|
1
|
|
|
<?php
|
|
2
|
|
|
|
|
3
|
|
|
namespace Swader\Diffbot\Traits;
|
|
4
|
|
|
|
|
5
|
|
|
/**
|
|
6
|
|
|
* Trait StandardEntity
|
|
7
|
|
|
* @package Swader\Diffbot\Traits
|
|
8
|
|
|
* @property $data array
|
|
9
|
|
|
*/
|
|
10
|
|
|
trait StandardEntity {
|
|
11
|
|
|
|
|
12
|
|
|
/**
|
|
13
|
|
|
* Alias for getLang()
|
|
14
|
|
|
* @see getLang()
|
|
15
|
|
|
* @return string
|
|
16
|
|
|
*/
|
|
17
|
1 |
|
public function getHumanLanguage()
|
|
18
|
|
|
{
|
|
19
|
1 |
|
return $this->getLang();
|
|
20
|
|
|
}
|
|
21
|
|
|
|
|
22
|
|
|
/**
|
|
23
|
|
|
* Returns the human language of the page as determined by Diffbot when looking at content.
|
|
24
|
|
|
* The code returned is a two-character ISO 639-1 code: http://en.wikipedia.org/wiki/List_of_ISO_639-1_codes
|
|
25
|
|
|
* @return string
|
|
26
|
|
|
*/
|
|
27
|
1 |
|
public function getLang()
|
|
28
|
|
|
{
|
|
29
|
1 |
|
return $this->data['humanLanguage'];
|
|
30
|
|
|
}
|
|
31
|
|
|
|
|
32
|
|
|
/**
|
|
33
|
|
|
* Returns the URL which was crawled
|
|
34
|
|
|
* @return string
|
|
35
|
|
|
*/
|
|
36
|
2 |
|
public function getPageUrl()
|
|
37
|
|
|
{
|
|
38
|
2 |
|
return $this->data['pageUrl'];
|
|
39
|
|
|
}
|
|
40
|
|
|
|
|
41
|
|
|
/**
|
|
42
|
|
|
* Returns page Url which was resolved by redirects, if any.
|
|
43
|
|
|
* For example, crawling a bitly link will make this method return the ultimate destination's URL
|
|
44
|
|
|
* @return string
|
|
45
|
|
|
*/
|
|
46
|
1 |
|
public function getResolvedPageUrl()
|
|
47
|
|
|
{
|
|
48
|
1 |
|
return (isset($this->data['resolvedPageUrl'])) ? $this->data['resolvedPageUrl'] : $this->getPageUrl();
|
|
49
|
|
|
}
|
|
50
|
|
|
|
|
51
|
|
|
/**
|
|
52
|
|
|
* Returns title of article as deduced by Diffbot
|
|
53
|
|
|
* @return string
|
|
54
|
|
|
*/
|
|
55
|
1 |
|
public function getTitle()
|
|
56
|
|
|
{
|
|
57
|
1 |
|
return $this->data['title'];
|
|
58
|
|
|
}
|
|
59
|
|
|
|
|
60
|
|
|
/**
|
|
61
|
|
|
* Returns an array of all links found on the crawled page.
|
|
62
|
|
|
*
|
|
63
|
|
|
* Only available in the standard APIs, not Custom APIs
|
|
64
|
|
|
* @return array|null
|
|
65
|
|
|
*/
|
|
66
|
1 |
|
public function getLinks()
|
|
67
|
|
|
{
|
|
68
|
1 |
|
return (isset($this->data['links'])) ? $this->data['links'] : null;
|
|
69
|
|
|
}
|
|
70
|
|
|
|
|
71
|
|
|
/**
|
|
72
|
|
|
* Returns a top-level object (meta) containing the full contents of page meta tags,
|
|
73
|
|
|
* including sub-arrays for OpenGraph tags, Twitter Card metadata, schema.org microdata,
|
|
74
|
|
|
* and -- if available -- oEmbed metadata.
|
|
75
|
|
|
*
|
|
76
|
|
|
* Only available in the standard APIs, not Custom APIs
|
|
77
|
|
|
* @return array|null
|
|
78
|
|
|
*/
|
|
79
|
1 |
|
public function getMeta()
|
|
80
|
|
|
{
|
|
81
|
1 |
|
return (isset($this->data['meta'])) ? $this->data['meta'] : null;
|
|
82
|
|
|
}
|
|
83
|
|
|
|
|
84
|
|
|
/**
|
|
85
|
|
|
* Returns any key/value pairs present in the URL querystring.
|
|
86
|
|
|
* Items without a discrete value will be returned as true.
|
|
87
|
|
|
*
|
|
88
|
|
|
* Only available in the standard APIs, not Custom APIs
|
|
89
|
|
|
* @return array|null
|
|
90
|
|
|
*/
|
|
91
|
1 |
|
public function getQueryString()
|
|
92
|
|
|
{
|
|
93
|
1 |
|
return (isset($this->data['queryString'])) ? $this->data['queryString'] : null;
|
|
94
|
|
|
}
|
|
95
|
|
|
|
|
96
|
|
|
/**
|
|
97
|
|
|
* Returns a top-level array (breadcrumb) of URLs and link text from page breadcrumbs.
|
|
98
|
|
|
*
|
|
99
|
|
|
* Only available in the standard APIs, not Custom APIs
|
|
100
|
|
|
* @return array|null
|
|
101
|
|
|
*/
|
|
102
|
1 |
|
public function getBreadcrumb()
|
|
103
|
|
|
{
|
|
104
|
1 |
|
return (isset($this->data['breadcrumb'])) ? $this->data['breadcrumb'] : null;
|
|
105
|
|
|
}
|
|
106
|
|
|
|
|
107
|
|
|
/**
|
|
108
|
|
|
* An internal identifier for Diffbot, used for indexing in their databases
|
|
109
|
|
|
* @return string
|
|
110
|
|
|
*/
|
|
111
|
1 |
|
public function getDiffbotUri()
|
|
112
|
|
|
{
|
|
113
|
1 |
|
return $this->data['diffbotUri'];
|
|
114
|
|
|
}
|
|
115
|
|
|
|
|
116
|
|
|
/**
|
|
117
|
|
|
* Returns the timestamp from the point in time the page was indexed by the engine
|
|
118
|
|
|
* Example date: "Wed, 18 Dec 2013 00:00:00 GMT"
|
|
119
|
|
|
* This will be a Carbon (https://github.com/briannesbitt/Carbon) instance if Carbon is installed.
|
|
120
|
|
|
* @return \Carbon\Carbon | string
|
|
121
|
|
|
*/
|
|
122
|
|
|
public function getTimestamp()
|
|
123
|
|
|
{
|
|
124
|
|
|
return (class_exists('\Carbon\Carbon')) ?
|
|
125
|
|
|
new \Carbon\Carbon($this->data['timestamp'], 'GMT') :
|
|
126
|
|
|
$this->data['timestamp'];
|
|
127
|
|
|
}
|
|
128
|
|
|
|
|
129
|
24 |
|
protected function getOrDefault($key, $default = null, $data = null)
|
|
130
|
|
|
{
|
|
131
|
24 |
|
$data = ($data !== null) ?: $this->data;
|
|
132
|
24 |
|
return (isset($data[$key]) ? $data[$key] : $default);
|
|
133
|
|
|
}
|
|
134
|
|
|
|
|
135
|
|
|
}
|
|
136
|
|
|
|
Swader /
diffbot-php-client
