1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
/* |
4
|
|
|
* This file is part of the ni-ju-san CMS. |
5
|
|
|
* |
6
|
|
|
* (c) Christian Gripp <[email protected]> |
7
|
|
|
* |
8
|
|
|
* For the full copyright and license information, please view the LICENSE |
9
|
|
|
* file that was distributed with this source code. |
10
|
|
|
*/ |
11
|
|
|
|
12
|
|
|
namespace Core23\LastFm\Crawler; |
13
|
|
|
|
14
|
|
|
use Core23\LastFm\Connection\ConnectionInterface; |
15
|
|
|
use Symfony\Component\DomCrawler\Crawler; |
16
|
|
|
|
17
|
|
|
abstract class AbstractCrawler |
18
|
|
|
{ |
19
|
|
|
const URL_PREFIX = 'http://last.fm'; |
20
|
|
|
|
21
|
|
|
/** |
22
|
|
|
* @var ConnectionInterface |
23
|
|
|
*/ |
24
|
|
|
protected $connection; |
25
|
|
|
|
26
|
|
|
/** |
27
|
|
|
* AbstractService constructor. |
28
|
|
|
* |
29
|
|
|
* @param ConnectionInterface $connection |
30
|
|
|
*/ |
31
|
|
|
public function __construct(ConnectionInterface $connection) |
32
|
|
|
{ |
33
|
|
|
$this->connection = $connection; |
34
|
|
|
} |
35
|
|
|
|
36
|
|
|
/** |
37
|
|
|
* Crawles a url. |
38
|
|
|
* |
39
|
|
|
* @param string $url |
40
|
|
|
* |
41
|
|
|
* @return Crawler |
42
|
|
|
*/ |
43
|
|
|
final protected function crawl($url) |
44
|
|
|
{ |
45
|
|
|
$content = $this->connection->loadPage($url); |
46
|
|
|
|
47
|
|
|
return new Crawler($content); |
48
|
|
|
} |
49
|
|
|
|
50
|
|
|
/** |
51
|
|
|
* Parses a url node. |
52
|
|
|
* |
53
|
|
|
* @param Crawler $node |
54
|
|
|
* @param string $attr |
55
|
|
|
* |
56
|
|
|
* @return null|string |
57
|
|
|
*/ |
58
|
|
|
final protected function parseUrl(Crawler $node, $attr = 'href') |
59
|
|
|
{ |
60
|
|
|
if ($node == null || $node->count() === 0) { |
61
|
|
|
return; |
62
|
|
|
} |
63
|
|
|
|
64
|
|
|
if ($url = $node->attr($attr)) { |
65
|
|
|
return preg_replace('/^\//', static::URL_PREFIX.'/', $url); |
66
|
|
|
} |
67
|
|
|
|
68
|
|
|
return; |
69
|
|
|
} |
70
|
|
|
|
71
|
|
|
/** |
72
|
|
|
* Parses an image node. |
73
|
|
|
* |
74
|
|
|
* @param Crawler $node |
75
|
|
|
* |
76
|
|
|
* @return null|string |
77
|
|
|
*/ |
78
|
|
|
final protected function parseImage(Crawler $node) |
79
|
|
|
{ |
80
|
|
|
return $this->parseUrl($node, 'src'); |
81
|
|
|
} |
82
|
|
|
|
83
|
|
|
/** |
84
|
|
|
* Parses a string node. |
85
|
|
|
* |
86
|
|
|
* @param Crawler $node |
87
|
|
|
* |
88
|
|
|
* @param bool $multiline |
89
|
|
|
* |
90
|
|
|
* @return null|string |
91
|
|
|
*/ |
92
|
|
|
final protected function parseString(Crawler $node, $multiline = false) |
93
|
|
|
{ |
94
|
|
|
if ($node == null || $node->count() === 0) { |
95
|
|
|
return; |
96
|
|
|
} |
97
|
|
|
|
98
|
|
|
$content = $node->attr('content'); |
99
|
|
|
|
100
|
|
|
if (!$content) { |
|
|
|
|
101
|
|
|
if ($multiline) { |
102
|
|
|
$content = strip_tags($node->html()); |
103
|
|
|
} else { |
104
|
|
|
$content = $node->text(); |
105
|
|
|
} |
106
|
|
|
} |
107
|
|
|
|
108
|
|
|
return trim($content); |
109
|
|
|
} |
110
|
|
|
|
111
|
|
|
/** |
112
|
|
|
* Parses a date note. |
113
|
|
|
* |
114
|
|
|
* @param Crawler $node |
115
|
|
|
* |
116
|
|
|
* @return \DateTime|null |
117
|
|
|
*/ |
118
|
|
|
final protected function parseDate(Crawler $node) |
119
|
|
|
{ |
120
|
|
|
$content = $this->parseString($node); |
121
|
|
|
|
122
|
|
|
if ($content) { |
|
|
|
|
123
|
|
|
return new \DateTime($content); |
124
|
|
|
} |
125
|
|
|
|
126
|
|
|
return; |
127
|
|
|
} |
128
|
|
|
} |
129
|
|
|
|
In PHP, under loose comparison (like
==
, or!=
, orswitch
conditions), values of different types might be equal.For
string
values, the empty string''
is a special case, in particular the following results might be unexpected: