|
1
|
|
|
<?php declare(strict_types=1); |
|
2
|
|
|
|
|
3
|
|
|
namespace Goose; |
|
4
|
|
|
|
|
5
|
|
|
use Goose\Text\StopWords; |
|
6
|
|
|
use Goose\Modules\ModuleInterface; |
|
7
|
|
|
|
|
8
|
|
|
/** |
|
9
|
|
|
* Configuration |
|
10
|
|
|
* |
|
11
|
|
|
* @package Goose |
|
12
|
|
|
* @license http://www.apache.org/licenses/LICENSE-2.0 Apache License 2.0 |
|
13
|
|
|
*/ |
|
14
|
|
|
class Configuration { |
|
15
|
|
|
/** @var mixed[] */ |
|
16
|
|
|
protected $options = [ |
|
17
|
|
|
'language' => 'en', |
|
18
|
|
|
'image_min_bytes' => 4500, |
|
19
|
|
|
'image_max_bytes' => 5242880, |
|
20
|
|
|
'image_min_width' => 120, |
|
21
|
|
|
'image_min_height' => 120, |
|
22
|
|
|
'image_fetch_best' => true, |
|
23
|
|
|
'image_fetch_all' => false, |
|
24
|
|
|
/** @see http://guzzle.readthedocs.io/en/stable/request-options.html */ |
|
25
|
|
|
'browser' => [ |
|
26
|
|
|
'timeout' => 60, |
|
27
|
|
|
'connect_timeout' => 30, |
|
28
|
|
|
'headers' => [ |
|
29
|
|
|
'User-Agent' => 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/41.0.2272.118 Safari/537.36', |
|
30
|
|
|
'Referer' => 'https://www.google.com/', |
|
31
|
|
|
], |
|
32
|
|
|
] |
|
33
|
|
|
]; |
|
34
|
|
|
|
|
35
|
|
|
/** @var mixed[] */ |
|
36
|
|
|
protected $modules = [ |
|
37
|
|
|
'cleaners' => [ |
|
38
|
|
|
'\Goose\Modules\Cleaners\DocumentCleaner', |
|
39
|
|
|
], |
|
40
|
|
|
'extractors' => [ |
|
41
|
|
|
'\Goose\Modules\Extractors\MetaExtractor', |
|
42
|
|
|
'\Goose\Modules\Extractors\ContentExtractor', |
|
43
|
|
|
'\Goose\Modules\Extractors\ImageExtractor', |
|
44
|
|
|
'\Goose\Modules\Extractors\PublishDateExtractor', |
|
45
|
|
|
'\Goose\Modules\Extractors\AdditionalDataExtractor', |
|
46
|
|
|
], |
|
47
|
|
|
'formatters' => [ |
|
48
|
|
|
'\Goose\Modules\Formatters\OutputFormatter', |
|
49
|
|
|
], |
|
50
|
|
|
]; |
|
51
|
|
|
|
|
52
|
|
|
/** |
|
53
|
|
|
* @param mixed[] $options |
|
54
|
|
|
*/ |
|
55
|
|
|
public function __construct(array $options = []) { |
|
56
|
|
|
$this->options = array_replace_recursive($this->options, $options); |
|
57
|
|
|
} |
|
58
|
|
|
|
|
59
|
|
|
/** |
|
60
|
|
|
* @param string $option |
|
61
|
|
|
* |
|
62
|
|
|
* @return mixed |
|
63
|
|
|
*/ |
|
64
|
|
|
public function get(string $option) { |
|
65
|
|
|
if (isset($this->options[$option])) { |
|
66
|
|
|
return $this->options[$option]; |
|
67
|
|
|
} |
|
68
|
|
|
|
|
69
|
|
|
return null; |
|
70
|
|
|
} |
|
71
|
|
|
|
|
72
|
|
|
/** |
|
73
|
|
|
* @param string $option |
|
74
|
|
|
* @param mixed $value |
|
75
|
|
|
* |
|
76
|
|
|
* @return self |
|
77
|
|
|
*/ |
|
78
|
|
|
public function set(string $option, $value): self { |
|
79
|
|
|
$this->options[$option] = $value; |
|
80
|
|
|
|
|
81
|
|
|
return $this; |
|
82
|
|
|
} |
|
83
|
|
|
|
|
84
|
|
|
/** |
|
85
|
|
|
* @param string $category |
|
86
|
|
|
* |
|
87
|
|
|
* @return mixed |
|
88
|
|
|
*/ |
|
89
|
|
|
public function getModules(string $category) { |
|
90
|
|
|
if (isset($this->modules[$category])) { |
|
91
|
|
|
return $this->modules[$category]; |
|
92
|
|
|
} |
|
93
|
|
|
|
|
94
|
|
|
return null; |
|
95
|
|
|
} |
|
96
|
|
|
|
|
97
|
|
|
/** |
|
98
|
|
|
* @param string $category |
|
99
|
|
|
* @param string[] $classes |
|
100
|
|
|
* |
|
101
|
|
|
* @return self |
|
102
|
|
|
*/ |
|
103
|
|
|
public function setModules(string $category, array $classes): self { |
|
104
|
|
|
if ($this->areValidModules($category, $classes)) { |
|
105
|
|
|
$this->modules[$category] = $classes; |
|
106
|
|
|
} |
|
107
|
|
|
|
|
108
|
|
|
return $this; |
|
109
|
|
|
} |
|
110
|
|
|
|
|
111
|
|
|
/** |
|
112
|
|
|
* @param string $category |
|
113
|
|
|
* @param string $class |
|
114
|
|
|
* |
|
115
|
|
|
* @return self |
|
116
|
|
|
*/ |
|
117
|
|
|
public function addModule(string $category, string $class): self { |
|
118
|
|
|
if ($this->isValidModule($category, $class)) { |
|
119
|
|
|
$this->modules[$category][] = $class; |
|
120
|
|
|
} |
|
121
|
|
|
|
|
122
|
|
|
return $this; |
|
123
|
|
|
} |
|
124
|
|
|
|
|
125
|
|
|
/** |
|
126
|
|
|
* @param string $category |
|
127
|
|
|
* @param string $class |
|
128
|
|
|
* |
|
129
|
|
|
* @return self |
|
130
|
|
|
*/ |
|
131
|
|
|
public function removeModule(string $category, string $class): self { |
|
132
|
|
|
if (isset($this->modules[$category])) { |
|
133
|
|
|
$key = array_search($class, $this->modules[$category]); |
|
134
|
|
|
|
|
135
|
|
|
if ($key !== false) { |
|
136
|
|
|
unset($this->modules[$category][$key]); |
|
137
|
|
|
} |
|
138
|
|
|
} |
|
139
|
|
|
|
|
140
|
|
|
return $this; |
|
141
|
|
|
} |
|
142
|
|
|
|
|
143
|
|
|
/** |
|
144
|
|
|
* @param string $category |
|
145
|
|
|
* @param string $class |
|
146
|
|
|
* |
|
147
|
|
|
* @return bool |
|
148
|
|
|
*/ |
|
149
|
|
|
public function isValidModule(string $category, string $class): bool { |
|
150
|
|
|
if (isset($this->modules[$category]) |
|
151
|
|
|
&& is_subclass_of($class, ModuleInterface::class)) { |
|
152
|
|
|
return true; |
|
153
|
|
|
} |
|
154
|
|
|
|
|
155
|
|
|
return false; |
|
156
|
|
|
} |
|
157
|
|
|
|
|
158
|
|
|
/** |
|
159
|
|
|
* @param string $category |
|
160
|
|
|
* @param string[] $classes |
|
161
|
|
|
* |
|
162
|
|
|
* @return bool |
|
163
|
|
|
*/ |
|
164
|
|
|
public function areValidModules(string $category, array $classes): bool { |
|
165
|
|
|
foreach ($classes as $class) { |
|
166
|
|
|
if (!$this->isValidModule($category, $class)) { |
|
167
|
|
|
return false; |
|
168
|
|
|
} |
|
169
|
|
|
} |
|
170
|
|
|
|
|
171
|
|
|
return true; |
|
172
|
|
|
} |
|
173
|
|
|
|
|
174
|
|
|
/** @var StopWords|null */ |
|
175
|
|
|
protected $stopWords; |
|
176
|
|
|
|
|
177
|
|
|
/** |
|
178
|
|
|
* @param StopWords|null $stopWords |
|
179
|
|
|
* |
|
180
|
|
|
* @return self |
|
181
|
|
|
*/ |
|
182
|
|
|
public function setStopWords(StopWords $stopWords = null): ?self { |
|
183
|
|
|
$this->stopWords = $stopWords; |
|
184
|
|
|
|
|
185
|
|
|
return $this; |
|
186
|
|
|
} |
|
187
|
|
|
|
|
188
|
|
|
/* |
|
189
|
|
|
* @return StopWords |
|
190
|
|
|
*/ |
|
191
|
|
|
public function getStopWords(): StopWords { |
|
192
|
|
|
if (is_null($this->stopWords)) { |
|
193
|
|
|
$this->stopWords = new StopWords($this); |
|
194
|
|
|
} |
|
195
|
|
|
|
|
196
|
|
|
return $this->stopWords; |
|
197
|
|
|
} |
|
198
|
|
|
} |
|
199
|
|
|
|