create DOM through Tidy

tidy-dom-xpath.php

<?php

$html = file_get_contents($argv[1]);

$dom = dom($html);

$xpath = new DOMXPath($dom);
foreach ($xpath->query('//a') as $e){
  var_dump($e->nodeValue);
}

function dom($html){
  $config = array(
    'output-xml' => true,
    'numeric-entities' => true,
  );

  $tidy = tidy_parse_string($html, $config, 'UTF8');
  $tidy->cleanRepair();
  $html = tidy_get_output($tidy);

  //var_dump($html);

  // PHP Warning:  DOMDocument::loadXML(): StartTag: invalid element name in Entity
  $html = preg_replace('/\<\!\[(.*?)\]\>/um', '<!--[\1]-->', $html);

  // loadXML() keeps multibyte strings even if partial html given
  $html = '<?xml version="1.0" encoding="UTF-8"?>' . "\n" . $html;

  //libxml_use_internal_errors(true);
  $doc = new DOMDocument();
  $doc->loadXML($html);
  //$doc->loadHTML($html);
  //libxml_clear_errors();
  //var_dump($doc->saveXML());
  return $doc;
}