|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace simplehtmldom; |
|
4
|
|
|
|
|
5
|
|
|
/** |
|
6
|
|
|
* Website: http://sourceforge.net/projects/simplehtmldom/ |
|
7
|
|
|
* Acknowledge: Jose Solorzano (https://sourceforge.net/projects/php-html/). |
|
8
|
|
|
* |
|
9
|
|
|
* Licensed under The MIT License |
|
10
|
|
|
* See the LICENSE file in the project root for more information. |
|
11
|
|
|
* |
|
12
|
|
|
* Authors: |
|
13
|
|
|
* S.C. Chen |
|
14
|
|
|
* John Schlick |
|
15
|
|
|
* Rus Carroll |
|
16
|
|
|
* logmanoriginal |
|
17
|
|
|
* |
|
18
|
|
|
* Contributors: |
|
19
|
|
|
* Yousuke Kumakura |
|
20
|
|
|
* Vadim Voituk |
|
21
|
|
|
* Antcs |
|
22
|
|
|
* |
|
23
|
|
|
* Version Rev. 2.0-RC2 (415) |
|
24
|
|
|
*/ |
|
25
|
|
|
include_once __DIR__ . '/HtmlDocument.php'; |
|
26
|
|
|
|
|
27
|
|
|
class HtmlWeb |
|
28
|
|
|
{ |
|
29
|
|
|
/** |
|
30
|
|
|
* @return HtmlDocument Returns the DOM for a webpage |
|
31
|
|
|
*/ |
|
32
|
|
|
public function load($url) |
|
33
|
|
|
{ |
|
34
|
|
|
if (!filter_var($url, FILTER_VALIDATE_URL)) { |
|
35
|
|
|
return null; |
|
36
|
|
|
} |
|
37
|
|
|
|
|
38
|
|
|
if ($scheme = parse_url($url, PHP_URL_SCHEME)) { |
|
39
|
|
|
switch (strtolower($scheme)) { |
|
40
|
|
|
case 'http': |
|
41
|
|
|
case 'https': |
|
42
|
|
|
break; |
|
43
|
|
|
default: |
|
44
|
|
|
return null; |
|
45
|
|
|
} |
|
46
|
|
|
|
|
47
|
|
|
if (extension_loaded('curl')) { |
|
48
|
|
|
return $this->load_curl($url); |
|
49
|
|
|
} elseif (ini_get('allow_url_fopen')) { |
|
50
|
|
|
return $this->load_fopen($url); |
|
51
|
|
|
} else { |
|
52
|
|
|
error_log(__FUNCTION__ . ' requires either the cURL extension or allow_url_fopen=On in php.ini'); |
|
53
|
|
|
} |
|
54
|
|
|
} |
|
55
|
|
|
|
|
56
|
|
|
return null; |
|
57
|
|
|
} |
|
58
|
|
|
|
|
59
|
|
|
/** |
|
60
|
|
|
* cURL implementation of load. |
|
61
|
|
|
*/ |
|
62
|
|
|
private function load_curl($url) |
|
63
|
|
|
{ |
|
64
|
|
|
$ch = curl_init(); |
|
65
|
|
|
|
|
66
|
|
|
curl_setopt($ch, CURLOPT_URL, $url); |
|
67
|
|
|
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); |
|
68
|
|
|
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true); |
|
69
|
|
|
|
|
70
|
|
|
// There is no guarantee this request will be fulfilled |
|
71
|
|
|
// -- https://www.php.net/manual/en/function.curl-setopt.php |
|
72
|
|
|
curl_setopt($ch, CURLOPT_BUFFERSIZE, MAX_FILE_SIZE); |
|
73
|
|
|
|
|
74
|
|
|
// There is no guarantee this request will be fulfilled |
|
75
|
|
|
$header = [ |
|
76
|
|
|
'Accept: text/html', // Prefer HTML format |
|
77
|
|
|
'Accept-Charset: utf-8', // Prefer UTF-8 encoding |
|
78
|
|
|
]; |
|
79
|
|
|
curl_setopt($ch, CURLOPT_HTTPHEADER, $header); |
|
80
|
|
|
|
|
81
|
|
|
$doc = curl_exec($ch); |
|
82
|
|
|
|
|
83
|
|
|
if (200 !== curl_getinfo($ch, CURLINFO_RESPONSE_CODE)) { |
|
84
|
|
|
return null; |
|
85
|
|
|
} |
|
86
|
|
|
|
|
87
|
|
|
curl_close($ch); |
|
88
|
|
|
|
|
89
|
|
|
if (strlen($doc) > MAX_FILE_SIZE) { |
|
|
|
|
|
|
90
|
|
|
return null; |
|
91
|
|
|
} |
|
92
|
|
|
|
|
93
|
|
|
return new HtmlDocument($doc); |
|
94
|
|
|
} |
|
95
|
|
|
|
|
96
|
|
|
/** |
|
97
|
|
|
* fopen implementation of load. |
|
98
|
|
|
*/ |
|
99
|
|
|
private function load_fopen($url) |
|
100
|
|
|
{ |
|
101
|
|
|
// There is no guarantee this request will be fulfilled |
|
102
|
|
|
$context = stream_context_create(['http' => [ |
|
103
|
|
|
'header' => [ |
|
104
|
|
|
'Accept: text/html', // Prefer HTML format |
|
105
|
|
|
'Accept-Charset: utf-8', // Prefer UTF-8 encoding |
|
106
|
|
|
], |
|
107
|
|
|
'ignore_errors' => true, // Always fetch content |
|
108
|
|
|
]]); |
|
109
|
|
|
|
|
110
|
|
|
$doc = file_get_contents($url, false, $context, 0, MAX_FILE_SIZE + 1); |
|
111
|
|
|
|
|
112
|
|
|
if (isset($http_response_header)) { |
|
113
|
|
|
foreach ($http_response_header as $rh) { |
|
114
|
|
|
// https://stackoverflow.com/a/1442526 |
|
115
|
|
|
$parts = explode(' ', $rh, 3); |
|
116
|
|
|
|
|
117
|
|
|
if (preg_match('/HTTP\/\d\.\d/', $parts[0])) { |
|
118
|
|
|
$code = $parts[1]; |
|
119
|
|
|
} |
|
120
|
|
|
} // Last code is final status |
|
121
|
|
|
|
|
122
|
|
|
if (!isset($code) || '200' !== $code) { |
|
123
|
|
|
return null; |
|
124
|
|
|
} |
|
125
|
|
|
} |
|
126
|
|
|
|
|
127
|
|
|
if (strlen($doc) > MAX_FILE_SIZE) { |
|
128
|
|
|
return null; |
|
129
|
|
|
} |
|
130
|
|
|
|
|
131
|
|
|
return new HtmlDocument($doc); |
|
132
|
|
|
} |
|
133
|
|
|
} |
|
134
|
|
|
|
dimaslanjaka /
universal-framework
Loading history...