kordoc 4.24.21 → 4.24.23

package/dist/{parser-I2C5DUXC.js → parser-U2WRX53G.js}

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   blocksToPages
-} from "./chunk-BTBFKDPH.js";
+} from "./chunk-SYJZYHBB.js";
 import {
   encodePng
 } from "./chunk-VKVDMH7X.js";
@@ -62,7 +62,7 @@ import {
   stripAbsorbedRunningLines,
   trimTrailingEmptyTableCols,
   wrapJoiner
-} from "./chunk-F6OHQLN7.js";
+} from "./chunk-NWNL5YYK.js";
 import {
   validateOcrOptions
 } from "./chunk-IQJKLIJV.js";
@@ -75,7 +75,7 @@ import {
   stripScriptTags,
   toTex,
   unframeLayoutTables
-} from "./chunk-F4GZFIUJ.js";
+} from "./chunk-3QWZCQI6.js";
 import {
   WINGDINGS,
   assignTableIds,
@@ -84,17 +84,17 @@ import {
   escapeLiteralDollar,
   mapPuaText,
   pairedNumber
-} from "./chunk-B3POEH7X.js";
+} from "./chunk-JDDKQD7W.js";
 import {
   ocrModelsCached,
   resolveOcrModels
-} from "./chunk-E6G4B457.js";
-import "./chunk-O5ZCJEZ4.js";
-import "./chunk-BV2YG7M6.js";
+} from "./chunk-5L32QFLQ.js";
+import "./chunk-AXBEO5NB.js";
+import "./chunk-2S4WERRK.js";
 import {
   KordocError,
   sanitizeHref
-} from "./chunk-WQC4XWVT.js";
+} from "./chunk-WTFXQGND.js";
 
 // src/pdf/image-extract.ts
 import { createHash } from "crypto";
@@ -1512,6 +1512,17 @@ function mergeSplitRow(table, owner, first, colXs, textOnly = false, cut = NO_CU
   const last = first - 1;
   const pairs = [];
   let carried = false;
+  const starts = [];
+  for (let c = 0; c < table.cols; ) {
+    const b = owner[first][c];
+    if (b && b.r === first) starts.push(b);
+    c = b ? b.c + b.cs : c + 1;
+  }
+  const tall = starts.filter((b) => b.rs > 1 && hasContent(b.cell));
+  const band = !textOnly && tall.length > 0 && starts.every((b) => b.rs > 1 || !hasContent(b.cell)) && tall.every((b) => {
+    const a = owner[last][b.c];
+    return !!a && a.rs > 1 && a.c === b.c && a.cs === b.cs && a.r + a.rs - 1 === last && hasContent(a.cell);
+  });
   for (let c = 0; c < table.cols; ) {
     const a = owner[last][c], b = owner[first][c];
     if (!a && !b) {
@@ -1522,7 +1533,7 @@ function mergeSplitRow(table, owner, first, colXs, textOnly = false, cut = NO_CU
     if (!a || !b) return false;
     if (a !== b) {
       if (a.c !== b.c || a.cs !== b.cs || a.r + a.rs - 1 !== last || b.r !== first) return false;
-      if (b.rs > 1 && (a.rs > 1 || !hasContent(a.cell))) return false;
+      if (b.rs > 1 && (a.rs > 1 || !hasContent(a.cell)) && !band) return false;
       pairs.push([a, b]);
     } else if (a.r === last && a.rs === 2) carried = true;
     c = a.c + a.cs;
@@ -1541,7 +1552,7 @@ function mergeSplitRow(table, owner, first, colXs, textOnly = false, cut = NO_CU
   if (cut.ruled && matchesIndependentRowPair(table, owner, first)) return false;
   const contentPairs = pairs.filter(([u, d]) => hasContent(u.cell) && hasContent(d.cell));
   const carriedSplit = !textOnly && carried && !newCell && (contentPairs.length <= 1 || contentPairs.every(([, d]) => d.rs > 1) || !cut.ruled && contentPairs.some(([u, d]) => wordOverflows(cell(u), cell(d), colXs[u.c], colXs[u.c + u.cs])));
-  if (!carriedSplit) {
+  if (!carriedSplit && !band) {
     const norm = (c) => flowText(c.text).replace(/\s+/g, "");
     if (pairs.some(([u, d]) => hasContent(d.cell) && norm(cell(d)) !== norm(cell(u)) && lineEnded(cell(u), colXs[u.c], colXs[u.c + u.cs], cell(d)))) return false;
     const oneLine = (c) => CELL_LINES.get(c)?.length === 1;
@@ -4147,7 +4158,7 @@ function mergeLinkRuns(markdown) {
 
 // src/pdf/formula-ocr.ts
 async function applyFormulaOcr(buffer, blocks, pageFilter, effectivePageCount, warnings, _onProgress) {
-  const formulaMod = await import("./formula-I4EA7ULD.js");
+  const formulaMod = await import("./formula-TI6JRF6D.js");
   const { FormulaPipeline, ensureFormulaModels } = formulaMod;
   await ensureFormulaModels((p) => {
     if (p.phase === "download" && p.total) {
@@ -4577,10 +4588,10 @@ async function parsePdfDocument(buffer, options) {
       }
       if (targets.size > 0) {
         try {
-          const { runPdfOcr } = await import("./pdf-ocr-U3I74Q4O.js");
+          const { runPdfOcr } = await import("./pdf-ocr-HUQ2RFMY.js");
           const mode = typeof options?.ocr === "function" ? options.ocr : "builtin";
           const regionPages = new Map([...ocrRegions].filter(([p]) => options?.ocr !== "force" && !isImageBased && !pageQuality.find((q) => q.page === p)?.needsOcr));
-          const { builtinOcrConfig } = await import("./pdf-ocr-U3I74Q4O.js");
+          const { builtinOcrConfig } = await import("./pdf-ocr-HUQ2RFMY.js");
           const ocrPageBlocks = await runPdfOcr(ocrBuffer, targets, mode, warnings, options?.onProgress, options?.tables !== false, vectorPageOps, regionPages, builtinOcrConfig(options), ocrLinesOut);
           for (const obs of ocrPageBlocks.values()) stripScriptTags({ blocks: obs });
           if (ocrPageBlocks.size > 0) {
@@ -4825,4 +4836,4 @@ export {
   parsePdfDocument,
   removeHeaderFooterBlocks
 };
-//# sourceMappingURL=parser-I2C5DUXC.js.map
+//# sourceMappingURL=parser-U2WRX53G.js.map