/** * Turns HTML into text browser visible text. Images * are converted to their alt text and tags are supressed. * Entities are converted to their visible representation. * @param string $html HTML to convert. * @return string Plain text. * @access public * @static */ function normalise($html) { $text = preg_replace('|<img.*?alt\\s*=\\s*"(.*?)".*?>|', ' \\1 ', $html); $text = preg_replace('|<img.*?alt\\s*=\\s*\'(.*?)\'.*?>|', ' \\1 ', $text); $text = preg_replace('|<img.*?alt\\s*=\\s*([a-zA-Z_]+).*?>|', ' \\1 ', $text); $text = preg_replace('|<.*?>|', '', $text); $text = SimpleSaxParser::decodeHtml($text); $text = preg_replace('|\\s+|', ' ', $text); return trim($text); }