gistfile1.php

<?php

# Unit test for paths imported by feeds
# Not amazing, but will gather 404 info

main();

function main() {
  $paths        = parse_csv('all-paths.csv',',');
  $compare_path = compare_paths($paths);
  render_results($compare_path['pathing'], 'Missing Path & Working Content');
  render_results($compare_path['node'], 'Missing Content & Working Path');
  render_results($compare_path['unpublished'], 'Working & Likely Unpublished');
  render_results($compare_path['misc'], 'Misc');
  render_results($compare_path['missing'], 'Missing Content');
  render_results($compare_path['working'], 'Fully Working');
}

function render_results($results, $title) {
  echo "<h2>" . $title . "</h2>";
  echo "<table>";

  foreach ($results as $result) {
    echo "<tr>";
    echo "<td>" . $result . "</td>";
    echo "</tr>";
  }

  echo "</table>";
}

function parse_csv($filename='', $delimiter=',') {
  $header = NULL;
  $data = array();

  if (($handle = fopen($filename, 'r')) !== FALSE) {
    while (($row = fgetcsv($handle, 1000, $delimiter)) !== FALSE) {
      if(!$header) {
        $header = $row;
      } else {
        $data[] = array_combine($header, $row);
      }
    }

    fclose($handle);
  }

  #print_r($data);
  return $data;
}

function compare_paths($paths) {
  foreach ($paths as $path) {
    $t['p'] = check_url($path['path']);
    $t['a'] = check_url($path['alias']);

    if (($t['p'] == $t['a']) AND ($t['p'] == '404')) {
      $pages['missing'][] = $path['alias'] . "(" . $t['a'] . ")"  . " -> " . $path['path'] . "(" . $t['p'] . ")";
    } elseif (($t['p'] == '200') AND ($t['a'] == '301')) {
      $pages['working'][] = $path['alias'] . "(" . $t['a'] . ")"  . " -> " . $path['path'] . "(" . $t['p'] . ")";
    } elseif (($t['p'] != $t['a']) AND ($t['a'] == '404')) {
      $pages['pathing'][] = $path['alias'] . "(" . $t['a'] . ")"  . " -> " . $path['path'] . "(" . $t['p'] . ")";
    } elseif (($t['p'] != $t['a']) AND ($t['p'] == '404')) {
      $pages['node'][] = $path['alias'] . "(" . $t['a'] . ")"  . " -> " . $path['path'] . "(" . $t['p'] . ")";
    } elseif (($t['a'] == '301') AND ($t['p'] == '403')) {
      $pages['unpublished'][] = $path['alias'] . "(" . $t['a'] . ")"  . " -> " . $path['path'] . "(" . $t['p'] . ")";
    } else {
      $pages['misc'][] = $path['alias'] . "(" . $t['a'] . ")"  . " -> " . $path['path'] . "(" . $t['p'] . ")";
    }
  } 

  #print_r($pages);
  return $pages;
}

function check_url($url) {

  $url = 'http://centerforfaithandwork.com/' . $url;

  $handle = curl_init($url);
  curl_setopt($handle,  CURLOPT_RETURNTRANSFER, TRUE);
  $response = curl_exec($handle);
  $http_code = curl_getinfo($handle, CURLINFO_HTTP_CODE);
  curl_close($handle);

  return $http_code; 
}