1 /* 2 ******************************************************************************* 3 * Copyright (C) 1997-2013, International Business Machines Corporation and others. 4 * All Rights Reserved. 5 ******************************************************************************* 6 */ 7 8 #ifndef RBNF_H 9 #define RBNF_H 10 11 #include "unicode/utypes.h" 12 13 /** 14 * \file 15 * \brief C++ API: Rule Based Number Format 16 */ 17 18 /** 19 * \def U_HAVE_RBNF 20 * This will be 0 if RBNF support is not included in ICU 21 * and 1 if it is. 22 * 23 * @stable ICU 2.4 24 */ 25 #if UCONFIG_NO_FORMATTING 26 #define U_HAVE_RBNF 0 27 #else 28 #define U_HAVE_RBNF 1 29 30 #include "unicode/coll.h" 31 #include "unicode/dcfmtsym.h" 32 #include "unicode/fmtable.h" 33 #include "unicode/locid.h" 34 #include "unicode/numfmt.h" 35 #include "unicode/unistr.h" 36 #include "unicode/strenum.h" 37 38 U_NAMESPACE_BEGIN 39 40 class NFRuleSet; 41 class LocalizationInfo; 42 43 /** 44 * Tags for the predefined rulesets. 45 * 46 * @stable ICU 2.2 47 */ 48 enum URBNFRuleSetTag { 49 URBNF_SPELLOUT, 50 URBNF_ORDINAL, 51 URBNF_DURATION, 52 URBNF_NUMBERING_SYSTEM, 53 URBNF_COUNT 54 }; 55 56 #if UCONFIG_NO_COLLATION 57 class Collator; 58 #endif 59 60 /** 61 * The RuleBasedNumberFormat class formats numbers according to a set of rules. This number formatter is 62 * typically used for spelling out numeric values in words (e.g., 25,3476 as 63 * "twenty-five thousand three hundred seventy-six" or "vingt-cinq mille trois 64 * cents soixante-seize" or 65 * "fünfundzwanzigtausenddreihundertsechsundsiebzig"), but can also be used for 66 * other complicated formatting tasks, such as formatting a number of seconds as hours, 67 * minutes and seconds (e.g., 3,730 as "1:02:10"). 68 * 69 * <p>The resources contain three predefined formatters for each locale: spellout, which 70 * spells out a value in words (123 is "one hundred twenty-three"); ordinal, which 71 * appends an ordinal suffix to the end of a numeral (123 is "123rd"); and 72 * duration, which shows a duration in seconds as hours, minutes, and seconds (123 is 73 * "2:03"). The client can also define more specialized <tt>RuleBasedNumberFormat</tt>s 74 * by supplying programmer-defined rule sets.</p> 75 * 76 * <p>The behavior of a <tt>RuleBasedNumberFormat</tt> is specified by a textual description 77 * that is either passed to the constructor as a <tt>String</tt> or loaded from a resource 78 * bundle. In its simplest form, the description consists of a semicolon-delimited list of <em>rules.</em> 79 * Each rule has a string of output text and a value or range of values it is applicable to. 80 * In a typical spellout rule set, the first twenty rules are the words for the numbers from 81 * 0 to 19:</p> 82 * 83 * <pre>zero; one; two; three; four; five; six; seven; eight; nine; 84 * ten; eleven; twelve; thirteen; fourteen; fifteen; sixteen; seventeen; eighteen; nineteen;</pre> 85 * 86 * <p>For larger numbers, we can use the preceding set of rules to format the ones place, and 87 * we only have to supply the words for the multiples of 10:</p> 88 * 89 * <pre> 20: twenty[->>]; 90 * 30: thirty[->>]; 91 * 40: forty[->>]; 92 * 50: fifty[->>]; 93 * 60: sixty[->>]; 94 * 70: seventy[->>]; 95 * 80: eighty[->>]; 96 * 90: ninety[->>];</pre> 97 * 98 * <p>In these rules, the <em>base value</em> is spelled out explicitly and set off from the 99 * rule's output text with a colon. The rules are in a sorted list, and a rule is applicable 100 * to all numbers from its own base value to one less than the next rule's base value. The 101 * ">>" token is called a <em>substitution</em> and tells the fomatter to 102 * isolate the number's ones digit, format it using this same set of rules, and place the 103 * result at the position of the ">>" token. Text in brackets is omitted if 104 * the number being formatted is an even multiple of 10 (the hyphen is a literal hyphen; 24 105 * is "twenty-four," not "twenty four").</p> 106 * 107 * <p>For even larger numbers, we can actually look up several parts of the number in the 108 * list:</p> 109 * 110 * <pre>100: << hundred[ >>];</pre> 111 * 112 * <p>The "<<" represents a new kind of substitution. The << isolates 113 * the hundreds digit (and any digits to its left), formats it using this same rule set, and 114 * places the result where the "<<" was. Notice also that the meaning of 115 * >> has changed: it now refers to both the tens and the ones digits. The meaning of 116 * both substitutions depends on the rule's base value. The base value determines the rule's <em>divisor,</em> 117 * which is the highest power of 10 that is less than or equal to the base value (the user 118 * can change this). To fill in the substitutions, the formatter divides the number being 119 * formatted by the divisor. The integral quotient is used to fill in the << 120 * substitution, and the remainder is used to fill in the >> substitution. The meaning 121 * of the brackets changes similarly: text in brackets is omitted if the value being 122 * formatted is an even multiple of the rule's divisor. The rules are applied recursively, so 123 * if a substitution is filled in with text that includes another substitution, that 124 * substitution is also filled in.</p> 125 * 126 * <p>This rule covers values up to 999, at which point we add another rule:</p> 127 * 128 * <pre>1000: << thousand[ >>];</pre> 129 * 130 * <p>Again, the meanings of the brackets and substitution tokens shift because the rule's 131 * base value is a higher power of 10, changing the rule's divisor. This rule can actually be 132 * used all the way up to 999,999. This allows us to finish out the rules as follows:</p> 133 * 134 * <pre> 1,000,000: << million[ >>]; 135 * 1,000,000,000: << billion[ >>]; 136 * 1,000,000,000,000: << trillion[ >>]; 137 * 1,000,000,000,000,000: OUT OF RANGE!;</pre> 138 * 139 * <p>Commas, periods, and spaces can be used in the base values to improve legibility and 140 * are ignored by the rule parser. The last rule in the list is customarily treated as an 141 * "overflow rule," applying to everything from its base value on up, and often (as 142 * in this example) being used to print out an error message or default representation. 143 * Notice also that the size of the major groupings in large numbers is controlled by the 144 * spacing of the rules: because in English we group numbers by thousand, the higher rules 145 * are separated from each other by a factor of 1,000.</p> 146 * 147 * <p>To see how these rules actually work in practice, consider the following example: 148 * Formatting 25,430 with this rule set would work like this:</p> 149 * 150 * <table border="0" width="100%"> 151 * <tr> 152 * <td><strong><< thousand >></strong></td> 153 * <td>[the rule whose base value is 1,000 is applicable to 25,340]</td> 154 * </tr> 155 * <tr> 156 * <td><strong>twenty->></strong> thousand >></td> 157 * <td>[25,340 over 1,000 is 25. The rule for 20 applies.]</td> 158 * </tr> 159 * <tr> 160 * <td>twenty-<strong>five</strong> thousand >></td> 161 * <td>[25 mod 10 is 5. The rule for 5 is "five."</td> 162 * </tr> 163 * <tr> 164 * <td>twenty-five thousand <strong><< hundred >></strong></td> 165 * <td>[25,340 mod 1,000 is 340. The rule for 100 applies.]</td> 166 * </tr> 167 * <tr> 168 * <td>twenty-five thousand <strong>three</strong> hundred >></td> 169 * <td>[340 over 100 is 3. The rule for 3 is "three."]</td> 170 * </tr> 171 * <tr> 172 * <td>twenty-five thousand three hundred <strong>forty</strong></td> 173 * <td>[340 mod 100 is 40. The rule for 40 applies. Since 40 divides 174 * evenly by 10, the hyphen and substitution in the brackets are omitted.]</td> 175 * </tr> 176 * </table> 177 * 178 * <p>The above syntax suffices only to format positive integers. To format negative numbers, 179 * we add a special rule:</p> 180 * 181 * <pre>-x: minus >>;</pre> 182 * 183 * <p>This is called a <em>negative-number rule,</em> and is identified by "-x" 184 * where the base value would be. This rule is used to format all negative numbers. the 185 * >> token here means "find the number's absolute value, format it with these 186 * rules, and put the result here."</p> 187 * 188 * <p>We also add a special rule called a <em>fraction rule </em>for numbers with fractional 189 * parts:</p> 190 * 191 * <pre>x.x: << point >>;</pre> 192 * 193 * <p>This rule is used for all positive non-integers (negative non-integers pass through the 194 * negative-number rule first and then through this rule). Here, the << token refers to 195 * the number's integral part, and the >> to the number's fractional part. The 196 * fractional part is formatted as a series of single-digit numbers (e.g., 123.456 would be 197 * formatted as "one hundred twenty-three point four five six").</p> 198 * 199 * <p>To see how this rule syntax is applied to various languages, examine the resource data.</p> 200 * 201 * <p>There is actually much more flexibility built into the rule language than the 202 * description above shows. A formatter may own multiple rule sets, which can be selected by 203 * the caller, and which can use each other to fill in their substitutions. Substitutions can 204 * also be filled in with digits, using a DecimalFormat object. There is syntax that can be 205 * used to alter a rule's divisor in various ways. And there is provision for much more 206 * flexible fraction handling. A complete description of the rule syntax follows:</p> 207 * 208 * <hr> 209 * 210 * <p>The description of a <tt>RuleBasedNumberFormat</tt>'s behavior consists of one or more <em>rule 211 * sets.</em> Each rule set consists of a name, a colon, and a list of <em>rules.</em> A rule 212 * set name must begin with a % sign. Rule sets with names that begin with a single % sign 213 * are <em>public:</em> the caller can specify that they be used to format and parse numbers. 214 * Rule sets with names that begin with %% are <em>private:</em> they exist only for the use 215 * of other rule sets. If a formatter only has one rule set, the name may be omitted.</p> 216 * 217 * <p>The user can also specify a special "rule set" named <tt>%%lenient-parse</tt>. 218 * The body of <tt>%%lenient-parse</tt> isn't a set of number-formatting rules, but a <tt>RuleBasedCollator</tt> 219 * description which is used to define equivalences for lenient parsing. For more information 220 * on the syntax, see <tt>RuleBasedCollator</tt>. For more information on lenient parsing, 221 * see <tt>setLenientParse()</tt>. <em>Note:</em> symbols that have syntactic meaning 222 * in collation rules, such as '&', have no particular meaning when appearing outside 223 * of the <tt>lenient-parse</tt> rule set.</p> 224 * 225 * <p>The body of a rule set consists of an ordered, semicolon-delimited list of <em>rules.</em> 226 * Internally, every rule has a base value, a divisor, rule text, and zero, one, or two <em>substitutions.</em> 227 * These parameters are controlled by the description syntax, which consists of a <em>rule 228 * descriptor,</em> a colon, and a <em>rule body.</em></p> 229 * 230 * <p>A rule descriptor can take one of the following forms (text in <em>italics</em> is the 231 * name of a token):</p> 232 * 233 * <table border="0" width="100%"> 234 * <tr> 235 * <td><em>bv</em>:</td> 236 * <td><em>bv</em> specifies the rule's base value. <em>bv</em> is a decimal 237 * number expressed using ASCII digits. <em>bv</em> may contain spaces, period, and commas, 238 * which are ignored. The rule's divisor is the highest power of 10 less than or equal to 239 * the base value.</td> 240 * </tr> 241 * <tr> 242 * <td><em>bv</em>/<em>rad</em>:</td> 243 * <td><em>bv</em> specifies the rule's base value. The rule's divisor is the 244 * highest power of <em>rad</em> less than or equal to the base value.</td> 245 * </tr> 246 * <tr> 247 * <td><em>bv</em>>:</td> 248 * <td><em>bv</em> specifies the rule's base value. To calculate the divisor, 249 * let the radix be 10, and the exponent be the highest exponent of the radix that yields a 250 * result less than or equal to the base value. Every > character after the base value 251 * decreases the exponent by 1. If the exponent is positive or 0, the divisor is the radix 252 * raised to the power of the exponent; otherwise, the divisor is 1.</td> 253 * </tr> 254 * <tr> 255 * <td><em>bv</em>/<em>rad</em>>:</td> 256 * <td><em>bv</em> specifies the rule's base value. To calculate the divisor, 257 * let the radix be <em>rad</em>, and the exponent be the highest exponent of the radix that 258 * yields a result less than or equal to the base value. Every > character after the radix 259 * decreases the exponent by 1. If the exponent is positive or 0, the divisor is the radix 260 * raised to the power of the exponent; otherwise, the divisor is 1.</td> 261 * </tr> 262 * <tr> 263 * <td>-x:</td> 264 * <td>The rule is a negative-number rule.</td> 265 * </tr> 266 * <tr> 267 * <td>x.x:</td> 268 * <td>The rule is an <em>improper fraction rule.</em></td> 269 * </tr> 270 * <tr> 271 * <td>0.x:</td> 272 * <td>The rule is a <em>proper fraction rule.</em></td> 273 * </tr> 274 * <tr> 275 * <td>x.0:</td> 276 * <td>The rule is a <em>master rule.</em></td> 277 * </tr> 278 * <tr> 279 * <td><em>nothing</em></td> 280 * <td>If the rule's rule descriptor is left out, the base value is one plus the 281 * preceding rule's base value (or zero if this is the first rule in the list) in a normal 282 * rule set. In a fraction rule set, the base value is the same as the preceding rule's 283 * base value.</td> 284 * </tr> 285 * </table> 286 * 287 * <p>A rule set may be either a regular rule set or a <em>fraction rule set,</em> depending 288 * on whether it is used to format a number's integral part (or the whole number) or a 289 * number's fractional part. Using a rule set to format a rule's fractional part makes it a 290 * fraction rule set.</p> 291 * 292 * <p>Which rule is used to format a number is defined according to one of the following 293 * algorithms: If the rule set is a regular rule set, do the following: 294 * 295 * <ul> 296 * <li>If the rule set includes a master rule (and the number was passed in as a <tt>double</tt>), 297 * use the master rule. (If the number being formatted was passed in as a <tt>long</tt>, 298 * the master rule is ignored.)</li> 299 * <li>If the number is negative, use the negative-number rule.</li> 300 * <li>If the number has a fractional part and is greater than 1, use the improper fraction 301 * rule.</li> 302 * <li>If the number has a fractional part and is between 0 and 1, use the proper fraction 303 * rule.</li> 304 * <li>Binary-search the rule list for the rule with the highest base value less than or equal 305 * to the number. If that rule has two substitutions, its base value is not an even multiple 306 * of its divisor, and the number <em>is</em> an even multiple of the rule's divisor, use the 307 * rule that precedes it in the rule list. Otherwise, use the rule itself.</li> 308 * </ul> 309 * 310 * <p>If the rule set is a fraction rule set, do the following: 311 * 312 * <ul> 313 * <li>Ignore negative-number and fraction rules.</li> 314 * <li>For each rule in the list, multiply the number being formatted (which will always be 315 * between 0 and 1) by the rule's base value. Keep track of the distance between the result 316 * the nearest integer.</li> 317 * <li>Use the rule that produced the result closest to zero in the above calculation. In the 318 * event of a tie or a direct hit, use the first matching rule encountered. (The idea here is 319 * to try each rule's base value as a possible denominator of a fraction. Whichever 320 * denominator produces the fraction closest in value to the number being formatted wins.) If 321 * the rule following the matching rule has the same base value, use it if the numerator of 322 * the fraction is anything other than 1; if the numerator is 1, use the original matching 323 * rule. (This is to allow singular and plural forms of the rule text without a lot of extra 324 * hassle.)</li> 325 * </ul> 326 * 327 * <p>A rule's body consists of a string of characters terminated by a semicolon. The rule 328 * may include zero, one, or two <em>substitution tokens,</em> and a range of text in 329 * brackets. The brackets denote optional text (and may also include one or both 330 * substitutions). The exact meanings of the substitution tokens, and under what conditions 331 * optional text is omitted, depend on the syntax of the substitution token and the context. 332 * The rest of the text in a rule body is literal text that is output when the rule matches 333 * the number being formatted.</p> 334 * 335 * <p>A substitution token begins and ends with a <em>token character.</em> The token 336 * character and the context together specify a mathematical operation to be performed on the 337 * number being formatted. An optional <em>substitution descriptor </em>specifies how the 338 * value resulting from that operation is used to fill in the substitution. The position of 339 * the substitution token in the rule body specifies the location of the resultant text in 340 * the original rule text.</p> 341 * 342 * <p>The meanings of the substitution token characters are as follows:</p> 343 * 344 * <table border="0" width="100%"> 345 * <tr> 346 * <td>>></td> 347 * <td>in normal rule</td> 348 * <td>Divide the number by the rule's divisor and format the remainder</td> 349 * </tr> 350 * <tr> 351 * <td></td> 352 * <td>in negative-number rule</td> 353 * <td>Find the absolute value of the number and format the result</td> 354 * </tr> 355 * <tr> 356 * <td></td> 357 * <td>in fraction or master rule</td> 358 * <td>Isolate the number's fractional part and format it.</td> 359 * </tr> 360 * <tr> 361 * <td></td> 362 * <td>in rule in fraction rule set</td> 363 * <td>Not allowed.</td> 364 * </tr> 365 * <tr> 366 * <td>>>></td> 367 * <td>in normal rule</td> 368 * <td>Divide the number by the rule's divisor and format the remainder, 369 * but bypass the normal rule-selection process and just use the 370 * rule that precedes this one in this rule list.</td> 371 * </tr> 372 * <tr> 373 * <td></td> 374 * <td>in all other rules</td> 375 * <td>Not allowed.</td> 376 * </tr> 377 * <tr> 378 * <td><<</td> 379 * <td>in normal rule</td> 380 * <td>Divide the number by the rule's divisor and format the quotient</td> 381 * </tr> 382 * <tr> 383 * <td></td> 384 * <td>in negative-number rule</td> 385 * <td>Not allowed.</td> 386 * </tr> 387 * <tr> 388 * <td></td> 389 * <td>in fraction or master rule</td> 390 * <td>Isolate the number's integral part and format it.</td> 391 * </tr> 392 * <tr> 393 * <td></td> 394 * <td>in rule in fraction rule set</td> 395 * <td>Multiply the number by the rule's base value and format the result.</td> 396 * </tr> 397 * <tr> 398 * <td>==</td> 399 * <td>in all rule sets</td> 400 * <td>Format the number unchanged</td> 401 * </tr> 402 * <tr> 403 * <td>[]</td> 404 * <td>in normal rule</td> 405 * <td>Omit the optional text if the number is an even multiple of the rule's divisor</td> 406 * </tr> 407 * <tr> 408 * <td></td> 409 * <td>in negative-number rule</td> 410 * <td>Not allowed.</td> 411 * </tr> 412 * <tr> 413 * <td></td> 414 * <td>in improper-fraction rule</td> 415 * <td>Omit the optional text if the number is between 0 and 1 (same as specifying both an 416 * x.x rule and a 0.x rule)</td> 417 * </tr> 418 * <tr> 419 * <td></td> 420 * <td>in master rule</td> 421 * <td>Omit the optional text if the number is an integer (same as specifying both an x.x 422 * rule and an x.0 rule)</td> 423 * </tr> 424 * <tr> 425 * <td></td> 426 * <td>in proper-fraction rule</td> 427 * <td>Not allowed.</td> 428 * </tr> 429 * <tr> 430 * <td></td> 431 * <td>in rule in fraction rule set</td> 432 * <td>Omit the optional text if multiplying the number by the rule's base value yields 1.</td> 433 * </tr> 434 * </table> 435 * 436 * <p>The substitution descriptor (i.e., the text between the token characters) may take one 437 * of three forms:</p> 438 * 439 * <table border="0" width="100%"> 440 * <tr> 441 * <td>a rule set name</td> 442 * <td>Perform the mathematical operation on the number, and format the result using the 443 * named rule set.</td> 444 * </tr> 445 * <tr> 446 * <td>a DecimalFormat pattern</td> 447 * <td>Perform the mathematical operation on the number, and format the result using a 448 * DecimalFormat with the specified pattern. The pattern must begin with 0 or #.</td> 449 * </tr> 450 * <tr> 451 * <td>nothing</td> 452 * <td>Perform the mathematical operation on the number, and format the result using the rule 453 * set containing the current rule, except: 454 * <ul> 455 * <li>You can't have an empty substitution descriptor with a == substitution.</li> 456 * <li>If you omit the substitution descriptor in a >> substitution in a fraction rule, 457 * format the result one digit at a time using the rule set containing the current rule.</li> 458 * <li>If you omit the substitution descriptor in a << substitution in a rule in a 459 * fraction rule set, format the result using the default rule set for this formatter.</li> 460 * </ul> 461 * </td> 462 * </tr> 463 * </table> 464 * 465 * <p>Whitespace is ignored between a rule set name and a rule set body, between a rule 466 * descriptor and a rule body, or between rules. If a rule body begins with an apostrophe, 467 * the apostrophe is ignored, but all text after it becomes significant (this is how you can 468 * have a rule's rule text begin with whitespace). There is no escape function: the semicolon 469 * is not allowed in rule set names or in rule text, and the colon is not allowed in rule set 470 * names. The characters beginning a substitution token are always treated as the beginning 471 * of a substitution token.</p> 472 * 473 * <p>See the resource data and the demo program for annotated examples of real rule sets 474 * using these features.</p> 475 * 476 * <p><em>User subclasses are not supported.</em> While clients may write 477 * subclasses, such code will not necessarily work and will not be 478 * guaranteed to work stably from release to release. 479 * 480 * <p><b>Localizations</b></p> 481 * <p>Constructors are available that allow the specification of localizations for the 482 * public rule sets (and also allow more control over what public rule sets are available). 483 * Localization data is represented as a textual description. The description represents 484 * an array of arrays of string. The first element is an array of the public rule set names, 485 * each of these must be one of the public rule set names that appear in the rules. Only 486 * names in this array will be treated as public rule set names by the API. Each subsequent 487 * element is an array of localizations of these names. The first element of one of these 488 * subarrays is the locale name, and the remaining elements are localizations of the 489 * public rule set names, in the same order as they were listed in the first arrray.</p> 490 * <p>In the syntax, angle brackets '<', '>' are used to delimit the arrays, and comma ',' is used 491 * to separate elements of an array. Whitespace is ignored, unless quoted.</p> 492 * <p>For example:<pre> 493 * < < %foo, %bar, %baz >, 494 * < en, Foo, Bar, Baz >, 495 * < fr, 'le Foo', 'le Bar', 'le Baz' > 496 * < zh, \\u7532, \\u4e59, \\u4e19 > > 497 * </pre></p> 498 * @author Richard Gillam 499 * @see NumberFormat 500 * @see DecimalFormat 501 * @stable ICU 2.0 502 */ 503 class U_I18N_API RuleBasedNumberFormat : public NumberFormat { 504 public: 505 506 //----------------------------------------------------------------------- 507 // constructors 508 //----------------------------------------------------------------------- 509 510 /** 511 * Creates a RuleBasedNumberFormat that behaves according to the description 512 * passed in. The formatter uses the default locale. 513 * @param rules A description of the formatter's desired behavior. 514 * See the class documentation for a complete explanation of the description 515 * syntax. 516 * @param perror The parse error if an error was encountered. 517 * @param status The status indicating whether the constructor succeeded. 518 * @stable ICU 3.2 519 */ 520 RuleBasedNumberFormat(const UnicodeString& rules, UParseError& perror, UErrorCode& status); 521 522 /** 523 * Creates a RuleBasedNumberFormat that behaves according to the description 524 * passed in. The formatter uses the default locale. 525 * <p> 526 * The localizations data provides information about the public 527 * rule sets and their localized display names for different 528 * locales. The first element in the list is an array of the names 529 * of the public rule sets. The first element in this array is 530 * the initial default ruleset. The remaining elements in the 531 * list are arrays of localizations of the names of the public 532 * rule sets. Each of these is one longer than the initial array, 533 * with the first String being the ULocale ID, and the remaining 534 * Strings being the localizations of the rule set names, in the 535 * same order as the initial array. Arrays are NULL-terminated. 536 * @param rules A description of the formatter's desired behavior. 537 * See the class documentation for a complete explanation of the description 538 * syntax. 539 * @param localizations the localization information. 540 * names in the description. These will be copied by the constructor. 541 * @param perror The parse error if an error was encountered. 542 * @param status The status indicating whether the constructor succeeded. 543 * @stable ICU 3.2 544 */ 545 RuleBasedNumberFormat(const UnicodeString& rules, const UnicodeString& localizations, 546 UParseError& perror, UErrorCode& status); 547 548 /** 549 * Creates a RuleBasedNumberFormat that behaves according to the rules 550 * passed in. The formatter uses the specified locale to determine the 551 * characters to use when formatting numerals, and to define equivalences 552 * for lenient parsing. 553 * @param rules The formatter rules. 554 * See the class documentation for a complete explanation of the rule 555 * syntax. 556 * @param locale A locale that governs which characters are used for 557 * formatting values in numerals and which characters are equivalent in 558 * lenient parsing. 559 * @param perror The parse error if an error was encountered. 560 * @param status The status indicating whether the constructor succeeded. 561 * @stable ICU 2.0 562 */ 563 RuleBasedNumberFormat(const UnicodeString& rules, const Locale& locale, 564 UParseError& perror, UErrorCode& status); 565 566 /** 567 * Creates a RuleBasedNumberFormat that behaves according to the description 568 * passed in. The formatter uses the default locale. 569 * <p> 570 * The localizations data provides information about the public 571 * rule sets and their localized display names for different 572 * locales. The first element in the list is an array of the names 573 * of the public rule sets. The first element in this array is 574 * the initial default ruleset. The remaining elements in the 575 * list are arrays of localizations of the names of the public 576 * rule sets. Each of these is one longer than the initial array, 577 * with the first String being the ULocale ID, and the remaining 578 * Strings being the localizations of the rule set names, in the 579 * same order as the initial array. Arrays are NULL-terminated. 580 * @param rules A description of the formatter's desired behavior. 581 * See the class documentation for a complete explanation of the description 582 * syntax. 583 * @param localizations a list of localizations for the rule set 584 * names in the description. These will be copied by the constructor. 585 * @param locale A locale that governs which characters are used for 586 * formatting values in numerals and which characters are equivalent in 587 * lenient parsing. 588 * @param perror The parse error if an error was encountered. 589 * @param status The status indicating whether the constructor succeeded. 590 * @stable ICU 3.2 591 */ 592 RuleBasedNumberFormat(const UnicodeString& rules, const UnicodeString& localizations, 593 const Locale& locale, UParseError& perror, UErrorCode& status); 594 595 /** 596 * Creates a RuleBasedNumberFormat from a predefined ruleset. The selector 597 * code choosed among three possible predefined formats: spellout, ordinal, 598 * and duration. 599 * @param tag A selector code specifying which kind of formatter to create for that 600 * locale. There are four legal values: URBNF_SPELLOUT, which creates a formatter that 601 * spells out a value in words in the desired language, URBNF_ORDINAL, which attaches 602 * an ordinal suffix from the desired language to the end of a number (e.g. "123rd"), 603 * URBNF_DURATION, which formats a duration in seconds as hours, minutes, and seconds, 604 * and URBNF_NUMBERING_SYSTEM, which is used to invoke rules for alternate numbering 605 * systems such as the Hebrew numbering system, or for Roman Numerals, etc. 606 * @param locale The locale for the formatter. 607 * @param status The status indicating whether the constructor succeeded. 608 * @stable ICU 2.0 609 */ 610 RuleBasedNumberFormat(URBNFRuleSetTag tag, const Locale& locale, UErrorCode& status); 611 612 //----------------------------------------------------------------------- 613 // boilerplate 614 //----------------------------------------------------------------------- 615 616 /** 617 * Copy constructor 618 * @param rhs the object to be copied from. 619 * @stable ICU 2.6 620 */ 621 RuleBasedNumberFormat(const RuleBasedNumberFormat& rhs); 622 623 /** 624 * Assignment operator 625 * @param rhs the object to be copied from. 626 * @stable ICU 2.6 627 */ 628 RuleBasedNumberFormat& operator=(const RuleBasedNumberFormat& rhs); 629 630 /** 631 * Release memory allocated for a RuleBasedNumberFormat when you are finished with it. 632 * @stable ICU 2.6 633 */ 634 virtual ~RuleBasedNumberFormat(); 635 636 /** 637 * Clone this object polymorphically. The caller is responsible 638 * for deleting the result when done. 639 * @return A copy of the object. 640 * @stable ICU 2.6 641 */ 642 virtual Format* clone(void) const; 643 644 /** 645 * Return true if the given Format objects are semantically equal. 646 * Objects of different subclasses are considered unequal. 647 * @param other the object to be compared with. 648 * @return true if the given Format objects are semantically equal. 649 * @stable ICU 2.6 650 */ 651 virtual UBool operator==(const Format& other) const; 652 653 //----------------------------------------------------------------------- 654 // public API functions 655 //----------------------------------------------------------------------- 656 657 /** 658 * return the rules that were provided to the RuleBasedNumberFormat. 659 * @return the result String that was passed in 660 * @stable ICU 2.0 661 */ 662 virtual UnicodeString getRules() const; 663 664 /** 665 * Return the number of public rule set names. 666 * @return the number of public rule set names. 667 * @stable ICU 2.0 668 */ 669 virtual int32_t getNumberOfRuleSetNames() const; 670 671 /** 672 * Return the name of the index'th public ruleSet. If index is not valid, 673 * the function returns null. 674 * @param index the index of the ruleset 675 * @return the name of the index'th public ruleSet. 676 * @stable ICU 2.0 677 */ 678 virtual UnicodeString getRuleSetName(int32_t index) const; 679 680 /** 681 * Return the number of locales for which we have localized rule set display names. 682 * @return the number of locales for which we have localized rule set display names. 683 * @stable ICU 3.2 684 */ 685 virtual int32_t getNumberOfRuleSetDisplayNameLocales(void) const; 686 687 /** 688 * Return the index'th display name locale. 689 * @param index the index of the locale 690 * @param status set to a failure code when this function fails 691 * @return the locale 692 * @see #getNumberOfRuleSetDisplayNameLocales 693 * @stable ICU 3.2 694 */ 695 virtual Locale getRuleSetDisplayNameLocale(int32_t index, UErrorCode& status) const; 696 697 /** 698 * Return the rule set display names for the provided locale. These are in the same order 699 * as those returned by getRuleSetName. The locale is matched against the locales for 700 * which there is display name data, using normal fallback rules. If no locale matches, 701 * the default display names are returned. (These are the internal rule set names minus 702 * the leading '%'.) 703 * @param index the index of the rule set 704 * @param locale the locale (returned by getRuleSetDisplayNameLocales) for which the localized 705 * display name is desired 706 * @return the display name for the given index, which might be bogus if there is an error 707 * @see #getRuleSetName 708 * @stable ICU 3.2 709 */ 710 virtual UnicodeString getRuleSetDisplayName(int32_t index, 711 const Locale& locale = Locale::getDefault()); 712 713 /** 714 * Return the rule set display name for the provided rule set and locale. 715 * The locale is matched against the locales for which there is display name data, using 716 * normal fallback rules. If no locale matches, the default display name is returned. 717 * @return the display name for the rule set 718 * @stable ICU 3.2 719 * @see #getRuleSetDisplayName 720 */ 721 virtual UnicodeString getRuleSetDisplayName(const UnicodeString& ruleSetName, 722 const Locale& locale = Locale::getDefault()); 723 724 725 using NumberFormat::format; 726 727 /** 728 * Formats the specified 32-bit number using the default ruleset. 729 * @param number The number to format. 730 * @param toAppendTo the string that will hold the (appended) result 731 * @param pos the fieldposition 732 * @return A textual representation of the number. 733 * @stable ICU 2.0 734 */ 735 virtual UnicodeString& format(int32_t number, 736 UnicodeString& toAppendTo, 737 FieldPosition& pos) const; 738 739 /** 740 * Formats the specified 64-bit number using the default ruleset. 741 * @param number The number to format. 742 * @param toAppendTo the string that will hold the (appended) result 743 * @param pos the fieldposition 744 * @return A textual representation of the number. 745 * @stable ICU 2.1 746 */ 747 virtual UnicodeString& format(int64_t number, 748 UnicodeString& toAppendTo, 749 FieldPosition& pos) const; 750 /** 751 * Formats the specified number using the default ruleset. 752 * @param number The number to format. 753 * @param toAppendTo the string that will hold the (appended) result 754 * @param pos the fieldposition 755 * @return A textual representation of the number. 756 * @stable ICU 2.0 757 */ 758 virtual UnicodeString& format(double number, 759 UnicodeString& toAppendTo, 760 FieldPosition& pos) const; 761 762 /** 763 * Formats the specified number using the named ruleset. 764 * @param number The number to format. 765 * @param ruleSetName The name of the rule set to format the number with. 766 * This must be the name of a valid public rule set for this formatter. 767 * @param toAppendTo the string that will hold the (appended) result 768 * @param pos the fieldposition 769 * @param status the status 770 * @return A textual representation of the number. 771 * @stable ICU 2.0 772 */ 773 virtual UnicodeString& format(int32_t number, 774 const UnicodeString& ruleSetName, 775 UnicodeString& toAppendTo, 776 FieldPosition& pos, 777 UErrorCode& status) const; 778 /** 779 * Formats the specified 64-bit number using the named ruleset. 780 * @param number The number to format. 781 * @param ruleSetName The name of the rule set to format the number with. 782 * This must be the name of a valid public rule set for this formatter. 783 * @param toAppendTo the string that will hold the (appended) result 784 * @param pos the fieldposition 785 * @param status the status 786 * @return A textual representation of the number. 787 * @stable ICU 2.1 788 */ 789 virtual UnicodeString& format(int64_t number, 790 const UnicodeString& ruleSetName, 791 UnicodeString& toAppendTo, 792 FieldPosition& pos, 793 UErrorCode& status) const; 794 /** 795 * Formats the specified number using the named ruleset. 796 * @param number The number to format. 797 * @param ruleSetName The name of the rule set to format the number with. 798 * This must be the name of a valid public rule set for this formatter. 799 * @param toAppendTo the string that will hold the (appended) result 800 * @param pos the fieldposition 801 * @param status the status 802 * @return A textual representation of the number. 803 * @stable ICU 2.0 804 */ 805 virtual UnicodeString& format(double number, 806 const UnicodeString& ruleSetName, 807 UnicodeString& toAppendTo, 808 FieldPosition& pos, 809 UErrorCode& status) const; 810 811 /** 812 * Formats the specified number using the default ruleset. 813 * @param obj The number to format. 814 * @param toAppendTo the string that will hold the (appended) result 815 * @param pos the fieldposition 816 * @param status the status 817 * @return A textual representation of the number. 818 * @stable ICU 2.0 819 */ 820 virtual UnicodeString& format(const Formattable& obj, 821 UnicodeString& toAppendTo, 822 FieldPosition& pos, 823 UErrorCode& status) const; 824 /** 825 * Redeclared Format method. 826 * @param obj the object to be formatted. 827 * @param result Output param which will receive the formatted string. 828 * @param status Output param set to success/failure code 829 * @return A reference to 'result'. 830 * @stable ICU 2.0 831 */ 832 UnicodeString& format(const Formattable& obj, 833 UnicodeString& result, 834 UErrorCode& status) const; 835 836 /** 837 * Redeclared NumberFormat method. 838 * @param number the double value to be formatted. 839 * @param output Output param which will receive the formatted string. 840 * @return A reference to 'output'. 841 * @stable ICU 2.0 842 */ 843 UnicodeString& format(double number, 844 UnicodeString& output) const; 845 846 /** 847 * Redeclared NumberFormat method. 848 * @param number the long value to be formatted. 849 * @param output Output param which will receive the formatted string. 850 * @return A reference to 'output'. 851 * @stable ICU 2.0 852 */ 853 UnicodeString& format(int32_t number, 854 UnicodeString& output) const; 855 856 /** 857 * Parses the specfied string, beginning at the specified position, according 858 * to this formatter's rules. This will match the string against all of the 859 * formatter's public rule sets and return the value corresponding to the longest 860 * parseable substring. This function's behavior is affected by the lenient 861 * parse mode. 862 * @param text The string to parse 863 * @param result the result of the parse, either a double or a long. 864 * @param parsePosition On entry, contains the position of the first character 865 * in "text" to examine. On exit, has been updated to contain the position 866 * of the first character in "text" that wasn't consumed by the parse. 867 * @see #setLenient 868 * @stable ICU 2.0 869 */ 870 virtual void parse(const UnicodeString& text, 871 Formattable& result, 872 ParsePosition& parsePosition) const; 873 874 875 /** 876 * Redeclared Format method. 877 * @param text The string to parse 878 * @param result the result of the parse, either a double or a long. 879 * @param status Output param set to failure code when a problem occurs. 880 * @stable ICU 2.0 881 */ 882 virtual inline void parse(const UnicodeString& text, 883 Formattable& result, 884 UErrorCode& status) const; 885 886 #if !UCONFIG_NO_COLLATION 887 888 /** 889 * Turns lenient parse mode on and off. 890 * 891 * When in lenient parse mode, the formatter uses a Collator for parsing the text. 892 * Only primary differences are treated as significant. This means that case 893 * differences, accent differences, alternate spellings of the same letter 894 * (e.g., ae and a-umlaut in German), ignorable characters, etc. are ignored in 895 * matching the text. In many cases, numerals will be accepted in place of words 896 * or phrases as well. 897 * 898 * For example, all of the following will correctly parse as 255 in English in 899 * lenient-parse mode: 900 * <br>"two hundred fifty-five" 901 * <br>"two hundred fifty five" 902 * <br>"TWO HUNDRED FIFTY-FIVE" 903 * <br>"twohundredfiftyfive" 904 * <br>"2 hundred fifty-5" 905 * 906 * The Collator used is determined by the locale that was 907 * passed to this object on construction. The description passed to this object 908 * on construction may supply additional collation rules that are appended to the 909 * end of the default collator for the locale, enabling additional equivalences 910 * (such as adding more ignorable characters or permitting spelled-out version of 911 * symbols; see the demo program for examples). 912 * 913 * It's important to emphasize that even strict parsing is relatively lenient: it 914 * will accept some text that it won't produce as output. In English, for example, 915 * it will correctly parse "two hundred zero" and "fifteen hundred". 916 * 917 * @param enabled If true, turns lenient-parse mode on; if false, turns it off. 918 * @see RuleBasedCollator 919 * @stable ICU 2.0 920 */ 921 virtual void setLenient(UBool enabled); 922 923 /** 924 * Returns true if lenient-parse mode is turned on. Lenient parsing is off 925 * by default. 926 * @return true if lenient-parse mode is turned on. 927 * @see #setLenient 928 * @stable ICU 2.0 929 */ 930 virtual inline UBool isLenient(void) const; 931 932 #endif 933 934 /** 935 * Override the default rule set to use. If ruleSetName is null, reset 936 * to the initial default rule set. If the rule set is not a public rule set name, 937 * U_ILLEGAL_ARGUMENT_ERROR is returned in status. 938 * @param ruleSetName the name of the rule set, or null to reset the initial default. 939 * @param status set to failure code when a problem occurs. 940 * @stable ICU 2.6 941 */ 942 virtual void setDefaultRuleSet(const UnicodeString& ruleSetName, UErrorCode& status); 943 944 /** 945 * Return the name of the current default rule set. If the current rule set is 946 * not public, returns a bogus (and empty) UnicodeString. 947 * @return the name of the current default rule set 948 * @stable ICU 3.0 949 */ 950 virtual UnicodeString getDefaultRuleSetName() const; 951 952 public: 953 /** 954 * ICU "poor man's RTTI", returns a UClassID for this class. 955 * 956 * @stable ICU 2.8 957 */ 958 static UClassID U_EXPORT2 getStaticClassID(void); 959 960 /** 961 * ICU "poor man's RTTI", returns a UClassID for the actual class. 962 * 963 * @stable ICU 2.8 964 */ 965 virtual UClassID getDynamicClassID(void) const; 966 967 /** 968 * Sets the decimal format symbols, which is generally not changed 969 * by the programmer or user. The formatter takes ownership of 970 * symbolsToAdopt; the client must not delete it. 971 * 972 * @param symbolsToAdopt DecimalFormatSymbols to be adopted. 973 * @stable ICU 49 974 */ 975 virtual void adoptDecimalFormatSymbols(DecimalFormatSymbols* symbolsToAdopt); 976 977 /** 978 * Sets the decimal format symbols, which is generally not changed 979 * by the programmer or user. A clone of the symbols is created and 980 * the symbols is _not_ adopted; the client is still responsible for 981 * deleting it. 982 * 983 * @param symbols DecimalFormatSymbols. 984 * @stable ICU 49 985 */ 986 virtual void setDecimalFormatSymbols(const DecimalFormatSymbols& symbols); 987 988 private: 989 RuleBasedNumberFormat(); // default constructor not implemented 990 991 // this will ref the localizations if they are not NULL 992 // caller must deref to get adoption 993 RuleBasedNumberFormat(const UnicodeString& description, LocalizationInfo* localizations, 994 const Locale& locale, UParseError& perror, UErrorCode& status); 995 996 void init(const UnicodeString& rules, LocalizationInfo* localizations, UParseError& perror, UErrorCode& status); 997 void dispose(); 998 void stripWhitespace(UnicodeString& src); 999 void initDefaultRuleSet(); 1000 void format(double number, NFRuleSet& ruleSet); 1001 NFRuleSet* findRuleSet(const UnicodeString& name, UErrorCode& status) const; 1002 1003 /* friend access */ 1004 friend class NFSubstitution; 1005 friend class NFRule; 1006 friend class FractionalPartSubstitution; 1007 1008 inline NFRuleSet * getDefaultRuleSet() const; 1009 Collator * getCollator() const; 1010 DecimalFormatSymbols * getDecimalFormatSymbols() const; 1011 1012 private: 1013 NFRuleSet **ruleSets; 1014 UnicodeString* ruleSetDescriptions; 1015 int32_t numRuleSets; 1016 NFRuleSet *defaultRuleSet; 1017 Locale locale; 1018 Collator* collator; 1019 DecimalFormatSymbols* decimalFormatSymbols; 1020 UBool lenient; 1021 UnicodeString* lenientParseRules; 1022 LocalizationInfo* localizations; 1023 }; 1024 1025 // --------------- 1026 1027 inline UnicodeString& 1028 RuleBasedNumberFormat::format(const Formattable& obj, 1029 UnicodeString& result, 1030 UErrorCode& status) const 1031 { 1032 // Don't use Format:: - use immediate base class only, 1033 // in case immediate base modifies behavior later. 1034 // dlf - the above comment is bogus, if there were a reason to modify 1035 // it, it would be virtual, and there's no reason because it is 1036 // a one-line macro in NumberFormat anyway, just like this one. 1037 return NumberFormat::format(obj, result, status); 1038 } 1039 1040 inline UnicodeString& 1041 RuleBasedNumberFormat::format(double number, UnicodeString& output) const { 1042 FieldPosition pos(0); 1043 return format(number, output, pos); 1044 } 1045 1046 inline UnicodeString& 1047 RuleBasedNumberFormat::format(int32_t number, UnicodeString& output) const { 1048 FieldPosition pos(0); 1049 return format(number, output, pos); 1050 } 1051 1052 inline void 1053 RuleBasedNumberFormat::parse(const UnicodeString& text, Formattable& result, UErrorCode& status) const 1054 { 1055 NumberFormat::parse(text, result, status); 1056 } 1057 1058 #if !UCONFIG_NO_COLLATION 1059 1060 inline UBool 1061 RuleBasedNumberFormat::isLenient(void) const { 1062 return lenient; 1063 } 1064 1065 #endif 1066 1067 inline NFRuleSet* 1068 RuleBasedNumberFormat::getDefaultRuleSet() const { 1069 return defaultRuleSet; 1070 } 1071 1072 U_NAMESPACE_END 1073 1074 /* U_HAVE_RBNF */ 1075 #endif 1076 1077 /* RBNF_H */ 1078 #endif 1079