Document::detectEncoding() - Code Metrics - Inspection of "Travis CI Bump" - scotteh/php-dom-wrapper - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( c33c36...393be6 )

by Andrew

created 2019-12-08 10:10 UTC

Document::detectEncoding() A

↳ Parent: Document

Complexity

Conditions	4
Paths	4

Size

Total Lines	14
Code Lines	7

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	0
CRAP Score	20

Importance

Changes

Metric	Value
eloc	7
c	0
b	0
f	0
dl	0
loc	14
ccs	0
cts	0
cp	0
rs	10
cc	4
nc	4
nop	1
crap	20

<?php declare(strict_types=1);

namespace DOMWrap;

use DOMWrap\Traits\{
    CommonTrait,
    TraversalTrait,
    ManipulationTrait
};

/**
 * Document Node
 *
 * @package DOMWrap
 * @license http://opensource.org/licenses/BSD-3-Clause BSD 3 Clause
 */
class Document extends \DOMDocument
{
    use CommonTrait;
    use TraversalTrait;
    use ManipulationTrait;

    /** @var int */
    protected $libxmlOptions = 0;

    /** @var string|null */
    protected $documentEncoding = null;

    public function __construct(string $version = '1.0', string $encoding = 'UTF-8') {
        parent::__construct($version, $encoding);

        $this->registerNodeClass('DOMText', 'DOMWrap\\Text');
        $this->registerNodeClass('DOMElement', 'DOMWrap\\Element');
        $this->registerNodeClass('DOMComment', 'DOMWrap\\Comment');
        $this->registerNodeClass('DOMDocument', 'DOMWrap\\Document');
        $this->registerNodeClass('DOMDocumentType', 'DOMWrap\\DocumentType');
        $this->registerNodeClass('DOMProcessingInstruction', 'DOMWrap\\ProcessingInstruction');
    }

    /**
     * Set libxml options.
     *
     * Multiple values must use bitwise OR.
     * eg: LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD
     *
     * @link http://php.net/manual/en/libxml.constants.php
     *
     * @param int $libxmlOptions
     */
    public function setLibxmlOptions(int $libxmlOptions): void {
        $this->libxmlOptions = $libxmlOptions;
    }

    /**
     * {@inheritdoc}
     */
    public function document(): ?\DOMDocument {
        return $this;
    }

    /**
     * {@inheritdoc}
     */
    public function collection(): NodeList {
        return $this->newNodeList([$this]);
    }

    /**
     * {@inheritdoc}
     */
    public function result(NodeList $nodeList) {
        if ($nodeList->count()) {
            return $nodeList->first();
        }

        return null;
    }

    /**
     * {@inheritdoc}
     */
    public function parent() {
        return null;
    }

    /**
     * {@inheritdoc}
     */
    public function parents() {
        return $this->newNodeList();
    }

    /**
     * {@inheritdoc}
     */
    public function replaceWith($newNode): self {
        $this->replaceChild($newNode, $this);

        return $this;
    }

    /**
     * {@inheritdoc}
     */
    public function _clone() {
        return null;
    }

    /**
     * {@inheritdoc}
     */
    public function getHtml(): string {
        return $this->getOuterHtml();
    }

    /**
     * {@inheritdoc}
     */
    public function setHtml($html): self {
        if (!is_string($html) || trim($html) == '') {
            return $this;
        }

        $internalErrors = libxml_use_internal_errors(true);
        $disableEntities = libxml_disable_entity_loader(true);

        $this->detectEncoding($html);

        $html = $this->convertToUtf8($html);

        $this->loadHTML($html, $this->libxmlOptions);

        // Remove <?xml ...> processing instruction.
        $this->contents()->each(function($node) {
            if ($node instanceof ProcessingInstruction && $node->nodeName == 'xml') {
                $node->remove();
            }
        });

        libxml_use_internal_errors($internalErrors);
        libxml_disable_entity_loader($disableEntities);

        return $this;
    }

    /**
     * @param string $html
     * @param int $options
     *
     * @return bool
     */
    public function loadHTML($html, $options = 0): bool {
        // Fix LibXML's crazy-ness RE root nodes
        // While importing HTML using the LIBXML_HTML_NOIMPLIED option LibXML insists
        //  on having one root node. All subsequent nodes are appended to this first node.
        // To counter this we will create a fake element, allow LibXML to 'do its thing'
        //  then undo it by taking the contents of the fake element, placing it back into
        //  the root and then remove our fake element.
        if ($options & LIBXML_HTML_NOIMPLIED) {
            $html = '<domwrap></domwrap>' . $html;
        }

        $html = '<?xml encoding="' . ($this->getEncoding() ?? 'UTF-8') . '">' . $html;

        $result = parent::loadHTML($html, $options);

        // Do our re-shuffling of nodes.
        if ($this->libxmlOptions & LIBXML_HTML_NOIMPLIED) {
            $this->children()->first()->contents()->each(function($node){
                $this->append($node);
            });

            $this->removeChild($this->children()->first());
        }

        return $result;
    }

    /*

     * @param $encoding string|null
     */
    public function setEncoding(string $encoding = null) {
        $this->documentEncoding = $encoding;
    }

    /*

     * @return string|null
     */
    public function getEncoding(): ?string {
        return $this->documentEncoding;
    }

    /*

     * @param $html string
     *
     * @return string|null
     */
    private function getCharset(string $html): ?string {
        $charset = null;

        if (preg_match('@<meta.*?charset=["\']?([^"\'\s>]+)@im', $html, $matches)) {
            $charset = mb_strtoupper($matches[1]);
        }

        return $charset;
    }

    /*
     * @param $html string
     */
    private function detectEncoding(string $html) {
        $charset = $this->getEncoding();

        if (is_null($charset)) {
            $charset = $this->getCharset($html);
        }

        $detectedCharset = mb_detect_encoding($html, mb_detect_order(), true);

        if ($charset === null && $detectedCharset == 'UTF-8') {
            $charset = $detectedCharset;
        }

        $this->setEncoding($charset);
    }

    /*
     * @param $html string
     *
     * @return string
     */
    private function convertToUtf8(string $html): string {
        $charset = $this->getEncoding();

        if ($charset !== null) {
            $html = preg_replace('@(charset=["]?)([^"\s]+)([^"]*["]?)@im', '$1UTF-8$3', $html);
            $mbHasCharset = in_array($charset, array_map('mb_strtoupper', mb_list_encodings()));

            if ($mbHasCharset) {
                $html = mb_convert_encoding($html, 'UTF-8', $charset);

            // Fallback to iconv if available.
            } elseif (extension_loaded('iconv')) {
                $htmlIconv = iconv($charset, 'UTF-8', $html);

                if ($htmlIconv !== false) {
                    $html = $htmlIconv;
                } else {
                    $charset = null;
                }
            }
        }

        if ($charset === null) {
            $html = mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8');
        }

        return $html;
    }
}


1		<?php declare(strict_types=1);
2
3		namespace DOMWrap;
4
5		use DOMWrap\Traits\{
6		CommonTrait,
7		TraversalTrait,
8		ManipulationTrait
9		};
10
11		/**
12		* Document Node
13		*
14		* @package DOMWrap
15		* @license http://opensource.org/licenses/BSD-3-Clause BSD 3 Clause
16		*/
17		class Document extends \DOMDocument
18		{
19		use CommonTrait;
20		use TraversalTrait;
21		use ManipulationTrait;
22
23	140	/** @var int */
24	140	protected $libxmlOptions = 0;
25
26	140	/** @var string\|null */
27	140	protected $documentEncoding = null;
28	140
29	140	public function __construct(string $version = '1.0', string $encoding = 'UTF-8') {
30	140	parent::__construct($version, $encoding);
31	140
32		$this->registerNodeClass('DOMText', 'DOMWrap\\Text');
33		$this->registerNodeClass('DOMElement', 'DOMWrap\\Element');
34		$this->registerNodeClass('DOMComment', 'DOMWrap\\Comment');
35		$this->registerNodeClass('DOMDocument', 'DOMWrap\\Document');
36	138	$this->registerNodeClass('DOMDocumentType', 'DOMWrap\\DocumentType');
37	138	$this->registerNodeClass('DOMProcessingInstruction', 'DOMWrap\\ProcessingInstruction');
38		}
39
40		/**
41		* Set libxml options.
42		*
43	134	* Multiple values must use bitwise OR.
44	134	* eg: LIBXML_HTML_NOIMPLIED \| LIBXML_HTML_NODEFDTD
45		*
46		* @link http://php.net/manual/en/libxml.constants.php
47		*
48		* @param int $libxmlOptions
49		*/
50	1	public function setLibxmlOptions(int $libxmlOptions): void {
51	1	$this->libxmlOptions = $libxmlOptions;
52		}
53
54		/**
55	1	* {@inheritdoc}
56		*/
57		public function document(): ?\DOMDocument {
58		return $this;
59		}
60
61	1	/**
62	1	* {@inheritdoc}
63		*/
64		public function collection(): NodeList {
65		return $this->newNodeList([$this]);
66		}
67
68	1	/**
69	1	* {@inheritdoc}
70		*/
71		public function result(NodeList $nodeList) {
72		if ($nodeList->count()) {
73		return $nodeList->first();
74		}
75
76		return null;
77		}
78
79		/**
80		* {@inheritdoc}
81		*/
82		public function parent() {
83		return null;
84	1	}
85	1
86		/**
87		* {@inheritdoc}
88		*/
89		public function parents() {
90		return $this->newNodeList();
91	2	}
92	2
93		/**
94		* {@inheritdoc}
95		*/
96		public function replaceWith($newNode): self {
97		$this->replaceChild($newNode, $this);
98	140
99	140	return $this;
100	1	}
101
102		/**
103	140	* {@inheritdoc}
104	140	*/
105		public function _clone() {
106	140	return null;
107	140	}
108
109	140	/**
110	1	* {@inheritdoc}
111	1	*/
112	1	public function getHtml(): string {
113		return $this->getOuterHtml();
114		}
115
116	140	/**
117		* {@inheritdoc}
118		*/
119	140	public function setHtml($html): self {
120		if (!is_string($html) \|\| trim($html) == '') {
121	140	return $this;
122	140	}
123
124	140	$internalErrors = libxml_use_internal_errors(true);
125		$disableEntities = libxml_disable_entity_loader(true);
126
127		$this->detectEncoding($html);
128
129		$html = $this->convertToUtf8($html);
130
131		$this->loadHTML($html, $this->libxmlOptions);
132
133		// Remove <?xml ...> processing instruction.
134		$this->contents()->each(function($node) {
135		if ($node instanceof ProcessingInstruction && $node->nodeName == 'xml') {
136		$node->remove();
137		}
138		});
139
140		libxml_use_internal_errors($internalErrors);
141		libxml_disable_entity_loader($disableEntities);
142
143		return $this;
144		}
145
146		/**
147		* @param string $html
148		* @param int $options
149		*
150		* @return bool
151		*/
152		public function loadHTML($html, $options = 0): bool {
153		// Fix LibXML's crazy-ness RE root nodes
154		// While importing HTML using the LIBXML_HTML_NOIMPLIED option LibXML insists
155		// on having one root node. All subsequent nodes are appended to this first node.
156		// To counter this we will create a fake element, allow LibXML to 'do its thing'
157		// then undo it by taking the contents of the fake element, placing it back into
158		// the root and then remove our fake element.
159		if ($options & LIBXML_HTML_NOIMPLIED) {
160		$html = '<domwrap></domwrap>' . $html;
161		}
162
163		$html = '<?xml encoding="' . ($this->getEncoding() ?? 'UTF-8') . '">' . $html;
164
165		$result = parent::loadHTML($html, $options);
166
167		// Do our re-shuffling of nodes.
168		if ($this->libxmlOptions & LIBXML_HTML_NOIMPLIED) {
169		$this->children()->first()->contents()->each(function($node){
170		$this->append($node);
171		});
172
173		$this->removeChild($this->children()->first());
174		}
175
176		return $result;
177		}
178
179		/*
		0 ignored issues – show Unused Code Comprehensibility introduced 2019-12-08 10:12 UTC by Report Bug Copy Issue Report `40%` of this comment could be valid code. Did you maybe forget this after debugging? Sometimes obsolete code just ends up commented out instead of removed. In this case it is better to remove the code once you have checked you do not need it. The code might also have been commented out for debugging purposes. In this case it is vital that someone uncomments it again or your project may behave in very unexpected ways in production. This check looks for comments that seem to be mostly valid code and reports them. Loading history...
180		* @param $encoding string\|null
181		*/
182		public function setEncoding(string $encoding = null) {
183		$this->documentEncoding = $encoding;
184		}
185
186		/*
		0 ignored issues – show Unused Code Comprehensibility introduced 2019-12-08 10:12 UTC by Report Bug Copy Issue Report `50%` of this comment could be valid code. Did you maybe forget this after debugging? Sometimes obsolete code just ends up commented out instead of removed. In this case it is better to remove the code once you have checked you do not need it. The code might also have been commented out for debugging purposes. In this case it is vital that someone uncomments it again or your project may behave in very unexpected ways in production. This check looks for comments that seem to be mostly valid code and reports them. Loading history...
187		* @return string\|null
188		*/
189		public function getEncoding(): ?string {
190		return $this->documentEncoding;
191		}
192
193		/*
		0 ignored issues – show Unused Code Comprehensibility introduced 2019-12-08 10:12 UTC by Report Bug Copy Issue Report `36%` of this comment could be valid code. Did you maybe forget this after debugging? Sometimes obsolete code just ends up commented out instead of removed. In this case it is better to remove the code once you have checked you do not need it. The code might also have been commented out for debugging purposes. In this case it is vital that someone uncomments it again or your project may behave in very unexpected ways in production. This check looks for comments that seem to be mostly valid code and reports them. Loading history...
194		* @param $html string
195		*
196		* @return string\|null
197		*/
198		private function getCharset(string $html): ?string {
199		$charset = null;
200
201		if (preg_match('@<meta.*?charset=["\']?([^"\'\s>]+)@im', $html, $matches)) {
202		$charset = mb_strtoupper($matches[1]);
203		}
204
205		return $charset;
206		}
207
208		/*
209		* @param $html string
210		*/
211		private function detectEncoding(string $html) {
212		$charset = $this->getEncoding();
213
214		if (is_null($charset)) {
215		$charset = $this->getCharset($html);
216		}
217
218		$detectedCharset = mb_detect_encoding($html, mb_detect_order(), true);
219
220		if ($charset === null && $detectedCharset == 'UTF-8') {
221		$charset = $detectedCharset;
222		}
223
224		$this->setEncoding($charset);
225		}
226
227		/*
228		* @param $html string
229		*
230		* @return string
231		*/
232		private function convertToUtf8(string $html): string {
233		$charset = $this->getEncoding();
234
235		if ($charset !== null) {
236		$html = preg_replace('@(charset=["]?)([^"\s]+)([^"]*["]?)@im', '$1UTF-8$3', $html);
237		$mbHasCharset = in_array($charset, array_map('mb_strtoupper', mb_list_encodings()));
238
239		if ($mbHasCharset) {
240		$html = mb_convert_encoding($html, 'UTF-8', $charset);
241
242		// Fallback to iconv if available.
243		} elseif (extension_loaded('iconv')) {
244		$htmlIconv = iconv($charset, 'UTF-8', $html);
245
246		if ($htmlIconv !== false) {
247		$html = $htmlIconv;
248		} else {
249		$charset = null;
250		}
251		}
252		}
253
254		if ($charset === null) {
255		$html = mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8');
256		}
257
258		return $html;
259		}
260		}
261

scotteh / php-dom-wrapper

Push — master ( c33c36...393be6 )

Document::detectEncoding() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like