kordoc 4.24.2 → 4.24.4

package/dist/{parser-OBGWEWSU.js → parser-MNNJSFEZ.js}

@@ -1,7 +1,7 @@
 import {
   blocksToPages,
   encodePng
-} from "./chunk-3HN65YAL.js";
+} from "./chunk-MC4DPT4I.js";
 import {
   hasRequestedPagesAfter,
   parsePageRange
@@ -61,7 +61,7 @@ import {
   stripAbsorbedRunningLines,
   trimTrailingEmptyTableCols,
   wrapJoiner
-} from "./chunk-JMOAMKAR.js";
+} from "./chunk-E7N6BUJT.js";
 import {
   CELL_BASELINES,
   CELL_EDGES,
@@ -78,12 +78,12 @@ import {
   stripScriptTags,
   unframeLayoutTables,
   validateOcrOptions
-} from "./chunk-TW3V6FSH.js";
-import "./chunk-E7B5CXYI.js";
+} from "./chunk-VBY3JQL2.js";
+import "./chunk-DTADAHJD.js";
 import {
   KordocError,
   sanitizeHref
-} from "./chunk-WJAFE3Z6.js";
+} from "./chunk-FK3JPTHI.js";
 
 // src/pdf/image-extract.ts
 import { createHash } from "crypto";
@@ -1183,9 +1183,13 @@ function joinSplitParts(prev, pcx, curr, ccx, dx = 0, prevBottom, lex) {
   const cols = U.length - 1;
   if (cols < 1) return null;
   const pa = anchorsOf(prev), ca = anchorsOf(curr);
+  const cutOpen = (() => {
+    const u = ROW_RULES.get(prev), d = ROW_RULES.get(curr);
+    return !!u && !!d && !u.bottom && !d.top;
+  })();
   let skip = 0, head = 0;
-  while (skip < Math.min(3, curr.rows - 1, prev.rows) && rowText(ca, skip) !== "" && rowText(ca, skip) === rowText(pa, skip)) skip++;
-  for (let h = 1; h <= 2 && !skip; h++) {
+  while (!cutOpen && skip < Math.min(3, curr.rows - 1, prev.rows) && rowText(ca, skip) !== "" && rowText(ca, skip) === rowText(pa, skip)) skip++;
+  for (let h = 1; h <= 2 && !skip && !cutOpen; h++) {
     head = h;
     while (skip < Math.min(3, curr.rows - 1, prev.rows - h) && hasText(rowText(ca, skip)) && rowText(ca, skip) === rowText(pa, h + skip)) skip++;
   }
@@ -4003,7 +4007,7 @@ function mergeLinkRuns(markdown) {
 
 // src/pdf/formula-ocr.ts
 async function applyFormulaOcr(buffer, blocks, pageFilter, effectivePageCount, warnings, _onProgress) {
-  const formulaMod = await import("./formula-U2LZ67KI.js");
+  const formulaMod = await import("./formula-MPBZ7OMN.js");
   const { FormulaPipeline, ensureFormulaModels } = formulaMod;
   await ensureFormulaModels((p) => {
     if (p.phase === "download" && p.total) {
@@ -4430,10 +4434,10 @@ async function parsePdfDocument(buffer, options) {
       }
       if (targets.size > 0) {
         try {
-          const { runPdfOcr } = await import("./pdf-ocr-K53REDNY.js");
+          const { runPdfOcr } = await import("./pdf-ocr-5HEVZKG5.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-K53REDNY.js");
+          const { builtinOcrConfig } = await import("./pdf-ocr-5HEVZKG5.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) {
@@ -4678,4 +4682,4 @@ export {
   parsePdfDocument,
   removeHeaderFooterBlocks
 };
-//# sourceMappingURL=parser-OBGWEWSU.js.map
+//# sourceMappingURL=parser-MNNJSFEZ.js.map