HTML parsing support (primarily for Drupal migrate framework, but could be used generally)

html.inc

<?php

/**
 * @file
 * HTML parsing support.
 */

class MigrateHtmlParser {

  protected $html;

  protected $dom;

  protected $elements;

  public function __construct($html) {
    $this->html = $html;
    // DOM can load HTML soup. But, HTML soup can throw warnings, suppress them.
    $old = libxml_use_internal_errors(TRUE);

    $dom = new DOMDocument();
    $dom->loadHTML($html);

    libxml_use_internal_errors($old);

    if ($dom) {
      // It's much easier to work with simplexml than DOM, luckily enough
      // we can just simply import our DOM tree.
      $this->elements = simplexml_import_dom($dom);
    }
    else {
      throw new UnexpectedValueException("The html could not be parsed");
    }
  }

  public function xpath($xpath) {
    $result = $this->elements->xpath($xpath);
    // Some combinations of PHP / libxml versions return an empty array
    // instead of the documented FALSE. Forcefully convert any falsish values
    // to an empty array to allow foreach(...) constructions.
    return $result ? $result : array();
  }

}