function load_html($html) { $tidy = tidy_parse_string($html); tidy_clean_repair($tidy); $html = tidy_get_html($tidy); phpQuery::unloadDocuments(); return phpQuery::newDocumentHTML($html); }
function getContent($url, $encording = 'utf8') { $config = array('output-xhtml' => true, 'quote-nbsp' => false, 'indent' => false, 'wrap' => 800, 'char-encoding' => 'raw'); $client = new Zend_Http_Client($url); $response = $client->request("GET"); if ($response->isSuccessful()) { $html = $response->getBody(); if ($encording != 'utf8') { $html = iconv("windows-1251", "utf-8", $html); } $tidy = new Tidy(); $tidy->parseString($html, $config, 'utf8'); $tidy = tidy_get_html($tidy); $thisxml = new SimpleXMLElement($tidy); } return $thisxml; }