HtmlDomParser::set_callback() - Code Metrics - Inspection of "[+]: added UTF-8 test" - voku/simple_html_dom - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 1483ee...c351f4 )

by Lars

created 2016-07-31 14:13 UTC

HtmlDomParser::set_callback() A

↳ Parent: HtmlDomParser

Complexity

Conditions	1
Paths	1

Size

Total Lines	4
Code Lines	2

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	0
CRAP Score	2

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
c	1
b	0
f	0
dl	0
loc	4
ccs	0
cts	3
cp	0
rs	10
cc	1
eloc	2
nc	1
nop	1
crap	2

<?php

namespace voku\helper;

use BadMethodCallException;
use DOMDocument;
use DOMXPath;
use InvalidArgumentException;
use RuntimeException;

/**
 * Class HtmlDomParser
 *
 * @package voku\helper
 *
 * @property-read string outerText Get dom node's outer html (alias for "outerHtml")
 * @property-read string outerHtml Get dom node's outer html
 * @property-read string innerText Get dom node's inner html (alias for "innerHtml")
 * @property-read string innerHtml Get dom node's inner html
 * @property-read string plaintext Get dom node's plain text
 *
 * @method string outerText() Get dom node's outer html (alias for "outerHtml()")
 * @method string outerHtml() Get dom node's outer html
 * @method string innerText() Get dom node's inner html (alias for "innerHtml()")
 * @method HtmlDomParser load() load($html) Load HTML from string
 * @method HtmlDomParser load_file() load_file($html) Load HTML from file
 *
 * @method static HtmlDomParser file_get_html() file_get_html($html, $libXMLExtraOptions = null) Load HTML from file
 * @method static HtmlDomParser str_get_html() str_get_html($html, $libXMLExtraOptions = null) Load HTML from string
 */
class HtmlDomParser
{
  /**
   * @var array
   */
  protected static $functionAliases = array(
      'outertext' => 'html',
      'outerhtml' => 'html',
      'innertext' => 'innerHtml',
      'innerhtml' => 'innerHtml',
      'load'      => 'loadHtml',
      'load_file' => 'loadHtmlFile',
  );

  /**
   * @var array
   */
  private static $domLinkReplaceHelper = array(
      'orig' => array('[', ']', '{', '}',),
      'tmp'  => array(
          '!!!!HTML_DOM__SQUARE_BRACKET_LEFT!!!!',
          '!!!!HTML_DOM__SQUARE_BRACKET_RIGHT!!!!',
          '!!!!HTML_DOM__BRACKET_LEFT!!!!',
          '!!!!HTML_DOM__BRACKET_RIGHT!!!!',
      ),
  );

  /**
   * @var array
   */
  protected static $domReplaceHelper = array(
      'orig' => array('&', '|', '+', '%'),
      'tmp'  => array(
          '!!!!HTML_DOM__AMP!!!!',
          '!!!!HTML_DOM__PIPE!!!!',
          '!!!!HTML_DOM__PLUS!!!!',
          '!!!!HTML_DOM__PERCENT!!!!',
      ),
  );

  /**
   * @var Callable
   */
  protected static $callback;

  /**
   * @var DOMDocument
   */
  protected $document;

  /**
   * @var string
   */
  protected $encoding = 'UTF-8';

  /**
   * @var bool
   */
  protected $isDOMDocumentCreatedWithoutHtml = false;

  /**
   * @var bool
   */
  protected $isDOMDocumentCreatedWithoutHtmlWrapper = false;

  /**
   * Constructor
   *
   * @param string|SimpleHtmlDom|\DOMNode $element HTML code or SimpleHtmlDom, \DOMNode
   */
  public function __construct($element = null)
  {
    $this->document = new \DOMDocument('1.0', $this->getEncoding());

    // DOMDocument settings
    $this->document->preserveWhiteSpace = true;
    $this->document->formatOutput = true;

    if ($element instanceof SimpleHtmlDom) {
      $element = $element->getNode();
    }

    if ($element instanceof \DOMNode) {
      $domNode = $this->document->importNode($element, true);

      if ($domNode instanceof \DOMNode) {
        $this->document->appendChild($domNode);
      }

      return;
    }

    if ($element !== null) {
      $this->loadHtml($element);
    }
  }

  /**
   * @param $name
   * @param $arguments
   *
   * @return bool|mixed
   */
  public function __call($name, $arguments)
  {
    $name = strtolower($name);

    if (isset(self::$functionAliases[$name])) {

      return call_user_func_array(array($this, self::$functionAliases[$name]), $arguments);
    }

    throw new BadMethodCallException('Method does not exist: ' . $name);
  }

  /**
   * @param $name
   * @param $arguments
   *
   * @return HtmlDomParser
   */
  public static function __callStatic($name, $arguments)
  {
    $arguments0 = null;
    if (isset($arguments[0])) {
      $arguments0 = $arguments[0];
    }

    $arguments1 = null;
    if (isset($arguments[1])) {
      $arguments1 = $arguments[1];
    }

    if ($name == 'str_get_html') {
      $parser = new self();

      return $parser->loadHtml($arguments0, $arguments1);
    }

    if ($name == 'file_get_html') {
      $parser = new self();

      return $parser->loadHtmlFile($arguments0, $arguments1);
    }

    throw new BadMethodCallException('Method does not exist');
  }

  /**
   * @param $name
   *
   * @return string
   */
  public function __get($name)
  {
    $name = strtolower($name);

    switch ($name) {
      case 'outerhtml':
      case 'outertext':
        return $this->html();
      case 'innerhtml':
      case 'innertext':
        return $this->innerHtml();
      case 'text':
      case 'plaintext':
        return $this->text();
    }

    return null;
  }

  /**
   * @param string $selector
   * @param int    $idx
   *
   * @return SimpleHtmlDom|SimpleHtmlDomNode|null
   */
  public function __invoke($selector, $idx = null)
  {
    return $this->find($selector, $idx);
  }

  /**
   * @return string
   */
  public function __toString()
  {
    return $this->html();
  }

  /**
   * does nothing (only for api-compatibility-reasons)
   *
   * @return bool
   */
  public function clear()
  {
    return true;
  }

  /**
   * @param string $html
   *
   * @return string
   */
  private function replaceToPreserveHtmlEntities($html)
  {
    preg_match_all("/(\bhttps?:\/\/[^\s()<>]+(?:\([\w\d]+\)|[^[:punct:]\s]|\/|\}|\]))/i", $html, $linksOld);

    $linksNew = array();
    if (!empty($linksOld[1])) {
      $linksOld = $linksOld[1];
      foreach ($linksOld as $linkKey => $linkOld) {
        $linksNew[$linkKey] = str_replace(
            self::$domLinkReplaceHelper['orig'],
            self::$domLinkReplaceHelper['tmp'],
            $linkOld
        );
      }
    }

    $linksNewCount = count($linksNew);
    if ($linksNewCount > 0 && count($linksOld) === $linksNewCount) {
      $search = array_merge($linksOld, self::$domReplaceHelper['orig']);
      $replace = array_merge($linksNew, self::$domReplaceHelper['tmp']);
    } else {
      $search = self::$domReplaceHelper['orig'];
      $replace = self::$domReplaceHelper['tmp'];
    }

    return str_replace($search, $replace, $html);
  }

  /**
   * @param string $html
   *
   * @return string
   */
  public static function putReplacedBackToPreserveHtmlEntities($html)
  {
    return str_replace(
        array_merge(
            self::$domLinkReplaceHelper['tmp'],
            self::$domReplaceHelper['tmp'],
            array('&#13;')
        ),
        array_merge(
            self::$domLinkReplaceHelper['orig'],
            self::$domReplaceHelper['orig'],
            array('')
        ),
        $html
    );
  }

  /**
   * create DOMDocument from HTML
   *
   * @param string   $html
   * @param int|null $libXMLExtraOptions
   *
   * @return \DOMDocument
   */
  private function createDOMDocument($html, $libXMLExtraOptions = null)
  {
    if (strpos($html, '<') === false) {
      $this->isDOMDocumentCreatedWithoutHtml = true;
    }

    if (strpos($html, '<html') === false) {
      $this->isDOMDocumentCreatedWithoutHtmlWrapper = true;
    }

    // set error level
    $internalErrors = libxml_use_internal_errors(true);
    $disableEntityLoader = libxml_disable_entity_loader(true);
    libxml_clear_errors();

    $options = LIBXML_DTDLOAD | LIBXML_DTDATTR | LIBXML_NONET;

    if (defined('LIBXML_COMPACT')) {
      $options |= LIBXML_COMPACT;
    }

    if (defined('LIBXML_HTML_NOIMPLIED')) {
      $options |= LIBXML_HTML_NOIMPLIED;
    }

    if (defined('LIBXML_HTML_NODEFDTD')) {
      $options |= LIBXML_HTML_NODEFDTD;
    }

    if ($libXMLExtraOptions !== null) {
      $options |= $libXMLExtraOptions;
    }

    $sxe = simplexml_load_string($html, 'SimpleXMLElement', $options);
    if ($sxe !== false && count(libxml_get_errors()) === 0) {
      $this->document = dom_import_simplexml($sxe)->ownerDocument;
    } else {

      // UTF-8 hack: http://php.net/manual/en/domdocument.loadhtml.php#95251
      $html = trim($html);
      $xmlHackUsed = false;
      if (stripos('<?xml', $html) !== 0) {
        $xmlHackUsed = true;
        $html = '<?xml encoding="' . $this->getEncoding() . '" ?>' . $html;
      }

      $html = $this->replaceToPreserveHtmlEntities($html);

      $this->document->loadHTML($html);

      // remove the "xml-encoding" hack
      if ($xmlHackUsed === true) {
        foreach ($this->document->childNodes as $child) {
          if ($child->nodeType == XML_PI_NODE) {
            $this->document->removeChild($child);
          }
        }
      }

      libxml_clear_errors();
    }

    // set encoding
    $this->document->encoding = $this->getEncoding();

    // restore lib-xml settings
    libxml_use_internal_errors($internalErrors);
    libxml_disable_entity_loader($disableEntityLoader);

    return $this->document;
  }

  /**
   * Return SimpleHtmlDom by id.
   *
   * @param string $id
   *
   * @return SimpleHtmlDom|SimpleHtmlDomNodeBlank
   */
  public function getElementById($id)
  {
    return $this->find("#$id", 0);
  }

  /**
   * Return SimpleHtmlDom by tag name.
   *
   * @param string $name
   *
   * @return SimpleHtmlDom|SimpleHtmlDomNodeBlank
   */
  public function getElementByTagName($name)
  {
    $node = $this->document->getElementsByTagName($name)->item(0);

    if ($node !== null) {
      return new SimpleHtmlDom($node);
    } else {
      return new SimpleHtmlDomNodeBlank();
    }
  }

  /**
   * Returns Elements by id
   *
   * @param string   $id
   * @param null|int $idx
   *
   * @return SimpleHtmlDomNode|SimpleHtmlDomNode[]|SimpleHtmlDomNodeBlank
   */
  public function getElementsById($id, $idx = null)
  {
    return $this->find("#$id", $idx);
  }

  /**
   * Returns Elements by tag name
   *
   * @param string   $name
   * @param null|int $idx
   *
   * @return SimpleHtmlDomNode|SimpleHtmlDomNode[]|SimpleHtmlDomNodeBlank
   */
  public function getElementsByTagName($name, $idx = null)

  {
    $nodesList = $this->document->getElementsByTagName($name);

    $elements = new SimpleHtmlDomNode();

    foreach ($nodesList as $node) {
      $elements[] = new SimpleHtmlDom($node);
    }

    if (null === $idx) {
      return $elements;
    } else {
      if ($idx < 0) {
        $idx = count($elements) + $idx;
      }
    }

    if (isset($elements[$idx])) {
      return $elements[$idx];
    } else {
      return new SimpleHtmlDomNodeBlank();
    }
  }

  /**
   * Find list of nodes with a CSS selector.
   *
   * @param string $selector
   * @param int    $idx
   *
   * @return SimpleHtmlDom|SimpleHtmlDom[]|SimpleHtmlDomNodeBlank
   */
  public function find($selector, $idx = null)
  {
    $xPathQuery = SelectorConverter::toXPath($selector);

    $xPath = new DOMXPath($this->document);
    $nodesList = $xPath->query($xPathQuery);
    $elements = new SimpleHtmlDomNode();

    foreach ($nodesList as $node) {
      $elements[] = new SimpleHtmlDom($node);
    }

    if (null === $idx) {
      return $elements;
class Author {
    private $name;

    public function __construct($name) {
        $this->name = $name;
    }

    public function getName() {
        return $this->name;
    }
}

abstract class Post {
    public function getAuthor() {
        return 'Johannes';
    }
}

class BlogPost extends Post {
    public function getAuthor() {
        return new Author('Johannes');
    }
}

class ForumPost extends Post { /* ... */ }

function my_function(Post $post) {
    echo strtoupper($post->getAuthor());
}
    } else {
      if ($idx < 0) {
        $idx = count($elements) + $idx;
      }
    }

    if (isset($elements[$idx])) {
      return $elements[$idx];
    } else {
      return new SimpleHtmlDomNodeBlank();
    }
  }

  /**
   * @param string $content
   *
   * @return string
   */
  protected function fixHtmlOutput($content)
  {
    // INFO: DOMDocument will encapsulate plaintext into a paragraph tag (<p>),
    //          so we try to remove it here again ...

    if ($this->isDOMDocumentCreatedWithoutHtmlWrapper === true) {
      $content = str_replace(
          array(
              "\n",
              "\r\n",
              "\r",
              '<simpleHtmlDomP>',
              '</simpleHtmlDomP>',
              '<body>',
              '</body>',
              '<html>',
              '</html>',
          ),
          '',
          $content
      );
    }

    if ($this->isDOMDocumentCreatedWithoutHtml === true) {
      $content = str_replace(
          array(
              '<p>',
              '</p>',
              '<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">'
          ),
          '',
          $content);
    }

    $content = UTF8::html_entity_decode($content);
    $content = trim($content);
    $content = UTF8::urldecode($content);

    $content = self::putReplacedBackToPreserveHtmlEntities($content);

    return $content;
  }

  /**
   * @return DOMDocument
   */
  public function getDocument()
  {
    return $this->document;
  }

  /**
   * Get the encoding to use
   *
   * @return string
   */
  private function getEncoding()
  {
    return $this->encoding;
  }

  /**
   * @return bool
   */
  public function getIsDOMDocumentCreatedWithoutHtml()
  {
    return $this->isDOMDocumentCreatedWithoutHtml;
  }

  /**
   * @return bool
   */
  public function getIsDOMDocumentCreatedWithoutHtmlWrapper()
  {
    return $this->isDOMDocumentCreatedWithoutHtmlWrapper;
  }

  /**
   * Get dom node's outer html
   *
   * @return string
   */
  public function html()
  {
    if ($this::$callback !== null) {
      call_user_func_array($this::$callback, array($this));
    }

    if ($this->getIsDOMDocumentCreatedWithoutHtmlWrapper()) {
      $content = $this->document->saveHTML($this->document->documentElement);
    } else {
      $content = $this->document->saveHTML();
    }

    return $this->fixHtmlOutput($content);
  }

  /**
   * Get the HTML as XML.
   *
   * @return string
   */
  public function xml()
  {
    $xml = $this->document->saveXML(null, LIBXML_NOEMPTYTAG);

    // remove the XML-header
    $xml = ltrim(preg_replace('/<\?xml.*\?>/', '', $xml));

    return $this->fixHtmlOutput($xml);
  }

  /**
   * Get dom node's inner html
   *
   * @return string
   */
  public function innerHtml()
  {
    $text = '';

    foreach ($this->document->documentElement->childNodes as $node) {
      $text .= $this->fixHtmlOutput($this->document->saveHTML($node));
    }

    return $text;
  }

  /**
   * Load HTML from string
   *
   * @param string   $html
   * @param int|null $libXMLExtraOptions
   *
   * @return HtmlDomParser
   *
   * @throws InvalidArgumentException if argument is not string
   */
  public function loadHtml($html, $libXMLExtraOptions = null)
  {
    if (!is_string($html)) {
      throw new InvalidArgumentException(__METHOD__ . ' expects parameter 1 to be string.');
    }

    $this->document = $this->createDOMDocument($html, $libXMLExtraOptions);

    return $this;
  }

  /**
   * Load HTML from file
   *
   * @param string   $filePath
   * @param int|null $libXMLExtraOptions
   *
   * @return HtmlDomParser
   */
  public function loadHtmlFile($filePath, $libXMLExtraOptions = null)
  {
    if (!is_string($filePath)) {
      throw new InvalidArgumentException(__METHOD__ . ' expects parameter 1 to be string.');
    }

    if (!preg_match("/^https?:\/\//i", $filePath) && !file_exists($filePath)) {
      throw new RuntimeException("File $filePath not found");
    }

    try {
      $html = file_get_contents($filePath);

    } catch (\Exception $e) {
      throw new RuntimeException("Could not load file $filePath");
    }

    if ($html === false) {
      throw new RuntimeException("Could not load file $filePath");
    }

    $this->loadHtml($html, $libXMLExtraOptions);

    return $this;
  }

  /**
   * Save dom as string
   *
   * @param string $filepath
   *
   * @return string
   */
  public function save($filepath = '')
  {
    $string = $this->innerHtml();
    if ($filepath !== '') {
      file_put_contents($filepath, $string, LOCK_EX);
    }

    return $string;
  }

  /**
   * @param $functionName
   */
  public function set_callback($functionName)
  {
    $this::$callback = $functionName;
  }

  /**
   * Get dom node's plain text
   *
   * @return string
   */
  public function text()
  {
    return $this->fixHtmlOutput($this->document->textContent);
  }
}


1			<?php
2
3			namespace voku\helper;
4
5			use BadMethodCallException;
6			use DOMDocument;
7			use DOMXPath;
8			use InvalidArgumentException;
9			use RuntimeException;
10
11			/**
12			* Class HtmlDomParser
13			*
14			* @package voku\helper
15			*
16			* @property-read string outerText Get dom node's outer html (alias for "outerHtml")
17			* @property-read string outerHtml Get dom node's outer html
18			* @property-read string innerText Get dom node's inner html (alias for "innerHtml")
19			* @property-read string innerHtml Get dom node's inner html
20			* @property-read string plaintext Get dom node's plain text
21			*
22			* @method string outerText() Get dom node's outer html (alias for "outerHtml()")
23			* @method string outerHtml() Get dom node's outer html
24			* @method string innerText() Get dom node's inner html (alias for "innerHtml()")
25			* @method HtmlDomParser load() load($html) Load HTML from string
26			* @method HtmlDomParser load_file() load_file($html) Load HTML from file
27			*
28			* @method static HtmlDomParser file_get_html() file_get_html($html, $libXMLExtraOptions = null) Load HTML from file
29			* @method static HtmlDomParser str_get_html() str_get_html($html, $libXMLExtraOptions = null) Load HTML from string
30			*/
31			class HtmlDomParser
32			{
33			/**
34			* @var array
35			*/
36			protected static $functionAliases = array(
37			'outertext' => 'html',
38			'outerhtml' => 'html',
39			'innertext' => 'innerHtml',
40			'innerhtml' => 'innerHtml',
41			'load' => 'loadHtml',
42			'load_file' => 'loadHtmlFile',
43			);
44
45			/**
46			* @var array
47			*/
48			private static $domLinkReplaceHelper = array(
49			'orig' => array('[', ']', '{', '}',),
50			'tmp' => array(
51			'!!!!HTML_DOM__SQUARE_BRACKET_LEFT!!!!',
52			'!!!!HTML_DOM__SQUARE_BRACKET_RIGHT!!!!',
53			'!!!!HTML_DOM__BRACKET_LEFT!!!!',
54			'!!!!HTML_DOM__BRACKET_RIGHT!!!!',
55			),
56			);
57
58			/**
59			* @var array
60			*/
61			protected static $domReplaceHelper = array(
62			'orig' => array('&', '\|', '+', '%'),
63			'tmp' => array(
64			'!!!!HTML_DOM__AMP!!!!',
65			'!!!!HTML_DOM__PIPE!!!!',
66			'!!!!HTML_DOM__PLUS!!!!',
67			'!!!!HTML_DOM__PERCENT!!!!',
68			),
69			);
70
71			/**
72			* @var Callable
73			*/
74			protected static $callback;
75
76			/**
77			* @var DOMDocument
78			*/
79			protected $document;
80
81			/**
82			* @var string
83			*/
84			protected $encoding = 'UTF-8';
85
86			/**
87			* @var bool
88			*/
89			protected $isDOMDocumentCreatedWithoutHtml = false;
90
91			/**
92			* @var bool
93			*/
94			protected $isDOMDocumentCreatedWithoutHtmlWrapper = false;
95
96			/**
97			* Constructor
98			*
99			* @param string\|SimpleHtmlDom\|\DOMNode $element HTML code or SimpleHtmlDom, \DOMNode
100			*/
101	115		public function __construct($element = null)
102			{
103	115		$this->document = new \DOMDocument('1.0', $this->getEncoding());
104
105			// DOMDocument settings
106	115		$this->document->preserveWhiteSpace = true;
107	115		$this->document->formatOutput = true;
108
109	115		if ($element instanceof SimpleHtmlDom) {
110	51		$element = $element->getNode();
111	51		}
112
113	115		if ($element instanceof \DOMNode) {
114	51		$domNode = $this->document->importNode($element, true);
115
116	51		if ($domNode instanceof \DOMNode) {
117	51		$this->document->appendChild($domNode);
118	51		}
119
120	51		return;
121			}
122
123	115		if ($element !== null) {
124	70		$this->loadHtml($element);
125	69		}
126	114		}
127
128			/**
129			* @param $name
130			* @param $arguments
131			*
132			* @return bool\|mixed
133			*/
134	32		public function __call($name, $arguments)
135			{
136	32		$name = strtolower($name);
137
138	32	View Code Duplication	if (isset(self::$functionAliases[$name])) {
			0 ignored issues – show Duplication introduced 2016-07-18 06:48 UTC by Report Bug Copy Issue Report This code seems to be duplicated across your project. Duplicated code is one of the most pungent code smells. If you need to duplicate the same code in three or more different places, we strongly encourage you to look into extracting the code into a single class or operation. You can also find more detailed suggestions in the “Code” section of your repository. Loading history...
139	31		return call_user_func_array(array($this, self::$functionAliases[$name]), $arguments);
140			}
141
142	1		throw new BadMethodCallException('Method does not exist: ' . $name);
143			}
144
145			/**
146			* @param $name
147			* @param $arguments
148			*
149			* @return HtmlDomParser
150			*/
151	13		public static function __callStatic($name, $arguments)
152			{
153	13		$arguments0 = null;
154	13		if (isset($arguments[0])) {
155	12		$arguments0 = $arguments[0];
156	12		}
157
158	13		$arguments1 = null;
159	13		if (isset($arguments[1])) {
160	1		$arguments1 = $arguments[1];
161	1		}
162
163	13		if ($name == 'str_get_html') {
164	9		$parser = new self();
165
166	9		return $parser->loadHtml($arguments0, $arguments1);
167			}
168
169	4		if ($name == 'file_get_html') {
170	3		$parser = new self();
171
172	3		return $parser->loadHtmlFile($arguments0, $arguments1);
173			}
174
175	1		throw new BadMethodCallException('Method does not exist');
176			}
177
178			/**
179			* @param $name
180			*
181			* @return string
182			*/
183	14		public function __get($name)
184			{
185	14		$name = strtolower($name);
186
187			switch ($name) {
188	14		case 'outerhtml':
189	14		case 'outertext':
190	8		return $this->html();
191	6		case 'innerhtml':
192	6		case 'innertext':
193	4		return $this->innerHtml();
194	2		case 'text':
195	2		case 'plaintext':
196	1		return $this->text();
197			}
198
199	1		return null;
200			}
201
202			/**
203			* @param string $selector
204			* @param int $idx
205			*
206			* @return SimpleHtmlDom\|SimpleHtmlDomNode\|null
207			*/
208	2		public function __invoke($selector, $idx = null)
209			{
210	2		return $this->find($selector, $idx);
211			}
212
213			/**
214			* @return string
215			*/
216	14		public function __toString()
217			{
218	14		return $this->html();
219			}
220
221			/**
222			* does nothing (only for api-compatibility-reasons)
223			*
224			* @return bool
225			*/
226	1		public function clear()
227			{
228	1		return true;
229			}
230
231			/**
232			* @param string $html
233			*
234			* @return string
235			*/
236	70		private function replaceToPreserveHtmlEntities($html)
237			{
238	70		preg_match_all("/(\bhttps?:\/\/[^\s()<>]+(?:\([\w\d]+\)\|[^[:punct:]\s]\|\/\|\}\|\]))/i", $html, $linksOld);
239
240	70		$linksNew = array();
241	70		if (!empty($linksOld[1])) {
242	49		$linksOld = $linksOld[1];
243	49		foreach ($linksOld as $linkKey => $linkOld) {
244	49		$linksNew[$linkKey] = str_replace(
245	49		self::$domLinkReplaceHelper['orig'],
246	49		self::$domLinkReplaceHelper['tmp'],
247			$linkOld
248	49		);
249	49		}
250	49		}
251
252	70		$linksNewCount = count($linksNew);
253	70		if ($linksNewCount > 0 && count($linksOld) === $linksNewCount) {
254	49		$search = array_merge($linksOld, self::$domReplaceHelper['orig']);
255	49		$replace = array_merge($linksNew, self::$domReplaceHelper['tmp']);
256	49		} else {
257	22		$search = self::$domReplaceHelper['orig'];
258	22		$replace = self::$domReplaceHelper['tmp'];
259			}
260
261	70		return str_replace($search, $replace, $html);
262			}
263
264			/**
265			* @param string $html
266			*
267			* @return string
268			*/
269	53		public static function putReplacedBackToPreserveHtmlEntities($html)
270			{
271	53		return str_replace(
272	53		array_merge(
273	53		self::$domLinkReplaceHelper['tmp'],
274	53		self::$domReplaceHelper['tmp'],
275	53		array(' ')
276	53		),
277	53		array_merge(
278	53		self::$domLinkReplaceHelper['orig'],
279	53		self::$domReplaceHelper['orig'],
280	53		array('')
281	53		),
282			$html
283	53		);
284			}
285
286			/**
287			* create DOMDocument from HTML
288			*
289			* @param string $html
290			* @param int\|null $libXMLExtraOptions
291			*
292			* @return \DOMDocument
293			*/
294	103		private function createDOMDocument($html, $libXMLExtraOptions = null)
295			{
296	103		if (strpos($html, '<') === false) {
297	6		$this->isDOMDocumentCreatedWithoutHtml = true;
298	6		}
299
300	103		if (strpos($html, '<html') === false) {
301	57		$this->isDOMDocumentCreatedWithoutHtmlWrapper = true;
302	57		}
303
304			// set error level
305	103		$internalErrors = libxml_use_internal_errors(true);
306	103		$disableEntityLoader = libxml_disable_entity_loader(true);
307	103		libxml_clear_errors();
308
309	103		$options = LIBXML_DTDLOAD \| LIBXML_DTDATTR \| LIBXML_NONET;
310
311	103		if (defined('LIBXML_COMPACT')) {
312	103		$options \|= LIBXML_COMPACT;
313	103		}
314
315	103		if (defined('LIBXML_HTML_NOIMPLIED')) {
316	103		$options \|= LIBXML_HTML_NOIMPLIED;
317	103		}
318
319	103		if (defined('LIBXML_HTML_NODEFDTD')) {
320	103		$options \|= LIBXML_HTML_NODEFDTD;
321	103		}
322
323	103		if ($libXMLExtraOptions !== null) {
324	1		$options \|= $libXMLExtraOptions;
325	1		}
326
327	103		$sxe = simplexml_load_string($html, 'SimpleXMLElement', $options);
328	103		if ($sxe !== false && count(libxml_get_errors()) === 0) {
329	35		$this->document = dom_import_simplexml($sxe)->ownerDocument;
330	35		} else {
331
332			// UTF-8 hack: http://php.net/manual/en/domdocument.loadhtml.php#95251
333	70		$html = trim($html);
334	70		$xmlHackUsed = false;
335	70		if (stripos('<?xml', $html) !== 0) {
336	70		$xmlHackUsed = true;
337	70		$html = '<?xml encoding="' . $this->getEncoding() . '" ?>' . $html;
338	70		}
339
340	70		$html = $this->replaceToPreserveHtmlEntities($html);
341
342	70		$this->document->loadHTML($html);
343
344			// remove the "xml-encoding" hack
345	70		if ($xmlHackUsed === true) {
346	70		foreach ($this->document->childNodes as $child) {
347	70		if ($child->nodeType == XML_PI_NODE) {
348	70		$this->document->removeChild($child);
349	70		}
350	70		}
351	70		}
352
353	70		libxml_clear_errors();
354			}
355
356			// set encoding
357	103		$this->document->encoding = $this->getEncoding();
358
359			// restore lib-xml settings
360	103		libxml_use_internal_errors($internalErrors);
361	103		libxml_disable_entity_loader($disableEntityLoader);
362
363	103		return $this->document;
364			}
365
366			/**
367			* Return SimpleHtmlDom by id.
368			*
369			* @param string $id
370			*
371			* @return SimpleHtmlDom\|SimpleHtmlDomNodeBlank
372			*/
373	2		public function getElementById($id)
374			{
375	2		return $this->find("#$id", 0);
376			}
377
378			/**
379			* Return SimpleHtmlDom by tag name.
380			*
381			* @param string $name
382			*
383			* @return SimpleHtmlDom\|SimpleHtmlDomNodeBlank
384			*/
385	1		public function getElementByTagName($name)
386			{
387	1		$node = $this->document->getElementsByTagName($name)->item(0);
388
389	1		if ($node !== null) {
390	1		return new SimpleHtmlDom($node);
391			} else {
392			return new SimpleHtmlDomNodeBlank();
393			}
394			}
395
396			/**
397			* Returns Elements by id
398			*
399			* @param string $id
400			* @param null\|int $idx
401			*
402			* @return SimpleHtmlDomNode\|SimpleHtmlDomNode[]\|SimpleHtmlDomNodeBlank
403			*/
404			public function getElementsById($id, $idx = null)
405			{
406			return $this->find("#$id", $idx);
407			}
408
409			/**
410			* Returns Elements by tag name
411			*
412			* @param string $name
413			* @param null\|int $idx
414			*
415			* @return SimpleHtmlDomNode\|SimpleHtmlDomNode[]\|SimpleHtmlDomNodeBlank
416			*/
417	3	View Code Duplication	public function getElementsByTagName($name, $idx = null)
			0 ignored issues – show Duplication introduced 2016-06-07 08:47 UTC by Report Bug Copy Issue Report This method seems to be duplicated in your project. Duplicated code is one of the most pungent code smells. If you need to duplicate the same code in three or more different places, we strongly encourage you to look into extracting the code into a single class or operation. You can also find more detailed suggestions in the “Code” section of your repository. Loading history...
418			{
419	3		$nodesList = $this->document->getElementsByTagName($name);
420
421	3		$elements = new SimpleHtmlDomNode();
422
423	3		foreach ($nodesList as $node) {
424	3		$elements[] = new SimpleHtmlDom($node);
425	3		}
426
427	3		if (null === $idx) {
428	2		return $elements;
429			} else {
430	1		if ($idx < 0) {
431			$idx = count($elements) + $idx;
432			}
433			}
434
435	1		if (isset($elements[$idx])) {
436	1		return $elements[$idx];
437			} else {
438			return new SimpleHtmlDomNodeBlank();
439			}
440			}
441
442			/**
443			* Find list of nodes with a CSS selector.
444			*
445			* @param string $selector
446			* @param int $idx
447			*
448			* @return SimpleHtmlDom\|SimpleHtmlDom[]\|SimpleHtmlDomNodeBlank
449			*/
450	75		public function find($selector, $idx = null)
451			{
452	75		$xPathQuery = SelectorConverter::toXPath($selector);
453
454	75		$xPath = new DOMXPath($this->document);
455	75		$nodesList = $xPath->query($xPathQuery);
456	75		$elements = new SimpleHtmlDomNode();
457
458	75		foreach ($nodesList as $node) {
459	71		$elements[] = new SimpleHtmlDom($node);
460	75		}
461
462	75		if (null === $idx) {
463	48		return $elements;
			0 ignored issues – show Bug Best Practice introduced 2016-07-09 21:10 UTC by Report Bug Copy Issue Report The return type of `return $elements;` (`voku\helper\SimpleHtmlDomNode`) is incompatible with the return type documented by `voku\helper\HtmlDomParser::find` of type `voku\helper\SimpleHtmlDo...\SimpleHtmlDomNodeBlank`. If you return a value from a function or method, it should be a sub-type of the type that is given by the parent type f.e. an interface, or abstract method. This is more formally defined by the Lizkov substitution principle, and guarantees that classes that depend on the parent type can use any instance of a child type interchangably. This principle also belongs to the SOLID principles for object oriented design. Let’s take a look at an example: class Author { private $name; public function __construct($name) { $this->name = $name; } public function getName() { return $this->name; } } abstract class Post { public function getAuthor() { return 'Johannes'; } } class BlogPost extends Post { public function getAuthor() { return new Author('Johannes'); } } class ForumPost extends Post { /* ... */ } function my_function(Post $post) { echo strtoupper($post->getAuthor()); } Our function `my_function` expects a `Post` object, and outputs the author of the post. The base class `Post` returns a simple string and outputting a simple string will work just fine. However, the child class `BlogPost` which is a sub-type of `Post` instead decided to return an `object`, and is therefore violating the SOLID principles. If a `BlogPost` were passed to `my_function`, PHP would not complain, but ultimately fail when executing the `strtoupper` call in its body. Loading history...
464			} else {
465	39		if ($idx < 0) {
466	11		$idx = count($elements) + $idx;
467	11		}
468			}
469
470	39		if (isset($elements[$idx])) {
471	36		return $elements[$idx];
472			} else {
473	5		return new SimpleHtmlDomNodeBlank();
474			}
475			}
476
477			/**
478			* @param string $content
479			*
480			* @return string
481			*/
482	44		protected function fixHtmlOutput($content)
483			{
484			// INFO: DOMDocument will encapsulate plaintext into a paragraph tag (<p>),
485			// so we try to remove it here again ...
486
487	44		if ($this->isDOMDocumentCreatedWithoutHtmlWrapper === true) {
488	19		$content = str_replace(
489			array(
490	19		"\n",
491	19		"\r\n",
492	19		"\r",
493	19		'<simpleHtmlDomP>',
494	19		'</simpleHtmlDomP>',
495	19		'<body>',
496	19		'</body>',
497	19		'<html>',
498	19		'</html>',
499	19		),
500	19		'',
501			$content
502	19		);
503	19		}
504
505	44		if ($this->isDOMDocumentCreatedWithoutHtml === true) {
506	5		$content = str_replace(
507			array(
508	5		'<p>',
509	5		'</p>',
510			'<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">'
511	5		),
512	5		'',
513	5		$content);
514	5		}
515
516	44		$content = UTF8::html_entity_decode($content);
517	44		$content = trim($content);
518	44		$content = UTF8::urldecode($content);
519
520	44		$content = self::putReplacedBackToPreserveHtmlEntities($content);
521
522	44		return $content;
523			}
524
525			/**
526			* @return DOMDocument
527			*/
528	36		public function getDocument()
529			{
530	36		return $this->document;
531			}
532
533			/**
534			* Get the encoding to use
535			*
536			* @return string
537			*/
538	115		private function getEncoding()
539			{
540	115		return $this->encoding;
541			}
542
543			/**
544			* @return bool
545			*/
546	7		public function getIsDOMDocumentCreatedWithoutHtml()
547			{
548	7		return $this->isDOMDocumentCreatedWithoutHtml;
549			}
550
551			/**
552			* @return bool
553			*/
554	32		public function getIsDOMDocumentCreatedWithoutHtmlWrapper()
555			{
556	32		return $this->isDOMDocumentCreatedWithoutHtmlWrapper;
557			}
558
559			/**
560			* Get dom node's outer html
561			*
562			* @return string
563			*/
564	32		public function html()
565			{
566	32		if ($this::$callback !== null) {
567			call_user_func_array($this::$callback, array($this));
568			}
569
570	32		if ($this->getIsDOMDocumentCreatedWithoutHtmlWrapper()) {
571	14		$content = $this->document->saveHTML($this->document->documentElement);
572	14		} else {
573	22		$content = $this->document->saveHTML();
574			}
575
576	32		return $this->fixHtmlOutput($content);
577			}
578
579			/**
580			* Get the HTML as XML.
581			*
582			* @return string
583			*/
584	1		public function xml()
585			{
586	1		$xml = $this->document->saveXML(null, LIBXML_NOEMPTYTAG);
587
588			// remove the XML-header
589	1		$xml = ltrim(preg_replace('/<\?xml.*\?>/', '', $xml));
590
591	1		return $this->fixHtmlOutput($xml);
592			}
593
594			/**
595			* Get dom node's inner html
596			*
597			* @return string
598			*/
599	13		public function innerHtml()
600			{
601	13		$text = '';
602
603	13		foreach ($this->document->documentElement->childNodes as $node) {
604	13		$text .= $this->fixHtmlOutput($this->document->saveHTML($node));
605	13		}
606
607	13		return $text;
608			}
609
610			/**
611			* Load HTML from string
612			*
613			* @param string $html
614			* @param int\|null $libXMLExtraOptions
615			*
616			* @return HtmlDomParser
617			*
618			* @throws InvalidArgumentException if argument is not string
619			*/
620	106		public function loadHtml($html, $libXMLExtraOptions = null)
621			{
622	106		if (!is_string($html)) {
623	3		throw new InvalidArgumentException(__METHOD__ . ' expects parameter 1 to be string.');
624			}
625
626	103		$this->document = $this->createDOMDocument($html, $libXMLExtraOptions);
627
628	103		return $this;
629			}
630
631			/**
632			* Load HTML from file
633			*
634			* @param string $filePath
635			* @param int\|null $libXMLExtraOptions
636			*
637			* @return HtmlDomParser
638			*/
639	11		public function loadHtmlFile($filePath, $libXMLExtraOptions = null)
640			{
641	11		if (!is_string($filePath)) {
642	2		throw new InvalidArgumentException(__METHOD__ . ' expects parameter 1 to be string.');
643			}
644
645	9		if (!preg_match("/^https?:\/\//i", $filePath) && !file_exists($filePath)) {
646	1		throw new RuntimeException("File $filePath not found");
647			}
648
649			try {
650	8		$html = file_get_contents($filePath);
651
652	8		} catch (\Exception $e) {
653	1		throw new RuntimeException("Could not load file $filePath");
654			}
655
656	7		if ($html === false) {
657			throw new RuntimeException("Could not load file $filePath");
658			}
659
660	7		$this->loadHtml($html, $libXMLExtraOptions);
661
662	7		return $this;
663			}
664
665			/**
666			* Save dom as string
667			*
668			* @param string $filepath
669			*
670			* @return string
671			*/
672	1		public function save($filepath = '')
673			{
674	1		$string = $this->innerHtml();
675	1		if ($filepath !== '') {
676			file_put_contents($filepath, $string, LOCK_EX);
677			}
678
679	1		return $string;
680			}
681
682			/**
683			* @param $functionName
684			*/
685			public function set_callback($functionName)
686			{
687			$this::$callback = $functionName;
688			}
689
690			/**
691			* Get dom node's plain text
692			*
693			* @return string
694			*/
695	2		public function text()
696			{
697	2		return $this->fixHtmlOutput($this->document->textContent);
698			}
699			}
700

voku / simple_html_dom

Push — master ( 1483ee...c351f4 )

HtmlDomParser::set_callback() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like