Updated

Author: Balearica

js/export/writeTabular.js

@@ -3,6 +3,20 @@ import ocr from '../objects/ocrObjects.js';
 import { inputData, opt } from '../containers/app.js';
 import { extractTableContent } from '../extractTables.js';
 
+/**
+ * Convert a 0-based column index to an Excel column reference (A, B, ..., Z, AA, AB, ...).
+ * @param {number} index
+ */
+function colIndexToRef(index) {
+  let ref = '';
+  let n = index;
+  do {
+    ref = String.fromCharCode(65 + (n % 26)) + ref;
+    n = Math.floor(n / 26) - 1;
+  } while (n >= 0);
+  return ref;
+}
+
 /**
  * @param {Object} params
  * @param {ReturnType<extractTableContent>} params.tableWordObj
@@ -206,3 +220,35 @@ export async function writeXlsx({
 
   return zipFileData;
 }
+
+/**
+ * Create a single-sheet xlsx workbook from plain string data.
+ * @param {Array<Array<string>>} rows - 2D array of cell values.
+ */
+export async function writeXlsxFromStrings(rows) {
+  const { xlsxStrings, sheetStart, sheetEnd } = await import('./resources/xlsxFiles.js');
+  const { Uint8ArrayWriter, TextReader, ZipWriter } = await import('../../lib/zip.js/index.js');
+
+  let sheetContent = sheetStart;
+  for (let i = 0; i < rows.length; i++) {
+    sheetContent += `<row r="${String(i + 1)}">`;
+    for (let j = 0; j < rows[i].length; j++) {
+      const cellText = ocr.escapeXml(rows[i][j] ?? '');
+      sheetContent += `<c r="${colIndexToRef(j)}${String(i + 1)}" t="inlineStr"><is><t xml:space="preserve">${cellText}</t></is></c>`;
+    }
+    sheetContent += '</row>';
+  }
+  sheetContent += sheetEnd;
+
+  const zipFileWriter = new Uint8ArrayWriter();
+  const zipWriter = new ZipWriter(zipFileWriter);
+
+  await zipWriter.add('xl/worksheets/sheet1.xml', new TextReader(sheetContent));
+
+  for (let i = 0; i < xlsxStrings.length; i++) {
+    await zipWriter.add(xlsxStrings[i].path, new TextReader(xlsxStrings[i].content));
+  }
+
+  await zipWriter.close();
+  return await zipFileWriter.getData();
+}

scribe.js

@@ -41,7 +41,7 @@ import {
   calcConf, checkOcrWordsAdjacent, mergeOcrWords, splitOcrWord,
 } from './js/utils/ocrUtils.js';
 import { assignParagraphs } from './js/utils/reflowPars.js';
-import { writeXlsx } from './js/export/writeTabular.js';
+import { writeXlsx, writeXlsxFromStrings } from './js/export/writeTabular.js';
 import { calcColumnBounds, detectTablesInPage, makeTableFromBbox } from './js/utils/detectTables.js';
 import { ca } from './js/canvasAdapter.js';
 import { addHighlights, clearHighlights } from './js/addHighlights.js';
@@ -234,6 +234,8 @@ class utils {
 
   static writeXlsx = writeXlsx;
 
+  static writeXlsxFromStrings = writeXlsxFromStrings;
+
   // Misc utils
   static calcBoxOverlap = calcBoxOverlap;
 

tests/module/exportPdf.spec.js

@@ -191,6 +191,65 @@ describe('Check addHighlights and clearHighlights.', function () {
     assert.throws(() => scribe.addHighlights([{ page: 0 }]));
   }).timeout(10000);
 
+  it('addHighlights with multi-line range highlights all lines including the last', async () => {
+    await scribe.importFiles([`${ASSETS_PATH_KARMA}/testocr.abbyy.xml`]);
+    const lineCount = scribe.data.ocr.active[0].lines.length;
+    assert.isAbove(lineCount, 3, 'Document should have more than 3 lines');
+
+    const result = scribe.addHighlights([{ page: 0, startLine: 0, endLine: lineCount - 1 }]);
+    assert.strictEqual(result.totalLinesHighlighted, lineCount);
+
+    // Verify the last line specifically has all words annotated
+    const lastLine = scribe.data.ocr.active[0].lines[lineCount - 1];
+    const annots = scribe.data.annotations.pages[0];
+    let lastLineAnnotCount = 0;
+    for (const word of lastLine.words) {
+      if (annots.some((a) => a.bbox === word.bbox)) lastLineAnnotCount++;
+    }
+    assert.strictEqual(lastLineAnnotCount, lastLine.words.length, 'All words on the last line should be annotated');
+    scribe.clearHighlights();
+  }).timeout(10000);
+
+  after(async () => {
+    await scribe.terminate();
+  });
+}).timeout(120000);
+
+describe('Check MCP highlight flow preserves data file.', function () {
+  this.timeout(30000);
+
+  it('Multi-line highlights should work when ensureFileLoaded is called without dataFile on already-loaded document', async () => {
+    const { ensureFileLoaded, resetState } = await import('../../mcp/index.js');
+    resetState();
+
+    const imgPath = `${ASSETS_PATH_KARMA}/testocr.png`;
+    const dataPath = `${ASSETS_PATH_KARMA}/testocr.abbyy.xml`;
+
+    await ensureFileLoaded(imgPath, dataPath);
+    const lineCount = scribe.data.ocr.active[0].lines.length;
+    assert.strictEqual(lineCount, 8);
+
+    await ensureFileLoaded(imgPath, undefined);
+    const lineCountAfter = scribe.data.ocr.active[0]?.lines?.length ?? 0;
+
+    assert.strictEqual(lineCountAfter, 8);
+
+    // Verify multi-line highlights work for the full range including the last line
+    const result = scribe.addHighlights([{ page: 0, startLine: 0, endLine: lineCount - 1 }]);
+    assert.strictEqual(result.totalLinesHighlighted, lineCount);
+
+    const lastLine = scribe.data.ocr.active[0].lines[lineCount - 1];
+    const annots = scribe.data.annotations.pages[0];
+    let lastLineAnnotCount = 0;
+    for (const word of lastLine.words) {
+      if (annots.some((a) => a.bbox === word.bbox)) lastLineAnnotCount++;
+    }
+    assert.strictEqual(lastLineAnnotCount, 7);
+
+    scribe.clearHighlights();
+    resetState();
+  }).timeout(20000);
+
   after(async () => {
     await scribe.terminate();
   });