CrawlerDetect::setCrawlers() - Code Metrics - Inspection of "FEATURE Injectable components to support custom de..." - JayBizzle/Crawler-Detect - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#384)

by Damian

created 2020-08-09 23:06 UTC

CrawlerDetect::setCrawlers() A

↳ Parent: CrawlerDetect

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	6
rs	10
c	0
b	0
f	0

<?php

/*
 * This file is part of Crawler Detect - the web crawler detection library.
 *
 * (c) Mark Beech <[email protected]>
 *
 * This source file is subject to the MIT license that is bundled
 * with this source code in the file LICENSE.
 */

namespace Jaybizzle\CrawlerDetect;

use Jaybizzle\CrawlerDetect\Fixtures\Crawlers;
use Jaybizzle\CrawlerDetect\Fixtures\Exclusions;
use Jaybizzle\CrawlerDetect\Fixtures\Headers;

class CrawlerDetect
{
    /**
     * The user agent.
     *
     * @var string
     */
    protected $userAgent = null;

    /**
     * Headers that contain a user agent.
     *
     * @var string[]
     */
    protected $httpHeaders = array();

    /**
     * Store regex matches.
     *
     * @var array
     */
    protected $matches = array();

    /**
     * Crawlers object.
     *
     * @var Crawlers
     */
    protected $crawlers;

    /**
     * Exclusions object.
     *
     * @var Exclusions
     */
    protected $exclusions;

    /**
     * Headers object.
     *
     * @var Headers
     */
    protected $uaHttpHeaders;

    /**
     * The compiled regex string.
     *
     * @var string
     */
    protected $compiledRegex;

    /**
     * The compiled exclusions regex string.
     *
     * @var string
     */
    protected $compiledExclusions;

    /**
     * Class constructor.
     *
     * @param string[] $headers   Map of HTTP header values
     * @param string   $userAgent Browser user agent to detect
     */
    public function __construct(array $headers = null, $userAgent = null)
    {
        $this->setCrawlers(new Crawlers());
        $this->setExclusions(new Exclusions());
        $this->setUaHttpHeaders(new Headers());
        $this->setHttpHeaders($headers);
        $this->setUserAgent($userAgent);
    }

    /**
     * Compile the regex patterns into one regex string.
     *
     * @param array
     *
     * @return string
     */
    public function compileRegex($patterns)
    {
        return '(' . implode('|', $patterns) . ')';
    }

    /**
     * Get HTTP Headers
     *
     * @return array
     */
    public function getHttpHeaders()
    {
        return $this->httpHeaders;
    }

    /**
     * Set HTTP headers.
     *
     * @param array|null $httpHeaders
     * @return $this
     */
    public function setHttpHeaders($httpHeaders)
    {
        // Use global _SERVER if $httpHeaders aren't defined.
        if (empty($httpHeaders)) {
            $httpHeaders = $_SERVER;
        }

        // Clear existing headers.
        $this->httpHeaders = [];

        // Only save HTTP headers. In PHP land, that means
        // only _SERVER vars that start with HTTP_.
        foreach ($httpHeaders as $key => $value) {
            if (strpos($key, 'HTTP_') === 0) {
                $this->httpHeaders[$key] = $value;
            }
        }

        return $this;
    }

    /**
     * Return user agent headers.
     *
     * @return string[]
     */
    public function getUaHttpHeaders()
    {
        return $this->uaHttpHeaders->getAll();
    }

    /**
     * Assign headers used to detect user agent
     *
     * @param Headers $headers
     * @return $this
     */
    public function setUaHttpHeaders(Headers $headers)
    {
        $this->uaHttpHeaders = $headers;
        return $this;
    }

    /**
     * Get, or detect, user agent string
     *
     * @return string
     */
    public function getUserAgent()
    {
        if ($this->userAgent) {
            return $this->userAgent;
        }

        // Detect from headers
        $userAgent = '';
        $headers = $this->getHttpHeaders();
        foreach ($this->getUaHttpHeaders() as $altHeader) {
            if (isset($headers[$altHeader])) {
                $userAgent .= $headers[$altHeader] . ' ';
            }
        }
        return $userAgent;
    }

    /**
     * Set the user agent.
     *
     * @param string $userAgent
     * @return $this
     */
    public function setUserAgent($userAgent)
    {
        $this->userAgent = $userAgent;
        return $this;
    }

    /**
     * Get crawlers list
     *
     * @return string[]
     */
    public function getCrawlers()
    {
        return $this->crawlers->getAll();
    }

    /**
     * @param Crawlers $crawlers
     * @return $this
     */
    public function setCrawlers(Crawlers $crawlers)
    {
        $this->crawlers = $crawlers;
        $this->compiledRegex = $this->compileRegex($this->getCrawlers());
        return $this;
    }

    /**
     * Get exclusions list
     *
     * @return string[]
     */
    public function getExclusions()
    {
        return $this->exclusions->getAll();
    }

    /**
     * @param Exclusions $exclusions
     * @return $this
     */
    public function setExclusions(Exclusions $exclusions)
    {
        $this->exclusions = $exclusions;
        $this->compiledExclusions = $this->compileRegex($this->getExclusions());
        return $this;
    }

    /**
     * Check user agent string against the regex.
     *
     * @param string|null $userAgent
     *
     * @return bool
     */
    public function isCrawler($userAgent = null)
    {
        $agent = trim(preg_replace(
            "/{$this->compiledExclusions}/i",
            '',
            $userAgent ?: $this->getUserAgent()
        ));

        if ($agent == '') {
            return false;
        }

        $result = preg_match("/{$this->compiledRegex}/i", $agent, $matches);

        if ($matches) {

            $this->matches = $matches;
        }

        return (bool)$result;
    }

    /**
     * Return the matches.
     *
     * @return string|null
     */
    public function getMatches()
    {
        return isset($this->matches[0]) ? $this->matches[0] : null;
    }
}


1			<?php
2
3			/*
4			* This file is part of Crawler Detect - the web crawler detection library.
5			*
6			* (c) Mark Beech <[email protected]>
7			*
8			* This source file is subject to the MIT license that is bundled
9			* with this source code in the file LICENSE.
10			*/
11
12			namespace Jaybizzle\CrawlerDetect;
13
14			use Jaybizzle\CrawlerDetect\Fixtures\Crawlers;
15			use Jaybizzle\CrawlerDetect\Fixtures\Exclusions;
16			use Jaybizzle\CrawlerDetect\Fixtures\Headers;
17
18			class CrawlerDetect
19			{
20			/**
21			* The user agent.
22			*
23			* @var string
24			*/
25			protected $userAgent = null;
26
27			/**
28			* Headers that contain a user agent.
29			*
30			* @var string[]
31			*/
32			protected $httpHeaders = array();
33
34			/**
35			* Store regex matches.
36			*
37			* @var array
38			*/
39			protected $matches = array();
40
41			/**
42			* Crawlers object.
43			*
44			* @var Crawlers
45			*/
46			protected $crawlers;
47
48			/**
49			* Exclusions object.
50			*
51			* @var Exclusions
52			*/
53			protected $exclusions;
54
55			/**
56			* Headers object.
57			*
58			* @var Headers
59			*/
60			protected $uaHttpHeaders;
61
62			/**
63			* The compiled regex string.
64			*
65			* @var string
66			*/
67			protected $compiledRegex;
68
69			/**
70			* The compiled exclusions regex string.
71			*
72			* @var string
73			*/
74			protected $compiledExclusions;
75
76			/**
77			* Class constructor.
78			*
79			* @param string[] $headers Map of HTTP header values
80			* @param string $userAgent Browser user agent to detect
81			*/
82			public function __construct(array $headers = null, $userAgent = null)
83			{
84			$this->setCrawlers(new Crawlers());
85			$this->setExclusions(new Exclusions());
86			$this->setUaHttpHeaders(new Headers());
87			$this->setHttpHeaders($headers);
88			$this->setUserAgent($userAgent);
89			}
90
91			/**
92			* Compile the regex patterns into one regex string.
93			*
94			* @param array
95			*
96			* @return string
97			*/
98			public function compileRegex($patterns)
99			{
100			return '(' . implode('\|', $patterns) . ')';
101			}
102
103			/**
104			* Get HTTP Headers
105			*
106			* @return array
107			*/
108			public function getHttpHeaders()
109			{
110			return $this->httpHeaders;
111			}
112
113			/**
114			* Set HTTP headers.
115			*
116			* @param array\|null $httpHeaders
117			* @return $this
118			*/
119			public function setHttpHeaders($httpHeaders)
120			{
121			// Use global _SERVER if $httpHeaders aren't defined.
122			if (empty($httpHeaders)) {
123			$httpHeaders = $_SERVER;
124			}
125
126			// Clear existing headers.
127			$this->httpHeaders = [];
128
129			// Only save HTTP headers. In PHP land, that means
130			// only _SERVER vars that start with HTTP_.
131			foreach ($httpHeaders as $key => $value) {
132			if (strpos($key, 'HTTP_') === 0) {
133			$this->httpHeaders[$key] = $value;
134			}
135			}
136
137			return $this;
138			}
139
140			/**
141			* Return user agent headers.
142			*
143			* @return string[]
144			*/
145			public function getUaHttpHeaders()
146			{
147			return $this->uaHttpHeaders->getAll();
148			}
149
150			/**
151			* Assign headers used to detect user agent
152			*
153			* @param Headers $headers
154			* @return $this
155			*/
156			public function setUaHttpHeaders(Headers $headers)
157			{
158			$this->uaHttpHeaders = $headers;
159			return $this;
160			}
161
162			/**
163			* Get, or detect, user agent string
164			*
165			* @return string
166			*/
167			public function getUserAgent()
168			{
169			if ($this->userAgent) {
170			return $this->userAgent;
171			}
172
173			// Detect from headers
174			$userAgent = '';
175			$headers = $this->getHttpHeaders();
176			foreach ($this->getUaHttpHeaders() as $altHeader) {
177			if (isset($headers[$altHeader])) {
178			$userAgent .= $headers[$altHeader] . ' ';
179			}
180			}
181			return $userAgent;
182			}
183
184			/**
185			* Set the user agent.
186			*
187			* @param string $userAgent
188			* @return $this
189			*/
190			public function setUserAgent($userAgent)
191			{
192			$this->userAgent = $userAgent;
193			return $this;
194			}
195
196			/**
197			* Get crawlers list
198			*
199			* @return string[]
200			*/
201			public function getCrawlers()
202			{
203			return $this->crawlers->getAll();
204			}
205
206			/**
207			* @param Crawlers $crawlers
208			* @return $this
209			*/
210			public function setCrawlers(Crawlers $crawlers)
211			{
212			$this->crawlers = $crawlers;
213			$this->compiledRegex = $this->compileRegex($this->getCrawlers());
214			return $this;
215			}
216
217			/**
218			* Get exclusions list
219			*
220			* @return string[]
221			*/
222			public function getExclusions()
223			{
224			return $this->exclusions->getAll();
225			}
226
227			/**
228			* @param Exclusions $exclusions
229			* @return $this
230			*/
231			public function setExclusions(Exclusions $exclusions)
232			{
233			$this->exclusions = $exclusions;
234			$this->compiledExclusions = $this->compileRegex($this->getExclusions());
235			return $this;
236			}
237
238			/**
239			* Check user agent string against the regex.
240			*
241			* @param string\|null $userAgent
242			*
243			* @return bool
244			*/
245			public function isCrawler($userAgent = null)
246			{
247			$agent = trim(preg_replace(
248			"/{$this->compiledExclusions}/i",
249			'',
250			$userAgent ?: $this->getUserAgent()
251			));
252
253			if ($agent == '') {
254			return false;
255			}
256
257			$result = preg_match("/{$this->compiledRegex}/i", $agent, $matches);
258
259			if ($matches) {
			0 ignored issues – show Bug Best Practice introduced 2015-11-25 13:32 UTC by Report Bug Copy Issue Report The expression `$matches` of type `string[]` is implicitly converted to a boolean; are you sure this is intended? If so, consider using `! empty($expr)` instead to make it clear that you intend to check for an array without elements. This check marks implicit conversions of arrays to boolean values in a comparison. While in PHP an empty array is considered to be equal (but not identical) to false, this is not always apparent. Consider making the comparison explicit by using `empty(..)` or `! empty(...)` instead. Loading history...
260			$this->matches = $matches;
261			}
262
263			return (bool)$result;
264			}
265
266			/**
267			* Return the matches.
268			*
269			* @return string\|null
270			*/
271			public function getMatches()
272			{
273			return isset($this->matches[0]) ? $this->matches[0] : null;
274			}
275			}
276

JayBizzle / Crawler-Detect

Pull Request — master (#384)

CrawlerDetect::setCrawlers() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like