MessageFormatter.php 12.7 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 29 30 31 32
 *   The fallback implementation only supports the following message formats:
 *   - plural formatting for english
 *   - 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 49
	private $_errorCode = 0;
	private $_errorMessage = '';

50
	/**
51
	 * Get the error code from the last operation
52
	 * @link http://php.net/manual/en/messageformatter.geterrorcode.php
53
	 * @return string Code of the last error.
54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
	 */
	public function getErrorCode()
	{
		return $this->_errorCode;
	}

	/**
	 * 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;
	}

	/**
71 72 73 74 75
	 * 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.
76
	 *
77 78
	 * @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.
79 80
	 * @param string $language The locale to use for formatting locale-dependent parts
	 * @return string|boolean The formatted pattern string or `FALSE` if an error occurred
81
	 */
82
	public function format($pattern, $params, $language)
83
	{
84 85 86 87
		$this->_errorCode = 0;
		$this->_errorMessage = '';

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

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

95
		if (version_compare(PHP_VERSION, '5.5.0', '<') || version_compare(INTL_ICU_VERSION, '4.8', '<')) {
96
			$pattern = $this->replaceNamedArguments($pattern, $params);
97 98
			$params = array_values($params);
		}
99
		$formatter = new \MessageFormatter($language, $pattern);
100
		if ($formatter === null) {
101 102
			$this->_errorCode = -1;
			$this->_errorMessage = "Message pattern is invalid.";
103 104 105 106 107 108 109 110 111
			return false;
		}
		$result = $formatter->format($params);
		if ($result === false) {
			$this->_errorCode = $formatter->getErrorCode();
			$this->_errorMessage = $formatter->getErrorMessage();
			return false;
		} else {
			return $result;
112
		}
113 114 115
	}

	/**
116
	 * Parses an input string according to an [ICU message format](http://userguide.icu-project.org/formatparse/messages) pattern.
117
	 *
118
	 * It uses the PHP intl extension's [MessageFormatter::parse()](http://www.php.net/manual/en/messageformatter.parsemessage.php)
Carsten Brandt committed
119
	 * and adds support for named arguments.
120 121 122 123
	 * Usage of this method requires PHP intl extension to be installed.
	 *
	 * @param string $pattern The pattern to use for parsing the message.
	 * @param string $message The message to parse, conforming to the pattern.
124
	 * @param string $language The locale to use for formatting locale-dependent parts
125
	 * @return array|boolean An array containing items extracted, or `FALSE` on error.
126
	 * @throws \yii\base\NotSupportedException when PHP intl extension is not installed.
127
	 */
128
	public function parse($pattern, $message, $language)
129
	{
130 131 132 133 134
		$this->_errorCode = 0;
		$this->_errorMessage = '';

		if (!class_exists('MessageFormatter', false)) {
			throw new NotSupportedException('You have to install PHP intl extension to use this feature.');
135 136
		}

137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
		// replace named arguments
		if (($tokens = $this->tokenizePattern($pattern)) === false) {
			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);

155 156
		$formatter = new \MessageFormatter($language, $pattern);
		if ($formatter === null) {
157 158
			$this->_errorCode = -1;
			$this->_errorMessage = "Message pattern is invalid.";
159
			return false;
160
		}
161 162 163 164 165 166
		$result = $formatter->parse($message);
		if ($result === false) {
			$this->_errorCode = $formatter->getErrorCode();
			$this->_errorMessage = $formatter->getErrorMessage();
			return false;
		} else {
167 168 169 170 171
			$values = [];
			foreach($result as $key => $value) {
				$values[$map[$key]] = $value;
			}
			return $values;
172
		}
173 174 175
	}

	/**
176
	 * Replace named placeholders with numeric placeholders and quote unused.
177
	 *
Alexander Makarov committed
178
	 * @param string $pattern The pattern string to replace things into.
179 180 181 182 183 184
	 * @param array $args The array of values to insert into the format string.
	 * @return string The pattern string with placeholders replaced.
	 */
	private static function replaceNamedArguments($pattern, $args)
	{
		$map = array_flip(array_keys($args));
185

186
		// parsing pattern based on ICU grammar:
187
		// http://icu-project.org/apiref/icu4c/classMessageFormat.html#details
188 189 190 191
		$parts = explode('{', $pattern);
		$c = count($parts);
		$pattern = $parts[0];
		$d = 0;
Alexander Makarov committed
192
		$stack = [];
193
		for ($i = 1; $i < $c; $i++) {
194
			if (preg_match('~^(\s*)([\d\w]+)(\s*)([},])(\s*)(.*)$~us', $parts[$i], $matches)) {
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
				// if we are not inside a plural or select this is a message
				if (!isset($stack[$d]) || $stack[$d] != 'plural' && $stack[$d] != 'select') {
					$d++;
					// replace normal arg if it is available
					if (isset($map[$matches[2]])) {
						$q = '';
						$pattern .= '{' . $matches[1] . $map[$matches[2]] . $matches[3];
					} else {
						// quote unused args
						$q = ($matches[4] == '}') ? "'" : "";
						$pattern .= "$q{" . $matches[1] . $matches[2] . $matches[3];
					}
					$pattern .= ($term = $matches[4] . $q . $matches[5] . $matches[6]);
					// store type of current level
					$stack[$d] = ($matches[4] == ',') ? substr($matches[6], 0, 6) : 'none';
					// if it's plural or select, the next bracket is NOT begin of a message then!
					if ($stack[$d] == 'plural' || $stack[$d] == 'select') {
						$i++;
						$d -= substr_count($term, '}');
					} else {
						$d -= substr_count($term, '}');
						continue;
					}
218
				}
219
			}
220 221 222 223
			$pattern .= '{' . $parts[$i];
			$d += 1 - substr_count($parts[$i], '}');
		}
		return $pattern;
224
	}
225 226 227

	/**
	 * Fallback implementation for MessageFormatter::formatMessage
228 229 230
	 * @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 $locale The locale to use for formatting locale-dependent parts
231 232
	 * @return string|boolean The formatted pattern string or `FALSE` if an error occurred
	 */
233
	protected function fallbackFormat($pattern, $args, $locale)
234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
	{
		if (($tokens = $this->tokenizePattern($pattern)) === false) {
			return false;
		}
		foreach($tokens as $i => $token) {
			if (is_array($token)) {
				if (($tokens[$i] = $this->parseToken($token, $args, $locale)) === false) {
					return false;
				}
			}
		}
		return implode('', $tokens);
	}

	/**
	 * Tokenizes a pattern by separating normal text from replaceable patterns
	 * @param string $pattern patter to tokenize
	 * @return array|bool array of tokens or false on failure
	 */
	private function tokenizePattern($pattern)
	{
		$depth = 1;
		if (($start = $pos = mb_strpos($pattern, '{')) === false) {
			return [$pattern];
		}
		$tokens = [mb_substr($pattern, 0, $pos)];
260
		while (true) {
261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
			$open = mb_strpos($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
	 * @param array $token the token to parse
	 * @param array $args arguments to replace
	 * @param string $locale the locale
	 * @return bool|string parsed token or false on failure
	 * @throws \yii\base\NotSupportedException when unsupported formatting is used.
	 */
	private function parseToken($token, $args, $locale)
	{
		$param = trim($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 '$type' is not supported. You have to install PHP intl extension to use this feature.");
316 317 318 319 320
			case 'number':
				if (is_int($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.");
321 322 323 324 325 326 327 328
			case 'none': return $arg;
			case 'select':
				/* http://icu-project.org/apiref/icu4c/classicu_1_1SelectFormat.html
				selectStyle = (selector '{' message '}')+
				*/
				$select = static::tokenizePattern($token[2]);
				$c = count($select);
				$message = false;
329
				for ($i = 0; $i + 1 < $c; $i++) {
330 331 332 333 334 335 336 337 338
					if (is_array($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) {
339
					return $this->fallbackFormat($message, $args, $locale);
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
				}
			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
				*/
				$plural = static::tokenizePattern($token[2]);
				$c = count($plural);
				$message = false;
				$offset = 0;
355
				for ($i = 0; $i + 1 < $c; $i++) {
356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
					if (is_array($plural[$i]) || !is_array($plural[$i + 1])) {
						return false;
					}
					$selector = trim($plural[$i++]);
					if ($i == 1 && substr($selector, 0, 7) == 'offset:') {
						$offset = (int) trim(mb_substr($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 == 'zero' && $arg - $offset == 0 ||
						$selector == 'one' && $arg - $offset == 1 ||
						$selector == 'two' && $arg - $offset == 2
					) {
						$message = implode(',', str_replace('#', $arg - $offset, $plural[$i]));
					}
				}
				if ($message !== false) {
374
					return $this->fallbackFormat($message, $args, $locale);
375 376 377 378 379
				}
				break;
		}
		return false;
	}
380
}