These results are based on our legacy PHP analysis, consider migrating to our new PHP analysis engine instead. Learn more
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 null |
||
24 | */ |
||
25 | protected $userAgent = null; |
||
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 | * Class constructor. |
||
64 | */ |
||
65 | public function __construct(array $headers = null, $userAgent = null) |
||
66 | { |
||
67 | $this->crawlers = new Crawlers(); |
||
68 | $this->exclusions = new Exclusions(); |
||
69 | $this->uaHttpHeaders = new Headers(); |
||
70 | |||
71 | $this->setHttpHeaders($headers); |
||
72 | $this->setUserAgent($userAgent); |
||
73 | } |
||
74 | |||
75 | /** |
||
76 | * Set HTTP headers. |
||
77 | * |
||
78 | * @param array $httpHeaders |
||
79 | */ |
||
80 | public function setHttpHeaders($httpHeaders = null) |
||
0 ignored issues
–
show
|
|||
81 | { |
||
82 | // use global _SERVER if $httpHeaders aren't defined |
||
83 | if (!is_array($httpHeaders) || !count($httpHeaders)) { |
||
84 | $httpHeaders = $_SERVER; |
||
85 | } |
||
86 | // clear existing headers |
||
87 | $this->httpHeaders = array(); |
||
88 | // Only save HTTP headers. In PHP land, that means only _SERVER vars that |
||
89 | // start with HTTP_. |
||
90 | foreach ($httpHeaders as $key => $value) { |
||
91 | if (substr($key, 0, 5) === 'HTTP_') { |
||
92 | $this->httpHeaders[$key] = $value; |
||
93 | } |
||
94 | } |
||
95 | } |
||
96 | |||
97 | /** |
||
98 | * Return user agent headers. |
||
99 | * |
||
100 | * @return array |
||
101 | */ |
||
102 | public function getUaHttpHeaders() |
||
103 | { |
||
104 | return $this->uaHttpHeaders->getAll(); |
||
105 | } |
||
106 | |||
107 | /** |
||
108 | * Set the user agent. |
||
109 | * |
||
110 | * @param string $userAgent |
||
111 | */ |
||
112 | public function setUserAgent($userAgent = null) |
||
113 | { |
||
114 | if (false === empty($userAgent)) { |
||
115 | return $this->userAgent = $userAgent; |
||
116 | } else { |
||
117 | $this->userAgent = null; |
||
118 | foreach ($this->getUaHttpHeaders() as $altHeader) { |
||
119 | if (false === empty($this->httpHeaders[$altHeader])) { // @todo: should use getHttpHeader(), but it would be slow. |
||
120 | $this->userAgent .= $this->httpHeaders[$altHeader].' '; |
||
121 | } |
||
122 | } |
||
123 | |||
124 | return $this->userAgent = (!empty($this->userAgent) ? trim($this->userAgent) : null); |
||
125 | } |
||
126 | } |
||
127 | |||
128 | /** |
||
129 | * Build the user agent regex. |
||
130 | * |
||
131 | * @return string |
||
132 | */ |
||
133 | public function getRegex() |
||
134 | { |
||
135 | return '('.implode('|', $this->crawlers->getAll()).')'; |
||
136 | } |
||
137 | |||
138 | /** |
||
139 | * Build the replacement regex. |
||
140 | * |
||
141 | * @return string |
||
142 | */ |
||
143 | public function getExclusions() |
||
144 | { |
||
145 | return '('.implode('|', $this->exclusions->getAll()).')'; |
||
146 | } |
||
147 | |||
148 | /** |
||
149 | * Check user agent string against the regex. |
||
150 | * |
||
151 | * @param string $userAgent |
||
152 | * |
||
153 | * @return bool |
||
154 | */ |
||
155 | public function isCrawler($userAgent = null) |
||
156 | { |
||
157 | $agent = is_null($userAgent) ? $this->userAgent : $userAgent; |
||
158 | |||
159 | $agent = preg_replace('/'.$this->getExclusions().'/i', '', $agent); |
||
160 | |||
161 | if (strlen(trim($agent)) == 0) { |
||
162 | return false; |
||
163 | } else { |
||
164 | $result = preg_match('/'.$this->getRegex().'/i', trim($agent), $matches); |
||
165 | } |
||
166 | |||
167 | if ($matches) { |
||
168 | $this->matches = $matches; |
||
169 | } |
||
170 | |||
171 | return (bool) $result; |
||
172 | } |
||
173 | |||
174 | /** |
||
175 | * Return the matches. |
||
176 | * |
||
177 | * @return string |
||
178 | */ |
||
179 | public function getMatches() |
||
180 | { |
||
181 | return $this->matches[0]; |
||
182 | } |
||
183 | } |
||
184 |
Instead of super-globals, we recommend to explicitly inject the dependencies of your class. This makes your code less dependent on global state and it becomes generally more testable: