Link::parseUrl() - Code Metrics - Inspection of "Próba naprawienia buga z nawiasami w linkach" - adam-boduch/coyote - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#529)

by Daniel

created 2020-04-25 16:35 UTC

Link::parseUrl() A

↳ Parent: Link

Complexity

Conditions	1
Paths	1

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	1
nc	1
nop	1
dl	0
loc	4
rs	10
c	0
b	0
f	0

<?php

namespace Coyote\Services\Parser\Parsers;

use Collective\Html\HtmlBuilder;
use Coyote\Repositories\Contracts\PageRepositoryInterface as PageRepository;

class Link extends Parser implements ParserInterface
{
    const LINK_TAG_REGEXP = "<a\s[^>]*href=(\"??)([^\" >]*?)\\1[^>]*>(.*)<\/a>";
    const LINK_INTERNAL_REGEXP = '\[\[(.*?)(\|(.*?))*\]\]';

    const REGEXP_EMAIL = '#(^|[\n \[\]\:<>&;]|\()([a-z0-9&\-_.]+?@[\w\-]+\.(?:[\w\-\.]+\.)?[\w]+)#i';

    /**
     * @var PageRepository
     */
    private $page;

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

    /**
     * @var HtmlBuilder|null
     */
    private $html;

    /** @var UrlFormatter */
    private $urlParser;

    /**
     * Link constructor.
     *
     * @param PageRepository $page
     * @param string $host
     * @param HtmlBuilder|null $html
     */
    public function __construct(PageRepository $page, string $host, HtmlBuilder $html = null)
    {
        $this->page = $page;
        $this->host = $host;
        $this->html = $html;
        $this->urlParser = new UrlFormatter($host, app('html'));
    }

    /**
     * @param string $text
     * @return string
     */
    public function parse($text)
    {
        // first, make <a> from plain URL's
        // -----------------------------------------
        $text = $this->hashBlock($text, ['code', 'a']);
        $text = $this->hashInline($text, 'img');

        $text = $this->parseUrl($text);
        $text = $this->parseEmail($text);

        $text = $this->unhash($text);
        // ------------------------------------------

        $text = $this->hashBlock($text, 'code');
        $text = $this->hashInline($text, 'img');

        // then, parse internal links and youtube video links
        // --------------------------------------------------
        $text = $this->parseLinks($text);

        $text = $this->hashBlock($text, 'a');

        // at last, parse coyote markup
        $text = $this->parseInternalAccessors($text);

        $text = $this->unhash($text);

        return $text;
    }

    /**
     * @param string $text
     * @return string
     */
    protected function parseLinks($text)
    {
        if (!preg_match_all('/' . self::LINK_TAG_REGEXP . '/siU', $text, $matches, PREG_SET_ORDER)) {

            return $text;
        }

        for ($i = 0, $count = count($matches); $i < $count; $i++) {
            $link = $matches[$i][2];
            $title = $matches[$i][3];
            $match = $matches[$i][0];

            $text = $this->parseInternalLink($text, $match, $link, $title);
            $text = $this->parseYoutubeLinks($text, $match, $link, $title);
        }

        return $text;
    }

    /**
     * @param string $text
     * @param string $match
     * @param string $url
     * @param string $title
     * @return string
     */
    protected function parseInternalLink($text, $match, $url, $title)
    {
        if (urldecode($title) === urldecode($url) && ($path = $this->getPathFromInternalUrl($url)) !== false) {
            $page = $this->page->findByPath($path);

            if ($page) {
                $text = str_replace($match, link_to($url, $page->title), $text);
<?php

/**
 * @property int $x
 * @property int $y
 * @property string $text
 */
class MyLabel
{
    private $properties;

    private $allowedProperties = array('x', 'y', 'text');

    public function __get($name)
    {
        if (isset($properties[$name]) && in_array($name, $this->allowedProperties)) {
            return $properties[$name];
        } else {
            return null;
        }
    }

    public function __set($name, $value)
    {
        if (in_array($name, $this->allowedProperties)) {
            $properties[$name] = $value;
        } else {
            throw new \LogicException("Property $name is not defined.");
        }
    }

}
            }
        }

        return $text;
    }

    /**
     * @param string $text
     * @param string $match
     * @param string $url
     * @param string $title
     * @return string
     */
    protected function parseYoutubeLinks($text, $match, $url, $title)
    {
        if ($this->html === null) {
            return $text;
        }

        if (urldecode($title) !== urldecode($url)) {
            return $text;
        }

        $components = parse_url($url);

        if ($this->isUrl($components)) {
            // get host without "www"
            $host = $this->getHost($components['host']);
            $path = trim($components['path'], '/');

            if ($host === 'youtube.com' && $path === 'watch') {
                parse_str($components['query'], $query);

                if (!empty($query['v'])) {
                    parse_str($components['fragment'] ?? '', $fragments);

                    $text = str_replace(
                        $match,
                        $this->makeIframe($query['v'], $this->timeToSeconds($fragments['t'] ?? null)),
                        $text
                    );
                }
            }

            if ($host === 'youtu.be' && $path !== '') {
                parse_str($components['query'] ?? '', $query);

                $text = str_replace($match, $this->makeIframe($path, $this->timeToSeconds($query['t'] ?? null)), $text);
            }
        }

        return $text;
    }

    /**
     * Parse "old" coyote links like [[Foo/Bar]] to http://4programmers.net/Foo/Bar
     *
     * @param string $text
     * @return string
     */
    protected function parseInternalAccessors($text)
    {
        if (!preg_match_all('/' . self::LINK_INTERNAL_REGEXP . '/i', $text, $matches, PREG_SET_ORDER)) {

            return $text;
        }

        for ($i = 0, $count = count($matches); $i < $count; $i++) {
            $origin = $matches[$i][0];

            $path = '/' . str_replace(' ', '_', trim($matches[$i][1], '/?&'));

            $title = $matches[$i][3] ?? null;
            $hash = $this->getHashFromPath($path);

            $page = $this->page->findByPath($path);
            $attr = [];

            if (empty($page)) {
                $attr = ['class' => 'link-broken', 'title' => 'Dokument nie istnieje'];
                $path = 'Create' . $path;

                if (empty($title)) {
                    $title = str_replace('_', ' ', last(explode('/', $path)));
                }
            } else {
                $path = $page->path;
                $title = $title ?: $page->title;
<?php

/**
 * @property int $x
 * @property int $y
 * @property string $text
 */
class MyLabel
{
    private $properties;

    private $allowedProperties = array('x', 'y', 'text');

    public function __get($name)
    {
        if (isset($properties[$name]) && in_array($name, $this->allowedProperties)) {
            return $properties[$name];
        } else {
            return null;
        }
    }

    public function __set($name, $value)
    {
        if (in_array($name, $this->allowedProperties)) {
            $properties[$name] = $value;
        } else {
            throw new \LogicException("Property $name is not defined.");
        }
    }

}
            }

            $text = str_replace($origin, link_to($path . ($hash ? '#' . $hash : ''), $title, $attr), $text);
        }

        return $text;
    }

    /**
     * @param string $text
     * @return string
     */
    protected function parseUrl(string $text): string
    {
        return $this->urlParser->parse($text);
    }

    /**
     * @param string $text
     * @return string
     */
    protected function parseEmail(string $text): string
    {
        return preg_replace(self::REGEXP_EMAIL, "\$1<a href=\"mailto:\$2\">$2</a>", $text);
    }

    /**
     * @param string|null $time
     * @return null|string
     */
    private function timeToSeconds($time)
    {
        if (!$time) {
            return null;
        }

        if (preg_match('/(\d+)m(\d+)s/', $time, $match)) {
            return ($match[1] * 60) + $match[2];
        }

        return $time;
    }

    /**
     * @param string $videoId
     * @param string $start
     * @return string
     */
    private function makeIframe(string $videoId, string $start = null): string
    {
        $iframe = (string) $this->html->tag('iframe', '', [
            'src'             => 'https://youtube.com/embed/' . $videoId . ($start !== null ? "?start=$start" : ''),
            'class'           => 'embed-responsive-item',
            'allowfullscreen' => 'allowfullscreen'
        ]);

        return (string) $this->html->tag('div', $iframe, ['class' => 'embed-responsive embed-responsive-16by9']);
    }

    /**
     * Get path from url only if it's internal link (false if it's NOT internal link)
     *
     * @example http://4programmers.net/Foo/Bar => /Foo/Bar
     * @param string $url
     * @return string|false
     */
    private function getPathFromInternalUrl($url)
    {
        $components = parse_url($url);
        $path = false;

        if ($this->isUrl($components)) {
            // sprawdzamy, czy mamy do czynienia z linkiem wewnetrznym
            if ($this->host === $this->getHost($components['host'])) {
                $path = urldecode($components['path']);
            }
        }

        return $path;
    }

    /**
     * @param array|false $components
     * @return bool
     */
    private function isUrl($components)
    {
        if (!is_array($components)) {
            return false;
        }

        return (!empty($components['path']) && !empty($components['host']));
    }

    /**
     * @param string $path
     * @return string
     */
    private function getHashFromPath(&$path)
    {
        $hash = '';

        if (($pos = strpos($path, '#')) !== false) {
            $hash = htmlspecialchars(substr($path, $pos + 1));
            $path = substr($path, 0, $pos);
        }

        return $hash;
    }

    /**
     * Get host without "www" at the beginning.
     *
     * @param string $host
     * @return string
     */
    private function getHost(string $host): string
    {
        $parts = explode('.', $host);

        if ($parts[0] === 'www') {
            array_shift($parts);
        }

        return implode('.', $parts);
    }
}


1			<?php
2
3			namespace Coyote\Services\Parser\Parsers;
4
5			use Collective\Html\HtmlBuilder;
6			use Coyote\Repositories\Contracts\PageRepositoryInterface as PageRepository;
7
8			class Link extends Parser implements ParserInterface
9			{
10			const LINK_TAG_REGEXP = "<a\s[^>]href=(\"??)([^\" >]?)\\1[^>]>(.)<\/a>";
11			const LINK_INTERNAL_REGEXP = '\[\[(.?)(\\|(.?))*\]\]';
12
13			const REGEXP_EMAIL = '#(^\|[\n \[\]\:<>&;]\|\()([a-z0-9&\-_.]+?@[\w\-]+\.(?:[\w\-\.]+\.)?[\w]+)#i';
14
15			/**
16			* @var PageRepository
17			*/
18			private $page;
19
20			/**
21			* @var string
22			*/
23			private $host;
24
25			/**
26			* @var HtmlBuilder\|null
27			*/
28			private $html;
29
30			/** @var UrlFormatter */
31			private $urlParser;
32
33			/**
34			* Link constructor.
35			*
36			* @param PageRepository $page
37			* @param string $host
38			* @param HtmlBuilder\|null $html
39			*/
40			public function __construct(PageRepository $page, string $host, HtmlBuilder $html = null)
41			{
42			$this->page = $page;
43			$this->host = $host;
44			$this->html = $html;
45			$this->urlParser = new UrlFormatter($host, app('html'));
46			}
47
48			/**
49			* @param string $text
50			* @return string
51			*/
52			public function parse($text)
53			{
54			// first, make <a> from plain URL's
55			// -----------------------------------------
56			$text = $this->hashBlock($text, ['code', 'a']);
57			$text = $this->hashInline($text, 'img');
58
59			$text = $this->parseUrl($text);
60			$text = $this->parseEmail($text);
61
62			$text = $this->unhash($text);
63			// ------------------------------------------
64
65			$text = $this->hashBlock($text, 'code');
66			$text = $this->hashInline($text, 'img');
67
68			// then, parse internal links and youtube video links
69			// --------------------------------------------------
70			$text = $this->parseLinks($text);
71
72			$text = $this->hashBlock($text, 'a');
73
74			// at last, parse coyote markup
75			$text = $this->parseInternalAccessors($text);
76
77			$text = $this->unhash($text);
78
79			return $text;
80			}
81
82			/**
83			* @param string $text
84			* @return string
85			*/
86			protected function parseLinks($text)
87			{
88			if (!preg_match_all('/' . self::LINK_TAG_REGEXP . '/siU', $text, $matches, PREG_SET_ORDER)) {
			0 ignored issues – show Duplication introduced 2016-05-18 13:49 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...
89			return $text;
90			}
91
92			for ($i = 0, $count = count($matches); $i < $count; $i++) {
93			$link = $matches[$i][2];
94			$title = $matches[$i][3];
95			$match = $matches[$i][0];
96
97			$text = $this->parseInternalLink($text, $match, $link, $title);
98			$text = $this->parseYoutubeLinks($text, $match, $link, $title);
99			}
100
101			return $text;
102			}
103
104			/**
105			* @param string $text
106			* @param string $match
107			* @param string $url
108			* @param string $title
109			* @return string
110			*/
111			protected function parseInternalLink($text, $match, $url, $title)
112			{
113			if (urldecode($title) === urldecode($url) && ($path = $this->getPathFromInternalUrl($url)) !== false) {
114			$page = $this->page->findByPath($path);
115
116			if ($page) {
117			$text = str_replace($match, link_to($url, $page->title), $text);
			0 ignored issues – show Documentation introduced 2019-01-22 06:27 UTC by Report Bug Copy Issue Report The property `title` does not exist on `object<Coyote\Page>`. Since you implemented `__get`, maybe consider adding a @property annotation. Since your code implements the magic getter `_get`, this function will be called for any read access on an undefined variable. You can add the `@property` annotation to your class or interface to document the existence of this variable. <?php /** * @property int $x * @property int $y * @property string $text */ class MyLabel { private $properties; private $allowedProperties = array('x', 'y', 'text'); public function __get($name) { if (isset($properties[$name]) && in_array($name, $this->allowedProperties)) { return $properties[$name]; } else { return null; } } public function __set($name, $value) { if (in_array($name, $this->allowedProperties)) { $properties[$name] = $value; } else { throw new \LogicException("Property $name is not defined."); } } } If the property has read access only, you can use the @property-read annotation instead. Of course, you may also just have mistyped another name, in which case you should fix the error. See also the PhpDoc documentation for @property. Loading history...
118			}
119			}
120
121			return $text;
122			}
123
124			/**
125			* @param string $text
126			* @param string $match
127			* @param string $url
128			* @param string $title
129			* @return string
130			*/
131			protected function parseYoutubeLinks($text, $match, $url, $title)
132			{
133			if ($this->html === null) {
134			return $text;
135			}
136
137			if (urldecode($title) !== urldecode($url)) {
138			return $text;
139			}
140
141			$components = parse_url($url);
142
143			if ($this->isUrl($components)) {
144			// get host without "www"
145			$host = $this->getHost($components['host']);
146			$path = trim($components['path'], '/');
147
148			if ($host === 'youtube.com' && $path === 'watch') {
149			parse_str($components['query'], $query);
150
151			if (!empty($query['v'])) {
152			parse_str($components['fragment'] ?? '', $fragments);
153
154			$text = str_replace(
155			$match,
156			$this->makeIframe($query['v'], $this->timeToSeconds($fragments['t'] ?? null)),
157			$text
158			);
159			}
160			}
161
162			if ($host === 'youtu.be' && $path !== '') {
163			parse_str($components['query'] ?? '', $query);
164
165			$text = str_replace($match, $this->makeIframe($path, $this->timeToSeconds($query['t'] ?? null)), $text);
166			}
167			}
168
169			return $text;
170			}
171
172			/**
173			* Parse "old" coyote links like [[Foo/Bar]] to http://4programmers.net/Foo/Bar
174			*
175			* @param string $text
176			* @return string
177			*/
178			protected function parseInternalAccessors($text)
179			{
180			if (!preg_match_all('/' . self::LINK_INTERNAL_REGEXP . '/i', $text, $matches, PREG_SET_ORDER)) {
			0 ignored issues – show Duplication introduced 2016-05-18 13:49 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...
181			return $text;
182			}
183
184			for ($i = 0, $count = count($matches); $i < $count; $i++) {
185			$origin = $matches[$i][0];
186
187			$path = '/' . str_replace(' ', '_', trim($matches[$i][1], '/?&'));
188
189			$title = $matches[$i][3] ?? null;
190			$hash = $this->getHashFromPath($path);
191
192			$page = $this->page->findByPath($path);
193			$attr = [];
194
195			if (empty($page)) {
196			$attr = ['class' => 'link-broken', 'title' => 'Dokument nie istnieje'];
197			$path = 'Create' . $path;
198
199			if (empty($title)) {
200			$title = str_replace('_', ' ', last(explode('/', $path)));
201			}
202			} else {
203			$path = $page->path;
204			$title = $title ?: $page->title;
			0 ignored issues – show Documentation introduced 2019-01-22 06:27 UTC by Report Bug Copy Issue Report The property `title` does not exist on `object<Coyote\Page>`. Since you implemented `__get`, maybe consider adding a @property annotation. Since your code implements the magic getter `_get`, this function will be called for any read access on an undefined variable. You can add the `@property` annotation to your class or interface to document the existence of this variable. <?php /** * @property int $x * @property int $y * @property string $text */ class MyLabel { private $properties; private $allowedProperties = array('x', 'y', 'text'); public function __get($name) { if (isset($properties[$name]) && in_array($name, $this->allowedProperties)) { return $properties[$name]; } else { return null; } } public function __set($name, $value) { if (in_array($name, $this->allowedProperties)) { $properties[$name] = $value; } else { throw new \LogicException("Property $name is not defined."); } } } If the property has read access only, you can use the @property-read annotation instead. Of course, you may also just have mistyped another name, in which case you should fix the error. See also the PhpDoc documentation for @property. Loading history...
205			}
206
207			$text = str_replace($origin, link_to($path . ($hash ? '#' . $hash : ''), $title, $attr), $text);
208			}
209
210			return $text;
211			}
212
213			/**
214			* @param string $text
215			* @return string
216			*/
217			protected function parseUrl(string $text): string
218			{
219			return $this->urlParser->parse($text);
220			}
221
222			/**
223			* @param string $text
224			* @return string
225			*/
226			protected function parseEmail(string $text): string
227			{
228			return preg_replace(self::REGEXP_EMAIL, "\$1<a href=\"mailto:\$2\">$2</a>", $text);
229			}
230
231			/**
232			* @param string\|null $time
233			* @return null\|string
234			*/
235			private function timeToSeconds($time)
236			{
237			if (!$time) {
238			return null;
239			}
240
241			if (preg_match('/(\d+)m(\d+)s/', $time, $match)) {
242			return ($match[1] * 60) + $match[2];
243			}
244
245			return $time;
246			}
247
248			/**
249			* @param string $videoId
250			* @param string $start
251			* @return string
252			*/
253			private function makeIframe(string $videoId, string $start = null): string
254			{
255			$iframe = (string) $this->html->tag('iframe', '', [
256			'src' => 'https://youtube.com/embed/' . $videoId . ($start !== null ? "?start=$start" : ''),
257			'class' => 'embed-responsive-item',
258			'allowfullscreen' => 'allowfullscreen'
259			]);
260
261			return (string) $this->html->tag('div', $iframe, ['class' => 'embed-responsive embed-responsive-16by9']);
262			}
263
264			/**
265			* Get path from url only if it's internal link (false if it's NOT internal link)
266			*
267			* @example http://4programmers.net/Foo/Bar => /Foo/Bar
268			* @param string $url
269			* @return string\|false
270			*/
271			private function getPathFromInternalUrl($url)
272			{
273			$components = parse_url($url);
274			$path = false;
275
276			if ($this->isUrl($components)) {
277			// sprawdzamy, czy mamy do czynienia z linkiem wewnetrznym
278			if ($this->host === $this->getHost($components['host'])) {
279			$path = urldecode($components['path']);
280			}
281			}
282
283			return $path;
284			}
285
286			/**
287			* @param array\|false $components
288			* @return bool
289			*/
290			private function isUrl($components)
291			{
292			if (!is_array($components)) {
293			return false;
294			}
295
296			return (!empty($components['path']) && !empty($components['host']));
297			}
298
299			/**
300			* @param string $path
301			* @return string
302			*/
303			private function getHashFromPath(&$path)
304			{
305			$hash = '';
306
307			if (($pos = strpos($path, '#')) !== false) {
308			$hash = htmlspecialchars(substr($path, $pos + 1));
309			$path = substr($path, 0, $pos);
310			}
311
312			return $hash;
313			}
314
315			/**
316			* Get host without "www" at the beginning.
317			*
318			* @param string $host
319			* @return string
320			*/
321			private function getHost(string $host): string
322			{
323			$parts = explode('.', $host);
324
325			if ($parts[0] === 'www') {
326			array_shift($parts);
327			}
328
329			return implode('.', $parts);
330			}
331			}
332

adam-boduch / coyote

Pull Request — master (#529)

Link::parseUrl() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like