MessageFormatter.php 16 KB
Newer Older
1 2 3 4 5 6 7 8 9
<?php
/**
 * @link http://www.yiiframework.com/
 * @copyright Copyright (c) 2008 Yii Software LLC
 * @license http://www.yiiframework.com/license/
 */

namespace yii\i18n;

10 11
use yii\base\Component;
use yii\base\NotSupportedException;
12

13
/**
14 15 16
 * MessageFormatter allows formatting messages via [ICU message format](http://userguide.icu-project.org/formatparse/messages)
 *
 * This class enhances the message formatter class provided by the PHP intl extension.
Qiang Xue committed
17 18
 *
 * The following enhancements are provided:
19
 *
20
 * - It accepts named arguments and mixed numeric and named arguments.
21 22
 * - Issues no error when an insufficient number of arguments have been provided. Instead, the placeholders will not be
 *   substituted.
Qiang Xue committed
23
 * - Fixes PHP 5.5 weird placeholder replacement in case no arguments are provided at all (https://bugs.php.net/bug.php?id=65920).
24 25 26
 * - Offers limited support for message formatting in case PHP intl extension is not installed.
 *   However it is highly recommended that you install [PHP intl extension](http://php.net/manual/en/book.intl.php) if you want
 *   to use MessageFormatter features.
27
 *
28
 *   The fallback implementation only supports the following message formats:
29
 *   - plural formatting for english ('one' and 'other' selectors)
30 31 32
 *   - select format
 *   - simple parameters
 *   - integer number parameters
33
 *
34 35 36
 *   The fallback implementation does NOT support the ['apostrophe-friendly' syntax](http://www.php.net/manual/en/messageformatter.formatmessage.php).
 *   Also messages that are working with the fallback implementation are not necessarily compatible with the
 *   PHP intl MessageFormatter so do not rely on the fallback if you are able to install intl extension somehow.
37
 *
38 39
 * @property string $errorCode Code of the last error. This property is read-only.
 * @property string $errorMessage Description of the last error. This property is read-only.
40
 *
41
 * @author Alexander Makarov <sam@rmcreative.ru>
42
 * @author Carsten Brandt <mail@cebe.cc>
43 44
 * @since 2.0
 */
45
class MessageFormatter extends Component
46
{
47 48
    private $_errorCode = 0;
    private $_errorMessage = '';
49

50

51 52 53 54 55 56 57 58 59
    /**
     * Get the error code from the last operation
     * @link http://php.net/manual/en/messageformatter.geterrorcode.php
     * @return string Code of the last error.
     */
    public function getErrorCode()
    {
        return $this->_errorCode;
    }
60

61 62 63 64 65 66 67 68 69
    /**
     * Get the error text from the last operation
     * @link http://php.net/manual/en/messageformatter.geterrormessage.php
     * @return string Description of the last error.
     */
    public function getErrorMessage()
    {
        return $this->_errorMessage;
    }
70

71 72 73 74 75 76 77
    /**
     * Formats a message via [ICU message format](http://userguide.icu-project.org/formatparse/messages)
     *
     * It uses the PHP intl extension's [MessageFormatter](http://www.php.net/manual/en/class.messageformatter.php)
     * and works around some issues.
     * If PHP intl is not installed a fallback will be used that supports a subset of the ICU message format.
     *
78 79 80
     * @param string $pattern The pattern string to insert parameters into.
     * @param array $params The array of name value pairs to insert into the format string.
     * @param string $language The locale to use for formatting locale-dependent parts
81 82 83 84 85 86
     * @return string|boolean The formatted pattern string or `FALSE` if an error occurred
     */
    public function format($pattern, $params, $language)
    {
        $this->_errorCode = 0;
        $this->_errorMessage = '';
87

88 89 90
        if ($params === []) {
            return $pattern;
        }
91

92 93 94
        if (!class_exists('MessageFormatter', false)) {
            return $this->fallbackFormat($pattern, $params, $language);
        }
95

96 97 98 99 100
        if (version_compare(PHP_VERSION, '5.5.0', '<') || version_compare(INTL_ICU_VERSION, '4.8', '<')) {
            // replace named arguments
            $pattern = $this->replaceNamedArguments($pattern, $params, $newParams);
            $params = $newParams;
        }
101

102 103 104 105
        $formatter = new \MessageFormatter($language, $pattern);
        if ($formatter === null) {
            $this->_errorCode = intl_get_error_code();
            $this->_errorMessage = "Message pattern is invalid: " . intl_get_error_message();
106

107 108 109 110 111 112
            return false;
        }
        $result = $formatter->format($params);
        if ($result === false) {
            $this->_errorCode = $formatter->getErrorCode();
            $this->_errorMessage = $formatter->getErrorMessage();
113

114 115 116 117 118
            return false;
        } else {
            return $result;
        }
    }
119

120 121 122 123 124 125 126
    /**
     * Parses an input string according to an [ICU message format](http://userguide.icu-project.org/formatparse/messages) pattern.
     *
     * It uses the PHP intl extension's [MessageFormatter::parse()](http://www.php.net/manual/en/messageformatter.parsemessage.php)
     * and adds support for named arguments.
     * Usage of this method requires PHP intl extension to be installed.
     *
127 128 129 130
     * @param string $pattern The pattern to use for parsing the message.
     * @param string $message The message to parse, conforming to the pattern.
     * @param string $language The locale to use for formatting locale-dependent parts
     * @return array|boolean An array containing items extracted, or `FALSE` on error.
131 132 133 134 135 136
     * @throws \yii\base\NotSupportedException when PHP intl extension is not installed.
     */
    public function parse($pattern, $message, $language)
    {
        $this->_errorCode = 0;
        $this->_errorMessage = '';
137

138 139 140
        if (!class_exists('MessageFormatter', false)) {
            throw new NotSupportedException('You have to install PHP intl extension to use this feature.');
        }
141

142 143 144 145
        // replace named arguments
        if (($tokens = self::tokenizePattern($pattern)) === false) {
            $this->_errorCode = -1;
            $this->_errorMessage = "Message pattern is invalid.";
146

147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
            return false;
        }
        $map = [];
        foreach ($tokens as $i => $token) {
            if (is_array($token)) {
                $param = trim($token[0]);
                if (!isset($map[$param])) {
                    $map[$param] = count($map);
                }
                $token[0] = $map[$param];
                $tokens[$i] = '{' . implode(',', $token) . '}';
            }
        }
        $pattern = implode('', $tokens);
        $map = array_flip($map);
162

163 164 165 166
        $formatter = new \MessageFormatter($language, $pattern);
        if ($formatter === null) {
            $this->_errorCode = -1;
            $this->_errorMessage = "Message pattern is invalid.";
167

168 169 170 171 172 173
            return false;
        }
        $result = $formatter->parse($message);
        if ($result === false) {
            $this->_errorCode = $formatter->getErrorCode();
            $this->_errorMessage = $formatter->getErrorMessage();
174

175 176 177 178 179 180 181 182 183 184 185 186 187 188
            return false;
        } else {
            $values = [];
            foreach ($result as $key => $value) {
                $values[$map[$key]] = $value;
            }

            return $values;
        }
    }

    /**
     * Replace named placeholders with numeric placeholders and quote unused.
     *
189 190 191 192
     * @param string $pattern The pattern string to replace things into.
     * @param array $givenParams The array of values to insert into the format string.
     * @param array $resultingParams Modified array of parameters.
     * @param array $map
193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223
     * @return string The pattern string with placeholders replaced.
     */
    private function replaceNamedArguments($pattern, $givenParams, &$resultingParams, &$map = [])
    {
        if (($tokens = self::tokenizePattern($pattern)) === false) {
            return false;
        }
        foreach ($tokens as $i => $token) {
            if (!is_array($token)) {
                continue;
            }
            $param = trim($token[0]);
            if (isset($givenParams[$param])) {
                // if param is given, replace it with a number
                if (!isset($map[$param])) {
                    $map[$param] = count($map);
                    // make sure only used params are passed to format method
                    $resultingParams[$map[$param]] = $givenParams[$param];
                }
                $token[0] = $map[$param];
                $quote = "";
            } else {
                // quote unused token
                $quote = "'";
            }
            $type = isset($token[1]) ? trim($token[1]) : 'none';
            // replace plural and select format recursively
            if ($type == 'plural' || $type == 'select') {
                if (!isset($token[2])) {
                    return false;
                }
224 225 226
                if (($subtokens = self::tokenizePattern($token[2])) === false) {
                    return false;
                }
227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244
                $c = count($subtokens);
                for ($k = 0; $k + 1 < $c; $k++) {
                    if (is_array($subtokens[$k]) || !is_array($subtokens[++$k])) {
                        return false;
                    }
                    $subpattern = $this->replaceNamedArguments(implode(',', $subtokens[$k]), $givenParams, $resultingParams, $map);
                    $subtokens[$k] = $quote . '{' . $quote . $subpattern . $quote . '}' . $quote;
                }
                $token[2] = implode('', $subtokens);
            }
            $tokens[$i] = $quote . '{' . $quote . implode(',', $token) . $quote . '}' . $quote;
        }

        return implode('', $tokens);
    }

    /**
     * Fallback implementation for MessageFormatter::formatMessage
245 246 247
     * @param string $pattern The pattern string to insert things into.
     * @param array $args The array of values to insert into the format string
     * @param string $locale The locale to use for formatting locale-dependent parts
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
     * @return string|boolean The formatted pattern string or `FALSE` if an error occurred
     */
    protected function fallbackFormat($pattern, $args, $locale)
    {
        if (($tokens = self::tokenizePattern($pattern)) === false) {
            $this->_errorCode = -1;
            $this->_errorMessage = "Message pattern is invalid.";

            return false;
        }
        foreach ($tokens as $i => $token) {
            if (is_array($token)) {
                if (($tokens[$i] = $this->parseToken($token, $args, $locale)) === false) {
                    $this->_errorCode = -1;
                    $this->_errorMessage = "Message pattern is invalid.";

                    return false;
                }
            }
        }

        return implode('', $tokens);
    }

    /**
     * Tokenizes a pattern by separating normal text from replaceable patterns
274
     * @param string $pattern patter to tokenize
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
     * @return array|boolean array of tokens or false on failure
     */
    private static function tokenizePattern($pattern)
    {
        $depth = 1;
        if (($start = $pos = mb_strpos($pattern, '{')) === false) {
            return [$pattern];
        }
        $tokens = [mb_substr($pattern, 0, $pos)];
        while (true) {
            $open = mb_strpos($pattern, '{', $pos + 1);
            $close = mb_strpos($pattern, '}', $pos + 1);
            if ($open === false && $close === false) {
                break;
            }
            if ($open === false) {
                $open = mb_strlen($pattern);
            }
            if ($close > $open) {
                $depth++;
                $pos = $open;
            } else {
                $depth--;
                $pos = $close;
            }
            if ($depth == 0) {
                $tokens[] = explode(',', mb_substr($pattern, $start + 1, $pos - $start - 1), 3);
                $start = $pos + 1;
                $tokens[] = mb_substr($pattern, $start, $open - $start);
                $start = $open;
            }
        }
        if ($depth != 0) {
            return false;
        }

        return $tokens;
    }

    /**
     * Parses a token
316 317 318 319
     * @param array $token the token to parse
     * @param array $args arguments to replace
     * @param string $locale the locale
     * @return bool|string parsed token or false on failure
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347
     * @throws \yii\base\NotSupportedException when unsupported formatting is used.
     */
    private function parseToken($token, $args, $locale)
    {
        // parsing pattern based on ICU grammar:
        // http://icu-project.org/apiref/icu4c/classMessageFormat.html#details

        $param = trim($token[0]);
        if (isset($args[$param])) {
            $arg = $args[$param];
        } else {
            return '{' . implode(',', $token) . '}';
        }
        $type = isset($token[1]) ? trim($token[1]) : 'none';
        switch ($type) {
            case 'date':
            case 'time':
            case 'spellout':
            case 'ordinal':
            case 'duration':
            case 'choice':
            case 'selectordinal':
                throw new NotSupportedException("Message format '$type' is not supported. You have to install PHP intl extension to use this feature.");
            case 'number':
                if (is_int($arg) && (!isset($token[2]) || trim($token[2]) == 'integer')) {
                    return $arg;
                }
                throw new NotSupportedException("Message format 'number' is only supported for integer values. You have to install PHP intl extension to use this feature.");
348 349
            case 'none':
                return $arg;
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393
            case 'select':
                /* http://icu-project.org/apiref/icu4c/classicu_1_1SelectFormat.html
                selectStyle = (selector '{' message '}')+
                */
                if (!isset($token[2])) {
                    return false;
                }
                $select = self::tokenizePattern($token[2]);
                $c = count($select);
                $message = false;
                for ($i = 0; $i + 1 < $c; $i++) {
                    if (is_array($select[$i]) || !is_array($select[$i + 1])) {
                        return false;
                    }
                    $selector = trim($select[$i++]);
                    if ($message === false && $selector == 'other' || $selector == $arg) {
                        $message = implode(',', $select[$i]);
                    }
                }
                if ($message !== false) {
                    return $this->fallbackFormat($message, $args, $locale);
                }
            break;
            case 'plural':
                /* http://icu-project.org/apiref/icu4c/classicu_1_1PluralFormat.html
                pluralStyle = [offsetValue] (selector '{' message '}')+
                offsetValue = "offset:" number
                selector = explicitValue | keyword
                explicitValue = '=' number  // adjacent, no white space in between
                keyword = [^[[:Pattern_Syntax:][:Pattern_White_Space:]]]+
                message: see MessageFormat
                */
                if (!isset($token[2])) {
                    return false;
                }
                $plural = self::tokenizePattern($token[2]);
                $c = count($plural);
                $message = false;
                $offset = 0;
                for ($i = 0; $i + 1 < $c; $i++) {
                    if (is_array($plural[$i]) || !is_array($plural[$i + 1])) {
                        return false;
                    }
                    $selector = trim($plural[$i++]);
394

395
                    if ($i == 1 && strncmp($selector, 'offset:', 7) === 0) {
396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
                        $offset = (int) trim(mb_substr($selector, 7, ($pos = mb_strpos(str_replace(["\n", "\r", "\t"], ' ', $selector), ' ', 7)) - 7));
                        $selector = trim(mb_substr($selector, $pos + 1));
                    }
                    if ($message === false && $selector == 'other' ||
                        $selector[0] == '=' && (int) mb_substr($selector, 1) == $arg ||
                        $selector == 'one' && $arg - $offset == 1
                    ) {
                        $message = implode(',', str_replace('#', $arg - $offset, $plural[$i]));
                    }
                }
                if ($message !== false) {
                    return $this->fallbackFormat($message, $args, $locale);
                }
                break;
        }

        return false;
    }
414
}