|
1
|
|
|
<?php |
|
2
|
|
|
/* |
|
3
|
|
|
* This file is part of dispositif/wikibot application (@github) |
|
4
|
|
|
* 2019-2023 © Philippe M./Irønie <[email protected]> |
|
5
|
|
|
* For the full copyright and MIT license information, view the license file. |
|
6
|
|
|
*/ |
|
7
|
|
|
|
|
8
|
|
|
declare(strict_types=1); |
|
9
|
|
|
|
|
10
|
|
|
namespace App\Domain\Publisher\Traits; |
|
11
|
|
|
|
|
12
|
|
|
use App\Domain\Enums\Language; |
|
13
|
|
|
use DateTime; |
|
14
|
|
|
use Exception; |
|
15
|
|
|
|
|
16
|
|
|
/** |
|
17
|
|
|
* Typo cleaning, email delete, convert date to d-m-Y, etc. |
|
18
|
|
|
*/ |
|
19
|
|
|
trait MapperConverterTrait |
|
20
|
|
|
{ |
|
21
|
|
|
protected function isAnArticle(?string $str): bool |
|
22
|
|
|
{ |
|
23
|
|
|
return in_array($str, ['article', 'journalArticle']); |
|
24
|
|
|
} |
|
25
|
|
|
|
|
26
|
|
|
protected function convertDCpage(array $meta): ?string |
|
27
|
|
|
{ |
|
28
|
|
|
if (isset($meta['citation_firstpage'])) { |
|
29
|
|
|
$page = $meta['citation_firstpage']; |
|
30
|
|
|
if (isset($meta['citation_lastpage'])) { |
|
31
|
|
|
$page .= '–' . $meta['citation_lastpage']; |
|
32
|
|
|
} |
|
33
|
|
|
|
|
34
|
|
|
return (string)$page; |
|
35
|
|
|
} |
|
36
|
|
|
|
|
37
|
|
|
return null; |
|
38
|
|
|
} |
|
39
|
|
|
|
|
40
|
|
|
/** |
|
41
|
|
|
* Note : à appliquer AVANT wikification (sinon bug sur | ) |
|
42
|
|
|
* |
|
43
|
|
|
* @param string|null $str |
|
44
|
|
|
* |
|
45
|
|
|
* @return string|null |
|
46
|
|
|
*/ |
|
47
|
|
|
protected function clean(?string $str = null): ?string |
|
48
|
|
|
{ |
|
49
|
|
|
if ($str === null) { |
|
50
|
|
|
return null; |
|
51
|
|
|
} |
|
52
|
|
|
$str = $this->stripEmailAdress($str); |
|
53
|
|
|
|
|
54
|
|
|
$str = str_replace( |
|
55
|
|
|
[ |
|
56
|
|
|
'|', |
|
57
|
|
|
"\n", |
|
58
|
|
|
"\t", |
|
59
|
|
|
"\r", |
|
60
|
|
|
''', |
|
61
|
|
|
''', |
|
62
|
|
|
''', |
|
63
|
|
|
''', |
|
64
|
|
|
"\n", |
|
65
|
|
|
" ", |
|
66
|
|
|
"é", |
|
67
|
|
|
'©', |
|
68
|
|
|
'{{', |
|
69
|
|
|
'}}', |
|
70
|
|
|
'[[', |
|
71
|
|
|
']]', |
|
72
|
|
|
], |
|
73
|
|
|
[ |
|
74
|
|
|
'/', |
|
75
|
|
|
' ', |
|
76
|
|
|
' ', |
|
77
|
|
|
'', |
|
78
|
|
|
"’", |
|
79
|
|
|
"'", |
|
80
|
|
|
"'", |
|
81
|
|
|
"'", |
|
82
|
|
|
'', |
|
83
|
|
|
' ', |
|
84
|
|
|
"é", |
|
85
|
|
|
'', |
|
86
|
|
|
'', |
|
87
|
|
|
'', |
|
88
|
|
|
'', |
|
89
|
|
|
'', |
|
90
|
|
|
], |
|
91
|
|
|
$str |
|
92
|
|
|
); |
|
93
|
|
|
|
|
94
|
|
|
$str = html_entity_decode($str); |
|
95
|
|
|
$str = strip_tags($str); |
|
96
|
|
|
|
|
97
|
|
|
return trim($str); |
|
98
|
|
|
} |
|
99
|
|
|
|
|
100
|
|
|
protected function stripEmailAdress(?string $str = null): ?string |
|
101
|
|
|
{ |
|
102
|
|
|
if ($str === null) { |
|
103
|
|
|
return null; |
|
104
|
|
|
} |
|
105
|
|
|
|
|
106
|
|
|
return preg_replace('# ?[^ ]+@[^ ]+\.[A-Z]+#i', '', $str); |
|
107
|
|
|
} |
|
108
|
|
|
|
|
109
|
|
|
protected function convertOGtype2format(?string $ogType): ?string |
|
110
|
|
|
{ |
|
111
|
|
|
if (empty($ogType)) { |
|
112
|
|
|
return null; |
|
113
|
|
|
} |
|
114
|
|
|
// og:type = default: website / video.movie / video.tv_show video.other / article, book, profile |
|
115
|
|
|
if (strpos($ogType, 'video') !== false) { |
|
116
|
|
|
return 'vidéo'; |
|
117
|
|
|
} |
|
118
|
|
|
if (strpos($ogType, 'book') !== false) { |
|
119
|
|
|
return 'livre'; |
|
120
|
|
|
} |
|
121
|
|
|
|
|
122
|
|
|
return null; |
|
123
|
|
|
} |
|
124
|
|
|
|
|
125
|
|
|
/** |
|
126
|
|
|
* https://developers.facebook.com/docs/internationalization#locales |
|
127
|
|
|
* @param string|null $lang |
|
128
|
|
|
* |
|
129
|
|
|
* @return string|null |
|
130
|
|
|
*/ |
|
131
|
|
|
protected function convertLangue(?string $lang = null): ?string |
|
132
|
|
|
{ |
|
133
|
|
|
if (empty($lang)) { |
|
134
|
|
|
return null; |
|
135
|
|
|
} |
|
136
|
|
|
// en_GB |
|
137
|
|
|
if (preg_match('#^([a-z]{2})_[A-Z]{2}$#', $lang, $matches)) { |
|
138
|
|
|
return $matches[1]; |
|
139
|
|
|
} |
|
140
|
|
|
|
|
141
|
|
|
return Language::all2wiki($lang); |
|
142
|
|
|
} |
|
143
|
|
|
|
|
144
|
|
|
protected function convertDate(?string $str): ?string |
|
145
|
|
|
{ |
|
146
|
|
|
if (empty($str)) { |
|
147
|
|
|
return null; |
|
148
|
|
|
} |
|
149
|
|
|
$str = str_replace(' 00:00:00', '', $str); |
|
150
|
|
|
$str = str_replace('/', '-', $str); |
|
151
|
|
|
|
|
152
|
|
|
// "2012" - "1775-1783" (Gallica) |
|
153
|
|
|
if (preg_match('#^[12]\d{3}$#', $str) || preg_match('#^[12]\d{3}-[12]\d{3}$#', $str)) { |
|
154
|
|
|
return $str; |
|
155
|
|
|
} |
|
156
|
|
|
|
|
157
|
|
|
return $this->tryFormatDateOrComment($str); |
|
158
|
|
|
} |
|
159
|
|
|
|
|
160
|
|
|
protected function tryFormatDateOrComment(string $str): string |
|
161
|
|
|
{ |
|
162
|
|
|
try { |
|
163
|
|
|
$date = new DateTime($str); |
|
164
|
|
|
} catch (Exception $e) { |
|
165
|
|
|
// 23/11/2015 00:00:00 |
|
166
|
|
|
if (isset($this->log) && method_exists($this->log, 'notice')) { |
|
167
|
|
|
$this->log->notice('tryFormatDateOrComment failed with ' . $str); |
|
168
|
|
|
} |
|
169
|
|
|
|
|
170
|
|
|
return sprintf('<!-- %s -->', $str); |
|
171
|
|
|
} |
|
172
|
|
|
|
|
173
|
|
|
return $date->format('d-m-Y'); |
|
174
|
|
|
} |
|
175
|
|
|
} |
|
176
|
|
|
|
Dispositif /
Wikibot
