1. <?php
  3. /**
  4. * Archives HTML content with manual overrides.
  5. *
  6. * @param string $url The URL of the HTML page to archive.
  7. * @param string $archiveDir The directory to store the archived content.
  8. * @param array $overrides An associative array of HTML elements to override.
  9. * Key: XPath expression, Value: Replacement HTML.
  10. *
  11. * @return bool True on success, false on failure.
  12. */
  13. function archiveHtmlPage(string $url, string $archiveDir, array $overrides = []): bool
  14. {
  15. // Validate inputs
  16. if (!is_readable($url)) {
  17. error_log("Error: URL is not readable: " . $url);
  18. return false;
  19. }
  21. if (!is_dir($archiveDir) || !is_ writable($archiveDir)) {
  22. error_log("Error: Archive directory is not valid or writable: " . $archiveDir);
  23. return false;
  24. }
  26. // Fetch HTML content
  27. try {
  28. $html = file_get_contents($url);
  29. if ($html === false) {
  30. error_log("Error: Failed to fetch HTML content from: " . $url);
  31. return false;
  32. }
  33. } catch (Exception $e) {
  34. error_log("Error fetching HTML content: " . $e->getMessage());
  35. return false;
  36. }
  39. // Apply overrides
  40. foreach ($overrides as $xpath => $replacement) {
  41. try {
  42. $html = preg_replace($xpath, $replacement, $html);
  43. } catch (Exception $e) {
  44. error_log("Error applying override for XPath '$xpath': " . $e->getMessage());
  45. return false;
  46. }
  47. }
  49. // Generate filename
  50. $filename = preg_replace("/[^a-zA-Z0-9._-]/", "_", basename($url)) . ".html";
  51. $filepath = rtrim($archiveDir, '/') . '/' . $filename;
  53. // Save the archived content
  54. if (file_put_contents($filepath, $html) === false) {
  55. error_log("Error: Failed to save archived content to: " . $filepath);
  56. return false;
  57. }
  59. return true;
  60. }
  62. // Example Usage (replace with your actual values)
  63. $url = 'https://www.example.com';
  64. $archiveDir = '/path/to/archive';
  65. $overrides = [
  66. '//div[@class="header"]' => '<div class="custom-header">Custom Header</div>', //Example override
  67. ];
  69. if (archiveHtmlPage($url, $archiveDir, $overrides)) {
  70. echo "Archive created successfully!\n";
  71. } else {
  72. echo "Archive creation failed.\n";
  73. }
  75. ?>

Add your comment