter((l) => l.type === "v");
+    if (hLines.length < 2 || vLines.length < 2) continue;
+    let gx1 = Infinity, gy1 = Infinity, gx2 = -Infinity, gy2 = -Infinity;
+    for (const l of vLines) {
+      if (l.x1 < gx1) gx1 = l.x1;
+      if (l.x1 > gx2) gx2 = l.x1;
+    }
+    for (const l of hLines) {
+      if (l.y1 < gy1) gy1 = l.y1;
+      if (l.y1 > gy2) gy2 = l.y1;
+    }
+    const groupBbox = {
+      x1: gx1 - CONNECT_TOL,
+      y1: gy1 - CONNECT_TOL,
+      x2: gx2 + CONNECT_TOL,
+      y2: gy2 + CONNECT_TOL
+    };
+    const groupVertices = vertices.filter(
+      (v) => v.x >= groupBbox.x1 && v.x <= groupBbox.x2 && v.y >= groupBbox.y1 && v.y <= groupBbox.y2
+    );
+    const groupRadius = groupVertices.length > 0 ? groupVertices.reduce((max, v) => Math.max(max, v.radius), 1) : globalRadius;
+    const coordMergeTol = Math.max(VERTEX_MERGE_FACTOR * groupRadius, MIN_COORD_MERGE_TOL);
+    const rawYs = [
+      ...hLines.map((l) => l.y1),
+      ...groupVertices.map((v) => v.y)
+    ];
+    const rowYs = clusterCoordinates(rawYs, coordMergeTol).sort((a, b) => b - a);
+    const rawXs = [
+      ...vLines.map((l) => l.x1),
+      ...groupVertices.map((v) => v.x)
+    ];
+    const colXs = clusterCoordinates(rawXs, coordMergeTol).sort((a, b) => a - b);
+    if (rowYs.length < 2 || colXs.length < 2) continue;
+    const validColXs = enforceMinWidth(colXs, MIN_COL_WIDTH);
+    const validRowYs = enforceMinHeight(rowYs, MIN_ROW_HEIGHT);
+    if (validRowYs.length < 2 || validColXs.length < 2) continue;
+    const bbox = {
+      x1: validColXs[0],
+      y1: validRowYs[validRowYs.length - 1],
+      x2: validColXs[validColXs.length - 1],
+      y2: validRowYs[0]
+    };
+    grids.push({ rowYs: validRowYs, colXs: validColXs, bbox, vertexRadius: groupRadius });
+  }
+  return mergeAdjacentGrids(grids);
+}
+function enforceMinWidth(colXs, minWidth) {
+  if (colXs.length <= 2) return colXs;
+  const result = [colXs[0]];
+  for (let i = 1; i < colXs.length; i++) {
+    const prevX = result[result.length - 1];
+    if (colXs[i] - prevX < minWidth && i < colXs.length - 1) {
+      continue;
+    }
+    result.push(colXs[i]);
+  }
+  return result;
+}
+function enforceMinHeight(rowYs, minHeight) {
+  if (rowYs.length <= 2) return rowYs;
+  const result = [rowYs[0]];
+  for (let i = 1; i < rowYs.length; i++) {
+    const prevY = result[result.length - 1];
+    if (prevY - rowYs[i] < minHeight && i < rowYs.length - 1) {
+      continue;
+    }
+    result.push(rowYs[i]);
+  }
+  return result;
+}
+function mergeAdjacentGrids(grids) {
+  if (grids.length <= 1) return grids;
+  const sorted = [...grids].sort((a, b) => b.bbox.y2 - a.bbox.y2);
+  const merged = [sorted[0]];
+  for (let i = 1; i < sorted.length; i++) {
+    const prev = merged[merged.length - 1];
+    const curr = sorted[i];
+    if (prev.colXs.length === curr.colXs.length) {
+      const mergeTol = Math.max(VERTEX_MERGE_FACTOR * Math.max(prev.vertexRadius, curr.vertexRadius), 6) * 3;
+      const colMatch = prev.colXs.every((x, ci) => Math.abs(x - curr.colXs[ci]) <= mergeTol);
+      const verticalGap = prev.bbox.y1 - curr.bbox.y2;
+      if (colMatch && verticalGap >= -CONNECT_TOL && verticalGap <= 20) {
+        const allRowYs = [.../* @__PURE__ */ new Set([...prev.rowYs, ...curr.rowYs])].sort((a, b) => b - a);
+        merged[merged.length - 1] = {
+          rowYs: allRowYs,
+          colXs: prev.colXs,
+          bbox: {
+            x1: Math.min(prev.bbox.x1, curr.bbox.x1),
+            y1: Math.min(prev.bbox.y1, curr.bbox.y1),
+            x2: Math.max(prev.bbox.x2, curr.bbox.x2),
+            y2: Math.max(prev.bbox.y2, curr.bbox.y2)
+          },
+          vertexRadius: Math.max(prev.vertexRadius, curr.vertexRadius)
+        };
+        continue;
+      }
+    }
+    merged.push(curr);
+  }
+  return merged;
+}
+function clusterCoordinates(values, tolerance) {
+  if (values.length === 0) return [];
+  const sorted = [...values].sort((a, b) => a - b);
+  const clusters = [{ sum: sorted[0], count: 1 }];
+  for (let i = 1; i < sorted.length; i++) {
+    const last = clusters[clusters.length - 1];
+    const avg = last.sum / last.count;
+    if (Math.abs(sorted[i] - avg) <= tolerance) {
+      last.sum += sorted[i];
+      last.count++;
+    } else {
+      clusters.push({ sum: sorted[i], count: 1 });
+    }
+  }
+  return clusters.map((c) => c.sum / c.count);
+}
+function groupConnectedLines(lines) {
+  const parent = lines.map((_, i) => i);
+  function find(x) {
+    while (parent[x] !== x) {
+      parent[x] = parent[parent[x]];
+      x = parent[x];
+    }
+    return x;
+  }
+  function union(a, b) {
+    const ra = find(a), rb = find(b);
+    if (ra !== rb) parent[ra] = rb;
+  }
+  for (let i = 0; i < lines.length; i++) {
+    for (let j = i + 1; j < lines.length; j++) {
+      if (linesIntersect(lines[i], lines[j])) {
+        union(i, j);
+      }
+    }
+  }
+  const groups = /* @__PURE__ */ new Map();
+  for (let i = 0; i < lines.length; i++) {
+    const root = find(i);
+    if (!groups.has(root)) groups.set(root, []);
+    groups.get(root).push(lines[i]);
+  }
+  return [...groups.values()];
+}
+function linesIntersect(a, b) {
+  if (a.type === b.type) {
+    if (a.type === "h") {
+      if (Math.abs(a.y1 - b.y1) > CONNECT_TOL) return false;
+      return Math.min(a.x2, b.x2) >= Math.max(a.x1, b.x1) - CONNECT_TOL;
+    } else {
+      if (Math.abs(a.x1 - b.x1) > CONNECT_TOL) return false;
+      return Math.min(a.y2, b.y2) >= Math.max(a.y1, b.y1) - CONNECT_TOL;
+    }
+  }
+  const h = a.type === "h" ? a : b;
+  const v = a.type === "h" ? b : a;
+  const tol = CONNECT_TOL;
+  return v.x1 >= h.x1 - tol && v.x1 <= h.x2 + tol && h.y1 >= v.y1 - tol && h.y1 <= v.y2 + tol;
+}
+function extractCells(grid, horizontals, verticals) {
+  const { rowYs, colXs } = grid;
+  const numRows = rowYs.length - 1;
+  const numCols = colXs.length - 1;
+  if (numRows <= 0 || numCols <= 0) return [];
+  const vBorders = Array.from(
+    { length: numRows },
+    (_, r) => Array.from(
+      { length: numCols + 1 },
+      (_2, c) => hasVerticalLine(verticals, colXs[c], rowYs[r], rowYs[r + 1], grid.vertexRadius)
+    )
+  );
+  const hBorders = Array.from(
+    { length: numRows + 1 },
+    (_, r) => Array.from(
+      { length: numCols },
+      (_2, c) => hasHorizontalLine(horizontals, rowYs[r], colXs[c], colXs[c + 1], grid.vertexRadius)
+    )
+  );
+  const occupied = Array.from({ length: numRows }, () => Array(numCols).fill(false));
+  const cells = [];
+  for (let r = 0; r < numRows; r++) {
+    for (let c = 0; c < numCols; c++) {
+      if (occupied[r][c]) continue;
+      let colSpan = 1;
+      let rowSpan = 1;
+      while (c + colSpan < numCols && !vBorders[r][c + colSpan]) {
+        let canExpand = true;
+        for (let dr = 0; dr < rowSpan; dr++) {
+          if (vBorders[r + dr][c + colSpan]) {
+            canExpand = false;
+            break;
+          }
+        }
+        if (!canExpand) break;
+        colSpan++;
+      }
+      while (r + rowSpan < numRows) {
+        let hasLine = false;
+        for (let dc = 0; dc < colSpan; dc++) {
+          if (hBorders[r + rowSpan][c + dc]) {
+            hasLine = true;
+            break;
+          }
+        }
+        if (hasLine) break;
+        rowSpan++;
+      }
+      for (let dr = 0; dr < rowSpan; dr++) {
+        for (let dc = 0; dc < colSpan; dc++) {
+          occupied[r + dr][c + dc] = true;
+        }
+      }
+      cells.push({
+        row: r,
+        col: c,
+        rowSpan,
+        colSpan,
+        bbox: {
+          x1: colXs[c],
+          y1: rowYs[r + rowSpan],
+          x2: colXs[c + colSpan],
+          y2: rowYs[r]
+        }
+      });
+    }
+  }
+  return cells;
+}
+function hasVerticalLine(verticals, x, topY, botY, vertexRadius) {
+  const tol = Math.max(VERTEX_MERGE_FACTOR * vertexRadius, 4);
+  for (const v of verticals) {
+    if (Math.abs(v.x1 - x) <= tol) {
+      const cellH = Math.abs(topY - botY);
+      if (cellH < 0.1) continue;
+      const overlapTop = Math.min(v.y2, topY);
+      const overlapBot = Math.max(v.y1, botY);
+      const overlap = overlapTop - overlapBot;
+      if (overlap >= cellH * 0.75) return true;
+    }
+  }
+  return false;
+}
+function hasHorizontalLine(horizontals, y, leftX, rightX, vertexRadius) {
+  const tol = Math.max(VERTEX_MERGE_FACTOR * vertexRadius, 4);
+  for (const h of horizontals) {
+    if (Math.abs(h.y1 - y) <= tol) {
+      const cellW = Math.abs(rightX - leftX);
+      if (cellW < 0.1) continue;
+      const overlapLeft = Math.max(h.x1, leftX);
+      const overlapRight = Math.min(h.x2, rightX);
+      const overlap = overlapRight - overlapLeft;
+      if (overlap >= cellW * 0.75) return true;
+    }
+  }
+  return false;
+}
+function mapTextToCells(items, cells) {
+  const result = /* @__PURE__ */ new Map();
+  for (const cell of cells) {
+    result.set(cell, []);
+  }
+  for (const item of items) {
+    const pad = CELL_PADDING;
+    let bestCell = null;
+    let bestScore = 0;
+    for (const cell of cells) {
+      const ix1 = Math.max(item.x, cell.bbox.x1 - pad);
+      const ix2 = Math.min(item.x + item.w, cell.bbox.x2 + pad);
+      const iy1 = Math.max(item.y, cell.bbox.y1 - pad);
+      const iy2 = Math.min(item.y + (item.h || item.fontSize), cell.bbox.y2 + pad);
+      if (ix1 >= ix2 || iy1 >= iy2) continue;
+      const intersectArea = (ix2 - ix1) * (iy2 - iy1);
+      const itemArea = Math.max(item.w, 1) * Math.max(item.h || item.fontSize, 1);
+      const score = intersectArea / itemArea;
+      if (score > bestScore) {
+        bestScore = score;
+        bestCell = cell;
+      }
+    }
+    if (bestCell && bestScore > 0.3) {
+      result.get(bestCell).push(item);
+    }
+  }
+  return result;
+}
+function cellTextToString(items) {
+  if (items.length === 0) return "";
+  if (items.length === 1) return items[0].text;
+  const sorted = [...items].sort((a, b) => b.y - a.y || a.x - b.x);
+  const lines = [];
+  let curLine = [sorted[0]];
+  let curY = sorted[0].y;
+  for (let i = 1; i < sorted.length; i++) {
+    const tol = Math.max(3, Math.min(sorted[i].fontSize, curLine[0].fontSize) * 0.6);
+    if (Math.abs(sorted[i].y - curY) <= tol) {
+      curLine.push(sorted[i]);
+    } else {
+      lines.push(curLine);
+      curLine = [sorted[i]];
+      curY = sorted[i].y;
+    }
+  }
+  lines.push(curLine);
+  const textLines = lines.map((line) => {
+    const s = line.sort((a, b) => a.x - b.x);
+    if (s.length === 1) return s[0].text;
+    const evenSpaced = detectEvenSpacedItems(s);
+    let result = s[0].text;
+    for (let j = 1; j < s.length; j++) {
+      if (evenSpaced[j]) {
+        result += s[j].text;
+        continue;
+      }
+      const gap = s[j].x - (s[j - 1].x + s[j - 1].w);
+      const avgFs = (s[j].fontSize + s[j - 1].fontSize) / 2;
+      const prevIsKorean = /[가-힣]$/.test(result);
+      const currIsKorean = /^[가-힣]/.test(s[j].text);
+      if (gap < avgFs * 0.15) {
+        result += s[j].text;
+      } else if (gap < avgFs * 0.35 && (prevIsKorean || currIsKorean)) {
+        result += s[j].text;
+      } else {
+        result += " " + s[j].text;
+      }
+    }
+    return result;
+  });
+  return mergeCellTextLines(textLines);
+}
+function detectEvenSpacedItems(items) {
+  const result = new Array(items.length).fill(false);
+  if (items.length < 3) return result;
+  let runStart = -1;
+  for (let i = 0; i < items.length; i++) {
+    const isShortKorean = /^[가-힣]{1}$/.test(items[i].text) || /^[\d]{1}$/.test(items[i].text);
+    if (isShortKorean && runStart >= 0 && i > 0) {
+      const gap = items[i].x - (items[i - 1].x + items[i - 1].w);
+      const maxRunGap = Math.max(items[i].fontSize * 3, 30);
+      if (gap > maxRunGap) {
+        if (i - runStart >= 3) markEvenRun(items, result, runStart, i);
+        runStart = i;
+        continue;
+      }
+    }
+    if (isShortKorean) {
+      if (runStart < 0) runStart = i;
+    } else {
+      if (runStart >= 0 && i - runStart >= 3) {
+        markEvenRun(items, result, runStart, i);
+      }
+      runStart = -1;
+    }
+  }
+  if (runStart >= 0 && items.length - runStart >= 3) {
+    markEvenRun(items, result, runStart, items.length);
+  }
+  return result;
+}
+function markEvenRun(items, result, start, end) {
+  const gaps = [];
+  for (let i = start + 1; i < end; i++) {
+    gaps.push(items[i].x - (items[i - 1].x + items[i - 1].w));
+  }
+  const posGaps = gaps.filter((g2) => g2 > 0);
+  if (posGaps.length < 2) return;
+  let minGap = Infinity, maxGap = -Infinity;
+  for (const g2 of posGaps) {
+    if (g2 < minGap) minGap = g2;
+    if (g2 > maxGap) maxGap = g2;
+  }
+  const avgFs = items[start].fontSize;
+  if (minGap >= avgFs * 0.1 && maxGap <= avgFs * 3 && maxGap / Math.max(minGap, 0.1) <= 3) {
+    for (let i = start + 1; i < end; i++) {
+      result[i] = true;
+    }
+  }
+}
+function mergeCellTextLines(textLines) {
+  if (textLines.length <= 1) return textLines[0] || "";
+  const merged = [textLines[0]];
+  for (let i = 1; i < textLines.length; i++) {
+    const prev = merged[merged.length - 1];
+    const curr = textLines[i];
+    if (/[가-힣]$/.test(prev) && /^[가-힣]+$/.test(curr) && curr.length <= 8 && !curr.includes(" ")) {
+      merged[merged.length - 1] = prev + curr;
+    } else if (curr.trim().length <= 3 && /^[)\]%}]/.test(curr.trim())) {
+      merged[merged.length - 1] = prev + curr.trim();
+    } else if (/[,(]$/.test(prev.trim()) && curr.trim().length <= 15) {
+      merged[merged.length - 1] = prev + curr.trim();
+    } else if (/[\d,]$/.test(prev) && /^[\d,]+[)\]]?$/.test(curr.trim()) && curr.trim().length <= 10) {
+      merged[merged.length - 1] = prev + curr.trim();
+    } else {
+      merged.push(curr);
+    }
+  }
+  return merged.join("\n");
+}
+
+// src/pdf/cluster-detector.ts
+var Y_TOL = 3;
+var COL_CLUSTER_TOL = 15;
+var MIN_ROWS = 3;
+var MIN_COLS = 2;
+var MIN_GAP_FACTOR = 2;
+var MIN_GAP_ABSOLUTE = 20;
+var MIN_COL_FILL_RATIO = 0.4;
+function detectClusterTables(items, pageNum) {
+  if (items.length < MIN_ROWS * MIN_COLS) return [];
+  const { merged, originMap } = mergeEvenSpacedClusters(items);
+  const rows = groupByBaseline(merged);
+  if (rows.length < MIN_ROWS) return [];
+  const results = [];
+  const headerResult = detectHeaderRow(rows);
+  if (headerResult) {
+    const { columns, headerIdx } = headerResult;
+    const headerRow = rows[headerIdx];
+    const headerItems = [...headerRow.items].sort((a, b) => a.x - b.x);
+    const headerAndBelow = rows.slice(headerIdx);
+    const mergedRows = mergeMultiLineRows(headerAndBelow, columns);
+    const tableRegions = findTableRegionsByHeader(mergedRows, columns, headerItems);
+    for (const region of tableRegions) {
+      const table = buildClusterTable(region.rows, columns, pageNum);
+      if (table) {
+        expandUsedItems(table.usedItems, originMap);
+        results.push(table);
+      }
+    }
+  }
+  if (results.length === 0) {
+    const suspiciousRows = rows.filter((row) => hasSuspiciousGaps(row));
+    if (suspiciousRows.length >= MIN_ROWS) {
+      const columns = extractColumnClusters(suspiciousRows);
+      if (columns.length >= MIN_COLS) {
+        const tableRegions = findTableRegions(rows, columns);
+        for (const region of tableRegions) {
+          const mergedRows = mergeMultiLineRows(region.rows, columns);
+          const table = buildClusterTable(mergedRows, columns, pageNum);
+          if (table) {
+            expandUsedItems(table.usedItems, originMap);
+            results.push(table);
+          }
+        }
+      }
+    }
+  }
+  return results;
+}
+function mergeEvenSpacedClusters(items) {
+  const originMap = /* @__PURE__ */ new Map();
+  const rows = groupByBaseline(items);
+  const merged = [];
+  for (const row of rows) {
+    const sorted = [...row.items].sort((a, b) => a.x - b.x);
+    let i = 0;
+    while (i < sorted.length) {
+      if (/^[가-힣\d]$/.test(sorted[i].text)) {
+        let runEnd = i + 1;
+        while (runEnd < sorted.length && /^[가-힣\d]$/.test(sorted[runEnd].text)) {
+          const gap = sorted[runEnd].x - (sorted[runEnd - 1].x + sorted[runEnd - 1].w);
+          const fs = sorted[runEnd].fontSize;
+          if (gap < fs * 0.1 || gap > fs * 3) break;
+          runEnd++;
+        }
+        if (runEnd - i >= 3) {
+          const gaps = [];
+          for (let g2 = i + 1; g2 < runEnd; g2++) {
+            gaps.push(sorted[g2].x - (sorted[g2 - 1].x + sorted[g2 - 1].w));
+          }
+          let minG = Infinity, maxG = -Infinity;
+          for (const g2 of gaps) {
+            if (g2 < minG) minG = g2;
+            if (g2 > maxG) maxG = g2;
+          }
+          if (minG > 0 && maxG / minG <= 3) {
+            const run = sorted.slice(i, runEnd);
+            const text = run.map((r) => r.text).join("");
+            const first = run[0], last = run[runEnd - i - 1];
+            const item = {
+              text,
+              x: first.x,
+              y: first.y,
+              w: last.x + last.w - first.x,
+              h: first.h,
+              fontSize: first.fontSize,
+              fontName: first.fontName
+            };
+            originMap.set(item, run);
+            merged.push(item);
+            i = runEnd;
+            continue;
+          }
+        }
+      }
+      merged.push(sorted[i]);
+      i++;
+    }
+  }
+  return { merged, originMap };
+}
+function expandUsedItems(usedItems, originMap) {
+  const toAdd = [];
+  for (const item of usedItems) {
+    const origins = originMap.get(item);
+    if (origins) for (const o of origins) toAdd.push(o);
+  }
+  for (const a of toAdd) usedItems.add(a);
+}
+function detectHeaderRow(rows) {
+  const allItems = rows.flatMap((r) => r.items);
+  if (allItems.length === 0) return null;
+  let allMinX = Infinity, allMaxX = -Infinity;
+  for (const i of allItems) {
+    if (i.x < allMinX) allMinX = i.x;
+    const r = i.x + i.w;
+    if (r > allMaxX) allMaxX = r;
+  }
+  const pageSpan = allMaxX - allMinX;
+  if (pageSpan <= 0) return null;
+  for (let ri = 0; ri < rows.length; ri++) {
+    const row = rows[ri];
+    if (row.items.length < MIN_COLS || row.items.length > 6) continue;
+    if (row.items.some((i) => i.text.length > 8)) continue;
+    if (!row.items.some((i) => /[가-힣]/.test(i.text))) continue;
+    if (row.items.some((i) => /^[□■○●·※▶▷◆◇\-]/.test(i.text))) continue;
+    const sorted = [...row.items].sort((a, b) => a.x - b.x);
+    const xSpan = sorted[sorted.length - 1].x + sorted[sorted.length - 1].w - sorted[0].x;
+    if (xSpan / pageSpan < 0.4) continue;
+    const avgFs = sorted.reduce((s, i) => s + i.fontSize, 0) / sorted.length;
+    let hasLargeGap = false;
+    for (let i = 1; i < sorted.length; i++) {
+      const gap = sorted[i].x - (sorted[i - 1].x + sorted[i - 1].w);
+      if (gap >= avgFs * 2.5) {
+        hasLargeGap = true;
+        break;
+      }
+    }
+    if (!hasLargeGap) continue;
+    const columns = sorted.map((item) => ({ x: item.x, count: 0 }));
+    let matchCount = 0;
+    for (let j = ri + 1; j < rows.length && matchCount < MIN_ROWS + 2; j++) {
+      const matched = countMatchedColumnsRange(rows[j], columns, sorted);
+      if (matched >= MIN_COLS) matchCount++;
+    }
+    if (matchCount < MIN_ROWS) continue;
+    return { columns, headerIdx: ri };
+  }
+  return null;
+}
+function mergeMultiLineRows(rows, columns) {
+  if (rows.length <= 1) return rows;
+  const result = [rows[0]];
+  const allFontSizes = rows.flatMap((r) => r.items).map((i) => i.fontSize);
+  const avgFontSize = allFontSizes.length > 0 ? allFontSizes.reduce((s, v) => s + v, 0) / allFontSizes.length : 12;
+  for (let i = 1; i < rows.length; i++) {
+    const prev = result[result.length - 1];
+    const curr = rows[i];
+    const yGap = Math.abs(prev.y - curr.y);
+    const matchedCols = countMatchedColumns(curr, columns);
+    if (yGap < avgFontSize * 1.8 && curr.items.length <= 2 && (matchedCols < MIN_COLS || curr.items.length === 1)) {
+      result[result.length - 1] = {
+        y: prev.y,
+        items: [...prev.items, ...curr.items]
+      };
+    } else {
+      result.push(curr);
+    }
+  }
+  return result;
+}
+function groupByBaseline(items) {
+  if (items.length === 0) return [];
+  const sorted = [...items].sort((a, b) => b.y - a.y || a.x - b.x);
+  const rows = [];
+  let curItems = [sorted[0]];
+  let curY = sorted[0].y;
+  for (let i = 1; i < sorted.length; i++) {
+    if (Math.abs(sorted[i].y - curY) <= Y_TOL) {
+      curItems.push(sorted[i]);
+    } else {
+      rows.push({ y: curY, items: curItems });
+      curItems = [sorted[i]];
+      curY = sorted[i].y;
+    }
+  }
+  if (curItems.length > 0) rows.push({ y: curY, items: curItems });
+  return rows;
+}
+function hasSuspiciousGaps(row) {
+  if (row.items.length < 2) return false;
+  const sorted = [...row.items].sort((a, b) => a.x - b.x);
+  if (sorted.length === 2 && sorted[1].text.length > 20) return false;
+  const avgFontSize = sorted.reduce((s, i) => s + i.fontSize, 0) / sorted.length;
+  const minGap = Math.max(avgFontSize * MIN_GAP_FACTOR, MIN_GAP_ABSOLUTE);
+  for (let i = 1; i < sorted.length; i++) {
+    const gap = sorted[i].x - (sorted[i - 1].x + sorted[i - 1].w);
+    if (gap >= minGap) return true;
+  }
+  return false;
+}
+function extractColumnClusters(rows) {
+  const allX = [];
+  for (const row of rows) {
+    for (const item of row.items) allX.push(item.x);
+  }
+  if (allX.length === 0) return [];
+  allX.sort((a, b) => a - b);
+  const clusters = [];
+  let clusterStart = 0;
+  for (let i = 1; i <= allX.length; i++) {
+    if (i === allX.length || allX[i] - allX[i - 1] > COL_CLUSTER_TOL) {
+      const slice = allX.slice(clusterStart, i);
+      const avg = Math.round(slice.reduce((s, v) => s + v, 0) / slice.length);
+      clusters.push({ x: avg, count: slice.length });
+      clusterStart = i;
+    }
+  }
+  const minCount = Math.max(2, Math.floor(rows.length * MIN_COL_FILL_RATIO));
+  return clusters.filter((c) => c.count >= minCount).sort((a, b) => a.x - b.x);
+}
+function findTableRegionsByHeader(allRows, columns, headerItems) {
+  const regions = [];
+  let currentRegion = [];
+  let missStreak = 0;
+  for (const row of allRows) {
+    const matchedCols = countMatchedColumnsRange(row, columns, headerItems);
+    if (matchedCols >= MIN_COLS) {
+      currentRegion.push(row);
+      missStreak = 0;
+    } else if (currentRegion.length > 0 && (row.items.length <= 2 || missStreak === 0)) {
+      currentRegion.push(row);
+      missStreak++;
+    } else {
+      while (currentRegion.length > 0) {
+        const last = currentRegion[currentRegion.length - 1];
+        if (countMatchedColumnsRange(last, columns, headerItems) >= MIN_COLS) break;
+        currentRegion.pop();
+      }
+      if (currentRegion.length >= MIN_ROWS) {
+        regions.push({ rows: [...currentRegion] });
+      }
+      currentRegion = [];
+      missStreak = 0;
+    }
+  }
+  while (currentRegion.length > 0) {
+    const last = currentRegion[currentRegion.length - 1];
+    if (countMatchedColumnsRange(last, columns, headerItems) >= MIN_COLS) break;
+    currentRegion.pop();
+  }
+  if (currentRegion.length >= MIN_ROWS) {
+    regions.push({ rows: currentRegion });
+  }
+  return regions;
+}
+function findTableRegions(allRows, columns) {
+  const regions = [];
+  let currentRegion = [];
+  for (const row of allRows) {
+    const matchedCols = countMatchedColumns(row, columns);
+    if (matchedCols >= MIN_COLS) {
+      currentRegion.push(row);
+    } else if (row.items.length === 1) {
+      if (currentRegion.length > 0) {
+        currentRegion.push(row);
+      }
+    } else {
+      if (currentRegion.length >= MIN_ROWS) {
+        regions.push({ rows: [...currentRegion] });
+      }
+      currentRegion = [];
+    }
+  }
+  if (currentRegion.length >= MIN_ROWS) {
+    regions.push({ rows: currentRegion });
+  }
+  return regions;
+}
+function countMatchedColumns(row, columns) {
+  const matched = /* @__PURE__ */ new Set();
+  for (const item of row.items) {
+    for (let ci = 0; ci < columns.length; ci++) {
+      if (Math.abs(item.x - columns[ci].x) <= COL_CLUSTER_TOL * 2) {
+        matched.add(ci);
+        break;
+      }
+    }
+  }
+  return matched.size;
+}
+function countMatchedColumnsRange(row, columns, headerItems) {
+  const boundaries = [];
+  for (let ci = 0; ci < headerItems.length; ci++) {
+    const left = ci === 0 ? 0 : (headerItems[ci - 1].x + headerItems[ci - 1].w + headerItems[ci].x) / 2;
+    const right = ci === headerItems.length - 1 ? Infinity : (headerItems[ci].x + headerItems[ci].w + headerItems[ci + 1].x) / 2;
+    boundaries.push({ left, right });
+  }
+  const matched = /* @__PURE__ */ new Set();
+  for (const item of row.items) {
+    for (let ci = 0; ci < boundaries.length; ci++) {
+      if (item.x >= boundaries[ci].left && item.x < boundaries[ci].right) {
+        matched.add(ci);
+        break;
+      }
+    }
+  }
+  return matched.size;
+}
+function assignRowItems(items, columns, numCols) {
+  if (items.length === 0) return [];
+  const sorted = [...items].sort((a, b) => a.x - b.x);
+  const colCenters = columns.map((c) => c.x);
+  const gaps = [];
+  for (let i = 1; i < sorted.length; i++) {
+    gaps.push({ idx: i, size: sorted[i].x - (sorted[i - 1].x + sorted[i - 1].w) });
+  }
+  const gapSizes = gaps.map((g2) => g2.size).sort((a, b) => a - b);
+  const medianGap = gapSizes.length > 0 ? gapSizes[Math.floor(gapSizes.length / 2)] : 0;
+  const gapThreshold = sorted.length <= numCols + 1 ? 12 : Math.max(medianGap * 2.5, 12);
+  const significantGaps = gaps.filter((g2) => g2.size >= gapThreshold).sort((a, b) => b.size - a.size).slice(0, numCols - 1).sort((a, b) => a.idx - b.idx);
+  const groups = [];
+  let start = 0;
+  for (const gap of significantGaps) {
+    groups.push(sorted.slice(start, gap.idx));
+    start = gap.idx;
+  }
+  groups.push(sorted.slice(start));
+  const result = [];
+  const usedCols = /* @__PURE__ */ new Set();
+  const groupCenters = groups.map((g2) => {
+    let minX = Infinity, maxX = -Infinity;
+    for (const i of g2) {
+      if (i.x < minX) minX = i.x;
+      const r = i.x + i.w;
+      if (r > maxX) maxX = r;
+    }
+    return (minX + maxX) / 2;
+  });
+  const assignments = [];
+  for (let gi = 0; gi < groups.length; gi++) {
+    for (let ci = 0; ci < numCols; ci++) {
+      assignments.push({ gi, ci, dist: Math.abs(groupCenters[gi] - colCenters[ci]) });
+    }
+  }
+  assignments.sort((a, b) => a.dist - b.dist);
+  const assignedGroups = /* @__PURE__ */ new Set();
+  for (const { gi, ci } of assignments) {
+    if (assignedGroups.has(gi) || usedCols.has(ci)) continue;
+    result.push({ col: ci, items: groups[gi] });
+    assignedGroups.add(gi);
+    usedCols.add(ci);
+  }
+  for (let gi = 0; gi < groups.length; gi++) {
+    if (assignedGroups.has(gi)) continue;
+    let bestCol = 0, bestDist = Infinity;
+    for (let ci = 0; ci < numCols; ci++) {
+      const d = Math.abs(groupCenters[gi] - colCenters[ci]);
+      if (d < bestDist) {
+        bestDist = d;
+        bestCol = ci;
+      }
+    }
+    result.push({ col: bestCol, items: groups[gi] });
+  }
+  return result;
+}
+function buildClusterTable(rows, columns, pageNum) {
+  const numCols = columns.length;
+  const numRows = rows.length;
+  if (numRows < MIN_ROWS || numCols < MIN_COLS) return null;
+  const cells = Array.from(
+    { length: numRows },
+    () => Array.from({ length: numCols }, () => ({ text: "", colSpan: 1, rowSpan: 1 }))
+  );
+  const usedItems = /* @__PURE__ */ new Set();
+  for (let r = 0; r < numRows; r++) {
+    const row = rows[r];
+    if (row.items.length === 1 && numCols > 1) {
+      cells[r][0] = { text: row.items[0].text, colSpan: numCols, rowSpan: 1 };
+      usedItems.add(row.items[0]);
+      continue;
+    }
+    const assignments = assignRowItems(row.items, columns, numCols);
+    for (const { col, items } of assignments) {
+      const text = items.map((i) => i.text).join(" ");
+      const existing = cells[r][col].text;
+      cells[r][col].text = existing ? existing + " " + text : text;
+      for (const item of items) usedItems.add(item);
+    }
+  }
+  let emptyRows = 0;
+  for (const row of cells) {
+    if (row.every((c) => c.text === "")) emptyRows++;
+  }
+  if (emptyRows > numRows * 0.5) return null;
+  for (let c = 0; c < numCols; c++) {
+    const hasValue = cells.some((row) => row[c].text !== "");
+    if (!hasValue) return null;
+  }
+  for (let r = numRows - 1; r >= 1; r--) {
+    const nonEmptyCols = cells[r].filter((c) => c.text.trim()).length;
+    if (nonEmptyCols !== 1) continue;
+    if (cells[r][0].text.trim() !== "") continue;
+    const contentText = _optionalChain([cells, 'access', _4 => _4[r], 'access', _5 => _5.find, 'call', _6 => _6((c) => c.text.trim()), 'optionalAccess', _7 => _7.text, 'access', _8 => _8.trim, 'call', _9 => _9()]) || "";
+    if (/^[○●▶\-·]/.test(contentText)) continue;
+    for (let pr = r - 1; pr >= 0; pr--) {
+      if (cells[pr].some((c) => c.text.trim())) {
+        for (let c = 0; c < numCols; c++) {
+          const prev = cells[pr][c].text.trim();
+          const curr = cells[r][c].text.trim();
+          if (curr) cells[pr][c].text = prev ? prev + " " + curr : curr;
+        }
+        for (let c = 0; c < numCols; c++) cells[r][c].text = "";
+        break;
+      }
+    }
+  }
+  for (let r = 0; r < cells.length - 1; r++) {
+    const row = cells[r];
+    const hasCol0 = row[0].text.trim() !== "";
+    const hasColLast = numCols > 1 && row[numCols - 1].text.trim() !== "";
+    const midEmpty = row.slice(1, numCols - 1).every((c) => c.text.trim() === "");
+    if (hasCol0 && hasColLast && midEmpty) {
+      const next = cells[r + 1];
+      if (next[0].text.trim() === "" && next.some((c) => c.text.trim())) {
+        for (let c = 1; c < numCols; c++) {
+          const curr = next[c].text.trim();
+          if (curr) row[c].text = row[c].text.trim() ? row[c].text.trim() + " " + curr : curr;
+        }
+        for (let c = 0; c < numCols; c++) next[c].text = "";
+      }
+    }
+  }
+  const filteredCells = cells.filter((row) => row.some((c) => c.text.trim()));
+  const finalRowCount = filteredCells.length;
+  if (finalRowCount < MIN_ROWS) return null;
+  const irTable = {
+    rows: finalRowCount,
+    cols: numCols,
+    cells: filteredCells,
+    hasHeader: finalRowCount > 1
+  };
+  const allItems = rows.flatMap((r) => r.items);
+  let minX = Infinity, minY = Infinity, maxX = -Infinity, maxY = -Infinity;
+  for (const i of allItems) {
+    if (i.x < minX) minX = i.x;
+    if (i.y < minY) minY = i.y;
+    if (i.x + i.w > maxX) maxX = i.x + i.w;
+    const h = i.h > 0 ? i.h : i.fontSize;
+    if (i.y + h > maxY) maxY = i.y + h;
+  }
+  return {
+    table: irTable,
+    bbox: { page: pageNum, x: minX, y: minY, width: maxX - minX, height: maxY - minY },
+    usedItems
+  };
+}
+
+// src/pdf/polyfill.ts
+var _pdfworkermjs = require('pdfjs-dist/legacy/build/pdf.worker.mjs'); var pdfjsWorker = _interopRequireWildcard(_pdfworkermjs);
+var g = globalThis;
+if (typeof g.DOMMatrix === "undefined") {
+  g.DOMMatrix = (_class = class DOMMatrix {
+    __init() {this.m = [1, 0, 0, 1, 0, 0]}
+    constructor(init) {;_class.prototype.__init.call(this);
+      if (init) this.m = init;
+    }
+  }, _class);
+}
+if (typeof g.Path2D === "undefined") {
+  g.Path2D = class Path2D {
+  };
+}
+g.pdfjsWorker = pdfjsWorker;
+
+// src/pdf/parser.ts
+
+_pdfmjs.GlobalWorkerOptions.workerSrc = "";
+var MAX_PAGES = 5e3;
+var MAX_TOTAL_TEXT = 100 * 1024 * 1024;
+var PDF_LOAD_TIMEOUT_MS = 3e4;
+async function loadPdfWithTimeout(buffer) {
+  const loadingTask = _pdfmjs.getDocument.call(void 0, {
+    data: new Uint8Array(buffer),
+    useSystemFonts: true,
+    disableFontFace: true,
+    isEvalSupported: false
+  });
+  let timer;
+  try {
+    return await Promise.race([
+      loadingTask.promise,
+      new Promise((_, reject) => {
+        timer = setTimeout(() => {
+          loadingTask.destroy();
+          reject(new (0, _chunkOJ4QR33Vcjs.KordocError)("PDF \uB85C\uB529 \uD0C0\uC784\uC544\uC6C3 (30\uCD08 \uCD08\uACFC)"));
+        }, PDF_LOAD_TIMEOUT_MS);
+      })
+    ]);
+  } finally {
+    if (timer !== void 0) clearTimeout(timer);
+  }
+}
+async function parsePdfDocument(buffer, options) {
+  const doc = await loadPdfWithTimeout(buffer);
+  try {
+    const pageCount = doc.numPages;
+    if (pageCount === 0) throw new (0, _chunkOJ4QR33Vcjs.KordocError)("PDF\uC5D0 \uD398\uC774\uC9C0\uAC00 \uC5C6\uC2B5\uB2C8\uB2E4.");
+    const metadata = { pageCount };
+    await extractPdfMetadata(doc, metadata);
+    const blocks = [];
+    const warnings = [];
+    let totalChars = 0;
+    let totalTextBytes = 0;
+    const effectivePageCount = Math.min(pageCount, MAX_PAGES);
+    const pageFilter = _optionalChain([options, 'optionalAccess', _10 => _10.pages]) ? _chunkMUOQXDZ4cjs.parsePageRange.call(void 0, options.pages, effectivePageCount) : null;
+    const totalTarget = pageFilter ? pageFilter.size : effectivePageCount;
+    const fontSizeFreq = /* @__PURE__ */ new Map();
+    const pageHeights = /* @__PURE__ */ new Map();
+    let parsedPages = 0;
+    for (let i = 1; i <= effectivePageCount; i++) {
+      if (pageFilter && !pageFilter.has(i)) continue;
+      try {
+        const page = await doc.getPage(i);
+        const tc = await page.getTextContent();
+        const viewport = page.getViewport({ scale: 1 });
+        pageHeights.set(i, viewport.height);
+        const rawItems = tc.items;
+        const items = normalizeItems(rawItems);
+        const { visible, hiddenCount } = filterHiddenText(items, viewport.width, viewport.height);
+        if (hiddenCount > 0) {
+          warnings.push({ page: i, message: `${hiddenCount}\uAC1C \uC228\uACA8\uC9C4 \uD14D\uC2A4\uD2B8 \uC694\uC18C \uD544\uD130\uB9C1\uB428`, code: "HIDDEN_TEXT_FILTERED" });
+        }
+        for (const item of visible) {
+          if (item.fontSize > 0) fontSizeFreq.set(item.fontSize, (fontSizeFreq.get(item.fontSize) || 0) + 1);
+        }
+        const opList = await page.getOperatorList();
+        const pageBlocks = extractPageBlocksWithLines(visible, i, opList, viewport.width, viewport.height);
+        for (const b of pageBlocks) blocks.push(b);
+        for (const b of pageBlocks) {
+          const t = b.text || "";
+          totalChars += t.replace(/\s/g, "").length;
+          totalTextBytes += t.length * 2;
+        }
+        if (totalTextBytes > MAX_TOTAL_TEXT) throw new (0, _chunkOJ4QR33Vcjs.KordocError)("\uD14D\uC2A4\uD2B8 \uCD94\uCD9C \uD06C\uAE30 \uCD08\uACFC");
+        parsedPages++;
+        _optionalChain([options, 'optionalAccess', _11 => _11.onProgress, 'optionalCall', _12 => _12(parsedPages, totalTarget)]);
+      } catch (pageErr) {
+        if (pageErr instanceof _chunkOJ4QR33Vcjs.KordocError) throw pageErr;
+        warnings.push({ page: i, message: `\uD398\uC774\uC9C0 ${i} \uD30C\uC2F1 \uC2E4\uD328: ${pageErr instanceof Error ? pageErr.message : "\uC54C \uC218 \uC5C6\uB294 \uC624\uB958"}`, code: "PARTIAL_PARSE" });
+      }
+    }
+    const parsedPageCount = parsedPages || (pageFilter ? pageFilter.size : effectivePageCount);
+    if (totalChars / Math.max(parsedPageCount, 1) < 10) {
+      if (_optionalChain([options, 'optionalAccess', _13 => _13.ocr])) {
+        try {
+          const { ocrPages } = await Promise.resolve().then(() => _interopRequireWildcard(require("./provider-YN2SSK4X.cjs")));
+          const ocrBlocks = await ocrPages(doc, options.ocr, pageFilter, effectivePageCount);
+          if (ocrBlocks.length > 0) {
+            const ocrMarkdown = ocrBlocks.map((b) => b.text || "").filter(Boolean).join("\n\n");
+            return { markdown: ocrMarkdown, blocks: ocrBlocks, metadata, warnings, isImageBased: true };
+          }
+        } catch (e) {
+        }
+      }
+      throw Object.assign(new (0, _chunkOJ4QR33Vcjs.KordocError)(`\uC774\uBBF8\uC9C0 \uAE30\uBC18 PDF (${pageCount}\uD398\uC774\uC9C0, ${totalChars}\uC790)`), { isImageBased: true });
+    }
+    if (_optionalChain([options, 'optionalAccess', _14 => _14.removeHeaderFooter]) !== false && parsedPageCount >= 3) {
+      const removed = removeHeaderFooterBlocks(blocks, pageHeights, warnings);
+      for (let ri = removed.length - 1; ri >= 0; ri--) {
+        blocks.splice(removed[ri], 1);
+      }
+    }
+    const medianFontSize = computeMedianFontSizeFromFreq(fontSizeFreq);
+    if (medianFontSize > 0) {
+      detectHeadings(blocks, medianFontSize);
+    }
+    detectMarkerHeadings(blocks);
+    const outline = blocks.filter((b) => b.type === "heading" && b.level && b.text).map((b) => ({ level: b.level, text: b.text, pageNumber: b.pageNumber }));
+    let markdown = cleanPdfText(_chunkOJ4QR33Vcjs.blocksToMarkdown.call(void 0, blocks));
+    return { markdown, blocks, metadata, outline: outline.length > 0 ? outline : void 0, warnings: warnings.length > 0 ? warnings : void 0 };
+  } finally {
+    await doc.destroy().catch(() => {
+    });
+  }
+}
+async function extractPdfMetadata(doc, metadata) {
+  try {
+    const result = await doc.getMetadata();
+    if (!_optionalChain([result, 'optionalAccess', _15 => _15.info])) return;
+    const info = result.info;
+    if (typeof info.Title === "string" && info.Title.trim()) metadata.title = info.Title.trim();
+    if (typeof info.Author === "string" && info.Author.trim()) metadata.author = info.Author.trim();
+    if (typeof info.Creator === "string" && info.Creator.trim()) metadata.creator = info.Creator.trim();
+    if (typeof info.Subject === "string" && info.Subject.trim()) metadata.description = info.Subject.trim();
+    if (typeof info.Keywords === "string" && info.Keywords.trim()) {
+      metadata.keywords = info.Keywords.split(/[,;]/).map((k) => k.trim()).filter(Boolean);
+    }
+    if (typeof info.CreationDate === "string") metadata.createdAt = parsePdfDate(info.CreationDate);
+    if (typeof info.ModDate === "string") metadata.modifiedAt = parsePdfDate(info.ModDate);
+  } catch (e2) {
+  }
+}
+function parsePdfDate(dateStr) {
+  const m = dateStr.match(/D:(\d{4})(\d{2})?(\d{2})?(\d{2})?(\d{2})?(\d{2})?/);
+  if (!m) return void 0;
+  const [, year, month = "01", day = "01", hour = "00", min = "00", sec = "00"] = m;
+  return `${year}-${month}-${day}T${hour}:${min}:${sec}`;
+}
+async function extractPdfMetadataOnly(buffer) {
+  const doc = await loadPdfWithTimeout(buffer);
+  try {
+    const metadata = { pageCount: doc.numPages };
+    await extractPdfMetadata(doc, metadata);
+    return metadata;
+  } finally {
+    await doc.destroy().catch(() => {
+    });
+  }
+}
+function filterHiddenText(items, pageWidth, pageHeight) {
+  let hiddenCount = 0;
+  const visible = [];
+  for (const item of items) {
+    if (item.isHidden) {
+      hiddenCount++;
+      continue;
+    }
+    const margin = Math.max(pageWidth, pageHeight) * 0.1;
+    if (item.x < -margin || item.x > pageWidth + margin || item.y < -margin || item.y > pageHeight + margin) {
+      hiddenCount++;
+      continue;
+    }
+    visible.push(item);
+  }
+  return { visible, hiddenCount };
+}
+function computeMedianFontSizeFromFreq(freq) {
+  if (freq.size === 0) return 0;
+  let total = 0;
+  for (const count of freq.values()) total += count;
+  const sorted = [...freq.entries()].sort((a, b) => a[0] - b[0]);
+  const mid = Math.floor(total / 2);
+  let cumulative = 0;
+  for (const [size, count] of sorted) {
+    cumulative += count;
+    if (cumulative > mid) return size;
+  }
+  return sorted[sorted.length - 1][0];
+}
+function detectHeadings(blocks, medianFontSize) {
+  for (const block of blocks) {
+    if (block.type !== "paragraph" || !block.text || !_optionalChain([block, 'access', _16 => _16.style, 'optionalAccess', _17 => _17.fontSize])) continue;
+    const text = block.text.trim();
+    if (text.length === 0 || text.length > 200) continue;
+    if (/^\d+$/.test(text)) continue;
+    const ratio = block.style.fontSize / medianFontSize;
+    let level = 0;
+    if (ratio >= _chunkOJ4QR33Vcjs.HEADING_RATIO_H1) level = 1;
+    else if (ratio >= _chunkOJ4QR33Vcjs.HEADING_RATIO_H2) level = 2;
+    else if (ratio >= _chunkOJ4QR33Vcjs.HEADING_RATIO_H3) level = 3;
+    if (level > 0) {
+      block.type = "heading";
+      block.level = level;
+      block.text = collapseEvenSpacing(text);
+    }
+  }
+}
+function collapseEvenSpacing(text) {
+  const tokens = text.split(" ");
+  const singleCharCount = tokens.filter((t) => t.length === 1).length;
+  if (tokens.length >= 3 && singleCharCount / tokens.length >= 0.7) {
+    return tokens.join("");
+  }
+  return text.replace(
+    /(?<![가-힣])[가-힣](?: [가-힣\d]){2,}(?![가-힣])/g,
+    (match) => match.replace(/ /g, "")
+  );
+}
+function shouldDemoteTable(table) {
+  const allCells = table.cells.flatMap((row) => row.map((c) => c.text.trim())).filter(Boolean);
+  const allText = allCells.join(" ");
+  if (table.rows <= 3 && table.cols <= 3) {
+    const totalCells2 = table.rows * table.cols;
+    const emptyCells2 = totalCells2 - allCells.length;
+    if (emptyCells2 >= totalCells2 * 0.3) return true;
+    if (/[□■◆○●▶ㅇ]/.test(allText)) return true;
+    if (/<[^>]+>/.test(allText)) return true;
+  }
+  if (allText.length > 200) return false;
+  if (/[□■◆○●▶]/.test(allText) && table.rows <= 3) return true;
+  const totalCells = table.rows * table.cols;
+  const emptyCells = totalCells - allCells.length;
+  if (table.rows <= 2 && emptyCells > totalCells * 0.5) return true;
+  if (table.rows === 1 && !/\d{2,}/.test(allText)) return true;
+  return false;
+}
+function demoteTableToText(table) {
+  const lines = [];
+  for (let r = 0; r < table.rows; r++) {
+    const cells = table.cells[r].map((c) => c.text.trim()).filter(Boolean);
+    if (cells.length === 0) continue;
+    if (table.cols === 2 && cells.length === 2) {
+      lines.push(`${cells[0]} : ${cells[1]}`);
+    } else {
+      lines.push(cells.join(" "));
+    }
+  }
+  return lines.join("\n");
+}
+function detectMarkerHeadings(blocks) {
+  for (let i = 0; i < blocks.length; i++) {
+    const block = blocks[i];
+    if (block.type !== "paragraph" || !block.text) continue;
+    const text = block.text.trim();
+    if (text.length < 50 && /^[□■◆◇▶]\s*[가-힣]/.test(text)) {
+      block.type = "heading";
+      block.level = 4;
+      continue;
+    }
+    if (/^[가-힣]{2,6}$/.test(text) && _optionalChain([block, 'access', _18 => _18.style, 'optionalAccess', _19 => _19.fontSize])) {
+      const prev = blocks[i - 1];
+      const next = blocks[i + 1];
+      const prevIsStructural = !prev || prev.type === "table" || prev.type === "heading" || prev.type === "separator";
+      const nextIsStructural = !next || next.type === "table" || next.type === "heading" || next.type === "paragraph" && next.text && /^[□■◆○●]/.test(next.text.trim());
+      if (prevIsStructural || nextIsStructural) {
+        block.type = "heading";
+        block.level = 3;
+      }
+    }
+  }
+}
+var MAX_XYCUT_DEPTH = 50;
+function xyCutOrder(items, gapThreshold, depth = 0) {
+  if (items.length === 0) return [];
+  if (items.length <= 2 || depth >= MAX_XYCUT_DEPTH) return [items];
+  const region = computeRegion(items);
+  const ySplit = findYSplit(items, region, gapThreshold);
+  if (ySplit !== null) {
+    const upper = items.filter((i) => i.y > ySplit);
+    const lower = items.filter((i) => i.y <= ySplit);
+    if (upper.length > 0 && lower.length > 0 && upper.length < items.length) {
+      return [...xyCutOrder(upper, gapThreshold, depth + 1), ...xyCutOrder(lower, gapThreshold, depth + 1)];
+    }
+  }
+  const xSplit = findXSplit(items, region, gapThreshold);
+  if (xSplit !== null) {
+    const left = items.filter((i) => i.x + i.w / 2 < xSplit);
+    const right = items.filter((i) => i.x + i.w / 2 >= xSplit);
+    if (left.length > 0 && right.length > 0 && left.length < items.length) {
+      return [...xyCutOrder(left, gapThreshold, depth + 1), ...xyCutOrder(right, gapThreshold, depth + 1)];
+    }
+  }
+  return [items];
+}
+function computeRegion(items) {
+  let minX = Infinity, minY = Infinity, maxX = -Infinity, maxY = -Infinity;
+  for (const i of items) {
+    if (i.x < minX) minX = i.x;
+    if (i.y < minY) minY = i.y;
+    if (i.x + i.w > maxX) maxX = i.x + i.w;
+    if (i.y + i.h > maxY) maxY = i.y + i.h;
+  }
+  return { items, minX, minY, maxX, maxY };
+}
+function findYSplit(items, _region, gapThreshold) {
+  const sorted = [...items].sort((a, b) => b.y - a.y);
+  let bestGap = gapThreshold;
+  let bestSplit = null;
+  for (let i = 1; i < sorted.length; i++) {
+    const prevBottom = sorted[i - 1].y - sorted[i - 1].h;
+    const currTop = sorted[i].y;
+    const gap = prevBottom - currTop;
+    if (gap > bestGap) {
+      bestGap = gap;
+      bestSplit = (prevBottom + currTop) / 2;
+    }
+  }
+  return bestSplit;
+}
+function findXSplit(items, _region, gapThreshold) {
+  const sorted = [...items].sort((a, b) => a.x - b.x);
+  let bestGap = gapThreshold;
+  let bestSplit = null;
+  for (let i = 1; i < sorted.length; i++) {
+    const prevRight = sorted[i - 1].x + sorted[i - 1].w;
+    const currLeft = sorted[i].x;
+    const gap = currLeft - prevRight;
+    if (gap > bestGap) {
+      bestGap = gap;
+      bestSplit = (prevRight + currLeft) / 2;
+    }
+  }
+  return bestSplit;
+}
+function extractPageBlocksWithLines(items, pageNum, opList, pageWidth, pageHeight) {
+  if (items.length === 0) return [];
+  let { horizontals, verticals } = extractLines(opList.fnArray, opList.argsArray);
+  ({ horizontals, verticals } = filterPageBorderLines(horizontals, verticals, pageWidth, pageHeight));
+  ({ horizontals, verticals } = preprocessLines(horizontals, verticals));
+  const grids = buildTableGrids(horizontals, verticals);
+  if (grids.length > 0) {
+    return extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals);
+  }
+  return extractPageBlocksFallback(items, pageNum);
+}
+function extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals) {
+  const blocks = [];
+  const usedItems = /* @__PURE__ */ new Set();
+  const sortedGrids = [...grids].sort((a, b) => b.bbox.y2 - a.bbox.y2);
+  for (const grid of sortedGrids) {
+    const numGridRows = grid.rowYs.length - 1;
+    const numGridCols = grid.colXs.length - 1;
+    if (numGridRows === 1 && numGridCols >= 2) continue;
+    const tableItems = [];
+    const pad = 3;
+    const gridW = grid.bbox.x2 - grid.bbox.x1;
+    for (const item of items) {
+      if (usedItems.has(item)) continue;
+      if (item.y < grid.bbox.y1 - pad || item.y > grid.bbox.y2 + pad) continue;
+      if (item.x < grid.bbox.x1 - pad || item.x + item.w > grid.bbox.x2 + pad) continue;
+      if (gridW < 120 && item.x + item.w > grid.bbox.x2 - 2) continue;
+      tableItems.push(item);
+      usedItems.add(item);
+    }
+    const cells = extractCells(grid, horizontals, verticals);
+    if (cells.length === 0) continue;
+    const textItems = tableItems.map((i) => ({
+      text: i.text,
+      x: i.x,
+      y: i.y,
+      w: i.w,
+      h: i.h,
+      fontSize: i.fontSize,
+      fontName: i.fontName
+    }));
+    const cellTextMap = mapTextToCells(textItems, cells);
+    const numRows = grid.rowYs.length - 1;
+    const numCols = grid.colXs.length - 1;
+    const irGrid = Array.from(
+      { length: numRows },
+      () => Array.from({ length: numCols }, () => ({ text: "", colSpan: 1, rowSpan: 1 }))
+    );
+    for (const cell of cells) {
+      const cellItems = cellTextMap.get(cell) || [];
+      let text = cellTextToString(cellItems);
+      text = text.replace(/^[\s]*[-–—]\s*\d+\s*[-–—][\s]*$/gm, "").trim();
+      text = text.split("\n").map((line) => collapseEvenSpacing(line)).join("\n");
+      irGrid[cell.row][cell.col] = {
+        text,
+        colSpan: cell.colSpan,
+        rowSpan: cell.rowSpan
+      };
+    }
+    const irTable = {
+      rows: numRows,
+      cols: numCols,
+      cells: irGrid,
+      hasHeader: numRows > 1
+    };
+    const hasContent = irGrid.some((row) => row.some((cell) => cell.text.trim() !== ""));
+    if (!hasContent) continue;
+    const tableBbox = {
+      page: pageNum,
+      x: grid.bbox.x1,
+      y: grid.bbox.y1,
+      width: grid.bbox.x2 - grid.bbox.x1,
+      height: grid.bbox.y2 - grid.bbox.y1
+    };
+    if (shouldDemoteTable(irTable)) {
+      const demoted = demoteTableToText(irTable);
+      if (demoted) {
+        const text = numGridRows === 1 ? "\n" + demoted + "\n" : demoted;
+        blocks.push({ type: "paragraph", text, pageNumber: pageNum, bbox: tableBbox, style: dominantStyle(tableItems) });
+      }
+      continue;
+    }
+    blocks.push({ type: "table", table: irTable, pageNumber: pageNum, bbox: tableBbox });
+  }
+  let remaining = items.filter((i) => !usedItems.has(i));
+  if (remaining.length > 0) {
+    remaining.sort((a, b) => b.y - a.y || a.x - b.x);
+    const clusterItems = remaining.map((i) => ({
+      text: i.text,
+      x: i.x,
+      y: i.y,
+      w: i.w,
+      h: i.h,
+      fontSize: i.fontSize,
+      fontName: i.fontName
+    }));
+    const clusterResults = detectClusterTables(clusterItems, pageNum);
+    if (clusterResults.length > 0) {
+      const ciToIdx = /* @__PURE__ */ new Map();
+      for (let ci = 0; ci < clusterItems.length; ci++) ciToIdx.set(clusterItems[ci], ci);
+      const usedClusterIndices = /* @__PURE__ */ new Set();
+      for (const cr of clusterResults) {
+        for (const ci of cr.usedItems) {
+          const idx = ciToIdx.get(ci);
+          if (idx !== void 0) usedClusterIndices.add(idx);
+        }
+        blocks.push({ type: "table", table: cr.table, pageNumber: pageNum, bbox: cr.bbox });
+      }
+      remaining = remaining.filter((_, idx) => !usedClusterIndices.has(idx));
+    }
+    if (remaining.length > 0) {
+      const allY = remaining.map((i) => i.y);
+      const pageH = _chunkOJ4QR33Vcjs.safeMax.call(void 0, allY) - _chunkOJ4QR33Vcjs.safeMin.call(void 0, allY);
+      const groups = xyCutOrder(remaining, Math.max(15, pageH * 0.03));
+      const textBlocks = [];
+      for (const group of groups) {
+        if (group.length === 0) continue;
+        const groupBlocks = extractPageBlocksFallback(group, pageNum);
+        for (const b of groupBlocks) textBlocks.push(b);
+      }
+      const finalTextBlocks = detectListBlocks(textBlocks);
+      for (const b of finalTextBlocks) blocks.push(b);
+    }
+    blocks.sort((a, b) => {
+      const ay = a.bbox ? a.bbox.y + a.bbox.height : 0;
+      const by = b.bbox ? b.bbox.y + b.bbox.height : 0;
+      return by - ay;
+    });
+    return mergeAdjacentTableBlocks(blocks);
+  }
+  return mergeAdjacentTableBlocks(blocks);
+}
+function mergeAdjacentTableBlocks(blocks) {
+  if (blocks.length <= 1) return blocks;
+  const result = [blocks[0]];
+  for (let i = 1; i < blocks.length; i++) {
+    const prev = result[result.length - 1];
+    const curr = blocks[i];
+    if (prev.type === "table" && curr.type === "table" && prev.table && curr.table && prev.table.cols === curr.table.cols) {
+      const merged = {
+        rows: prev.table.rows + curr.table.rows,
+        cols: prev.table.cols,
+        cells: [...prev.table.cells, ...curr.table.cells],
+        hasHeader: prev.table.hasHeader
+      };
+      result[result.length - 1] = { ...prev, table: merged };
+    } else {
+      result.push(curr);
+    }
+  }
+  return result;
+}
+function extractPageBlocksFallback(items, pageNum) {
+  if (items.length === 0) return [];
+  const blocks = [];
+  const clusterItems = items.map((i) => ({
+    text: i.text,
+    x: i.x,
+    y: i.y,
+    w: i.w,
+    h: i.h,
+    fontSize: i.fontSize,
+    fontName: i.fontName
+  }));
+  const clusterResults = detectClusterTables(clusterItems, pageNum);
+  if (clusterResults.length > 0) {
+    const ciToIdx = /* @__PURE__ */ new Map();
+    for (let ci = 0; ci < clusterItems.length; ci++) ciToIdx.set(clusterItems[ci], ci);
+    const usedIndices = /* @__PURE__ */ new Set();
+    for (const cr of clusterResults) {
+      for (const ci of cr.usedItems) {
+        const idx = ciToIdx.get(ci);
+        if (idx !== void 0) usedIndices.add(idx);
+      }
+      blocks.push({ type: "table", table: cr.table, pageNumber: pageNum, bbox: cr.bbox });
+    }
+    const remaining = items.filter((_, idx) => !usedIndices.has(idx));
+    if (remaining.length > 0) {
+      const yLines = groupByY(remaining);
+      for (const line of yLines) {
+        const text = mergeLineSimple(line);
+        if (!text.trim()) continue;
+        const bbox = computeBBox(line, pageNum);
+        blocks.push({ type: "paragraph", text, pageNumber: pageNum, bbox, style: dominantStyle(line) });
+      }
+    }
+    blocks.sort((a, b) => {
+      const ay = a.bbox ? a.bbox.y + a.bbox.height : 0;
+      const by = b.bbox ? b.bbox.y + b.bbox.height : 0;
+      return by - ay;
+    });
+  } else {
+    const allYLines = groupByY(items);
+    const columns = detectColumns(allYLines);
+    if (columns && columns.length >= 3) {
+      const tableText = extractWithColumns(allYLines, columns);
+      const bbox = computeBBox(items, pageNum);
+      blocks.push({ type: "paragraph", text: tableText, pageNumber: pageNum, bbox, style: dominantStyle(items) });
+    } else {
+      const allY = items.map((i) => i.y);
+      const pageHeight = _chunkOJ4QR33Vcjs.safeMax.call(void 0, allY) - _chunkOJ4QR33Vcjs.safeMin.call(void 0, allY);
+      const gapThreshold = Math.max(15, pageHeight * 0.03);
+      const orderedGroups = xyCutOrder(items, gapThreshold);
+      for (const group of orderedGroups) {
+        if (group.length === 0) continue;
+        const yLines = groupByY(group);
+        const groupColumns = detectColumns(yLines);
+        if (groupColumns && groupColumns.length >= 3) {
+          const tableText = extractWithColumns(yLines, groupColumns);
+          const bbox = computeBBox(group, pageNum);
+          blocks.push({ type: "paragraph", text: tableText, pageNumber: pageNum, bbox, style: dominantStyle(group) });
+        } else {
+          for (const line of yLines) {
+            const text = mergeLineSimple(line);
+            if (!text.trim()) continue;
+            const bbox = computeBBox(line, pageNum);
+            blocks.push({ type: "paragraph", text, pageNumber: pageNum, bbox, style: dominantStyle(line) });
+          }
+        }
+      }
+    }
+  }
+  return detectSpecialKoreanTables(blocks);
+}
+function computeBBox(items, pageNum) {
+  let minX = Infinity, minY = Infinity, maxX = -Infinity, maxY = -Infinity;
+  for (const i of items) {
+    if (i.x < minX) minX = i.x;
+    if (i.y < minY) minY = i.y;
+    if (i.x + i.w > maxX) maxX = i.x + i.w;
+    const effectiveH = i.h > 0 ? i.h : i.fontSize;
+    if (i.y + effectiveH > maxY) maxY = i.y + effectiveH;
+  }
+  return { page: pageNum, x: minX, y: minY, width: maxX - minX, height: maxY - minY };
+}
+function dominantStyle(items) {
+  if (items.length === 0) return void 0;
+  const freq = /* @__PURE__ */ new Map();
+  let maxCount = 0, dominantSize = 0;
+  for (const i of items) {
+    if (i.fontSize <= 0) continue;
+    const count = (freq.get(i.fontSize) || 0) + 1;
+    freq.set(i.fontSize, count);
+    if (count > maxCount) {
+      maxCount = count;
+      dominantSize = i.fontSize;
+    }
+  }
+  if (dominantSize === 0) return void 0;
+  const fontName = _optionalChain([items, 'access', _20 => _20.find, 'call', _21 => _21((i) => i.fontSize === dominantSize), 'optionalAccess', _22 => _22.fontName]) || void 0;
+  return { fontSize: dominantSize, fontName };
+}
+function normalizeItems(rawItems) {
+  const items = [];
+  const spacePositions = [];
+  for (const i of rawItems) {
+    if (typeof i.str !== "string") continue;
+    const x = Math.round(i.transform[4]);
+    const y = Math.round(i.transform[5]);
+    if (!i.str.trim()) {
+      spacePositions.push({ x, y });
+      continue;
+    }
+    const scaleY = Math.abs(i.transform[3]);
+    const scaleX = Math.abs(i.transform[0]);
+    const fontSize = Math.round(Math.max(scaleY, scaleX));
+    const w = Math.round(i.width);
+    const h = Math.round(i.height);
+    const isHidden = fontSize === 0 || i.width === 0 && i.str.trim().length > 0;
+    let text = i.str.trim();
+    if (/^[\d\s\-().·,☎]+$/.test(text) && /\d/.test(text) && / /.test(text)) {
+      text = text.replace(/ /g, "");
+    }
+    const split = splitEvenSpacedItem(text, x, w, fontSize);
+    if (split) {
+      for (const s of split) {
+        items.push({ text: s.text, x: s.x, y, w: s.w, h, fontSize, fontName: i.fontName || "", isHidden });
+      }
+    } else {
+      items.push({ text, x, y, w, h, fontSize, fontName: i.fontName || "", isHidden });
+    }
+  }
+  const sorted = items.sort((a, b) => b.y - a.y || a.x - b.x);
+  const deduped = [];
+  for (let i = 0; i < sorted.length; i++) {
+    let isDup = false;
+    for (let j = deduped.length - 1; j >= 0; j--) {
+      const prev = deduped[j];
+      if (prev.y - sorted[i].y > 3) break;
+      if (Math.abs(prev.y - sorted[i].y) <= 3 && prev.text === sorted[i].text && Math.abs(prev.x - sorted[i].x) <= 3) {
+        isDup = true;
+        break;
+      }
+    }
+    if (!isDup) deduped.push(sorted[i]);
+  }
+  if (spacePositions.length > 0) {
+    for (const item of deduped) {
+      for (const sp of spacePositions) {
+        if (Math.abs(sp.y - item.y) <= 3) {
+          const dist = item.x - sp.x;
+          if (dist >= 0 && dist <= 20) {
+            item.hasSpaceBefore = true;
+            break;
+          }
+        }
+      }
+    }
+  }
+  return deduped;
+}
+function splitEvenSpacedItem(text, itemX, itemW, fontSize) {
+  if (!/^[가-힣\d](?: [가-힣\d]){2,}$/.test(text)) return null;
+  const chars = text.split(" ");
+  if (chars.length < 3) return null;
+  const charW = itemW / chars.length;
+  if (charW > fontSize * 2) return null;
+  return chars.map((ch, idx) => ({
+    text: ch,
+    x: Math.round(itemX + idx * charW),
+    w: Math.round(charW * 0.8)
+    // 실제 글자 폭은 간격보다 좁음
+  }));
+}
+function groupByY(items) {
+  if (items.length === 0) return [];
+  const lines = [];
+  let curY = items[0].y;
+  let curLine = [items[0]];
+  for (let i = 1; i < items.length; i++) {
+    if (Math.abs(items[i].y - curY) > 3) {
+      lines.push(curLine);
+      curLine = [];
+      curY = items[i].y;
+    }
+    curLine.push(items[i]);
+  }
+  if (curLine.length > 0) lines.push(curLine);
+  return lines;
+}
+function isProseSpread(items) {
+  if (items.length < 4) return false;
+  const sorted = [...items].sort((a, b) => a.x - b.x);
+  const gaps = [];
+  for (let i = 1; i < sorted.length; i++) {
+    gaps.push(sorted[i].x - (sorted[i - 1].x + sorted[i - 1].w));
+  }
+  const maxGap = _chunkOJ4QR33Vcjs.safeMax.call(void 0, gaps);
+  const avgLen = items.reduce((s, i) => s + i.text.length, 0) / items.length;
+  return maxGap < 40 && avgLen < 5;
+}
+function detectColumns(yLines) {
+  const allItems = yLines.flat();
+  if (allItems.length === 0) return null;
+  const pageWidth = _chunkOJ4QR33Vcjs.safeMax.call(void 0, allItems.map((i) => i.x + i.w)) - _chunkOJ4QR33Vcjs.safeMin.call(void 0, allItems.map((i) => i.x));
+  if (pageWidth < 100) return null;
+  let bigoLineIdx = -1;
+  for (let i = 0; i < yLines.length; i++) {
+    if (yLines[i].length <= 2 && yLines[i].some((item) => item.text === "\uBE44\uACE0")) {
+      bigoLineIdx = i;
+      break;
+    }
+  }
+  const tableYLines = bigoLineIdx >= 0 ? yLines.slice(0, bigoLineIdx) : yLines;
+  const CLUSTER_TOL = 22;
+  const xClusters = [];
+  for (const line of tableYLines) {
+    if (isProseSpread(line)) continue;
+    for (const item of line) {
+      let found = false;
+      for (const c of xClusters) {
+        if (Math.abs(item.x - c.center) <= CLUSTER_TOL) {
+          c.center = Math.round((c.center * c.count + item.x) / (c.count + 1));
+          c.minX = Math.min(c.minX, item.x);
+          c.count++;
+          found = true;
+          break;
+        }
+      }
+      if (!found) {
+        xClusters.push({ center: item.x, count: 1, minX: item.x });
+      }
+    }
+  }
+  const peaks = xClusters.filter((c) => c.count >= 3).sort((a, b) => a.minX - b.minX);
+  if (peaks.length < 3) return null;
+  const MERGE_TOL = 40;
+  const merged = [peaks[0]];
+  for (let i = 1; i < peaks.length; i++) {
+    const prev = merged[merged.length - 1];
+    if (peaks[i].minX - prev.minX < MERGE_TOL) {
+      if (peaks[i].count > prev.count) {
+        prev.center = peaks[i].center;
+      }
+      prev.count += peaks[i].count;
+      prev.minX = Math.min(prev.minX, peaks[i].minX);
+    } else {
+      merged.push({ ...peaks[i] });
+    }
+  }
+  const rawColumns = merged.filter((c) => c.count >= 3).map((c) => c.minX);
+  if (rawColumns.length < 3) return null;
+  const MIN_DETECT_COL_WIDTH = 30;
+  const columns = [rawColumns[0]];
+  for (let i = 1; i < rawColumns.length; i++) {
+    if (rawColumns[i] - columns[columns.length - 1] < MIN_DETECT_COL_WIDTH) continue;
+    columns.push(rawColumns[i]);
+  }
+  return columns.length >= 3 ? columns : null;
+}
+function findColumn(x, columns) {
+  for (let i = columns.length - 1; i >= 0; i--) {
+    if (x >= columns[i] - 10) return i;
+  }
+  return 0;
+}
+function extractWithColumns(yLines, columns) {
+  const result = [];
+  const colMin = columns[0];
+  const colMax = columns[columns.length - 1];
+  let bigoIdx = -1;
+  for (let i = 0; i < yLines.length; i++) {
+    if (yLines[i].length <= 2 && yLines[i].some((item) => item.text === "\uBE44\uACE0")) {
+      bigoIdx = i;
+      break;
+    }
+  }
+  let tableStart = -1;
+  for (let i = 0; i < (bigoIdx >= 0 ? bigoIdx : yLines.length); i++) {
+    const usedCols = new Set(yLines[i].map((item) => findColumn(item.x, columns)));
+    if (usedCols.size >= 3) {
+      tableStart = i;
+      break;
+    }
+  }
+  const tableEnd = bigoIdx >= 0 ? bigoIdx : yLines.length;
+  for (let i = 0; i < (tableStart >= 0 ? tableStart : tableEnd); i++) {
+    result.push(mergeLineSimple(yLines[i]));
+  }
+  if (tableStart >= 0) {
+    const tableLines = yLines.slice(tableStart, tableEnd);
+    const gridLines = [];
+    for (const line of tableLines) {
+      const inRange = line.some(
+        (item) => item.x >= colMin - 20 && item.x <= colMax + 200
+      );
+      if (inRange && !isProseSpread(line)) {
+        gridLines.push(line);
+      } else {
+        if (gridLines.length > 0) {
+          result.push(buildGridTable(gridLines.splice(0), columns));
+        }
+        result.push(mergeLineSimple(line));
+      }
+    }
+    if (gridLines.length > 0) {
+      result.push(buildGridTable(gridLines, columns));
+    }
+  }
+  if (bigoIdx >= 0) {
+    result.push("");
+    for (let i = bigoIdx; i < yLines.length; i++) {
+      result.push(mergeLineSimple(yLines[i]));
+    }
+  }
+  return result.join("\n");
+}
+function buildGridTable(lines, columns) {
+  const numCols = columns.length;
+  const yRows = lines.map((items) => {
+    const row = Array(numCols).fill("");
+    for (const item of items) {
+      const col = findColumn(item.x, columns);
+      row[col] = row[col] ? row[col] + " " + item.text : item.text;
+    }
+    return row;
+  });
+  const dataColStart = Math.max(2, Math.floor(numCols / 2));
+  const merged = [];
+  for (const row of yRows) {
+    if (row.every((c) => c === "")) continue;
+    if (merged.length === 0) {
+      merged.push([...row]);
+      continue;
+    }
+    const prev = merged[merged.length - 1];
+    const filledCols = row.map((c, i) => c ? i : -1).filter((i) => i >= 0);
+    const filledCount = filledCols.length;
+    let isNewRow = false;
+    if (row[0] && row[0].length >= 3) {
+      isNewRow = true;
+    }
+    if (!isNewRow && numCols > 1 && row[1]) {
+      isNewRow = true;
+    }
+    if (!isNewRow) {
+      const hasData = row.slice(dataColStart).some((c) => c !== "");
+      const prevHasData = prev.slice(dataColStart).some((c) => c !== "");
+      if (hasData && prevHasData) {
+        isNewRow = true;
+      }
+    }
+    if (isNewRow && filledCount === 1 && row[0] && row[0].length <= 2) {
+      isNewRow = false;
+    }
+    if (isNewRow) {
+      merged.push([...row]);
+    } else {
+      for (let c = 0; c < numCols; c++) {
+        if (row[c]) {
+          prev[c] = prev[c] ? prev[c] + " " + row[c] : row[c];
+        }
+      }
+    }
+  }
+  if (merged.length < 2) {
+    return merged.map((r) => r.filter((c) => c).join(" ")).join("\n");
+  }
+  let headerEnd = 0;
+  for (let r = 0; r < merged.length; r++) {
+    const hasDataValues = merged[r].slice(dataColStart).some((c) => c && /\d/.test(c));
+    if (hasDataValues) break;
+    headerEnd = r + 1;
+  }
+  if (headerEnd > 1) {
+    const headerRow = Array(numCols).fill("");
+    for (let r = 0; r < headerEnd; r++) {
+      for (let c = 0; c < numCols; c++) {
+        if (merged[r][c]) {
+          headerRow[c] = headerRow[c] ? headerRow[c] + " " + merged[r][c] : merged[r][c];
+        }
+      }
+    }
+    merged.splice(0, headerEnd, headerRow);
+  }
+  for (const row of merged) {
+    for (let c = 0; c < row.length; c++) {
+      if (row[c]) row[c] = collapseEvenSpacing(row[c]);
+    }
+  }
+  const totalCells = merged.length * numCols;
+  const filledCells = merged.reduce((s, row) => s + row.filter((c) => c).length, 0);
+  if (filledCells < totalCells * 0.35 || merged.length < 2 || merged.length <= 3 && numCols >= 7) {
+    return merged.map((r) => r.filter((c) => c).join("	")).join("\n");
+  }
+  const md = [];
+  md.push("| " + merged[0].join(" | ") + " |");
+  md.push("| " + merged[0].map(() => "---").join(" | ") + " |");
+  for (let r = 1; r < merged.length; r++) {
+    md.push("| " + merged[r].join(" | ") + " |");
+  }
+  return md.join("\n");
+}
+function mergeLineSimple(items) {
+  if (items.length <= 1) return _optionalChain([items, 'access', _23 => _23[0], 'optionalAccess', _24 => _24.text]) || "";
+  const sorted = [...items].sort((a, b) => a.x - b.x);
+  const isEvenSpaced = detectEvenSpacedItems(sorted);
+  let result = sorted[0].text;
+  for (let i = 1; i < sorted.length; i++) {
+    const gap = sorted[i].x - (sorted[i - 1].x + sorted[i - 1].w);
+    const avgFs = (sorted[i].fontSize + sorted[i - 1].fontSize) / 2;
+    const tabThreshold = Math.max(avgFs * 2, 30);
+    if (gap > tabThreshold) {
+      result += "	";
+      result += sorted[i].text;
+      continue;
+    }
+    if (isEvenSpaced[i]) {
+      result += sorted[i].text;
+      continue;
+    }
+    if (sorted[i].hasSpaceBefore && gap >= avgFs * 0.05) {
+      result += " ";
+      result += sorted[i].text;
+      continue;
+    }
+    if (/[□■○●▶◆◇ㅇ]$/.test(sorted[i - 1].text) && /^[가-힣]/.test(sorted[i].text) && gap > 1) {
+      result += " ";
+      result += sorted[i].text;
+      continue;
+    }
+    if (gap < avgFs * 0.15) {
+    } else if (gap < avgFs * 0.35 && (/[가-힣]$/.test(result) || /^[가-힣]/.test(sorted[i].text))) {
+    } else if (gap > 3) result += " ";
+    result += sorted[i].text;
+  }
+  return result;
+}
+function cleanPdfText(text) {
+  return mergeKoreanLines(
+    text.replace(/^\d{1,4}\n/, "").replace(/^[\s]*[-–—]\s*[-–—]?\d+[-–—]?[\s]*[-–—]?[\s]*$/gm, "").replace(/^\s*\d+\s*\/\s*\d+\s*$/gm, "").replace(/\n\d{1,4}\n/g, "\n").replace(/\n\d{1,4}$/, "").replace(/^#{1,6}\s*\d{1,4}\s*$/gm, "")
+  ).replace(/^(?!\| ---).*$/gm, (line) => collapseEvenSpacing(line)).replace(/([□■◆○●▶ㅇ])\s+([가-힣])\s+([가-힣])/g, "$1 $2$3").replace(/\n{3,}/g, "\n\n").trim();
+}
+function startsWithMarker(line) {
+  const t = line.trimStart();
+  return /^[가-힣ㄱ-ㅎ][.)]/.test(t) || /^\d+[.)]/.test(t) || /^\([가-힣ㄱ-ㅎ\d]+\)/.test(t) || /^[○●※▶▷◆◇■□★☆\-·]\s/.test(t) || /^제\d+[조항호장절]/.test(t);
+}
+function isStandaloneHeader(line) {
+  return /^제\d+[조항호장절](\([^)]*\))?(\s+\S+){0,7}$/.test(line.trim());
+}
+function detectListBlocks(blocks) {
+  const result = [];
+  for (let i = 0; i < blocks.length; i++) {
+    const block = blocks[i];
+    if (block.type === "paragraph" && block.text) {
+      const text = block.text.trim();
+      if (/^\d+\.\s/.test(text)) {
+        result.push({ ...block, type: "list", listType: "ordered", text: block.text });
+        continue;
+      }
+      if (/^[○●·※▶▷◆◇\-]\s/.test(text)) {
+        result.push({ ...block, type: "list", listType: "unordered", text: block.text });
+        continue;
+      }
+    }
+    result.push(block);
+  }
+  return result;
+}
+var KOREAN_TABLE_HEADER_RE = /^\(?(구분|항목|종류|분류|유형|대상|내용|기간|금액|비율|방법|절차|요건|조건|근거|목적|범위|기준)\)?[:\s]/;
+var KV_FALSE_POSITIVE_RE = /\d{1,2}:\d{2}|:\/\/|\d+:\d+/;
+function detectSpecialKoreanTables(blocks) {
+  const result = [];
+  let kvLines = [];
+  const flushKvTable = () => {
+    if (kvLines.length < 2) {
+      for (const kv of kvLines) result.push(kv.block);
+      kvLines = [];
+      return;
+    }
+    const cells = kvLines.map((kv) => {
+      if (kv.value) {
+        return [
+          { text: kv.key, colSpan: 1, rowSpan: 1 },
+          { text: kv.value, colSpan: 1, rowSpan: 1 }
+        ];
+      }
+      return [
+        { text: kv.key, colSpan: 2, rowSpan: 1 },
+        { text: "", colSpan: 1, rowSpan: 1 }
+      ];
+    });
+    const irTable = {
+      rows: cells.length,
+      cols: 2,
+      cells,
+      hasHeader: true
+    };
+    const firstBlock = kvLines[0].block;
+    result.push({
+      type: "table",
+      table: irTable,
+      pageNumber: firstBlock.pageNumber,
+      bbox: firstBlock.bbox
+    });
+    kvLines = [];
+  };
+  for (const block of blocks) {
+    if (block.type !== "paragraph" || !block.text) {
+      flushKvTable();
+      result.push(block);
+      continue;
+    }
+    const text = block.text.trim();
+    if (KOREAN_TABLE_HEADER_RE.test(text)) {
+      const colonIdx = text.indexOf(":");
+      if (colonIdx >= 0) {
+        kvLines.push({
+          key: text.slice(0, colonIdx).trim(),
+          value: text.slice(colonIdx + 1).trim(),
+          block
+        });
+      } else {
+        const spaceIdx = text.search(/\s/);
+        if (spaceIdx > 0) {
+          kvLines.push({
+            key: text.slice(0, spaceIdx).trim(),
+            value: text.slice(spaceIdx + 1).trim(),
+            block
+          });
+        } else {
+          kvLines.push({ key: text, value: "", block });
+        }
+      }
+      continue;
+    }
+    if (kvLines.length > 0 && text.includes(":")) {
+      if (!KV_FALSE_POSITIVE_RE.test(text) && !text.includes("(") && !text.includes(")")) {
+        const colonIdx = text.indexOf(":");
+        const key = text.slice(0, colonIdx).trim();
+        if (/^[가-힣]+$/.test(key) && key.length >= 2 && key.length <= 8) {
+          kvLines.push({
+            key,
+            value: text.slice(colonIdx + 1).trim(),
+            block
+          });
+          continue;
+        }
+      }
+    }
+    flushKvTable();
+    result.push(block);
+  }
+  flushKvTable();
+  return result;
+}
+function removeHeaderFooterBlocks(blocks, pageHeights, warnings) {
+  const ZONE_RATIO = 0.1;
+  const MIN_REPEAT = 3;
+  const headerTexts = /* @__PURE__ */ new Map();
+  const footerTexts = /* @__PURE__ */ new Map();
+  for (let bi = 0; bi < blocks.length; bi++) {
+    const b = blocks[bi];
+    if (!b.bbox || !b.pageNumber || !_optionalChain([b, 'access', _25 => _25.text, 'optionalAccess', _26 => _26.trim, 'call', _27 => _27()])) continue;
+    const ph = pageHeights.get(b.bbox.page) || pageHeights.get(b.pageNumber);
+    if (!ph) continue;
+    const blockTop = ph - (b.bbox.y + b.bbox.height);
+    const blockBottom = ph - b.bbox.y;
+    if (blockBottom <= ph * ZONE_RATIO) {
+      const arr = footerTexts.get(b.pageNumber) || [];
+      arr.push(b.text.trim());
+      footerTexts.set(b.pageNumber, arr);
+    } else if (blockTop >= ph * (1 - ZONE_RATIO)) {
+      const arr = headerTexts.get(b.pageNumber) || [];
+      arr.push(b.text.trim());
+      headerTexts.set(b.pageNumber, arr);
+    }
+  }
+  const repeatedPatterns = /* @__PURE__ */ new Set();
+  for (const textsMap of [headerTexts, footerTexts]) {
+    const patternCount = /* @__PURE__ */ new Map();
+    for (const [, texts] of textsMap) {
+      for (const t of texts) {
+        const normalized = t.replace(/\d+/g, "#");
+        patternCount.set(normalized, (patternCount.get(normalized) || 0) + 1);
+      }
+    }
+    for (const [pattern, count] of patternCount) {
+      if (count >= MIN_REPEAT) repeatedPatterns.add(pattern);
+    }
+  }
+  if (repeatedPatterns.size === 0) return [];
+  const removeIndices = [];
+  for (let bi = 0; bi < blocks.length; bi++) {
+    const b = blocks[bi];
+    if (!b.bbox || !b.pageNumber || !_optionalChain([b, 'access', _28 => _28.text, 'optionalAccess', _29 => _29.trim, 'call', _30 => _30()])) continue;
+    const ph = pageHeights.get(b.bbox.page) || pageHeights.get(b.pageNumber);
+    if (!ph) continue;
+    const blockTop = ph - (b.bbox.y + b.bbox.height);
+    const blockBottom = ph - b.bbox.y;
+    const inZone = blockBottom <= ph * ZONE_RATIO || blockTop >= ph * (1 - ZONE_RATIO);
+    if (!inZone) continue;
+    const normalized = b.text.trim().replace(/\d+/g, "#");
+    if (repeatedPatterns.has(normalized)) {
+      removeIndices.push(bi);
+    }
+  }
+  if (removeIndices.length > 0) {
+    warnings.push({ message: `${removeIndices.length}\uAC1C \uBA38\uB9AC\uAE00/\uBC14\uB2E5\uAE00 \uC694\uC18C \uC81C\uAC70\uB428`, code: "HIDDEN_TEXT_FILTERED" });
+  }
+  return removeIndices;
+}
+function mergeKoreanLines(text) {
+  if (!text) return "";
+  const lines = text.split("\n");
+  if (lines.length <= 1) return text;
+  const result = [lines[0]];
+  for (let i = 1; i < lines.length; i++) {
+    const prev = result[result.length - 1];
+    const curr = lines[i];
+    const currTrimmed = curr.trim();
+    if (/^#{1,6}\s/.test(prev) || /^#{1,6}\s/.test(curr) || /^\|/.test(currTrimmed) || /^---/.test(currTrimmed)) {
+      result.push(curr);
+      continue;
+    }
+    if (/,$/.test(prev.trim()) && currTrimmed.length > 0) {
+      result[result.length - 1] = prev + "\n" + curr;
+      continue;
+    }
+    if (/^\(※/.test(currTrimmed)) {
+      result[result.length - 1] = prev + " " + currTrimmed;
+      continue;
+    }
+    if (/[가-힣·,\-]$/.test(prev) && /^[가-힣(]/.test(curr) && !startsWithMarker(curr) && !isStandaloneHeader(prev) && !startsWithMarker(prev)) {
+      result[result.length - 1] = prev + " " + curr;
+    } else {
+      result.push(curr);
+    }
+  }
+  return result.join("\n");
+}
+
+
+
+
+exports.cleanPdfText = cleanPdfText; exports.extractPdfMetadataOnly = extractPdfMetadataOnly; exports.parsePdfDocument = parsePdfDocument;
+//# sourceMappingURL=parser-CYBX5MP4.cjs.map