*/
abstract class AbstractHtmlProcessor
{
/**
* @var string
*/
const DEFAULT_DOCUMENT_TYPE = '';
/**
* @var string
*/
const CONTENT_TYPE_META_TAG = '';
/**
* @var string Regular expression part to match tag names that PHP's DOMDocument implementation is not aware are
* self-closing. These are mostly HTML5 elements, but for completeness (obsolete) and
* (deprecated) are also included.
*
* @see https://bugs.php.net/bug.php?id=73175
*/
const PHP_UNRECOGNIZED_VOID_TAGNAME_MATCHER = '(?:command|embed|keygen|source|track|wbr)';
/**
* @var \DOMDocument
*/
protected $domDocument = null;
/**
* @param string $unprocessedHtml raw HTML, must be UTF-encoded, must not be empty
*
* @throws \InvalidArgumentException if $unprocessedHtml is anything other than a non-empty string
*/
public function __construct($unprocessedHtml)
{
if (!\is_string($unprocessedHtml)) {
throw new \InvalidArgumentException('The provided HTML must be a string.', 1515459744);
}
if ($unprocessedHtml === '') {
throw new \InvalidArgumentException('The provided HTML must not be empty.', 1515763647);
}
$this->setHtml($unprocessedHtml);
}
/**
* Sets the HTML to process.
*
* @param string $html the HTML to process, must be UTF-8-encoded
*
* @return void
*/
private function setHtml($html)
{
$this->createUnifiedDomDocument($html);
}
/**
* Provides access to the internal DOMDocument representation of the HTML in its current state.
*
* @return \DOMDocument
*/
public function getDomDocument()
{
return $this->domDocument;
}
/**
* Renders the normalized and processed HTML.
*
* @return string
*/
public function render()
{
$htmlWithPossibleErroneousClosingTags = $this->domDocument->saveHTML();
return $this->removeSelfClosingTagsClosingTags($htmlWithPossibleErroneousClosingTags);
}
/**
* Renders the content of the BODY element of the normalized and processed HTML.
*
* @return string
*/
public function renderBodyContent()
{
$htmlWithPossibleErroneousClosingTags = $this->domDocument->saveHTML($this->getBodyElement());
$bodyNodeHtml = $this->removeSelfClosingTagsClosingTags($htmlWithPossibleErroneousClosingTags);
return \str_replace(['', ''], '', $bodyNodeHtml);
}
/**
* Eliminates any invalid closing tags for void elements from the given HTML.
*
* @param string $html
*
* @return string
*/
private function removeSelfClosingTagsClosingTags($html)
{
return \preg_replace('%%', '', $html);
}
/**
* Returns the BODY element.
*
* This method assumes that there always is a BODY element.
*
* @return \DOMElement
*/
private function getBodyElement()
{
return $this->domDocument->getElementsByTagName('body')->item(0);
}
/**
* Creates a DOM document from the given HTML and stores it in $this->domDocument.
*
* The DOM document will always have a BODY element and a document type.
*
* @param string $html
*
* @return void
*/
private function createUnifiedDomDocument($html)
{
$this->createRawDomDocument($html);
$this->ensureExistenceOfBodyElement();
}
/**
* Creates a DOMDocument instance from the given HTML and stores it in $this->domDocument.
*
* @param string $html
*
* @return void
*/
private function createRawDomDocument($html)
{
$domDocument = new \DOMDocument();
$domDocument->strictErrorChecking = false;
$domDocument->formatOutput = true;
$libXmlState = \libxml_use_internal_errors(true);
$domDocument->loadHTML($this->prepareHtmlForDomConversion($html));
\libxml_clear_errors();
\libxml_use_internal_errors($libXmlState);
$this->domDocument = $domDocument;
}
/**
* Returns the HTML with added document type, Content-Type meta tag, and self-closing slashes, if needed,
* ensuring that the HTML will be good for creating a DOM document from it.
*
* @param string $html
*
* @return string the unified HTML
*/
private function prepareHtmlForDomConversion($html)
{
$htmlWithSelfClosingSlashes = $this->ensurePhpUnrecognizedSelfClosingTagsAreXml($html);
$htmlWithDocumentType = $this->ensureDocumentType($htmlWithSelfClosingSlashes);
return $this->addContentTypeMetaTag($htmlWithDocumentType);
}
/**
* Makes sure that the passed HTML has a document type.
*
* @param string $html
*
* @return string HTML with document type
*/
private function ensureDocumentType($html)
{
$hasDocumentType = \stripos($html, '/i', '' . static::CONTENT_TYPE_META_TAG, $html);
} elseif ($hasHtmlTag) {
$reworkedHtml = \preg_replace(
'//i',
'' . static::CONTENT_TYPE_META_TAG . '',
$html
);
} else {
$reworkedHtml = static::CONTENT_TYPE_META_TAG . $html;
}
return $reworkedHtml;
}
/**
* Makes sure that any self-closing tags not recognized as such by PHP's DOMDocument implementation have a
* self-closing slash.
*
* @param string $html
*
* @return string HTML with problematic tags converted.
*/
private function ensurePhpUnrecognizedSelfClosingTagsAreXml($html)
{
return \preg_replace(
'%<' . static::PHP_UNRECOGNIZED_VOID_TAGNAME_MATCHER . '\\b[^>]*+(?)%',
'$0/',
$html
);
}
/**
* Checks that $this->domDocument has a BODY element and adds it if it is missing.
*
* @return void
*/
private function ensureExistenceOfBodyElement()
{
if ($this->domDocument->getElementsByTagName('body')->item(0) !== null) {
return;
}
$htmlElement = $this->domDocument->getElementsByTagName('html')->item(0);
$htmlElement->appendChild($this->domDocument->createElement('body'));
}
}