GlHtml::getLinksByTagAttribute() - Code Metrics - Inspection of "Merge pull request #2 from matyo91/composer_requir..." - emmanuelroecker/php-simply-html - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 192986...577fb2 )

by Emmanuel

created 2016-04-11 20:51 UTC

GlHtml::getLinksByTagAttribute() A

↳ Parent: GlHtml

Complexity

Conditions	4
Paths	3

Size

Total Lines	10
Code Lines	6

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
c	1
b	0
f	0
dl	0
loc	10
rs	9.2
cc	4
eloc	6
nc	3
nop	3

<?php
/**
 * Main Class to manipulate dom
 *
 * PHP version 5.4
 *
 * @category  GLICER
 * @package   GlHtml
 * @author    Emmanuel ROECKER
 * @author    Rym BOUCHAGOUR
 * @copyright 2015 GLICER
 * @license   MIT
 * @link      http://dev.glicer.com/
 *
 * Created : 19/02/15
 * File : GlHtml.php
 *
 */


namespace GlHtml;

use Symfony\Component\CssSelector\CssSelector;
use Symfony\Component\CssSelector\CssSelectorConverter;

/**
 * Class GlHtml
 * @package GlHtml
 */
class GlHtml
{
    /**
     * @var \DOMDocument
     */
    private $dom;

    /**
     * @var string
     */
    private $html;

    /**
     * @param string $html
     */
    public function __construct($html)
    {
        $html      = static::fixNewlines($html);
        $this->dom = new \DOMDocument();

        $libxml_previous_state = libxml_use_internal_errors(true); //disable warnings
        $this->dom->loadHTML(mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8'));
        libxml_clear_errors();
        libxml_use_internal_errors($libxml_previous_state);

        $this->html = $html;
    }

    /**
     * Unify newlines
     *
     * @param string $text
     *
     * @return string the fixed text
     */
    static function fixNewlines($text)

    {
        $text = str_replace("\r\n", "\n", $text);
        $text = str_replace("\r", "\n", $text);

        return $text;
    }

    /**
     * return one dom element with $selector css filter
     *
     * @param string $selector CSS 3 Selector
     *
     * @return GlHtmlNode[]
     */
    public function get($selector)
    {
        $xpath = new \DOMXPath($this->dom);

        if(class_exists('Symfony\Component\CssSelector\CssSelector')) {
            $expression = CssSelector::toXPath($selector);
        } else {
            $converter = new CssSelectorConverter();
            $expression = $converter->toXPath($selector);
        }
        $nodes = $xpath->query($expression);

        $glnodes = [];
        foreach ($nodes as $node) {
            $glnodes[] = new GlHtmlNode($node);
        }

        return $glnodes;
    }

    /**
     * set a list of attributes
     *
     * @param string $selector
     * @param array  $attributes
     * @param string $value
     */
    public function set($selector, array $attributes, $value = null)
    {
        $nodes = $this->get($selector);

        foreach ($nodes as $node) {
            $node->set($attributes, $value);

        }
    }

    /**
     * @param string $selector
     */
    public function delete($selector)
    {
        $nodes = $this->get($selector);
        foreach ($nodes as $node) {
            $node->delete();
        }
    }

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

    public function getText()
    {
        $body = $this->get("body")[0];

        return $body->getText();
    }

    /**
     * @param string $tagname
     * @param string $attribute
     * @param array  $links
     */
    private function getLinksByTagAttribute($tagname, $attribute, array &$links)
    {
        $tagslink = $this->get($tagname);
        foreach ($tagslink as $taglink) {
            $href = $taglink->getAttribute($attribute);
            if (isset($href) && (strlen(trim($href)) > 0)) {
                $links[$href] = $href;
            }
        }
    }

    /**
     * @param bool $all if true get url in text and params
     *
     * @return array
     */
    public function getLinks($all = false)
    {
        $links = [];

        $this->getLinksByTagAttribute("link", "href", $links);
        $this->getLinksByTagAttribute("a", "href",$links);
        $this->getLinksByTagAttribute("script", "src", $links);
        $this->getLinksByTagAttribute("iframe", "src", $links);
        $this->getLinksByTagAttribute("img", "src", $links);

        //get all string started with http
        $regexUrl = '/[">\s]+((http|https|ftp|ftps)\:\/\/(.*?))["<\s]+/';
        $urls     = null;
        if (preg_match_all($regexUrl, $this->html, $urls) > 0) {
            $matches = $urls[1];
            foreach ($matches as $url) {
                if (filter_var($url, FILTER_VALIDATE_URL)) {
                    $links[$url] = $url;
                }
            }
        }

        if ($all) {
            //get all params which can be a url
            $regexParam = '/["](.*?)["]/';
            $params     = [];
            if (preg_match_all($regexParam, $this->html, $params) > 0) {
                $urls = $params[1];
                foreach ($urls as $url) {
                    $url = trim($url);
                    if ((strpbrk($url, "/.") !== false) && (strpbrk($url, " ") === false)) {
                        $links[$url] = $url;
                    }
                }
            }
        }

        foreach ($links as $link) {
            $url = parse_url($link);
            if (!((isset($url['host']) && isset($url['scheme'])) || (isset($url['path'])))) {
                unset($links[$link]);
            }
        }

        return $links;
    }

    public function getSentences()
    {
        $sentences = [];

        $body = $this->get("body");
        if (count($body) > 0) {
            $sentences = $body[0]->getSentences();
        }

        $description = $this->get('meta[name="description"]');
        if (count($description) > 0) {

            $description = trim($description[0]->getAttribute("content"));
            if (strlen($description) > 0) {
                array_unshift($sentences, $description);
            }
        }

        $title = $this->get('title');
        if (count($title) > 0) {

            $title = trim($title[0]->getText());
            if (strlen($title) > 0) {
                array_unshift($sentences, $title);
            }
        }

        return $sentences;
    }

    /**
     * @return GlHtmlSummary[]
     */
    public function getSummary()
    {
        $body = $this->get("body")[0];

        $summary  = [];
        $callback = function (GlHtmlNode $childNode) use (&$summary) {
            $nodeName = $childNode->getName();

            if (preg_match('/^h(\d+)$/', $nodeName, $matches)) {
                $summary[] = new GlHtmlSummary($childNode, $matches[1]);
            }
        };

        $body->callChild($callback);

        return $summary;
    }
}


1		<?php
2		/**
3		* Main Class to manipulate dom
4		*
5		* PHP version 5.4
6		*
7		* @category GLICER
8		* @package GlHtml
9		* @author Emmanuel ROECKER
10		* @author Rym BOUCHAGOUR
11		* @copyright 2015 GLICER
12		* @license MIT
13		* @link http://dev.glicer.com/
14		*
15		* Created : 19/02/15
16		* File : GlHtml.php
17		*
18		*/
19
20
21		namespace GlHtml;
22
23		use Symfony\Component\CssSelector\CssSelector;
24		use Symfony\Component\CssSelector\CssSelectorConverter;
25
26		/**
27		* Class GlHtml
28		* @package GlHtml
29		*/
30		class GlHtml
31		{
32		/**
33		* @var \DOMDocument
34		*/
35		private $dom;
36
37		/**
38		* @var string
39		*/
40		private $html;
41
42		/**
43		* @param string $html
44		*/
45		public function __construct($html)
46		{
47		$html = static::fixNewlines($html);
48		$this->dom = new \DOMDocument();
49
50		$libxml_previous_state = libxml_use_internal_errors(true); //disable warnings
51		$this->dom->loadHTML(mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8'));
52		libxml_clear_errors();
53		libxml_use_internal_errors($libxml_previous_state);
54
55		$this->html = $html;
56		}
57
58		/**
59		* Unify newlines
60		*
61		* @param string $text
62		*
63		* @return string the fixed text
64		*/
65		static function fixNewlines($text)
		0 ignored issues – show Best Practice introduced 2016-01-03 11:22 UTC by Report Bug Copy Issue Report It is generally recommended to explicitly declare the visibility for methods. Adding explicit visibility (`private`, `protected`, or `public`) is generally recommend to communicate to other developers how, and from where this method is intended to be used. Loading history...
66		{
67		$text = str_replace("\r\n", "\n", $text);
68		$text = str_replace("\r", "\n", $text);
69
70		return $text;
71		}
72
73		/**
74		* return one dom element with $selector css filter
75		*
76		* @param string $selector CSS 3 Selector
77		*
78		* @return GlHtmlNode[]
79		*/
80		public function get($selector)
81		{
82		$xpath = new \DOMXPath($this->dom);
83
84		if(class_exists('Symfony\Component\CssSelector\CssSelector')) {
85		$expression = CssSelector::toXPath($selector);
86		} else {
87		$converter = new CssSelectorConverter();
88		$expression = $converter->toXPath($selector);
89		}
90		$nodes = $xpath->query($expression);
91
92		$glnodes = [];
93		foreach ($nodes as $node) {
94		$glnodes[] = new GlHtmlNode($node);
95		}
96
97		return $glnodes;
98		}
99
100		/**
101		* set a list of attributes
102		*
103		* @param string $selector
104		* @param array $attributes
105		* @param string $value
106		*/
107		public function set($selector, array $attributes, $value = null)
108		{
109		$nodes = $this->get($selector);
110
111		foreach ($nodes as $node) {
112		$node->set($attributes, $value);
		0 ignored issues – show Bug introduced 2016-01-03 11:22 UTC by Report Bug Copy Issue Report The method `set()` does not seem to exist on `object<GlHtml\GlHtmlNode>`. This check looks for calls to methods that do not seem to exist on a given type. It looks for the method on the type itself as well as in inherited classes or implemented interfaces. This is most likely a typographical error or the method has been renamed. Loading history...
113		}
114		}
115
116		/**
117		* @param string $selector
118		*/
119		public function delete($selector)
120		{
121		$nodes = $this->get($selector);
122		foreach ($nodes as $node) {
123		$node->delete();
124		}
125		}
126
127		/**
128		* @return string
129		*/
130		public function html()
131		{
132		return $this->dom->saveHTML();
133		}
134
135		public function getText()
136		{
137		$body = $this->get("body")[0];
138
139		return $body->getText();
140		}
141
142		/**
143		* @param string $tagname
144		* @param string $attribute
145		* @param array $links
146		*/
147		private function getLinksByTagAttribute($tagname, $attribute, array &$links)
148		{
149		$tagslink = $this->get($tagname);
150		foreach ($tagslink as $taglink) {
151		$href = $taglink->getAttribute($attribute);
152		if (isset($href) && (strlen(trim($href)) > 0)) {
153		$links[$href] = $href;
154		}
155		}
156		}
157
158		/**
159		* @param bool $all if true get url in text and params
160		*
161		* @return array
162		*/
163		public function getLinks($all = false)
164		{
165		$links = [];
166
167		$this->getLinksByTagAttribute("link", "href", $links);
168		$this->getLinksByTagAttribute("a", "href",$links);
169		$this->getLinksByTagAttribute("script", "src", $links);
170		$this->getLinksByTagAttribute("iframe", "src", $links);
171		$this->getLinksByTagAttribute("img", "src", $links);
172
173		//get all string started with http
174		$regexUrl = '/[">\s]+((http\|https\|ftp\|ftps)\:\/\/(.*?))["<\s]+/';
175		$urls = null;
176		if (preg_match_all($regexUrl, $this->html, $urls) > 0) {
177		$matches = $urls[1];
178		foreach ($matches as $url) {
179		if (filter_var($url, FILTER_VALIDATE_URL)) {
180		$links[$url] = $url;
181		}
182		}
183		}
184
185		if ($all) {
186		//get all params which can be a url
187		$regexParam = '/["](.*?)["]/';
188		$params = [];
189		if (preg_match_all($regexParam, $this->html, $params) > 0) {
190		$urls = $params[1];
191		foreach ($urls as $url) {
192		$url = trim($url);
193		if ((strpbrk($url, "/.") !== false) && (strpbrk($url, " ") === false)) {
194		$links[$url] = $url;
195		}
196		}
197		}
198		}
199
200		foreach ($links as $link) {
201		$url = parse_url($link);
202		if (!((isset($url['host']) && isset($url['scheme'])) \|\| (isset($url['path'])))) {
203		unset($links[$link]);
204		}
205		}
206
207		return $links;
208		}
209
210		public function getSentences()
211		{
212		$sentences = [];
213
214		$body = $this->get("body");
215		if (count($body) > 0) {
216		$sentences = $body[0]->getSentences();
217		}
218
219		$description = $this->get('meta[name="description"]');
220	View Code Duplication	if (count($description) > 0) {
		0 ignored issues – show Duplication introduced 2016-01-03 11:22 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...
221		$description = trim($description[0]->getAttribute("content"));
222		if (strlen($description) > 0) {
223		array_unshift($sentences, $description);
224		}
225		}
226
227		$title = $this->get('title');
228	View Code Duplication	if (count($title) > 0) {
		0 ignored issues – show Duplication introduced 2016-01-03 11:22 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...
229		$title = trim($title[0]->getText());
230		if (strlen($title) > 0) {
231		array_unshift($sentences, $title);
232		}
233		}
234
235		return $sentences;
236		}
237
238		/**
239		* @return GlHtmlSummary[]
240		*/
241		public function getSummary()
242		{
243		$body = $this->get("body")[0];
244
245		$summary = [];
246		$callback = function (GlHtmlNode $childNode) use (&$summary) {
247		$nodeName = $childNode->getName();
248
249		if (preg_match('/^h(\d+)$/', $nodeName, $matches)) {
250		$summary[] = new GlHtmlSummary($childNode, $matches[1]);
251		}
252		};
253
254		$body->callChild($callback);
255
256		return $summary;
257		}
258		}
259

emmanuelroecker / php-simply-html

GitHub Access Token became invalid

Push — master ( 192986...577fb2 )

GlHtml::getLinksByTagAttribute() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like