1 /* 2 ******************************************************************************* 3 * Copyright (C) 2004-2016, International Business Machines Corporation and * 4 * others. All Rights Reserved. * 5 * Copyright (C) 2009 , Yahoo! Inc. * 6 ******************************************************************************* 7 */ 8 package com.ibm.icu.text; 9 10 import java.io.IOException; 11 import java.io.ObjectInputStream; 12 import java.text.FieldPosition; 13 import java.text.Format; 14 import java.text.ParsePosition; 15 16 import com.ibm.icu.impl.PatternProps; 17 18 /** 19 * <p><code>SelectFormat</code> supports the creation of internationalized 20 * messages by selecting phrases based on keywords. The pattern specifies 21 * how to map keywords to phrases and provides a default phrase. The 22 * object provided to the format method is a string that's matched 23 * against the keywords. If there is a match, the corresponding phrase 24 * is selected; otherwise, the default phrase is used. 25 * 26 * <h3>Using <code>SelectFormat</code> for Gender Agreement</h3> 27 * 28 * <p>Note: Typically, select formatting is done via <code>MessageFormat</code> 29 * with a <code>select</code> argument type, 30 * rather than using a stand-alone <code>SelectFormat</code>. 31 * 32 * <p>The main use case for the select format is gender based inflection. 33 * When names or nouns are inserted into sentences, their gender can affect pronouns, 34 * verb forms, articles, and adjectives. Special care needs to be 35 * taken for the case where the gender cannot be determined. 36 * The impact varies between languages: 37 * 38 * <ul> 39 * <li>English has three genders, and unknown gender is handled as a special 40 * case. Names use the gender of the named person (if known), nouns referring 41 * to people use natural gender, and inanimate objects are usually neutral. 42 * The gender only affects pronouns: "he", "she", "it", "they". 43 * 44 * <li>German differs from English in that the gender of nouns is rather 45 * arbitrary, even for nouns referring to people ("Mädchen", girl, is neutral). 46 * The gender affects pronouns ("er", "sie", "es"), articles ("der", "die", 47 * "das"), and adjective forms ("guter Mann", "gute Frau", "gutes Mädchen"). 48 * 49 * <li>French has only two genders; as in German the gender of nouns 50 * is rather arbitrary - for sun and moon, the genders 51 * are the opposite of those in German. The gender affects 52 * pronouns ("il", "elle"), articles ("le", "la"), 53 * adjective forms ("bon", "bonne"), and sometimes 54 * verb forms ("allé", "allée"). 55 * 56 * <li>Polish distinguishes five genders (or noun classes), 57 * human masculine, animate non-human masculine, inanimate masculine, 58 * feminine, and neuter. 59 * </ul> 60 * 61 * <p>Some other languages have noun classes that are not related to gender, 62 * but similar in grammatical use. 63 * Some African languages have around 20 noun classes. 64 * 65 * <p><b>Note:</b>For the gender of a <i>person</i> in a given sentence, 66 * we usually need to distinguish only between female, male and other/unknown. 67 * 68 * <p>To enable localizers to create sentence patterns that take their 69 * language's gender dependencies into consideration, software has to provide 70 * information about the gender associated with a noun or name to 71 * <code>MessageFormat</code>. 72 * Two main cases can be distinguished: 73 * 74 * <ul> 75 * <li>For people, natural gender information should be maintained for each person. 76 * Keywords like "male", "female", "mixed" (for groups of people) 77 * and "unknown" could be used. 78 * 79 * <li>For nouns, grammatical gender information should be maintained for 80 * each noun and per language, e.g., in resource bundles. 81 * The keywords "masculine", "feminine", and "neuter" are commonly used, 82 * but some languages may require other keywords. 83 * </ul> 84 * 85 * <p>The resulting keyword is provided to <code>MessageFormat</code> as a 86 * parameter separate from the name or noun it's associated with. For example, 87 * to generate a message such as "Jean went to Paris", three separate arguments 88 * would be provided: The name of the person as argument 0, the gender of 89 * the person as argument 1, and the name of the city as argument 2. 90 * The sentence pattern for English, where the gender of the person has 91 * no impact on this simple sentence, would not refer to argument 1 at all: 92 * 93 * <pre>{0} went to {2}.</pre> 94 * 95 * <p><b>Note:</b> The entire sentence should be included (and partially repeated) 96 * inside each phrase. Otherwise translators would have to be trained on how to 97 * move bits of the sentence in and out of the select argument of a message. 98 * (The examples below do not follow this recommendation!) 99 * 100 * <p>The sentence pattern for French, where the gender of the person affects 101 * the form of the participle, uses a select format based on argument 1: 102 * 103 * <pre>{0} est {1, select, female {allée} other {allé}} à {2}.</pre> 104 * 105 * <p>Patterns can be nested, so that it's possible to handle interactions of 106 * number and gender where necessary. For example, if the above sentence should 107 * allow for the names of several people to be inserted, the following sentence 108 * pattern can be used (with argument 0 the list of people's names, 109 * argument 1 the number of people, argument 2 their combined gender, and 110 * argument 3 the city name): 111 * 112 * <pre>{0} {1, plural, 113 * one {est {2, select, female {allée} other {allé}}} 114 * other {sont {2, select, female {allées} other {allés}}} 115 * }à {3}.</pre> 116 * 117 * <h4>Patterns and Their Interpretation</h4> 118 * 119 * <p>The <code>SelectFormat</code> pattern string defines the phrase output 120 * for each user-defined keyword. 121 * The pattern is a sequence of (keyword, message) pairs. 122 * A keyword is a "pattern identifier": [^[[:Pattern_Syntax:][:Pattern_White_Space:]]]+ 123 * 124 * <p>Each message is a MessageFormat pattern string enclosed in {curly braces}. 125 * 126 * <p>You always have to define a phrase for the default keyword 127 * <code>other</code>; this phrase is returned when the keyword 128 * provided to 129 * the <code>format</code> method matches no other keyword. 130 * If a pattern does not provide a phrase for <code>other</code>, the method 131 * it's provided to returns the error <code>U_DEFAULT_KEYWORD_MISSING</code>. 132 * <br> 133 * Pattern_White_Space between keywords and messages is ignored. 134 * Pattern_White_Space within a message is preserved and output. 135 * 136 * <pre>Example: 137 * MessageFormat msgFmt = new MessageFormat("{0} est " + 138 * "{1, select, female {allée} other {allé}} à Paris.", 139 * new ULocale("fr")); 140 * Object args[] = {"Kirti","female"}; 141 * System.out.println(msgFmt.format(args)); 142 * </pre> 143 * <p> 144 * Produces the output:<br> 145 * <code>Kirti est allée à Paris.</code> 146 * 147 * @stable ICU 4.4 148 */ 149 150 public class SelectFormat extends Format{ 151 // Generated by serialver from JDK 1.5 152 private static final long serialVersionUID = 2993154333257524984L; 153 154 /* 155 * The applied pattern string. 156 */ 157 private String pattern = null; 158 159 /** 160 * The MessagePattern which contains the parsed structure of the pattern string. 161 */ 162 transient private MessagePattern msgPattern; 163 164 /** 165 * Creates a new <code>SelectFormat</code> for a given pattern string. 166 * @param pattern the pattern for this <code>SelectFormat</code>. 167 * @stable ICU 4.4 168 */ SelectFormat(String pattern)169 public SelectFormat(String pattern) { 170 applyPattern(pattern); 171 } 172 173 /* 174 * Resets the <code>SelectFormat</code> object. 175 */ reset()176 private void reset() { 177 pattern = null; 178 if(msgPattern != null) { 179 msgPattern.clear(); 180 } 181 } 182 183 /** 184 * Sets the pattern used by this select format. 185 * Patterns and their interpretation are specified in the class description. 186 * 187 * @param pattern the pattern for this select format. 188 * @throws IllegalArgumentException when the pattern is not a valid select format pattern. 189 * @stable ICU 4.4 190 */ applyPattern(String pattern)191 public void applyPattern(String pattern) { 192 this.pattern = pattern; 193 if (msgPattern == null) { 194 msgPattern = new MessagePattern(); 195 } 196 try { 197 msgPattern.parseSelectStyle(pattern); 198 } catch(RuntimeException e) { 199 reset(); 200 throw e; 201 } 202 } 203 204 /** 205 * Returns the pattern for this <code>SelectFormat</code> 206 * 207 * @return the pattern string 208 * @stable ICU 4.4 209 */ toPattern()210 public String toPattern() { 211 return pattern; 212 } 213 214 /** 215 * Finds the SelectFormat sub-message for the given keyword, or the "other" sub-message. 216 * @param pattern A MessagePattern. 217 * @param partIndex the index of the first SelectFormat argument style part. 218 * @param keyword a keyword to be matched to one of the SelectFormat argument's keywords. 219 * @return the sub-message start part index. 220 */ findSubMessage(MessagePattern pattern, int partIndex, String keyword)221 /*package*/ static int findSubMessage(MessagePattern pattern, int partIndex, String keyword) { 222 int count=pattern.countParts(); 223 int msgStart=0; 224 // Iterate over (ARG_SELECTOR, message) pairs until ARG_LIMIT or end of select-only pattern. 225 do { 226 MessagePattern.Part part=pattern.getPart(partIndex++); 227 MessagePattern.Part.Type type=part.getType(); 228 if(type==MessagePattern.Part.Type.ARG_LIMIT) { 229 break; 230 } 231 assert type==MessagePattern.Part.Type.ARG_SELECTOR; 232 // part is an ARG_SELECTOR followed by a message 233 if(pattern.partSubstringMatches(part, keyword)) { 234 // keyword matches 235 return partIndex; 236 } else if(msgStart==0 && pattern.partSubstringMatches(part, "other")) { 237 msgStart=partIndex; 238 } 239 partIndex=pattern.getLimitPartIndex(partIndex); 240 } while(++partIndex<count); 241 return msgStart; 242 } 243 244 /** 245 * Selects the phrase for the given keyword. 246 * 247 * @param keyword a phrase selection keyword. 248 * @return the string containing the formatted select message. 249 * @throws IllegalArgumentException when the given keyword is not a "pattern identifier" 250 * @stable ICU 4.4 251 */ format(String keyword)252 public final String format(String keyword) { 253 //Check for the validity of the keyword 254 if (!PatternProps.isIdentifier(keyword)) { 255 throw new IllegalArgumentException("Invalid formatting argument."); 256 } 257 // If no pattern was applied, throw an exception 258 if (msgPattern == null || msgPattern.countParts() == 0) { 259 throw new IllegalStateException("Invalid format error."); 260 } 261 262 // Get the appropriate sub-message. 263 int msgStart = findSubMessage(msgPattern, 0, keyword); 264 if (!msgPattern.jdkAposMode()) { 265 int msgLimit = msgPattern.getLimitPartIndex(msgStart); 266 return msgPattern.getPatternString().substring(msgPattern.getPart(msgStart).getLimit(), 267 msgPattern.getPatternIndex(msgLimit)); 268 } 269 // JDK compatibility mode: Remove SKIP_SYNTAX. 270 StringBuilder result = null; 271 int prevIndex = msgPattern.getPart(msgStart).getLimit(); 272 for (int i = msgStart;;) { 273 MessagePattern.Part part = msgPattern.getPart(++i); 274 MessagePattern.Part.Type type = part.getType(); 275 int index = part.getIndex(); 276 if (type == MessagePattern.Part.Type.MSG_LIMIT) { 277 if (result == null) { 278 return pattern.substring(prevIndex, index); 279 } else { 280 return result.append(pattern, prevIndex, index).toString(); 281 } 282 } else if (type == MessagePattern.Part.Type.SKIP_SYNTAX) { 283 if (result == null) { 284 result = new StringBuilder(); 285 } 286 result.append(pattern, prevIndex, index); 287 prevIndex = part.getLimit(); 288 } else if (type == MessagePattern.Part.Type.ARG_START) { 289 if (result == null) { 290 result = new StringBuilder(); 291 } 292 result.append(pattern, prevIndex, index); 293 prevIndex = index; 294 i = msgPattern.getLimitPartIndex(i); 295 index = msgPattern.getPart(i).getLimit(); 296 MessagePattern.appendReducedApostrophes(pattern, prevIndex, index, result); 297 prevIndex = index; 298 } 299 } 300 } 301 302 /** 303 * Selects the phrase for the given keyword. 304 * and appends the formatted message to the given <code>StringBuffer</code>. 305 * @param keyword a phrase selection keyword. 306 * @param toAppendTo the selected phrase will be appended to this 307 * <code>StringBuffer</code>. 308 * @param pos will be ignored by this method. 309 * @throws IllegalArgumentException when the given keyword is not a String 310 * or not a "pattern identifier" 311 * @return the string buffer passed in as toAppendTo, with formatted text 312 * appended. 313 * @stable ICU 4.4 314 */ format(Object keyword, StringBuffer toAppendTo, FieldPosition pos)315 public StringBuffer format(Object keyword, StringBuffer toAppendTo, 316 FieldPosition pos) { 317 if (keyword instanceof String) { 318 toAppendTo.append(format( (String)keyword)); 319 }else{ 320 throw new IllegalArgumentException("'" + keyword + "' is not a String"); 321 } 322 return toAppendTo; 323 } 324 325 /** 326 * This method is not supported by <code>SelectFormat</code>. 327 * @param source the string to be parsed. 328 * @param pos defines the position where parsing is to begin, 329 * and upon return, the position where parsing left off. If the position 330 * has not changed upon return, then parsing failed. 331 * @return nothing because this method is not supported. 332 * @throws UnsupportedOperationException thrown always. 333 * @stable ICU 4.4 334 */ parseObject(String source, ParsePosition pos)335 public Object parseObject(String source, ParsePosition pos) { 336 throw new UnsupportedOperationException(); 337 } 338 339 /** 340 * {@inheritDoc} 341 * @stable ICU 4.4 342 */ 343 @Override equals(Object obj)344 public boolean equals(Object obj) { 345 if(this == obj) { 346 return true; 347 } 348 if(obj == null || getClass() != obj.getClass()) { 349 return false; 350 } 351 SelectFormat sf = (SelectFormat) obj; 352 return msgPattern == null ? sf.msgPattern == null : msgPattern.equals(sf.msgPattern); 353 } 354 355 /** 356 * {@inheritDoc} 357 * @stable ICU 4.4 358 */ 359 @Override hashCode()360 public int hashCode() { 361 if (pattern != null) { 362 return pattern.hashCode(); 363 } 364 return 0; 365 } 366 367 /** 368 * {@inheritDoc} 369 * @stable ICU 4.4 370 */ 371 @Override toString()372 public String toString() { 373 return "pattern='" + pattern + "'"; 374 } 375 readObject(ObjectInputStream in)376 private void readObject(ObjectInputStream in) 377 throws IOException, ClassNotFoundException { 378 in.defaultReadObject(); 379 if (pattern != null) { 380 applyPattern(pattern); 381 } 382 } 383 } 384