Neah/lib/utils/email-content.ts

import DOMPurify from 'dompurify';
import { detectTextDirection, applyTextDirection } from './text-direction';
import { sanitizeHtml } from './email-utils';

/**
 * Format and standardize email content for display following email industry standards.
 * This function handles various email content formats and ensures proper display
 * including support for HTML emails, plain text emails, RTL languages, and email client quirks.
 */
export function formatEmailContent(email: any): string {
  if (!email) {
    console.log('formatEmailContent: No email provided');
    return '';
  }
  
  try {
    // Get the content in order of preference with proper fallbacks
    let content = '';
    let isHtml = false;
    let textContent = '';
    
    // Extract content based on standardized property hierarchy
    if (email.content && typeof email.content === 'object') {
      isHtml = !!email.content.html;
      content = email.content.html || '';
      textContent = email.content.text || '';
    } else if (typeof email.content === 'string') {
      // Check if the string content is HTML
      isHtml = email.content.trim().startsWith('<') && 
               (email.content.includes('<html') || 
                email.content.includes('<body') || 
                email.content.includes('<div') || 
                email.content.includes('<p>'));
      content = email.content;
      textContent = email.content;
    } else if (email.html) {
      isHtml = true;
      content = email.html;
      textContent = email.text || '';
    } else if (email.text) {
      isHtml = false;
      content = '';
      textContent = email.text;
    } else if (email.formattedContent) {
      // Assume formattedContent is already HTML
      isHtml = true;
      content = email.formattedContent;
      textContent = '';
    }
    
    // Log what we found for debugging
    console.log(`Email content detected: isHtml=${isHtml}, contentLength=${content.length}, textLength=${textContent.length}`);
    
    // If we have HTML content, sanitize and standardize it
    if (isHtml && content) {
      // CRITICAL FIX: Check for browser environment since DOMParser is browser-only
      const hasHtmlTag = content.includes('<html');
      const hasBodyTag = content.includes('<body');
      
      // Extract body content if we have a complete HTML document and in browser environment
      if (hasHtmlTag && hasBodyTag && typeof window !== 'undefined' && typeof DOMParser !== 'undefined') {
        try {
          // Create a DOM parser to extract just the body content
          const parser = new DOMParser();
          const doc = parser.parseFromString(content, 'text/html');
          const bodyContent = doc.body.innerHTML;
          
          if (bodyContent) {
            content = bodyContent;
          }
        } catch (error) {
          // If extraction fails, continue with the original content
          console.error('Error extracting body content:', error);
        }
      }
      
      // Use the centralized sanitizeHtml function
      const sanitizedContent = sanitizeHtml(content);
      
      // Fix common email client quirks
      let fixedContent = sanitizedContent
        // Fix for Outlook WebVML content
        .replace(/<!--\[if\s+gte\s+mso/g, '<!--[if gte mso')
        // Fix for broken image paths that might be relative
        .replace(/(src|background)="(?!(?:https?:|data:|cid:))/gi, '$1="https://')
        // Fix for base64 images that might be broken across lines
        .replace(/src="data:image\/[^;]+;base64,\s*([^"]+)\s*"/gi, (match, p1) => {
          return `src="data:image/png;base64,${p1.replace(/\s+/g, '')}"`;
        });
      
      // Use the centralized text direction utility
      const styledContent = `<div style="font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Helvetica, Arial, sans-serif; line-height: 1.6; color: #333; max-width: 100%; overflow-x: auto; overflow-wrap: break-word; word-wrap: break-word;">${fixedContent}</div>`;
      
      // Apply correct text direction using the centralized utility
      return applyTextDirection(styledContent, textContent);
    } 
    // If we only have text content, format it properly
    else if (textContent) {
      // Escape HTML characters to prevent XSS
      const escapedText = textContent
        .replace(/&/g, '&amp;')
        .replace(/</g, '&lt;')
        .replace(/>/g, '&gt;')
        .replace(/"/g, '&quot;')
        .replace(/'/g, '&#039;');
      
      // Format plain text with proper line breaks and paragraphs
      const formattedText = escapedText
        .replace(/\r\n|\r|\n/g, '<br>') // Convert all newlines to <br>
        .replace(/((?:<br>){2,})/g, '</p><p>') // Convert multiple newlines to paragraphs
        .replace(/<br><\/p>/g, '</p>') // Fix any <br></p> combinations
        .replace(/<p><br>/g, '<p>'); // Fix any <p><br> combinations
      
      const styledPlainText = `<div style="font-family: -apple-system, BlinkMacSystemFont, Menlo, Monaco, Consolas, 'Courier New', monospace; white-space: pre-wrap; line-height: 1.5; color: #333; padding: 15px; max-width: 100%; overflow-wrap: break-word;"><p>${formattedText}</p></div>`;
      
      // Apply correct text direction using the centralized utility
      return applyTextDirection(styledPlainText, textContent);
    }
    
    // Default case: empty or unrecognized content
    return '<div class="email-content-empty" style="padding: 20px; text-align: center; color: #666;">No content available</div>';
  } catch (error) {
    console.error('formatEmailContent: Error formatting email content:', error);
    return `<div class="email-content-error" style="padding: 15px; color: #721c24; background-color: #f8d7da; border: 1px solid #f5c6cb; border-radius: 4px;"><p>Error displaying email content</p><p style="font-size: 12px; margin-top: 10px;">${error instanceof Error ? error.message : 'Unknown error'}</p></div>`;
  }
}