kordoc 4.6.0 → 4.7.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (182) hide show
  1. package/LICENSE +21 -21
  2. package/NOTICE +176 -176
  3. package/README.md +962 -958
  4. package/THIRD_PARTY/apache-2.0.LICENSE +201 -201
  5. package/THIRD_PARTY/claw-hwp.LICENSE +21 -21
  6. package/THIRD_PARTY/hml-equation-parser.LICENSE +201 -201
  7. package/THIRD_PARTY/hml-equation-parser.txt +35 -35
  8. package/THIRD_PARTY/opendataloader-pdf.txt +32 -32
  9. package/THIRD_PARTY/pix2text.txt +17 -17
  10. package/THIRD_PARTY/rhwp-forms.txt +35 -35
  11. package/dist/{-DEVLQKIT.js → -MK2223RL.js} +12 -12
  12. package/dist/{chunk-3FFPHTXD.js → chunk-3H5X63BQ.js} +1 -1
  13. package/dist/chunk-3H5X63BQ.js.map +1 -0
  14. package/dist/{chunk-IBVMJIHR.js → chunk-3VTMSSDW.js} +2 -2
  15. package/dist/chunk-3VTMSSDW.js.map +1 -0
  16. package/dist/{chunk-HPRGVU6A.cjs → chunk-5QTHHQQE.cjs} +6 -6
  17. package/dist/chunk-5QTHHQQE.cjs.map +1 -0
  18. package/dist/{chunk-6GBLFQMA.js → chunk-5RPYBC2Q.js} +1 -1
  19. package/dist/chunk-5RPYBC2Q.js.map +1 -0
  20. package/dist/{chunk-7BTO44Q7.js → chunk-6PIR74EA.js} +13 -13
  21. package/dist/chunk-6PIR74EA.js.map +1 -0
  22. package/dist/{chunk-W34QSJFF.js → chunk-7FCNOKOR.js} +1 -1
  23. package/dist/chunk-7FCNOKOR.js.map +1 -0
  24. package/dist/{chunk-MSMSHU2F.js → chunk-AEYNY6FG.js} +3 -3
  25. package/dist/chunk-AEYNY6FG.js.map +1 -0
  26. package/dist/{chunk-KC63W72C.js → chunk-AFPD74JJ.js} +2 -2
  27. package/dist/chunk-AFPD74JJ.js.map +1 -0
  28. package/dist/{chunk-IHK5LCYX.js → chunk-BCG2PKQU.js} +2 -2
  29. package/dist/chunk-BCG2PKQU.js.map +1 -0
  30. package/dist/{chunk-I42DND3B.cjs → chunk-C27PPXRZ.cjs} +342 -29
  31. package/dist/chunk-C27PPXRZ.cjs.map +1 -0
  32. package/dist/{chunk-X34YWRL5.cjs → chunk-DCZVOIEO.cjs} +1 -1
  33. package/dist/chunk-DCZVOIEO.cjs.map +1 -0
  34. package/dist/{chunk-O5EK52TG.js → chunk-DZIXKL7E.js} +1 -1
  35. package/dist/chunk-DZIXKL7E.js.map +1 -0
  36. package/dist/{chunk-AIQ3ISQU.js → chunk-GE43BE46.js} +1 -1
  37. package/dist/chunk-GE43BE46.js.map +1 -0
  38. package/dist/{chunk-2CENWSAW.js → chunk-GIWKBPII.js} +335 -22
  39. package/dist/chunk-GIWKBPII.js.map +1 -0
  40. package/dist/chunk-GS7T56RP.cjs.map +1 -1
  41. package/dist/{chunk-6XPXIG57.js → chunk-GZDOWYJM.js} +2 -2
  42. package/dist/chunk-GZDOWYJM.js.map +1 -0
  43. package/dist/{chunk-OPL22SYF.js → chunk-IDHBO6AG.js} +3 -3
  44. package/dist/chunk-IDHBO6AG.js.map +1 -0
  45. package/dist/{chunk-3TBUDJDE.js → chunk-MOL7MDBG.js} +1 -1
  46. package/dist/chunk-MOL7MDBG.js.map +1 -0
  47. package/dist/{chunk-PSAHY5XW.js → chunk-MVTX6ODZ.js} +1 -1
  48. package/dist/chunk-MVTX6ODZ.js.map +1 -0
  49. package/dist/{chunk-2CABQYTT.js → chunk-P2NWNJMR.js} +3 -3
  50. package/dist/chunk-P2NWNJMR.js.map +1 -0
  51. package/dist/{chunk-PQDNCWAZ.js → chunk-PEVPKD6Z.js} +336 -23
  52. package/dist/chunk-PEVPKD6Z.js.map +1 -0
  53. package/dist/{chunk-UQ4NSK66.js → chunk-PV2EXXGI.js} +3 -3
  54. package/dist/chunk-PV2EXXGI.js.map +1 -0
  55. package/dist/{chunk-WX7WQTOT.cjs → chunk-TUU535RB.cjs} +2 -2
  56. package/dist/chunk-TUU535RB.cjs.map +1 -0
  57. package/dist/{chunk-YNJDNP44.js → chunk-UAFSPWU7.js} +1 -1
  58. package/dist/chunk-UAFSPWU7.js.map +1 -0
  59. package/dist/{chunk-CARNWWGI.cjs → chunk-W5KRKBYR.cjs} +4 -4
  60. package/dist/chunk-W5KRKBYR.cjs.map +1 -0
  61. package/dist/{chunk-VNEV6VKD.js → chunk-XZDJXYUC.js} +2 -2
  62. package/dist/chunk-XZDJXYUC.js.map +1 -0
  63. package/dist/{chunk-UGKXJA3W.js → chunk-YUVKU5CW.js} +2 -2
  64. package/dist/chunk-YUVKU5CW.js.map +1 -0
  65. package/dist/{chunk-KQOTL4WI.js → chunk-YXXHH6WC.js} +2 -2
  66. package/dist/chunk-YXXHH6WC.js.map +1 -0
  67. package/dist/{chunk-MGGU3ZRK.cjs → chunk-ZV5OKQ3Z.cjs} +1 -1
  68. package/dist/chunk-ZV5OKQ3Z.cjs.map +1 -0
  69. package/dist/{chunk-ZYQ5X6PG.js → chunk-ZYVULRYF.js} +2 -2
  70. package/dist/chunk-ZYVULRYF.js.map +1 -0
  71. package/dist/chunks-4HEG6ZPG.js +10 -0
  72. package/dist/cli.js +24 -24
  73. package/dist/cli.js.map +1 -1
  74. package/dist/{formula-WKGP3D3Y.cjs → formula-CG4S7TWO.cjs} +8 -8
  75. package/dist/formula-CG4S7TWO.cjs.map +1 -0
  76. package/dist/{formula-ETB6MESN.js → formula-JLYS73QL.js} +2 -2
  77. package/dist/formula-JLYS73QL.js.map +1 -0
  78. package/dist/{formula-KW4NI5IO.js → formula-LIVQSCMM.js} +2 -2
  79. package/dist/formula-LIVQSCMM.js.map +1 -0
  80. package/dist/{image-ocr-C6U7YHRC.js → image-ocr-3A3HZX6O.js} +5 -5
  81. package/dist/image-ocr-3A3HZX6O.js.map +1 -0
  82. package/dist/{image-ocr-YIBJCOUQ.js → image-ocr-4MHEX4MN.js} +7 -7
  83. package/dist/image-ocr-4MHEX4MN.js.map +1 -0
  84. package/dist/{image-ocr-YID6TDN5.cjs → image-ocr-OB5PCQUU.cjs} +11 -11
  85. package/dist/image-ocr-OB5PCQUU.cjs.map +1 -0
  86. package/dist/index.cjs +185 -185
  87. package/dist/index.cjs.map +1 -1
  88. package/dist/index.js +7 -7
  89. package/dist/index.js.map +1 -1
  90. package/dist/mcp.js +20 -20
  91. package/dist/mcp.js.map +1 -1
  92. package/dist/{models-6VMAVZR6.js → models-6SIKLKRH.js} +3 -3
  93. package/dist/page-range-737B4EZW.js +8 -0
  94. package/dist/page-range-LNMXJU6W.js +7 -0
  95. package/dist/page-range-P7SDW6LR.cjs +8 -0
  96. package/dist/page-range-P7SDW6LR.cjs.map +1 -0
  97. package/dist/{parser-GFXZV2VJ.js → parser-EIVCD4NS.js} +10 -10
  98. package/dist/parser-EIVCD4NS.js.map +1 -0
  99. package/dist/{parser-SHM3JEH6.js → parser-EQN4YKU4.js} +8 -8
  100. package/dist/parser-EQN4YKU4.js.map +1 -0
  101. package/dist/{parser-EYCVDUJA.cjs → parser-VACM3DFR.cjs} +28 -28
  102. package/dist/parser-VACM3DFR.cjs.map +1 -0
  103. package/dist/pdf-ocr-N62EWEX4.js +12 -0
  104. package/dist/pdf-ocr-NPYDUIC5.js +15 -0
  105. package/dist/pdf-ocr-OBPQ4F4D.cjs +13 -0
  106. package/dist/pdf-ocr-OBPQ4F4D.cjs.map +1 -0
  107. package/dist/{profile-io-FQZNNA3O.js → profile-io-L5EH4JAD.js} +3 -3
  108. package/dist/{rasterize-QGKFNTXO.js → rasterize-7AR662AL.js} +2 -2
  109. package/dist/rasterize-7AR662AL.js.map +1 -0
  110. package/dist/{redact-W5K3ILQH.js → redact-7ILEAUTS.js} +2 -2
  111. package/dist/render-CAFQT7S4.js +10 -0
  112. package/dist/seal-ESAC3IZR.js +10 -0
  113. package/dist/{setup-I4I75QJU.js → setup-GUQT5ELJ.js} +1 -1
  114. package/dist/setup-GUQT5ELJ.js.map +1 -0
  115. package/dist/{watch-OIVXH2RN.js → watch-MONT6CGI.js} +12 -12
  116. package/dist/watch-MONT6CGI.js.map +1 -0
  117. package/package.json +103 -103
  118. package/templates//352/260/204/354/235/264/352/270/260/354/225/210/353/254/270_/354/230/210/354/213/234/352/260/222.json +9 -9
  119. package/templates//354/235/274/353/260/230/352/270/260/354/225/210/353/254/270_/354/230/210/354/213/234/352/260/222.json +24 -24
  120. package/dist/chunk-2CABQYTT.js.map +0 -1
  121. package/dist/chunk-2CENWSAW.js.map +0 -1
  122. package/dist/chunk-3FFPHTXD.js.map +0 -1
  123. package/dist/chunk-3TBUDJDE.js.map +0 -1
  124. package/dist/chunk-6GBLFQMA.js.map +0 -1
  125. package/dist/chunk-6XPXIG57.js.map +0 -1
  126. package/dist/chunk-7BTO44Q7.js.map +0 -1
  127. package/dist/chunk-AIQ3ISQU.js.map +0 -1
  128. package/dist/chunk-CARNWWGI.cjs.map +0 -1
  129. package/dist/chunk-HPRGVU6A.cjs.map +0 -1
  130. package/dist/chunk-I42DND3B.cjs.map +0 -1
  131. package/dist/chunk-IBVMJIHR.js.map +0 -1
  132. package/dist/chunk-IHK5LCYX.js.map +0 -1
  133. package/dist/chunk-KC63W72C.js.map +0 -1
  134. package/dist/chunk-KQOTL4WI.js.map +0 -1
  135. package/dist/chunk-MGGU3ZRK.cjs.map +0 -1
  136. package/dist/chunk-MSMSHU2F.js.map +0 -1
  137. package/dist/chunk-O5EK52TG.js.map +0 -1
  138. package/dist/chunk-OPL22SYF.js.map +0 -1
  139. package/dist/chunk-PQDNCWAZ.js.map +0 -1
  140. package/dist/chunk-PSAHY5XW.js.map +0 -1
  141. package/dist/chunk-UGKXJA3W.js.map +0 -1
  142. package/dist/chunk-UQ4NSK66.js.map +0 -1
  143. package/dist/chunk-VNEV6VKD.js.map +0 -1
  144. package/dist/chunk-W34QSJFF.js.map +0 -1
  145. package/dist/chunk-WX7WQTOT.cjs.map +0 -1
  146. package/dist/chunk-X34YWRL5.cjs.map +0 -1
  147. package/dist/chunk-YNJDNP44.js.map +0 -1
  148. package/dist/chunk-ZYQ5X6PG.js.map +0 -1
  149. package/dist/chunks-Y2VE2WCJ.js +0 -10
  150. package/dist/formula-ETB6MESN.js.map +0 -1
  151. package/dist/formula-KW4NI5IO.js.map +0 -1
  152. package/dist/formula-WKGP3D3Y.cjs.map +0 -1
  153. package/dist/image-ocr-C6U7YHRC.js.map +0 -1
  154. package/dist/image-ocr-YIBJCOUQ.js.map +0 -1
  155. package/dist/image-ocr-YID6TDN5.cjs.map +0 -1
  156. package/dist/page-range-CIRRJPXJ.js +0 -7
  157. package/dist/page-range-OF5I4PQY.js +0 -8
  158. package/dist/page-range-TPIRSA3J.cjs +0 -8
  159. package/dist/page-range-TPIRSA3J.cjs.map +0 -1
  160. package/dist/parser-EYCVDUJA.cjs.map +0 -1
  161. package/dist/parser-GFXZV2VJ.js.map +0 -1
  162. package/dist/parser-SHM3JEH6.js.map +0 -1
  163. package/dist/pdf-ocr-TMAWPOHC.cjs +0 -13
  164. package/dist/pdf-ocr-TMAWPOHC.cjs.map +0 -1
  165. package/dist/pdf-ocr-UKFYODLP.js +0 -15
  166. package/dist/pdf-ocr-ZG4UVUIT.js +0 -12
  167. package/dist/rasterize-QGKFNTXO.js.map +0 -1
  168. package/dist/render-YFIWQYVB.js +0 -10
  169. package/dist/seal-GZ7NDWLR.js +0 -10
  170. package/dist/setup-I4I75QJU.js.map +0 -1
  171. package/dist/watch-OIVXH2RN.js.map +0 -1
  172. /package/dist/{-DEVLQKIT.js.map → -MK2223RL.js.map} +0 -0
  173. /package/dist/{chunks-Y2VE2WCJ.js.map → chunks-4HEG6ZPG.js.map} +0 -0
  174. /package/dist/{models-6VMAVZR6.js.map → models-6SIKLKRH.js.map} +0 -0
  175. /package/dist/{page-range-CIRRJPXJ.js.map → page-range-737B4EZW.js.map} +0 -0
  176. /package/dist/{page-range-OF5I4PQY.js.map → page-range-LNMXJU6W.js.map} +0 -0
  177. /package/dist/{pdf-ocr-UKFYODLP.js.map → pdf-ocr-N62EWEX4.js.map} +0 -0
  178. /package/dist/{pdf-ocr-ZG4UVUIT.js.map → pdf-ocr-NPYDUIC5.js.map} +0 -0
  179. /package/dist/{profile-io-FQZNNA3O.js.map → profile-io-L5EH4JAD.js.map} +0 -0
  180. /package/dist/{redact-W5K3ILQH.js.map → redact-7ILEAUTS.js.map} +0 -0
  181. /package/dist/{render-YFIWQYVB.js.map → render-CAFQT7S4.js.map} +0 -0
  182. /package/dist/{seal-GZ7NDWLR.js.map → seal-ESAC3IZR.js.map} +0 -0
@@ -3,11 +3,11 @@ import {
3
3
  HEADING_RATIO_H1,
4
4
  HEADING_RATIO_H2,
5
5
  HEADING_RATIO_H3
6
- } from "./chunk-PSAHY5XW.js";
6
+ } from "./chunk-MVTX6ODZ.js";
7
7
  import {
8
8
  safeMax,
9
9
  safeMin
10
- } from "./chunk-ZYQ5X6PG.js";
10
+ } from "./chunk-ZYVULRYF.js";
11
11
 
12
12
  // src/pdf/image-regions.ts
13
13
  import { OPS } from "pdfjs-dist/legacy/build/pdf.mjs";
@@ -2684,6 +2684,149 @@ function findVerticalCut(items) {
2684
2684
  return { position, gap: largestGap };
2685
2685
  }
2686
2686
 
2687
+ // src/pdf/two-column.ts
2688
+ var GUTTER_SCAN_LO = 0.3;
2689
+ var GUTTER_SCAN_HI = 0.7;
2690
+ var GUTTER_MAX_SAMPLES = 400;
2691
+ var WIDE_RECT_RATIO = 0.55;
2692
+ var GUTTER_MAX_COVER_RATIO = 0.06;
2693
+ var SIDE_MIN_COUNT = 5;
2694
+ var SIDE_MIN_HEIGHT_RATIO = 0.3;
2695
+ var SIDE_MIN_VSPAN_RATIO = 0.45;
2696
+ var SIDE_MIN_WIDTH_SYMMETRY = 0.5;
2697
+ var MIN_CONTENT_SPAN = 300;
2698
+ var TEXT_RECT_MAX_H = 20;
2699
+ var ROW_Y_TOL = 2;
2700
+ var SIDE_MIN_TEXT_ROWS = 12;
2701
+ var CROSS_ROW_TINY_GAP = 12;
2702
+ var MAX_TINY_GAP_ROWS = 1;
2703
+ var MAX_BOTH_SIDE_ROW_RATIO = 0.65;
2704
+ var BLOCK_DOMINANT_TEXT_RECTS = 10;
2705
+ function detectColumnGutter(rects) {
2706
+ if (rects.length < 8) return null;
2707
+ let minX = Infinity;
2708
+ let maxX = -Infinity;
2709
+ for (const r of rects) {
2710
+ if (r.x < minX) minX = r.x;
2711
+ if (r.x + r.w > maxX) maxX = r.x + r.w;
2712
+ }
2713
+ const span = maxX - minX;
2714
+ if (!Number.isFinite(span) || span < MIN_CONTENT_SPAN) return null;
2715
+ const narrow = rects.filter((r) => r.w < span * WIDE_RECT_RATIO && r.w > 0 && r.h > 0);
2716
+ if (narrow.length < SIDE_MIN_COUNT * 2) return null;
2717
+ let totalH = 0;
2718
+ for (const r of narrow) totalH += r.h;
2719
+ if (totalH <= 0) return null;
2720
+ const lo = minX + span * GUTTER_SCAN_LO;
2721
+ const hi = minX + span * GUTTER_SCAN_HI;
2722
+ const step = Math.max(2, (hi - lo) / GUTTER_MAX_SAMPLES);
2723
+ let gutterX = 0;
2724
+ let bestCover = Infinity;
2725
+ for (let x = lo; x <= hi; x += step) {
2726
+ let cover = 0;
2727
+ for (const r of narrow) {
2728
+ if (r.x < x && r.x + r.w > x) cover += r.h;
2729
+ }
2730
+ if (cover < bestCover) {
2731
+ bestCover = cover;
2732
+ gutterX = x;
2733
+ }
2734
+ }
2735
+ if (bestCover > totalH * GUTTER_MAX_COVER_RATIO) return null;
2736
+ const left = [];
2737
+ const right = [];
2738
+ for (const r of narrow) {
2739
+ if (r.x + r.w <= gutterX) left.push(r);
2740
+ else if (r.x >= gutterX) right.push(r);
2741
+ }
2742
+ if (left.length < SIDE_MIN_COUNT || right.length < SIDE_MIN_COUNT) return null;
2743
+ const sideStats = (side) => {
2744
+ let sMinX = Infinity, sMaxR = -Infinity, sMinY = Infinity, sMaxT = -Infinity, hSum = 0;
2745
+ for (const r of side) {
2746
+ if (r.x < sMinX) sMinX = r.x;
2747
+ if (r.x + r.w > sMaxR) sMaxR = r.x + r.w;
2748
+ if (r.y < sMinY) sMinY = r.y;
2749
+ if (r.y + r.h > sMaxT) sMaxT = r.y + r.h;
2750
+ hSum += r.h;
2751
+ }
2752
+ return { width: sMaxR - sMinX, vspan: sMaxT - sMinY, minY: sMinY, maxT: sMaxT, hSum };
2753
+ };
2754
+ const L = sideStats(left);
2755
+ const R = sideStats(right);
2756
+ if (Math.min(L.hSum, R.hSum) / Math.max(L.hSum, R.hSum) < SIDE_MIN_HEIGHT_RATIO) return null;
2757
+ const unionVspan = Math.max(L.maxT, R.maxT) - Math.min(L.minY, R.minY);
2758
+ if (unionVspan <= 0) return null;
2759
+ if (L.vspan / unionVspan < SIDE_MIN_VSPAN_RATIO) return null;
2760
+ if (R.vspan / unionVspan < SIDE_MIN_VSPAN_RATIO) return null;
2761
+ if (Math.min(L.width, R.width) / Math.max(L.width, R.width) < SIDE_MIN_WIDTH_SYMMETRY) return null;
2762
+ const textRects = (side) => side.filter((r) => r.h <= TEXT_RECT_MAX_H);
2763
+ const lText = textRects(left);
2764
+ const rText = textRects(right);
2765
+ if (lText.length + rText.length >= BLOCK_DOMINANT_TEXT_RECTS) {
2766
+ const rows = /* @__PURE__ */ new Map();
2767
+ const rowKey = (y) => Math.round(y / ROW_Y_TOL);
2768
+ for (const r of lText) {
2769
+ const k = rowKey(r.y);
2770
+ const e = rows.get(k) ?? { lEnd: -Infinity, rStart: Infinity };
2771
+ if (r.x + r.w > e.lEnd) e.lEnd = r.x + r.w;
2772
+ rows.set(k, e);
2773
+ }
2774
+ const lRowCount = rows.size;
2775
+ for (const r of rText) {
2776
+ const k = rowKey(r.y);
2777
+ const e = rows.get(k) ?? { lEnd: -Infinity, rStart: Infinity };
2778
+ if (r.x < e.rStart) e.rStart = r.x;
2779
+ rows.set(k, e);
2780
+ }
2781
+ let rRowCount = 0;
2782
+ let bothRows = 0;
2783
+ let tinyGapRows = 0;
2784
+ for (const e of rows.values()) {
2785
+ if (e.rStart < Infinity) rRowCount++;
2786
+ if (e.lEnd > -Infinity && e.rStart < Infinity) {
2787
+ bothRows++;
2788
+ if (e.rStart - e.lEnd < CROSS_ROW_TINY_GAP) tinyGapRows++;
2789
+ }
2790
+ }
2791
+ if (lRowCount < SIDE_MIN_TEXT_ROWS || rRowCount < SIDE_MIN_TEXT_ROWS) return null;
2792
+ if (tinyGapRows > MAX_TINY_GAP_ROWS) return null;
2793
+ if (bothRows / Math.min(lRowCount, rRowCount) > MAX_BOTH_SIDE_ROW_RATIO) return null;
2794
+ }
2795
+ return gutterX;
2796
+ }
2797
+ function orderByGutter(units, rectOf, gutterX) {
2798
+ const left = [];
2799
+ const right = [];
2800
+ const cross = [];
2801
+ for (const u of units) {
2802
+ const r = rectOf(u);
2803
+ const top = r.y + r.h;
2804
+ if (r.x < gutterX && r.x + r.w > gutterX) cross.push({ u, top });
2805
+ else if (r.x + r.w <= gutterX) left.push({ u, top });
2806
+ else right.push({ u, top });
2807
+ }
2808
+ const byTopDesc = (a, b) => b.top - a.top;
2809
+ cross.sort(byTopDesc);
2810
+ left.sort(byTopDesc);
2811
+ right.sort(byTopDesc);
2812
+ const bandOf = (top) => {
2813
+ let k = 0;
2814
+ while (k < cross.length && cross[k].top > top) k++;
2815
+ return k;
2816
+ };
2817
+ const ordered = [];
2818
+ for (let k = 0; k <= cross.length; k++) {
2819
+ for (const t of left) {
2820
+ if (bandOf(t.top) === k) ordered.push(t.u);
2821
+ }
2822
+ for (const t of right) {
2823
+ if (bandOf(t.top) === k) ordered.push(t.u);
2824
+ }
2825
+ if (k < cross.length) ordered.push(cross[k].u);
2826
+ }
2827
+ return ordered;
2828
+ }
2829
+
2687
2830
  // src/pdf/columns.ts
2688
2831
  function isProseSpread(items) {
2689
2832
  if (items.length < 4) return false;
@@ -2900,6 +3043,123 @@ function buildGridTable(lines, columns) {
2900
3043
  return md.join("\n");
2901
3044
  }
2902
3045
 
3046
+ // src/pdf/underline.ts
3047
+ var UNDER_MAX_THICKNESS = 2;
3048
+ var UNDER_BELOW_EM = 0.72;
3049
+ var UNDER_BELOW_MIN_PT = 3;
3050
+ var UNDER_ABOVE_PT = 1;
3051
+ var UNDER_MIN_OVERLAP_RATIO = 0.6;
3052
+ var UNDER_OWNER_PAD_EM = 0.75;
3053
+ var UNDER_OWNER_PAD_MIN_PT = 4;
3054
+ var UNDER_MIN_COVERAGE = 0.6;
3055
+ var UNDER_GRID_EPS = 2;
3056
+ var UNDER_COLUMN_GAP_EM = 2;
3057
+ var UNDER_REPEATED_SPAN_LEVELS = 3;
3058
+ var UNDER_SPAN_OVERLAP_RATIO = 0.8;
3059
+ var UNDER_BOX_PAIR_MIN_EM = 0.5;
3060
+ var UNDER_BOX_PAIR_MAX_EM = 2.2;
3061
+ var UNDER_BOX_PAIR_OVERLAP = 0.8;
3062
+ function markUnderlineItems(items, horizontals, verticals) {
3063
+ if (items.length === 0 || horizontals.length === 0) return;
3064
+ for (const line of horizontals) {
3065
+ if (line.lineWidth > UNDER_MAX_THICKNESS) continue;
3066
+ if (touchesVertical(line, verticals)) continue;
3067
+ if (isRepeatedSpanRule(line, horizontals)) continue;
3068
+ const matches = [];
3069
+ for (const item of items) {
3070
+ const h = item.h > 0 ? item.h : item.fontSize;
3071
+ if (h <= 0 || item.w <= 0 || !item.text.trim()) continue;
3072
+ const below = Math.max(h * UNDER_BELOW_EM, UNDER_BELOW_MIN_PT);
3073
+ if (line.y1 < item.y - below || line.y1 > item.y + UNDER_ABOVE_PT) continue;
3074
+ const overlap = Math.min(line.x2, item.x + item.w) - Math.max(line.x1, item.x);
3075
+ if (overlap / item.w < UNDER_MIN_OVERLAP_RATIO) continue;
3076
+ matches.push(item);
3077
+ }
3078
+ if (matches.length === 0) continue;
3079
+ let x1 = Infinity, x2 = -Infinity, maxH = 0, covered = 0;
3080
+ for (const m of matches) {
3081
+ x1 = Math.min(x1, m.x);
3082
+ x2 = Math.max(x2, m.x + m.w);
3083
+ maxH = Math.max(maxH, m.h > 0 ? m.h : m.fontSize);
3084
+ covered += m.w;
3085
+ }
3086
+ const pad = Math.max(maxH * UNDER_OWNER_PAD_EM, UNDER_OWNER_PAD_MIN_PT);
3087
+ if (line.x1 < x1 - pad || line.x2 > x2 + pad) continue;
3088
+ if (covered < (line.x2 - line.x1) * UNDER_MIN_COVERAGE) continue;
3089
+ if (hasBoxTopPair(line, maxH, horizontals)) continue;
3090
+ matches.sort((a, b) => a.x - b.x);
3091
+ let hole = false;
3092
+ for (let i = 1; i < matches.length; i++) {
3093
+ if (matches[i].x - (matches[i - 1].x + matches[i - 1].w) > maxH * UNDER_COLUMN_GAP_EM) {
3094
+ hole = true;
3095
+ break;
3096
+ }
3097
+ }
3098
+ if (hole) continue;
3099
+ for (const m of matches) m.underline = true;
3100
+ }
3101
+ }
3102
+ function touchesVertical(line, verticals) {
3103
+ for (const v of verticals) {
3104
+ if (v.x1 < line.x1 - UNDER_GRID_EPS || v.x1 > line.x2 + UNDER_GRID_EPS) continue;
3105
+ const lo = Math.min(v.y1, v.y2), hi = Math.max(v.y1, v.y2);
3106
+ if (lo <= line.y1 + UNDER_GRID_EPS && hi >= line.y1 - UNDER_GRID_EPS) return true;
3107
+ }
3108
+ return false;
3109
+ }
3110
+ function hasBoxTopPair(line, maxH, horizontals) {
3111
+ const w = line.x2 - line.x1;
3112
+ if (w <= 0) return false;
3113
+ for (const o of horizontals) {
3114
+ if (o === line) continue;
3115
+ const dy = o.y1 - line.y1;
3116
+ if (dy < maxH * UNDER_BOX_PAIR_MIN_EM || dy > maxH * UNDER_BOX_PAIR_MAX_EM) continue;
3117
+ const ow = o.x2 - o.x1;
3118
+ if (ow <= 0) continue;
3119
+ const overlap = Math.min(line.x2, o.x2) - Math.max(line.x1, o.x1);
3120
+ if (overlap / Math.max(w, ow) >= UNDER_BOX_PAIR_OVERLAP) return true;
3121
+ }
3122
+ return false;
3123
+ }
3124
+ function isRepeatedSpanRule(line, horizontals) {
3125
+ const w = line.x2 - line.x1;
3126
+ if (w <= 0) return false;
3127
+ const ys = [];
3128
+ for (const o of horizontals) {
3129
+ const ow = o.x2 - o.x1;
3130
+ if (ow <= 0) continue;
3131
+ const overlap = Math.min(line.x2, o.x2) - Math.max(line.x1, o.x1);
3132
+ if (overlap / Math.max(w, ow) < UNDER_SPAN_OVERLAP_RATIO) continue;
3133
+ if (!ys.some((y) => Math.abs(y - o.y1) < 2)) ys.push(o.y1);
3134
+ }
3135
+ return ys.length >= UNDER_REPEATED_SPAN_LEVELS;
3136
+ }
3137
+ function wrapUnderlineRuns(items) {
3138
+ const marked = items.filter((i) => i.underline);
3139
+ if (marked.length === 0) return;
3140
+ const lines = /* @__PURE__ */ new Map();
3141
+ for (const item of marked) {
3142
+ const key = Math.round(item.y / 3);
3143
+ const arr = lines.get(key) || [];
3144
+ arr.push(item);
3145
+ lines.set(key, arr);
3146
+ }
3147
+ for (const arr of lines.values()) {
3148
+ arr.sort((a, b) => a.x - b.x);
3149
+ let runStart = 0;
3150
+ for (let i = 1; i <= arr.length; i++) {
3151
+ const prev = arr[i - 1];
3152
+ const em = Math.max(prev.h > 0 ? prev.h : prev.fontSize, 1);
3153
+ const gap = i < arr.length ? arr[i].x - (prev.x + prev.w) : Infinity;
3154
+ if (gap > em) {
3155
+ arr[runStart].text = "<u>" + arr[runStart].text;
3156
+ prev.text = prev.text + "</u>";
3157
+ runStart = i;
3158
+ }
3159
+ }
3160
+ }
3161
+ }
3162
+
2903
3163
  // src/pdf/page-blocks.ts
2904
3164
  function extractPageBlocksWithLines(items, pageNum, opList, pageWidth, pageHeight, extraLines, detectTables = true) {
2905
3165
  if (items.length === 0) return [];
@@ -2915,6 +3175,8 @@ function extractPageBlocksWithLines(items, pageNum, opList, pageWidth, pageHeigh
2915
3175
  verticals = bridgeSplitColumnVerticals(horizontals, verticals);
2916
3176
  markStrikethroughItems(items, horizontals);
2917
3177
  wrapStrikethroughRuns(items);
3178
+ markUnderlineItems(items, horizontals, verticals);
3179
+ wrapUnderlineRuns(items);
2918
3180
  const grids = detectTables ? buildTableGrids(horizontals, verticals) : [];
2919
3181
  if (grids.length > 0) {
2920
3182
  return extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals);
@@ -3117,6 +3379,9 @@ function extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals) {
3117
3379
  blocks.push({ type: "table", table: irTable, pageNumber: pageNum, bbox: tableBbox });
3118
3380
  }
3119
3381
  let remaining = items.filter((i) => !usedItems.has(i));
3382
+ const groupSizes = [];
3383
+ let finalTextBlocks = [];
3384
+ let gutterX = null;
3120
3385
  if (remaining.length > 0) {
3121
3386
  remaining.sort((a, b) => b.y - a.y || a.x - b.x);
3122
3387
  const clusterItems = remaining.map((i) => ({
@@ -3143,9 +3408,37 @@ function extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals) {
3143
3408
  }
3144
3409
  remaining = remaining.filter((_, idx) => !usedClusterIndices.has(idx));
3145
3410
  }
3146
- const groupSizes = [];
3147
- let finalTextBlocks = [];
3148
- if (remaining.length > 0) {
3411
+ }
3412
+ {
3413
+ const rects = remaining.map((i) => ({ x: i.x, y: i.y, w: i.w, h: i.h > 0 ? i.h : i.fontSize }));
3414
+ for (const b of blocks) {
3415
+ if (b.bbox) rects.push({ x: b.bbox.x, y: b.bbox.y, w: b.bbox.width, h: b.bbox.height });
3416
+ }
3417
+ gutterX = detectColumnGutter(rects);
3418
+ }
3419
+ if (remaining.length > 0) {
3420
+ if (gutterX !== null) {
3421
+ const gx = gutterX;
3422
+ const allY = remaining.map((i) => i.y);
3423
+ const pageH = safeMax(allY) - safeMin(allY);
3424
+ const gapThreshold = Math.max(15, pageH * 0.03);
3425
+ const sides = [
3426
+ remaining.filter((i) => i.x + i.w <= gx),
3427
+ remaining.filter((i) => i.x < gx && i.x + i.w > gx),
3428
+ remaining.filter((i) => i.x >= gx)
3429
+ ];
3430
+ const textBlocks = [];
3431
+ for (const side of sides) {
3432
+ if (side.length === 0) continue;
3433
+ for (const group of xyCutOrder(side, gapThreshold)) {
3434
+ if (group.length === 0) continue;
3435
+ const groupBlocks = extractPageBlocksFallback(group, pageNum);
3436
+ for (const b of groupBlocks) textBlocks.push(b);
3437
+ groupSizes.push(groupBlocks.length);
3438
+ }
3439
+ }
3440
+ finalTextBlocks = detectListBlocks(textBlocks);
3441
+ } else {
3149
3442
  const allY = remaining.map((i) => i.y);
3150
3443
  const pageH = safeMax(allY) - safeMin(allY);
3151
3444
  const groups = xyCutOrder(remaining, Math.max(15, pageH * 0.03));
@@ -3158,26 +3451,43 @@ function extractBlocksWithGrids(items, pageNum, grids, horizontals, verticals) {
3158
3451
  }
3159
3452
  finalTextBlocks = detectListBlocks(textBlocks);
3160
3453
  }
3161
- const units = blocks.map((b) => [b]);
3162
- let off = 0;
3163
- for (const size of groupSizes) {
3164
- const unit = finalTextBlocks.slice(off, off + size);
3165
- off += size;
3166
- if (unit.length > 0) units.push(unit);
3454
+ }
3455
+ const units = blocks.map((b) => [b]);
3456
+ let off = 0;
3457
+ for (const size of groupSizes) {
3458
+ const unit = finalTextBlocks.slice(off, off + size);
3459
+ off += size;
3460
+ if (unit.length > 0) units.push(unit);
3461
+ }
3462
+ const unitTopY = (u) => {
3463
+ let top = 0;
3464
+ for (const b of u) {
3465
+ if (b.bbox && b.bbox.y + b.bbox.height > top) top = b.bbox.y + b.bbox.height;
3167
3466
  }
3168
- const unitTopY = (u) => {
3169
- let top = 0;
3467
+ return top;
3468
+ };
3469
+ if (gutterX !== null && units.length > 1) {
3470
+ const gx = gutterX;
3471
+ const unitRect = (u) => {
3472
+ let minX = Infinity, minY = Infinity, maxR = -Infinity, maxT = -Infinity;
3170
3473
  for (const b of u) {
3171
- if (b.bbox && b.bbox.y + b.bbox.height > top) top = b.bbox.y + b.bbox.height;
3172
- }
3173
- return top;
3474
+ if (!b.bbox) continue;
3475
+ if (b.bbox.x < minX) minX = b.bbox.x;
3476
+ if (b.bbox.y < minY) minY = b.bbox.y;
3477
+ if (b.bbox.x + b.bbox.width > maxR) maxR = b.bbox.x + b.bbox.width;
3478
+ if (b.bbox.y + b.bbox.height > maxT) maxT = b.bbox.y + b.bbox.height;
3479
+ }
3480
+ if (!Number.isFinite(minX)) return { x: gx - 1, y: 0, w: 2, h: 0 };
3481
+ return { x: minX, y: minY, w: maxR - minX, h: maxT - minY };
3174
3482
  };
3175
- units.sort((a, b) => unitTopY(b) - unitTopY(a));
3176
- const ordered = [];
3177
- for (const u of units) for (const b of u) ordered.push(b);
3178
- return mergeAdjacentTableBlocks(ordered);
3483
+ const ordered2 = [];
3484
+ for (const u of orderByGutter(units, unitRect, gx)) for (const b of u) ordered2.push(b);
3485
+ return mergeAdjacentTableBlocks(ordered2);
3179
3486
  }
3180
- return mergeAdjacentTableBlocks(blocks);
3487
+ units.sort((a, b) => unitTopY(b) - unitTopY(a));
3488
+ const ordered = [];
3489
+ for (const u of units) for (const b of u) ordered.push(b);
3490
+ return mergeAdjacentTableBlocks(ordered);
3181
3491
  }
3182
3492
  var NEIGHBOR_TABLE_EPSILON = 0.2;
3183
3493
  function mergeCrossPageTables(blocks) {
@@ -3326,7 +3636,10 @@ function extractPageBlocksFallback(items, pageNum, fullPage = false, detectTable
3326
3636
  return by - ay;
3327
3637
  });
3328
3638
  } else {
3329
- const proseCutX = fullPage ? findTwoColumnProseCutX(items) : null;
3639
+ let proseCutX = fullPage ? findTwoColumnProseCutX(items) : null;
3640
+ if (proseCutX === null && fullPage) {
3641
+ proseCutX = detectColumnGutter(items.map((i) => ({ x: i.x, y: i.y, w: i.w, h: i.h > 0 ? i.h : i.fontSize })));
3642
+ }
3330
3643
  const allYLines = mergeSuperscriptLines(groupByY(items));
3331
3644
  const columns = proseCutX !== null || !detectTables ? null : detectColumns(allYLines);
3332
3645
  if (columns && columns.length >= 3) {
@@ -3374,4 +3687,4 @@ export {
3374
3687
  extractPageBlocksWithLines,
3375
3688
  mergeCrossPageTables
3376
3689
  };
3377
- //# sourceMappingURL=chunk-PQDNCWAZ.js.map
3690
+ //# sourceMappingURL=chunk-PEVPKD6Z.js.map