|
1
|
|
|
<?php |
|
2
|
|
|
/** |
|
3
|
|
|
* vipnytt/RobotsTxtParser |
|
4
|
|
|
* |
|
5
|
|
|
* @link https://github.com/VIPnytt/RobotsTxtParser |
|
6
|
|
|
* @license https://github.com/VIPnytt/RobotsTxtParser/blob/master/LICENSE The MIT License (MIT) |
|
7
|
|
|
*/ |
|
8
|
|
|
|
|
9
|
|
|
namespace vipnytt\RobotsTxtParser\Parser; |
|
10
|
|
|
|
|
11
|
|
|
use vipnytt\RobotsTxtParser\RobotsTxtInterface; |
|
12
|
|
|
|
|
13
|
|
|
/** |
|
14
|
|
|
* Class HeaderParser |
|
15
|
|
|
* |
|
16
|
|
|
* @link https://tools.ietf.org/html/rfc7231 |
|
17
|
|
|
* @link https://tools.ietf.org/html/rfc7230 |
|
18
|
|
|
* @link https://tools.ietf.org/html/rfc2616 |
|
19
|
|
|
* |
|
20
|
|
|
* @package vipnytt\RobotsTxtParser\Parser |
|
21
|
|
|
*/ |
|
22
|
|
|
class HeaderParser implements RobotsTxtInterface |
|
23
|
|
|
{ |
|
24
|
|
|
/** |
|
25
|
|
|
* ANSI C's asctime() format |
|
26
|
|
|
*/ |
|
27
|
|
|
const DATE_ASCTIME = 'D M j h:i:s Y'; |
|
28
|
|
|
|
|
29
|
|
|
/** |
|
30
|
|
|
* HTTP date formats |
|
31
|
|
|
* @link https://tools.ietf.org/html/rfc7231#section-7.1.1 |
|
32
|
|
|
* @link https://tools.ietf.org/html/rfc2616#section-3.3 |
|
33
|
|
|
*/ |
|
34
|
|
|
const DATE_HTTP = [ |
|
35
|
|
|
DATE_RFC1123, |
|
36
|
|
|
DATE_RFC850, |
|
37
|
|
|
self::DATE_ASCTIME, |
|
38
|
|
|
]; |
|
39
|
|
|
|
|
40
|
|
|
/** |
|
41
|
|
|
* cURL resource |
|
42
|
|
|
* @var resource |
|
43
|
|
|
*/ |
|
44
|
|
|
protected $curlHandler; |
|
45
|
|
|
|
|
46
|
|
|
/** |
|
47
|
|
|
* Headers |
|
48
|
|
|
* @var string[] |
|
49
|
|
|
*/ |
|
50
|
|
|
private $headers; |
|
51
|
|
|
|
|
52
|
|
|
/** |
|
53
|
|
|
* HeaderParser constructor. |
|
54
|
|
|
* |
|
55
|
|
|
* @param resource $handler |
|
56
|
|
|
*/ |
|
57
|
|
|
public function __construct($handler) |
|
58
|
|
|
{ |
|
59
|
|
|
$this->curlHandler = $handler; |
|
60
|
|
|
} |
|
61
|
|
|
|
|
62
|
|
|
/** |
|
63
|
|
|
* cURL CURLOPT_HEADERFUNCTION callback |
|
64
|
|
|
* @link https://tools.ietf.org/html/rfc7230#section-3.2.4 |
|
65
|
|
|
* |
|
66
|
|
|
* This callback function must return the number of bytes actually taken care of. |
|
67
|
|
|
* If that amount differs from the amount passed in to your function, it'll signal an error to the library. |
|
68
|
|
|
* This will cause the transfer to get aborted and the libcurl function in progress will return CURLE_WRITE_ERROR. |
|
69
|
|
|
* @link https://curl.haxx.se/libcurl/c/CURLOPT_HEADERFUNCTION.html |
|
70
|
|
|
* |
|
71
|
|
|
* @param resource $handler - cURL resource |
|
72
|
|
|
* @param string $line - cURL header line string |
|
73
|
|
|
* @return int - the number of bytes written |
|
74
|
|
|
*/ |
|
75
|
|
|
public function curlCallback($handler, $line) |
|
76
|
|
|
{ |
|
77
|
|
|
$this->curlHandler = $handler; |
|
78
|
|
|
$split = explode(':', $line, 2); |
|
79
|
|
|
$this->headers[strtolower($split[0])] = trim(end($split)); |
|
80
|
|
|
return strlen($line); |
|
81
|
|
|
} |
|
82
|
|
|
|
|
83
|
|
|
/** |
|
84
|
|
|
* Content-Type encoding HTTP header |
|
85
|
|
|
* @link https://tools.ietf.org/html/rfc2616#section-14.17 |
|
86
|
|
|
* |
|
87
|
|
|
* @return string |
|
88
|
|
|
*/ |
|
89
|
|
|
public function getCharset() |
|
90
|
|
|
{ |
|
91
|
|
|
if (isset($this->headers['content-type']) && |
|
92
|
|
|
($value = $this->getInlineValue($this->headers['content-type'], 'charset', ';')) !== false |
|
93
|
|
|
) { |
|
94
|
|
|
return $value; |
|
95
|
|
|
} |
|
96
|
|
|
return self::ENCODING; |
|
97
|
|
|
} |
|
98
|
|
|
|
|
99
|
|
|
/** |
|
100
|
|
|
* Get inline header variable value |
|
101
|
|
|
* |
|
102
|
|
|
* @param string $header |
|
103
|
|
|
* @param string $part |
|
104
|
|
|
* @param string $delimiter |
|
105
|
|
|
* @return string|false |
|
106
|
|
|
*/ |
|
107
|
|
|
private function getInlineValue($header, $part, $delimiter = ";") |
|
108
|
|
|
{ |
|
109
|
|
|
foreach (array_map('trim', explode($delimiter, $header)) as $string) { |
|
110
|
|
|
if (stripos($string, $part . '=') === 0) { |
|
111
|
|
|
return trim(explode('=', $string, 2)[1]); |
|
112
|
|
|
} |
|
113
|
|
|
} |
|
114
|
|
|
return false; |
|
115
|
|
|
} |
|
116
|
|
|
|
|
117
|
|
|
/** |
|
118
|
|
|
* Cache-Control max-age HTTP header |
|
119
|
|
|
* @link https://tools.ietf.org/html/rfc7234#section-5.2.1.1 |
|
120
|
|
|
* @link https://tools.ietf.org/html/rfc7234#section-5.2.2.8 |
|
121
|
|
|
* @link https://tools.ietf.org/html/rfc2616#section-14.9.3 |
|
122
|
|
|
* |
|
123
|
|
|
* @return int |
|
124
|
|
|
*/ |
|
125
|
|
|
public function getMaxAge() |
|
126
|
|
|
{ |
|
127
|
|
|
if (isset($this->headers['cache-control']) && |
|
128
|
|
|
($value = $this->getInlineValue($this->headers['cache-control'], 'max-age', ',')) !== false |
|
129
|
|
|
) { |
|
130
|
|
|
return intval($value); |
|
131
|
|
|
} |
|
132
|
|
|
return 0; |
|
133
|
|
|
} |
|
134
|
|
|
|
|
135
|
|
|
/** |
|
136
|
|
|
* Cache-Control Retry-After HTTP header |
|
137
|
|
|
* @link https://tools.ietf.org/html/rfc2616#section-14.37 |
|
138
|
|
|
* |
|
139
|
|
|
* @param int $requestTime |
|
140
|
|
|
* @return int |
|
141
|
|
|
*/ |
|
142
|
|
|
public function getRetryAfter($requestTime) |
|
143
|
|
|
{ |
|
144
|
|
|
if (isset($this->headers['retry-after'])) { |
|
145
|
|
|
if (is_numeric($this->headers['retry-after'])) { |
|
146
|
|
|
return intval($this->headers['retry-after']); |
|
147
|
|
|
} elseif (($time = $this->parseHttpDate($this->headers['retry-after'])) !== false) { |
|
148
|
|
|
return max(0, $time - $requestTime); |
|
149
|
|
|
} |
|
150
|
|
|
} |
|
151
|
|
|
// If no valid Retry-after header is found, retry after 15 minutes |
|
152
|
|
|
return 900; |
|
153
|
|
|
} |
|
154
|
|
|
|
|
155
|
|
|
/** |
|
156
|
|
|
* Parse HTTP-date |
|
157
|
|
|
* @link https://tools.ietf.org/html/rfc7231#section-7.1.1 |
|
158
|
|
|
* @link https://tools.ietf.org/html/rfc2616#section-3.3 |
|
159
|
|
|
* |
|
160
|
|
|
* @param string $string |
|
161
|
|
|
* @return int|false |
|
162
|
|
|
*/ |
|
163
|
|
|
private function parseHttpDate($string) |
|
164
|
|
|
{ |
|
165
|
|
|
foreach (self::DATE_HTTP as $format) { |
|
166
|
|
|
if (($dateTime = date_create_from_format($format, $string, new \DateTimeZone('UTC'))) !== false) { |
|
167
|
|
|
return (int)date_format($dateTime, 'U'); |
|
168
|
|
|
} |
|
169
|
|
|
} |
|
170
|
|
|
return false; |
|
171
|
|
|
} |
|
172
|
|
|
} |
|
173
|
|
|
|
VIPnytt /
RobotsTxtParser
