From e72b86336217299ba0a8ca602bd1832f477769fc Mon Sep 17 00:00:00 2001
From: Amit Haridas <amit.wh@gmail.com>
Date: Tue, 30 Jun 2026 13:57:29 +0530
Subject: [PATCH] fix(word): strip HTML style blocks and alignment divs from
 DOCX input

Pre-process markdown before Word/DOCX export to remove raw HTML artifacts

(<style> blocks, HTML comments, and <div align=...> tags) that were

visible in the generated document. Applies to single and batch DOCX exports

via both Pandoc and WordTemplateExporter paths.
---
 src/main.js                          | 60 +++++++++++++++++++++++++--
 src/wordTemplateExporter.js          | 23 ++++++++++-
 tests/word-template-exporter.test.js | 61 ++++++++++++++++++++++++++++
 3 files changed, 139 insertions(+), 5 deletions(-)
 create mode 100644 tests/word-template-exporter.test.js

diff --git a/src/main.js b/src/main.js
index fa93015..3385f54 100644
--- a/src/main.js
+++ b/src/main.js
@@ -2562,7 +2562,23 @@ function performExportWithOptions(format, options) {
 
       // Use pandoc for export with advanced options
 
-      let pandocCmd = `${getPandocPath()} "${currentFile}" -o "${outputFile}"`;
+      let inputFile = currentFile;
+      let tempInputFile = null;
+
+      // Pre-process markdown for Word output so that <style> blocks, HTML
+      // comments, and alignment <div> tags do not appear as visible text.
+      if (format === 'docx') {
+        const content = fs.readFileSync(currentFile, 'utf-8');
+        const cleanedContent = WordTemplateExporter.preprocessMarkdownForWordExport(content);
+        tempInputFile = path.join(
+          require('os').tmpdir(),
+          `mc_export_${Date.now()}_${path.basename(currentFile)}`
+        );
+        fs.writeFileSync(tempInputFile, cleanedContent, 'utf-8');
+        inputFile = tempInputFile;
+      }
+
+      let pandocCmd = `${getPandocPath()} "${inputFile}" -o "${outputFile}"`;
 
       // Add template if specified
       if (options.template && options.template !== 'default') {
@@ -2657,7 +2673,15 @@ function performExportWithOptions(format, options) {
         });
       } else if (format === 'docx') {
         pandocCmd += ' -t docx';
-        exportWithPandoc(pandocCmd, outputFile, format);
+        exportWithPandoc(pandocCmd, outputFile, format, () => {
+          if (tempInputFile) {
+            try {
+              fs.unlinkSync(tempInputFile);
+            } catch {
+              // Ignore cleanup errors
+            }
+          }
+        });
       } else if (format === 'pptx') {
         // Add PowerPoint footer if enabled
         if (headerFooterSettings.enabled && headerFooterSettings.footer.center) {
@@ -2859,7 +2883,7 @@ function showExportSuccess(outputFile) {
 }
 
 // Helper function to export with pandoc (general) - uses runPandocCmd for safety
-function exportWithPandoc(pandocCmd, outputFile, format) {
+function exportWithPandoc(pandocCmd, outputFile, format, onComplete) {
   runPandocCmd(pandocCmd, async (error, stdout, stderr) => {
     if (error) {
       console.error(`Pandoc error for ${format}:`, error);
@@ -2924,6 +2948,9 @@ function exportWithPandoc(pandocCmd, outputFile, format) {
       }
       showExportSuccess(outputFile);
     }
+    if (typeof onComplete === 'function') {
+      onComplete(error);
+    }
   });
 }
 
@@ -3749,7 +3776,23 @@ async function performBatchConversion(inputFolder, outputFolder, format, options
     }
 
     // Build pandoc command for other formats
-    let pandocCmd = `${getPandocPath()} "${inputFile}" -o "${outputFile}"`;
+    let pandocInputFile = inputFile;
+    let batchTempInputFile = null;
+
+    // Pre-process markdown for Word output so that <style> blocks, HTML
+    // comments, and alignment <div> tags do not appear as visible text.
+    if (format === 'docx') {
+      const content = fs.readFileSync(inputFile, 'utf-8');
+      const cleanedContent = WordTemplateExporter.preprocessMarkdownForWordExport(content);
+      batchTempInputFile = path.join(
+        require('os').tmpdir(),
+        `mc_batch_export_${Date.now()}_${path.basename(inputFile)}`
+      );
+      fs.writeFileSync(batchTempInputFile, cleanedContent, 'utf-8');
+      pandocInputFile = batchTempInputFile;
+    }
+
+    let pandocCmd = `${getPandocPath()} "${pandocInputFile}" -o "${outputFile}"`;
 
     // Add template if specified
     if (options.template && options.template !== 'default') {
@@ -3850,6 +3893,15 @@ async function performBatchConversion(inputFolder, outputFolder, format, options
 
     // Execute conversion (using runPandocCmd for safety)
     runPandocCmd(pandocCmd, async (error, _stdout, stderr) => {
+      // Clean up temporary pre-processed input file
+      if (batchTempInputFile) {
+        try {
+          fs.unlinkSync(batchTempInputFile);
+        } catch {
+          // Ignore cleanup errors
+        }
+      }
+
       if (error) {
         console.error(`Batch: Failed to convert ${path.basename(inputFile)}:`, error.message, stderr);
       } else {
diff --git a/src/wordTemplateExporter.js b/src/wordTemplateExporter.js
index 8047466..84c61de 100644
--- a/src/wordTemplateExporter.js
+++ b/src/wordTemplateExporter.js
@@ -15,6 +15,26 @@ class WordTemplateExporter {
     this.pageSettings = pageSettings; // Page size and orientation settings
   }
 
+  /**
+   * Strip HTML artifacts that Pandoc / Word cannot render and that would
+   * otherwise appear as visible text in DOCX output.
+   * Removes HTML comments, <style> blocks, and alignment <div> tags.
+   */
+  static preprocessMarkdownForWordExport(markdown) {
+    if (typeof markdown !== 'string') return markdown;
+    return (
+      markdown
+        // HTML comments (multi-line)
+        .replace(/<!--[\s\S]*?-->/g, '')
+        // <style> blocks (case-insensitive, multi-line)
+        .replace(/<style\b[\s\S]*?<\/style>/gi, '')
+        // Opening <div align="..."> tags
+        .replace(/<div\b[^>]*?\balign\s*=\s*["'][^"']*["'][^>]*>/gi, '')
+        // Closing </div> tags
+        .replace(/<\/div\s*>/gi, '')
+    );
+  }
+
   /**
    * Convert markdown to Word document using template
    */
@@ -33,7 +53,8 @@ class WordTemplateExporter {
       }
 
       // Parse markdown and generate Word XML
-      const newContentXml = this.markdownToWordXml(markdownContent);
+      const cleanedContent = WordTemplateExporter.preprocessMarkdownForWordExport(markdownContent);
+      const newContentXml = this.markdownToWordXml(cleanedContent);
 
       // Insert new content after the specified start page
       const modifiedXml = this.insertContentAfterPage(documentXml, newContentXml, this.startPage);
diff --git a/tests/word-template-exporter.test.js b/tests/word-template-exporter.test.js
new file mode 100644
index 0000000..8cb8b7e
--- /dev/null
+++ b/tests/word-template-exporter.test.js
@@ -0,0 +1,61 @@
+/**
+ * Tests for WordTemplateExporter
+ */
+
+const WordTemplateExporter = require('../src/wordTemplateExporter');
+
+describe('WordTemplateExporter.preprocessMarkdownForWordExport', () => {
+  test('removes HTML style blocks', () => {
+    const input = `<style>
+<!-- sneh-a4-print v1 -->
+@media print { body { font-size: 8pt; } }
+</style>
+
+# Heading
+`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).not.toContain('<style');
+    expect(output).not.toContain('</style>');
+    expect(output).not.toContain('sneh-a4-print');
+    expect(output).not.toContain('@media print');
+    expect(output).toContain('# Heading');
+  });
+
+  test('removes HTML comments outside style blocks', () => {
+    const input = `<!-- comment -->
+Hello world
+`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).not.toContain('<!--');
+    expect(output).not.toContain('-->');
+    expect(output).toContain('Hello world');
+  });
+
+  test('removes alignment div tags', () => {
+    const input = `<div align="center">
+Centered content
+</div>
+`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).not.toContain('<div align="center">');
+    expect(output).not.toContain('</div>');
+    expect(output).toContain('Centered content');
+  });
+
+  test('preserves regular markdown content', () => {
+    const input = `# Title
+
+| A | B |
+|---|---|
+| 1 | 2 |
+`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).toBe(input);
+  });
+
+  test('handles content with no HTML artifacts', () => {
+    const input = 'Plain text paragraph.';
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).toBe(input);
+  });
+});