public function testHtmlNoFollowLinks()
{
$html = '<HTML>' . '<HEAD><TITLE>Page title</TITLE></HEAD>' . '<BODY>' . 'Document body.' . '<a href="link1.html">Link 1</a>.' . '<a href="link2.html" rel="nofollow">Link 1</a>.' . '</BODY>' . '</HTML>';
$oldNoFollowValue = Document\HTML::getExcludeNoFollowLinks();
Document\HTML::setExcludeNoFollowLinks(false);
$doc1 = Document\HTML::loadHTML($html);
$this->assertTrue($doc1 instanceof Document\HTML);
$this->assertTrue(array_values($doc1->getLinks()) == array('link1.html', 'link2.html'));
Document\HTML::setExcludeNoFollowLinks(true);
$doc2 = Document\HTML::loadHTML($html);
$this->assertTrue($doc2 instanceof Document\HTML);
$this->assertTrue(array_values($doc2->getLinks()) == array('link1.html'));
}
/**
* Highlight matches in $inputHTMLFragment and return it (without HTML header and body tag)
*
* @param string $inputHTMLFragment
* @param string $encoding Input HTML string encoding
* @param \Zend\Search\Lucene\Search\Highlighter|null $highlighter
* @return string
*/
public function htmlFragmentHighlightMatches($inputHTMLFragment, $encoding = 'UTF-8', $highlighter = null)
{
if ($highlighter === null) {
$highlighter = new Highlighter\DefaultHighlighter();
}
$inputHTML = '<html><head><META HTTP-EQUIV="Content-type" CONTENT="text/html; charset=UTF-8"/></head><body>' . iconv($encoding, 'UTF-8//IGNORE', $inputHTMLFragment) . '</body></html>';
$doc = Document\HTML::loadHTML($inputHTML);
$highlighter->setDocument($doc);
$this->_highlightMatches($highlighter);
return $doc->getHTMLBody();
}
