Passed
Push — 2.1 ( ab9191...f7f27f )
by Greg
06:37
created

I18N::translate()   A

Complexity

Conditions 1
Paths 1

Size

Total Lines 5
Code Lines 2

Duplication

Lines 0
Ratio 0 %

Importance

Changes 0
Metric Value
cc 1
eloc 2
nc 1
nop 2
dl 0
loc 5
rs 10
c 0
b 0
f 0
1
<?php
2
3
/**
4
 * webtrees: online genealogy
5
 * Copyright (C) 2022 webtrees development team
6
 * This program is free software: you can redistribute it and/or modify
7
 * it under the terms of the GNU General Public License as published by
8
 * the Free Software Foundation, either version 3 of the License, or
9
 * (at your option) any later version.
10
 * This program is distributed in the hope that it will be useful,
11
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
 * GNU General Public License for more details.
14
 * You should have received a copy of the GNU General Public License
15
 * along with this program. If not, see <https://www.gnu.org/licenses/>.
16
 */
17
18
declare(strict_types=1);
19
20
namespace Fisharebest\Webtrees;
21
22
use Closure;
23
use Collator;
24
use Exception;
25
use Fisharebest\Localization\Locale;
26
use Fisharebest\Localization\Locale\LocaleEnUs;
27
use Fisharebest\Localization\Locale\LocaleInterface;
28
use Fisharebest\Localization\Translation;
29
use Fisharebest\Localization\Translator;
30
use Fisharebest\Webtrees\Module\ModuleCustomInterface;
31
use Fisharebest\Webtrees\Module\ModuleLanguageInterface;
32
use Fisharebest\Webtrees\Services\ModuleService;
33
34
use function array_merge;
35
use function class_exists;
36
use function html_entity_decode;
37
use function in_array;
38
use function mb_strtolower;
39
use function mb_strtoupper;
40
use function mb_substr;
41
use function ord;
42
use function sprintf;
43
use function str_contains;
44
use function str_replace;
45
use function strcmp;
46
use function strip_tags;
47
use function strlen;
48
use function strtr;
49
use function var_export;
50
51
/**
52
 * Internationalization (i18n) and localization (l10n).
53
 */
54
class I18N
55
{
56
    // MO files use special characters for plurals and context.
57
    public const PLURAL  = "\x00";
58
    public const CONTEXT = "\x04";
59
60
    // Digits are always rendered LTR, even in RTL text.
61
    private const DIGITS = '0123456789٠١٢٣٤٥٦٧٨٩۰۱۲۳۴۵۶۷۸۹';
62
63
    // These locales need special handling for the dotless letter I.
64
    private const DOTLESS_I_LOCALES = [
65
        'az',
66
        'tr',
67
    ];
68
69
    private const DOTLESS_I_TOLOWER = [
70
        'I' => 'ı',
71
        'İ' => 'i',
72
    ];
73
74
    private const DOTLESS_I_TOUPPER = [
75
        'ı' => 'I',
76
        'i' => 'İ',
77
    ];
78
79
    // The ranges of characters used by each script.
80
    private const SCRIPT_CHARACTER_RANGES = [
81
        [
82
            'Latn',
83
            0x0041,
84
            0x005A,
85
        ],
86
        [
87
            'Latn',
88
            0x0061,
89
            0x007A,
90
        ],
91
        [
92
            'Latn',
93
            0x0100,
94
            0x02AF,
95
        ],
96
        [
97
            'Grek',
98
            0x0370,
99
            0x03FF,
100
        ],
101
        [
102
            'Cyrl',
103
            0x0400,
104
            0x052F,
105
        ],
106
        [
107
            'Hebr',
108
            0x0590,
109
            0x05FF,
110
        ],
111
        [
112
            'Arab',
113
            0x0600,
114
            0x06FF,
115
        ],
116
        [
117
            'Arab',
118
            0x0750,
119
            0x077F,
120
        ],
121
        [
122
            'Arab',
123
            0x08A0,
124
            0x08FF,
125
        ],
126
        [
127
            'Deva',
128
            0x0900,
129
            0x097F,
130
        ],
131
        [
132
            'Taml',
133
            0x0B80,
134
            0x0BFF,
135
        ],
136
        [
137
            'Sinh',
138
            0x0D80,
139
            0x0DFF,
140
        ],
141
        [
142
            'Thai',
143
            0x0E00,
144
            0x0E7F,
145
        ],
146
        [
147
            'Geor',
148
            0x10A0,
149
            0x10FF,
150
        ],
151
        [
152
            'Grek',
153
            0x1F00,
154
            0x1FFF,
155
        ],
156
        [
157
            'Deva',
158
            0xA8E0,
159
            0xA8FF,
160
        ],
161
        [
162
            'Hans',
163
            0x3000,
164
            0x303F,
165
        ],
166
        // Mixed CJK, not just Hans
167
        [
168
            'Hans',
169
            0x3400,
170
            0xFAFF,
171
        ],
172
        // Mixed CJK, not just Hans
173
        [
174
            'Hans',
175
            0x20000,
176
            0x2FA1F,
177
        ],
178
        // Mixed CJK, not just Hans
179
    ];
180
181
    // Characters that are displayed in mirror form in RTL text.
182
    private const MIRROR_CHARACTERS = [
183
        '('  => ')',
184
        ')'  => '(',
185
        '['  => ']',
186
        ']'  => '[',
187
        '{'  => '}',
188
        '}'  => '{',
189
        '<'  => '>',
190
        '>'  => '<',
191
        '‹ ' => '›',
192
        '› ' => '‹',
193
        '«'  => '»',
194
        '»'  => '«',
195
        '﴾ ' => '﴿',
196
        '﴿ ' => '﴾',
197
        '“ ' => '”',
198
        '” ' => '“',
199
        '‘ ' => '’',
200
        '’ ' => '‘',
201
    ];
202
203
    // Punctuation used to separate list items, typically a comma
204
    public static string $list_separator;
205
206
    private static ModuleLanguageInterface $language;
207
208
    private static LocaleInterface $locale;
209
210
    private static Translator $translator;
211
212
    private static ?Collator $collator = null;
213
214
    /**
215
     * The preferred locales for this site, or a default list if no preference.
216
     *
217
     * @return array<LocaleInterface>
218
     */
219
    public static function activeLocales(): array
220
    {
221
        $locales = app(ModuleService::class)
222
            ->findByInterface(ModuleLanguageInterface::class, false, true)
223
            ->map(static function (ModuleLanguageInterface $module): LocaleInterface {
224
                return $module->locale();
225
            });
226
227
        if ($locales->isEmpty()) {
228
            return [new LocaleEnUs()];
229
        }
230
231
        return $locales->all();
232
    }
233
234
    /**
235
     * What format is used to display dates in the current locale?
236
     *
237
     * @return string
238
     */
239
    public static function dateFormat(): string
240
    {
241
        /* I18N: This is the format string for full dates. See https://php.net/date for codes */
242
        return self::$translator->translate('%j %F %Y');
243
    }
244
245
    /**
246
     * Convert the digits 0-9 into the local script
247
     * Used for years, etc., where we do not want thousands-separators, decimals, etc.
248
     *
249
     * @param string|int $n
250
     *
251
     * @return string
252
     */
253
    public static function digits($n): string
254
    {
255
        return self::$locale->digits((string) $n);
256
    }
257
258
    /**
259
     * What is the direction of the current locale
260
     *
261
     * @return string "ltr" or "rtl"
262
     */
263
    public static function direction(): string
264
    {
265
        return self::$locale->direction();
266
    }
267
268
    /**
269
     * Initialise the translation adapter with a locale setting.
270
     *
271
     * @param string $code
272
     * @param bool   $setup
273
     *
274
     * @return void
275
     */
276
    public static function init(string $code, bool $setup = false): void
277
    {
278
        self::$locale = Locale::create($code);
279
280
        // Load the translation file
281
        $translation_file = __DIR__ . '/../resources/lang/' . self::$locale->languageTag() . '/messages.php';
282
283
        try {
284
            $translation  = new Translation($translation_file);
285
            $translations = $translation->asArray();
286
        } catch (Exception $ex) {
287
            // The translations files are created during the build process, and are
288
            // not included in the source code.
289
            // Assuming we are using dev code, and build (or rebuild) the files.
290
            $po_file      = Webtrees::ROOT_DIR . 'resources/lang/' . self::$locale->languageTag() . '/messages.po';
291
            $translation  = new Translation($po_file);
292
            $translations = $translation->asArray();
293
            file_put_contents($translation_file, "<?php\n\nreturn " . var_export($translations, true) . ";\n");
294
        }
295
296
        // Add translations from custom modules (but not during setup, as we have no database/modules)
297
        if (!$setup) {
298
            $module_service = app(ModuleService::class);
299
300
            $translations = $module_service
301
                ->findByInterface(ModuleCustomInterface::class)
302
                ->reduce(static function (array $carry, ModuleCustomInterface $item): array {
303
                    return array_merge($carry, $item->customTranslations(self::$locale->languageTag()));
304
                }, $translations);
305
306
            self::$language = $module_service
307
                ->findByInterface(ModuleLanguageInterface::class, true)
308
                ->first(fn (ModuleLanguageInterface $module): bool => $module->locale()->languageTag() === $code);
309
        }
310
311
        // Create a translator
312
        self::$translator = new Translator($translations, self::$locale->pluralRule());
313
314
        /* I18N: This punctuation is used to separate lists of items */
315
        self::$list_separator = self::translate(', ');
316
317
        // Create a collator
318
        try {
319
            // Symfony provides a very incomplete polyfill - which cannot be used.
320
            if (class_exists('Collator')) {
321
                // Need phonebook collation rules for German Ä, Ö and Ü.
322
                if (str_contains(self::$locale->code(), '@')) {
323
                    self::$collator = new Collator(self::$locale->code() . ';collation=phonebook');
324
                } else {
325
                    self::$collator = new Collator(self::$locale->code() . '@collation=phonebook');
326
                }
327
                // Ignore upper/lower case differences
328
                self::$collator->setStrength(Collator::SECONDARY);
329
            }
330
        } catch (Exception $ex) {
331
            // PHP-INTL is not installed?  We'll use a fallback later.
332
        }
333
    }
334
335
    /**
336
     * Translate a string, and then substitute placeholders
337
     * echo I18N::translate('Hello World!');
338
     * echo I18N::translate('The %s sat on the mat', 'cat');
339
     *
340
     * @param string $message
341
     * @param string ...$args
342
     *
343
     * @return string
344
     */
345
    public static function translate(string $message, ...$args): string
346
    {
347
        $message = self::$translator->translate($message);
348
349
        return sprintf($message, ...$args);
350
    }
351
352
    /**
353
     * @return string
354
     */
355
    public static function languageTag(): string
356
    {
357
        return self::$locale->languageTag();
358
    }
359
360
    /**
361
     * @return LocaleInterface
362
     */
363
    public static function locale(): LocaleInterface
364
    {
365
        return self::$locale;
366
    }
367
368
    /**
369
     * @return ModuleLanguageInterface
370
     */
371
    public static function language(): ModuleLanguageInterface
372
    {
373
        return self::$language;
374
    }
375
376
    /**
377
     * Translate a number into the local representation.
378
     * e.g. 12345.67 becomes
379
     * en: 12,345.67
380
     * fr: 12 345,67
381
     * de: 12.345,67
382
     *
383
     * @param float $n
384
     * @param int   $precision
385
     *
386
     * @return string
387
     */
388
    public static function number(float $n, int $precision = 0): string
389
    {
390
        return self::$locale->number(round($n, $precision));
391
    }
392
393
    /**
394
     * Translate a fraction into a percentage.
395
     * e.g. 0.123 becomes
396
     * en: 12.3%
397
     * fr: 12,3 %
398
     * de: 12,3%
399
     *
400
     * @param float $n
401
     * @param int   $precision
402
     *
403
     * @return string
404
     */
405
    public static function percentage(float $n, int $precision = 0): string
406
    {
407
        return self::$locale->percent(round($n, $precision + 2));
408
    }
409
410
    /**
411
     * Translate a plural string
412
     * echo self::plural('There is an error', 'There are errors', $num_errors);
413
     * echo self::plural('There is one error', 'There are %s errors', $num_errors);
414
     * echo self::plural('There is %1$s %2$s cat', 'There are %1$s %2$s cats', $num, $num, $colour);
415
     *
416
     * @param string $singular
417
     * @param string $plural
418
     * @param int    $count
419
     * @param string ...$args
420
     *
421
     * @return string
422
     */
423
    public static function plural(string $singular, string $plural, int $count, ...$args): string
424
    {
425
        $message = self::$translator->translatePlural($singular, $plural, $count);
426
427
        return sprintf($message, ...$args);
428
    }
429
430
    /**
431
     * UTF8 version of PHP::strrev()
432
     * Reverse RTL text for third-party libraries such as GD2 and googlechart.
433
     * These do not support UTF8 text direction, so we must mimic it for them.
434
     * Numbers are always rendered LTR, even in RTL text.
435
     * The visual direction of characters such as parentheses should be reversed.
436
     *
437
     * @param string $text Text to be reversed
438
     *
439
     * @return string
440
     */
441
    public static function reverseText(string $text): string
442
    {
443
        // Remove HTML markup - we can't display it and it is LTR.
444
        $text = strip_tags($text);
445
        // Remove HTML entities.
446
        $text = html_entity_decode($text, ENT_QUOTES, 'UTF-8');
447
448
        // LTR text doesn't need reversing
449
        if (self::scriptDirection(self::textScript($text)) === 'ltr') {
450
            return $text;
451
        }
452
453
        // Mirrored characters
454
        $text = strtr($text, self::MIRROR_CHARACTERS);
455
456
        $reversed = '';
457
        $digits   = '';
458
        while ($text !== '') {
459
            $letter = mb_substr($text, 0, 1);
460
            $text   = mb_substr($text, 1);
461
            if (str_contains(self::DIGITS, $letter)) {
462
                $digits .= $letter;
463
            } else {
464
                $reversed = $letter . $digits . $reversed;
465
                $digits   = '';
466
            }
467
        }
468
469
        return $digits . $reversed;
470
    }
471
472
    /**
473
     * Return the direction (ltr or rtl) for a given script
474
     * The PHP/intl library does not provde this information, so we need
475
     * our own lookup table.
476
     *
477
     * @param string $script
478
     *
479
     * @return string
480
     */
481
    public static function scriptDirection(string $script): string
482
    {
483
        switch ($script) {
484
            case 'Arab':
485
            case 'Hebr':
486
            case 'Mong':
487
            case 'Thaa':
488
                return 'rtl';
489
            default:
490
                return 'ltr';
491
        }
492
    }
493
494
    /**
495
     * Identify the script used for a piece of text
496
     *
497
     * @param string $string
498
     *
499
     * @return string
500
     */
501
    public static function textScript(string $string): string
502
    {
503
        $string = strip_tags($string); // otherwise HTML tags show up as latin
504
        $string = html_entity_decode($string, ENT_QUOTES, 'UTF-8'); // otherwise HTML entities show up as latin
505
        $string = str_replace([
506
            Individual::NOMEN_NESCIO,
507
            Individual::PRAENOMEN_NESCIO,
508
        ], '', $string);
509
        $pos    = 0;
510
        $strlen = strlen($string);
511
        while ($pos < $strlen) {
512
            // get the Unicode Code Point for the character at position $pos
513
            $byte1 = ord($string[$pos]);
514
            if ($byte1 < 0x80) {
515
                $code_point = $byte1;
516
                $chrlen     = 1;
517
            } elseif ($byte1 < 0xC0) {
518
                // Invalid continuation character
519
                return 'Latn';
520
            } elseif ($byte1 < 0xE0) {
521
                $code_point = (($byte1 & 0x1F) << 6) + (ord($string[$pos + 1]) & 0x3F);
522
                $chrlen     = 2;
523
            } elseif ($byte1 < 0xF0) {
524
                $code_point = (($byte1 & 0x0F) << 12) + ((ord($string[$pos + 1]) & 0x3F) << 6) + (ord($string[$pos + 2]) & 0x3F);
525
                $chrlen     = 3;
526
            } elseif ($byte1 < 0xF8) {
527
                $code_point = (($byte1 & 0x07) << 24) + ((ord($string[$pos + 1]) & 0x3F) << 12) + ((ord($string[$pos + 2]) & 0x3F) << 6) + (ord($string[$pos + 3]) & 0x3F);
528
                $chrlen     = 3;
529
            } else {
530
                // Invalid UTF
531
                return 'Latn';
532
            }
533
534
            foreach (self::SCRIPT_CHARACTER_RANGES as $range) {
535
                if ($code_point >= $range[1] && $code_point <= $range[2]) {
536
                    return $range[0];
537
                }
538
            }
539
            // Not a recognised script. Maybe punctuation, spacing, etc. Keep looking.
540
            $pos += $chrlen;
541
        }
542
543
        return 'Latn';
544
    }
545
546
    /**
547
     * A closure which will compare strings using local collation rules.
548
     *
549
     * @return Closure
550
     */
551
    public static function comparator(): Closure
552
    {
553
        $collator = self::$collator;
554
555
        if ($collator instanceof Collator) {
556
            return static fn (string $x, string $y): int => (int) $collator->compare($x, $y);
557
        }
558
559
        return static fn (string $x, string $y): int => strcmp(self::strtolower($x), self::strtolower($y));
560
    }
561
562
563
564
    /**
565
     * Convert a string to lower case.
566
     *
567
     * @param string $string
568
     *
569
     * @return string
570
     */
571
    public static function strtolower(string $string): string
572
    {
573
        if (in_array(self::$locale->language()->code(), self::DOTLESS_I_LOCALES, true)) {
574
            $string = strtr($string, self::DOTLESS_I_TOLOWER);
575
        }
576
577
        return mb_strtolower($string);
578
    }
579
580
    /**
581
     * Convert a string to upper case.
582
     *
583
     * @param string $string
584
     *
585
     * @return string
586
     */
587
    public static function strtoupper(string $string): string
588
    {
589
        if (in_array(self::$locale->language()->code(), self::DOTLESS_I_LOCALES, true)) {
590
            $string = strtr($string, self::DOTLESS_I_TOUPPER);
591
        }
592
593
        return mb_strtoupper($string);
594
    }
595
596
    /**
597
     * What format is used to display dates in the current locale?
598
     *
599
     * @return string
600
     */
601
    public static function timeFormat(): string
602
    {
603
        /* I18N: This is the format string for the time-of-day. See https://php.net/date for codes */
604
        return self::$translator->translate('%H:%i:%s');
605
    }
606
607
    /**
608
     * Context sensitive version of translate.
609
     * echo I18N::translateContext('NOMINATIVE', 'January');
610
     * echo I18N::translateContext('GENITIVE', 'January');
611
     *
612
     * @param string $context
613
     * @param string $message
614
     * @param string ...$args
615
     *
616
     * @return string
617
     */
618
    public static function translateContext(string $context, string $message, ...$args): string
619
    {
620
        $message = self::$translator->translateContext($context, $message);
621
622
        return sprintf($message, ...$args);
623
    }
624
}
625