fix for missing table cells in incomplete tables

#282
html-to-text · Feb 10, 2023 · c049ac6 · c049ac6
1 parent 6336d9b
commit c049ac6
Show file tree

Hide file tree

Showing 3 changed files with 51 additions and 17 deletions.
diff --git a/.eslintrc.cjs b/.eslintrc.cjs
@@ -90,7 +90,7 @@ module.exports = {
     'max-lines-per-function': ['error', 80],
     'max-nested-callbacks': ['error', 4],
     'max-params': ['error', 6],
-    'max-statements': ['error', 35],
+    'max-statements': ['error', 41],
     'max-statements-per-line': ['error', { 'max': 2 }],
     'multiline-ternary': ['error', 'always-multiline'],
     'newline-per-chained-call': ['error', { 'ignoreChainWithDepth': 3 }],

diff --git a/packages/html-to-text/src/table-printer.js b/packages/html-to-text/src/table-printer.js
@@ -18,9 +18,11 @@ function transposeInPlace (matrix, maxSize) {
     const rowI = getRow(matrix, i);
     for (let j = 0; j < i; j++) {
       const rowJ = getRow(matrix, j);
-      const temp = rowI[j];
-      rowI[j] = rowJ[i];
-      rowJ[i] = temp;
+      if (rowI[j] || rowJ[i]) {
+        const temp = rowI[j];
+        rowI[j] = rowJ[i];
+        rowJ[i] = temp;
+      }
     }
   }
 }
@@ -34,10 +36,17 @@ function putCellIntoLayout (cell, layout, baseRow, baseCol) {
   }
 }
 
+function getOrInitOffset (offsets, index) {
+  if (offsets[index] === undefined) {
+    offsets[index] = (index === 0) ? 0 : 1 + getOrInitOffset(offsets, index - 1);
+  }
+  return offsets[index];
+}
+
 function updateOffset (offsets, base, span, value) {
   offsets[base + span] = Math.max(
-    offsets[base + span] || 0,
-    offsets[base] + value
+    getOrInitOffset(offsets, base + span),
+    getOrInitOffset(offsets, base) + value
   );
 }
 
@@ -82,19 +91,27 @@ function tableToString (tableRows, rowSpacing, colSpacing) {
   for (let x = 0; x < colNumber; x++) {
     let y = 0;
     let cell;
-    while (y < rowNumber && (cell = layout[x][y])) {
-      if (!cell.rendered) {
-        let cellWidth = 0;
-        for (let j = 0; j < cell.lines.length; j++) {
-          const line = cell.lines[j];
-          const lineOffset = rowOffsets[y] + j;
-          outputLines[lineOffset] = (outputLines[lineOffset] || '').padEnd(colOffsets[x]) + line;
-          cellWidth = (line.length > cellWidth) ? line.length : cellWidth;
+    const rowsInThisColumn = Math.min(rowNumber, layout[x].length);
+    while (y < rowsInThisColumn) {
+      cell = layout[x][y];
+      if (cell) {
+        if (!cell.rendered) {
+          let cellWidth = 0;
+          for (let j = 0; j < cell.lines.length; j++) {
+            const line = cell.lines[j];
+            const lineOffset = rowOffsets[y] + j;
+            outputLines[lineOffset] = (outputLines[lineOffset] || '').padEnd(colOffsets[x]) + line;
+            cellWidth = (line.length > cellWidth) ? line.length : cellWidth;
+          }
+          updateOffset(colOffsets, x, cell.colspan, cellWidth + colSpacing);
+          cell.rendered = true;
         }
-        updateOffset(colOffsets, x, cell.colspan, cellWidth + colSpacing);
-        cell.rendered = true;
+        y += cell.rowspan;
+      } else {
+        const lineOffset = rowOffsets[y];
+        outputLines[lineOffset] = (outputLines[lineOffset] || '');
+        y++;
       }
-      y += cell.rowspan;
     }
   }
 

diff --git a/packages/html-to-text/test/tags.js b/packages/html-to-text/test/tags.js
@@ -869,6 +869,23 @@ describe('tags', function () {
       expect(htmlToText(html, options)).to.equal(expected);
     });
 
+    it('should not miss content in tables with variable number of cells per row', function () {
+      const html = `
+      <table>
+      <tr><td>a</td></tr>
+      <tr><td>b</td><td>c</td></tr>
+      <tr></tr>
+      <tr><td>d</td></tr>
+      </table>`;
+      const expected = 'a\nb   c\n\nd';
+      const options = {
+        selectors: [
+          { selector: 'table', format: 'dataTable' }
+        ]
+      };
+      expect(htmlToText(html, options)).to.equal(expected);
+    });
+
   });
 
   describe('custom formatting', function () {