|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* Entity. |
|
5
|
|
|
*/ |
|
6
|
|
|
|
|
7
|
|
|
namespace PiedWeb\UrlHarvester; |
|
8
|
|
|
|
|
9
|
|
|
use Symfony\Component\DomCrawler\Crawler as DomCrawler; |
|
10
|
|
|
|
|
11
|
|
|
class Link |
|
12
|
|
|
{ |
|
13
|
12 |
|
private $url; |
|
14
|
|
|
private $anchor; |
|
15
|
12 |
|
private $element; |
|
16
|
12 |
|
|
|
17
|
12 |
|
/** |
|
18
|
|
|
* @param int |
|
19
|
12 |
|
*/ |
|
20
|
12 |
|
private $type; |
|
21
|
|
|
|
|
22
|
12 |
|
const LINK_A = 1; |
|
23
|
|
|
const LINK_SRC = 4; |
|
24
|
12 |
|
const LINK_3XX = 2; |
|
25
|
|
|
const LINK_301 = 3; |
|
26
|
12 |
|
|
|
27
|
|
|
public function __construct(string $url, \DOMElement $element = null, int $type = null) |
|
28
|
|
|
{ |
|
29
|
12 |
|
$this->url = self::normalizeUrl($url); |
|
30
|
|
|
if (null !== $element) { |
|
31
|
9 |
|
$this->setAnchor($element); |
|
32
|
|
|
} |
|
33
|
9 |
|
$this->element = $element; |
|
34
|
|
|
|
|
35
|
|
|
$this->type = $type ?? (null !== $element ? $this->getTypeFromElement($element) : null); |
|
36
|
3 |
|
} |
|
37
|
|
|
|
|
38
|
3 |
|
public static function createRedirection(string $url, int $redirType = null): self |
|
39
|
|
|
{ |
|
40
|
|
|
return new self($url, null, $redirType ?? self::LINK_3XX); |
|
41
|
3 |
|
} |
|
42
|
|
|
|
|
43
|
3 |
|
public static function create(\DOMElement $element) |
|
44
|
|
|
{ |
|
45
|
|
|
if ('a' != $element->tagName || !$element->getAttribute('href')) { |
|
46
|
3 |
|
throw new \Exception('no link found in '.$element->nodeValue); |
|
47
|
|
|
} |
|
48
|
3 |
|
|
|
49
|
|
|
return new self($element->getAttribute('href'), $element, self::LINK_A); |
|
50
|
|
|
} |
|
51
|
6 |
|
|
|
52
|
|
|
protected function getTypeFromElement(\DomElement $element) |
|
53
|
6 |
|
{ |
|
54
|
3 |
|
return 'a' == $element->tagName && $element->getAttribute('href') ? self::LINK_A : null; |
|
55
|
3 |
|
} |
|
56
|
|
|
|
|
57
|
|
|
protected static function normalizeUrl(string $url) |
|
58
|
|
|
{ |
|
59
|
3 |
|
$url = trim($url); |
|
60
|
|
|
|
|
61
|
|
|
if ('' == preg_replace('@(.*\://?([^\/]+))@', '', $url)) { |
|
62
|
|
|
$url = $url.'/'; |
|
63
|
|
|
} |
|
64
|
|
|
|
|
65
|
|
|
return $url; |
|
66
|
|
|
} |
|
67
|
|
|
|
|
68
|
|
|
protected function getType(): ?int |
|
69
|
|
|
{ |
|
70
|
|
|
return $this->type; |
|
71
|
|
|
} |
|
72
|
|
|
|
|
73
|
|
|
protected function setAnchor(\DomElement $element) |
|
74
|
|
|
{ |
|
75
|
|
|
// Get classic text anchor |
|
76
|
|
|
$this->anchor = $element->textContent; |
|
77
|
|
|
|
|
78
|
|
|
// If get nothing, then maybe we can get an alternative text (eg: img) |
|
79
|
|
|
if (empty($this->anchor)) { |
|
80
|
|
|
$alt = (new DomCrawler($element))->filter('*[alt]'); |
|
81
|
|
|
if ($alt->count() > 0) { |
|
82
|
|
|
$this->anchor = $alt->eq(0)->attr('alt') ?? ''; |
|
83
|
|
|
} |
|
84
|
|
|
} |
|
85
|
|
|
|
|
86
|
|
|
// Limit to 100 characters |
|
87
|
|
|
// Totally subjective |
|
88
|
|
|
$this->anchor = substr(Helper::clean($this->anchor), 0, 99); |
|
89
|
|
|
|
|
90
|
|
|
return $this; |
|
91
|
|
|
} |
|
92
|
|
|
|
|
93
|
|
|
public function getUrl() |
|
94
|
|
|
{ |
|
95
|
|
|
return $this->url; |
|
96
|
|
|
} |
|
97
|
|
|
|
|
98
|
|
|
public function getPageUrl() |
|
99
|
|
|
{ |
|
100
|
|
|
return preg_replace('/(\#.*)/si', '', $this->url); |
|
101
|
|
|
} |
|
102
|
|
|
|
|
103
|
|
|
public function getPageUrlWithoutBase(string $base, bool $page = false) |
|
104
|
|
|
{ |
|
105
|
|
|
$url = $this->url; |
|
106
|
|
|
|
|
107
|
|
|
if (0 === strpos($this->url, $base)) { |
|
108
|
|
|
$url = substr($url, strlen($base)); |
|
109
|
|
|
} |
|
110
|
|
|
|
|
111
|
|
|
return true === $page ? preg_replace('/(\#.*)/si', '', $url) : $url; |
|
112
|
|
|
} |
|
113
|
|
|
|
|
114
|
|
|
public function getAnchor() |
|
115
|
|
|
{ |
|
116
|
|
|
return $this->anchor; |
|
117
|
|
|
} |
|
118
|
|
|
|
|
119
|
|
|
public function getElement() |
|
120
|
|
|
{ |
|
121
|
|
|
return $this->element; |
|
122
|
|
|
} |
|
123
|
|
|
|
|
124
|
|
|
/** |
|
125
|
|
|
* @return bool |
|
126
|
|
|
*/ |
|
127
|
|
|
public function mayFollow() |
|
128
|
|
|
{ |
|
129
|
|
|
if (null !== $this->element && null !== $this->element->getAttribute('rel')) { |
|
130
|
|
|
if (false !== strpos($this->element->getAttribute('rel'), 'nofollow')) { |
|
131
|
|
|
return false; |
|
132
|
|
|
} |
|
133
|
|
|
} |
|
134
|
|
|
|
|
135
|
|
|
return true; |
|
136
|
|
|
} |
|
137
|
|
|
|
|
138
|
|
|
/** |
|
139
|
|
|
* @return string |
|
140
|
|
|
*/ |
|
141
|
|
|
public function getRel(): ?string |
|
142
|
|
|
{ |
|
143
|
|
|
return $this->element->getAttribute('rel'); |
|
|
|
|
|
|
144
|
|
|
} |
|
145
|
|
|
} |
|
146
|
|
|
|
PiedWeb /
UrlHarvester
This check looks for calls to methods that do not seem to exist on a given type. It looks for the method on the type itself as well as in inherited classes or implemented interfaces.
This is most likely a typographical error or the method has been renamed.