#1224

Build # #1224

Build Type

push

Committed by

Commit Message

Merge pull request #494 from sunnydavis/master

Run Details

30 of 31 relevant lines covered (96.77%)

8.58 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

96.67

/src/CrawlerDetect.php

<?php

/*
 * This file is part of Crawler Detect - the web crawler detection library.
 *
 * (c) Mark Beech <m@rkbee.ch>
 *
 * 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|null
     */
    protected $userAgent;

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

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

    /**
     * Crawlers object.
     *
     * @var \Jaybizzle\CrawlerDetect\Fixtures\Crawlers
     */
    protected $crawlers;

    /**
     * Exclusions object.
     *
     * @var \Jaybizzle\CrawlerDetect\Fixtures\Exclusions
     */
    protected $exclusions;

    /**
     * Headers object.
     *
     * @var \Jaybizzle\CrawlerDetect\Fixtures\Headers
     */
    protected $uaHttpHeaders;

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

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

    /**
     * Class constructor.
     */
    public function __construct(array $headers = null, $userAgent = null)
    {
        $this->crawlers = new Crawlers();
        $this->exclusions = new Exclusions();
        $this->uaHttpHeaders = new Headers();

        $this->compiledRegex = $this->compileRegex($this->crawlers->getAll());
        $this->compiledExclusions = $this->compileRegex($this->exclusions->getAll());

        $this->setHttpHeaders($headers);
        $this->setUserAgent($userAgent);
    }

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

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

        // Clear existing headers.
        $this->httpHeaders = array();

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

    /**
     * Set the user agent.
     *
     * @param string|null $userAgent
     */
    public function setUserAgent($userAgent)
    {
        if (is_null($userAgent)) {
            foreach ($this->getUaHttpHeaders() as $altHeader) {
                if (isset($this->httpHeaders[$altHeader])) {
                    $userAgent .= $this->httpHeaders[$altHeader].' ';
                }
            }
        }

        return $this->userAgent = $userAgent;
    }

    /**
     * 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->userAgent ?: ''
        ));

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

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

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


    /**
     * @return string|null
     */
    public function getUserAgent()
    {
        return $this->userAgent;
    }
}

1	<?php
2
3	/*
4	* This file is part of Crawler Detect - the web crawler detection library.
5	*
6	* (c) Mark Beech <m@rkbee.ch>
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\|null
24	*/
25	protected $userAgent;
26
27	/**
28	* Headers that contain a user agent.
29	*
30	* @var array
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 \Jaybizzle\CrawlerDetect\Fixtures\Crawlers
45	*/
46	protected $crawlers;
47
48	/**
49	* Exclusions object.
50	*
51	* @var \Jaybizzle\CrawlerDetect\Fixtures\Exclusions
52	*/
53	protected $exclusions;
54
55	/**
56	* Headers object.
57	*
58	* @var \Jaybizzle\CrawlerDetect\Fixtures\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	public function __construct(array $headers = null, $userAgent = null)
80	{
81	$this->crawlers = new Crawlers();	10✔
82	$this->exclusions = new Exclusions();	10✔
83	$this->uaHttpHeaders = new Headers();	10✔
84
85	$this->compiledRegex = $this->compileRegex($this->crawlers->getAll());	10✔
86	$this->compiledExclusions = $this->compileRegex($this->exclusions->getAll());	10✔
87
88	$this->setHttpHeaders($headers);	10✔
89	$this->setUserAgent($userAgent);	10✔
90	}
91
92	/**
93	* Compile the regex patterns into one regex string.
94	*
95	* @param array
96	*
97	* @return string
98	*/
99	public function compileRegex($patterns)
100	{
101	return '('.implode('\|', $patterns).')';	10✔
102	}
103
104	/**
105	* Set HTTP headers.
106	*
107	* @param array\|null $httpHeaders
108	*/
109	public function setHttpHeaders($httpHeaders)
110	{
111	// Use global _SERVER if $httpHeaders aren't defined.
112	if (! is_array($httpHeaders) \|\| ! count($httpHeaders)) {	10✔
113	$httpHeaders = $_SERVER;	8✔
114	}
115
116	// Clear existing headers.
117	$this->httpHeaders = array();	10✔
118
119	// Only save HTTP headers. In PHP land, that means
120	// only _SERVER vars that start with HTTP_.
121	foreach ($httpHeaders as $key => $value) {	10✔
122	if (strpos($key, 'HTTP_') === 0) {	10✔
123	$this->httpHeaders[$key] = $value;	2✔
124	}
125	}
126	}
127
128	/**
129	* Return user agent headers.
130	*
131	* @return array
132	*/
133	public function getUaHttpHeaders()
134	{
135	return $this->uaHttpHeaders->getAll();	9✔
136	}
137
138	/**
139	* Set the user agent.
140	*
141	* @param string\|null $userAgent
142	*/
143	public function setUserAgent($userAgent)
144	{
145	if (is_null($userAgent)) {	10✔
146	foreach ($this->getUaHttpHeaders() as $altHeader) {	9✔
147	if (isset($this->httpHeaders[$altHeader])) {	9✔
148	$userAgent .= $this->httpHeaders[$altHeader].' ';	2✔
149	}
150	}
151	}
152
153	return $this->userAgent = $userAgent;	10✔
154	}
155
156	/**
157	* Check user agent string against the regex.
158	*
159	* @param string\|null $userAgent
160	*
161	* @return bool
162	*/
163	public function isCrawler($userAgent = null)
164	{
165	$agent = trim(preg_replace(	10✔
166	"/{$this->compiledExclusions}/i",	10✔
167	'',	10✔
168	$userAgent ?: $this->userAgent ?: ''	10✔
169	));	10✔
170
171	if ($agent === '') {	10✔
172	return false;	2✔
173	}
174
175	return (bool) preg_match("/{$this->compiledRegex}/i", $agent, $this->matches);	9✔
176	}
177
178	/**
179	* Return the matches.
180	*
181	* @return string\|null
182	*/
183	public function getMatches()
184	{
185	return isset($this->matches[0]) ? $this->matches[0] : null;	4✔
186	}
187
188
189	/**
190	* @return string\|null
191	*/
192	public function getUserAgent()
193	{
194	return $this->userAgent;	×
195	}
196	}

JayBizzle / Crawler-Detect / #1224

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous