1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
declare(strict_types=1); |
4
|
|
|
|
5
|
|
|
namespace XoopsModules\Publisher; |
6
|
|
|
|
7
|
|
|
/* |
8
|
|
|
You may not change or alter any portion of this comment or credits |
9
|
|
|
of supporting developers from this source code or any supporting source code |
10
|
|
|
which is considered copyrighted (c) material of the original comment or credit authors. |
11
|
|
|
|
12
|
|
|
This program is distributed in the hope that it will be useful, |
13
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of |
14
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. |
15
|
|
|
*/ |
16
|
|
|
|
17
|
|
|
/** |
18
|
|
|
* @copyright The XUUPS Project http://sourceforge.net/projects/xuups/ |
19
|
|
|
* @license http://www.fsf.org/copyleft/gpl.html GNU public license |
20
|
|
|
* @since 1.0 |
21
|
|
|
* @author trabis <[email protected]> |
22
|
|
|
* @author The SmartFactory <www.smartfactory.ca> |
23
|
|
|
*/ |
24
|
|
|
|
25
|
|
|
use XoopsModules\Publisher\Helper; |
26
|
|
|
|
27
|
|
|
require_once \dirname(__DIR__) . '/include/common.php'; |
28
|
|
|
|
29
|
|
|
/** |
30
|
|
|
* Class Metagen |
31
|
|
|
*/ |
32
|
|
|
class Metagen |
33
|
|
|
{ |
34
|
|
|
/** |
35
|
|
|
* @var Helper |
36
|
|
|
*/ |
37
|
|
|
public $helper; |
38
|
|
|
/** |
39
|
|
|
* @var \MyTextSanitizer |
40
|
|
|
*/ |
41
|
|
|
public $myts; |
42
|
|
|
/** |
43
|
|
|
* @var string |
44
|
|
|
*/ |
45
|
|
|
public $title; |
46
|
|
|
/** |
47
|
|
|
* @var string |
48
|
|
|
*/ |
49
|
|
|
public $originalTitle; |
50
|
|
|
/** |
51
|
|
|
* @var string |
52
|
|
|
*/ |
53
|
|
|
public $keywords; |
54
|
|
|
/** |
55
|
|
|
* @var string |
56
|
|
|
*/ |
57
|
|
|
public $categoryPath; |
58
|
|
|
/** |
59
|
|
|
* @var string |
60
|
|
|
*/ |
61
|
|
|
public $description; |
62
|
|
|
/** |
63
|
|
|
* @var int |
64
|
|
|
*/ |
65
|
|
|
public $minChar = 4; |
66
|
|
|
|
67
|
|
|
/** |
68
|
|
|
* @param string $title |
69
|
|
|
* @param string $keywords |
70
|
|
|
* @param string $description |
71
|
|
|
* @param string $categoryPath |
72
|
|
|
*/ |
73
|
|
|
public function __construct($title, $keywords = '', $description = '', $categoryPath = '') |
74
|
|
|
{ |
75
|
|
|
/** @var Helper $this- >helper */ |
76
|
|
|
$this->helper = Helper::getInstance(); |
77
|
|
|
$this->myts = \MyTextSanitizer::getInstance(); |
78
|
|
|
$this->setCategoryPath($categoryPath); |
79
|
|
|
$this->setTitle($title); |
80
|
|
|
$this->setDescription($description); |
81
|
|
|
if ('' == $keywords) { |
82
|
|
|
$keywords = $this->createMetaKeywords(); |
83
|
|
|
} |
84
|
|
|
$this->setKeywords($keywords); |
85
|
|
|
} |
86
|
|
|
|
87
|
|
|
/** |
88
|
|
|
* @param string $title |
89
|
|
|
*/ |
90
|
|
|
public function setTitle($title) |
91
|
|
|
{ |
92
|
|
|
$this->title = $this->html2text($title); |
93
|
|
|
$this->originalTitle = $this->title; |
94
|
|
|
$titleTag = []; |
95
|
|
|
$titleTag['module'] = $this->helper->getModule()->getVar('name'); |
96
|
|
|
if (isset($this->title) && ('' != $this->title) && (\mb_strtoupper($this->title) != \mb_strtoupper($titleTag['module']))) { |
|
|
|
|
97
|
|
|
$titleTag['title'] = $this->title; |
98
|
|
|
} |
99
|
|
|
if (isset($this->categoryPath) && ('' != $this->categoryPath)) { |
100
|
|
|
$titleTag['category'] = $this->categoryPath; |
101
|
|
|
} |
102
|
|
|
$ret = $titleTag['title'] ?? ''; |
103
|
|
|
if (isset($titleTag['category']) && '' != $titleTag['category']) { |
104
|
|
|
if ('' != $ret) { |
105
|
|
|
$ret .= ' - '; |
106
|
|
|
} |
107
|
|
|
$ret .= $titleTag['category']; |
108
|
|
|
} |
109
|
|
|
if (isset($titleTag['module']) && '' != $titleTag['module']) { |
110
|
|
|
if ('' != $ret) { |
111
|
|
|
$ret .= ' - '; |
112
|
|
|
} |
113
|
|
|
$ret .= $titleTag['module']; |
114
|
|
|
} |
115
|
|
|
$this->title = $ret; |
116
|
|
|
} |
117
|
|
|
|
118
|
|
|
/** |
119
|
|
|
* @param string $keywords |
120
|
|
|
*/ |
121
|
|
|
public function setKeywords($keywords) |
122
|
|
|
{ |
123
|
|
|
$this->keywords = $keywords; |
124
|
|
|
} |
125
|
|
|
|
126
|
|
|
/** |
127
|
|
|
* @param string $categoryPath |
128
|
|
|
*/ |
129
|
|
|
public function setCategoryPath($categoryPath) |
130
|
|
|
{ |
131
|
|
|
$categoryPath = $this->html2text($categoryPath); |
132
|
|
|
$this->categoryPath = $categoryPath; |
133
|
|
|
} |
134
|
|
|
|
135
|
|
|
/** |
136
|
|
|
* @param string $description |
137
|
|
|
*/ |
138
|
|
|
public function setDescription($description) |
139
|
|
|
{ |
140
|
|
|
$description = $this->html2text($description); |
141
|
|
|
$description = $this->purifyText($description); |
142
|
|
|
$this->description = $description; |
143
|
|
|
} |
144
|
|
|
|
145
|
|
|
/** |
146
|
|
|
* Does nothing |
147
|
|
|
*/ |
148
|
|
|
public function createTitleTag() |
149
|
|
|
{ |
150
|
|
|
} |
151
|
|
|
|
152
|
|
|
/** |
153
|
|
|
* @param int $maxWords |
154
|
|
|
* |
155
|
|
|
* @return string |
156
|
|
|
*/ |
157
|
|
|
public function createMetaDescription($maxWords = 30) |
|
|
|
|
158
|
|
|
{ |
159
|
|
|
$description = $this->purifyText($this->description); |
160
|
|
|
$description = $this->html2text($description); |
161
|
|
|
$words = \explode(' ', $description); |
162
|
|
|
$ret = ''; |
163
|
|
|
$i = 1; |
164
|
|
|
$wordCount = \count($words); |
165
|
|
|
foreach ($words as $word) { |
166
|
|
|
$ret .= $word; |
167
|
|
|
if ($i < $wordCount) { |
168
|
|
|
$ret .= ' '; |
169
|
|
|
} |
170
|
|
|
++$i; |
171
|
|
|
} |
172
|
|
|
|
173
|
|
|
return $ret; |
174
|
|
|
} |
175
|
|
|
|
176
|
|
|
/** |
177
|
|
|
* @param string $text |
178
|
|
|
* @param int $minChar |
179
|
|
|
* |
180
|
|
|
* @return array |
181
|
|
|
*/ |
182
|
|
|
public function findMetaKeywords($text, $minChar) |
183
|
|
|
{ |
184
|
|
|
$keywords = []; |
185
|
|
|
$text = $this->purifyText($text); |
186
|
|
|
$text = $this->html2text($text); |
187
|
|
|
$originalKeywords = \explode(' ', $text); |
188
|
|
|
foreach ($originalKeywords as $originalKeyword) { |
189
|
|
|
$secondRoundKeywords = \explode("'", $originalKeyword); |
190
|
|
|
foreach ($secondRoundKeywords as $secondRoundKeyword) { |
191
|
|
|
if (\mb_strlen($secondRoundKeyword) >= $minChar) { |
192
|
|
|
if (!\in_array($secondRoundKeyword, $keywords, true)) { |
193
|
|
|
$keywords[] = \trim($secondRoundKeyword); |
194
|
|
|
} |
195
|
|
|
} |
196
|
|
|
} |
197
|
|
|
} |
198
|
|
|
|
199
|
|
|
return $keywords; |
200
|
|
|
} |
201
|
|
|
|
202
|
|
|
/** |
203
|
|
|
* @return string |
204
|
|
|
*/ |
205
|
|
|
public function createMetaKeywords() |
206
|
|
|
{ |
207
|
|
|
$keywords = $this->findMetaKeywords($this->originalTitle . ' ' . $this->description, $this->minChar); |
208
|
|
|
$moduleKeywords = $this->helper->getConfig('seo_meta_keywords'); |
209
|
|
|
if ('' != $moduleKeywords) { |
210
|
|
|
$moduleKeywords = \explode(',', $moduleKeywords); |
211
|
|
|
$keywords = \array_merge($keywords, \array_map('\trim', $moduleKeywords)); |
212
|
|
|
} |
213
|
|
|
$ret = \implode(',', $keywords); |
214
|
|
|
|
215
|
|
|
return $ret; |
216
|
|
|
} |
217
|
|
|
|
218
|
|
|
/** |
219
|
|
|
* Does nothing |
220
|
|
|
*/ |
221
|
|
|
public function autoBuildMetaKeywords() |
222
|
|
|
{ |
223
|
|
|
} |
224
|
|
|
|
225
|
|
|
/** |
226
|
|
|
* Build Metatags |
227
|
|
|
*/ |
228
|
|
|
public function buildAutoMetaTags() |
229
|
|
|
{ |
230
|
|
|
$this->keywords = $this->createMetaKeywords(); |
231
|
|
|
$this->description = $this->createMetaDescription(); |
232
|
|
|
//$this->title = $this->createTitleTag(); |
233
|
|
|
} |
234
|
|
|
|
235
|
|
|
/** |
236
|
|
|
* Creates meta tags |
237
|
|
|
*/ |
238
|
|
|
public function createMetaTags() |
239
|
|
|
{ |
240
|
|
|
global $xoopsTpl, $xoTheme; |
241
|
|
|
if ('' != $this->keywords) { |
242
|
|
|
$xoTheme->addMeta('meta', 'keywords', $this->keywords); |
243
|
|
|
} |
244
|
|
|
if ('' != $this->description) { |
245
|
|
|
$xoTheme->addMeta('meta', 'description', $this->description); |
246
|
|
|
} |
247
|
|
|
if ('' != $this->title) { |
248
|
|
|
$xoopsTpl->assign('xoops_pagetitle', $this->title); |
249
|
|
|
} |
250
|
|
|
} |
251
|
|
|
|
252
|
|
|
/** |
253
|
|
|
* Return true if the string is length > 0 |
254
|
|
|
* |
255
|
|
|
* @credit psylove |
256
|
|
|
* @param mixed $var |
257
|
|
|
* @return bool |
258
|
|
|
*/ |
259
|
|
|
public static function emptyString($var) |
260
|
|
|
{ |
261
|
|
|
return ('' !== $var); |
262
|
|
|
} |
263
|
|
|
|
264
|
|
|
/** |
265
|
|
|
* Create a title for the short_url field of an article |
266
|
|
|
* |
267
|
|
|
* @credit psylove |
268
|
|
|
* |
269
|
|
|
* @param string|array $title title of the article |
270
|
|
|
* @param bool $withExt do we add an html extension or not |
271
|
|
|
* |
272
|
|
|
* @return string short url for article |
273
|
|
|
*/ |
274
|
|
|
public static function generateSeoTitle($title = '', $withExt = true) |
275
|
|
|
{ |
276
|
|
|
// Transformation de la chaine en minuscule |
277
|
|
|
// Codage de la chaine afin d'éviter les erreurs 500 en cas de caractères imprévus |
278
|
|
|
$title = \rawurlencode(\mb_strtolower($title)); |
|
|
|
|
279
|
|
|
// Transformation des ponctuations |
280
|
|
|
|
281
|
|
|
$pattern = [ |
282
|
|
|
'/%09/', // Tab |
283
|
|
|
'/%20/', // Space |
284
|
|
|
'/%21/', // ! |
285
|
|
|
'/%22/', // " |
286
|
|
|
'/%23/', // # |
287
|
|
|
'/%25/', // % |
288
|
|
|
'/%26/', // & |
289
|
|
|
'/%27/', // ' |
290
|
|
|
'/%28/', // ( |
291
|
|
|
'/%29/', // ) |
292
|
|
|
'/%2C/', // , |
293
|
|
|
'/%2F/', // / |
294
|
|
|
'/%3A/', // : |
295
|
|
|
'/%3B/', // ; |
296
|
|
|
'/%3C/', // < |
297
|
|
|
'/%3D/', // = |
298
|
|
|
'/%3E/', // > |
299
|
|
|
'/%3F/', // ? |
300
|
|
|
'/%40/', // @ |
301
|
|
|
'/%5B/', // [ |
302
|
|
|
'/%5C/', // \ |
303
|
|
|
'/%5D/', // ] |
304
|
|
|
'/%5E/', // ^ |
305
|
|
|
'/%7B/', // { |
306
|
|
|
'/%7C/', // | |
307
|
|
|
'/%7D/', // } |
308
|
|
|
'/%7E/', // ~ |
309
|
|
|
"/\./", // . |
310
|
|
|
]; |
311
|
|
|
$repPat = ['-', '-', '-', '-', '-', '-100', '-', '-', '-', '-', '-', '-', '-', '-', '-', '-', '-', '-', '-at-', '-', '-', '-', '-', '-', '-', '-', '-', '-']; |
312
|
|
|
$title = \preg_replace($pattern, $repPat, $title); |
313
|
|
|
// Transformation des caractères accentués |
314
|
|
|
$pattern = [ |
315
|
|
|
'/%B0/', // ° |
316
|
|
|
'/%E8/', // è |
317
|
|
|
'/%E9/', // é |
318
|
|
|
'/%EA/', // ê |
319
|
|
|
'/%EB/', // ë |
320
|
|
|
'/%E7/', // ç |
321
|
|
|
'/%E0/', // à |
322
|
|
|
'/%E2/', // â |
323
|
|
|
'/%E4/', // ä |
324
|
|
|
'/%EE/', // î |
325
|
|
|
'/%EF/', // ï |
326
|
|
|
'/%F9/', // ù |
327
|
|
|
'/%FC/', // ü |
328
|
|
|
'/%FB/', // û |
329
|
|
|
'/%F4/', // ô |
330
|
|
|
'/%F6/', // ö |
331
|
|
|
]; |
332
|
|
|
$repPat = ['-', 'e', 'e', 'e', 'e', 'c', 'a', 'a', 'a', 'i', 'i', 'u', 'u', 'u', 'o', 'o']; |
333
|
|
|
$title = \preg_replace($pattern, $repPat, $title); |
334
|
|
|
$tableau = \explode('-', $title); // Transforms the string in table //Transforme la chaine de caractères en tableau |
335
|
|
|
$tableau = \array_filter($tableau, [__CLASS__, 'emptyString']); // Remove empty strings of the table //Supprime les chaines vides du tableau |
336
|
|
|
$title = \implode('-', $tableau); // Transforms a character string in table separated by a hyphen //Transforme un tableau en chaine de caractères séparé par un tiret |
337
|
|
|
if ($title && !(\is_array($title))) { |
338
|
|
|
if ($withExt) { |
339
|
|
|
$title .= '.html'; |
340
|
|
|
} |
341
|
|
|
|
342
|
|
|
return $title; |
343
|
|
|
} |
344
|
|
|
|
345
|
|
|
return ''; |
346
|
|
|
} |
347
|
|
|
|
348
|
|
|
/** |
349
|
|
|
* @param $text |
350
|
|
|
* @param bool $keyword |
351
|
|
|
* |
352
|
|
|
* @return mixed |
353
|
|
|
*/ |
354
|
|
|
public function purifyText($text, $keyword = false) |
355
|
|
|
{ |
356
|
|
|
// $text = str_replace([' ', ' '], ['<br>', ' '], $text); //for php 5.4 |
357
|
|
|
$text = \str_replace(' ', ' ', $text); |
358
|
|
|
$text = \str_replace('<br>', ' ', $text); |
359
|
|
|
$text = \strip_tags($text); |
360
|
|
|
$text = \html_entity_decode($text); |
361
|
|
|
$text = $this->myts->undoHtmlSpecialChars($text); |
362
|
|
|
|
363
|
|
|
$text = \str_replace(')', ' ', $text); |
364
|
|
|
$text = \str_replace('(', ' ', $text); |
365
|
|
|
$text = \str_replace(':', ' ', $text); |
366
|
|
|
$text = \str_replace('&euro', ' euro ', $text); |
367
|
|
|
$text = \str_replace('&hellip', '...', $text); |
368
|
|
|
$text = \str_replace('&rsquo', ' ', $text); |
369
|
|
|
$text = \str_replace('!', ' ', $text); |
370
|
|
|
$text = \str_replace('?', ' ', $text); |
371
|
|
|
$text = \str_replace('"', ' ', $text); |
372
|
|
|
$text = \str_replace('-', ' ', $text); |
373
|
|
|
$text = \str_replace('\n', ' ', $text); |
374
|
|
|
|
375
|
|
|
// $text = str_replace([')','(',':','&euro','&hellip','&rsquo','!','?','"','-','\n'], [' ' , ' ', ' ', ' euro ', '...', ' ', ' ', ' ', ' ', ' ', ' '], $text); //for PHP 5.4 |
376
|
|
|
|
377
|
|
|
if ($keyword) { |
378
|
|
|
$text = \str_replace('.', ' ', $text); |
379
|
|
|
$text = \str_replace(',', ' ', $text); |
380
|
|
|
$text = \str_replace('\'', ' ', $text); |
381
|
|
|
// $text = str_replace(['.', ' '], [',', ' '], ['\'', ' '], $text); //for PHP 5.4 |
382
|
|
|
} |
383
|
|
|
$text = \str_replace(';', ' ', $text); |
384
|
|
|
|
385
|
|
|
return $text; |
386
|
|
|
} |
387
|
|
|
|
388
|
|
|
/** |
389
|
|
|
* @param string $document |
390
|
|
|
* |
391
|
|
|
* @return mixed |
392
|
|
|
*/ |
393
|
|
|
public function html2text($document) |
394
|
|
|
{ |
395
|
|
|
if (empty($document)) { |
396
|
|
|
return ''; |
397
|
|
|
} |
398
|
|
|
// PHP Manual:: function preg_replace |
399
|
|
|
// $document should contain an HTML document. |
400
|
|
|
// This will remove HTML tags, javascript sections |
401
|
|
|
// and white space. It will also convert some |
402
|
|
|
// common HTML entities to their text equivalent. |
403
|
|
|
// Credits : newbb2 |
404
|
|
|
$search = [ |
405
|
|
|
"'<script[^>]*?>.*?</script>'si", // Strip out javascript |
406
|
|
|
"'<img.*?>'si", // Strip out img tags |
407
|
|
|
"'<[\/\!]*?[^<>]*?>'si", // Strip out HTML tags |
408
|
|
|
"'([\r\n])[\s]+'", // Strip out white space |
409
|
|
|
"'&(quot|#34);'i", // Replace HTML entities |
410
|
|
|
"'&(amp|#38);'i", |
411
|
|
|
"'&(lt|#60);'i", |
412
|
|
|
"'&(gt|#62);'i", |
413
|
|
|
"'&(nbsp|#160);'i", |
414
|
|
|
"'&(iexcl|#161);'i", |
415
|
|
|
"'&(cent|#162);'i", |
416
|
|
|
"'&(pound|#163);'i", |
417
|
|
|
"'&(copy|#169);'i", |
418
|
|
|
]; // evaluate as php |
419
|
|
|
|
420
|
|
|
$replace = [ |
421
|
|
|
'', |
422
|
|
|
'', |
423
|
|
|
'', |
424
|
|
|
'\\1', |
425
|
|
|
'"', |
426
|
|
|
'&', |
427
|
|
|
'<', |
428
|
|
|
'>', |
429
|
|
|
' ', |
430
|
|
|
\chr(161), |
431
|
|
|
\chr(162), |
432
|
|
|
\chr(163), |
433
|
|
|
\chr(169), |
434
|
|
|
]; |
435
|
|
|
|
436
|
|
|
$text = \preg_replace($search, $replace, $document); |
437
|
|
|
|
438
|
|
|
\preg_replace_callback( |
439
|
|
|
'/&#(\d+);/', |
440
|
|
|
static function ($matches) { |
441
|
|
|
return \chr($matches[1]); |
442
|
|
|
}, |
443
|
|
|
$document |
444
|
|
|
); |
445
|
|
|
|
446
|
|
|
return $text; |
447
|
|
|
} |
448
|
|
|
} |
449
|
|
|
|